Evaluation of survival in HIV-positive patients
Hamish McManus
Submitted for the degree of Doctor of Philosophy
July 2014
The Kirby Institute for infection and immunity in society
UNSW Australia PLEASE TYPE THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet
Surname or Family name: McManus
First name: Hamish Other name/s:
Abbreviation for degree as given in the University calendar: PhD
School: The Kirby Institute for infection and immunity in society Faculty: Medicine
Title: Evaluation of survival in HIV-positive patients
Abstract 350 words maximum: (PLEASE TYPE)
Life expectancy (LE) for people living with HIV in Australia has increased and this is largely attributable to the suppression of the virus by potent combination antiretroviral therapy (cART). The extent to which LE for people living with HIV approaches that of the general population continues to be revealed as the horizon of survival expands. Because the epidemic is still relatively young and the nature of long-term survival with HIV over lifetimes is unclear, this thesis describes long term survival in HIV+ cART initiated patients in Australia. We used a prospective observational cohort of HIV+ patients (the Australian HIV Observational Database (AHOD)). We evaluated trends in population averages of surrogates of outcome stratified by era of treatment initiation across sufficient durations of treatment to show asymptotic plateaus. We assessed mortality using survival models and validated results by using linkage to the National Death Index to ascertain mortality in Lost to follow-up (L TFU). A study of determinants of suicide or accidental or violent death in HIV+ patients using a case control design was conducted.
Higher proportions of early era HIV+ populations were found to have experienced significantly higher durations with detectable VL after treatment initiation. Mortality in cART initiated patients was close to general population levels in patients with high CD4+ (especially above 500 cells/ L). Duration of treatment was not associated with survival when adjusted for CD4+ level. Decreases in 10 year survival probabilities over a range of ages were similar to those of the Australian general population in patients with high CD4+, low VL, prior ADI and non IDU-exposure. There was similar mortality in L TFU and non-LTFU patients. We found association between immunological status and risk of suicide and accidental or violent death.
Overall, the treatment and prognosis for HIV+ patients was seen to be improving and survival close to that of the general population may be possible. Extensive empirical evaluation of aged long-term HIV+ populations is not yet possible because of the age distribution of infected patients. The distal effects of the disease, exacerbated by early era suboptimal therapy and non-adherence may not yet be manifest.
Declaration relating to disposition of project thesis/dissertation
I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known , subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation .
I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only) . .. z6./P9. (zpn ·· · ··· ~ · ~···Signature · ·········· ···· ······ ··· · · ······~····.(J ... .. ~t:~~ ·~~ · ··· ·· · ··········· · ··· Di~·
The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.
FOR OFFICE USE ONLY Date of completion of requirements for Award :
THIS SHEET IS TO BE GLUED TO THE INSIDE FRONT COVER OF THE THESIS ORIGINALITY STATEMENT
'I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.'
Signed .. b ' ~ .>
Date ...... ?::~/ . C?.J./ <~ft Declaration
RE:
1. McManus H, Hay JF, Woolley I, Boyd MA, Kelly MD, Mulhall B, et al. Recent trends in early stage response to combination antiretroviral therapy in Australia. Antivir Ther. 2014.
2. McManus H, O'Connor CC, Boyd M, Broom J, Russell D, Watson K, et al. Long-term survival in HIV positive patients with up to 15 Years of antiretroviral therapy. PLoS One. 2012;7(11 ):e48839.
3. McManus H, Petoumenos K, Brown K, Baker D, Russell D, Read T, et al. Loss to Follow-up in the Australian HIV Observational Database. [submitted for second review].
4. McManus H, Petoumenos K, Franic T, Kelly MD , Watson J, O'Connor CC, et al. Determinants of suicide and accidental or violent death in the Australian HIV Observational Database. PLoS One. 2014;9(2):e89089.
I certify that the above publications were direct results of my research towards this PhD, and that reproduction in this thesis does not breach copyright regulations .
.. ~ .·--~·························· Hamish McManus [PhD candidate] Date SUPERVISOR STATEMENT
I hereby certify that all co-authors of the published or submitted papers agree to the candidate Hamish McManus submitting those papers as part of his Doctoral Thesis. Signed ~~ ················· oate ...... :?: 4./.:; ~(?./ ~ .t 't ...... Contents
Abstract……………………………………………………………………………………………. 1
Acknowledgements……………………………………………………………………………… 2
List of abbreviations…………………………………………………………………………….. 3
Thesis overview………………………………………………………………………………….. 5
Chapter 1: Long term cART and survival in HIV-positive patients……………………... 7
1.1. Trends in early stage response of surrogates of treatment outcome…… 7
1.2. Life expectancy…………………………………………………………………….. 10
1.3. Loss to follow-up in high resource settings………………………………….. 15
1.4. Suicidal ideation and risk of suicide in HIV-positive population….……… 18
1.5. References………………………………………………………………………….. 22
Chapter 2: Methodological overview…………………………………………………………. 36
2.1. The Australian HIV Observational database………………………………….. 36
2.2. Longitudinal studies………………………………………………………………. 37
2.3. Statistical models………………………………………………………………….. 43
2.4. Notes on particular covariate specifications used in AHOD………………. 44
2.5. References………………………………………………………………………….. 46
Chapter 3: Recent trends in early stage response to combination antiretroviral therapy in Australia……………………………………………………………………………… 48 Chapter 4: Long-term survival in HIV-positive patients with up to 15 years of antiretroviral therapy…………………………………………………………………..………... 67
Chapter 5: Loss to Follow-up in the Australian HIV Observational Database………... 77
Chapter 6: Determinants of suicide and accidental or violent death in the Australian HIV Observational Database……………………………………………………... 99
Chapter 7: Conclusions and recommendations…………………………………………… 108
7.1. Trends in early stage response of surrogates of treatment outcome…… 108 7.2. Long-term survival in HIV-positive patients………………………………….. 108
7.3. Loss to follow-up in the Australian HIV Observational Database………… 109
7.4. Determinants of suicide and accidental or violent death in the Australian HIV Observational Database……………………………………………………... 110
7.5. Recommendations………………………………………………………………… 111
7.6. References………………………………………………………………………….. 114
Appendix: Suicide in HIV-positive populations literature review……………………….. 117
Abstract Life expectancy (LE) for people living with HIV in Australia has increased and this is largely attributable to the suppression of the virus by potent combination antiretroviral therapy (cART). The extent to which LE for people living with HIV approaches that of the general population continues to be revealed as the horizon of survival expands. Because the epidemic is still relatively young and the nature of long-term survival with HIV over lifetimes is unclear, this thesis describes long-term survival in HIV-positive (HIV+) cART initiated patients in Australia.
We used a prospective observational cohort of HIV+ patients (the Australian HIV Observational Database). We evaluated trends in population averages of surrogates of outcome stratified by era of treatment initiation across sufficient durations of treatment to show asymptotic plateaus. We assessed mortality using survival models and validated results by using linkage to the National Death Index to ascertain mortality in lost to follow-up (LTFU). A study of determinants of suicide or accidental or violent death in HIV+ patients using a case control design was conducted.
Higher proportions of early era HIV+ populations were found to have experienced significantly higher durations with detectable viral load (VL) after treatment initiation. Mortality in cART initiated patients was close to general population levels in patients with high CD4+ (especially above 500 cells/µL). Duration of treatment was not associated with survival when adjusted for CD4+ level. Decreases in 10 year survival probabilities over a range of ages were similar to those of the Australian general population in patients with high CD4+, low VL, prior AIDS defining illness and with modes of exposure other than injecting drug use. There was similar mortality in LTFU and non-LTFU patients. We found association between reduced immunological status and risk of suicide and accidental or violent death.
Overall, the treatment and prognosis for HIV+ patients was seen to be improving and survival close to that of the general population may be possible. Extensive empirical evaluation of aged long-term HIV+ populations is not yet possible because of the age distribution of infected patients. The long-term effects of the disease are yet to be fully revealed.
Page 1 of 127 Acknowledgements I would like to thank and acknowledge in particular, my supervisor Dr Kathy Petoumenos and my co-supervisor Professor Matthew Law for their invaluable experience and expertise, insight, advice, time and (above all) patience. I am honoured to have been given the opportunity to work with them both on this project.
I would also like to thank and acknowledge the AHOD community of collaborators whose ongoing scientific, clinical, analytical and administrative expertise has been exceptionally important to the preparation of data, analyses and manuscripts and more broadly to the study of HIV.
I would also like to acknowledge those patients who have ever contributed to AHOD. Because of this generosity, I have in turn felt a very real responsibility of stewardship during my time working on this thesis and as an AHOD coordinator.
I would like to thank my colleagues at the Kirby Institute and the biostatistics group in particular for their friendship, support and energy. I especially thank Steve W for allowing me to incessantly bounce ideas off, and more often than not for his forthright (and severe) appraisals.
And I would also like to thank my wife Sharon and daughters Tara and Kirsten for all putting up with me.
Page 2 of 127 List of abbreviations
AHOD Australian HIV Observational Database ADI AIDS defining illness AIDS Acquired immuno deficiency syndrome ALT Alanine aminotransferase ART Antiretroviral therapy AST Aspartate aminotreansferase cART Combination antiretroviral therapy CASCADE Concerted Action on SeroConversion to AIDS and Death in Europe CD4+ count CD4 positive T-cell lymphocyte count CHIC UK Collaborative HIV Cohort CNS Central nervous system CoDe Cause of death COHERE Collaboration of Observational HIV Epidemiological Research in Europe CPE Central nervous system penetration-effectiveness CVL Community viral load D:A:D Data Collection on Adverse Events of Anti-HIV Drugs GRADE Gradings recommendations assessment, development and evaluation HAD HIV associated dementia HAND HIV associated neurocognitive disorder HBV Hepatitits B virus HCV Hepatitis C virus HIV Human immuno deficiency virus HIV- Human immuno deficiency virus negative HIV+ Human immuno deficiency virus positive IDSA Infectious Diseases Society of America IDU Injecting drug use ITT Intention-to-treat LE Life expectancy LTFU Loss to follow-up / Lost to follow-up MAR Missing at random MCAR Missing completely at random MSM Men who have sex with men NDI National Death Index NMAR Not missing at random SE Standard error SMART Strategies for management of antiretroviral therapy SMR Standardised mortality ratio SNAE Serious non-AIDS event START Strategic timing of antiretroviral treatment STI Sexually transmitted infection TAsP Treatment as prevention TI Treatment interruption
Page 3 of 127 VACS Veterans Aging Cohort Study VF Viral failure VL Viral load
Page 4 of 127 Thesis overview
Life expectancy for people living with HIV in Australia has increased and this is largely attributable to the suppression of the virus by potent combination antiretroviral therapy (cART). Many HIV-positive (HIV+) patients have up to twenty years of cART experience and HIV is now widely regarded as a treatable chronic condition. From an epidemiological perspective, evaluation of long-term survival with HIV can now draw on extended durations of observed treatment at both the individual level and at the population level. However, the extent to which whole of life expectancy (LE) for people living with HIV is approaching that of the general population is not clear because the duration of the epidemic is still relatively short and much less than a normal lifespan. This thesis describes long-term survival in HIV+ cART initiated patients in Australia.
In the first chapter of this thesis, the literature of aspects of long-term survival with HIV is discussed. A preliminary review of reported trends in early stage response to cART is conducted to contextualise the early treatment background of HIV+ patients, which is highly prognostic of outcome. Following on from this, studies of LE of cART initiated patients are discussed and thereafter, studies of loss to follow-up (LTFU) in HIV+ cohorts from industrialised nations are discussed, in particular with a view to how this might affect observable treatment outcome including survival. Finally, to presage Chapter 6, studies of suicide in HIV+ populations are reviewed.
In the second chapter a methodological overview is provided which briefly describes the Australian HIV Observational Database (AHOD) and discusses considerations arising from use of longitudinal cohort data, especially with regard to AHOD.
The body of this thesis (chapters 3 to 6) is a series of peer reviewed publications based on epidemiological studies using the Australian HIV Observational Database (AHOD). Therefore these chapters are in manuscript form. They are formatted according to the style of each respective publishing journal, except for Chapter 4 which was under second review and had not been published at time of thesis submission.
In the third chapter, trends in surrogates of treatment outcome in Australian HIV+ populations are evaluated. Population average immunologic and virologic outcomes are compared across eras of treatment initiation. Trends in time to regimen switch are then evaluated by era of treatment initiation.
In Chapter 4 rates and determinants of survival in HIV+ patients on cART are evaluated. Standardised mortality ratios (SMR) are developed for a range of prognostic groups including by duration of treatment and immunologic and virologic status. Prognostic models of survival are also developed and relative survival compared for a range of risk groups as well as against the general Australian population.
Page 5 of 127 In Chapter 5, causes of LTFU in HIV+ patients in Australia are evaluated. Survival in LTFU HIV+ patients is evaluated by using the National Death Index (NDI) and relative rates of survival for LTFU and in follow-up patients compared. Overall rates of survival in HIV+ patients who have commenced cART, including LTFU patients are developed.
Chapter 6 is an evaluation of determinants of risk factors for suicide and accidental or violent death in an Australian HIV+ population. Both traditional psycho-social risk factors, as well as HIV specific risk factors are considered.
The final chapter summarises findings and their significance. Directions for further study are proposed in light of the conclusions that are drawn.
Page 6 of 127 Chapter 1: Long term cART and survival in HIV-positive patients
1.1. Trends in early stage response of surrogates of treatment outcome
Initial immunologic response to combination antiretroviral therapy (cART) as measured by change in CD4 cell count (CD4+) is important because it is predictive of clinical outcomes, specifically AIDS and death (1, 2). At early stages of treatment, response is strongly influenced by pre-initiation conditions (including CD4+ and VL and other prognostic characteristics such as age, socio demographic status, alcohol and injecting drug use (IDU), cardiovascular, liver and renal comorbidity and disease stage), as well as by the potency of implemented antiretroviral therapy (ART), and is also predictive of later outcomes including long-term CD4+ and VL, AIDS and death (3-6).
Generally, significant levels of virologic control (low level or undetectable VL) are reached within a year following treatment initiation (7-10). Phillips et al have shown attainment of at least one VL measure less than 500 copies/ml in 85% of patients within 32 weeks after initiating therapy for early cART era (from 1996 to 2000 as used in that study) patients, but that there were subsequently increasing rates of viral rebound (2 or more consecutive VL greater than 500 copies/ml as used in that study) over the duration of that study (up to 192 weeks) (7). A recent study by Torian et al has shown improved rates of virologic suppression (VL less than 400 copies/ml as used in that study) by year of diagnosis for diagnoses between 2006 and 2009 although that study did not look at time from cART initiation (6). In a recent meta-analysis of studies of treatment efficacy from 2005 or later and including 76 randomised trials and 6 prospective cohorts, Lee et al found mean patient proportions with (study defined) undetectable VL at 48 weeks to be 77% (standard deviation (SD) 8.6) (11).
Subsequent immunologic recovery (CD4+ increase towards stable levels depending on initial CD4+) has been shown to be more gradual (between 4 and 8 years in patients with sustained virologic control), and is strongly influenced by baseline levels (5). There is strong evidence that CD4+ trends towards different plateaus according to pre-cART (i.e. prior to commencement of cART) levels suggesting that long-term mortality is associated with early uncontrolled viral replication and immune activation (5, 12, 13). The association between low CD4+ as a surrogate for poorer outcome including AIDS and death is well established (14-16). In Chapter 3 of this thesis trends in surrogates (CD4+ and VL) of early stage response to treatment are resolved by era of cART initiation to explore changes in treatment effectiveness.
Prognostic factors of response to combination antiretroviral therapy Responses to cART of surrogates of outcome (specifically CD4+ and VL) are strongly influenced by a complex interaction of patient specific factors including baseline clinical status and treatment history, genotypic variation, as well as lifestyle and socio-demographic factors (4-6, 17, 18). Studies stratified by baseline CD4+ and VL have shown better virologic response
Page 7 of 127 in older patients attributable to better adherence, but that CD4+ repletion is slower, and older patients have increased risk of adverse events (19, 20). In addition to baseline CD4+ and CD8 cell counts and HIV RNA levels, recent US guidelines for management of HIV+ patients strongly recommend early general testing for drug resistance, complete blood counts, lipids, creatinine, alanine aminotransferase and aspartate aminotransferase, hepatitis B virus infection (HBV), hepatitis C virus infection (HCV) and sexually transmitted infections (STI) (21). Treatment guidelines are discussed in more detail in the next section.
The Veterans Aging Cohort Study (VACS) has developed an index based on most of these measures (and including age and gender specific adjustments) which is strongly prognostic of mortality up to 6 years after ART initiation (17). The same guidelines also strongly recommend comprehensive medical history, physical examination, medication/social/family history be obtained upon initiation of care (21).
In addition, response to cART is strongly characterised by duration of treatment, particularly by the first months after cART initiation when there is often a rapid decrease in HIV VL (7, 8). Outcomes are also influenced by broader changes to available ART, and health care resourcing, such as the availability of most effective ART, guidelines and interventions associated with any given time period (9, 10).
Changes in combination antiretroviral therapy and treatment guidelines in Australia There have been incremental improvements in cART over the last 15 years, including increased potency, reduced pill burden, combination formulations and new drug classes (22- 26). HIV drugs are classified according to the method by which they inhibit HIV-viral replication. Detailed description of the pharmacology of antiretrovirals is beyond the scope of this thesis, but briefly, prior to the licensure of protease inhibitors (PI) in the mid 1990’s, antiretroviral monotherapy/dual therapy used reverse transcriptase inhibitors (RT) comprising various Nucleoside/nucleotide RT inhibitors (NRTIs) which corrupt HIV DNA sequence and thereby impede HIV DNA synthesis, and Non-nucleoside RT inhibitors (NNRTIs) which impede RT conversion of HIV RNA to HIV DNA (27). PIs inhibit the action of the HIV enzyme protease from assembling virus particles. The combination of use of PIs with existing drug classes provided improved potency which largely suppressed the virus and thereby sustained viral suppression by lowering the rate of development of HIV drug resistant mutations (28). More recently available drug classes include entry inhibitors (EI) (which inhibit the virus’ ability to bind to the outer surface of host cells) and fusion inhibitors (FI) (which inhibit the virus’ ability to fuse with the cellular membrane and thereby prevent the virus from entering the cell.
Following the inception of cART in 1996, Australian treatment guidelines supported aggressive early stage initiation, in part motivated by the possibility of eradicating HIV infection (29, 30). This was followed by a shift towards relatively delayed treatment initiation in asymptomatic patients with lower CD4+ (<200 cells/µL) (2000-2007) following the identification of persistent
Page 8 of 127 cellular reservoirs of HIV, necessitating indefinite continuation of treatment using available regimens (31, 32).
From 2008 onwards, guidelines have supported earlier treatment initiation (at ≤350 cells/µL in asymptomatic patients) (31) and contemporaneous studies have demonstrated the importance of initiation prior to CD4+ dropping below 350 cells/µL (33, 34). However the effectiveness of long-term ART has been shown to be limited by increased co-morbidity including cardio vascular disease (CVD) in ART experienced patients, adverse events, low adherence and emerging HIV drug resistance (35-38). To potentially counter this, treatment sparing strategies including CD4+ guided structured treatment interruptions were investigated by the Strategies for Management of Antiretroviral Therapy (SMART) study. While observing a generally increased proportion of non-AIDS events compared to AIDS, that study found that non-AIDS events occurred most frequently in treatment interrupted groups, in addition to increased mortality and opportunistic infection (39, 40). Consequently, treatment interruption has been discounted as a serious option for mitigation of inherent risks associated with lifelong therapy and 2009 guidelines adjudged benefits of ART to outweigh risks associated with long term exposure (in particular with relevance to CVD)(41) . In light of this and to identify changes in treatment durability, trends in time to first treatment switch by era of treatment commencement are evaluated in Chapter 3 of this thesis.
Generally, studies have shown increased tolerability (i.e. reduced incidence of adverse effects arising from treatment) of later era regimens (22, 42, 43). The objective of recent studies, including the Strategic Timing of Anti-Retroviral Treatment (START) trial, is the evaluation of the benefit: risk ratio of earlier ART initiation (at 500 cells/µL) (44). The rationale for earlier treatment initiation is based on mitigation of potentially higher risk associated with CD4+ decrease even at relatively high CD4+ levels outweighing side effects associated with increased effective duration of treatment (45, 46). This in part is influenced by the relatively increased efficacy and tolerability of current first-line ART which makes the reasons for deferral less strong. Benefits associated with early treatment initiation have also been investigated by the HPTN 052 trial, which looked at HIV transmissions in sero-discordant couples, and which found significantly reduced risk of transmission associated with earlier initiation, supporting the broad strategy of ‘treatment as prevention’ (47, 48).
Primary care guidelines for the management of persons infected with HIV were recently updated by the Infectious Diseases Society of America (IDSA) in 2013 to replace 2009 guidelines (21). Guidelines now used the Gradings Recommendations Assessment, Development and Evaluation (GRADE) system which grades recommendations as being either strong or weak, and the quality of supporting evidence as being either high, moderate, low or very low (49, 50). Latest guidelines with Australian commentary support the initiation of treatment in all HIV-positive (HIV+) individuals at CD4+ above 500 cells/µl. However this recommendation is based on evidence rated by that panel as moderate, as well as being
Page 9 of 127 based on expert opinion. These recommendations now have a twofold objective, being to reduce disease progression in patients, as well as to lower the risk of transmission of HIV (51).
Given shifting strategy and availability of ART, the individual treatment histories of individual patients may be quite diverse. The complex interaction between treatment history and prognostic clinical factors is difficult to explicitly model without limiting the scope of results, and may lead to differences between the estimated and actual utility of treatment. However, few analyses of population level trends in treatment response simultaneously resolve both temporal changes in individual patient HIV disease histories, as well as broader trends and developments in ART. This may in part be because many studies do not have both long-term follow up and ongoing recruitment, which are necessary to measure these changes. In Chapter 3 trends in surrogates of outcome are investigated by adjusting for both of these potentially confounding time scales.
1.2. Life expectancy
LE with HIV is increasingly being revealed through the study of populations of patients with extended durations of exposure (10, 52-55). Long-term survival in HIV+ populations with access to effective treatment has improved over time (56-59) and appears to be approaching that of the general population (55, 60). Studies have shown declining rates of AIDS related death compared to non-AIDS related death since the introduction of cART (59, 61) and describe a need for increasing focus on chronic disease management and health promotion (62, 63).
Importance of life expectancy Life expectancy (LE) in HIV+ patients is important at both the individual and institutional level (64, 65). Improved prognosis and quality of life, aside from strong psychic reassurance, also necessitates longer term practical planning and provisioning at the individual level on core matters such as home and family, employment and lifestyle (64, 66). LE is integral to organisational assessment of risk of mortgage default, insurance or pension annuity funding. As the mortality in HIV+ populations approaches that of general populations, capital reserves required by organisations to cover the increased risk associated with HIV disease, and often passed on through premium levies, can be reduced (64, 65). Also, from the perspective of patients as well as health service providers, the cost of ART is a substantial proportion of total health care cost (67), and cost effectiveness and affordability of ART is increasingly important over extended lifetimes (68-70).
While evaluation of the impacts of changes in LE in HIV+ patients (e.g. psychosocial and financial costs and benefits) is beyond the scope of this thesis, mortality in HIV+ populations compared to general populations is directly associated with the magnitude of such impacts. Mortality in HIV+ populations compared to general populations is explored in Chapter 4 of this thesis.
Page 10 of 127 Improvement in life expectancy Mortality has decreased for newly diagnosed HIV+ patients since the inception and availability of cART (10, 53, 58, 71). In a study of mortality in the pre-cART era, Babiker et al showed median survival time from seroconversion to be 10.9 years (95% CI: 10.6-11.3) in 25-34 years old patients and less for patients at older ages (72). To contrast, in a 2010 study, Van Sighem et al found LE amongst recently diagnosed 25 year old males in the cART era was 52.7 years (IQR 44.2-59.3) compared to 53.1 years in the general population, although there may have been some under ascertainment of death in that study because of loss to follow-up (LTFU) (55). Ongoing improvement in LE since the inception of cART was demonstrated by analysis from the UK collaborative HIV Cohort Study which showed an increase in LE of a 20 year old from 30 years (standard error (SE) 1.2) in 1996-99 to 45.8 years (SE 1.7) in 2006-08 (54). However, in that study, the average LE for a 20 year old with HIV infection was over 18 years less than that for a 20 year male without HIV infection. In a 2007 study, Lewden et al showed that in patients initiating cART and maintaining high levels of CD4+, mortality rates reached those of the general population after 6 years treatment (73). Ongoing increase in LE in the cART era has been attributed to a variety of factors, including reduced exposure to suboptimal therapy and hence less drug resistance, lower toxicity and associated adverse events (74, 75), improved robustness and potency of ART (76) and improved adherence through reduced pill burdens and combination formulations(22-26).
However, patients without fully suppressed virus across extended durations of survival are increasingly likely to develop resistance to specific regimens and drug classes especially with respect to cross-resistant mutations of the virus (74, 77). The number of patients with extensive triple class failure has been shown to have increased, especially in patients with exposure to suboptimal therapies including mono/dual therapy where low potency ART has prevented complete viral suppression (74). While there are now a variety of drug classes to draw from, the prospect of exhausting effective therapy remains a real possibility for such patients although this risk depends on the ongoing development of new drugs.
Prognostic factors - non-HIV related Prognostic factors used in models of LE can be divided into non-HIV and HIV related categories (64). The first category includes socio-demographic, lifestyle and clinical risk factors present in the general population such as male gender, smoking, IDU, alcohol consumption, hepatitis co-infection, sexually transmitted infection, and other co morbidities, ethnicity and income level (17, 36, 64, 78-81). Generally, HIV+ populations exhibit higher lifestyle related risk factors than general populations and certain of these are also associated with risk of acquisition (81-87). Also, studies have shown increased severity of risk attributable to many of these factors in HIV+ populations compared to the general population. For example, Helleberg et al found a doubling of population attributable risk of death in HIV+ smokers compared to non-HIV+ smokers (81). Also, HIV+ populations have increased risk of many common clinical conditions (88, 89), and increased relative risk of serious non-AIDS
Page 11 of 127 events in HIV+ populations has been attributed to long term effects of immune dysfunction attributable directly to the virus, as well as to the action of cART itself (88, 90).
Prognostic factors – HIV related Of HIV specific risk factors, immunosuppression, especially at the start of ART, has been shown to be strongly predictive of outcome, and low CD4+ is the dominant predictor of LE (53, 67, 78, 86). All-cause mortality in patients who have achieved high CD4+ levels (above 500 counts/µl) approaches that of the general population over time (55, 60), although there is strong evidence that CD4+ trends towards different plateaus according to pre cART levels (5, 91). The COHERE collaboration has shown standardised mortality ratios (SMR) of 1.8 for CD4+ in the range of 350–499 cells/mL and 1.5 for CD4+ of above 500 cells/mL) (60). Of note in this cohort, while observed survival approached that of the general population, average rates at increased CD4+ levels remained slightly higher than those of the general population (60, 73). The change in guidelines and effect of late presentation on outcome has been discussed above, but to emphasise, at the start of ART CD4+ is prognostic of rate of immunologic recovery and survival even at extended durations (4, 36, 92) and models of LE must therefore, account for immunologic depletion associated with late initiation of treatment or interruption to treatment, sometimes in accordance with guidelines (54, 93).
Prognostic factors - age The average age of the Australian HIV+ population has been shown to be increasing (94). In Australia the epidemic is concentrated in men who have sex with men (MSM) (95). In this population Murray et al estimated the average age of HIV+ MSM for 2010 to exceed 44 years of age and showed a 12% annual increase in numbers of HIV+ MSM over 60 years of age for the period from 1995 to 2005 (94). By 2015 over 50% of the US HIV+ population will be over 50 years old (96). The age structure of HIV+ populations is affected by increased LE associated with improved ART as described above, but also by increased HIV disease incidence in older individuals (93).
Detailed description of the pathogenesis of ageing and of HIV disease is beyond the scope of this thesis, but conceptual models of survival with HIV incorporate interactions between ageing and HIV disease and comorbidity related pathophysiologic processes, chronic inflammation and treatment toxicity leading to organ system damage, vascular harm and multimorbidity, likely to manifest in old age as a variety of geriatric syndromes (90, 93).
Page 12 of 127 Residual viral replication Persistent virus expression (in lymph nodes) Loss of immunoregulatory cells Collagen deposition Microbial translocation High pathogen load (CMV, HBV, HCV) Thymic dysfunction
Suboptimal Residual Hyper- CD4+ gains Inflammation coagulation
Non–AIDS-related events and premature mortality
Figure 1: Factors associated with increased risk of premature age associated non-AIDS related events in treated HIV-disease (adapted from Deeks et al, Top HIV Med. 2009)
There is strong evidence associating immunologic resilience with age, and increased age at cART initiation has been associated with reduced rate and extent of immunologic recovery (19, 20, 96, 97) and time from HIV-infection to the development of AIDS is shorter in older patients (98). HIV+ patients receiving long term cART have increased risk of age-related serious non-AIDS events (SNAE) compared to the general population (99) and some studies have associated HIV-infection with the occurrence of earlier frailty phenotypes (100, 101). A 2008 workshop on HIV infection and ageing proposed that “HIV infection could serve as a model of accelerated ageing in a multiply morbid patient that could be broadly applicable to geriatrics and gerontology” (96).
While the spectrum of inflammatory and coagulation abnormalities seen in HIV+ populations is similar to that expected in much older general populations (90), recent literature does not strongly support the concept of HIV-associated “accelerated ageing”. It instead describes disease related functional compromise or vulnerability in conjunction with ageing, and a complex cascade of interacting conditions requiring increasingly individualised consideration (93, 99). Deeks et al propose a model of multiple HIV-associated factors leading to inflammation, low immunologic repletion and hypercoagulable state and subsequent age- associated non-AIDS events and associated mortality, but acknowledge that the interactions and mechanism leading to these outcomes is not known (Figure 1) (99) Justice et al present a compatible model which also explicitly incorporates ageing as an element contributing to functional decline, organ system failure and death (93). From the perspective of LE estimation, the effects of age and ageing on survival need to be considered in addition to just the effects Page 13 of 127 of increased duration of illness. A major limitation is that studies of overall LE in HIV+ populations are often limited by insufficient data in older age groups (69).
Model limitations Estimates of LE are developed either through the assembly of life tables or mathematical models, and both methods are reliant on cohort study collections of clinical and socio demographic data over sufficiently large populations and durations to estimate mortality (69). However, cohort study data has numerous limitations, including limited representativeness of entire lifespans (69, 86). In HIV+ populations, mortality at ages over 70 is still a relatively rare event because of low numbers in this age bracket, as well as low mortality associated with effective cART and healthy survivor bias, and models generally extrapolate beyond the range of observable data to encompass these age groups. Some studies have accounted for insufficient data at these ages by applying assumed relative rates of mortality (54). However, there is high risk of error in doing so when comparing survival relative to the general population because at these ages, general mortality rates increase rapidly and can magnify errors in model assumptions.
This is generally true of statistically modelling risk factors which lie beyond range of observed cohort experience and therefore models require some level of assumption to be built into model parameters (69). Given the relatively limited history of HIV infection and at-risk population, there are sparse data at extended durations of infection. To compare, type-1 childhood-onset diabetes patients have relatively low and stable mortality rates at earlier durations (up to 15 years) but increased rates at extended durations (up to 45 years) (102). Concomitant illnesses are well documented in diabetic populations and SMRs for ischaemic heart disease and renal disease in particular, are much higher than those of the general population (103) and there are poorer prognoses compared to similar events in the general population (104). In Chapter 4 of this thesis 10 year survival probabilities for HIV+ patients are developed, which are based on observable durations of treatment and do not rely on assumptions about mortality at older ages. These rates are compared against those of the general population for a range of scenarios and over a range of observable ages.
Page 14 of 127 1.3. Loss to follow-up in high resource settings
Significance of loss to follow-up Loss to follow-up (LTFU) in HIV+ cohorts is an important surrogate for interrupted clinical care which can potentially influence the assessment of HIV disease status and outcomes. Interrupted clinical follow-up of HIV-positive patients can delay the timely initiation of antiretroviral therapy (ART) in ART-naive patients, as well as disrupt ongoing ART in treatment experienced patients and thereby impair treatment response. The evaluation of determinants of LTFU in HIV+ cohorts is therefore important at both the patient level to prospectively guide pre-emptive intervention, and at the program level to retrospectively identify and adjust for potential biases introduced by different outcomes compared to patients in follow-up. The identification of specific patients who are at increased risk of LTFU can prompt interventions that are pre-emptive of discontinuous clinical attendance and treatment adherence (105, 106).
Inaccurate assumptions about outcomes in LTFU patients can bias findings derived from in- care populations (107). Prognostic characteristics of patients LTFU are discussed in the following section (Prognostic factors of loss to follow-up), but differences in prevalence of such characteristics according to LTFU status represent a form of selection bias for analyses based on just populations under follow-up. In particular, survival of LTFU patients might be poor compared to patients not LTFU if there is significant disease resurgence during LTFU episodes, and the ascertainment of mortality risk in patients LTFU remains an important objective of similar studies (108, 109). Predicted outcomes in LTFU patients might also be biased by unreported patient linkage to other health care providers (108). By identifying date of death using national death registries, reliable rates of survival in LTFU patients can be compared to patients in routine care which might also allow inference to be made about the extent of patients who are truly disengaged from care (110-112).
Definitions of loss to follow-up and limitations Definitions of LTFU tend to vary and comparison of different studies can therefore, be difficult. In high resource settings recorded failure to attend scheduled visits, is rarely used to define LTFU although this is an intuitive patient specific measure (both Karcher et al and Chi et al used this definition in low resource settings (113, 114)). This is likely because this data are seldom collected. Instead, in high resource settings, a significant interval between clinical attendances (often CD4+ test dates), usually collected in the setting of an observational cohort, is often used as a surrogate measure of LTFU (108, 109, 115-117). However, the application of general durations of non-attendance may be inappropriate for any given patient whose scheduling may vary according to treatment guidelines, year, health status, clinician or site. Most studies in high resource settings use an interval of over a year between attendances to define LTFU (108, 109, 115, 116).
Guidelines recommend scheduling of clinical visits for patients under routine care every two to three months (51) although there are recent recommendations to extend this interval and
Page 15 of 127 frequency of monitoring (for example of CD4+ to annually in those with high CD4+) (118). HIV prescriptions in Australia are generally made for similar spans and are invalid after one year duration (119, 120). A 1 year interval between attendances, therefore, represents significant departure from guidelines (between 4-6 missed visits depending on setting) and patients may be non-adherent or disengaged from care for a significant prior duration. In cohort studies, where data are often only collected annually, this definition may limit type-1 errors (i.e. incorrect rejection of a test hypothesis in this case the hypothesis that certain patients are not LTFU) arising from the incorrect assignment of LTFU status to patients who are actually in routine care based on shorter intervals between clinic visits.
There are certain other limitations evident in studies of LTFU. For example, many studies do not incorporate more than one episode of LTFU per patient into definitions used in analyses, and this might significantly underestimate the true level of patient disengagement from care (115-117, 121). In Chapter 5 of this thesis, LTFU including multiple episodes per patients is evaluated. Also, many studies may not be able to ascertain the vital status of patients who have disengaged from care which can lead to misallocation of LTFU status and underestimation of mortality in LTFU (108, 109). In Chapter 5, vital status in is ascertained based on linkage to the National Death Index and rates of mortality in patients LTFU are compared to patients who are not LTFU.
Levels of loss to follow-up In a 2009 study from the UK Collaborative HIV Cohort (CHIC) a relatively high rate of LTFU (inclusive of episodes resulting in return to care) was observed (16.7/100 person years (95% CI: 16.4 -17.2)) (108) despite access to free ART in the UK. However that analysis defined follow-up by duration between CD4+ test dates rather than all clinical visits and included multiple episodes per patient. Rates excluding episodes of return to care are typically close to 5/100 person years for high resource settings although there is reasonably wide regional variation in these rates by region and prognostic grouping (109, 115, 117, 122). For example, in the aforementioned study, Hill et al observed episodes of LTFU (defined as interval between CD4+ tests) in 43.6% of patients (108), compared to 22% observed by Mocroft et al (2008) who used similar definitions in a multiregional study (109), and 16% observed by Cohen et al in a 1998 US study (although in that study the definition of LTFU was not clear) (123).
Prognostic factors of loss to follow-up The above comparison shows regional variation in LTFU and the reasons for this are likely multifactorial and include patient mobility, availability of other clinics, local resourcing, and differences in clinical practice (109). Demographically, risk of LTFU in HIV+ populations has been associated with gender (males), younger age and mode of HIV exposure (including IDU) (108, 109, 116, 117). These characteristics are associated with residential transience and are consistent with shorter term engagement with localised healthcare, as well as with relatively poor adherence to ART (124-129) and higher transmission risk behaviours (127, 130-132).
Page 16 of 127 Low socio economic status, and associated indicators such as employment status, income, housing, insurance cover and ethnicity and illegal immigrant status have also been associated with LTFU (108, 115, 117, 121, 127).
The association between LTFU and clinical risk factors is less clear and there is some indication that better long-term prognosis is associated inversely with risk of LTFU. For example, while both Mocroft and Hill found lower CD4+ to be associated with permanent LTFU, Hill found that patients with higher CD4+ and low VL were more likely to have episodic LTFU (108, 109). Certain risk factors associated with poorer prognosis but requiring increased levels of clinical care, for example prior AIDS and psychiatric diagnoses, have been shown to be associated with decreased risk of LTFU (115, 122). Similarly cART initiation has been associated with increased likelihood of return to care following episodic LTFU (108, 109). In a recent US study, Udeagu et al found that the most commonly given reason for disengaging from care was that patients ‘felt well’ (121). This was given by 41% of LTFU, although a range of other reasons associated with personal well-being, as well as housing and social services, health care and service providers and personal beliefs about diagnosis and treatment were also recorded by that study (121). In Chapter 5 of this thesis determinants of LTFU in HIV+ patients in a high resource setting are evaluated.
Outcomes of loss to follow-up In this thesis we only consider LTFU in high resource settings with relatively high availability of ART. In limited resource settings LTFU is associated with poorer outcome including markedly increased mortality (133-136). In Australia there are low barriers to continuous engagement or re-engagement with care providers because of accessible subsidised treatment (119). As mentioned above, there is some indication that relatively good health (increased CD4+, undetectable VL, no prior ADI), is prognostic of episodic LTFU, but that, of patients who become intermittently LTFU, patients with better health are more likely to return to care (108, 109). While this is consistent with poorer outcomes in patients who remain LTFU, those studies were not able to completely verify vital status of LTFU patients. In a 2009 study of HIV+ patients, Lanoy et al found statistically significant increase in mortality in LTFU using linkage to a national mortality databases (116). However that study was limited to linkage using gender, month and year of birth, transmission group and place of residence or death. Conversely, Edwards et al found only marginally increased mortality in patients LTFU and this was not significant after adjustment for socio-demographic factors (sex, age, race, ethnicity, sexual orientation and IDU use) as well as HIV diseases specific factors (including CD4+, VL, ART initiation date and AIDS status) (137).
Where LTFU is associated with non-adherence and hence poor viral control, (as well as being associated with increased transmission risk) presents risk of disease resistance described earlier and hence reduced treatment options or even exhaustion of treatment options (77).
Page 17 of 127 However disease resistance is not directly evaluable in AHOD and beyond the scope of this thesis.
In Chapter 5 of this thesis vital status of LTFU is ascertained by linkage to the National Death Index (NDI) using matching based on name code, date of birth and gender as well as validation of matching using database date variables. Relative rates of mortality in patients LTFU and patients not LTFU are developed.
1.4. Suicidal ideation and risk of suicide in HIV-positive populations
High rates of suicide and accidental or violent death have been described in HIV+ populations, including in those receiving effective cART (138-140). The Data Collection on Adverse Events of Anti-HIV Drugs (D:A:D) cohort reported suicide as the cause of death in 4% of deaths, with a further 2.5% attributed to drug overdose and 1.5% as accident (141). In the Concerted Action on SeroConversion to AIDS and Death in Europe (CASCADE), suicide was reported in 6.4% of deaths, violence in 3.3%, and 5.7% of deaths were attributed to substance abuse (142). Both of these cohorts are discussed further below but are comprised of patients from predominantly well-resourced settings.
Prevalence of general risk factors in HIV-positive populations Models of suicide and suicidality in HIV+ populations incorporate synergies between general risk factors (i.e. risk factors also observed in the general population) and HIV specific risk factors (i.e. risk factors associated specifically with HIV virus and discussed in the next section). In general populations there is well documented association between completed suicide and risk factors such as male gender, lower age, drug use disorders, alcohol use, mental illness, socio economic status and homelessness (143, 144). Of specific interest to HIV+ populations is that many of these general risk factors are more prevalent in HIV+ populations and are associated with route of infection (145-150) including via increased transmission risk behaviours (151). In the Swiss HIV Cohort Study, the majority (75%) of HIV+ patients who committed suicide in the cART era had a diagnosis of mental illness, with depression being the most common (80%) (138). The Swiss HIV Cohort Study is a longitudinal cohort (i.e. data collection and analyses of the same group of individuals over time) comprising an estimated 15,275 patients (estimated 40% of all HIV+ patients in Switzerland). A significant proportion (23%) of patients who died by suicide in the cART era had untreated mental illness (138). In this cohort, suicide rates were shown to decline with increasing CD4+ (138). Advanced clinical stage (using the US Centres for Disease Control and Prevention classification system (152)) was significantly associated with suicide risk in both pre- and post-cART eras, after adjustment for other socio-demographic factors and history of psychiatric treatment. However, CD4+ and other HIV related factors were not included in this risk factor analysis.
Page 18 of 127 In the CASCADE study, latest CD4+ was not significantly associated with violent causes of death for patients followed in the cART era (142). CASCADE is a collaboration of 23 HIV+ cohorts and approximately 10,000 patients in Europe, Canada and Australia. The study in question examined specific causes of death as an outcome. As with the Swiss HIV Cohort Study, the CASCADE study did not assess risk factors for suicide which take into account both HIV related factors and general risk factors for suicide in one model.
HIV specific risk In the early years of the HIV epidemic, poor prognosis was a key contributing factor to high rates of suicide. Experiences of HIV+ people such as stigmatisation, discrimination and social isolation, were also identified as contributory factors to suicide risk (139, 153). Despite significant improvement in prognosis since the introduction of cART, the rates of suicide remain high. This was demonstrated in the Swiss HIV Cohort Study where rates of suicide decreased substantially in the cART era compared to the pre cART era but still remained well above that observed in the general population (138). As discussed above, increased all-cause mortality in HIV+ populations has been shown to be attributable to risk factors which are often identifiable prior to, or at early stages of cART, and which are often themselves prognostic of poorer outcome (81-85).
Models of suicide risk in HIV+ populations generally incorporate complex causal pathways. In particular, psychosocial risk factors have been shown to predispose patients to non-adherence (84, 154) hence poorer HIV disease control. There are also reciprocal causal effects of immunologic status on severity of neuropsychiatric symptoms as demonstrated by Warriner et al (155). The extent to which HIV infection is also associated with increased risk of suicide or violent death is not well documented. Recent qualitative models of suicidality in HIV+ populations in the post cART era also incorporate synergistic effects of HIV related factors and ageing (156). The association of increased CD4+ and reduced risk has been demonstrated by Keiser et al (138), although that analysis did not adjust for other risk factors. An objective of Chapter 6 of this thesis is to assess possible association between HIV specific risk factors and suicide, particularly after adjustment for accepted, non-HIV related risk factors.
Neuropsychiatric effects Possible neuropsychiatric effects resulting from HIV infection of the central nervous system (CNS) (broadly termed HIV associated neurocognitive disorders (HAND)), as well as neurotoxicity from ART (discussed later) may be associated with increased risk of suicide in HIV+ populations. HIV associated Dementia (HAD) is associated with an increased risk of mortality (157-160) and the optimal antiretroviral treatment for HAD remains controversial but there is evidence to suggest that use of cART regimens with good CNS penetration is superior Page 19 of 127 to the use of regimens with poor CNS penetration (161-164). Letendre et al. have assigned antiretroviral agents individual CNS penetration-effectiveness (CPE) ranks (164, 165) and patients using antiretroviral regimens with a CPE rank greater than cohort median levels (neurocART) were significantly less likely to have a detectable cerebrospinal fluid HIV VL (165). A recent paper by Smurzynski et al. (166) showed an adjusted association between increases in CPE score and neuropsychological test scores when accounting for an interaction with the number of antiretrovirals per regimen. Lanoy et al (167) showed all-cause mortality in neuroAIDS diagnoses was associated with CPE score for each of the periods 1992–1995 and 1996–1998 but not for 1999–2004. In that study, the authors attributed the lack of an associated effect in the period 1999–2004 to improved control of plasma VL (which was not adjusted for in initial models) by cART regimens in general. The use of neurocART has been shown to improve survival after diagnosis of HIV encephalopathy in perinatally infected children and adolescents (168), but survival effects are less clear in general HIV+ populations (169). An analysis from AHOD did not confirm the association between neurocART use and improved survival in a broader population of HIV-positive adults, with findings being robust to changes in model assumptions (170).
There is also evidence that neurotoxicity resulting directly from specific ART may lead to side effects, and toxicity associated with efavirenz use, in particular has been associated with neuropsychiatric adverse events including increased risk of attempted or completed suicide (148, 171-175). Risk associated with neurocART as well as for a range of antiretrovirals is also considered in Chapter 6 of this thesis.
Suicide study endpoints Most HIV-related studies of suicide focus on suicidality as an endpoint (176). This obviates difficulties associated with low prevalence of suicide cases, case finding and retrospective data collection. However, there are important differences between risk factors depending on level of intent, for example between suicidal ideation and attempts and completed suicide (177, 178). In a study of British private households, Gunnell et al found differences in the risk pattern of suicidal thoughts compared to completed suicide according to gender and age (178). In that study the incidence of suicidal thoughts was seen to be over 200 times greater than the incidence of suicide.
In addition to low prevalence, reporting of completed suicide may be incomplete. Recently Bohnert et al have shown increased likelihood of US medical examiners to classify overdose deaths in cases with substance use disorders as indeterminate intent or unintentional compared to other psychiatric disorders and suggested that this might therefore indicate misclassification of suicide deaths for this group (179). It is also possible that an additional proportion of accidental or violent deaths are in fact true suicide deaths but where official pronouncement of suicide might have been discouraged because of moral and legal implications (180).
Page 20 of 127 Systematic review of research into suicide in HIV-positive populations Preliminary overview of the literature indicated that research into HIV related suicide used a variety of outcome measures and rarely used completed suicide alone as a study outcome. To confirm this, a more systematic search of the literature was conducted.
In a review of papers published from 2000 onwards (Appendix: Suicide in HIV-positive populations), using pubmed and with keywords containing “HIV” & “suicid”, we identified 48 papers which investigated suicide or suicidality in HIV+ populations. This was based on a preliminary review of all papers containing “HIV” or “suicid” (749) of which 318 listed these as keywords and of which 146 were published from 2000 onwards. Of these, 97 papers were excluded because they were not medical studies (i.e. they were reviews, comment, qualitative models or opinions) or because they were case studies. This method excludes some potentially relevant papers where suicide or suicidality were incidental but not key findings such as certain studies of all-cause mortality in HIV+ populations.
The majority of papers reviewed (37 (77%)) used intermediate endpoints such as suicidality, ideation and psychiatric diagnoses. A further 7 (15%) included attempted suicide as an outcome while 6 (13%) of papers reported prevalence of completed suicide of which one also looked at associated risk factors.
Of the papers, 7 (15%) expressly examined the association between outcome and ART including one RCT investigating risk associated with efavirenz use. Most papers were based on US populations (19 (40%)), 6 (13%) were based on exclusively male populations compared to 8 (17%) based on female populations. Other target populations were IDU and older age groups.
Of the papers, 2 (4%) were Australian studies which looked at suicidality in predominantly MSM populations as well as the association between this and HIV status.
From review, it is apparent that there are relatively limited data investigating the association between HIV specific risk and completed suicide. Those papers which did investigate completed suicide were generally database queries, which could readily access increased numbers of cases but which had limited capacity to adjust for a wide range of clinical risk factors. To contrast, in Chapter 6, risk factors associated with completed suicide and accidental or violent death are investigated. Risk is evaluated for a range of socio- demographic, psychosocial and clinical factors (including HIV specific clinical factors).
Page 21 of 127 1.5. References
1. Piketty C, Weiss L, Thomas F, Mohamed AS, Belec L, Kazatchkine MD. Long-term clinical outcome of human immunodeficiency virus-infected patients with discordant immunologic and virologic responses to a protease inhibitor-containing regimen. The Journal of infectious diseases. 2001;183(9):1328-35. 2. Grabar S, Le Moing V, Goujard C, Leport C, Kazatchkine MD, Costagliola D, et al. Clinical outcome of patients with HIV-1 infection according to immunologic and virologic response after 6 months of highly active antiretroviral therapy. Annals of internal medicine. 2000;133(6):401-10. 3. Wright S, Petoumenos K, Boyd M, Carr A, Downing S, O'Connor C, et al. Ageing and long-term CD4 cell count trends in HIV-positive patients with 5 years or more combination antiretroviral therapy experience. HIV medicine. 2013;14(4):208-16. 4. Egger M, May M, Chene G, Phillips AN, Ledergerber B, Dabis F, et al. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet. 2002;360(9327):119-29. 5. Hughes R, Sterne J, Walsh J, Bansi L, Gilson R, Orkin C, et al. Long-term trends in CD4 cell counts and impact of viral failure in individuals starting antiretroviral therapy: UK Collaborative HIV Cohort (CHIC) study. HIV medicine. 2011;12(10):583-93. 6. Torian LV, Xia Q. Achievement and maintenance of viral suppression in persons newly diagnosed with HIV, New York City, 2006-2009: using population surveillance data to measure the treatment part of "test and treat". Journal of acquired immune deficiency syndromes. 2013;63(3):379-86. 7. Phillips AN, Staszewski S, Weber R, Kirk O, Francioli P, Miller V, et al. HIV viral load response to antiretroviral therapy according to the baseline CD4 cell count and viral load. JAMA. 2001;286(20):2560-7. 8. Bart PA, Rizzardi GP, Tambussi G, Chave JP, Chapuis AG, Graziosi C, et al. Immunological and virological responses in HIV-1-infected adults at early stage of established infection treated with highly active antiretroviral therapy. Aids. 2000;14(13):1887-97. 9. Bartlett JA, Johnson J, Herrera G, Sosa N, Rodriguez A, Liao Q, et al. Long-term results of initial therapy with abacavir and Lamivudine combined with Efavirenz, Amprenavir/Ritonavir, or Stavudine. Journal of acquired immune deficiency syndromes. 2006;43(3):284-92. 10. May MT, Sterne JA, Costagliola D, Sabin CA, Phillips AN, Justice AC, et al. HIV treatment response and prognosis in Europe and North America in the first decade of highly active antiretroviral therapy: a collaborative analysis. Lancet. 2006;368(9534):451-8. 11. Lee FJ, Amin J, Carr A. Efficacy of initial antiretroviral therapy for HIV-1 infection in adults: a systematic review and meta-analysis of 114 studies with up to 144 weeks' follow-up. PloS one. 2014;9(5):e97482.
Page 22 of 127 12. Wright ST, Carr A, Woolley I, Giles M, Hoy J, Cooper DA, et al. CD4 cell responses to combination antiretroviral therapy in patients starting therapy at high CD4 cell counts. Journal of acquired immune deficiency syndromes. 2011;58(1):72-9. 13. Jiang W, Lederman MM, Hunt P, Sieg SF, Haley K, Rodriguez B, et al. Plasma levels of bacterial DNA correlate with immune activation and the magnitude of immune restoration in persons with antiretroviral-treated HIV infection. The Journal of infectious diseases. 2009;199(8):1177-85. 14. Crowe SM, Carlin JB, Stewart KI, Lucas CR, Hoy JF. Predictive value of CD4 lymphocyte numbers for the development of opportunistic infections and malignancies in HIV- infected persons. Journal of acquired immune deficiency syndromes. 1991;4(8):770-6. 15. Phillips AN, Pezzotti P, Lepri AC, Rezza G. CD4 lymphocyte count as a determinant of the time from HIV seroconversion to AIDS and death from AIDS: evidence from the Italian Seroconversion Study. Aids. 1994;8(9):1299-305. 16. Mocroft A, Johnson MA, Phillips AN. Factors affecting survival in patients with the acquired immunodeficiency syndrome. Aids. 1996;10(10):1057-65. 17. Tate JP, Justice AC, Hughes MD, Bonnet F, Reiss P, Mocroft A, et al. An internationally generalizable risk index for mortality after one year of antiretroviral therapy. Aids. 2013;27(4):563-72. 18. Haaland RE, Johnson JA, Tang J. Recent Advances in Research of HIV Infection: Implications of Viral and Host Genetics on Treatment and Prevention. Public Health Genomics. 2013;16(1-2):31-6. 19. Silverberg MJ, Leyden W, Horberg MA, DeLorenze GN, Klein D, Quesenberry CP, Jr. Older age and the response to and tolerability of antiretroviral therapy. Arch Intern Med. 2007;167(7):684-91. 20. Grabar S, Kousignian I, Sobel A, Le Bras P, Gasnault J, Enel P, et al. Immunologic and clinical responses to highly active antiretroviral therapy over 50 years of age. Results from the French Hospital Database on HIV. Aids. 2004;18(15):2029-38. 21. Aberg JA, Gallant JE, Ghanem KG, Emmanuel P, Zingman BS, Horberg MA. Primary care guidelines for the management of persons infected with HIV: 2013 update by the HIV Medicine Association of the Infectious Diseases Society of America. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2014;58(1):1- 10. 22. Boyd MA. Improvements in antiretroviral therapy outcomes over calendar time. Current opinion in HIV and AIDS. 2009;4(3):194-9. 23. Willig JH, Abroms S, Westfall AO, Routman J, Adusumilli S, Varshney M, et al. Increased regimen durability in the era of once-daily fixed-dose combination antiretroviral therapy. Aids. 2008;22(15):1951-60. 24. Parienti JJ, Bangsberg DR, Verdon R, Gardner EM. Better adherence with once-daily antiretroviral regimens: a meta-analysis. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2009;48(4):484-8. Page 23 of 127 25. Llibre JM, Clotet B. Once-daily single-tablet regimens: a long and winding road to excellence in antiretroviral treatment. AIDS Rev. 2012;14(3):168-78. 26. Bartlett JA, Fath MJ, Demasi R, Hermes A, Quinn J, Mondou E, et al. An updated systematic overview of triple combination therapy in antiretroviral-naive HIV-infected adults. Aids. 2006;20(16):2051-64. 27. HIV/AIDS Antiretroviral Drugs Classes: National Institute of Allergy and Infectious Diseases; 2014 [updated 23/09/2013; cited 2014 4/12/2014]. 28. Gulick RM, Mellors JW, Havlir D, Eron JJ, Gonzalez C, McMahon D, et al. Treatment with indinavir, zidovudine, and lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral therapy. The New England journal of medicine. 1997;337(11):734-9. 29. Ho DD. Time to hit HIV, early and hard. The New England journal of medicine. 1995;333(7):450-1. 30. Perelson AS, Essunger P, Cao YZ, Vesanen M, Hurley A, Saksela K, et al. Decay characteristics of HIV-1-infected compartments during combination therapy. Nature. 1997;387(6629):188-91. 31. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. In: Services DoHaH, editor. 2006. 32. Hammer SM, Saag MS, Schechter M, Montaner JS, Schooley RT, Jacobsen DM, et al. Treatment for adult HIV infection: 2006 recommendations of the International AIDS Society-- USA panel. Top HIV Med. 2006;14(3):827-43. 33. Kitahata MM, Gange SJ, Abraham AG, Merriman B, Saag MS, Justice AC, et al. Effect of early versus deferred antiretroviral therapy for HIV on survival. The New England journal of medicine. 2009;360(18):1815-26. 34. Sterne JA, May M, Costagliola D, de Wolf F, Phillips AN, Harris R, et al. Timing of initiation of antiretroviral therapy in AIDS-free HIV-1-infected patients: a collaborative analysis of 18 HIV cohort studies. Lancet. 2009;373(9672):1352-63. 35. Palella FJ, Jr., Baker RK, Moorman AC, Chmiel JS, Wood KC, Brooks JT, et al. Mortality in the highly active antiretroviral therapy era: changing causes of death and disease in the HIV outpatient study. Journal of acquired immune deficiency syndromes. 2006;43(1):27- 34. 36. May M, Sterne JA, Sabin C, Costagliola D, Justice AC, Thiebaut R, et al. Prognosis of HIV-1-infected patients up to 5 years after initiation of HAART: collaborative analysis of prospective studies. Aids. 2007;21(9):1185-97. 37. Group DADS, Friis-Moller N, Reiss P, Sabin CA, Weber R, Monforte A, et al. Class of antiretroviral drugs and the risk of myocardial infarction. The New England journal of medicine. 2007;356(17):1723-35.
Page 24 of 127 38. Weber R, Sabin CA, Friis-Moller N, Reiss P, El-Sadr WM, Kirk O, et al. Liver-related deaths in persons infected with the human immunodeficiency virus: the D:A:D study. Arch Intern Med. 2006;166(15):1632-41. 39. El-Sadr WM, Lundgren JD, Neaton JD, Gordin F, Abrams D, Arduino RC, et al. CD4+ count-guided interruption of antiretroviral treatment. The New England journal of medicine. 2006;355(22):2283-96. 40. Strategies for Management of Antiretroviral Therapy Study G, Lundgren JD, Babiker A, El-Sadr W, Emery S, Grund B, et al. Inferior clinical outcome of the CD4+ cell count-guided antiretroviral treatment interruption strategy in the SMART study: role of CD4+ Cell counts and HIV RNA levels during follow-up. The Journal of infectious diseases. 2008;197(8):1145-55. 41. Aberg JA, Kaplan JE, Libman H, Emmanuel P, Anderson JR, Stone VE, et al. Primary care guidelines for the management of persons infected with human immunodeficiency virus: 2009 update by the HIV medicine Association of the Infectious Diseases Society of America. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2009;49(5):651-81. 42. Ortiz R, Dejesus E, Khanlou H, Voronin E, van Lunzen J, Andrade-Villanueva J, et al. Efficacy and safety of once-daily darunavir/ritonavir versus lopinavir/ritonavir in treatment- naive HIV-1-infected patients at week 48. Aids. 2008;22(12):1389-97. 43. Gallant JE, DeJesus E, Arribas JR, Pozniak AL, Gazzard B, Campo RE, et al. Tenofovir DF, emtricitabine, and efavirenz vs. zidovudine, lamivudine, and efavirenz for HIV. New England Journal of Medicine. 2006;354(3):251-60. 44. Babiker AG, Emery S, Fatkenheuer G, Gordin FM, Grund B, Lundgren JD, et al. Considerations in the rationale, design and methods of the Strategic Timing of AntiRetroviral Treatment (START) study. Clin Trials. 2013;10(1 Suppl):S5-S36. 45. Phillips AN, Gazzard B, Gilson R, Easterbrook P, Johnson M, Walsh J, et al. Rate of AIDS diseases or death in HIV-infected antiretroviral therapy-naive individuals with high CD4 cell count. Aids. 2007;21(13):1717-21. 46. Guiguet M, Porter K, Phillips A, Costagliola D, Babiker A. Clinical progression rates by CD4 cell category before and after the initiation of combination antiretroviral therapy (cART). Open AIDS J. 2008;2:3-9. 47. Cohen MS, Gay CL. Treatment to prevent transmission of HIV-1. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2010;50 Suppl 3:S85-95. 48. Cohen MS, Chen YQ, McCauley M, Gamble T, Hosseinipour MC, Kumarasamy N, et al. Prevention of HIV-1 infection with early antiretroviral therapy. The New England journal of medicine. 2011;365(6):493-505. 49. Kristiansen A, Brandt L, Agoritsas T, Akl EA, Berge E, Bondi J, et al. Adaptation of trustworthy guidelines developed using the GRADE methodology: a novel five-step process. Chest. 2014;146(3):727-34.
Page 25 of 127 50. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. Bmj. 2008;336(7650):924-6. 51. 2013 Antiretroviral Guidelines with Australian Commentary: Australasian Society for HIV Medicine; 2014 [updated 2/12/2013; cited 2014 11/03/2014]. Available from: http://arv.ashm.org.au/arv-guidelines/initiating-art-in-treatment-naive-patients. 52. Mahy M, Stover J, Stanecki K, Stoneburner R, Tassie JM. Estimating the impact of antiretroviral therapy: regional and global estimates of life-years gained among adults. Sex Transm Infect. 2010;86 Suppl 2:ii67-71. 53. Harrison KM, Song R, Zhang X. Life expectancy after HIV diagnosis based on national HIV surveillance data from 25 states, United States. Journal of acquired immune deficiency syndromes. 2010;53(1):124-30. 54. May M, Gompels M, Delpech V, Porter K, Post F, Johnson M, et al. Impact of late diagnosis and treatment on life expectancy in people with HIV-1: UK Collaborative HIV Cohort (UK CHIC) Study. BMJ. 2011;343:d6016. 55. van Sighem AI, Gras LA, Reiss P, Brinkman K, de Wolf F. Life expectancy of recently diagnosed asymptomatic HIV-infected patients approaches that of uninfected individuals. Aids. 2010;24(10):1527-35. 56. Walensky RP, Paltiel AD, Losina E, Mercincavage LM, Schackman BR, Sax PE, et al. The survival benefits of AIDS treatment in the United States. The Journal of infectious diseases. 2006;194(1):11-9. 57. Lima VD, Hogg RS, Harrigan PR, Moore D, Yip B, Wood E, et al. Continued improvement in survival among HIV-infected individuals with newer forms of highly active antiretroviral therapy. Aids. 2007;21(6):685-92. 58. Life expectancy of individuals on combination antiretroviral therapy in high-income countries: a collaborative analysis of 14 cohort studies. Lancet. 2008;372(9635):293-9. 59. Lohse N, Hansen AB, Pedersen G, Kronborg G, Gerstoft J, Sorensen HT, et al. Survival of persons with and without HIV infection in Denmark, 1995-2005. Annals of internal medicine. 2007;146(2):87-95. 60. Lewden C, the Mortality Working Group of COHERE, editors. Time with CD4 Cell Count above 500 cells.mm3 Allows HIV-Infected Men, but not Women, to Reach Similar Mortality Rates to Those of the General Population 17th CROI; 2010; San Francisco. 61. d'Arminio Monforte A, Sabin CA, Phillips A, Sterne J, May M, Justice A, et al. The changing incidence of AIDS events in patients receiving highly active antiretroviral therapy. Archives of Internal Medicine. 2005;165(4):416-23. 62. Petoumenos K, Worm S, Reiss P, de Wit S, d'Arminio Monforte A, Sabin C, et al. Rates of cardiovascular disease following smoking cessation in patients with HIV infection: results from the D:A:D study(*). HIV medicine. 2011;12(7):412-21. 63. Denholm JT, Yong MK, Elliott JH. Long term management of people with HIV. Aust Fam Physician. 2009;38(8):574-7. Page 26 of 127 64. May MT, Ingle SM. Life expectancy of HIV-positive adults: a review. Sexual health. 2011;8(4):526-33. 65. Lohse N, Hansen AB, Gerstoft J, Obel N. Improved survival in HIV-infected persons: consequences and perspectives. The Journal of antimicrobial chemotherapy. 2007;60(3):461- 3. 66. Cooper DA. Life and death in the cART era. Lancet. 2008;372(9635):266-7. 67. Sloan CE, Champenois K, Choisy P, Losina E, Walensky RP, Schackman BR, et al. Newer drugs and earlier treatment: impact on lifetime cost of care for HIV-infected adults. Aids. 2012;26(1):45-56. 68. Beck EJ, Mandalia S, Youle M, Brettle R, Fisher M, Gompels M, et al. Treatment outcome and cost-effectiveness of different highly active antiretroviral therapy regimens in the UK (1996-2002). Int J STD AIDS. 2008;19(5):297-304. 69. Nakagawa F, May M, Phillips A. Life expectancy living with HIV: recent estimates and future implications. Current opinion in infectious diseases. 2013;26(1):17-25. 70. Walensky RP, Sax PE, Nakamura YM, Weinstein MC, Pei PP, Freedberg KA, et al. Economic savings versus health losses: the cost-effectiveness of generic antiretroviral therapy in the United States. Annals of internal medicine. 2013;158(2):84-92. 71. Mocroft A, Brettle R, Kirk O, Blaxhult A, Parkin JM, Antunes F, et al. Changes in the cause of death among HIV positive subjects across Europe: results from the EuroSIDA study. Aids. 2002;16(12):1663-71. 72. Babiker A, Darby S, De Angelis D, Kwart D, Porter K, Beral V, et al. Time from HIV-1 seroconversion to AIDS and death before widespread use of highly-active antiretroviral therapy: a collaborative re-analysis. Lancet. 2000;355(9210):1131-7. 73. Lewden C, Chene G, Morlat P, Raffi F, Dupon M, Dellamonica P, et al. HIV-infected adults with a CD4 cell count greater than 500 cells/mm3 on long-term combination antiretroviral therapy reach same mortality rates as the general population. Journal of acquired immune deficiency syndromes. 2007;46(1):72-7. 74. Bansi L, Sabin C, Delpech V, Hill T, Fisher M, Walsh J, et al. Trends over calendar time in antiretroviral treatment success and failure in HIV clinic populations. HIV medicine. 2010;11(7):432-8. 75. Lampe FC, Smith CJ, Madge S, Kinloch-de Loes S, Tyrer M, Sabin CA, et al. Success of clinical care for human immunodeficiency virus infection according to demographic group among sexually infected patients in a routine clinic population, 1999 to 2004. Archives of Internal Medicine. 2007;167(7):692-700. 76. Lodwick RK, Smith CJ, Youle M, Lampe FC, Tyrer M, Bhagani S, et al. Stability of antiretroviral regimens in patients with viral suppression. Aids. 2008;22(9):1039-46. 77. Bansi L, Geretti AM, Dunn D, Hill T, Green H, Fearnhill E, et al. Impact of transmitted drug-resistance on treatment selection and outcome of first-line Highly Active Antiretroviral Therapy (HAART). Journal of acquired immune deficiency syndromes. 2010;53(5):633-9.
Page 27 of 127 78. Nakagawa F, Lodwick RK, Smith CJ, Smith R, Cambiano V, Lundgren JD, et al. Projected life expectancy of people with HIV according to timing of diagnosis. Aids. 2012;26(3):335-43. 79. Cockerham L, Scherzer R, Zolopa A, Rimland D, Lewis CE, Bacchetti P, et al. Association of HIV infection, demographic and cardiovascular risk factors with all-cause mortality in the recent HAART era. Journal of acquired immune deficiency syndromes. 2010;53(1):102-6. 80. Mimiaga MJ, Reisner SL, Grasso C, Crane HM, Safren SA, Kitahata MM, et al. Substance use among HIV-infected patients engaged in primary care in the United States: findings from the Centers for AIDS Research Network of Integrated Clinical Systems cohort. Am J Public Health. 2013;103(8):1457-67. 81. Helleberg M, Afzal S, Kronborg G, Larsen CS, Pedersen G, Pedersen C, et al. Mortality attributable to smoking among HIV-1-infected individuals: a nationwide, population- based cohort study. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2013;56(5):727-34. 82. Galvan FH, Bing EG, Fleishman JA, London AS, Caetano R, Burnam MA, et al. The prevalence of alcohol consumption and heavy drinking among people with HIV in the United States: results from the HIV Cost and Services Utilization Study. J Stud Alcohol. 2002;63(2):179-86. 83. Murray M, Hogg RS, Lima VD, May MT, Moore DM, Abgrall S, et al. The effect of injecting drug use history on disease progression and death among HIV-positive individuals initiating combination antiretroviral therapy: collaborative cohort analysis. HIV medicine. 2012;13(2):89-97. 84. Obel N, Omland LH, Kronborg G, Larsen CS, Pedersen C, Pedersen G, et al. Impact of non-HIV and HIV risk factors on survival in HIV-infected patients on HAART: a population- based nationwide cohort study. PloS one. 2011;6(7):e22698. 85. Duval X, Baron G, Garelik D, Villes V, Dupre T, Leport C, et al. Living with HIV, antiretroviral treatment experience and tobacco smoking: results from a multisite cross- sectional study. Antiviral therapy. 2008;13(3):389-97. 86. Sabin CA. Do people with HIV infection have a normal life expectancy in the era of combination antiretroviral therapy? BMC Med. 2013;11:251. 87. Hightow-Weidman LB, Golin CE, Green K, Shaw EN, MacDonald PD, Leone PA. Identifying people with acute HIV infection: demographic features, risk factors, and use of health care among individuals with AHI in North Carolina. AIDS and behavior. 2009;13(6):1075-83. 88. Deeks SG, Phillips AN. HIV infection, antiretroviral treatment, ageing, and non-AIDS related morbidity. Bmj. 2009;338:a3172. 89. Phillips AN, Neaton J, Lundgren JD. The role of HIV in serious diseases other than AIDS. Aids. 2008;22(18):2409-18.
Page 28 of 127 90. Deeks SG, Tracy R, Douek DC. Systemic effects of inflammation on health during chronic HIV infection. Immunity. 2013;39(4):633-45. 91. Wright ST, Carr A, Woolley I, Giles M, Hoy J, Cooper DA, et al. CD4 cell responses to combination antiretroviral therapy in patients starting therapy at high CD4 cell counts. J Acquir Immune Defic Syndr. 2011;58(1):72-9. 92. antiretroviral therapy cohort c, Lanoy E, May M, Mocroft A, Phillip A, Justice A, et al. Prognosis of patients treated with cART from 36 months after initiation, according to current and previous CD4 cell count and plasma HIV-1 RNA measurements. Aids. 2009;23(16):2199- 208. 93. Justice AC. HIV and aging: time for a new paradigm. Current HIV/AIDS reports. 2010;7(2):69-76. 94. Murray JM, McDonald AM, Law MG. Rapidly ageing HIV epidemic among men who have sex with men in Australia. Sexual health. 2009;6(1):83-6. 95. HIV, viral hepatitis and sexually transmissible infections in Australia Annual Surveillance
Report 2012. The Kirby Institute: 2012. 96. Effros RB, Fletcher CV, Gebo K, Halter JB, Hazzard WR, Horne FM, et al. Aging and infectious diseases: workshop on HIV infection and aging: what is known and future research directions. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2008;47(4):542-53. 97. Li X, Margolick JB, Jamieson BD, Rinaldo CR, Phair JP, Jacobson LP. CD4+ T-cell counts and plasma HIV-1 RNA levels beyond 5 years of highly active antiretroviral therapy. Journal of acquired immune deficiency syndromes. 2011;57(5):421-8. 98. Schneider MF, Gange SJ, Williams CM, Anastos K, Greenblatt RM, Kingsley L, et al. Patterns of the hazard of death after AIDS through the evolution of antiretroviral therapy: 1984-2004. Aids. 2005;19(17):2009-18. 99. Deeks SG. Immune dysfunction, inflammation, and accelerated aging in patients on antiretroviral therapy. Top HIV Med. 2009;17(4):118-23. 100. Desquilbet L, Jacobson LP, Fried LP, Phair JP, Jamieson BD, Holloway M, et al. HIV-1 infection is associated with an earlier occurrence of a phenotype related to frailty. J Gerontol A Biol Sci Med Sci. 2007;62(11):1279-86. 101. Desquilbet L, Jacobson LP, Fried LP, Phair JP, Jamieson BD, Holloway M, et al. A frailty-related phenotype before HAART initiation as an independent risk factor for AIDS or death after HAART among HIV-infected men. J Gerontol A Biol Sci Med Sci. 2011;66(9):1030- 8. 102. Secrest AM, Becker DJ, Kelsey SF, LaPorte RE, Orchard TJ. All-cause mortality trends in a large population-based cohort with long-standing childhood-onset type 1 diabetes: the Allegheny County type 1 diabetes registry. Diabetes Care. 2010;33(12):2573-9.
Page 29 of 127 103. Skrivarhaug T, Bangstad HJ, Stene LC, Sandvik L, Hanssen KF, Joner G. Long-term mortality in a nationwide cohort of childhood-onset type 1 diabetic patients in Norway. Diabetologia. 2006;49(2):298-305. 104. Ouhoummane N, Abdous B, Emond V, Poirier P. Impact of diabetes and gender on survival after acute myocardial infarction in the Province of Quebec, Canada--a population- based study. Diabet Med. 2009;26(6):609-16. 105. Jordan WC. Reaching lost-to-care populations. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2007;45 Suppl 4:S275-80. 106. Harding R, Krakauer EL, Sithole Z, De Lima L, Selman L. The 'lost' HIV population: time to refocus our clinical and research efforts. AIDS. 2009;23(1):145-6. 107. Porter K, Johnson AM, Phillips AN, Darbyshire JH. The practical significance of potential biases in estimates of the AIDS incubation period distribution in the UK register of HIV seroconverters. Aids. 1999;13(14):1943-51. 108. Hill T, Bansi L, Sabin C, Phillips A, Dunn D, Anderson J, et al. Data linkage reduces loss to follow-up in an observational HIV cohort study. Journal of clinical epidemiology. 2010;63(10):1101-9. 109. Mocroft A, Kirk O, Aldins P, Chies A, Blaxhult A, Chentsova N, et al. Loss to follow-up in an international, multicentre observational study. HIV medicine. 2008;9(5):261-9. 110. Simard EP, Pfeiffer RM, Engels EA. Mortality due to cancer among people with AIDS: a novel approach using registry-linkage data and population attributable risk methods. Aids. 2012;26(10):1311-8. 111. Fonseca MG, Coeli CM, de Fatima de Araujo Lucena F, Veloso VG, Carvalho MS. Accuracy of a probabilistic record linkage strategy applied to identify deaths among cases reported to the Brazilian AIDS surveillance database. Cad Saude Publica. 2010;26(7):1431-8. 112. Centers for Disease C, Prevention. Electronic record linkage to identify deaths among persons with AIDS--District of Columbia, 2000-2005. MMWR Morb Mortal Wkly Rep. 2008;57(23):631-4. 113. Karcher H, Omondi A, Odera J, Kunz A, Harms G. Risk factors for treatment denial and loss to follow-up in an antiretroviral treatment cohort in Kenya. Trop Med Int Health. 2007;12(5):687-94. 114. Chi BH, Cantrell RA, Mwango A, Westfall AO, Mutale W, Limbada M, et al. An empirical approach to defining loss to follow-up among patients enrolled in antiretroviral treatment programs. Am J Epidemiol. 2010;171(8):924-31. 115. Lebouche B, Yazdanpanah Y, Gerard Y, Sissoko D, Ajana F, Alcaraz I, et al. Incidence rate and risk factors for loss to follow-up in a French clinical cohort of HIV-infected patients from January 1985 to January 1998. HIV medicine. 2006;7(3):140-5. 116. Lanoy E, Lewden C, Lievre L, Tattevin P, Boileau J, Aouba A, et al. How does loss to follow-up influence cohort findings on HIV infection? A joint analysis of the French hospital database on HIV, Mortalite 2000 survey and death certificates. HIV medicine. 2009;10(4):236- 45. Page 30 of 127 117. Schepens T, Morreel S, Florence E, Koole O, Colebunders R. Incidence and risk factors associated with lost to follow-up in a Belgian cohort of HIV-infected patients treated with highly active antiretroviral therapy. Int J STD AIDS. 2010;21(11):765-9. 118. Guidelines forthe Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. In: Services DoHaH, editor. 2014. 119. Medicare_Australia. Highly Specialised Drugs (HSD) program: Commonwealth of Australia; 2014 [updated 19 December, 2013; cited 2014 12 May, 2014]. Available from: http://www.medicareaustralia.gov.au/provider/pbs/highly-specialised-drugs/#N100E3. 120. Medicare_Australia. Browse by Section 100 Item List: Commonwealth of Australia; 2014 [cited 2014 12 May, 2014]. Available from: http://pbs.gov.au/browse/section100- hb?initial=a. 121. Udeagu CC, Webster TR, Bocour A, Michel P, Shepard CW. Lost or just not following up: public health effort to re-engage HIV-infected persons lost to follow-up into HIV medical care. Aids. 2013;27(14):2271-9. 122. Lanoy E, Mary-Krause M, Tattevin P, Dray-Spira R, Duvivier C, Fischer P, et al. Predictors identified for losses to follow-up among HIV-seropositive patients. Journal of clinical epidemiology. 2006;59(8):829-35. 123. Cohen CJ, Iwane MK, Palensky JB, Levin DL, Meagher KJ, Frost KR, et al. A national HIV community cohort: design, baseline, and follow-up of the AmFAR Observational Database. American Foundation for AIDS Research Community-Based Clinical Trials Network. Journal of clinical epidemiology. 1998;51(9):779-93. 124. Chesney MA. Factors affecting adherence to antiretroviral therapy. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2000;30 Suppl 2:S171-6. 125. Keiser O, Spycher B, Rauch A, Calmy A, Cavassini M, Glass TR, et al. Outcomes of antiretroviral therapy in the Swiss HIV Cohort Study: latent class analysis. AIDS and behavior. 2012;16(2):245-55. 126. Andrews JR, Wood R, Bekker LG, Middelkoop K, Walensky RP. Projecting the benefits of antiretroviral therapy for HIV prevention: the impact of population mobility and linkage to care. The Journal of infectious diseases. 2012;206(4):543-51. 127. Rachlis B, Brouwer KC, Mills EJ, Hayes M, Kerr T, Hogg RS. Migration and transmission of blood-borne infections among injection drug users: understanding the epidemiologic bridge. Drug and alcohol dependence. 2007;90(2-3):107-19. 128. South SJ, Crowder KD. Residential mobility between cities and suburbs: race, suburbanization, and back-to-the-city moves. Demography. 1997;34(4):525-38. 129. Wood E, Yip B, Gataric N, Montaner JS, O'Shaughnessy MV, Schechter MT, et al. Determinants of geographic mobility among participants in a population-based HIV/AIDS drug treatment program. Health Place. 2000;6(1):33-40.
Page 31 of 127 130. Madden HC, Phillips-Howard PA, Hargreaves SC, Downing J, Bellis MA, Vivancos R, et al. Access to HIV community services by vulnerable populations: evidence from an enhanced HIV/AIDS surveillance system. AIDS Care. 2011;23(5):542-9. 131. German D, Davey MA, Latkin CA. Residential transience and HIV risk behaviors among injection drug users. AIDS and behavior. 2007;11(6 Suppl):21-30. 132. Strathdee SA, Hallett TB, Bobrova N, Rhodes T, Booth R, Abdool R, et al. HIV and risk environment for injecting drug users: the past, present, and future. Lancet. 2010;376(9737):268-84. 133. Brinkhof MW, Pujades-Rodriguez M, Egger M. Mortality of patients lost to follow-up in antiretroviral treatment programmes in resource-limited settings: systematic review and meta- analysis. PloS one. 2009;4(6):e5790. 134. Dalal RP, Macphail C, Mqhayi M, Wing J, Feldman C, Chersich MF, et al. Characteristics and outcomes of adult patients lost to follow-up at an antiretroviral treatment clinic in johannesburg, South Africa. Journal of acquired immune deficiency syndromes. 2008;47(1):101-7. 135. Brennan AT, Maskew M, Sanne I, Fox MP. The importance of clinic attendance in the first six months on antiretroviral treatment: a retrospective analysis at a large public sector HIV clinic in South Africa. Journal of the International AIDS Society. 2010;13:49. 136. Egger M, Spycher BD, Sidle J, Weigel R, Geng EH, Fox MP, et al. Correcting mortality for loss to follow-up: a nomogram applied to antiretroviral treatment programmes in sub- Saharan Africa. PLoS Med. 2011;8(1):e1000390. 137. Edwards JK, Cole SR, Westreich D, Moore R, Mathews C, Geng E, et al. Loss to clinic and five-year mortality among HIV-infected antiretroviral therapy initiators. PloS one. 2014;9(7):e102305. 138. Keiser O, Spoerri A, Brinkhof MW, Hasse B, Gayet-Ageron A, Tissot F, et al. Suicide in HIV-infected individuals and the general population in Switzerland, 1988-2008. Am J Psychiatry. 2010;167(2):143-50. 139. Carrico AW. Elevated suicide rate among HIV-positive persons despite benefits of antiretroviral therapy: implications for a stress and coping model of suicide. Am J Psychiatry. 2010;167(2):117-9. 140. Rice BD, Smith RD, Delpech VC. HIV infection and suicide in the era of HAART in England, Wales and Northern Ireland. Aids. 2010;24(11):1795-7. 141. Smith C, Sabin CA, Lundgren JD, Thiebaut R, Weber R, Law M, et al. Factors associated with specific causes of death amongst HIV-positive individuals in the D:A:D Study. Aids. 2010;24(10):1537-48. 142. Marin B, Thiebaut R, Bucher HC, Rondeau V, Costagliola D, Dorrucci M, et al. Non- AIDS-defining deaths and immunodeficiency in the era of combination antiretroviral therapy. Aids. 2009;23(13):1743-53.
Page 32 of 127 143. Wilcox HC, Conner KR, Caine ED. Association of alcohol and drug use disorders and completed suicide: an empirical review of cohort studies. Drug and alcohol dependence. 2004;76 Suppl:S11-9. 144. Logan J, Hall J, Karch D. Suicide categories by patterns of known risk factors: a latent class analysis. Arch Gen Psychiatry. 2011;68(9):935-41. 145. Carrico AW, Johnson MO, Morin SF, Remien RH, Charlebois ED, Steward WT, et al. Correlates of suicidal ideation among HIV-positive persons. Aids. 2007;21(9):1199-203. 146. Chikezie UE, Otakpor AN, Kuteyi OB, James BO. Suicidality among individuals with HIV/AIDS in Benin City, Nigeria: a case-control study. AIDS Care. 2012. 147. Jin H, Atkinson JH, Duarte NA, Yu X, Shi C, Riggs PK, et al. Risks and predictors of current suicidality in HIV-infected heroin users in treatment in Yunnan, China: a controlled study. Journal of acquired immune deficiency syndromes. 2013;62(3):311-6. 148. Journot V, Chene G, De Castro N, Rancinan C, Cassuto JP, Allard C, et al. Use of efavirenz is not associated with a higher risk of depressive disorders: a substudy of the randomized clinical trial ALIZE-ANRS 099. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2006;42(12):1790-9. 149. Preau M, Bouhnik AD, Peretti-Watel P, Obadia Y, Spire B. Suicide attempts among people living with HIV in France. AIDS Care. 2008;20(8):917-24. 150. Ploderl M, Wagenmakers EJ, Tremblay P, Ramsay R, Kralovec K, Fartacek C, et al. Suicide risk and sexual orientation: a critical review. Arch Sex Behav. 2013;42(5):715-27. 151. Carrico AW, Neilands TB, Johnson MO. Suicidal ideation is associated with HIV transmission risk in men who have sex with men. Journal of acquired immune deficiency syndromes. 2010;54(4):e3-4. 152. Schneider E, Whitmore S, Glynn KM, Dominguez K, Mitsch A, McKenna MT. Revised surveillance case definitions for HIV infection among adults, adolescents, and children aged <18 months and for HIV infection and AIDS among children aged 18 months to <13 years-- United States, 2008. MMWR Recomm Rep. 2008;57(RR-10):1-12. 153. Heckman TG, Miller J, Kochman A, Kalichman SC, Carlson B, Silverthorn M. Thoughts of suicide among HIV-infected rural persons enrolled in a telephone-delivered mental health intervention. Ann Behav Med. 2002;24(2):141-8. 154. Jia CX, Mehlum L, Qin P. AIDS/HIV infection, comorbid psychiatric illness, and risk for subsequent suicide: a nationwide register linkage study. J Clin Psychiatry. 2012;73(10):1315- 21. 155. Warriner EM, Rourke SB, Rourke BP, Rubenstein S, Millikin C, Buchanan L, et al. Immune activation and neuropsychiatric symptoms in HIV infection. J Neuropsychiatry Clin Neurosci. 2010;22(3):321-8. 156. Vance DE, Moneyham L, Fordham P, Struzick TC. A model of suicidal ideation in adults aging with HIV. J Assoc Nurses AIDS Care. 2008;19(5):375-84.
Page 33 of 127 157. Mayeux R, Stern Y, Tang MX, Todak G, Marder K, Sano M, et al. Mortality risks in gay men with human immunodeficiency virus infection and cognitive impairment. Neurology. 1993;43(1):176-82. 158. Sacktor NC, Bacellar H, Hoover DR, Nance-Sproson TE, Selnes OA, Miller EN, et al. Psychomotor slowing in HIV infection: a predictor of dementia, AIDS and death. J Neurovirol. 1996;2(6):404-10. 159. Wilkie FL, Goodkin K, Eisdorfer C, Feaster D, Morgan R, Fletcher MA, et al. Mild cognitive impairment and risk of mortality in HIV-1 infection. J Neuropsychiatry Clin Neurosci. 1998;10(2):125-32. 160. Price RW, Yiannoutsos CT, Clifford DB, Zaborski L, Tselis A, Sidtis JJ, et al. Neurological outcomes in late HIV infection: adverse impact of neurological impairment on survival and protective effect of antiviral therapy. AIDS Clinical Trial Group and Neurological AIDS Research Consortium study team. Aids. 1999;13(13):1677-85. 161. Brew BJ. HIV, the brain, children, HAART and 'neuro-HAART': a complex mix. Aids. 2009;23(14):1909-10. 162. Tozzi V, Balestra P, Salvatori MF, Vlassi C, Liuzzi G, Giancola ML, et al. Changes in cognition during antiretroviral therapy: comparison of 2 different ranking systems to measure antiretroviral drug efficacy on HIV-associated neurocognitive disorders. Journal of acquired immune deficiency syndromes. 2009;52(1):56-63. 163. Cysique LA, Maruff P, Brew BJ. Antiretroviral therapy in HIV infection: are neurologically active drugs important? Arch Neurol. 2004;61(11):1699-704. 164. Letendre S, Marquie-Beck J, Capparelli E, Best B, Clifford D, Collier AC, et al. Validation of the CNS Penetration-Effectiveness rank for quantifying antiretroviral penetration into the central nervous system. Arch Neurol. 2008;65(1):65-70. 165. Letendre S, Ellis R, Deutsch R, Clifford D, Collier AC, Gelman BB, et al. Correlates of Time-to-Loss-of-Viral-Response in CSF and Plasma in the CHARTER Cohort. 2010. 166. Smurzynski M, Wu K, Letendre S, Robertson K, Bosch RJ, Clifford DB, et al. Effects of central nervous system antiretroviral penetration on cognitive functioning in the ALLRT cohort. Aids. 2011;25(3):357-65. 167. Lanoy E, Guiguet M, Bentata M, Rouveix E, Dhiver C, Poizot-Martin I, et al. Survival after neuroAIDS: association with antiretroviral CNS Penetration-Effectiveness score. Neurology. 2011;76(7):644-51. 168. Patel K, Ming X, Williams PL, Robertson KR, Oleske JM, Seage GR, 3rd. Impact of HAART and CNS-penetrating antiretroviral regimens on HIV encephalopathy among perinatally infected children and adolescents. Aids. 2009;23(14):1893-901. 169. Garvey L, Winston A, Sabin C, Group UCS. Does antiretroviral combination therapies with greater central nervous system (CNS) penetration prevent the development of CNS opportunistc diseases. CROI 2010, Session 88-Poster Abstracts, Infection and Immune Activation of CNS Compartments. [Poster]. In press 2010.
Page 34 of 127 170. McManus H, Li P, Nolan D, Bloch M, Kiertiburanakul S, Choi J, et al. Does use of antiretroviral therapy regimens with high central nervous system penetration improve survival in HIV-infected adults? HIV medicine. 2011;12(10):610-9. 171. MOLLAN K, SMURZYNSKI M, NA L, ROBERTSON K, CAMPBELL T, SAX P, et al. Hazard of Suicidality in Patients Randomly Assigned to Efavirenz for Initial Treatment of HIV- 1: a Cross-Study Analysis Conducted by the AIDS Clinical Trials Group (ACTG) IDWeek 2013; October 4, 2013; The Moscone Centre, San Francisco, CA2013. 172. Shubber Z, Calmy A, Andrieux-Meyer I, Vitoria M, Renaud-Thery F, Shaffer N, et al. Adverse events associated with nevirapine and efavirenz-based first-line antiretroviral therapy: a systematic review and meta-analysis. Aids. 2013;27(9):1403-12. 173. Leutscher PD, Stecher C, Storgaard M, Larsen CS. Discontinuation of efavirenz therapy in HIV patients due to neuropsychiatric adverse effects. Scand J Infect Dis. 2013;45(8):645-51. 174. Rihs TA, Begley K, Smith DE, Sarangapany J, Callaghan A, Kelly M, et al. Efavirenz and chronic neuropsychiatric symptoms: a cross-sectional case control study. HIV medicine. 2006;7(8):544-8. 175. Raffi F, Pozniak AL, Wainberg MA. Has the time come to abandon efavirenz for first- line antiretroviral therapy? The Journal of antimicrobial chemotherapy. 2014. 176. Catalan J, Harding R, Sibley E, Clucas C, Croome N, Sherr L. HIV infection and mental health: suicidal behaviour--systematic review. Psychol Health Med. 2011;16(5):588-611. 177. Nock MK, Kessler RC. Prevalence of and risk factors for suicide attempts versus suicide gestures: analysis of the National Comorbidity Survey. J Abnorm Psychol. 2006;115(3):616-23. 178. Gunnell D, Harbord R, Singleton N, Jenkins R, Lewis G. Factors influencing the development and amelioration of suicidal thoughts in the general population. Cohort study. Br J Psychiatry. 2004;185:385-93. 179. Bohnert AS, McCarthy JF, Ignacio RV, Ilgen MA, Eisenberg A, Blow FC. Misclassification of suicide deaths: examining the psychiatric history of overdose decedents. Inj Prev. 2013. 180. Cantor CH, Neulinger K, De Leo D. Australian suicide trends 1964-1997: youth and beyond? The Medical journal of Australia. 1999;171(3):137-41.
Page 35 of 127 Chapter 2: Methodological overview
All analyses in this thesis are based on routinely collected source data from the Australian HIV Observational Database (AHOD) except for Chapter 5 which also uses National Death Index (NDI) data on date and fact of death (see Chapter 5 for detail), and Chapter 6 which also uses additional data collected from patient records (see Chapter 6 for detail).
Detailed methodologies for specific studies of this thesis are described in each relevant chapter (chapters 3-6). However, this chapter provides a broad overview of the methodology used, especially considerations arising from the analysis of longitudinal data and with reference to AHOD in particular which was the primary data source of the thesis.
2.1. The Australian HIV Observational Database
AHOD is a large national prospective observational cohort of HIV+ persons in Australia and has been running since 1999. AHOD is a multicohort collaboration with currently 29 sites within Australia (and recently 2 sites from New Zealand) including sites from specialised general practitioners, hospitals and sexual health clinics (1, 2).
The primary objectives of AHOD are to monitor patterns of antiretroviral treatment use including the effectiveness of different treatment regimens related to demographic factors and markers of HIV disease stage; to monitor how often people with HIV are changing antiretroviral treatments, and the reasons for treatment changes; to monitor patterns of toxicities/adverse events associated with antiretroviral treatment use; to monitor HIV related and non-HIV related causes of death; and to monitor treatment use and long term health outcomes among different cultural and ethnic groups, including Aboriginal and Torres Strait Islanders.
Routine clinical data is transferred and combined electronically via standardised formats every six months. Data collected includes:
• patient identifiers – name code (first two letters of surname and given name), sex, date of birth, cultural/ethnic group, Aboriginal and/or Torres Strait Islander status • clinical history – exposure category, date first HIV positive test, date last HIV negative test, hepatitis B virus surface antigen and core antibody status, hepatitis C virus antibody PCR status • stage of disease - CD4 counts, CD8 counts, HIV viral load, AIDS defining illnesses (ADI)
Page 36 of 127 • antiretroviral treatments including reasons for stopping/changing antiretroviral treatment, including treatment failure, clinical progression and adverse events • cause of death • cardiovascular risk factors - smoking, lipids, height and weight to calculate body mass index (BMI) • laboratory parameters routinely used to assess liver function (Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) and kidney function (S-creatinine) Initial recruitment to AHOD occurred primarily between 1999 and 2002. By July 2008 AHOD was a highly treatment experienced cohort (over 90% treated), of whom more than 75% were on their third plus regimen and was therefore limited in exploring outcomes of first line use of newer antiretroviral agents and classes. To assess the short and long-term effects of newer antiretroviral agents, recruitment commenced from July 2008 of an additional 1,000 HIV+ patients who were antiretroviral treatment naïve or recently treated.
2.2. Longitudinal studies
Study design Analyses in this thesis are based on repeated measurements of medical data from HIV+ patients collected at times of routine visit to collaborating clinics. Such data are observational and in principle involve no intervention (experimental or otherwise) on the part of the investigator other than routine treatment of patients. This form of study design allows evaluation of the association between incident outcomes (such as death) over the duration of the study and of a range of prognostic factors (exposures) (3). Specifically, studies of repeated measures data (longitudinal studies) can estimate change in individuals over durations of interest and factors influencing differences in change among individuals over these durations (3).
AHOD is a prospective clinical cohort, in that HIV related clinical data is compiled from time of enrolment. However, with the advent of widespread electronic medical record systems which has allowed transfer of large amounts of pre-enrolment (retrospective) data, AHOD data also comprises extensive retrospective clinical data and hence also has features associated with historical cohort studies.
The analytical advantages of the prospective cohort design include the potential to evaluate incidence (i.e. absolute risk) for a range of outcomes and association with given exposures (unlike cross-sectional studies which are limited to investigation of prevalence); the logic is very similar to the gold standard randomised controlled trials in that the effect of exposure on
Page 37 of 127 disease incidence is investigated; exposure/prognostic factors can be measured without ascertainment bias (see Error, chance and bias) arising from use of pre-occurring (prevalent) outcomes; standardised methods of data collection can be implemented to improve efficiency of data collection and generalisability of analyses (4, 5). Retrospective data collection in AHOD also allows increased numbers of patients to be included in analyses which include baselines (start of analysis time at-risk) which are not always prospectively recorded, although, the biases inherent in this methodology are described below (see Error, chance and bias).
However, cohort designs are susceptible to confounding and bias especially associated with loss to follow-up (LTFU) (see Error, chance and bias) because of the large numbers of patients and extended durations required to measure often rare outcomes. Further, analyses are generally limited to consideration of prognostic factors collected at the outset of the study. Generalisation of results may be limited by the characteristics of the cohort population (geographical, treatment centre type, exposure group, treatment status etc.). Another important limitation of cohort studies is that even the largest studies may be underpowered to measure important but rare outcomes such as cardiovascular disease risk, and analysis of long term conditions may be limited by attrition bias associated with LTFU of long term patients.
AHOD is a relatively small cohort compared to many international HIV+ cohorts from similar resource settings, and is similar in size to single site cohorts such as the Frankfurt HIV cohort (6) or the St Pierre HIV cohort (7), although these studies were commenced at earlier dates (see Table 1). However high levels of follow-up in AHOD including amongst patients who were recruited at database inception facilitates analysis of long-term outcomes despite otherwise relatively low patient numbers overall. This characteristic is discussed in Chapter 3. As with most of the cohorts shown in Table 1, AHOD is multi-site and hospital/clinic based and hence sites may be clustered around high-prevalence geographical locations and often there will be overlaps of attendance across geographically close sites. In AHOD for example there is strong clustering of sites around Sydney and Melbourne. However, AHOD also has wide geographical distribution across most of Australia and patients often move across sites.
Page 38 of 127 Table 1: Examples of international HIV cohorts in high resource settings by size and year of inception Country Cohort (ref) Started Multisite N Belgium St Pierre (7) 1982 N >5,000 Germany Frankfurt (6) 1987 N >3,600 Switzerland Swiss HIV Cohort Study (8) 1988 Y >18,000 France FHDH (9) 1989 Y >120,000 Denmark EuroSIDA (10) 1994 Y >18,500 Italy ICONA (11) 1997 Y >6,000 USA VACS (12) 1997 Y >40,0001 Australia AHOD (1) 1999 Y >3,800 Austria AHIVCOS (13) 2001 Y >7,500
UK CHIC (14) 2001 Y >50,000 Netherlands ATHENA (15) 2009 Y >15,000 1. Cohort also includes >80,000 HIV-negative controls matched on age, race and site.
Time at-risk, truncation and censoring Throughout this thesis time at-risk (analysis time) is selected so that subjects at equivalent durations and with the same prognostic features (collected in AHOD data) are likely to experience the same risk. In this thesis onset of risk is often set at time of cART commencement which is associated with rapid disease control (although certain analyses employ different specifications described in detail in respective chapters) but with appropriate baseline covariates being considered to adjust for individual differences in disease stage (e.g. baseline CD4+), time infected (i.e. duration of infection at date of cART commencement) or contemporary treatment and guidelines (e.g. calendar year category). Because many AHOD patients commenced cART prior to enrolment, their inclusion into at- risk groups may be lagged by the duration between commencement of cART and enrolment in AHOD (left truncation) in analyses which specify prospective follow-up as part of the inclusion criteria. While baseline clinical covariates and time-updated covariates associated with these patients are likely to reflect level of risk consistent with other patients, this method may still be subject to biases, such as survivor bias which is described in more detail below. Where possible, sensitivity analyses using only patients with prospective baseline dates (i.e. who were enrolled before or at time of cART commencement) were conducted to verify results. However, in AHOD such analyses can be limited by low numbers of patients and outcomes especially for endpoints such as death.
Patients are followed from entry up to a maximum calendar date equal to the administrative deadline applied to all AHOD data in any given year by data coordinators (right censoring) which removes site specific variations in reported time at-risk. Subject specific right censoring is also applied at instances of earlier times of LTFU from the cohort assigned according to analyses specific rules unless analyses draw on data linkage to ascertain patient outcomes irrespective of LTFU status. In certain instances interval
Page 39 of 127 censoring is also applied, for example in Chapter 5 where LTFU is considered as episodic and patients re-enter analysis at time of return from LTFU and with relevant time updated covariates.
Correlation and causality in longitudinal data Longitudinal data are characterised by the temporal ordering of within-subject observations as well as by correlation associated with clustering of observations about any given individual subject (3). That is, there is normalised linear dependence of responses for repeated measures on an individual. While longitudinal study design provides potentially more precise estimates of change and of the effect of covariates than other study designs, the assumption of independence used by many standard analytical techniques does not hold and must also be accounted for (discussed in following sections).
Correlation among repeated measures on an individual can be affected by variability arising from any given individual’s propensity to have a certain level of characteristic response and response trajectory for all repeated measures attributed to genetic, environmental, social and behavioural factors. The random variation of individual level response around a given propensity response may also be attributable to subject specific underlying biological processes or measurement error (3). Additionally, inaccurate statistical modelling of a mean response (misspecification) will influence the correlation (to the extent that these are defined in terms of the model for the mean response) and result in overdispersal of observations.
In this thesis, inferences about causality are drawn based on strength of observed associations between covariates and outcome, in conjunction with expert clinical interpretation and evidence available from prior studies and published literature. While model covariates exhibiting strong associations with outcome may be considered to be strong markers of risk within the context of any given analysis, causal or deterministic mechanisms cannot necassarily be attributed to these associations. Generally accepted risk factors, often with clinically established causal pathways, are included in analyses as covariates wherever possible to facilitate comparison of results with other studies. However, many analyses involve a complex response of multiple components but do not explicitly establish causal pathways. In particular any given model covariate may not be necessary to entail a given observed outcome (16). Further, certain component causes may not be available in AHOD data especially for less specific risk factors not proximal to outcome. Therefore discussion and interpretation of observed association is cautious and draws extensively on existing literature.
Page 40 of 127 Missing Data and loss to follow-up Longitudinal data in health sciences and especially in observational studies rarely comprises a complete sequence of observation times for all individuals (i.e. data are unbalanced) and are subject to missing data. For example, patients may miss scheduled visits or leave the study prior to completion especially over prolonged durations of follow-up associated with cohort studies such as AHOD (see Chapter 5) where non-response can occur on any given occasion. Such incompleteness of data entails loss of information, efficiency and precision of estimation. Importantly missing data can introduce bias in estimates and needs careful consideration.
Often the reasons for missing data are not completely understood and as such assumptions about the reasons for missing data need to be made. In certain analyses data is assumed to be missing completely at random (MCAR) and hence no bias in estimates of mean response is assumed to arise because the observed data are a random subset of overall responses. However, where the probability of a response being missing is related to the response values then bias may potentially be introduced to sample means. In these analyses we often assume that the probability of a response being missing is unrelated to the likely response values (i.e. data are missing at random (MAR)) and restrict estimates and discussion to observable strata of response (e.g. survival analyses are often restricted to below 70 years of age because of low numbers of patients at older ages). However, in any analysis, we consider the possibility of data being not missing at random (NMAR) and that therefore assumptions in any modelling of non-response are unverifiable from the data at hand. Generally in these instances resultant potential bias in estimates is addressed by reference to relevant published literature and expert opinion.
In this thesis “loss to follow-up” (LTFU) describes the prolonged and unexpected absence of a patient from clinic beyond scheduled treating durations and in AHOD is often causative of missing data (including both intervals of missing data as well as permanent drop-out from study). In Chapter 4 we investigate survival under the assumption that episodes of LTFU can be treated as MCAR in part because the low numbers of LTFU are unlikely to influence model accuracy or precision regardless of the strength of the MCAR assumption especially for most model strata. In Chapter 5 we verify the strength of this assumption by investigating survival in LTFU using precise definitions of LTFU.
Error, chance and bias Instances and consequences of chance or random error resulting from biological variation, measurement error, as well as low precision resulting from reduced sample size attributable to missing data (sampling error) have been discussed. Generally estimators are presented
Page 41 of 127 with 95% confidence intervals to indicate the precision of results. Briefly, we expect the confidence interval to represent a range of plausible values which in repeated sampling from the population includes the population value 95% of the time. However, we accept that non- representative observations may have arisen by chance alone especially for marginal p- values and for studies with relatively small sample size or degrees of freedom. For example, in Chapter 6 we model outcomes of suicide and accidental or violent death using low numbers of endpoints (<100). In this study results are therefore taken as indicative and contrasted with other studies and also the need for cautious interpretation is emphasised. The actual results are discussed more fully in this light in Chapter 6.
In the preceding we have also indicated how systematic error or poor accuracy (bias) can be introduced to estimates including where there is association between observed response values and missing response values; where data is not missing at random; and where there is systematic measurement error (information bias).
In cohort studies the selection of non-representative subsets of the population according to any given prognostic factor (selection bias) can also lead to inaccurate estimation. Specifically in AHOD we consider selection bias arising from volunteerism where participants might be more likely to exhibit propensity to certain levels of health; LTFU (discussed in Chapter 5); and ascertainment bias including where patients with improved health are more likely to have survived long enough to be recruited to the cohort (healthy survivor bias) (4).
In AHOD and this thesis in general, methodological approaches are taken to minimise inaccuracy arising from bias. As discussed above, recruitment to AHOD has been from a diverse range of health care providers across a long period and wide geographic range and substantial number of diverse HIV+ participants. Participation in the study requires only signed consent and no change to routine treatment and hence attrition through demands of participation is likely minimal. In this regard selection bias associated with volunteerism and LTFU are somewhat mitigated. AHOD in particular has relatively low rates of LTFU although the effects of LTFU on estimates of survival are investigated using linkage to the National Death Index in Chapter 5. While this does not necessarily remove ascertainment bias associated with migration of people who are more likely to die (including salmon bias where less healthy people might return to home nations to die), these effects are likely minimal given low overall rates of LTFU as well as relatively low numbers of participants born outside of Australia.
Page 42 of 127 Confounding In this thesis measures are taken to control potential confounding by stratification of analyses on, or inclusion in multivariate models of, potential or known confounders (such as age or site). However, consideration is given to the possible distortion of the observed association between response and exposure by independent exogenous factors. Therefore, in all analyses consistency of observed estimators is confirmed using available literature and prior studies.
2.3. Statistical models
Descriptive statistics In this thesis, descriptive overview of the at-risk population is generally provided as a preliminary procedure for any given analysis using various descriptive statistics. Distributions of prognostic factors are shown as patient counts per covariate category at a common baseline (usually cART commencement) or common intervals. Crude incidence rates of response (count of events per covariate category over analysis time at-risk) are also used. For survival analyses unadjusted description of overall survivor experience is provided using crude estimates of the survivorship function using the Kaplan-Meier (KM) method (17). In certain analyses mean or median levels of response are analysed over time at-risk according to different prognostic categories, but differences in response according to categorical level are also compared using regression methods.
Regression methods In this chapter, regression modelling refers to the development of mathematical models to describe the association of the mean of some form of a response variable such as death, to a set of covariates (17). Ordinary least squares regression methods of modelling time to event as response cannot generally be used in analyses used in this thesis because the assumption of normality of time to event does not hold. Instead survival analyses used in this thesis develop models of response which do not use assumed normal distribution of times to event. In particular assumptions about changes in hazard over time at-risk are applied and the hazard function (the conditional failure rate) associated with categorical levels of prognostic factors is modelled.
Cox proportional hazards (CPH) models are used frequently throughout this thesis to explore relative differences in hazard of prognostic categories. Unlike parametric models of survival which specify the density functions of outcome, CPH models (classified as semi parametric models) are based on the hazard function, specifically the hazard ratio, and consequently a reduced set of assumptions is required to develop clinically meaningful estimators (hazard ratios) compared to parametric models (17). The key assumptions of CPH models are that
Page 43 of 127 the effect of prognostic factors on hazard is multiplicative and does not change over time. The validity of the proportional hazards assumption is tested in analyses to verify validity of use of CPH for any given analysis where used.
Parametric regression models are also used in this thesis to model outcome for given prognostic scenarios, after preliminary investigation of covariate hazard ratios using CPH models. Exponential type distributions, in particular, Weibull and Poisson models are used to explore survival times for a range of prognostic scenarios in this thesis and detail of the model fitting procedures used are described in relevant chapters (e.g. see Chapter 4). The parameterisations used in this thesis assume the effect of prognostic factors to be multiplicative on the time scale (accelerated failure time models) but with proportional hazards (3, 17).
In certain analyses used in this thesis, especially where there is a focus on changes in mean response (such as Chapter 3), longitudinal AHOD data is converted to panel form where covariates and outcomes are recorded for/at given intervals of time for each patient and the mean response modelled using regression models. This form of model specification can make use of relatively balanced AHOD data for given panel intervals to develop robust estimates of mean response. In particular generalised estimating equations (GEE) are used to model the mean response as a function of prognostic factors alone (and not of any previous responses or random effects). These models specify conditional expectation and variance of the mean response for given prognostic factors which is similar to methodology used for standard generalised linear models but without making distributional assumptions about the response (18). GEE models yield consistent parameter estimates independent of incorrect specification of the model of covariance among repeated measures and standard errors are robust to incorrect model specification for within-subject associations for large sample sizes (i.e. are asymptotically robust) (18).
2.4. Notes on particular covariate specifications used in AHOD
Analyses conducted in this thesis consider a range of demographic and clinical prognostic factors (covariates). Covariates are specified with consideration to generalizability of results and to facilitate comparison with other studies. Analyses take the approach of using all available data. Baseline covariate specifications are used which have low rates of missing data.
Intervals between CD4+ and VL testing for any given patient in AHOD vary according to a variety of component factors including particular patient, testing site, calendar year and disease stage. Generally analyses in this thesis incorporating CD4+ and VL as time updated
Page 44 of 127 covariates carry forward CD4+ and VL values until next date of measurement unless that interval exceeds a specified maximum analysis specific duration (often 180 days). CD4+ and VL are then usually treated as categorical variables using analysis specific cut points. The specifications for these covariates are described in detail in relevant chapters according to the particular analysis.
Repeated measures of hepatitis testing (HCV and HBV) are not recorded in AHOD. Instead “positive ever” of HCV or HBV is used rather than a time-updated specification. While this might be less informative than using a time-updated specification, this specification accords with the method of data collection. For HCV and HBV “no record” and “negative result” are often conflated into one category (“never positive”). This is based on clinical feedback that there would otherwise be detection bias associated with the increased rates of testing and recording results in symptomatic or high risk patients, while low risk patients are not routinely tested. Importantly, these specifications are consistent and are unbiased subject to correct interpretation and discussion.
Unless explicitly stated, treatment is categorised as a binary variable (having commenced cART or not) and analyses employ intention-to-treat (ITT) type methodology (19). Specifically, patients are categorised as having commenced cART at recorded commencement date (usually defined as the commencement of therapy comprising 3 or more drugs from 2 or more drug classes) regardless of ongoing adherence, treatment interruption (TI) or cessation, regimen type or dose strength. Under this specification analyses do not compare differences in prognosis arising from different adherences or types of cART. This methodology overcomes problems arising from missing data or measurement error associated with ART therapy which is otherwise complex and susceptible to such measurement errors especially in cohort studies over extended durations which might miss detail on adherence, regimen switching, dose strength and potentially introduce bias to estimates. However there may be some dilution in estimates of the effectiveness of cART because intervals of TI are specified as continued treatment.
In Chapter 3, descriptive analysis of time to first treatment switch is conducted using the KM method (17). In this analysis treatment switch is defined as the commencement of 2 or more new antiretrovirals and/or the commencement of a new drug class. Under the ITT-type specification of cART status discussed above, TI and cessations are not specified in models and such intervals are treated as continuations of cART treatment. Under this specification, the commencement of 2 or more new antiretrovirals or new class after a TI compared to the last regimen immediately prior to that TI is considered a treatment switch. Analyses are not censored at time of treatment switch and therefore incidence of first treatment switch is
Page 45 of 127 independent of TI under this specification. A consequence of this is that TI are not treated as competing risks for switching. Cumulative incidence (CI) estimates of switching based on data censored at time of TI will be larger than estimates derived from a joint distribution of competing risks (20). Conversely, for similar reasons described in the previous paragraph, our estimates will tend to be lower than estimates using TI censored data (i.e. because of time at-risk including TI).
In AHOD detailed information has been collected on cause of death since study inception using standard cause of death forms (CoDe) (21) completed by site clinicians and forwarded to AHOD coordinators. Additionally in Chapter 5, linkage to the National Death Index was conducted to establish fact of death for all AHOD patients including LTFU. This is described in detail in Chapter 7. Inherent in that method is assumed low rate of emigration (discussed in that chapter) which would otherwise alter results if mortality was particularly high amongst this group (e.g. salmon bias).
2.5. References
1. The Australian HIV Observational Database: The Kirby Institute; 2014 [cited 2014]. Available from: https://kirby.unsw.edu.au/projects/australian-hiv-observational-database- ahod. 2. Australian HIV Observational Database Annual Report. In: Institute TK, editor.: The University of New South Wales; 2013. 3. Fitzmaurice GM, Laird NM, Ware JH. Applied longitudinal analysis Part 1. 2nd ed. Hoboken, N.J.: Wiley; 2011. xxv, 701 p. 4. Fletcher RH, Fletcher SW, Fletcher GS. Clinical epidemiology : the essentials. 5th ed. Philadelphia: Wolters Kluwer/Lippincott Williams & Wilkins Health; 2014. 253 p. 5. Porta MS, Greenland S, Hernán M, Silva IdS, Last JM, International Epidemiological Association. A dictionary of epidemiology. Six edition / ed. Oxford: Oxford University Press; 2014. xxxii, 343 p. 6. Frankfurt HIV Cohort: JW Goethe -Universität; [cited 2014 12/12/2014]. Available from: http://www.hivforum.org/storage/documents/Cohorts/frankfurt%20hiv%20cohort.pdf. 7. Brussels Saint-Pierre Cohort: Infectious Diseases Department – Saint-Pierre University Hospital Brussels 2014 [cited 2014 12/12/2104]. Available from: http://www.infectio- saintpierre.be/web/index.php?option=com_content&view=category&layout=blog&id=22&lang =en. 8. Swiss Cohort Study: SHCS; [cited 2014 12/12/2014]. Available from: http://www.shcs.ch/93-shcs-tables#Demographical_characteristics.
Page 46 of 127 9. French Hospitals database: UPMC; 2014 [cited 2014 12/12/2014]. Available from: http://www.ccde.fr/main.php?main_file=fl-1309272043-794.html. 10. EuroSIDA: EuroSIDA; [cited 2014 12/12/2014]. Available from: http://www.eurocoord.net/partners/founding_networks/eurosida.aspx. 11. ICONA: ICONA; [cited 2014 12/12/2014]. Available from: http://www.fondazioneicona.org/_new/. 12. VACS: Yale School of Medicine; [cited 2014 12/12/2014]. Available from: http://medicine.yale.edu/intmed/vacs/. 13. 23rd Report of the Austrian HIV Cohort Study: ISSUU; [cited 2014 12/12/2014]. Available from: http://issuu.com/agesnews/docs/ages_hiv-bericht_2013_ebook/25. 14. CHIC: UCL; [cited 2014 12/12/2014]. Available from: http://www.ucl.ac.uk/iph/research/hivbiostatistics/CHIC. 15. ATHENA: ATHENA Network; [cited 2014 12/12/2014]. Available from: http://www.athenanetwork.org/. 16. Pearl J. Causality : models, reasoning, and inference. Cambridge, U.K. ; New York: Cambridge University Press; 2000. xvi, 384 p. 17. Hosmer DW, Lemeshow S, May S. Applied survival analysis : regression modeling of time-to-event data. 2nd ed. Hoboken, N.J.: Wiley-Interscience; 2008. xiii, 392 p. 18. Fitzmaurice GM, Laird NM, Ware JH. Applied longitudinal analysis, Part 2. 2nd ed. Hoboken, N.J.: Wiley; 2011. xxv, 701 p. 19. Piantadosi S. Clinical trials : a methodologic perspective. 2nd ed. Hoboken, N.J.: Wiley-Interscience; 2005. xxvii, 687 p. 20. Bakoyannis G, Touloumi G. Practical methods for competing risks data: a review. Statistical methods in medical research. 2012;21(3):257-72. 21. Kowalska JD, Friis-Moller N, Kirk O, Bannister W, Mocroft A, Sabin C, et al. The Coding Causes of Death in HIV (CoDe) Project: initial results and evaluation of methodology. Epidemiology. 2011;22(4):516-23.
Page 47 of 127 Chapter 3: Recent trends in early stage response to combination antiretroviral therapy in Australia
McManus H, Hoy JF, Woolley I, Boyd MA, Kelly MD, Mulhall B, et al. Recent trends in early stage response to combination antiretroviral therapy in Australia. Antivir Ther. 2014 10.3851/IMP2774
Page 48 of 127
Recent trends in early stage response to combination antiretroviral therapy in Australia
Hamish McManus, Jennifer F Hoy, Ian Woolley, Mark A Boyd, Mark D Kelly, Brian Mulhall, Norman J Roth, Kathy Petoumenos, Matthew G Law, the Australian HIV Observational Database
Antiviral Therapy 2014; 10.3851/IMP2774
Submission date 21st January 2014 Acceptance date 16th March 2014 Publication date 4th April 2014
For information about publishing your article in Antiviral Therapy go to http://www.intmedpress.com/index.cfm?pid=12
©2014 International Medical Press ISSN 1359-6535
Page 49 of 127 Publication: Antiviral Therapy; Type: Original article
Recent trends in early stage response to combination antiretroviral therapy in Australia Hamish McManus1*, Jennifer F Hoy2,3, Ian Woolley3,4,5, Mark A Boyd1, Mark D Kelly6, Brian Mulhall1,7, Norman J Roth8, Kathy Petoumenos1, Matthew G Law1, the Australian HIV Observational Database 1The Kirby Institute, University of New South Wales, Sydney, Australia 2The Alfred Hospital, Melbourne, Australia 3Department of Medicine, Monash University, Melbourne, Australia 4Department of Infectious Diseases, Monash University, Melbourne, Australia 5Infectious Diseases, Monash Medical Centre, Melbourne, Australia 6Brisbane Sexual Health Service, Brisbane, Australia 7School of Public Health, University of Sydney, Camperdown, Australia 8Prahran Market Clinic, Prahran, Australia *Corresponding author e-mail: [email protected]
Abstract Background: There have been improvements in combination antiretroviral therapy (cART) over the last 15 years. The aim of this analysis was to assess whether improvements in ART have resulted in improvements in surrogates of HIV outcome. Methods: Patients in the Australian HIV Observational Database who initiated treatment using mono/duo therapy prior to 1996, or using cART from 1996 onwards, were included in the analysis. Patients were stratified by era of ART initiation. Median changes in CD4+ and the proportion of patients with detectable HIV viral load (>400 copies/ml) were calculated over the first 4 years of treatment. Probabilities of treatment switch were estimated using the Kaplan-Meier method. Results: 2,753 patients were included in the analysis: 28% initiated treatment <1996 using mono/duo therapy; and 72% initiated treatment .1996 using cART (30% 1996-99; 12% 2000-03; 11% 2004-07; and 19% .2008). Overall CD4 response improved by later era of initiation (p<0.001), although 2000-03 CD4 response was less than that for 1996-99 (p=0.007). The average proportion with detectable viral load from 2 to 4 years post treatment commencement by era was: <1996 mono/duo 0.69 (0.67-0.71); 1996-99 cART 0.29 (0.28-0.30); 2000-03 cART 0.22 (0.20-0.24); 2004-07 cART 0.09 (0.07-0.10); .2008 cART 0.04 (0.03- 0.05). Probability of treatment switch at 4 years after initiation decreased from 53% in 1996-99 to 29% after 2008 (p<0.001). Conclusions: Across the five time-periods examined, there have been incremental improvements for patients initiated on cART, as measured by overall response (viral load and CD4 count), and also increased durability of first-line ART regimens.
Page 50 of 127 Publication: Antiviral Therapy; Type: Original article
Accepted 16 March 2014, published online 4 April 2014 Running Head: Trends in HIV treatment response in Australia
Introduction
There have been incremental improvements in combination antiretroviral therapy (cART) over the last 15 years, including increased potency, reduced pill burden, combination formulations and new drug classes [1–5]. Determining the extent to which these are reflected by improvements in surrogates of HIV outcome, such as median CD4 cell count and the proportion of patients with an undetectable HIV viral load, is important for developing health care management strategy but analysis is challenging at the epidemiological level.
Responses to cART can be affected by individual patient factors including baseline clinical status and treatment history, genotypic variation, drug interactions, toxicities and tolerability, as well as lifestyle [6]. The complex interaction of these factors is difficult to explicitly model, especially without limiting the scope of results, and can lead to differences between the estimated and actual utility of treatment. Descriptions of population level trends based on large-scale longitudinal studies can produce generalizable results by avoiding confounding inherent in smaller studies of discrete populations.
Conversely, large-scale long-term longitudinal studies (often observational) can be confounded by temporal trends associated with both treatment stage and calendar year. The pattern of surrogate response to cART is strongly characterised by duration of treatment, particularly by the first months after treatment initiation when there is often a rapid decrease in HIV viral load [7,8]. Outcomes are also influenced by broader changes to available ART, contemporary resourcing, guidelines and interventions associated with calendar year [9,10]. For example, Australian cART initiation guidelines have shifted from early stage intervention (“hit hard, hit early”) (1996-99), to delayed intervention (at CD4 cell counts <200 cells/µL in asymptomatic patients) (2000), and back to relatively early stage intervention (at CD4 cell counts 350 cells/µL in asymptomatic patients) (2008) [11–13]. However, few analyses of population level trends in treatment response simultaneously resolve both of these time scales which may in part be because many studies do not have both long- term follow up and ongoing recruitment necessary to reflect these changes.
In particular, early post-treatment initiation response is strongly indicative of overall treatment utility. At early stages (less than 4 years after treatment initiation), response is most strongly influenced by pre-initiation conditions and baseline disease stage as well as by the potency of implemented ART, and is also predictive of later outcomes [14–17]. Generally, significant levels of virological control are reached within a year of treatment [7,8,18,19]. Phillips et al have shown attainment of at least one viral load measure less than 500 copies/ml in 85% of patients within 32 weeks after initiating therapy for early cART era patients, but that there were subsequently increasing rates of viral rebound over the duration of that study (up to 192 weeks) [7]. Subsequent immunological
Page 51 of 127 Publication: Antiviral Therapy; Type: Original article repletion has been shown to be more gradual (between 4 and 8 years in patients with sustained virological control) and strongly influenced by baseline levels [16].
We investigated trends in response to ART over the first 4 years of treatment by duration of treatment and calendar year of treatment initiation using the Australian HIV Observational database (AHOD). AHOD is a large observational cohort study with extensive treatment and clinical data, and ongoing patient recruitment. The aim of this analysis was to assess whether improvements in ART over calendar time have resulted in improvements in surrogates of HIV outcome.
Methods
Study population The Australian HIV Observational Database (AHOD) is an observational clinical cohort study of patients with HIV infection seen at 27 clinical sites throughout Australia. AHOD uses methodology which has been described in detail elsewhere [20]. Data are transferred electronically to The Kirby Institute, University of New South Wales every 6 months. Prospective data collection commenced in 1999, with retrospective data provided where available.
Ethics approval for the AHOD study was granted by the University of New South Wales Human Research Ethics Committee, and all other relevant institutional review boards. Written informed consent was obtained from participating individuals. All study procedures were developed in accordance with the revised 1975 Helsinki Declaration.
This analysis included all patients who had been recruited to AHOD as part of general AHOD recruitment prior to 31 March 2013 who initiated treatment using mono therapy/dual therapy (mono/duo therapy) prior to 1996, or using cART from 1996 onwards. Patient follow-up was from treatment initiation until censoring at the earlier of death, lost to follow-up, 4 years after treatment commencement, or cohort censoring date (31 March 2013). Patients were defined as lost to follow-up at date of last clinical visit if they had no recorded visit in the year prior to cohort censoring date (31 March 2013). Patients were retained in the analysis until censoring regardless of interruptions or changes to treatment using an intention-to-treat approach. Patients were stratified by time period of treatment initiation as follows: <1996 mono/duo, 1996-99 cART; 2000-03 cART; 2004-07 cART, 2008 cART. Patients were excluded if they did not have a record of CD4 or HIV-RNA tests after commencing ART.
Covariates analysed were: sex; age (“<20”, “20-29”, “30-39”, “40-49”, “50-59”, “60-69”, “70- 79”); mode of HIV exposure (men who have sex with men (MSM), heterosexual, injecting drug user (IDU), other/unknown); time updated instance of first AIDS defining illness (ADI); HCV antibody (no/not tested, positive); HBV surface antigen (HBVsAg) (no/not tested, positive); CD4 cell count (from the closest test date between 180 days prior to a given analysis time-point and up to 30 days after that time-point – “0-199”,”200-349”, “350–499”, “500” cells/ml); viral load (from the closest test date between 180 days prior to a given analysis time-point and up to 30 days after that time-point –
Page 52 of 127 Publication: Antiviral Therapy; Type: Original article
“400”, “401-1000”, “1001-10,000”, >10000 copies/ml) and drug class (“Nucleoside/Nucleotide Reverse Transcriptase Inhibitors (NRTI)”, “Non-Nucleoside Reverse Transcriptase Inhibitors (NNRTI)”, “Protease Inhibitors (PI)”, “Integrase Inhibitors (II)”).
Statistical Analysis Patient characteristics at start of ART were summarised by era of ART commencement.
Median and interquartile range for CD4 cell count and viral load at start of treatment was summarised by era. Kruskal-Wallis rank test of equality of baseline CD4 cell count by era was performed. Median changes in CD4 cell counts for the first 4 years of treatment were calculated using the closest CD4 test result to each 90 day cut point following first ART per patient.
Percentage of patients with detectable viral load (>400 copies/ml) for the first 4 years of treatment was calculated using the closest viral load test result to each 90 day cut point following first ART per patient. The threshold of >400 copies/ml was used to allow comparison of all treatment eras including earlier treatment eras when more sensitive assays were not available. Kruskal-Wallis rank test of equality of the log10 of baseline viral load by era was performed.
Equality of response by era for the first 4 years after cART was assessed using multivariate generalised estimating equation modelling with assumed exchangeable variance structure, and robust calculated variances. Models were adjusted for era, age group, IDU mode of exposure, prior ADI and baseline CD4. For immune response, a normal distribution of change in CD4 cell count per unit time was assumed, while for virological response a binomial distribution of detectable viral load with logistic link was used. Wald tests of the overall significance of era (excluding the mono/duo era) and pairwise comparison of the predicted probability of CD4 cell response conditional on era category levels (excluding the mono/duo era) were also conducted. The mono/duo era was excluded from tests of response by era because mono/duo regimen response has been shown to be minimal in comparison to cART response [21] and would have otherwise masked the differentiation of cART era response using overall tests of the association of era with response.
Rates of first switch of treatment were calculated by era for all switches; for switches with recent virological failure (from the closest viral load test within 180 days prior to switch and up to 30 days after switch >1000 copies/ml); for switches without recent virological failure (closest viral load test within 180 days prior to switch and up to 30 days after switch 1000 copies/ml); and for switches without a recent viral load reading (closest viral load test was over 180 days prior to switch and/or over 30 days after switch). A switch was defined as the commencement of 2 or more new antiretroviral drugs and/or the commencement of a new class.
Kaplan-Meier curves by era were generated from treatment start until first change of regimen for up to 4 years after baseline for specific treatment switch types (all switches, switch with recent virological failure, and switch without recent virological failure). A log rank test of equality of curves was conducted but excluding the mono/duo era for reasons discussed above.
Page 53 of 127 Table 1 Patient characteristics at start of treatment by treatment era and type Era Early (Mono/Duo) Early Middle Middle/Late Late (<1996) (1996-99) (2000-03) (2004-07) (2008+) N (%) N (%) N (%) N (%) N (%) N 767 (100) 817 (100) 330 (100) 310 (100) 529 (100) Gender F 41(5) 39(5) 21(6) 26(8) 42(8) M 726(95) 778(95) 309(94) 284(92) 487(92) Age (years) mean (SD) 37(8.9) 38(9.8) 40(10.4) 43(9.9) 42(10.9) <20 7(1) 7(1) 1(0) 1(0) (0) 20-29 176(23) 155(19) 48(15) 23(7) 70(13) 30-39 335(44) 349(43) 136(41) 114(37) 168(32) 40-49 181(24) 200(24) 90(27) 104(34) 179(34) 50-59 62(8) 87(11) 41(12) 48(15) 79(15) 60-69 6(1) 15(2) 11(3) 18(6) 27(5) 70-79 0(0) 4(0) 3(1) 2(1) 6(1) Mode of exposure MSM 642(84) 632(77) 221(67) 202(65) 376(71) Heterosexual 63(8) 88(11) 71(22) 84(27) 107(20) IDU 38(5) 67(8) 21(6) 14(5) 22(4) Other 24(3) 30(4) 17(5) 10(3) 24(5) Prior ADI No 706(92) 718(88) 278(84) 270(87) 491(93) Yes 61(8) 99(12) 52(16) 40(13) 38(7) HCV ever No 585(76) 654(80) 263(80) 251(81) 421(80) Yes 105(14) 92(11) 33(10) 26(8) 50(9) Not recorded 77(10) 71(9) 34(10) 33(11) 58(11) HBV ever No 653(85) 670(82) 257(78) 239(77) 415(78) Yes 42(5) 43(5) 14(4) 6(2) 11(2) Not recorded 72(9) 104(13) 59(18) 65(21) 103(19) CD4 cell count (cells/µl) Mean (SD) 330(219.8) 342(232.1) 307(257.8) 298(217.2) 338(189.0) Median (IQR) 294(190-435) 320(170-475) 250(120-420) 250(170-373) 320(220-430)
Log 10 viral load Mean (SD) 5.16 (0.78) 4.75 (0.95) 4.88 (0.89) 4.62 (1.14) 4.53 (0.98) Median (IQR) 5.16 (4.61-5.71) 4.89 (4.28-5.43) 5.00 (4.60-5.51) 4.99 (4.26-5.35) 4.78 (4.13-5.04) Regimen class1 NNRTI 5/767 (1) 346/817 (42) 195/330 (59) 205/310 (66) 350/529 (66)
NRTI 766/767 (100) 815/817 (100) 330/330 (100) 310/310 (100) 526/529 (99)
PI 4/767 (1) 472/817 (58) 113/330 (34) 104/310 (34) 138/529 (26)
II 0/767 (0) 0/817 (0) 0/330 (0) 1/310 (0) 74/529 (14) 1. NNRTI= Non-Nucleoside Reverse Transcriptase Inhibitors, NRTI=Nucleoside/Nucleotide Reverse Transcriptase Inhibitors, PI=Protease Inhibitors, II=Integrase Inhibitors
Page 54 of 127 (lnfSII8~) OOv 09£ lUno~ 00£ 092: 118~ OOG P08 09 U! ~ a5ueL.p 00 ~ 09 UE!P8W 0 0 N ...... i+= +-' +-' +-' +-' I- I- I- I- c Cl) 0 Cl) Cl) cu E cu E tf) tf) >. - I- I- (0 ('() 0 0 0 0 0 c.o <( <( mo m..;:t m m m m n:: n:: ~ ~ ~ N V I I :::J c 0 0 0 C) C) I i • • • • • I I - - I- I- 1'-- ('() 0 0 0 0 CO <( <( n:: n:: ~ N N I I I C) C) I I I -
Page 55 of 127
Figure 1 Median change in CD4 cell count after first treatment by first treatment type and era of treatment commencement1 1 Closest patient CD4 cell count test date to each 90 day interval following first treatment; global Wald test for era (excluding mono/duo era) p<0.001 (based on GEE model using normal family, identity link, adjusted for age, IDU exposure, prior ADI and baseline CD4) Publication: Antiviral Therapy; Type: Original article
Results
Patient characteristics By March 2013, 3,607 patients had been recruited to AHOD as part of general recruitment, of whom 854 were excluded from the analysis: 327 had no record of treatment initiation; 481 had treatment commencement prior to or later than analysis treatment era censoring dates; and 46 had no record of subsequent CD4 cell count or viral load test. The remaining 2,753 patients were included in the analysis (Table 1): 767 (28%) from the mono/duo treatment era (<1996); 817 (30%) from the early cART era (1996-99); 330 (12%) from the middle cART era (2000-03); 310 (11%) from the middle/late cART era (2004-07); and 529 (19%) from the late cART era (2008).
Patients were predominantly male (94%), and mode of exposure was predominantly via MSM (75%). The mean age at ART initiation was 39.4 years (standard deviation (SD) 10.08), which increased by later era (38.4 (SD 9.79) 1996-99, 42.1 (SD 10.9) 2008). Over cART eras the proportion of patients commencing with NNRTI containing regimens increased from 42% (1996-99) to 66% (2008), and the number commencing with PI based regimens fell from 58% (1996-99) to 26% (2008).
Immunology Base line CD4 cell counts were relatively similar by era of treatment commencement but were slightly lower for middle era commencements. Median CD4 cell counts at commencement were: 294 cells/µl (IQR 190-435) for <1996 mono/duo; 320 cells/µl (IQR 170-475) for 1996-99 cART; 250 cells/µl (IQR 120-420) for 2000-03 cART; 250 cells/µl (IQR 170-373) for 2004-07 cART; and 320cells/µl (IQR 220- 430) for 2008 cART (Table 1). Kruskal-Wallis rank test of baseline CD4 cell count by era (excluding mono/duo) showed significant change over time (p<0.001).
Median CD4 cell counts increased by year from first treatment for cART regimens. Median CD4 cell count change from baseline was greatest in the first year after treatment (Figure 1).
A global Wald test indicated that era of commencement was a significant predictor of change from baseline CD4 (p<0.001) in an adjusted model. Pairwise comparison of predicted change in CD4 cell count by different eras of treatment initiation (excluding the mono/duo era) showed difference between 2000-03 cART compared to all other eras (1996-99 cART p=0.007, 2004-07 cART p=0.006, 2008-13 cART p<0.001), and marginal difference between 1996-99 cART compared to 2008-13 cART (p=0.044). There was no difference for other pairwise comparisons (1996-99 cART compared to 2004-07 cART p=0.603; and 2004-07 cART compared to 2008-13 cART p=0.252).
Virology
Baseline viral loads were elevated across all eras of treatment commencement. Medians of log10 of viral load copies/ml were similar by era of commencement: 5.16 (IQR 4.61-5.71) for <1996 mono/duo; 4.89 (IQR 4.28-5.43) for 1996-99 cART; 5.00 (IQR 4.60-5.51) for 2000-03 cART; 4.99 (IQR 4.26-5.35)
Page 56 of 127 PBOIIBJ!A ~ 6' 8 ... ' .,.. a1qBl~alaP L' .. .-- ...... g · 9' J •• Lll!M /: v· I • • ' • I • I I • • • s· I i I • • I • • • f • \ • • ' • • I • a6BlUa~Jad • • \ • I • \ l ' • z:· ~· 0 0 N ...... ~ +-' +-' +-' +-' c en en 0 E E Q) Q) Q) ro ro \- \- \- \- >- 1-1- 1- ...-o ...-N 01- 0)0 0)0 -o::: <000 0)...- ('f) (0 0:::0::: C) 0 0 0 0 0::: <(<( mo <( m-.:::r ~ N N V I I :::J g (.) 0 0 (.) I i • • • • • I I - - 1'- ('f) 0 <( I I I (.) (.) I I I - Page 57 of 127
Figure 2 Percentage of tests with undetectable viral load by first treatment type and era of treatment commencement1 1 Closest patient viral load test date to each 90 day interval following first treatment. Viral loads less than or equal to 400 cells/ul categorised as undetectable; global Wald test p-value for era p<0.001 (based on GEE model using normal family, identity link, adjusted for age, IDU exposure, prior ADI and baseline CD4) Publication: Antiviral Therapy; Type: Original article
for 2004-07 cART; and 4.78 (IQR 4.13-5.04) for 2008 cART. Kruskal-Wallis rank test of log10 of baseline viral load by era (excluding mono/duo era) showed a significant lower viral load at initiation in later eras (p<0.001).
Median viral load dropped to below detectable thresholds after the first year of treatment for patients commencing on cART from relatively similarly high baseline levels. The proportion of patients with detectable viral load more rapidly reached lower thresholds by later era of treatment commencement (Figure 2). The average proportion with detectable viral load per quarter from 2 years to 4 years post treatment commencement by era was: <1996 mono/duo 0.69 (0.67-0.71); 1996-99 cART 0.29 (0.28-0.30); 2000-03 cART 0.22 (0.20-0.24); 2004-07 cART 0.09 (0.07-0.10); 2008 cART 0.04 (0.03-0.05).
Treatment Switching The rate of first treatment switches by era and viral measure category is shown in Table 2. The overall rate of switching was 17.3 (95% CI 16.4-18.3) per 100 person years. There were declines by cART era in the rate of switching overall (p<0.001); with recent virological failure (p<0.001); and with unrecorded recent viral record (p<0.001). There was no change in the observed rate of switching without recent viral failure (p=0.428).
Kaplan Meier probability of having switched treatment (all switches) decreased by later era of cART commencement (Figure 3A). Kaplan Meier probability of having switched treatment (all switches) after 4 years was 0.58 (95% CI 0.54-0.61) for <1996 mono/duo; 0.53 (95% CI 0.50-0.57) for 1996-99 cART; 0.44 (95% CI 0.39-0.50) for 2000-03 cART; 0.39 (95% CI 0.34-0.45) for 2004-07 cART; and 0.29 (95% CI 0.25-0.35) for 2008 cART. Log rank test of time to switch by era (excluding the mono duo era) showed significant difference (p<0.001).
Kaplan Meier probability of having switched treatment with recent virological failure also decreased by later era of cART commencement (p<0.001) (Figure 3B) but there was not a difference for switches without recent virological failure (p=0.424) (Figure 3C).
Discussion
We measured trends in early stage response to treatment for a large HIV-positive cohort. Over the first 4 years of treatment CD4 cell count response was marginally improved by later cART era of treatment commencement although 2000-03 cART commencements were seen to have relatively less CD4 cell count recovery. However, there were significantly improved rates of virological suppression by later era of treatment commencement for all eras. We found improving durability of ART over the same period of follow-up.
Median baseline CD4 cell counts were low across all eras of treatment initiation (<350 counts/µL), although levels were relatively lower for middle eras (2000-2007) (p<0.001). There was indication of difference in rates of immunological recovery across different eras of treatment initiation. Plots of median change in CD4 cell count and pairwise comparison of CD4 cell count response by era
Page 58 of 127
Page Kaplan-Meier probabilities of time to first switch by era. A, All switches. B, Switches with recent virological failure. C, Switches without recent 59 virological failure1 of
127 1 Virological failure defined as viral load >1000 copies/ml at most recent test within 180 days prior to and up to 30 days after switch; Log rank test of era for (excluding mono/duo era) A: p<0.001; B: p<0.001; C: p=0.424 Table 2 Rate of first regimen switch in first 4 years of treatment by switch type and treatment era Early Early Middle Middle/ Late Mono/Duo cART cART Late cART cART p2 All eras
Switch per 100 pyrs per 100 pyrs per 100 pyrs per 100 pyrs per 100 pyrs per 100 pyrs Category1 (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Without recent 0.9 (0.6-1.4) 7.4 (6.3-8.6) 7.2 (5.7-9.2) 8.5 (6.8-10.7) 6.7 (5.3-8.4) 0.428 5.4 (4.9-5.9) virological failure With recent 7.2 (6.2-8.4) 9.9 (8.6-11.3) 7.4 (5.8-9.4) 4.2 (3.1-5.9) 3.5 (2.6-4.8) <0.001 7.1 (6.5-7.7) virological failure No recent 11.4 (10.1-12.9) 2.8 (2.2-3.6) 1.3 (0.8-2.3) 1.0 (0.5-2.0) 1.0 (0.5-1.8) <0.001 4.8 (4.3-5.3) viral record 19.5 (17.8-21.5) 20.1 (18.3-22.1) 16.0 (13.6-18.8) 13.8 (11.5-16.5) 11.2 (9.4-13.3) <0.001 17.3 (16.4-18.3) Overall
1. Virological failure defined as viral load >1000 copies/ml at most recent test within 180 days prior to and up to 30 days after switch. 2. log-rank test (excludes Mono/Duo era)
Page 60 of 127 Publication: Antiviral Therapy; Type: Original article of treatment initiation suggested that later era responses were marginally improved. However, there was relatively reduced recovery of early middle era counts (2000-03) compared to early era cART (1996-99). Baseline CD4 levels were lower for 2000-03 compared to 1996-99 but were similar to those for 2004-07 (where there was improved CD4 cell count response). This suggests that level of response by era may be influenced by factors other than initiation stage related to era for example, although investigation of these potential causal factors is beyond the scope of this study. The observed trends in baseline CD4 cell count and CD4 cell count response are consistent with findings from similarly resourced countries although baseline levels in this study are generally higher than elsewhere [19]. Generally, while there is indication of improved later era CD4 cell count response, we found that CD4 cell count measures alone are likely to be insufficient surrogates to differentiate HIV treatment outcomes. In particular, observed changes in CD4 cell count were not stratified by baseline CD4 cell count, baseline viral load, or rate of attainment of viral suppression or patient age which have been shown to significantly modify immune response [22,23]. In this study the improvement in immune response by era of initiation was less evident than that observed for virological suppression.
Changes in proportions with detectable viral load following treatment were clearly different by era despite generally high baseline levels. In particular, we observed increasing rate of on-treatment community return to virological control, and a clear trend towards lower asymptotic proportions with detectable viral loads for later eras. The average proportion with detectable viral load over the period from 2 years (when viral loads had reached asymptotic thresholds for all eras) to 4 years post ART fell from 69% (95% CI 66.7%-71.6%) in the mono/duo era and 29% (95% CI 27.8%-30.0%) in the early cART era to 4% (95% CI 2.8%-4.5%) in the late cART era. This is consistent with increasing effective ART potency from the suppression of HIV-RNA aspect although it should be noted that this study uses an intention-to-treat approach and findings are not stratified by potentially different levels of adherence per era which might also modify rates of suppression. Our results are qualitatively consistent with those of Torian et al who have shown improved rates of virological suppression by year of diagnosis for diagnoses between 2006 and 2009 [17] although that study did not look at time from cART initiation. We have found limited publication of comparable measures for other in-care cohorts.
We observed a clear, measureable decrease in all first line cART switches by later era of commencement (p<0.001). This was attributable to decreased rates of switching with associated virological failure in recent eras (p<0.001). Only moderate changes in rates of switching were observed for switches without associated virological failure (p=0.428) and there were very low rates of observed switching with missing recent viral load tests, particularly after 2000. This suggests improved durability of first line regimens and that later era patients who commence ART early will not be disadvantaged by frequent treatment switches and less treatment choices in the future. While we did not observe any difference in time to first switch without virological failure across cART eras, other studies have shown increased tolerability of many later era regimens [1,24,25]. It is possible that recent reductions in switching associated with increased tolerability might have been offset by other factors such as increases in relative rates of switching associated with the increased availability of
Page 61 of 127 Publication: Antiviral Therapy; Type: Original article regimens with improved potency, reduced toxicity and reduced pill burden, although this was not examined by this analysis [26].
Generally, while we found improving overall therapeutic utility at the on-treatment community level, there were not consistent trends in observed clinical baselines. Baseline CD4 cell count dropped to a minimum of 247cells/µl in the middle eras (2000-2007) before increasing again in the later era (from 2008). Similarly, baseline viral load did not uniformly fall over calendar time. This is consistent with contemporary Australian treatment guidelines which have shifted from aggressive early stage initiation immediately following the inception of cART in 1996, towards relatively delayed treatment initiation in asymptomatic patients with lower CD4 cell counts (<200 cells/µL) over the middle eras of this study (2000-2007) [11,27]. From 2008 onwards guidelines supported earlier treatment initiation (at 350 cells/µL in asymptomatic patients) [11]. Given the relatively advanced stage of immunological decline and virological failure observed at baseline across all periods in this study and improved treatment response, then this supports a focus of HIV care on treatment naive populations, including on improved early risk assessment, symptom awareness and diagnosis.
A limitation of this study is the generalizability of findings to the wider on-treatment community, as well as to the wider HIV-positive population. AHOD represents a significant subset (15- 20%) of the Australian on-treatment population [28,29], which has access to publicly funded ART. While AHOD data may be reasonably representative of the Australian HIV-positive population who are in care, key subsets of the broader HIV-positive population are not included in this analysis. In particular trends in off-treatment, as well as of in-care but irregular or chaotic treatment groups with likely significantly increased viral loads are difficult to assess.
The selection bias represented by cohort data in general has been described in detail by other papers [30–32]. In particular, likely higher rates of detectable viral load associated with out-of- care patients are not collected in AHOD. Generally, this study focuses on on-treatment community level trends and measures do not incorporate off-treatment subpopulations that are more difficult to reliably capture. Although our results do highlight off-treatment subpopulations as relatively important targets of intervention because of trends towards more highly effective treatment, extrapolation of our results to broader HIV-positive communities should be made with caution.
Secondly, in this study, observations from early treatment eras (<2000) are likely to reflect a healthy survivor bias to an extent. Response in surrogates of HIV outcome for these eras would otherwise be lower than those observed by this study because patients with advanced disease who died or were lost to clinical follow-up before 2000 were not included in AHOD enrolment. However, in this study virological response was found to be significantly improved in later eras regardless of survivor bias and similarly, CD4 response was marginally improved for 2004-13 cART compared to earlier eras (1996-99 and 2000-03).Overall, this study demonstrates continuing improvement in surrogates of HIV treatment outcomes by later era of treatment initiation.
A third limitation of our findings is that community level observation cannot necessarily be attributed to causal effects at the individual level [30]. In terms of immunological and virological
Page 62 of 127 Publication: Antiviral Therapy; Type: Original article response in this study, we describe treatment response in broad or net terms, and conversely, have examined a relatively uniform, stratified on-treatment community. However, results are likely also attributable to unaccounted for exogenous factors for any given patient, and or a causal structure of determinants not necessarily described in these analyses [30].
A strength of our study is that the population size, its extended follow up and further recruitment have facilitated estimation of generalizable indicators of treatment response. In particular, our measures resolve both treatment duration and calendar year related effects. Very few studies adjust population level indicators for temporal confounding in this way and therefore may have relatively limited generalizability.
Conclusions
Across the five time-periods examined, there have been incremental improvements for patients initiated on cART, as measured by overall response of surrogates of HIV outcome (viral load and CD4 count) across treatment eras, and also increased durability of first-line regimens. There were inconsistent changes in clinical stage of initiation which might be explained by evolving attitudes of patients and prescribers to the risk/benefit of cART initiation as the era of cART has unfurled. Population level indicators of treatment response should be adjusted for both calendar time and treatment duration to properly assess these outcomes.
Acknowledgements The Australian HIV Observational Database is funded as part of the Asia Pacific HIV Observational Database, a program of The Foundation for AIDS Research, amfAR, and is supported in part by a grant from the U.S. National Institutes of Health’s National Institute of Allergy and Infectious Diseases (NIAID) (Grant No. U01-AI069907) and by unconditional grants from Merck Sharp & Dohme; Gilead; Bristol-Myers Squibb; Boehringer Ingelheim; Roche; Pfizer; GlaxoSmithKline; Janssen-Cilag. The Kirby Institute is funded by the Australian Government Department of Health and Ageing, and is affiliated with the Faculty of Medicine, The University of New South Wales. The views expressed in this publication do not necessarily represent the position of the Australian Government.
Disclosure statement The authors declare no competing interests.
Additional file A list of the AHOD Collaborators can be found in the additional file.
References 1. Boyd MA. Improvements in antiretroviral therapy outcomes over calendar time. Current opinion in HIV and AIDS. 2009;4 [3]:194-9. Epub 2009/06/18. 2. Willig JH, Abroms S, Westfall AO, et al. Increased regimen durability in the era of once-daily fixed-dose combination antiretroviral therapy. AIDS 2008; 22:1951–1960 Epub 2008/09/12. 3. Parienti JJ, Bangsberg DR, Verdon R, Gardner EM. Better adherence with once-daily antiretroviral regimens: a meta-analysis. Clin Infect Dis 2009; 48:484–488 Epub 2009/01/15. 4. Llibre JM, Clotet B. Once-daily single-tablet regimens: a long and winding road to excellence in antiretroviral treatment. AIDS Rev 2012; 14:168–178 Epub 2012/07/27.
Page 63 of 127 Publication: Antiviral Therapy; Type: Original article
5. Bartlett JA, Fath MJ, Demasi R, et al. An updated systematic overview of triple combination therapy in antiretroviral-naive HIV-infected adults. AIDS 2006; 20:2051–2064 Epub 2006/10/21. 6. Robertson GR, Grant DM, Piquette-Miller M. Pharmacogenetics of pharmacoecology: which route to personalized medicine? Clin Pharmacol Ther 2009; 85:343–348 Epub 2009/03/20. 7. Phillips AN, Staszewski S, Weber R, et al. HIV viral load response to antiretroviral therapy according to the baseline CD4 cell count and viral load. JAMA 2001; 286:2560–2567 Epub 2001/11/28. 8. Bart PA, Rizzardi GP, Tambussi G, et al. Immunological and virological responses in HIV-1- infected adults at early stage of established infection treated with highly active antiretroviral therapy. AIDS 2000; 14:1887–1897 Epub 2000/09/21. 9. Frieden TR, Das-Douglas M, Kellerman SE, Henning KJ. Applying public health principles to the HIV epidemic. N Engl J Med 2005; 353:2397–2402 Epub 2005/12/02. 10. Marconi VC, Grandits GA, Weintrob AC, et al. Outcomes of highly active antiretroviral therapy in the context of universal access to healthcare: the U.S. Military HIV Natural History Study. AIDS Res Ther 2010; 7:14 Epub 2010/05/29. 11. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. In: Services DoHaH, editor. 2006. 12. Hammer SM, Saag MS, Schechter M, et al. Treatment for adult HIV infection: 2006 recommendations of the International AIDS Society-USA panel. JAMA 2006; 296:827–843 Epub 2006/08/15. 13. 2013 Antiretroviral Guidelines with Australian Commentary. Australasian Society for HIV Medicine; 2014 [updated 2/12/2013; cited 2014 11/03/2014]; Available from: http://arv.ashm.org.au/arv-guidelines/initiating-art-in-treatment-naive-patients. 14. Wright S, Petoumenos K, Boyd M, et al. Ageing and long-term CD4 cell count trends in HIV- positive patients with 5 years or more combination antiretroviral therapy experience. HIV Med 2013; 14:208–216 Epub 2012/10/06. 15. Egger M, May M, Chene G, et al. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet 2002; 360:119–129 Epub 2002/07/20. 16. Hughes R, Sterne J, Walsh J, et al. Long-term trends in CD4 cell counts and impact of viral failure in individuals starting antiretroviral therapy: UK Collaborative HIV Cohort (CHIC) study. HIV Med 2011; 12:583–593 Epub 2011/05/17. 17. Torian LV, Xia Q. Achievement and maintenance of viral suppression in persons newly diagnosed with HIV, New York City, 2006-2009: using population surveillance data to measure the treatment part of “test and treat”. J Acquir Immune Defic Syndr 2013; 63:379–386 Epub 2013/03/29. 18. Bartlett JA, Johnson J, Herrera G, et al. Long-term results of initial therapy with abacavir and Lamivudine combined with Efavirenz, Amprenavir/Ritonavir, or Stavudine. J Acquir Immune Defic Syndr 2006; 43:284–292 Epub 2006/09/13. 19. May MT, Sterne JA, Costagliola D, et al. HIV treatment response and prognosis in Europe and North America in the first decade of highly active antiretroviral therapy: a collaborative analysis. Lancet 2006; 368:451–458 Epub 2006/08/08. 20. The Australian HIV Observational Database. Rates of combination antiretroviral treatment change in Australia, 1997-2000. HIV Med 2002; 3:28–36 Epub 2002/06/13. 21. Fischl MA, Richman DD, Hansen N, et al. The safety and efficacy of zidovudine (AZT) in the treatment of subjects with mildly symptomatic human immunodeficiency virus type 1 (HIV) infection. A double-blind, placebo-controlled trial. The AIDS Clinical Trials Group. Ann Intern Med 1990; 112:727– 737 Epub 1990/05/15. 22. Staszewski S, Miller V, Sabin C, et al. Determinants of sustainable CD4 lymphocyte count increases in response to antiretroviral therapy. AIDS 1999; 13:951–956 Epub 1999/06/17.
Page 64 of 127 Publication: Antiviral Therapy; Type: Original article
23. Li TS, Tubiana R, Katlama C, Calvez V, Ait Mohand H, Autran B. Long-lasting recovery in CD4 T-cell function and viral-load reduction after highly active antiretroviral therapy in advanced HIV- 1 disease. Lancet 1998; 351:1682–1686 Epub 1998/09/12. 24. Ortiz R, Dejesus E, Khanlou H, et al. Efficacy and safety of once-daily darunavir/ritonavir versus lopinavir/ritonavir in treatment-naive HIV-1-infected patients at week 48. AIDS 2008; 22:1389– 1397 Epub 2008/07/11. 25. Gallant JE, DeJesus E, Arribas JR, et al. Tenofovir DF, emtricitabine, and efavirenz vs. zidovudine, lamivudine, and efavirenz for HIV. N Engl J Med 2006; 354:251–260. 26. Davidson I, Beardsell H, Smith B, et al. The frequency and reasons for antiretroviral switching with specific antiretroviral associations: the SWITCH study. Antiviral Res 2010; 86:227–229 Epub 2010/03/10. 27. Hammer SM, Saag MS, Schechter M, et al. Treatment for adult HIV infection: 2006 recommendations of the International AIDS Society--USA panel. Top HIV Med 2006; 14:827–843 Epub 2006/10/05. 28. Australian HIV Observational Database Annual Report. In: Institute TK, editor.: The University of New South Wales; 2013. 29. Falster K, Gelgor L, Shaik A, et al. Trends in antiretroviral treatment use and treatment response in three Australian states in the first decade of combination antiretroviral treatment. Sex Health 2008; 5:141–154 Epub 2008/07/01. 30. Miller WC, Powers KA, Smith MK, Cohen MS. Community viral load as a measure for assessment of HIV treatment as prevention. Lancet Infect Dis 2013; 13:459–464 Epub 2013/03/30. 31. Xia Q, Wiewel EW, Torian LV. Revisiting the methodology of measuring HIV community viral load. J Acquir Immune Defic Syndr 2013; 63:e82–e84 Epub 2013/05/15. 32. Law MG, Woolley I, Templeton DJ, et al. Trends in detectable viral load by calendar year in the Australian HIV observational database. J Int AIDS Soc 2011; 14:10 Epub 2011/02/25.
Page 65 of 127 AHOD Collaborators Australian HIV Observational Database contributors Asterisks indicate steering committee members in 2013. New South Wales: D Ellis, General Medical Practice, Coffs Harbour; M Bloch, S Agrawal, T Vincent, Holdsworth House Medical Practice, Darlinghurst; D Allen, JL Little, Holden Street Clinic, Gosford; D Smith, C Mincham, Lismore Sexual Health & AIDS Services, Lismore; D Baker*, V Ieroklis, East Sydney Doctors, Surry Hills; DJ Templeton*, CC O’Connor, S Phan, RPA Sexual Health Clinic, Camperdown; E Jackson, K McCallum, Blue Mountains Sexual Health and HIV Clinic, Katoomba; M Grotowski, S Taylor, Tamworth Sexual Health Service, Tamworth; D Cooper, A Carr, F Lee, K Hesse, K Sinn, R Norris, St Vincent’s Hospital, Darlinghurst; R Finlayson, I Prone, A Patel, Taylor Square Private Clinic, Darlinghurst; R Varma, J Shakeshaft, Nepean Sexual Health and HIV Clinic, Penrith; K Brown, C McGrath, V McGrath, S Halligan, Illawarra Sexual Health Service, Warrawong; L Wray, P Read, H Lu, Sydney Sexual Health Centre, Sydney; D Couldwell, Parramatta Sexual Health Clinic; DE Smith*, V Furner, Albion Street Centre; Clinic 16 – Royal North Shore Hospital, S Fernando; Dubbo Sexual Health Centre, Dubbo; Holdsworth House Medical Practice, Byron Bay, J Chuah*; J Watson*, National Association of People living with HIV/AIDS; C Lawrence*, National Aboriginal Community Controlled Health Organisation; B Mulhall*, Department of Public Health and Community Medicine, University of Sydney; M Law*, K Petoumenos*, S Wright*, H McManus*, C Bendall*, M Boyd*, The Kirby Institute, University of NSW. Northern Territory: N Ryder, R Payne, Communicable Disease Centre, Royal Darwin Hospital, Darwin. Queensland: D Russell, S Doyle-Adams, Cairns Sexual Health Service, Cairns; D Sowden, K Taing, K McGill, Clinic 87, Sunshine Coast-Wide Bay Health Service District, Nambour; D Orth, D Youds, Gladstone Road Medical Centre, Highgate Hill; M Kelly, A Gibson, H Magon, Brisbane Sexual Health and HIV Service, Brisbane; B Dickson*, CaraData. South Australia: W Donohue,O’Brien Street General Practice, Adelaide. Victoria: R Moore, S Edwards, R Liddle, P Locke, Northside Clinic, North Fitzroy; NJ Roth*, H Lau, Prahran Market Clinic, South Yarra; T Read, J Silvers*, W Zeng, Melbourne Sexual Health Centre, Melbourne; J Hoy*, K Watson*, M Bryant, S Price, The Alfred Hospital, Melbourne; I Woolley, M Giles*, T Korman, J Williams*, Monash Medical Centre, Clayton. Western Australia: D Nolan, J Robinson, Department of Clinical Immunology, Royal Perth Hospital, Perth. Coding of Death Form (CoDe) reviewers: D Sowden, J Hoy, L Wray, I Woolley, K Morwood, N Roth, K Choong, CC O'Connor, MA Boyd.
Page 66 of 127 Chapter 4: Long-term survival in HIV-positive patients with up to 15 years of antiretroviral therapy
McManus H, O'Connor CC, Boyd M, Broom J, Russell D, Watson K, et al. Long-term survival in HIV positive patients with up to 15 Years of antiretroviral therapy. PLoS One. 2012;7(11):e48839.
Page 67 of 127 Long-Term Survival in HIV Positive Patients with up to 15 Years of Antiretroviral Therapy
Hamish McManus1*, Catherine C. O’Connor2,3,4, Mark Boyd1, Jennifer Broom5, Darren Russell6, Kerrie Watson7, Norman Roth8, Phillip J. Read9, Kathy Petoumenos1, Matthew G. Law1, on Behalf of the Australian HIV Observational Database" 1 The Kirby Institute, UNSW, Sydney, New South Wales, Australia, 2 RPA Sexual Health, Royal Prince Alfred Hospital, Sydney, New South Wales, Australia, 3 South Western Clinical School, UNSW, Sydney, New South Wales, Australia, 4 Central Clinical School, Sydney University, Sydney, New South Wales, Australia, 5 Department of Infectious Diseases, Nambour General Hospital, Nambour, Queensland, Australia, 6 Cairns Sexual Health Service, Cairns, Queensland, Australia, 7 The Alfred Hospital, Melbourne, Victoria, Australia, 8 Prahran Market Clinic, Prahran, Victoria, Australia, 9 Sydney Sexual Health Centre, Sydney, New South Wales, Australia
Abstract
Background: Life expectancy has increased for newly diagnosed HIV patients since the inception of combination antiretroviral treatment (cART), but there remains a need to better understand the characteristics of long-term survival in HIV-positive patients. We examined long-term survival in HIV-positive patients receiving cART in the Australian HIV Observational Database (AHOD), to describe changes in mortality compared to the general population and to develop longer-term survival models.
Methods: Data were examined from 2,675 HIV-positive participants in AHOD who started cART. Standardised mortality ratios (SMR) were calculated by age, sex and calendar year across prognostic characteristics using Australian Bureau of Statistics national data as reference. SMRs were examined by years of duration of cART by CD4 and similarly by viral load. Survival was analysed using Cox-proportional hazards and parametric survival models.
Results: The overall SMR for all-cause mortality was 3.5 (95% CI: 3.0–4.0). SMRs by CD4 count were 8.6 (95% CI: 7.2–10.2) for CD4,350 cells/ml; 2.1 (95% CI: 1.5–2.9) for CD4 = 350–499 cells/ml; and 1.5 (95% CI: 1.1–2.0) for CD4$500 cells/ml. SMRs for patients with CD4 counts ,350 cells/mL were much higher than for patients with higher CD4 counts across all durations of cART. SMRs for patients with viral loads greater than 400 copies/ml were much higher across all durations of cART. Multivariate models demonstrated improved survival associated with increased recent CD4, reduced recent viral load, younger patients, absence of HBVsAg-positive ever, year of HIV diagnosis and incidence of ADI. Parametric models showed a fairly constant mortality risk by year of cART up to 15 years of treatment.
Conclusion: Observed mortality remained fairly constant by duration of cART and was modelled accurately by accepted prognostic factors. These rates did not vary much by duration of treatment. Changes in mortality with age were similar to those in the Australian general population.
Citation: McManus H, O’Connor CC, Boyd M, Broom J, Russell D, et al. (2012) Long-Term Survival in HIV Positive Patients with up to 15 Years of Antiretroviral Therapy. PLoS ONE 7(11): e48839. doi:10.1371/journal.pone.0048839 Editor: Michael Alan Polis, National Institute of Allergy and Infectious Diseases, United States of America Received April 3, 2012; Accepted October 1, 2012; Published November 7, 2012 Copyright: ß 2012 McManus et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: The Australian HIV Observational Database is funded as part of the Asia Pacific HIV Observational Database, a program of The Foundation for AIDS Research, amfAR, and is supported in part by a grant from the U.S. National Institutes of Health’s National Institute of Allergy and Infectious Diseases (NIAID) (Grant No. U01-AI069907) and by unconditional grants from Merck Sharp & Dohme; Gilead; Bristol-Myers Squibb; Boehringer Ingelheim; Roche; Pfizer; GlaxoSmithKline; Janssen-Cilag. The Kirby Institute is funded by The Australian Government Department of Health and Ageing, and is affiliated with the Faculty of Medicine, The University of New South Wales. The views expressed in this publication do not necessarily represent the position of the Australian Government. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: Matthew Law has received research grants, consultancy and/or travel grants from: Boehringer Ingelheim; Bristol-Myers Squibb; Gilead; GlaxoSmithKline; Janssen-Cilag; Johnson & Johnson; Merck Sharp & Dohme; Pfizer; Roche; and CSL Ltd. Mark Boyd has been a board member for Merck and BMS received research grants, consultancy and/or travel grants from: Merck Sharp & Dohme; Abbott; Gilead; Janssen-Cilag; and from Boehringer Ingelheim. Darren Russell has received research grants, consultancy and/or travel grants from: Janssen-Cilag; Merck Sharp & Dohme; Abbott; and Gilead. Norman Roth has received research grants, consultancy and/or travel grants from Boehringer Ingelheim; ViiV; Merck Sharp & Dohme; and Janssen-Cilag. This does not alter the authors’ adherence to all the PLOS ONE policies on sharing data and materials. * E-mail: [email protected] " Membership of the Australian HIV Observational Database is provided in the Acknowledgments.
Introduction manageable chronic condition. The nature of long-term survival with HIV is increasingly being revealed through the study of Mortality has decreased for newly diagnosed HIV-positive populations of patients with extended durations of exposure [3,4]. patients since the inception of combination antiretroviral therapy Treatment can be complex with chronic pathologies associated (cART) [1,2] and HIV infection can now be characterised as a with immunodeficiency, chronic viral infection and sociobeha-
PLOS ONE | www.plosone.org 1 November 2012 | Volume 7 | Issue 11 | e48839 Page 68 of 127 Long-Term Survival in HIV Positive Patients vioural factors. The accurate description of survival in HIV- Briefly, AHOD uses standardised cause of death (CoDe) forms positive populations today is therefore increasingly important in based on versions of the Data Collection on Adverse Events of HIV management. Anti-HIV Drugs (D:A:D) cohort study [16] (www.cphiv.dk/ Long-term survival in HIV-positive populations with access to CoDe). These are reviewed by an independent HIV specialist effective treatment appears to be approaching that of the general clinician at the Kirby Institute who determines the primary and population [5]. Studies have shown declining rates of AIDS related secondary causes of death or recommends additional information death compared to non-AIDS related death since the introduction be provided by the study site as required. of cART [6,7] and describe a need for increasing focus on chronic The covariates considered were: sex; age; mode of HIV disease management and health promotion [8,9]. All-cause exposure (men who have sex with men (MSM), heterosexual, mortality in patients who have achieved high CD4 cell count injecting drug user (IDU), other/unknown); time updated instance levels approaches that of the general population over time [10], of first AIDS defining illness (ADI); mono or dual antiretroviral although there is strong evidence that CD4 cell counts trend treatment prior to first cART; HCV antibody (no/not tested, ever towards different plateaus according to pre cART levels [11,12]. positive); HBV surface antigen (HBVsAg) (no/not tested, ever This suggests that long-term mortality can be associated with early positive); time updated CD4 cell count (closest prior or up to 30 uncontrolled viral replication and immune activation, and has led days after 2,350, 350–499, $500 cells/ml); time updated viral to contention about threshold levels of CD4 cell counts for load (closest prior or up to 30 days after 2#400, .400 copies/ treatment initiation. ml); year of first cART (#1995, 1996–99, 2000–2003, $2004); Further, there is strong evidence associating immunologic and time updated number of regimens received (a new regimen resilience with age, and age at cART initiation has been associated was defined as the commencement of 1 or more new antiretroviral with the rate and extent of immunologic recovery [13]. The effects drugs and for more than 14 days). of ageing on survival therefore need to be considered in addition to just the effects of increased duration of illness. However, studies of Statistical Analysis overall life expectancy in HIV-positive populations are often Follow up was calculated from the start of cART and censored limited by insufficient data in older age groups [4] where rapid at the earlier of death, lost to follow up or 31 March 2011 (cohort increases in general population mortality are observed. There censoring date). Lost to follow up date was defined as the most remains a need to better understand long-term survival in ageing recent clinic visit if no clinic visit had been recorded in the 12 HIV-positive patients after prolonged cART. months prior to 31 March 2011. We used an intent-to-continue The primary objective of this analysis is to measure all-cause treatment approach and ignored any changes to treatment after mortality in adult HIV-positive patients receiving cART in baseline including periods off treatment. This approach avoids Australia. Specifically we want to compare mortality rates in problems of modelling causal effects of patients stopping cART. these patients with those of the general population, over the long- Standardised mortality ratios (SMR) across prognostic charac- term, and examine how these rates are affected by duration of teristics were calculated referent to the Australian Bureau of treatment when adjusted for prognostic factors. A secondary Statistics (ABS) general population mortality rates by age group, objective of this analysis is to examine the effects of ageing on gender and calendar year [17,18]. SMRs were also calculated by mortality in HIV-positive populations relative to the general years of cART for given levels of time updated CD4 cell count and population. for given levels of time updated viral load. Cox proportional hazards models were initially developed to Methods identify important survival covariates. Initially univariate models were developed and all significant univariate predictors were Study population considered for inclusion in multivariate models. Final multivariate The Australian HIV Observational Database (AHOD) is an models were developed using a backwards stepwise approach to observational clinical cohort study of patients with HIV infection reduce to a parsimonious set of statistically significant (2p,0.05) seen at 27 clinical sites throughout Australia. AHOD utilises covariates. A sensitivity analysis was conducted using Cox methodology which has been described in detail elsewhere [14]. proportional hazards models of survival in patients who had Briefly, data are transferred electronically to the Kirby Institute at commenced cART after 1 January 1999. the University of New South Wales every 6 months. Core data Parametric survival-time Weibull models using clustered vari- variables include: sex; date of birth; date of most recent visit; HIV ance estimators were developed based on these covariates, and exposure; hepatitis B virus (HBV) surface antigen status; hepatitis used to investigate survival trends by age. Probability of 10 year C virus (HCV) antibody status; CD4 and CD8 counts; HIV viral survival was calculated for a range of representative prognostic load; antiretroviral treatment data; AIDS-defining illnesses; and covariate values. Results were compared with 10 year survival date and cause of death. Prospective data collection commenced in rates for general population Australian males 2007–2009 using 1999, with retrospective data provided where available. ABS data [19]. Ethics approval for the study was granted by the University of Data were analysed using Stata version 12 (Stata Corporation, New South Wales Human Research Ethics Committee, and all College Station, Texas, USA). other relevant institutional review boards. Written informed consent was obtained from participating individuals. All study Results procedures were developed in accordance with the revised 1975 Helsinki Declaration. Patient characteristics This analysis included all patients who had been recruited to Of 3,173 people in AHOD, 2,675 had commenced cART and AHOD prior to 31 March 2011 and who had at least 1 subsequent had at least 1 subsequent clinical visit or result recorded post clinical visit after the date of first recorded cART (defined as the therapy. During 15,936 patient years of follow-up 206 deaths were use of 3 or more antiretrovirals from more than 1 class). observed. The rate of loss to follow-up of patients was 40.4 per The study endpoint was mortality. The method of collection of 1000 person years of follow-up (95% CI: 37.3–43.6). The number mortality in AHOD has been described in detail elsewhere [15]. of patients at first cART by selected characteristics is shown in
PLOS ONE | www.plosone.org 2 November 2012 | Volume 7 | Issue 11 | e48839 Page 69 of 127 Long-Term Survival in HIV Positive Patients
Table 1. At first cART the population was predominantly male Table 1. Patient characteristics at cART commencement. (94%), aged from 30–49 (70%) and the main exposure category was MSM (75%). The median CD4 cell count at cART commencement was 282 cells/mL (interquartile range (IQR): N = 2,675 (%) 151–427) and viral load was 58,942 copies/ml (IQR: 1,000– 199,263). The majority of patients (60%) had no prior treatment at Sex cART commencement. Female 149 (6) Male 2,526 (94) Mortality rates Age at first cART (years) Observed mortality by patient characteristics is shown in Median (IQR) 42 (36–49) Table 2. The overall mortality rate was 12.9 deaths/1000 patient ,30 364 (14) years (95% CI: 11.3–14.8) and the overall SMR was 3.5 (95% CI: 3.0–4.0). An increase was observed in mortality rates with 30–39 1074 (40) increasing age from 10.8 deaths/1000 patient years (95% CI: 40–49 791 (30) 4.0–28.7) in patients aged ,30, to 21.6 deaths/1000 patient years 50–59 354 (13) (95% CI: 15.6–30.0) in patients aged $60. A decrease towards 60–69 81 (3) general population level was observed in SMRs by increasing age 70–79 11 (0) from patients aged ,30 (SMR 12.4; 95% CI 4.7–33.0) to patients aged $60 (SMR 1.4; 95% CI 1.0–2.0). An increased SMR was Mode of HIV exposure observed for patients with mode of HIV exposure through IDU MSM 2,001 (75) (SMR 11.3; 95% CI 7.3–17.5). SMRs were increased for patients IDU 154 (6) with recorded HCV ever (SMR 7.2; 95% CI 5.1–10.0) compared HET 248 (9) to those without HCV ever (SMR 3.1; 95% CI 2.7–3.7), and Other 272 (10) similarly for patients with recorded HBVsAg ever (SMR 10.6; 95% CI 7.0–15.9) compared to those without HBVsAg ever (SMR ADI 3.2; 95% CI 2.8–3.7). No 2,271 (85) A large difference was observed in mortality rates for patients Yes 404 (15) with CD4 cell counts of ,350 cells/ml (Mortality rate (MR) 32.0; HCV (ever) 95% CI 27.0–38.0) compared to patients with CD4 cell counts of No/not tested 2390 (89) 350–499 cells/ml (MR 7.7; 95% CI 5.5–11.0); and CD4 cell counts Yes 285 (11) of $500 cells/ml (MR 5.6; 95% CI 4.1–7.5). Similarly the SMR for patients with CD4 cell counts of ,350 cells/ml (SMR 8.6; 95% HBVsAg (ever) CI 7.2–10.2) was much higher than for patients with CD4 cell No/not tested 2549 (95) counts of 350–499 cells/ml (SMR 2.1; 95% CI 1.5–2.9); and for Yes 126 (5) patients with CD4 cell counts of $500 cells/ml (SMR 1.5; 95% CI CD4 (cells/mL) 1.1–2.0) which was slightly higher than population level mortality. Median (IQR) 282 (151–427) An increased SMR was observed for patients with viral loads of .400 copies/ml (SMR 9.0; 95% CI 7.5–10.9) compared to ,350 1,316 (49) patients with viral loads of #400 copies/ml (SMR 2.1; 95% CI 351–499 408 (15) 1.7–2.5). $500 379 (14) SMRs by time updated CD4 cell count category by years that a Missing 572 (21) patient has been receiving cART are shown in Figure 1. SMRs for Viral Load (copies/ml) patients with CD4 cell counts ,350 cells/mL were much higher than for patients with higher CD4 cell counts across all durations Median (IQR) 58,942 (10,000–199,263) of cART. SMRs for patients with CD4 cell counts of 350–499 #400 166 (6) cells/mL were in the range 1.2 to 2.4. SMRs for patients with CD4 .400 1,722 (64) cell counts of $500 cells/mL were consistently closer to population Missing 787 (29) level in the range 1.3 to 1.9. Year of first cART SMRs by time updated viral load category by years that a #1995 138 (5) patient has been receiving cART are shown in Figure 2. SMRs for patients with viral loads of #400 copies/ml were relatively 1996–1999 1593 (60) constant by years of cART (in the range 1.8 to 2.4). SMRs for 2000–2003 390 (15) patients with viral loads of .400 copies/ml were much higher (in $2004 554 (21) the range 7.6 to 12.4) across all durations of cART and showed an Regimen number at first cART upward trend after 3 years of cART. 1st 1,600 (60) nd rd Prognostic model 2 or 3 824 (31) th Univariate and multivariate Weibull model predictors of 4 or more 251 (9) survival are shown in Table 3. Univariate predictors associated doi:10.1371/journal.pone.0048839.t001 with survival were time updated age (p = 0.003); mode of HIV exposure through IDU (p = 0.011); time updated instance of first number of regimens received (p,0.001). Predictors retained in the ADI (p,0.001); HCV ever (p = 0.019); HBVsAg ever (p,0.001); multivariate model were time updated age (p,0.001); mode of time updated CD4 cell count (p,0.001); time updated viral load HIV exposure through IDU (p = 0.041); time updated instance of (p,0.001); treatment prior to cART (p,0.001); and time updated
PLOS ONE | www.plosone.org 3 November 2012 | Volume 7 | Issue 11 | e48839 Page 70 of 127 Long-Term Survival in HIV Positive Patients
Table 2. Mortality by patient characteristics.
Deaths Expected PYs (000’s) Mortality rate(95% CI) SMR (95% CI)
All mortality 206 59.4 15.9 12.9 (11.3, 14.8) 3.5 (3.0, 4.0) Sex Female 8 1.3 0.9 9.3 (4.6, 18.5) 6.0 (3.0, 11.9) Male 198 58 15.1 13.1 (11.4, 15.1) 3.4 (3.0, 3.9) Age 0–29 4 0.3 0.4 10.8 (4.0, 28.7) 12.4 (4.7, 33.0) 30–39 38 4.4 3.8 9.9 (7.2, 13.6) 8.6 (6.2, 11.8) 40–49 73 12.4 6.2 11.8 (9.4, 14.8) 5.9 (4.7, 7.4) 50–59 55 17.2 3.9 14.2 (10.9, 18.5) 3.2 (2.5, 4.2) 60–90 36 25.1 1.7 21.6 (15.6, 30.0) 1.4 (1.0, 2.0) Mode of HIV exposure MSM 153 46.4 12.2 12.6 (10.7, 14.7) 3.3 (2.8, 3.9) IDU 20 1.8 0.9 22.5 (14.5, 34.9) 11.3 (7.3, 17.5) HET 12 4.9 1.4 8.8 (5.0, 15.5) 2.5 (1.4, 4.4) OTHER 21 6.4 1.5 13.8 (9.0, 21.1) 3.3 (2.1, 5.0) ADI prior to cART No 155 48.5 13.3 11.6 (9.9, 13.6) 3.2 (2.7, 3.7) Yes 51 10.9 2.6 19.5 (14.8, 25.6) 4.7 (3.6, 6.2) HCV (ever) No/no tested 172 54.6 14.2 12.2 (10.5, 14.1) 3.1 (2.7, 3.7) Yes 34 4.8 1.8 19.1 (13.6, 26.7) 7.2 (5.1, 10.0) HBVsAg (ever) No/not tested 183 57.2 15.2 12.1 (10.4, 13.9) 3.2 (2.8, 3.7) Yes 23 2.2 0.8 30.0 (20.0, 45.2) 10.6 (7.0, 15.9) CD4 (cells/ml)1 ,350 131 15.3 4.1 32.0 (27.0, 38.0) 8.6 (7.2, 10.2) 350–499 32 15.6 4.1 7.7 (5.5, 11.0) 2.1 (1.5, 2.9) $500 43 28.5 7.7 5.6 (4.1, 7.5) 1.5 (1.1, 2.0) Viral Load (copies/ml)1 #400 99 47.5 11.8 8.4 (6.9, 10.2) 2.1 (1.7, 2.5) .400 107 11.9 4.2 25.8 (21.3, 31.1) 9.0 (7.5, 10.9) Treatment prior to cART No 81 29.4 8.4 9.7 (7.8, 12) 2.8 (2.2, 3.4) Yes 125 30 7.6 16.5 (13.9, 19.7) 4.2 (3.5, 5.0) Year of cART commencement #1995 17 4.3 1.1 16.1 (10.0, 25.9) 3.9 (2.4, 6.3) 1996–1999 159 44.7 11.7 13.6 (11.6, 15.9) 3.6 (3.0, 4.2) 2000–2003 24 6.8 2.1 11.7 (7.8, 17.5) 3.5 (2.4, 5.3) $2004 6 3.5 1.1 5.3 (2.4, 11.8) 1.7 (0.8, 3.8) Regimen number1 1st 12 6.0 2.1 5.7 (3.2, 10.0) 2.0 (1.1, 3.5) 2nd or 3rd 29 16.6 4.7 6.1 (4.3, 8.8) 1.7 (1.2, 2.5) 4th or more 165 36.8 9.1 18.1 (15.6, 21.1) 4.5 (3.9, 5.2)
1Time updated. doi:10.1371/journal.pone.0048839.t002 first ADI (p,0.001); HBVsAg ever(p,0.001); time updated CD4 Weibull model 10 year survival probabilities for specified ages cell count (p,0.001); time updated viral load (p,0.001); and time by CD4 cell count are shown in Figure 3. For illustrative purposes updated number of regimens received (p,0.001). we present 10 year survival probabilities of patients on their 4th or
PLOS ONE | www.plosone.org 4 November 2012 | Volume 7 | Issue 11 | e48839 Page 71 of 127 Long-Term Survival in HIV Positive Patients
Figure 2. SMRs and 95% confidence intervals by years of cART and time updated HIV viral load. Blue/square markers represent patients with viral load #400 copies/ml. Green/round markers represent patients with HIV viral load .400 copies/ml. Grey/horizontal Figure 1. SMRs and 95% confidence intervals by years of cART dashed line represents SMR of 1. and time updated CD4 cell count. Blue/square markers represent doi:10.1371/journal.pone.0048839.g002 patients with CD4,350 cells/ml. Green/round markers represent patients with CD4 from 350 to 499 cells/ml. Red/triangular markers general population level mortality, but mortality rates were close to represent patients with CD4$500 cells/ml. Grey/horizontal dashed line general population levels in patients with high CD4 cell counts represents SMR of 1. (especially above 500 cells/mL). There was no evidence of doi:10.1371/journal.pone.0048839.g001 increased mortality by duration of cART when controlling for CD4 cell count level. There was no observed evidence of changes greater regimen with no HBVsAg ever and mode of exposure in mortality associated with age relative to the general population other than IDU because these were the most representative after controlling for time updated incidence of prior AIDS, mode observed states based on duration of analysis time. Scenarios are of HIV exposure, time updated CD4 cell count, time updated viral presented variously controlling for incidence of prior ADI and load and time updated regimen number. viral load. Survival probabilities for patients with CD4 cell counts Prognostic indicators in our study are consistent with other of $350 cells/mL did not differ significantly from those of the studies [14,20,21]. In particular, our results confirm the impor- general population for all age groups when controlling for other tance of immunological reconstitution as a biomarker for long- model covariates. Survival probabilities for patients maintaining term survival in HIV-positive populations and show the similarity CD4 cell counts of $500 cells/mL and controlling for other model between mortality in patients with CD4 cell counts above 350 covariates were similar to general population levels for all age cells/mL (350–499 cells/mL and $500cells/mL) and large increases groups: the 10 year survival probability for a 20 year old HIV- in mortality below this level. Our results are very similar to those positive person controlling for other model covariates was 0.98 published recently in a study by the COHERE collaboration: (95% CI 0.96–1.00) compared to 0.99 for 20 year olds in the male SMR was 1.8 for CD4 cell count of 350–499 cells/mL and 1.5 for general population [19]; and for a 60 year old was 0.91 (95% CI CD4 cell count of $500 cells/mL) [10]. We found that, at 0.85–0.98) compared to 0.89 for 60 year olds in the male general increased CD4 cell count levels, observed survival approaches that population [19]. The decrease in survival probabilities associated of the general population, although average rates at these levels with increase in age was less for patients with CD4 cell counts remained slightly higher than those of the general population. This $350 cells/mL compared to patients with CD4 cell counts of reflects the findings of recent studies [10,22] and, given the ,350 cells/mL. Prior ADI, increased viral load, and mode of HIV association between CD4 cell count levels at commencement of exposure through IDU were also associated with significantly cART and capacity for immunological recovery, reinforces the reduced 10 year survival probabilities compared to the general notion that CD4 cell count of 350 cells/ml is an important population. Survival probabilities given the presence of 1 or more treatment threshold. Our results support the importance of of these conditions decreased proportionally more with increasing ongoing immunological reconstitution as a treatment priority. age and relative to the general population compared to survival There was no observed association between duration of without these conditions. treatment and survival when adjusted for CD4 cell count level. In a sensitivity analysis, 10 year survival probabilities for a At CD4 cell counts less than: 350 cells/mL there was a mild multivariate model of survival in patients who had commenced decrease in SMRs by duration of cART but rates remained much cART after 1 January 1999 were similar to those developed in the higher than population level; while above this level, average rates primary analysis (results not shown). remained fairly constant, and just above population level. Although we observed a decrease in SMRs associated with recent Discussion cART commencement and year of cART commencement was a significant univariate predictor of survival, it was not a significant In this analysis from AHOD, overall mortality in adult HIV- covariate in multivariate models. We also observed slightly positive patients receiving cART was found to be higher than decreased SMRs for patients with less than 3 years of cART
PLOS ONE | www.plosone.org 5 November 2012 | Volume 7 | Issue 11 | e48839 Page 72 of 127 Long-Term Survival in HIV Positive Patients
Table 3. Weibull model predictors of survival for patients commencing cART1.
Univariate Multivariate2
Hazard (95% CI) p r3 Hazard (95% CI) p r3
Sex Female ref Male 1.41 (0.69, 2.87) 0.343 Age4 1.02 (1.01, 1.04) 0.003 1.04 (1.02, 1.05) ,0.001 Mode of HIV exposure Not IDU ref ref IDU 1.83 (1.15, 2.90) 0.011 1.63 (1.02, 2.60) 0.041 ADI4 No ref ref Yes 2.44 (1.84, 3.23) ,0.001 1.62 (1.22, 2.14) ,0.001 HCV (ever) No/not tested ref Yes 1.56 (1.07, 2.25) 0.019 HBVsAg (ever) No/not tested ref ref Yes 2.49 (1.61, 3.86) ,0.001 2.31 (1.51, 3.53) ,0.001 CD4(cells/mL)4 ,350 ref ,0.001 ref ,0.001 350–499 0.24 (0.16, 0.36) ,0.001 0.33 (0.23, 0.49) ,0.001 $500 0.17 (0.12, 0.24) ,0.001 0.29 (0.20, 0.40) ,0.001 Viral Load(count/ml)4 #400 ref ref .400 3.17 (2.41, 4.18) ,0.001 2.57 (1.94, 3.41) ,0.001 Treatment prior to cART No ref Yes 1.67 (1.25, 2.23) ,0.001 Year of cART commencement #1995 ref 0.173 1996–1999 0.85 (0.50, 1.42) 0.531 2000–2003 0.73 (0.37, 1.45) 0.371 $2004 0.33 (0.12, 0.91) 0.032 Regimen number4 1st ref ,0.001 ref ,0.001 2nd or 3rd 1.26 (0.63, 2.54) 0.515 1.10 (0.54, 2.21) 0.800 4th or more 4.19 (2.19, 8.02) ,0.001 2.66 (1.38, 5.13) 0.004
12675 patients, 15936 years of follow up, 206 deaths. 2Covariates selected by backwards stepwise selection from significant univariate predictors. 3Wald test for homogeneity for categorical covariates. 4Time updated. doi:10.1371/journal.pone.0048839.t003 experience compared to those with 3–6 years cART experience cell count levels, and included earlier periods of follow-up than when stratified by CD4 cell count level, and these rates are those used in this study. weighted heavily by more recent cART starters. However these This analysis provides a robust assessment of the association differences were not significant. These findings can be contrasted between age and survival for this population. Decreasing SMRs with a recent study by the Antiretroviral Therapy Cohort were observed with age, which adjust for different distributions of Collaboration of decreasing mortality by period of cART initiation age, gender and temporal structure of AHOD compared to the (1996–99, 2000–02, 2003–05) [4]. In that study duration of general Australian population. However these ratios do not adjust treatment was associated with decreasing SMRs for each period of for additional covariate factors, such as changes in CD4 cell count cART initiation, but these differences were not adjusted for CD4 and viral load, because of limited sample size. Instead, survival
PLOS ONE | www.plosone.org 6 November 2012 | Volume 7 | Issue 11 | e48839 Page 73 of 127 Long-Term Survival in HIV Positive Patients
mortality compared to general population rates associated with age. Some caution should be used when extrapolating such trends given the relatively limited history of HIV infection. To compare, type 1 childhood-onset diabetes patients have relatively low and stable mortality rates at earlier durations (up to 15 years) but increased rates at extended durations (up to 45 years) [23]. Concomitant illnesses are well documented in diabetic populations and SMRs for ischaemic heart disease and renal disease in particular, are much higher than those of the general population [24] and there are poorer prognoses compared to similar events in the general population [25]. In contrast, our findings are based on shorter duration of illness and it is possible that significant changes in morbidity and mortality in HIV–positive populations compared to the general population will emerge with extended follow-up. The findings of this study apply generally to the AHOD population and are subject to the real life prospect that patients do not always adhere to therapy. In AHOD, while time off treatment is well reported, there are limited measures of the extent to which this is associated with adherence. However, the use of intent-to- continue treatment principles in this report reflects the applied utility of cART and results can be reasonably simply interpreted as describing the mortality risk in all patients who start cART. Generally AHOD adherence is expected to be high for several reasons including low rates of loss-to-follow-up, cohort effect, and resourcing. There are some limitations to our analyses. First, our data had insufficient observations to robustly assess survival at ages over 70. Some studies have accounted for insufficient data at these ages by applying assumed relative rates of mortality [26]. We believe there is high risk of error in doing so when comparing survival relative to the general population because at these ages, general mortality rates increase rapidly and can magnify errors in model assump- tions. Further, the projection of whole of life experience to recent infections is likely to omit the effects of future developments in HIV specific and general treatments. Instead we estimated 10 year survival probabilities by age and relative to the general population. We believe this is a much more robust comparison as it does not require extensive extrapolation outside the range of the analysis dataset. This comparison found that estimated ten-year survival probabilities are close to population rates when controlling for other prognostic factors (age, mode of HIV exposure through IDU, ADI, HBVsAg, CD4 cell count, viral load and regimen number), and survival rates do not decrease relative to those of the general population at increased ages (at least up to 70 years of age). Figure 3. 10 year survival probabilities and 95% confidence However, increased data on aged patients is required to more intervals by age and time updated CD4 count. Blue/square markers represent patients with CD4 less than 350 cells/ml. Green/round broadly describe survival experience in AHOD. markers represent patients with CD4 from 350 to 499 cells/ml. Red/ Second, the generalisation of these findings may be limited by triangular markers represent patients with CD4$500 cells/ml. Grey/ likely survivor bias in our analysis associated especially with horizontal dashed line represents Australian males 2008–09. These plots patients who commenced cART prior to 1999. AHOD com- apply to patients on 4th or greater regimen, no prior HBVsAg, non-IDU menced in 1999 nearly 5 years after the inception of cART and mode of exposure and specified existing viral load and incidence of ADI. Plot (A) shows 10 year survival probabilities for patients with viral load does not include mortality from before this time in patients (who #400 copies/ml and with no prior ADI. Plot (B) shows 10 year survival were also receiving suboptimal treatment). However, a sensitivity probabilities for patients with viral load #400 copies/ml and with prior analysis of patients who had commenced cART after 1999 was ADI. Plot (C) shows 10 year survival probabilities for patients with viral qualitatively similar to the model used in the primary analysis load .400 copies/ml and with prior ADI. suggesting that the effects of survivor bias are limited. Also, the doi:10.1371/journal.pone.0048839.g003 year of cART commencement and duration of treatment were not important as covariates in survival models for the study population models of the data demonstrated that values of, and decreases in especially when adjusting for recent CD4 cell count. 10 year survival probabilities over the range of ages included in Third, the omission of the effects of prognostic lifestyle factors these analyses (up to 70 years old) were similar to those of the may reduce the accuracy of analyses, especially where the Australian general population when adjusting for other factors distribution of these factors differs from that of the general (maintaining high CD4 cell count, low viral load, no prior ADI population. In particular, smoking is not widely recorded in and non IDU-exposure). We found no evidence of increasing AHOD. In a subsample of the cohort for which these rates were
PLOS ONE | www.plosone.org 7 November 2012 | Volume 7 | Issue 11 | e48839 Page 74 of 127 Long-Term Survival in HIV Positive Patients collected, 25% of patients had reported as a smoker at most recent Taylor Square Private Clinic, Darlinghurst; E Jackson, J Shakeshaft, visit, and 40% as ex-smokers, compared to 22% and 29% of males Nepean Sexual Health and HIV Clinic, Penrith; K Brown, C McGrath, V in the general Australian population 2007–2008 [27]. Had McGrath, S Halligan, Illawarra Sexual Health Service, Warrawong; L Wray, P Read, H Lu, Sydney Sexual Health Centre, Sydney; D Couldwell, analyses been adjusted for smoking in this study, estimated Parramatta Sexual Health Clinic; D Smith,V Furner, Albion Street Centre; survival probabilities could possibly exceed those seen in the Dubbo Sexual Health Centre, Dubbo; J Watson*, National Association of general population when also controlling for other prognostic People living with HIV/AIDS; C Lawrence*, National Aboriginal factors. Similarly this might be expected when adjusting for other Community Controlled Health Organisation; B Mulhall*, Department of lifestyle factors such as recent injecting drug use, alcohol Public Health and Community Medicine, University of Sydney; M Law*, consumption etc. K Petoumenos*, S Wright*, H McManus*, C Bendall*, M Boyd*, The A strength of our analyses is the long-term follow-up compared Kirby Institute, University of NSW. to other studies, as few cohorts have data on patients who have Northern Territory: A Kulatunga, P Knibbs, Communicable Disease Centre, Royal Darwin Hospital, Darwin. received cART for 15+ years. Also, while randomised clinical trials Queensland: J Chuah*, M Ngieng, B Dickson, Gold Coast Sexual Health have been used to investigate mortality among adults infected with Clinic, Miami; D Russell, S Downing, Cairns Sexual Health Service, HIV receiving cART [28,29], they are generally limited to the Cairns; D Sowden, J Broom, K Taing, C Johnston, K McGill, Clinic 87, examination of short-term effects. This analysis was able to Sunshine Coast-Wide Bay Health Service District, Nambour; D Orth, D examine survival in HIV-positive patients across long-term cART Youds, Gladstone Road Medical Centre, Highgate Hill; M Kelly, A experience, and wide age ranges. Gibson, H Magon, Brisbane Sexual Health and HIV Service, Brisbane. In conclusion, mortality rates in this large cohort of Australian South Australia: W Donohue,O’Brien Street General Practice, Adelaide. HIV-positive individuals were close to those of the Australian Victoria: R Moore, S Edwards, R Liddle, P Locke, Northside Clinic, North Fitzroy; NJ Roth*, J Nicolson*, H Lau, Prahran Market Clinic, population, particularly for patients with CD4 counts above 500 South Yarra; T Read, J Silvers*, W Zeng, Melbourne Sexual Health cells/mL. These rates did not vary much by duration of treatment. Centre, Melbourne; J Hoy*, K Watson*, M Bryant, S Price, The Alfred Changes in mortality with age were similar to those in the Hospital, Melbourne; I Woolley, M Giles, T Korman, J Williams, Monash Australian general population. Further long-term follow-up is Medical Centre, Clayton. needed to characterize survival and other outcomes in HIV- Western Australia: D Nolan, J Skett, J Robinson, Department of Clinical positive patients on cART at older ages. Immunology, Royal Perth Hospital, Perth. Asterisks indicate steering committee members in 2011. Acknowledgments CoDe reviewers: AHOD reviewers: D Sowden, DJ Templeton, J Hoy, L Wray, J Chuah, The members of the Australian HIV Observational Database are: K Morwood, T Read, N Roth, I Woolley, M Kelly, J Broom. New South Wales: D Ellis, General Medical Practice, Coffs Harbour; M TAHOD reviewers: PCK Li, MP Lee, S Vanar, S Faridah, A Bloch, T Franic*, S Agrawal, L McCann, N Cunningham, T Vincent, Kamarulzaman, JY Choi, B Vannary, R Ditangco, K Tsukada, SH Holdsworth House General Practice, Darlinghurst; D Allen, JL Little, Han, S Pujari, A Makane, OT Ng, AJ Sasisopin. Holden Street Clinic, Gosford; D Smith, C Gray, Lismore Sexual Health & Independent reviewers: F Drummond, M Boyd. AIDS Services, Lismore; D Baker*, R Vale, East Sydney Doctors, Surry Hills; DJ Templeton*, CC O’Connor, C Dijanosic, RPA Sexual Health Author Contributions Clinic, Camperdown; E Jackson, K McCallum, Blue Mountains Sexual Health and HIV Clinic, Katoomba; M Grotowski, S Taylor, Tamworth Conceived and designed the experiments: HM KP MGL CCO MB PJR Sexual Health Service, Tamworth; D Cooper, A Carr, F Lee, K Hesse, K NR KW JB DR. Analyzed the data: HM. Wrote the paper: HM KP MGL Sinn, R Norris, St Vincent’s Hospital, Darlinghurst; R Finlayson, I Prone, CCO MB PJR NR KW JB DR.
References 1. Harrison KM, Song R, Zhang X (2010) Life expectancy after HIV diagnosis 11. Hughes R, Sterne J, Walsh J, Bansi L, Gilson R, et al. (2011) Long-term trends based on national HIV surveillance data from 25 states, United States. J Acquir in CD4 cell counts and impact of viral failure in individuals starting Immune Defic Syndr 53: 124–130. antiretroviral therapy: UK Collaborative HIV Cohort (CHIC) study. HIV 2. May MT, Sterne JA, Costagliola D, Sabin CA, Phillips AN, et al. (2006) HIV Med 12: 583–593. treatment response and prognosis in Europe and North America in the first 12. Wright ST, Carr A, Woolley I, Giles M, Hoy J, et al. (2011) CD4 cell responses decade of highly active antiretroviral therapy: a collaborative analysis. Lancet to combination antiretroviral therapy in patients starting therapy at high CD4 368: 451–458. cell counts. J Acquir Immune Defic Syndr 58: 72–79. 3. Mahy M, Stover J, Stanecki K, Stoneburner R, Tassie JM (2010) Estimating the 13. Li X, Margolick JB, Jamieson BD, Rinaldo CR, Phair JP, et al. (2011) CD4+ T- impact of antiretroviral therapy: regional and global estimates of life-years cell counts and plasma HIV-1 RNA levels beyond 5 years of highly active gained among adults. Sex Transm Infect 86 Suppl 2: ii67–71. antiretroviral therapy. J Acquir Immune Defic Syndr 57: 421–428. 4. Hogg R, Lima V, Sterne JA, Grabar S, Battegay M, et al. (2008) Life expectancy 14. The Australian HIV Observational Database (2002) Rates of combination of individuals on combination antiretroviral therapy in high-income countries: a antiretroviral treatment change in Australia, 1997–2000. HIV Med 3: 28–36. collaborative analysis of 14 cohort studies. Lancet 372: 293–299. 15. Falster K, Choi JY, Donovan B, Duncombe C, Mulhall B, et al. (2009) AIDS- 5. van Sighem AI, Gras LA, Reiss P, Brinkman K, de Wolf F (2010) Life related and non-AIDS-related mortality in the Asia-Pacific region in the era of expectancy of recently diagnosed asymptomatic HIV-infected patients ap- combination antiretroviral treatment. AIDS 23: 2323–2336. proaches that of uninfected individuals. AIDS 24: 1527–1535. 16. Kowalska JD, Friis-Moller N, Kirk O, Bannister W, Mocroft A, et al. (2011) The 6. Lohse N, Hansen AB, Pedersen G, Kronborg G, Gerstoft J, et al. (2007) Survival Coding Causes of Death in HIV (CoDe) Project: initial results and evaluation of of persons with and without HIV infection in Denmark, 1995–2005. Annals of methodology. Epidemiology 22: 516–523. Internal Medicine 146: 87–95. 17. (2010) Table 1.9 Deaths, Summary, Australia–1999 to 2009. Australian Bureau 7. d’Arminio Monforte A, Sabin CA, Phillips A, Sterne J, May M, et al. (2005) The of Statistics. changing incidence of AIDS events in patients receiving highly active 18. (2011) Table 59. Estimated Resident Population By Single Year Of Age, antiretroviral therapy. Archives of Internal Medicine 165: 416–423. Australia. Australian Bureau of Statistics. 8. Denholm JT, Yong MK, Elliott JH (2009) Long term management of people 19. (2011) Table 1: Life Tables Australia 2007–2009. Australian Bureau of Statistics. with HIV. Aust Fam Physician 38: 574–577. 20. Zhou J, Kumarasamy N, Ditangco R, Kamarulzaman A, Lee CK, et al. (2005) 9. Petoumenos K, Worm S, Reiss P, de Wit S, d’Arminio Monforte A, et al. (2011) The TREAT Asia HIV Observational Database: baseline and retrospective Rates of cardiovascular disease following smoking cessation in patients with HIV data. J Acquir Immune Defic Syndr 38: 174–179. infection: results from the D:A:D study(*). HIV Med 12: 412–421. 21. Baillargeon J, Borucki M, Black SA, Dunn K (1999) Determinants of survival in 10. Lewden C, the Mortality Working Group of COHERE. Time with CD4 Cell HIV-positive patients. Int J STD AIDS 10: 22–27. Count above 500 cells.mm3 Allows HIV-Infected Men, but not Women, to 22. Lewden C, Chene G, Morlat P, Raffi F, Dupon M, et al. (2007) HIV-infected Reach Similar Mortality Rates to Those of the General Population 2010; San adults with a CD4 cell count greater than 500 cells/mm3 on long-term Francisco.
PLOS ONE | www.plosone.org 8 November 2012 | Volume 7 | Issue 11 | e48839 Page 75 of 127 Long-Term Survival in HIV Positive Patients
combination antiretroviral therapy reach same mortality rates as the general 27. (2011) National Health Survey: Summary of Results, 2007–2008, Table 1.3: population. J Acquir Immune Defic Syndr 46: 72–77. Health risk factors: Age and sex, Proportion of persons. Australian Bureau of 23. Secrest AM, Becker DJ, Kelsey SF, Laporte RE, Orchard TJ (2010) Cause- Statistics. specific mortality trends in a large population-based cohort with long-standing 28. Cameron DW, Heath-Chiozzi M, Danner S, Cohen C, Kravcik S, et al. (1998) childhood-onset type 1 diabetes. Diabetes 59: 3216–3222. Randomised placebo-controlled trial of ritonavir in advanced HIV-1 disease. 24. Skrivarhaug T, Bangstad HJ, Stene LC, Sandvik L, Hanssen KF, et al. (2006) The Advanced HIV Disease Ritonavir Study Group. Lancet 351: 543–549. Long-term mortality in a nationwide cohort of childhood-onset type 1 diabetic 29. Hammer SM, Squires KE, Hughes MD, Grimes JM, Demeter LM, et al. (1997) patients in Norway. Diabetologia 49: 298–305. A controlled trial of two nucleoside analogues plus indinavir in persons with 25. Ouhoummane N, Abdous B, Emond V, Poirier P (2009) Impact of diabetes and human immunodeficiency virus infection and CD4 cell counts of 200 per cubic gender on survival after acute myocardial infarction in the Province of Quebec, millimeter or less. AIDS Clinical Trials Group 320 Study Team. N Engl J Med Canada–a population-based study. Diabet Med 26: 609–616. 337: 725–733. 26. May M, Gompels M, Delpech V, Porter K, Post F, et al. (2011) Impact of late diagnosis and treatment on life expectancy in people with HIV-1: UK Collaborative HIV Cohort (UK CHIC) Study. BMJ 343: d6016.
PLOS ONE | www.plosone.org 9 November 2012 | Volume 7 | Issue 11 | e48839 Page 76 of 127 Chapter 5: Loss to Follow-up in the Australian HIV Observational Database
(Manuscript resubmitted to Antiviral Therapy for second review with incorporated response to reviewer comment)
Page 77 of 127 Loss to Follow-up in the Australian HIV Observational Database
Hamish McManus1, Kathy Petoumenos1, Katherine Brown2, David Baker3, Darren
Russell4,5,6, Tim Read7, Don Smith8,9, Lynne Wray10, Michelle Giles11,12 , Jennifer Hoy13,
Andrew Carr14, Matthew Law1 on behalf of the Australian HIV Observational Database
1The Kirby Institute, UNSW Australia, Sydney, Australia
2Illawarra Sexual Health Service, Warrawong, Australia
3East Sydney Doctors, Sydney, NSW, Australia
4Cairns Sexual Health Service, Cairns, QLD, Australia
5James Cook University, Cairns, Australia
6The University of Melbourne, Melbourne, Australia
7Melbourne Sexual Health Centre, Alfred Health, Melbourne, Australia
8The Albion Centre, Sydney, Australia
9School of Public Health and Community Medicine, University of New South Wales, Sydney,
Australia
10Sydney Sexual Health Centre, Sydney, Australia
11Infectious Diseases, Monash Medical Centre, Melbourne, Australia
12Department of Infectious Diseases, Monash University, Melbourne, Australia
13Department of Infectious Diseases, The Alfred Hospital and Monash University,
Melbourne, Australia
14HIV, Immunology and Infectious Diseases Unit, St Vincent’s Hospital, Sydney, NSW,
Australia
Page 78 of 127 Abstract
Background: Loss to follow-up (LTFU) in HIV-positive cohorts is an important surrogate for
interrupted clinical care which can potentially influence the assessment of HIV disease
status and outcome. After preliminary evaluation of characteristics and risk factors for LTFU,
we evaluated the relative risk of mortality by LTFU status in a high resource setting.
Methods: Rates of LTFU were measured in the Australian HIV Observational Database for
a range of patient characteristics. Multivariate repeated measures regression methods were
used to identify determinants of LTFU. Mortality by LTFU status was ascertained using
linkage to the National Death Index. Survival following combination antiretroviral therapy
initiation was investigated using the Kaplan-Meier (KM) method and Cox proportional hazards models.
Results: Of 3,413 patients included in this analysis, 1,632 (47.8%) had at least one episode of LTFU after enrolment. In 1,283 (54.6%) of the 2,349 episodes of LTFU, patients returned
to care. Multivariate predictors of LTFU included viral load (VL)>10,000 copies/ml (1.63
(1.45-1.84) (ref ≤400)), time under follow-up (per year) (1.03 (1.02-1.04)) and prior LTFU
(per episode) (1.15 (1.06-1.24)). KM curves were similar by LTFU status (p=0.484). LTFU
was not associated with survival in Cox proportional hazards models (univariate hazard ratio
(HR) 0.93 (0.69-1.26) and multivariate HR 1.04 (0.77-1.43)).
Conclusions: Increased risk of LTFU was identified amongst patients with potentially higher
infectiousness. We did not find significant mortality risk associated with LTFU. This is
consistent with timely re-engagement with treatment, possibly via high levels of unreported
linkage to other health care providers.
Page 79 of 127 Introduction
Loss to follow-up (LTFU) in HIV-positive cohorts is an important surrogate for interrupted
clinical care which can potentially influence the assessment of HIV disease status and
outcome. Interrupted clinical follow-up of HIV-positive patients can delay timely initiation of therapy and impair treatment responses.
Prior studies have reported an association between episodes of LTFU and poorer outcome in HIV-positive patients in both low and high resource settings [1-7]. In particular, survival of
LTFU patients might be poor compared to patients in care if there is significant disease
resurgence during episodes of LTFU. The ascertainment of survival by LTFU status is an
important objective of this study as well as of similar studies of HIV-positive populations in
high resource settings [5, 8, 9].
Inaccurate assumptions about outcomes in LTFU patients can bias findings derived from in-
care populations [10]. Evaluation of risk of LTFU can assist in identification and adjustment
of biases introduced by different outcomes compared with patients in follow-up. However,
predicted outcomes in LTFU patients might also be confounded by unreported patient
linkage to other health care providers. By identifying mortality using national death
registries, reliable rates of survival in LTFU patients can be compared to patients in routine
care which might also allow some inference to be made about the extent to which patients
are truly disengaged from care [11-13].
Patient populations with extended durations of LTFU are also of importance because they
may include groups with relatively low treatment adherence who are more likely to have viral
rebound and who are, therefore, potentially a source of ongoing HIV transmission.
Identification of specific patients who may be at increased risk of LTFU can prompt
preventative strategies and can direct the introduction of supports to pre-empt discontinuous
Page 80 of 127 clinical attendance and improve treatment adherence [14, 15]. Determination of risk of LTFU is, therefore, important at the patient level to allow early intervention to prevent LTFU.
After preliminary investigation of rates and determinants of LTFU in a high resource setting we compared mortality in patients by LTFU status. For this we used the Australian HIV
Observational Database which has detailed long-term attendance data, a large patient population and wide regional coverage. We used linkage to the National Death Index to compare mortality in LTFU with that of patients under routine care.
Methods
The Australian HIV Observational Database (AHOD) is an observational clinical cohort study of patients with HIV infection seen at 29 clinical sites throughout Australia. AHOD utilises methodology which has been described in detail elsewhere [16]. Briefly, data are transferred electronically to The Kirby Institute, University of New South Wales every 6 months.
Prospective data collection commenced in 1999, with retrospective data provided where available.
Ethics approval for the AHOD study was granted by the University of New South Wales
Human Research Ethics Committee, and all other relevant institutional review boards.
Written informed consent was obtained from participating individuals. All study procedures were developed in accordance with the revised 1975 Helsinki Declaration. The Australian
Institute of Health and Welfare and all relevant institutional review boards granted specific ethics approval for this particular study which included linkage of consented AHOD patient identifiers to be made to the National Death Index to identify fact of death and date of death.
Study population
This analysis included patients who had been recruited to AHOD as part of general AHOD recruitment prior to 31 March 2013 with at least one recorded clinical visit thereafter.
Page 81 of 127 Patients from non-Australian sites were excluded as were patients recruited as part of AHOD
sub-studies which often target specific non representative patient groups.
We used all recorded site visits from enrolment in AHOD until the earlier of identified date of death or cohort censoring date (31 March 2103) or site specific censoring date as a
surrogate for clinical follow-up. Episodes of LTFU were defined as greater than 365 days between any recorded treatment visit and the earlier of censor date and next recorded visit.
A 365 day gap represents more than 2 missed visits based on routine clinical attendance
(according to patient, site and era) as well as 2 consecutive AHOD reporting periods without updated data, and has been used by cohort studies from similar resource settings [8, 9, 17,
18]. Episodes of LTFU commenced at 180 days after the last relevant recorded visit.
Covariates analysed were: sex; age (“<20”, “20-29”, “30-39”, “40-49”, “50-59”, “60-69”, “70-
79”); mode of HIV exposure (“men who have sex with men (MSM)”, “heterosexual”,
“intravenous drug use (IDU)”, “other/unknown”); hepatitis C virus (HCV) antibody (“never positive/never tested”, “positive ever”); hepatitis B virus surface antigen (HBsAg) (“never
positive/never tested”, “positive ever”); year of first HIV positive test (“<1996”, “≥1996”; this
categorisation was based on dichotomisation about the population mean), cART naïve
(“yes”, “no”), calendar year (“≤2004”, “>2004”; this categorisation was based on
dichotomisation about the population mean), time under follow-up, prior episodes of LTFU,
CD4 cell count (CD4+) (closest from 180 days prior or up to 30 days after – “0-199”,”200-
349”, “350–499”, “≥500” cells/µl); and viral load (VL) (closest 180 days prior or up to 30 days
after – “≤400”, “401-1000”, “1001-10,000”, >10000 copies/ml).
Categorical cut points for continuous variables, unless otherwise stated, were based on
widely used clinical thresholds to facilitate generalizability of results, but with sufficient
observations per category level to facilitate meaningful statistical analysis. Time dependent
risk factors were updated in the analysis at the time of visit.
Statistical Methods
Determinants of loss to follow-up
Page 82 of 127 Rates of LTFU episodes were calculated by demographic and clinical patient characteristics
Patients enrolled over 1 year prior to censoring (who were therefore at risk of becoming lost to follow-up) were followed from enrolment until the earlier of death, cohort censoring date
(31 March 2013) or site specific censoring date (31 March 2007 for one site subgroup of patients and 31 March 2009 for another site). Factors associated with LTFU were examined using repeated measures Poisson regression, with generalised estimating equations methodology. This allowed for multiple episodes of LTFU per patient and accounted for within and between patient variability. Exchangeable variance structure was assumed, but with robust calculated variances to minimise possible incorrectly assumed variance structure. For GEE models, time was coarsened to annual intervals. Time dependent variables were updated at the most recent visit except for CD4+ and VL where the median of test results over the year was used.
Data linkage to the National Death Index (NDI) To develop estimates of mortality adjusted for mortality in LTFU patients, the subgroup of
AHOD general recruitment with identified approval for data linkage was linked to the NDI using definite linkage for a range of linkage keys for each patient. Linkage was made using first two letters of given name, first two letters of surname, the day, month, and year of birth, and sex after preliminary quality assessment of the performance of a range of linkage keys by the Australian Institute of Health and Welfare (AIHW) linkage team[19].
Where possible we manually assessed linkage matches using comparison of linked date of death against AHOD date variables (including enrolment date, recent visit date and date of death).
We estimated sensitivity of matched NDI mortality status against AHOD mortality status
using linkage results for patients with a recorded AHOD death as indicated by the
completion of an AHOD Cause of Death (CoDe) form. CoDe forms are completed by a site
clinician and have detailed information on clinical status at time of death as well as attached
autopsy reports where relevant. These are reviewed by AHOD coordinators and if required
further reviewed by HIV specialist clinicians to verify primary and secondary causes of death.
Page 83 of 127 Because of the rigour of this process, recorded AHOD death was taken to indicate confirmed
death. In addition we assumed 100% specificity of matching after validating all linked deaths
against AHOD date variables to verify the absence of false positive matches.
We then developed estimates for the number of extra deaths identified by linkage using the
sensitivity from the analysis above to incorporate rates of true positives and of false
negatives in LTFU patients arising from the linkage process.
Finally, we resolved estimates by LTFU status and calculated a likely adjustment multiplier
as the ratio of estimated extra deaths versus linkage deaths in LTFU to apply to rates of
mortality amongst patients LTFU to adjust for missing deaths associated with patients LTFU.
Mortality estimates We compared crude rates of death for patients who had commenced cART and who had
consented to data linkage by time updated LTFU status. Follow-up status was categorised
as either “not LTFU”, “returned to care” or “LTFU”. Analysis baseline was cART initiation but
with left truncation applied such that entry was delayed until enrolment in instances where
patients had commenced cART prior to enrolment. Patients were followed until cohort
censor date or site specific censor date or death. An administrative censoring time of 17
years after cART commencement was also applied.
Univariate and multivariate Cox proportional hazards models were developed to evaluate the
relationship between LTFU and mortality while controlling for potential confounders. Forward
stepwise selection of covariates was used with forced inclusion of LTFU status to develop a
parsimonious model but which was inclusive of important predictors of survival. Results from
model fitting using backward selection methods were compared. Baseline covariates
examined were age, sex, mode of exposure, year HIV positive (<1996/≥1996), year of first
cART (<2000, ≥2000), prior AIDS defining illness (ADI), HCV (ever), HBsAg (ever), prior mono/dual (Mono/Duo) treatment and CD4+. Analyses were stratified by treatment centre.
Analyses were conducted using Stata version 12.1 (Stata Corp LP, College Station, TX,
United States).
Page 84 of 127 Results
Total AHOD recruitment was 3, 894 patients of whom 286 were omitted from analyses because they were from non-Australian sites or were part of targeted AHOD sub studies.
Of the remaining 3,608 patients, those who were recruited at least 1 year prior to censor date were included in analysis of determinants of LTFU (3,413 patients (94.6%)) and followed for a total of 23,922 person years from enrolment. Of these patients 1,632 (47.8%)
had at least one episode of LTFU.
Patients were predominantly male (3,208 (94.0%)), with mean age of 42.3 years at
enrolment and predominant route of HIV transmission was via MSM (2,581 (75.6%)) (Table
1). The median duration between attendances was 62 days (interquartile range (IQR) 17-
100) with 42.5% of visits being between 62 and 180 days apart, 5.7% of visits being between
180 and 365 days apart and 2.0% of visits being over 365 days apart. The overall crude rate
of LTFU (including episodes resulting in return to care) was 9.82 episodes per 100 person
years (95% confidence interval (CI):9.43-10.22)
Predictors of lost to follow-up
Increased crude rates of LTFU were associated with younger patients (<30 years (21.95
(18.64-25.84)), and 30-39 years (14.06 (13.08-15.12)), IDU mode of exposure (13.22 (11.39-
15.36)), VL>10,000 copies/ml (11.6 (10.51-12.81)), prior episodes of LTFU (20.63 (18.91-
22.51) with 1 prior episode), and earlier calendar year periods of follow-up (≤2004) (11.08
(10.38-11.83)) (Table 1).
In a multivariate model, increased risk of LTFU was associated with heterosexual mode of
exposure (HR 1.17 (95% CI 1.04-1.33) (ref MSM), p=0.012), patients who had initiated cART
(1.76 (1.44-2.16), p<0.001), increased VL (copies/ml) (1,001-10,000 1.34 (1.16-1.55),
>10,000 1.63 (1.45-1.84) (ref ≤400), p<0.001), time under follow-up (per additional year)
(1.03 (1.02-1.04), p<0.001) and prior episodes of LTFU (per additional episode) (1.15 ( 1.06-
1.24), p<0.001). Female sex (0.68 (0.55-0.83), p<0.001) and increased age (per year older)
Page 85 of 127 Table 1 Patient numbers, follow-up time, episodes of loss to follow-up and crude rates of loss to follow-up (LTFU) by characteristic categories for patients enrolled in the Australian HIV Observational database between 1999 and 2013 and with consent to data linkage N (%) Person years (PY) LTFU Rate (95% CI) per 100 PY All 3,413 (100.00) 23,922 2,349 9.82 (9.43-10.22) Sex Male 3,208 (94.0) 22,620 2,241 9.91 (9.51-10.33) Female 205 (6.0) 1,302 108 8.30 (6.87-10.02) Age1 Mean (SD) 42.3 (10.2) <30 321 (9.4) 656 144 21.95 (18.64-25.84) 30-39 1,247 (36.5) 5,240 737 14.06 (13.08-15.12) 40-49 1,126 (33.0) 9,089 881 9.69 (9.07-10.36) 50-59 519 (15.2) 5,997 427 7.12 (6.48-7.83) 60-69 164 (4.8) 2,407 128 5.32 (4.47-6.32) ≥70 36 (1.1) 532 32 6.01 (4.25-8.5) Exposure MSM 2,581 (75.6) 18,566 1,783 9.60 (9.17-10.06) IDU 209 (6.1) 1,301 172 13.22 (11.39-15.36) Heterosexual 499 (14.6) 3,236 333 10.29 (9.24-11.46) Other 124 (3.6) 820 61 7.44 (5.79-9.56) HCV (ever) No 3,026 (88.7) 21,252 2,066 9.72 (9.31-10.15) Yes 387 (11.3) 2,670 283 10.6 (9.43-11.91) HBsAg (ever) No 3,269 (95.8) 22,906 2,237 9.77 (9.37-10.18) Yes 144 (4.2) 1,015 112 11.03 (9.17-13.28) Year HIV Positive <1996 1,735 (50.8) 14,435 1,341 9.29 (8.81-9.80) ≥1996 1,508 (44.2) 8,419 888 10.55 (9.88-11.26) Missing 170 (5.0) 1,067 120 11.25 (9.40-13.45) Prior cART1 No 831 (24.4) 1,435 117 8.15 (6.80-9.77) Yes 2,582 (75.7) 22,486 2,232 9.93 (9.52-10.35) CD4 cell count1 median (IQR) 480 (320-665) <200 374 (11.0) 1,669 147 8.81 (7.49-10.35) 200-349 574 (16.8) 3,258 307 9.42 (8.43-10.54) 350-499 807 (23.6) 4,821 398 8.26 (7.48-9.11) ≥500 1509 (44.2) 12,300 996 8.10 (7.61-8.62) Missing 149 (4.4) 1,873 501 26.75 (24.51-29.2) Viral load1 median (IQR) ≤400 (≤400-8,000) ≤400 1910 (56.0) 16,601 1,198 7.22 (6.82-7.64) 401-1,000 175 (5.1) 545 50 9.18 (6.96-12.11) 1,001-10,000 431 (12.6) 1,599 167 10.45 (8.98-12.16) >10,000 759 (22.2) 3,379 392 11.60 (10.51-12.81) Missing 138 (4.0) 1,799 542 30.13 (27.70-32.78) Calendar Year1 ≤2004 2,250 (65.9) 8,157 904 11.08 (10.38-11.83) >2004 1,163 (34.1) 15,765 1,445 9.17 (8.71-9.65) Episodes of LTFU2 0 1,781 (52.2) 20,448 1,632 7.98 (7.60-8.38) 1 1,126 (33.0) 2,452 506 20.63 (18.91-22.51) ≥2 506 (14.8) 1,021 211 20.66 (18.05-23.64) 1. N=number at enrolment, PY & LTFU & Rate based on time updated value 2. N=total number ever, PY & LTFU & Rate based on time updated value
Page 86 of 127 (0.74 (0.71-0.78), p<0.001) were associated with decreased risk of LTFU. Missing recent
CD4+ (cells/µl) (1.67 (1.33-2.08) ref <200), p<0.001) and missing recent VL (3.35 (2.89-
3.88) (ref ≤400), p<0.001) were also associated with increased risk of LTFU in this model
(Table 2).
Data linkage Of 3,608 general recruitment AHOD patients, 3,404 (94.3%) were consented and linked to the NDI. Of these patients 2,529 (74.3%) had a current AHOD mortality status, while 875
(25.7%) did not. The estimated linkage sensitivity was 84.5% based on 246 of 291 confirmed
AHOD deaths being matched to NDI deaths.
Of 2,501 linked patients not LTFU, 263 (10.5%) had a death recorded in AHOD. Of the
remaining 903 patients who were LTFU, 28 (3.1%) had an AHOD death recorded, and 42
(4.7%) extra deaths were identified by linkage. A likely extra 8 (0.9%) deaths had occurred
but were missed by linkage based on the above sensitivity. This suggests that, for AHOD,
estimates of mortality in patients LTFU should be revised upwards by a factor of 1.11 (i.e.
78/70).
Mortality Of 3,404 AHOD patients linked to the NDI, 3,030 (89.0%) had initiated cART prior to censor
date and were included in survival analyses. Over the duration of survival analysis (from cART commencement until censoring) 323 deaths were observed: 228 (70.6%) in patients with no prior LTFU, 32 (9.9%) in patients returned to follow-up, and 63 (19.5%) in patients
LTFU. Overall mortality was 12.87 deaths/1000 person years (95% CI: 11.54-14.35). Crude rates of mortality by time updated LTFU status for enrolled patients were similar by follow-up
status: 13.35 (11.73-15.20) for patients with no prior LTFU; 11.41 (8.07-16.14) for patients
returned to follow-up; and 12.08 (9.44-15.47) for patients LTFU. Kaplan Meier curves of the
unadjusted relationship between LTFU and survival showed similar survival by LTFU status
(Figure 1) (log rank p=0.484).
Page 87 of 127 Table 2 Predictors of loss to follow-up (LTFU) after enrolment for univariate and multivariate Poisson regression analyses for patients enrolled in the Australian HIV Observational database between 1999 and 2013 and with consent to data linkage1 Univariate Multivariate IRR (95% CI) p P IRR (95% CI) p P Sex Male 1 1 Female 0.82 (0.67-1.00) 0.051 0.68 (0.55-0.83) <0.001 Age2 Per 10 years older 0.78 (0.75-0.82) <0.001 0.74 (0.71-0.78) <0.001 Exposure MSM 1 <0.001 1 0.018 IDU 1.38 (1.15-1.64) <0.001 1.16 (0.99-1.36) 0.059 Heterosexual 1.04 (0.92-1.18) 0.505 1.17 (1.04-1.33) 0.012 Other 0.76 (0.57-1.01) 0.063 0.87 (0.66-1.16) 0.350 HCV (ever) No 1 Yes 1.10 (0.95-1.26) 0.199 HBsAg (ever) No 1 Yes 1.16 (0.93-1.45) 0.200 Year HIV Positive <1996 1 0.547 ≥1996 1.04 (0.94-1.14) 0.429 Missing 1.10 (0.89-1.37) 0.371 Started cART2 No 1 1 Yes 1.46 (1.20-1.77) <0.001 1.76 (1.44-2.16) <0.001 Recent CD43 <200 1 <0.001 1 <0.001 200-349 1.02 (0.86-1.22) 0.822 1.17 (0.98-1.39) 0.090 350-499 0.89 (0.75-1.06) 0.200 1.05 (0.88-1.26) 0.587 ≥500 0.91 (0.77-1.08) 0.283 1.08 (0.90-1.29) 0.400 Missing 3.37 (2.81-4.04) <0.001 1.67 (1.33-2.08) <0.001 Recent viral load3 ≤400 1 <0.001 1 <0.001 401-1,000 1.13 (0.88-1.44) 0.334 1.20 (0.94-1.53) 0.139 1,001-10,000 1.27 (1.10-1.47) 0.001 1.34 (1.16-1.55) <0.001 >10,000 1.57 (1.41-1.74) <0.001 1.63 (1.45-1.84) <0.001 Missing 4.77 (4.34-5.25) <0.001 3.35 (2.89-3.88) <0.001 Calendar Year2 ≤2004 1 >2004 1.08 (1.01-1.16) 0.032 Time under follow-up2 Per additional year 1.03 (1.02-1.04) <0.001 1.03 (1.02-1.04) <0.001 Episode of LTFU2 Per additional episode 1.25 (1.16-1.34) <0.001 1.15 (1.06-1.24) <0.001 1. Loss to follow-up defined as ≥365 days until next clinic visits or prior to death or cohort/site censoring. GEE models, Poisson family, log link. Patients excluded if censoring or death occurred less than 1 year after enrolment. 2. Time updated variable based on most recent available measure 3. Time updated variable based on median annual value
Page 88 of 127 LTFU status was not associated with mortality in a univariate cox proportional hazard model
(p=0.893) (Table 3). Baseline covariates associated with increased mortality were increased age, IDU mode of exposure, prior ADI, HCV or HBV ever, earlier year of infection, earlier year of first cART and lower CD4+. LTFU status was not associated with survival in a multivariate model which was adjusted for Age, IDU exposure status, HBV, year of first cART and CD4+ (p=0.844).
Discussion
Characteristics of patients in the Australian HIV Observational Database who were at increased risk of LTFU were consistent with groups experiencing increased viral load and exhibiting higher infectiousness. We found no difference between mortality rates in patients according to follow-up status. This might indicate high levels of timely and often unreported re-engagement in care amongst these patients.
Overall, the observed rate of episodes of LTFU (which included multiple episodes per
patient) was relatively low (9.82/100 person years (95% CI: 9.43-10.22)). In the UK
Collaborative HIV Cohort (CHIC) study a higher rate of LTFU was observed (16.7/100
person years (95% CI: 16.4 -17.2)) [8]. However that analysis defined follow-up by duration
between CD4+ test dates rather than all clinical visits which might decrease observable
attendance. Compared to many other cohorts, AHOD has wide national coverage of the
epidemic and comprises a large proportion of Australian patients under care (approximately
15-20% [20, 21]), and internal linkage is used to capture duplicated cohort recruitment
amongst participating sites. Also, in Australia there are relatively low barriers to continuous
engagement with care providers because of accessible subsidised treatment that might
reduce financially related attrition, although in the UK for example ART is free. It is also
reasonable to expect lower rates of LTFU via emigration in AHOD patients given relatively
high Australian resourcing compared to other regional national health services. These
characteristics increase the likelihood and identification of re-enrolment of transient patients
across AHOD sites.
Page 89 of 127 Table 3 Univariate and multivariate hazard ratios (HR) of mortality following cART commencement for patients enrolled in the Australian HIV Observational Database between 1999 and 2013 with consent to data linkage1 Univariate models Multivariate model HR (95% CI) p P HR (95% CI) P P 2,3 LTFU No 1 0.893 1 0.844 Returned 0.95 (0.64-1.41) 0.803 1.12 (0.75-1.69) 0.570 Yes 0.93 (0.69-1.26) 0.649 1.04 (0.77-1.43) 0.782 Gender Male 1 Female 0.93 (0.57-1.51) 0.759 Age4 Per year 1.03 (1.02-1.04) <0.001 1.04 (1.03-1.05) <0.001 IDU5 No 1 1 Yes 1.73 (1.21-2.48) 0.003 1.95 (1.34-2.84) <0.001 Prior ADI4 No I Yes 1.50 (1.14-1.97) 0.003 HCV (ever) No I Yes 1.45 (1.09-1.94) 0.012 HBsAg (ever) No 1 1 Yes 1.71 (1.14-2.55) 0.009 1.71 (1.14-2.57) 0.009 Year HIV Positive <1996 1 1 ≥1996 0.57 (0.43-0.76) <0.001 0.67 (0.50-0.91) 0.009 Year of first cART <2000 1 1 ≥2000 0.64 (0.46-0.90) 0.009 0.66 (0.45-0.96) 0.028 CD4 cell count4 0-199 1 <0.001 1 <0.001 200-349 0.59 (0.44-0.79) <0.001 0.57(0.42-0.77) <0.001 ≥350 0.50 (0.38-0.67) <0.001 0.49 (0.36-0.66) <0.001 Prior Mono/Dual4 No 1 Yes 1.43 (1.13-1.79) 0.002 1. Cox proportional hazards ratios. Baseline= cART commencement. Patients left truncated at enrolment. Multivariate model selection using forward stepwise selection with forced inclusion of LTFU status. 2. LTFU if no recorded AHOD date other than AHOD death or NDI death within 1 year of cohort/site censor date or date of death. Lost to follow-up status time updated from 180 days after date of last AHOD visit date other than death. 1 subset of patients censored at 31 March 2007. 1 site censored at 31 March 2010. Administrative censored at 17 years of cART. Analyses stratified by treatment centre. 3. Time updated variable 4. At cART initiation 5. Mode of exposure via intravenous drug use
Demographically, risk of LTFU was associated with males, younger age and mode of exposure (heterosexual and marginally also IDU). These characteristics have been associated with residential transience [22, 23] and are consistent with shorter term engagement with localised healthcare as well as with relatively poor adherence to ART [24-
29] and higher transmission risk behaviours [27, 30-32]. This suggests that LTFU events are likely to correlate with increased risk of viral rebound and have serious implications for the
HIV epidemic with higher community VL and infectiousness, and consequent ongoing HIV transmission.
Page 90 of 127 We also observed higher risk of LTFU by higher recent VL which suggests that possibly less adherent patients are more likely to become LTFU. In this study there was no difference in risk associated with level of recently tested CD4+, but instead we observed increased risk associated with missing CD4+ tests (and similarly, missing VL tests). This was facilitated by defining LTFU based on durations between any recorded clinical attendances rather than just durations between attendances with recorded CD4+ testing. Our results suggest that at- risk patients are less likely to engage in a structured or consistent approach to treatment, to the extent that this is reflected by routine CD4+/VL monitoring. To contrast, Hill et al observed a moderate decrease in risk associated with decreasing CD4+ [8] in a similarly resourced setting using a definition that incorporated episodic LTFU. However, that analysis was based on follow-up defined by duration between CD4+ test dates rather than all clinical visits which may affect the association with risk of LTFU, as well as preclude comparison of relative rates of attendance without CD4+ testing.
Figure 1 KM survival probability by years of cART and follow-up status for patients enrolled in the Australian HIV Observational Database
Page 91 of 127 We also found cART initiation, prior follow-up and prior episodes of LTFU to be associated with increased risk. Conversely, Hill et al observed decreased risk associated with cART initiation [8, 9]. This might to an extent be attributable to relative differences in the recruitment to both cohorts, with over 40% of follow-up time in the CHIC study being of cART naïve patients compared to 6% in this study. Our findings describe more experienced patients and show that there is persistent habitual LTFU amongst this population. It is possible that these patients re-present at the same centre following periods of low treatment adherence which can result in viral rebound. This is consistent with AHOD patients generally being at risk of deterioration in health during LTFU episodes.
The median duration between all attendances in this study was 62 days (IQR 17-100) which accords with accepted Australian guidelines over the period of study [33], although this figure is likely strongly weighted by patients requiring more intensive care, whereas longer term, stable patients are likely to be seen at more extended intervals (42.5% of visits were between 2 and 6 months apart). Generally, observed attendance patterns in this study reflect that LTFU, as defined, is associated with strong departure from recommended and normal therapy. It is likely that many patients are not adherent to treatment for sizeable proportions of these episodes given that HIV prescriptions in Australia are generally made for much
shorter intervals and are invalid after one year duration [34, 35].
In this study we ascertained vital status of all patients by linkage to the NDI to investigate
potentially poorer survival in LTFU patients. We internally validated linkage against known
AHOD deaths and estimated linkage sensitivity as 84.5%, and given relatively low migration
as described above, we propose this as a reliable estimate of true patient mortality. We
found that true mortality was likely to be over 17% higher than recorded mortality. Overall
mortality was observed to be around 13 deaths per 1000 patient years (incorporating
adjustment for possible false negatives from NDI linkage) and we observed similar rates of
mortality by LTFU status. In particular, LTFU status was not a significant predictor of survival
Page 92 of 127 in multivariate analyses adjusting for age, IDU exposure, HBsAg status, year positive, year of first cART and baseline CD4+. This suggests that additional risk associated with potential disease re-emergence during these episodes is often able to be mitigated. This may be via delayed re-engagement with the same treating centre (as suggested via high rates of episodic LTFU), or via unreported linkage to other health care although this was not investigated by this study.
A limitation of this study is that more detailed attendance data was not incorporated into analyses. In particular recorded failure to attend scheduled visits is likely to correlate strongly with at risk patients. Use of this predictor would obviate the use of duration based definitions of LTFU which may be inappropriate for any given patient specific schedule of attendance.
This data is not collected in AHOD although it is recommended that this aspect of clinical attendance be investigated further. The consideration of non-attendance over intervals less than 365 days, which might potentially be informative, was limited by the bi-annual period of
AHOD reporting. In practice non-reporting of true attendances for any single reporting period
(half year intervals) was seen to be sufficiently common in preparatory analysis that calculated rates of non-attendance might be incorrectly and significantly increased.
However, this is likely to be mitigated at durations above a year, which permit sites to respond to quality assurance procedures, and generally rates of calculated LTFU in AHOD are low compared to comparable cohorts using the 365 day definition of LTFU.
Also, some socio-demographic risk factors were not able to be included in this analysis. In particular there are strong posited reasons for ethnicity being associated with risk of LTFU, although in the comparable studies of Mocroft and Hill results are conflicting [8, 9]. In AHOD, the population is predominantly Caucasian and any given ethnic minority category may have insufficient numbers to include in adjusted analyses. There are also reasonably high levels of underreporting of ethnicity, often by site policy, which might introduce bias to analyses incorporating this variable. Generally however, our results can be taken to be representative of the broader in treatment population in Australia.
Page 93 of 127 A strength of this analysis was the ability to link AHOD data to the National Death Index, which permitted the ascertainment of mortality in patients LTFU. We have therefore been able to develop rates of mortality which are unbiased by patient attrition and we have also shown that this bias is actually likely to be quite low in this cohort. Many similar studies have listed linkage to death registries as an important but unattained goal.
In summary, increased risk of LTFU was identified amongst patients with increased viral load and these patients might therefore have higher infectiousness than other groups. However, we did not find a significant mortality risk associated with LTFU suggesting that there is relatively low detriment to individuals that is associated with LTFU events. This is consistent with timely re-engagement with treatment possibly via high levels of unreported linkage to other health care providers.
Funding Source
The Australian HIV Observational Database is funded as part of the Asia Pacific HIV
Observational Database, a program of The Foundation for AIDS Research, amfAR, and is supported in part by a grant from the U.S. National Institutes of Health’s National Institute of
Allergy and Infectious Diseases (NIAID) (Grant No. U01-AI069907) and by unconditional grants from Merck Sharp & Dohme; Gilead Sciences; Bristol-Myers Squibb; Boehringer
Ingelheim; Janssen-Cilag; ViiV Healthcare. The Kirby Institute is funded by the Australian
Government Department of Health and Ageing, and is affiliated with the Faculty of Medicine,
UNSW Australia. The views expressed in this publication do not necessarily represent the position of the Australian Government.
Acknowledgements
Australian HIV Observational Database contributors
Asterisks indicate steering committee members in 2014.
New South Wales: D Ellis, General Medical Practice, Coffs Harbour; M Bloch, S Agrawal, T
Vincent, Holdsworth House Medical Practice, Darlinghurst; D Allen, JL Little, Holden Street
Clinic, Gosford; D Smith, R Hawkins, K Allardice, Lismore Sexual Health & AIDS Services,
Page 94 of 127 Lismore; D Baker*, V Ieroklis, East Sydney Doctors, Surry Hills; DJ Templeton*, CC
O’Connor, S Phan, RPA Sexual Health Clinic, Camperdown; E Jackson, K McCallum, Blue
Mountains Sexual Health and HIV Clinic, Katoomba; M Grotowski, S Taylor, Tamworth
Sexual Health Service, Tamworth; D Cooper, A Carr, F Lee, K Hesse, St Vincent’s Hospital,
Darlinghurst; R Finlayson, S Gupta, Taylor Square Private Clinic, Darlinghurst; R Varma, J
Shakeshaft, Nepean Sexual Health and HIV Clinic, Penrith; K Brown, V McGrath, S Halligan,
N Arvela Illawarra Sexual Health Service, Warrawong; L Wray, R Foster, H Lu, Sydney
Sexual Health Centre, Sydney; D Couldwell, Parramatta Sexual Health Clinic; DE Smith*, V
Furner Albion Street Centre; Clinic 16 – Royal North Shore Hospital, S Fernando; Dubbo
Sexual Health Centre, Dubbo; J Watson*, National Association of People living with
HIV/AIDS; C Lawrence*, National Aboriginal Community Controlled Health Organisation; B
Mulhall*, Department of Public Health and Community Medicine, University of Sydney; M
Law*, K Petoumenos*, S Wright*, H McManus*, C Bendall*, M Boyd*, The Kirby Institute,
University of NSW. Northern Territory: N Ryder, R Payne, Communicable Disease Centre,
Royal Darwin Hospital, Darwin. Queensland: M O’Sullivan, S White, Gold Coast Sexual
Health Clinic, Miami; D Russell, S Doyle-Adams, C Cashman, Cairns Sexual Health Service,
Cairns; D Sowden, K Taing, K McGill, Clinic 87, Sunshine Coast-Wide Bay Health Service
District, Nambour; D Orth, D Youds, Gladstone Road Medical Centre, Highgate Hill; M Kelly,
D Rowling, N Latch, Brisbane Sexual Health and HIV Service, Brisbane; B Dickson*,
CaraData. South Australia: W Donohue, O’Brien Street General Practice, Adelaide. Victoria:
R Moore, S Edwards, R Woolstencroft Northside Clinic, North Fitzroy; NJ Roth*, H Lau,
Prahran Market Clinic, South Yarra; T Read, J Silvers*, W Zeng, Melbourne Sexual Health
Centre, Melbourne; J Hoy*, K Watson*, M Bryant, S Price, The Alfred Hospital, Melbourne; I
Woolley, M Giles*, T Korman, J Williams*, Monash Medical Centre, Clayton. Western
Australia: D Nolan, J Robinson, Department of Clinical Immunology, Royal Perth Hospital,
Perth. New Zealand: G Mills, C Wharry, Waikato District Hospital Hamilton; N Raymond, K
Bargh, Wellington Hospital, Wellington.
Page 95 of 127 Coding of Death Form (CoDe) reviewers:
AHOD reviewers: D Sowden, J Hoy, L Wray, I Woolley, K Morwood, N Roth, K Choong, CC O'Connor, MA Boyd.
References 1. Brinkhof MW, Pujades-Rodriguez M, Egger M. Mortality of patients lost to follow-up in antiretroviral treatment programmes in resource-limited settings: systematic review and meta-analysis. PLoS One. 2009;4(6):e5790. 2. Dalal RP, Macphail C, Mqhayi M, Wing J, Feldman C, Chersich MF, et al. Characteristics and outcomes of adult patients lost to follow-up at an antiretroviral treatment clinic in johannesburg, South Africa. J Acquir Immune Defic Syndr. 2008;47(1):101-7. 3. Brennan AT, Maskew M, Sanne I, Fox MP. The importance of clinic attendance in the first six months on antiretroviral treatment: a retrospective analysis at a large public sector HIV clinic in South Africa. J Int AIDS Soc. 2010;13:49. 4. Egger M, Spycher BD, Sidle J, Weigel R, Geng EH, Fox MP, et al. Correcting mortality for loss to follow-up: a nomogram applied to antiretroviral treatment programmes in sub-Saharan Africa. PLoS Med. 2011;8(1):e1000390. 5. Edwards JK, Cole SR, Westreich D, Moore R, Mathews C, Geng E, et al. Loss to clinic and five-year mortality among HIV-infected antiretroviral therapy initiators. PLoS One. 2014;9(7):e102305. 6. Horberg MA, Hurley LB, Silverberg MJ, Klein DB, Quesenberry CP, Mugavero MJ. Missed office visits and risk of mortality among HIV-infected subjects in a large healthcare system in the United States. AIDS Patient Care STDS. 2013;27(8):442-9. 7. Lanoy E, Lewden C, Lievre L, Tattevin P, Boileau J, Aouba A, et al. How does loss to follow-up influence cohort findings on HIV infection? A joint analysis of the French hospital database on HIV, Mortalite 2000 survey and death certificates. HIV Med. 2009;10(4):236-45. 8. Hill T, Bansi L, Sabin C, Phillips A, Dunn D, Anderson J, et al. Data linkage reduces loss to follow-up in an observational HIV cohort study. J Clin Epidemiol. 2010;63(10):1101-9. 9. Mocroft A, Kirk O, Aldins P, Chies A, Blaxhult A, Chentsova N, et al. Loss to follow- up in an international, multicentre observational study. HIV Med. 2008;9(5):261-9. 10. Porter K, Johnson AM, Phillips AN, Darbyshire JH. The practical significance of potential biases in estimates of the AIDS incubation period distribution in the UK register of HIV seroconverters. AIDS. 1999;13(14):1943-51. 11. Simard EP, Pfeiffer RM, Engels EA. Mortality due to cancer among people with AIDS: a novel approach using registry-linkage data and population attributable risk methods. AIDS. 2012;26(10):1311-8.
Page 96 of 127 12. Fonseca MG, Coeli CM, de Fatima de Araujo Lucena F, Veloso VG, Carvalho MS. Accuracy of a probabilistic record linkage strategy applied to identify deaths among cases reported to the Brazilian AIDS surveillance database. Cad Saude Publica. 2010;26(7):1431- 8. 13. Centers for Disease C, Prevention. Electronic record linkage to identify deaths among persons with AIDS--District of Columbia, 2000-2005. MMWR Morb Mortal Wkly Rep. 2008;57(23):631-4. 14. Jordan WC. Reaching lost-to-care populations. Clin Infect Dis. 2007;45 Suppl 4:S275-80. 15. Harding R, Krakauer EL, Sithole Z, De Lima L, Selman L. The 'lost' HIV population: time to refocus our clinical and research efforts. AIDS. 2009;23(1):145-6. 16. Friis-Moller N, Sabin CA, Weber R, d'Arminio Monforte A, El-Sadr WM, Reiss P, et al. Combination antiretroviral therapy and the risk of myocardial infarction. N Engl J Med. 2003;349(21):1993-2003. 17. Cullen BL, Codere G, Wallace LA, Baguley S, Clutterbuck DJ. British Association for Sexual Health and HIV Scottish investigation 2010-2011: reasons for non-attendance in individuals lost to follow-up for HIV care for more than 12 months in Scotland. Int J STD AIDS. 2013;24(6):481-4. 18. Lebouche B, Yazdanpanah Y, Gerard Y, Sissoko D, Ajana F, Alcaraz I, et al. Incidence rate and risk factors for loss to follow-up in a French clinical cohort of HIV-infected patients from January 1985 to January 1998. HIV Med. 2006;7(3):140-5. 19. EO2013-2-11: Risk in Loss to Follow-up in the Australian HIV Observational Database Linkage Report. The Australian Institute of Health and wellfare; 2013. 20. Falster K, Gelgor L, Shaik A, Zablotska I, Prestage G, Grierson J, et al. Trends in antiretroviral treatment use and treatment response in three Australian states in the first decade of combination antiretroviral treatment. Sexual health. 2008;5(2):141-54. 21. Australian HIV Observational Database Annual Report. In: Institute TK, editor.: The University of New South Wales; 2013. 22. Pollio DE. The relationship between transience and current life situation in the homeless services-using population. Social Work. 1997;42(6):541-51. 23. Tulloch AD, Fearon P, David AS. Timing, prevalence, determinants and outcomes of homelessness among patients admitted to acute psychiatric wards. Soc Psych Psych Epid. 2012;47(7):1181-91. 24. Chesney MA. Factors affecting adherence to antiretroviral therapy. Clin Infect Dis. 2000;30 Suppl 2:S171-6.
Page 97 of 127 25. Keiser O, Spycher B, Rauch A, Calmy A, Cavassini M, Glass TR, et al. Outcomes of antiretroviral therapy in the Swiss HIV Cohort Study: latent class analysis. AIDS Behav. 2012;16(2):245-55. 26. Andrews JR, Wood R, Bekker LG, Middelkoop K, Walensky RP. Projecting the benefits of antiretroviral therapy for HIV prevention: the impact of population mobility and linkage to care. J Infect Dis. 2012;206(4):543-51. 27. Rachlis B, Brouwer KC, Mills EJ, Hayes M, Kerr T, Hogg RS. Migration and transmission of blood-borne infections among injection drug users: understanding the epidemiologic bridge. Drug Alcohol Depend. 2007;90(2-3):107-19. 28. South SJ, Crowder KD. Residential mobility between cities and suburbs: race, suburbanization, and back-to-the-city moves. Demography. 1997;34(4):525-38. 29. Wood E, Yip B, Gataric N, Montaner JS, O'Shaughnessy MV, Schechter MT, et al. Determinants of geographic mobility among participants in a population-based HIV/AIDS drug treatment program. Health Place. 2000;6(1):33-40. 30. Madden HC, Phillips-Howard PA, Hargreaves SC, Downing J, Bellis MA, Vivancos R, et al. Access to HIV community services by vulnerable populations: evidence from an enhanced HIV/AIDS surveillance system. AIDS Care. 2011;23(5):542-9. 31. German D, Davey MA, Latkin CA. Residential transience and HIV risk behaviors among injection drug users. AIDS Behav. 2007;11(6 Suppl):21-30. 32. Strathdee SA, Hallett TB, Bobrova N, Rhodes T, Booth R, Abdool R, et al. HIV and risk environment for injecting drug users: the past, present, and future. Lancet. 2010;376(9737):268-84. 33. 2013 Antiretroviral Guidelines with Australian Commentary: Australasian Society for HIV Medicine; 2014 [updated 2/12/2013; cited 2014 11/03/2014]. Available from: http://arv.ashm.org.au/arv-guidelines/initiating-art-in-treatment-naive-patients. 34. Medicare_Australia. Highly Specialised Drugs (HSD) program: Commonwealth of Australia; 2014 [updated 19 December, 2013; cited 2014 12 May, 2014]. Available from: http://www.medicareaustralia.gov.au/provider/pbs/highly-specialised-drugs/#N100E3. 35. Medicare_Australia. Browse by Section 100 Item List: Commonwealth of Australia; 2014 [cited 2014 12 May, 2014]. Available from: http://pbs.gov.au/browse/section100- hb?initial=a.
Page 98 of 127 Chapter 6: Determinants of suicide and accidental or violent death in the Australian HIV Observational Database
McManus H, Petoumenos K, Franic T, Kelly MD, Watson J, O'Connor CC, et al. Determinants of suicide and accidental or violent death in the Australian HIV Observational Database. PLoS One. 2014;9(2):e89089.
Page 99 of 127 Determinants of Suicide and Accidental or Violent Death in the Australian HIV Observational Database
Hamish McManus1*, Kathy Petoumenos1, Teo Franic2, Mark D. Kelly3, Jo Watson4, Catherine C. O’Connor1,5,6, Mark Jeanes7, Jennifer Hoy7, David A. Cooper1, Matthew G. Law1, on behalf of the Australian HIV Observational Database" 1 The Kirby Institute, University of New South Wales, Sydney, Australia, 2 Holdsworth House GP, Sydney, Australia, 3 Brisbane Sexual Health and HIV Service, Brisbane, Queensland, Australia, 4 National Association of People with HIV Australia, Sydney, Australia, 5 RPA Sexual Health, Royal Prince Alfred Hospital, Sydney, Australia, 6 Central Clinical School, Sydney University, Sydney, Australia, 7 Department of Medicine, Monash University, Melbourne, Victoria, Australia
Abstract
Background: Rates of suicide and accidental or violent death remain high in HIV-positive populations despite significantly improved prognosis since the introduction of cART.
Methods: We conducted a nested case-control study of suicide and accidental or violent death in the Australian HIV Observational Database (AHOD) between January 1999 and March 2012. For each case, 2 controls were matched by clinic, age, sex, mode of exposure and HIV-positive date to adjust for potential confounding by these covariates. Risk of suicide and accidental or violent death was estimated using conditional logistic regression.
Results: We included 27 cases (17 suicide and 10 violent/accidental death) and 54 controls. All cases were men who have sex with men (MSM) or MSM/ injecting drug use (IDU) mode of exposure. Increased risk was associated with unemployment (Odds Ratio (OR) 5.86, 95% CI: 1.69–20.37), living alone (OR 3.26, 95% CI: 1.06–10.07), suicidal ideation (OR 6.55, 95% CI: 1.70–25.21), and .2 psychiatric/cognitive risk factors (OR 4.99, 95% CI: 1.17–30.65). CD4 cell count of .500 cells/mL (OR 0.25, 95% CI: 0.07–0.87) and HIV-positive date $1990 (1990–1999 (OR 0.31, 95% CI: 0.11–0.89), post-2000 (OR 0.08, 95% CI: 0.01– 0.84)) were associated with decreased risk. CD4 cell count $500 cells/mL remained a significant predictor of reduced risk (OR 0.15, 95% CI: 0.03–0.70) in a multivariate model adjusted for employment status, accommodation status and HIV-positive date.
Conclusions: After adjustment for psychosocial factors, the immunological status of HIV-positive patients contributed to the risk of suicide and accidental or violent death. The number of psychiatric/cognitive diagnoses contributed to the level of risk but many psychosocial factors were not individually significant. These findings indicate a complex interplay of factors associated with risk of suicide and accidental or violent death.
Citation: McManus H, Petoumenos K, Franic T, Kelly MD, Watson J, et al. (2014) Determinants of Suicide and Accidental or Violent Death in the Australian HIV Observational Database. PLoS ONE 9(2): e89089. doi:10.1371/journal.pone.0089089 Editor: Gurudutt Pendyala, University of Nebraska Medical Center, United States of America Received July 30, 2013; Accepted December 23, 2013; Published February 19, 2014 Copyright: ß 2014 McManus et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: The Australian HIV Observational Database is funded as part of the Asia Pacific HIV Observational Database, a program of The Foundation for AIDS Research, amfAR, and is supported in part by a grant from the U.S. National Institutes of Health’s National Institute of Allergy and Infectious Diseases (NIAID) (Grant No. U01-AI069907) and by unconditional grants from Merck Sharp & Dohme; Gilead; Bristol-Myers Squibb; Boehringer Ingelheim; Roche; Pfizer; GlaxoSmithKline; Janssen-Cilag. The Kirby Institute is funded by The Australian Government Department of Health and Ageing, and is affiliated with the Faculty of Medicine, The University of New South Wales. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have the following interests: the Australian HIV Observational Database is funded in part by unconditional grants from Merck Sharp & Dohme; Gilead; Bristol-Myers Squibb; Boehringer Ingelheim; Roche; Pfizer; GlaxoSmithKline; Janssen-Cilag. MGL has received research grants, consultancy and/or travel grants from Boehringer Ingelheim; Bristol-Myers Squibb; Gilead; GlaxoSmithKline; Janssen-Cilag; Johnson & Johnson; Merck Sharp & Dohme; Pfizer; Roche; and CSL Ltd. JH’s institution has received funding for investigator-initiated research, advisory board honorarium and conference sponsorship from Janssen- Cilag, Gilead Sciences, Merck Sharp & Dohme, and ViiV Healthcare. This does not alter their adherence to all the PLOS ONE policies on sharing data and materials. * E-mail: [email protected] "Membership of the Australian HIV Observational Database is provided in the Acknowledgments.
Introduction receiving effective cART [6–8]. The Data Collection on Adverse Events of Anti-HIV Drugs (D:A:D) cohort reported suicide as the In the era of effective combination antiretroviral treatment cause of death in 4% of deaths, with a further 2.5% attributed to (cART) more than half the causes of death in HIV positive patients drug overdose and 1.5% as accident [9]. In CASCADE, suicide are non-AIDS related [1–5], the most common being non-AIDS was reported in 6.4% of deaths, violence in 3.3%, and 5.7% of defining cancers, cardiovascular disease and liver disease. deaths were attributed to substance abuse [10]. High rates of suicide and accidental or violent death have also In the early years of the HIV epidemic, poor prognosis was a been described in HIV infected populations including in those key contributing factor to high rates of suicide. Common
PLOS ONE | www.plosone.org 1 February 2014 | Volume 9 | Issue 2 | e89089 Page 100 of 127 Suicide and Accidental or Violent Death in AHOD experiences of HIV positive people such as stigmatisation, date and cause of death. Prospective data collection commenced in discrimination, social isolation, anxiety and depression, as well as 1999, with retrospective data provided where available. frequent substance abuse, were also identified as contributory Ethics approval for the AHOD study was granted by the factors to suicide risk. Yet, despite the significant improvement in University of New South Wales Human Research Ethics prognosis since the introduction of cART the rates of suicide Committee, and all other relevant institutional review boards. remain high. This was demonstrated in the Swiss HIV Cohort Written informed consent was obtained from participating Study where rates of suicide decreased substantially in the cART individuals. All study procedures were developed in accordance era compared to the pre cART era but still remained well above with the revised 1975 Helsinki Declaration. Further specific ethics that observed in the general population [6]. approval was granted for this particular study by the University of In the Swiss HIV Cohort Study, the majority (.75%) of New South Wales Human Research Ethics Committee, and all patients who committed suicide in the cART era had a diagnosis other relevant institutional review boards. of mental illness with depression being the most common (.80%). A significant proportion (23%) of patients who died by suicide in Classification of causes of death the cART era had untreated mental illness. In this cohort, suicide The primary endpoint for this study was mortality from suicide rates were shown to decline with increasing CD4 cell counts. and accidental or other violent causes and has been classified as Advanced clinical stage (using the US Centres for Disease Control ‘Unnatural’ cause of death in a previous AHOD study [15]. and Prevention classification system [11]) was significantly Collection of data on cause of death in AHOD has been described associated with suicide risk in both pre- and post-cART eras, in detail elsewhere [3,15]. Briefly, AHOD has collected detailed after adjustment for other socio-demographic factors and history information on cause of death (COD) for all deaths occurring since of psychiatric treatment. However, CD4 cell count and other HIV study inception in 1999: from 2001 until 2002 collected by the related factors were not included in this risk factor analysis. In the study coordinator via direct contact with relevant sites and Concerted Action on SeroConversion to AIDS and Death in included study participant deaths occurring prior to that period; Europe (CASCADE) study, latest CD4 cell count was not thereafter until 2005 using a standardized COD form based on the significantly associated with violent causes of death. As with the Data collection on adverse events of anti-HIV Drugs (D:A:D) Swiss HIV Cohort Study, the CASCADE study did not assess risk cohort original COD form [16]; and from 2005 using, the more factors for suicide which take into account both HIV related detailed D:A:D cohort CoDe case report form (CRF) for coding factors and general risk factors for suicide in one model. causes of death and adapted from ICD-10 codes (http://www. The extent to which HIV infection is also associated with cphiv.dk/Portals/0/files/CRF2012v2.pdf) [17]. Both the initial increased risk of suicide or violent death is not well documented. COD form and later the CoDe are completed by a clinician at the Specifically the relative contribution of HIV related factors such as study site with details of autopsy reports appended where relevant immune deficiency, as well as aspects of treatment neuroprotec- and then forwarded to the Kirby Institute for review by AHOD tion/ neurotoxicity (e.g. cerebro-spinal fluid (CSF) penetrating/ coordinators. If required, an independent HIV specialist clinician non penetrating ART) compared to accepted psychosocial risk verifies the primary and secondary causes of death and in cases factors has not been described in detail. Recent qualitative models where inadequate information is provided to determine the COD, of suicidality in HIV populations in the post cART era also the study coordinator contacts personnel at the study site for incorporate synergistic effects of HIV related factors and ageing further information. Patients are classified as having an unknown [12]. COD if no further information is obtained. Most HIV-related studies of suicide focus on suicidality as an For this study, consistency between recorded cause of death and endpoint [13]. This obviates difficulties associated with low observed medical records was verified during the site visit by the prevalence of suicide cases, case finding and retrospective data AHOD study coordinator at the time of study specific data collection. However, there are important differences between collection. determinants of suicidal ideation and suicide. For example, in a study of British private households, Gunnell et al found differences Statistical methods in the risk pattern of suicidal thoughts compared to completed This study is a nested case control study from those sites with suicide according to gender and age [14]. In that study the confirmed cases of suicide and accidental or violent death for incidence of suicidal thoughts was seen to be over 200 times patients consented and recruited to AHOD over the period from greater than the incidence of suicide. In this study we analysed study inception in 1999 until 31 March 2012. For each case 2 determinants of confirmed cases of suicide as well as of accidental controls were randomly selected from AHOD patients matched to or violent death to develop adjusted models of risk associated with cases by treating clinic, sex, 10-year age group at the time of case both psychosocial and HIV-related factors. death (‘‘20–29’’/’’30–39’’/’’40–49’’/’’50–59’’) and mode of expo- sure (‘‘men who have sex with men (MSM)’’/’’MSM and IDU’’/’’ Methods IDU’’/’’ Heterosexual’’/’’Receipt of blood products’’/’’Other’’) to control for confounder effects. Controls were alive and were HIV- Study Population positive at the time of case death. Mode of exposure was not The Australian HIV Observational Database (AHOD) is an included in analyses as an independent predictor, to avoid possible observational clinical cohort study of patients with HIV infection large discrepancy in empiric distributions of modes of exposure seen at 27 clinical sites throughout Australia. AHOD utilises between cases and controls. This would substantially reduce the methodology which has been described in detail elsewhere [14]. statistical efficiency of analyses incorporating this covariate. Briefly, data are transferred electronically to The Kirby Institute, Specific data on possible risk factors associated with suicide or University of New South Wales every 6 months. Core data accidental or violent cause of death was collected using a study variables include: sex; date of birth; date of most recent visit; HIV Case Record Form (CRF) by a single AHOD coordinator (HM) at exposure; hepatitis B virus (HBV) surface antigen status; hepatitis site visits. An additional 2 case CRFs for 1 site were completed by C virus (HCV) antibody status; CD4 and CD8 counts; HIV viral site psychologists using medical records and previous study load; antiretroviral treatment data; AIDS-defining illnesses; and documents not available at the time of site visit by the AHOD
PLOS ONE | www.plosone.org 2 February 2014 | Volume 9 | Issue 2 | e89089 Page 101 of 127 Suicide and Accidental or Violent Death in AHOD coordinator. An additional 1 case CRF and 2 associated control accurately model immunological/virological response than using CRFs for 1 other site were completed by the principal investigator most recent CD4 cell counts and HIV viral load values. for that site following determination by the research ethics Specifically, constant rates of change in CD4 cell counts and committee for that site that the AHOD investigator not be HIV viral load values between test dates were assumed over the granted direct access to those medical records. All CRF data was interval between consecutive tests and these gradients used to uploaded to an MS Access database at the Kirby Institute by calculate duration weighted average CD4 cell counts and HIV electronic form entry. viral load values for year prior to case death. Specific CRF variables used in this analysis included socio- Determinants of suicide and accidental or violent death were demographic and lifestyle factors, specifically: Employment in the analysed using univariate conditional logistic regression, matching year prior to case death (dichotomised to ‘‘Full time employment cases to controls and using the variables listed above as (FTE)’’/’’Not FTE’’); Accommodation in the year prior to case independent variables. Robust Huber/White variance estimation death (dichotomised to ‘‘Alone’’/’’Not alone’’); Record of prior was used. suicide attempts (‘‘Yes’’/’’No’’); Recorded alcohol consumption in Multivariate analysis was conducted using backward stepwise the year prior to case death (‘‘No Record’’ (NR)/’’ No selection from the set of significant predictors based on univariate consumption’’/’’Consumption’’); Specified alcohol amount con- analyses but was limited by low numbers of observations. sumed in year prior to case death; Recorded smoking prior to case Interaction between covariates was examined using bivariate death (‘‘NR’’/’’Never’’/’’Currently’’/’’Previously’’); Recorded interaction models of significant predictors in multivariate models. recreational/ illicit drug use (‘‘NR’’/’’Never’’/ ‘‘YES-non A sensitivity analysis investigating determinants of suicide-only IDU’’/’’Yes-IDU’’) and Record of recreational/illicit drug use in death was conducted using univariate conditional logistic regres- the year prior to case death. sion as above. Data were analysed using Stata version 12 (Stata CRF variables also included the incidence (ever and in year Corporation, College Station, Texas, USA). prior to case death) of recorded mental and cognitive and other risk factors, specifically: suicidal ideation or intent; depression/ Results anxiety; agitation; dysthymia; bipolar disorder; schizophrenia; other mood disorders; anorexia nervosa; chronic pain; epilepsy; Patient characteristics and dementia. Data was also collected on recorded psychiatric A total of 81 patients were included in this analysis – 27 cases medications (type, use ever, use in the year prior to case death). and 54 controls. Of cases, 17 (63%) were classified as death by CRF detail on specific psychiatric, cognitive and other risk suicide with mode of suicide by hanging (5 deaths) and drug factors was determined and corroborated where possible by overdose (5 deaths) being the most common (Table 1). A further repeated clinician entries in medical records, entries by different 10 cases (37%) were classified as violent or accidental death, with clinicians, by specialist referrals and diagnoses including psychi- death with associated drug overdose (8 deaths) being the most atrist notes and psychologist/counselor notes, and by recorded common. All cases were male, with a mean age of 41.6 years prescriptions of associated medications/ therapies. compared with 41.1 years for controls. Most deaths occurred prior An aggregated count of mental and cognitive risk factors (‘‘0’’/ to 2006 (61.6%) and mode of HIV exposure was either MSM ’’1–2’’/’’.2’’) was also calculated as the count of first incidence of (92.6%) or MSM/IDU (7.4%) (Table 1). these conditions. Patient characteristics prior to case death are shown in Table 2. Aggregated quantity of alcohol intake (Alcohol) was determined Cases generally recorded higher levels of risk factors for suicide or by ranking reported alcohol consumption quantities for the year violent death. Of social and demographic risk factors, cases were prior to case death and classifying as ‘‘Low‘‘, ’’Moderate’’, ‘‘High’’ less likely to be in full time employment (59.3% vs. 35.3%), and or ‘‘Not stated’’. Specifically, ‘‘Low’’ was categorised as up to and more likely to live alone or in an institutional environment (55.6% including 3 standard drinks/month, ‘‘Moderate’’ was up to and vs. 25.9%). Of mental health, neurocognitive and other medical including 2–5 standard drinks/night while ‘‘High’’ was all risk factors, cases were more likely to have a record of suicidal consumption greater than this. Where there was reported alcohol ideation (55.6% vs. 20.4%) and to have an increased number of consumption without specified quantity of alcohol consumption, recorded mental and cognitive and other clinical risk factors (.2 alcohol intake was assigned the median value (‘‘Moderate’’). conditions ever recorded (29.6% vs. 11.1%)). Of HIV specific Other core HIV related variables routinely collected in AHOD factors, cases were less likely to have recent CD4 count $500 and included in this analysis were CD4 cell count, HIV viral load, cells/mL (18.5% vs. 46.3%). Cases were more likely to have been AIDS defining illnesses (ADI) (‘‘Yes’’/’’No’’) and antiretroviral HIV-positive prior to 1990 (44.4% vs. 22.2%). treatments (efavirenz, nevirapine, abacavir, stavudine, lamivudine and ritonavir). The duration weighted use of cART with cerebro- Univariate analysis spinal fluid penetrative effectiveness scores (CPE) greater than the Factors significantly associated with increased risk of suicide and study median (nerocART) in the year prior to case death was also accidental or violent death (Table 2) were employment (not in full included in analyses. CPE scores were based on the 2010 ranks time employment (p = 0.005), not stated (p = 0.003)), living alone published by Letendre [18]. NeurocART status was assigned to or in an institution (p = 0.021), record of suicidal ideation regimens with CPE greater than the study median for that year. (p = 0.006) and more than two recorded psychiatric/cognitive risk Neurocognitive impairment (NCI) was included in analyses factors (p = 0.031)). Factors significantly associated with a reduced based on medical record of dementia, or detail listing memory loss, risk of suicide or accidental or violent death were recent CD4 cell intellectual impariment, encephalopathy or brain damage, or count greater than or equal to 500 cells/ml (p = 0.022) and year incidence of HIV encephalopathy/AIDS dementia complex HIV-positive after 1990 (1990–1999 (p = 0.03), $2000 (p = 0.035). (ADC) or progressive multifocal leukoenephalopathy (PML) recorded in AHOD. Multivariate analysis Duration weighted average of CD4 cell counts and HIV viral To facilitate fitting of the multivariate model given low numbers load values from the year prior to case death as predictors of of observations, certain inclusion variables were dichotomised. A suicide risk were used in analyses. These variables might more multivariate model was initially fit with recorded recent full time
PLOS ONE | www.plosone.org 3 February 2014 | Volume 9 | Issue 2 | e89089 Page 102 of 127 Suicide and Accidental or Violent Death in AHOD
Table 1. Characteristics used for case control selection1. Discussion This study found multivariate association between increased risk of suicide and accidental or violent death in HIV-positive patients, CASE CONTROL psychosocial factors (employment and accommodation status) and n=27 % n=54 % recent immunological status while controlling for age, sex, mode of Suicide death (n = 17) exposure, treating clinic and calendar year of HIV diagnosis. We found an increased prevalence of observed psychosocial risk hanging 5 18.5 factors in cases which reflects well documented prognostic drug overdose 5 18.5 indicators [7,19]. In particular, employment, accommodation wrist cutting 1 3.7 status and suicidal ideation were associated with increased risk. In gas inhalation 1 3.7 this study previously identified risk factors were not individually jump from height 1 3.7 prognostic (alcohol intake, smoking status, recent recreational/ not specified 4 14.8 illicit drug use, depression/anxiety, chronic pain and NCI). This is attributable to insufficient difference in prevalence between cases Violent/accidental death (n = 10) and controls for the given statistical power of this analysis. These drug overdose (direct or associated) 8 29.6 results are also consistent with a high prevalence of psychosocial motor vehicle accident 2 7.4 risk factors in HIV-positive populations in particular, but also in Gender medically ill populations in general [20-23]. Male 27 100 54 100 The observed significance of increased number of psychiatric/ cognitive risk factors in analyses is consistent with the exacerbation Age at case death of overall risk by cumulative burden of illnesses. Of recorded mean (SD) 41.6 (1.54) 41.1 (1.11) psychiatric/cognitive risk factors, only record of suicidal ideation 20–29 1 3.7 2 3.7 was statistically significant, however, number of diagnoses of 30–39 13 48.1 26 48.1 psychiatric/cognitive risk factors (#2/.2 risk factors) was 40–49 8 29.6 16 29.6 prognostic in univariate models. This variable may also more 50–59 5 18.5 10 18.5 accurately reflect psychiatric/cognitive status by mitigating specific omissions or oversights as logged in and obtained from medical Year of Death records. 1999–2002 11 40.7 –– This analysis found a recent CD4 cell count of $500 cells/mLto 2003–2005 7 25.9 ––be a significant predictor of reduced risk of death by suicide, 2006–2008 5 18.5 ––violence or accident. This remained significant in multivariate 2009–2011 4 14.8 ––models after adjustment for socio-demographic factors (employ- Mode of Exposure ment and accommodation status). The association of increased CD4 cell count and reduced risk has been demonstrated by Keiser MSM 25 92.6 50 92.6 et al [24], although that analysis did not adjust for other risk MSM/IDU 2 7.4 4 7.4 factors. Our study uses average recent CD4 cell count which
1. Controls were also matched to cases by treating clinic and were HIV-positive reduces the influence of anomalous or non-characteristic mea- prior to case death. surements compared to single test values. doi:10.1371/journal.pone.0089089.t001 Models of suicide risk in HIV-positive populations generally incorporate complex causal pathways. In particular, psychosocial employment (FTE) (‘‘Y’’/’’N’’), recorded recent accommodation risk factors have been shown to predispose patients to non- status of alone/institution (‘‘Y’’/N’’), psychiatric/cognitive risk adherence [25,26] hence poorer HIV disease control. There are factor count (‘‘#2’’/’’.2’’), year HIV positive (‘‘,2000’’/ also reciprocal causal effects of immunological status on severity of ’’$2000’’) and duration weighted average CD4 cell count neuropsychiatric symptoms as demonstrated by Warriner et al (‘‘,500’’/’’$500’’/’’Missing’’). Record of suicidal ideation was [27]. In our analysis there was no evidence of strong interaction excluded from model fitting because of collinearity associated with between covariates, although robustness of multivariate models prior inclusion of psychiatric/cognitive risk factor count. The was limited by statistical power. In these models reduced level of variables retained in the final model were average recent CD4 immunological recovery can be seen to add to the overall risk (p = 0.013), recent full time employment (p = 0.021), recent associated with just psychosocial factors and may also indicate accommodation status of alone/institution (p = 0.005) and year comorbidity effects specifically associated with HIV-disease. HIV positive (p = 0.004) (Table 3). We found earlier calendar year of HIV diagnosis to be associated with increased risk. This study only looks at post cART Suicide-only death era mortality, when improved prognosis has been associated with Low numbers of endpoints prevented detailed examination of some alleviation of psychosocial burden and reduced suicide risk suicide-only death, particularly for non-binary categorical vari- [24]. However, increased all-cause mortality in HIV populations ables with unreported/missing values. Models were otherwise has been shown to be mainly attributable to risk factors identifiable prior to, or at early stages of cART [25]. While our results in part qualitatively similar to those developed for the primary analysis reflect increased experience of suboptimal therapy amongst high (results not shown). Record of suicidal ideation ever was the only risk groups, they also suggest that extended duration of infection significant predictor of risk of suicide (OR 7.88, 95% CI 1.44- increases risk as has been observed in other chronic medical 43.06, p = 0.020). conditions [28]. Posited mechanisms for suicide risk in HIV-positive populations include the extent of neuro-protection and neuro-toxicity of ART
PLOS ONE | www.plosone.org 4 February 2014 | Volume 9 | Issue 2 | e89089 Page 103 of 127 Suicide and Accidental or Violent Death in AHOD
Table 2. Patient characteristics prior to case death and risk of suicide and accidental/violent death1.
Case Control
n=27 % n=54 % OR 95% CI p p2
Full time employment3 Yes 6 22.2 33 61.1 1.00 - - 0.002 No 16 59.3 19 35.2 5.86 [1.69,20.37] 0.005 Not stated (NS) 5 18.5 2 3.7 28.54 [3.12,261.23] 0.003 Accommodation3 Not alone 7 25.9 23 42.6 1.00 - - 0.021 Alone/Institution 15 55.6 14 25.9 3.26 [1.06,10.07] 0.040 Not stated 5 18.5 17 31.5 0.63 [0.20,1.96] 0.425 Alcohol intake3 Moderate or less 13 48.1 36 66.7 1.00 - - 0.397 High 5 18.5 6 11.1 2.31 [0.46, 11.54] 0.308 NS 9 33.3 12 22.2 1.72 [0.61,4.91] 0.196 Smoking3 Never 11 40.7 18 33.3 1.00 - - 0.284 Prior 2 7.4 10 18.5 0.33 [0.08,1.42] 0.136 Current 12 44.4 25 46.3 0.76 [0.20,2.87] 0.681 Not stated 2 7.4 1 1.9 3.50 [0.25,49.36] 0.354 Recreational/Illicit drug use ever No 4 14.8 4 7.4 Non IDU 10 37 11 20.4 IDU 3 11.1 5 9.3 Not stated 10 37 34 63 Recreational/Illicit drug use3 No 21 77.8 47 87 1.00 - - Yes 6 22.2 7 13 2.35 [0.64,8.70] 0.200 Suicide attempts No 21 77.8 48 88.9 1.00 - - Yes 6 22.2 6 11.1 2.73 [0.70,10.69] 0.149 Suicidal ideation No 12 44.4 43 79.6 1.00 - - Yes 15 55.6 11 20.4 6.55 [1.70,25.21] 0.006 Depression/anxiety No 9 33.3 27 50.0 1.00 - - Yes 18 66.7 27 50.0 1.87 [0.67,5.25] 0.234 Neurocognitive impairment No 24 88.9 47 87.0 1.00 - - Yes 3 11.1 7 13.0 0.84 [0.22,3.26] 0.804 Chronic pain No 21 77.8 49 90.7 1.00 - - Yes 6 22.2 5 9.3 2.68 [0.74,9.68] 0.132 Psych/cognitive risks #21970.4 48 88.9 1.00 - - .2829.6 6 11.1 4.99 [1.17,30.65] 0.031 Recent psych/cognitive risks3, 4 #22488.9 50 92.6 1.00 - - .2311.1 4 7.4 1.62 [0.23,11.30] 0.628 Psychotropic medications #21763 42 77.8 1.00 - - .21037 12 22.2 1.76 [0.61,5.08] 0.297 Recent psychotropic medications3 #21451.9 42 77.8 1.00 - - .21348.1 12 22.2 2.47 [0.95,6.41] 0.062 AIDS defining illness No 25 92.6 42 77.8 1.00 - - Yes 2 7.4 12 22.2 0.29 [0.06,1.34] 0.112 AIDS dementia complex No 27 100 51 94.4 1.00 - - Yes 0 0 3 5.6 Does not -- converge NeurocART5, 6 No 16 59.3 38 70.4 1.00 - - Yes 11 40.7 16 29.6 1.74 [0.68,4.48] 0.251 CD4 cell count (cells/mL)6 ,500 17 63.0 26 48.1 1.00 - - 0.022 $500 5 18.5 25 46.3 0.25 [0.07,0.87] 0.030 Missing 5 18.5 3 5.6 1.68 [0.33, 8.48] 0.533 HIV viral load (copies/ml)6 #400 11 40.7 30 55.6 1.00 - - 0.291
PLOS ONE | www.plosone.org 5 February 2014 | Volume 9 | Issue 2 | e89089 Page 104 of 127 Suicide and Accidental or Violent Death in AHOD
Table 2. Cont.
Case Control
n=27 % n=54 % OR 95% CI p p2
.400 11 40.7 21 38.9 1.63 [0.61,4.36] 0.368 Missing 5 18.5 3 5.6 3.91 [0.69,22.2] 0.124 First HIV positive result ,1990 12 44.4 12 22.2 1.00 - - 0.04 1990-1999 10 37 28 51.9 0.31 [0.11,0.89] 0.030 $2000 5 18.5 14 25.9 0.08 [0.01,0.84] 0.035 Efavirenz No 20 74.1 49 90.7 1.00 - - Yes 7 25.9 5 9.3 3.78 [0.96,14.87] 0.057 Nevirapine No 23 85.2 36 66.7 1.00 - - Yes 4 14.8 18 33.3 0.37 [0.12,1.10] 0.073 Abacavir No 25 92.6 42 77.8 1.00 - - Yes 2 7.4 12 22.2 0.27 [0.05,1.38] 0.115 Stavudine No 23 85.2 46 85.2 1.00 - - Yes 4 14.8 8 14.8 1.00 [0.24,4.11] 1.000 Lamivudine No 17 63 28 51.9 1.00 - - Yes 10 37 26 48.1 0.53 [0.17,1.72] 0.293 Ritonavir No 22 81.5 40 74.1 1.00 - - Yes 5 18.5 14 25.9 0.66 [0.20, 2.14] 0.488
1. Controls matched to cases by treating clinic, sex, age (10 year age category at case death), mode of exposure, HIV-positive at case death. Univariate conditional logistic regression used to determine OR. 2. Overall Wald test for categorical variables. 3. Record in year prior to case death. 4. Count of recorded suicidal ideation, depression/anxiety, dysthymia, mood and personality disorders, anorexia nervosa, schizophrenia, neurocognitive impairment. 5. NeurocART regimens are those with CPE greater study median per calendar year of case death. 6. Duration weighted average in year prior to case death. doi:10.1371/journal.pone.0089089.t002 although this was not supported by this analysis. In this study the antiretrovirals were associated with risk of suicide or death by recent use of neurocART was not significantly associated with risk violence or accident. in univariate analyses. Generally, there was no association between In this analysis all observed cases had reported mode of risk and neurocognitive impairment although increased observa- exposure as ‘‘MSM’’ or ‘‘MSM/IDU’’. This reflects the compo- tions are required to support these findings. Similarly, no specific sition of the AHOD cohort (.75% of exposures were via these modes) and of the wider epidemic in Australia, rather than
Table 3. Multivariate conditional logistic regression of determinants of suicide and accidental/violent death1.
Case Control
n=27 % n=54 % OR 95% CI PP2
Full time employment3 Yes 6 22.2 33 61.1 1.00 - - No/NS 21 77.8 21 38.9 14.25 (1.49, 136.17) 0.021 Accommodation (alone/institution)3 No/NS 12 44.4 40 74.1 1.00 - - Yes 15 55.6 14 25.9 4.66 (1.59, 13.68) 0.005 CD4 cell count (cells/mL)4 ,500 17 63.0 25 46.3 1.00 - - 0.013 $500 5 18.5 24 44.4 0.15 (0.03, 0.70) 0.016 Missing 5 18.5 5 9.3 2.40 (0.37, 15.44) 0.358 First HIV positive result ,2000 22 81.5 40 74.1 1.00 - - $2000 5 18.5 14 25.9 0.07 (0.01, 0.41) 0.004
P.chi2 = 0.019, Pseudo R2 = 0.467. 1. Controls matched to cases by treating clinic, sex, age (10 year age category at case death), mode of exposure, HIV-positive at case death. 2. Overall Wald test for categorical variables. 3. Record in year prior to case death. 4. Duration weighted average in year prior to case death. doi:10.1371/journal.pone.0089089.t003
PLOS ONE | www.plosone.org 6 February 2014 | Volume 9 | Issue 2 | e89089 Page 105 of 127 Suicide and Accidental or Violent Death in AHOD necessarily the relative risk compared to other modes of exposure. violent death after adjustment for psychosocial factors. Number of In this study, cases and controls were matched on ‘‘MSM’’ or psychiatric/cognitive diagnoses contributed to the level of risk but ‘‘MSM/IDU’’ mode of exposure and analyses therefore exclude many psychosocial factors were not individually significant confounding effects on risk arising from other modes of exposure predictors of risk. These findings indicate a complex interplay of in controls. factors associated with risk of suicide and accidental or violent Generally, for case-control studies estimated relative risks are death. influenced by any causal association with case matching variables. Age, gender, MSM status and IDU history, which were categories Acknowledgments used for case matching, have well documented association with the prevalence of general risk factors such as recent drug use, alcohol Disclaimer: The views expressed in this publication do not necessarily consumption, smoking status and psychiatric and cognitive risks. represent the position of the Australian Government. The members of the Australian HIV Observational Database are: Given relatively low case numbers, this was an efficient way to New South Wales: D Ellis, General Medical Practice, Coffs Harbour; M conduct adjusted analyses of epidemic specific and less strongly Bloch, T Franic*, S Agrawal, L McCann, N Cunningham, T Vincent, associated predictors such as HIV related factors. Our results Holdsworth House General Practice, Darlinghurst; D Allen, JL Little, however, do not reflect risk associated with different categories of Holden Street Clinic, Gosford; D Smith, C Gray, Lismore Sexual Health & case matching variables. AIDS Services, Lismore; D Baker*, R Vale, East Sydney Doctors, Surry This study has a number of limitations. First, primary analyses Hills; DJ Templeton*, CC O’Connor, C Dijanosic, RPA Sexual Health were based on suicide and accidental or violent death although Clinic, Camperdown; E Jackson, K McCallum, Blue Mountains Sexual correlates of accidental or violent death potentially differ from Health and HIV Clinic, Katoomba; M Grotowski, S Taylor, Tamworth Sexual Health Service, Tamworth; D Cooper, A Carr, F Lee, K Hesse, K those of suicide only death. In this study, CoDe forms, which Sinn, R Norris, St Vincent’s Hospital, Darlinghurst; R Finlayson, I Prone, record detail of cause of death, facilitated the use of this endpoint Taylor Square Private Clinic, Darlinghurst; E Jackson, J Shakeshaft, to capture extra cases with likely similar associated risk behaviours. Nepean Sexual Health and HIV Clinic, Penrith; K Brown, C McGrath, V Specifically, eight of ten accidental or violent deaths were by McGrath, S Halligan, Illawarra Sexual Health Service, Warrawong; L overdose and the associated degree of intentionality could not be Wray, P Read, H Lu, Sydney Sexual Health Centre, Sydney; D Couldwell, determined. While this may represent some loss of specificity, Parramatta Sexual Health Clinic; D Smith,V Furner, Albion Street Centre; there is well documented association between drug use disorders Dubbo Sexual Health Centre, Dubbo; J Watson*, National Association of and completed suicide [29]. Recently Bohnert et al have shown People living with HIV/AIDS; C Lawrence*, National Aboriginal Community Controlled Health Organisation; B Mulhall*, Department of increased likelihood of US medical examiners to classify overdose Public Health and Community Medicine, University of Sydney; M Law*, deaths in cases with substance use disorders as indeterminate K Petoumenos*, S Wright*, H McManus*, C Bendall*, M Boyd*, The intent or unintentional compared to other psychiatric disorders Kirby Institute, University of NSW. and suggested that this might therefore indicate misclassification of Northern Territory: A Kulatunga, P Knibbs, Communicable Disease suicide deaths for this group [30]. It is also possible that an Centre, Royal Darwin Hospital, Darwin. additional proportion of accidental or violent deaths are in fact Queensland: J Chuah*, M Ngieng, B Dickson, Gold Coast Sexual Health true suicide deaths but where official pronouncement of suicide Clinic, Miami; D Russell, S Downing, Cairns Sexual Health Service, might have been discouraged because of moral and legal Cairns; D Sowden, J Broom, K Taing, C Johnston, K McGill, Clinic 87, implications [31]. Specifically in this analysis, deaths are very Sunshine Coast-Wide Bay Health Service District, Nambour; D Orth, D Youds, Gladstone Road Medical Centre, Highgate Hill; M Kelly, A likely due to suicide or to have associated suicidal ideation. The Gibson, H Magon, Brisbane Sexual Health and HIV Service, Brisbane. increase in statistical power associated with the expanded endpoint South Australia: W Donohue,O’Brien Street General Practice, Adelaide. has permitted more detailed analysis than possible by using suicide Victoria: R Moore, S Edwards, R Liddle, P Locke, Northside Clinic, only death and should be considered in further analyses. North Fitzroy; NJ Roth*, J Nicolson*, H Lau, Prahran Market Clinic, Second, there is a low absolute number of observed suicide South Yarra; T Read, J Silvers*, W Zeng, Melbourne Sexual Health deaths in AHOD. In this study we examined the expanded Centre, Melbourne; J Hoy*, K Watson*, M Bryant, S Price, The Alfred endpoint of suicide to include accidental or violent death and used Hospital, Melbourne; I Woolley, M Giles, T Korman, J Williams, Monash efficient statistical methods including a case control design nested Medical Centre, Clayton. on certain accepted prognostic factors. However, results need to be Western Australia: D Nolan, J Skett, J Robinson, Department of Clinical Immunology, Royal Perth Hospital, Perth. carefully interpreted and the chance selection of non-associated Asterisks indicate steering committee members in 2011. variables in models cannot entirely be discounted. However, it is of CoDe reviewers: interest that our findings are consistent with other analyses, as AHOD reviewers: D Sowden, DJ Templeton, J Hoy, L Wray, J Chuah, described [7,19,21–24]. K Morwood, T Read, N Roth, I Woolley, M Kelly, J Broom. Third, there may be site specific differences in comprehensive- TAHOD reviewers: PCK Li, MP Lee, S Vanar, S Faridah, A ness of reporting of certain risk factors. For example, certain Kamarulzaman, JY Choi, B Vannary, R Ditangco, K Tsukada, SH treating centres restrict detailing of illicit drug use in patient Han, S Pujari, A Makane,, OT Ng, AJ Sasisopin. records because of possible legal implications. Omitted data is not Independent reviewers: F Drummond, M Boyd. expected to introduce bias into analyses because of the case- control design of the study which matched cases and controls at Author Contributions the same treating centre. Performed the experiments: HM KP MGL TF MDK JW CCO MJ JH In conclusion, immunological status of HIV-positive patients DAC. Analyzed the data: HM. Wrote the paper: HM KP MGL TF MDK was found to add to the level of risk of suicide and accidental or JW CCO MJ JH DAC.
References 1. Bonnet F, Lewden C, May T, Heripret L, Jougla E, et al. (2004) Malignancy- 2. Chiao EY, Krown SE (2003) Update on non-acquired immunodeficiency related causes of death in human immunodeficiency virus-infected patients in the syndrome-defining malignancies. Curr Opin Oncol 15: 389–397. era of highly active antiretroviral therapy. Cancer 101: 317–324. 3. Petoumenos K, Law MG (2006) Risk factors and causes of death in the Australian HIV Observational Database. Sex Health 3: 103–112.
PLOS ONE | www.plosone.org 7 February 2014 | Volume 9 | Issue 2 | e89089 Page 106 of 127 Suicide and Accidental or Violent Death in AHOD
4. Monforte AD, Abrams D, Pradier C, Weber R, Bonnet F, et al. (2007) HIV- 17. Kowalska JD, Friis-Moller N, Kirk O, Bannister W, Mocroft A, et al. (2011) The induced immunodeficiency and risk of fatal AIDS-defining and non-AIDS Coding Causes of Death in HIV (CoDe) Project: initial results and evaluation of defining malignancies: Results from the D:A:D Study. Abstract No: 84, CROI methodology. Epidemiology 22: 516–523. 5. Lewden C, Salmon D, Morlat P, Bevilacqua S, Jougla E, et al. (2005) Causes of 18. Letendre S, Ellis R, Deutsch R, Clifford D, Collier AC, et al. (2010) Correlates of death among human immunodeficiency virus (HIV)-infected adults in the era of Time-to-Loss-of-Viral-Response in CSF and Plasma in the CHARTER Cohort. potent antiretroviral therapy: emerging role of hepatitis and cancers, persistent 19. Preau M, Bouhnik AD, Peretti-Watel P, Obadia Y, Spire B, et al. (2008) Suicide role of AIDS. Int J Epidemiol 34: 121–130. attempts among people living with HIV in France. Aids Care-Psychological and 6. Keiser O, Spoerri A, Brinkhof MWG, Hasse B, Gayet-Ageron A, et al. (2010) Socio-Medical Aspects of Aids/Hiv 20: 917–924. Suicide in HIV-Infected Individuals and the General Population in Switzerland, 20. Komiti A, Judd F, Grech P, Mijch A, Hoy J, et al. (2001) Suicidal behaviour in 1988-2008. American Journal of Psychiatry 167: 143–150. people with HIV/AIDS: a review. Aust N Z J Psychiatry 35: 747–757. 7. Carrico AW (2010) Elevated suicide rate among HIV-positive persons despite 21. Logan J, Hall J, Karch D (2011) Suicide categories by patterns of known risk benefits of antiretroviral therapy: implications for a stress and coping model of factors: a latent class analysis. Arch Gen Psychiatry 68: 935–941. suicide. Am J Psychiatry 167: 117–119. 22. Nock MK, Kessler RC (2006) Prevalence of and risk factors for suicide attempts 8. Rice BD, Smith RD, Delpech VC (2010) HIV infection and suicide in the era of versus suicide gestures: analysis of the National Comorbidity Survey. J Abnorm HAART in England, Wales and Northern Ireland. AIDS 24: 1795–1797. Psychol 115: 616–623. 9. Smith C, Sabin CA, Lundgren JD, Thiebaut R, Weber R, et al. (2010) Factors 23. Schroecksnadel S, Kurz K, Weiss G, Fuchs D (2012) Immune activation and associated with specific causes of death amongst HIV-positive individuals in the neuropsychiatric symptoms in human immunodeficiency virus type 1 infection. D:A:D Study. AIDS 24: 1537–1548. Neurobehavioral HIV Medicine 4: 1–13. 10. Marin B, Thiebaut R, Bucher HC, Rondeau V, Costagliola D, et al. (2009) Non- 24. Keiser O, Spoerri A, Brinkhof MW, Hasse B, Gayet-Ageron A, et al. (2010) AIDS-defining deaths and immunodeficiency in the era of combination Suicide in HIV-infected individuals and the general population in Switzerland, antiretroviral therapy. AIDS 23: 1743–1753. 1988–2008. Am J Psychiatry 167: 143–150. 11. Schneider E, Whitmore S, Glynn KM, Dominguez K, Mitsch A, et al. (2008) 25. Obel N, Omland LH, Kronborg G, Larsen CS, Pedersen C, et al. (2011) Impact Revised surveillance case definitions for HIV infection among adults, of non-HIV and HIV risk factors on survival in HIV-infected patients on adolescents, and children aged ,18 months and for HIV infection and AIDS HAART: a population-based nationwide cohort study. PLoS One 6: e22698. among children aged 18 months to ,13 years—United States, 2008. MMWR 26. Jia CX, Mehlum L, Qin P (2012) AIDS/HIV infection, comorbid psychiatric Recomm Rep 57: 1–12. illness, and risk for subsequent suicide: a nationwide register linkage study. J Clin 12. Vance DE, Moneyham L, Fordham P, Struzick TC (2008) A model of suicidal Psychiatry 73: 1315–1321. ideation in adults aging with HIV. J Assoc Nurses AIDS Care 19: 375–384. 27. Warriner EM, Rourke SB, Rourke BP, Rubenstein S, Millikin C, et al. (2010) 13. Catalan J, Harding R, Sibley E, Clucas C, Croome N, et al. (2011) HIV Immune activation and neuropsychiatric symptoms in HIV infection. J infection and mental health: suicidal behaviour—systematic review. Psychol Neuropsychiatry Clin Neurosci 22: 321–328. Health Med 16: 588–611. 28. Harris EC, Barraclough BM (1994) Suicide as an Outcome for Medical 14. Gunnell D, Harbord R, Singleton N, Jenkins R, Lewis G (2004) Factors Disorders. Medicine 73: 281–296. influencing the development and amelioration of suicidal thoughts in the general 29. Wilcox HC, Conner KR, Caine ED (2004) Association of alcohol and drug use population. Cohort study. Br J Psychiatry 185: 385–393. disorders and completed suicide: an empirical review of cohort studies. Drug 15. Falster K, Choi JY, Donovan B, Duncombe C, Mulhall B, et al. (2009) AIDS- Alcohol Depend 76 Suppl: S11–19. related and non-AIDS-related mortality in the Asia-Pacific region in the era of 30. Bohnert AS, McCarthy JF, Ignacio RV, Ilgen MA, Eisenberg A, et al. (2013) combination antiretroviral treatment. AIDS 23: 2323–2336. Misclassification of suicide deaths: examining the psychiatric history of overdose 16. Friis-Moller N, Sabin CA, Weber R, d’Arminio Monforte A, El-Sadr WM, et al. decedents. Inj Prev. (2003) Combination antiretroviral therapy and the risk of myocardial infarction. 31. Cantor CH, Neulinger K, De Leo D (1999) Australian suicide trends 1964– N Engl J Med 349: 1993–2003. 1997: youth and beyond? Med J Aust 171: 137–141.
PLOS ONE | www.plosone.org 8 February 2014 | Volume 9 | Issue 2 | e89089 Page 107 of 127 Chapter 7: Conclusions and recommendations In this thesis, long-term survival in HIV+ populations who have initiated treatment is investigated. Results indicate strong and often complex associations between levels of disease control and improved outcome after adjusting for other prognostic factors. By evaluating long-term survival with HIV in Australia, findings presage high level of future burden of disease associated with long-term treatment of ageing HIV+ populations.
7.1. Trends in early stage response of surrogates of treatment outcome We measured trends in surrogates of response to treatment. Specifically, we measured changes in CD4+ and VL following cART commencement, as well as time to first treatment switch following cART commencement. While we observed improved later era CD4+ response, we found that CD4+ measures alone are likely to be insufficient surrogates to differentiate HIV treatment outcomes. Changes in proportions with detectable VL following treatment were clearly different by era. In particular, we observed increasing rates of return to virological control following treatment commencement, and a clear trend towards lower proportions of patients with detectable VL for later eras. This suggests increasing effective ART potency from the suppression of HIV-RNA aspect. We did not observe consistent trends in CD4+ or VL at cART initiation which might be explained by evolving attitudes of patients and prescribers to the risk/benefit of cART initiation. We also observed clear, measureable decrease in time to first line cART switches by later eras of commencement which was attributable to decreased rates of switching with associated viral failure indicating improved regimen durability.
We have found limited publication of comparable measures which differentiate initiation date over durations of treatment for other HIV+ cohorts and population level measures of changes in the effectiveness of cART are seldom reported. We were able to conduct these analyses in AHOD because of high levels of long-term follow-up and ongoing recruitment which are necessary to make these adjustments.
7.2. Long-term survival in HIV-positive patients Overall mortality in adult HIV-positive patients receiving cART was found to be higher than general population level mortality, but mortality rates were close to general population levels in patients with high CD4+ (especially above 500 cells/µL). Results confirm the importance of immunological reconstitution as a biomarker for long-term survival in HIV+ populations and show the similarity between mortality in patients with CD4+ above 350 cells/µL (350-499 cells/µL and ≥500cells/µL) and large increases in mortality below this level. This reflects the findings of recent studies (1, 2) and, given the association between CD4+ levels at
Page 108 of 127 commencement of cART and capacity for immunological recovery, reinforces the notion that CD4+ of 350 cells/µl is an important treatment threshold.
There was no observed association between duration of treatment and survival when adjusted for CD4+ level. At CD4+ less than 350 cells/µL there was a mild decrease in standardise mortality ratios by duration of cART but rates remained much higher than population level, while above this level, average rates remained fairly constant, and just above population level. Although we observed a decrease in SMRs associated with recent cART commencement and year of cART commencement was a significant univariate predictor of survival, it was not a significant covariate in multivariate models. These findings can be contrasted with a recent study by the Antiretroviral Therapy Cohort Collaboration of decreasing mortality by period of cART initiation (1996-99, 2000-02, 2003-05) (3). In that study duration of treatment was associated with decreasing SMRs for each period of cART initiation, but these differences were not adjusted for CD4+ levels, and included earlier periods of follow-up than those used in this study.
A robust assessment of the association between age and survival was conducted. Survival models demonstrate that values of, and decreases in 10 year survival probabilities over the range of ages included in these analyses (up to 70 years old) were similar to those of the Australian general population when adjusting for other factors (maintaining high CD4+, low VL, no prior ADI and non IDU-exposure). We found no evidence of increasing mortality compared to general population rates associated with age. Prognostic indicators in our study are consistent with other studies (4-6).
7.3. Loss to follow-up in the Australian HIV Observational Database Characteristics of patients who were at increased risk of LTFU were consistent with groups experiencing increased VL and exhibiting higher infectiousness but we found no difference between mortality rates in patients according to follow-up status.
The observed rate of LTFU (which included multiple episodes per patient) was relatively low (9.82/100 person years (95% CI: 9.43-10.22)). In the UK Collaborative HIV Cohort (CHIC) study a higher rate of LTFU was observed (16.7/100 person years (95% CI: 16.4 -17.2)) (7). Generally, attendance patterns in this study reflect that LTFU, as defined, is associated with strong departure from recommended and normal therapy. It is likely that many patients are not adherent to treatment for sizeable proportions of these episodes given that HIV prescriptions in Australia are generally made for much shorter intervals and are invalid after one year duration (8, 9).
Page 109 of 127 Demographically, risk of LTFU was associated with males, younger age and mode of exposure (heterosexual and marginally also IDU). These characteristics have been associated with residential transience (10, 11) and are consistent with shorter term engagement with localised healthcare as well as with relatively poor adherence to ART (12- 17) and higher transmission risk behaviours (15, 18-20). This suggests that LTFU events are likely to correlate with increased risk of viral rebound and have serious implications for the HIV epidemic with higher community VL and infectiousness, and consequent ongoing HIV transmission.
We also observed higher risk of LTFU by higher recent VL which suggests that possibly less adherent patients are more likely to become LTFU. In this study there was no difference in risk associated with level of recently tested CD4+, but instead we observed increased risk associated with missing CD4+ tests (and similarly, missing VL tests). Our results suggest that at-risk patients are less likely to engage in a structured or consistent approach to treatment, to the extent that this is reflected by routine of CD4+/VL monitoring.
In this study we found that LTFU status was not a significant predictor of survival. This suggests that additional risk associated with potential disease re-emergence during these episodes is often able to be mitigated. This may be via delayed re-engagement with the same treating centre (as suggested via high rates of episodic LTFU), or via unreported linkage to other health care although this was not investigated by this study. Very few studies in similarly resourced settings have been able to adjust survival estimates for potentially different rates in LTFU although this has often been a stated key objective (7, 21). Importantly, these results indicate that analyses of survival such as those presented in chapter 4 are unlikely to be biased by LTFU patients.
7.4. Determinants of suicide and accidental or violent death in the Australian HIV Observational Database We found increased risk of suicide and accidental or violent death in HIV+ patients, psychosocial factors (employment and accommodation status) and recent immunological status while controlling for age, sex, mode of exposure, treating clinic and calendar year of HIV diagnosis.
As discussed in Chapter 1.4 as well as Chapter 6, in general populations there is well documented association between completed suicide and risk factors such as male gender, lower age, drug use disorders, alcohol use, mental illness, socio economic status and homelessness (22, 23) and many of these general risk factors are more prevalent in HIV+ populations and are associated with route of infection (24-29) including via increased
Page 110 of 127 transmission risk behaviours (30). In this analysis, we found lack of full time employment and living alone or in an institution to be associated with increased risk but many well established risk factors were not individually prognostic (alcohol intake, smoking status, recent recreational/illicit drug use, depression/anxiety, chronic pain and NCI). This is in part attributable to insufficient difference in prevalence between cases and controls for the given statistical power of this analysis. However, results are also consistent with a high prevalence of psychosocial risk factors in HIV+ populations in particular, but also in medically ill populations in general (23, 31-33).
This analysis found recent CD4+ of ≥500 cells/µL to be a significant predictor of reduced risk of death by suicide, violence or accident, after adjustment for socio-demographic factors (employment and accommodation status). The association of increased CD4+ and reduced risk has been demonstrated by Keiser et al (34), although that analysis did not adjust for other risk factors.
We found earlier calendar year of HIV diagnosis to be associated with increased risk. This study only looks at post cART era mortality, when improved prognosis has been associated with some alleviation of psychosocial burden and reduced suicide risk (34). However, increased all-cause mortality in HIV+ populations has been shown to be mainly attributable to risk factors identifiable prior to, or at early stages of cART (35). While our results in part reflect increased experience of suboptimal therapy amongst high risk groups, they also suggest that extended duration of infection increases risk as has been observed in other chronic medical conditions (36).
However, in this study there were low absolute numbers of cases. This means that results, especially for multivariate models, need to be carefully interpreted and the chance selection of non-associated variables in models cannot entirely be discounted.
7.5. Recommendations We contextualised treatment history of HIV+ patients over an extended duration by looking at surrogates of outcome stratified by era of treatment initiation, as well as across sufficient durations of individual patient observation to show often different plateaus of outcome. We propose that this method of assessment of ART effectiveness is a useful summary guide, especially given recent policy focus on community level measures of infectiousness (37). Findings show that there have been incremental improvements for patients initiated on cART, but that patients who have commenced treatment in early eras are likely to have experienced significantly longer durations with detectable VL. The long-term effects of this should be monitored closely, especially given that the HIV+ population is aging
Page 111 of 127 While the cohort used for analyses (AHOD) may be reasonably representative of the Australian HIV+ population who are in care, key subsets of the broader HIV+ population are not included in this analysis. In particular, trends in off-treatment, as well as of in-care but irregular or chaotic treatment groups with likely significantly increased VL are difficult to assess. Given relatively advanced disease stage at cART commencement across all eras, our findings highlight increased relative importance of off-treatment populations. Future intervention and study should therefore be directed at pre-initiation levels of infectiousness, as well as at viral rebound and immunological decline in non-adherent population groups.
Results show that survival can approach that of the general population at high levels of CD4+ and support the importance of CD4+ increase as a treatment priority. As with most HIV+ cohorts, our data had insufficient observations to robustly assess survival at ages over 70. Some studies have accounted for insufficient data at these ages by applying assumed relative rates of mortality (38). We believe there is high risk of error in doing so when comparing survival relative to the general population because at these ages, general mortality rates increase rapidly and can magnify errors in model assumptions. Further, the projection of whole of life experience to recent infections is likely to omit the effects of future developments in HIV specific and general treatments. Instead we estimated 10 year survival probabilities by age and relative to the general population. We believe this is a much more robust comparison as it does not require extensive extrapolation outside the range of the analysis dataset. However, increased data on aged patients are required to more broadly describe survival experience in AHOD, and future studies should draw on increasing numbers of elderly HIV+ patients to accurately assess survival in patients over the age of 70. The LE of HIV+ patients cannot yet be empirically determined, but increased accuracy will be achieved by ongoing analysis of older populations with access to the current and developing array of treatment options.
We in part validated survival estimates by showing similar outcomes in LTFU and non-LTFU patients. This may indicate that there are high levels of unreported linkage to other health care providers or that episodic LTFU has not resulted in critical deterioration in health over the duration of study. This might also in part alleviate concerns about outcome in such patients. However, that patients with higher VL were more likely to become LTFU, suggests that these episodes might be associated with increased infectiousness. Investigation of the extent of unreported engagement with other health care providers as well as of methods to limit LTFU should be conducted to both evaluate and minimise risk of LTFU. As discussed in chapter 5, recorded failure to attend scheduled visits is likely to correlate strongly with at risk patients. Use of this predictor would obviate the use of duration based definitions of LTFU
Page 112 of 127 which may be inappropriate for any given patient specific schedule of attendance and exploration of this definition of LTFU should be explored by further study.
Finally, chronic HIV infection and treatment, is inextricably linked to numerous non-AIDS morbidities. In this thesis we investigated suicide and accidental or violent death in HIV+ populations in particular, and found association between immunological status, year of infection and risk. HIV disease specific factors may perhaps be useful indicators of risk in HIV+ populations, which have generally high background levels of risk and where conventional flags therefore may be less useful at differentiating acute risk. Because of a low number of suicide deaths, we examined the expanded endpoint of suicide to include accidental or violent death, and used efficient statistical methods including a case control design nested on certain accepted prognostic factors. While the chance selection of non- associated variables in models cannot entirely be discounted, it is of interest that our findings are consistent with other analyses albeit of intermediate endpoints (23, 32-34, 39, 40). However, further study with increased endpoints is required to validate these results.
Overall, the prognosis for HIV+ patients was seen to be improving and survival close to that of the general population may be possible. There is some cause for caution however. Firstly, extensive empirical evaluation of aged long-term HIV+ populations is not yet possible because of the age distribution of infected patients, and because of the relatively recent emergence of the disease. Secondly, as survival is prolonged, the availability of cost effective, durable regimens as well as of HIV-specific gerontological care is of increased relative importance. From this aspect, this thesis flags the likely necessity of ongoing treatment development and provision.
Page 113 of 127 7.6. References 1. Lewden C, Chene G, Morlat P, Raffi F, Dupon M, Dellamonica P, et al. HIV-infected adults with a CD4 cell count greater than 500 cells/mm3 on long-term combination antiretroviral therapy reach same mortality rates as the general population. J Acquir Immune Defic Syndr. 2007;46(1):72-7. 2. Lewden C, the Mortality Working Group of COHERE, editors. Time with CD4 Cell Count above 500 cells.mm3 Allows HIV-Infected Men, but not Women, to Reach Similar Mortality Rates to Those of the General Population 17th CROI; 2010; San Francisco. 3. Hogg R, Lima V, Sterne JA, Grabar S, Battegay M, Bonarek M, et al. Life expectancy of individuals on combination antiretroviral therapy in high-income countries: a collaborative analysis of 14 cohort studies. Lancet. 2008;372(9635):293-9. 4. The Australian HIV Observational Database. Rates of combination antiretroviral treatment change in Australia, 1997-2000. HIV Med. 2002;3(1):28-36. 5. Zhou J, Kumarasamy N, Ditangco R, Kamarulzaman A, Lee CK, Li PC, et al. The TREAT Asia HIV Observational Database: baseline and retrospective data. J Acquir Immune Defic Syndr. 2005;38(2):174-9. 6. Baillargeon J, Borucki M, Black SA, Dunn K. Determinants of survival in HIV-positive patients. Int J STD AIDS. 1999;10(1):22-7. 7. Hill T, Bansi L, Sabin C, Phillips A, Dunn D, Anderson J, et al. Data linkage reduces loss to follow-up in an observational HIV cohort study. Journal of clinical epidemiology. 2010;63(10):1101-9. 8. Medicare_Australia. Highly Specialised Drugs (HSD) program: Commonwealth of Australia; 2014 [updated 19 December, 2013; cited 2014 12 May, 2014]. Available from: http://www.medicareaustralia.gov.au/provider/pbs/highly-specialised-drugs/#N100E3. 9. Medicare_Australia. Browse by Section 100 Item List: Commonwealth of Australia; 2014 [cited 2014 12 May, 2014]. Available from: http://pbs.gov.au/browse/section100- hb?initial=a. 10. Pollio DE. The relationship between transience and current life situation in the homeless services-using population. Social Work. 1997;42(6):541-51. 11. Tulloch AD, Fearon P, David AS. Timing, prevalence, determinants and outcomes of homelessness among patients admitted to acute psychiatric wards. Social Psychiatry and Psychiatric Epidemiology. 2012;47(7):1181-91. 12. Chesney MA. Factors affecting adherence to antiretroviral therapy. Clin Infect Dis. 2000;30 Suppl 2:S171-6.
Page 114 of 127 13. Keiser O, Spycher B, Rauch A, Calmy A, Cavassini M, Glass TR, et al. Outcomes of antiretroviral therapy in the Swiss HIV Cohort Study: latent class analysis. AIDS and behavior. 2012;16(2):245-55. 14. Andrews JR, Wood R, Bekker LG, Middelkoop K, Walensky RP. Projecting the benefits of antiretroviral therapy for HIV prevention: the impact of population mobility and linkage to care. J Infect Dis. 2012;206(4):543-51. 15. Rachlis B, Brouwer KC, Mills EJ, Hayes M, Kerr T, Hogg RS. Migration and transmission of blood-borne infections among injection drug users: understanding the epidemiologic bridge. Drug Alcohol Depend. 2007;90(2-3):107-19. 16. South SJ, Crowder KD. Residential mobility between cities and suburbs: race, suburbanization, and back-to-the-city moves. Demography. 1997;34(4):525-38. 17. Wood E, Yip B, Gataric N, Montaner JS, O'Shaughnessy MV, Schechter MT, et al. Determinants of geographic mobility among participants in a population-based HIV/AIDS drug treatment program. Health Place. 2000;6(1):33-40. 18. Madden HC, Phillips-Howard PA, Hargreaves SC, Downing J, Bellis MA, Vivancos R, et al. Access to HIV community services by vulnerable populations: evidence from an enhanced HIV/AIDS surveillance system. AIDS Care. 2011;23(5):542-9. 19. German D, Davey MA, Latkin CA. Residential transience and HIV risk behaviors among injection drug users. AIDS and behavior. 2007;11(6 Suppl):21-30. 20. Strathdee SA, Hallett TB, Bobrova N, Rhodes T, Booth R, Abdool R, et al. HIV and risk environment for injecting drug users: the past, present, and future. Lancet. 2010;376(9737):268-84. 21. Mocroft A, Kirk O, Aldins P, Chies A, Blaxhult A, Chentsova N, et al. Loss to follow- up in an international, multicentre observational study. HIV medicine. 2008;9(5):261-9. 22. Wilcox HC, Conner KR, Caine ED. Association of alcohol and drug use disorders and completed suicide: an empirical review of cohort studies. Drug Alcohol Depend. 2004;76 Suppl:S11-9. 23. Logan J, Hall J, Karch D. Suicide categories by patterns of known risk factors: a latent class analysis. Arch Gen Psychiatry. 2011;68(9):935-41. 24. Carrico AW, Johnson MO, Morin SF, Remien RH, Charlebois ED, Steward WT, et al. Correlates of suicidal ideation among HIV-positive persons. AIDS. 2007;21(9):1199-203. 25. Chikezie UE, Otakpor AN, Kuteyi OB, James BO. Suicidality among individuals with HIV/AIDS in Benin City, Nigeria: a case-control study. AIDS Care. 2012. 26. Jin H, Atkinson JH, Duarte NA, Yu X, Shi C, Riggs PK, et al. Risks and predictors of current suicidality in HIV-infected heroin users in treatment in Yunnan, China: a controlled study. J Acquir Immune Defic Syndr. 2013;62(3):311-6.
Page 115 of 127 27. Journot V, Chene G, De Castro N, Rancinan C, Cassuto JP, Allard C, et al. Use of efavirenz is not associated with a higher risk of depressive disorders: a substudy of the randomized clinical trial ALIZE-ANRS 099. Clin Infect Dis. 2006;42(12):1790-9. 28. Preau M, Bouhnik AD, Peretti-Watel P, Obadia Y, Spire B. Suicide attempts among people living with HIV in France. AIDS Care. 2008;20(8):917-24. 29. Ploderl M, Wagenmakers EJ, Tremblay P, Ramsay R, Kralovec K, Fartacek C, et al. Suicide risk and sexual orientation: a critical review. Arch Sex Behav. 2013;42(5):715-27. 30. Carrico AW, Neilands TB, Johnson MO. Suicidal ideation is associated with HIV transmission risk in men who have sex with men. J Acquir Immune Defic Syndr. 2010;54(4):e3-4. 31. Komiti A, Judd F, Grech P, Mijch A, Hoy J, Lloyd JH, et al. Suicidal behaviour in people with HIV/AIDS: a review. Aust N Z J Psychiatry. 2001;35(6):747-57. 32. Nock MK, Kessler RC. Prevalence of and risk factors for suicide attempts versus suicide gestures: analysis of the National Comorbidity Survey. J Abnorm Psychol. 2006;115(3):616-23. 33. Schroecksnadel S, Kurz K, Weiss G, Fuchs D. Immune activation and neuropsychiatric symptoms in human immunodeficiency virus type 1 infection. Neurobehavioral HIV Medicine. 2012;4:1-13. 34. Keiser O, Spoerri A, Brinkhof MW, Hasse B, Gayet-Ageron A, Tissot F, et al. Suicide in HIV-infected individuals and the general population in Switzerland, 1988-2008. Am J Psychiatry. 2010;167(2):143-50. 35. Obel N, Omland LH, Kronborg G, Larsen CS, Pedersen C, Pedersen G, et al. Impact of non-HIV and HIV risk factors on survival in HIV-infected patients on HAART: a population- based nationwide cohort study. PLoS One. 2011;6(7):e22698. 36. Harris EC, Barraclough BM. Suicide as an Outcome for Medical Disorders. Medicine. 1994;73(6):281-96. 37. Miller WC, Powers KA, Smith MK, Cohen MS. Community viral load as a measure for assessment of HIV treatment as prevention. Lancet Infect Dis. 2013;13(5):459-64. 38. May M, Gompels M, Delpech V, Porter K, Post F, Johnson M, et al. Impact of late diagnosis and treatment on life expectancy in people with HIV-1: UK Collaborative HIV Cohort (UK CHIC) Study. Bmj. 2011;343:d6016. 39. Carrico AW. Elevated suicide rate among HIV-positive persons despite benefits of antiretroviral therapy: implications for a stress and coping model of suicide. Am J Psychiatry. 2010;167(2):117-9. 40. Preau M, Bouhnik AD, Peretti-Watel P, Obadia Y, Spire B, Grp A-E-V. Suicide attempts among people living with HIV in France. Aids Care-Psychological and Socio- Medical Aspects of Aids/Hiv. 2008;20(8):917-24.
Page 116 of 127 Appendix: Suicide in HIV-positive populations literature review
hotic hotic inflicted injury was -
isk factors for remain suicide Noncompliance with medical medical Noncompliance with with associated treatment was psychotic disorders. Suicide attempt/self associated mood, with adjustment, anxiety, personality, and psyc disorders. rates decreased Suicide thesignificantly with they but introduction HAART, of remain the rate above observed general and population, thein r similar. Findings HIV+, Increased in risk particularly recentdiagnosis, intensive and frequent hospitalisation, psychiatric comorbidity cause uncommon an is Suicide rates reported of and death overdoses. by be inflated may
suicide suicide d 2008) 2008) -
HAART -
1995) and and 1995) - HAART (1996 the and eras HIV+ in general population in Switzerland between HIV+between an suicide cause death of Evaluate in HIV+ the association Evaluate of disorders psychiatric among hospitalized the USA in HIV+ women with their lack of adherence to medical treatment suicide and attempt. Evaluate time trends of and predictors thein pre (1988 Objectives associationEvaluate
- - -
Population & HIV & HIV & HIV
HIV+ HIV+ women HIV+ Target HIV+
95,075 35,641 15,275 N 198,937
Prospective cohort Retrospective databaseRetrospective query Multicentre sample Design Multicentre registry query -
- 2004." - 1438. - infected infected - -
L, Qin P. Qin P. L,
21. -
150.
- (10):1315 itle (2010). "Suicide in HIV"Suicide in (2010). general the and individuals Switzerland, 1988 population in 167(2): J Psychiatry 2008." Am 143 diagnoses among HIV diagnoses among 1994 the in USA, women AIDS Care 1432 21(11): al. et O.,Keiser, Spoerri, A. England, Wales Northern England, and 1795 Ireland." 24(11): AIDS 1797. Bansil, P., J. D. Jamieson, et al. (2009). "Trends in with psychiatrichospitalizations nationwide register nationwide linkage Psychiatry. 2012; Clin study. J 73 Rice, D., B. Smith, D. R. et al. "HIV(2010). and infection in HAART the of in suicide era T CX,Jia Mehlum AIDS/HIV infection, comorbid psychiatric illness, riskand for subsequent suicide: a
Page 117 of 127 Title Design N Target Population Objectives Findings Preau, M., A. D. Bouhnik, et al. Survey 2,932 HIV+ Evaluate the prevalence Findings indicate a high level of (2008). "Suicide attempts and characteristics of attempted suicide among HIV+ among people living with HIV in attempted suicide and emphasize the multiple France." AIDS Care 20(8): 917- among a representative roles of factors associated with 924. sample of HIV+ living with HIV, together with socio-demographic factors. Carrico, A. W., M. O. Johnson, Randomized behavioural 2,909 HIV+ Evaluate risk factors Suicidal ideation among HIV+ et al. (2007). "Correlates of prevention trial associated with suicidal is relatively common and suicidal ideation among HIV- ideation in HIV+ associated with multiple factors. positive persons." AIDS 21(9): individuals. 1199-1203. Scheer, S., M. McQuitty, et al. Cross-sectional 1,959 HIV+ & HIV- Evaluate whether AIDS HIV+ decedents were more (2001). "Undiagnosed and surveillance misses a likely to be men, <45 years old, unreported AIDS deaths: results substantial number of and less likely to be from the San Francisco Medical persons who die with Asian/Pacific Islander or Native Examiner." J Acquir Immune unreported AIDS. American. They were more Defic Syndr 27(5): 467-471. likely to have died of suicide (p <.05) or drug abuse/overdose (p <.001). Badiee, J., D. J. Moore, et al. Prospective cohort 1,560 HIV+ Evaluate prevalence of Mood disruption is still (2012). "Lifetime suicidal suicidal ideation and prevalent in HIV+ ideation and attempt are attempt among HIV+ in common among HIV+ the cART era individuals." J Affect Disord 136(3): 993-999. Le Coeur, S., M. Khlat, et al. Register study 1,309 HIV+ & HIV- Evaluate the association HIV prevalence among violent (2008). "Increased HIV infection between violent deaths deaths was 37%, significantly rate among violent deaths: a and HIV in Africa higher than 10% among mortuary study in the Republic accidental deaths
Page of Congo." AIDS 22(13): 1675- 1676. 118 of 127 Title Design N Target Population Objectives Findings Abelson, J., S. Lambevski, et al. Questionnaire 1,300 HIV+ & HIV- Evaluate factors Many risk factors were (2006). "Factors associated with Men associated with feeling associated with suicidality only 'feeling suicidal': the role of suicidal in a large through their association with sexual identity." J Homosex sample of urban men being gay/bisexual. 51(1): 59-80. Lawrence, S. T., J. H. Willig, et Questionnaire 1,216 HIV+ Evaluate suicidal Suicidal ideation was al. (2010). "Routine, self- ideation and associated associated with current administered, touch-screen, conditions using substance abuse and computer-based suicidal computerized patient- depression ideation assessment linked to reported outcome automated response team notification in an HIV primary care setting." Clin Infect Dis 50(8): 1165-1173. May, T., C. Lewden, et al. National survey 964 HIV+ & HIV- Evaluate the causes of Suicide accounted for 11% of (2004). "[Causes and death among HIV- deaths characteristics of death among infected adults in France HIV-1 infected patients with in the year 2000 immunovirologic response to antiretroviral treatment]." Presse Med 33(21): 1487-1492. Sherr, L., F. Lampe, et al. Cross-sectional design 778 HIV+ Evaluate suicidal 31% prevalence of suicidal (2008). "Suicidal ideation in UK ideation in HIV clinic ideation. Independent HIV clinic attenders." AIDS attenders in the UK. predictors of suicidal ideation 22(13): 1651-1658. were very similar among the subgroup of 492 patients who had commenced ART. Page 119 of 127 Title Design N Target Population Objectives Findings Quintana-Ortiz, R. A., M. A. Retrospective cohort 714 HIV+ Evaluate the profile and Prevalence of suicide attempts Gomez, et al. (2008). "Suicide trends of attempted increased over 5 years, from attempts among Puerto Rican suicide in a sample of 9.0% to 22.0%. Men were more men and women with HIV patients in likely to attempt suicide. After HIV/AIDS: a study of Bayamon, Puerto Rico adjusting for age, sex, prevalence and risk factors." HIV/AIDS status and IDU, risk Ethn Dis 18(2 Suppl 2): S2-219- factors were stresses on filial 224. relationships, psychoactive substance use, and isolation. . Gielen, A. C., K. A. McDonnell, Structured interview 611 HIV+ women Evaluate association HIV+ and abused fared worse. et al. (2005). "Suicide risk and between women's HIV+ Abused HIV- had elevated risk mental health indicators: Do serostatus, intimate suggesting abuse is linked to they differ by abuse and HIV partner violence (IPV), negative outcomes. status?" Womens Health Issues and risk of suicide and 15(2): 89-95. other mental health indicators. Peng, E. Y., C. Y. Yeh, et al. Questionnaire 535 HIV+ men Evaluate the prevalence SI was associated with a) (2010). "Prevalence and and correlates of recent psychological distress, correlates of lifetime suicidal suicidal ideation (SI) b) lifetime experience of ideation among HIV-infected among HIV+ inmates. depression, serious anxiety, or male inmates in Taiwan." AIDS tension, or hallucinations. Care 22(10): 1212-1220. Jin, H., J. H. Atkinson, et al. Case Control cross- 406 HIV+ & HIV- IDU Evaluate the frequency Suicidal ideation is frequent in (2013). "Risks and predictors of sectional and predictors of IDUs in China, regardless of current suicidality in HIV- suicidal ideas in HIV+ HIV status, and is not fully infected heroin users in and HIV- heroin IDU accounted for by past treatment in Yunnan, China: a psychiatric history controlled study." J Acquir Immune Defic Syndr 62(3): 311-
Page 120 of 127 316. Title Design N Target Population Objectives Findings Atkinson, J. H., H. Jin, et al. Case control 401 HIV+ & HIV- Evaluate psychiatric and Increased disorders in HIV+. (2011). "Psychiatric context of substance use disorders Onset often preceded date of human immunodeficiency virus of former plasma donors infection. HIV+ had a higher infection among former plasma proportion of lifetime substance donors in rural China." J Affect use diagnoses. Disord 130(3): 421-428. Gaynes, B. N., B. W. Pence, et Structured Interview 400 HIV+ Evaluate the High observed rates of major al. (2012). "Prevalence and epidemiology of depressive disorders (MDD) but predictors of major depression depression in HIV+ low rates of treatment. The in HIV-infected patients on patients on ART number of prior episodes and antiretroviral therapy in HIV symptoms were the Bamenda, a semi-urban center strongest predictors of past- in Cameroon." PLoS One 7(7): year MDD. e41699. Journot, V., G. Chene, et al. Randomized control trial 355 HIV+ Evaluates whether the Depressive disorders were high (2006). "Use of efavirenz is not use of efavirenz is in this population, but disorders associated with a higher risk of associated with the risk were not related to efavirenz depressive disorders: a of depression or suicide. treatment. substudy of the randomized clinical trial ALIZE-ANRS 099." Clin Infect Dis 42(12): 1790- 1799. Grassi, L., D. Mondardini, et al. Case control 295 HIV+ & HIV- Evaluate the prevalence No difference was found (2001). "Suicide probability and and characteristics of between the groups (HIV+) vs psychological morbidity suicide ideation and (HIV-/HCV+) vs (HIV-/HCV-) secondary to HIV infection: a psychological morbidity control study of HIV- associated with HIV and seropositive, hepatitis C virus HCV infection in IDUs (HCV)-seropositive and
Page HIV/HCV-seronegative injecting drug users." J Affect Disord 121 64(2-3): 195-202. of 127 Title Design N Target Population Objectives Findings Gibbie, T. M., A. Mijch, et al. Convenience sample 250 HIV+ men Evaluate the association Negative affect, suicidality and (2012). "High levels of between psychological amphetamine use were psychological distress in MSM distress (PD), HIV associated with PD. There were are independent of HIV status." status and substance no differences in PD across J Health Psychol 17(5): 653- use HIV diagnostic groups 663. Robertson, K., T. D. Parsons, et Cross-sectional design 246 HIV+ Evaluate the prevalence A third of HIV+ reported current al. (2006). "Thoughts of death of death thoughts and suicidal ideation. Suicidal and suicidal ideation in suicidality in HIV ideation did not increase with nonpsychiatric human infection advancing disease. immunodeficiency virus seropositive individuals." Death Stud 30(5): 455-469. Carrico, A. W., T. B. Neilands, Questionnaire 232 HIV+ MSM Evaluate transmission Suicidal ideation may be et al. (2010). "Suicidal ideation risk behaviour in HIV+ associated with HIV is associated with HIV with suicidal ideation transmission risk behaviour transmission risk in men who have sex with men." J Acquir Immune Defic Syndr 54(4): e3- 4. Cooperman, N. A. and J. M. Survey 207 HIV+ women Evaluate prevalence, 26% of the women reported Simoni (2005). "Suicidal timing, and predictors of attempting suicide since ideation and attempted suicide suicidal ideation and diagnosis. AIDS diagnosis, among women living with attempted suicide psychiatric symptoms, and HIV/AIDS." J Behav Med 28(2): physical/sexual abuse were 149-156. predictors of suicidal ideation and attempts. Having children and being employed were also predictors. Spirituality was
Page negatively associated with suicidal ideation. 122 of 127 Title Design N Target Population Objectives Findings Heckman, T. G., J. Miller, et al. Self-administered survey 201 HIV+ Evaluate rates and 38% of HIV+ rural persons had (2002). "Thoughts of suicide predictors of suicidal engaged in thoughts of suicide among HIV-infected rural thoughts among HIV+ during the past week. Risk persons enrolled in a persons living in rural factors were depressive telephone-delivered mental communities symptoms, less coping self- health intervention." Ann Behav efficacy, worry about Med 24(2): 141-148. transmission of HIV, and stress from AIDS-related stigma. Govender, R. D. and L. Questionnaire 190 HIV+ Evaluate suicidal At 72 hours suicidal ideation Schlebusch (2012). "Suicidal ideation in patients who was present in 17.1%, and at 6 ideation in seropositive patients were referred to a HIV weeks in 24.1% seen at a South African HIV counselling and testing voluntary counselling and clinic and who were testing clinic." Afr J Psychiatry found to be seropositive (Johannesbg) 15(2): 94-98. Haller, D. L. and D. R. Miles Medical assessment and 190 HIV+ Evaluate suicidality At-risk individuals, especially (2003). "Suicidal ideation structured interview among participants in an those with "dual disorders," among psychiatric patients with HIV+ mental health unstable interpersonal HIV: psychiatric morbidity and clinic relations, and a restricted social quality of life." AIDS Behav 7(2): environment, should be 101-108. carefully screened for suicidality Schlebusch, L. and R. D. Cross-sectional 189 HIV+ Evaluate prevalence of HIV+ status, young age and Govender (2012). "Age, gender suicidal ideation in those male gender were associated and suicidal ideation following tested for HIV-infection with increased risk. voluntary HIV counseling and and whether HIV+ testing." Int J Environ Res status, age and gender Public Health 9(2): 521-530. was associated with suicidal ideation. Page 123 of 127 Title Design N Target Population Objectives Findings Lau, J. T., X. N. Yu, et al. Structured Interview 176 HIV+ Evaluate the prevalence HIV-related variables and (2010). "Suicidal ideation of suicidal ideation and perceived discrimination and among HIV+ former blood associated factors HIV status of the spouse, were and/or plasma donors in rural among HIV+ who were not associated with suicidal China." AIDS Care 22(8): 946- former blood and/or ideation. 954. plasma donors in a rural county in central China. Goggin, K., M. Sewell, et al. Structured interview 167 HIV+ men Evaluate the prevalence A small number of HIV+ men (2000). "Plans to hasten death and nature of thoughts (17%) reported serious among gay men with HIV/AIDS: and future plans to end thoughts or plans to end their relationship to psychological one's life lives. adjustment." AIDS Care 12(2): 125-136. Capron, D. W., A. Gonzalez, et Cross-sectional 164 HIV+ Evaluate anxiety Cognitive concerns are al. (2012). "Suicidality and sensitivity and cognitive, associated with suicidality in anxiety sensitivity in adults with physical, and social HIV+ after adjustment for HIV." AIDS Patient Care STDS subfactors in relation to demographics, HIV relevant 26(5): 298-303. suicidality in HIV+ factors, and negative affectivity Chikezie, U. E., A. N. Otakpor, Case control 150 HIV+ Evaluate the prevalence Higher rates of suicidal ideation et al. (2012). "Suicidality among of suicidal ideation and in HIV+ than controls. Higher individuals with HIV/AIDS in attempt among HIV+ risk in females, unemployed, Benin City, Nigeria: a case- co-morbidity, living alone and control study." AIDS Care 24(7): having a partner with the 843-845. disease. Roy, A. (2003). "Characteristics Questionnaire 149 HIV+ Evaluate risk factors for Both distal and proximal risk of HIV patients who attempt suicidal behaviour in factors are involved in suicidal suicide." Acta Psychiatr Scand HIV+ behaviour in HIV+ substance 107(1): 41-44. dependent patients. Page 124 of 127 Title Design N Target Population Objectives Findings Kalichman, S. C., T. Heckman, Questionnaire 113 HIV+ elderly Evaluate the prevalence With the exceptions of physical et al. (2000). "Depression and and characteristics of functioning and coping thoughts of suicide among suicidal ideation among strategies, differences between middle-aged and older persons middle-aged and older those who had contemplated living with HIV-AIDS." Psychiatr persons who have HIV suicide and those who had not Serv 51(7): 903-907. infection or AIDS. remained unchanged after controlling for symptoms of depression Schlebusch, L. and N. Vawda Questionnaire 112 HIV+ Evaluate variables HIV-related attempted suicide (2010). "HIV-infection as a self- associated with recently rate of 67.2/100 000 and reported risk factor for diagnosed HIV+ as a increased risk for attempted attempted suicide in South self-reported attempted suicide of 13.33% to 18.87% Africa." Afr J Psychiatry suicide risk factor. were calculated (Johannesbg) 13(4): 280-283. Summers, J., S. Zisook, et al. Case Control cross- 93 HIV+ Evaluate the Women presented with (2004). "Gender, AIDS, and sectional bereavement intensified bereavement, bereavement: a comparison of experience, psychiatric increased generalized anxiety women and men living with morbidity, and suicidality disorder, and suicidality HIV." Death Stud 28(3): 225- in bereaved men and compared to men 241. women living with HIV Pompili, M., A. Pennica, et al. Structured Interview 88 HIV+ Evaluate affective Patients with a poorer HRQoL (2013). "Depression and temperaments in HIV+, were more depressed and at affective temperaments are the impact of increased risk of suicide. These associated with poor health- hopelessness on were also more likely to have related quality of life in patients HRQoL, and depressive affective with HIV infection." J Psychiatr associations among temperaments. Pract 19(2): 109-117. HRQoL, hopelessness, and affective temperaments. Page 125 of 127 Title Design N Target Population Objectives Findings Lewis, C. F. (2005). "Post- Structured Interview 81 HIV+ women Evaluate a sample of HIV+ female inmates with traumatic stress disorder in HIV+ incarcerated PTSD are a complex population HIV-positive incarcerated women with and without who are likely to need careful women." J Am Acad Psychiatry a post-traumatic stress psychiatric assessment, and Law 33(4): 455-464. disorder (PTSD). medical and mental health treatment Lewis, E. L., M. Mosepele, et al. Cross-sectional design 62 HIV+ women Evaluate screening The screening measures used (2012). "Depression in HIV- methods for depression (an inventory of daily activities, positive women in Gaborone, and suicide ideation in and questionnaire of cognitive Botswana." Health Care HIV+ women. function) are useful for Women Int 33(4): 375-386. detecting depression in HIV+ women in resource-limited countries. Malbergier, A. and A. G. de Structured Interview 60 HIV+ & HIV- IDU Evaluate psychiatric HIV+ status was not associated Andrade (2001). "Depressive disorders in HIV+ with depressive disorders and disorders and suicide attempts compared to HIV- IDU suicide attempts in injecting drug users with and without HIV infection." AIDS Care 13(1): 141-150. Borzecki, A., M. Salaga-Pylak, Questionnaire 60 HIV+ Evaluate selected Confirmed the existence of the et al. (2002). "Psycho-social psycho-social problems psycho/social problems in the problems of HIV carriers." Ann among HIV carriers carriers and their families Univ Mariae Curie Sklodowska Med 57(1): 564-568. Stevens, P. E. and E. Longitudinal study 55 HIV+ Women Evaluate the diagnosis There were severe short and Hildebrandt (2006). "Life experiences of HIV+ long term reactions. It usually changing words: women's women took months months and responses to being diagnosed sometimes years before with HIV infection." ANS Adv women could extricate
Page Nurs Sci 29(3): 207-221. themselves from these patterns of response. 126 of 127 Title Design N Target Population Objectives Findings Shelton, A. J., J. Atkinson, et al. Convenience sample 54 HIV+ men Evaluate suicidal 59% of the sample reported (2006). "The prevalence of behaviour in HIV ever thinking about suicide, of suicidal behaviours in a group positive men. which half reported attempting of HIV-positive men." AIDS suicide. White participants had Care 18(6): 574-576. increased risk. Jin, H., J. Hampton Atkinson, et Case Control, cross- 51 HIV+ Evaluate rates of major 79% of HIV+ reported lifetime al. (2006). "Depression and sectional depression and major depression compared to suicidality in HIV/AIDS in suicidality and their 4% of HIV- groups. Most HIV+ China." J Affect Disord 94(1-3): associations with daily developed depression within 6 269-275. functioning in HIV+ and months after testing HIV+. HIV- Dabaghzadeh, F., P. Ghaeli, et Randomized control trial 51 HIV+ Evaluate Cyprohepradine reduces risk in al. (2013). "Cyproheptadine for cyproheptadine HIV+ after initiation of ART prevention of neuropsychiatric prevention of the including efavirenz adverse effects of efavirenz: a neuropsychiatric randomized clinical trial." AIDS adverse effects of an Patient Care STDS 27(3): 146- antiretroviral regimen 154. including efavirenz Kwalombota, M. (2002). "The Questionnaire 45 HIV+ Women Evaluate the mental Women who discover their HIV effect of pregnancy in HIV- health of pregnant HIV+ status during the course of their infected women." AIDS Care women pregnancy are more liable to 14(3): 431-433. develop major depressive illness and somatic disorders. Lavery, J. V., J. Boyle, et al. Questionnaire 32 HIV+ Evaluate why people Desire for euthanasia and (2001). "Origins of the desire for desire euthanasia or assisted suicide was affected euthanasia and assisted suicide assisted suicide by 1) disintegration and 2) loss in people with HIV-1 or AIDS: a of community. These factors qualitative study." Lancet resulted in perceived loss of 358(9279): 362-367. self.
Page 127 of 127