Quality of Life Outcomes of Saquinavir, Zalcitabine

Home , Saquinavir

Antiviral Therapy 4: 35–44 Quality of life outcomes of saquinavir, zalcitabine and combination saquinavir plus zalcitabine therapy for adults with advanced HIV infection with CD4 counts between 50 and 300 cells/mm3

Dennis A Revicki1*, Cassandra Swartz1, Albert W Wu2, Richard Haubrich3 and Ann C Collier4

1MEDTAP International Inc., Bethesda, Md., USA 2Johns Hopkins University, Baltimore, Md., USA 3University of California-San Diego, San Diego, Calif., USA 4University of Washington, Seattle, Wash., USA

*Corresponding author: Tel: +1 301 654 9729; Fax: +1 301 654 9864; E-mail: [email protected]

Background: Benefits in patient health-related quality of demonstrated significantly greater decreases in PHS life (HRQL) have not yet been demonstrated for combi- scores (–4.4±0.6; saquinavir: –1.3±0.6; zalcitabine plus nation antiretroviral therapy with protease inhibitors and saquinavir: –1.7±0.6; P<0.0001) and MHS scores nucleoside analogues. This double-blind study evaluated (–2.2±0.5; saquinavir: –1.0±0.5; zalcitabine plus zalcitabine or saquinavir monotherapy and combination saquinavir: –0.5±0.5; P=0.032) compared to saquinavir saquinavir plus zalcitabine therapy on HRQL of human and zalcitabine plus saquinavir treated patients. No immunodeficiency virus (HIV)-infected adults. differences were observed on the VAS (P=0.172). Nine of Methods: 940 HIV-infected patients (CD4 counts 10 MOS-HIV subscales demonstrated results consistent 50–300 cells/mm3) who had discontinued zidovudine with the primary endpoints. After 48 weeks, a statistically therapy (for intolerance or treatment failure) were significant difference between the saquinavir-treated randomized to one of three regimens: zalcitabine 0.75 groups and the zalcitabine monotherapy group was mg every 8 h; saquinavir 600 mg every 8 h; or combi- observed for PHS scores (zalcitabine: –5.8±0.6; nation zalcitabine 0.75 mg plus saquinavir 600 mg saquinavir: –4.1±0.6; zalcitabine plus saquinavir: every 8 hours. HRQL was measured at baseline, 24 and –3.5±0.6; P=0.014). 48 weeks using the Medical Outcome Study HIV Health Conclusions: Saquinavir monotherapy and combination Survey (MOS-HIV). The primary endpoints were the saquinavir plus zalcitabine demonstrated a benefit in physical and mental health summary scores (PHS; MHS) HRQL relative to zalcitabine monotherapy in patients of the MOS-HIV as well as a global visual analogue with prior zidovudine therapy. The HRQL findings are scale (VAS) score. concordant with improved survival and reduced clinical Results: After 24 weeks, the zalcitabine-treated patients progression of HIV infection found in this study.

Introduction

Human immunodeficiency virus (HIV) infection results efficacy [14,15]. For some patients treated with in a chronic disease with progressive deterioration of zidovudine, the delay in disease progression may be the immune system, frequent opportunistic infections offset by the decrease in HRQL attributable to side- and death. The introduction of reverse transcriptase effects [16]. The quality of life impact of treatments for inhibitors, including zidovudine [1–3], didanosine HIV-infected patients is an increasingly important [4,5] and zalcitabine [6,7], prophylactic therapies consideration in treatment selection [17–19]. against opportunistic infections [8,9] and most recently In the past few years, therapeutic options for HIV the protease inhibitors [10] and combination therapy disease have expanded rapidly; 12 different antiretro- [11] has extended survival in patients with sympto- viral therapies are now approved for use in the US [20]. matic HIV infection and AIDS. Although treatments Antiretroviral therapies in combination, with or can preserve and improve health, they may also have a without protease inhibitors, have demonstrated negative impact on the patient’s health-related quality improvements in laboratory markers of HIV and delay of life (HRQL). For example, zidovudine monotherapy disease progression [21–23]. Collier et al. [22] found has significant toxicity [12,13] and limited duration of that combination saquinavir, zalcitabine and zidovu-

dine therapy resulted in greater impact on CD4 cell regimen. Randomization was stratified by screening counts and plasma HIV RNA compared with CD4 cell count (50 to 99, 100 to 300 cells/mm3). Strata saquinavir plus zidovudine or zidovudine plus were weighted 3:1 with more patients entering the zalcitabine treatment. Combination antiretroviral >100 CD4 cell count group. therapy extends survival benefits to patients with The research protocol was approved by each clinical symptomatic HIV disease and AIDS [24–26]. It is centre’s institutional review board and patients provided assumed that protease inhibitors in combination with written informed consent before entering the study. other antiretroviral therapy have a positive impact on patient functioning and well-being, but few systemati- Treatment regimen cally collected data are available. Ritonavir combined Patients were randomly assigned to one of three treat- with a variety of antiretroviral therapies demonstrated ment groups: (1) zalcitabine 0.75 mg every 8 h; (2) HRQL benefits in patients with advanced AIDS [27]. saquinavir 600 mg every 8 h; or (3) zalcitabine 0.75 However, the ritonavir study did not specify a partic- mg plus saquinavir 600 mg every 8 h. Placebo ular combination therapy regimen. medications were added to the monotherapy groups This study was designed to compare the effect of so that each patient was required to take equivalent saquinavir monotherapy, zalcitabine monotherapy and numbers of tablets or capsules. Patients who devel- combination saquinavir plus zalcitabine therapy on oped an AIDS-defining event were taken off blinded HRQL in HIV-infected patients with CD4 counts therapy and were offered open label saquinavir 600 between 50 and 300 cells/mm3 with previous zidovu- mg three times daily combined with other available dine treatment. The primary clinical efficacy and safety agents. The median time of study follow-up was 12.7 results from this clinical trial have been reported previ- months. The median time spent on blinded protocol ously [24]; this report focuses on the HRQL outcomes. therapy was 11.2 months for the zalcitabine To improve understanding of the clinical importance of monotherapy group, 12.9 months for the saquinavir the quality of life scores, we examined the magnitude monotherapy group and 13 months for the of change in HRQL scores in relation to changes in zalcitabine plus saquinavir group (P=0.088). Patients CD4 count and occurrence of AIDS-defining events. in the HRQL analysis were evaluated while receiving blinded treatment. Therapy was stopped for AIDS- Methods defining events, treatment limiting toxicity and patient or investigator preference. Concomitant Study design and patient population medications were allowed with the exception of other This was a multicentre, three group, double-blind, antiretroviral agents and drugs with toxicities that randomized clinical trial involving advanced HIV- overlap with zalcitabine or other investigational infected patients with CD4 cell counts between 50 and drugs. Prophylactic and chronic maintenance thera- 300 cells/mm3. A total of 940 patients were recruited pies were allowed for Pneumocystis carinii from 48 clinical centres located throughout the US and pneumonia and other opportunistic diseases. Canada. One clinical centre was dropped from the analysis owing to administrative irregularities; patients Clinical endpoints from this centre were excluded from the intent-to-treat The primary clinical endpoint was the first new or analysis of HRQL outcomes. Male or female patients recurrent AIDS-defining event or death that occurred with documented HIV infection were eligible to partic- during follow-up. Events were included in the analysis ipate if they were aged 18 years or older, had if they occurred at least 2 weeks after the start of study discontinued zidovudine (because of intolerance, therapy, whether or not the patient remained on disease progression or patient choice) and had blinded medication. All AIDS-defining events were Karnofsky Performance Status scores ≥60. Patients confirmed by an independent review committee [24]. were excluded if they had had previous treatment with a protease inhibitor, a history or current evidence of HRQL outcome measures peripheral neuropathy, prior treatment with any anti- HRQL is a multidimensional patient outcome which retroviral agent except zidovudine, evidence of includes physical, psychological and social functioning, malabsorption, severe laboratory abnormalities, active and general well-being [28–30]. For this study, the opportunistic infections, unexplained fever, malig- Medical Outcomes Study HIV (MOS-HIV) [31] and a nancy, or non-Hodgkin’s lymphoma, were receiving visual analogue scale (VAS) were selected to assess radiation or antineoplastic therapy or who were preg- HRQL based on their applicability for measuring HRQL nant or breast feeding. in HIV-infected patients, match with study objectives, Patients were required to have stopped zidovudine psychometric characteristics and ease of administration. at least 28 days prior to starting the study treatment The MOS-HIV was developed to assess HRQL in

Table 1. Baseline demographic and clinical characteristics*

Zalcitabine plus saquinavir Characteristic Zalcitabine (n=314) Saquinavir (n=318) (n=308)

Mean (±SD) age, years 38.7±8.2 38.8±8.4 38.4±8.7 Male, n (%) 284 (90) 289 (91) 291 (94) White, n (%) 237 (76) 232 (73) 245 (80) High school graduate, n (%) 283 (92) 295 (95) 274 (91) Mean (±SD) duration of prior zidovudine, months 23±20 22±20 21±19 Mean (±SD) CD4 count cells/mm3, n (%) 177±91 163±81 170±83 3 CD4 count <100 cells/mm , n (%) 71 (23) 80 (25) 77 (25)

Mean (±SD) HIV RNA, log10 copies/ml 5.0±0.6 5.1±0.6 5.0±0.6 Mean (±SD) Karnofsky Performance Status score 94.4±6.6 93.4±7.1 94.3±6.8 *There were no statistically significant differences among the three treatment groups.

patients with HIV infection and contains 10 subscales: tics (for example, CD4 count, Karnofsky performance general health perceptions, physical, social, role and status scores, weight) by treatment group. There were cognitive functioning, pain, energy, mental health, very few missing item-level data (<1%). MOS-HIV health distress and quality of life [31]. For each subscales with less than 50% missing item responses subscale, responses to questions are summed and were treated as missing. scores are converted to a 0 to 100 scale, with 100 indi- The primary outcomes for the HRQL data analysis cating better health status. The subscales have were baseline to 24-week and 48-week endpoint acceptable internal consistency, reliability and validity, changes in PHS, MHS and VAS scores. The study had are able to detect clinically important differences greater than 90% statistical power to detect clinically between patients with asymptomatic and symptomatic meaningful differences in PHS and MHS scores HIV infection, and the subscales are correlated to HIV- [34,37]. Changes in HRQL scores after 24 weeks and related symptoms [19,31–33]. Revicki and colleagues 48 weeks of treatment were identified a priori as [34] developed physical health summary (PHS) and endpoints for the HRQL data analysis to allow the mental health summary (MHS) scores based on the 10 evaluation of both short-term and long-term impact of MOS-HIV subscales. The PHS measures physical func- the treatments on HRQL. The 10 MOS-HIV subscale tioning and activities and pain, and the MHS measures scores were designated secondary HRQL endpoints. mental health and psychological functioning. There is Change scores were constructed at 24 weeks and 48 evidence of reliability and validity of the PHS and weeks for all HRQL measures by subtracting baseline MHS [34]. The PHS and MHS scores were used as the scores. Intent-to-treat principles, with the last observa- primary HRQL outcomes in this study. tion carried forward, were used for the between-group A VAS, adapted from the EuroQol measure [35] was analyses. Only patients with at least one follow-up used to obtain subjective ratings of the patient’s overall HRQL assessment were included in the statistical quality of life. The scale is anchored at 0, the worst analysis. Within-group changes in HRQL scores were imaginable health state, and 100, the best imaginable evaluated using a paired t-test. health state. Higher scores indicate more positive eval- Analysis of covariance (ANCOVA) was used to uations of quality of life. VAS have been used evaluate changes in the three primary HRQL extensively to assess patient outcomes and are easy to endpoints. The ANCOVA models were specified a complete, reliable and valid [32,36]. priori and included terms for the CD4 stratum, treat- The HRQL instruments were administered at base- ment group, geographic region, the treatment line and 24 and 48 weeks after enrolment. All group-stratum interaction, treatment group–geo evaluations were performed within 7 days of the sched- graphic region interaction and the relevant baseline uled follow-up visit. Patients who discontinued HRQL score. Since there were no statistically signifi- prematurely from the clinical trial were administered cant interaction terms in the ANCOVA models the HRQL measures at the time of discontinuation. (P>0.10), these terms were dropped from the final models. Baseline HRQL scores were included because Statistical methodology of differences observed between patients with and Baseline MOS-HIV subscale and summary scores without follow-up HRQL scores. Statistically signifi- among the three treatment groups were compared cant overall between-group differences (P<0.05) were using analysis of variance. We summarized descriptive followed by pair-wise comparisons of treatment data on demographic and selected clinical characteris- groups. Post hoc t-tests of differences in mean change

Antiviral Therapy 4:1 37 D Revicki et al.

Table 2. Baseline and 24-week change in HRQL scores by treatment group Baseline* Change from baseline to 24 week† Zalcitabine Saquinavir Zalcitabine/saquinavir Zalcitabine Saquinavir Zalcitabine/saquinavir Scale mean±SE mean±SE mean±SE mean±SE mean±SE mean±SE P-Value‡ N§ 304 308 295 258 259 251 Primary Outcomes Physical health summary 49.9±0.6 49.6±0.6 50.3±0.6 –4.4±0.6¶||** –1.3±0.6¶|| –1.7±0.6¶** <0.001 Mental health summary 49.7±0.6 49.5±0.6 50.6±0.6 –2.2±0.5¶ –1.0±0.5¶ –0.5±0.5 0.032 Visual analogue scale 81.0±0.9 78.9±0.9 80.6±0.9 –3.6±0.9¶ –1.9±0.9¶ –1.4±0.9 0.172 Secondary Outcomes Physical function 80.9±1.3 79.9±1.3 80.2±1.4 –6.0±1.4¶||** –1.3±1.4|| –1.2±1.4** 0.013 Role function 73.2±2.3 72.9±2.3 72.3±2.3 –13.0±2.3¶ –5.0±2.3¶ –4.1±2.3 0.004 Social function 84.7±1.2 82.3±1.3 86.4±1.1 –8.2±1.5¶|| –3.5±1.5¶|| –4.0±1.5¶ 0.025 Cognitive function 83.5±1.2 82.4±1.2 82.6±1.2 –4.5±1.1¶|| –0.8±1.1|| –1.4±1.1 0.026 General health 55.5±1.3 55.9±1.3 58.6±1.5 –7.8±1.3 –3.8±1.3¶ –1.7±1.3** 0.001 Energy 61.5±1.1 60.6±1.1 63.0±1.2 –6.4±1.1¶|| –2.7±1.1¶|| –2.5±1.1¶ 0.013 Pain 64.3±1.6 64.8±1.6 63.8±1.6 –11.1±1.7¶||** –3.2±1.7|| –5.1±1.7¶** 0.001 Mental health 69.4±1.1 69.0±1.1 68.8±1.1 –3.7±1.0¶ –1.5±1.0 –0.6±1.0 0.050 Health distress 69.8±1.4 70.8±1.3 72.0±1.4 –1.6±1.2 –1.6±1.2 0.8±1.3 0.233 Quality of life 69.5±1.2 67.8±1.1 70.6±1.1 –5.7±1.2¶ –3.7±1.2¶ –0.7±1.2 0.010

*There were no statistically significant differences in baseline mean scores among the three treatment groups. †Least-squares mean from analysis of covariance model with baseline HRQL score, treatment group, region and CD4 strata. Intent-to-treat, last observation carried forward. ‡Two-tailed P-value for treatment group from analysis of covariance model adjusting for baseline health-related quality of life score, region, and CD4 strata. §At baseline, the zalcitabine group sample size ranged from 304–310, the saquinavir group sample size ranged from 308–315 and the zalcitabine plus saquinavir sample size ranged from 295–302 depending on the specific scale. At 24 weeks, the zalcitabine group sample size ranged from 258–273, the saquinavir group sample size ranged from 259–275, and the zalcitabine plus saquinavir group sample size ranged from 251–261 depending on the specific scale. ¶P<0.05, paired t-test for within-group change in HRQL scores. ||P<0.05, t-test between-group zalcitabine and saquinavir. **P<0.05, t-test between-group zalcitabine and zalcitabine plus saquinavir.

between pairs of treatment groups were adjusted for treatment group and 308 in the zalcitabine plus multiple comparisons using the Bonferroni correction saquinavir treatment group. There were no statistically (P<0.017). In the text, 95% confidence intervals (CI) significant differences among the three treatment associated with between-group differences are groups on baseline demographic or clinical characteris- reported. tics (Table 1). The baseline to 24 week and baseline to 48 week changes in secondary HRQL endpoints were also Clinical endpoints analysed using ANCOVA models. Statistically signifi- There were 223 new AIDS-defining events or deaths in cant overall between-group differences were followed the study, 88 with zalcitabine therapy, 84 with by pair-wise comparisons of treatment groups. No saquinavir therapy and 51 with combination adjustments were made for multiple comparisons. A P zalcitabine plus saquinavir therapy. The time to first value of 0.05 was used to assess statistically significant AIDS-defining event or death was prolonged for the differences in the secondary HRQL endpoints. combination therapy group compared to zalcitabine The relationship between changes in CD4 cell count monotherapy (P=0.0006). The time to AIDS-defining and changes in the HRQL measures was examined event or death was not different between the two using ordinary least-squares regression analysis. monotherapy groups (P=0.51). After 12 months of Student’s t-tests were used to compare mean follow-up follow-up, the proportion of patients surviving without HRQL scores between patients experiencing an AIDS- an AIDS-defining event was 0.79 for the zalcitabine defining event and patients not experiencing an group, 0.83 for the saquinavir group and 0.87 for AIDS-defining event during the course of the study. combination zalcitabine plus saquinavir therapy. For the zalcitabine group, the main reasons for Results protocol discontinuation were toxicity (24.9%), AIDS- defining event (15.7%) and patient request (11.1%). Study population For the saquinavir group, the main reasons for Nine hundred and forty patients were randomized to protocol discontinuation were toxicity (16.9%), AIDS- one of the three treatment groups, with 314 in the defining event (11.3%) and patient request (17.1%). zalcitabine treatment group, 318 in the saquinavir For the zalcitabine plus saquinavir group, the main

Table 3. Baseline to 48-week endpoint change in HRQL by treatment group

Baseline to 48-week endpoint change* Zalcitabine Saquinavir Zalcitabine/saquinavir Scale mean±SE mean±SE mean±SE P-value† N‡ 247 252 241 Primary Outcomes Physical health summary –5.8±0.6§|| –4.1±0.6§ –3.5±0.6§|| 0.014 Mental health summary –3.1±0.6§ –2.5±0.6§ –2.1±0.6§ 0.406 Visual analogue scale –5.3±1.0§ –5.1±1.0§ –2.4±1.0§ 0.066 Secondary Outcomes Physical function –7.9±1.6§ –8.4±1.5§ –3.7±1.6§ 0.048 Role function –17.0±2.4§¶ –7.4±2.4§¶ –9.2±2.5§ 0.008 Social function –12.4±1.6§|| –9.8±1.6§ –6.2±1.6§|| 0.014 Cognitive function –5.8±1.2§ –3.4±1.2§ –3.4±1.2§ 0.206 General health –9.6±1.4§ –7.5±1.4§ –6.5±1.4§ 0.208 Energy –8.5±1.3§ –6.8±1.2§ –6.6±1.3§ 0.460 Pain –12.8±1.7§ –6.7±1.6§ –9.2±1.7§ 0.019 Mental health –5.0±1.1§ –3.4±1.1§ –2.6±1.1§ 0.227 Health distress –3.1±1.3§ –3.6±1.3§ –1.2±1.3 0.385 Quality of Life –7.2±1.3§ –5.4±1.3§ –4.7±1.3§ 0.359 *Least-squares mean from analysis of covariance model adjusting for baseline HRQL score, treatment group, region and CD4 strata. Intent-to-treat, last observation carried forward. †Two-tailed P-value from analysis of covariance model adjusting for baseline health-related quality of life score, region and CD4 strata. ‡The zalcitabine group sample size ranged from 247–261, the saquinavir group sample size ranged from 252–268 and the zalcitabine plus saquinavir group sample size ranged from 241–253 depending on the specific scale. §P<0.05, paired t-test for within-group change in health-related quality of life score. ¶P<0.05, t-test between-group zalcitabine plus saquinavir. ||P<0.05, t-test between-group zalcitabine and zalcitabine plus saquinavir. reasons for protocol discontinuation were toxicity cells/mm3 (P=0.011), have higher baseline viral load (15.4%), AIDS-defining event (8.8%) and patient (P=0.010) and lower Karnofsky performance status request (17.6%). More complete clinical efficacy and scores (P=0.0001). For all HRQL scales, the baseline safety data from this clinical trial have been reported scores of patients with missing endpoint HRQL previously [24]. outcomes were significantly worse than those of patients with endpoint scores (P=0.037 to P<0.0001). HRQL Ninety-seven percent of the patients (916 of 940) had Primary HRQL outcomes: 24 weeks baseline HRQL scores and there were no statistically In general, quality of life outcomes declined over time. significant differences in baseline mean HRQL scores For the primary endpoint summary scores (PHS, MHS), between the treatment groups (Table 2). Eighty-two to 24 week scores were significantly lower than baseline 86% (depending on specific subscale) of the patients scores (within-group comparison) except for the had a baseline and one or more follow-up MOS-HIV zalcitabine plus saquinavir group where MHS and VAS scale scores and 768 of 940 patients (82%) had a base- were not significantly lower than baseline (Table 2). line and at least one follow-up PHS or MHS score. Statistically significant differences in 24 week endpoint There were no statistically significant differences changes (between-group comparison) were seen for PHS between the three treatment groups in percentage of scores (P<0.0001) and for MHS scores (P=0.032), but missing endpoint HRQL scores (P=0.501). For the not for VAS scores (P=0.172). Mean PHS and MHS zalcitabine group, missing endpoint HRQL scores were scores by treatment group over time are summarized in more likely owing to toxicity (40.2%) or an AIDS- Figures 1 and 2. The decrease in PHS scores (adjusted) defining event (16.5%), compared with the saquinavir observed in the zalcitabine group (–4.4) was signifi- group (missing endpoint owing to toxicity, 21.9%; cantly larger than the decrease seen in the saquinavir AIDS-defining event, 16.1%) or the zalcitabine plus group (–1.3, P<0.05) and the zalcitabine plus saquinavir saquinavir group (missing endpoint owing to toxicity, group (–1.7, P<0.05). Statistically significant differences 26.7%; AIDS-defining event, 9.3%) (P=0.007). For all were seen between the saquinavir and zalcitabine groups groups, patients completing protocol therapy were in PHS scores (unadjusted differences 2.97, 95% CI: more likely to have endpoint HRQL outcomes. 1.17, 4.73, P<0.05). Significant differences (unadjusted) The demographic and clinical characteristics of of 2.35 points were observed between the zalcitabine patients with and without endpoint HRQL scores were plus saquinavir and zalcitabine groups (95% CI: 0.57, comparable, except for CD4 cell count and viral load. 4.14, P<0.05). The decrease in MHS scores (adjusted) Patients with missing endpoint HRQL outcomes were were larger in the zalcitabine group (–2.2) compared more likely to have CD4 counts less than 100 with the saquinavir group (–1.0) or the zalcitabine plus

Antiviral Therapy 4:1 39 D Revicki et al.

Figure 1. Mean physical health summary (PHS) scores by tion therapy group (P<0.05) but not compared to the treatment group over 48 weeks saquinavir monotherapy group (P>0.05). Statistically significant differences of 2.10 points (unadjusted) were 55 observed between the zalcitabine plus saquinavir and zalcitabine groups on PHS scores (95% CI: 0.13, 4.07, P<0.05). This lower rate of change in mean PHS scores 50 over 48 weeks for the zalcitabine plus saquinavir group is depicted in Figure 1. No statistically significant between-group differences in changes in MHS scores 45 Zalcitabine (P=0.406), or VAS scores (P=0.066) were observed after

Physical health summary (PHS) score Saquinavir Saquinavir/zalcitabine 48 weeks of follow-up.

40 Secondary HRQL outcomes: 48 weeks 0 24 48 Time (weeks) Decreases in MOS-HIV subscale scores were observed in all three treatment groups (Table 3). Statistically significant between-group differences were seen for physical function (P=0.048), role function (P=0.008), social func- saquinavir group (–0.5), but this was not statistically tion (P=0.014), and pain subscale scores (P=0.019) after significant (P>0.05). 48 weeks of follow-up. Pair-wise group comparisons demonstrated that the saquinavir monotherapy group Secondary HRQL outcomes: 24 weeks reported smaller decreases (–7.4±2.4) in role function Individual secondary endpoint scores were signifi- scores compared with the zalcitabine monotherapy cantly lower in nine of 10 subscales in the zalcitabine group (–17.0±2.4, P<0.05). The saquinavir plus group, five of 10 subscales in the saquinavir group and zalcitabine combination therapy group had smaller three of 10 in the combination zalcitabine plus decreases in social function scores (–6.2±1.6) compared saquinavir group (Table 2). Statistically significant with the zalcitabine treatment group (–12.4±1.6, differences between the three treatment groups were P<0.05). For all the remaining subscales, the zalcitabine observed among baseline to 24 week changes in nine of treatment group reported numerically larger decreases the 10 MOS-HIV scale scores (P=0.05 to P<0.001). compared to the saquinavir-treated groups (Table 3). The zalcitabine monotherapy group reported significantly larger decreases in physical function and pain Relationship between CD4 count, clinical events and scores compared with both the saquinavir and HRQL zalcitabine plus saquinavir treatment groups. The Statistically significant relationships were observed saquinavir monotherapy group had significantly between baseline to 24 week changes in CD4 count smaller decreases in physical function, social function, and changes in the PHS (r=0.10, P=0.016), MHS cognitive function, pain, and energy scores compared scores (r=0.12, P =0.002), and the VAS (r=0.16, to the zalcitabine monotherapy group (all P<0.05). The P=0.0001). Changes in CD4 count were also signifi- zalcitabine plus saquinavir treatment group showed cantly associated with changes in physical function significantly smaller changes in role function, general (r=0.13, P=0.0004), general health perceptions health perceptions, and pain scores compared to the (r=0.13, P=0.0005), energy (r=0.12, P=0.002), health zalcitabine monotherapy group (all P<0.05). No statis- distress (r=0.13, P=0.0005) and overall ratings of tically significant differences in changes in MOS-HIV quality of life (r=0.08, P=0.037). A decrease of 100 subscale scores were observed between the saquinavir CD4 cells/mm3 was associated with a decrease of 2 monotherapy and the zalcitabine plus saquinavir points in the physical health summary score and combination therapy groups. mental health summary score, a decrease of 4 points in the MOS-HIV physical function scale and a decrease of Primary HRQL outcomes: 48 weeks 3.5 points in the energy scale. Changes in log RNA Statistically significant decreases in scores from baseline were correlated with changes in PHS scores (r=0.10, to 48 weeks (within group comparisons) were observed P=0.059) and VAS scores (r=0.10, P=0.008). for the three primary quality of life endpoints (Table 3). At the 24 week follow-up, patients who experienced Between-group differences in changes in PHS scores an AIDS-defining event over the course of the clinical persisted after 48 weeks of treatment (P=0.014). Pair- trial reported statistically significantly lower PHS wise comparisons show that the zalcitabine (P=0.0001), MHS (P=0.0001) and VAS scores monotherapy group had larger decreases in PHS scores (P=0.0001) compared to patients not experiencing a compared to the zalcitabine plus saquinavir combina- clinical endpoint. Mean (±SD) endpoint PHS scores

Figure 2. Mean mental health summary (MHS) scores by treatment group over 48 weeks this study extend those observed by Testa et al. [41] on the impact of saquinavir combined with zidovudine or

55 zidovudine plus zalcitabine in advanced HIV-infected patients. Both the current study and the study by Testa et al. [41] demonstrated a significant impact of 50 saquinavir-containing regimens over 20–24 weeks on measures of physical functioning and activities. The saquinavir monotherapy and zalcitabine plus 45 Saquinavir/zalcitabine saquinavir combination therapy were comparable in Zalcitabine Saquinavir impact of HRQL outcomes over the initial 24 weeks of Mental health summary (MHS) score the study. 40 0 24 48 The effects of saquinavir on HRQL were attenuated Time (weeks) after 48 weeks of follow-up, but significant differences favouring the saquinavir monotherapy and saquinavir and zalcitabine combination therapy remained on the were 41.8±11.5 for patients with a clinical event PHS measure. These longer-term effects on HRQL compared to 48.5±10.9 for those without a clinical were most apparent for saquinavir plus zalcitabine event. Mean (±SD) endpoint MHS scores were 44.9±9.8 combination therapy. Saquinavir plus zalcitabine and 49.4±10.2 for patients with a clinical event and for combination therapy showed greater impact on clinical patients without a clinical event, respectively. endpoints and survival compared to the monotherapy Statistically significant differences favouring patients groups [24]. Combination therapy with saquinavir has without AIDS-defining clinical events were observed a beneficial impact on both survival and HRQL on the physical function (mean difference 9.1 points, outcomes in advanced HIV infection. P=0.0007), energy (mean difference 13.2 points, The HRQL findings are consistent with a recently P=0.0001), general health perceptions (mean differ- reported clinical trial of ritonavir combined with two ence 14.0 points, P=0.0001), social function (mean or more antiretroviral therapies in patients with difference 16.0 points, P=0.0001), role function (mean advanced HIV infection [27]. As with the saquinavir difference 16.2 points, P=0.002), body pain (mean combination therapy, the greatest benefits were difference 10.8 points, P=0.004) and overall ratings of observed for ritonavir combination therapy on quality of life (mean difference 10.0 points, P=0.0001). measures of physical functioning and well-being. Ritonavir-treated patients demonstrated smaller decre- Discussion ments in PHS scores and measures of physical functioning, vitality and health status than the placebo- This randomized clinical trial evaluated the impact of treated patients [27]. zalcitabine monotherapy, saquinavir monotherapy, and Previous studies have suggested that a 5 to 10 point zalcitabine and saquinavir combination therapy on change (depending on the particular subscale) in MOS- HRQL outcomes in HIV-infected patients who were based health status subscales is clinically significant intolerant or who had failed zidovudine therapy. The [31,32,37,42]. Recent research on the summary MHS findings demonstrate that, over 24 weeks, patients and PHS scores based on the MOS-HIV indicated that treated with saquinavir or combination saquinavir plus a 3 point change is clinically important [34,37]. In this zalcitabine therapy showed significantly less decline in clinical trial, mean PHS scores decreased more than 4 HRQL than patients treated with zalcitabine. These points by 24 weeks and more than 5 points after 48 HRQL effects persisted over 48 weeks of therapy for weeks of zalcitabine treatment. These decreases were measures of physical functioning, but not for measures only 1.3 and 1.7 points in the saquinavir and of psychological well-being. Saquinavir’s impact on zalcitabine plus saquinavir treatment groups, respec- slowing this decrease in HRQL in advanced HIV infec- tively, at 24 weeks, and 3.5 to 4.1 points at 48 weeks. tion represents an encouraging finding, especially when Therefore, the zalcitabine-treated group demonstrated viewed in conjunction with its effects on prolonging clinically meaningful decreases in mean PHS score by survival in this patient population. 24 weeks and over the 48 week study, while the The zalcitabine-treated patients had decreases in saquinavir-treated patients showed smaller decreases HRQL scores that were greater than those of over the course of the study. saquinavir-treated patients. Bozzette et al. [40], The clinical endpoint study found a significant comparing zidovudine and zalcitabine monotherapy, advantage of combination zalcitabine plus saquinavir demonstrated significant impairments in health status therapy over both monotherapy groups and no differ- and functioning in the zalcitabine group. The results of ences between saquinavir and zalcitabine monotherapy

Antiviral Therapy 4:1 41 D Revicki et al.

[24]. HRQL outcomes reflect both the impact of treat- current study was started (February 1994), treatment ment and disease progression. The short-term quality options for patients intolerant to or failing zidovudine of life benefits of saquinavir alone are in part driven by included switching to either zalcitabine or didanosine the absence of serious side-effects, while the long-term [6]. Triple combination therapies were not yet in benefits of combination therapy are attributable to common use in early 1994. Combination protease lower rates of disease progression. In addition, both inhibitor, sometimes with multiple different protease the combination of zalcitabine and saquinavir and the inhibitors and other antiretroviral therapies are now saquinavir monotherapy groups had reduced inci- the current approach to clinical management. Because dences of toxicities commonly seen with zalcitabine of the development of resistance to monotherapy, three than those treated with zalcitabine alone. The differ- drug combination therapy is now common [20,26]. ences in HRQL scores between the combination Regardless, the primary objectives of the therapeutic therapy and monotherapy groups are attenuated regimen are to reduce viral load to undetectable levels, overall since patients discontinuing the study without maintain immune function and enhance patient endpoint HRQL scores were more severely impaired at survival and quality of life [20]. baseline compared with patients with endpoint HRQL In summary, the beneficial effects on laboratory scores. Therefore, the differences in changes in patient markers of HIV [22,24], clinical events and survival function and well-being between monotherapy and [24] and absence of major toxicities [22,24] support zalcitabine plus saquinavir therapy may actually be the benefits of saquinavir-containing regimens for larger than those observed in this study. patients with advanced HIV infection. Based on these More than 80% of the patients entered into the quality of life findings, the clinical effects are associ- study had at least one HRQL follow-up assessment, ated with significant effects on patient physical although there are missing HRQL endpoints for some functioning and well-being. Saquinavir’s primary patients who experienced an AIDS-defining event. impact on HRQL outcomes occurred during the Missing data attributable to patient death or disease initial six months of therapy, although impact on progression is always a problem for HRQL studies in patient physical functioning was demonstrated over patients with advanced HIV infection or other life- longer time periods for combination therapy. threatening conditions. Missing HRQL data was more Saquinavir therapy had clear benefits for patients likely in patients experiencing an AIDS-defining event with HIV disease who were intolerant of or who had or toxicity. Therefore, the differences between failed zidovudine. In clinical practice, saquinavir is saquinavir combination therapy and zalcitabine generally used in combination with other antiretro- monotherapy may be attenuated, since more viral therapies [20]. Studies are necessary to evaluate zalcitabine patients experienced clinical events or the impact of saquinavir in combination with reverse toxcity. transcriptase inhibitors and other protease inhibitors In this study, we found significant and consistent on clinical events, survival and patient health-related relationships between changes in CD4 cell count and quality of life. changes in the primary HRQL endpoints. Patients who experienced a clinical endpoint during treatment Acknowledgements reported lower PHS and MHS scores at follow-up compared to study patients without clinical endpoints. This research was supported by F Hoffmann-La Patients who had AIDS-defining illnesses had mean Roche, Basel, Switzerland. MOS-HIV subscale scores that were from 3 to 16 points lower than the means of patients without AIDS- References defining endpoints. These findings suggest that the 1. Fischl MA, Richman DD, Grieco MH, Gottlieb MS, occurrence of a clinical endpoint results in a decrement Volberding PA, Laskin OL, Leedom JM, Groopman JE, of about 5 points in MHS scores and a decrement of 7 Mildvan D, Schooley RT et al. The efficacy of azidothymi- dine (AZT) in the treatment of patients with AIDS and points in PHS scores and that differences of this magni- AIDS-related complex. New England Journal of Medicine tude are clinically important. HRQL is impacted by 1987; 317:185–191. both disease progression and treatment toxicity. The 2. Fischl MA, Richman DD, Hansen N, Collier AC, Carey JT, Para MF, Hardy WD, Dolin R, Powderly WG, Allan JD et occurrence of a clinical event was strongly associated al. The safety and efficacy of zidovudine (AZT) in the with impaired functioning and well-being in HIV- treatment of subjects with mildly symptomatic human immunodeficiency virus type 1 (HIV) infection. Annals of infected patients. Internal Medicine 1990; 112:727–737. Recently published guidelines for the clinical 3. Fischl MA, Parker CB, Pettinelli C, Wulfsohn M, Hirsch management of HIV infection [20] do not recommend MS, Collier AC, Antoniskis D, Ho M, Richman DD, Fuchs E et al. A randomized controlled trial of a reduced daily the use of antiretroviral monotherapy, even in patients dose of zidovudine in patients with the acquired discontinuing or intolerant to zidovudine. When the immunodeficiency syndrome. New England Journal of

Medicine 1990; 323:1009–1014. in asymptomatic human immunodeficiency virus infec- 4. Kahn JO, Lagakos SW, Richman DD, Cross A, Pettinelli C, tion. New England Journal of Medicine 1994; Liou SH, Brown M, Volberding PA, Crumpacker CS, Beall 330:738–743. G et al. A controlled trial comparing continued zidovudine 17. de Boer JB, van Dam FS, Sprangers MA, Frissen PH & with didanosine in human immunodeficiency virus infec- Lange JM. Longitudinal study on the quality of life of tion. New England Journal of Medicine 1992; symptomatic HIV-infected patients in a trial of zidovudine 327:582–587. versus zidovudine and interferon-alpha. AIDS 1993; 5. Collier AC, Coombs RW, Fischl MA, Skolnik PR, Northfelt 7:947–953. D, Boutin P, Hooper CJ, Kaplan LD, Volberding PA, Davis 18. Stanton DL, Wu AW, Moore RD, Rucker SC, Piazza MP, LG et al. Combination therapy with zidovudine and Abrams JE & Chaisson RE. Functional status of persons didanosine compared to zidovudine alone in human with HIV infection in an ambulatory setting. Journal of immunodeficiency virus type one infection. Annals of Acquired Immune Deficiency Syndromes 1994; Internal Medicine 1993; 119:786–793. 7:1050–1056. 6. Abrams DI, Goldman AI, Launer C, Korvick JA, Neaton 19. Wu AW, Rubin HR, Mathews WC, Brysk LM, Bozzette JD, Crane LR, Grodesky M, Wakefield S, Muth K, SA, Hardy WD, Atkinson JH, Grant I, Spector SA, Kornegay S et al. A comparative trial of didanosine or McCutchan JA et al. Functional status and well-being in a zalcitabine after treatment with zidovudine in patients with placebo-controlled trial of zidovudine in early symptomatic human immunodeficiency virus infection. New England HIV infection. Journal of Acquired Immune Deficiency Journal of Medicine 1994; 330:657–662. Syndromes 1993; 6:452–458. 7. Fischl MA, Olson RM, Follansbee SE, Lalezari JP, Henry 20. The Department of Health and Human Services Panel on DH, Frame PT, Remick SC, Salgo MP, Lin AH, Nauss- Clinical Practices for the Treatment of HIV Infection. Karol C et al. Zalcitabine compared with zidovudine in Guidelines for use of antiretroviral agents in HIV-infected patients with advanced HIV-1 infection who received adults and adolescents. Annals of Internal Medicine 1998; previous zidovudine therapy. Annals of Internal Medicine 128:1079–1100. 1993; 118:762–769. 21. Hammer SM, Katzenstein DA, Hughes MD, Gundacker H, 8. Public Health Service Task Force on Antipneumocystis Schooley RT, Haubrich RH, Henry WK, Lederman MM, Prophylaxis in Patients with Human Immunodeficiency Phair JP, Niu M, Hirsch MS & Merigan TC. A trial Virus Infection. Recommendations for prophylaxis against comparing nucleoside monotherapy with combination Pneumocystis caranii pneumonia for persons infected with therapy in HIV-infected adults with CD4 cell counts from human immunodeficiency virus. Journal of Acquired 200 to 500 per cubic millimeter. New England Journal of Immune Deficiency Syndromes 1993; 6:46–55. Medicine 1996; 335:1081–1090. 9. Public Health Service Task Force on Prophylaxis and 22. Collier AC, Coombs RW, Schoenfeld DA, Bassett RL, Therapy for Mycobacterium avium complex. Timpone J, Baruch A, Jones M, Facey K, Whitacre C, Recommendations on prophylaxis for disseminated McAuliffe VJ, Friedman HM, Merigan TC, Reichman RC, Mycobacterium avium complex disease in patients infected Hooper C & Corey L. Treatment of human with the human immunodeficiency virus. New England immunodeficiency virus infection with saquinavir, zidovu- Journal of Medicine 1993; 329:898–904. dine and zalcitabine. New England Journal of Medicine 10. Deeks SG, Smith M, Holodniy M & Kahn JO. HIV-1 1996; 334:1011–1017. protease inhibitors: a review for clinicians. Journal of the 23. CAESAR Coordinating Committee. Randomized trial of American Medical Association 1997; 277:145–153. addition of lamivudine or lamivudine plus loviride to 11. Carpenter CC, Fischl MA, Hammer SM, Hirsch MS, zidovudine-containing regimens for patients with HIV-1 Jacobsen DM, Katzenstein DA, Montaner JS, Richman infection: the CAESAR trial. Lancet 1997; 349:1413–1421. DD, Saag MS, Schooley RT, Thompson MA, Vella S, Yeni 24. Haubrich R, Lalezari J, Follansbee FE, Gill MJ, Hirsch M, PG & Volberding PA. Antiretroviral therapy for HIV infec- Richman D, Mildvan D, Burger HU, Beattie D, Donatacci tion in 1997: updated recommendations of the L, Salgo MP and the NV14256 Study Team. Improved International AIDS Society-USA Panel. Journal of the survival and reduced clinical progression in HIV-infected American Medical Association 1997; 277:1962–1969. patients with advanced disease treated with saquinavir plus 12. Richman DD, Fischl MA, Grieco MH, Gottlieb MS, zalcitabine. Antiviral Therapy 1998; 3:33–42. Volberding PA, Laskin OL, Leedom JM, Groopman JE, 25. Cameron DW, Heath-Chiozzi M, Danner S, Cohen C, Mildvan D, Hirsch MS et al. The toxicity of azidothymi- Kravcik S, Maurath C, Sun E, Henry D, Rode R, Potthoff dine (AZT) in the treatment of patients with AIDS and A & Leonard J. Randomised placebo-controlled trial of AIDS-related complex. New England Journal of Medicine ritonavir in advanced HIV-1 disease. Lancet 1998; 1987; 317:192–197. 351:543–549. 13. Volberding PA, Lagakos SW, Koch MA, Pettinelli C, Myers 26. Hammer SM, Squires KE, Hughes MD, Grimes JM, MW, Booth DK, Balfour HH Jr, Reichman RC, Bartlett JA, Demeter LM, Currier JS, Eron JJ, Feinberg JE, Balfour Hirsch MS et al. Zidovudine in asymptomatic human HH, Deyton LR, Chodakewitz JA & Fischl MA. A immunodeficiency virus infection: a controlled trial in controlled trial of two nucleoside analogs plus indinavir in persons with fewer than 500 CD4-positive cells per cubic persons with human immunodeficiency virus infection and millimeter. New England Journal of Medicine 1990; CD4 cell counts of 200 per cubic millimeter or less. New 322:91–949. England Journal of Medicine 1997; 337:725–733. 14. Fischl MA, Richman DD, Causey DM, Grieco MH, Bryson 27. Cohen C, Revicki DA, Nabulsi A, Sarocco PW, Jiang P and Y, Mildvan D, Laskin OL, Groopman JE, Volberding PA, the Advance HIV Disease Ritonavir Study Group. A Schooley RT et al. Prolonged zidovudine therapy in randomized trial of the effect of ritonavir in maintaining patients with AIDS and advanced AIDS-related complex. quality of life in advanced HIV disease. AIDS Journal of the American Medical Association 1989; 1998;12:1495–1502. 262:2405–2410. 28. Testa MA & Simonson DC. Assessment of quality-of-life 15. Volberding PA, Lagakos SW, Grimes JM, Stein DS, Rooney outcomes. New England Journal of Medicine 1996; J, Meng TC, Fischl MA, Collier AC, Phair JP, Hirsch MS et 334:835–840. al. A comparison of immediate with deferred zidovudine 29. Revicki DA. Health care technology assessment and health- therapy for asymptomatic HIV-infected adults with CD4 related quality of life. In Health Care Technology and its cell counts of 500 or more per cubic millimeter. New Assessment: an International Perspective, 1993, England Journal of Medicine 1995; 333:401–407. pp.114–131. Edited by D Banta & BR Luce. New York: 16. Lenderking WR, Gelber RD, Cotton DJ, Cole BF, Oxford University Press. Goldhirsch A, Volberding PA & Testa MA. Evaluation of 30. Testa MA & Lenderking WR. Quality of life considera- the quality of life associated with zidovudine treatment tions in clinical trials of persons with HIV infection. In

Antiviral Therapy 4:1 43 D Revicki et al.

Aids Clinical Trials, 1995, pp. 213–241. Edited by D 37. Wu AW, Revicki DA, Jacobson D & Malitz FE. Evidence Finkelstein & D Schoenfeld. New York: Wiley-Liss. for the reliability, validity and usefulness of the MOS- 31. Wu AW, Rubin HR, Mathews WC, Ware JE Jr, Brysk LT, HIV Health Survey. Quality of Life Research 1997; Hardy WD, Bozzette SA, Spector SA & Richman DD. A 6:481–493. health status questionnaire using 30 items from the 38. Lubeck DP & Fries JF. Health status among persons Medical Outcomes Study: preliminary validation in persons infected with human immunodeficiency virus: a commu- with early HIV infection. Medical Care 1991; 29:786–798. nity-based study. Medical Care 1993; 31:269–276. 32. Revicki DA, Wu AW, Murray M. Change in clinical status, 39. Lubeck DP & Fries JF. Changes in quality of life among health status and health utility outcomes in HIV-infected persons with HIV infection. Quality of Life Research patients. Medical Care 1995; 33:AS173–AS182. 1992; 1:359–366. 33. Burgess A, Dayer M, Catalan J, Hawkins D & Gazzard B. 40. Bozzette SA, Kanouse DE, Berry S & Duan N. Health The reliability and validity of two HIV-specific health- status and function with zidovudine or zalcitabine as related quality of life measures. AIDS 1993; 7:1001–1008. initial therapy for AIDS: randomized controlled trial. 34. Revicki DA, Sorensen S & Wu AW. Reliability and validity Journal of the American Medical Association 1995; 273: of physical and mental health summary scores from the 295–301. MOS HIV Health Survey. Medical Care 1998; 36:126–137. 41. Testa MA, Lenderking WR, Fischer L et al. Effects of 35. Kind P. The EuroQol instrument: an index of health- combination therapy with saquinavir, zidovudine and related quality of life. In Quality of Life and zalcitabine on quality of life. 11th International Conference Pharmacoeconomics in Clinical Trials, 2nd edition, on AIDS. 7–12 July, 1996, Vancouver, BC, Canada. 1996, pp. 191–201. Edited by B Spilker. Philadelphia: Abstract Lippincott-Raven. 42. Ware JE, Snow KK, Kosinski M & Gandek B. SF-36 36. Froberg DG & Kane RL. Methodology for measuring Health Survey. In Manual and Interpretation Guide. health-state preferences: II: scaling methods. Journal of Boston, MA, USA, 1993. The Health Institute, New Clinical Epidemiology 1989; 52:459–471. England Medical Center.

Received 24 August 1998; accepted 14 January 1999