Improving the Diagnosis and Treatment of Smear-Negative Pulmonary And

Stop TB Department Improving the HIV/AIDS Department diagnosis and HIV/AIDS, Tuberculosis and Malaria Cluster treatment of WORLD HEALTH ORGANIZATION 20 Av. Appia, CH-1211 Geneva 27, Switzerland smear-negative Website: http://www.who.int/tb Website: http://www.who.int/hiv pulmonary and extrapulmonary For further information about tuberculosis tuberculosis or other communicable diseases, please contact: among adults and Information Resource Centre HTM/STB World Health Organization adolescents 20 Av. Appia, CH-1211 Geneva 27, Switzerland E-mail: [email protected] Recommendations

For further information about HIV/AIDS, please contact: for HIV-prevalent and Information Resource Centre resource-constrained HIV/AIDS Department, World Health Organization settings 20 av. Appia, CH-1211 Geneva 27, Switzerland E-mail: hiv [email protected] WHO/HTM/TB/2007.379 WHO/HIV/2007.01

Improving the diagnosis and treatment of smear-negative pulmonary and extrapulmonary tuberculosis among adults and adolescents Recommendations for HIV-prevalent and resource-constrained settings

STOP TB DEPARTMENT DEPARTMENT OF HIV/AIDS © World Health Organization 2007 All rights reserved. The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement. The mention of speciﬁ c companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters. All reasonable precautions have been taken by the World Health Organization to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The respon- sibility for the interpretation and use of the material lies with the reader. In no event shall the World Health Organization be liable for damages arising from its use. Contents

Acknowledgements iv Abbreviations v Part I. Improving the diagnosis and treatment of smear-negative tuberculosis 1 Background 3 Target audience 3 Process of formulation 3 Strength of the recommendations 4 Implementation and evaluation 4 Recommendations 5 Algorithms for the diagnosis of smear-negative tuberculosis 8 Part II. Simpliﬁ ed and standardized clinical management guidelines for extrapulmonary tuberculosis 15 Background 17 Target audience 17 Diagnosis and management 17 Further reading 23 References 24 Annex. Protocol for operational evaluation of the revised recommendations and algorithms for improving the diagnosis of tuberculosis in HIV-prevalent settings 27 Background 29 Objectives of the evaluation 29 Purpose of the protocol 29 Hypotheses 30 Study design and procedure 30 References 36

iii Acknowledgements

Prepared by TB (PAG) of KNCV, the German Leprosy and TB Relief Association and the Damien Founda- Members of the WHO Expert Group on Smear- tion have also provided their comments on the Negative TB: Getachew Aderaye (Addis Ababa earlier version. The document was reviewed by University, Ethiopia), Ludwig Apers (Institute of members of the Core Group of the global TB/ Tropical Medicine, Antwerp, Belgium), Leopold HIV Working Group of the Stop TB Partnership Blanc (World Health Organization, Switzerland), and the Strategic and Technical Advisory Group Amy Bloom (United States Agency for Interna- for Tuberculosis (STAG-TB) and the Strategic tional Development (USAID), United States of and Technical Advisory Committee for HIV America), Jermiah Chakaya (Ministry of Health, (STAC-HIV) of the World Health Organiza- Kenya), Liz Corbett (London School of Tropi- tion. cal Medicine and Hygiene, United Kingdom), Haileyesus Getahun (World Health Organiza- Valuable comments were also provided by the tion, Switzerland), Charlie Gilks (World Health following individuals: Raimond Armengol (Pan Organization, Switzerland), Jeroen van Gorkom American Health Organization), Ramzi Asfour (KNCV Tuberculosis Foundation, the Nether- (WHO Headquarters), Daniel Chin (WHO, lands), Mark Harrington (Treatment Action China), Mirtha Del Granado (Pan American Group, United States of America), Pierre-Yves Health Organization), Reuben Granich (Offi ce Norval (World Health Organization, Switzer- of the Global AIDS Coordinator, United States land), Paul Nunn (World Health Organiza- of America), Christy Hanson (USAID, United tion, Switzerland), Rick O’Brien (Foundation States of America), Michael Kimerling (Uni- for Innovative New Diagnostics, Switzerland), versity of Alabama, United States of America), T. Santha (Ministry of Health, India) and Jay Nani Nair (WHO Regional Offi ce for South-East Varma (United States Centers for Disease Con- Asia), Lisa Nelson (Centers for Disease Control, trol and Prevention, Thailand). United States of America), Wilfred Nkhoma (WHO Regional Offi ce for Africa), Pilar Ramon- Pardo (Pan American Health Organization), Acknowledgements Mario Raviglione (WHO Headquarters), Fabio Scano (WHO Headquarters), Akhiro Seita Useful and detailed feedback was obtained on (WHO Regional Offi ce for the Eastern Mediter- an earlier version of the document from more ranean), Sahu Suvanand (WHO, India), Patrick than 130 national tuberculosis and HIV pro- van der Stuyft (Institute of Tropical Medicine, gramme managers, WHO regional and country Belgium), Marco Vitoria (WHO Headquarters), staff, researchers, clinicians, nongovernmental Fraser Wares (WHO, India). organizations and other health workers from all regions through global web-based consultations. All leading international organizations work- Overall coordination ing on tuberculosis, including the International Union Against Tuberculosis and Lung Disease Haileyesus Getahun. (UNION), the Programme Advisory Group for

iv Abbreviations

AFB acid-fast bacillus CPT co-trimoxazole preventive therapy CXR chest X-ray ETB extrapulmonary tuberculosis HIV human immunodeﬁ ciency virus IRIS immune reconstitution inﬂ ammatory syndrome PCP Pneumocystis carinii pneumonia WHO World Health Organization

v PART I Improving the diagnosis and treatment of smear- negative tuberculosis

PART I. IMPROVING THE DIAGNOSIS AND TREATMENT OF SMEAR-NEGATIVE TUBERCULOSIS

Background management of smear-negative pulmonary and extrapulmonary tuberculosis. The recom- Rates of smear-negative pulmonary and mendations and algorithms are designed for extrapulmonary tuberculosis have been rising use by national tuberculosis and HIV/AIDS in countries with HIV epidemics. The mortality control programmes and service providers. rate among HIV-infected tuberculosis patients HIV-prevalent settings are defi ned as countries, is higher than that of noninfected tuberculosis subnational administration units (e.g. districts, patients, particularly for those with smear-neg- counties) or selected facilities (e.g. referral hos- ative pulmonary and extrapulmonary tubercu- pitals, drug rehabilitation centres) where the losis. Delayed diagnosis may be an important adult HIV prevalence rate among pregnant cause of excess mortality in people living with women is ≥1% or HIV prevalence among tuber- HIV who have smear-negative pulmonary and culosis patients is ≥5%. In those countries where extrapulmonary tuberculosis. In the absence national HIV prevalence is below 1%, national of rapid, simple, and accurate diagnostic tools tuberculosis and HIV control authorities should for smear-negative pulmonary and extrapul- identify and defi ne HIV-prevalent settings (sub- monary tuberculosis, diagnostic algorithms national administrative units or facilities) based have been recommended. Earlier algorithms on the epidemiology of the HIV epidemic and and recommendations have been developed the magnitude of HIV-associated tuberculo- through consensus and expert opinion, without sis, and develop appropriate guidance for the a fi rm evidence base. These algorithms extend a implementation of these recommendations. patient’s evaluation over a period of time, dur- The recommendations and revised algorithms ing which HIV-infected patients may die from are intended for immediate implementation in undiagnosed tuberculosis or from advanced sub-Saharan Africa and other HIV-prevalent HIV complications. The Stop TB Strategy now settings, as defi ned by national tuberculosis and emphasizes the timely diagnosis and treatment HIV control authorities, to guide the expedited of all cases of tuberculosis, including smear- diagnosis and management of tuberculosis. negative pulmonary and extrapulmonary tuberculosis. The existing guidelines for diagnosis of smear- Process of formulation negative pulmonary tuberculosis were published In September 2005, WHO convened an expert by WHO in 2003 (1) and codifi ed in 2006 in group to review currently recommended the International standards for tuberculosis care approaches to the diagnosis of smear-negative (2), a publication of organizations, including tuberculosis in HIV-prevalent settings and to WHO, which are members of the Stop TB Part- propose revisions to existing WHO guidelines. nership. The International standards generally The Expert Group has reviewed existing evi- maintained the 2003 WHO recommendations, dence in each of the relevant areas and made but recognize the importance of “fl exibil- recommendations and has revised the existing ity” when applying these guidelines to smear- diagnostic algorithms. The recommendations negative patients who are seriously ill, such as and revised diagnostic algorithms were then patients with HIV infection. It also highlights posted on the WHO Stop TB Department’s the absence of evidence showing how well these web site for an open consultation. Feedback was guidelines perform in HIV-infected patients. obtained from national programme managers, researchers, clinicians and other health workers throughout the world, and from all the leading Target audience international organizations working on tuber- This document is intended for those deal- culosis. The Expert Group subsequently revised ing with tuberculosis and HIV at all levels in the recommendations and algorithms in the HIV-prevalent and resource-constrained set- light of the feedback from the global consultation tings. It is intended to assist development of and from presentations at various international national policies to improve the diagnosis and scientifi c meetings. The Strategic and Technical

3 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

Advisory Group for Tuberculosis (STAG-TB) cost-effectiveness, although the realities of bur- and the Strategic and Advisory Committee for den of disease, human resources, health system HIV (STAC-HIV), the two independent bod- infrastructure and socioeconomic issues need ies that advise WHO on tuberculosis and HIV to be taken into account when adapting these respectively, endorsed the recommendations. recommendations to regional and country programmes. Strength of the recommendations Implementation and evaluation The recommendations contained in these guidelines are based on evidence from randomized In the absence of complete evidence, the rec- clinical trials, high-quality scientifi c studies, ommendations were built on consensus and observational cohort data and, where suffi cient iterative global expert opinion. It is believed evidence is not available, on expert opinion (see that they will provide a reasonable response to Table 1). When appropriate, the level of evi- the catastrophe posed by the dual tuberculosis dence used to formulate the recommendations is and HIV epidemics. These recommendations included in the text of the document and shown should, therefore, be implemented in HIV-prev- in Table 1. The strength of each recommenda- alent settings in order to improve and expedite tion is stated when appropriate, along with the the diagnosis of tuberculosis among people liv- level of evidence, to provide a general indica- ing with HIV. The implementation of the rec- tion of the extent to which regional and country ommendations requires a reasonably effi cient programmes should consider implementing the health system, including quality assurance for recommendations. laboratories and effective supply management and training for programme staff. Moreover, For example, a recommendation marked as A II depending on country-specifi c-factors, it may is a recommendation that should be followed require revision of national guidelines, logistic and is based on evidence from at least one high- and technical arrangements including human quality study or several adequate studies with resources, training and infrastructure develop- clinical, laboratory or programmatic end-points. ment. While the recommendations are being Those recommendations which are based on implemented, it is essential to build up the evi- well established clinical practice are presented as dence base required to assess their effectiveness such, without any indication of the level of evi- and feasibility. Careful evaluations by national dence. For example, the recommendation that authorities, research groups and interested par- calls for an increased level of clinical awareness ties are needed to assess the likely benefi ts and and competence in managing extrapulmonary responsiveness of the recommendations for tuberculosis at fi rst-level health facilities is not the dual tuberculosis and HIV epidemics. The linked with a particular level of evidence. The fi ndings of these evaluations will inform policy recommendations do not explicitly consider

Table 1. Grading of recommendations and levels of evidence

Strength of the recommendations Level of evidence available for the recommendations A. Recommended – should be followed I. At least one randomized controlled trial with clinical, laboratory or programmatic end-points B. Consider – applicable in most situations II. At least one high-quality study or several adequate studies with clinical, laboratory or programmatic end-points C. Optional III. Observational cohort data, one or more case-controlled or analytical studies adequately conducted IV. Expert opinion based on evaluation of other evidence

Sources: adapted from (3), (4), (5), (6).

4 PART I. IMPROVING THE DIAGNOSIS AND TREATMENT OF SMEAR-NEGATIVE TUBERCULOSIS

change designed to improve programme per- • Laboratory confi rmation of HIV infection or formance both globally and nationally. A proto- • Strong clinical evidence of HIV infection1 col that provides generic guidance on evaluation and of the recommendations to improve the diagnosis of tuberculosis in HIV-prevalent settings is • A decision by a clinician to treat with a full annexed to this document. course of antituberculosis chemotherapy. Strength of recommendation: A Recommendations Revised case defi nitions Antibiotics trial The following are suggested case defi nitions for Context: There is limited evidence for the use use in HIV-prevalent settings: of empirical antibiotic treatment to rule out Smear-positive pulmonary tuberculosis tuberculosis as a cause of cough in HIV-infected persons. Although non-response to antibiot- • One sputum smear examination positive for ics increases the likelihood of tuberculosis, the acid-fast bacilli (AFB) and converse is not true; response to antibiotics does • Laboratory confi rmation of HIV infection or not exclude tuberculosis in tuberculosis suspects living in HIV-prevalent settings. Inappropriate • Strong clinical evidence of HIV infection.1 use of broad-spectrum antibiotics may also lead Smear-negative pulmonary tuberculosis to drug resistance, treatment delay and death of patients because of prolonged symptoms. • At least two sputum specimens negative for AFB and Recommendations: • Radiographical abnormalities consistent with • The primary role of antibiotics should not active tuberculosis and be as a diagnostic aid; they should be used to treat concomitant bacterial infection in peo- • Laboratory confi rmation of HIV infection or ple living with HIV/AIDS with cough or seri- • Strong clinical evidence of HIV infection1 ous illness (Strength: A–IV). and • Antibiotic treatment is appropriate for HIV- • Decision by a clinician to treat with a full infected patients with cough, because bac- course of antituberculosis chemotherapy terial infections are common both with and OR without tuberculosis (Strength: A–II). • A patient with AFB smear-negative sputum • Seriously ill patients with symptoms sug- which is culture-positive for Mycobacterium gestive of tuberculosis should be treated tuberculosis. empirically with broad-spectrum antibiotics because the benefi ts outweigh the risks Extrapulmonary tuberculosis (Strength: A–II). • One specimen from an extrapulmonary site • When indicated, one course of broad-spec- culture-positive for Mycobacterium tubercu- trum antibiotics, including coverage for losis or smear-positive for AFB typical and atypical causes of community- OR acquired pneumonia, should be used to reduce the time delay for tuberculosis diagno- • Histological or strong clinical evidence con- sis (Strength: A–IV). In such circumstances, sistent with active extrapulmonary tubercu- fl uoroquinolones should be avoided, as they losis and may cause undue delay in the diagnosis of tuberculosis (Strength: A–II). 1 Depending on clinical assessment and national and/ or local policy, a person of unknown HIV status may • More research about the effectiveness and be classifi ed as HIV-positive for the purposes of diag- use of an antibiotic trial in the diagnostic nosis and management. algorithm and the choice of antibiotics, par-

5 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

ticularly for people living with HIV, is needed Sputum culture (Strength: A). Context: Sputum culture is the gold standard for the diagnosis of tuberculosis. However, Chest radiograph mycobacteria are slow-growing organisms and culture takes several weeks and requires Context: Although chest X-ray abnormalities relatively sophisticated facilities and technical are common in HIV-infected persons without expertise. Sputum culture of HIV-infected indi- tuberculosis, the chest X-ray plays an impor- viduals requires more incubation time than for tant role in the diagnosis of tuberculosis among non-HIV-infected patients, although it is still of people living with HIV. The chest X-ray can also value. There are major challenges to ensuring be an important entry point to diagnosing non- access to high-quality sputum culture in HIV- tubercular chest diseases, which are common prevalent and resource-constrained settings. among people living with HIV. Recommendations: Recommendations: • Careful feasibility studies are needed, partic- • Chest X-ray presentations of tuberculosis in ularly for liquid culture systems that are more HIV patients are now well characterized and sensitive and rapid than solid culture, and should no longer be considered “atypical” have the potential for expanded use, includ- for tuberculosis in HIV-prevalent settings ing in HIV-prevalent and resource-limited (Strength: A–IV). settings (Strength: A–II). • Chest X-rays play a signifi cant role in short- • In patients with negative sputum smears, ening delays in diagnosis and should be per- sputum culture should be encouraged as formed early in the course of investigation of part of the diagnostic procedure for people a tuberculosis suspect (Strength: A–II). living with HIV who are being evaluated for • Sound clinical judgement is needed to put AFB smear-negative tuberculosis, since it will a seriously ill patient with negative sputum improve the quality of care and assist the con- smear results on antituberculosis treatment fi rmation of the diagnosis (Strength: A–I). using only suggestive radiographical fi ndings. • Existing capacity for the use of conven- In such circumstances, the clinical response tional culture systems in countries should of the patient has to be monitored and tuber- be explored, encouraged and strengthened. culosis diagnosis should be confi rmed at least Decentralization of sputum culture services by clinical response to antituberculosis treat- with an effi cient quality assurance system ment and preferably by culture (Strength: is essential. Establishment of an effective B–II). transport system for sputum is also essential • The limitations that exist on the wider use (Strength: A). of chest X-rays, such as nonavailability at peripheral health facilities and the diffi - Immune reconstitution infl ammatory culty of interpreting results, even by trained syndrome (IRIS) and tuberculosis physicians, need to be addressed, including diagnosis through training (Strength: A). Context: Immune recovery usually occurs rap- • Research is needed to identify innovative ways idly in HIV-infected adults who are started on to enhance the ability of clinicians, includ- antiretroviral treatment (ART). Occasionally, ing nonphysicians, to interpret chest X-rays recovery of the immune system leads to clini- accurately, to assess the feasibility and added cal signs and symptoms of active tuberculosis. value of peer reviewing of chest X-rays and to This may be because patients had subclinical evaluate novel imaging techniques that might tuberculosis before the antiretroviral treatment replace conventional radiography (Strength: began, or because a latent tuberculosis infection A). has been reactivated. The condition, which is

6 PART I. IMPROVING THE DIAGNOSIS AND TREATMENT OF SMEAR-NEGATIVE TUBERCULOSIS

known as immune reconstitution infl ammatory Recommendations: syndrome (IRIS), usually occurs within three • There should be an increased level of clini- months of initiation of antiretroviral treatment. cal awareness and competence in managing It can also appear as exacerbation of tuberculosis extrapulmonary tuberculosis at fi rst-level when initiating antiretroviral treatment in HIV- health facilities, including earlier referral of infected tuberculosis patients who are already patients when appropriate (Strength: A). undergoing antituberculosis treatment, similar • In peripheral health facilities in HIV-preva- to the well documented paradoxical reactions lent settings, health care workers should ini- seen in some patients without underlying HIV tiate empirical antituberculosis treatment infection. IRIS is commonly associated with early in patients with serious illness thought tuberculosis, although it can also occur with to be due to extrapulmonary tuberculosis. other pathogens. Every effort should then be made to confi rm the diagnosis of tuberculosis, including mon- Recommendations: itoring the clinical response of the patient, • Tuberculosis should be diagnosed and treated to ensure that the patient’s illness is being before initiation of antiretroviral treatment managed appropriately. If additional diag- and whenever there is clinical suspicion of nostic tests are unavailable, and if referral IRIS (Strength: A–IV). to a higher level facility for confi rmation of • IRIS is not a reason to switch patients on to the diagnosis is not possible, antituberculo- second-line antiretroviral treatment, although sis treatment should be continued and com- adjustment to the treatment regimen may be pleted (Strength: B–IV). needed to ensure compatibility with antitu- • Empirical trials of treatment with incomplete berculosis treatment (Strength: A–IV). regimens of antituberculosis drugs should • Health care workers should be aware of para- not be performed (Strength: A–I). doxical worsening of tuberculosis on starting • If a patient is treated with empirical antituber- antiretroviral treatment, and both antiretro- culosis drugs, treatment should be with stand- viral and antituberculosis treatments should ardized, fi rst-line regimens, which should be be continued (Strength: A–IV). used for the entire duration of antituberculosis treatment. Empirical treatment should Diagnosis of extrapulmonary tuberculosis be stopped only if there is bacteriological, histological or strong clinical evidence of an Context: Extrapulmonary tuberculosis is more alternative diagnosis (Strength: A). strongly HIV-related than pulmonary tuberculosis, with a combination of the two being especially suggestive of underlying HIV-infection. Recording and reporting HIV-related extrapulmonary tuberculosis is a Context: The recording and reporting of smear- WHO clinical stage 4 (advanced AIDS) diag- negative pulmonary and extrapulmonary nosis, and patients with HIV-related extrapul- tuberculosis by national tuberculosis control monary tuberculosis often have disseminated programmes needs strengthening. Information disease and are at high risk of rapid clinical from case-reporting should increasingly be used deterioration and death. The accurate diagno- to inform changes in programme performance. sis of extrapulmonary tuberculosis is complex and diffi cult, particularly in peripheral health Recommendations: facilities with limited support and diagnostic • The 2003 recommendation that cases without infrastructure. Simplifi ed, standardized clinical smear results should be reported as smear- management guidelines for most common and negative pulmonary cases should be revised serious forms of extrapulmonary tuberculosis (Strength: A). are included in this document to assist health care workers at the district hospital level in HIV- • The revised standard tuberculosis recording prevalent settings (see Part II below). and reporting formats should be used to gen-

7 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

erate sound case-notifi cation and treatment sputum produced after an overnight sleep. One outcome data for smear-negative pulmo- positive AFB smear will be suffi cient to classify nary and extrapulmonary cases. This should a patient as a smear-positive case if the patient is inform policy and programme performance HIV-infected or if there is strong clinical suspi- both nationally and globally (Strength: A). cion of HIV infection. HIV testing: HIV testing should be routinely Algorithms for the diagnosis of offered along with sputum examination for AFB smear-negative tuberculosis in HIV-prevalent settings for patients presenting with cough of 2–3 weeks’ duration. A per- In the absence of rapid and simple tools to diag- son with unknown HIV status (e.g. because of nose tuberculosis, the main aim of these algo- unavailability of HIV test kits or refusal to be rithms is to assist clinical decision-making in tested) can be classifi ed as HIV-positive if there HIV-prevalent and resource-constrained set- is strong clinical evidence of HIV infection. tings, to expedite the diagnostic process and HIV assessment: This includes clinical staging minimize incorrect diagnosis and mortality. The of HIV infection (see Table 2), immunologi- algorithms will have signifi cant implications for cal staging (CD4 count), referral for HIV care both tuberculosis and HIV/AIDS service provid- including antiretroviral treatment, long-term ers in these settings, and will catalyse the inte- follow-up and chronic management, including gration of HIV and tuberculosis interventions co-trimoxazole preventive therapy. The clinical at the point of service delivery. The algorithms staging is important, as some patients with pul- are aimed at adult and adolescent patients pre- monary tuberculosis may also have concurrent senting with cough of 2–3 weeks’ duration and stage 4 disease requiring more rapid initiation differ according to the clinical condition of the of antiretroviral treatment. patient (ambulatory or seriously ill). Clinical assessment: This is a critical step in the Guiding principles diagnostic process, particularly in the absence of any bacteriological confi rmation of tuber- Target group: The newly revised algorithms culosis. It must be based, as far as possible, on (Figures 1 and 2) are targeted at adults living with supportive investigations and sound clinical HIV/AIDS and those considered to be at high judgement in order to arrive at a correct diag- risk of HIV infection on clinical and epidemio- nosis without undue delay and prevent excess logical grounds, as laid down in national and/or mortality from undiagnosed tuberculosis. It is local policy. The diagnostic procedure for HIV- also useful for the diagnosis and management of negative patients and those who are less likely nontubercular conditions during all evaluations to be HIV-infected should follow the codifi ed of the patient. Sound clinical judgement will be algorithm (based on WHO’s 2003 recommenda- essential for: classifying the patient as ambula- tions) included in the International standards for tory or seriously ill on the basis of danger signs; tuberculosis care, 2006 (2) (Figure 3). classifying the patient of unknown HIV status as Danger signs: The adult patient will be classi- HIV-positive or negative; starting the patient on fi ed as seriously ill if one or more of the follow- broad-spectrum antibiotics or antituberculosis ing danger signs are present: drugs on the basis of his/her clinical condition and presentation; assessing, managing and/or • unable to walk unaided referring the patient for treatment for other dis- • respiratory rate over 30 per minute eases. Because performing these activities is part of basic clinical practice, it is not possible to be • fever of more than 39 °C more instructive in these recommendations. • pulse rate of over 120 per minute. Clinical response: For patients in whom tuber- AFB microscopy: At least two sputum speci- culosis is less likely and who are treated empiri- mens should be taken and examined for AFB. cally for bacterial pneumonia or Pneumocystis One of the specimens should be early-morning carinii pneumonia (PCP), clinical response

8 PART I. IMPROVING THE DIAGNOSIS AND TREATMENT OF SMEAR-NEGATIVE TUBERCULOSIS

FIGURE 1 Algorithm for the diagnosis of tuberculosis in ambulatory HIV-positive patient

Ambulatory patient with cough 2–3 weeks and no danger signsa

AFB 1 b

st HIV test

ii 2 visit

HIV+ or status unknownc

AFB-positive d AFB-negative d nd

ii 3 visit

Treat for TB TB likely CXR g CPTe Sputum AFB and culture g HIV assessment f Clnical assessment g

rd TB unlikely

ii 4 visit

Treat for PCP i Treat for bacterial infection h HIV assessment f HIV assessment f CPTe

j j th Response No or partial response Response

visit

Reassess for TB

a The danger signs include any one of: respiratory rate >30/minute, fever >39 °C, pulse rate >120/min and unable to walk unaided. b For countries with adult HIV prevalence rate ≥1% or prevalence rate of HIV among tuberculosis patients ≥5%. c In the absence of HIV testing, classifying HIV status unknown ias HIV-positive depends on clinical assessment or national and/or local policy. d AFB-positive is deﬁ ned at least one positive and AFB-negative as two or more negative smears. e CPT = Co-trimoxazole preventive therapy. f HIV assessment includes HIV clinical staging, determination of CD4 count if available and referral for HIV care. g The investigations within the box should be done at the same time wherever possible in order to decrease the number of visits and speed up the diagnosis. h Antibiotics (except ﬂ uoroquinolones) to cover both typical and atypical bacteria should be considered. i PCP: Pneumocystis carinii pneumonia, also known as Pneumocystis jirovecii pneumonia. j Advise to return for reassessment if symptoms recur.

9 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

should not automatically exclude the diagno- tive therapy provided according to national sis of tuberculosis. Acute bacterial pneumonia guidelines. or PCP may occur in patients with underlying • Third visit: Results of the second-visit inves- tuberculosis and patients should, therefore, be tigations (except culture) should be avail- re-evaluated for tuberculosis, particularly if able during the patient’s third visit. Patients respiratory symptoms persist after treatment. suspected of having tuberculosis after these Follow-up assessment of these patients can take investigations (e.g. compatible radiograph place under either tuberculosis services or HIV plus symptoms) should be treated for tuber- services, according to country-specifi c guidance culosis. Patients who are not treated for and practice. tuberculosis should receive either a broad- based antibiotic (not a fl uoroquinolone) to Algorithm for the ambulatory patient treat bacterial infection or treatment for PCP. HIV assessment should also be performed This algorithm is used for a tuberculosis sus- and co-trimoxazole preventive therapy pro- pect without the danger signs defi ned above vided according to national guidelines. (an ambulatory patient). The diagnostic process should be expedited if the patient is HIV- • Fourth visit: The patient’s response is positive, or likely to be so. The total number assessed and a clinical follow-up mechanism of visits for separate evaluations from the time is established (in either the tuberculosis or of initial presentation to a health facility to the the HIV services). For patients with immedi- time of diagnosis should not exceed four. The ate response to PCP or antibiotic treatment, number of days involved between evaluations continued vigilance is necessary to exclude will vary depending on several country-spe- superimposed tuberculosis. Those patients cifi c factors, and appropriate measures should with an unsatisfactory response to treatment be instituted by national and local tuberculosis for PCP or bacterial pneumonia should be and HIV authorities to minimize the time and reassessed both clinically and bacteriologi- the number of visits required to establish the cally for tuberculosis. diagnosis. Shortening the turnaround time for sputum smear examinations is crucial. Algorithm for seriously ill patient The following principles should be followed A seriously ill patient with one of the danger when applying the algorithms for the ambula- signs should be immediately referred to a higher- tory patient in order to expedite the diagnosis of level health facility. When immediate referral is smear-negative pulmonary tuberculosis. not possible, the following measures should be • First visit: HIV testing should be offered and undertaken in the peripheral health facility. AFB sputum examination should be per- • Immediately start with broad-spectrum formed. If AFB test is positive, treat for tuber- parenteral antibiotics for bacterial infec- culosis. tion and perform HIV test and sputum AFB • Second visit: If the AFB examination is nega- examination. Safe injection practices should tive, the patient should be provided with all be strictly followed. If the indications laid available investigations during the second down in national guidelines are present, PCP visit. The second visit should ideally take treatment should be considered. If the HIV place on the second day following fi rst pres- test is negative or there is less clinical suspi- entation at the health facility. The investiga- cion of HIV infection, or if the national or tions include: repeated sputum AFB, sputum local guidelines do not classify the area as culture and chest X-ray. Clinical assessment HIV-prevalent, continue management of the is also important for deciding whether to put HIV-negative patient according to national the patient on antituberculosis treatment practice and guidelines. If the HIV test is at this stage. HIV assessment should also positive, or there is high clinical suspicion of be performed and co-trimoxazole preven- HIV infection, follow the algorithm.

10 PART I. IMPROVING THE DIAGNOSIS AND TREATMENT OF SMEAR-NEGATIVE TUBERCULOSIS

FIGURE 2 Algorithm for the diagnosis of tuberculosis in seriously ill HIV-positive patient

Seriously ill patient with cough 2–3 weeks and danger signsa

Referral to higher level Immediate referral facility not possible

Parenteral antibiotic treatment for Parenteral antibiotics for bacterial bacterial infection b,d infection b,d Sputum AFB and culture b Consider treatment for PCP e HIV test b,c Sputum AFB and culture b CXR b HIV test b,c

HIV+ or unknown f

No Treat AFB-positive g AFB-negative g tuberculosis tuberculosis

Improvement No improvement after 3–5 days after 3–5 days

Reassess for other TB unlikely Reassess Start TB treatment HIV-related disease for tuberculosis h Complete antibiotics Refer for HIV and tuberculosis care

a The danger signs include any one of: respiratory rate >30/min, fever >39 °C, pulse rate >120/min and unable to walk unaided. b The investigations within the box should be done at the same time wherever possible in order to decrease the number of visits and speed up the diagnosis. c For countries with adult HIV prevalence rate ≥1% or prevalence rate of HIV among tuberculosis patients ≥5%. d Antibiotics (except ﬂ uoroquinolones) to cover both typical and atypical bacteria should be considered. e PCP: Pneumocystis carinii pneumonia, also known as Pneumocystis jirovecii pneumonia. f In the absence of HIV testing, classify HIV status unknown into HIV-positive depends on clinical assessment or national and/or local policy. g AFB-positive is deﬁ ned as at least one positive and AFB-negative as two or more negative smears. h Reassessment for tuberculosis includes AFB examination and clinical assessment.

11 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

FIGURE 3 Algorithm for the diagnosis of tuberculosis in HIV-negative patients (International standards for tuberculosis care, 2006)

All patients suspected of having pulmonary tuberculosis

Sputum microscopy for AFB

Three negative smears

Broad-spectrum antibiotics (excluding antituberculosis drugs and ﬂ uoroquinolones)

No improvement Improved

Repeat sputum microscopy

One or more All smears negative positive smears

Chest radiograph and physician’s judgement

Tuberculosis No tuberculosis

Source: Adapted from (1)

12 PART I. IMPROVING THE DIAGNOSIS AND TREATMENT OF SMEAR-NEGATIVE TUBERCULOSIS

Table 2. Revised WHO clinical staging of HIV/AIDS for adults and adolescents with conﬁ rmed HIV infection CLINICAL STAGE 1 Asymptomatic Persistent generalized lymphadenopathy

CLINICAL STAGE 2 Moderate unexplaineda weight loss (<10% of presumed or measured body weight)b Recurrent respiratory tract infections (sinusitis, tonsillitis, otitis media, pharyngitis) Herpes zoster Angular cheilitis Recurrent oral ulceration Papular pruritic eruptions Seborrhoeic dermatitis Fungal nail infections

CLINICAL STAGE 3 Unexplaineda severe weight loss (>10% of presumed or measured body weight)b Unexplaineda chronic diarrhoea for longer than one month Unexplaineda persistent fever (above 37.5 °C intermittent or constant for longer than one month) Persistent oral candidiasis Oral hairy leukoplakia Pulmonary tuberculosis Severe bacterial infections (e.g. pneumonia, empyema, pyomyositis, bone or joint infection, meningitis, bacteraemia) Acute necrotizing ulcerative stomatitis, gingivitis or periodontitis Unexplained anaemia (<8 g/dl ), neutropenia (<0.5 x 109/L) and or chronic thrombocytopenia (<50 X 109/L3)

CLINICAL STAGE 4c HIV wasting syndrome Pneumocystis pneumonia Recurrent severe bacterial pneumonia Chronic herpes simplex infection (orolabial, genital or anorectal of more than one month’s duration or visceral at any site) Oesophageal candidiasis (or candidiasis of trachea, bronchi or lungs) Extrapulmonary tuberculosis Kaposi’s sarcoma Cytomegalovirus infection (retinitis or infection of other organs) Central nervous system toxoplasmosis Extrapulmonary cryptococcosis including meningitis Disseminated nontuberculous mycobacteria infection Progressive multifocal leukoencephalopathy Chronic cryptosporidiosis Chronic isosporiasis Disseminated mycosis (extrapulmonary histoplasmosis, coccidiomycosis) Recurrent septicaemia (including nontyphoidal Salmonella) Lymphoma (cerebral or B cell nonHodgkin) Invasive cervical carcinoma Atypical disseminated leishmaniasis Symptomatic HIV-associated nephropathy or symptomatic HIV-associated cardiomyopathy

Source: Adapted from (7). a. Unexplained indicates that the condition is not explained by any other condition. b. Assessment of body weight in pregnant women needs to take into account the expected weight gain of pregnancy. c. Some additional speciﬁ c conditions can also be included in regional classiﬁ cations (e.g. reactivation of American trypanosomiasis (meningoencephalitis and/or myocarditis) in the WHO Region of the Americas and penicilliosis in Asia).

13 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

• If the diagnosis of tuberculosis is conﬁ rmed by AFB smear examination, start antituberculosis treatment. The antibiotic treatment should be continued and completed. • If the AFB smear is negative, response to parenteral antibiotics should be assessed 3–5 days into treatment, and, if there is no improvement, antituberculosis treatment should be initiated. The initial antibiotic course should be continued and completed. HIV assessment and clinical staging should be performed. Patients should be referred to the next level of care to conﬁ rm the diagnosis of tuberculosis and for HIV care. If referral is not possible, antituberculosis treatment should be completed. • If referral to a higher-level facility is possible, the patient should be managed as an emer- gency and all available investigations, including HIV testing, should be performed at one time for the diagnosis of tuberculosis.

14 PART II Simpliﬁ ed and standardized clinical management guidelines for extrapulmonary tuberculosis

PART II. SIMPLIFIED AND STANDARDIZED CLINICAL MANAGEMENT GUIDELINES

Background ment should be assessed after one month. If there is no improvement, a clinical reassessment One in fi ve registered tuberculosis patients has should be performed and an alternative diagno- extrapulmonary tuberculosis (8, 9, 10). The sis sought. commonest forms include lymph node (especially in the neck or under the arms), pleural HIV testing should be offered to all patients sus- (usually one-sided pleural effusion) and dissem- pected of extrapulmonary tuberculosis. This is inated tuberculosis (disease that is not limited to because HIV-related extrapulmonary tubercu- one site in the body). Pericardial and meningeal losis is an indication for early commencement tuberculosis are less frequent forms of extrapul- of antiretroviral treatment (clinical stage 4 of monary tuberculosis that are also covered in HIV disease). For HIV-related extrapulmonary these guidelines. About one-third of deaths in tuberculosis, the following interventions should HIV-positive Africans are due to disseminated be carried out: tuberculosis (11, 12, 13) but only about half of • refer for HIV care or start antiretroviral treat- HIV-positive patients who die from dissemi- ment according to national guidelines nated tuberculosis are diagnosed before death (12, 13, 14). With the exception of lymph node • start co-trimoxazole preventive therapy tuberculosis, which can usually be confi rmed • remain vigilant for clinical deterioration of through aspiration of affected lymph nodes, extrapulmonary tuberculosis after the start most patients with extrapulmonary tubercu- of antiretroviral treatment (immune recon- losis are managed without bacteriological or stitution infl ammatory syndrome – IRIS) histological confi rmation (15). Therefore, it is and take appropriate measures. important for health care workers to have simplifi ed, standardized guidelines for the prompt diagnosis and management of extrapulmonary Tuberculous lymphadenitis tuberculosis. Tuberculous lymphadenitis should be suspected in any patient with enlarged lymph nodes that Target audience are fi rm, asymmetrical, more than 2 cm in diameter, or where a node has become fl uctu- These guidelines are intended to assist the ant or developed a fi stula over several months. prompt diagnosis and management of extrapul- It most commonly affects the nodes in the neck monary tuberculosis by physicians and other (cervical region) and is diffi cult to distinguish clinicians working in district hospitals of HIV- clinically from other causes of enlarged nodes, prevalent and resource-constrained settings as such as reactive and/or HIV-related lymphad- part of national tuberculosis control programme enopathy, malignancies and other lymph node activities. infections, which are also common. Therefore, needle aspiration using recommended techniques (see box “Guidelines for lymph node Diagnosis and management aspiration” below) should be carried out at the The indications for suspected extrapulmonary fi rst outpatient visit for all patients. tuberculosis and the key signs to look for in Needle aspiration with cytology and tuber- the commonest forms of the disease are sum- culosis microscopy of aspirated material has a marized in Figure 4. Table 3 summarizes the high diagnostic yield, with confi rmation of over essential investigations required for diagnosis 85% of patients with tuberculous lymphadenitis and key steps for immediate management of in some (16, 17, 18, 19) but not all (20) reports, suspected extrapulmonary tuberculosis cases. suggesting that the technique may be impor- For a patient with suspected extrapulmonary tant. If a fi stula has formed, then microscopy tuberculosis who is started on antituberculosis of discharging pus is likely to show AFB. Cytol- treatment without bacteriological or histologi- ogy, if available, can identify most other impor- cal confi rmation, the clinical response to treat- tant causes of enlarged lymph nodes, including

17 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

FIGURE 4 Suggested clinical characteristics to assist the diagnosis of extrapulmonary tuberculosis (ETB)

Suspect ETB in patients with Look and listen for Cough for two weeks or more or • Lymph nodes swelling in the neck or armpits • Unintentional weight loss with (if present with other types of • Night sweats and ETB it may provide the only • Temperature >37.5 °C or way to confi rm the diagnosis) feels feverish Possible tuberculosis • Breathlessness (effusion/ lymphadenitis pericarditis) or • Enlarged glands in neck/ • Signs of fl uid in the chest armpit or • Absent breath sounds • Chest X-ray • Reduced chest wall • Miliary or diffuse shadowing movement • Large heart (especially if • Dull to percussion symmetrical and rounded) Possible tuberculosis pleural • Pleural effusion effusion • Enlarged lymph nodes inside the chest • Signs of fl uid around the heart • Chronic headache or altered • Heart sounds distant mental state • Swollen legs and/or abdomen Suspect disseminated • Neck and hand veins tuberculosis in all people living distended with arm held with HIV who experience rapid above the shoulder or marked weight loss, fever and Possible tuberculosis night sweats pericarditis

• Signs of meningitis • neck stiffness Establish HIV status if ETB • confusion is suspected • abnormal eye movements • Advise and arrange for Possible tuberculosis rapid HIV testing if status meningitis is unknown or last test was negative • Explain that this will affect the way that this illness is investigated and treated • Discuss the need for antiretroviral treatment if HIV-related tuberculosis is diagnosed • If consent is given, try to arrange testing on the same day

18 PART II. SIMPLIFIED AND STANDARDIZED CLINICAL MANAGEMENT GUIDELINES

malignancies and other infections. Follow-up to receive the results should be within seven days. If Guidelines for lymph node aspiration the aspirate does not yield a diagnosis, then exci- Equipment needed: topical antiseptic, gloves, sion biopsy for gross examination, Ziehl-Neelsen 5 ml syringe and 18 to 21 gauge needle,* 3 glass microscopy slides, cytological fi xative microscopy, mycobacterial culture and, if availa- (e.g. absolute alcohol or methanol) if cytology ble, histological examination can be considered. available However, antituberculosis treatment should be Steps started immediately if: 1. Prepare the microscopy slides with the patient’s name and identifi cation number. • the patient is HIV-infected and has clinical 2. Apply a topical antiseptic to the skin overly- features of disseminated tuberculosis (such ing the enlarged lymph node. as marked weight loss, rapid clinical deterio- 3. Attach the needle and expel all air from the ration or multiple sites of suspected tubercu- syringe. losis) or 4. With the nondominant hand, take the gland between the thumb and the index fi nger to • tuberculous lymphadenitis is considered the make it stand out and hold it steady. most likely clinical diagnosis, but logistic or 5. Taking the syringe in the dominant hand, economic barriers are likely to delay excision insert the needle through healthy skin into biopsy for two weeks or longer. the centre of the node or at the point of maximum fl uctuance, and pull back on the syringe piston. If no aspirate is obtained, move the needle in and out of the centre of Pleural effusion the node while pulling back on the syringe piston. Gently compress the node with the Tuberculosis is t he li kely cause of uni latera l pleu- nondominant hand and revolve the needle ral effusion in countries with a high tuberculosis in both directions. Small amounts of lymph burden. It was the diagnosis reached in 95% of node tissue will collect in the needle and needle hub, even if there is no visible aspi- patients in two recent case-series from Uganda rate inside the syringe. and Zimbabwe (21, 22). Pleural effusion is the 6. Withdraw the needle and syringe and most common form of HIV-related extrapul- spread aspirate onto each slide. It may be monary tuberculosis, with high mortality (over necessary to disconnect the syringe and 20%) in the fi rst two months of antituberculosis introduce a small amount of air in order to expel the contents. A separate aspirate treatment (21, 22). may be needed for each slide. The following key steps should be undertaken. 7. Allow slides to air- dr y. If pus is obtained, send one slide for Gram stain and one • The management of tuberculous pleural for tuberculosis microscopy. If no pus is effusion should aim at starting antitubercu- obtained, then send both slides for tuberculosis microscopy. losis treatment and identifying underlying 8. If available, spray the remaining slide with HIV infection without delay. Pleural biopsy cytological fi xative, and send for cytology has a high diagnostic yield (21, 22), but is not when dry. recommended because it is unnecessarily * Reported yields are better with larger needle sizes (18 or 19G: wide-needle aspiration), but fi ne-needle invasive and has the potential to introduce aspiration with a standard phlebotomy (21G) needle diagnostic delay. can be used if that is all that is available. Lymph node needle-core biopsy is an acceptable alternative • Suspected pleural effusion should be con- for facilities with appropriate equipment. fi rmed by chest radiography and immediate aspiration of fl uid whenever possible (see Table 3), placing aliquots of the aspirate into ing, unless there is clinical concern about one plain and two anticoagulated tubes. bacterial pneumonia. • Treatment with broad-spectrum antibiotics • Patients with unusual fi ndings, such as bilat- is not required before antituberculosis treat- eral effusions and cloudy or bloody aspirates, ment in patients with unilateral effusions if should undergo the additional investigations the pleural fl uid is clear and clots on stand- detailed in Table 3. If visible clots form in

19 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS so so d AFB)/ tube, tube, and and

tococcal) r India ose ose cerebrospinal cerebrospinal stain an stain t a ococcal disease disease ococcal

t sweats, fever lture) –ve in in –ve lture) rculosis elsewhere rculosis tigen (o tigen low gluc low tube n-tuberculosis n-tuberculosis berculosis only berculosis robably cryprobably sugges e (crypt

rculosis treatment rculosis nosis made nosis cerebrospinal fl uid uid fl cerebrospinal anticoagulated r cryptococcal disease if onset onset rebrospinal fl uid clear with with clear uid fl rebrospinal rebrospinal fl uid cloudy or or cloudy uid fl rebrospinal into an into

HIV test (rapid if possible) if (rapid test HIV puncture Lumbar (Gram Microscopy ink and fungal cu fungal and ink uid ﬂ cerebrospinal Evidence of lymphocytes Cryptococcal an tube Start protein/glucose in protein/glucose ﬂ u i d Cryptococcal antigen/stain coughing if smears Sputum Weight loss, nigh Ce high protein, Admit Treat fo no and HIV+ if or crypto +ve tests other diag HIV test +v test HIV tuberculosis) than likely more Ce (probably microscopy on neutrophils bacterial) Cryptococcal +ve tests Rapid Very high (p pressure

TUBERCULOUS MENINGITIS TUBERCULOUS investigations Essential • • • • • thatFindings management Immediate Features of tu • • • tuberculosis if: of suspicion High • • • Features no of diagnosis • non-tuberculosis diagnosis • • • • • and and lve lve ld be placed be ld not not art failure) (urea

ests ests ultrasound re, va re, ape ape ultrasound not sion)

ble he ble and no other other no and

t sweats, fever rculosis elsewhere rculosis if coughing if nd (ideally) nd liquot shou n-tuberculosis diagnosis. diagnosis. n-tuberculosis ing of lung lung of ing

ram sugg ram rculosis if if rculosis heart sh e for enlarged heart ral effu ral tube n-tuberculosis n-tuberculosis berculosis only berculosis other causes rculosis treatment treatment rculosis blood pressu trasound) cal (proba cal Streaky shadow Streaky fi e l d s asymmetri n d / o r Cardiac ultrasou Electrocardiogram if available have may (but clear elds fi Lung pleubilateral Weight loss, nigh Evidence of pressure blood High Electrocardiog caus another HIV test (rapid if possible) if (rapid test HIV CXR smearsSputum (e.g. high Refer for urgent aspiration if very if aspiration urgent for Refer breathless/unwell Start tube Start Investigate disease, dilated cardiomyopathy) dilated disease, disease) valvular (probable Murmur Rigors (probable bacterial pericarditis) cardiac ul cardiac Treat for tube confi rms effusion rms confi diagnosis made by 7 days 7 diagnosis made by

Findings that suggest that Findings a non-tuberculosis diagnosis • • • tuberculosis if: of suspicion High • • • • • PERICARDIAL EFFUSION investigations Essential • • • • Features no of • diagnosis • • • management Immediate Features of tu • rate rate

od count count od d cough ch can include include can ch t a espiratory critically ill critically ulosis treatmentulosis

if coughing if

ver an ver blood culture blood r or likely pathogen critically ill) critically , malaria, es, full blo full es, n-tuberculosis n-tuberculosis berculosis only berculosis sugges er Salmonella, Salmonella, er

rculosis treatment (add (add treatment rculosis HIV test (rapid if possible) if (rapid test HIV CXR lm ﬁ blood Malaria smearsSputum cultur Blood cryptococcal antigen and Weight loss, fe (whi CXR Abnormal miliary pattern) spleen/liver Large sweats Night Anaemia Rigors Very breathless (r >30/min) diarrhoea Severe stool in Blood Positive cryptococcal antigen, malaria smea isolated from tube Start if antibiotics causes other Investigate tuberc both Start if antibiotics and

spected tuberculosis extrapulmonary DISSEMINATED TUBERCULOSIS DISSEMINATED investigations Essential • • • • • tuberculosis if: of suspicion High • • • • • thatFindings non-tuberculosis diagnosis consid HIV+ if pneumococcus Cryptococcus • • • • • management Immediate Features of tu • Features no of diagnosis • • for for ly ly s by 7 days ﬂ u i d the on the b tial cell cell tial

cytes and cytes and lignancy n (if possible) of if available, available, if ﬂ u i d

rculosis diagnosisrculosis differential cell differential cell (possible heart rculosis elsewhere rculosis

t sweats, fever is:- but send send but if coughing if

b e is failure of aspirate rculosis if rculosis other diagnosi spect tube

berculosis only berculosis /other ma anticoagulants anticoagulants and differen and 50% lympho 50% a 0 g/L suggests tuberculosis rculosis treatment rculosis ≥ rculosis, rculosis, empyema) tube without without protein protein count, and consider heart consider and count, failure) n tube (with no anticoagulant) in order to observe its appearance and clotting. A second a Evidence for effusion Bilateral failure or pneumonia) or failure KS Clinical Unilateral effusion Unilateral uid ﬂ of Aspirate aspirate HIV test (rapid if possible) if (rapid test HIV CXR smearsSputum and count cell blood white Differential determinatio protein (probable — Cloudy/pus — Fails to clot (does not exclude Aspirate & in & Aspirate is:- uid ﬂ of Aspirate — Clear and straw coloured and — Clots on standing in a tube Weight loss, nigh unusual featur unusual no or clot, to

Start tube Start for aspirate Send and, protein count, Treat for tube cytology: >3 protein

• suggest that Findings a non-tuberculosis diagnosis • • High suspicion of tuberculosis if: of suspicion High • •

PLEURAL EFFUSION investigations Essential • • • • • • • • Features non-tube of • Immediate management Immediate Features of tu • d cell count and protein determination can bendings requested to suggest if there a no are any ﬁ put in a plai a in put AFB AFB

ite bloo ite lent lent B ( 18 to weeks and and sis and immediate management of su t a in 2 2 in d biopsy not if coughing coughing if ly clinical clinical ly if aspirate aspirate if ations

tology ed, puru ed, BERCULOSIS BERCULOSIS (probable night sweats, s sarcoma an cervicalan rate for AF for rate sugges re in size in re

vestig rapid rapid rculosis considered considered rculosis in skin or mouth or skin in

disseminated tuberculosis (e.g. deterioration) clinical the most like most the diagnosis, an diagnosis, available with available a that a differential wh differential a that KS-Kaposi’ be should aspirate The RIPHERAL) Aspirate for cy for Aspirate HIV test (rapid if possible) if (rapid test HIV smears Sputum 21 gauge) moor2 cm Asymmetrical/localized Painless swelling stulated uctuant/fi Firm/fl location Cervical Weight loss, fever Needle aspi Needle (probable KS nodes) Symmetrical HIV or lymphoma lymphadenopathy) Tender,am infl (bacterial fungal) or th other Site microscopy biopsy Excision non-diagnostic unless — HIV+ with possible — Tube

management Immediate • • • tuberculosis if: of suspicion High • • • • • • thatFindings TableDiagno 3. LYMPH NODE TU (PE in Essential • a b • • • •

non-tuberculosis diagnosis •

20 PART II. SIMPLIFIED AND STANDARDIZED CLINICAL MANAGEMENT GUIDELINES

the aspirate within a few minutes of its being assessments shown in Table 3. The attending placed into a plain tube (no anticoagulant), health care worker should carefully consider then this confi rms the high protein content the need for additional investigations and treat- of the fl uid, which indicates tuberculosis. No ment (such as antibiotics) if a diagnosis other further investigations are needed if the aspi- than tuberculosis is suspected. However, it is rate is clear and straw-coloured and there are not necessary to give broad-spectrum antibiot- no other features suggestive of a diagnosis ics routinely before considering antituberculo- other than tuberculosis. sis treatment. • Failure of the aspirate to clot does not exclude Tuberculosis treatment should be started as soon tuberculosis, and such patients can still be as other common conditions that can cause a started on antituberculosis treatment imme- similar clinical picture have been excluded (see diately if there are no other unusual fi ndings Table 3 for essential investigations) in patients (Table 3), but laboratory analysis of fl uid presenting with the following conditions. is needed to determine the protein content • Pericardial effusion: tuberculosis is the cause (expect ≥30 g/L in patients with a tuber- of about 90% of HIV-related pericardial effu- culous effusion, but it can be lower in very sion, but a lower percentage (50% to 70%) of wasted patients) and differential cell count pericardial effusions in HIV-negative indi- (expect ≥50% lymphocytes in a tuberculous viduals (23, 24, 25). effusion). The aim should be to start antituberculosis treatment within seven days unless • Meningitis with features of the cerebrospinal another diagnosis has been made. fl uid suggestive of tuberculosis (see Table 3). • If thoracentesis is not available, antitubercu- • Suspected disseminated tuberculosis in losis treatment should be started immediately, febrile patients presenting with HIV wast- particularly if the patient is HIV-infected, ing syndrome. High rates of undiagnosed unless there are clinical or radiological fea- disseminated tuberculosis have been con- tures suggestive of a diagnosis other than sistently identifi ed in febrile, HIV-positive tuberculosis. inpatients and in postmortem series from several countries (11, 12, 13, 14, 26, 27, 28, 29, 30). Other forms of extrapulmonary tuberculosis Patients with clinical features or investigation results that suggest a diagnosis other than Most patients with other forms of extrapulmo- extrapulmonary tuberculosis (listed in Table 3) nary tuberculosis present in a suffi ciently charac- need more extensive investigation before antitu- teristic way to allow antituberculosis treatment berculosis treatment is considered, but with to be started without attempting to confi rm the aim of keeping diagnostic delays to a mini- the disease bacteriologically or histologically. mum. Although extrapulmonary tuberculosis can be confi rmed in the majority of patients through invasive biopsy and/or multiple cultures, these Adjuvant corticosteroids investigations are not routinely recommended, Corticosteroids started at the time of tuberculo- as they are expensive and may result in lengthy sis diagnosis and given for the fi rst two months diagnostic delays that can reduce the chances of of treatment signifi cantly improve survival a good treatment response (19). from tuberculous meningitis in HIV-negative Taking specimens for culture increases the patients, and are now recommended for such chances that tuberculosis will be confi rmed, but patients (31). For other forms of extrapulmo- treatment should not generally be delayed until nary tuberculosis and for HIV-related tubercu- culture results are available. Instead, antituber- lous meningitis, the effects of steroids are still culosis treatment should be started promptly, if uncertain. The results of small trials on tuber- indicated after the essential investigations and culous pericarditis are promising (32). There

21 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

appears to be no beneﬁ t in adding steroids to the treatment of tuberculous pleural effusion, with some suggestion of possible harm to HIV- positive patients (33). Recommendations may change when the results of larger randomized clinical trials become available within the next few years.

22 Further reading

• Antiretroviral therapy for HIV infection in adults • Revised TB recording and reporting forms and and adolescents in resource-limited settings: registers. Geneva, World Health Organization towards universal access. Recommendations (in press). for a public health approach. Geneva, World • TB/HIV: a clinical manual, 2nd ed. (document Health Organization, 2006, 2006. http://www. WHO/HTM/TB/2004.329). Geneva, World who.int/hiv/pub/guidelines/adult/en/index. Health Organization, 2004. http://www.who. html (accessed 29 October 2006). int/tb/publications/who_htm_tb_2004_329/ • Guidance for national tuberculosis pro- en/index.html (accessed 29 October 2006). grammes on the management of tuberculo- • The Stop TB strategy (document WHO/ sis in children (WHO/HTM/TB/2006.371/ HTM/TB/2006.368). Geneva, World Health WHO/FCH/CAH/2006.7). Geneva, World Organization, 2006. Health Organization, 2006. • WHO case deﬁ nitions of HIV for surveillance • Guidelines on co-trimoxazole prophylaxis for and revised clinical staging and immunological HIV-related infections among children, ado- classiﬁ cation of HIV-related disease in adults and lescents and adults in resource-limited settings. children. Geneva, World Health Organization, Recommendations for a public health approach. 2006. http://www.who.int/hiv/pub/guide- Geneva, World Health Organization, 2006. lines/hivstaging/en/index.html (accessed 29 http://www.who.int/hiv/pub/guidelines/ctx/ October 2006). en/index.html (accessed 29 October 2006).

• Interim policy on collaborative TB/HIV activities (document WHO/HTM/TB/2004.330/ WHO/HTM/HIV/2004.1). Geneva, World Health Organization, 2004. http://www.who. int/hiv/pub/tb/tbhiv/en/ (accessed 29 Octo- ber 2006).

23 References

1. Treatment of tuberculosis: guidelines for national 9. Harries AD et al. The scourge of HIV-related programmes, 3rd ed. Geneva, World Health tuberculosis: a cohort study in a district gen- Organization, 2003 (WHO/CDS/TB/2003.313). eral hospital in Malawi. Annals of Tropical http://www.who.int/tb/publications/cds_tb_ Medicine and Parasitology, 1997, 91:771–776. 2003_313/en/index.html (accessed 29 October 10. Tam CM et al. Tuberculosis in Hong Kong – 2006). patient characteristics and treatment outcome. 2. International standards for tuberculosis care. Hong Kong Medical Journal, 2003, 9:90. The Hague, Tuberculosis Coalition for Techni- 11. Lucas SB et al. The mortality and pathology cal Assistance, 2006. of HIV infection in a west African city. AIDS, 3. BHIVA guidelines for the treatment of HIV- 1993, 7:1569–1579. infected adults with antiretroviral therapy. 12. Rana FS et al. Autopsy study of HIV-1-positive London, British HIV Association, 2005. http:// and HIV-1-negative adult medical patients in www.bhiva.org/guidelines/2005/BHIVA- Nairobi, Kenya. Journal of Acquired Immune guidelines (accessed 30 October 2006). Defi ciency Syndrome, 2000, 24:23–29. 4. Briss PA et al. Developing an evidence-based 13. Ansari NA et al. Pathology and causes of death guide to community preventive services in a group of 128 predominantly HIV-positive – methods. The Task Force on Community patients in Botswana, 1997–1998. International Preventive Services. American Journal of Pre- Journal of Tuberculosis and Lung Disease, 2002, ventive Medicine, 2000, 18(1 Suppl):35–43. 6:55–63. 5. WHO Health Evidence Network. What is the 14. Archibald LK et al. Fatal Mycobacterium evidence for the effectiveness of interventions to tuberculosis bloodstream infections in febrile reduce hepatitis co-infection and the associated hospitalized adults in Dar es Salaam, Tanzania. morbidity? Copenhagen, WHO Regional Offi ce Clinical Infectious Diseases, 1998, 26:290–296. for Europe, 2005. www.euro.who.int/HEN/ Syntheses/hepatitisC/20050412_1 (accessed 15. Richter C et al. Extrapulmonary tuberculosis 30 October 2006). – a simple diagnosis? Tropical and Geographical Medicine, 1991, 43:375–378. 6. EBM guidelines: evidence-based medicine [online database]. New York, Wiley. http:// 16. Bem C et al. The value of wide-needle aspira- ebmg.wiley.com/ebmg/ltk.koti (accessed 16 tion in the diagnosis of tuberculous lymphad- June 2006). enitis in Africa. AIDS, 1993, 7:1221–1225. 7. Revised WHO clinical staging and immunologi- 17. Pithie AD, Chicksen B. Fine-needle extratho- cal classifi cation of HIV and case defi nition of racic lymph-node aspiration in HIV-associ- HIV for surveillance. Geneva, World Health ated sputum-negative tuberculosis. Lancet, Organization (in press 2006). 1992, 340:1504–1505. 8. Nunn P et al. Cross-sectional survey of HIV 18. Wilson D et al. Diagnostic yield of periph- infection among patients with tuberculosis eral lymph node needle core biopsies in HIV- in Nairobi, Kenya. Tubercle and Lung Disease, infected adults with suspected smear-negative 1992, 73:45–51. tuberculosis. International Journal of Tubercu- losis and Lung Disease, 2005, 9:220–222.

24 REFERENCES

19. Hudson CP, Wood R, Maartens G. Diagnosing 27. Archibald LK et al. A hospital-based prevalence HIV-associated tuberculosis: reducing costs survey of bloodstream infections in febrile and diagnostic delay. International Journal of patients in Malawi: implications for diagnosis Tuberculosis and Lung Disease, 2000, 4:240– and therapy. Journal of Acquired Immune Defi - 245. ciency Syndromes and Human Retrovirology, 2000, 181:1414–1420. 20. Bekedam HJ et al. Tuberculous lymphadenitis, a diagnostic problem in areas of high preva- 28. Bell M et al. Seasonal variation in the etiology lence of HIV and tuberculosis. Transactions of bloodstream infections in a febrile inpatient of the Royal Society of Tropical Medicine and population in a developing country. Inter- Hygiene, 1997, 91:294–297. national Journal of Infectious Diseases, 2001, 5:63–69. 21. Luzze H et al. Evaluation of suspected tuberculous pleurisy: clinical and diagnostic fi ndings 29. McDonald LC et al. Unrecognised Mycobacte- in HIV-1-positive and HIV-negative adults in rium tuberculosis bacteraemia among hospital Uganda. International Journal of Tuberculosis inpatients in less developed countries. Lancet, and Lung Disease, 2001, 5:746–753. 1999, 354:1159–1163. 22. Heyderman R et al. Pleural tuberculosis in 30. Lewis DK et al. Clinical indicators of myco- Harare, Zimbabwe: the relationship between bacteraemia in adults admitted to hospital human immunodefi ciency virus, CD4 lym- in Blantyre, Malawi. International Journal of phocyte count, granuloma formation and Tuberculosis and Lung Disease, 2002, 6:1067– disseminated disease. Tropical Medicine and 1074. International Health, 1998:3:14–20. 31. Thwaites GE et al. Dexamethasone for the 23. Maher D, Harries AD. Tuberculous pericardial treatment of tuberculous meningitis in adoles- effusion: a prospective clinical study in a low- cents and adults. New England Journal of Medi- resource setting – Blantyre, Malawi. Interna- cine, 2004, 351:1741–1751. tional Journal of Tuberculosis and Lung Disease, 32. Mayosi BM et al. Interventions for treating 1997, 1:358–364. tuberculous pericarditis. Cochrane Database of 24. Cegielski JP et al. Tuberculous pericarditis Systematic Reviews, 2002, (4):CD000526. in Tanzanian patients with and without HIV 33. Elliott AM et al. A randomized, double-blind, infection. Tubercle and Lung Disease, 1994, placebo-controlled trial of the use of pred- 75:429–434. nisolone as an adjunct to treatment in HIV- 25. Reuter H, Burgess LJ, Doubell AF. Epidemiol- 1-associated pleural tuberculosis. Journal of ogy of pericardial effusions at a large academic Infectious Diseases, 2004, 190:869–878. hospital in South Africa. Epidemiological Infec- tion, 2005, 133:393–399. 26. Archibald LK et al. Fever and human immunodefi ciency virus infection as sentinels for emerging mycobacterial and fungal bloodstream infections in hospitalised patients ≥15 years old, Bangkok. Journal of Infectious Dis- eases, 1999, 180:87–92.

ANNEX Protocol for operational evaluation of the revised recommendations and algorithms for improving the diagnosis of tuberculosis in HIV-prevalent settings

ANNEX. PROTOCOL FOR OPERATIONAL EVALUATION OF THE REVISED RECOMMENDATIONS

Background 6. Sputum culture for Mycobacterium tuberculosis should be performed in patients who are In 1991, WHO fi rst published guidelines for sputum smear-negative to confi rm the diag- national tuberculosis control programmes that nosis of tuberculosis and improve the quality included criteria for the diagnosis of smear- of care. positive and smear-negative pulmonary and extrapulmonary tuberculosis. These were sub- In the absence of complete evidence, the recom- sequently revised in 1997 and 2003. In response mendations are based on consensus and iterative to concerns that the 2003 guidelines (1) did global expert opinion in order to respond to the not adequately refl ect the diagnostic and treat- catastrophe posed by the HIV epidemic. They ment challenges of HIV-associated tuberculosis, should, therefore, be implemented in HIV-prev- WHO has revised its recommendations for the alent settings. However, it is equally important diagnosis of tuberculosis in HIV-prevalent set- to build up the evidence base simultaneously, in tings. settings where this is possible, in order to assess the effectiveness and feasibility of the guidelines Major changes between the revised recommen- and thus inform changes in policy and practice. dations presented in this document, and the previous guidelines (2003) are as follows. 1. The revised 2006 guidelines apply only to: Objectives of the evaluation (a) patients suspected of having tuberculosis The primary intent of the evaluation is to meas- and living in settings (geographical area ure the performance of tuberculosis programmes or health facility) with an HIV prevalence that implement the revised recommendations >1% in pregnant women or an HIV-prev- and generate knowledge for improving those alence ≥5% in tuberculosis patients. specifi c programmes. The evaluation involves measuring different indicators of input, proc- (b) patient’s age >15 (guidelines for child- ess, output, outcome and impact in settings hood tuberculosis are separate). that implement the revised recommendations (c) note that, for all other populations, the and comparing them with settings that have existing guidelines laid down in the Inter- not implemented the recommendations. The national standards for tuberculosis care (2) evaluation provides information for interna- should be followed. tional and national health policy-makers and public health offi cials about the strengths and 2. All tuberculosis suspects should be routinely weaknesses of the revised guidelines in order to offered HIV counselling and testing. This inform changes in international and national differs from the existing WHO recommen- policy and practice. Therefore, the evaluations dation that these should be offered only to should be conducted in close collaboration with tuberculosis patients. national tuberculosis and HIV control pro- 3. A “trial” of antibiotics is not required to diag- grammes in the countries concerned. nose smear-negative pulmonary tuberculosis. Purpose of the protocol 4. Two sputum specimens, with one collected in the morning, are suffi cient for the initial This protocol provides generic guidance on diagnostic evaluation of tuberculosis in HIV conducting evaluations of the revised recom- patients. This differs from the 2003 WHO rec- mendations to improve the diagnosis of tuber- ommendation that “at least” three specimens culosis in HIV-prevalent settings. It is intended should be AFB-negative before diagnosing to standardize the minimum information that smear-negative pulmonary tuberculosis. needs to be generated by the evaluation in order to inform changes in policy at both national and 5. A patient is considered to have smear-positive global levels. The protocol will also provide a tuberculosis if at least one specimen is posi- fl exible platform for research groups and inter- tive for AFB.

29 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

ested stakeholders to evaluate these recommen- testing, or withholding the results of those tests dations and, at the same time, enable them to from patients and clinicians. Therefore, a more contribute to changes in policy and improve- practical, ethically acceptable approach to vali- ments in programme performance both glo- dating the revised guidelines is to measure only bally and nationally. operational performance, rather than diagnostic test performance, and randomize only facilities or populations, not individuals. Hypotheses Depending on country-specifi c factors, imple- Compared with settings implementing the 2003 mentation of the revised guidelines may require guidelines (existing practice), settings imple- revision of national guidelines and new logistic menting the revised guidelines of 2006 are and technical arrangements, including human expected to display the following characteris- resources and infrastructure development. tics. Therefore, in many countries, the revised rec- 1. A larger proportion of tuberculosis suspects ommendations will probably be implemented diagnosed with smear-negative tuberculosis. in stages. This will facilitate evaluation by interested parties and stakeholders and help to 2. A smaller proportion of smear-negative pul- defi ne intervention and nonintervention set- monary and extrapulmonary tuberculosis tings within a country as part of the scale-up patients who die before treatment is com- process. pleted. 3. A smaller proportion of tuberculosis suspects who die before completion of the diagnostic Overall design evaluation or within two months of initial The suggested overall design for the study is a contact with health services for tuberculosis prospective, observational study. The justifi ca- diagnosis. tion for this approach is as follows. 4. A shorter time-lag between onset of cough 1. Prospective: Historical data, such as medi- and treatment for tuberculosis, and between cal chart reviews, could be used to compare date of initial contact with health services outcomes, but such data may lack the detail for tuberculosis diagnosis and beginning of required to answer important public health antituberculosis treatment. questions about patient and provider sat- 5. A larger proportion of patients and providers isfaction or to compare costs. Similarly, in satisfi ed with the speed and quality of diag- many countries, rapid advances are occur- nostic services. ring in HIV care and treatment, which could 6. A larger proportion of patients with a com- affect the frequency of diagnosis of tubercu- plete diagnostic evaluation, including two losis, the types of disease diagnosed and the sputum smears, chest radiography and spu- outcomes of patients treated for tuberculosis. tum culture. Conducting this study prospectively will help control for these factors.

Study design and procedure 2. Observational: The study will involve no experimental diagnostic tests or medicines. The highest-quality evidence would come from Clinical care will be implemented according a randomized clinical trial in which individual to existing national guidelines. Even if a coun- patients would be enrolled and then randomly try wanted to implement the revised guide- assigned to either the revised or the existing rec- lines nationwide, implementation would be ommendations and algorithms. However, even likely to occur in stages. The observational if such a trial was technically and ﬁ nancially study design therefore allows for both imple- feasible, it would be extremely difﬁ cult to jus- mentation of guidelines and a quasi-experi- tify ethically, as it would involve failing to per- mental assessment of impact. For the reasons form chest radiography, sputum culture or HIV described above, a randomized clinical trial

30 ANNEX. PROTOCOL FOR OPERATIONAL EVALUATION OF THE REVISED RECOMMENDATIONS

is not an ethically appropriate design. Involv- sites implementing the revised guidelines), all ing multiple centres will both be useful and clinicians (physicians, clinical ofﬁ cers, nurses allow for comparison of the impact on pro- and other clinical staff) will receive training grammes. on the revised guidelines, basic diagnosis and management of tuberculosis and HIV, and com- Study description pletion of study documents. Supplies and equipment necessary for the implementation of the A “setting” denotes either an individual health revised guidelines will be installed and relevant facility or an administrative area (e.g. district) staff trained in their use. To reduce the potential that contains multiple health facilities. Set- bias in study outcomes that may be associated tings will be included in the study only if they with such training, nonintervention sites will are already following the WHO-recommended undergo similar training focusing on the 2003 Stop TB strategy, which includes standardized guidelines. recording and reporting of tuberculosis cases, and if they are implementing either the existing (2003) or the revised (2006) recommenda- Suggested studies tions. The settings implementing the revised Within each setting (intervention and nonin- recommendations should also be required to tervention), a number of evaluations will be implement the revised recording and report- conducted to measure input, process, output, ing formats according to national guidelines. outcome and impact of the revised guidelines. (The relevant documents are available from the These suggested studies can be conducted either WHO web site: http://www.who.int/tb/err/en/ as independent studies or as part of a larger index.html). Ideally, settings will be allocated in study, depending on the local context and the a concealed, randomized process; but because interests of research groups. Table A1 below the study is being conducted in a programme describes the types of study and the indicators context, other factors may need to be consid- they will measure. ered when selecting sites, including availability of personnel, infrastructure and budget. For example, the availability of necessary tests and Study 1: assessment of costs services (e.g. HIV testing, chest X-ray, sputum A standardized instrument to measure costs culture, etc.) would affect the selection of an should be developed. The purpose of this intervention site. instrument should be to measure the cost of the Likewise, it may not be possible to select non- human resources development and infrastruc- intervention sites randomly. There may be sub- ture required to implement the revised guide- stantial variations in practice across and within lines. As far as possible, identical information settings that need to be considered for the will be obtained from the sites implementing selection. In those countries that have already the 2003 guidelines, in order to permit an esti- adopted the revised guidelines as their national mate of costs associated with routine practice. guidelines, those settings which have not imple- The study instrument will be also used, if neces- mented the revised guidelines for the duration sary, to abstract data from ﬁ nancial records. of the evaluation may be selected as noninter- Minimum cost components to be measured vention sites. The ideal nonintervention site will in this instrument should include costs related have a standardized approach to tuberculosis to: training, study materials, personnel time diagnosis: its practice should be consistent with (including health care workers and trainers), national guidelines and, if possible, with the transportation of persons and specimens, con- 2003 WHO guidelines. struction, equipment, supplies, reagents and Before data collection begins, both intervention consumables, standard diagnostic procedures sites and nonintervention sites should undergo (e.g. microscopy, radiography, culture, HIV test- training in conducting the evaluation of the ing), antituberculosis treatment, antibiotics pre- revised guidelines. For intervention sites (i.e. scribed for bacterial infection.

31 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

Table A1. Study tools to be used during evaluation, populations to be studied, and indicators to be measured Study Study Indicator no. Tool population type Indicators measured 1 Questionnaire Public health offi cers Input Cost of human resources development to implementing study; implement guidelines (e.g. training, clerks maintaining staffi ng, materials) fi n a n c ial records Input Cost of infrastructure needed to implement guidelines (e.g. equipment, supplies, construction, transportation) 2 Questionnaire Tuberculosis Process Satisfaction of patients with speed and suspects attending quality of services, as measured through health facilities patient survey 3 Questionnaire Health care providers Process Satisfaction of providers with clinical working at health practice guidelines, as measured through facilities provider survey 4 Case-report Tuberculosis Output Proportion of pulmonary tuberculosis form suspects attending suspects with at least two sputum smears health facilities collected, a chest radiograph performed, a sputum culture performed Output Proportion of pulmonary tuberculosis suspects who complete diagnostic evaluation Output Proportion of pulmonary tuberculosis suspects (a)who die before diagnostic evaluation completed; (b) before beginning tuberculosis treatment; (c) within two months of initial contact with health services for tuberculosis diagnosis Output Proportion of pulmonary tuberculosis suspects with known HIV status Output Proportion of pulmonary tuberculosis suspects given a diagnosis other than tuberculosis 5 Case-report Tuberculosis patients Outcome Proportion of pulmonary tuberculosis form treated in health cases diagnosed as smear-negative facilities Outcome Proportion of pulmonary tuberculosis cases with known HIV status Outcome Days between onset of cough and treatment for tuberculosis; days between initial contact with health services for tuberculosis diagnosis and initiation of tuberculosis treatment Impact Proportion of pulmonary tuberculosis cases who have died two months into treatment and six months into treatment (or at end of treatment), stratifi ed by smear status Impact Proportion of pulmonary tuberculosis cases who complete treatment, stratifi ed by smear status

32 ANNEX. PROTOCOL FOR OPERATIONAL EVALUATION OF THE REVISED RECOMMENDATIONS

No statistical sampling will be performed for this 2. Study staff will present tuberculosis suspects study. No informed consent will be obtained, with a consent card or form, informing them because no personal, sensitive or health-related that a survey is being conducted about patient information will be collected. satisfaction. Staff will explain to patients that they will be asked to provide contact information so that they can be contacted in one Study 2: patient satisfaction survey month to determine how satisfi ed they were A survey of tuberculosis suspects will be con- with their tuberculosis evaluation. Patient ducted to determine their satisfaction with the consent or refusal to participate in the survey diagnostic process in facilities implementing will be recorded in the register. the revised guidelines, compared with those 3. After two months, study staff will review the implementing the 2003 guidelines. list of consenting patients and, according to Case defi nition: A pulmonary tuberculosis sus- the established sampling criteria, attempt pect will be defi ned as any person not currently to contact tuberculosis suspects who were receiving antituberculosis treatment and with- evaluated during the fi rst month of the study. out a current diagnosis of tuberculosis, with Subsequent reviews of the list will occur every cough >2 weeks duration or sputum collected month. for AFB smear microscopy at the request of a 4. Study staff will attempt to contact patients clinician. to perform the survey. The survey may be Inclusion criteria: All persons meeting the case administered by telephone or in person, defi nition who seek health care at a participat- depending on the logistics at each site. If a ing facility during the enrolment period, who patient cannot be contacted, patients will agree to be contacted one month after their fi rst continue to be selected from the list of con- diagnostic evaluation as a tuberculosis suspect. senting tuberculosis suspects, following the same selection procedure. If that procedure Exclusion criteria: Persons not meeting the cannot be followed, an alternative procedure case defi nition and inclusion criteria, or persons for randomly selecting patients will be identi- meeting both the case defi nition and inclusion fi ed; that procedure will be documented. criteria who are aged <15 years, cannot be contacted or refuse to participate, or who die before Consent: Patients will be asked to consent to par- they can be interviewed. ticipating in the survey. They have to be assured that consenting or refusing to participate in the Estimated number of participants and sam- survey will have no effect on the treatment they pling: At least 200 persons in total will be inter- receive. Consent will be verbal and brief, because viewed, 100 from intervention and 100 from no sensitive health care information is being nonintervention sites over a six-month period. collected and such surveys are part of routine This sample size and study period are suggested health care practice in many settings. Contact for convenience and practicality. However, more information for tuberculosis suspects, such as accurate sample-size estimations and study peri- address or telephone, may already be collected ods may be employed, based on the local tuber- routinely for fi nancial, administrative or public culosis and HIV epidemiology. If the estimated health reasons. In that situation, individual sites number of suspects is >200, an appropriate and will decide whether it is ethically appropriate to uniform method for sampling the population use that information for selecting tuberculosis will be chosen for both intervention and nonin- suspects for the satisfaction survey. tervention sites. Study instruments: A study instrument will Enrolment procedure: be developed to meet the needs of the site con- 1. Persons presenting at the health care facil- cerned. Minimum data elements to be included ity will be identifi ed as tuberculosis suspects are patient ratings of: speed of service; compre- using routine criteria for that facility, and a hensiveness of service; affordability of service; register will be maintained of all tuberculosis perception of attention to individual needs; suspects (see below). overall satisfaction with service.

33 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

Study 3: health care provider survey about satisfaction with the process for diagnosis A survey of health care providers will be con- of tuberculosis in HIV-infected patients. ducted to determine their satisfaction with the clinical practice guidelines. Study 4: tuberculosis suspect outcome Case defi nition: A health care provider will review be defi ned as any person employed by a health Case-report forms will be completed for all care facility and involved in the clinical care of tuberculosis suspects to measure programme tuberculosis patients. outputs from implementation of the clini- Inclusion criteria: All health care providers, cal practice guidelines. Depending on existing including nurses, clinical offi cers, physicians or practices at participating health facilities, it other care providers, who work in facilities par- is possible that no new data collection instru- ticipating in this study during the study period ments or procedures, other than those already and are involved in the diagnosis or treatment used routinely, will be needed for this compo- of tuberculosis. nent of the study. Estimated number of participants and statisti- Case defi nition: The defi nition of a tuberculosis cal sampling: The number of participants will suspect is given in the section “Patient satisfac- vary depending on the size of the participating tion survey” above. facilities. If possible, all health care providers Inclusion criteria: All persons meeting the case will be studied and there will be no statistical defi nition who seek health care at a participat- sampling. ing facility during the enrolment period. Enrolment procedure: Exclusion criteria: None. 1. A supervisor at the health care facility will Estimated number of participants and sta- prepare a list of all health care providers tistical sampling: The number of participants meeting the inclusion criteria. will vary depending on the size of participat- 2. At the beginning of the study period, the ing facilities. The minimum sample size needed questionnaire will be sent to all health care to demonstrate a difference between facilities providers. implementing the 2003 guidelines (i.e. existing practice) and the 2006 guidelines will be calcu- 3. At the end of the study period (i.e. after six lated on the basis of the epidemiological situa- months), the same questionnaire will be tion and baseline programme performance in readministered. participating sites. Consent: Health care providers will be asked to consent to participating in the survey. Consent Enrolment procedure: will be written. It will be brief, because no sensi- 1. Persons presenting at the health care facil- tive information is being collected. ity will be identifi ed as tuberculosis suspects using routine criteria for that facility and a Study instruments: A study instrument will be register will be maintained of all tuberculosis developed. Minimum data elements will include: suspects. basic questions about tuberculosis knowledge, attitudes and practices; basic questions about 2. Patients will be informed generally by written HIV knowledge, attitudes and practices; ranked signs or posters that the clinic is participating measurement of availability and quality of smear in a study, but individual patients will not be microscopy and chest radiography; patient per- asked to provide informed consent (see justi- ception of the diagnostic process; availability, fi cation below). speed, quality; individual perception of cur- 3. Case-report forms will be collected for all rent guidelines in the facility for smear-negative tuberculosis suspects, using a separate study tuberculosis diagnosis; feasibility, speed, qual- form or a modifi cation of existing clinical ity, overall satisfaction; open-ended questions records (see below).

34 ANNEX. PROTOCOL FOR OPERATIONAL EVALUATION OF THE REVISED RECOMMENDATIONS

4. For patients who do not complete the diag- clinical practice guidelines. Depending on exist- nostic evaluation, health care facilities will ing practices at participating health facilities, it use existing contact information (e.g. tel- is possible that no new data collection instru- ephone number, address, treatment sup- ments or procedures, other than those already porters) to contact patients two months after used routinely, will be needed for this compo- their initial contact with health services for nent of the study. tuberculosis diagnosis to determine whether Case defi nition: A tuberculosis patient will be they are still alive and, if they have died, when defi ned as any person diagnosed with tuber- death occurred. culosis and advised to begin antituberculosis Consent: Tuberculosis control programmes medication. The defi nitions of smear-positive vs routinely collect and review the medical records smear-negative tuberculosis will depend on the of tuberculosis suspects to assess programme guidelines being implemented at the study site performance, e.g. measuring the number of spu- (the 2003 guidelines – i.e. existing practice – or tum specimens collected. This process does not the 2006 guidelines). involve informed consent, because no patient Inclusion criteria: All persons meeting the case identifi ers are collected and data are used spe- defi nition who are diagnosed with tuberculosis cifi cally to evaluate and improve programme at a participating facility during the enrolment performance. Although the number of data period. components being reviewed is likely to be greater than usual, the process and intent are similar. Exclusion criteria: Patients will be excluded if they are registered as “transfer in”, “treat- Study instruments: Health care facilities rou- ment after default”, “treatment after failure” or tinely collect standard data about patients. The “chronic”. extent of such routine data collection varies depending on the facility and provider practice. Estimated number of participants and sta- For this evaluation, health care forms will be tistical sampling: The number of participants modifi ed (or nationally revised recording and will vary depending on the size of participat- reporting formats will be used) to include the ing facilities. The minimum sample size needed following minimum data elements, or a separate to demonstrate a difference between facilities study-specifi c form will be used, depending on implementing the 2003 guidelines (i.e. existing each site’s preference: unique tuberculosis sus- practice) and the 2006 guidelines will be calcu- pect identifying number; age; sex; district; date lated on the basis of the epidemiological situa- when fi rst presented at clinic; cough; date when tion and baseline programme performance in cough began; other symptoms; HIV diagnosis participating sites. (status: positive/negative/unknown and date of Enrolment procedure: HIV diagnosis); sputum smears (date, results); 1. Patients will be registered and begin anti- sputum culture (date, result); chest radiograph tuberculosis treatment following the routine (date, fi ndings); antibiotics taken before visit practice at the health care facility concerned. and after initial visit (prescribed/self-procured/ name of antibiotic and dosage); date of tubercu- 2. Patients will be informed generally by written losis diagnosis; fi nal diagnosis (if not diagnosed signs or posters that the clinic is participating with tuberculosis) and date of last evaluation in in a study, but individual patients will not be clinic. asked to provide informed consent (see justifi cation below). 3. Data from the tuberculosis register and Study 5: tuberculosis patient outcome patient records will be abstracted during the review course of the study. No specifi c procedures Case-report forms will be completed for all will be used for enrolment or withdrawal tuberculosis patients to measure programme from the study, since patients will be treated outcome and impact of implementation of the according to routine practice.

35 RECOMMENDATIONS FOR HIV-PREVALENT AND RESOURCE-CONSTRAINED SETTINGS

Consent: See section on “Consent” under Study study to evaluate the impact of implement- 4 above. ing the revised recommendations, but other changes may occur in the health system at the Study instruments: Health care facilities rou- same time, e.g. wider availability of antiretrovi- tinely collect standard data about patients. The ral treatment or active case-fi nding for HIV or extent of such routine data collection varies tuberculosis that identifi es patients at an earlier depending on the facility and provider practice. stage of either disease. Such events will need to If necessary, health care forms will be modi- be considered when interpreting fi ndings from fi ed to include the following minimum data the studies. elements, or a separate study-specifi c form will be used, depending on each site’s preference: unique, random tuberculosis suspect identify- Identifying, managing and reporting ing number, tuberculosis patient registration adverse events identifying number; tuberculosis registration date; HIV clinical stage; CD4 count; dates and Adverse events are common during treatment of dosage of co-trimoxazole prescribed; dates, HIV-associated tuberculosis, including adverse dosage and regimen of antiretroviral treatment; drug reactions, hospitalization and death. As antituberculosis treatment regimen; presence part of routine public health practice, tuber- or absence of adverse events during antituberculosis control programmes maintain direct culosis treatment; sputum conversion result (if communication with patients throughout the smear-positive at the beginning of treatment); course of treatment. During this observational treatment outcome and date of treatment out- study, such events will be handled according to come. routine public health and clinical practice. Par- ticipating study sites are all public facilities to which the patients have access because they have Study timeline been registered for antituberculosis treatment. Otherwise, patients will not be exposed to any The exact duration of the study will depend on physical or psychological risks beyond those the number of settings involved, the total volume normally encountered during routine clinical of patients in each setting and local formalities care. (such as ethical clearance) of the stakeholders carrying out the evaluation. Expedited implementation of the evaluation is highly recom- References mended and desirable. • Treatment of tuberculosis: guidelines for national programmes, 3rd ed. Geneva, World Health Reimbursements and incentives Organization, 2003 (WHO/CDS/TB/2003.313). http://www.who.int/tb/publications/cds_tb_ Participants will receive no formal reimburse- 2003_313/en/index.html (accessed 29 October ments or incentives as part of t h i s st udy, a lt houg h 2006). sites are permitted to provide modest incentives, such as food, for patients who attend follow-up • International standards for tuberculosis care. The Hague, Tuberculosis Coalition for Techni- visits, if such incentives are part of routine care. cal Assistance, 2006.

Data handling and analysis Statistical methodology, data collection, planned tables and ﬁ gures may vary depending on the investigators conducting the evaluation. Pub- lic health programmes related to tuberculosis and HIV are changing rapidly throughout the world. The study is designed as an observational