Introduction
HIV/AIDS is now the leading cause of death in sub-Saharan Africa with 2.4 million deaths in 2005 for a worldwide toll of 3.1 million [1]. Whereas the mortality has significantly decreased in the last decade in developed countries thanks to the use of HAART, access to HAART is still low in Africa owing mainly to low access to treatment (only 11% of patients in urgent need were receiving it as of June 2005) [2,3]. Mortality in patients on antiretroviral therapy in Africa has been mostly reported for the first year of treatment [4–9]. In Africa, the Senegalese antiretroviral drug access initiative (Isaarv) was one of the first to begin in August 1998 and previous reports have underlined similar results in terms of virological and immunologic responses, adherence and emergence of resistant strains to those of Western countries [10,11]. However, a higher mortality rate than in European patients enrolled in the same period was reported [2,12]. In addition, data on causes of deaths of patients on HAART in Africa are scarce and, in the Senegalese initiative, have only been reported among the very first 96 patients included [11].
We explore here in detail the mortality pattern and likely causes of death on a longer follow-up and a larger number of patients enrolled in the Isaarv cohort.
Methods
Design
Between August 1998 and April 2002, 404 HIV-1-infected adult patients (age > 15 years) were enrolled in the Senegalese antiretroviral drug access initiative (Isaarv) in Dakar and included in an observational cohort. Of these, 80 patients participated for a period of 18 months in two clinical trials and were enrolled between January 2000 and April 2001 if they were antiretroviral therapy (ART)-naive and presented a CD4 cell count < 350 cells/μl, a plasma viral load > 30 000 copies/ml and a Karnofsky score ≥ 70% [13,14]. The 324 other patients were enrolled between August 1998 and April 2002 if they fell into one of the following categories: (1) for asymptomatic patients: CD4 cell count < 350 cells/μl and plasma viral load > 105 copies/ml and, from October 2000 CD4 cell count < 200 cells/μl; (2) for paucisymptomatic patients: CD4 cell count < 350 cells/μl and plasma viral load > 104 copies/ml and, from October 2000 CD4 cell count < 200 cells/μl; and (3) for symptomatic patients: Karnofsky score ≥ 70%, irrespective of CD4 cell count or plasma viral load, and free of major opportunistic infections. The initial ART regimen was a triple-drug combination [two nucleoside reverse transcriptase inhibitors (NRTIs) + one non-nucleoside reverse transcriptase inhibitors (NNRTI) or one protease inhibitor (PI)] but 18 patients received two NRTIs alone before May 2000 (4.4%). ART was provided free of charge for patients in clinical trials or commensurate with their incomes for the other patients, but became free for all patients as of December 2003. A pharmacist delivered the drugs every month or 2 months and recorded adherence. After the pre-enrolment and enrolment visits, which included routine laboratory evaluations, CD4 cell count and HIV plasma viral load, patients were examined 1 month later and thereafter at least every 2 months. Follow-up biological evaluations were carried out every 6 months.
Laboratory procedure
The CD4 cell counts were assessed using the FACSCount System (Becton Dickinson, Le Pont de Claix, France) and plasma viral load using either the Amplicor HIV-1 1.5 or 2.0 assay (Roche Molecular Systems, Meylan, France) or the Bayer bDNA HIV-1 Quantiplex assay 2.0 or 3.0 (Bayer Diagnostics, Aulnay s/Bois, France). The hepatitis B surface antigen (HBs Ag) detection with confirmatory test (IMx HBs Ag; Abbott, Rungis, France) and two hepatitis C virus (HCV) enzyme immunoassay screening tests (IMx HCV version 3.0; Abbott) were performed at the enrolment visit. All of these tests were performed in Dakar.
Monitoring of the cohort
Research assistants met periodically the physicians taking care of the patients to fill standardized research files. All recorded data were made anonymous. Data were subject to a double-keyboard entry with validation checks using custom-developed software. Every month, the research assistants, the pharmacist, one anthropologist, social workers and members of associations for people living with AIDS met to trace patients not seen at a scheduled visit for more than 2 months or patients who were 2 weeks late at drug dispensation. Repeated phone calls were made and several home visits were organized to identify deaths or to discuss reasons for non-adherence. These discussions helped to restart treatment among several missing patients. After 6 months without news or willingness to resume treatment, contacts were ceased. The confidentiality regarding HIV status was strictly observed during any contact.
Ascertainment of the deaths and their likely underlying causes
For deaths that occurred in the three main national hospitals in Dakar where patients were followed or, for a few cases, in remote health centres, information regarding the circumstances and likely causes of death was retrieved from the medical charts, except in two cases.
If death occurred at home with no informative data on the circumstances of the death in the research file, attempts were made to identify the likely causes of death by verbal autopsy, a method that relies on a post-mortem interview of the close relatives who assisted the patient. The questionnaire was derived from the all-causes of death questionnaire already in use in rural Senegal and adapted by infectious diseases specialists [15]. It comprised open questions on signs just before death and closed questions on symptoms with a particular orientation towards AIDS-related causes and adverse effects of ART. The respondent(s) was/were identified by the social workers and the questionnaires passed by a physician along with the social worker, either at home or at the hospital, according to the respondent's preference. The questionnaires were then merged with the data available from the follow-up database to build a clinical history file which was submitted independently to two clinicians to assign a likely underlying cause of death. In case of discordance, a review committee, made up of five different clinicians, reviewed the files to reach a consensus on a final diagnosis. For analytical purpose, morbid entities of infectious origin were grouped by the main organic system involved. However, due to their importance, mycobacterial infections were kept in a single category.
Data analysis
Data were censored at the date of death or at the date of the most recent visit (medical examination, biological assessment, ART dispensation) as of 30 September 2005. Survival probabilities up to 60 months after starting HAART were estimated by the Kaplan–Meier method and risk factors of death by Cox proportional hazards models with fixed covariates at baseline. The proportional hazards assumption has been checked using the Schoenfeld residuals (tests and graphs). Stata 9 was used for analyses (College Station, Texas, USA).
The national ethics committee approved this study and the patients gave their informed consent and received a complete information sheet. We were careful never to disclose the serologic status of the deceased patients during interviews with their relatives.
Results
As of 30 September 2005, the 404 patients had been followed up for a median of 46 months [interquartile range (IQR), 32–57 months], accruing 17 980 person-months of observation.
The characteristics of the patients are summarized in Table 1. Ten were dually infected with HIV-1 and HIV-2. They were included at an advanced stage of HIV disease. At baseline, more than half of the patients were already at CDC stage C, 70% presented less than 200 CD4 cells/μl and more than one-third were anaemic and presented less than 1200 total lymphocytes/μl. CD4 cell counts and total lymphocytes counts were correlated (r = 0.4, F test: P < 10−3). Viral load exceeded 5 log copies/ml in 61% of the patients. At initiation of HAART, most of the patients had no previous ART treatment (94.5%), 78.5% received a cotrimoxazole prophylaxis and 11.6% anti-tuberculosis drugs. After 1 and 2 years of HAART, the proportion of patients receiving cotrimoxazole fell to 68 and 49%, respectively, whereas the proportion of anti-tuberculosis treatment increased to 13.4 and 21.1%, respectively.
;)
Table 1: Baseline characteristics and univariate survival analysis, 404 HIV-1-infected patients under HAART, Senegal, 1998–2005.
The median increase in CD4 cell count reached 104 cells/μl (IQR, 40–158; 261 measures) 6 months after starting HAART and 300 cells/μl (IQR, 174–444; 102 measures) after 4 years of treatment. The median decrease in viral load attained was −3.0 log copies/ml (IQR, −3.6 to −1.9; 240 measures) 6 months after starting HAART and did not change substantially thereafter (−2.9 log copies/ml at 4 years; IQR, −3.6 to −1.4; 95 measures). HIV-1 RNA was undetectable in 60% [95% confidence interval (CI), 54–65%; 287 measures] and 56% (95% CI, 45–65%; 97 measures) of the patients at 6 months and 4 years after starting HAART.
Overall, 93 deaths were recorded. Half of the deaths occurred within the first 12 months (IQR, 5–23) after starting HAART and one-third during the first semester. Seven patients were lost to follow-up during the first year (four voluntary withdrawals, one out-country leave, two with no information) and nine during the second year (two voluntary withdrawals, four out-country leave, three with no information). The probability of dying reached 7.2% (95% CI, 5.1–10.2%), 11.7% (95% CI, 8.9–15.3%), 17.4% (95% CI, 13.9–21.5%) and 24.6% (95% CI, 20.4–29.4%) at 6, 12, 24 and 60 months, respectively. Under the hypothesis that the patients lost to follow-up were dead, probabilities of dying were respectively 13.4% (95% CI, 10.4–17.1%) and 21.0% (95% CI, 17.4–25.4%) at 12 and 24 months of follow-up. Mortality rate decreased from 12.5 deaths/100 person-years (95% CI, 9.4–16.7) during the first year of treatment to 6.6/100 person-years (95% CI, 4.3–10.0) during the second year [hazard ratio (HR), 1.9; 95% CI, 1.1–3.1; P = 0.01] and kept decreasing thereafter (4.5, 2.3 and 2.2/100 person-years for years 3, 4 and 5, respectively). As detailed in Fig. 1, the peak in mortality rates was reached within the second semester after starting HAART. Survival functions differed by level of CD4 cells at baseline (log-rank test: P < 10−2) and the cumulative probability of dying at 12 months attained 17.9% (95% CI, 11.5–27.2%), 13.1% (95% CI, 8.9–19.0%) and 5.8% (95% CI, 2.8–11.9%) for less than 50, 50–199 and more than 200 CD4 cells/μl, respectively. Between baseline and last assessment, the median gain in CD4 cell count was null for deceased patients (median = 0; IQR, 0–61), while alive/lost to follow-up patients gained a median of 236 CD4 cells/μl (IQR, 127–397). Their median virologic response was much weaker than among alive/lost to follow-up ones (−0.9 log; IQR, −2.7 to 0 versus −3.0 log; IQR, −3.6 to −1.4) (Mann–Whitney U test: P < 10−4). Log-rank analysis and univariate Cox's model of baseline characteristics showed that age, gender, financial participation, non-naive ART status, cotrimoxazole prophylaxis, protease inhibitor containing regimen, an HIV viral load exceeding 5 log and carriage of HBs Ag or HCV antibodies were not significantly associated with a differential survival (Table 1). Conversely, a clinical stage C versus A/B, a body mass index (BMI) less than 19 kg m−2, a haemoglobin level less than 10 g/dl, a total lymphocyte count less than 1200 cells/μl and a CD4 cell count less than 200 cells/μl were associated with a worse survival (Table 1). A multivariate Cox's model retained a BMI ≥ 19 kg m−2, a haemoglobin level ≥ 10 g/dl and a CD4 cell count ≥ 200 cells/μl as baseline independent predictors of survival (Table 2). Indeed, BMI and haemoglobin were associated with clinical stage (chi-squared test: p < 10−3 and P = 0.02, respectively). A model without CD4 cell count at baseline did not retain the total lymphocytes count as a predictor but BMI and haemoglobin as markers of clinical stage. Figure 2 displays the cumulative probabilities of dying according to baseline predictors and CD4 cell count adjusted for body mass index and haemoglobin level.
Fig. 1: Smoothed hazard rate after starting HAART, 404 HIV-1 patients, Senegal, 1998–2005. CI, confidence interval.
Table 2: Adjusted mortality ratios (95% confidence interval) from Cox's proportional hazards models, HIV-1-infected patients under HAART, Senegal, 1998–2005.
Fig. 2: Kaplan–Meier failure estimates by predictors at baseline: haemoglobin level. (a); body mass index (BMI) (b); CD4 cell count (c); and CD4 cell count adjusted for haemoglobin level and body mass index (d); 404 patients HIV-1 patients on HAART, Senegal, 1998–2005.
Likely underlying causes of death
Forty-seven patients died in the hospital where they were followed (50%), seven in a remote health structure (8%) and 39 at home (42%). In 60 cases (64%) the likely causes of death were abstracted from hospital charts. Twenty verbal autopsies were done (18 deaths at home plus two deaths in a hospital and a remote health structure) with a median recall period of 21 months (IQR, 16–28). Half needed a consensus. In 13 cases, verbal autopsy could not be performed for different reasons [distance (n = 7), no contact (n = 5) or ethics (n = 1)]. A likely cause of death has therefore been assigned in 80 of the 93 deaths (86%) but only 13 diagnoses were biologically confirmed (Table 3). Respiratory tuberculosis and mycobacterial infections explained 18% of the deaths (17 cases, four confirmed). Eight of these deaths occurred during the first year on HAART (47%) and a history of tuberculosis before HAART initiation was found in nine cases (53%). The difference in proportion of anti-tuberculosis treatment at HAART initiation between tuberculosis and non-tuberculosis-related death was not significant (23.5 versus 15.3%). Infections of the central nervous system accounted for an equivalent proportion of deaths (17 cases) and comprised five cases of cerebral cryptococcosis (one confirmed), three cases of toxoplasmosis and three suspected HIV encephalitis. Septicaemia (eight cases, miscellaneous origin) and unspecified infectious diseases accounted also for the same proportion (17 cases). Five patients died from acute liver failure (one chronically infected with hepatitis B, one presenting hepatitis C antibodies and three with unknown hepatitis markers). Three patients on protease inhibitors died from metabolic disorders (diabetes, hyperamylasaemia in a patient receiving didanosine, metabolic acidosis).
Table 3: Likely causes of death, 404 HIV-1-infected patients on HAART, Senegal, 1998–2005.
Discussion
Mortality
This 7-year prospective cohort study provides for the first time rates and causes of death in patients receiving HAART over a long time and underlines the treatment's effectiveness in reducing the mortality in Africa. Following a peak of death in the first year of treatment (50% of the deaths before 12 months), the mortality rate decreased regularly from 12.5 per 100 person-years in the first year to 0.9 per 100 person-years in the fifth year. Efforts were made to trace the patients lost to follow-up and the status of only two and four patients remained unknown during the first and second year of follow-up; therefore, the ascertainment of deaths was almost complete. Upper bounds of mortality considering all patients lost-to-follow up as death were 13.4 and 21% at 1 and 2 years, respectively, after HAART initiation and should be considered as conservative estimates.
During the early phase of access to ART in Africa (Drug Access Initiative), a higher mortality after the first year of treatment has been reported from Uganda (26%) or Ivory Coast (16%); however, all the patients were not on HAART at that time [4,6]. Later, among patients receiving exclusively HAART, a lower mortality was reported after 1 year of treatment in Cameroon (8%) or after 2 years in South Africa (13.6%) [7,8]. Recently, the analysis of pooled databases on response to HAART from 18 developing countries (ART-LINC collaboration), provided a first-year risk of dying of 5.8%, which is twice as low as in the Senegalese cohort but 3.5 times as high as in developed countries [16,17]. However the mortality in ART-LINC is likely to be higher when the status of patients lost to follow-up is assessed. Indeed, a higher attrition rate has been observed among the most immuno-deficient patients [18]. As observed in ART-LINC and other cohorts, the mortality was strongly correlated with the CD4 cell count or clinical stage at baseline [6,19]. The median CD4 cell count at baseline in the ART-LINC and Isaarv cohorts did not differ 108 cells/μl; IQR, (37–210) versus 128 cells/μl; IQR, (54–217)] but the probability of dying by level of CD4 cell count were roughly from two to three times as high in the Senegalese cohort as in the pooled ART-LINC cohorts. Hence, the advanced stage of the Senegalese patients at HAART initiation only partially explains the first-year mortality and, as previously noted by Laurent et al., the limited access to diagnosis and treatment of infectious diseases had a negative impact on survival [12].
As already reported in the Isaarv cohort, anaemia at baseline appeared to be a strong predictor of mortality, independent of CD4 cell count [12]. While total lymphocyte count was correlated with CD4 cell count at baseline and could be used as a surrogate to enrol patients in the absence of CD4 cell count, it did not appear to be a predictor of mortality after adjustment on haemoglobin level. A fall in viral load of more than 0.9 log without a gain in CD4 cells count was observed in half of the deceased patients, indicating a sub-optimal immuno-virological response. However, most of the dead patients remained under HAART until the last quarter and adherence among the first patients included in the Isaarv cohort exceeded 90% during the first year, warranting a more detailed analysis of adherence at the initial stage of HAART [20]. We did not specifically study the role of an immune reconstitution inflammatory syndrome (IRIS) in the early deaths after HAART initiation, and particularly tuberculosis-related deaths. However, IRIS could have played a substantial role since patients with a low CD4 cell count at baseline, an underlying viral or mycobacterial opportunistic infection and a fall in viral load are at the greatest risk of developing such a syndrome [21]. There is not yet a standard definition of IRIS but a rise in CD4 cell count is not considered to be a key element due to the delay between virologic and immunological responses. In our cohort, one-third of the deaths occurred within the first semester of HAART, before a rise in CD4 cell count could be recorded.
Causes of death
In the absence of medical records documenting the cause of death, verbal autopsy can be helpful to classify deaths in broad categories. It does not permit the assignment of a precise cause of death, but instead helps to indicate the most likely cause of death given the available information [22]. In the context of AIDS, verbal autopsy has been used to estimate the distribution of AIDS and non-AIDS deaths in rural Africa with sufficient confidence but has never been used to identify causes of death among HIV patients receiving HAART [23]. Therefore, the results presented here must be interpreted with caution. A likely cause of death has been assigned in 86% of the deaths and in 75% of these deaths, information came from hospital records. Verbal autopsy was therefore used to assign a likely cause of death in only 21% of the deceased patients.
Tuberculosis and atypical mycobacteriosis were leading causes of death, as seen in Ivory Coast or West Indies [6,24–26]. The number of deaths attributable to tuberculosis was likely to be underestimated given the late stage of HIV infection when patients were enrolled in the ART programme, increasing the probability of smear-negative sputum and atypical chest X-ray appearance. Indeed, Agyei, in Ghana, showed that 41% of the confirmed tuberculosis-related deaths were not suspected before death [27]. The availability of new and automated mycobacterial culture systems will certainly improve the management of the patients by considerably reducing the delay in detection. In our study, it is noteworthy that a past history of tuberculosis was found in 53% of the tuberculosis-related deaths, opening the discussion for a secondary tuberculosis chemoprophylaxis (isoniazid) to decrease the risk of recurrence [25,28,29]. The role of the immune reconstitution disease in the high incidence of fatal mycobacterial infections remains to be established [21].
Infectious diseases with a neurological tropism appeared as frequent as mycobacterial infections. Cerebral cryptococcosis was the most cited neurological infection, as it was in AIDS patients in Dakar before the HAART era [30]. In Uganda, this infection appeared as the first cause of death and the second reason of admission after tuberculosis [31].
End-stage liver diseases of chronic hepatitis infections appear now to be a leading cause of death in developed countries but its rank in the sub-Saharan settings needs to be clarified [32]. Hepatitis B is highly endemic in Senegal with 15–20% of the population carrying HBs Ag and less than 1% of blood donors in Dakar exhibited antibodies to HCV in 2001 [33,34]. The proportion of co-infected patients with HBV is roughly similar in the Isaarv cohort and the co-infected patients, receiving lamivudine and potential hepatotoxic drugs, should be given particular attention. A switch to a tenofovir–emtricitabine-containing regimen providing two agents with dual activity against HIV and HBV should be considered [35].
This study underlines the early mortality pattern after HAART initiation, reports long-term trend in mortality and highlights the leading role of mycobacterial and neurological infections in the causes of death. Efforts towards an earlier initiation of HAART in the course of HIV infection and a better approach to diagnosis and management of tuberculosis and opportunistic infections will be the next step to limit the still too high mortality.
Acknowledgements
We sincerely thank all the relatives of the deceased patients who agreed to answer sensitive questions. We are grateful to M. Ndiaye, A. Thiam, D. Bal who organized the interviews with the relatives and M. Diouf Lakh for her assistance with the medical charts (Department of Infectious Diseases, Fann University Teaching Hospital, Dakar, Senegal).
Sponsorship: This study was funded by the National Agency for AIDS and Hepatitis Research (ANRS, France) the European Union UE (contract B7-6211/99/005) and the Research Institute for Development (IRD, France).
References
1. UNAIDS/WHO.
AIDS Epidemic Update: 2005. Geneva: UNAIDS/WHO; 2005.
2. Mocroft A, Ledergerber B, Katlama C, Kirk O, Reiss P, d'Arminio MA,
et al. Decline in the AIDS and death rates in the EuroSIDA study: an observational study. Lancet 2003; 362:22–29.
4. Weidle PJ, Malamba S, Mwebaze R, Sozi C, Rukundo G, Downing R,
et al. Assessment of a pilot antiretroviral drug therapy programme in Uganda: patients' response, survival, and drug resistance. Lancet 2002; 360:34–40.
5. Djomand G, Roels T, Ellerbrock T, Hanson D, Diomande F, Monga B,
et al. Virologic and immunologic outcomes and programmatic challenges of an antiretroviral treatment pilot project in Abidjan, Cote d'Ivoire. AIDS 2003; 17(suppl 3):S5–S15.
6. Seyler C, Anglaret X, Dakoury-Dogbo N, Messou E, Toure S, Danel C,
et al. Medium-term survival, morbidity and immunovirological evolution in HIV-infected adults receiving antiretroviral therapy, Abidjan, Côte d'Ivoire. Antivir Ther 2003; 8:385–393.
7. Coetzee D, Boulle A, Hildebrand K, Asselman V, Van Cutsem G, Goemaere E. Promoting adherence to antiretroviral therapy: the experience from a primary care setting in Khayelitsha, South
Africa. AIDS 2004; 18(suppl 3):S27–S31.
8. Laurent C, Kouanfack C, Koulla-Shiro S, Nkoue N, Bourgeois A, Calmy A,
et al. Effectiveness and safety of a generic fixed-dose combination of nevirapine, stavudine, and lamivudine in HIV-1-infected adults in Cameroon: open-label multicentre trial. Lancet 2004; 364:29–34.
9. Bourgeois A, Laurent C, Mougnutou R, Nkoue N, Lactuock B, Ciaffi L,
et al. Field assessment of generic antiretroviral drugs: a prospective cohort study in Cameroon. Antivir Ther 2005; 10:335–341.
10. Laurent C, Diakhate N, Gueye NF, Toure MA, Sow PS, Faye MA,
et al. The Senegalese government's highly active antiretroviral therapy initiative: an 18-month follow-up study. AIDS 2002; 16:1363–1370.
11. Desclaux A, Lanièce I, Ndoye I, Taverne B.
The Senegalese Antiretroviral Drug Access Initiative. An Economic, Social, Behavioural and Biomedical Analysis. Paris: ANRS, UNAIDS, WHO; 2004.
12. Laurent C, Ngom Gueye NF, Ndour CT, Gueye PM, Diouf M, Diakhate N,
et al. Long-term benefits of highly active antiretroviral therapy in Senegalese HIV-1-infected adults. J Acquir Immune Defic Syndr 2005; 38:14–17.
13. Landman R, Schiemann R, Thiam S, Vray M, Canestri A, Mboup S,
et al. Once-a-day highly active antiretroviral therapy in treatment-naive HIV-1-infected adults in Senegal. AIDS 2003; 17:1017–1022.
14. Canestri A, Ndir A, Vray M, Sow PS, Mboup S, Niang AF,
et al.
Long-term evaluation of d4t, dl and EFV in antiretroviral naive patients in Senegal. ANRS 12-06/IMEA 012 trial. 2003.
Second IAS Conference on HIV Pathogenesis and Treatment, Paris, July 2003 [abstract 569].
15. Etard JF, Le Hesran JY, Diallo A, Diallo JP, Ndiaye JL, Delaunay V. Childhood mortality and probable causes of death using verbal autopsy in Niakhar, Senegal, 1989–2000. Int J Epidemiol 2004; 33:1286–1292.
16. The Antiretroviral Therapy in Lower Income Countries (ART-LINC) Study Group.
Cohort profile: antiretroviral therapy in lower income countries (ART-LINC): international collaboration of treatment cohorts.
Int J Epidemiol 2005;
34:979–986.
17. Braitstein P, Brinkhof MW, Dabis F, Schechter M, Boulle A, Miotti P,
et al. Mortality of HIV-1-infected patients in the first year of antiretroviral therapy: comparison between low-income and high-income countries. Lancet 2006; 367:817–824.
18. Balestre E, Brinkhof M, Dabis F, Schechter M, Braitstein P, Egger M,
et al.
Short-term incidence and risk factors for being lost to treatment in 16 antiretroviral treatment cohorts in low income countries.
Third IAS Conference on HIV Pathogenesis and Treatment, Rio de Janeiro, July 2005 [abstract no. MoPe11.10C13].
19. Casseb J, Fonseca LA, Veiga AP, de Almeida A, Bueno A, Ferez AC,
et al. AIDS incidence and mortality in a hospital-based cohort of HIV-1-seropositive patients receiving highly active antiretroviral therapy in Sao Paulo, Brazil. AIDS Patient Care STDs 2003; 17:447–452.
20. Lanièce I, Ciss M, Desclaux A, Diop K, Mbodj F, Ndiaye B,
et al. Adherence to HAART and its principal determinants in a cohort of Senegalese adults. AIDS 2003; 17(suppl 3):S103–S108.
21. Shelburne SA, Montes M, Hamill RJ.
Immune reconstitution inflammatory syndrome: more answers, more questions.J Antimicrob Chemother; Advance access published on December 14, 2005.
22. Chandramohan D, Maude GH, Rodrigues LC, Hayes RJ. Verbal autopsies for adult deaths: issues in their development and validation. Int J Epidemiol 1994; 23:213–222.
23. Doctor HV, Weinreb AA. Estimation of AIDS adult mortality by verbal autopsy in rural Malawi. AIDS 2003; 17:2509–2513.
24. Bonard D, Messou E, Seyler C, Vincent V, Gabillard D, Anglaret X. High incidence of atypical mycobacteriosis in African HIV-infected adults with low CD4 cell counts: a 6-year cohort study in Cote d'Ivoire. AIDS 2004; 18:1961–1964.
25. Seyler C, Toure S, Messou E, Bonard D, Gabillard D, Anglaret X. Risk factors for active tuberculosis following antiretroviral treatment initiation in Abidjan. Am J Respir Crit Care Med 2005; 172:123–127.
26. Lewden C, Sobesky M, Cailleton V, Couppie P, Boulard F, May T,
et al. [Causes of death among HIV infected adults in French Guyana and the French West Indies in the era of highly active antiretroviral therapy (HAART)]. Med Mal Infect 2004; 34:286–292.
27. Agyei A, Lartey M.
Spectrum of opportunistic infections causing death in patients with aids in ghana: correlation of clinical diagnoses with autopsy results.XI Conference on Retrovirus and Opportunistic Infections, San Francisco, 8–11 February 2004 [abstract 574].
28. Churchyard GJ, Fielding K, Charalambous S, Day JH, Corbett EL, Hayes RJ,
et al. Efficacy of secondary isoniazid preventive therapy among HIV-infected Southern Africans: time to change policy? AIDS 2003; 17:2063–2070.
29. Harries A, Maher D, Graham S, Eds.
TB/HIV: a clinical manual. Geneva: WHO; 2004.
30. Hovette P, Camara P, Raphenon G, Soko T, Burgel P. [Prevalence of cerebro-meningeal cryptococcosis associated with AIDS in Dakar.]. Presse Med 1999; 28:741–742.
31. Otim TW, Mugyenyi P, Ssali FS, Kityo CM.
Survival analysis of AIDS inpatients at joint clinical research center in Kampala.
XV International AIDS conference. Bangkok, July 2004 [abstract MoPeC3515].
32. Lewden C, Salmon D, Morlat P, Bevilacqua S, Jougla E, Bonnet F,
et al. Causes of death among human immunodeficiency virus (HIV)-infected adults in the era of potent antiretroviral therapy: emerging role of hepatitis and cancers, persistent role of AIDS. Int J Epidemiol 2005; 34:121–130.
33. Coursaget P, Leboulleux D, Soumare M, le Cann P, Yvonnet B, Chiron JP,
et al. Twelve-year follow-up study of hepatitis B immunization of Senegalese infants. J Hepatol 1994; 21:250–254.
34. Etard JF, Colbachini P, Dromigny J, Perrier-Gros-Claude J. Hepatitis C antibodies among blood donors, Senegal. Emerg Infect Dis 2003; 9:1492–1493.
35. Thio CL, Sulkowski MS, Thomas DL. Treatment of chronic hepatitis B in HIV-infected persons: thinking outside the black box. Clin Infect Dis 2005; 41:1035–1040.
Appendix
ANRS 1290 study group
I. Ndoye (Multisectorial AIDS Program, Dakar, Senegal) and E. Delaporte, J.-F. Etard, C. Laurent, B. Taverne, M. Thierry-Mieg, [UMR 145, Research Institute for Development (IRD)/University of Montpellier, Dakar, Senegal and Montpellier, France], M.M. Bitèye, A.B. Dieng, A. Diouf, A. Sarr, L. Zié (Regional Research and Training Center for HIV/AIDS, Fann University Teaching Hospital, Dakar, Senegal), V. Cilote, I. Lanièce (French Ministry of Foreign Affairs, Dakar, Senegal), I. Ndiaye, A. Ndir, C.T. Ndour, C.S. Senghor, P.S. Sow, (Department of Infectious Diseases, Fann University Teaching Hospital, Dakar, Senegal), M. Basty Fall, N.M. Dia Badiane, N. Diakhaté, M. Diallo, L.M. Diouf, N.F. Ngom Guèye, (Ambulatory Care Unit-Croix-Rouge, Fann University Teaching Hospital, Dakar, Senegal), K. Ba Fall, P.M. Guèye, C. Périno (Military Hospital of Dakar, Senegal), P.A. Diaw, H. Diop Ndiaye, S. Mboup, N.C. Touré Kane (Le Dantec University Teaching Hospital, Virology and Bacteriology Laboratory, Dakar, Senegal), K. Sow (AIDS Control Division, Ministry of Health, Dakar, Senagal), K. Diop, B. Ndiaye (Central Pharmacy, Fann University Teaching Hospital, Dakar, Senegal), M. Ciss (Ministry of Health, National Drug Laboratory, Dakar, Senegal), A. Desclaux (Paul Cezanne University, CReCSS, Aix-en-Provence, France).
