Risk factors for active tuberculosis following antiretroviral treatment
initiation in Abidjan
Catherine Seyler 1, Siaka Toure 1, Eugène Messou 1, Dominique Bonard 1,2,
Delphine Gabillard 3, Xavier Anglaret 1,3
Author Affiliations :
1 Programme PAC-CI, Abidjan, Côte d’Ivoire
2 Centre de Diagnostic et de Recherches sur le SIDA (CeDReS), Centre Hospitalier Universitaire
de Treichville, Abidjan
3 INSERM U.593, Université Victor Segalen Bordeaux 2, Bordeaux, France
Reprints and Correspondence :
Dr Xavier Anglaret
INSERM U593 ?
Université Victor Segalen Bordeaux 2
146 rue Léo Saignat
e-mail : Xavier.Anglaret@isped.u-bordeaux2.fr
Running head : tuberculosis under HAART in Abidjan
Descriptor number : 121
This study was supported by the Agence Nationale de Recherches sur le SIDA (ANRS, France),
and the Programme National de Lutte contre le SIDA, les maladies sexuellement transmissibles
et la tuberculose (PNLS/MST/TUB, Côte d’Ivoire) within the collaborative Programme PAC-CI.
AJRCCM Articles in Press. Published on March 31, 2005 as doi:10.1164/rccm.200410-1342OC
Copyright (C) 2005 by the American Thoracic Society.
Rationale: In sub-Saharan Africa: (i) tuberculosis is the first cause of HIV-related mortality; (ii)
the incidence of tuberculosis in adults receiving highly active antiretroviral therapy (HAART) is
lower than in untreated HIV-infected adults but higher than in HIV-negative adults; (iii) factors
associated with the occurrence of tuberculosis in patients receiving HAART have never been
Objective: To look for the risk factors for active tuberculosis in HIV-infected adults receiving
HAART in Abidjan.
Methods: Seven-year prospective cohort of HIV-infected adults, with standardized procedures
for documenting morbidity. We analyzed: (i) the incidence of active tuberculosis in patients who
started HAART; (ii) the association between the occurrence of tuberculosis and the
characteristics of these patients at HAART initiation.
Main results: 129 adults (median baseline CD4 count 125/mm3) started HAART and were then
followed up during 270 person-years (P-Y). At HAART initiation, 31 had a past history of
tuberculosis and none had current active tuberculosis. During follow-up, the incidence of active
tuberculosis was 4.8/100 P-Y (95%CI 2.5-8.3) overall, 3.0/100 P-Y (95%CI 1.1-6.6) in patients
with no tuberculosis history, and 11.3/100 P-Y (95%CI 4.1-24.5) in patients with a history of
tuberculosis (Adjusted Hazard Ratio 4.64, 95%CI, 1.29-16.62, p=0.02).
Conclusion: The risk of tuberculosis after HAART initiation was significantly higher in patients
with a past history of tuberculosis than in those with no tuberculosis history. If confirmed by
others, this finding could lead to assessment of new patterns of time-limited tuberculosis
secondary chemoprophylaxis during the period of initiation of HAART in sub-Saharan African
adults. 248 words
Keywords : HIV; Africa South of the Sahara; HAART; risk-factors; adults; tuberculosis; incidence
In non-industrialized countries the incidence of opportunistic diseases in HIV-infected adults
is likely to decrease with the increasing access to antiretroviral drugs, as previously reported
in industrialized countries (1-3). In sub-Saharan Africa, where tuberculosis (TB) has long been
identified as the first cause of HIV-related mortality in adults (4), one study already showed a
dramatic decrease in the TB incidence in adults receiving highly active antiretroviral therapy
(HAART) as compared to those not receiving it (5). However, though reduced by HAART, the
incidence of active TB continues to be higher in HIV-infected adults receiving HAART than
in HIV-negative patients (5), and TB has been reported to remain the most common cause of
intermediate-term severe morbidity in adults receiving HAART (6). In the HAART era, TB
will thus continue to be an everyday problem for the sub-Saharan African teams in charge of
The aim of this study was to look for the factors associated with the occurrence of active TB
in a cohort of HIV-infected adults receiving HAART in Côte d’Ivoire.
From March 1996 to July 31 2003, 723 HIV-infected adults have been followed up in Abidjan
in the Cotrame ANRS 1203 cohort, the procedures of which have been described previously
(6). Additional detail is provided in an online data supplement. Patients described here are all
patients from the Cotrame cohort who have started HAART. Baseline was the day of HAART
initiation. The protocol of the Cotrame study has been approved by the ethics committee of
the Ivoirian Ministry of Health.
Diagnosis and treatment of tuberculosis
On admission to the Cotrame cohort, patients reporting a history of TB were asked to return
the national program registration card. On admission and during follow-up in the cohort,
including before HAART and after HAART was started, patients with symptoms consistent
with TB were investigated with at least three sputum direct examinations and cultures and one
chest X-ray. Other procedures for investigating signs and symptoms in the Cotrame cohort
centre have been described previously (7). When an episode of active TB was diagnosed, the
treatment was given according to the national TB program recommendations. Additional
detail is provided in the online data supplement.
Definition of TB
The following criteria were required for the diagnosis of active TB.
“Definitive” TB: (i) consistent clinical picture, and (ii) positive culture for Mycobacterium
tuberculosis of sputum sample or bronchoalveolar lavage (“pulmonary” TB) or from normally
sterile body fluid or tissue from a site other than lungs (“extra-pulmonary” TB);
“Presumptive” pulmonary TB: (i) Consistent clinical picture > 30 days, (ii) positive smear for
acid-fast bacilli in sputum or bronchoalveolar lavage sample and/or absence of
microbiological evidence of pneumonia due to any other known pathogen, (iii) unsuccessful
response to standard antibiotic therapy and (iv) successful response to standard
“Presumptive” extra-pulmonary TB: (i) Consistent clinical picture > 30 days, (ii) positive
smear for acid-fast bacilli in normally sterile body fluid or tissue from a site other than lungs
and/or no other microbiological explanation, and (iii) successful response to standard
Definition of prior tuberculosis
Patients were considered to have: (i) a past history of active TB at HAART initiation if they
reported a history of TB on admission to the Cotrame cohort and/or if they had an episode of
active TB which was considered to be cured between their admission to the Cotrame cohort
and HAART initiation; (ii) a prevalent episode of active TB if they had an episode of active
TB which was under treatment on HAART initiation, or if an episode of active TB was
documented after HAART initiation and the date of the first symptoms recorded in the
Cotrame centre medical record preceded the date of HAART initiation; (iii) an incident
episode of active TB if TB was documented after HAART initiation and the date of the first
symptoms was posterior to HAART initiation.
The incidence rate of tuberculosis was defined as the number of patients with incident active
tuberculosis per 100 patient-years (/100 P-Y) of at-risk follow-up. Univariate Cox
proportional hazard regression models for first events were used to study the association
between incident tuberculosis and baseline characteristics (8). A multivariate model was then
used to study the association between incident tuberculosis and all characteristics associated
with tuberculosis with a p<0.25 in a univariate analysis. We decided a priori that the baseline
CD4 count and the past history of TB would be included in the multivariate model even if not
associated with tuberculosis with a p<0.25 in a univariate analysis. Additional detail is
provided in the online data supplement.
During the study period, 129 of the 723 patients participating in the Cotrame cohort started
HAART. Baseline characteristics, initial treatment regimens and follow-up characteristics are
reported in table 1.
History of tuberculosis
At HAART initiation, no patient had current active tuberculosis and 31 patients had a past
history of active tuberculosis, including 27 with one previous episode and four with two
episodes. The last episode was pulmonary in 21 (68%), extra-pulmonary in five (16%) and
both pulmonary and extra-pulmonary in five (16%). The median period between diagnosis of
the last episode and HAART initiation was 2.8 years (Interquartile range [IQR] 1.4-4.7). Of
the 31 previous episodes, 12 were diagnosed and treated in one of the national tuberculosis
program centres before the patients were included in the Cotrame cohort, and 19 were
diagnosed and treated in the Cotrame cohort centre while patients were already participating
in the Cotrame cohort but not yet receiving HAART. All of the 31 previous episodes were
considered as cured after a standard 6-month rifampin-based regimen, according to national
Incident tuberculosis episodes
During follow-up under HAART, 12 patients had 14 episodes of tuberculosis, an overall
incidence of 4.8/100 person-years [P-Y] (95% CI 2.5-8.3). Among the 12 first episodes, the
first symptoms appeared before the sixth month following HAART initiation in three,
between six and 12 months in two, between 12 and 24 months in four, and after 24 months in
three. In the two patients who had two episodes, the symptoms of the second episode
appeared 14 and 15 months after the first episode, respectively. Of the 12 first episodes, six
occurred in patients with no history of tuberculosis (incidence rate 3.0/100 P-Y, 95% CI 1.1-
6.6) and six in patients with a history of tuberculosis (incidence rate 11.3/100 P-Y, 95% CI
4.1-24.5). The probability of remaining free from active tuberculosis was 0.98 (95% CI 0.95–
0.99) at 12 months and 0.95 (0.90–0.99) at 24 months in patients with no TB history, versus
0.90 (0.78–1.00) at 12 months and 0.80 (0.63–0.96) at 24 months in patients with a past
history of TB.
Of the 14 episodes, eight were diagnosed as definitive (pulmonary: seven; pulmonary and
renal: one) and six as presumptive (miliary: one; pleural and pulmonary: one; pulmonary
only: four), including two with positive sputum no smear and culture performed. An
antibiogram was performed in seven of the eight episodes of definitive tuberculosis. The
strain was isoniazide sensitive in six of the seven strains, and rifampicin and ethambutol
sensitive in all strains. Table 2 shows for the 12 first episodes: the date of HAART initiation,
the presence of a past history of tuberculosis together with the interval since the last episode,
the interval from HAART initiation to the first symptoms of the incident episode, the most
recent CD4 count, the pulmonary and/or extra-pulmonary localization, and the diagnosis
certainty. Eleven episodes were considered to be cured after a standard 6-month regimen, and
one patient died before the end of the antituberculous treatment.
In univariate analysis, age, sex, schooling, professional activity, number of people at home,
electricity at home, WHO clinical stage, absolute neutrophil count, platelet count and marital
status were not significantly associated with the risk of tuberculosis with a p≥0.25. These
variables were thus not included in the multivariate analysis. The following baseline variables
were associated with the risk of tuberculosis with a p<0.25 and were therefore included in the
multivariate analysis (table 3) : haemoglobin count (p=0.24), body mass index (p=0.16),
baseline HAART regimen (p=0.17), past history of tuberculosis (p=0.03) and running water at
home (p=0.05). Finally, the baseline CD4 count was not associated with the risk of
tuberculosis in univariate analysis (p=0.91). However, it was one of the variables considered
a priori to be included in the multivariate model, as defined in the method section. In
multivariate analysis, the only variable which remained associated with the occurrence of TB
was a past history of TB. The Adjusted Hazard Ratio of TB was 4.64 in patients with a past
history of TB compared with those with no TB history (95% CI 1.29-16.62, p=0.02) (table 3).
Additional detail on analysis is provided in the online data supplement
This study is the first to look for factors associated with active TB in adults receiving HAART
in sub-Saharan Africa. We followed a cohort of 129 HIV-infected adults receiving HAART
during a median of 26 months in Abidjan. Our main finding was that the risk of active TB
after HAART initiation was significantly higher in patients with a previous history of TB than
in those with no TB history.
There are several points to be made about this finding.
First, all patients with a history of TB had previously been considered to be cured after a
standard 6-month regimen, according to the national program criteria which include
microbiological and radiological control. An active search for tuberculosis prior to initiating
HAART was done in all symptomatic patients, according to the procedures of the study
center. However, such an active search was not done in asymptomatic patients, and one
cannot totally exclude that some cases of incident TB were in fact preexisting cases. Now,
only three of the 12 TB episodes began within the 6 months following HAART initiation,
including two in patients with no TB history. If these three episodes would have been
preexisting cases, this would have tended to minimize the association between newly active
TB and history of TB.
Second, 33% of our TB cases were smear-negative and culture-negative TB, and one cannot
exclude that some of these were not true TB cases. However, acid-fast smear has long been
reported to be more frequently negative in HIV-infected patients than in HIV-negative
subjects (9), and this rate of presumptive cases is in the range of those usually reported in
studies of HIV-associated TB in sub-Saharan African adults (5,10).
Third, the cohort study was not designed to collect information on all HIV diseases risk
factors. Some important TB risk factors, such as the exposition/exposure to TB at home or at
work have not been documented. When finding an association between the new active TB and
a TB history, there could be a confounding effect.
Together with these limitations, this study has several strong points.
The Cotrame cohort was designed to study in the long term the causes of morbidity in HIV-
infected adults. At inclusion, the past history of TB before the inclusion in the cohort was
carefully documented through the forms of the national TB program. Patients were followed a
median of 41 months before HAART was started. Standardized procedures of investigations
and diagnostic criteria were applied. An event documentation committee approved all
diagnoses (6,7). As a result of all these points, the diagnosis of active TB under HAART and
the past history of TB before starting HAART could be documented over a long period and
under favorable conditions which are uncommon in most HIV care centers in sub-Saharan
Africa. The association between TB incidence and past history of TB that we observed is thus
unlikely to be due to misdiagnoses or misclassifications. We believe that this association
should be considered as a serious hypothesis to be borne out by more specific studies.
If confirmed, this could have implications in terms of TB secondary prophylaxis. Though with
proven medium-term benefits (11-13), TB secondary prophylaxis is currently not routinely
applied in sub-Saharan Africa, partly because of the fear of potential long-term risks
associated with a lifelong treatment with one or various antituberculous drugs, both for the
individual and for the population. The current increasing access to HAART could either end
the need for TB secondary prophylaxis, or to document the need for it. On the one hand,
HAART leads to a dramatic increase in the CD4 count and to a consecutive noticeable
decrease in the incidence of TB. Thus, treating HIV-positive TB patients with HAART could
now be considered as the most appropriate intervention to decrease the risk of TB recurrence
(14). and prophylaxis for TB in HIV-infected patients might be seen as outdated. On the other
hand, if the risk of TB recurrence remains high in patients who start HAART, prior to an
increased CD4 count (13), TB secondary prophylaxis could be restricted to the period of time
during which patients are at higher risk of TB, and prophylaxis could be stopped as soon as
the CD4 raise over a given threshold. To explore the latter hypothesis, further therapeutic
trials should assess: (i) whether HIV-infected immunocompromised adults with active TB
could benefit from post treatment INH prophylaxis until CD4 count is increased; (ii) whether
HIV-infected adults who present for the first time with criteria for starting HAART and a past
history of successfully treated active TB could benefit from receiving isoniazid 300 mg a day
in combination with HAART until CD4 count is increased. In these trials, the best CD4
threshold to stop TB prophylaxis should be studied by comparing various arms (e.g. stopping
isoniazid when the CD4 count went over 200, 350, or 500 CD4/mm3).
Reporting a higher incidence of TB in patients with past history of TB inevitably leads to
question if the TB episodes would have been more frequently relapses due to persistent bacilli
rather than re-infection with new bacilli occurring since the previous episode. Studies
including RFLP fingerprinting have now clearly shown that both mechanisms may be equally
involved in TB recurrence in sub-Saharan HIV-infected adults (15-17). One can also imagine
that re-infection with new TB bacilli could be more frequent in HIV-infected patients through
the same mechanisms as non-tuberculous mycobacteriosis have been shown to be more
frequent in patients with a history of TB
(18). However our study did not include
fingerprinting, and therefore does not contribute to this debate.
Finally, it is important to highlight that this study did not address the association between
HAART and active TB, and that our findings do not imply that HAART is associated with an
increase in the overall TB incidence. Before the HAART era in the Cotrame Cohort, the
incidence of active tuberculosis was estimated at 10.1 /100 P-Y (19). Important studies have
now shown a decrease in the incidence of opportunistic infections in patients receiving
HAART (1-3), including studies performed in sub-Saharan African and focusing on the
incidence of TB (5). Though it is difficult to compare incidences reported in studies from
different settings, the global incidence of TB that we observed in our cohort is consistent with
that reported in South Africa in HIV-infected adults receiving HAART (5) and lower than the
incidence reported in HIV-infected adults not receiving HAART in Côte d’Ivoire (19) and in
other sub-Saharan African countries (20). The incidence of TB in our group of patients with
history of TB is in the upper limit of the range of incidences of TB recurrence that have been
reported in HIV-infected adults who do not receive HAART in sub-Saharan Africa (21). In our
cohort of adults receiving HAART, it is thus just as if the incidence of TB would have been
reduced more in patients with no TB history than in those with TB history.
In conclusion, in this cohort of adults starting HAART, active TB was more frequent in
patients with a previous history of TB than in patients with no TB history. Globally, the TB
incidence appeared to be reduced compared with the incidence observed in the same setting
and the same population before the HAART era. Cohorts of adults receiving HAART in sub-
Saharan Africa should carefully record data on TB history at baseline, and further studies
should explore the causal mechanisms and other risk factors. If confirmed, this could be a first
step toward new patterns of time-limited TB secondary chemoprophylaxis in Africa.
1. Girardi E, Antonucci G, Vanacore P, Libanore M, Errante I, Matteelli A, et al. Impact of
combination antiretroviral therapy on the risk of tuberculosis among persons with HIV
infection. AIDS 2000;14:1985-91.
2. Palella FJ, Jr., Delaney KM, Moorman AC, Loveless MO, Fuhrer J, Satten GA, et al.
Declining morbidity and mortality among patients with advanced human
immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med
3. Kirk O, Gatell JM, Mocroft A, Pedersen C, Proenca R, Brettle RP, et al. Infections with
Mycobacterium tuberculosis and Mycobacterium avium among HIV-infected patients
after the introduction of highly active antiretroviral therapy. EuroSIDA Study Group
JD. Am J Respir Crit Care Med 2000;162:865-72.
4. Lucas SB, Hounnou A, Peacock C, Beaumel A, Djomand G, N'Gbichi J-M, et al. The
mortality and pathology of HIV infection in a West African city. AIDS 1993;7:1569-
5. Badri M, Wilson D, Wood R. Effect of highly active antiretroviral therapy on incidence of
tuberculosis in South Africa: a cohort study. Lancet 2002;359:2059-64.
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, Cote d'Ivoire. Antivir Ther 2003;8:385-93.
7. Anglaret X, Dakoury-Dogbo N, Bonard D, Toure S, Combe P, Ouassa T, et al. Causes and
empirical treatment of fever in HIV-infected adult outpatients, Abidjan, Cote d'Ivoire.
8. Cox D. Regression models and life-tables. J R Statist Soc. B;34:187-220.
9. Klein NC, Duncanson FP, Lenox TH, 3rd, Pitta A, Cohen SC, Wormser GP. Use of
mycobacterial smears in the diagnosis of pulmonary tuberculosis in AIDS/ARC
patients. Chest 1989;95:1190-2.
10. Siddiqi K, Lambert ML, Walley J. Clinical diagnosis of smear-negative pulmonary
tuberculosis in low-income countries: the current evidence. Lancet Infect Dis
11. Perriëns JH, St Louis ME, Mukadi YB, Brown C, Prignot J, Pouthier F, et al. Pulmonary
tuberculosis in HIV-infected patients in Zaïre. A controlled trial of treatment for either
6 or 12 months. N Engl J Med 1995;332:779-84.
12. Fitzgerald DW, Desvarieux M, Severe P, Joseph P, Johnson WD, Jr., Pape JW. Effect of
post-treatment isoniazid on prevention of recurrent tuberculosis in HIV-1-infected
individuals: a randomised trial. Lancet 2000;356:1470-4.
13. 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;17:2063-70.
14. Harries AD, Chimzizi RB, Nyirenda TE, van Gorkom J, Salaniponi FM. Preventing
recurrent tuberculosis in high HIV-prevalent areas in sub-Saharan Africa: what are the
options for tuberculosis control programmes? Int J Tuberc Lung Dis 2003;7:616-22.
15. Gilks CF, Godfrey-Faussett P, Batchelor BI, Ojoo JC, Ojoo SJ, Brindle RJ, et al. Recent
transmission of tuberculosis in a cohort of HIV-1-infected female sex workers in
Nairobi, Kenya. AIDS 1997;11:911-8.
16. Sonnenberg P, Murray J, Glynn JR, Shearer S, Kambashi B, Godfrey-Faussett P. HIV-1
and recurrence, relapse, and reinfection of tuberculosis after cure: a cohort study in
South African mineworkers. Lancet 2001;358:1687-93.
17. Glynn JR, Yates MD, Crampin AC, Ngwira BM, Mwaungulu FD, Black GF, et al. DNA
fingerprint changes in tuberculosis: reinfection, evolution, or laboratory error ? J Infect
18. Corbett EL, Churchyard GJ, Clayton T, Herselman P, Williams B, Hayes R, et al. Risk
factors for pulmonary mycobacterial disease in South African gold miners. A case-
control study. Am J Respir Crit Care Med 1999;159:94-9.
19. Attia A, Huet C, Anglaret X, Toure S, Ouassa T, Gourvellec G, et al. HIV-1-related
morbidity in adults, Abidjan, Cote d'Ivoire: a nidus for bacterial diseases. J Acquir
Immune Defic Syndr 2001;28:478-86.
20. van der Sande MA, Schim van der Loeff MF, Bennett RC, Dowling M, Aveika AA,
Togun TO, et al. Incidence of tuberculosis and survival after its diagnosis in patients
infected with HIV-1 and HIV-2. AIDS 2004;18:1933-41.
21. Korenromp EL, Scano F, Williams BG, Dye C, Nunn P. Effects of human
immunodeficiency virus infection on recurrence of tuberculosis after rifampin-based
treatment: an analytical review. Clin Infect Dis 2003;37:101-12.