Tropical Medicine and International Health

volume 10 no 8 pp 723–733 august 2005

Efficacy and safety of two dosages of cotrimoxazole as

preventive treatment for HIV-infected Malawian adults
with new smear-positive tuberculosis
Martin J. Boeree1,2, Delphine Sauvageot3, Hastings T. Banda1, Anthony D. Harries4 and Eduard E. Zijlstra1

1 Department of Medicine, University of Malawi, College of Medicine, Blantyre, Malawi

2 Department of Respiratory Diseases, Radboud University Nijmegen and University Lung Centre Dekkerswald, Nijmegen,
the Netherlands
3 Institut National Statistique Epidémiologie de Recherche Médicale, Bordeaux, France
4 The National Tuberculosis Control Programme, Community Health Science Unit, Lilongwe, Malawi

Summary objective To assess the efficacy and safety of two different dosages of cotrimoxazole (CTX) in
prophylaxis in HIV-positive new smear-positive pulmonary tuberculosis (TB) patients in Blantyre,
Malawi.
method Randomized, double-blind trial using 480 and 960 mg of CTX given to new TB patients,
who were followed up until the end of the tuberculosis treatment. The primary outcome was survival.
The outcome in the two groups was also compared with an unselected cohort of similar patients
registered in Zomba, Malawi in 1995 and new smear-positive patients registered in the National
Tuberculosis Programme in 1999. The secondary outcome was the occurrence of (opportunistic) events,
especially bacterial pneumonia.
results There were no statistically significant differences in mortality and bacterial pneumonia
between the groups receiving the two different dosages. The case fatality rate at the end of the
tuberculosis treatment was 15.4% in the 480 mg group and 14.0% in the 960 mg group. This was lower
than the case fatality rate in the Zomba cohort (19.2%, P ¼ 0.10) and lower than the case fatality rate in
the national programme (21.0%, P < 0.001). CTX was well tolerated. Compliance was fair.
conclusions CTX prophylaxis may have a beneficial effect on mortality and morbidity in HIV-
infected smear-positive tuberculosis patients in Malawi. The efficacy of both dosages is not significantly
different. The intervention is cheap and easy to implement. These results would support implementation
of CTX in this patient group until better strategies are available or evidence is convincingly presented to
suggest that its benefit is marginal.

keywords Africa, clinical trials, opportunistic infection, HIV infections and mortality, TB drug therapy,
TB epidemiology, TB mortality

the reported incidence of P. carinii pneumonia (PCP) is

Introduction
The mortality in Sub-Saharan Africa of patients who are
treated for tuberculosis has increased dramatically in the
last decades because of HIV (Mukadi et al. 2001), and
there is an urgent need for measures to reduce it. At
present, the availability of antiretroviral treatment in
Africa is limited. Cotrimoxazole (CTX) prophylaxis has
several potential benefits. Before the introduction of
antiretroviral therapy in the industrialized world CTX
prophylaxis was widely used in HIV-infected patients with
impaired immunity. The decrease in morbidity and mor-
tality was mainly due to preventing opportunistic infection
with Pneumocystis carinii (Hardy et al. 1992). In Africa,
much lower, although increasing or more common than
previously thought (Elvin et al. 1989; Gilks et al. 1990;
Batungwanayo et al. 1994; Malin et al. 1995; Gordon
et al. 2001a). However, CTX prophylaxis may still be
effective in decreasing mortality and morbidity through the
prevention of other CTX-susceptible (opportunistic)
infections. This led to trials to assess the efficacy of CTX in
Africa. In Ivory Coast, the use of CTX prophylaxis reduced
mortality in HIV-positive individuals and morbidity in
HIV-positive individuals co-infected with TB (Anglaret
et al. 1999; Wiktor et al. 1999). In Senegal (Maynart et al.
2001), a study in HIV-positive patients without tuber-
culosis showed no beneficial effect. In Malawi, a package

ª 2005 Blackwell Publishing Ltd 723


Tropical Medicine and International Health
M. J. Boeree et al. Cotrimoxazole prophylaxis in HIV/TB in Malawi
of counselling, HIV testing and CTX to those who were
HIV-positive showed a reduction in mortality in HIV-
positive patients with smear-negative pulmonary tubercu-
losis (PTB) and extrapulmonary TB, but not in smear-
positive PTB in one trial (Zachariah et al. 2003) and a clear
reduction in mortality, especially in smear-positive PTB
patients in another trial (Mwaungulu et al. 2004). Extra-
polation of the Ivory Coast trial to other regions within
Sub-Saharan Africa may not be justified because of the
relatively low bacterial resistance rates to CTX in Ivory
Coast. It has been suggested that these results should be
investigated in regions with high bacterial CTX resistance
(Boeree et al. 1999). Finally, while WHO/UNAIDS (UN-
AIDS 2000) recommend CTX 960 mg daily, the optimal
dosage of CTX still needs to be established in terms of
efficacy, safety and cost.
We investigated the efficacy of CTX in preventing death
and disease in HIV-infected patients with active smear
positive tuberculosis in Blantyre, Malawi. Smear-positive
patients are a well-defined group within the National
Tuberculosis Control Programme (NTP) with known
outcome parameters. The original design was a random-
ized placebo-controlled trial. Soon after implementation of
the trial the results of the study in Ivory Coast became
available (Wiktor et al. 1999). Ethical considerations,
especially in the light of a vigorous debate at that time
about ethical aspects of placebo-controlled studies in
Africa (Taylor 1995) led to our trial being terminated.
The trial was then redesigned to compare two dosages of
CTX (480 and 960 mg). We expected a dosage of CTX
480 mg daily to be as effective as a CTX 960 mg, as
demonstrated in a study in the Netherlands in preventing
PCP (Schneider et al. 1995). Use of a single dose would
reduce the costs by more than half. We compared case
fatality rates, specific morbidity and frequency of drug
reactions during anti-TB treatment in patients who
received the two different dosages of CTX. We also
compared case fatality rates in patients on CTX with those
observed in a cohort of HIV-positive new smear-positive
PTB patients recruited to a study in Zomba, Malawi, in
1995 and with the reported case fatality rates in new
smear-positive PTB patients registered in the NTP in 1999
(Harries et al. 1998; Kang’ombe et al. 2000).
Methods
The setting
The study was conducted at Queen Elizabeth Central
Hospital (QECH) in Blantyre, Malawi, from April 1998
until September 2001. QECH is a large 1050-bedded
teaching hospital and functions as a primary, secondary
724
13653156, 2005, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-3156.2005.01433.x, Wiley Online Library on [30/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
volume 10 no 8 pp 723–733 august 2005
and tertiary facility. It is the only government hospital in
Blantyre and health care is free of charge; the hospital
serves the population of the city as well as of that of
Blantyre District (estimated population of 830 000 in
mid-1998). The NTP has an office at QECH managed by a
District Tuberculosis Officer.
Patients
Patients who were diagnosed with smear-positive PTB
were screened before starting treatment for inclusion in the
trial. If the patient was willing to participate he/she was
counselled for HIV testing. Patients who fulfilled the
following eligibility criteria were candidates for inclusion:
older than 15 years, serological evidence of HIV infection,
PTB diagnosed by sputum microscopy reconfirmed in our
research laboratory, resident and planning to stay in
Blantyre District and surrounding areas for 3 years, and
having given written or verbal informed consent. Patients
were excluded on the basis of the following criteria: a
diagnosis of active symptomatic bacterial infection other
than tuberculosis or protozoal infection, having had TB
treatment for longer than 8 days, known hypersensitivity
to or intolerance to CTX, liver disease as evidenced by
clinically apparent jaundice, inability to comply with
protocol requirements, malignancy, inability to leave the
bed unassisted to be weighed, being a prisoner, a vendor or
a refugee (because of anticipated problems with compli-
ance), pregnancy or a previous episode of TB.
Study methods
Each patient underwent a thorough anamnesis and clinical
examination. Blood and sputum samples were taken.
Eligible patients were randomized to receive either CTX
480 or 960 mg during TB treatment. Randomization was
done using a computer-generated randomization list, using
Microsoft Excel. An association of two letters was
randomly assigned according to this list to each consecu-
tive patient in order of enrolment. Each combination of
syllables corresponded to a simple dose of CTX except for
one of them, which was a placebo. One group received
once daily 480 mg CTX in the morning (one active capsule
of 480 mg and one placebo capsule); the second group
received once daily two capsules of 480 mg CTX in the
morning.
All study medication was identical in appearance and
was manufactured by Genpharmaceutical, Johannesburg,
South Africa. The enrolment of the patients and the
distribution of the capsules were made by the study clinical
officers. Patients and clinical officers (and all other mem-
bers of the study group) were blinded to the treatment.
ª 2005 Blackwell Publishing Ltd

Tropical Medicine and International Health
M. J. Boeree et al. Cotrimoxazole prophylaxis in HIV/TB in Malawi
During the inclusion process patients started TB treatment
with 1 month of daily streptomycin (intramuscular), and
rifampin, isoniazid, and pyrazinamide (oral), 1 month of
the same four drugs three times a week (supervised),
followed by a continuation phase of 6 months with daily
unsupervised isoniazid and ethambutol (1SHRZ/
1S3H3R3Z3/6HE). The total duration of TB treatment
and CTX prophylaxis was therefore 8 months.
The patients were followed up every 4 weeks until the
end of tuberculosis treatment. At each follow-up visit we
took a brief history, examined the patient, counted the pills
and provided a new supply of CTX. Samples of blood and
sputum were taken according to protocol. Follow-up
discontinued if a patient was near death or died, withdrew
voluntarily or moved out of the district. Patients who did
not show up for their scheduled visit or their relatives were
visited at home by the study nurse and encouraged to
resume attending the clinic. Patients were considered lost
to follow-up if they missed two consecutive visits. If
patients had died an attempt was made to establish the
cause of death using available clinical information or in
case of a death occurring at home by interviewing a
relative.
Compliance of CTX prophylaxis was checked by a
monthly pill count at each visit. A patient was considered
compliant if the ratio of the total number of study drugs
actually taken and prescribed was between 0.8 and 1.0.
Laboratory tests
At enrolment HIV serology was performed using a HIV
rapid assay and if positive confirmed by ELISA. Haemo-
globin, haematocrit, platelets, white blood cell count with
a differential count, CD4 lymphocyte count, aspartate
transaminase (ASAT), alanine transaminase (ALAT), direct
bilirubin levels and Toxoplasma IgG levels were measured.
Two sputum samples were taken and stained with the
Ziehl–Neelsen staining method and microscopically
examined for acid-fast bacilli. Sputum samples were also
cultured for Mycobacterium tuberculosis using Löwen-
stein–Jensen medium.
The CD4 count was repeated after 2 months, sputum
microscopy and culture were repeated after 5 and
8 months, while full blood count (FBC), ASAT, ALAT and
bilirubin were repeated after 2 months and every 3 months
thereafter until the end of the study.
Outcome
The primary efficacy outcome was death during anti-TB
treatment. The secondary efficacy outcome was an occur-
rence of an infectious clinical event, defined as definite if
ª 2005 Blackwell Publishing Ltd
13653156, 2005, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-3156.2005.01433.x, Wiley Online Library on [30/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
volume 10 no 8 pp 723–733 august 2005
confirmed bacteriologically (by blood or sputum culture),
as probable if a patient had a clinical diagnosis of the
particular event and responding to antibiotic treatment or
as possible if there was a clinical diagnosis but no response
to treatment. Safety was assessed by the occurrence of
adverse events using a structured evaluation form, especi-
ally those that are known to be side-effects associated with
the use of CTX (skin, bone marrow and gastric intoler-
ance).
The Zomba cohort and patients registered in the Malawi
NTP
An observational cohort of tuberculosis patients in Zomba,
Malawi, in whom mortality was the main parameter
recorded, has been previously described (Harries et al. 1998;
Kang’ombe et al. 2000). We compared the outcome of our
study patients with patients in this cohort who were HIV
seropositive and sputum smear positive. We also compared
outcomes with patients who had new smear-positive PTB
registered by the NTP throughout the country in 1999.
Statistical analysis
In the original study design a sample size of 600 patients
was calculated to detect a difference of 20% mortality with
80% power and a two-sided significance of 5%; this was
left unchanged after the study was redesigned. The analysis
was conducted as an intention-to-treat analysis. Efficacy of
the two dosages of CTX was evaluated by comparing the
two groups with respect to mortality, incidence of events of
pneumonia and other clinical events. The safety of the two
dosages was assessed by comparing the incidence of
adverse events. Furthermore, Kaplan–Meier survival ana-
lysis was used with and without stratification by age, CD4
count and confirmation of tuberculosis by culture. The
Cox proportional hazards model was used to analyse the
impact of several variables on the outcomes and was also
used for comparison with the observational cohort in
Zomba. The comparisons of the case fatality rates with
Zomba and the NTP were made using a chi-square test.
The study was approved by the National Health Science
and Research Committee of Malawi and the UNAIDS
Ethical Review Board.
Results
Recruitment and follow-up
From March 1998 to January 2001, 1795 patients were
screened (Figure 1). Of these, 579 were eligible for analysis
and randomized; 272 patients received 480 mg CTX and
725
 13653156, 2005, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-3156.2005.01433.x, Wiley Online Library on [30/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Tropical Medicine and International Health volume 10 no 8 pp 723–733 august 2005

M. J. Boeree et al. Cotrimoxazole prophylaxis in HIV/TB in Malawi

Number of smear positive

PTB patients screened for
eligibility (n = 1795)

Excluded (n = 948)

Not meeting inclusion

criteria (n = 784)

Ineligible according to
exclusion criteria
(n = 54)

Refused to participate
(n = 162)

Other reasons (n = 28)

Randomised (n = 767)

Allocated to CTX 480 mg (n = 378) Allocated to CTX 960 mg (n = 391)

Received CTX 480 mg (n = 272) Received CTX 960 mg (n = 307)

Did not receive CTX 480 mg (n = 106) Did not receive CTX 480 mg (n = 84)
Found to be smear negative by Found to be smear negative by
study laboratory (n = 97) study laboratory (n = 77)
Other reasons (n = 9) Other reasons (n = 7)

Lost to follow-up: missed two consecutive Lost to follow-up: missed two

visits (n = 20) consecutive visits (n = 24)

Analysed (n = 272) Analysed (n = 307)

None excluded from analysis None excluded from analysis

Figure 1 Flow diagram of the trial.

726 ª 2005 Blackwell Publishing Ltd

 13653156, 2005, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-3156.2005.01433.x, Wiley Online Library on [30/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Tropical Medicine and International Health volume 10 no 8 pp 723–733 august 2005

M. J. Boeree et al. Cotrimoxazole prophylaxis in HIV/TB in Malawi

Table 1 Demographical, clinical and bio-

logical characteristics of randomized CTX 480 mg arm CTX 960 mg arm
patients according to cotrimoxazole Characteristics at inclusion (n ¼ 272) (n ¼ 307)
prophylaxis treatment arms
Age in year, mean (SD) 32.4 (30.4) 32.4 (30.8)
Age (year)
15–24 19.1 (52) 19.3 (59)
25–34 47.1 (128) 45.4 (139)
35–44 24.0 (66) 26.1 (80)
45+ 9.6 (26) 9.2 (28)
Male 50.0 (136) 49.5 (152)
Body mass index (kg/m2), mean (SD) 17.6 (2.4) 17.3 (2.1)
<16 22.8 (62) 24.4 (75)
16–17 16.5 (45) 17.9 (55)
17–18.5 26.1 (71) 25.7 (79)
18.5–25 26.1 (71) 24.4 (75)
>25 1.8 (5) 0.3 (1)
Unknown 6.6 (18) 7.2 (22)
CD4 count (cells/mm3), mean (SD) 238 (214) 221 (215)
0–99 25.7 (70) 28.7 (88)
100–199 27.2 (74) 28.0 (86)
200–349 20.2 (55) 19.9 (61)
‡350 21.7 (59) 17.3 (53)
Unknown 5.1 (14) 6.2 (19)
Haemoglobin level (g/dl), mean (SD) 9.66 (2.00) 9.52 (1.95)
Tb culture results
Mycobacterium tuberculosis 79.8 (217) 78.8 (242)
Culture negative 18.8 (51) 19.9 (61)
Unknown 1.5 (4) 1.3 (4)

Values are presented as % (n) unless indicated.

307 patients received 960 mg CTX. The groups were
comparable (Table 1). At the end of TB treatment 42
patients (7.2%) were lost to follow-up: 24 in the CTX
480 mg group and 20 in the 960 mg group. Furthermore,
74 patients had been withdrawn from the study at the end
of TB treatment, because they missed two consecutive visits
(22 patients), because they moved out of the district (32
patients), and because they wished to withdraw from the
study (20 patients).
Death rates in the two groups
The number of patients who died during tuberculosis
treatment until 30 September 2001 was 85. The overall
case fatality rate at the end of tuberculosis treatment was
14.7%. The incidence rates for death were 23.3/100 person
years in the CTX 480 mg arm and 21.0/100 person years in
the CTX 960 mg arm. The probability of survival was not
significantly different between groups (log rank P ¼ 0.63)
with a hazard ratio (CTX 480/960 mg) of 1.11 (95% CI
0.72–1.71). These results remained unchanged after
adjustment on age, gender, baseline CD4 count, baseline
toxoplasmosis serology, haemoglobin at 8 month and
bacterial pneumonia during TB treatment. The adjusted
hazard ratio was 1.00 (95% CI 0.62–1.63).
Increasing age was significantly associated with risk of
death; there was no significant association with risk of
death and sex or CD4 count (Table 2). The established
causes of death did not differ between the two groups
(Table 3).
Pneumonia and other clinical events
The number of patients with one or more events of
pneumonia or other clinical events and the adjusted hazard
ratios are shown in Table 4. There were 62 events of
definite or probable bacterial pneumonia in 37 patients,
eight of them were definite and 54 probable. Thirty events
in 18 patients occurred in the 480-mg group and 32 events
in 19 patients occurred in the 960-mg group. Toxoplasma
serology was negative in 90.0%, positive in 8.9% and
indeterminate in 1.0% of patients. We did not diagnose
any case of Toxoplasma encephalitis (TE).
Death rates compared with the Zomba cohort and the
NTP
Specific outcome parameters of our study are summarized
in Table 5 and compared with the outcome parameters
in the Zomba cohort and the general outcome parameters

ª 2005 Blackwell Publishing Ltd 727

 13653156, 2005, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-3156.2005.01433.x, Wiley Online Library on [30/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Tropical Medicine and International Health volume 10 no 8 pp 723–733 august 2005

M. J. Boeree et al. Cotrimoxazole prophylaxis in HIV/TB in Malawi

Table 2 Incidence rate and rate ratio for

Dead* Alive* Hazard ratio  Hazard ratio mortality during TB treatment period
Mortality (n ¼ 85) (n ¼ 494) (95% CI) adjusted  (95% CI)

Age (years)
15–24 7.1 (6) 21.3 (105) Ref Ref
25–34 47.1 (40) 46.0 (227) 3.0 (1.27–7.09) 2.74 (1.14.-6.61)
35–44 27.1 (23) 24.9 (123) 3.18 (1.30–7.82) 2.69 (1.04–6.97)
45+ 18.8 (16) 7.7 (38) 6.72 (2.63–17.2) 5.22 (1.88–14.54)
CD4 (cells/mm3)
0–199 66.7 (50) 56.9 (268) 1.51 (0.80–2.83) 1.38 (0.72–2.62)
200–349 17.3 (13) 21.9 (103) 1.03 (0.47–2.25) 1.06 (0.48–2.35)
>350 16.0 (12) 21.2 (100) Ref Ref
Sex
Male 51.8 (44) 49.4 (244) Ref Ref
Female 48.2 (41) 50.6 (250) 0.89 (0.58–1.36) 1.11 (0.68–1.81)

* Values are presented as % (n).

  Cox regression model.

Table 3 Causes of death during TB treatment according to the two groups. No patient received anti-HIV treatment,
cotrimoxazole (CTX) prophylaxis arms because it was hardly available in Malawi during the
period of the study.
Causes of death CTX 480 mg CTX 960 mg

Total number of death 42/272 43/307 Discussion

during TB treatment
Tuberculosis 23.8 (10) 27.9 (12) This study showed no significant differences in mortality
Gastrointestinal disease 26.2 (11) 18.6 (8) and occurrence of clinical events between 480 and
Pneumonia 7.1 (3) 4.7 (2) 960 mg CTX prophylaxis in HIV-positive smear-positive
Meningitis 7.1 (3) 18.6 (8)
tuberculosis patients in Malawi. Furthermore, both
Kaposi’s sarcoma 2.4 (1) 0
Anaemia 2.4 (1) 0 dosages of CTX were well tolerated with equal compli-
Malaria 2.4 (1) 0 ance. In addition, the overall case fatality rate was
Septicaemia 2.4 (1) 0 reduced significantly when compared with the case
Respiratory events other 4.8 (2) 0 fatality rate in the National Tuberculosis Programme
than pneumonia although not when compared with the observational
Other 11.9 (5) 16.3 (7) cohort in Zomba.
Unknown 9.5 (4) 9.3 (4)
We hypothesized that CTX prophylaxis may be effective
Values are presented as % (n). in decreasing mortality and morbidity through the pre-
vention of CTX-susceptible (opportunistic) infections. This
assumption was based on existing knowledge of the causes
of the NTP. The case fatality rate of the NTP was of mortality and morbidity in HIV-infected patients in
significantly higher (P < 0.001) compared with our study Africa. Mortality is high because of infections with
groups. Streptococcus pneumoniae and non-typhoidal salmonellae
(NTS) (Gilks et al. 1990; Gordon et al. 2001a,b). Other
causes of death and serious illness include cerebral infec-
Adverse effects and compliance
tions with Toxoplasma gondii (Lucas et al. 1993; Eholie
There were 47 adverse events occurring in 37 patients. All et al. 2000) and chronic diarrhoea caused by Cryptospor-
but one were mild or moderate. There were no significant idium parvum, Isospora belli and Cyclospora cayetanensis
differences between the two groups. Most events, which (Kelly et al. 1997). Evidence from studies in the industri-
were attributed to CTX, were rash and gastrointestinal alized world showed that the incidence of pneumococcal
discomfort. One patient in the CTX 960 group had to stop disease and TE was lower in patients who received CTX
the prophylaxis indefinitely because she developed Stevens– prophylaxis for PCP (Zangerle & Allerberger 1991; Navin
Johnson syndrome. The patient recovered. Fourteen et al. 2000). Furthermore, there is a known, recently
patients had a compliance ratio of <0.6, and 45 a low confirmed (Omar et al. 2001), beneficial effect of CTX on
compliance ratio of 0.6–0.8 with no differences between Plasmodium falciparum.

728 ª 2005 Blackwell Publishing Ltd

 13653156, 2005, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-3156.2005.01433.x, Wiley Online Library on [30/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Tropical Medicine and International Health volume 10 no 8 pp 723–733 august 2005

M. J. Boeree et al. Cotrimoxazole prophylaxis in HIV/TB in Malawi

Table 4 Incidence rates of first clinical events during tuberculosis treatment according to cotrimoxazole prophylaxis arms

CTX 480 mg arm CTX 960 mg arm Hazard ratio* Hazard ratio
(n ¼ 272) (n ¼ 307) (95% CI) adjusted  (95% CI) P-value

Patients with at least 127 136

one clinical event
Bacterial pneumonia (n) 18 19 1.07 (0.56–2.06) 1.08 (0.55–2.11) 0.83
Incidence (/100 py) 12.3 10.9 (ref)
Diarrhoea (n) 18 28 0.62 (0.33–1.16) 0.64 (0.34–1.22) 0.18
Incidence (/100 py) 12.0 16.2 (ref)
Malaria (n) 18 30 0.69 (0.39–1.24) 0.71 (0.40–1.29) 0.26
Incidence (/100 py) 12.1 17.6 (ref)
Meningitis (n) 3 6 0.58 (0.15–2.34) 0.59 (0.15–2.36) 0.45
Incidence (/100 py) 2.0 3.4 (ref)
UTI/STD (n) 11 13 0.99 (0.44–2.20) 1.34 (0.57–3.16) 0.51
Incidence (/100 py) 7.4 7.5 (ref)
PCP (n) 2 2 0.59 (0.05–6.52) 0.47 (0.02–9.07) 0.62
Incidence (/100 py) 1.3 1.1 (ref)
Septicemia (n) 12 8 1.76 (0.72–4.31) 1.92 (0.77–4.74) 0.16
Incidence (/100 py) 8.0 4.5 (ref)

* Cox regression model.

  Cox regression model with adjustment for age/sex/baseline CD4 count/Toxoplasma serology/baseline TB confirmation/CTX arms.

Table 5 Outcome parameters of our study

compared with the Zomba observational CTX 480 mg CTX 960 mg NTP Zomba
cohort and the NTP for new smear-positive 1998–2001 1998–2001 1999 1995
PTB patients (n ¼ 272) (n ¼ 307) (n ¼ 8185) (n ¼ 255)

Cured or completed 62.9 58.6 71 69.4

treatment (%)
Died (%) 15.4 14.0 21* 18.0**
Failed (%) 3.7 3.9 1 1.2
Transferred (%) 3 5.5
Defaulted  (%) 8.8 6.5 4 5.9

  In the study groups default are those patients who were lost to follow-up for the study.
* P < 0.001 when compared with case fatality rate in CTX 480/960 mg (v2 test).
** P < 0.10 when compared with case fatality rate in CTX 480/960 mg (v2 test).

The trials to assess the efficacy of CTX in Africa do not
provide enough evidence to justify the provisional recom-
mendation of WHO/UNAIDS that ‘CTX should be used
for prophylaxis in adults and children living with HIV/
AIDS in Africa as part of a minimum package of care’
(UNAIDS 2000). The trial in Ivory Coast showed with the
use of CTX prophylaxis a reduction in mortality in patients
with tuberculosis by 46% (Wiktor et al. 1999). In HIV-
positive patients without tuberculosis there was no reduc-
tion in mortality, but a 43% reduction in hospital
admissions (Anglaret et al. 1999). In Senegal (Maynart
et al. 2001), a study in HIV-positive patients without
tuberculosis showed no beneficial effect, but had small
numbers of patients. In a recently completed study in
Thyolo district, Malawi, CTX showed a significant
reduction in mortality in HIV-positive patients with
smear-negative PTB and extrapulmonary TB, but not in
smear-positive PTB (Zachariah et al. 2003). In a recently
published paper in a cohort of 362 HIV-positive TB
patients who received CTX 960 mg in Karonga, North
Malawi, when compared with a historical cohort there was
a significant reduction in mortality, especially in smear-
positive patients (Mwaungulu et al. 2004). In two recent
studies from Uganda and Zambia, where rates of CTX
resistance to bacterial pathogens are high, CTX was
associated with a significant reduction in mortality in HIV-
positive adults in Uganda and in HIV-positive children in
Zambia, particularly those with lower CD4-lymphocyte
counts (Chintu et al. 2004; Mermin et al. 2004).
In our study, tuberculosis, gastrointestinal disease and
respiratory infection were the most common causes of
death. Most of the deaths occurred within the first

ª 2005 Blackwell Publishing Ltd 729


Tropical Medicine and International Health
M. J. Boeree et al. Cotrimoxazole prophylaxis in HIV/TB in Malawi
2 months, which has been observed in the Zomba cohort
as well and could be a potential target for other interven-
tions such as antiretroviral treatment (Dean et al. 2002). In
line with other reports in Africa (Harries et al. 1998; van
den Broek et al. 1998; Steen et al. 2001), mortality was
higher in older age groups.
In comparison with the overall countrywide case fatality
rate of the NTP for patients with smear-positive tuber-
culosis, our study patients had a significant lower case
fatality rate. This difference may in fact be more pro-
nounced taking into account that in the NTP data both
HIV-positive and -negative patients are included; correct-
ing for the lower mortality rate in HIV-negative patients
(10%) (Harries et al. 1998) and assuming that 72–80% of
smear-positive patients are HIV-positive, the NTP mor-
tality figure for HIV-positive patients would be 23–25%.
These data need to be interpreted with caution as patient
characteristics in the various groups differ and our study
patients were at the start of the study not very ill and had
the benefit of being in a study. Similarly, although all
patients were treated according to NTP guidelines, we have
no data with regard to compliance to treatment in the
Zomba cohort and the patients registered by the NTP.
There were no differences in clinical outcomes between
the two study groups; bacterial pneumonia, diarrhoea,
bacteraemia and malaria were the most frequently identi-
fied opportunistic infections, whereas PCP was rare.
Although this study was not designed to arrive at a firm
diagnosis of clinical events, these findings are in keeping
with data from other studies in Africa (Gilks et al. 1990;
Greenberg et al. 1995; van den Broek et al. 1998; Grant
et al. 2001; Gordon et al. 2001a, 2002). Streptococcus
pneumoniae is one of the commonest pathogens found in
patients with respiratory infection in our wards (Gordon
et al. 2002) and this has been found elsewhere (Gilks et al.
1990). There are no published studies on the causes of
diarrhoea in Malawi. However, a recent study in patients
with and without diarrhoea found a prevalence of
C. parvum and I. belli of 11% and 2.5% respectively
(Cranendonk et al. 2003). In contrast to findings in
Malawian children (Graham et al. 2000), PCP appeared
uncommon in adults; and similar low prevalence rates
have been found in other studies in Africa (Batungwanayo
et al. 1994; Hargreaves et al. 2001). However, there are
recent reports from Zimbabwe, Kenya, Ethiopia and South
Africa showing an increasing or higher prevalence of PCP
in HIV-positive patients than in earlier studies (Malin et al.
1995; Robberts et al. 2002; Aderaye et al. 2003; Chakaya
et al. 2003).
No case of cerebral toxoplasmosis was diagnosed; it is
possible that cases were missed as no CT scan was
available and patients may have died in the ward or at
730
13653156, 2005, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-3156.2005.01433.x, Wiley Online Library on [30/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
volume 10 no 8 pp 723–733 august 2005
home of rapidly fatal toxoplasmosis encephalitis.
However, the seroprevalence of anti-toxoplasmosis IgG
antibody was low (9.6%), suggesting a relatively low risk.
The incidence of TE in HIV-infected patients directly
correlates with the prevalence of T. gondii antibodies
among the general HIV-infected population, the degree of
immune suppression, and the institution of prophylaxis
against TE (Luft & Remington 1992). CTX may also have
benefit through its well-recognized effect in treating
falciparum malaria, and CTX preventive therapy may
reduce the frequency of episodes of malaria. Due to the
absence of a placebo-controlled arm this hypothesis
remains speculation.
We were specifically interested in the use of the single-
dose CTX because side-effects may be less in HIV-infected
patients, in whom adverse effects to CTX are known to be
more frequent (van der Ven et al. 1996). Currently no data
are available on its use in African patients. We are
confident that the use of CTX in HIV-infected patients
with tuberculosis in this trial was safe. Only one patient
had a serious side-effect, which led to withdrawal. This
patient developed Stevens–Johnson syndrome from which
she recovered. The other 36 events of side-effects were mild
and transient. This result is in line with reports from Ivory
Coast (Wiktor et al. 1999), Senegal (Maynart et al. 2001),
and Thyolo District in Malawi (Zachariah et al. 2003). In
the group of patients who used 960 mg CTX there were
slightly more side-effects than in the group who used
480 mg, but this was not statistically significant. There was
a fair compliance of CTX use and a good tolerance of the
drug in combination with the anti-tuberculous drugs. Most
patients chose to continue the prophylaxis after the
termination of the study. A considerable number of
patients were lost to follow-up or chose to withdraw from
the study. The reasons for this withdrawal were mainly
the burden of too many drugs (TB treatment in addition
with the two capsules of CTX) and the alleged stigmat-
ization of the study towards HIV infection.
Although there are important methodological differences
between studies, this study suggests that CTX primary
prophylaxis may have a potential beneficial effect on
mortality in HIV-infected new smear-positive TB patients
and that low-dose CTX 480 mg daily may be as effective as
CTX 960 mg, which is the current WHO/UNAIDS
recommendation. Although the study has limited power to
show equal efficacy, in the sample size that we investigated
a difference of 9.5% would be statistically significant. We
found a difference of 1.4% between the two groups.
The use of a single dose would reduce the costs by more
than half. At the International Dispensary Association
(IDA, Amsterdam, the Netherlands), which is the most
frequently used supplier of drugs in Africa, a container of
ª 2005 Blackwell Publishing Ltd

Tropical Medicine and International Health
M. J. Boeree et al. Cotrimoxazole prophylaxis in HIV/TB in Malawi
1000 capsules of CTX 480 mg costs €7.19 (0.72 cent per
tablet), while 500 capsules with 960 mg cost €8.24 (1.65
cent per tablet), which means a cost reduction of 56.4%
(IDA Formularium 2003).
What can be learned from this study? Clearly, as other
effective and affordable interventions are still beyond the
horizon for the majority of patients, one would like to err
on the side where patients who otherwise have a poor
prognosis may benefit. In this context it is regrettable that
randomized-controlled trials are still not available in
southern Africa and that most studies that have been
carried out have methodological problems. High levels of
resistance to CTX of common bacteria such as
S. pneumoniae and low prevalence of other infections to
which CTX is targeted (PCP, toxoplasmosis, isosporiasis)
should caution against over interpretation of the currently
available data. In addition, implementation of a strategy of
widespread use of CTX prophylaxis needs to be balanced
against unwanted long-term effects. The emergence of
antibiotic resistance to pathogens is a potential problem
after widespread implementation of CTX. In Malawi, the
rate of resistance of NTS and S. pneumoniae to CTX is
already very high with 73% and 91% respectively (Gordon
et al. 2001b, 2002). In a recent study there was an increase
of drug resistance to Escherichia coli for CTX in Thyolo
District in Malawi after the implementation of routine
CTX prophylaxis (Zachariah et al. 2002). The clinical
consequences of this observation are not clear and need to
be investigated. Another concern is the emergence of cross
resistance of P. falciparum to a similar anti-folate combi-
nation, sulphadoxine–pyrimethamine (SP), which is now
the first-line antimalarial drug in several of the African
countries with the highest rates of HIV infection (Iyer et al.
2001; Omar et al. 2001). As the only effective and
affordable successor to SP, namely the combination of
lapudrine and dapsone (Lapdap) that will be marketed
soon is also an anti-folate this may have severe conse-
quences for malaria control in the community as well as for
treatment of the individual. Clearly these issues should be
addressed urgently.
Taking all these considerations into account, a meeting
was held in 2002 between the College of Medicine, the
NTP, interested stakeholders and the Ministry of Health of
Malawi to discuss these results as well as operational
research data results from Lilongwe, Karonga and Thyolo.
The outcome of the meeting was a policy statement which
(i) endorsed that Voluntary Counselling and Testing for
HIV (VCT) and CTX prophylaxis for HIV-seropositive TB
patients be continued in those districts where it has been
started, (ii) encouraged VCT plus CTX prophylaxis for
HIV-seropositive TB patients to be expanded in other
districts in a phased approach with results carefully
ª 2005 Blackwell Publishing Ltd
13653156, 2005, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-3156.2005.01433.x, Wiley Online Library on [30/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
volume 10 no 8 pp 723–733 august 2005
monitored by the NTP, (iii) urged that the NTP encourage
and work with its partners to conduct a randomized-
controlled trial to determine the efficacy of CTX prophy-
laxis with and without antiretroviral therapy in
HIV-seropositive TB patients and (iv) requested all parties
to keep up to date with any new information about the
efficacy of CTX prophylaxis in the region and act
accordingly if evidence suggests that CTX has harmful
consequences.
Acknowledgements
We thank UNAIDS for the financial and intellectual
support they gave, J.J. Perriens for the initial writing of the
protocol and the supervisory role within UNAIDS, B. Samb
for the initial monitoring and later the supervisory role
within UNAIDS, J.J. Wirima, College of Medicine, for his
initial advise and support, V. Leroy, INSERM. for super-
vision of the data monitoring and analysis, G. Mateyu,
A. Silungwe, A. Kamenya, R. Sikwese, H. Chilonda,
R. Gondwe, S. Chirambo, E. Makata for the hard clinical
work and the data management, N. Chilewani,
W. Dzinyemba, R. Banda, M. Ndileke, College of
Medicine for the laboratory work, the Wellcome Trust
Laboratories, Blantyre for letting us use the equipment
and staff for laboratory and microbiological tests, the
Blantyre Christian Centre counsellors for their counselling
work for HIV infection.
References
Aderaye G, Bruchfeld J, Olsson M & Lindquist L (2003) Occurrence
of Pneumocystis carinii in HIV-positive patients with suspected
pulmonary tuberculosis in Ethiopia. AIDS 17, 435–440.
Anglaret X, Chêne G, Attia A et al. (1999) Early chemoprophy-
laxis with trimethoprim-sulphamethoxazole for HIV-1-infected
adults in Abidjan, Cote d’Ivoire: a randomised trial. Cotrimo-CI
Study Group. Lancet 353, 1463–1468.
Batungwanayo J, Taelman H, Lucas S et al. (1994) Pulmonary
disease associated with the human immunodeficiency virus in
Kigali, Rwanda. A fiberoptic bronchoscopic study of 111 cases
of undetermined etiology. American Journal of Respiratory and
Critical Care Medicine 149, 1591–1596.
Boeree MJ, Harries AD, Zijlstra EE, Taylor TE & Molyneux ME
(1999) Cotrimoxazole in HIV-1infection. Lancet 354, 334.
van den Broek J, Mfinanga S, O’Brien R, Mugomela A & Lefi M
(1998) Impact of human immunodeficiency virus infection on
the outcome of treatment and survival of tuberculosis patients in
Mwanza, Tanzania. International Journal of Tuberculosis and
Lung Disease 2, 547–552.
Chakaya JM, Bii C, Ng’ang’a L et al. (2003) Pneumocystis carinii
pneumonia in HIV/AIDS patients at an urban district hospital
in Kenya. East African Medical Journal 80, 30–35.
731

Tropical Medicine and International Health
M. J. Boeree et al. Cotrimoxazole prophylaxis in HIV/TB in Malawi
Chintu PC, Bhat G, Walker A et al. (2004) Cotrimoxazole as
prophylaxis against opportunistic infections in HIV-infected
Zambian children (CHAP): a double-blind randomised placebo-
controlled trial. Lancet 364, 1865–1871.
Cranendonk RJ, Kodde CJ, Chipeta D, Zijlstra EE & Sluiters JF
(2003) Prevalence of Cryptosporidium parvum and Isospora
belli infections among patients with and without diarrhoea in
the medical wards at Queen Elizabeth Central Hospital, Blan-
tyre, Malawi. East African Medical Journal 80, 398–401.
Dean GL, Edwards SG, Ives NJ et al. (2002) Treatment of tuber-
culosis in HIV-infected persons in the era of highly active anti-
retroviral therapy. AIDS 16, 75–83.
Eholie SP, Adou-Brynh D, Domoua K et al. (2000) Adult non-viral
lymphocytic meningitis in Abidjan (Cote d’Ivoire). Bulletin de la
Societé de Pathologie Exotique 93, 50–54.
Elvin KM, Lumbwe CM, Luo NP, BkOrkman A, Källenius G &
Linder E (1989) Pneumocystis carinii is not a major cause of
pneumonia in HIV infected patients in Lusaka, Zambia.
Transactions of the Royal Society of Tropical Medicine and
Hygiene 83, 553–555.
Gilks CF, Brindle RJ, Otieno LS et al. (1990) Life-threatening
bacteraemia in HIV-1 seropositive adults admitted to hospital in
Nairobi, Kenya. Lancet 336, 545–549.
Gordon MA, Walsh AL, Chaponda M et al. (2001a) Bacteraemia
and mortality among adult medical admissions in Malawi –
predominance of non-typhi salmonellae and Streptococcus
pneumoniae. Journal of Infection 42, 44–49.
Gordon SB, Molyneux ME, Boeree MJ et al. (2001b) Opsonic
phagocytosis of Streptococcus pneumoniae by alveolar macr-
ophages is not impaired in human immunodeficiency virus-
infected Malawian adults. Journal of Infectious Diseases 184,
1345–1349.
Gordon SB, Chaponda M, Walsh AL et al. (2002) Pneumococcal
disease in HIV-infected Malawian adults: acute mortality and
long-term survival. AIDS 16, 1409–1417.
Graham SM, Mtitimila EI, Kamanga HS, Walsh AL, Hart CA &
Molyneux ME (2000) Clinical presentation and outcome of
Pneumocystis carinii pneumonia in Malawian children. Lancet
355, 369–373.
Grant AD, Kaplan JE & De Cock KM (2001) Preventing
opportunistic infections among human immunodeficiency virus-
infected adults in African countries. American Journal of
Tropical Medicine and Hygiene 65, 810–821.
Greenberg AE, Lucas S, Tossou O et al. (1995) Autopsy-proven
causes of death in HIV-infected patients treated for tuberculosis
in Abidjan, Cote d’Ivoire. AIDS 9, 1251–1254.
Hardy WD, Feinberg J, Finkelstein DM et al. (1992) A controlled
trial of trimethoprim-sulfamethoxazole or aerosolized pen-
tamidine for secondary prophylaxis of Pneumocystis carinii
pneumonia in patients with the acquired immunodeficiency
syndrome. AIDS Clinical Trials Group Protocol 021. New
England Journal of Medicine 327, 1842–1848.
Hargreaves NJ, Kadzakumanja O, Phiri S et al. (2001) Pneumo-
cystis carinii pneumonia in patients being registered for smear-
negative pulmonary tuberculosis in Malawi. Transactions of the
Royal Society of Tropical Medicine and Hygiene 95, 402–408.
732
13653156, 2005, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-3156.2005.01433.x, Wiley Online Library on [30/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
volume 10 no 8 pp 723–733 august 2005
Harries AD, Nyangulu DS, Kang’ombe C et al. (1998) Treatment
outcome of an unselected cohort of tuberculosis patients in
relation to human immunodeficiency virus serostatus in Zomba
Hospital, Malawi. Transactions of the Royal Society of Tropical
Medicine and Hygiene 92, 343–347.
IDA Formularium (2003) Catalogue. http://www.ida.nl
Iyer JK, Milhous WK, Cortese JF, Kublin JG & Plowe CV (2001)
Plasmodium falciparum cross-resistance between trimethoprim
and pyrimethamine. Lancet 358, 1066–1067.
Kang’ombe C, Harries AD, Banda H et al. (2000) High
mortality rates in tuberculosis patients in Zomba Hospital,
Malawi, during 32 months of follow-up. Transactions of the
Royal Society of Tropical Medicine and Hygiene 94, 305–309.
Kelly P, Davies SE, Mandanda B et al. (1997) Enteropathy in
Zambians with HIV related diarrhoea: regression modelling of
potential determinants of mucosal damage. Gut 41, 811–816.
Lucas SB, Hounnou A, Peacock C et al. (1993) The mortality and
pathology of HIV infection in a west African city. AIDS 7,
1569–1579.
Luft BJ & Remington JS (1992) Toxoplasmic encephalitis in AIDS.
Clinical Infectious Diseases 15, 211–222.
Malin AS, Gwanzura LK, Klein S, Robertson VJ, Musvaire P &
Mason PR (1995) Pneumocystis carinii pneumonia in Zim-
babwe. Lancet 346, 1258–1261.
Maynart M, Lièvre L, Sow PS et al. (2001) Primary prevention
with cotrimoxazole for HIV-1-infected adults: results of the
pilot study in Dakar, Senegal. Journal of Acquired Immune
Deficiency Syndromes 26, 130–136.
Mermin J, Lule J, Ekwaru JP et al. (2004) Effect of cotrimoxazole
prophylaxis on morbidity, mortality, CD4-cell count, and viral
load in HIV infection in rural Uganda. Lancet 364, 1428–1434.
Mukadi YD, Maher D & Harries A (2001) Tuberculosis case
fatality rates in high HIV prevalence populations in Sub-Saharan
Africa. AIDS 15, 143–152.
Mwaungulu FB, Floyd S, Crampin AC et al. (2004) Cotrimoxazole
prophylaxis reduces mortality in human immunodeficiency
virus-positive tuberculosis patients in Karonga District, Malawi.
Bulletin of the World Health Organization 82, 354–363.
Navin TR, Rimland D, Lennox JL et al. (2000) Risk factors for
community-acquired pneumonia among persons infected with
human immunodeficiency virus. Journal of Infectious Diseases
181, 158–164.
Omar SA, Bakari A, Owiti A, Adagu IS & Warhurst DC (2001)
Cotrimoxazole compared with sulfadoxine-pyremathamine in
the treatment of uncomplicated malaria in Kenyan children.
Transactions of the Royal Society of Tropical Medicine and
Hygiene 95, 657–660.
Robberts FJ, Chalkley LJ & Liebowitz LD (2002) Pneumocystis
pneumonia. Journal of the South African Dental Association 57,
451–453.
Schneider MM, Nielsen TL, Nelsing S et al. (1995) Efficacy and
toxicity of two doses of trimethoprim-sulfamethoxazole as
primary prophylaxis against Pneumocystis carinii pneumonia in
patients with human immunodeficiency virus. Dutch AIDS
Treatment Group. Journal of Infectious Diseases 171, 1632–
1636.
ª 2005 Blackwell Publishing Ltd
 13653156, 2005, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-3156.2005.01433.x, Wiley Online Library on [30/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Tropical Medicine and International Health volume 10 no 8 pp 723–733 august 2005

M. J. Boeree et al. Cotrimoxazole prophylaxis in HIV/TB in Malawi

Steen TW, Aruwa JE & Hone NM (2001) The epidemiology of
adult lung disease in Botswana. International Journal of
Tuberculosis and Lung Disease 5, 775–782.
Taylor HR (1995) The use of placebo controls. New England
Journal of Medicine 332, 60–61.
UNAIDS (2000) Provisional WHO/UNAIDS Secretariat Recom-
mendations on the Use of Cotrimoxazole Prophylaxis in Adults
and Children Living with HIV/AIDS in Africa. UNAIDS, Gen-
eva, Switzerland.
van der Ven A, Vree TB, Koopmans PP & van der Meer JW (1996)
Adverse reactions to cotrimoxazole in HIV infection: a
reappraisal of the glutathione-hydroxylamine hypothesis. Jour-
nal of Antimicrobial Chemotherapy 37 (Suppl. B), 55–60.
Wiktor SZ, Sassan-Morokro M, Grant AD et al. (1999) Efficacy
of trimethoprim-sulphamethoxazole prophylaxis to decrease
morbidity and mortality in HIV-1-infected patients with tu-
berculosis in Abidjan, Cote d’Ivoire: a randomised controlled
trial. Lancet 353, 1469–1475.
Zachariah R, Harries AD, Spielmann MP et al. (2002) Changes in
Escherichia coli resistance to cotrimoxazole in tuberculosis
patients and in relation to cotrimoxazole prophylaxis in Thyolo,
Malawi. Transactions of the Royal Society of Tropical Medicine
and Hygiene 96, 202–204.
Zachariah R, Spielmann ML, Chingi C et al. (2003) Voluntary
counselling, HIV testing and adjunctive cotrimoxazole testing
reduces mortality in tuberculosis patients in Thyolo, Malawi.
AIDS 17, 1053–1061.
Zangerle R & Allerberger F (1991) Effect of prophylaxis against
Pneumocystis carinii on Toxoplasma encephalitis. Lancet 337,
1232.

Authors
M. J. Boeree, Department of Respiratory Diseases, Radboud University Nijmegen and University Lung Centre Dekkerswald,
PO Box 66, 6560 GB Groesbeek, the Netherlands. Tel.: +31 24 6859911; Fax: +31 24 6859; E-mail: m.boeree@ulc.umcn.nl
(corresponding author).
Delphine Sauvageot, Epicentre, 42 bis, bd Richard Le noir, 75011 Paris, France. Tel.: +33 1 40 21 28 15;
E-mail: delphine.sauvageot@paris.msf.org
Hastings T. Banda and Eduard E. Zijlstra, College of Medicine, Private Bag 360, Chichiri, Blantyre 3, Malawi.
E-mail: htbanda@hotmail.com; eezijlstra@malawi.net
Anthony D. Harries, HIV/AIDS Unit, Ministry of Health, PO Box 30377, Lilongwe, Malawi. E-mail: adharries@malawi.net

ª 2005 Blackwell Publishing Ltd 733