Share this article on:

Effectiveness of cotrimoxazole prophylaxis on mortality in adults with tuberculosis in rural South Africa

Grimwade, Katea,b,*; Sturm, A Willemc; Nunn, Andrew Jd; Mbatha, Daniela; Zungu, Dawna; Gilks, Charles Fe

Clinical Science: Concise Communication

Background: Adults with dual tuberculosis (TB) and HIV infection have a poor outcome. Studies in West Africa suggest that cotrimoxazole prophylaxis may reduce this mortality.

Objective: To evaluate the effectiveness of cotrimoxazole in reducing mortality in adults with active TB, irrespective of HIV status, in a high prevalence setting.

Design: Cohort study using historical controls.

Methods: Adults treated for TB between 1998 and 2000 were traced and vital status at 6 months ascertained (2004: control group). All adults starting treatment for TB between June 2001 and June 2002 were offered cotrimoxazole prophylaxis 960 mg once daily for 6 months during TB treatment irrespective of HIV status (1321: intervention group). Mortality, adverse reactions and adherence were compared between intervention and control groups.

Results: HIV seroprevalence in patients with TB at the start of the intervention was estimated to be 78%. Mortality at 6 months was 29% lower in the group given cotrimoxazole than in the control group. The number needed to treat to prevent one death during the period of TB treatment was 24. The benefit was seen across all types of TB but was only evident in new patients; patients being retreated had similar outcomes in both groups. Adverse events were infrequent and minor, with only two participants having treatment stopped for this reason.

Conclusion: Cotrimoxazole prophylaxis for all adults with TB, irrespective of HIV status, in an area of high HIV seroprevalence may be a feasible, safe and effective way to reduce mortality for the duration of treatment.

From the aHlabisa Hospital, Kwazulu Natal, South Africa

bSchool of Tropical Medicine, Liverpool, UK

cDepartment of Medical Microbiology, University of KwaZulu Natal, Durban, South Africa

dMRC Clinical Trials Unit, London, UK

eImperial College of Science, Technology and Medicine, London, UK.

Present address: Department of Infection and Tropical Medicine, Heartlands Hospital, Birmingham, UK.

Received 17 June, 2004

Revised 28 October, 2004

Accepted 2 November, 2004

Correspondence to Dr K. Grimwade, Department of Infection and Tropical Medicine, Birmingham Heartlands Hospital, Bordesley Green East, Birmingham, B95SS, UK. E-mail:

Back to Top | Article Outline


HIV-1 infection and tuberculosis (TB) are leading causes of morbidity and mortality in sub-Saharan Africa [1]. Mortality rates in HIV-infected individuals with TB are high [2–4] because of pneumonia, bacterial septicaemia and other opportunistic infections [4,5]. Two randomized controlled trials in West Africa have suggested that prophylactic treatment with the broad-spectrum antibiotic cotrimoxazole may be of benefit to HIV-infected people particularly in those coinfected with TB [6,7]. These studies showed low rates of cotrimoxazole resistance amongst common bacterial pathogens. One small placebo-controlled study (100 participants) of cotrimoxazole prophylaxis in an area of high drug resistance [8] failed to show a beneficial effect. Some of the benefit in the West Africa studies was as a result of prevention of malaria, which has become recognized as an HIV-related infection [9,10]. No data exist for countries with low or no malaria burden. There are serious concerns regarding the possible effect of a widespread prophylaxis programme on antibiotic resistance [11,12].

South Africa has one of the highest rates of dual HIV–TB infection in the world [13]. High levels of resistance to cotrimoxazole exist in many pathogens [14]. In November 2000, the South African AIDS programme published national guidelines for prophylactic treatment of people with HIV [World Health Organization (WHO) stage 2 and above] with cotrimoxazole [15] in line with provisional WHO/UNAIDS recommendations [16]. Few patients complete the full voluntary counselling and testing process at the time of diagnosis of TB and, therefore, rarely receive cotrimoxazole prophylaxis [17] (and unpublished data Hlabisa Hospital). Offering cotrimoxazole to all patients with TB in a high HIV prevalence setting at the onset of anti-TB treatment is simple, may increase uptake of voluntary counselling and testing and may provide substantial benefits. This study was designed to evaluate the effectiveness of cotrimoxazole prophylaxis in preventing mortality in adults with active TB irrespective of HIV status in comparison to historical controls.

Back to Top | Article Outline


The study was conducted in Hlabisa subdistrict, Northern KwaZulu–Natal, South Africa, a rural area with high rates of both TB and HIV. TB had increased from 312 cases in 1991 to over 2200 by the end of 2001 (unpublished TB programme data). HIV had also increased dramatically; seroprevalence in antenatal surveys increased from 4 to 35% [18,19]. TB is managed through a directly observed treatment (DOTS) programme [20] and evidence-based TB diagnosis protocols [21]. All patients were treated with the standard national TB regimen (2 months of rifampin, isoniazid, pyrazinamide and ethambutol followed by 4 months of rifampin and isoniazid). Malaria is uncommon but recent epidemics have occurred [10].

The initial study was approved by ethics committees of the University of Natal, South Africa and the Liverpool School of Tropical Medicine, UK. Study revision was reviewed and approved by the same bodies.

Back to Top | Article Outline

Historical controls

All adult patients with TB (aged 13 years or above) treated in the community from January 1998 to December 2000 were included. Patients were traced to their homes and vital status at 6 and 12 months following the commencement of treatment determined. If the patient had died an attempt was made to document date of death.

Back to Top | Article Outline

Pilot study

During a pilot period (205 patients), when cotrimoxazole was offered only to those testing positive for HIV, it was found that although > 95% of patients received counselling < 10% accepted an HIV test. The main reason for this was difficulty in accepting both diagnoses at once. Given the high HIV seroprevalence in patients with TB, it was decided to offer prophylaxis to all starting anti-TB treatment.

Back to Top | Article Outline

Intervention group

Adults (aged 13 years or above) were enrolled and followed up. Verbal explanations as well as a written information sheet in Zulu were given and all treatment and follow-up was provided free of charge. Cotrimoxazole was provided separately from the TB–DOTS drugs.

Adherence to cotrimoxazole was based on participant reporting and clinic registers, which were signed monthly by staff as more tablets were collected. For the purpose of this analysis, those reporting to be not taking prophylaxis, not collecting new tablets since last follow-up or not attending follow-up were assumed to be non-adherent. Non-adherent participants still attending follow-up were asked for their main reason for stopping cotrimoxazole.

At 12 months, participants were traced to their home and their vital status determined. If they had died, an attempt was made to document date of death.

Back to Top | Article Outline

HIV testing

Those participants who chose to test at any time during the study underwent pre- and post-test counselling and were tested using a two-test algorithm. At the start of enrolment, an anonymous HIV survey of adults being discharged on TB treatment was conducted, with their consent.

Back to Top | Article Outline

Statistical methods

Data were entered and analysed using Epi Info 6.04d (Centers for Disease Control and Prevention, Atlanta, Georgia, USA), linked to the TB programme database. Mortality analysis was done using two-by-two tables and chi-squared test with significance level set at P = 0.05. A two-by-n chi-squared test was used in subgroup analysis for age groups and types of TB.

Back to Top | Article Outline


The homes of 97% (1944/2004) of the control individuals were visited. Fig. 1 details the enrolment and follow-up of the intervention group. Anonymous surveys of patients with TB in Hlabisa showed that HIV rates rose from 36% in 1993 [22] to 58% in 1995 [23]. During the intervention period, HIV seroprevalence was 78% [95% confidence interval (CI), 71–84] in 150 adult patients with TB surveyed anonymously.

Fig. 1

Fig. 1

The characteristics of the two groups and the effect of cotrimoxazole are shown in Tables 1 and 2. There were significant differences in sex, type of TB and category of patient.

Table 1

Table 1

Table 2

Table 2

At 6 months, there was a 29% reduction in death rate between the intervention and control groups (95% CI, 13–45; P < 0.001), from 15% in the control group to 10% in the cotrimoxazole group. Odds ratio for mortality by 6 months in the cotrimoxazole group was 0.68 (95% CI, 0.55–0.86). The number needed to treat to prevent one death was 24. There was no change in mortality rates in the two groups between 6 and 12 months. HIV status was known in < 10% of participants; therefore, analysis by HIV status was not possible.

The greatest effect from prophylaxis was seen in those aged 35–44 years, where the mortality at 6 months fell by 41%, from 21% to 12%. This was more marked in women than men.

Overall, 58% (743 patients) were adherent to cotrimoxazole at 3 months, and 43% (523) patients at 6 months. Female patients adhered to therapy better than male patients, both at 3 months [449 (63%) versus 294 (51%); P < 0.001] and at 6 months [316 (47%) versus 207 (38%); P = 0.02]. Adherence to cotrimoxazole prophylaxis at 3 months was highly predictive of survival at 6 months: 12 (1.8%) of adherent patients were dead at 6 months compared with 27 (6%) of non-adherent patients (P < 0.001).

The most common reasons for non-adherence related to problems in collection of treatment: financial, transport or physical constraints; or clinics were too far away to attend monthly to pick up tablets. Perceived adverse reactions were given as the reason for stopping prophylaxis by 23 patients but only two significant adverse reactions were identified. One patient developed Stevens–Johnson syndrome while still taking TB treatment; this improved on cessation of both prophylaxis and TB drugs. A second patient (known to be HIV positive) developed early exfoliative dermatitis at 9 months after TB diagnosis. Other adverse reactions were minor, mostly itching or nausea, and settled with symptomatic treatment, allowing prophylaxis to be continued.

During TB treatment, 245 (19%) participants chose to have further counselling; 91(37%) of these proceeded to HIV testing and 68 (72%) of those tested were found to be HIV positive. Together with those found to be positive at the start of treatment (47 patients), cotrimoxazole prophylaxis was continued beyond the end of anti-TB treatment. By 1 year, however, only 70 (61%) of these 115 patients were still taking cotrimoxazole.

Back to Top | Article Outline


Mortality was reduced by 29% in comparison with historical controls in a cohort of adult patients with TB taking cotrimoxazole, irrespective of HIV status. The number needed to treat to prevent one death was 24. There was no difference in mortality rates in the two groups between 6 and 12 months probably because few patients (less than 10%) still took prophylaxis. A significant reduction in mortality was seen in patients who had never received TB treatment before, but there was little or no effect in those being retreated. Much of this mortality may be failure to respond to anti-TB treatment rather than secondary bacterial or parasitic infection. This would not be prevented by cotrimoxazole and would be similar in both control and intervention group.

The limited available data suggest that deaths in the control group occurred earlier than those in the cotrimoxazole group. This supports a policy of starting prophylaxis as early as possible, and the approach taken in this study to start cotrimoxazole without knowing HIV status.

The beneficial effect of prophylaxis was more pronounced in women, who were also found to adhere to treatment better than men. The beneficial effect may reflect this better adherence as adherence at 3 months was related to a better outcome at 6 months.

Cotrimoxazole was safe with few adverse reactions identified. Adherence overall was poor, the main reasons being related to problems getting the tablets because of distance to the clinics or lack of understanding of the long-term nature of treatment. These could be addressed by linking the distribution of tablets to TB treatment, either providing tablets via the DOTS supervisor or providing all drugs together for the duration of TB treatment. Additional education may still be required to explain the idea of preventative treatment, which is a difficult concept to grasp in this setting.

With the emergence of programmes to provide antiretroviral therapy in resource-poor settings, the TB service is a priority entry point for treatment. However, establishing the requisite counselling, testing and provision infrastructure will take time. This approach is an immediate way to provide cotrimoxazole safely and simply when services are set up. Once routine counselling and testing are available to new patients with TB, uptake may still be poor, despite the offer of antiretroviral treatment, because of the problems described here of accepting two diagnoses at once. This strategy would allow cotrimoxazole to start immediately and act as an incentive for HIV testing by linking prophylaxis to a complete package of care for newly identified infected individuals that would include antiretroviral treatment.

Historically controlled trials have drawbacks in interpretation. It is possible that the better outcome seen in the intervention arm is related to a higher standard of care. This study implemented the prophylaxis programme with minimal extra resources while formalizing preexisting follow-up to collect the necessary data. Thus the same level of care was maintained across groups.

The HIV epidemic evolved over the course of the study, leading to a higher HIV seroprevalence amongst patients with TB in our intervention group. Probably as a result of this, the type of TB varied, with a reduction in the proportion of those with sputum-positive disease compared with sputum-negative or extrapulmonary TB. Despite these changes, mortality in the intervention group decreased, strengthening the argument for a beneficial effect from prophylaxis.

It is unlikely that the effect of cotrimoxazole was through the prevention of malaria. A malaria epidemic occurred during the summer of 1999–2000. Examination of dates of death in the control group, where available, showed no clustering during the outbreak. Repeating the analysis having excluded those in the control group taking TB treatment during this epidemic did not alter the results.

These results suggest that widespread use of cotrimoxazole may help to reduce mortality in a patient population with TB with an area of high HIV prevalence, pending the introduction of antiretroviral therapy.

Back to Top | Article Outline


We are indebted to Nokukhanya Buthelezi, Nonhlanhla Buthelezi, Jabu Dlamini, Bongi Khumalo, Dumo Simelane and Muzi Sithole for their contributions to enrollment, counseling and follow-up of participants and data entry in Hlabisa; to David Lalloo and Martin Dedicoat in Liverpool for their comments on the paper; and to the staff and patients of Hlabisa health district.

Sponsorship: This study was financed by the UK Department For International Development (DFID) HIV/AIDS/STI Knowledge Programme at the Liverpool School of Tropical Medicine. However, DFID does not accept responsibility for any information supplied or views expressed in this article.

Back to Top | Article Outline


1 UNAIDS. AIDS Epidemic Update: July 2004. Geneva: UNAIDS; 2004.
2 Colebunders RL, Ryder RW, Nzilambi N, Dikulu K, Willame JC, Kaboto M, et al. HIV infection in patients with tuberculosis in Kinshasa, Zaire. Am Rev Respir Dis 1989; 139:1082–1085.
3 de Cock KM, Soro B, Coulibaly IM, Lucas SB. Tuberculosis and HIV in sub-Saharan Africa. JAMA 1992; 268:1581–1587.
4 Brindle RJ, Nunn PP, Batchelor BI, Gathua SN, Kimari JN, Newnham RS, et al. Infection and morbidity in patients with TB in Nairobi, Kenya. AIDS 1993; 7:1469–14674.
5 Churchyard GJ, Kleinschmidt I, Corbett EL, Murray J, Smit J, de Cock KM. Factors associated with an increased case-fatality rate in HIV-infected and non-infected South African gold miners with pulmonary tuberculosis. Int J Tuberc Lung Dis 2000; 4:705–712.
6 Wiktor SZ, Sasson-Morokro M, Grant AD, Abouya L, Karon JM, Maurice C, et al. Efficacy of trimethoprim-sulfamethoxazole prophylaxis to decrease morbidity and mortality in HIV-infected patients with tuberculosis in Abidjan, Côte d’Ivoire: a randomized controlled trial. Lancet 1999; 353:1469–14675.
7 Anglaret X, Chêne G, Attia A, Toure S, Lafont S, Combe P, et al. Early chemoprophylaxis with trimethoprim-sulfamethoxazole for HIV-1 infected adults in Abidjan, Côte d’Ivoire: a randomized trial. Lancet 1999; 353:1463–1468.
8 Maynart M, Lièvre L, Sow PS, Kony S, Gueye NF, Bassene E, et al. Primary prevention with cotrimoxazole for HIV-1 infected adults: results of the pilot study in Dakar, Senegal. J Aquir Immune Defic Syndr 2001; 26:130–136.
9 French N, Nakiyingi J, Lugadda E, Watera C, Whitworth JAG, Gilks CF. Increasing rates of malarial fever with deteriorating immune status in HIV-1-infected Ugandan adults. AIDS 2001; 15:899–906.
10 Grimwade K, French N, Mbatha D, Zungu D, Dedicoat M, Gilks C. HIV-infection as a co-factor for severe complicated falciparum malaria in adults living in a region of unstable malaria transmission in South Africa. AIDS 2004; 18:547–554.
11 Abdel-Haq N, Abuhammour W, Asmar B, Thomas R, Dabbagh S, Gonzalez R. Nasopharyngeal colonization with Streptococcus pneumoniae in children receiving trimethoprim-sulfamethoxazole prophylaxis. Pediatr Infect Dis J 1999; 18:647–649.
12 Martin JN, Rose DA, Hadley WK, Perdeau-Remington F, Lam PK, Gerberding JL. Emergence of trimethoprim–sulfamethoxazole resistance in the AIDS era. J Infect Dis 1999; 180:1809–1818.
13 World Health Organization. Stop TB Partnership. Progress report on the global plan to Stop Tuberculosis. Geneva: World Health Organization; 2004.
14 Liebowitz LD, Slabbert M, Huisamen A. National surveillance programme on susceptibility patterns of respiratory pathogens in South Africa: moxifloxacin compared with eight other antimicrobial agents. J Clin Pathol 2003; 56:344–347.
15 Department of Health, South Africa. HIV/AIDS Policy Guideline. Pretoria: Department of Health, 2000.
16 World Health Organization/UNAIDS. Provisional WHO/UNAIDS Secretariat Recommendations on the Use of Cotrimoxazole Prophylaxis in Adults and Children living with HIV/AIDS in Africa. Geneva: UNAIDS; 2000.
17 Hausler H. Lessons learned in South African TB/HIV pilot districts. S Afr Resp J 2002; 8:17–22.
18 Floyd K, Reid A, Wilkinson D, Gilks C. Admission trends in a rural South African hospital during the early years of the HIV epidemic. JAMA 1999; 11:1087–1091.
19 Rollins N, Dedicoat M, Danaviah S, Page T, Bishop K, Coovadia H, et al. Prevalence, incidence, and mother-to-child transmission of HIV-1 in rural South Africa. Lancet 2002; 360:389–390.
20 Davies GR, Connolly C, Sturm AW, McAdam KP, Wilkinson D. Twice-weekly, directly observed treatment for HIV-infected and uninfected tuberculosis patients: cohort study in rural South Africa. AIDS 1999; 13:811–817.
21 Wilkinson D, Newman W, Reid A, Squire SB, Sturm AW, Gilks CF. Trial-of-antibiotic algorithm for the diagnosis of tuberculosis in a district hospital in a developing country with high HIV prevalence. Int J Tuberc Lung Dis 2000; 4:513–518.
22 Wilkinson D, Moore DA. HIV-related tuberculosis in South Africa: clinical features and outcome. S Afr Med J 1996; 1:64–67.
23 Davies GR, Wilkinson D, Colvin M. HIV and tuberculosis. S Afr Med J 1996; 86:91.

cotrimoxazole prophylaxis; HIV disease; outcome; death; Africa tuberculosis

© 2005 Lippincott Williams & Wilkins, Inc.