Study design and population

From April 2000 to April 2005, adults with pulmonary TB were enrolled in a randomized, double-blind, placebo-controlled trial to examine the effect of micronutrient supplements on TB treatment outcomes and mortality. Details of the study design have been published previously [Reference Villamor5]. Participants were recruited from five outpatient TB clinics in Dar es Salaam, Tanzania. Eligible participants were men and non-pregnant women aged between 18 and 65 years who had at least two positive sputum smears for acid-fast bacilli (AFB). Patients with a Karnofsky performance score of <40% or haemoglobin concentration of <7·0 g/dl were excluded from the study. HIV testing was conducted using two sequential enzyme-linked immunosorbent assays (ELISA) in an alternative confirmatory strategy and discrepant results were resolved by Western blot assay. Participants were stratified by HIV status and randomly assigned to receive a daily dose of micronutrients (5000 IU retinol, 20 mg vitamin B1, 20 mg vitamin B2, 100 mg niacin, 25 mg vitamin B6, 50 μg vitamin B12, 500 mg vitamin C, 200 mg vitamin E, 0·8 mg folic acid, 100 μg selenium) or placebo at the time of initiation of TB treatment. As recommended by the Tanzania National TB and Leprosy Programme at the time of the study, patients received a daily combination of rifampicin, isoniazid, pyrazinamide, and ethambutol at the health facility under direct observation of a healthcare worker for the first 2 months. In the subsequent 6 months, patients self-administered isoniazid and ethambutol at home. Antiretroviral therapy was not available to most HIV-positive participants at the time of the study in Tanzania.

At the baseline visit, trained research assistants obtained information about demographic and socioeconomic characteristics. Follow-up was conducted every month at the study clinics where research nurses inquired about the health of the patient. A physician carried out a complete physical examination every 3 months. Sputum for AFB smears and cultures was collected. A blood sample was obtained to measure haemoglobin concentrations and CD4+, CD8+, and CD3+ T cells using the FACScount or FACScan systems (BD Biosciences, USA). HIV viral load was also measured using the Roche Amplicor v. 1.5 assay (Roche Diagnostics, USA). The HIV disease stage was assessed according to the World Health Organization (WHO) staging system [6].

The study protocol was approved by the Institutional Review Boards at the Muhimbili University of Health and Allied Sciences, the Tanzanian National AIDS Control Program, and the Harvard School of Public Health. All patients provided written informed consent to participate in the trial.

Assessment of lymphocyte proliferation assay response

In vitro lymphocyte proliferation is frequently used as a marker for cell-mediated immune function. Lymphocyte proliferation was measured by [³H]thymidine incorporation after stimulation with T-cell mitogens or TB antigens using the whole-blood culture method. For each patient, 5 ml blood was collected into heparin-containing Vacutainers for the whole-blood assay. The blood was diluted at 1:10 with complete Roswell Park Memorial Institute Medium (RPMI 1640 culture medium supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mmol/l l-glutamine, 25 mmol/l Hepes). Diluted blood was distributed into 96-well tissue culture plates, and then stimulated with concanavalin A (ConA) at 5, 25, and 50 mg/l, phytohaemagglutinin (PHA) at 2, 10, and 50 mg/l, or purified protein derivatives (PPD) at 2, 5, and 20 mg/l. The plates were incubated for 72 h at 37°C in a 5% CO₂ humidified atmosphere. Cultures were pulsed with [³H]thymidine during the final 4 h of incubation. At the end of incubation, the plates were first frozen at –70°C overnight and then stored at –20°C before harvest. After 1 h of incubation at 37°C in trypsin/EDTA solution, cells were harvested using an automatic harvester. Cell proliferation was quantified as the amount of [³H]thymidine incorporation into DNA as determined by a liquid scintillation counter. The results of in vitro stimulation of lymphocytes were reported as counts per minute (cpm). We report results generated with use of the optimal concentration for each stimulant: 25 mg/l ConA, 10 mg/l PHA and 5 mg/l PPD. Similar results were found when other concentrations were used for stimulation (data not shown).

Statistical analysis

Of 887 adults with pulmonary TB (416 HIV-negative and 471 HIV-positive patients) enrolled in the trial, at least one measurement of lymphocyte proliferation at 2 or 8 months was made for 423 patients (267 HIV-negative and 156 HIV-positive patients). Due to logistics issues, lymphocyte proliferation could not be measured in all participants, especially HIV-positive patients. We compared the distribution of baseline characteristics between treatment groups within the strata of HIV status by using the Wilcoxon rank sum test for continuous variables and the χ ² test for categorical variables. An intention-to-treat analysis was performed. To examine the effect of micronutrient supplements on lymphocyte proliferative responses, we conducted an analysis of parallel mean response profiles with adjustment for baseline using generalized linear models for repeated measurements [Reference Fitzmaurice, Laird and Ware7]. The robust empirical variance structure was used. All statistical analyses were conducted using SAS v. 9.1 (SAS Institute Inc., USA).