In a cross-sectional study, 137 ambulatory participants 18 years or older residing in Kampala district or 20 km from the study site if residence was outside Kampala in Uganda were contacted to participate in the study. Participants with prior tuberculosis treatment and those who declined written consent were excluded. The study was conducted at the National Tuberculosis and Leprosy Program (NTLP) Clinic of Mulago tertiary teaching hospital complex between November 2007 and May 2008. The tuberculosis patients were recruited at Mulago NTLP Clinic; HIV sero-positive participants without tuberculosis at the Infectious Disease Institute Clinic (IDI) located 300 meters away from the Mulago NTLP Clinic; and HIV sero-negative participants without tuberculosis from the community where enrolled tuberculosis patients resided. The community individuals were selected through random pick of consenting adults from the immediate neighborhood households to those for tuberculosis patients who had presented at the hospital.
The institutional review boards at Case Western Reserve University and Joint Clinical Research Center approved the study, with final approval by the Uganda National Council for Science and Technology. All participants provided written informed consent to the study.
All subjects in the study were given appropriate pre- and post-test HIV counseling and AIDS education. HIV-1 infection was diagnosed on the basis of a positive enzyme-linked immunosorbent assay for HIV-1 antibodies (Recombigen; Cambridge Biotech, Cambridge, MA). At enrollment, basic demographic information and a medical history were collected, and a standardized physical examination was conducted by a medical officer. Active pulmonary tuberculosis was confirmed by sputum smear microscopy and culture. Patients with active tuberculosis were treated with standard four-drug chemotherapy for tuberculosis per guidelines of the Ugandan Ministry of Health. HIV sero-positive patients eligible for antiretroviral therapy were started on treatment and cotrimoxazole prophylaxis at the IDI clinic.
Nutrient intake assessment was made with a single 24-hour dietary recall using a pre-tested questionnaire. Assessment was conducted by four trained study staff nurses and supervised by a nutritionist using local food photographs, portion-size images, and volumetric vessels to increase the accuracy of the recall from the previous 24 hours. The nutritive value of ingredients was computed using the East African food composition table database whose database was imported into the NutriSurvey software (
http://www.nutrisurvey.de) to perform computations. When the East African food composition table was found deficient in certain food items, the United States Department of Agriculture database and the African composition table were used.
We defined body wasting of participants using body mass index (BMI) and height-normalized indices (adjusted for height) for body composition of lean mass index (LMI) and fat mass index (FMI). BMI can be partitioned into height-normalized LMI and FMI as explained as elsewhere
[10–12]. Height-normalized LMI and FMI eliminate differences in fat and fat-free mass associated with height
[12, 13]. The LMI and FMI have the advantages of compensating for differences in height and age
; the use of the LMI and FMI also eliminates some of the differences between population groups. We defined body wasting according to sex. For men, wasting was present when the LMI was below 16.7 (kg/m2) and/or the FMI was below 1.8 (kg/m2). For women, wasting was present with the LMI below 14.6 (kg/m2) and/or FMI below 3.9 (kg/m2). These cutoffs correlate with the BMI according to the WHO cutoff for malnutrition
[10, 11, 15] among adults.
Anthropometric measurements included height and weight. BMI was computed using the relationship of weight in kilograms divided by height in meters squared (kg/m2). Weight was taken using Hanson digital electronic scales to the nearest 100g. Height was measured to the nearest cm of standing height using a stadiometer. All anthropometric measurement values were means of duplicates.
The single-frequency bioelectrical impedance analyzer (BIA Detroit, MI, RJL Systems) performing at 50 kHz and 800 mA was used for BIA measures with detecting electrodes placed on the wrist and ankle and signal introduction electrodes placed on the first joint of the middle finger and behind the middle toe. Before performing measurements on each subject, the BIA instrument was calibrated using the manufacturer’s recalibration device. Measurements were taken at rest in participants with voided bladder, ambient temperature, and had not exercised or taken alcohol. The resistance and reactance were based on measures of a series circuit
. BIA measurements were performed in duplicate for each subject. Lean tissue mass was calculated from BIA measurements using equations that were previously cross-validated in a sample of patients (white, black and Hispanic) with and without HIV infection
 and have been applied elsewhere in African studies
[17–19]. Fat mass was calculated as body weight minus lean tissue mass.
We classified tuberculosis patients into two categories of clinical disease severity using the clinical TBscore
. The clinical TBscore is a low-cost tool that has been developed and validated to assess severity of clinical tuberculosis disease in resource limited settings
. The score has a maximum of 13 clinical variables each scoring 1 point as follows: self-reported symptoms of cough, hemoptysis, dyspnea, chest pain, and night sweating; anemic conjunctivae, tachycardia, positive finding at lung auscultation, axillary temperature >37.00C, BMI <18, BMI <16, middle upper arm circumference (MUAC) <220 mm, and MUAC <200 mm. Two clinical severity categories were generated using the following cutoffs as reported elsewhere
: mild 0 – 5 and moderate-to-severe > 5. Higher scores are associated with severe disease. HIV sero-negative participants without tuberculosis were classified as no disease category.
All study participants in the analysis were categorized into 1) 4 mutually independents groups: HIV sero-positive patients with and without tuberculosis disease, HIV sero-negative patients with and without tuberculosis, 2) wasted and non-wasted using cutoff of low BMI and low LMI, and 3) mild and moderate-to-severe TB disease using the clinical TBscore
. Measures of central tendency and variability were compared between women and men across 4 mutually exclusive groups, between patients with and patients without wasting, and between patients with mild and with moderate-to-severe tuberculosis disease using Wilcoxon-Mann Whitney test for average weight, height, BMI, LMI, FMI, and nutrient intake parameters due to lack of normality. A p-value of <0.05 was considered significant in all analyses. A sample size of 14 participants was computed per group to detect 5 mega joules (MJ) difference in energy intake between wasted and participants without wasting or between patients with mild and patients with moderate-severe disease assuming type I error of 5%, 80% power, and normal distribution with standard deviation of 4.5. All analyses were performed using SAS version 9.2 (Cary software, North Carolina SAS Institute Inc 2004).