Study design and setting
From September 2015 to February 2016, a cross-sectional study conducted at Mulago National TB Treatment Centre, in Kampala, Uganda, which has 60 and 40 beds for patients with DS-TB and DR-TB, respectively.
Participants
Out and in-patients diagnosed with TB and seeking treatment at Mulago National TB Treatment Centre were screened. We consecutively sampled in-patients and used convenience sampling for out-patients. Participants were referred by clinical staff. A research assistant stationed at the Center from Monday to Saturday provided Study information and obtained written informed consent prior to enrollment.
We verified drug resistance and HIV status by reviewing medical files. We excluded patients without Xpert® MTB/RIF or drug susceptibility testing (DST) results, those with history of steroid use in the prior 5 days, fluconazole use, pregnancy, diabetes mellitus, extra pulmonary TB or inability to provide informed consent (refusal, language barrier or unconscious with no attendant).
Variables
We obtained HIV and treatment status, TB past and current treatment, drug-susceptibility test results and verified primary or secondary drug-resistance from the participant’s treatment card. Clinical examination was performed to assess hyperpigmentation of buccal mucosa, palms, and scarred skin, postural hypotension to assess for signs of adrenocortical failure. Early morning blood draws for serum cortisol level, electrolytes and complete blood count were performed.
Data collection and outcome measures
Functional Adrenal Insufficiency, the main outcome variable was defined as early morning serum cortisol ≤ 414 nmol/L [2]. Blood was drawn between 0700 and 0930 hours for serum cortisol, potassium, sodium, calcium, and complete blood count. Total serum cortisol was measured using Cortisol ELISA Test (Diagnostic Automation Inc., California) with detection range 1-100 ng/ml (0.32-31.45 nmol/L). Diagnosis of rifampicin resistant-TB was performed using the Xpert® MTB/RIF assay (Cepheid, Sunnyvale, CA, USA). Drug susceptibility testing for multidrug resistance was performed using the BACTEC MGIT 960 system (Becton Dickinson Microbiology System, Sparks, NV, USA) and Löwenstein-Jensen (L-J) media. An interviewer-administered questionnaire used to collect data on HIV and treatment status, TB past and current treatment, drug-susceptibility test results, patient demographics, and medical history.
Sample size estimation was guided by the difference between 2 proportions formula, assuming a 5% standard error, previous prevalence of FAI of 46% in DR-TB and 36% in DS-TB [21], giving a size of 191 for each group.
Statistical Analysis
Data were entered into Epi-Data version 3.1 and exported to Stata version 13.0 (StataCorp, College Station, TX, USA). A Student t-test was used to assess differences in serum cortisol for DR-TB and DS-TB participants. Associations with FAI were modeled using logistic regression. All significant variables at bivariate analysis were used in multivariate logistic regression model. All analyses were performed using Stata. Two-sided p-values ≤0.05 were considered statistically significant.