Study design
This randomized, double-blind, placebo-controlled phase IV clinical trial was performed at Dourados State Penitentiary in Mato Grosso do Sul, Brazil. From September 2017 to June 2018, we enrolled male inmates between 18 and 45 years of age, serving time in Dourados State Penitentiary with a negative QFT test. We excluded all participants testing QFT positive (Antigen—nil ≥0.35 IU/ml), previously treated for tuberculosis or currently with tuberculosis, alcohol abuse, a history of epilepsy or current epilepsy, HIV positive, hepatitis B or C positive, or demonstrating serum concentrations of glutamic-oxaloacetic transaminase (SGOT) or serum concentrations of glutamic-pyruvic transaminase (SGPT) three times above the upper limit and the Alcohol Use Disorders Identification Test (AUDIT) greater than or equal to 15.
In the first three months of recruitment, we included individuals with a negative test as defined by the manufacturer (<0.35 IU/ml). However, during the study, we used a stricter definition of a negative test (<0.20 IU/ml), based on emerging evidence that QFT positivity thresholds require re-evaluation.[16–18]
The study protocol was approved by our institutional review board (#2.758.659, August 2016, State University of Mato Grosso do Sul, Dourados, Brazil). All participants provided written informed consent and, study data were collected and managed using electronic data capture tools.
Randomization and blinding
Participants were randomized into two arms, intervention and control, using 1:1 allocation. We performed a simple randomization in a single block with 728 elements using alphanumeric codes, generated by the principal investigator. For each participant, the unblinded study pharmacist assigned one of these codes, which placed it in one arm. No other team member (including laboratory staff) had access to the code identification key, thus ensuring continuity of blindness. The codes were only opened in case of proven need; e.g., clinical management of a serious adverse event. Blinding occurred immediately following randomization. Specifically, after randomization, the pharmacist prepared identical bottles containing isoniazid or placebo and referred them to the field staff. The vials were identified by labels with the participant’s name, the study identification number, and the randomization code.
Study outcomes
The planned primary study outcome was QFT conversion at the 12-month visit, defined as an interferon-γ value ≥0.35 IU/ml, for either tube. However, since the trial was stopped early, we analyzed QFT conversion at the 6-month visit. As a secondary analysis, we analyzed QFT conversion at any time point in the study. Exploratory analyses of QFT conversions at alternative interferon-γ value cutoffs (≥0.7, ≥2.0, and ≥4.0 IU/ml) were performed, based on recent evidence linking high quantitative QFT values and disease.[16–18]
We used definitions from the Guideline by International Conference Harmonization[19] for adverse events and severe adverse events. Adverse event severity was assessed using a toxicity table from the Common Terminology Criteria for Adverse Events[20] and the grouping in body systems according the Medical Dictionary for Regulatory Activities (MedDRA). We defined hepatotoxicity as SGOT or SGPT >5 times the normal and symptoms of hepatitis: unexplained anorexia, nausea, vomiting, dark urine, icterus, rash, persistent paresthesia of the hands and feet, persistent fatigue, weakness or fever of greater than three days duration/or abdominal tenderness, especially right upper quadrant discomfort.[21, 22]
Study activities
The intervention arm received 900mg of isoniazid orally twice weekly under directly observed therapy (DOT). The control arm received placebo under the same conditions. The study was conducted between October 2017 and July 2018. All participants received two weekly visits, which, in addition to DOT, included questioning and observation of signs and symptoms of tuberculosis and adverse events.
Additional visits were performed in the third month of follow-up, in which participants were clinically evaluated and had a blood sample collected for SGOT and SGPT testing. In addition to clinical evaluations, liver enzyme testing, QFT, and sputum culture were performed in the sixth month and at the termination of the study for those who had not reached six months of follow-up.
Statistical analysis
We estimated from previous studies[12] that 560 individuals would be required to achieve 90% power to detect a 50% difference in QFT conversion between the intervention group and the control group. The level of significance was set at 5%. Accounting for a 30% loss to follow-up, the enrollment goal was 728 individuals.
After recruitment of at least 100 participants in the 6th month, an interim analysis was performed by an independent data and safety monitoring board. We performed this evaluation to verify the feasibility of continuity of the study and adequacy of the sample calculation. If the primary outcome difference between arms was less than 2.5%, discontinuation of the study was recommended.
The complete analysis was performed including all participants who were still on DOT, as well as participants completing six months of follow-up. For adverse events, we calculated the percent using the number of participants who received at least one dose, and in the calculation of the severity percent, we used the total number of events recorded.
The outcomes are reported as percentages when converted, and group comparisons are reported as intervention efficacy (1-RR, where RR is the relative risk) with confidence intervals. The P value was estimated using a two-sample test for equivalence of proportions and statistical significance was set at P<0.05. For all expected values less than 5, Fisher’s exact test was used. Otherwise, chi-square significance testing was used. For continuous variables, 2-sided t-tests were used for independent samples. Analyses were performed with SPSS 23 (IBM SPSS, Chicago, USA) and SAS 9.4 (SAS Institute Inc., Cary, USA).