**Background:** The World Health Organization estimates that one-fourth of the world population is infected with Mycobacterium tuberculosis, and diabetes mellitus (DM) triples the risk of active tuberculosis. In Mexico, TB rates have increased 13% since 2015, and one-fifth of TB patients also have type 2 DM. Tuberculosis preventive treatment (TPT) is recommended to prevent progression from infection to disease, but safety data in patients with DM are limited. This study aimed to assess adverse events (AEs) of TPT in patients with type 2 DM.
**Methods:** This open-label, parallel-group, randomized controlled trial was conducted at a tertiary care center in Mexico City from July 2017 to May 2019. Eligible participants were adults ≥18 years with type 2 DM and a positive tuberculin skin test (≥10 mm). Exclusion criteria included active TB, HIV, immunosuppressive therapy, severe peripheral neuropathy, hepatic or kidney disease, excessive alcohol use, or known allergy to study drugs. Participants were randomized 1:1 to six months of daily isoniazid (INH) 300 mg plus pyridoxine 75 mg or three months of daily rifampicin (RIF) 600 mg. Follow-up visits occurred on days 15, 30, 60, 90 (both groups), 120, and 180 (INH group). AEs were graded using standard criteria; hepatotoxicity was defined as ALT 5-10× ULN (grade 3) or >10× ULN (grade 4). An independent Adverse Event Safety Panel, blinded to treatment allocation, adjudicated AEs as possibly or probably related to study drugs. The primary outcome was grade 1-2 rash, recurrent grade 2 hepatotoxicity, or grade 3-5 AEs leading to permanent treatment cessation. Secondary outcomes included grade 3-4 hepatotoxicity and gastrointestinal intolerance. The study was designed to enroll 403 participants but was halted early due to safety concerns.
**Key Results:** A total of 131 subjects were randomized (68 INH, 63 RIF); 130 were analyzed (68 INH, 62 RIF). Median age was 57 years (IQR 50-62), 41.2% were men, and median glycated hemoglobin was 8.2% (IQR 6.6-10). The INH group had a longer time since DM diagnosis (13 vs. 10 years, p=0.045) and more retinopathy (27.9% vs. 8.0%, p=0.007). The Safety Panel judged 18 events as possibly or probably related to study drugs: 9 (13.23%) in the INH group and 9 (14.51%) in the RIF group (RR 1.09, 95% CI 0.47-2.59). However, all hepatotoxicity events (1 recurrent grade 2 and 6 grade 3-4) occurred in the INH group (8.8% vs. 0% in RIF). In contrast, grade 3-4 gastrointestinal intolerance was more frequent in the RIF group (14.51% vs. 2.94%; RR 4.93, 95% CI 1.11-21.96). Multivariate Cox regression showed that chronic kidney disease (HR 6.44, 95% CI 1.56-26.58, p=0.010) and female sex (HR 4.53, 95% CI 1.25-16.39, p=0.021) were significantly associated with AEs, while treatment arm was not (HR 0.62, 95% CI 0.24-1.57, p=0.311). Overall adherence (≥80% of doses) was 72.31%, with no difference between groups (76.41% INH vs. 67.74% RIF, p=0.26).
**Clinical Implications:** This study provides important safety data on TPT in patients with type 2 DM, a population at high risk for TB reactivation. The high rate of grade 3-4 hepatotoxicity with INH (8.8%) suggests that six months of daily INH may not be safe enough for universal use in this population, especially given the high prevalence of polypharmacy, non-alcoholic fatty liver disease, and poor glycemic control. Conversely, RIF was associated with significant gastrointestinal intolerance, which may limit adherence. The findings underscore the need for alternative TPT regimens with better safety profiles, such as rifapentine-based regimens. The study's limitations include its small sample size due to early termination, open-label design, and baseline imbalances (longer DM duration and more retinopathy in the INH group). Nonetheless, the results are generalizable to similar settings with high DM and TB burden and older patients with uncontrolled diabetes.