Determinants of treatment interruption among patients with multidrug-resistant tuberculosis in Addis Ababa using an unmatched case-control study

Résumé fourni par la source

Multidrug-resistant tuberculosis (MDR-TB) requires prolonged, complex regimens frequently associated with substantial adverse effects, making treatment interruption a major barrier to successful outcomes. Interruption may increase the risk of mortality, acquired drug resistance, and ongoing transmission. Evidence on factors associated with treatment interruption among MDR-TB patients in Ethiopia remains limited, particularly at specialized referral centers. This study aimed to identify factors associated with treatment interruption among patients receiving MDR-TB treatment at Saint Peter Specialized Hospital (SPSH), Addis Ababa, Ethiopia. We conducted a hospital-based, census-based observational study using an unmatched case-control analytic approach at Saint Peter Specialized Hospital. All eligible MDR-TB patients registered over five years (2020–2024) were considered for inclusion. A total of 267 participants (53 cases, 214 controls) were enrolled. Cases were patients who missed at least one prescribed dose for at least one day but less than two consecutive months; controls completed treatment as prescribed without such interruption. Data were collected via structured interviews and medical-record review using KoboToolbox and analyzed in Stata 14.0. Variables with p ≤ 0.25 in bivariate analysis were entered into multivariable logistic regression; associations are reported as adjusted odds ratios (AOR) with 95% confidence intervals (CI), with significance set at p < 0.05. Of 343 patients registered during the study period, 285 were eligible, and 267 were analyzed (53 cases and 214 controls). In multivariable analysis, HIV co-infection (AOR = 3.90; 95% CI: 1.45–10.50), history of alcohol consumption (AOR = 3.32; 95% CI: 1.50–7.30), and male sex (AOR = 2.89; 95% CI: 1.28–6.50) were associated with higher odds of treatment interruption, while bacteriological (versus clinical) diagnosis was associated with lower odds (AOR = 0.29; 95% CI: 0.12–0.67). In this study, HIV co-infection, male sex, and alcohol consumption were associated with higher odds of treatment interruption, whereas bacteriologically confirmed diagnosis was associated with lower odds. Strengthening integrated TB–HIV care, addressing alcohol use, implementing sex-responsive adherence interventions, and expanding access to bacteriological diagnosis may help reduce treatment interruption and improve MDR-TB outcomes in Ethiopia.