Syrian civil war and assessment of tuberculosis among Syrian refugees and local citizens in Mardin

Abstract

Background: We compared tuberculosis (TB) characteristics and outcomes between Syrian refugees and local citizens in Mardin, Turkey (2016–2023), a border province with substantial population mobility. Methods: Retrospective, registry-based cross-sectional analysis of 491 patients (locals n = 456; refugees n = 35). Descriptive comparisons used χ2/Fisher (categorical) and Mann–Whitney U (age). Annual incidence per 100,000 used mid-year denominators (locals: ABPRS/NVI; refugees: DGMM/PMM and UNHCR). For outcomes with significant crude differences (treatment success, BCG scar, transferred-out), age- and sex-adjusted bias-reduced (Firth) logistic regression was applied; p-values from penalized likelihood-ratio (PLR) tests. Results: BCG-scar positivity was lower in refugees than locals (62.9% vs. 93.2%, p < 0.001). Microbiological confirmation remained below WHO targets in both groups. Crude treatment success was lower in refugees (68.6%) than locals (90.4%, p = 0.03), while transferred-out was higher (25.7% vs. 5.3%, p = 0.001). In adjusted Firth models including all cases, refugee status was associated with lower odds of success (aOR 0.224, 95% CI 0.103–0.488; PLR p < 0.001); after excluding transferred-out cases the association attenuated and was not significant (aOR 0.562, 95% CI 0.121–2.605; PLR p = 0.42). In pulmonary-only analyses, the association persisted (aOR 0.216, 95% CI 0.083–0.567; PLR p = 0.002). Refugee incidence dipped in 2020–2021 and rebounded in 2022–2023. Conclusion: Differences likely reflect operational barriers—especially transfers disrupting continuity—rather than intrinsic factors. Refugee-inclusive TB services with robust inter-provincial transfer tracking, patient navigation, and expanded bacteriological testing (notably for extrapulmonary disease) should be prioritized. Given the small refugee subgroup and denominator uncertainties, findings are hypothesis-generating.