Diagnostic shifts in extrapulmonary tuberculosis during COVID-19: evidence of vulnerability among migrants in a border province.

backgroundThe diagnosis of extrapulmonary tuberculosis (EPTB) often depends on advanced imaging, invasive tissue sampling, and multidisciplinary referral pathways. These complex, procedure-dependent diagnostic processes are particularly vulnerable to disruption during periods of health system strain. During the COVID-19 pandemic, the reallocation of healthcare resources, restrictions on elective procedures, and delays in diagnostic services may have disproportionately affected EPTB detection, leading to missed or delayed diagnoses. Migrant populations, who already face structural barriers to healthcare access related to language, socioeconomic factors, and referral processes, may be especially vulnerable to such diagnostic disruptions during public health emergencies.

methodsWe conducted a retrospective observational study using complete provincial tuberculosis dispensary records from Mardin, a border province in southeastern Turkey. All tuberculosis cases diagnosed between 2017 and 2025 were included. Diagnostic periods were defined as pre-COVID (2017-2019), COVID (2020-2022), and post-COVID (2023-2025). Cases with concomitant pulmonary and extrapulmonary involvement were classified as pulmonary tuberculosis (PTB). Distributions of PTB and EPTB by diagnostic period and nationality (local vs. migrant) were compared using chi-square or exact tests, as appropriate. For microbiological analyses, hypothesis testing was restricted to cases with available results. Factors associated with EPTB were evaluated using Firth penalized logistic regression to reduce small-sample bias and separation.

resultsA total of 525 tuberculosis cases diagnosed between 2017 and 2025 were included (mean age 37.8 ± 19.3 years; 51.2% male), of whom 49 (9.3%) were migrants. Overall, 235 cases (44.8%) were classified as extrapulmonary tuberculosis (EPTB). In the local population, EPTB proportions varied across periods (47.4% pre-COVID, 35.2% during COVID, and 54.9% post-COVID; p = 0.0034). Among migrants, EPTB proportions also differed across periods (26.9%, 16.7%, and 63.6%, respectively; exact p = 0.0425), although estimates were imprecise due to small strata sizes. Among locals, AFB positivity among performed tests and culture positivity among determinate results varied across periods (p = 0.0041 and p = 0.0001, respectively). In Firth penalized logistic regression male sex and the COVID period were associated with lower odds of EPTB. Migrant status showed lower odds but did not reach statistical significance, and the post-COVID period was not statistically significant. These findings reflect changes in the PTB/EPTB case-mix among notified TB cases rather than population incidence.

conclusionWe observed a transient shift in the PTB/EPTB case-mix during the COVID period with a post-COVID rebound. While these patterns are compatible with pandemic-related changes in diagnostic pathways, mechanisms cannot be directly assessed in registry data without denominators and service-level diagnostic volumes. Maintaining access to referral- and procedure-dependent diagnostic pathways remains important, particularly for vulnerable populations. CLINICAL TRAIL: Not applicable.