We retrospectively studied all PLHIV who initiated ART during 2021–2022, using data from Mozambique’s national ART data warehouse, MozART (Appendix, https://wwwnc.cdc.gov/EID/article/32/3/25-1349-App1.pdf); as of December 2024, the warehouse contained deidentified clinical-encounter data from 620 facilities and 1.6 million PLHIV on ART. We excluded PLHIV with previous TB treatment or TB diagnosis within 90 days of ART initiation; we also excluded those without a documented HIV viral load test within 2 years or ART pickup within 3 months of ART initiation, as a proxy for nonengagement in care.

The outcome of interest was any TB diagnosis within 2 years of ART initiation, as documented in medical records or by enrollment in the TB service ward. The primary independent variable was TPT completion, categorized as did not start TPT, incomplete TPT (<170 days for INH; <80 days for 3HP), and completed TPT (>170 days for INH; >80 days for 3HP). Control variables were age at ART initiation (<15, 15–49, >50 years); sex; facility setting (urban/rural) and region; most recent CD4 count (<200 cells/mm³, >200 cells/mm³) and World Health Organization (WHO) clinical stage (I–IV) at ART initiation; and most recent HIV viral load result (suppressed, <1,000 copies/mL; unsuppressed, >1,000 copies/mL) before or after TB diagnosis or <24 months on ART and time of TPT initiation (at ART start or delayed).

We used Poisson regression with generalized estimating equations to calculate incidence rate ratios (IRRs) and to account for potential correlation between observations within health facilities. We used IRRs to compare completion categories. We calculated incidence of diagnosed TB as the number of PLHIV with TB diagnosis per total person-years divided by 1,000.

During 2021–2022, a total of 505,098 PLHIV were newly initiated on ART; of those, 28,498 (5.6%) were <15 years of age. We included a total of 341,844 (67.7%) in the analysis; 215,357 (63%) were female and 126,487 (37%) male, median age was 30.0 (interquartile range [IQR] 24–39) years, and 82,418 (24%) received ART in the Southern Region (Table 1). Of the 163,254 PLHIV excluded from the analysis, 120,555 (23.9%) did not have a registered viral load, 28,620 (5.7%) did not have an ART pickup within 30 days of ART initiation, and 14,079 (2.8%) received their diagnosis before initiating ART.

Of 341,844 PLHIV analyzed, 266,977 (78%) completed TPT, 68,594 (20%) had incomplete TPT, and 6,273 (1.8%) did not start TPT (Table 1). Incidence of diagnosed TB was 5.0/1,000 person-years. Compared with those who completed TPT (3.1/1,000 person-years), diagnosed TB was higher among PLHIV who did not start TPT (21.6/1,000 person-years; IRR 6.9 [95% CI 5.7–8.3]) and those with incomplete TPT (11.0/1,000 person-years; IRR 3.5 [95% CI 3.2–3.8]) (Table 2).

Median time between ART initiation and TB diagnosis was shortest for PLHIV who did not start TPT (182 [IQR 118–323] days), followed by those with incomplete TPT (216 [IQR 141–408] days) and those who completed TPT (418 IQR 294–560] days). Among those who completed TPT, diagnosed TB was higher in male PLHIV (IRR 1.5 [95% CI 1.4–1.7]), younger and older age groups (<15 years, IRR 1.5 [95% CI 1.4–1.7]; >50 years, IRR 1.7 [95% CI 1.5–2.0]), and in the Southern Region (IRR 1.6 [95% CI 1.3–1.9]) (Table 3). Other risk factors were unsuppressed viral load (IRR 2.3 [95% CI 2.0–2.7]), CD4 count <200 cells/mm³ (IRR 1.9 [95% CI 1.5–2.3]), and WHO clinical stage II–IV at ART initiation (stage II, IRR 1.6 [95% CI 1.3–1.8]; stage III, IRR 2.5 [95% CI 2.1–2.9]; stage IV, IRR 1.7 [95% CI 1.1–2.7]).

In Mozambique, TPT was associated with reduced incidence of TB disease among PLHIV, including among those with incomplete TPT. TPT likely averted TB among PLHIV, potentially saving lives and reducing costs to the health system. Despite that success, certain groups had an elevated risk of developing TB disease, even after TPT completion: men, younger and older PLHIV, and those with poor clinical status (unsuppressed HIV viral load, WHO clinical stages II–IV, and low CD4 count). Our findings are consistent with studies on TB incidence after TPT among PLHIV in sub-Saharan Africa (7,8). In addition, PLHIV in the Southern Region had elevated incidence, which aligns with higher overall TB incidence rates in the region; it might also reflect regional differences in TB services, including TB case detection (9; D. Respeito et al.).

PLHIV who completed TPT had a longer median time to TB diagnosis than did those who did not complete TPT. Our results aligned with similar studies in sub-Saharan Africa and India (10–12). Because TB diagnosis among those completing TPT occurred »14 months after ART initiation, a second course of TPT after a year could be beneficial, particularly among groups with elevated incidence. However, a recent randomized controlled trial did not find additional benefit from a repeat course of 3HP in Ethiopia, Mozambique, and South Africa (2,13). There is insufficient evidence for additional courses of TPT for PLHIV who completed TPT without new TB exposures (2).

Our findings are strengthened by data from a large national patient cohort containing 85% of PLHIV on ART. However, routine data are prone to quality concerns (i.e., data entry errors, poor completion of clinical tools); continual quality assurance activities and rigorous study exclusion criteria helped mitigate potential data-quality bias. We analyzed available programmatic data, which did not include other variables affecting TB (e.g., underlying conditions, body mass index). Misclassification could have occurred, given reliance on PLHIV self-reporting as newly initiated on ART and data cleaning assumptions. Despite those limitations, our study shows high concordance with other studies analyzing the effectiveness of TPT among PLHIV (14) and underscores the value of using a national data warehouse for analyzing programmatic outcomes.

In summary, TPT was associated with reduced TB incidence among PLHIV in Mozambique, including those who had not completed TPT. Improving TPT initiation and completion, along with focusing on men, younger and older PLHIV, and those with poor clinical status, could further reduce TB incidence.