C28-04 The Significant Burden of Post-tuberculosis Lung Disease in America: Evidence From the Trinetx Federated Database

Abstract Rationale Tuberculosis (TB) is the leading cause of infectious death with over 10 million cases and 1 million deaths yearly, predominantly in low-and-middle income countries (LMIC’s). In LMIC’s 40-60% of TB survivors develop Post-Tuberculosis Lung Disease (PTLD), defined as symptoms, radiologic abnormality or pulmonary dysfunction following cure. PTLD is associated with increased morbidity and mortality in LMIC cohorts, but the burden of PTLD in high-income settings is unknown. The cause of PTLD is unknown, however, healthcare access and quality barriers are implicated. Using proxy definitions and the TriNetX EHR network we examined the burden and associated outcomes of PTLD in the USA. Methods We leveraged the TriNetX federated EHR network, to identify American patients without pre-existing lung disease, diagnosed with TB who completed RIPE (rifampin, isoniazid, pyrazinamide and ethambutol) AND a visit ≥ 6 months after RIPE. PTLD was defined ≥6 months from TB diagnosis using sensitive (radiologic, oxygen dependence, dyspnea, cough or hypoxemia) and specific (radiologic only) definitions. Propensity matching (1:1 nearest-neighbor without replacement) compared TB and PTLD cohorts to controls (identical characteristics with non-TB-pneumonia, no TB and no RIPE). Matching covariates included age, sex, race, diabetes, chronic kidney disease, heart failure, coronary artery disease, HIV, smoking and economic insecurity. Post-match analyses calculated odds ratios (ORs) and hazard ratios (HRs). Results Our TB cohort included 3,194 individuals from 122,200,614 subjects at 68 HCO’s. Compared to matched pneumonia controls (3,075 per arm), the TB cohort showed no difference in ER visits or lung cancer incidence, but had higher risk of hospital admission, higher all-cause mortality and increased odds of mechanical ventilation, respiratory failure and new oxygen dependence with OR/HR ranging from 1.17 for admission to 1.945 for mechanical ventilation. Within the TB cohort 38.9% (n = 1,243) met the sensitive PTLD definition and 21.7% (n = 692) met the specific definition. Common criteria for PTLD classification included: Dyspnea (56.6%), abnormal imaging (52.8%), and cough (48.7%). The sensitive PTLD cohort was compared with non-TB-pneumonia controls (1,209 per arm), the risk for all outcomes was higher in PTLD with OR/HR’s ranging from 1.51 for mortality to 2.48 for mechanical ventilation. Conclusion In American TB patients, PTLD develops in 21.7-38.9% after cure. Individuals with prior TB, particularly those with PTLD, have increased healthcare utilization, mortality and respiratory complications compared to a well-matched cohort with non-tuberculous pneumonia. This supports the need for ongoing investigations of PTLD and implicates factors beyond healthcare access and quality in PTLD development. This abstract is funded by: None