BACKGROUND: Uncertainty exists regarding patient outcomes when using TNF inhibitors versus other biological and targeted synthetic disease-modifying antirheumatic drugs (DMARDs) in rheumatoid arthritis-associated interstitial lung disease (ILD). We compared survival and respiratory hospitalisation outcomes following initiation of TNF-inhibitor or non-TNF inhibitor biological or targeted synthetic DMARDs for treatment of rheumatoid arthritis-associated ILD. METHODS: We did a retrospective, active-comparator, new-user, observational cohort study with propensity score matching following the target trial emulation framework using US Department of Veterans Affairs (VA) electronic and administrative health records. VA health-care enrollees with rheumatoid arthritis-associated ILD and no previous receipt of ILD-directed therapies (eg, antifibrotics) who initiated a TNF inhibitor or non-TNF inhibitor between Jan 1, 2006, and Dec 31, 2018, were included. Propensity score matching was performed using demographics, health-care use, health behaviours, comorbidity burden, rheumatoid arthritis-related severity factors, and ILD-related severity factors, including baseline forced vital capacity. Study outcomes were respiratory hospitalisation, all-cause mortality, and respiratory-related death over follow-up of up to 3 years, from VA, Medicare, and National Death Index data. People with lived experience of rheumatoid arthritis-associated ILD were not involved in the design or conduct of this study. FINDINGS: Of 1047 patients with rheumatoid arthritis-associated-ILD who initiated biological or targeted synthetic DMARDs, we matched 237 patients who had initiated TNF inhibitors and 237 who had initiated non-TNF inhibitors (mean age 68 years [SD 9]); 434 (92%) of 474 were male and 40 (8%) were female. Death and respiratory hospitalisation did not significantly differ between groups (adjusted hazard ratio 1·21 [95% CI 0·92-1·58]). Respiratory hospitalisation (1·27 [0·91-1·76]), all-cause mortality (1·15 [0·83-1·60]), and respiratory mortality (1·38 [0·79-2·42]) did not differ between groups. Secondary, sensitivity, and subgroup analyses supported the primary findings. INTERPRETATION: In US veterans with rheumatoid arthritis-associated ILD, no difference in outcomes were seen between those who started TNF inhibitors compared to those starting non-TNF biological or targeted synthetic DMARDs. These data do not support systematic avoidance of TNF inhibitors in all people with rheumatoid arthritis-associated ILD. Comparative efficacy trials in patients with rheumatoid arthritis-associated ILD are needed given the potential for residual confounding and selection bias in observational studies. FUNDING: US Department of Veterans Affairs.