Multimorbidity and Its Impact in Older U.S. Veterans Newly Treated for Advanced Non-Small Cell Lung Cancer

Rationale: Older adults make up the majority of patients with advanced non-small cell lung cancer (NSCLC) and often carry multiple other comorbidities (multimorbidity) when initiating treatment. The nature and impact of multimorbidity remain largely unknown, given the limitations of standard count-based comorbidity indices in aging patients and their exclusion from clinical trials. Objectives: Our objective is to identify and define multimorbidity patterns in older U.S. veterans newly treated for advanced NSCLC in the national Veterans Affairs healthcare system between 2002 to 2020, and whether they are associated with mortality and healthcare use. Methods: We measured 63 chronic conditions in 10,160 veterans aged ⩾65 years newly treated for NSCLC in the national Veterans Affairs healthcare system from 2002 to 2020. Latent class analysis was used to identify patterns of multimorbidity among these conditions, with final patterns determined on the basis of model fit and clinical meaningfulness. Kaplan-Meier and Cox proportional hazards regression analyses were used to evaluate the association of multimorbidity patterns with overall survival (primary outcome) and with emergency department visits and unplanned hospitalizations (secondary outcomes). Results: Five multimorbidity patterns arose from the latent class analysis, with overall survival varying across patterns (log-rank two-sided P < 0.001). Veterans with metabolic diseases (24.7% of all patients; hazard ratio [HR] [95% confidence interval (CI)], 1.10 [1.04-1.16]), psychiatric and substance use disorders (16.0%; HR [95% CI], 1.17 [1.10-1.24]), cardiovascular disease (14.4%; HR [95% CI], 1.22 [1.15-1.30]), and multisystem impairment (10.7%; HR [95% CI], 1.36 [1.26-1.46]) had a higher hazard of death than veterans with common conditions of aging beyond their NSCLC (34.2%, reference), controlling for age, sex, race, days between diagnosis and treatment, date of diagnosis, and NSCLC stage and histology. Associations held after adjusting for the count-based Charlson comorbidity index. Multimorbidity patterns were also independently associated with emergency department visits and unplanned hospitalizations. Conclusions: Our findings reveal that the numerous chronic conditions present in older veterans with late-stage NSCLC cluster together into distinct multimorbidity patterns; the nature of conditions in these patterns carries value beyond their number.