Frailty Linked to Incident Cancer in Veterans With RA
Key Takeaways
- Among veterans with rheumatoid arthritis in the Veterans Affairs setting, age-adjusted new-cancer incidence increased from robust through prefrail to frail categories.
- Prefrailty and frailty each had a significant adjusted association with incident cancer compared with robust status.
- Overall frailty had no statistically significant adjusted association with cancer-related mortality.
Researchers retrospectively analyzed 2,723 eligible veterans with RA in the Veterans Affairs Rheumatoid Arthritis Registry. Registry data spanned 2002–2023; those dates describe data coverage, not each participant’s follow-up. The Veterans Affairs Frailty Index (VA-FI) classified baseline status as robust, prefrail, or frail. Participants were older and predominantly male, and Black veterans were represented. Investigators evaluated incident cancer and cancer-related mortality as distinct outcomes using competing-risk models adjusted for demographics, lifestyle factors including smoking, RA characteristics, medications, and comorbidities. They also examined domain-specific frailty exposures and linear frailty modeling. No intervention was assigned.
Age-adjusted incident-cancer rates per 1,000 person-years were 12.3 among robust participants, 15.5 among prefrail participants, and 19.9 among frail participants. These descriptive rates differ from the fully adjusted associations estimated by the competing-risk models.
With robust status as the reference, fully adjusted incident-cancer associations were significant for both prefrailty and frailty. The subdistribution hazard ratio (sHR) for prefrailty was 1.37 (95% confidence interval [CI] 1.02–1.84); for frailty, the sHR was 1.83 (95% CI 1.23–2.71). These estimates concern new diagnoses, not cancer-related deaths.
Cancer-related mortality rates per 1,000 person-years were 7.8 for robust participants, 10.6 for prefrail participants, and 9.8 for frail participants. Overall frailty had no statistically significant adjusted association with cancer-related mortality, and cause-specific Cox models, linear frailty modeling, and a sensitivity analysis substituting the Rheumatic Disease Comorbidity Index for individual comorbidities were consistent with this result.
Cause-specific Cox models and linear frailty modeling gave consistent incident-cancer results. However, in a sensitivity analysis substituting the Rheumatic Disease Comorbidity Index for individual comorbidities, the frail-versus-robust association attenuated (sHR 1.50, 95% CI 1.02–2.21) and the prefrail association was no longer significant (sHR 1.24, 95% CI 0.93–1.66). Among frailty domains, only the geriatric/other domain was significantly associated with incident cancer (sHR 1.12 per 10% increase in domain score).
The retrospective design supports an association, not a causal claim that frailty produces cancer. Because participants were older, predominantly male U.S. veterans with RA, applicability to other RA populations remains uncertain. The descriptive mortality rates should not be interpreted as adjusted effects.
The authors concluded that baseline frailty was independently associated with incident cancer and may identify higher malignancy risk within this RA cohort.
Clinician Questions
Which rheumatoid arthritis characteristics were included in the adjusted frailty–cancer analysis of veterans?
In veterans with rheumatoid arthritis, competing-risk models accounted for RA seropositivity and Disease Activity Score-28, along with demographics, smoking and other lifestyle factors, medication use, and comorbidities.
Has frailty-guided cancer screening improved outcomes for veterans with rheumatoid arthritis?
The retrospective Veterans Affairs rheumatoid arthritis cohort evaluated baseline frailty in relation to incident cancer and cancer-related deaths. It did not test a frailty-guided screening intervention or whether changed cancer-screening intervals improved outcomes.