How common is frailty within clinical trials? An analysis of individual participant data from industry-sponsored clinical trials

Problem

Frailty is common in clinical practice. Many clinical guidelines warn that clinical trial findings may not be applicable to frail patients. However, few trials explicitly quantify or report frailty. In this study we quantified frailty using individual patient data from industry-sponsored clinical trials of pharmacological interventions. We included three exemplar conditions: type 2 diabetes mellitus (T2DM), chronic obstructive pulmonary disease (COPD), and rheumatoid arthritis (RA).

Approach

Trials were identified from the Clinical Study Data Request repository. Frailty index (FI) was calculated using a previously published standard method. Forty deficits (comorbidities, laboratory deficits, functional limitations and symptoms) meeting standard FI criteria were identified from baseline data. The FI was calculated for each participant as the total number of deficits present, divided by the total number of possible deficits. Participants with an FI >0.24 were considered ‘frail’. FI 0.12-0.24 was considered ‘pre-frailty’. Baseline disease severity was assessed using HbA1c for T2DM, % predicted FEV1 for COPD, and Disease Activity Score-28 (DAS-28) for RA. Using generalised gamma regression, we modelled FI on age, sex and disease severity. In negative binomial regression we modelled serious adverse event rates on FI.

Findings

26 studies were identified from the CSDR repository, of which 13 contained sufficient data to calculate a 40-item FI (5 T2DM, 4 COPD, 4 RA). The remaining trials redacted or did not record individual-level participant data (IPD) on baseline functional measures, medical history and/or laboratory deficits. The prevalence of frailty ranged from 7-22% in T2DM trials, 15-21% in COPD trials, and 32-46% in RA trials. Increased disease severity (HbA1c, FEV1 or DAS-28, respectively) and female sex were associated with higher FI in all trials. Older participants had a higher FI in T2DM and RA trials, however there was no significant association with age in the COPD trials. After adjusting for age, sex and disease severity, in 10 out of the 13 trials, a higher FI was associated with a higher rate of serious adverse events (e.g. incidence rate ratio 1.56 (95% confidence interval 1.05-2.40) per 0.1-point increase in FI).

Consequences

Using a standard method, frailty among clinical trial participants was quantifiable, and generally showed the expected associations with age, sex, disease severity and serious adverse event rates. The prevalence of frailty varied between trials and between index conditions, but all trials included some frail participants. Clinical trials therefore appear to be an underused resource for determining treatment effects in people with frailty. Understanding treatment effects in people with frailty is essential to inform decision making and maximise wellness across the life course.