Trends in obesity among patients registered to SWOG cancer clinical treatment trials.

Abstract

126 Background: Rising obesity rates have been documented in the United States population since the 1980s. Several studies have shown links between obesity and the incidence of – and the outcomes of – specific cancer types. Few studies have examined the prevalence of obesity among cancer patients at the time of diagnosis or among those who participate in clinical trials. We examined the prevalence of obesity over a 35-year period among patients enrolled in clinical treatment trials conducted by the SWOG Cancer Research Network. Methods: We analyzed body mass index (BMI) at registration among patients enrolled in phase II or III clinical trials conducted by SWOG. Adult patients (age≥18) participating in trials in obesity-related cancers between 1985 and 2020 were included. Obesity was classified as BMI ≥ 30 kg/m2. Multivariable logistic regression was used to examine time trends, adjusting for age, sex, race, and treatment type. Time was defined as year of enrollment and examined in separate models as continuous years and as 5-year intervals. We examined trends among patient subgroups by sex and by treatment type (chemotherapy, immunotherapy, and targeted therapy). Results: 16,374 patients enrolled in SWOG clinical trials conducted between 1985 and 2020 were analyzed. Patients were most commonly enrolled in trials for breast cancer (n = 6,327, 38.6%), colon cancer (n = 2,408, 14.7%), multiple myeloma (n = 2,112, 12.9%), and rectal cancer (n = 1,785, 10.9%). Overall, 32% of patients were obese (n = 5,252). Unadjusted obesity rates among trial participants increased from 18% in 1985-1990 to 42% in 2016-2020, compared to an increase from 23% to 42% among US adults over a similar period. Obesity prevalence rates increased over time for all sex, age, and race subgroups. Overall, there was an increasing linear trend in obesity (OR = 1.36 for each 5-year increase, 95% CI: 1.33-1.39, p < .0001). These trends persisted with adjustments for age, sex, race, and treatment type (OR = 1.35 per 5-year increase, 95% CI: 1.31-1.39, p < .0001). Positive linear trends in obesity prevalence were observed for all subgroups by sex and treatment type. There was no evidence of difference in trends between males and females or between the three treatment types. Conclusions: We observed an increasing trend in obesity among patients participating in clinical trials for obesity-related cancers, mirroring US adult obesity rates. Our findings indicate good representation of obese patients in clinical treatment trials, which has favorable implications for the applicability of trial findings to this group of patients.