CARDIAC DISEASE PREDICTION FOLLOWING H ODGKIN LYMPHOMA: AN EORTC LYMPHOMA GROUP AND GELA FOLLOW‐UP STUDY

Abstract

Introduction: Hodgkin lymphoma (HL) survivors are known to suffer from an excess risk of cardiac disease (CD) due to radiation and anthracyclines. However, CD is also associated with other risk factors. The purpose of the study was to evaluate the impact of treatment and patient characteristics on the risk of subsequent CD in the EORTC (European Organisation for Research and Treatment of Cancer)-GELA (Groupe d'Étude des Lymphomes de l'Adulte, now LYSA) cohort of HL patients treated in 9 randomized trials from 1964 to 2004 (n = 6658). Methods: Incidence of CD was reported during follow-up and through a patient-reported questionnaire (LSQ), mailed in 2009 to 2010 (56.7% of all patients alive with known address). Comorbidities were registered at LSQ completion. A multivariate Cox proportional hazards regression model on first CD, stratified by trial, was fitted on treatment variables (mean radiation dose to the heart and cumulative doses of anthracyclines and vinca-alkaloids) as well as patient characteristics at treatment start (age, gender, country, gender∗age, smoking, family history of CD, BMI). Model reduction was performed with the backward selection procedure at a 5% threshold level. Estimated hazard ratios (HRs) were presented with their 95% confidence intervals (CI) and tested at the 5% two-sided significance level. Data from H1 to H8 trials were used to build the model; the model validation was performed based on H9 trial data using measures of discrimination (C-index). Results: A total of 1919 patients responded to the LSQ, 49% were males and the median age at treatment start was 29 years (range, 10-69). A total of 85.9% had early-stage disease, and 1293 patients received both radiotherapy and anthracyclines. The median duration of follow-up was 14 years (range, 5-44). A total of 416 patients (21.7%) reported CD events. The model built included mean radiation dose to the heart (HR = 1.02 per Gy increase; 95% CI, 1.00-1.03; P = .019), cumulative dose of anthracyclines (HR = 1.17 per 100 mg/ m2 increase; 95% CI, 1.00-1.37; P = 0.05), age (HR = 1.04 per year increase; 95% CI, 1.03-1.05; P < .001), and BMI (HR = 1.05; 95% CI, 1.02-1.09; P < .001). The resulting linear predictor was LP = [0.017 × heart dose] + [0.002 × dose anthracyclines] + [0.037 × age] + [0.053 × BMI]. The C-index was moderate at 0.57 (95% CI, 0.39-0.74). There was a strong association betweenCD and post-treatment smoking and comorbidities such as high blood pressure, high cholesterol levels, diabetes, and elevated BMI. Conclusions: Following treatment for HL, the subsequent risk of CD is influenced by radiation dose to the heart, the cumulative dose of anthracycline but not vinca-alkaloids, and age and BMI at treatment start. However in this cohort of HL patients, the overall cardiac risk is only partly explained by treatment exposure. The poor model performance (low C-index) may be explained by important posttreatment events such as later occurrence of established cardiac risk factors.