Cardiovascular disease risk scores as predictors of cardiovascular disease and heart failure in allogeneic hematopoietic cell transplant survivors

Abstract

Abstract Survivors of allogeneic hematopoietic cell transplant (allo-HCT) have a significantly higher risk of developing atherosclerotic cardiovascular disease (ASCVD) and a lower age of onset compared to the general population (PMID: 27590105). Risk calculators are used to estimate risk and aid clinical decision-making for primary prevention of CVD, but their utility in HCT recipients is unclear. The Pooled Cohort Equation (PCE) calculates the 10-year risk of ASCVD for individuals 40-79 years using age, sex, race, cholesterol, blood pressure, diabetes, smoking, and hypertension treatment status. The Predicting Risk of CVD EVENTs (PREVENT) Equations improve upon the PCE with inclusion of glomerular filtration rate, lipid-lowering medication status, body mass index, and hemoglobin A1C. The PREVENT equations provide 10-year risk estimates of ASCVD and heart failure (HF) for individuals 30-79 years. This retrospective study identified patients without pre-existing ASCVD or HF who underwent allo-HCT for any indication between 10/2016-7/2023 and survived at least 1 year post HCT without relapse. Risk estimates were calculated at 1 year post HCT. Patients were grouped into low (<5%), borderline (5 to <7.5%), intermediate (7.5% to <20%), and high (≥20%) risk groups according to American Heart Association guidelines. Primary outcome was time to first ASCVD or HF event from 1-year post-HCT. Univariate and multivariate Cox proportional hazards regression was used to evaluate PCE and PREVENT scores as predictors of cardiac events alone and in combination with patient, disease and transplant characteristics. Due to patients with parameters outside of criteria of risk scores (age, BMI, or cholesterol values), sensitivity analyses were conducted with and without inclusion of these patients. 316 patients who met inclusion criteria were identified, and of those, 250 patients had complete data for risk score calculation. Median age was 61 (range: 22 – 78) and 49% were female. The most frequent indication for transplant was AML (N=108, 43%) and most patients (N=132, 53%) received reduced intensity conditioning. Most patients (N = 157, 63%) received grafts from matched unrelated donors (MUD). 121 patients (48%) received PTCy-based GVHD prophylaxis. Median follow up time was 30.3 months (IQR: 16.1, 49.6). Of 250 patients, 187 (75%) had data within validated parameters of PCE and 191 (77%) for PREVENT. According to PCE, 65 (35%) of 187 patients were low risk, 26 (14%) borderline, 58 (31%) intermediate, and 38 (20%) high. For PREVENT ASCVD risk, 103 (54%) of 191 were low risk, 41 (22%) borderline, 47 (25%) intermediate, and none were high. For PREVENT HF risk, 106 (56%) patients were low risk, 33 (17%) borderline, 50 (26%) intermediate, and 2 (1%) high risk. The median follow-up was 30.3 months. A total of 13 ASCVD events occurred during follow-up. The estimated event rate for ASCVD was 0.8%. (95% CI 0%, 2%) at 1 year, and 1.18% (95% CI 4.2%, 18.8%) at 5 years. PCE risk groups were not significantly associated with ASCVD. In multivariate analysis, intermediate PREVENT risk group was associated with higher risk of ASCVD than low risk (p=0.029), but borderline was not. In multivariate analysis, patients with a MUD donor were at significantly lower risk of ASCVD (HR = 0.22, 95% CI 0.05, 0.88, p = 0.032). Sensitivity analysis including patients who were outside parameters of risk scores for age, BMI or cholesterol did not impact the predictive value of PCE or PREVENT. A total of 15 HF events occurred during follow-up. The estimated event rate for HF at 1 year was 1.7% (95% CI 0, 34%) and 11.4% (95%CI 4.3%, 18%) at 5 years. In multivariate analysis, the intermediate risk was associated with significantly higher risk of HF compared to the low-risk group (HR = 6.4, 95%CI 1.22, 33.69). Sensitivity analysis demonstrated a difference with inclusion of patients with parameters outside of PREVENT parameters for HF, with the high-risk group associated with increased risk of HF (p = 0.001) but not the intermediate group (p = 0.0999). The results of this study suggest that the PREVENT risk calculator may be a useful tool to predict ASCVD and HF risk in an allo-HCT population. The significance of donor type is unclear but could potentially be related to donor age and clonal hematopoiesis, and further analysis is ongoing. Given limitations of sample size and length of follow up, additional study is needed to validate these models in HCT recipients.