Background: Cardiovascular disease (CVD) now ranks as a leading cause of non-cancer mortality among cancer patients (pts). However, most published studies have reported on hematological malignancies (HMs) as a group. Because treatment regimens vary widely across HMs, it is important to assess the risk of CVD-related mortality for individual HMs. In this study we assessed the risk of CVD-related mortality across eight specific HMs in US cancer pts by leveraging the NIH Surveillance, Epidemiology, and End Results (SEER8) Program.

Methods: We collected clinical data from the SEER8 Program registry for cancer pts between 1975-2022 diagnosed with a primary HM between the ages of 15-84 years (yrs). Using ICD-10 codes, we classified HMs into eight sites: myeloma; Hodgkin’s lymphoma (HL); non-Hodgkin's lymphoma (NHL); acute lymphoid leukemia (ALL); acute myeloid leukemia (AML); chronic lymphocytic leukemia (CLL); chronic myeloid leukemia (CML); and myelodysplastic syndrome, myeloproliferative neoplasm, and chronic myelomonocytic leukemia (MDS/MPN/CMML). We classified causes of death into three categories: CVD-related, cancer, and all others. We calculated CVD-related standardized mortality ratios (SMR) of HMs and age groups relative to the matched general US population. For each HM, we modeled temporal proportion trends for different causes of death and competing mortality risk models for 5-, 10-, and 20-yr survivors. We generated age-stratified cause-specific Kaplan-Meier (K-M) survival curves for three demographic cohorts – adolescent and young adults (AYA: 15-39 yrs), middle-aged adults (40-64 yrs), and elderly adults (65-84 yrs). Hazard ratios (HRs) were also calculated to compare CVD-related mortality risk between HM subgroups. Analyses were conducted in R 4.2.2 & SAS 9.4.

Results: In a cohort of 217,894 HM pts, with a total of 2,440,386 person-yrs of follow-up (mean: 7.07 person-yrs), we observed 28,139 CVD-related deaths. The relative proportion of CVD-related mortalities increased steadily over time, while cancer mortalities decreased. In all malignancies except HL, the proportion of CVD-related deaths (PCVD) matched or exceeded cancer-related mortality (PCAN) on average at 28 yrs post-diagnosis [range: 12.5 – 40]. PCVD reached 60% at 40 yrs post-diagnosis for myeloma, CLL, and MDS/MPN/CMML. Across all HMs over a period of five decades, CVD-specific mortality accounted for 50% of all deaths at 38.3 yrs and 19.3 yrs post-diagnosis in the middle-aged and elderly cohorts respectively, while remaining below 20% in the AYA cohort. Age-specific K-M curves showed a steady rise in CVD-specific mortality among elderly patients exceeding cancer-specific mortality in myeloma and MDS/MPN/CMML at 32.5 and 20 yrs post-diagnosis respectively. In contrast, cancer was the leading cause of death among AYA pts, with the exception of MDS/MPN/CMML, where CVD-specific mortality surpassed cancer mortality at 40 yrs post-diagnosis. SMR for CVD-related death was significantly higher for all HMs except ALL and CLL – myeloma (SMR = 1.65 [95% CI: 1.60-1.70]), HL (SMR = 1.98 [95% CI: 1.89-2.07]), NHL (SMR = 1.10 [95% CI: 1.08-1.12]), AML (SMR = 2.17 [95% CI: 2.04-2.30]), CML (SMR = 1.86 [95% CI: 1.73-1.99]), and MDS/MPN/CMML (SMR = 1.76 [95% CI: 1.70-1.82]). CVD-related mortality risk for 20-yr HM survivors in AYA (HR = 0.132 [95% CI: 0.116-0.150], p<0.0001) and middle-aged (HR = 0.378 [95% CI: 0.342-0.419], p<0.0001) cohorts was particularly low compared to 20-yr survivors in the elderly cohort.

Conclusion: CVD-related mortality poses a growing burden for HM pts, especially beyond 20 yrs post-diagnosis. Risk of CVD-related death often exceeds cancer mortality, particularly in elderly and MDS/MPN/CMML pts. For AYA pts, CVD-related mortality is seen in long-term survivors. These results underscore the need for age- and disease-specific survivorship strategies and suggest that CVD-related mortality prevention must become a central component of long-term care for HM pts.

This content is only available as a PDF.
2025
Sign in via your Institution