Cause‐specific mortality following polycythemia vera, essential thrombocythemia, and primary myelofibrosis in the US population, 2001–2017

Contemporary, population-based, cause-specific mortality risks within the heterogeneous subtypes of classical myeloproliferative neoplasms (MPNs)–polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis (PMF)–in the United States are needed to inform patient care. Therefore, we leveraged MPN cases diagnosed in the surveillance, epidemiology, and end results (SEER) Program in the 21st century to quantify cause-specific mortality risks and associated burden of death separately for PV, ET, and PMF, overall and by age and sex. We included all cases of first primary PV, ET, and PMF diagnosed during 2001–2016 (follow-up through 2017) among adult (ages 20–84 years) residents of 17 SEER cancer registry areas. Relative risk of death was estimated using standardized mortality ratios (SMRs; observed/expected), calculated by MPN subtype and stratified by age at diagnosis and sex. Differences in SMRs between subgroups were tested for significance using Poisson regression modeling and likelihood methods. We also calculated excess absolute risks (EAR) per 10 000 person-years and estimated MPN subtype-specific cumulative mortality. See the Methods section in Supporting Information for further details describing the SEER Program, study population, disease groupings, and statistical methods. We identified 13 340 individuals with PV, 12 346 individuals with ET, and 3230 individuals with PMF (Table 1). Mean follow-up time was 6.5, 6.0, and 4.0 years after a diagnosis of PV, ET, and PMF, respectively. MPNs accounted for < 10% of cause-specific death following a diagnosis of PV and ET and 34% following PMF (Table S1). All-cause mortality for each MPN exceeded that in the similarly aged general population, with 10-year cumulative mortality reaching 18.3%, 12.5%, and 48.5% for PV, ET, and PMF, respectively, for individuals < 60 years (general population: 6.8%) and 46.7%, 44.7%, and 83.7%, respectively, for those ≥ 60 years (general population: 29.1%) (Figure S1, Table S2). For patients with PV or ET, cumulative mortality for noncancer deaths far exceeded that for cancer deaths throughout 10 years of follow-up. In contrast, for PMF, cumulative incidence of deaths from all noncancers, all hematologic neoplasms (excluding MPN), and MPN were generally similarly elevated, whereas deaths from all solid tumors remained low throughout the follow-up period. Patients with PV had a 90% increased risk of death from all causes (excluding MPN), when compared to that expected in the general population (SMR = 1.9, 95% CI 1.8–2.0, EAR = 172.1; Table 1). Nearly 75% of these deaths (n = 2279; SMR = 2.0, EAR = 131.1) were due to noncancer causes, with 1.8–2.7-fold significantly increased SMRs observed for infections, including septicemia and respiratory infections; digestive diseases, including liver disease; cerebrovascular disease; cardiovascular diseases, including heart disease; and respiratory diseases, including chronic obstructive pulmonary disease (COPD); and a 4.3-fold elevated risk for benign hematologic disorders. The greatest burden of noncancer excess deaths/10 000 person-years was due to heart disease (EAR = 42.7), with lesser contributions from COPD, cerebrovascular disease, and infections (EARs = 11–21). Most of the relative and excess risks of deaths from neoplasms were accounted for by MDS/AML (SMR = 12.9, EAR = 17.6). In age and sex-specific analyses for PV, relative risks for all causes (excluding MPN) of death were higher among younger (SMR<60 = 2.8) compared with older individuals (SMR≥60 = 1.7, p < .001), but older patients experienced a disproportionate number of excess deaths (EAR<60 = 126.3 vs. EAR≥60 = 221.8; Table S3). Significantly higher SMRs due to cerebrovascular disease, heart disease, and COPD occurred among those diagnosed at ages < 60 versus ≥ 60 years, and significantly higher SMRs from heart disease, COPD, adverse events, and MDS/AML were observed among females than males. Heart disease had the highest EAR of non-MPN deaths regardless of age (EAR<60 = 36.0; EAR≥60 = 50.1) and sex (EARmales = 41.9; EARfemales = 44.2). Significantly increased risks of non-MPN deaths also were seen following a diagnosis of ET (SMR = 1.8, 95% CI = 1.8–1.9, EAR = 149.1). SMRs for noncancer causes were generally similar to PV but smaller in magnitude for most outcomes, with SMRs ranging from 1.7 to 2.2 for infections, digestive diseases, cerebrovascular disease, heart disease, COPD, nephritis/nephrosis, and adverse events. However, notably higher risk for death from benign hematologic disorders was observed (SMRET = 12.5; pPV vs ET < .001) and lower risk from COPD (SMRET = 1.8; pPV vs ET < .001) for ET compared with PV. Deaths due to solid cancers were not elevated beyond expectation after ET (SMR = 1.0), although the SMR for MDS/AML was especially high (SMR = 24.8) and significantly different from PV patients (pPV vs ET < .001). Notably, MDS/AML (EAR = 32.2) and cardiovascular diseases (EAR = 32.0) accounted for a similar number of excess deaths. ET-specific analyses by age and sex revealed risk patterns similar to PV. Specifically, younger ET patients had higher SMRs from all causes (excluding MPN) than older patients (SMR<60 vs ≥60 = 2.3 vs 1.7, p < .001), but larger absolute risks were observed in the ≥ 60 age group (EAR<60 vs ≥60 = 74.8 vs. 223.1; Table S4); the magnitude of risks for specific causes generally paralleled those for PV. In noteworthy exceptions, heart disease SMRs were comparable for younger versus older patients (SMR<60 vs. ≥60 = 1.8 vs. 1.6, p > .50), while EARs were substantially higher for older patients (EAR<60 vs. ≥60 = 9.5 vs. 46.3). Males (SMR = 1.9) had slightly higher risk for all noncancer deaths than females (SMR = 1.7, p = .038). Individuals with PMF experienced 4.7-times higher risk of all-cause mortality (excluding MPN) (SMR = 4.7, 95% CI = 4.5–5.0, EAR = 768.0) compared to the general population, with noncancer causes accounting for 47% (n = 586) of deaths (SMR = 3.2, EAR = 313.1). SMRs for noncancer deaths from infections, benign hematologic disorders, digestive diseases, renal diseases, and adverse events were increased > 4-fold following PMF, while approximately 2-fold elevated SMRs were seen for endocrine disorders, cerebrovascular disease, heart diseases, and COPD. The greatest noncancer mortality EARs were from heart disease (EAR = 71.3), infections (EAR = 65.6), and benign hematologic disorders (EAR = 43.0). Exceptionally high SMRs for all hematopoietic neoplasms (excluding MPNs) were observed, with MDS/AML (SMR = 188.3) accounting for most of the excess deaths (EAR = 302.2). PMF-specific analyses stratified by age and sex demonstrated that SMRs for all causes (excluding MPN), all noncancers, all hematopoietic neoplasms (excluding MPN), and MDS/AML were significantly higher among PMF patients < 60 years versus ≥ 60 years of age (Table S5); however, the older age group sustained the greatest burden of deaths from each of these causes. Infections, heart disease, and benign hematologic disorders accounted for most noncancer excess deaths among both age groups. Relative risk of death from all causes (excluding MPN) were significantly increased among males (SMR = 4.7) and females (SMR = 4.8; p = > .50). The SMRs from heart disease, adverse events, and MDS/AML were significantly higher among females than males. With the exception of heart disease (EARmales = 62.7; EARfemales = 81.9) and adverse events (EARmales = 21.0; EARfemales = 30.4), the death burden from other specific causes of noncancer deaths, all solid neoplasms, lung and bronchus cancers, all hematopoietic neoplasms (excluding MPN), and MDS/AML predominated among males. Overall, we found that individuals in the United States with PV, ET, and PMF experienced significantly increased mortality risks from noncancer causes and hematologic neoplasms, with heart disease, cerebrovascular disease, COPD, benign hematologic disorders, and infections, accounting for the majority of excess noncancer deaths. PMF patients had a substantially higher mortality burden overall and for many specific causes of death than those with PV or ET. While age- and sex-specific patterns across MPNs are highlighted, all subgroups had higher noncancer- and MDS/AML-related mortality compared to the general population. We found significantly increased risks for heart disease following PV, ET, and PMF, with the latter experiencing the highest mortality burden. Overall risk of death from cerebrovascular disease was also increased after an MPN diagnosis, although with fewer excess deaths than observed for heart disease. Chronic inflammation and overproduction of inflammatory cytokines in the classical MPNs is thought to initiate and drive the development of premature atherosclerosis through JAK and other MPN-associated mutation pathways.1-3 We found significantly increased SMRs for respiratory diseases and COPD for all MPNs although patients with PV and PMF contributed disproportionately to these excess deaths. Our study augments findings from a Danish study that found increased risks of similar magnitude for respiratory mortality among all classical MPNs combined.4 While the mechanisms underlying increased respiratory mortality in MPNs are uncertain, activation of inflammatory pathways and therapies administered for MPN may contribute to these findings.3-5 Mortality from benign hematologic disorders was increased across all MPNs, with disproportionately increased SMRs and excess deaths following PMF. Bleeding related to cytopenias from cytoreductive therapy or marrow fibrosis, platelet dysfunction from anti-platelet therapy, and acquired coagulopathies (e.g., von Willebrand syndrome, therapeutic anticoagulation) could contribute to these deaths. Infection-related deaths were increased for all MPNs with the highest risks following PMF. Risk of hospitalization from infection was also identified among MPN patients in Sweden, irrespective of common cytoreductive treatments, with the highest risks of infections following PMF.6 Immune dysfunction and dysregulated chronic inflammation inherent to MPNs could play a role in the infection-related deaths we observed. Strengths of our study included the large number of individuals diagnosed with MPNs during a contemporary era encompassing driver mutation testing. Our study is limited by likely under-reporting of MPNs to cancer registries, misclassification, and lack of information on MPN treatment. Furthermore, we cannot exclude potential misclassification of cause of death based on limitations in coding of death certificates. In summary, we noted some distinct mortality patterns by MPN subtype, age, and sex. The increased risk of death following an MPN diagnosis during a contemporary era suggests that additional measures are needed to decrease mortality. The authors thank Jeremy Miller (Information Management Services, Inc., Rockville, MD) for computing support. This study was supported by the Intramural Research Program, National Cancer Institute, National Institutes of Health (NIH). The authors have no conflicts of interest to declare. Conception and design: all authors; collection and assembly of data: GMD; data analysis and interpretation: all authors; manuscript writing and revision for important intellectual content: all authors; final approval of manuscript: all authors; financial support: LMM. The opinions and information in this article are those of the authors, and do not represent the views and/or policies of the U.S. Food and Drug Administration. The data that support the findings of this study are available at http:www.seer.cancer.gov with Data Use Agreement. Appendix S1 Supporting Information Figure S1 Cumulative mortality among a simulated general U.S. population and 13 340, 12 346, and 3230 individuals diagnosed with polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis (PMF), respectively, at ages 20–84 years, 17 SEER registry areas, 2000–2016 (followed through 2017). All-cause mortality panels depict cumulative mortality from all causes of death in the general population and PV, ET, and PMF populations by age group (<60 vs. ≥60 years); cause-specific mortality panels depict cumulative mortality from all noncancers, all solid neoplasms, hematologic neoplasms (excluding myeloproliferative neoplasm [MPN]), and MPNs among patients diagnosed with PV, ET, and PMF by age group (<60 vs. ≥60 years). *Note difference in y-axes between all-cause mortality and cause-specific mortality panels. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.