Introduction
Polycythemia Vera (PV) is derived from a Latin origin and means “true increase in multiple blood cells”. It is a part of a greater category of pathologies called Myeloproliferative Neoplasms (MPN). It is characterized by an increase in the production of red blood cells, granulocytes, and platelets through stimulating multipotent progenitor cells [1]. This exclusively happens without any identifiable physiologic triggers. Triggers such as dehydration and diuretics could cause hemoconcentration and, in turn, relative erythrocytosis. Also, hypoxia, sleep apnea, high altitudes, renal disease, and certain drugs could be among the causes of absolute erythrocytosis [2]. On the other hand, PV is mainly associated with mutations in JAK2 V617F and JAK2 exon 12 to a lesser extent. These mutations are known to cause halting of the autoinhibitory mechanisms in the JAK-STAT signaling pathway, resulting in an uncontrolled hematopoiesis [3].
The incidence of PV was reported to be 1.57 cases per 100,000 person-years in the United States between 2002 and 2016 in a study using the Surveillance, Epidemiology, and End Results (SEER) database [4]. In a Cohort study between two large health plans in the USA between 2008 and 2010, prevalence was estimated to be 44-57 per 100,000 persons [5]. Although it has been challenging to obtain a true and consistent estimate of PV prevalence, recent improvements in diagnostic methods and improvements in survival have made it possible.
PV usually manifests by causing headaches, visual changes, thrombotic events, splenomegaly, and many other vascular complications as a result of blood hyperviscosity [6]. Erythromelalgia, which is redness and burning pain in the extremities, is often part of PV’s clinical findings. It’s thought to be due to ongoing erythrocytosis and thrombocytosis [7]. More importantly, PV can turn into more dreadful complications, like myelofibrosis or acute leukemia, especially in patients with previous exposure to cytotoxic agents or radiation [8].
PV is often an incidental finding when elevated hemoglobin/hematocrit is found in the lab work. Aquagenic pruritus is a hallmark finding in PV and can often be an eye-opener to help clinicians pursue diagnostic routes to confirm PV. Based on the revised WHO Classification of Hematolymphoid tumors in 2022, the three major criteria to diagnose PV are: elevated hemoglobin/hematocrit levels, accompanied by trilineage hyperplasia (panmyelosis), with pleomorphic mature megakaryocytes in the bone marrow, and JAK2 V617F or JAK2 exon 12 mutations. The minor criterion is subnormal serum erythropoietin levels, indicating negative feedback. To diagnose PV, you need to have the three aforementioned major criteria combined, or two major criteria along with the minor criterion [9].
The mainstay of PV treatment is frequent phlebotomies with a goal to reduce red blood cell mass and get hematocrit to 45%, which in turn reduces hyperviscosity and helps create an iron-deficient state to slow down the process of erythropoiesis. It is reported that patients older than 60 years old or with a history of thrombosis seem to have a higher risk of developing PV-related thrombosis compared to those who are younger than 60 and those with no history of thrombosis. Low-dose Aspirin has been shown to be effective in reducing thrombotic events in all risk groups, with some nuances related to the use of twice-daily Aspirin or anticoagulation in patients with a previous thrombosis history. Cytoreductive agents such as Hydroxyurea are also used to decrease the risk of thrombosis in high-risk groups, or pegylated interferon-\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$\alpha$\end{document}α in Hydroxyurea-resistant cases or cases where Hydroxyurea was not well tolerated [10].
The data on PV mortality trends and disparities remain scarce compared to how much we know about its clinical features. In this study, using the CDC WONDER database for adults between 1999 and 2023, we investigate mortality trends and disparities, aiming to understand the behavior of the disease and its trends based on certain parameters like age, sex, ethnicity, and geographical differences. Understanding this will hopefully allow us to observe certain biases and factors that might help us in early detection and overcoming disparities related to PV.
Results
Overall Trends
Throughout the study timeframe, from 1999 to 2023, a total of 24,236 deaths occurred among adults with polycythemia vera. The AAMR observed an overall notable decrease during the study period, from 1.15 in 1999 to 0.73 in 2023, with an AAPC of -1.97 (95% CI: -2.29 to -1.73; p < 0.000001).
From 1999 to 2014, AAMR showed a significant drop, from 1.15 to 0.64 (APC: -3.83; 95% Cl: -5.22 to -2.36; p = 0.020). This was followed by a minute yet non-significant increase to 0.68 in 2018 (APC: 0.92; 95% CI: -4.25 to 2.97; p = 0.880), a notable rise to 0.83 in 2021 (APC: 7.36; 95% Cl: 4.28 to 9.86; p = 0.001), and a significant decrease to 0.73 in 2023 (APC: -6.81; 95% Cl: -10.60 to -2.35; p = 0.001) (Supplemental Table 1, 2, 3) .
Gender Trends
From 1999 to 2023, women exhibited higher mortality than men (12,207 vs. 12,029 deaths). In 1999, the AAMR for men was 1.38 compared to 1 in women. These values decreased to 0.87 and 0.62 by the year 2023. Accordingly, men exhibited a less negative AAPC of -1.88 (95% Cl: -2.36 to -1.37; p < 0.000001), compared to women with an AAPC of -2.03 (95% Cl: -2.66 to -1.43; p < 0.000001).
AAMR among women saw a notable decline, from 1 in 1999 to 0.55 in 2018 (APC: -3.72; 95% Cl: -4.42 to -3.17; p = 0.020), followed by a non-significant rise to 0.72 in 2021 (APC: 13.42; 95% CI: -4.46 to 17.36; p = 0.120), and a non-significant decrease to 0.62 in 2023 (APC: -7.16; 95% CI: -14.71 to 5.47; p = 0.154). Among men, AAMR exhibited a significant decline from 1.38 in 1999 to 0.74 in 2015 (APC: -3.78; 95% Cl: -4.52 to -3.06; p = 0.016). This was followed by a non-significant rise to 0.99 in 2021 (APC: 5.07; 95% CI: -4.19 to 11.04; p = 0.097), and a non-significant decrease to 0.87 in 2023 (APC: -6.54; 95% CI: -12.63 to 2.72; p = 0.131) (Supplemental Table 1, 2, 3) .
Racial Trends
When stratified by race, NH White individuals exhibited a higher mortality as compared to NH Black/African American individuals (21,963 vs 1,072 deaths). Deaths among other predefined racial/ethnic groups (Hispanic, NH American Indian/Alaska Native, NH Asian/Pacific Islander) were omitted from trend analysis due to small numbers and data suppression.
Consequently, NH White individuals demonstrated a greater overall AAMR as compared to NH Black/African American individuals (0.93 vs 0.42). NH White individuals showed a greater overall decline in AAMR from 1.23 in 1999 to 0.84 in 2023 (AAPC: -1.69; 95% Cl: -2.15 to -1.37; p < 0.000001), as compared to NH Black/African American individuals, which showed a non-significant decrease from 0.57 in 1999 to 0.42 in 2023 (AAPC: -1.03; 95% Cl: -2.61 to 0.28; p = 0.128).
In brief, NH Black/African American individuals showed a significant decrease in AAMR, from 0.57 in 1999 to 0.29 in 2016 (APC: -3.46; 95% CI: -11.40 to -1.79; p = 0.020), followed by a non-significant increase, reaching 0.42 in 2023 (APC: 5.12; 95% Cl: -0.71 to 26.87; p = 0.098). NH White individuals saw a significant decrease from 1.23 in 1999 to 0.75 in 2016 (APC: -3.48; 95% Cl: -3.97 to - 3.07; p = 0.004), a significant rise to 0.98 in 2021 (APC: 6.85; 95% Cl: 4.63 to 12.75; p = 0.018), and a steady yet significant drop to 0.84 in 2023 (APC: -6.70; 95% Cl: -12.22 to -0.43; p = 0.035) (Supplemental Table 1, 2, 4) .
Age-specific Trends
Between 1999 and 2023, older adults experienced a higher mortality than middle-aged adults (20,960 vs 3,276 deaths). Overall, CMR was higher among older adults (1.96) as compared to middle-aged adults (0.17). CMR among older adults saw a steeper decline throughout the study timeframe, from 2.80 in 1999 to 1.58 in 2023 (AAPC: -2.39; 95% CI: -2.81 to -1.98; p < 0.000001), as compared to CMR among middle-aged adults, which fell from 0.19 in 1999 to 0.18 in 2023 (AAPC: -0.41; 95% Cl: -1.48 to 0.67; p = 0.457).
CMR among middle-aged adults observed a significant decrease from 0.19 in 1999 to 0.13 in 2012 (APC: -2.99; 95% CI: -9.48 to -0.90; p = 0.010), subsequently rising significantly to 0.18 in 2023 (APC: 2.72; 95% CI: 0.51 to 9.80; p = 0.019). CMR among older adults showed a notable decrease from 2.80 in 1999 to 1.51 in 2016 (APC: -3.87; 95% Cl: -4.51 to -3.25; p = 0.022), followed by a non-significant rise to 1.84 in 2021 (APC: 4.32; 95% Cl: -4.97 to 9.81; p = 0.129), and a non-significant drop to 1.58 in 2023 (APC: -5.91; 95% Cl: -11.24 to 1.68; p = 0.137) (Supplemental Table 2, 5) .
Regional Trends
Between 1999 and 2023, the Southern region reported the highest number of deaths (7,254), followed by the Midwest (6,283), the West (5,829), and the Northeast (4,870). Overall, AAMRs were highest in the Midwest (0.95), the West (0.93), the Northeast (0.83), and the South (0.69).
All regions demonstrated an overall decrease in AAMR between 1999 and 2023. In the Northeast, AAMR decreased from 1.18 to 0.71 (AAPC: -1.77; 95% CI: -2.53 to - 1.09; p < 0.000001), while the Midwest saw a decrease from 1.27 to 0.69 (AAPC: -2.66; 95% CI: -3.24 to -2.06; p < 0.000001). The South experienced a drop from 1.0 to 0.64 (AAPC: -1.83; 95% Cl: -2.30 to -1.26; p < 0.000001) over the same period. The West experienced a drop from 1.29 to 0.89 (AAPC: -1.30; 95% Cl: -1.91 to -0.67; p < 0.000001). AAMR in the Northeast exhibited a notable decline, from 1.18 in 1999 to 0.65 in 2014 (APC: -3.26; 95% Cl: -7.75 to -2.25; p = 0.007), followed by a non-significant rise to 0.71 in 2023 (APC: 0.76; 95% Cl: -1.45 to 8.90; p = 0.471). In the Midwest, AAMR declined significantly from 1.27 in 1999 to 0.71 in 2016 (APC: -3.49; 95% Cl: -4.22 to -2.87; p = 0.012), followed by a non-significant rise to 0.90 in 2021 (APC: 3.99; 95% CI: -3.40 to 10.45; p = 0.068), and a significant decrease to 0.69 in 2023 (APC: -11.23; 95% CI: -18.39 to - 0.06; p = 0.049).
In the Southern region, AAMR decreased significantly from 1.0 in 1999 to 0.48 in 2015 (APC: -4.58; 95% Cl: -5.82 to - 2.98; p = 0.032), followed by a non-significant rise, reaching 0.72 in 2021 (APC: 6.61; 95% CI: -7.18 to 12.86; p = 0.197), and a non-significant drop to 0.64 in 2023 (APC: -3.76; 95% Cl: -9.65 to 4.65; p = 0.471). The Western region saw a notable decrease in AAMR from 1.29 in 1999 to 0.72 in 2013 (APC: -4.04; 95% Cl: -6.06 to -2.84; p < 0.000001), subsequently rising significantly to 0.89 in 2023 (APC: 2.67; 95% Cl: 0.84 to 6.16; p = 0.007). (Supplemental Table 2, 6)
Place of Death
Between 1999 and 2023, place of death data were available for 24,228 cases. The majority of deaths occurred in medical facilities (39.36%), followed by decedents’ residences (34.12%), nursing homes/long-term care facilities (18.45%), and hospice facilities (3.88%). A small proportion of deaths (0.20%) occurred in unspecified locations, while the remaining proportion of deaths (4.01%) occurred in other locations (Supplemental Table 7).
Urbanization Trends
Urbanization data is available between 1999 and 2020; metropolitan areas accounted for a higher mortality (16,690 deaths) compared to non-metropolitan areas (4,154). However, the overall AAMR remained higher in non-metropolitan areas (0.91) relative to metropolitan areas (0.82). From 1999 to 2020, both metropolitan and non-metropolitan areas observed a drop in AAMR; 1.15 to 0.78 (AAPC: -2.29; 95% CI: -2.63 to -1.96; p < 0.000001), compared to non-metropolitan areas, which observed a decrease from 1.09 to 0.98 (AAPC: -1.09; 95% CI: -1.90 to -0.56; p < 0.000001).
Metropolitan regions saw a notable decrease in AAMR from 1.15 in 1999 to 0.60 in 2015 (APC: -4.01; 95% CI: -4.52 to -3.60; p < 0.000001), followed by a significant increase to 0.78 in 2020 (APC: 3.42; 95% CI: 1.30 to 7.24; p = 0.002). AAMR dropped significantly from 1.09 in 1999 to 0.76 in 2017 (APC: -2.84; 95% CI: -3.75 to -2.25; p = 0.001), followed by a notable increase to 0.98 in 2020 (APC: 10.09; 95% CI: 2.02 to 21.43; p = 0.015) (Supplemental Table 2, 8) .
Discussion
The study reveals that, nationwide, deaths attributable to polycythemia vera have steadily decreased from 1999 to 2023. Although the trend has variations across sex, race, geography, and urban-rural divisions, the absolute death counts were greater among women, and men had consistently increased AAMRs. The analyses of different races show that the non-Hispanic White patients experienced higher rates, whereas the non-Hispanic Black/African American patients exhibited smaller, less stable downward shifts in mortality. Regional disparities are also noted as the Midwest and Western states report elevated AAMRs, in contrast to persistently lower mortality rates from the Southern states, indicating the differences in healthcare accessibility and patient care. The non-metropolitan regions show higher AAMRs, consistent with the hypothesis that structural inequalities are responsible for mortality risk. Age stratification identifies older patients aged 65 years and above as the principal pool of mortality burden, thus reflecting a decline in organ system function, biological frailty, and the prevalence of multiple comorbidities. These results collectively highlight the necessity of interventions that concurrently address demographic, geographic, and systemic healthcare differences in order to control the death rates.
PV predominantly presents around the age of sixties, yet cases appear at nearly every age [19]. Between fifty and sixty, the incidence climbs sharply; the most pronounced acceleration begins at seventy-five. [20,21,22] Older populations bear heavier symptom burdens, heightened by the malignancy’s innate aggressiveness and the aggregation of other underlying diseases. Guideline-driven treatment presently stratifies vascular hazard along age and past thrombosis, labelling everyone beyond sixty as a high-risk cohort warranting cytoreductive intervention [23,21,24]. Age also connects to a growing failure or intolerance to hydroxyurea, complicating care and possibly escalating PV-driven mortality [25].
Throughout the period of 1999-2023, women had more deaths than men. However, men consistently experienced higher AAMRs. Both males and females had their mortality rates significantly reduced over time, but men followed with a less steep decrease compared to women. Moreover, recent reductions after 2018 were not significant. These observed disparities in mortality rates by sex may be partly explained by underlying biological and clinical differences in PV progression. Women develop PV less frequently, and at the time of diagnosis, the prevalence of the homozygous JAK2 mutation is modest (61% versus 80% in men) [26]. Long-term studies indicate that JAK2 allele burden diverges, a difference that may foster heightened rates of mitotic recombination and clonal evolution into homozygous expansions, which in turn correlate with virulent disease [27,26,28]. Additional evidence shows that men more frequently harbour adverse co-mutations, specifically ASXL1, SRSF2, U2AF1, EZH2, IDH1, and IDH2, while women bear comparatively few secondary lesions [29]. These genetic features translate into different clinical landscapes; men experience a higher burden of cardiovascular sequelae and show markedly elevated frequencies of myocardial infarction and peripheral arterial occlusion, whereas women tend to manifest a more indolent disease course [30]. The interplay of somatic, clinical, and sex-related factors ultimately elevates PV-specific mortality in the male cohort.
Our analysis also notes racial and ethnic differences in PV-related mortality, with trends generally highest in non-Hispanic White and non-Hispanic Black populations. These disparities may stem from complex socioeconomic, systemic, and healthcare access factors, as observed in other hematologic malignancies. For instance, studies in related conditions have highlighted substantially reduced access to stem cell transplantation among Black patients, even after adjusting for socioeconomic factors [31,32,33]. Such observations suggest the potential influence of structural impediments or disparate treatment pathways at the provider or institutional level. Similarly, delayed receipt of chemotherapy, suboptimal treatment adherence, and missed follow-up visits, as documented in broader oncological contexts, could contribute to diminished survival among minority populations [34,35,32,36]. While the exact impact of these factors on PV-related mortality specifically requires further research, their documented influence in other cancers suggests they warrant consideration. Additionally, the persistent underrepresentation of minority patients in clinical trials may limit the generalizability of treatment guidelines, potentially exacerbating disparities by failing to adequately address the needs of diverse populations [37,38,39].
Mortality rates remained elevated in thinly populated districts, with both metropolitan and rural certainties initially declining, then later regaining upward inflexions, thereby underscoring a persistent, geographic, therapeutically related disparity. Understanding how rurality intersects with insurance features, coverage limits, co-insurance, co-pays, and the delivery of cancer care still needs thorough examination [40]. Few previous studies show that rural areas tend to attract fewer commercial plans under the Medicare Advantage marketplace, which results in decreased overall coverage rates among older adults with cancer living in rural areas. Meanwhile, the persistent obstacles of distance to reach healthcare facilities from their homes and unreliable transportation continue to affect timely care. Although due to the COVID-19 pandemic, there was an adoption of telehealth strategies, telecolposcopy, remote genetic risk evaluations, and teledermatology. However, certain rural patients still face major hurdles, including limited broadband access, uncertainty about which modality preference, and a lack of knowledge and data on how different visit types affect clinical outcomes [41,42,43]. Even traditional travel to regional oncology clinics and the influence of variable health literacy have received far too little attention [44,45]. Equally consequential, lifestyle risk profiles, namely rural-urban gradients in smoking, obesity, physical inactivity, and dietary patterns, remain under-studied in the cancer literature [46]. Strategic structural interventions, combined with transplant teams and primary oncology teams in multidisciplinary venues, prove to be promising pathways for narrowing the persistent disparities in cancer outcomes between rural and urban patients [47].
During the study period, distinct regional patterns emerged, illustrating a broad, albeit oscillating, decline in PV-attributed AAMR. The Northeast, Midwest, and Southern tiers first reflected sizable drops, later followed by erratic ascents and final stable plateaus. The West recorded the steepest early descent, yet it later exhibited a pronounced upswing ending in 2023.
Limitations
This analysis has a number of important limitations. First, mortality data derived from death certificates have the risk of misclassification of polycythaemia vera-related deaths, partly due to inconsistent coding and variations in cause-of-death documentation, which may hide the true epidemiological burden. Second, the absence of clinical parameters such as therapy regimens, resistance profiles, comorbidities, and molecular characterisations prevents the clarification of present-day trends from the effects of underlying biological and treatment-related determinants. For instance, differences in resistance to hydroxyurea, the frequency of phlebotomy, and access to targeted agents such as ruxolitinib remain unevaluated. Third, suppressed data of races other than Non-Hispanic Whites and African Americans in the national mortality databases may lead to biased subgroup assessments. Similarly, comparisons made between different regions and urban-rural stratifications may fail to characterise heterogeneity at deeper geographic levels. Lastly, while the long-term secular trends observed are robust, the possibility of confounding by socioeconomic factors, healthcare accessibility, and lifestyle and behavioral factors persists. These shortcomings emphasise the critical need for parallel, registry-based, prospective investigations that collect detailed clinical, molecular, genetic, and sociodemographic data to interpret observed mortality trends accurately.