# paper.pdf ## Metadata - PDFFormatVersion=1.7 - Language=en - IsLinearized=true - IsAcroFormPresent=false - IsXFAPresent=false - IsCollectionPresent=false - IsSignaturesPresent=false - CreationDate=D:20260519015958+05'30' - Subject=BLOOD ADVANCES, 10 (2026) 3746-3755. doi:10.1182/bloodadvances.2025019152 - Author=Orly Leiva - Creator=Elsevier - Producer=PDFlib+PDI 10.0.2p2 (C++/Win64) - Custom.EPSprocessor=PStill version 1.84.42 - Custom.ElsevierWebPDFSpecifications=7.0.1 - Custom.robots=noindex - Custom.doi=10.1182/bloodadvances.2025019152 - ModDate=D:20260519020807+05'30' - Title=Low serum cholesterol is a biomarker of advanced disease and poor outcomes in myeloproliferative neoplasms - dc:format=application/pdf - dc:identifier=doi:10.1182/bloodadvances.2025019152 - dc:title=Low serum cholesterol is a biomarker of advanced disease and poor outcomes in myeloproliferative neoplasms - dc:creator=Orly Leiva,Steven Soo,Olivia C. Liu,Victor You,Andrew Palmer,Justin Kahla,Yasmeen Murtaza,Olatoyosi Odenike,Anand Patel,Jeanne DeCara,Patrycja M. Dubielecka,Max Petersen,Michelle Hyunju Lee,Joan How,Gabriela Hobbs - dc:description=BLOOD ADVANCES, 10 (2026) 3746-3755. doi:10.1182/bloodadvances.2025019152 - dc:publisher=American Society of Hematology - prism:aggregationtype=journal - prism:publicationname=BLOOD ADVANCES - prism:copyright=© 2026 American Society of Hematology. Published by Elsevier Inc. Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0), permitting only noncommercial, nonderivative use with attribution. All other rights reserved. - prism:issn=2473-9529 - prism:volume=10 - prism:number=10 - prism:coverdisplaydate=26 May 2026 - prism:coverdate=2026-05-26 - prism:pagerange=3746-3755 - prism:startingpage=3746 - prism:endingpage=3755 - prism:doi=10.1182/bloodadvances.2025019152 - prism:url=https://doi.org/10.1182/bloodadvances.2025019152 - jav:journal_article_version=VoR - pdfx:elsevierwebpdfspecifications=7.0.1 - pdfx:doi=10.1182/bloodadvances.2025019152 - pdfx:robots=noindex - pdfx:epsprocessor=PStill version 1.84.42 - xap:creatortool=Elsevier - xap:createdate=2026-05-19T01:59:58+05:30 - xap:modifydate=2026-05-19T02:08:07+05:30 - xap:metadatadate=2026-05-19T02:08:07+05:30 - xaprights:marked=True - tdm:reservation=1 - tdm:policy=https://www.elsevier.com/tdm/tdmrep-policy.json - pdf:producer=PDFlib+PDI 10.0.2p2 (C++/Win64) - xapmm:documentid=uuid:01281606-c980-4dd0-ba9c-e39b0f1aa3b3 - xapmm:instanceid=uuid:779b3588-c3b2-425f-9309-743cb6bf014d - ali:license_ref=https://creativecommons.org/licenses/by-nc-nd/4.0/legalcode ## Contents ### Page 1 ![Figure]() Low serum cholesterol is a biomarker of advanced disease andpoor outcomes in myeloproliferative neoplasms Orly Leiva,1 Steven Soo,2 Olivia C. Liu,2 Victor You,3 Andrew Palmer,4 Justin Kahla,4 Yasmeen Murtaza,4 Olatoyosi Odenike,5Anand Patel,5 Jeanne DeCara,4 Patrycja M. Dubielecka,6 Max Petersen,6 Michelle Hyunju Lee,3 Joan How,7 and Gabriela Hobbs3 1 Gill Heart & Vascular Institute, University of Kentucky College of Medicine, Lexington, KY;2 Division of Cardiology, Department of Medicine, New York University GrossmanSchool of Medicine, New York, NY;3 Division of Hematology & Oncology, Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston,MA;4 Department of Medicine, Section of Cardiology and5 Department of Medicine, Section of Hematology & Oncology, University of Chicago, Chicago, IL;6 Division ofHematology/Oncology, Rhode Island Hospital and Warren Alpert Medical School of Brown University, Providence, RI; and7 Division of Hematology & Oncology, Departmentof Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA ![Figure]() Myeloproliferative neoplasms (MPNs), including essential thrombocythemia (ET),polycythemia vera (PV), and myelofibrosis (MF), are a diverse group of hematopoietic stemcell neoplasms associated with an increased risk of hematologic progression to secondaryMF or leukemia and cardiovascular disease (CVD). Elevated total cholesterol levels aretraditional CVD risk factors. Patients with MPNs may have reduced cholesterol levels,potentially reflecting an underlying hypermetabolic state. The impact of serum totalcholesterol on clinical outcomes in patients with MPNs has not yet been evaluated. Weperformed a multicenter retrospective cohort study of patients with MPNs who hadtransthoracic echocardiogram (TTE) with available serum lipid levels at time of TTE not onstatin therapy. Patients were categorized by total cholesterol level (≥150, 120-149, and<120 mg/dL). Outcomes evaluated were MPN disease progression to secondary MF or acuteleukemia, major adverse cardiovascular events, and all-cause mortality. Multivariablecompeting-risk regression modeling was performed. A total of 305 patients were includedof whom 54.4% had total cholesterol ≥150 mg/dL, 21.3% had 120 to 149 mg/dL, and 24.3%had <120 mg/dL. After a median follow-up of 50.8 months, total cholesterol <120 mg/dL wasassociated with increased risk of MPN disease progression (adjusted subhazard ratio, 3.00;95% confidence interval [CI], 1.36-6.63), heart failure hospitalization (aSHR, 3.76; 95% CI,1.77-7.96), and all-cause death (adjusted hazard ratio, 1.87; 95% CI, 1.06-3.30) vs ≥150 mg/dL. Among patients with MPN, low total cholesterol levels were associated with increasedrisk of MPN disease progression and death especially among patients with ET or PV. Introduction Philadelphia chromosome-negative myeloproliferative neoplasms (MPNs), including essential thrombocythemia (ET), polycythemia vera (PV), and myelofibrosis (MF), are a heterogenous group ofhematologic neoplasms associated with an increased risk of cardiovascular disease (CVD) in additionto progression to secondary MF or acute leukemia.1-3 Elevated total cholesterol level is a traditional risk Submitted 19 November 2025; accepted 3 February 2026; prepublished online onBlood Advances First Edition 4 March 2026; final version published online 26 May2026. Deidentified data are available from the corresponding authors, Orly Leiva (;) and Gabriela Hobbs (), on reasonablerequest. 3746 The full-text version of this article contains a data supplement. © 2026 American Society of Hematology. Published by Elsevier Inc. Licensed under, permitting only noncommercial, nonderivative use with attribution. All other rights reserved. 26 MAY 2026 • VOLUME 10, NUMBER 10 ### Page 2 factor for CVD, including myocardial infarction and stroke.4,5However, evidence suggests that patients with certain malignancies, including MPNs, may have reduced serum total cholesterol levels, which are potentially reflective of an underlyinghypermetabolic state or malnutrition.6-9 In 1 case-control study,patients with PV had on average lower serum total cholesterollevels compared with patients with secondary erythrocytosis.7 In asmall study of patients with MF, patients with low total cholesterolhad higher rates of CVD.10 However, despite the describedassociation between MPNs and low serum cholesterol levels, theimpact of low serum total cholesterol levels on MPN outcomes,including cardiovascular (CV) outcomes and MPN disease progression, has not been well-characterized. Therefore, we aimed toleverage a multicenter retrospective registry of patients with MPNsto investigate the role of total cholesterol level as a potentialbiomarker for adverse MPN and CV outcomes. Furthermore, weexplored associations between serum total cholesterol levels andclinical biomarkers of inflammation and hematologic proliferation. Methods Patient population We performed a retrospective analysis of a registry of patientsaged ≥18 years meeting World Health Organization 2016 criteriafor MPN (ET, PV, or MF), who had at least 1 transthoracic echocardiogram (TTE) after their MPN diagnosis from 1 January 2010to 1 January 2024 at 3 academic hospitals (MassachusettsGeneral Hospital, New York University Langone Health, and University of Chicago).11-14 Patients without serum lipid levels at timeof first TTE or were on statin therapy at the time of first TTE wereexcluded. Baseline characteristics at the time of first TTE includingage, comorbidities, MPN type, driver mutation (JAK2, CALR, MPL,or triple negative), presence of nonphenotypic driver mutations,and spleen size (as measured by largest cranio-caudal dimensionon cross-sectional imaging or ultrasound if performed within 3 months of TTE or MPN diagnosis) were captured. Laboratoryvalues, including white blood count (WBC), hemoglobin, plateletcount, lactate dehydrogenase (LDH), and creatinine werecaptured. Markers of inflammation including C-reactive protein andneutrophil/lymphocyte ratio (NLR), were also captured at time offirst TTE, if available. Outcomes Outcomes were identified via chart review. The primary hematologic outcome was progression of MPN disease defined as progression to secondary MF or acute leukemia. The primary CVoutcome was major adverse CV events (MACE), which was acomposite outcome of CV death, arterial thrombosis (includingischemic stroke, acute myocardial infarction, and peripheral orsplanchnic arterial thromboembolism), venous thromboembolism,and heart failure (HF) hospitalization. CV death was determined viachart review and discharge summaries and deaths were deemedto be CV deaths if the primary cause of death was due to MACE orarrhythmias. Secondary outcomes were all-cause death and individual components of MACE and hematologic progression. Statistical analysis Patients were categorized by total cholesterol levels (≥150, 120149, and <120 mg/dL). Categorical variables were compared between groups using χ2 test and continuous variables werecompared using 1-way analysis of variance tests with post hocBonferroni test. Multivariable competing-risk regression modelingwas performed to estimate risk of hematologic progression andMACE, and Cox proportional hazards regression was performed toestimate risk of all-cause death with patients with total cholesterollevels of ≥150 mg/dL as reference group. Hematologic progression models were adjusted for age, sex, race, MPN type, canonicaldriver mutation (JAK2, MPL, CALR), nonphenotypic driver mutation, spleen size, and WBC. MACE, HF hospitalization, arterialthrombosis, and all-cause death models were adjusted for thesame variables in addition to prior atherosclerotic CV disease,prior HF hospitalization, atrial fibrillation, anticoagulation use,chronic kidney disease, and MPN treatment at time of first TTE.,, Given the differences in clinical outcomes and MPN disease trajectory, we performed an analysis of patients with ET/PV and thosewith MF separately. Multivariable models were adjusted for thesame variables as primary analysis for the analysis of patients withET or PV. Given the smaller number of patients with MF, models forhematologic progression among patients with MF were adjustedfor age and mutation-enhanced international prognostic scoresystem (MIPSS70) score. Models for MACE and all-cause deathwere adjusted for age, MIPSS70, prior atherosclerotic CV disease,prior HF, atrial fibrillation, aspirin use, and anticoagulation use forpatients with MF. To explore an association between serum total cholesterol levelsand markers of inflammation we examined correlation betweentotal cholesterol levels and C-reactive protein and total cholesterollevels and NLR using Spearman rank correlation. The NLR is abiomarker that is associated with chronic inflammation andelevated NLR has been associated with development and worseoutcomes of CV disease as well as increased risk of thrombosisand poor outcomes in patients with MPNs.18-21 We also exploredan association between serum total cholesterol levels and markersof hematopoietic proliferation and MPN disease progressionincluding serum LDH levels, spleen size, and driver mutationvariant allele fraction. Statistical analyses were performed usingSPSS version 29.0 (IBM, Armonk, NY) and STATA version 15(StataCorp, College Station, TX). Study was approved by institutional review boards at Massachusetts General Hospital, University of Chicago, and New YorkUniversity and conducted in accordance with the Declaration ofHelsinki. Results Baseline patient characteristics of patients withMPNs by total cholesterol level A total of 305 patients were included with a median age of 69years at time of first TTE, of whom 163 (53.4%) were female, 252(82.6%) were White, 239 (78.4%) had mutation in JAK2, 118(38.7%) had PV, 100 (32.8%) had ET, and 87 (28.5%) had MF.The median time from MPN diagnosis to first TTE was 39.6 months(interquartile range [IQR], 12.4-93.2). The median serum totalcholesterol level was 156 mg/dL (IQR, 121-191). A total of 166(54.4%) patients had total cholesterol ≥150 mg/dL, 65 (21.3%)had 120 to 149 mg/dL, and 74 (24.3%) had <120 mg/dL. There 26 MAY 2026 • VOLUME 10, NUMBER 10 LOW CHOLESTEROL AND WITH POOR OUTCOMES IN MPNS 3747 ### Page 3 | All patientsN = 305 | >150 mg/dLN = 166 | 120-149 mg/dLN = 65<120 mg/dLN = 74 | P value | | - | - | - | - | | Age at TTE, median (IQR) | 69 (60-79) | 67 (58-75) | 72 (62-80) | 71 (64-81) | .13 | | Female sex, N (%) | 163 (53.4) | 106 (63.9) | 29 (44.6) | 28 (37.8) | <.001 | | Race, N (%) | .36 | | White | 252 (82.6) | 139 (83.7) | 49 (75.4) | 64 (86.5) | | Black | 32 (10.5) | 13 (7.8) | 12 (18.5) | 7 (9.5) | | Hispanic | 5 (1.6) | 4 (2.4) | 0 1 (1.4) | | Asian | 9 (3.0) | 6 (3.6) | 2 (3.1) | 1 (1.4) | | Other/unknown | 7 (2.3) | 4 (2.4) | 2 (3.1) | 1 (1.4) | | BMI, median (IQR) | 25.5 (22.6-28.9) | 25.9 (22.6-29.6) | 25.8 (24.1-29.8)25.2 (22.1-28.5) | .064 | | Time from MPN diagnosis to first TTE, medianmonths (IQR) | 39.6 (12.4-93.2) | 40.2 (12.1-91.6) | 37.4 (13.2-91.5)55.0 (18.1-138.7).020 | | Follow-up since first TTE, median months (IQR) | 50.8 (28.0-81.4) | 56.3 (38.1-82.3) | 43.2 (26.2-61.6)41.2 (16.7-74.4) | .012 | | MPN characteristics, N (%) | | MPN type at first TTE | <.001 | | PV | 118 (38.7) | 67 (40.4) | 26 (40.0) | 25 (33.8) | | ET | 100 (32.8) | 71 (42.8) | 18 (27.7) | 11 (14.9) | | MF | 87 (28.5) | 28 (16.9) | 21 (32.3) | 38 (51.4) | | Driver mutation | .59 | | JAK2 | 239 (78.4) | 131 (78.9) | 46 (70.8) | 62 (83.8) | | CALR | 39 (12.8) | 22 (13.3) | 11 (16.9) | 6 (8.1) | | MPL | 8 (2.6) | 3 (1.8) | 3 (4.6) | 2 (2.7) | | Triple negative | 19 (6.2) | 10 (6.0) | 5 (7.7) | 4 (5.4) | | MPN driver mutations VAF, median (IQR)N available | 46.0 (20.9-75.9)N = 122 | 39.5 (20.3-60.2)N = 70 | 42.9 (13.6-76.5)N = 2661.7 (43.9-88.2)N = 26 | .005 | | Nonphenotypic driver mutations, n (%) | | Any | 107 (35.1) | 49 (29.5) | 24 (36.9) | 34 (45.9) | .045 | | HMR | 50 (16.4) | 19 (11.4) | 11 (16.9) | 20 (27.0) | .011 | | Treatment for MPN | | Any | 225 (73.8) | 120 (72.3) | 49 (75.4) | 56 (75.7) | .81 | | Hydroxyurea | 129 (42.3) | 70 (42.2) | 32 (49.2) | 27 (36.5) | .32 | | Ruxolitinib | 41 (13.4) | 15 (9.0) | 9 (13.8) | 17 (23.0) | .014 | Spleen size, median cm (IQR) 13.5 (10.5-17.5) 11.4 (9.7-14.1) 13.5 (10.3-15.6) 18.9 (14.8-22.2) < .001 Comorbidities | Prior HF | 29 (9.5) | 11 (6.6) | 5 (7.7) | 13 (17.6) | .024 | | - | - | - | - | - | - | | Prior ASCVD | 76 (24.9) | 35 (21.1) | 20 (30.8) | 21 (28.4) | .23 | | Hypertension | 173 (56.7) | 90 (54.2) | 40 (61.5) | 43 (58.1) | .58 | | Atrial fibrillation | 85 (18.0) | 26 (15.7) | 11 (16.9) | 18 (24.3) | .26 | | CKD | 40 (13.1) | 17 (10.2) | 10 (15.4) | 13 (17.6) | .25 | | Prior VTE | 46 (15.1) | 16 (9.6) | 12 (18.5) | 18 (24.3) | .009 | | Diabetes | 30 (9.8) | 7 (4.2) | 11 (16.9) | 12 (16.2) | .002 | | Current or former smoking | 125 (41.0) | 66 (39.8) | 25 (38.5) | 34 (45.9) | .60 | | Medications at time of TTE, N (%) | | Aspirin | 189 (62.0) | 116 (69.9) | 40 (61.5) | 33 (44.6) | <.001 | | P2Y12 inhibitor | 10 (3.3) | 3 (1.8) | 4 (6.2) | 3 (4.1) | .23 | ASCVD, atherosclerotic CVD; CALR, calreticulin; CKD, chronic kidney disease; HDL, high-density lipoprotein; HMR, high molecular risk; LDL, low-density lipoprotein; OSA, obstructivesleep apnea; VAF, variant allele fraction; VTE, venous thromboembolism. *P < .05 vs cholesterol >150. **P < .001 vs cholesterol >150. †Presence of either ASXL1, SRSF2, EZH2, LDH1/2, or U2AF1Q157. ‡Prior coronary artery disease, peripheral vascular disease, or ischemic stroke. 3748 LEIVA et al 26 MAY 2026 • VOLUME 10, NUMBER 10 ### Page 4 | All patientsN = 305 | >150 mg/dLN = 166 | 120-149 mg/dLN = 65 | <120 mg/dLN = 74 | P value | | - | - | - | - | - | | Anticoagulation | 66 (21.6) | 25 (15.1) | 14 (21.5) | 27 (36.5) | <.001 | | Antihypertensive | 183 (60.0) | 90 (54.2) | 47 (72.3) | 46 (62.2) | .038 | | Laboratory values at TTE, median (IQR) | | LDL, mg/dL | 85 (61-110) | 106 (90-122) | 73 (66-80) | 51 (38-55) | <.001 | | HDL, mg/dL | 44 (32-57) | 52 (42-67) | 37 (29-51) | 27 (21-37) | <.001 | | Triglycerides, mg/dL | 107 (77-148) | 109 (83-156) | 98 (69-131) | 103 (73-146) | .013 | | WBC, K/μL | 9.6 (6.8-13.8) | 8.5 (6.6-11.2) | 9.0 (6.3-13.6) | 13.8 (9.5-21.3) | <.001 | | Hemoglobin, g/dL | 12.5 (10.3-14.2) | 13.3 (11.6-14.6) | 12.4 (10.5-13.8) | 10.6 (9.4-13.3) | <.001 | | Platelet count, K/μL | 422 (247-634) | 450 (338-640) | 425 (255-711) | 350 (189-505) | .43 | | Creatinine, mg/dL | 1.0 (0.8-1.2) | 00.9 (0.7-1.1) | 1.0 (0.8-1.2) | 1.1 (0.9-1.5) | .76 | | NLR | 4.5 (2.6-8.3) | 3.6 (2.2-5.6) | 4.4 (3.0-8.3) | 7.3 (3.6-12.9) | <.001 | | LDH, U/L | 388 (242-602) | 320 (207-516) | 376 (248-658) | 460 (278-830) | .004 | ASCVD, atherosclerotic CVD; CALR, calreticulin; CKD, chronic kidney disease; HDL, high-density lipoprotein; HMR, high molecular risk; LDL, low-density lipoprotein; OSA, obstructivesleep apnea; VAF, variant allele fraction; VTE, venous thromboembolism. *P < .05 vs cholesterol >150. **P < .001 vs cholesterol >150. †Presence of either ASXL1, SRSF2, EZH2, LDH1/2, or U2AF1Q157. ‡Prior coronary artery disease, peripheral vascular disease, or ischemic stroke. was no difference in age between the groups. Patients with totalcholesterol 120 to 149 (44.6% vs 63.9%; P < .05) and <120 mg/dL (37.8% vs 63.9%, P < .001) were less likely to be femalecompared with patients with total cholesterol ≥150 mg/dL. Therewas stepwise increase in the proportion of patients that had MFacross groups (16.9% in total cholesterol ≥150 mg/dL group,32.3% in 120-149 mg/dL group, and 51.4% in <120 mg/dLgroup; P < .001). Patients with total cholesterol <120 also hadhigher rates of prior HF (17.6% vs 6.6%) and diabetes (16.2% vs4.2%) compared with those with total cholesterol ≥150 mg/dL.There was a stepwise increase in proportion of high molecular riskmutations (P = .011) across groups, nonphenotypic mutationsamong groups described in supplemental Table 1. There was astepwise increase in median WBC (P < .001), NLR (P < .001),and LDH (P = .004) and a stepwise decrease in median hemoglobin (P < .001) across groups (). Outcomes of patients with MPNs by total cholesterollevel After a median follow-up of 50.8 months (IQR, 28.0-81.4), 42(13.8%) had MPN disease progression to secondary MF or acuteleukemia, 112 (36.7%) had MACE, 102 (33.4%) all-cause death,and 51 (16.7%) had HF hospitalization. There was a stepwiseincrease in incidence of MPN progression (9.0%, 15.4%, 23.0%in patients with total cholesterol of ≥150, 120-149, and <120 mg/dL, respectively; P = .014), MACE (28.3%, 40.0%, 52.7% inpatients with total cholesterol of ≥150, 120-149, and <120 mg/dL, respectively; P = .001), and all-cause death (9.0%, 15.4%,23.0% in patients with total cholesterol of ≥150, 120-149, and<120 mg/dL, respectively; P < .001;). After multivariable competing-risk regression, total cholesterol<120 mg/dL group (adjusted subdistribution hazard ratio [SHR],3.00; 95% confidence interval [95% CI], 1.36-6.63) but not 120to 149 mg/dL group (adjusted SHR, 1.93; 95% CI, 0.82-4.55)was associated with increased risk of MPN disease progression to secondary MF or acute leukemia compared with total cholesterol - ≥150 mg/dL group. There was no difference in risk of MACEamong patients with total cholesterol <120 mg/dL (adjusted SHR,1.47; 95% CI, 0.87-2.49) and those with total cholesterol 120 to149 mg/dL (adjusted SHR, 1.48; 95% CI, 0.83-2.62). However,patients with total cholesterol <120 mg/dL (adjusted SHR, 3.76;95% CI, 1.77-7.96) and 120 to 149 mg/dL (adjusted SHR, 3.05;95% CI, 1.29-7.19) had higher risk of HF hospitalization. Therewas also an association between total cholesterol levels <120 mg/dL (adjusted hazard ratio [HR], 1.87; 95% CI, 1.06-03.30) and120 to 149 mg/dL (adjusted HR, 2.02; 95% CI, 1.13-3.63)groups and all-cause death compared with total cholesterol - ≥150 mg/dL group (). Kaplan-Meier cumulative incidenceand survival curves of outcomes are depicted in. Effect of total cholesterol levels on outcomes ofpatients with MPNs stratified by subtype of MPN A total of 218 patients had ET or PV in our cohort. Median age was70 years (IQR, 60-79), 125 (57.3%) were female, 185 (84.9%)were White, 118 (54.1%) had PV, 138 (63.3%) had totalcholesterol levels ≥150 mg/dL, 44 (20.2%) had 120 to 149mg/dL, and 36 (16.5%) had <120 mg/dL. There were morepatients with total cholesterol <120 mg/dL who had PV (69.4% vs48.6%, P < .05) compared with patients with total cholesterol ≥150 mg/dL. Patients with total cholesterol <120 mg/dL also hadhigher median spleen size (16.0 vs 11.3 cm, P < .001) comparedwith those with total cholesterol ≥150 mg/dL. Similarly, patientswith total cholesterol <120 mg/dL had higher WBC (median, 13.8vs 8.3 × 103/μL; P < .001) and NLR (median 8.0 vs 3.5, P < .05)compared with patients with total cholesterol ≥150 mg/dL,supplemental Table 2. After multivariable competing-risk regression, patients with total cholesterol <120 mg/dL had increasedrisk of MPN disease progression (adjusted SHR, 3.89; 95% CI,1.47-10.34), HF hospitalization (adjusted SHR, 6.26; 95%CI, 2.47-15.82), and all-cause death (adjusted HR, 2.15; 95% CI, 26 MAY 2026 • VOLUME 10, NUMBER 10 LOW CHOLESTEROL AND WITH POOR OUTCOMES IN MPNS 3749 ### Page 5 <120 mg/dL Unadjusted SHR or HR (95% CI) Adjusted SHR or HR (95% CI) >150 mg/dL 120-149 mg/dL <120 mg/dL >150 mg/dL 120-149 mg/dL <120 mg/dL N = 74 P value 120-149 mg/dLN = 65 >150 mg/dLN = 166 All patientsN = 305 MPN disease progression 42 (13.8) 15 (9.0) 10 (15.4) 17 (23.0) .014 Ref 1.94 (0.89-4.23) 2.81 (1.39-5.66) Ref 1.93 (0.82-4.55) 3.00 (1.36-6.63) MACE 112 (36.7) 47 (28.3) 26 (40.0) 39 (52.7) .001 Ref 1.66 (1.03-2.69) 2.21 (1.44-3.39) Ref 1.48 (0.83-2.62) 1.47 (0.87-2.49) HF hospitalization 51 (16.7) 17 (10.2) 13 (20.0) 21 (28.4) .002 Ref 2.36 (1.17-4.78) 3.14 (1.66-5.97) Ref 3.05 (1.29-7.19) 3.76 (1.77-7.96) Arterial thrombosis 53 (17.4) 25 (15.1) 16 (24.6) 12 (16.2) .22 Ref 1.78 (0.95-3.33) 1.08 (0.54-2.15) Ref 1.38 (0.60-3.18) 00.88 (0.36-2.13) All-cause death 102 (33.4) 40 (24.1) 23 (35.4) 39 (52.7) <.001 Ref 2.03 (1.22-3.40) 2.71 (1.74-4.22) Ref 2.02 (1.13-3.63) 1.87 (1.06-3.30) ASCVD, atherosclerotic CVD; CKD, chronic kidney disease. *MPN disease progression adjusted for age, sex, MPN type, MPN driver mutation, nonphenotypic driver mutations, WBC and spleen size; MACE, HF hospitalization, arterial thrombosis, all-cause death adjusted for same variables inaddition to prior ASCVD, prior heart failure, atrial fibrillation, aspirin use, anticoagulation, MPN treatment, and CKD. †Composite of CV death, myocardial infarction, stroke, peripheral arterial thrombosis, VTE, or HF hospitalization. 1.07-4.34) compared with patients with total cholesterol ≥150 mg/dL. Patients with total cholesterol 120 to 149 mg/dLalso had increased risk of HF (adjusted SHR, 3.17; 95% CI, 1.138.89;A). A total of 87 patients had MF, of whom 28 (32.2%) had totalcholesterol ≥150 mg/dL, 21 (24.1%) had 120 to 149 mg/dL, and38 (43.7%) had <120 mg/dL. The median age was 69 years (IQR,62-78), 38 (43.7%) were female, and 67 (77.0%) were White.There was no difference in MIPSS70 risk category or MF type(primary, after PV, or after ET). Patients with total cholesterol<120 mg/dL had larger median spleen size compared withpatients with total cholesterol ≥150 mg/dL (median 21.0 vs14.8 cm, P < .001; supplemental Tables 3 and 4). After multivariable adjustment, there was no difference in rates of leukemiaprogression, MACE, HF hospitalization, or all-cause death amongpatients with total cholesterol <120 or 120 to 149 mg/dLcompared with those with total cholesterol ≥150 mg/dL(B). Multivariable regression modeling shown insupplemental Table 5. Relationships between serum total cholesterol levelsand markers of inflammation and diseaseprogression There was a modest correlation between total cholesterol level andWBC among all patients (r = −0.28, P < .001) and those with ETor PV (r = −0.30, P < .001) but not MF (r = −0.19, P = .072).Similarly, there was a modest correlation between total cholesterollevel and NLR among all patients (r = −0.30, P < .001) and thosewith ET or PV (r = −0.31, P < .001) but not MF (r = −0.16,P = .14). There was also a modest correlation between totalcholesterol levels and LDH levels (r = −0.24, P < .001), but thiswas not seen after stratifying by MPN type. There was a moderatecorrelation between total cholesterol and spleen size among allpatients (r = −0.52, P < .001), ET or PV patients (r = −0.36, P <.001), and MF patients (r = −0.50, P < .001). There was a modestcorrelation between driver mutation variant allele fraction and totalcholesterol among all patients (r = −0.27, P = .003) but notamong patients ET or PV (r = −0.19, P = .072) or MF (r = −0.07,P = .72; and). Discussion In our multicenter cohort study of patients with MPN, we demonstrated an association between lower total cholesterol andincreased risk of MPN disease progression, HF hospitalization,and all-cause death especially among patients with ET or PV. Ourstudy also demonstrated an association between decreasingserum total cholesterol levels and elevated markers of inflammationand MPN disease progression, especially NLR and spleen size. An interesting observation from our study is the associationbetween low serum total cholesterol and hematologic progressionand MACE. This suggests that total cholesterol among patientsnot on statin therapy with MPN may be a biomarker for higher riskdisease. Cholesterol metabolism is a complex process that isaltered in MPNs and other chronic inflammatory conditions. Insepsis, serum cholesterol levels are decreased throughout thecourse of illness and lower cholesterol levels have been associated with worse outcomes.22,23 In our study, WBC and NLR were 3750 LEIVA et al 26 MAY 2026 • VOLUME 10, NUMBER 10 ### Page 6 negatively correlated with serum total cholesterol. Additionally,patients with MF had lower serum total cholesterol levelscompared with ET or PV. It is well established that patients withMPNs have elevated circulating proinflammatory cytokines.24Therefore, an increased inflammatory milieu likely plays a role inreducing serum total cholesterol levels in patients with MPN, andserum total cholesterol levels may be useful biomarkers inassessing degree of inflammation. Additionally, recent studiessuggest that chronic inflammation is important in the progressionof MPNs to secondary MF, which may explain the associationbetween hypocholesterolemia and increased risk of MPN progression in our cohort.25,26 Our study also showed a strong negative correlation between totalcholesterol and spleen size. Increased spleen size is a marker ofpotential disease progression in patients with MPN, in particularfor patients with PV and ET, who do not often have splenomegaly.The association between total cholesterol and spleen size may suggest that increased cellular turnover and proliferation, andtherefore increased turnover of cell membranes, play a role in totalcholesterol levels in MPNs.27 This hypothesis is supported by priorliterature that demonstrated an association between decreasinglipid levels and lymphoid malignancy in the years leading up tocancer diagnosis.27-29 Mendelian randomization studies did notfind a causal link between low serum cholesterol levels andmalignancy, though it is likely that low serum cholesterol is abiomarker of progression rather than the cause of progression.30Indeed, in 1 study that included patients with and without lipidlowering medications, the association between lower serumcholesterol level and increased risk of lymphoid malignancy wasseen more strongly among patients not on lipid-lowering medications.27 Cholesterol homeostasis reflects a balance betweencholesterol synthesis, driven by β-hydroxy β-methylglutaryl coenzyme A (HMG-CoA) reductase activity, and cellular efflux, mediated by transporters such as ATP-binding cassette transporter(ABCA) A1 and ABCG1.31 In MPNs, inflammatory cytokines and Time from first TTE (months) ### Page 7 clonal hematopoiesis may disrupt this equilibrium, leading toabnormal efflux, intracellular retention and/or overconsumption, alllowering circulating cholesterol and signaling aggressive diseasebiology. This metabolic dysregulation underscores cholesterol as both a biomarker and a potential mechanistic axis in MPN pathophysiology. Therefore, low serum cholesterol levels are most likelybiomarkers of disease progression but not causal. Additionally,prior studies have suggested statin use is associated with 1 ### Page 8 improved outcomes in patients with MPN.32-35 Given theincreased risk of CVD among patients with MPN and prior studiessuggesting benefit of statin use, clinicians should continue toconsider statin initiation per guidelines and more data is needed toevaluate the impact of statins on MPN outcomes.32-36 An increase in cellular turnover and spleen size may also explainthe observed higher risk of HF hospitalization among patients withlower total cholesterol levels in our cohort. Patients with MPNs,particularly MF, are at risk of high-output HF due to increased cellturnover and splenomegaly.37 In our prior study of patients withMPN, spleen size was associated with increased risk of HF.13,38However, our study did not find an association between totalcholesterol levels and outcomes among patients with MF. This islikely due to the small number of patients and events. Therefore,further studies are needed to better understand the interplaybetween cholesterol metabolism and MPN pathology. Our findings may suggest that hypocholesterolemia correlates with hypercatabolic states, systemic inflammation, and reduced survivalindependent of other risk factors. There are several limitations to consider when interpreting ourresults including its retrospective nature. Our cohort of patientsonly included patients with MPN who underwent TTE after MPNdiagnosis, thus likely introducing selection and indication biasesand enriching our patient population with a higher risk of CV disease compared to the general MPN population. Patients withstatins were excluded from our study to reduce their confoundingeffect on total cholesterol levels and outcomes, however furtherstudies are needed to evaluate the role of statins in MPN outcomes. Additionally, too few patients had C-reactive protein valuesavailable and serum cytokine levels were unavailable, thus furtherstudies are needed to determine the correlation between inflammatory cytokines and cholesterol levels in patients with MPN. Our 100 ### Page 9 study included few patients with MF, thus further study with alarger population of MF is needed. Conclusions Lower serum total cholesterol levels were associated withincreased risk of MPN disease progression, HF hospitalization,and all-cause death and were correlated with markers of inflammation and MPN disease activity and cellular proliferation/turnover.These findings suggest that low cholesterol levels may be a usefulmarker in identifying patients with MPN at risk of adverse outcomes. Further studies are needed to better understand theinterplay between cholesterol metabolism and MPN diseasebiology to improve outcomes and prognostication. Authorship Contribution: O.L., G.H. contributed to the conceptualization, datacuration, investigation, and methodology; O.L. analyzed the dataand performed statistical analyses; S.S., O.C.L., V.Y., A.P., J.K., Y.M. contributed to data curation and investigation; O.O., A. Patel,J.D., P.M.D., M.P., M.H.L., and J.H. contributed to the methodologyand provided critical review and edits of the manuscript; O.L.drafted the first manuscript draft; and all authors reviewed andapproved of the final manuscript draft. Conflict of interest disclosure: The authors declare nocompeting financial interests. ORCID profiles: O.C.L.,; V.Y.,; A. Palmer,; J.K.,; A. Patel,;P.M.D.,; M.H.L.,;J.H.,; G.H.,. Correspondence: Orly Leiva, Gill Heart & Vascular Institute,University of Kentucky College of Medicine, 800 Rose St, Lexington, KY 40508; email:;; andGabriela Hobbs, Division of Hematology and Oncology, Massachusetts General Hospital and Harvard Medical School, 55 FruitSt, Boston, MA 02114; email:. References - 1.. - 2.. - 3.. - 4.. - 5.. - 6.. - 7.. - 8.. - 9.. - 10.. - 11.. - 12.. - 13.. - 14.. - 15.. - 16.. 3754 LEIVA et al 26 MAY 2026 • VOLUME 10, NUMBER 10 ### Page 10 - 17.. - 18.. - 19.. - 20.. - 21.. - 22.. - 23.. - 24.. - 25.. - 26.. - 27.. - 28.. - 29.. - 30.. - 31.. - 32.. - 33.. - 34.. - 35.. - 36.. - 37.. - 38.. 26 MAY 2026 • VOLUME 10, NUMBER 10 LOW CHOLESTEROL AND WITH POOR OUTCOMES IN MPNS 3755