Clonal hematopoiesis of indeterminate potential (CHIP) is a common risk factor for hematologic malignancies and cardiovascular diseases. This study aimed to investigate the association between CHIP-related mutations and symptomatic heart failure (HF) in patients diagnosed with acute myeloid leukemia (AML). A total of 563 patients with newly diagnosed AML who underwent DNA sequencing of bone marrow before treatment were retrospectively investigated. Cox proportional hazard regression models and Fine and Gray’s subdistribution hazard regression models were used to assess the association between CHIP-related mutations and symptomatic HF. A total of 79.0% patients had at least 1 CHIP-related mutation; the most frequent mutations were DNMT3A, ASXL1, and TET2. A total of 51 patients (9.1%) developed symptomatic HF. The incidence of symptomatic HF was more frequent in patients with DNMT3A mutations (p <0.01), with a 1-year cumulative incidence of symptomatic HF in patients with DNMT3A mutations of 11.4%, compared with 3.9% in patients with wild-type DNMT3A (p <0.01). After adjustment for age and anthracyclines dose, DNMT3A mutations remained independently correlated with HF (hazard ratio 2.32, 95% confidence interval 1.26 to 4.29, p = 0.01). In conclusion, in patients with AML, the presence of DNMT3A mutations was associated with a twofold increased risk for symptomatic HF, irrespective of age and anthracyclines use.
Patients with acute leukemia have a higher prevalence of symptomatic heart failure (HF) than patients with other cancers. , Shared risk factors between acute leukemia and cardiovascular (CV) diseases may contribute to the high incidence of symptomatic HF in this specific cohort. Clonal hematopoiesis of indeterminate potential (CHIP) is a common age-associated phenomenon characterized by any clonal outgrowth of hematopoietic cells in patients without evidence of hematologic malignancy, dysplasia, or cytopenia. The genetic variants associated with CHIP (e.g., DNMT3A, TET2, and ASXL1) are known driver mutations for hematologic malignancies. In population-based studies, CHIP mutations were associated with an increased risk of a first episode of hospitalized HF, independently of traditional CV risk factors. , In patients with chronic HF, the presence of mutated TET2 or DNMT3A is correlated with the progression and poor prognosis of HF. , In addition, in patients with acute myeloid leukemia (AML), mutations in CHIP-related genes are associated with a higher risk for a composite outcome of all CV events. However, the association of symptomatic HF in patients with AML and CHIP-related mutations has not been widely explored. Therefore, this study aimed to characterize the relation of CHIP-related mutations with the occurrence of symptomatic HF in a large contemporary population of patients with AML.
Methods
All consecutive newly diagnosed AML adult patients (aged ≥18 years) with next-generation sequencing of DNA of bone marrow treated at the Penn Medicine system between January 2012 and November 2019 were included. Patients with a history of HF or without available follow-up record after the first hospital admission were excluded. The patients were monitored based on a recommended protocol that included quarterly visits throughout the first to third years and biannual visits thereafter, ensuring that the events of interest could be successfully captured. Preexisting CV risk factors and CV diseases were extracted electronically from the Epic electronic medical record (Epic systems Corporation, Verona, Wisconsin) using the relevant International Classification of Diseases Ninth and/or Tenth Revision Clinical Modification codes. The confirmation of all the preexisting diseases or risk factors required the presence of diagnostic codes and either relevant pharmacologic therapies or objective findings or signs, including laboratory evidence supporting the diagnosis. Anthracyclines dose was calculated based on the following doxorubicin hematologic toxicity equivalence: daunorubicin, 1.0; idarubicin, 5.0; and mitoxantrone, 4.0.
The primary outcome was new-onset symptomatic HF. To identify the occurrence of HF, each chart was reviewed individually. Symptomatic HF was identified as defined by the Standardized Data Collection for Cardiovascular Trials Initiative and the US Food and Drug Administration. Briefly, HF was diagnosed when 3 or more of the following 4 criteria were met: (1) symptoms of HF; (2) clinical signs of HF; (3) diagnostic testing results consistent with the diagnosis of HF (B-type natriuretic peptide or N-terminal pro–B-type natriuretic peptide, Kerley B-lines or pulmonary edema, pleural effusion, decreased left ventricular ejection fraction), and (4) initiation of new treatment for HF (pharmacologic therapies, e.g., diuretic agents and/or mechanical support). A single or more symptoms of HF and 2 or more signs on physical examination were necessary for the diagnosis. Symptoms of HF were defined as dyspnea at rest or during exercise, decreased exercise capacity, and symptoms of volume overload. Clinical signs on physical examination were defined as peripheral edema, ascites in the absence of hepatic disease, pulmonary crackles or rales, increased jugular venous pressure, S3 gallop, and significant and rapid weight gain related to fluid retention. Symptomatic HF was adjudicated by 2 independent cardiologists (YK and BL). Disagreement between the 2 readers was adjudicated by a third cardiologist (MSC). HF concomitant with sepsis was not considered a primary outcome.
If patients had no encounters over the preceding 12 months before the end of the study (March 2021) and no date of death was found in the chart, published obituaries were searched. Incomplete follow-up was defined as no encounters within 12 months of the end of the study and no notice of death.
Genomic DNA was extracted from bone marrow samples before initiation of the leukemia treatment and fragmented using sonication. PennSeq Hematological Malignancies Sequencing Panel (PennSeq Heme) was used to analyze 116 genes ( Supplemental Appendix 1 ). Variants are reported according to Human Genome Variation Society nomenclature and classified into the following reported categories: disease-associated variants and variants of uncertain significance. The variant allele frequency (VAF) was defined as the proportion of variant reads detected at a locus.
All oncogenic mutations with VAFs >5% in the following genes were considered as CHIP-related mutations: DNMT3A, TET2, ASXL1, TP53, JAK2, SRSF2, and SF3B1. Only disease-associated variants were considered as CHIP-related mutations; variants of uncertain significance were excluded. Mutations in these genes are the most common variants in patients with CHIP and known as driver mutations for hematologic malignancies.
Values were expressed as mean ± SD or counts (percentages). Follow-up time was presented as median and interquartile range (interquartile range [IQR] [twenty-fifth to seventy-fifth percentile]). Differences between patients with HF or non-HF or between patients with or without oncogenic mutations were determined using one-way analysis of variance or the Kruskal–Wallis test. Categorical variables were compared using the chi-square test. The primary analysis was the association of CHIP-related mutations with the occurrence of HF. Time-on-study was used as time scale in all survival analyses. Follow-up began at the diagnosis of AML and ended in case of death, event of interest, or the date of last encounter, whichever came first. The cumulative incidence function was used to estimate the incidence of HF, with non-HF-related death as a competing event, and the Fine and Gray’s method was applied to determine whether there were significant differences between groups. Univariable Cox proportional hazard regression models were used to analyze the association between baseline variables and HF. We developed a multivariable Cox proportional hazard model (model 1) through the inclusion of variables with a p <0.10 in univariable analysis and according to their hypothesized clinical relevance while avoiding collinearity. Because of competing risks, univariable and multivariable Fine and Gray’s subdistribution hazard regression models were also used to determine the association between clinical parameters and HF. The results of Cox proportional hazard models and Fine–Gray subdistribution hazard models are presented as hazard ratios (HRs) with 95% confidence interval (CI).
To understand the interaction between individual CV risk factors (hypertension, diabetes, obesity, and hypercholesteremia) or preexisting CV diseases (atrial fibrillation/flutter, coronary artery disease, cerebrovascular disease, chronic kidney disease, and peripheral artery disease), DNMT3A mutation and subsequent HF, we created separate models to evaluate the differential risk of HF in the following patients subgroups: (1) no CV risk factors or previous CV diseases and no DNMT3A mutation (reference group), (2) CV risk factors or previous CV diseases and no DNMT3A mutation, (3) no CV risk factors or previous CV diseases and DNMT3A mutation, and (4) CV risk factors or previous CV diseases and DNMT3A mutation. The cumulative incidence of HF for each category was calculated as described previously and multivariable Fine–Gray subdistribution hazard models were used to adjust for potential confounders.
Data were analyzed by SPSS version 16.0 (SPSS, Chicago, Illinois) and R version ii386 3.5.0 (Vienna, Austria). A p <0.05 was considered statistically significant.
Results
A total of 582 consecutive patients with AML and next-generation sequencing of DNA of bone marrow were identified. After patients with incomplete follow-up or history of HF were excluded (n = 19), 563 patients (298, 52.9% men, age range of the cohort: 19 to 92 years, IQR 54 to 71 years) were included in the study cohort (median follow-up period 445 days, range 1 to 5,023 days, IQR 150 to 1,061 days).
The baseline clinical characteristics of the patients are listed in Table 1 . A total of 51 patients (9.1%) developed symptomatic HF, including 46 patients (90.2%) with HF with reduced ejection fraction and 5 (9.8%) with HF with preserved ejection fraction. The median time to HF was 121 days (range 1 to 1,596 days, IQR 42 to 371 days). A total of 341 patients (60.6%) died of non-HF–related causes. The median time to death was 261 days (range 1 to 3,118 days, IQR 73 to 504 days). Patients who developed HF had a higher prevalence of hypertension and hypercholesterolemia ( Table 1 ).
| All (n=563) | HF (n=51) | Non-HF (n=512) | P Value | |
|---|---|---|---|---|
| Demographics | ||||
| Age, years | 62±14 | 62±14 | 62±14 | 0.96 |
| Age>60y, n(%) | 350 (64.8) | 30 (58.8) | 320 (62.5) | 0.65 |
| Male sex, n(%) | 298 (52.9) | 24 (47.1) | 274 (53.5) | 0.38 |
| BMI, kg/m 2 | 28.2±6.6 | 26.8±5.4 | 28.3±6.7 | 0.07 |
| Duration of follow-up,d,[Q1,Q3] | 445,[148,1066] | 448,[239,1019] | 442,[144,1068] | 0.24 |
| Cardiovascular risks and disease history | ||||
| Hypercholesterolemia, n(%) | 139 (24.7) | 20 (39.2) | 119 (23.2) | 0.01 |
| Hypertension, n(%) | 169 (30.0) | 23 (45.1) | 146 (28.5) | 0.01 |
| Diabetes, n(%) | 58 (10.3) | 6 (11.8) | 52 (10.2) | 0.43 |
| Obesity, n(%) | 160 (28.4) | 13 (25.5) | 147 (28.7) | 0.75 |
| Current/previous smoker, n(%) | 215 (38.2) | 23 (451.) | 192 (37.5) | 0.53 |
| ≥2 CV risk factors * , n(%) | 148 (26.3) | 19 (37.3) | 129 (25.2) | 0.07 |
| Coronary heart disease, n(%) | 66 (11.7) | 10 (19.6) | 56 (10.4) | 0.06 |
| Atrial fibrillation/flutter, n(%) | 37 (6.6) | 2 (3.9) | 35 (6.8) | 0.33 |
| Peripheral artery disease, n(%) | 32 (5.7) | 3 (5.9) | 29 (5.7) | 0.57 |
| Chronic kidney disease, n(%) | 18 (3.2) | 2 (3.9) | 16 (3.1) | 0.50 |
| Cerebrovascular disease, n(%) | 36 (6.4) | 4 (7.8) | 32 (6.3) | 0.42 |
| DVT/pulmonary embolism, n(%) | 29 (5.2) | 1 (2.0) | 28 (5.5) | 0.24 |
| Pre-existing CV diseases † , n (%) | 130 (23.1) | 14 (27.5) | 116 (22.7) | 0.27 |
| Cardiovascular medicine | ||||
| Beta-blockers, n(%) | 60 (10.7) | 12 (23.5) | 48 (9.4) | <0.01 |
| ACEI, n(%) | 60 (10.7) | 7 (13.7) | 53 (10.4) | 0.29 |
| ARBs, n(%) | 53 (9.4) | 4 (7.8) | 49 (9.6) | 0.46 |
| Aspirin, n(%) | 138 (24.5) | 11 (21.6) | 127 (24.8) | 0.38 |
| Anticoagulants, n(%) | 53 (9.4) | 2 (3.9) | 51 (10.0) | 0.12 |
| Statins, n(%) | 132 (23.4) | 13 (25.5) | 119 (23.2) | 0.42 |
| Cancer related treatment | ||||
| Anthracycline, n(%) | 481 (85.4) | 47 (92.2) | 434 (84.8) | 0.11 |
| Anthracycline dose, mg/m 2 , [Q1,Q3] | 180,[177,300] | 255,[180,330] | 180,[173,300] | 0.20 |
| Anthracycline dose >250mg/m 2 , n(%) | 164 (29.1) | 19 (37.3) | 145 (28.3) | 0.06 |
⁎ CV risk factors include obesity, hypertension, hypercholesterolemia, and diabetes.
† CV diseases include coronary artery disease, atrial fibrillation/flutter, cerebrovascular disease, and peripheral artery disease; p value: compared between patients with and with no HF.
A total of 445 patients (79.0%) had at least 1 CHIP-related mutation and 197 patients (35.0%) had more than 1 CHIP-related mutations. Patients with CHIP-related mutations were older (64 ± 13 vs 55 ± 16 years, p <0.01),\ and had a higher cumulative incidence of death over the total duration of follow-up (70.1% vs 54.5%, p <0.01) ( Supplemental Table 1 ). The cumulative incidence of death in patients with CHIP-related mutations was higher, with a 1-year cumulative incidence of 41.7% compared with 24.2% in patients with no CHIP-related mutations (p <0.01). There was no difference in the cumulative incidence of HF between patients with and with no CHIP-related mutations (p=0.31) or between patients with 1 and multiple CHIP-related mutations (p = 0.47). There was no difference in the cumulative incidence of death between patients with 1 or with multiple mutations (p = 0.09). DNMT3A (36.1%), ASXL1 (24.0%), and TET2 (22.2%) were the most common mutated genes ( Table 2 ). Patients with DNMT3A mutations were older, more likely to be treated with anthracyclines, and more likely to receive higher dose of anthracyclines than patients without the DNMT3A mutation ( Table 3 ). The median VAF and first to third quartiles of DNMT3A, TET2, and ASXL1 were 42.1% [26.3% to 45.8%], 44.6% [30.3% to 48.3%], and 40.6% [25.2% to 45.4%], respectively.
| All (n=563) | HF (n=51) | Non-HF (n=512) | P Value | |
|---|---|---|---|---|
| DNMT3A, n(%) | 203 (36.1) | 28 (54.9) | 175 (34.2) | <0.01 * |
| TET2, n(%) | 125 (22.2) | 10 (19.6) | 115 (22.5) | 0.73 |
| ASXL1, n(%) | 135 (24.0) | 11 (21.6) | 124 (24.2) | 0.73 |
| JAK2, n(%) | 37 (6.6) | 3 (5.9) | 34 (6.6) | 1.00 |
| TP53, n(%) | 116 (20.6) | 10 (19.6) | 106 (20.7) | 1.00 |
| SF3B1, n(%) | 22 (9.7) | 1(2.0) | 21(4.1) | 0.71 |
| SRSF2, n(%) | 57(10.1) | 3(5.9) | 54(10.5) | 0.46 |
| Presence of CHIP-related mutation, n(%) | 445(79.0) | 45(88.2) | 400(78.1) | 0.11 |
| Presence of ≥2 CHIP-related mutations, n(%) | 197(35.0) | 18(35.3) | 179(35.0) | 1.00 |
| Non-DNMT3A (n=360) | DNMT3A (n=203) | P Value | |
|---|---|---|---|
| Demographics | |||
| Age (y) | 61±15 | 65±13 | <0.01 |
| Age >60y, n(%) | 213 (59.2) | 137 (67.5) | 0.06 |
| Male sex, n(%) | 209 (58.1) | 89 (43.4) | <0.01 |
| BMI (kg/m2) | 28.7±6.9 | 27.2±5.8 | 0.01 |
| Duration of follow-up, d,[Q1,Q3] | 423,[144,1099] | 483,[171,980] | 0.97 |
| Cardiovascular risks and disease history | |||
| Hypercholesterolemia, n(%) | 90 (25.0) | 49 (24.1) | 0.45 |
| Hypertension, n(%) | 118 (32.8) | 51 (25.1) | 0.04 |
| Diabetes, n(%) | 43 (11.9) | 15 (7.4) | 0.06 |
| Obesity, n(%) | 109 (30.3) | 51 (25.1) | 0.11 |
| Current/previous smoker, n(%) | 133 (36.9) | 82 (40.4) | 0.36 |
| ≥2 CV risk factors * , n(%) | 103 (28.6) | 45 (22.2) | 0.06 |
| Coronary heart disease, n(%) | 42 (11.7) | 24 (11.8) | 0.53 |
| Atrial fibrillation/flutter, n(%) | 31 (8.6) | 6 (3.0) | 0.01 |
| Peripheral artery disease, n(%) | 23 (6.4) | 9 (4.4) | 0.22 |
| Chronic kidney disease, n(%) | 14 (3.9) | 4 (2.0) | 0.16 |
| Cerebrovascular disease, n(%) | 25 (6.9) | 11 (5.4) | 0.30 |
| DVT/pulmonary embolism, n(%) | 20 (5.6) | 9 (4.4) | 0.36 |
| Pre-existing CV diseases † , n(%) | 89 (24.7) | 41 (20.2) | 0.13 |
| Chemotherapy | |||
| With Anthracycline, n(%) | 279 (77.5) | 202 (99.5) | <0.01 |
| Anthracycline dose, mg/m 2 ,[Q1,Q3] | 180,[134,276] | 194,[180,340] | <0.01 |
| Anthracycline >250mg/m 2 , n(%) | 103 (28.6) | 62 (30.5) | 0.05 |
| Outcome | |||
| Heart failure, n(%) | 23 (6.4) | 28 (13.8) | <0.01 |
| Time to heart failure, d,[Q1,Q3] | 414,[104,917] | 401,[141,1083] | 0.40 |
| Death, n(%) | 247 (68.6) | 130 (64.0) | 0.16 |
| Time to death, d,[Q1,Q3] | 271,[74,484] | 277,[83,546] | 0.64 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
