Abstract
Aims
The prognostic significance of D‐dimer in hospitalized heart failure (HF) patients is incompletely characterized. We aimed to assess the association of D‐dimer levels on admission with adverse events at follow‐up in patients hospitalized with HF across all ejection fraction (EF) phenotypes.
Methods and results
Consecutive patients hospitalized from December 2006 to December 2017 for HF with D‐dimer and EF values available (n = 1795) were enrolled. Associations between D‐dimer and all‐cause death were examined at 1‐year follow‐up. Median age was 57 years, 73.4% were male, and the majority (72.1%) were in New York Heart Association Classes III–IV. EF was reduced in 53.3% (HFrEF), mildly reduced in 16.3% (HFmrEF), and preserved in 30.4% (HFpEF). Median (interquartile range) D‐dimer on admission was 0.56 (0.27–1.295) μg/mL FEU (fibrinogen‐equivalent unit) in the whole cohort, 0.64 (0.28–1.48) μg/mL FEU in HFrEF, 0.50 (0.27–1.03) μg/mL FEU in HFmrEF, and 0.495 (0.25–1.10) μg/mL FEU in HFpEF (P = 0.001). At 1‐year follow‐up, higher D‐dimer (D‐dimer ≥0.56 μg/mL FEU) independently predicted all‐cause death in total cohort [hazard ratio (HR) 1.55; 95% confidence interval (CI), 1.15–2.1], in HFrEF (HR, 1.49; P = 0.039), and in HFpEF (HR, 2.06; P = 0.033). However, no relationship was found for HFrEF or HFmrEF when D‐dimer was treated as quartiles. In sensitivity analysis, quantitatively similar but more pronounced association between D‐dimer and all‐cause death was observed in total cohort and HFpEF cohort.
Conclusions
In hospitalized HF patients, higher D‐dimer concentration was a significant and independent predictor of 1‐year all‐cause mortality. Across all HF phenotypes, this effect was most evident in HFpEF patients.
Keywords: D‐dimer, Heart failure, Prognosis, HFrEF, HFmrEF, HFpEF
Introduction
Heart failure (HF) is a life‐threatening and highly prevalent disease. Despite establishment of novel therapeutic methods, long‐term prognosis associated with HF remains poor, 1 , 2 causing a major strain on healthcare systems.
HF is associated with an increased risk of thrombosis, leading to sudden death, cardio‐embolic stroke, systemic thromboembolism, and venous thromboembolism, which poses serious adverse effects on quality of life and prognosis of patients. 3 This is true of HF patients in sinus rhythm, and even more so in the many HF patients with concomitant atrial fibrillation. 4 , 5
D‐dimer, a specific degradation product of XIIIa‐crosslinked fibrin, is an early marker of in vivo coagulation activation and thrombogenesis. 6 Measurement of D‐dimer plasma concentration is commonly used in the diagnosis of pulmonary embolism, 7 deep vein thromboembolism, 8 acute coronary syndrome, 9 and acute aortic dissection. 10 With respect to HF, increased D‐dimer concentrations have been found in acute HF, and in outpatients 11 , 12 ; however, the prognostic significance of this finding has been scarcely investigated. In particular, unexplored is the possible prognostic role of D‐dimer measurements in HF patients with different phenotypes.
Given the heterogeneity of HF syndrome with respect to the degree of cardiac contractile impairment as manifest by left ventricular ejection fraction (LVEF)—the simplest and most widely used tool to ‘phenotype’ HF patients—it is relevant to investigate the prognostic value of D‐dimer across EF phenotypes within HF, spanning from HF with reduced LVEF (HFrEF), mildly reduced LVEF (HFmrEF), to HF with preserved LVEF (HFpEF). This differentiation of HF patients based on LVEF is important, as it may entail different characteristics with respect to aetiology, demographics, co‐morbidities, and response to therapy. 13 , 14 , 15 , 16 , 17 , 18 , 19
Accordingly, we performed a retrospective study aiming to assess the association of D‐dimer levels on admission with adverse events at 1‐year follow‐up in patients hospitalized with HF across all HF phenotypes.
Materials and methods
Study population
In this single‐centre study, data were prospectively collected and retrospectively analysed. We analysed a total of 1795 consecutive patients who were admitted to the Heart Failure Care Unit of Fuwai Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (CAMS&PUMC), with a clinical diagnosis of HF from December 2006 to December 2017 and in whom D‐dimer and LVEF values were available. Diagnosis of HF was based on the Framingham study 20 and Chinese guidelines for the diagnosis and treatment of heart failure, 21 and confirmed by two cardiologists. HF phenotypes were determined according to LVEF. Those with a significant reduction in LVEF (LVEF <40%) were designated as HFrEF. Patients with an LVEF between 40 and 49% were designated as HFmrEF, and those with an LVEF ≥50% were designated as HFpEF.
We excluded patients in whom no follow‐up data were available; other exclusion criteria were (i) acute coronary syndrome; (ii) suspected or confirmed pulmonary thromboembolism identified in CT scans of the chest or pulmonary ventilation/perfusion imaging; (iii) deep vein thromboembolism or thromboembolic events in other organs identified in imaging reports (e.g. CT scans of the head/abdomen, lower extremity ultrasound, and echocardiography); (iv) cancer; (v) acute aortic dissection; (vi) recent trauma; and (vii) severe renal dysfunction. Patients<14 years were also excluded (Figure S1 ).
Informed consent was obtained from all patients, the study complied with the Declaration of Helsinki, and it was approved by the institutional ethics committee of Fuwai Hospital.
Study endpoint
Patients were followed up at 1, 6, and 12 months through outpatient visits or telephone communication. Primary endpoint was all‐cause mortality at 1‐year follow‐up. All‐cause mortality was defined as death from any cause (including heart transplantation and left ventricular assist device requirement) during follow‐up.
Clinical and laboratory assessments
Blood samples were collected within 24 h of admission. Whole‐blood samples were collected into tubes that were then centrifuged. Haematologic tests were performed in the clinical laboratory of Fuwai Hospital. D‐dimer was measured by the PATHFAST™ D‐Dimer assay (Mitsubishi Kagaku Iatron, Inc., Tokyo, Japan) based on chemiluminescence immune assay (CLEIA) and MAGTRATION® methodology. Alkaline phosphatase‐labelled anti‐D‐dimer monoclonal antibody and anti‐D‐dimer monoclonal antibody‐coated magnetic particles are mixed with the sample. D‐dimer binds to the anti‐D‐dimer antibodies forming an immunocomplex with enzyme‐labelled antibody and antibody coated with magnetic particles. After removing the unbound enzyme‐labelled antibody, a chemiluminescent substrate is added. After incubation, the intensity of the measured luminescence generated by the enzyme reaction is proportional to the D‐dimer concentration of the sample and expressed as FEU (fibrinogen‐equivalent unit).
Baseline data such as demographics, clinical characteristics, medical treatments, and laboratory examination values were collected from medical records, prior medication, and self‐reports. Cardiac function data were collected from the first transthoracic echocardiography performed after admission. LVEF was calculated by modified Simpson's method, left atrium diameter (LAD), and left ventricular end‐diastolic diameter (LVEDD) were obtained by M‐mode echocardiography.
Co‐morbidities (e.g. hypertension and diabetes mellitus) were diagnosed according to the World Health Organization International Classification of Disease. Severe renal dysfunction was defined as an estimated glomerular rate of less than 30 mL/min/1.73m2. The estimated glomerular rate was calculated using baseline serum creatinine concentration. 22
Statistical analysis
Continuous variables are expressed as means and standard deviations (normal distribution) or medians and interquartile ranges (skewed distribution); categorical variables were presented as frequencies and percentages. N‐terminal pro‐B‐type natriuretic peptide (NT‐proBNP) was log transformed because of its right‐skewed distribution.
D‐dimer levels of the whole population were grouped into quartiles [bottom quartile (Q1): D‐dimer < 0.27 μg/mL FEU; second quartile (Q2): 0.27 μg/mL FEU ≤ D‐dimer < 0.56 μg/mL FEU; third quartile (Q3): 0.56 μg/mL FEU ≤ D‐dimer < 1.295 μg/mL FEU; top quartile (Q4) D‐dimer ≥ 1.295 μg/mL FEU]. We also grouped D‐dimer levels of the whole population into higher D‐dimer (≥0.56 μg/mL FEU) and lower D‐dimer (<0.56 μg/mL FEU), based on median value of D‐dimer.
The differences of baseline characteristics were analysed based on dichotomized D‐dimer and HF phenotypes. Student's t‐test (Wilcoxon rank‐sum test for skewed variables) was used for continuous variables, and chi‐square test or Fisher's exact test was used for categorical variables.
To visually assess non‐linear relationships between log‐transformed D‐dimer levels and mortality risk, restricted cubic spline curves before and after adjustment were performed. A thousand‐fold D‐dimer was log transformed, and the value corresponding to 0.56 was set as hazard ratio (HR) = 1.
The cumulative incidence of all‐cause death was estimated by the Kaplan–Meier method among D‐dimer quartiles and compared with log‐rank test.
Cox proportional hazards regression models assessed the relationship between D‐dimer and all‐cause mortality in total cohort and different HF phenotype cohorts. Established prognosticators were included in multivariate Cox proportional hazards model. The regression results were reported according to dichotomized D‐dimer using the lower range group as the reference and D‐dimer quartiles using the bottom quartile as the reference. Trend tests were based on D‐dimer level quartiles as a continuous variable. Testing for interaction between D‐dimer and ejection fraction groups on outcome was conducted through the use of likelihood ratio tests. The adjusted HRs with their respective 95% confidence intervals (CIs) were calculated. Proportionality assumptions were assessed using graphical approach based on Schoenfeld residuals.
As a sensitivity analysis to test the robustness of the relationship, an inverse propensity of treatment weighting (IPTW) was used. What is more, we conducted multiple imputation to maximize statistical power and minimize bias that might occur if patients with missing data were excluded from analysis. Cox proportional hazards regression models were again constructed based on cohorts after imputation.
All statistical tests were two‐tailed, and P < 0.05 was considered to be statistically significant. All analyses were performed using R 3.6.2 (R Foundation for Statistical Computing, Vienna, Austria).
Results
Patient population
The final cohort consisted of 1795 patients. Characteristics of the study population are shown in Tables 1 and S1 . Patients were classified according to dichotomized D‐dimer and HF phenotypes. The median age was 57 (47–67.5) years, and 73.4% were male. Overall, 1188 patients (72.1%) were in New York Heart Association (NYHA) functional Class III or IV. Ischaemic aetiology was present in 778 (43.3%) of patients. HFrEF was found in 947 (53.3%) patients, HFmrEF in 289 (16.3%), and HFpEF in 540 (30.4%).
Table 1.
Characteristics of the study participants dichotomized according to D‐dimer concentrations at baseline
| All population | D‐dimer < 0.56 μg/mL FEU | D‐dimer ≥ 0.56 μg/mL FEU | P value | |
|---|---|---|---|---|
| (n = 1795) | (n = 887) | (n = 908) | ||
| Demographic characteristics | ||||
| Age (years) | 57.00 [47.00, 67.50] | 57.00 [47.00, 65.00] | 58.50 [46.00, 70.00] | 0.001 |
| Sex (male), n (%) | 1317 (73.4) | 676 (76.2) | 641 (70.6) | 0.008 |
| BMI (kg/m2) | 24.45 [21.93, 27.20] | 25.02 [22.72, 27.76] | 23.70 [20.97, 26.55] | <0.001 |
| Heart rate (bpm) | 78.00 [67.00, 90.00] | 75.00 [66.00, 86.00] | 82.00 [70.00, 95.00] | <0.001 |
| Systolic BP (mmHg) | 119.42 (20.14) | 122.77 (20.36) | 116.15 (19.39) | <0.001 |
| Diastolic BP (mmHg) | 70.00 [63.00, 80.00] | 72.00 [65.00, 80.00] | 70.00 [61.00, 80.00] | <0.001 |
| NYHA functional Class III/IV, n (%) | 1188 (72.1) | 517 (63.0) | 671 (81.2) | <0.001 |
| Co‐morbidities | ||||
| Coronary heart disease, n (%) | 778 (43.3) | 385 (43.4) | 393 (43.3) | 0.996 |
| Hypertension, n (%) | 842 (46.9) | 448 (50.5) | 394 (43.4) | 0.003 |
| Diabetes, n (%) | 527 (29.4) | 257 (29.0) | 270 (29.7) | 0.762 |
| Hypertrophic cardiomyopathy, n (%) | 53 (3.0) | 36 (4.1) | 17 (1.9) | 0.009 |
| Dilated cardiomyopathy, n (%) | 575 (32.0) | 267 (30.1) | 308 (33.9) | 0.092 |
| OSAS, n (%) | 154 (8.6) | 90 (10.1) | 64 (7.0) | 0.024 |
| COPD, n (%) | 150 (8.4) | 61 (6.9) | 89 (9.8) | 0.031 |
| Laboratory variables | ||||
| White blood cell (*10^9/L) | 7.02 [5.75, 8.59] | 6.92 [5.77, 8.33] | 7.20 [5.74, 8.84] | 0.022 |
| Haemoglobin (g/L) | 139.00 (21.77) | 142.85 (19.27) | 135.24 (23.37) | <0.001 |
| Haematocrit (%) | 41.35 (6.12) | 42.15 (5.47) | 40.57 (6.61) | <0.001 |
| Platelet count (*10^9/L) | 194.00 [156.00, 241.00] | 194.00 [158.00, 238.00] | 195.00 [153.00, 245.00] | 0.939 |
| Prothrombin activity (%) | 88.00 [75.00, 99.00] | 94.00 [81.00, 103.00] | 84.00 [72.00, 93.00] | <0.001 |
| International normalized ratio (R) | 1.08 [1.01, 1.20] | 1.04 [0.98, 1.14] | 1.11 [1.04, 1.24] | <0.001 |
| Albumin (g/L) | 39.87 (5.27) | 41.50 (4.75) | 38.27 (5.26) | <0.001 |
| ALT (IU/L) | 23.00 [14.00, 37.00] | 23.00 [15.00, 36.00] | 23.00 [14.00, 40.00] | 0.805 |
| AST (IU/L) | 24.00 [18.00, 32.00] | 23.00 [18.00, 30.00] | 25.00 [19.00, 34.00] | <0.001 |
| Alkaline phosphatase (IU/L) | 66.00 [54.00, 84.00] | 63.00 [52.00, 77.00] | 71.00 [56.00, 90.00] | <0.001 |
| γ‐Glutamyl transpeptidase (IU/L) | 45.00 [27.50, 85.50] | 40.00 [26.00, 72.00] | 50.00 [29.00, 95.00] | <0.001 |
| Total bilirubin (μmol/L) | 20.00 [14.30, 29.79] | 18.50 [13.50, 25.87] | 22.20 [15.10, 34.10] | <0.001 |
| Serum sodium (mmol/L) | 138.00 [135.40, 140.00] | 138.68 [136.00, 140.63] | 137.52 [135.00, 140.00] | <0.001 |
| Serum potassium (mmol/L) | 3.94 [3.66, 4.26] | 3.92 [3.67, 4.20] | 4.00 [3.65, 4.30] | 0.012 |
| Serum creatinine (μmol/L) | 90.28 [75.40, 108.15] | 87.20 [73.54, 103.79] | 93.84 [77.22, 112.50] | <0.001 |
| Blood urea nitrogen (mmol/L) | 7.20 [5.47, 9.31] | 6.80 [5.28, 8.80] | 7.50 [5.69, 9.86] | <0.001 |
| Triglyceride (mmol/L) | 1.28 [0.95, 1.83] | 1.48 [1.05, 2.08] | 1.18 [0.89, 1.58] | <0.001 |
| Total cholesterol (mmol/L) | 3.99 [3.30, 4.74] | 4.11 [3.43, 4.86] | 3.88 [3.17, 4.66] | <0.001 |
| High‐density lipoprotein (mmol/L) | 0.95 [0.78, 1.18] | 0.98 [0.82, 1.23] | 0.91 [0.73, 1.13] | <0.001 |
| Low‐density lipoprotein (mmol/L) | 2.45 [1.92, 3.06] | 2.49 [1.94, 3.12] | 2.42 [1.89, 3.02] | 0.193 |
| Hs‐CRP (mg/L) | 3.32 [1.46, 9.55] | 2.22 [1.10, 4.92] | 5.30 [2.26, 11.66] | <0.001 |
| NT‐proBNP (pg/mL) | 1788.00 [710.00, 4480.00] | 1061.00 [466.50, 2385.50] | 3101.50 [1297.50, 6581.25] | <0.001 |
| Log‐transformed NT‐proBNP | 3.22 (0.63) | 2.99 (0.61) | 3.45 (0.57) | <0.001 |
| Echocardiographic evaluation | ||||
| LAD (mm) | 44.00 [39.00, 50.00] | 44.00 [38.00, 49.00] | 45.00 [40.00, 51.00] | <0.001 |
| LVEDD (mm) | 61.51 (12.57) | 61.29 (12.55) | 61.73 (12.59) | 0.461 |
| LVPWT (mm) | 9.00 [8.00, 10.00] | 9.50 [8.05, 10.00] | 9.00 [8.00, 10.00] | 0.183 |
| IVST (mm) | 10.00 [8.50, 11.00] | 10.00 [8.60, 11.00] | 9.90 [8.25, 11.00] | 0.255 |
| LVEF (%) | 38.00 [29.00, 53.00] | 40.00 [30.00, 55.00] | 35.00 [27.00, 50.00] | <0.001 |
| Medication | ||||
| Digoxin, n (%) | 899 (50.8) | 417 (47.3) | 482 (54.3) | 0.003 |
| Beta blockers, n (%) | 1524 (86.2) | 773 (87.6) | 751 (84.7) | 0.082 |
| ACEi/ARB, n (%) | 1036 (58.6) | 582 (66.0) | 454 (51.2) | <0.001 |
| Aldosterone antagonists, n (%) | 1278 (72.2) | 629 (71.3) | 649 (73.2) | 0.414 |
| Diuretics, n (%) | 1520 (86.5) | 759 (87.0) | 761 (86.0) | 0.565 |
ACEi, angiotensin‐converting enzyme inhibitor; ALT, alanine aminotransferase; ARB, angiotensin II receptor blocker; AST, aspartate aminotransferase; BMI, body mass index; BP, blood pressure; COPD, chronic obstructive pulmonary disease; hs‐CRP, high‐sensitivity C‐reactive protein; IVST, interventricular septal thickness; LAD, left atrium diameter; LVEDD, left ventricular end‐diastolic diameter; LVEF, left ventricular ejection fraction; LVPWT, left ventricular posterior wall thickness; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; NYHA, New York Heart Association; OSAS, obstructive sleep apnoea syndrome.
Continuous variables are presented as mean (SD) or median (interquartile range). Categorical variables are presented as numbers and frequencies [n (%)].
Overall, patients with higher D‐dimer tended to be older, have lower body mass index (BMI), higher NYHA functional class, higher levels of NT‐proBNP, and lower LVEF. They were more frequently complicated by dyslipidaemia and impaired liver and renal function (Table 1 ).
Impact of dichotomized D‐dimer and D‐dimer quartiles on adverse events in the total cohort
The median (interquartile range) D‐dimer concentrations in the whole cohort was 0.56 (0.27–1.295) μg/mL FEU. All‐cause mortality at 1‐year follow‐up occurred in 324 (18.1%) patients.
The restricted cubic spline curves showed that log‐transformed D‐dimer has an S‐shaped relationship with all‐cause mortality over 1‐year follow‐up period (Figure S2A ), even after adjustment of age, sex, and BMI (Figure S2B ).
Kaplan–Meier survival analysis for quartiles of D‐dimer yielded diverging curves, with the comparison showing pronounced risk stratification (log‐rank test P < 0.0001) (Figure 1 A ).
Figure 1.
Kaplan–Meier survival curves according to D‐dimer quartiles in total cohort and different HF phenotypes. 1‐year all‐cause mortality categorized by D‐dimer quartiles in total cohort (A), HFrEF (B), HFmrEF (C), and HFpEF (D).
Results of the univariate and multivariate Cox proportional hazards analysis for all‐cause death are depicted in Table 2 . We simultaneously extracted the results from two adjusted models: Model 1 was adjusted for age, sex, and BMI; Model 2 was further adjusted for other well‐established prognosticators such as systolic blood pressure, heart rate, serum creatine, serum sodium, total bilirubin, total cholesterol, haemoglobin, log‐transformed NT‐proBNP, LVEF, CHD, NYHA functional Class III or IV, diagnosis of diabetes, renin–angiotensin system inhibitor (RASi) prescription, beta‐blocker prescription, and aldosterone antagonist prescription. Higher D‐dimer remained markedly associated with all‐cause death at 1‐year follow‐up after adjustment (HR, 1.55; 95% CI 1.15–2.1); a similar trend was seen when higher quartiles were compared with lower quartiles.
Table 2.
Risk of 1‐year all‐cause death by D‐dimer
| Unadjusted | Model 1 | Model 2 | |||||
|---|---|---|---|---|---|---|---|
| HR (95% CI) | P value | HR (95% CI) | P value | HR (95% CI) | P value | ||
| D‐dimer ≥ 0.56 vs. <0.56 μg/mL FEU | |||||||
| Overall | 3.12 (2.41,4.03) | <0.001 | 2.62 (2,3.44) | <0.001 | 1.55 (1.15,2.1) | 0.004 | |
| 1.64 (1.11,2.43) a | 0.014 a | ||||||
| HFrEF | 2.79 (2.02,3.84) | <0.001 | 2.43 (1.74,3.4) | <0.001 | 1.49 (1.02,2.17) | 0.039 | |
| 1.42 (0.87,2.30) a | 0.158 a | ||||||
| HFmrEF | 2.64 (1.25,5.62) | 0.011 | 1.77 (0.76,4.15) | 0.188 | 0.74 (0.26,2.1) | 0.570 | |
| 0.55 (0.14,2.21) a | 0.400 a | ||||||
| HFpEF | 4.18 (2.42,7.21) | <0.001 | 3.65 (2.07,6.42) | <0.001 | 2.06 (1.06,4) | 0.033 | |
| 2.62 (1.01,6.77) a | 0.048 a | ||||||
| P value for interaction | 0.419 | 0.686 | 0.716 | ||||
| D‐dimer quartiles | |||||||
| Overall | |||||||
| Q1:D‐dimer < 0.27 μg/mL FEU | Reference | P for trend < 0.001 | Reference | P for trend < 0.001 | Reference | P for trend 0.017 | |
| Q2:0.27 ≤ D‐dimer < 0.56 μg/mL FEU | 1.52 (0.97,2.39) | 0.066 | 1.5 (0.95,2.39) | 0.082 | 0.96 (0.59,1.58) | 0.882 | |
| Q3:0.56 ≤ D‐dimer < 1.295 μg/mL FEU | 3.31 (2.22,4.93) | <0.001 | 2.85 (1.88,4.31) | <0.001 | 1.53 (0.99,2.37) | 0.054 | |
| Q4:D‐dimer ≥ 1.295 μg/mL FEU | 4.59 (3.11,6.77) | <0.001 | 3.71 (2.48,5.54) | < 0.001 | 1.5 (0.97,2.33) | 0.068 | |
| HFrEF | |||||||
| Q1:D‐dimer < 0.27 μg/mL FEU | Reference | P for trend < 0.001 | Reference | P for trend < 0.001 | Reference | P for trend 0.244 | |
| Q2:0.27 ≤ D‐dimer < 0.56 μg/mL FEU | 1.09 (0.63,1.9) | 0.754 | 1.05 (0.59,1.86) | 0.868 | 0.77 (0.42,1.43) | 0.408 | |
| Q3:0.56 ≤ D‐dimer < 1.295 μg/mL FEU | 2.67 (1.69,4.24) | < 0.001 | 2.38 (1.48,3.82) | < 0.001 | 1.45 (0.87,2.41) | 0.152 | |
| Q4:D‐dimer ≥ 1.295 μg/mL FEU | 3.14 (2.01,4.91) | < 0.001 | 2.6 (1.63,4.13) | < 0.001 | 1.18 (0.7,1.98) | 0.532 | |
| HFmrEF | |||||||
| Q1:D‐dimer < 0.27 μg/mL FEU | Reference | P for trend < 0.001 | Reference | P for trend 0.026 | Reference | P for trend 0.921 | |
| Q2:0.27 ≤ D‐dimer < 0.56 μg/mL FEU | 2.02 (0.52,7.8) | 0.309 | 1.55 (0.39,6.2) | 0.537 | 0.41 (0.06,2.58) | 0.340 | |
| Q3:0.56 ≤ D‐dimer < 1.295 μg/mL FEU | 2.83 (0.77,10.46) | 0.118 | 1.21 (0.29,5.11) | 0.791 | 0.25 (0.04,1.46) | 0.124 | |
| Q4:D‐dimer≥1.295 μg/mL FEU | 6.12 (1.73,21.68) | 0.005 | 4.08 (1.11,14.94) | 0.034 | 0.76 (0.14,4.02) | 0.744 | |
| HFpEF | |||||||
| Q1:D‐dimer < 0.27 μg/mL FEU | Reference | P for trend < 0.001 | Reference | P for trend < 0.001 | Reference | P for trend 0.003 | |
| Q2:0.27 ≤ D‐dimer < 0.56 μg/mL FEU | 3.4 (1.11,10.42) | 0.032 | 3.67 (1.19,11.32) | 0.024 | 2.91 (0.88,9.65) | 0.081 | |
| Q3:0.56 ≤ D‐dimer < 1.295 μg/mL FEU | 7.38 (2.56,21.27) | <0.001 | 6.92 (2.36,20.31) | <0.001 | 3.17 (1.01,9.92) | 0.047 | |
| Q4:D‐dimer ≥ 1.295 μg/mL FEU | 11.13 (3.92,31.61) | <0.001 | 9.86 (3.41,28.54) | <0.001 | 5.61 (1.77,17.81) | 0.003 | |
| P value for interaction | 0.27 | 0.243 | 0.292 | ||||
CI, confidence interval; HFmrEF, heart failure with mildly reduced ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; HR, hazard ratio.
Model 1, adjusted for age, sex, and BMI; Model 2, further adjusted for systolic blood pressure, heart rate, serum creatine, serum sodium, total bilirubin, total cholesterol, haemoglobin, log‐transformed NT‐proBNP, LVEF, CHD, NYHA functional Class III or IV, diagnosis of diabetes, RASi prescription, beta‐blocker prescription, and aldosterone antagonist prescription.
Results of inverse propensity of treatment weighting.
Impact of dichotomized D‐dimer and D‐dimer quartiles on survival prediction in different HF phenotype cohorts
The median (interquartile range) D‐dimer concentrations in the three EF phenotypes were 0.64 (0.28–1.48) μg/mL FEU in HFrEF, 0.50 (0.27–1.03) μg/mL FEU in HFmrEF, and 0.495 (0.25–1.10) μg/mL FEU in HFpEF (P = 0.001).
Disparity exists in D‐dimer's prognostic effects in three HF phenotype cohorts. Though Kaplan–Meier survival analysis for quartiles of D‐dimer yielded diverging curves in different HF phenotypes individually (all log‐rank test P < 0.05) (Figure 1 B–D ), prognostic power of D‐dimer varied through the whole spectrum of EF phenotypes (Table 2 ). In HFpEF, at 1‐year follow‐up, patients with higher D‐dimer had 2.06 times higher risk after adjustment (P = 0.033, 95% CI 1.06–4); with regard to quartiles, patients with D‐dimer values in Q2, Q3, and Q4 had 2.91 (P = 0.081, 95% CI 0.88–9.65), 3.17 (P = 0.047, 95% CI 1.01–9.92), and 5.61 (P = 0.003, 95% CI 1.77–17.81) higher risk compared with patients in Q1, respectively (P for trend 0.003). In HFrEF or HFmrEF patients, higher D‐dimer quartiles were not related to increased risk of all‐cause death.
Sensitivity analysis
IPTW allows a pseudo‐population to be created through assigning individuals with weights that corresponded to the inverse of their probability of being allocated to dichotomized D‐dimer given observed covariates; it was carried out for sensitivity analysis. In line with previous findings, it showed pronounced prognostic effect of baseline dichotomized D‐dimer in whole cohort (HR, 1.64; P = 0.014) and in HFpEF (HR, 2.62; P = 0.048) (Table 2 ). In HFrEF, higher D‐dimer was prognostic of 1‐year mortality (HR, 1.49; P = 0.039); however, after IPTW, this effect was attenuated (HR, 1.42; P = 0.158). After multiple imputation of missing data, we repeated Cox proportional hazards analysis in the total cohort and HF phenotype cohorts; quantitatively similar but more pronounced results were found as in the prior analysis. When the total cohort was categorized according to D‐dimer quartiles, we observed a notable association between D‐dimer and 1‐year all‐cause mortality; patients with D‐dimer values in Q2, Q3, and Q4 had 1.16 (P = 0.524, 95% CI 0.73–1.83), 1.7 (P = 0.012, 95% CI 1.12–2.56), and 1.68 (P = 0.013, 95% CI 1.11–2.55) higher risk compared with patients in Q1, respectively (P for trend 0.004). Results are shown in Table S2 .
What is more, because we excluded 284 patients in whom follow‐up data were absent, a comparison of baseline characteristics for patients lost in follow‐up and patients included in final analysis was conducted to avoid potential risk of bias (we excluded a total of 82 patients in cohort of patients lost to follow‐up according to previously stated exclusion criteria). Obvious difference was not discerned between groups (Table S3 ).
Discussion
This single‐centre study evaluated the associations between D‐dimer concentrations and 1‐year mortality in 1795 consecutive patients hospitalized with HF from 2006 to 2017. It reveals that D‐dimer was associated with a higher risk of 1‐year mortality in the whole hospitalized HF patient cohort. Across the ejection fraction phenotypes, a more pronounced association was found in patients with HFpEF. Sensitivity analyses reinforced the finding.
Thromboembolism in heart failure
Thromboembolism is a major complication in hospitalized HF patients. 23 , 24 , 25 , 26 The ARIC (Atherosclerosis Risk in Communities) cohort found 27% incidence rate of venous thromboembolism in HF outpatients over 22 years follow‐up. 27 Patients with HF are characterized by ‘Virchow triad’, that is, blood stasis, hypercoagulable state, endocardium, and endothelial dysfunction, all resulting in a pro‐thrombotic state. 28 , 29 Patients complicated with thromboembolism have poor prognosis according to the Worcester Venous Thromboembolism Study. 30 The disease burden and poor prognosis indicate a significant role of pro‐thrombotic biomarkers in HF prognosis.
D‐dimer and its unexplored role in heart failure
D‐dimer is a simple, widely employed biomarker reflecting activation of coagulation and fibrinolysis, and it is applied to the diagnosis of venous thromboembolism, identification of patients at high risk of venous thromboembolism, identification of patients who may benefit from rivaroxaban for stroke reduction, prediction of stroke risk in atrial fibrillation, and independent prediction of cardiovascular disease mortality and other fatal events across multiple diseases. 31 , 32 However, despite abundant evidence of elevated pro‐thrombotic biomarkers in HF patients, including D‐dimer, 11 , 12 their prognostic value has not been recognized until recent years. 33 , 34 , 35 , 36 Previous studies on this topic are characterized by limitations, including relatively small sample size, 33 , 34 , 36 or short follow‐up. 34 Furthermore, study populations were at times complicated by multiple co‐morbidities that may affect D‐dimer levels, such as malignancies, aortic dissection, and deep vein thromboembolism. 35 Thus, a more accurate assessment is needed.
Different HF phenotypes are distinct in underlying diseases, pathophysiological mechanisms, clinical manifestations, and disease prognosis. In this respect, biomarker‐based risk stratification is not only feasible and easily accessible in clinical practice, but it could also prove important for tailoring treatment and long‐term management. However, most widely accepted biomarkers, such as NT‐proBNP, troponin, soluble suppression of tumorigenicity 2 (sST2), galectin‐3, and growth differentiation factor‐15 (GDF‐15), tend to have prognostic data derived from studies in unselected HF populations, or restricted to HFrEF, whereas valuable prognostic biomarkers in hospitalized HFpEF patients, with a few exceptions, 37 , 38 are poorly investigated.
Prognostic importance of D‐dimer in hospitalized heart failure patients and across the range of heart failure phenotypes
In the present study, higher D‐dimer independently predicted all‐cause death at 1‐year follow‐up. Sensitivity analysis revealed more pronounced prognostic value of D‐dimer, underlining D‐dimer's prognostic effects at 1 year.
The precise mechanisms of baseline D‐dimer's prognostic effects on 1‐year mortality are unclear. D‐dimer is a symbol of hypercoagulable state that is, by most reckoning, transient. However, in recent study, it is found that hypercoagulable state can be persistent in medically ill hospitalized patients. 39 It is conceivable that, in addition to reversible and correctable pro‐thrombotic causes (such as immobilization and hemodynamic alterations in hospitalized HF patients), endothelial injury, chronic systemic inflammatory condition, and persistent hemodynamic abnormalities, may be long‐term risk factors of thrombosis in HF patients.
What is more, elevated D‐dimer is not merely associated with HF prognosis as frequently noted, that it may be the consequence of haemodynamic alterations and impaired blood flow in HF, reflecting the cardiac functional status and disease severity, and that elevated D‐dimer in turn results in inflammatory reactions by inducing synthesis and release of inflammatory cytokines, 40 adding to the disease burden.
With respect of different HF phenotypes, the present study shows that D‐dimer has particular prognostic value in HFpEF patients. This is not entirely surprising, as over the last decade, our understanding of HFpEF has evolved from a mere left ventricular diastolic dysfunction to a multi‐organ syndrome, in which obesity, diabetes, and hypertension may predispose to systemic inflammation and microcirculatory alterations affecting the myocardium. 13 , 41 Cluster analysis of HFpEF based on clinical characteristics and biomarker profiles has been developed to explain heterogeneity in pathophysiology and to identify ‘phenogroups’ with distinctive prognosis. 17 , 18 , 42 , 43 , 44 However, apart from platelet activation biomarkers, thrombotic biomarkers are under‐represented in those studies. A study on SERPINE1 and ageing has pointed out that plasminogen activator inhibitor 1 was associated with HFpEF, 45 which indicated that coagulation and fibrinolysis biomarkers might be essential for clustering of HFpEF phenogroups and understanding of pathophysiological pathways, which merits further investigation.
Limitations
This is a single‐centre study, performed in a Chinese population; hence, findings should be extrapolated cautiously to populations with different ethnicity. Due to its observational nature, there are possible confounders that we were unable to control. Of note, however, patient enrolment was concluded well before onset of SARS‐Cov‐2 pandemic, and therefore, potential confounding effects of D‐dimer elevation in patients with concomitant COVID‐19 and HF can be excluded. 46 With respect to different EF phenotypes, in our analysis, HFmrEF patients were substantially fewer, in line with previous reports, 14 , 15 , 19 and this may have contributed to the inability to detect a significant effect in this sub‐group. Finally, patients in this study were enrolled prior to the extensive prescription of ARNI and SGLT‐2 inhibitors.
Conclusions
In patients hospitalized with HF, D‐dimer was a significant and independent predictor of 1‐year all‐cause mortality. Across all HF phenotypes, D‐dimer was independently associated with all‐cause mortality at 1‐year follow‐up period in HFpEF patients. Future studies exploring treatment with statins and anti‐thrombotic agents might be considered in these patients.
Conflict of interest
None declared.
Funding
This work was supported by the Key Projects in the National Science and Technology Pillar Program of the 13th Five‐Year Plan Period, Beijing, People's Republic of China (grant number 2017YFC1308300 to Jian Zhang).
Supporting information
Figure S1. Study Flow Chart.
Figure S2. D‐dimer Concentrations and All‐Cause Mortality.
Table S1. Characteristics of the Study Participants Categorized by HF Phenotypes and D‐dimer Concentrations.
Table S2. Risk of All‐Cause Death by D‐dimer (After Imputation).
Table S3. Characteristics of the Patients Lost in Follow‐up and Patients Included in Final Analysis.
Acknowledgements
We are grateful to all members who contributed to the study.
Huang, L. , Liang, L. , Tian, P. , Zhao, L. , Chen, Y. , Huang, Y. , Zhou, Q. , Zhai, M. , Zhang, Y. , Ambrosio, G. , and Zhang, J. (2022) D‐dimer and outcomes in hospitalized heart failure patients across the ejection fraction phenotypes. ESC Heart Failure, 9: 3060–3070. 10.1002/ehf2.14049.
All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation.
Contributor Information
Yuhui Zhang, Email: yuhuizhangjoy@126.com.
Giuseppe Ambrosio, Email: giuseppe.ambrosio@ospedale.perugia.it.
References
- 1. Bloom MW, Greenberg B, Jaarsma T, Januzzi JL, Lam CSP, Maggioni AP, Trochu JN, Butler J. Heart failure with reduced ejection fraction. Nat Rev Dis Primers 2017; 3: 17058. [DOI] [PubMed] [Google Scholar]
- 2. Ponikowski P, Anker SD, AlHabib KF, Cowie MR, Force TL, Hu S, Jaarsma T, Krum H, Rastogi V, Rohde LE, Samal UC, Shimokawa H, Budi Siswanto B, Sliwa K, Filippatos G. Heart failure: Preventing disease and death worldwide. ESC Heart Fail 2014; 1: 4–25. [DOI] [PubMed] [Google Scholar]
- 3. Lip GY, Piotrponikowski P, Andreotti F, Anker SD, Filippatos G, Homma S, Morais J, Pullicino P, Rasmussen LH, Marín F, Lane DA. Thromboembolism and antithrombotic therapy for heart failure in sinus rhythm: An executive summary of a joint consensus document from the ESC Heart Failure Association and the ESC working group on thrombosis. Thromb Haemost 2012; 108: 1009–1022. [DOI] [PubMed] [Google Scholar]
- 4. Ambrosio G, Camm AJ, Bassand JP, Corbalan R, Kayani G, Carluccio E, Mantovani LG, Virdone S, Kakkar AK. Characteristics, treatment, and outcomes of newly diagnosed atrial fibrillation patients with heart failure: GARFIELD‐AF. ESC Heart Fail 2021; 8: 1139–1149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Corbalan R, Bassand JP, Illingworth L, Ambrosio G, Camm AJ, Fitzmaurice DA, Fox KAA, Goldhaber SZ, Goto S, Haas S, Kayani G, Mantovani LG, Misselwitz F, Pieper KS, Turpie AGG, Verheugt FWA, Kakkar AK. Analysis of outcomes in ischemic vs nonischemic cardiomyopathy in patients with atrial fibrillation: A report from the GARFIELD‐AF registry. JAMA Cardiol 2019; 4: 526–548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Adam SS, Key NS, Greenberg CS. D‐dimer antigen: Current concepts and future prospects. Blood 2009; 113: 2878–2887. [DOI] [PubMed] [Google Scholar]
- 7. van der Hulle T, den Exter PL, Erkens PG, van Es J, Mos IC, ten Cate H, Kamphuisen PW, Hovens MM, Büller HR, Klok FA, Huisman MV. Variable D‐dimer thresholds for diagnosis of clinically suspected acute pulmonary embolism. J Thromb Haemost 2013; 11: 1986–1992. [DOI] [PubMed] [Google Scholar]
- 8. Faller N, Limacher A, Méan M, Righini M, Aschwanden M, Beer JH, Frauchiger B, Osterwalder J, Kucher N, Lämmle B, Cornuz J, Angelillo‐Scherrer A, Matter CM, Husmann M, Banyai M, Staub D, Mazzolai L, Hugli O, Rodondi N, Aujesky D. Predictors and causes of long‐term mortality in elderly patients with acute venous thromboembolism: A prospective cohort study. Am J Med 2017; 130: 198–206. [DOI] [PubMed] [Google Scholar]
- 9. Bayes‐Genis A, Mateo J, Santaló M, Oliver A, Guindo J, Badimon L, Martínez‐Rubio A, Fontcuberta J, Schwartz RS, De Luna AB. D‐dimer is an early diagnostic marker of coronary ischemia in patients with chest pain. Am Heart J 2000; 140: 379–384. [DOI] [PubMed] [Google Scholar]
- 10. Suzuki T, Distante A, Zizza A, Trimarchi S, Villani M, Salerno Uriarte JA, De Luca Tupputi Schinosa L, Renzulli A, Sabino F, Nowak R, Birkhahn R, Hollander JE, Counselman F, Vijayendran R, Bossone E, Eagle K. Diagnosis of acute aortic dissection by D‐dimer: The international registry of acute aortic dissection substudy on biomarkers (IRAD‐Bio) experience. Circulation 2009; 119: 2702–2707. [DOI] [PubMed] [Google Scholar]
- 11. Davis CJ, Gurbel PA, Gattis WA, Fuzaylov SY, Nair GV, O'Connor CM, Serebruany VL. Hemostatic abnormalities in patients with congestive heart failure: Diagnostic significance and clinical challenge. Int J Cardiol 2000; 75: 15–21. [DOI] [PubMed] [Google Scholar]
- 12. Jafri SM, Ozawa T, Mammen E, Levine TB, Johnson C, Goldstein S. Platelet function, thrombin and fibrinolytic activity in patients with heart failure. Eur Heart J 1993; 14: 205–212. [DOI] [PubMed] [Google Scholar]
- 13. Paulus WJ, Tschöpe C. A novel paradigm for heart failure with preserved ejection fraction: Comorbidities drive myocardial dysfunction and remodeling through coronary microvascular endothelial inflammation. J Am Coll Cardiol 2013; 62: 263–271. [DOI] [PubMed] [Google Scholar]
- 14. Chioncel O, Lainscak M, Seferovic PM, Anker SD, Crespo‐Leiro MG, Harjola VP, Parissis J, Laroche C, Piepoli MF, Fonseca C, Mebazaa A, Lund L, Ambrosio GA, Coats AJ, Ferrari R, Ruschitzka F, Maggioni AP, Filippatos G. Epidemiology and one‐year outcomes in patients with chronic heart failure and preserved, mid‐range and reduced ejection fraction: An analysis of the ESC heart failure long‐term registry. Eur J Heart Fail 2017; 19: 1574–1585. [DOI] [PubMed] [Google Scholar]
- 15. Koh AS, Tay WT, Teng THK, Vedin O, Benson L, Dahlstrom U, Savarese G, Lam CSP, Lund LH. A comprehensive population‐based characterization of heart failure with mid‐range ejection fraction. Eur J Heart Fail 2017; 19: 1624–1634. [DOI] [PubMed] [Google Scholar]
- 16. Senni M, Paulus WJ, Gavazzi A, Fraser AG, Díez J, Solomon SD, Smiseth OA, Guazzi M, Lam CS, Maggioni AP, Tschöpe C, Metra M, Hummel SL, Edelmann F, Ambrosio G, Stewart Coats AJ, Filippatos GS, Gheorghiade M, Anker SD, Levy D, Pfeffer MA, Stough WG, Pieske BM. New strategies for heart failure with preserved ejection fraction: The importance of targeted therapies for heart failure phenotypes. Eur Heart J 2014; 35: 2797–2815. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Napoli C, Benincasa G, Donatelli F, Ambrosio G. Precision medicine in distinct heart failure phenotypes: Focus on clinical epigenetics. Am Heart J 2020; 224: 113–128. [DOI] [PubMed] [Google Scholar]
- 18. Stienen S, Ferreira JP, Kobayashi M, Preud'homme G, Dobre D, Machu JL, Duarte K, Bresso E, Devignes MD, López N, Girerd N, Aakhus S, Ambrosio G, Brunner‐La Rocca HP, Fontes‐Carvalho R, Fraser AG, van Heerebeek L, Heymans S, de Keulenaer G, Marino P, McDonald K, Mebazaa A, Papp Z, Raddino R, Tschöpe C, Paulus WJ, Zannad F, Rossignol P. Enhanced clinical phenotyping by mechanistic bioprofiling in heart failure with preserved ejection fraction: Insights from the MEDIA‐DHF study (the metabolic road to diastolic heart failure). Biomarkers 2020; 25: 201–211. [DOI] [PubMed] [Google Scholar]
- 19. Ambrosio G, Leiro MGC, Lund LH, Coiro S, Cardona A, Filippatos G, Ferrari R, Piepoli MF, Coats AJS, Anker SD, Laroche C, Almenar‐Bonet L, Poder P, Valero DB, Frisinghelli A, Maggioni AP. Serum uric acid and outcomes in patients with chronic heart failure through the whole spectrum of ejection fraction phenotypes: Analysis of the ESC‐EORP heart failure long‐term (HF LT) registry. Eur J Intern Med 2021; 89: 65–75. [DOI] [PubMed] [Google Scholar]
- 20. McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of congestive heart failure: The Framingham study. N Engl J Med 1971; 285: 1441–1446. [DOI] [PubMed] [Google Scholar]
- 21. Chinese Society of Cardiology of Chinese Medical Association , Editorial Board of Chinese Journal of Cardiology . Chinese guidelines for the diagnosis and treatment of heart failure 2014. Zhonghua Xin Xue Guan Bing Za Zhi 2014; 42: 98–122. [PubMed] [Google Scholar]
- 22. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009; 150: 604–612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Beemath A, Stein PD, Skaf E, Al Sibae MR, Alesh I. Risk of venous thromboembolism in patients hospitalized with heart failure. Am J Cardiol 2006; 98: 793–795. [DOI] [PubMed] [Google Scholar]
- 24. Cohen AT, Davidson BL, Gallus AS, Lassen MR, Prins MH, Tomkowski W, Turpie AG, Egberts JF, Lensing AW. Efficacy and safety of fondaparinux for the prevention of venous thromboembolism in older acute medical patients: Randomised placebo controlled trial. BMJ 2006; 332: 325–329. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Leizorovicz A, Cohen AT, Turpie AG, Olsson CG, Vaitkus PT, Goldhaber SZ. Randomized, placebo‐controlled trial of dalteparin for the prevention of venous thromboembolism in acutely ill medical patients. Circulation 2004; 110: 874–879. [DOI] [PubMed] [Google Scholar]
- 26. Ng TM, Tsai F, Khatri N, Barakat MN, Elkayam U. Venous thromboembolism in hospitalized patients with heart failure: Incidence, prognosis, and prevention. Circ Heart Fail 2010; 3: 165–173. [DOI] [PubMed] [Google Scholar]
- 27. Fanola CL, Norby FL, Shah AM, Chang PP, Lutsey PL, Rosamond WD, Cushman M, Folsom AR. Incident heart failure and long‐term risk for venous thromboembolism. J Am Coll Cardiol 2020; 75: 148–158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Zannad F, Stough WG, Regnault V, Gheorghiade M, Deliargyris E, Gibson CM, Agewall S, Berkowitz SD, Burton P, Calvo G, Goldstein S, Verheugt FW, Koglin J, O'Connor CM. Is thrombosis a contributor to heart failure pathophysiology? Possible mechanisms, therapeutic opportunities, and clinical investigation challenges. Int J Cardiol 2013; 167: 1772–1782. [DOI] [PubMed] [Google Scholar]
- 29. Aispuru GR, Clavier MM, Cardone AJ, Gilberto DO, Barousse AP. Thrombotic biomarkers and left ventricle characteristics as short‐term predictors of thrombotic events in patients hospitalized for acute decompensated heart failure. Eur J Intern Med 2012; 23: 545–551. [DOI] [PubMed] [Google Scholar]
- 30. Piazza G, Goldhaber SZ, Lessard DM, Goldberg RJ, Emery C, Spencer FA. Venous thromboembolism in heart failure: Preventable deaths during and after hospitalization. Am J Med 2011; 124: 252–259. [DOI] [PubMed] [Google Scholar]
- 31. Weitz JI, Fredenburgh JC, Eikelboom JW. A test in context: D‐dimer. J Am Coll Cardiol 2017; 70: 2411–2420. [DOI] [PubMed] [Google Scholar]
- 32. Ferreira JP, Lam CSP, Anker SD, Mehra MR, van Veldhuisen DJ, Byra WM, La Police DA, Cleland JGF, Greenberg B, Zannad F. Plasma D‐dimer concentrations predicting stroke risk and rivaroxaban benefit in patients with heart failure and sinus rhythm: An analysis from the COMMANDER‐HF trial. Eur J Heart Fail 2021; 23: 648–656. [DOI] [PubMed] [Google Scholar]
- 33. Huang B, Li YJ, Shen J, Yang Y, Liu G, Luo SX. D‐dimer level and long‐term outcome in patients with end‐stage heart failure secondary to idiopathic dilated cardiomyopathy. J Geriatr Cardiol 2019; 16: 621–629. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Minami Y, Haruki S, Jujo K, Itani R, Shimazaki K, Arashi H, Watanabe E, Hagiwara N. Elevated D‐dimer levels predict an adverse outcome in hospitalized patients with acute decompensated heart failure. Int J Cardiol 2016; 204: 42–44. [DOI] [PubMed] [Google Scholar]
- 35. Alehagen U, Dahlström U, Lindahl TL. Elevated D‐dimer level is an independent risk factor for cardiovascular death in out‐patients with symptoms compatible with heart failure. Thromb Haemost 2004; 92: 1250–1258. [DOI] [PubMed] [Google Scholar]
- 36. Zorlu A, Yilmaz MB, Yucel H, Bektasoglu G, Refiker Ege M, Tandogan I. Increased d‐dimer levels predict cardiovascular mortality in patients with systolic heart failure. J Thromb Thrombolysis 2012; 33: 322–328. [DOI] [PubMed] [Google Scholar]
- 37. de Boer RA, Nayor M, deFilippi CR, Enserro D, Bhambhani V, Kizer JR, Blaha MJ, Brouwers FP, Cushman M, Lima JAC, Bahrami H, van der Harst P, Wang TJ, Gansevoort RT, Fox CS, Gaggin HK, Kop WJ, Liu K, Vasan RS, Psaty BM, Lee DS, Hillege HL, Bartz TM, Benjamin EJ, Chan C, Allison M, Gardin JM, Januzzi JL Jr, Shah SJ, Levy D, Herrington DM, Larson MG, van Gilst WH, Gottdiener JS, Bertoni AG, Ho JE. Association of cardiovascular biomarkers with Incident heart failure with preserved and reduced ejection fraction. JAMA Cardiol 2018; 3: 215–224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Blanco R, Ambrosio G, Belziti C, Lucas L, Arias A, D'Antonio A, Oberti P, Carluccio E, Pizarro R. Prognostic value of NT‐proBNP, and echocardiographic indices of diastolic function, in hospitalized patients with acute heart failure and preserved left ventricular ejection fraction. Int J Cardiol 2020; 317: 111–120. [DOI] [PubMed] [Google Scholar]
- 39. Amin AN, Varker H, Princic N, Lin J, Thompson S, Johnston S. Duration of venous thromboembolism risk across a continuum in medically ill hospitalized patients. J Hosp Med 2012; 7: 231–238. [DOI] [PubMed] [Google Scholar]
- 40. Robson SC, Shephard EG, Kirsch RE. Fibrin degradation product D‐dimer induces the synthesis and release of biologically active IL‐1 beta, IL‐6 and plasminogen activator inhibitors from monocytes in vitro. Br J Haematol 1994; 86: 322–326. [DOI] [PubMed] [Google Scholar]
- 41. Shah SJ, Kitzman DW, Borlaug BA, van Heerebeek L, Zile MR, Kass DA, Paulus WJ. Phenotype‐specific treatment of heart failure with preserved ejection fraction: A multiorgan roadmap. Circulation 2016; 134: 73–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Hedman ÅK, Hage C, Sharma A, Brosnan MJ, Buckbinder L, Gan LM, Shah SJ, Linde CM, Donal E, Daubert JC, Mälarstig A, Ziemek D, Lund L. Identification of novel pheno‐groups in heart failure with preserved ejection fraction using machine learning. Heart 2020; 106: 342–349. [DOI] [PubMed] [Google Scholar]
- 43. Cohen JB, Schrauben SJ, Zhao L, Basso MD, Cvijic ME, Li Z, Yarde M, Wang Z, Bhattacharya PT, Chirinos DA, Prenner S, Zamani P, Seiffert DA, Car BD, Gordon DA, Margulies K, Cappola T, Chirinos JA. Clinical phenogroups in heart failure with preserved ejection fraction: Detailed phenotypes, prognosis, and response to spironolactone. JACC Heart Fail 2020; 8: 172–184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Woolley RJ, Ceelen D, Ouwerkerk W, Tromp J, Figarska SM, Anker SD, Dickstein K, Filippatos G, Zannad F, Metra M, Ng L, Samani N, van Veldhuisen DJ, Lang C, Lam CS, Voors AA. Machine learning based on biomarker profiles identifies distinct subgroups of heart failure with preserved ejection fraction. Eur J Heart Fail 2021; 23: 983–991. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Khan SS, Shah SJ, Klyachko E, Baldridge AS, Eren M, Place AT, Aviv A, Puterman E, Lloyd‐Jones DM, Heiman M, Miyata T, Gupta S, Shapiro AD, Vaughan DE. A null mutation in SERPINE1 protects against biological aging in humans. Sci Adv 2017; 3: eaao1617. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Zhang Y, Coats AJS, Zheng Z, Adamo M, Ambrosio G, Anker SD, Butler J, Xu D, Mao J, Khan MS, Bai L, Mebazaa A, Ponikowski P, Tang Q, Ruschitzka F, Seferovic P, Tschöpe C, Zhang S, Gao C, Zhou S, Senni M, Zhang J, Metra M. Management of heart failure patients with COVID‐19: A joint position paper of the Chinese Heart Failure Association & National Heart Failure Committee and the Heart Failure Association of the European Society of Cardiology. Eur J Heart Fail 2020; 22: 941–956. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1. Study Flow Chart.
Figure S2. D‐dimer Concentrations and All‐Cause Mortality.
Table S1. Characteristics of the Study Participants Categorized by HF Phenotypes and D‐dimer Concentrations.
Table S2. Risk of All‐Cause Death by D‐dimer (After Imputation).
Table S3. Characteristics of the Patients Lost in Follow‐up and Patients Included in Final Analysis.