Summary
Background:
We previously showed that participants in the population-based Atherosclerosis Risk in Communities (ARIC) cohort with elevated blood biomarkers of inflammation or cardiac disease were at increased risk of venous thromboembolism (VTE).
Objective:
We hypothesized now that ARIC participants with larger 6-year increases in three biomarkers – C-reactive protein (CRP), N-terminal pro B-type natriuretic peptide (NT-proBNP), and troponin T – also would have increased subsequent risk of VTE.
Methods:
We measured changes in these biomarkers in 9,844 participants from 1990–1992 to 1996–1998 and then identified VTEs through 2015.
Results:
A greater 6-year rise in NT-proBNP, but not CRP or troponin T, was significantly associated with increased VTE incidence over a median of 17.6 years of follow-up. Adjusted for other VTE risk factors, those whose NT-proBNP rose from <100 to ≥100 pg/mL had a hazard ratio of VTE of 1.44 (95% CI = 1.15–1.80), compared to the reference group with a NT-proBNP concentration at both times below 100 pg/mL. This hazard ratio was slightly stronger [1.66 (95% CI = 1.19–2.31)] during the first 10 years of follow-up but was attenuated [1.24 (95% CI = 0.99–1.56)] with adjustment for prevalent and incident coronary heart disease, heart failure, and atrial fibrillation.
Conclusions:
The two most likely explanations for our result are that (a) rising NT-proBNP reflects increasing subclinical volume overload and potentially increased venous stasis or subclinical PE that went unrecognized over time or (b) rising NT-proBNP is a risk marker for impending cardiac disease that places patients at risk for VTE.
Keywords: Atherosclerosis Risk in Communities Study, C-reactive protein, NT-proBNP, troponin, venous thromboembolism
Introduction
Patients with major risk factors—such as surgery, immobility, cancer—are at substantial risk of venous thromboembolism (VTE), that is, deep vein thrombosis (DVT) or pulmonary embolism (PE) (1). To a lesser degree, patients with inflammatory (2) or cardiac diseases (3, 4), or even just elevations of blood biomarkers of inflammation or cardiac disease, also have an elevated risk of VTE. For example, we showed in the Atherosclerosis Risk in Communities (ARIC) Study that VTE incidence was approximately doubled in those in the top 10–20% of the distributions of baseline concentrations of either the inflammatory marker C-reactive protein (CRP), the cardiac myonecrosis marker troponin T, or the biomarker of myocardial stretch N-terminal pro B-type natriuretic peptide (NT-proBNP), compared to those with low concentrations of these biomarkers (5, 6). The associations of these three biomarkers with VTE are likely not causal, but the identification of patients at increased risk may be relevant to VTE prevention.
Also of scientific interest, but not yet studied, is whether a rise in the blood concentrations of these inflammatory or cardiovascular biomarkers over time further identifies patients at risk of VTE. By analogy, several studies have clearly shown that rises in concentrations of CRP, troponin T, or NT-proBNP are associated with increased risk of non-venous cardiovascular diseases (7–9). We tested whether a greater increase in these three biomarkers over 6 years is associated with increased subsequent 20-year risk of VTE in the ARIC Study.
Methods
Study sample
The ARIC Study enrolled 15,792 men and women aged 45 to 64 years from four U.S. communities in 1987–1989 (10). ARIC conducted five subsequent examinations in 1990–1992 (ARIC Visit 2, 93% response), 1993–1995, 1996–1998 (ARIC Visit 4, 80% response), 2011–2013, and 2016–2017. Between examinations, ARIC interviewers contacted cohort members annually or semi-annually via telephone to help find VTE events. The institutional review committees at each study center approved the methods, and staff obtained informed participant consent. This analysis examined biomarker changes from Visit 2 to Visit 4 and related those changes to VTE incidence after Visit 4.
Biomarkers
Trained field center technicians collected blood samples, immediately processed them, and stored serum and plasma at −80°C until assayed. The biological and analytic variability of these biomarkers and their stability during frozen storage have been published (11–13). In 2010, a research laboratory at the Baylor College of Medicine measured CRP, troponin T, and NT-proBNP in plasma samples from Visit 4. In 2011–13, a research laboratory at the University of Minnesota measured the same analytes in serum samples from Visit 2. An ARIC validation study showed that measured serum and plasma concentrations were nearly identical as determined by Deming regression analyses (14).
The Baylor laboratory measured troponin T and NT-proBNP on a Cobas e411 analyzer using the Elecsys high sensitivity Troponin T assay (15) and the Elecsys proBNP II immunoassay (both from Roche Diagnostics, Indianapolis, IN) (16). The reliability coefficient for blind replicate measurements from a single Visit 4 blood draw was r = 0.98 for troponin T and r = 0.99 for NT-proBNP. The laboratory measured CRP using a nephelometric assay on the Siemens (Dade Behring) BNII analyzer (Dade Behring, Deerfield, Ill). The Visit 4 reliability for CRP was r = 0.99.
The University of Minnesota laboratory measured troponin T and NT-proBNP on a Roche Elecsys 2010 using sandwich immunoassays from Roche Diagnostics. It measured CRP with an immunoturbidimetric assay on the Roche Modular P analyzer. Visit 2 reliability coefficients were similar to those for the Baylor laboratory.
Numerous troponin T values were below the lower detection limit of 5 ng/L (66% in Visit 2 and 48% in Visit 4). A few NT-proBNP values below the lower detection limit of 2.5 pg/mL (3% in both Visits 2 and 4).
Other variables
ARIC measured Visit 4 body mass index (BMI), cigarette use, and postmenopausal hormone replacement therapy by standard methods. We estimated Visit 4 glomerular filtration rate (eGFR) using the Chronic Kidney Disease Collaboration algorithm involving creatinine and cystatin C (17). We defined prevalent (at Visit 4) and incident (after Visit 4) coronary heart disease, heart failure, and atrial fibrillation, as previously described (6, 18).
VTE classification
ARIC followed participants through 2015 to identify possible hospitalized VTE events (DVT or PE). Two physicians reviewed the discharge summaries, imaging and consult reports, and other information and classified VTEs via standardized criteria (19). The physicians also subclassified VTEs as unprovoked or provoked (associated with recent hospitalization for major trauma, cancer, surgery, or with marked immobility) and as PE (with or without DVT) and DVT alone. Because upper extremity DVTs were mostly due to indwelling catheters, we excluded them from the VTE definition for all analyses.
Analysis
As shown in Figure 1, from the 11,656 men and women who came to ARIC Visit 4, we excluded those with a history of VTE, current anticoagulant use, race/ethnicity other than black or white (due to small numbers), a few without follow-up after Visit 4, and those with any of the three biomarkers missing to yield a total of 9,854 participants for analysis. We also excluded the participants with the 5 largest positive and 5 largest negative changes in each of the three biomarkers, because these values were obvious outliers. This left 9,844 participants for the analysis associating change in each biomarker with incident VTE.
Fig. 1.
Exclusions to form the analysis cohort.
Our aim was to associate change in each biomarker between Visits 2 and 4 with VTE incidence. However, because many values for troponin T were below the limit of detection, we could not readily calculate absolute change in troponin T. In addition, published ARIC papers had established a convention of examining change using categories with cutpoints of <5 ng/L, 5–13 ng/L, and ≥14 ng/L for troponin T and a clinically-based cutpoint of 3 mg/L for CRP (7, 9). We chose a cutpoint of 100 pg/mL for NT-proBNP. We chose the 100 pg/mL cutoff for NT-proBNP change a priori (i.e., before exploring associations with VTE). We picked 100 pg/mL because (a) our previous paper showed VTE risk rose above baseline NT-proBNP > 100 pg/mL (6) and (b) 100 pg/mL was the upper tertile cutpoint for NT-proBNP change and therefore yielded sufficient numbers in NT-proBNP change categories. The final categories for Visits 2 to 4 change are shown in the tables.
We used SAS software for analysis. To describe the relation of each biomarker to participant characteristics, we computed means and standard deviations for continuous characteristics and percentages for categorical characteristics within the biomarker change categories. We used Cox proportional hazards models to compute hazard ratios for incident VTE according to biomarker change categories. Model 1 adjusted for age, sex, and race (white, African American). Model 2 adjusted for age, race, BMI, eGFR, cigarette smoking (never, former, current), and postmenopausal hormone use (woman currently using hormones, women not using, and men). We ran Model 2 for the full follow-up and then restricted to the first 10 years of follow-up. In addition to examining total VTE, we ran Model 2 using full follow-up for provoked and unprovoked VTE separately and for isolated deep vein thrombosis and pulmonary embolism separately; we censored follow-up time for those who had the complementary outcomes at their incidence dates. Then, we ran Model 3 for total VTE, further adjusting, as a time dependent covariable, for coronary heart disease, heart failure, or atrial fibrillation at baseline or incident during follow-up; the purpose was to determine whether the biomarker changes were associated with VTE independent of clinically recognized cardiac disease. In addition, we explored the relation between the change in NT-proBNP, as a continuous variable, and incident VTE using a restricted cubic spline, adjusted for the variables in Model 2. Spline knots were placed at recommended percentiles of 5, 27.5, 50, 72.5 and 95 (20). When using a restricted cubic spline, one obtains a continuous smooth function that is linear before the first knot, a piecewise cubic polynomial between adjacent knots, and linear again after the last knot.
Results
As shown in Table 1 for NT-proBNP and in Supplemental Tables 1 and 2 for CRP and troponin T, a sizable proportion of participants had rises in biomarker concentrations over the six years; fewer had biomarker falls. Although the relations with sex and race varied among the three biomarkers, most VTE risk factors at Visit 4 were higher in those with the highest biomarkers concentrations and in those who experienced biomarker increases between Visits 2 and 4.
Table 1.
Participant characteristics according to categories of 6-year change in NT-proBNP concentration between Visit 2 (1990–1992) and Visit 4 (1996–1998), ARIC
| Baseline NT-proBNP Level (Visit 2) | <100 pg/mL |
≥100 pg/mL |
||
|---|---|---|---|---|
| 6-year follow-up NT-proBNP Level (Visit 4) | <100 pg/mL | ≥100 pg/mL | <100 pg/mL | ≥100 pg/mL |
| N | 6024 | 1821 | 466 | 1533 |
| Demographics | ||||
| Age in years, Mean (SD) | 61.7 (5.4) | 63.9 (5.6) | 62.3 (5.5) | 64.9 (5.7) |
| Men, N (%) | 3011 (50.0) | 721 (39.6) | 107 (23.0) | 422 (27.5) |
| African American, N (%) | 1506 (25.0) | 311 (17.1) | 117 (25.1) | 214 (14.0) |
| Body Mass Index, kg/m2, Mean (SD) | 29.1 (5.4) | 28.4 (5.6) | 28.7 (5.9) | 27.7 (5.9) |
| Medical Information | ||||
| Current Smoker, N (%) | 826 (13.8) | 276 (15.2) | 76 (16.4) | 249 (16.3) |
| Former Smoker, N (%) | 2614 (43.7) | 796 (43.9) | 188 (40.5) | 635 (41.6) |
| Current Hormone Use among women, N (%) | 771 (25.6) | 332 (30.2) | 85 (23.7) | 323 (29.1) |
| Laboratory Data | ||||
| eGFR category, N (%) | ||||
| ≥90 mL/min per 1.73 m2 | 2971 (49.3) | 617 (33.9) | 196 (42.1) | 455 (29.7) |
| 60–89 mL/min per 1.73 m2 | 2817 (46.8) | 1039 (57.1) | 240 (51.5) | 881 (57.5) |
| <60 mL/min per 1.73 m2 | 236 (3.9) | 165 (9.1) | 30 (6.4) | 197 (12.9) |
eGFR = estimated glomerular filtration rate
Change in NT-proBNP and VTE incidence
Over a median of 17.6 years (max 19.9), we identified 527 incident VTEs. Consistent with our hypothesis, those whose NT-proBNP rose from <100 to ≥100 pg/mL had 1.4 fold greater risk of VTE (p<0.05), compared to the reference group with a NT-proBNP concentration at both Visits 2 and 4 below 100 pg/mL (Model 2, Table 2). Those whose NT-proBNP was consistently ≥100 pg/mL had over a 1.6 fold greater risk of VTE (p<0.05). These hazard ratios were slightly stronger during the first 10 years of follow-up (Table 2). The few people whose NT-proBNP concentration dropped from ≥100 to <100 pg/mL had no increase in VTE risk compared to the reference group. These patterns were largely similar for unprovoked and provoked VTE as well as for DVT and PE subcategories, although confidence intervals for hazard ratios often overlapped 1 because of smaller numbers.
Table 2.
Incidence rate and hazard ratio (95% confidence interval) for incident venous thromboembolism (VTE) by 6-year change in NT-proBNP concentration between Visit 2 (1990–1992) and Visit 4 (1996–1998), ARIC
| Baseline NT-proBNP Level (Visit 2) | <100 pg/mL |
≥100 pg/mL |
||
|---|---|---|---|---|
| 6-year follow-up NT-proBNP Level (Visit 4) | <100 pg/mL | ≥100 pg/mL | <100 pg/mL | ≥100 pg/mL |
| N | 6024 | 1821 | 466 | 1533 |
| Total VTE Events (n = 527) | ||||
| N Events | 296 | 111 | 24 | 96 |
| Incidence Rate* | 3.0 | 4.1 | 3.3 | 4.5 |
| Hazard Ratio (95% CI) | ||||
| Model 1 | 1 | 1.43 (1.14–1.78) | 1.12 (0.74–1.71) | 1.62 (1.27–2.07) |
| Model 2 | 1 | 1.44 (1.15–1.80) | 1.13 (0.74–1.72) | 1.68 (1.31–2.14) |
| Model 3 | 1 | 1.24 (0.99–1.56) | 1.07 (0.70–1.62) | 1.34 (1.04–1.72) |
| Total VTE Events in first 10 years (n = 230) | ||||
| N Events | 113 | 54 | 10 | 53 |
| Model 2 Hazard Ratio (95% CI) | 1 | 1.66 (1.19–2.31) | 1.22 (0.64–2.34) | 2.11 (1.49–2.99) |
| Unprovoked VTE Events (n = 206) | ||||
| N Events | 115 | 42 | 13 | 36 |
| Model 2 Hazard Ratio (95% CI) | 1 | 1.36 (0.95–1.96) | 1.61 (0.90–2.87) | 1.56 (1.05–2.32) |
| Provoked VTE Events (n = 321) | ||||
| N Events | 181 | 69 | 11 | 60 |
| Model 2 Hazard Ratio (95% CI) | 1 | 1.49 (1.12–1.99) | 0.84 (0.45–1.55) | 1.75 (1.28–2.39) |
| DVT alone (n = 226) | ||||
| N Events | 122 | 52 | 11 | 41 |
| Model 2 Hazard Ratio (95% CI) | 1 | 1.69 (1.21–2.36) | 1.34 (0.72–2.49) | 1.82 (1.25–2.66) |
| PE (n = 301) | ||||
| N Events | 174 | 59 | 13 | 55 |
| Model 2 Hazard Ratio (95% CI) | 1 | 1.27 (0.94–1.73) | 1.00 (0.57–1.76) | 1.57 (1.14–2.17) |
per 1000 person-years
Model 1: adjusted for age, sex and race
Model 2: adjusted for age, race, BMI, eGFR, cigarette smoking and sex/hormone use
Model 3: adjusted in addition for time dependent prevalent and incident coronary heart disease, heart failure, and atrial fibrillation.
DVT = deep vein thrombosis; PE = pulmonary embolism
To further examine the shape of the association between NT-proBNP and VTE we used a restricted cubic spline model and Model 2 adjustments. Within the majority of the distribution of NT-proBNP change, there was a strong positive association with VTE incidence (Figure 2).
Fig. 2.
Association of 6-year change in NT-proBNP concentration and incident venous thromboembolism, ARIC.
Footnote: For ease of plotting, we show NT-proBNP change values between −200 and +400 pg/mL. N=71 values were lower and 246 were higher.
Because NT-proBNP may be elevated by cardiac disease, we reran Model 2 while also adjusting for time-dependent prevalent or incident coronary heart disease, heart failure, or atrial fibrillation (Model 3). The cumulative incidence percentages were approximately 16% for coronary heart disease, 20% for heart failure, and 18% for atrial fibrillation. After this adjustment, the hazard ratio associating NT-proBNP rise with total VTE was 1.24 (95% confidence interval 0.99–1.56) (Table 2, Model 3).
Change in CRP and VTE incidence
Those whose CRP was elevated ≥3 mg/L at both Visits 2 and 4 had 25% greater risk in Model 2 compared with the reference group whose CRP was consistently <3 mg/L (Table 3). This association was most apparent for unprovoked VTE and for DVT rather than PE. In contrast, and contrary to our hypothesis, those whose CRP rose from <3 to ≥3 mg/L during the 6 years between visits did not have increased risk of VTE. Likewise, those whose CRP concentration fell between visits did not have increased risk of VTE compared with the reference group. Adjustment for coronary heart disease or heart failure (Model 3) eliminated all associations of CRP with total VTE. A restricted cubic spline analysis (not shown) also showed continuous CRP change not materially associated with VTE incidence.
Table 3.
Incidence rate and hazard ratio (95% confidence interval) for incident venous thromboembolism (VTE) by 6-year change in CRP concentration between Visit 2 (1990–1992) and Visit 4 (1996–1998), ARIC
| Baseline CRP Level (Visit 2) | <3 mg/L (Low/Moderate) |
≥3 mg/L (Elevated) |
||
|---|---|---|---|---|
| 6-year follow-up CRP Level (Visit 4) | <3 mg/L | ≥3 mg/L | <3 mg/L | ≥3 mg/L |
| N | 4766 | 1466 | 827 | 2785 |
| Total VTE Events (n = 527) | ||||
| N Events | 217 | 74 | 48 | 188 |
| Incidence Rate* | 2.8 | 3.3 | 3.9 | 4.5 |
| Hazard Ratio (95% CI) | ||||
| Model 1 | 1 | 1.19 (0.92–1.56) | 1.33 (0.97–1.82) | 1.56 (1.28–1.92) |
| Model 2 | 1 | 1.08 (0.83–1.42) | 1.23 (0.90–1.69) | 1.25 (1.00–1.56) |
| Model 3 | 1 | 1.02 (0.78–1.34) | 1.18 (0.86–1.62) | 1.15 (0.92–1.44) |
| Total VTE Events in first 10 years (n = 230) | ||||
| N Events | 86 | 37 | 23 | 84 |
| Model 2 Hazard Ratio (95% CI) | 1 | 1.29 (0.87–1.90) | 1.36 (0.86–2.17) | 1.25 (0.89–1.76) |
| Unprovoked VTE Events (n = 206) | ||||
| N Events | 81 | 33 | 14 | 78 |
| Model 2 Hazard Ratio (95% CI) | 1 | 1.32 (0.87–2.00) | 0.97 (0.55–1.71) | 1.42 (1.00–2.03) |
| Provoked VTE Events (n = 321) | ||||
| N Events | 136 | 41 | 34 | 110 |
| Model 2 Hazard Ratio (95% CI) | 1 | 0.95 (0.66–1.35) | 1.39 (0.95–2.03) | 1.14 (0.86–1.53) |
| DVT alone (n = 226) | ||||
| N Events | 80 | 35 | 18 | 93 |
| Model 2 Hazard Ratio (95% CI) | 1 | 1.45 (0.96–2.17) | 1.28 (0.76–2.14) | 1.78 (1.27–2.50) |
| PE (n = 301) | ||||
| N Events | 137 | 39 | 30 | 95 |
| Model 2 Hazard Ratio (95% CI) | 1 | 0.87 (0.61–1.25) | 1.21 (0.81–1.80) | 0.95 (0.70–1.28) |
per 1000 person-years
Model 1: adjusted for age, sex and race
Model 2: adjusted for age, race, BMI, eGFR, cigarette smoking and sex/hormone use
Model 3: adjusted in addition for time dependent prevalent and incident coronary heart disease, heart failure, and atrial fibrillation.
DVT= deep vein thrombosis; PE = pulmonary embolism
Change in troponin T and VTE incidence
Consistent with our hypothesis, people who had a 6-year increase in troponin T from 5–13 ng/L to ≥14 ng/L had a 1.53-fold greater risk of VTE (p<0.05), compared with a reference group of <5 ng/L at both visits (Table 4). This seemed most apparent for DVT compared with PE and for provoked VTE. Contrary to our hypothesis, those whose troponin T increased from <5 to ≥ 5 ng/L had no elevation of VTE risk. The relatively few people who had troponin T ≥14 ng/L at Visit 2 had increased VTE risk regardless of their troponin T concentration at Visit 4, but the 95% confidence intervals for these hazard ratios overlapped 1. No hazard ratio for troponin T was significant after adjustment for cardiac disease in Model 3.
Table 4.
Incidence rate and hazard ratio (95% confidence interval) for incident venous thromboembolism (VTE) by 6-year change in Troponin T concentration between Visit 2 (1990–1992) and Visit 4 (1996–1998), ARIC
| Baseline Troponin T Level (Visit 2) | <5 ng/L |
5–13 ng/L |
≥14 ng/L |
||||
|---|---|---|---|---|---|---|---|
| 6-year follow-up Troponin T Level (Visit 4) | <5 ng/L | ≥5 ng/L | <5 ng/L | 5–13 ng/L | ≥14 ng/L | <14 ng/L | ≥14 ng/L |
| N | 4127 | 2411 | 632 | 1881 | 472 | 73 | 248 |
| Total VTE Events (n = 527) | |||||||
| N Events | 206 | 117 | 28 | 122 | 29 | 7 | 18 |
| Incidence Rate* | 3.0 | 3.2 | 2.7 | 4.3 | 5.4 | 6.5 | 6.4 |
| Hazard Ratio (95% CI) | |||||||
| Model 1 | 1 | 1.03 (0.81–1.30) | 0.85 (0.57–1.26) | 1.29 (1.01–1.65) | 1.73 (1.15–2.60) | 2.0 (0.94–4.28) | 1.86 (1.13–3.08) |
| Model 2 | 1 | 0.97 (0.76–1.23) | 0.84 (0.57–1.25) | 1.19 (0.93–1.53) | 1.53 (1.01–2.32) | 1.81 (0.85–3.88) | 1.61 (0.96–2.68) |
| Model 3 | 1 | 0.90 (0.71–1.14) | 0.80 (0.54–1.19) | 1.07 (0.83–1.37) | 1.21 (0.80–1.84) | 1.50 (0.70–3.23) | 1.26 (0.75–2.12) |
| Total VTE Events in first 10 years (n = 230) | |||||||
| N Events | 79 | 48 | 13 | 62 | 17 | 2 | 9 |
| Model 2 Hazard Ratio (95% CI) | 1 | 0.90 (0.62–1.32) | 0.97 (0.54–1.75) | 1.34 (0.92–1.95) | 1.53 (0.86–2.71) | 1.12 (0.27–4.58) | 1.31 (0.62–2.75) |
| Unprovoked VTE Events (n = 206) | |||||||
| N Events | 80 | 45 | 15 | 52 | 6 | 4 | 4 |
| Model 2 Hazard Ratio (95% CI) | 1 | 0.88 (0.60–1.29) | 1.11 (0.64–1.93) | 1.15 (0.78–1.70) | 0.67 (0.28–1.60) | 2.36 (0.85–6.54) | 0.76 (0.27–2.16) |
| Provoked VTE Events (n = 321) | |||||||
| N Events | 126 | 72 | 13 | 70 | 23 | 3 | 14 |
| Model 2 Hazard Ratio (95% CI) | 1 | 1.03 (0.76–1.40) | 0.66 (0.37–1.17) | 1.21 (0.87–1.68) | 2.23 (1.37–3.63) | 1.37 (0.43–4.34) | 2.30 (1.27–4.18) |
| DVT alone (n = 226) | |||||||
| N Events | 75 | 59 | 10 | 52 | 16 | 4 | 10 |
| Model 2 Hazard Ratio (95% CI) | 1 | 1.29 (0.90–1.85) | 0.80 (0.41–1.56) | 1.31 (0.88–1.94) | 1.99 (1.10–3.58) | 2.63 (0.95–7.29) | 1.99 (0.97–4.05) |
| PE (n = 301) | |||||||
| N Events | 131 | 58 | 18 | 70 | 13 | 3 | 8 |
| Model 2 Hazard Ratio (95% CI) | 1 | 0.78 (0.57–1.08) | 0.87 (0.53–1.43) | 1.13 (0.81–1.56) | 1.21 (0.66–2.22) | 1.29 (0.41–4.10) | 1.32 (0.63–2.79) |
per 1000 person-years
Model 1: adjusted for age, sex and race
Model 2: adjusted for age, race, BMI, eGFR, cigarette smoking and sex/hormone use
Model 3: adjusted in addition for time dependent prevalent and incident coronary heart disease, heart failure, and atrial fibrillation.
DVT = deep vein thrombosis; PE = pulmonary embolism
Discussion
We reported previously that baseline concentrations of NT-proBNP, CRP, and troponin T in the general population were associated positively with incidence of VTE (5, 6). In the present analysis, we found that a 6-year rise in only NT-proBNP, and not the other two biomarkers, was associated with subsequent VTE incidence over a median of 17.6 years and even stronger for the first 10 years of follow-up. This positive association was independent of several VTE risk factors, generally consistent across VTE subtypes, but was attenuated by adjustment for the presence or development of non-venous cardiac disease. The longitudinal nature of the analysis strengthens confidence that the association may be real.
We are uncertain why single measures of CRP and troponin T were associated previously with incident VTE (5, 6), yet changes in these biomarkers were not. One possibility is that the original findings were spurious, due to some unrecognized confounding variable. Notably, in the original reports (5, 6), baseline NT-proBNP appeared more strongly related to VTE in ARIC than did troponin T or CRP. A second possibility is that, compared with NT-proBNP, within-person variability may be greater for CRP and troponin T, and thus our measurements of changes in those two biomarkers were suboptimal. Overall, the lack of longitudinal associations between CRP and troponin T changes and VTE incidence suggest they, or the fundamental processes they represent (low-level inflammation and cardiac injury), are not risk factors for VTE.
To our knowledge, no other study has documented in the general population that a greater NT-proBNP concentration or an increase in NT-proBNP is associated with increased VTE incidence. Several clinical studies have however documented that acute PE patients with elevated NT-proBNP or BNP concentrations have increased rates of VTE recurrence or death, reflecting the fact that these are biomarkers of right ventricular dysfunction (21–23). However, right ventricular dysfunction is relatively rare in the general population and few ARIC participants had NT-proBNP concentrations elevated to levels seen in acute PE with right ventricular dysfunction. Clinical studies have also documented that not only the presence of heart failure (3), but also elevated NT-proBNP concentrations in hospitalized heart failure patients (24), are associated with increased risk of VTE. Likewise, atrial fibrillation is associated positively with NT-proBNP and with VTE (18). A significant proportion of ARIC participants developed heart failure, coronary heart disease, or atrial fibrillation, and our adjustment for cardiac disease (Model 3) did weaken the association between NT-proBNP rise and VTE incidence. One may ask why these hypothesized pathophysiological pathways would apply to NT-proBNP and not CRP and troponin T. Although CRP and troponin T are also positively associated with atrial fibrillation, NT-proBNP is more strongly associated (18, 25, 26). Although troponin T is associated with heart failure comparably to NT-proBNP, CRP is not an independent predictor of heart failure in ARIC (27, 28). In addition, in ARIC, baseline NT-proBNP was a slightly stronger predictor of VTE than was troponin T or CRP (5, 6). Drawbacks of our study warrant consideration. Firstly, the laboratory precision for the three biomarkers was high, but we measured Visit 2 and Visit 4 biomarkers in separate batches and in samples frozen long-term at −80°C. Although previous evidence suggests these biomarkers are stable in frozen samples (11–13), lab drift or sample deterioration, if not uniform across the samples, may have affected our results in unknown ways. Secondly, biomarkers are subject to within-person variability (11, 29, 30), either biological or random. Random variability may have obscured or weakened the estimated associations between biomarker changes and VTE. Thirdly, although we adjusted for VTE risk factors identified previously in ARIC, we may have overlooked some important confounding factors. Hemostatic factors related to VTE, such as factor VIII, von Willebrand factor, or D-dimer, may be correlated with the three biomarkers studied, but ARIC unfortunately did not measure hemostatic markers at visits 2 or 4. Genetic variants for VTE are unlikely to be related to the biomarkers and should not confound our results. Major precipitants for VTE, such as cancer, surgery, trauma, and immobility, are difficult to document in cohort studies over decades of follow-up. We addressed these potential confounders by stratifying VTE cases into provoked or unprovoked, rather than by statistical adjustment for them. A final drawback is that we identified only VTE patients who were hospitalized, but unpublished ARIC data suggest that the vast majority of participants with first VTEs were hospitalized.
In conclusion, a 6-year rise in NT-proBNP was associated with increased subsequent incidence of VTE in the general population. The two most likely explanations for our result are that (a) rising NT-proBNP reflects increasing subclinical volume overload and potentially increased venous stasis or subclinical PE that went unrecognized over time or (b) rising NT-proBNP is a risk marker for impending cardiac disease that places patients at risk for VTE.
Supplementary Material
Essentials.
Inflammatory and cardiac diseases are associated with increased venous thromboembolism (VTE) risk.
Our prospective study assessed rise in inflammatory or cardiac biomarkers and VTE risk.
A greater 6-year rise in N-terminal natriuretic peptide is associated with increased VTE incidence.
Volume overload or impending cardiac disease may contribute to VTE occurrence.
Acknowledgements
The authors thank the staff and participants of the ARIC Study for their important contributions. The National Heart, Lung, and Blood Institute provided support for venous thromboembolism identification via R01 HL059367 and for the Atherosclerosis Risk in Communities Study via contracts HHSN268201700001I, HHSN268201700002I, HHSN268201700003I, HHSN268201700004I, HHSN268201700005I. The National Institute of Diabetes and Digestive and Kidney Disease provided funding to measure plasma biomarkers via R01 DK089174 and R01 DK076770. Roche Diagnostics, Inc., also provided funding for laboratory reagents.
Footnotes
Addendum
A. R. Folsom, P. L. Lutsey, and M. Cushman designed the research. A. R. Folsom, P. L. Lutsey, S. R. Heckbert, R. C. Hoogeveen, M. Cushman, and C. M. Ballantyne conducted research. K. Poudel, S. Basu, and P. L. Lutsey analyzed data or performed statistical analysis. A. R. Folsom drafted the paper. All authors made critical comments on the paper. All authors read and approved the final manuscript.
Disclosure of Conflict of Interests
C. M. Ballantyne and Hoogeveen have received grant support from Roche Diagnostics, Inc. and are co-inventors on a patent filed on their behalf by Roche Diagnostics, Inc. and Baylor College of Medicine (patent no. 61/721475).
References
- 1.Lijfering WM, Rosendaal FR, Cannegieter SC. Risk factors for venous thrombosis – current understanding from an epidemiological point of view. Br J Haematol 2010; 149: 824–33. [DOI] [PubMed] [Google Scholar]
- 2.Lee JJ, Pope JE. A meta-analysis of the risk of venous thromboembolism in inflammatory rheumatic diseases. Arthritis Res Thera 2014; 16: 435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Tang L, Wu YY, Lip GY, Yin P, Hu Y. Heart failure and risk of venous thromboembolism: a systematic review and meta-analysis. Lancet Haematol 2016; 3: e30–44. [DOI] [PubMed] [Google Scholar]
- 4.Lijfering WM, Flinterman LE, Vandenbroucke JP, Rosendaal FR, Cannnegieter SC. Relationship between venous and arterial thrombosis: a review of the literature from a causal perspective. Semin Thromb Hemost 2011; 37: 885–96. [DOI] [PubMed] [Google Scholar]
- 5.Folsom AR, Lutsey PL, Astor BC, Cushman M. C-reactive protein and venous thromboembolism. A prospective investigation in the ARIC cohort. Thromb Haemost 2009; 102: 615–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Folsom AR, Lutsey PL, Nambi V, DeFilippi CR, Heckbert SR, Cushman M, Ballantyne CM. Troponin T, NT-pro BNP, and venous thromboembolism: The Longitudinal Investigation of Thromboembolism Etiology (LITE). Vasc Med 2014; 19: 33–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Parrinello CM, Lutsey PL, Ballantyne CM, Folsom AR, Pankow JS, Selvin E. Six-year change in high-sensitivity C-reactive protein and risk of diabetes, cardiovascular disease, and mortality. Am Heart J 2015; 170: 380–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Mishra RK, Judson G, Christenson RH, DeFilippi C, Wu AHB, Whooley MA. The association of five-year changes in the levels of N-terminal fragment of the prohormone brain-type natriuretic peptide (NT-proBNP) with subsequent heart failure and death in patients with stable coronary artery disease: The Heart and Soul Study. Cardiology 2017; 137: 201–6. [DOI] [PubMed] [Google Scholar]
- 9.McEvoy JW, Chen Y, Ndumele CE, Solomon SD, Nambi V, Ballantyne CM, Blumenthal RS, Coresh J, Selvin E. Six-year change in high-sensitivity cardiac troponin T and risk of subsequent coronary heart disease, heart failure, and death. AMA Cardiol 2016; 1: 519–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.The ARIC Investigators (Folsom AR, Principal Investigator). The Atherosclerosis Risk in Communities (ARIC) Study: Design and objectives. Am J Epidemiol 1989; 129: 687–702. [PubMed] [Google Scholar]
- 11.Agarwal SK, Avery CL, Ballantyne CM, Catellier D, Nambi V, Saunders J, Sharrett AR, Coresh J, Heiss G, Hoogeveen RC. Sources of variability in measurements of cardiac troponin T in a community-based sample: The Atherosclerosis Risk in Communities Study. Clin Chem 2011; 57: 891–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Nowatzke WL, Cole TG. Stability of N-terminal pro-brain natriuretic peptide after storage frozen for one year and after multiple freeze-thaw cycles. Clin Chem 2003; 49: 1560–2. [DOI] [PubMed] [Google Scholar]
- 13.Vasile VD, Saenger AK, Kroning JM, Jaffe AS. Biological and analytical variability of a novel high-sensitivity cardiac troponin T assay. Clin Chem 2010; 56: 1086–90. [DOI] [PubMed] [Google Scholar]
- 14.Parrinello CM, Grams ME, Couper D, Ballantyne CM, Hoogeveen RC, Eckfeldt JH, Selvin E, Coresh J. Recalibration of blood analytes over 25 years in the Atherosclerosis Risk in Communities Study: Impact of recalibration on chronic kidney disease prevalence and incidence. Clin Chem 2015; 61: 938–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Saunders JT, Nambi V, de Lemos JA, Chambless LE, Virani SS, Boerwinkle E, Hoogeveen RC, Liu X, Astor BC, Mosley TH, Folsom AR, Heiss G, Coresh J, Ballantyne CM. Cardiac troponin T measured by a highly sensitive assay predicts coronary heart disease, heart failure, and mortality in the Atherosclerosis Risk in Communities Study. Circulation 2011; 123: 1367–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Agarwal SK, Chambless LE, Ballantyne CM, Astor B, Bertoni AG, Chang PP, Folsom AR, He M, Hoogeveen RC, Ni H, Quibrera PM, Rosamond WD, Russell SD, Shahar E, Heiss G. Prediction of incident heart failure in general practice: the Atherosclerosis Risk in Communities (ARIC) Study. Circ Heart Fail 2012; 5: 422–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Inker LA, Schmid CH, Tighiouart H, Eckfeldt JH, Feldman HI, Green T, Kusek JW, Manzi J, Van Lente F, Zhang YL, Coresh J, Levey AS, for the CKD-EPI Investigators. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med 2012; 367: 20–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lutsey PL, Norby FL, Alonso A, Cushman M, Chen LY, Michos ED, Folsom AR. Atrial fibrillation and venous thromboembolism: Evidence of bidirectionality in the Atherosclerosis Risk in Communities Study. J Thromb Haemost 2018. February 12. doi: 10.1111/jth.13974. [Epub ahead of print]. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Cushman M, Tsai AW, White RH, Heckbert SR, Rosamond WD, Enright P, Folsom AR. Deep vein thrombosis and pulmonary embolism in two cohorts: the Longitudinal Investigation of Thromboembolism Etiology. Am J Med 2004; 117: 19–25. [DOI] [PubMed] [Google Scholar]
- 20.Harrell FE Jr. Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis. New York, Springer, 2001. [Google Scholar]
- 21.Coutance G, Le Page O, Lo T, Hamon M. Prognostic value of brain natriuretic peptide in acute pulmonary embolism. Crit Care 2008; 12: R109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Wang Y, Liu Z-h, Zhang H-l, Luo Q, Zhao Z-h, Zhao Q. Association of elevated NTproBNP with recurrent thromboembolic events after acute pulmonary embolism. Thromb Res 2012; 129: 688–92. [DOI] [PubMed] [Google Scholar]
- 23.Söhne M, Ten Wolde M, Boomsma F, Reitsma JB, Douketis JD, Büller HR. Brain natriuretic peptide in hemodynamically stable acute pulmonary embolism. J Thromb Haemost 2006; 4: 552–6. [DOI] [PubMed] [Google Scholar]
- 24.Mebazaa A, Spiro TE, Büller HR, Haskell L, Hu D, Hull R, Merli G, Schellong SW, Spyropoulos AC, Tapson VF, De Sanctis Y, Cohen AT. Predicting the risk of venous thromboembolism in patients hospitalized with heart failure. Circulation 2014; 130: 410–8. [DOI] [PubMed] [Google Scholar]
- 25.Sinner MF, Stepas KA, Moser CB, Krijthe BP, Aspelund T, Sotoodehnia N, Fontes JD, Janssens AC, Kronmal RA, Magnani JW, Witteman JC, Chamberlain AM, Lubitz SA, Schnabel RB, Vasan RS, Wang TJ, Agarwal SK, McManus DD, Franco OH, Yin X, Larson MG, Burke GL, Launer LJ, Hofman A, Levy D, Gottdiener JS, Kääb S, Couper D, Harris TB, Astor BC, Ballantyne CM, Hoogeveen RC, Arai AE, Soliman EZ, Ellinor PT, Stricker BH, Gudnason V, Heckbert SR, Pencina MJ, Benjamin EJ, Alonso A. B-type natriuretic peptide and C-reactive protein in the prediction of atrial fibrillation risk: the CHARGE-AF Consortium of community-based cohort studies. Europace 2014; 16: 1426–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Filion KB, Agarwal SK, Ballantyne CM, Ebert M, Hoogeveen RC, Huxley RR, Loehr LR, Nambi V, Soliman EZ, Alonso A. High-sensitivity cardiac troponin T and the risk of incident atrial fibrillation: the Atherosclerosis Risk in Communities (ARIC) study. Am Heart J 2015; 169: 31–8. e3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Nambi V, Liu X, Chambless LE, de Lemos JA, Virani SS, Agarwal S, Boerwinkle E, Hoogeveen RC, Aguilar D, Astor BC, Srinivas PR, Deswal A, Mosley TH, Coresh J, Folsom AR, Heiss G, Ballantyne CM. Troponin T and N-terminal pro-B-type natriuretic peptide: a biomarker approach to predict heart failure risk -- the Atherosclerosis Risk in Communities Study. Clin Chem 2013; 59: 1802–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Agarwal SK, Chambless LE, Ballantyne CM, Astor B, Bertoni AG, Chang PP, Folsom AR, He M, Hoogeveen RC, Ni H, Quibrera PM, Rosamond WD, Russell SD, Shahar E, Heiss G. Prediction of incident heart failure in general practice: the Atherosclerosis Risk in Communities (ARIC) Study. Circ Heart Fail 2012; 5 : 422–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Bower JK, Lazo M, Juraschek SP, Selvin E. Within-person variability in high-sensitivity C-reactive protein. Arch Intern Med 2012; 172: 1519–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Fradley MG, Larson MG, Cheng S, McCabe E, Coglianese E, Shah RV, Levy D, Vasan RS, Wang TJ. Reference limits for N-terminal-pro-B-type natriuretic peptide in healthy individuals (from the Framingham Heart Study). Am J Cardiol 2011; 108: 1341–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.