Sex- and age-related differences in the long-term cardiovascular predictive value of circulating biomarkers and arterial stiffness assessed by pulse wave velocity in apparently healthy individuals from the MONI10 cohort

Background
The Systematic COronary Risk Evaluation (SCORE) and the Framingham Risk Score (FRS) are traditionally used to estimate an individual’s 10-year risk of cardiovascular mortality (CVM), or cardiovascular disease (CVD), respectively, based on clinical characteristics and laboratory tests in apparently healthy individuals. However, these risk assessment tools are limited in their discriminative capability. Therefore, several additional circulating and non-circulating biomarkers have been investigated for refining risk stratification. These biomarkers include, among others, high-sensitivity C-reactive protein (hs-CRP), soluble urokinase plasminogen activator receptor (suPAR), N-terminal pro-brain natriuretic peptide (NT-proBNP), and carotid to femoral pulse wave velocity (cfPWV). When stratified by age and sex, the exact prognostic value of hs-CRP, NTproBNP, suPAR, and cfPWV, used in addition to traditional risk factors, remains unclear.

Aims
1) We hypothesize that suPAR and NT-proBNP, but not hs-CRP, carry incremental prognostic value in predicting CVD, CVM, or all-cause mortality (ACM) in apparently healthy subjects beyond traditional risk factors. 2) We hypothesize that hs-CRP primarily carries prognostic information in addition to traditional cardiovascular risk factors in young subjects while suPAR and NT-proBNP primarily carry prognostic information in addition to traditional cardiovascular risk factors in older subjects. 3) We hypothesize that NT-proBNP is associated with cfPWV, and that cfPWV does not add further prognostic information to NT-proBNP and traditional cardiovascular risk factors in healthy individuals.

Methods
Study Population
From 1982 to 1984, a randomized sample of 4807 men and women at 30, 40, 50, or 60 years of age that resided in the Greater Copenhagen Area were invited to participate in the MONItoring of trends and determinants in CArdiovascular disease (MONICA) project. Out of the original numbers asked, 3971 (82.6%) individuals chose to participate. When asked to participate in a follow-up examination ten years later (MONI10), 2656 subjects, out of the 3785 subjects contacted, participated (70.2%). Subjects with a history of CVD or diabetes (n = 142), or those who received anti-diabetic, anti-hypertensive, or lipid-lowering therapy (n = 312) were excluded. Then, participants who had missing data for any of the circulating biomarkers (n = 251) were excluded. This led us to a sample size of 1951 subjects for the first two studies. In our final study, we excluded subjects with CVD, diabetes, or those receiving treatment for lipid-lowering or antihypertensive treatment (n = 494) along with participants missing laboratory data, blood pressure measurements, or cfPWV evaluations (n = 290). This left 1872 eligible participants.

Baseline Evaluation
A questionnaire containing a complete medical history (i.e., medical conditions, medications, lifestyle, and smoking habits) was collected from each subject. The subjects' heights and weights were used to determine their body mass index (BMI). Mean systolic and diastolic blood pressures were calculated from two blood pressure measurements at heart level on the arm using a random zero mercury sphygmomanometer after sitting for 5 minutes. Participants' heart rates were counted over 15 seconds. After an overnight fast, blood samples were obtained to measure fasting plasma glucose, serum total cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycerides. Finally, low-density lipoprotein (LDL) cholesterol was calculated.

Cardiovascular Biomarkers
 Shortly after blood sample collection, samples were frozen at -20 Cº. In July 2003, they were thawed for analysis. Values of plasma suPAR and serum hs-CRP and NT-proBNP were measured between 0.1-20 mg/l, 5.1-34,9 pg/ml, and 0.6-22.0 ng/ml, respectively. CfPWV was calculated as the time measured between 2 piezoelectrical pressure transducers placed over the common carotid artery and the femoral artery divided by the distance. A value of cfPWV > 12 m/s was defined as increased arterial stiffness. 

Endpoint Classification
Cardiovascular events were obtained from national registries. Cause of death, migration status, and hospitalization were procured from the Danish Civil Registration System, the Danish Register of Causes of Death, and the Danish National Patient Register, respectively. The last follow-up was on December 31, 2010.

Results
After a median follow-up of 18.5 years (interquartile range: 18.1–19.0), 177 (9.1%) subjects died from a cardiovascular cause (CVM). Hs-CRP (HR: 1.37, 95% CI: 1.17–1.60), suPAR (HR: 1.35, 95% CI: 1.17–1.57), and NT-proBNP (HR: 1.90, 95% CI: 1.58–2.29) were all significantly associated with CVM after adjustment for traditional risk factors (p < 0.001). Additionally, hs-CRP, suPAR, and NT-proBNP were associated with significant net reclassification improvement (NRI). However, only NT-proBNP was able to significantly impact Harrell’s concordance index (C-index) in predicting CVM after adjusting for age, sex, smoking status, systolic blood pressure, and total cholesterol (C-index 0.860 versus 0.847; p = 0.02).

After stratification for age and sex, NT-proBNP was independently associated with CVM in all participants (p ≤ 0.02) except for in younger women (p = 0.70). Hs-CRP was associated with CVM in both younger and older men (p ≤ 0.007) while suPAR was associated with CVM only in older men (p < 0.001). None of the biomarkers improved discrimination ability beyond traditional risk factors (p ≥ 0.07) when stratified for age and sex. However, NT-proBNP was able to aid significantly in the NRI of CVM in all groups except in younger women. Our classification and regression tree (CART) analysis showed that NT-proBNP seemed to be of greater value for men while suPAR seemed to have greater significance among women. 

After adjustment for traditional cardiovascular risk factors, cfPWV was not found to be associated with NT-proBNP (β = -0.01, p = 0.67). Additionally, no significant interactions were found between these two biomarkers at any of our outcomes: major adverse cardiovascular events (MACE), CVM, coronary artery disease (CAD), heart failure (HF) or atrial fibrillation (AF). While NT-proBNP was associated with all four outcomes: (HRMACE = 1.33, 95% CI: 1.16-1.52, HRCVM = 2.02, 95% CI: 1.65-2.48, HRCAD = 1.29, 95% CI: 1.07-1.55, and HRHF or AF = 1.79, 95% CI: 1.40- 2.28), CVM was the only outcome significantly associated with cfPWV (HRCVM = 1.20, 95% CI: 1.01-1.41). CfPWV was unable to aid in the NRI, alone or in addition to NT-proBNP, for any endpoint.

Conclusion
Hs-CRP, suPAR, NT-proBNP, and cfPWV may enhance prognostication of CVM in individuals without known CVD beyond traditional risk factors. Due to its ability to significantly impact both the C-index and the NRI, NT-proBNP may have increased value in the prediction of CVM in healthy individuals. Furthermore, hs-CRP, suPAR, and NT-proBNP may have different roles in predicting cardiovascular events in younger and older men and women without known CVD. When evaluating NT-proBNP and arterial stiffness, assessed by cfPWV, we found that these two modalities independently predicted cardiovascular events.