Abstract P334: N-terminal Pro B-type Natriuretic Peptide Improves Stratification of Cardiovascular Disease and Specifically Stroke Risk in South Asians

Abstract

Introduction: N-terminal pro B-type natriuretic peptide (NT-proBNP) has been widely explored as a biomarker for improving cardiovascular disease (CVD) risk stratification in populations of European ancestry. Evidence on whether NT-proBNP adds value to CVD risk prediction in South Asians (SA) is limited. Hypothesis: We hypothesized that NT-proBNP would improve CVD and stroke-specific risk stratification in SA. Methods: In the London-based multi-ethnic Southall and Brent Revisited study 1710 first generation SA were randomly recruited from primary care lists and workplaces in 1988-91. Of these, 1149 (mean age 51±7y, 85% men) were free of prevalent CVD and had complete data on baseline CVD risk factors and NT-proBNP (median 31pg/ml, 1 st -3 rd quartile 19-55pg/ml). Follow-up for mortality and fatal/non-fatal coronary heart disease (CHD) and stroke events was available to 2008-11. We assessed change in discrimination (Harrell’s C-index [95% CI]) and integrated discrimination improvement (IDI) on adding standardized log-NT-proBNP to Cox models containing age, sex, body mass index, cholesterol:HDL ratio, systolic blood pressure, antihypertensive use, smoking, diabetes, chronic kidney disease and social class; for predicting (1) risk of composite-CVD (CHD or stroke) over (a) 10y and (b) maximum follow-up of 22y; and (2) risk of stroke over 22y. Net reclassification improvement (NRI) for events (NRIe) and non-events (NRIne) was calculated across high, intermediate and low-risk categories defined by cut-offs corresponding to observed (Kaplan-Meier) risk and half-the-observed risk of each endpoint: CVD 15% and 7.5% over 10y; 44% and 22% over 22y; and stroke 10% and 5% over 22y. Results: Over 10y, 175 SA experienced a first CVD event. NT-proBNP increased C-index from 0.7269 [0.6920-0.7618] to 0.7323 [0.6973-0.7674] with IDI 0.006, NRIe -0.5% and NRIne 3%. Over 22y, 455 participants experienced a first CVD event. NT-proBNP increased C-index from 0.6759 [0.6521-0.6998] to 0.6795 [0.6555-0.7035] with IDI 0.006, NRIe -0.2% and NRIne 2.5%. In both instances, NT-proBNP correctly downgraded 2% of participants without incident CVD from high to intermediate risk with little change in other risk categories. Additionally, 106 participants experienced a first stroke event over 22y. For this model, NT-proBNP increased C-index from 0.7610 [0.7174-0.8046] to 0.7700 [0.7262-0.8138] with IDI 0.018, NRIe 3.8% and NRIne 4.9%. This correctly downgraded 3.5% of participants without incident stroke to low risk. And upgraded 5% of those with incident stroke from intermediate to high risk, while incorrectly downgrading 1% of these participants to low risk. Conclusions: In middle-aged SA, NT-proBNP modestly improves 10y and long-term composite-CVD risk prediction but improves to a greater extent long-term stroke risk prediction leading to sizeable reclassifications into more appropriate risk categories.