We have undertaken an individual participant data (IPD) meta-analysis of carotid-femoral pulse wave velocity (cf-PWV) with all cause mortality, CHD, stroke and combined CVD events using data from 14 studies (2 unpublished). Unlike a previous report, which only used published data, we were able to undertake standardised analyses with and without adjustment for cardiovascular risk factors and test, a priori, for potential interactions between cf-PWV and age group, gender, diabetic or hypertensive status on the various outcomes. We calculated discrimination statistics for models with and without cf-PWV, specifically focussing on individuals at intermediate (25–75^th percentile) risk of CVD after adjustment for conventional Framingham risk factors. Fourteen studies provided data on 16,358 subjects with 1700 combined CVD events. We derived within study z-scores of log transformed cf-PWV (pooled SD = 3.3m/s). Risk of all outcomes was associated with increased cf-PWV (Table 1) and was linear across the range of cf-PWV values with no evidence of interaction except for age group (see Figure, p-value for trend = 0.0095). The additional benefit of measuring cf-PWV to reclassify intermediate risk individuals was assessed using the net reclassification index. 18.6% (p<0.001) and 22.4% (p<0.001) were appropriately reclassified into higher or lower quartiles of risk for CHD and stroke outcomes respectively. These findings highlight the added value of cf-PWV as an independent predictor, over and above existing risk factors, in intermediate risk groups and for younger subjects. Assessment of PWV should better identify high risk populations that may benefit from more aggressive risk factor management.

Figure:

Pooled hazard ratios for cardiovascular disease for a one standard deviation increase in pulse wave velocity (log transformed) stratified by age group.