Struthers 3P Screening for Residual Risk in Hypertension 237 has been shown to further regress LVH. 14 This suggests that a much more targeted approach is needed where the precise cause of the residual risk is identified and treated specifically. The next key question is whether it would be worth identifying the 34% of treated hypertensives who are harboring the above silent cardiac abnormalities (and if so, how). To answer the second question first, identifying LVH, LVDD, LAE, and LVSD would normally require echocardiography in all treated hypertensives, and there is general agreement that this would not be cost effective. Furthermore, identifying SMI would require a different test, which would involve either a stress test or a radiation exposure (computerized tomography coronary calcium); these latter tests would both be daunting as routine tests, especially considering that SMI is only present in 6% of such asymptomatic individuals. However, recent data from our group suggest that B-type natriuretic peptide (BNP)(±high sensitivity troponin) screening would be a cost-efficient way to identify which treated hypertensive patient is likely to have any of LVH, LVDD, LAE, LVSD, or SMI. 10 The c-statistic was 0.78 for this (0.81 for BNP+ hs troponin). The area under curve for the receiver operating curve for hs troponin on its own was lower (0.70). The sensitivity and specificity of BNP varied at different BNP levels, but at 10 pg/mL, it was 85% and 51%; maximizing sensitivity to avoid false negatives is best where a more definitive test follows (as here). The sensitivity and specificity of combining BNP (15 pg/mL) and hs cTnT (5· 9 ng/L) was 87% and 65%. Of particular importance was the fact that BNP accurately identified the worst cases (ie, BNP accurately identified> 93% of those with≥ 2 cardiac abnormalities or with LVSD on its own or with SMI on its own). Many earlier studies in various different populations had already shown that BNP could identify 1 or 2 isolated abnormalities like LVH, LAE, LV dysfunction, or silent ischemia. However, the importance of the new study described above is that all forms of cardiac abnormalities were looked for in the 1 study. 10 This is very important because, patients harboring forms of cardiac abnormalities not assessed in earlier studies will mistakenly appear as false positives for BNP; this will not occur if the search for cardiac abnormalities is comprehensive, as in this recent study. 10 Of course, BNP levels can be altered by renal dysfunction and obesity and future research should address the effect of these complicating factors and whether BNP testing can be applied to these subgroups (and if so, at what levels).

How do we know that, if we used BNP to identify those with LVH, LVDD, LAE, LVSD, or SMI, we would be identifying those at the highest risk? There are 2 strong indicators that this is the case. First, Tsang et al15 showed that each of LVH, LVDD, LAE, and LVSD independently increases CV risk by≥ 40%, whereas SMI is well recognized as a cause of sudden cardiac death in apparently healthy individuals. 16 Second, Nadir et al10 showed that those with any of LVH, LVDD, LAE, LVSD, and SMI have an elevated level of BNP, whereas Paget et al17 showed that a high N-terminal proBNP independently increased the risk of total mortality 3-fold in treated hypertensives. Overall, therefore, Paget et al17 clearly indicate that N-terminal proBNP can be used to identify those with the most residual risk, whereas Nadir et al10 show what known cardiac culprits are present in those with a high BNP