Comparing methods for assessing cardiovascular risk

There is a generalized agreement that quantification of total cardiovascular risk is necessary as a guide to initiate preventive treatment as well as to decide on its intensity and type. There are, however, different opinions as to which approach to follow to obtain a risk quantification as precise as possible, given that the probabilistic nature of the data on which quantification is based make whatever approach is selected inevitably somewhat inexact. Three approaches have gained popularity over the years and are thus predominantly used, the historical approach based on an algorithm derived from the Framingham data [1], the approach termed as SCORE which is based on the mortality data obtained from about 200 000 individuals from Europe [2]; and the approach originally proposed by the 1999 WHO/International Society of Hypertension guidelines [3] and expanded by the guidelines on hypertension issued by the European Society of Hypertension (ESH) and the European Society of Cardiology (ESC) [4]. Compared with the other two approaches the last one has the disadvantage that, with the exception of blood pressure (BP), cardiovascular risk factors are categorized in a yes and no fashion that does not allow their severity to be taken into account. It has, however, the advantage that quantification of cardiovascular risk makes use also of the presence of cardiac, renal and vascular subclinical damage via markers of documented adverse prognostic significance. This was thought to be potentially important because, when adequately searched for, subclinical organ damage can be found quite often in the hypertensive population [5,6] leading to a clear-cut increase in relative risk both when BP is elevated and when it is not [7]. The above potential advantages of the ESH/ESC approach have been only rarely tested by direct comparison with the other approaches, however. This is what was done by Sehestedt et al. [8] in the study that appears in the current issue of the Journal of Hypertension, based on individuals with a BP greater than optimal after exclusion of those with a history of cardiovascular disease, diabetes or intake of cardiovascular or lipid-lowering drugs. The comparison was limited to the ESH/ESC and the SCORE approaches, prognostic validation of the respective data being obtained by assessment of the rate of cardiovascular fatal and nonfatal events over 8 years. The results show a fair agreement between the two approaches, the use of which led to similar recommendations on the need of antihypertensive treatment in most individuals. The agreement would possibly have been even closer if patients with obvious indications for treatment, such as those with a previous cardiovascular event or diabetes were included. In the remaining small number of individuals, however, the two approaches disagreed, the SCORE approach being surprisingly the one that led to a cardiovascular risk quantification compatible with antihypertensive treatment recommendations despite lack of subclinical organ damage assessment. Does this imply that measuring subclinic organ damage only makes the procedure to quantify cardiovascular risk more complex and expensive, without any real advantage as compared with other approaches in which information on organ damage is given no or only little importance? This is not entirely implausible because, as mentioned by Sehestedt et al. [8], the advantage of having information on subclinical organ damage may be offset by the reduced precision of using the cardiovascular risk related to age and lipids only dichotomously. Yet, Sehestedt et al. [8] wisely take a cautious position on the issue. One, in several studies on hypertensive patients the systematic testing for microalbuminuria, left ventricular hypertrophy and carotid atherosclerosis led to frequent reclassification of patients from lower to higher risk categories [4,9,10] and documented an underestimation of risk by SCORE [11], although admittedly without the direct prospective validation of the data available in the present study [8]. Two, a recent review of the available prospective evidence [7] shows that subclinical cardiac, renal and vascular damage can definitively move mild hypertensive patients, without additional risk factors, into the high-risk category, a category not infrequently approached even when BP is in the high normal range. Three, the study by Sehestedt et al. [8] has limitations that were properly acknowledged by the authors. For example, because of the exclusion criteria that were adopted, the number of events observed during follow-up was limited, making the attribution to a given risk category by either approach not entirely certain. Furthermore, although commendably including several important markers of organ damage (echographic left ventricular hypertrophy, pulse wave velocity, carotid intima–media thickness and urine protein excretion) other markers were not considered. This is particularly relevant for the estimated glomerular filtration rate, a reduction of which is prognostically important and not at all uncommon [12,13]. Finally, Sehestedt et al. [8] correctly point out that assessment of subclinical organ damage may have an additional crucial advantage. Because its change by treatment has been shown to correlate with outcome [7], it may allow knowledge, albeit again on a probabilistic prospective, of whether the treatment originally selected is protecting the patient.