Assessing effective and balanced twenty-four-hour blood pressure reduction by treatment

In the Journal of Hypertension, as well as in a number of other journals, more and more papers are addressing the issue of the homogeneity of the blood pressure reduction obtained by antihypertensive treatment. The question is whether the blood pressure reduction induced by a given drug or drug regimen is distributed in a balanced fashion over the interdose interval, or whether the antihypertensive effect is markedly evident in the first few hours after drug(s) assumption only, and then wanes or disappears at a later time. Assessing the homogeneity of the blood pressure effect of an antihypertensive drug is a clinically relevant issue, because current emphasis on drugs to be administered once a day has increased the risk of a loss of efficacy at the end of the dosing interval when the pharmacokinetic properties of the selected compound do not guarantee therapeutic coverage over the entire 24 h [1–3]. To protect patients from this risk, some years ago, the United States Food and Drug Administration recommended that for each new drug, the trough treatment effect should be at least 50% of that at peak, that is, each new drug should have a trough: peak ratio of ≥ 0.5 [2,4]. This also protects the patient from another possible risk, that of taking an overdose of a drug with a relatively short duration of action, so that a trough blood pressure reduction would be obtained at the price of an excessive blood pressure fall at peak, with the risk of organ underperfusion [5–7]. Calculation of the trough: peak ratio was originally based on standardized sphygmomanometric blood pressure readings taken by the traditional Riva-Rocci/Korotkoff method, with repeated blood pressure measurements obtained in a controlled setting at different times after drug administration. The availability of 24 h ambulatory blood pressure monitoring has now made it possible to determine this value under normal daily life conditions [1,3,5,6]. Thus the homogenity of the antihypertensive effect can now be assessed in the presence of the continuous and pronounced blood pressure variations induced by a wide range of normal human behavior during the day and the night [8]. However, assessing the time-distribution of the blood pressure lowering effect of antihypertensive drugs over 24 h is a much more difficult task, because behavior-induced blood pressure changes are superimposed on those induced by the drug [5]. In particular, calculation of the trough: peak ratio from ambulatory blood pressure data appears to be less reliable because: (1) is based only on blood pressure values derived from two relatively narrow time windows (usually 1 or 2 h), thus disregarding most of the information obtained by the 24 h ambulatory blood pressure monitoring approach; (2) derives values from pretreatment and treatment periods that are often characterized by different behavioral conditions; and (3) refers to blood pressure changes that can be much smaller than those induced by behavioral factors. As a result, the trough: peak ratio values so obtained have often been characterized by a poor reproducibility [9]. Furthermore, trough: peak values with no biological significance (e.g. well above 1 or below 0) have often been calculated, and value distributions in a population of treated hypertensive patients have displayed a far from normal shape [10]. A number of solutions have been proposed to overcome these problems. First, trough: peak ratios may be computed not as averages from individual trough: peak values, but from the average troughs and average peaks of groups of subjects [9,10], because average peak and average trough values are normally distributed and are characterized by fair reproducibility [9,10]. This, however, prevents the assessment of individual responses to treatment, limiting the value of the trough: peak ratio to clinical pharmacology studies. Another solution is to use indices that extend the database from which the homogeneity of the treatment-induced blood pressure reduction is estimated, for example, by including all 24 h blood pressure values in the calculation. One potential index is the ‘surface ratio’ [11], which does focus on the whole blood pressure profile, but provides only a marginal improvement in some of the problems typical of the trough: peak ratio. Another method is to use the ‘smoothness index’ [12,13] which, in contrast to the trough: peak ratio, offers an evaluation of the daily effects of treatment based on the treatment-induced changes in ambulatory blood pressure over the full 24 h. The smoothness index is defined as the ratio between the average of the 24 hourly changes in blood pressure induced by treatment and the standard deviation of this average. Its advantages are: (1) a normal distribution in the group of subjects under evaluation and (2) a much more reproducible estimate of the homogeneity of the blood pressure effects of treatment over 24 h than the surface ratio or the trough: peak ratio [13]. These characteristics explain why some biological phenomena, presumably linked to the homogeneity of the effect of an antihypertensive drug, are much better reflected by the smoothness index than by the trough: peak ratio. For example, while the trough: peak ratio bears no relationship to 24 h blood pressure variability (an independent additional risk factor for cardiovascular complications in hypertension) [8,14–17], the smoothness index has been found to be inversely correlated with the standard deviation of the 24 h average blood pressure value [13]. Furthermore, while trough: peak ratio values have not shown any relationship with the effects of treatment on target-organ damage, that is, with the clinical outcome of antihypertensive therapy, the smoothness index has been significantly correlated with an improvement in clinically relevant target-organ damage, specifically with treatment-induced regression of left ventricular hypertrophy [13]. The smoothness index thus appears to be an effective measure of the homogeneity of the antihypertensive effect of a given treatment. It is certainly not a perfect measure, however. One limitation, for example, is that with this index the time-course of the blood pressure effect from drug assumption is necessarily lost. Also, whenever the results are based on measurements taken at intervals of 15 min or more, which is commonly the case with clinical pharmacological studies using ambulatory blood pressure monitoring, important information on the effect of treatment on blood pressure variations at smaller intervals may be lost [5,8]. Although this is a limitation of the automatic blood pressure monitoring technique rather than of data analysis, it emphasizes the potential remaining for improvement in this field. Further research is needed to find an even better way of describing the 24 h distribution of the antihypertensive effect in an individual subject, and of verifying the ability of a given antihypertensive treatment to effectively lower blood pressure throughout the different behavioral conditions of daily life.