The relationship between systolic and diastolic blood pressure: a clinically meaningful slope?

The traditional paradigm of hypertension as a product of increased stroke volume and peripheral resistance¹ has been progressively replaced over the last few decades by a more complex interpretative model, in which the left ventricle and peripheral arteries are functionally coupled with large arteries. The growing scientific interest in the functional properties of the large arteries has been driven primarily by an improved understanding of waveforms and of their modifications in physiological and pathological conditions.² Additional factors include the availability of non-invasive devices that are able to accurately estimate large-artery and muscular-artery stiffness,³ and the publication of several clinical studies that have shown the role of arterial stiffness as a predictor of cardiovascular damage and complications.^{4, 5} Aortic pulse-wave velocity, an index of large-artery stiffness and an independent predictor of cardiovascular complications in different clinical settings,^{4, 5} has thus been added to the list of factors influencing prognosis in hypertensive subjects,⁶ although the need for dedicated equipment and trained personnel has hindered its application in daily clinical practice. A technique to estimate arterial stiffness without the use of special equipment or qualified observers would thus be welcome.

By measuring blood pressure (BP) several times in a given individual, possibly under different circumstances, systolic BP (SBP) values can be plotted against diastolic BP (DBP) values. The slope of the relationship between SBP and DBP has been proposed as a theoretically attractive means of investigating large-artery functional properties.⁷ In principle, the link between SBP-on-DBP slope and arterial functional properties is rather straightforward.⁸ A change in DBP in a given subject is generally accompanied by a change in SBP that goes in the same direction. However, for a given increase in DBP, SBP might be expected to increase more in a stiffer artery than in a more pliable one. Thus, the increase in SBP for a given increase in DBP (SBP-on-DBP slope) could be related to arterial stiffness. The opposite holds for the increase in DBP for a given increase in SBP (DBP-on-SBP slope), which could be considered as a measure of arterial compliance. A slope-derived parameter obtained from 24-h ambulatory BP monitoring, namely the ambulatory arterial stiffness index (AASI),⁷ has received great attention from the scientific community, despite the need for caution inherent to the use of a surrogate for more direct measures of arterial rigidity.^{9, 10}

In the present issue of Hypertension Research, Gavish et al.¹¹ report the findings of a post-hoc analysis of three small studies that examined the effects of season, salt intake and paced breathing on 24-h BP and the related SBP-on-DBP slope. In one study, 13 hypertensive subjects were examined before and after 8 weeks of 15-min daily sessions of slow breathing guided by a device used to lower the high BP. In another study, 17 elderly subjects living in a nursing home underwent 24-h BP measurement before and 6 months after replacing dietary salt with low-sodium, high-potassium salt. In a third study, BP was measured in 12 subjects during the winter and the summer. Notably, slope-related measures were calculated using two different regression procedures (symmetric vs. asymmetric, see below) and three different BP-averaging methods (average of individual raw BP data over 24 h, mean of individual hourly average values and hourly averaged BP further averaged over the patient population).

The study shows that paced breathing and the summer season are both associated with a reduction in 24-h SBP and DBP as compared with baseline and winter values, respectively. More importantly, the 24-h SBP-on-DBP regression slope was significantly lower during summer and after paced breathing. Moreover, the authors found that changes in slope-related measures were sensitive to the methods used for analysis. Indexes based on symmetric regression (symmetric AASI, symmetric SBP-on-DBP slope or its equivalent, and the BP variability ratio¹²) showed a larger and more significant response to intervention or a change in condition than those based on asymmetric regression (‘traditional’ AASI). Also, a comparison of 24-h BP profiles based on hourly averaged BP, further averaged over the general population, yielded larger and more significant differences than the average of individual raw BP data.

The effect of treatment on SBP-on-DBP slope has received little attention thus far. In 188 hypertensive subjects, treatment with an angiotensin-receptor blocker was associated with a reduction in AASI, whereas patients treated with calcium-channel blockers showed no significant changes, despite a similar reduction in BP.¹³ In a randomized study, the impact of 1-year treatment with perindopril/indapamide on AASI and aortic pulse-wave velocity was not significantly different from that of atenolol.¹⁴ In a small study carried out in treated hypertensive patients, both asymmetric and symmetric AASI were lower in the subjects with high adherence to treatment.¹⁵ Overall, the present study suggests for the first time that non-drug interventions and changes in conditions known to affect BP have the potential to favorably modify the SBP–DBP slope over a period of 24 h.

How to measure SBP-on-DBP regression slope?

Several different indexes have been proposed to assess the slope of the relationship between SBP and DBP (Table 1). AASI is defined as 1 minus the DBP-on-SBP regression slope, the latter obtained using a standard asymmetric regression of DBP-on-SBP readings during a 24-h ambulatory BP monitoring session.⁷ The clinical relevance of AASI is suggested by its cross-sectional correlation with pre-clinical target-organ damage^{16, 17, 18} and its predictive value for future cardiovascular morbidity and mortality,^{19, 20, 21, 22} although the superiority of AASI over more widely accepted surrogate measures of arterial stiffness such as 24-h pulse pressure²³ has not been established.^{24, 25} Some methodological drawbacks of AASI need to be considered, however.¹⁰ First, AASI is strongly dependent on the degree of nocturnal BP drop.^{26, 27, 28} By definition, subjects with a low nocturnal BP fall (non-dippers) tended to have a narrower range of DBP values over the 24-h period (that is, the dependent variable in the regression equation that generates AASI). As a mathematical consequence, the coefficient of regression B of DBP on SBP tends to decrease over the 24 h, and its complement (AASI, or 1−B) increases.²⁶ Notably, adjustment for nocturnal BP reduction eliminates the relationship of AASI with measures of target-organ damage, such as left ventricular hypertrophy²⁶ and microalbuminuria.²⁹ Second, the ratio of nocturnal over diurnal BP readings, determined by the nighttime and daytime between-measurement time intervals, has an effect on AASI—the lower the number of nocturnal readings as compared with day-time readings, the higher the AASI.³⁰ This confounding effect is due to the above relationship between day–night BP reduction and AASI.³¹ Third, AASI is intrinsically dependent on the correlation coefficient between SBP and DBP.¹² SBP-on-DBP slope and DBP-on-SBP slope calculated using standard asymmetric regression are reciprocal to each other only when the correlation coefficient between SBP and DBP is 1. In all the other cases, using asymmetric regression leads to artifactual dependence of AASI on the SBP–DBP correlation coefficient.¹²

Table 1 Parameters based on the slope of the relation between systolic and diastolic blood pressure

A modified AASI, calculated with symmetric regression, may reduce its dependence on day–night BP changes and on the correlation between SBP and DBP, and has been associated with all-cause mortality.^{32, 33} The method, however, has not been compared with direct measures of arterial stiffness, and its clinical and prognostic relevance awaits further confirmation. Notably, symmetric AASI is mathematically connected to the ratio between the standard deviation of 24-h SBP and 24-h DBP (BP variability ratio)¹² (Table 1).

A cheaper and more universally available alternative way to obtain a large number of BP readings may be represented by home BP monitoring. Home arterial stiffness index, an index of the dynamic behavior of the relationship between SBP and DBP in a given individual obtained from home BP monitoring, has shown a direct correlation with age and pulse pressure, although such associations were invariably weaker than those observed with AASI.³⁴ Possible reasons for the inferiority of home arterial stiffness index as compared with AASI are a lower number of BP readings and the fact that home arterial stiffness index is based on BP measurements taken at rest and in a standardized position, whereas AASI includes measurements taken during the execution of typical daily activities.

SBP-on-DBP slope: a measure of vascular stiffness?

AASI, the most widely investigated measure of DBP-on-SBP slope, has been related to carotid-to-femoral pulse-wave velocity, a direct measure of aortic stiffness,⁷ although most studies did not confirm this finding in larger cohorts.^{26, 35, 36, 37} In these studies, the relation between AASI and aortic pulse-wave velocity was weak, and, importantly, was affected or entirely negated after adjustment for age and other factors.^{26, 35, 36, 37} Additionally, 24-h pulse pressure showed a stronger relationship with aortic pulse-wave velocity than AASI.^{36, 37}

The physiological significance of slope-related parameters is currently debated. Based on a computer model, Craiem et al.³⁸ concluded that AASI is higher in stiffer arteries due to the nonlinearity of arterial elasticity. Westerhof³⁹ suggested that AASI expresses ventriculo-arterial coupling rather than arterial stiffness. Gavish⁴⁰ suggested that the SBP-on-DBP slope could reflect the increase in arterial stiffness from diastolic to systolic values (‘systolic stiffening’) rather than diastolic stiffness, the latter typically measured by pulse-wave velocity. It was demonstrated over half a century ago that the slope of the change in pulse pressure over the change in diastolic pressure after inhalation of amyl nitrite (the so-called ‘arterial rigidity index’) increases with age and with clinical evidence of atherosclerosis.^{41, 42} Thus, the relationship between SBP and DBP might describe a physiological process that is different and only partially related to diastolic arterial stiffness as measured by pulse-wave velocity. The study by Gavish et al.¹¹ suggests that paced breathing and the summer season might have a positive effect on arterial functional properties as expressed by the slope of SBP-on-DBP over the 24-h period.

Study limitations and perspectives

The findings of the article by Gavish et al.¹¹ should be interpreted within the context of its limitations. Data from the salt restriction study were quite unexpected. Replacing common salt with low-sodium, potassium-enriched salt had no effect on slope-related measures and, most importantly, was followed by a paradoxical increase in SBP and pulse pressure. These data are in sharp contrast with the current literature, which shows a beneficial effect of salt restriction on arterial stiffness.^{43, 44} The authors provide no explanation for this puzzling finding. The examined population was very small, with approximately one dozen patients enrolled in each of the three studies. The small study sample limits the strength of the conclusions and prevented more refined data analyses. Also, the absence of a control group should be noted.

Given the methodological limitations of this study, these results would need to be confirmed in a larger group of patients and with a more robust methodological approach. Nevertheless, its conclusions may stimulate further investigations on the impact of lifestyle and drug treatment of hypertension on the relation between SBP and DBP.