Blood pressure variability and cardiovascular prognosis: implications for clinical practice

This editorial refers to ‘Visit-to-visit variability of blood pressure and cardiovascular outcomes in patients with stable coronary heart disease: insights from the STABILITY trial’†, by E. Vidal-Petiot et al., on page 2813. Blood pressure (BP) is not constant, and changes frequently in response to the physiological stimuli of everyday living. This has been documented by the use of ambulatory BP monitoring, a technique becoming more widely available that measures BP at regular intervals—typically each 15–30 min—throughout a 24-h period. This technique has demonstrated that differences among the daylong multiple readings—labelled as short-term variability—are predictive of cardiovascular outcomes independent of absolute BP levels.1 Subsequently, long-term variability, based on differences among BPs measured at multiple clinic visits during months or years of observation, also has been documented to predict fatal and non-fatal cardiovascular events. This visit-to-visit variability has been reported in patients with previous histories of cerebrovascular events2 and in patients with hypertension at high cardiovascular risk.3 Now, this concept has been explored further in a cohort of patients with coronary heart disease (CHD). Using data from the STABILITY trial, Vidal-Petiot and colleagues—calculating variability as the standard deviation of BP readings across five clinic visits during the first year of the trial—report in this issue of the journal that patients in the highest tertile of variability were significantly more likely than those in the lowest tertile to suffer major cardiovascular events and death.4 This finding applied to both systolic and diastolic BP measurements. An additional observation, perhaps of even greater clinical relevance, was that both the highest and the lowest recorded systolic and diastolic measurements during the study’s first year were also predictive of later cardiovascular events. The fact that visit-to-visit BP variability predicts events is a statistical truth, but is this observation clinically useful? It seems most reasonable to conclude that factors which increase BP variability can, for separate reasons, lead to adverse cardiovascular outcomes. So, while high visit-to-visit variability serves as a marker of cardiovascular risk, there is no evidence that it causes cardiovascular events. Admittedly, high spikes of BP such as the physiological early morning surges in BP associated with arousal from sleep, if excessive, are associated with increases in early morning strokes and magnetic resonance imaging (MRI) evidence for cerebrovascular infarcts.5 However, this is short-term rather than long-term variability. There are several factors that can affect BP and by separate mechanisms increase the probability of cardiovascular events. Some of these factors include the following. This is a likely cause since even when clinicians focus closely on drug adherence, many patients—perhaps 50%—take less than their full antihypertensive regimens, and sometimes take no medications at all.6 This problem can explain large swings in BP as well as a heightened risk of cardiovascular events. Disordered regulation of BP by autonomic mechanisms, including abnormal baroreceptor function, leads to large swings of BP and high variability.7 Many clinical situations, and even some hypertension drugs, can activate the sympathetic nervous system, leading to inconsistent BP control and increased cardiovascular events. This represents the increase in BP that occurs in about one-third of hypertensive patients when practitioners measure their BP. The magnitude of this effect changes over time as patients and clinicians become more familiar with each other. Compared with normal controls, white coat individuals have more reactive sympathetic and renin–angiotensin systems, early changes in left ventricular structure and function, and a greater likelihood of metabolic abnormalities, all of which represent cardiovascular risk factors.8 Not only do high-salt diets raise BP over time, acute ingestion of salty meals creates temporary periods of higher BP due to the pressure natriuresis required for renal elimination of excess sodium. Aso, high salt diets are associated with increased cardiovascular events. Body weight can contribute to BP variability since attempts at weight loss are often temporarily successful but are followed by increases in body weight and increases in BP. High body weight, in turn, is often linked to the risk factors of the cardiometabolic syndrome. Increased age is associated with progressive stiffening of major arteries, and in a cohort such as STABILITY with an average age of 65 and previous histories of CHD, typically there is reduced arterial compliance that increases both BP variability and the risk of cardiovascular events.9 Sleep apnoea, in particular, is common in hypertension, often linked to excess weight in these individuals, and is a cause of increased BP and thus could increase its variability. A result of poor sleep is that the physiologicall fall in BP during night-time hours does not occur, an effect known as ‘non-dipping’. As nocturnal BP measurements are more critical in determining cardiovascular outcomes than BPs at other times, the effects of sleep disorders on BP also increase cardiovascular risk.10 Common over-the-counter agents such as non-steroidal anti-inflammatory drugs (NSAIDs) and cold remedies can raise BP and contribute to variability. Many older patients who use NSAIDs have arthritis or chronic pain conditions, and so apart from the effects on BP it is possible that the need for these drugs points to concomitant inflammatory conditions that contribute to cardiovascular risk. Blood pressure is not constant throughout the day. For instance, there are physiological decreases in BP in the post-prandial periods with associated low BP measurements; whereas BPs measured at an early morning clinic visit might reflect the morning BP surge and be relatively high.5 Thus, patients whose follow-up clinic visits occur at differing times of day could appear to have increased BP variability. BPs also change according to the season of the year, particularly in countries with sharp contrasts in temperature among seasons.11 Figure 1 lists these causes of variability, many of which—separately from BP effects—can heighten cardiovascular risk. These explanations should be considered in patients whose BPs are unexpectedly inconsistent and prompt a re-evaluation of a patient’s cardiovascular risk status. Coronary Disease and BP Variability. Several situations associated with both BP variability and increased cardiovascular risk are listed at the top, but there is also a ‘hypothesized connection’ whereby myocardial ischaemia due to progressive or unstable coronary disease can intermittently trigger neuroendocrine vasoconstriction and increase BP variability. Despite the plausible mechanisms that underlie BP variability and cardiovascular events, there still appears to be a need to postulate a mechanism that directly connects these two findings. Since the STABILITY cohort had established CHD, it could be hypothesized that episodes of myocardial ischaemia in these patients could provoke sympathetic, angiotensin or other reactive responses that mediate systemic vasoconstriction and affect BP, thus more closely linking cardiovascular events and BP variability. Although this connection must be regarded as speculative, the occurrence of otherwise unexplained BP changes in vulnerable patients should raise suspicion of new, progressing, or unstable coronary lesions. Considering cardiovascular outcomes as a function of the maximum and minimum BP readings during the first 12 months of STABILITY, Vidal-Petiot and colleagues report that both the highest and the lowest pressures are associated with excess event rates. The SPRINT study recently reported that intensive treatment of office systolic BPs down to 130 mmHg (as adapted from the unique BP measurement method used in that trial) are associated with major cardiovascular benefits in high risk hypertensive patients.12 Vidal-Petiot and colleagues report that a systolic BP of 140 mmHg was a clear dividing line between high and low cardiovascular event rates, suggesting that a less intensive approach than SPRINT to systolic targets might be warranted in these CHD patients. Of note, the data from the STABILITY cohort importantly confirm the long-standing belief that diastolic values >90 mmHg are associated with excess events. Consideration of low BPs is critical since there have been concerns that excessive BP reductions during treatment might increase cardiovascular event rates—the so-called J-curve effect. In the ACCOMPLISH trial in high risk hypertension, achieved systolic BPs below 120 mmHg were related to increased coronary events, especially in patients with diabetes.13 Likewise, analysis of the CLARIFY Registry of patients with stable CHD found that systolic BPs below 120 mmHg predicted increased cardiovascular events.14 The finding from the current STABILITY cohort confirms the potential danger of low systolic pressures in CHD patients.4 However, since these recent studies were all observational, the finding that low systolic BPs were associated with adverse events could be due to reverse causality whereby low BPs reflect ongoing pathology rather than causing it. Problems with low diastolic BPs are easier to understand since coronary filling occurs during diastole, making patients with atherosclerotic coronary disease vulnerable when diastolic BPs are reduced excessively. The finding by Vidal-Petiot and colleagues of increased coronary events at values below 70 mmHg supports this understanding. Even so, this conclusion can be potentially misleading. The Framingham Heart Study similarly reported increased events in CHD patients with diastolic BPs below 70 mmHg, but only if patients had isolated systolic hypertension with pulse pressures (difference between systolic and diastolic values) of >68 mmHg.15 This finding could be explained as follows: the reduced coronary flow at low diastolic BPs in patients with CHD may be inadequate to support the major heart work required to maintain high systolic pressures; whereas, when systolic BPs are effectively reduced, low diastolic BPs can still adequately supply the myocardium. SPRINT is a good example of this: the optimal reductions in mortality and cardiovascular events produced by intensive treatment occurred with systolic BPs of ∼130 mmHg and diastolic BPs of <70 mmHg.12 Cardiovascular prognosis is optimal when both systolic and diastolic BPs are well controlled. Changes in BP, labelled as visit-to-visit variability, predict cardiovascular prognosis, and the current report demonstrates that increased variability is associated with high event rates in patients with CHD. The illustration lists several causes of BP variability, many of which also increase cardiovascular risk. We have no evidence that long-term BP variability directly causes cardiovascular outcomes, although it is tempting to speculate on a coronary/BP relationship. When unexpected BP changes occur, clinicians should consider correctable underlying causes, and—if appropriate—re-study the coronary circulation. Conflict of interest: the author served as a member of the independent Data Monitoring Committee for the STABILITY trial. The author has had no connection with the work of Vidal-Petiot et al.4 discussed in this Editorial and reports no conflicts of interest associated with this work.