Variability May Be the “Law of Life,” but Blood Pressure Variability May Forebode a Shorter Life

Variability is the law of life, and as no two faces are the same, so no two bodies are alike, and no two individuals react alike and behave alike under the abnormal conditions which we know as disease.

-Sir William Osler (1849–1919)¹

Genetic, phenotypic, and physiologic variability complicate the understanding, diagnosis, and treatment of almost all disease states. Historically, blood pressure (BP) variations across clinic visits were viewed as random fluctuations without clinical or prognostic significance. However, emerging evidence suggests that these long-term fluctuations, termed BP visit-to-visit variability, portend a range of adverse health events in the general and kidney disease populations.^2–4 Higher BP visit-to-visit variability has been associated with acute myocardial infarction (MI), stroke, and left ventricular dysfunction.⁵ Such adverse clinical outcomes have proved particularly difficult to modify among patients with chronic kidney disease due to the complex interplay of traditional and kidney disease–specific risk factors involved in cardiovascular disease pathogenesis. The prospect of long-term BP visit-to-visit variability as a novel modifiable cardiovascular risk factor has generated enthusiasm among some because it potentially represents a new therapeutic target through which to improve patient outcomes.

WHAT DOES THIS IMPORTANT STUDY SHOW?

A recent study published in the Annals of Internal Medicine adds to the growing BP variability evidence base by providing new data linking BP visit-to-visit variability with cardiovascular events and death.⁶ Muntner et al⁶ performed a post hoc analysis of ALLHAT (Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial) to assess the association between BP visit-to-visit variability and cardiovascular outcomes. Briefly, ALLHAT was a randomized, double-blind, multicenter, clinical trial designed to determine whether treatment with a calcium channel blocker (amlodipine), angiotensin-converting enzyme inhibitor (lisinopril), or α-blocker (doxazosin), each compared to treatment with a diuretic (chlorthalidone), reduces the risk for fatal coronary heart disease (CHD) and nonfatal MI among high-risk patients with hypertension. ⁷ ALLHAT participants were 55 years or older and had hypertension and at least one other cardiovascular risk factor. Muntner et al studied 25,814 ALLHAT participants assigned to the amlodipine, lisinopril, and chlorthalidone arms who were alive and without a new cardiovascular event at study week 28. Individuals randomly assigned to doxazosin treatment were excluded due to limited available follow-up time.

In this secondary analysis, visit-to-visit BP variability was estimated from BP measurements captured at 7 follow-up study visits occurring from 6 to 28 months postrandomization. The primary exposure of interest was systolic BP variability, defined by the intraindividual standard deviation of BP across visits. This metric of BP visit-to-visit variability is the average absolute distance between observed systolic BP measurements and their mean value. Alternative metrics of systolic BP variability and diastolic BP variability were also assessed. Outcomes of interest included fatal CHD or nonfatal MI, all-cause death, stroke, and heart failure. Eligible study participants were followed up from the week 28 study visit until the occurrence of a study outcome or the end of active ALLHAT follow-up.

During a mean follow-up of 2.8 years, 1,194 fatal CHD or nonfatal MI events, 1,948 all-cause deaths, 606 strokes, and 921 cases of heart failure occurred. In multivariable analyses adjusting for several potential confounding variables, higher BP visit-to-visit variability was associated with a greater hazard of cardiovascular events and mortality. Specifically, comparing participants in the highest versus lowest quintile of systolic BP standard deviation (≥14.4 vs <6.5 mm Hg), hazard ratios were 1.30 (95% confidence interval [CI], 1.06–1.59) for fatal CHD or nonfatal MI, 1.58 (95% CI, 1.32–1.90) for all-cause mortality, 1.46 (95% CI, 1.06–2.01) for stroke, and 1.25 (95% CI, 0.97–1.61) for heart failure. These findings were consistent across alternative systolic BP variability metrics and in analyses assessing diastolic BP visit-to-visit variability.

HOW DOES THIS STUDY COMPARE WITH PRIOR STUDIES?

BP variability has received more attention during the last 5 years with the advent of detailed patient data from electronic medical records and the increased availability of cardiovascular clinical research study data to investigators. BP variability is characterized as either long or short term, depending on the timing between measurements. Short-term BP variability is typically measured by 24-hour ambulatory BP monitoring, whereas long-term BP variability, as evaluated by Muntner et al, is characterized by visit-to-visit BP changes across intervals of days, weeks, or months. A variety of metrics can be used to characterize BP variability. A key consideration in metric selection is the metric’s ability to distinguish BP fluctuations (ie, variability) from ambient BP (ie, BP mean). In studies assessing BP variability and clinical outcomes, standard deviation, standard deviation independent of the mean, and average real variability have been the most commonly used measures (Fig 1).

Figure 1.

Blood pressure (BP) variability metrics. Each figure considers one hypothetical patient’s systolic BP measurements over the 7 study visits occurring during the 22-month exposure assessment period of ALLHAT (Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial). Figures are not drawn to scale. (Top panel) In the standard deviation (SD) graph, mean systolic BP during the exposure assessment period is depicted by the solid line, and distances between individual BP measurements (black dots), dashed lines. (Middle panel) To calculate standard deviation independent of the mean, the standard deviation of systolic BP is plotted against mean systolic BP to estimate the parameter a. The shape of the resultant curve is dependent on the raw intraindividual BP data and thus will vary in form. (Bottom panel) In the average real variability graph, the distance between each systolic BP measurement (black dots) and the next is depicted by dashed lines.

The clinical importance of long-term BP variability was first well-described by Rothwell et al⁸ in a 2010 Lancet publication demonstrating BP visit-to-visit variability as a risk factor for stroke. This study was followed by many others, finding associations between long-term BP variability and outcomes such as diastolic dysfunction,⁹ acute MI,² and all-cause and cardiovascular mortality.² Notably, BP visit-to-visit variability has been shown to be clinically significant among patients with chronic kidney disease. Long-term BP variability has been linked to moderately and severely increased albuminuria and chronic kidney disease progression.^10,11 Among hemodialysis patients, BP visit-to-visit variability has been associated with cardiovascular morbidity and mortality.¹² However, conflicting results exist. A post hoc analysis of the European Lacidipine Study on Atherosclerosis (ELSA) found no association between BP variability and carotid intima-media thickness or cardiovascular outcomes,¹³ and an observational cohort study of elderly patients revealed no link between long-term BP variability and all-cause mortality, CHD-related mortality, or stroke.¹⁴ A 2014 meta-analysis considering 41 cohorts and more than 175,000 patients reported modest associations between BP visit-to-visit variability and cardiovascular morbidity and mortality, noting the limitation of the lack of a standardized approach to BP variability characterization.²

Long-term BP variability is thought to influence outcomes through complex interactions among numerous humoral, neural, pharmacologic, behavioral, and environmental factors. Medication adherence and dose titration, seasonal changes, and emotional stress likely exert influence on BP visit-to-visit variability.⁵ Studies linking greater BP visit-to-visit variability to subclinical organ damage such as increased arterial stiffness,¹⁵ albuminuria,^10,11 white matter hyperintensity, ¹⁶ and endothelial dysfunction¹⁷ all provide insight into the potential mechanistic underpinnings of the BP variability–mortality association. Furthermore, a recent meta-analysis of antihypertensive clinical trials suggests that treatment with calcium channel blockers, agents with known vasodilatory and vascular remodeling properties, may reduce BP visit-to-visit variability, reinforcing the important role of the vascular system in BP variability pathogenesis.¹⁸

WHAT SHOULD CLINICIANS AND RESEARCHERS DO?

Clinical and population studies, including the recent publication from Muntner et al in the Annals of Internal Medicine, support BP visit-to-visit variability as a prognostic indicator. However, there are several important barriers to incorporating BP variability into clinical practice. Further research with the following foci will be helpful in clarifying the clinical applicability and importance of BP visit-to-visit variability. First, BP variability is a complex construct and is not readily available at clinic visits. If simpler metrics such as standard deviation remain the mainstay of BP variability assessment, one might imagine integration of this calculation into the electronic medical record. However, it is not clear which BP variability metric carries the greatest prognostic significance. Other metric-related issues such as minimum number of BP measurements needed to calculate BP visit-to-visit variability and optimal intermeasurement interval must be determined. When an optimal measure is identified, population-based studies establishing normal ranges of BP visit-to-visit variability and treatment targets for the general population and important clinical subgroups such as patients with chronic kidney disease will be necessary prior to incorporation of BP variability into routine clinical practice. Finally, the existing observational data are insufficient to answer the question of whether BP visit-to-visit variability is a risk marker, commonly associated with hypertension, or an independent risk factor for adverse cardiovascular outcomes. If the latter is true, prospective trials considering BP variability as an outcome are needed to identify optimal antihypertensive regimens that mitigate BP variability–induced risk.

Until these uncertainties are addressed and a clinically interpretable metric of BP variability is identified, BP visit-to-visit variability should be relegated to the research arena. In the interim, clinicians should follow current clinical practice guidelines, selecting BP targets and evidence-based antihypertensive therapy based on patient histories and risk factors, always cognizant of Osler’s warning that “no two individuals react alike and behave alike under the abnormal conditions which we know as disease.”¹