Blood pressure variability and risk of stroke in chronic kidney disease

Chronic kidney disease (CKD), defined as structural [increased albumin-to-creatinine ratio (ACR), abnormal imaging, histology, or urine sediment] or functional damage [estimated glomerular filtration rate (eGFR) lower than 60 ml/min per 1.73 m2] for at least 3 months [1], is a major public health problem with an estimated worldwide prevalence of approximately 10% [2–4]. Cardiovascular disease is the leading cause of death in patients with CKD [5]. In the last years, much focus has been given to the high rate of cardiac complications in CKD patients [6]; nonetheless, more attention should be given to the debilitating burden of cerebrovascular disease among CKD patients. Indeed, the risk of stroke in CKD has been reported to be significantly greater than that in the general population and a graded and independent relationship between eGFR and stroke risk has been clearly reported in several studies [7–10]. The prospective Atherosclerosis Risk in Communities study [8] demonstrated that CKD, defined as an eGFR less than 60 ml/min per 1.73 m2, increased the risk for stroke about two-fold [hazard ratio 1.81, 95% confidence interval (CI) 1.26--2.02], even after adjustment for traditional cardiovascular risk factors [8]. Similar results were documented by the multicenter Prevention Regimen for Effectively Avoiding Second Strokes Trial [10] and by a systematic review and meta-analysis of prospective studies for a total of 284 672 participants (follow-up 3.2–15 years, 7863 stroke events) [9]. These reports [9,10] showed a 43% increased risk of stroke among participants with an eGFR less than 60 ml/min per 1.73 m2 [relative risk (RR) 1.43, 95% CI 1.31--1.57; P < 0.001; normal reference: eGFR >60 ml/min per 1.73 m2 or >90 ml/min per 1.73 m2] [9], and very high risks among patients with an eGFR less than 40 ml/min per 1.73 m2 (RR: 1.77, 95% CI 1.32--2.38) [9,10]. As such, averting future cerebrovascular events in patients with a low eGFR should be a primary goal. Among modifiable risk factors for cerebrovascular accidents in CKD, hypertension is the major contributor for both ischemic and hemorrhagic stroke with risk increasing with worsening blood pressure (BP) control [11,12]. In this context, current guidelines [13,14] recommend multiple readings before making a diagnosis of hypertension as well as before initiation and titration of BP-lowering drugs. Thus, BP variability recorded over multiple time-frames adds to the complexity in managing hypertension. The picture is further complicated by the evidence that elevated long-term (visit-to-visit) BP variability may be associated with adverse outcomes in CKD [15–18]. Higher visit-to-visit BP variability in patients with CKD has been associated with mortality and incident cardiovascular events [15–18]. In a recent retrospective analysis from Kaiser Permanente Northern California population [15], visit-to-visit BP variability was assessed over 6 months in a community-based cohort of 114 900 adults with CKD stages 3 and 4 (eGFR 15–59 ml/min per 1.73 m2). The highest versus the lowest quintile of the coefficient of variation was associated with higher adjusted rates of death (hazard ratio 1.22; 95% CI 1.11--1.34) and hemorrhagic stroke (hazard ratio 1.91; 95% CI 1.36--2.68). On the contrary; BP variability was inconsistently associated with heart failure and end-stage renal disease, and was not significantly associated with acute coronary syndrome or ischemic stroke. For the limited amount of available data regarding the relationship between BP variability and stroke in CKD, results of the analysis by Li et al.[19] published in the current issue of the Journal are welcome. Briefly, the present post hoc analysis of the China Stroke Primary Prevention Trial (CSPPT) [19] evaluated the relation of visit-to-visit BP variability with the risk of first stroke, and examined any possible effect modifiers in hypertensive patients with mild-to-moderate CKD. Overall, data from 3091 hypertensive patients with eGFR 30–60 ml/min per 1.73 m2 and/or proteinuria at baseline were included. BP measurements were collected at baseline, randomization, and every 3 months thereafter. Standard deviation (SD) and coefficient of variation (calculated as SD divided by the mean) of BP were used as visit-to-visit parameters and were determined using all available BP measurements [19]. During the median follow-up time of 3.7 years, a first stroke occurred in 92 (3.0%) participants. After multivariable adjustment, including baseline SBP and mean SBP during the first 12-month treatment period, there was a significant direct relationship of SBP variability with the risk of first stroke (hazard ratio 1.41; 95% CI 1.17--1.69, per 1-SD increment). Moreover, the results were consistent for SD of DBP, and for coefficient of variation of SBP/DBP [19]. As acknowledged by the authors, this study suffers from some limitations, including its design and statistical plan; indeed, post hoc observations from a clinical trial should be treated with extreme caution irrespective of their statistical significance [20,21]. With this background in mind, some information and strengths of the study need to be mentioned. Exploring the prognostic relevance of visit-to-visit BP variability, all BP data preceding the occurrence of the outcome or censoring are generally used to compute BP variability. This is a main concern on the validity of results for the overlap of the variable-length exposure and outcome periods and a time-varying approach appears more appropriate and convenient. At this regard, Li et al.[19] included some sensitivity analyses to evaluate the impact of BP variability in some subgroups as defined by participants with greater than two BP measurements during the first 6 months or the first 18 months from randomization, and by the available number of BP measurements. Of note, results remained consistent across all these subgroups. Furthermore, Li et al.[19] demonstrated that visit-to-visit BP variability is related to the risk of incident stroke independent of some confounding factors, including age, sex, diabetes, number of concomitant antihypertensive medications used, and treatment compliance. Even more important, estimates of long-term BP variability proved to be useful for prognosis after adjustment for the corresponding SBP treated as time-dependent variable in the survival models. Notwithstanding, it is not entirely clear whether the two metrics (SD and coefficient of variation) used to estimate BP variability add comparable predictive power to a model formed by simple, easily available clinical variables in patients with CKD. Finally, long-term BP variability might also result from poor adherence to antihypertensive therapy. Kronish et al.[22] assessed the association between antihypertensive medication adherence and visit-to-visit BP variability in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). Participants who reported taking less than 80% of their antihypertensive medication at at least one study visit were categorized as nonadherent. SD independent of mean of BP was higher for nonadherent versus adherent participants (11.4 ± 4.9 versus 10.5 ± 4.5 for SBP; 6.8 ± 2.8 versus 6.2 ± 2.6 for DBP, each P < 0.001). SD independent of mean of BP remained higher among nonadherent than among adherent participants after multivariable adjustment. Of note, adjustment for nonadherence did not explain the association of visit-to-visit BP variability with higher fatal coronary heart disease or nonfatal myocardial infarction, stroke, heart failure, or mortality risk. Such results suggest that BP variability is associated with cardiovascular outcomes independent of medication adherence. Li et al.[19] further highlight this issue in CKD patients. In their analysis, BP measurements were taken at baseline, randomization, and every 3 months thereafter. BP variability was determined using all available BP measurements following the same prescription from the randomization visit to the 12-month visit. Specifically, Li and co-workers [19] removed from the determination of variability the BP values of the visit at which antihypertensive prescription were modified. Such attempt may have a role in limiting the confounding effect of nonadherence on prognosis. In conclusion, as stroke is a leading cause of mortality and morbidity worldwide [11], and BP control may reduce subsequent cerebrovascular disease in patients with CKD [12], it is important to identify the CKD patients at increased risk of stroke and treat them more intensively [23–29]. To date, estimates of long-term BP variability may be considered as potentially useful for stroke risk stratification in CKD patients. However, available data are relatively limited and do not support conclusive recommendations. Such aspects need to be addressed in future studies in patients with CKD, also to clarify whether lowering BP variability reduces adverse outcomes. Last but not least, the biological mechanisms underlying the association between BP variability, CKD and stroke need to be more clearly determined (Fig. 1). Various mechanisms may be involved in the association between kidney dysfunction, high BP variability and cerebral blood flow [30–33] (Fig. 1). The brain and kidney share similar microvasculature and vasoregulation behavior. Both organs are perfused by low-resistance vascular circuits, which permit continuous high-volume blood flow during systole and diastole. Small vessels in other organs are protected by upstream vasoconstriction; conversely, small arteries in the brain and kidney are constantly exposed to fluctuations in pressure, and flow because of low vascular resistance and upstream vasodilation [30–33]. To maintain relatively constant blood flow to the brain with variable systemic BP, the brain vasculature displays cerebral autoregulation to minimize hypoperfusion during low BP states and hyperperfusion during high BP states [30–33]. Thus, elevated BP variability may increase susceptibility of brain and kidney to microvascular dysfunction, leading to endothelial dysfunction. Of note, cerebral autoregulation is a complex intrinsic control mechanism maintaining a constant blood flow by changing cerebral vascular resistance in response to changing BP, cerebral perfusion pressure, or metabolic needs, and depends on preserved endothelial function [30–33]. Similarly, endothelial dysfunction in the kidney holds a key position as a central link of the intricate and intertwining pathways involved in the pathogenesis of CKD [34–38]. Finally, both CKD and high BP variability promotes carotid atherosclerosis, platelet dysfunction, oxidative stress, inflammation, reduced cerebral oxygenation and development of atrial fibrillation [34–38] (Fig. 1). More specifically, the augmented mechanical stress on the vascular system (which leads to vascular remodeling) and the increased variability of blood flow by augmented BP variability increases sheer stress on endothelial cells advancing atherosclerosis [34–38]. Sheer stress-induced platelet activation at atherosclerotic stenotic sites, and subsequent hypercoagulability may lead to cerebrovascular events [34–38]. Thus, when CKD and elevated BP variability coexist, the risk of stroke is expected to dramatically increase.FIGURE 1: Potential mechanisms linking elevated blood pressure variability, chronic kidney disease and risk of stroke (see text for details). BP, blood pressure variability; CBF, cerebral blood flow; CKD, chronic kidney disease.ACKNOWLEDGEMENTS Conflicts of interest There are no conflicts of interest.