Why Is Low Blood Pressure Related to Increased Cardiovascular Risk in CKD?

U-shaped association between BP and mortality in CKD

The researchers analyzed data from The Study of Heart and Renal Protection (SHARP)—a seminal trial in which 9270 patients with CKD were randomly assigned to ezetimibe/simvastatin versus placebo. The principal investigators of the SHARP Study (www.sharpinfo.org) were Colin Baigent, FRCP, FFPH, and Martin J. Landray, PhD, FRCP, also of CTSU.

The main SHARP results—published in The Lancet in 2011—showed that cholesterol-lowering therapy can substantially reduce the risk of major atherosclerotic events in CKD. Subsequent analyses of the SHARP data have yielded further insights on the outcomes and prognostic factors among people with CKD. In this new analysis, the SHARP investigators explored the paradoxical relationship between BP and cardiovascular risk in patients with CKD.

In apparently healthy adults, as BP increases so does the risk of death from ischemic heart disease, stroke, or heart failure. Risk is approximately doubled for each 20 mm Hg increase in “usual” systolic BP and each 10 mm Hg increase in diastolic BP—there is no threshold below which lower SBP is not associated with lower risk.

However, in CKD, the association curve is often U-shaped—cardiovascular risk is increased at both higher and lower BP values, including low-normal BP. One suggested reason is reverse causality: longstanding hypertension may lead to changes in cardiac structure and function, thus lowering BP while at the same time increasing cardiovascular risk.

Previous studies have found that at least half of patients with stage 4 to 5 CKD show cardiac structural abnormalities, often without signs or symptoms. In the Chronic Renal Insufficiency Cohort (CRIC) study, 75% of patients with an estimated glomerular filtration rate less than 30 mL/min per 1.73 m² had left ventricular hypertrophy on echocardiography.

Herrington and colleagues tested the hypothesis that the association between BP and cardiovascular risk might be confounded by the presence of such cardiac damage—patients who have CKD but have not yet developed cardiac disease might exhibit a positive loglinear association similar to that observed in apparently healthy adults.

To do this, the researchers needed a marker of cardiovascular risk. “The investigative trick was to use blood troponin to identify those at lowest risk of subclinical heart disease,” Herrington explained. “This was based on several previous studies showing that troponin-I is positively correlated with left ventricular mass and negatively correlated with cardiac function.”

In the SHARP cohort, higher baseline troponin-I was associated with male sex, older age, higher systolic BP, a higher prevalence of diabetes, and worse renal function. During a median follow-up of nearly five years, 2188 subjects had one or more cardiovascular events—a rate of 6.7% per year.

On adjusted analysis, higher baseline troponin-I was a strong predictor of future cardiovascular events. Risk was increased 61% for CKD patients with baseline troponin-I over 0.01 ng/mL and 182% for those over 0.03 ng/mL (compared to the reference value of 0.01 ng/mL or less). This association was apparent in both dialysis and non-dialysis CKD patients.

In the full cohort, the association between systolic BP and cardiovascular risk was U-shaped. However, among the 7278 patients without previous cardiovascular disease, there was a positive loglinear association. On adjusted analysis, each 10 mm Hg increment in usual systolic BP was associated with a 16% increase in cardiovascular risk. This risk increased to 27% per 10 mm Hg when analyses were further restricted to those patients without evidence of subclinical cardiac disease—i.e., baseline troponin of 0.01 ng/mL or less. The association was little affected by adjustment for baseline albumin:creatinine ratio, was about the same for atherosclerotic and nonatherosclerotic events, in dialysis and nondialysis patients, and among patients younger than 62 (the study median age) and those 62 years or older.

In the full cohort, however, there were also U-shaped associations for diastolic BP (but not pulse pressure). Associations with diastolic BP remained U-shaped among patients with a low troponin-I.

Support for studies of lower BP targets in CKD

The findings add to previous data on the complex relationship between BP and cardiovascular risk in CKD. Herrington and coauthors write: “The presence of a clear positive loglinear relationship between SBP (or pulse pressure) and cardiovascular events in patients with CKD at lowest risk of cardiac disease in SHARP suggests that reverse causality is a plausible explanation for previously observed U-shaped associations among patients with moderate-to-advanced CKD.”

Herrington commented: “This suggests that guidelines should not be using observational analyses of BP to define optimum BP targets in diseased populations, as such analyses may wrongly conclude that lower BP is dangerous, when the opposite may be the case.”

Randomized trials, which control for such confounding, have supported the effectiveness of lowering BP in other population groups where U-shaped associations between BP and cardiovascular risk have been observed, including patients with prior cardiovascular disease and older adults (e.g., the Systolic Blood Pressure Intervention Trial, or SPRINT). The same may therefore be true in CKD.

In the absence of sufficiently large trials, the optimal BP target in CKD remains unknown, and current recommendations vary widely. Recent studies, including SPRINT, “taken together with the evidence of reverse causality in the present analysis in the SHARP trial, suggest that trials of lower BP targets in patients with CKD are indicated,” the researchers write. Such studies would also address the potential harms as well as benefits of lower BP targets; in SPRINT, more intensive BP control was associated with an increased risk of acute kidney injury.

“The findings in this paper probably are most applicable to people with CKD not on dialysis,” commented Rajiv Agarwal, MBBS, of Indiana University School of Medicine, Indianapolis. In a 2004 review in Hemodialysis International, Agarwal hypothesized that reverse causality might account for the U-shaped association between blood pressure and cardiovascular risk in CKD.

To understand the association between BP and cardiovascular risk would require home or ambulatory BP recordings—which weren’t available in the SHARP data.

“Nonetheless, observational studies show that low BP associates with higher mortality in dialysis, while meta-analyses of randomized trials suggest the opposite,” Agarwal said. ”While we don’t have a definitive trial on BP level and outcomes in dialysis, I believe that the meta-analyses trump the observational data. Clearly there is room for research in this important area.”

Baigent noted: “The observational data in SHARP only appear to show that lower BP associates with higher risk of cardiovascular events. We argue that, if correctly analyzed with due regard to the presence of confounding by subclinical cardiac disease, the true association between BP and risk of major cardiovascular events is positive throughout the range studied.”