Serum Potassium and Mortality in High-Risk Patients: SPRINT

Abstract

A U-shaped association between serum potassium (s-potassium) and short-term mortality has been reported for patients with hypertension. Less is known about the long-term prognostic implications of s-potassium and whether this relationship is modified by intensive blood pressure (BP) control. SPRINT (Systolic Blood Pressure Intervention Trial) was a randomized, controlled trial of 9361 high-risk patients aged ≥50 years without diabetes, who were allocated to intensive versus standard BP control. We investigated associations between baseline and on-treatment s-potassium and death, using Cox proportional hazards regression (including s-potassium as a time-dependent covariate) and restricted cubic splines. We further explored the effects of intensive BP control across the s-potassium spectrum. Baseline s-potassium was available in 9336 individuals, and 8473 had a measurement at 12 months. Mean baseline s-potassium was similar between the 2 treatment groups (intensive 4.21 mmol/L versus standard 4.20 mmol/L; P=0.74), but on-treatment s-potassium was lower in the intensive group (4.14 mmol/L versus 4.18 mmol/L; P=0.001). Median follow-up was 3.3 years, with 365 all-cause deaths (3.9%) and 102 cardiovascular deaths (1.1%). Baseline s-potassium had a linear association with both types of death events (P<0.05). On-treatment potassium also had a linear association with all-cause death (P=0.04) but not with cardiovascular death (P=0.13). None of the associations remained significant after multivariable adjustment (P≥0.05). S-potassium did not modify the effect of intensive BP control (P≥0.05). In SPRINT, neither baseline nor on-treatment s-potassium levels were independently associated with death, and the effect of intensive BP control was not modified by s-potassium. Careful monitoring of patients on antihypertensive medications may eliminate the risks associated with abnormal s-potassium.