Background The impact of hyperkalemia on mortality in cohort studies is difficult to assess because of fluctuations in serum potassium (S-K) levels, differences in the intensity of intervention for S-K, and the reliability of S-K levels measured across multiple facilities. Because our corporation has intervened for hyperkalemia when centrally measured S-K levels are ≥ 6.5 mEq/L, we investigated whether average S-K levels of 5.5–6.4 mEq/L were associated with higher mortality. Methods This retrospective observational study included 678 patients undergoing dialysis who were divided into low (3.5–4.4 mEq/L), intermediate (4.5–5.4 mEq/L), and high (5.5–6.4 mEq/L) groups according to average S-K level, excluding those with levels ≥ 6.5 mEq/L at least once during the 3 months before the start of the study. The 2-year all-cause mortality was compared using a propensity score-matching (PSM) model with annually censoring of patients who had S-K levels of ≥ 6.5 mEq/L at least once after 10–12 months. Kaplan–Meier survival curves were compared using the log-rank test, after which Cox regression analysis was performed with adjustments. Results After PSM, S-K levels were 4.1 ± 0.2 mEq/L in the low group and 5.7 ± 0.2 in the high group (P < 0.001). Mortality was not significantly different between the two groups. After PSM, S-K levels were 4.9 ± 0.3 mEq/L in the intermediate group and 5.7 ± 0.2 in the high group (P < 0.001). Mortality was significantly lower in the high group compared with the intermediate group (hazard ratio 0.120, 95% confidence interval 0.015–0.956, P = 0.016, log-rank test). However, the adjusted Cox regression analysis showed no significant difference between the groups. Conclusions These results suggested that mortality is not higher in dialysis patients with average S-K levels of 5.5–6.4 mEq/L than in those with average 3.5–4.4 mEq/L or 4.5–5.4 mEq/L when the interventional S-K levels were standardized to ≥ 6.5 mEq/L, and patients with S-K levels of ≥ 6.5 mEq/L were excluded or censored. Trial registration: UMIN Clinical Trials Registry, UMIN000056789. Registered 22 January, 2025, Prospectively registered, <a href="https://center6.umin.ac.jp/cgi-bin/ctr/ctr_view_reg.cgi?recptno=R000064682">https://center6.umin.ac.jp/cgi-bin/ctr/ctr_view_reg.cgi?recptno=R000064682</a>.

Average hyperkalemia between 5.5 and 6.4 mEq/L does not increase mortality in patients undergoing dialysis

期刊论文
出版模式：

开放获取
发表日期：

2025年12月6日

查看全文

Kazuyoshi Okada,
Daisuke Hara,
Tomoko Inoue,
Manabu Tashiro,
Jun Minakuchi

摘要

Background

The impact of hyperkalemia on mortality in cohort studies is difficult to assess because of fluctuations in serum potassium (S-K) levels, differences in the intensity of intervention for S-K, and the reliability of S-K levels measured across multiple facilities. Because our corporation has intervened for hyperkalemia when centrally measured S-K levels are ≥ 6.5 mEq/L, we investigated whether average S-K levels of 5.5–6.4 mEq/L were associated with higher mortality.

Methods

This retrospective observational study included 678 patients undergoing dialysis who were divided into low (3.5–4.4 mEq/L), intermediate (4.5–5.4 mEq/L), and high (5.5–6.4 mEq/L) groups according to average S-K level, excluding those with levels ≥ 6.5 mEq/L at least once during the 3 months before the start of the study. The 2-year all-cause mortality was compared using a propensity score-matching (PSM) model with annually censoring of patients who had S-K levels of ≥ 6.5 mEq/L at least once after 10–12 months. Kaplan–Meier survival curves were compared using the log-rank test, after which Cox regression analysis was performed with adjustments.

Results

After PSM, S-K levels were 4.1 ± 0.2 mEq/L in the low group and 5.7 ± 0.2 in the high group (P < 0.001). Mortality was not significantly different between the two groups. After PSM, S-K levels were 4.9 ± 0.3 mEq/L in the intermediate group and 5.7 ± 0.2 in the high group (P < 0.001). Mortality was significantly lower in the high group compared with the intermediate group (hazard ratio 0.120, 95% confidence interval 0.015–0.956, P = 0.016, log-rank test). However, the adjusted Cox regression analysis showed no significant difference between the groups.

Conclusions

These results suggested that mortality is not higher in dialysis patients with average S-K levels of 5.5–6.4 mEq/L than in those with average 3.5–4.4 mEq/L or 4.5–5.4 mEq/L when the interventional S-K levels were standardized to ≥ 6.5 mEq/L, and patients with S-K levels of ≥ 6.5 mEq/L were excluded or censored.

Trial registration: UMIN Clinical Trials Registry, UMIN000056789. Registered 22 January, 2025, Prospectively registered, https://center6.umin.ac.jp/cgi-bin/ctr/ctr_view_reg.cgi?recptno=R000064682.