Admission plasma potassium and length of hospital stay: a meta-analysis.

OBJECTIVE

Hypokalaemia and hyperkalaemia ('dyskalaemia') are commonly seen in patients requiring emergency hospital admission. The adverse effect of dyskalaemia on mortality is well described but there are few data for the effect on hospital length of stay. We sought to determine the association of serum potassium concentration with in-hospital length of stay.

DESIGN

Systematic review and meta-analysis.

DATA SOURCES

A structured search of MEDLINE, PubMed and SCOPUS databases to 19 March 2021.

ELIGIBILITY CRITERIA

Observational cohort studies defining exposure of interest as serum potassium levels (at admission or within the first 72 hours) and with outcome of interest as length of hospital stay. Studies had to provide estimates of length of stay as a comparison between normokalaemia and defined ranges of hyperkalaemia or hypokalaemia.

DATA EXTRACTION AND SYNTHESIS

We identified 39 articles published to March 2021 that met the inclusion and exclusion criteria. Study selection, data extraction and quality assessment were carried out by two reviewers working independently and in duplicate, to assessed eligibility and risk of bias, and extract data from eligible studies. Random effects models were used to pool estimates across the included studies. Meta-analyses were performed using Cochrane-RevMan.

RESULTS

Five studies were included in the meta-analysis. Compared with the reference group (3.5-5.0 mmol/L), the pooled raw differences of medians were 4.45 (95% CI 2.71 to 6.91), 1.99 (95% CI 0.03 to 3.94), 0.98 (95% CI 0.91 to 1.05), 1.51 (95% CI 1.03 to 2.0), 1 (95% CI 0.75 to 1.25) and 2.76 (95% CI 1.24 to 4.29) for patients with potassium levels of <2.5, 2.5 to <3.0, 3.0 to <3.5, <5 to 5.5, <5.5 to 6 and >6.0 mmol/L, respectively.

CONCLUSION

Hospital length of stay follows a U-shaped distribution, with duration of admission being twofold greater at the extremes of the potassium range.

© Author(s) (or their employer(s)) 2023. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.