[The influence of hypernatremia on mortality in intensive care unit patients: a meta-analysis]

To evaluate the incidence of hypernatremia and its influence on mortality in intensive care unit (ICU) patients.

PubMed, EMBASE, Cochrane central registration of controlled trials, CNKI and Wanfang Data were searched, and relevant data of the study regarding hypernatremia in ICU patients published before September 2013 were retrieved. Meta analysis was conducted with STATA 12.0, and the effect of hypernatremia on hospital mortality and ICU mortality was evaluated.

The pooled results from 14 articles from 449 ICU in 9 countries involving 365,103 patients showed that the morbidity rate of hypernatremia in ICU was 6.9% (25,326/365,103), and the mortality rate was 32.7% (8,291/25,326). The mortality of non-hypernatremic patients was 11.9% (40,588/339,777), but the mortality of hypernatremic patients was about 2.75 times higher than non-hypernatremic patients. In patients with hypernatremia, the mortality of iatrogenic hypernatremia was 31.6% (3 790/11,976), which was similar to that of non-iatrogenic hypernatremia [33.7% (4 501/13 350)]. Further Meta analysis showed that hypernatremia was an independent risk factor of hospital mortality [relative risk (RR) 1.86, 95% confidence interval (95%CI) 1.57-2.20], and ICU mortality (RR 2.58, 95%CI 2.00-3.33), but the heterogeneity was high. The source of heterogeneity was found with subgroup analysis according to sample size (>1000 cases or ≤1000 cases), ICU type (surgical ICU or other types), study type (retrospective cohort study, prospective cohort study or case-control study), hypernatremia type (iatrogenic hypernatremia or not). Funnel plot showed no publication bias.

Hypernatremia could obviously raise the mortality rate in ICU patients, and it was an independent risk factor for critically ill patients.