Severe hyponatremia can lead to hyponatremic encephalopathy, which requires emergency treatment with hypertonic saline. Otherwise, permanent neurological damage or death could occur.
Current guidelines recommend limiting correction of severe hyponatremia for the first 24 hours to prevent osmotic demyelination syndrome (ODS). US guidelines suggest limits of 10-12 mEq/L or less in any 24-hour period and 18 mEq/L or less in any 48-hour period. For patients at high risk for ODS, the suggested limit is 8 mEq/L per 24-hour period. However, the effect that limiting hyponatremia correction has on mortality is not well understood.
To address this knowledge gap, Juan Carlos Ayus, MD, and colleagues performed a meta-analysis of cohort studies to assess the associations of different rates of correction of severe hyponatremia with mortality, hospital length of stay (LOS), and ODS. Their findings were published in JAMA Internal Medicine.
The researchers searched for randomized and nonrandomized clinical trials and observational comparative studies. There were no restrictions on language or publication status. Participants comprised adults hospitalized for severe hyponatremia (serum sodium <120 mEq/L) or with severe symptomatic hyponatremia (serum sodium <125 mEq/L plus severe symptoms, such as cardiorespiratory distress, seizures, Glasgow Coma Scale ≤8, or decreased level of consciousness).
Based on various rates of sodium correction reported in the included studies, the researchers named four categories of sodium correction rate: (1) very rapid (>12 mEq/L per 24 hours); (2) rapid (≥8-10 mEq/L per 24 hours); (3) slow (<8 or 6-10 mEq/L per 24 hours); and (4) very slow (<4-6 mEq/L per 24 hours).
The primary comparisons were made between the rapid versus slow rates and rapid versus very slow rates. However, each category was compared with every other category to examine dose-response gradients. The primary outcome of the study was mortality (both in-hospital and 30-day mortality). The secondary outcomes included hospital and intensive care unit (ICU) LOS and 90-day incidence of ODS.
The authors used Cochrane methods to perform meta-analyses for each comparison and used the random effects meta-analysis for the primary analysis. They calculated risk ratios (RRs) or odds ratios (ORs) with 95% CIs for dichotomous outcomes and mean difference for continuous outcomes.
The analysis included single-arm zero events studies and double-arm zero events studies; therefore, the Peto OR was not the optimal option. The authors used the Mantel-Haenszel risk difference, an empirical correction, and a continuity correction. Adjusted effect measures were used when available.
An I2 greater than 60% was considered substantial statistical heterogeneity. Sources of heterogeneity were studied through prespecified subgroup analyses by admission sodium levels, sex, Charlson Comorbidity Index, alcohol use, desmopressin use, setting, and cause of hyponatremia. A funnel plot was used to find and correct publication and other reporting biases when there were eight or more studies in the meta-analysis for a given comparison. Sensitivity analyses were performed by excluding high risk of bias studies or by using the fixed-effect model.
The researchers retrieved 5,010 records in their search for trials and evaluated the full text of 38 publications. Sixteen studies including a total of 11,811 patients met the inclusion criteria. In 14 of the studies included, the mean participant age was >60 years. The mean percentage of women was 56.7% in 15 of the included studies that reported sex. All but one study took place in high-income countries. Fifteen of the included studies reported in-hospital mortality, but only six reported adjusted in-hospital mortality of rapid correction compared to slow or very slow sodium correction. Eleven studies reported 30-day mortality, 14 reported ODS, 10 reported hospital LOS, and six reported ICU LOS.
Moderate-certainty evidence found that rapid sodium correction was associated with 32 (OR, 0.67; 95% CI, 0.55-0.82) and 221 (OR, 0.29; 95% CI, 0.11-0.79) fewer in-hospital deaths per 1,000 treated patients compared to slow and very slow correction, respectively. Low-certainty evidence implied that rapid correction was associated with 61 (RR, 0.55; 95% CI, 0.45-0.67) and 134 (RR, 0.35; 95% CI, 0.28-0.44) fewer deaths per 1,000 treated patients at 30 days.
A faster correction was consistently associated with shorter LOS, which suggests a potential dose-response effect. However, the difference in LOS was slight for very rapid compared to slow or very slow correction. ICU LOS was insignificantly shorter and there was no estimated dose-response effect when comparing rapid sodium correction to slow or very slow sodium correction. Low-certainty evidence suggested that rapid correction was associated with a reduction in LOS of 1.20 (95% CI, 0.51-1.89) and 3.09 (95% CI, 1.21-4.94) days, compared to slow and very slow correction, respectively.
Due to sparse data, only the alcohol use disorder subgroup analysis could be completed. Rapid sodium correction was not associated with a statistically significant increased risk of ODS when compared to slow or very slow correction in patients with and without alcohol use disorder, with an I2 of 0% in the test for subgroup differences.
Study limitations included the heterogeneity of inclusion and exclusion criteria, correction rate comparisons, cointerventions, and definitions of ODS among the included studies. In addition, ODS may have been underreported in the included studies because confirmatory imaging was required.
The authors concluded, “In this systematic review and meta-analysis, available evidence suggests that slow correction of sodium was associated with an increased risk of mortality and longer hospital LOS. These findings are further supported by dose-response effects with no statistically significant higher risk of ODS, suggesting a very favorable net benefit with rapid correction.”
Source: JAMA Internal Medicine
© 2025 Mashup Media, LLC, a Formedics Property. All Rights Reserved.