Hypernatremia is often corrected cautiously due to fear of cerebral edema. But what if faster correction—especially in the first 24 hours—actually improved survival? A recent systematic review challenges the old dogma.
Last updated 6/3/25
‣
The Paper
“Systematic review and meta-analysis of the treatment of hypernatremia in adult hospitalized patients: impact on mortality, morbidity, and treatment-related side effects”
Kitisin et al., Journal of Critical Care, 2025 đź”—Â PMID: 3975195 DOI:Â 10.1016/j.jcrc.2024.155012
‣
Study Design
- Design: PRISMA-compliant systematic review + meta-analysis
- Databases: Ovid MEDLINE, Embase, CENTRAL (inception–Aug 2024)
- Inclusion
- Population: Adults (≥18 years) hospitalized with hypernatremia (Na ≥145 mmol/L)
- Intervention: Any active treatment correcting hypernatremia (IV fluids, enteral water, diuretics, renal replacement)
- Exposure of interest: Documented sodium correction rate
- Outcome reporting: Must include mortality, sodium change over time, or treatment side effects
- Exclusion: Head trauma, case reports, narrative reviews
- Primary Outcome: Mortality
- Secondary Outcomes: Hospital length of stay (LOS), adverse effects (neurologic or fluid overload)
‣
Results
‣
Key Analysis
- 12 studies reviewed with > 10,000 patients
- Faster correction (>0.5 mmol/L/hr) vs. slower correction:
- Odds Ratio (OR) 0.68 [95% CI: 0.38–1.24] → no statistically significant mortality benefit
- I² = 95% → studies varied a lot (bad signal-to-noise ratio)
‣
Rapid Biostats Primer
‣
- OR = 1.0 → no difference in odds of death between fast and slow correction
- OR < 1.0 → patients in the fast correction group had lower odds of dying
- OR > 1.0 → patients in the fast correction group had higher odds of dying
- If the 95% confidence interval around the OR does not cross 1.0, the result is statistically significant
‣
- I² measures inconsistency - How much variation between studies is due to actual differences (vs. random chance).
- >75% = high heterogeneity = pooled results less trustworthy
- In plain speak: if I² is 95%, the studies are likely apples and oranges.
‣
Subgroup Analyses
- Admission Hypernatremia (Hypernatremia at arrival)
- OR 0.48 [0.35–0.68], I² = 2%
- Interpretation: Patients who were already hypernatremic on admission had a 52% lower odds of death if corrected faster. Low I² means the results across studies were consistent.
- Fast Correction Within the First 24 Hours
- OR 0.48 [0.31–0.73], I² = 65%
- Interpretation: Early, fast correction was associated with a 52% lower odds of dying. More variation between studies, but effect still likely real.
- Severe Hypernatremia (Na ≥155 mmol/L)
- OR 0.55 [0.33–0.92], I² = 79%
- Interpretation: Patients with very high sodium had a 45% lower odds of death when corrected faster. Still beneficial, but more inconsistency between studies.
‣
Author Conclusion
“ Faster sodium correction appears safe and may benefit patients with severe admission-related hypernatremia, particularly within the first 24 h. Further studies are needed to refine correction protocols.”
‣
Critical Appraisal
RAMMBO Breakdown (First10EM)
Element | Assessment | Summary |
Recruitment | âś… | Large, comprehensive search with well-defined criteria |
Allocation | ❌ | No RCTs on correction rate (all retrospective cohorts) |
Maintenance | ⚠️ | Variable correction rate definitions and protocols |
Measurement | âś… | Hard endpoints (mortality); neurologic data underreported |
Blinding | ❌ | Not applicable for included retrospective studies |
Other Biases | ⚠️ | High heterogeneity, publication bias likely, limited neuro data |
BEEM Systematic Review Appraisal (đź”—Â Link)
‣
- Yes. Hypernatremia management is a frequent ED/ICU issue. Traditional correction guidelines (≤0.5 mmol/L/hr) are based on pediatric and chronic cases, not acutely ill adults.
‣
- Yes. The authors searched MEDLINE, Embase, and CENTRAL up to August 2024, registered in PROSPERO, followed PRISMA, and hand-searched references.
‣
- Yes. Risk of bias was assessed using:
- Newcastle-Ottawa Scale (for cohort studies)
- Cochrane RoB tool (for the single RCT)
- GRADE for outcome-level certainty - They rated the mortality data from admission hypernatremia as moderate certainty; others were low to very low.
‣
- Yes. Two reviewers independently screened and extracted data. Disagreements were resolved by a third reviewer.
‣
- No. Overall mortality analysis had very high heterogeneity (I² = 95%), limiting confidence in the global estimate.
- But subgroup analyses (e.g., admission hypernatremia: I² = 2%) had low heterogeneity, making them more reliable.
‣
- Mixed.
- Overall effect was imprecise (CI crossed 1.0).
- Subgroup effects were statistically and clinically significant (e.g., ORs ~0.48–0.55), suggesting real benefit.
- These results challenge current guidelines and are clinically actionable for certain patient groups.
‣
Study Strengths and Limitations
Strengths
‣
- Most prior reviews lumped together treatment strategies—fluids, diuretics, dialysis—but this one specifically asks about correction speed. That’s the actual variable we care about at the bedside.
‣
- They pooled >10,000 patients, giving this review serious statistical weight.
- Even better: Subgroup analyses (e.g., admission hypernatremia, first 24 hours, Na ≥155) showed:
- Consistently lower mortality
- Low heterogeneity in the best-performing groups (I² = 2%)
- Those subgroup signals are strong enough to raise eyebrows and challenge dogma.
‣
- This study uses a clear, objective, patient-centered primary outcome: all-cause mortality. That’s what actually matters in critically ill hypernatremic patients.
‣
- PRISMA-guided
- PROSPERO-registered
- Used GRADE, NOS, and appropriate meta-analytic models
- It checks all the boxes methodologically, even if the underlying data isn’t perfect.
‣
- The authors don’t oversell. They acknowledge that this is not definitive, and explicitly call for prospective trials to determine safe and effective correction thresholds.
Limitations
‣
- Despite being a meta-analysis, this paper included zero randomized controlled trials directly comparing fast vs. slow sodium correction. Every included study that actually evaluated correction speed was retrospective, and most were database reviews.
- This opens the door wide to confounding: Patients who are corrected faster may be less sick, better resourced, or had earlier fluid resuscitation—any of which could lower mortality independent of sodium.
‣
- This is a meta-analysis, but the definition of "fast" correction ranged from: 0.134 mmol/L/hr to 0.78 mmol/L/hr. The pooled data incorporate a wide range of correction rate thresholds, which limits the interpretability of the overall effect.
‣
- The primary meta-analysis showed high heterogeneity (I² = 95%), meaning the studies didn’t agree much—likely because they defined “fast correction” differently and involved very different patient populations.
- Subgroup analyses helped bring the variability down (I² dropped to 2–65%) and findings are exploratory. They’re useful for generating hypotheses, not writing new protocols.
‣
- Only a handful of studies even mentioned neurologic complications, and none did any kind of systematic neuro follow-up or imaging. Most just said something like “no adverse events noted” and moved on.
‣
- One study used CVVH to correct sodium at nearly 0.8 mmol/L/hr. That’s great for a controlled ICU setting—not exactly what we’re doing in a busy ED.
- Most of the included studies focused on ICU patients with invasive monitoring, renal backup, and well-staffed teams. That’s a different world. In the ED, our patients are often undifferentiated, under-monitored, and seen intermittently. So while the findings are interesting, applying them to real-world ED resuscitation needs caution.
‣
Bottom Line
Faster sodium correction—up to ~1 mmol/L/hr—may be safe and associated with lower mortality in adults with severe or admission hypernatremia, especially when initiated early in resuscitation. The strongest support come from ICU settings with close monitoring, and while no neurologic harm was reported at these rates, neurologic outcomes weren’t systematically assessed. For now, this challenges the idea that 0.5 mmol/L/hr is a hard ceiling—but it doesn’t justify aggressive correction across the board. Individualize based on clinical context, monitor carefully, and know that this evidence is hypothesis-generating. We still need high-quality RCTs to truly define safe and effective correction targets.
Related Posts
This post is for education only and not medical advice.