Big practice change coming? How fast do you correct hyponatremia?

When managing hyponatremia, the primary concern (at least as far as I have always been taught) is osmotic demyelination syndrome. We are warned not to correct the sodium too quickly, because the neurologic outcome can be awful. However, demyelination is very rare, and apparently there is a lot of data that suggests correcting sodium too slowly can increase mortality!

The paper

Ayus JC, Moritz ML, Fuentes NA, Mejia JR, Alfonso JM, Shin S, Fralick M, Ciapponi A. Correction Rates and Clinical Outcomes in Hospitalized Adults With Severe Hyponatremia: A Systematic Review and Meta-Analysis. JAMA Intern Med. 2024 Nov 18:e245981. doi: 10.1001/jamainternmed.2024.5981. PMID: 39556338

The Methods

This is a systematic review and meta-analysis, following Cochrane methodology and PRISMA reporting guidelines. They searched for randomized and nonrandomized clinical trials and observational comparative studies, with no restrictions based on publication status or language. 

Patients

Hospitalized adults with severe hyponatremia, defined as a sodium less than 120 mEq/L, or severe symptomatic hyponatremia, defined as a sodium less than 125 mEq/L plus “severe symptoms, including cardiorespiratory distress, seizures, Glasgow Coma Scale ≤8, or decreased level of consciousness).”

Intervention / comparison

The looked at 4 categories of sodium correction rate:

• Very rapid (more than 12 mEq/ L per 24 hours)
• Rapid: (≥8-10 mEq/L per 24 hours)
• Slow (<8 or 6-10 mEq/L per 24 hours)
• Very slow (<4-6 mEq/L per 24 hours).

The ranges are because different studies used different definitions, and they don’t have individual patient level data. 

Outcome

The primary outcome was mortality (in hospital or 30 day). 

The Results

They found 16 studies that fit their inclusion criteria, encompassing 11,811 patients. 14 were retrospective cohorts, 1 was a prospective cohort, and one was an RCT, but not designed to answer this question (and so used as a prospective cohort). The mean sodium concentration was 115 mEq/L. 

There was an association between lower in hospital mortality and faster sodium correction. Comparing rapid correction to slow/very slow, the odds ratio was 0.59 (95% CI 0.45-0.76). They think there is a dose response relationship, with worse outcomes in the very slow (OR 0.29) than slow (OR 0.67) group. For every 1000 patients treated, this would equate to 32 extra deaths in the slow group and 221 extra deaths in the very slow group. The results were similar when looking at 30 day mortality. 

There probably aren’t enough cases of osmotic demyelination syndrome to comment accurately. There were 2 cases (out of 594; 0.3%) in the very rapid correction group, 18 cases (out of 3842; 0.5%) in the rapid group, 10 cases (out of 5652; 0.2%) in the slow group, and 1 case (out of 2466; <0.1%) in the very slow group. Although these differences are not statistically significant, it looks like there might be a trend in the direction we would expect. This would equate to 0.5 cases of demyelination syndrome per 1000 patients treated when comparing rapid to slow treatment, or 2 cases per 1000 when comparing rapid to very slow. 

Hospital length of stay was also shorter with more rapid correction (1.2 day less when comparing rapid to slow, and 3 days less when comparing rapid to very slow), which definitely makes sense if you are using specific sodium thresholds to determine safety for discharge. 

My thoughts

In medicine we often lose sight of the forest for the trees. Rare consequences or recent misses overshadow common and important outcomes. One case of torsades results in a black box causing thousands of patients to receive inadequate symptom relief. Our drive for a zero miss culture gives rise to a decision rule culture, resulting in worse overall decision making, and ultimately harming our patients. I never expected that to be the case in the management of hyponatremia, but the boogieman of demyelination may have caused us to lose track of reality. Unfortunately, reality might include many patients dying because of our phobias.

Of course, this is far from a perfect data set, and a meta-analysis is only as good as the data included.

Using “in hospital mortality” could be a biased outcome if your treatment keeps patients in hospital longer, and therefore gives them more chance to die in hospital. I don’t think that is an issue here, despite the discrepancy in admission length, because 30 day mortality outcomes were the same. 

Strictly speaking, this is all observational data, mostly not designed to answer this question, using a variety of definitions, and with significant risk of confounding. We have no data on how quickly these patients became hyponatremic, nor the underlying cause of hyponatremia, and it is very likely that clinicians acting with more information than we have are treating these patients differently. They might opt to go slowly in patients with severe cardiac disease, and it might be the underlying disease rather than the treatment of hyponatremia that is driving the mortality. Indeed, although I have not read all the underlying literature myself, the slow correction groups do seem to have higher rates of cirrhosis, heart failure, and metastatic cancer. We therefore need to be cautious in how we apply this data clinically. However, there are many theoretical reasons that slow correction of hyponatremia could result in increased mortality, and, unknown to me, this is a consistent and long known association.

At this point, we clearly need large RCTs comparing different rates of sodium correction, powered for both mortality and rates of demyelination. As always, clinicians need to decide how to act while waiting for appropriate scientific data. Personally, I think this data is strong enough to suggest a change in treatment pattern, focusing on faster sodium correction in these severe hyponatremia patients, while still trying to stay within the guideline limits for now. The US guidelines suggest a maximum correction of 12 mEq/L in any 24 hour period. (Verbalis 2013) The European guidelines are a little more strict, with a maximum correction of 10 mEq/L in the first 24 hours, and 8 thereafter. (Spasovski  2014) As rapid here was defined as 8-12 mEq/L, I think a target anywhere in that range is reasonable.

That being said, we often lack all the information needed to diagnose the underlying cause of hyponatremia in the first hour in the emergency department, and aside from significant hypovolemia, the first line treatment is often fluid restriction. Therefore, it isn’t clear how much this will change ED management, unless you are holding these patients for long periods of time without consultation. 

I include this nice graphic from the nephrology journal club as a nice overview of hyponatremia management:

Bottom line

This meta-analysis of observational data finds an association between slower correction of hyponatremia and increased mortality. This finding warrants follow-up RCTs. We need to be cautious interpreting this association, but the signal is large enough that it probably warrants some practice change while waiting on the follow-up research. 

I would love to remove advertisements from this website. My plan is to drop them when I get to 50 supporters on patreon, and I am currently at 48, so this is your chance to push us over the edge and help the broader community enjoy content without ads. 

Other FOAMed

NephJC Severe Hyponatremia Correction, Mortality, and Central Pontine Myelinolysis

AJKDblog #NephMadness 2024: Hyponatremia Correction Region

Evidence based medicine is easy

The EBM bibliography

Evidence based medicine resources

EBM deep dives

References

Ayus JC, Moritz ML, Fuentes NA, Mejia JR, Alfonso JM, Shin S, Fralick M, Ciapponi A. Correction Rates and Clinical Outcomes in Hospitalized Adults With Severe Hyponatremia: A Systematic Review and Meta-Analysis. JAMA Intern Med. 2024 Nov 18:e245981. doi: 10.1001/jamainternmed.2024.5981. PMID: 39556338

MacMillan TE, Shin S, Topf J, Kwan JL, Weinerman A, Tang T, Raissi A, Koppula R, Razak F, Verma AA, Fralick M. Osmotic Demyelination Syndrome in Patients Hospitalized with Hyponatremia. NEJM Evid. 2023 Apr;2(4):EVIDoa2200215. doi: 10.1056/EVIDoa2200215. Epub 2023 Mar 23. PMID: 38320046

Spasovski G, Vanholder R, Allolio B, Annane D, Ball S, Bichet D, Decaux G, Fenske W, Hoorn EJ, Ichai C, Joannidis M, Soupart A, Zietse R, Haller M, van der Veer S, Van Biesen W, Nagler E; Hyponatraemia Guideline Development Group. Clinical practice guideline on diagnosis and treatment of hyponatraemia. Nephrol Dial Transplant. 2014 Apr;29 Suppl 2:i1-i39. doi: 10.1093/ndt/gfu040. Epub 2014 Feb 25. Erratum in: Nephrol Dial Transplant. 2014 Jun;40(6):924. PMID: 24569496

Verbalis JG, Goldsmith SR, Greenberg A, Korzelius C, Schrier RW, Sterns RH, Thompson CJ. Diagnosis, evaluation, and treatment of hyponatremia: expert panel recommendations. Am J Med. 2013 Oct;126(10 Suppl 1):S1-42. doi: 10.1016/j.amjmed.2013.07.006. PMID: 24074529