A recent study published in the NEJM this month by Seethapathy et al is the latest in a recent slew of articles and editorials this year that have reignited the controversy of whether rapid correction of hyponatraemia increases the risk of osmotic demyelination syndrome. The retrospective cohort study looked at 3274 patients with severe hyponatraemia at admission (<120mmol/L) and looked at the sodium correction rate in the first 24 hours.
The principal finding suggested that patients with rapid sodium correction rates of > 10mmol/L over 24 hours had a lower in-hospital mortality and shorter length of stay than those with restricted sodium correction rates of 6-10mmol/L over 24 hours and <6mmol/L over 24h. (But is this really so?)
This was shown to also be the case in sub cohorts of patients with underlying malignancy or congestive cardiac failure. In addition, there were seven patients identified to have central pontine myelinolysis (CPM, which is a subset of osmotic demyelination syndrome) by MRI findings and this was used to show evidence that CPM is rare, occurring in only 1.3% of patients with sodium levels < 105mmol/L and in 1.2% of those with sodium of 106-110mmol/L and in only 0.1% in patients with sodiums >116mmol/L. Of those patients only 2 had correction rates of >10mmol/L in the first 24 hours.
The authors from this study suggest that rapid correction is in fact desirable, having lower in hospital mortality rates and that the guidelines (both US and EU) are overly restrictive and conservative without much evidence. But there are a few issues in this study that need further examination.
Firstly, the study only looked at the initial sodium values on admission. There was no stratification or assessment of chronicity, i.e. an attempt to see whether sodium values were low > 48 hours prior or if this was an acute drop. In addition, this would also fail to provide information about whether upon admission the sodium had already risen a certain amount prior to initiating any treatment.
Secondly the assessment of only the first 24 hours and not the subsequent period of the patient's stay also omits assessment of whether the patient's sodium rise was kept within 16-18mmol/L over 48 hours as per current guidelines or whether these were exceeded after the initial 24 hour assessment period. Hence the CPM cases in those with moderate sodium rise of <10mmol/L in the first 24 hours may have had more rapid rises preceding or after the 24 hours of measurement.
Thirdly, though this is difficult to do in a study that includes so many patients, accurate diagnosis of CPM requires careful documentation of complete neurological findings and patients may have only subtle clinical findings, normal MRIs, or only develop MRI findings many weeks later. Hence an assessment using only MRI reports is likely an underestimation of actual CPM rates.
In March a similar study published in the NEJM by Macmilan et al. looked at a large multicenter cohort of 22,000 patients with hyponatraemia (<130mmol/L) admitted over an 11 year period. The study was again fraught with similar issues and additionally, largely looked at patients with only moderate hyponatraemia who had a relatively lower risk of developing CPM.
The article was followed up by an editorial in the same issue by Ayus and Moritz titled "Hyponatremia Treatment Guidelines - Have They Gone Too Far?" This article suggested that other risk factors for CPM/ODS such as hypokalaemia, alcohol misuse, liver disease, malnutrition, and CNS hypoxia are more important risk factors than rapid sodium correction and than a limit of 15-20mmol/L over 48 hours would be a more useful guideline. Interestingly, the suggestion of 15-20mmol/L over 48 hours does not differ much from that of current guideline suggestions of 16-18mmol/L and in fact the existing guideline figures are conservative to allow for slight overcorrection since sodium levels are often difficult to titrate so accurately. The editorial suggests instead to omit the 24h criteria allowing for a more rapid initial correction in the first 24 hours followed by less rapid correction in the following 24 hours such that the overall correction rate is still similar over 48 hours. To definitively answer whether this would be better, a similar study to the one conducted by Seethapathy should perhaps be done looking at the first and second 24 hour periods to determine if the initial rapid correction is indeed more beneficial.
A balanced review on this topic is provided by Sterns et al. in "Treatment Guidelines for Hyponatremia - Stay the Course" in the Clinical Journal of the American Society of Nephrology in June 2023. This review details extensive evidence from studies done between 1987 and 2018 supporting that rapid sodium corrections of > 10-12mmol/L over 24 hours and > 18mmol/L over 48 hours do increase the risk of CPM. The review also highlights that CPM symptoms may develop even a week post discharge and lead to readmissions which would not have been picked up in the studies done by Macmilan or Seethapathy. The review also highlights that in the patients who had CPM with < 8mmol/L correction over 24 hours, the patients went on subsequently to develop periods of significant hypernatraemia with sodium levels ranging > 150mmol/L suggesting that these patients were in fact exposed to large variations in serum and hence CNS osmolality which is likely to have triggered the CPM. On balance, the review concludes that the existing guidelines are evidence backed and should still be followed not because rapid correction of sodium will definitely cause CPM but because we know that it can and maintaining those limits may be safer.
My thoughts on the matter
Personally, I think the articles are an interesting read. I was piqued by the initial abstract wondering if all those sleepless nights spent fretting over a sodium rise of 2mmol/L more than my target was really required during my residency. But after reviewing the evidence and reading the reviews on both sides - I feel justified in the evidence and the guidelines to continue to advocate for keeping to those targets. What did interest me was that the recent Seethapathy study showed that on average sodium correction rates have decreased from 12.1 mmol/L over 24 hours in 1993 to 7.8 mmol/L over 24 hours in 2018 suggesting that guidelines have altered our clinical practices significantly!
Commentaires