Salty dog, the highest sodium I ever saw.

A series of sodiums from 176-188 mmol/L

Those are not glucoses. They are Sodiums. And, except for maybe an infant with congenital adrenal hyperplasia when I was a resident, those are the highest sodiums I have ever seen.

The primary management concern was the speed of correction. The first Na at 188 was drawn at 4:32 pm. Four hours later it was 177, a change of 11 mEq. Too fast. Here is the salient section from UpToDate:

Rate of correction in chronic hypernatremia — There are no definitive clinical trials, but data in children (particularly infants) suggest that the maximum safe rate at which the serum sodium concentration should be lowered in patients with chronic hypernatremia is 12 meq/L per day. To be safe, we suggest a maximum rate of correction of the serum sodium of 10 meq/L per day in patients who have had hypernatremia for at least 24 hours. The following findings provide support for this conclusion:

  • A retrospective case control study included 97 children with hypernatremia and dehydration; the mean baseline serum sodium was 165 meq/L. The rate of reduction in serum sodium was significantly faster in the children who developed cerebral edema compared with children who had no complications following correction of the hypernatremia (1.0 versus 0.5 meq/L per hour).
  • Similar findings were noted in another report in which the rate of reduction in serum sodium was 1.0 meq/L per hour in the nine infants who developed seizures compared with 0.6 meq/L per hour or less in 31 infants who did not develop seizures.
So Bud Rose, the dean of electrolytes, says 12 mEq per day.

Burton Rose

My patient moves 11 mEq in 4 hours after receiving 500 mL of normal saline. Now what? I was convinced that continuing normal saline would perpetuate the overly rapid correction of the sodium and put the patient at risk of cerebral edema. But since the patient was still in hypovolemic shock, I couldn’t just stop the fluids.

I ordered 3% saline. I ran it at 100 mL/hour and 5 hours later the sodium was back up to 186. I was going in circles. I then changed back to normal saline and over the subsequent 48 hours we corrected the sodium at roughly 0.5 mEq/l/hr. The whole time I was going through this I was wondering is it all necessary? Is rapid correction of hypernatremia as dangerous as Rose said it was? Is half a miliequivalent/L/hr a real evidence based speed limit?

I put this question to a fourth year medical student and he did an excellent job diving into the evidence (or lack there of) on the topic. Here is his analysis:

docx | pdf 
I think he is a little hard on Rose’s guidelines. The student’s analysis criticizes one of 4 references that are provided in UpToDate (though his criticism is appropriate and is the sole reference in Rose’s landmark Clinical Physiology of Acid Bas Disorders) to support of the 0.5 mEq/L/hr speed limit.
The four UpToDate references are:
  1. Rose’s own textbook, Clinical Physiology of Acid Bas Disorders. The page to look at in my 5th edition is 777.
  2. Kahn et al. Controlled fall in natremia and risk of seizures in hypertonic dehydration. Intensive Care Med (1979) vol. 5 (1) pp. 27-31
  3. Fang et al. Fluid management of hypernatraemic dehydration to prevent cerebral oedema: a retrospective case control study of 97 children in China. Journal of Paediatrics and Child Health (2010) vol. 46 (6) pp. 301-3
  4. Blum et al. Safe oral rehydration of hypertonic dehydration. J Pediatr Gastroenterol Nutr (1986) vol. 5 (2) pp. 232-5
I will look at each reference in detail.
The section in Rose’s text is very similar to the UpToDate card. The relevant paragraph is on page 777 in my 5th edition (black cover). The sole reference in this paragraph, is the 1986 Blum article (number 4 from the above list).
iPhone, a surprisingly effective photocopier.
Kahn retrospectively looked at the care of infants. They used half normal saline and gave it at 160 mL/kg/day (6.7 mL/kg/hr). All of the patients were 0-5 months, so maintenance fluids would have been 4 mL/kg/hr.

What book did this great page on maintenance fluids come from?

Nine of the first 47 patients developed seizures that could not be explained by other etiologies (fever, hypocalcemia, hypoglycemia) and so were ascribed to rapid fluid restoration, Group I. They matched these nine to 22 age-matched children who were treated contemporaneously but did not have a seizure, Group II. The investigators then changed the protocol for treating infants with hypernatremia to 120 mL/kg/day and included data on 9 patients under 5 months who were treated this way, Group III.

Here is the primary data on the three groups:

All three groups had similar sodiums but Group I had significantly higher BUNs than Group II, with I +II vs III and I vs III being non-signifigant.

The protocol called for patients in Group I and II to get only 160 mL/kg/day. Both groups significantly exceeded this. Group III hit its fluid goal nearly on the nose. They blamed the excess fluid on oral replacement and volume resuscitation with sodium bicarbonate.

The net result was a wide spread in the rate of correction of sodium:

  • The kids that seized: 1.02 mEq/L/hr
  • The kids that did not seize but were on the same treatment protocol: 0.62 mEq/L/hr
  • The kids on the conservative protocol, also without seizures: 0.35 mEq/L/hr

Fang looked at 97 children with hypernatremia. Mean sodium was 164.5. Mean age 13 months. He performed a case-control study with the cases being patients who developed cerebral edema. Manifestations included seizures, eyelid edema, papilledema in all the patients and bulging fontanel in 36 and pupillary abnormalities in 9 cases.

The data shows cerebral edema was more common with bolus therapy, especially when the bolus was faster, higher sodiums were associated with cerebral edema but much of that disappeared in multivariate analysis as higher sodiums were also associated with increased rate of correction and ind increased bolus rates. Using ROC the investigators found a rate of fluid administration of 6.8 ml/kg/hr was safest. And the average decrease in serum sodium in the cerebral edema group was 1 mEq/L/hr and 0.5 mEq/L/hr without cerebral edema.

The final reference is Blum’s study of oral rehydration, this is the reference my med student was upset with. As he outlined, this was a study of oral rehydration rather than a study of rates of treatment. The cohort was composed of eighteen infants, 6 months or younger admitted with hypernatremia (Na > 150) and treated with oral fluid resuscitation. They compared the hypernatremia outcomes to a second cohort of 26 infants who received IV rehydration for hypernatremia. Average sodium for both groups was 160.

In both groups the reduction of sodium was slow (0.3 mEq/L/hr) and no patient developed seizures.

In Burton Rose’s Clinical Physiology of Acid-Base and Electrolyte Disorders this observational study with no seizures is the sole reference behind the recommendation for a slow restoration of normal sodium. Weak sauce. Of note Androgue’s review of hypernatremia in the NEJM from 2000 references the same Blum and Khan articles to support its recommendation of slow treatment.

So in the end, the recommendation for slow normalization of sodium is based on a handful of studies in infants with no randomized or even prospective studies. What is unbelievable to me is no one references a study that reviews the functional/neurological outcomes of patients with the highest sodiums admitted to a large hospital based on the speed of correction. Seems like an easy study and in its absence we are left to trust in the physiology of babies.