Statistics from Altmetric.com
Editor—We read with interest the excellent adverse drug reaction report by Belton and Thomas.1 They state “Treatment of severe or persistent hyponatremia is controversial” and highlight the danger of central pontine myelinolysis due to sodium infusion, particularly in patients with chronic hyponatraemia. In the learning points they mention “Sodium infusion is controversial and should only be considered in severe cases”.
We do not agree. Earlier it has been suggested that much of the brain damage associated with chronic hyponatraemia may be a consequence of improper treatment rather than hyponatraemic encephalopathy.2 In general, studies of patients suffering from chronic hyponatraemia lack information regarding whether the patients were symptomatic.3
Recent studies have demonstrated that the major criterion for use of intravenous sodium chloride treatment in patients with acute symptomatic hyponatraemia is the presence of central nervous system symptoms, regardless of the concentration of plasma sodium.4 It has been suggested that treatment with sodium chloride might be appropriate in patients with chronic symptomatic hyponatraemia.4 When pondering treatment options in patients with hyponatraemia, the pivotal issue is to determine if the patient displays encephalopathic symptoms of hyponatraemia. Ayus and Arieff have reported 53 postmenopausal women with chronic hyponatraemia who were all encephalopathic.4 Seventeen of these patients were seen before evidence of hypoxia or respiratory failure supervened. The remainder had already become hypoxic and required intubation and ventilatory assistance. The 17 patients without respiratory failure and 22 of the 36 patients with hypoxia were treated with saline or hypertonic saline to increase their plasma sodium concentration by no more than 0.8 mmol/l per hour but not greater than an end point of 133 mmol/l. The average absolute change in plasma sodium concentration was 22 and 30 mmol/l in patients with and without hypoxia respectively. Patients with seizures were given sufficient hypertonic sodium chloride to increase serum sodium concentration by 8 mmol/l during the first hour. The remaining 14 hypoxic patients who were treated conservatively with water deprivation showed an average increase of plasma concentration of 3 mmol/l during the entire course of their treatment. All patients without hypoxia recovered completely. None of the patients available for magnetic resonance imaging one year later showed cerebral abnormalities. Six of the 11 patients treated with saline who had imaging studies performed before treatment showed evidence of cerebral oedema.
Of the saline treated patients who were hypoxic, two recovered completely, six recovered partially, and the remainder either developed irreversible disabling brain damage or died. In contrast, each hypoxic patient treated with simple water deprivation either died or experienced permanently disabling brain damage. None of the 18 patients treated with saline examined with imaging techniques after treatment showed demyelination. This finding suggests that hypertonic saline, if used appropriately, does not impose that risk.
The presence of cerebral oedema on cerebral imaging before treatment in six of 11 patients treated with saline implies that cerebral adaptation to hypotonicity does not always occur, and further strengthens the justification for using hypertonic saline solution. The outcome was universally tragic in patients treated with simple water deprivation and indicates that this treatment should be abandoned in patients with symptomatic hyponatraemia. Thus, central nervous system symptoms would appear to be an indicator of the need for initiation of treatment with hypertonic sodium chloride in chronic hyponatraemia.5
Elderly patients with chronic hyponatraemia have a very high mortality rate: among 295 such patients, the mortality rate was 25%. Of these, neither the percentage of those with encephalopathy or who died of hyponatraemia can be ascertained, but at least part of this total appears to be associated with reluctance to treat chronic hyponatraemic patients with intravenous sodium chloride.4
The authors respond:
We are grateful to Kashyap and Kashyap for their remarks on our paper and for allowing us to comment on the report of Ayus and Arieff ,which was published after our adverse drug reaction report was accepted for publication.1-1
This study compared outcomes in postmenopausal hyponatraemic women treated in three different ways. Hyponatraemia was of mixed aetiology and only five of the 53 patients had inappropriate antidiuretic hormone secretion. Other causes were polydipsia (21), thiazide diuretics (17), and being postoperative (15). Group 1 (17 patients) were treated with intravenous sodium chloride before the onset of respiratory insufficiency. Group 2 (n = 22) received intravenous sodium chloride after the onset of respiratory insufficiency and group 3 (14 patients) were treated with fluid restriction only. All patients had neurological features. Outcome in group 1 was good for the nine patients successfully followed up but eight patients were lost to follow up. For group 2, two patients recovered completely, six recovered partially and were able to live independently, but 14 either died or developed neurological impairment severe enough to require institutionalisation. For group 3 all 14 patients either died or experienced permanent neurological damage and 10 died within 24 hours.
As Drs A and S Kashyap suggest, cerebral oedema was commonly present on cranial imaging before saline infusion was given, and no evidence of central pontine myelinolysis was found in the 18 patients who were evaluated at least four months after recovery.
At face value these results do suggest a very poor outcome for patients treated with fluid restriction alone. However, outcome was also poor in group 2 with only two of 22 patients recovering completely in spite of saline infusion. Furthermore, it is important to appreciate that this was not a randomised comparison. Group 3 were preselected as patients who had not responded to conservative treatment and were referred to a specialist team by the general physician. We do not know how many patients who would meet entry criteria for group 3 actually recovered with simple fluid replacement and as a result were not referred for more intensive treatment. It is also important to note that there was no relationship between the rate of correction of plasma sodium and the outcome. Furthermore, as Ayus and Arieff point out, over-rapid correction of plasma sodium also contributes to brain injury.
We accept that the data of Ayus and Arieff indicate that in severe hyponatraemia associated with neurological symptoms the benefits of sodium infusion at an appropriate rate are likely to outweigh the potential risks. However, as we suggested in our original paper, sodium infusion should only be advocated for patients with severe symptoms associated with hyponatraemia. All patients with neurological features would fall into this category. Drs A and S Kashyap's implication that sodium infusion should be considered in non-severe cases in not something that we would agree with. Under these circumstances the risks of sodium infusion are likely to outweigh benefits in this group of patients who often do well with conservative treatment.1-2