Potentially life threatening profound hypokalaemia with metabolic acidosis may not be adequately dealt with by current treatment recommendations
Abnormalities of serum potassium are associated with well described clinical features: lassitude when potassium <3.5 mmol/l, possible muscle necrosis at < 2.5 mmol/l, and a flaccid paralysis with respiratory compromise at <2 mmol/l.1 World wide, hypokalaemia is most often caused by diarrhoea, although specific treatment of hypokalaemia is not mentioned in international guidelines for managing gastroenteritis.2 Furthermore, a recent case made us concerned that the potassium replacement recommended in medical texts (a maximum rate of infusion of 0.3-0.5 mmol/kg/hour and a maximum daily replacement of 3-5 mmol/kg) may be inadequate for profound hypokalaemia (⩽1.5 mmol/l).
Patients, methods, and results
The patient (case 1, table) was an 8 month old child with gastroenteritis who was too weak to respond appropriately to pain, with reduced respiratory effort, metabolic acidosis, intermittent sinoatrial block, and an inappropriately low heart rate (72 beats/min) given the degree of dehydration. The risk of inadequate treatment seemed to outweigh the risk of aggressive fluid and potassium replacement as mechanical ventilation and inotropic support were not available. The maximum recommended administration rate and total daily dose for intravenous potassium were therefore exceeded by at least 4 and 3 fold respectively without adverse effects.
We identified further cases with potassium concentrations ⩽1.5 mmol/l (614 Na+/K+ Analyser, Chiron Diagnostics) from paediatric admissions to the high dependency unit of Kilifi District Hospital in1993-2000. Data were extracted from the case records and examined for blood gas and potassium values at 4-8 hours and 18-30 hours after admission (early and late resuscitation phases). The maximum and average hourly rates and total of potassium infusion during resuscitation were calculated.
Thirteen patients, seven of whom died, were identified (table). In four death was too rapid to allow evaluation, and in one survivor data were inadequate. Strikingly, nine out of 11 patients with data on blood gases on admission were markedly acidotic. Although acidosis was persistent (possibly confounding potassium measurements) and continued stool losses could not be measured, there was a significant correlation between the late phase change in potassium and the average rate of potassium replacement over 24 hours (Spearman's τ 0.78, P=0.02). The only child developing a potassium value>5.6 mmol/l during admission was a child with a Gram negative septicaemia (case 8): potassium rose to 6.6 mmol/l (pH 6.99, base excess –23.9 mmol) at 48 hours, shortly before death.
Comment
Current guidelines for potassium replacement may not deal adequately with the rare but life threatening situation of profound hypokalaemia (⩽1.5 mmol/l) associated with metabolic acidosis seen in our developing country setting. Furthermore, recent prospective data suggest that half the children admitted with gastroenteritis have a base excess ⩽–10 mmol and 7% a potassium <2 mmol/l (PS, unpublished data) even though acidosis would normally be expected to increase potassium concentrations (due to efflux of intracellular potassium in exchange for extracellular hydrogen). About 500 children with gastroenteritis and 300 with severe malnutrition are admitted to our hospital annually, so the problem of hypokalaemia with acidosis is important.
Globally, lack of resources makes it likely that such hypokalaemia is rarely recognised. Paradoxically, therefore, children with severe gastroenteritis, perhaps at highest risk of hypokalaemia, may receive intravenous fluids2 with little or no potassium (0.9% saline, Ringer's lactate, or Hartmann's). In fact by ameliorating any associated acidosis through correcting hypovolaemia or direct alkalinisation (by lactate) cells may import potassium in exchange for intracellular hydrogen ions, further lowering serum potassium with possibly disastrous consequences.3 Further research on potassium replacement in severe gastroenteritis is required, particularly in potentially life threatening, profound hypokalaemia.
Table.
Characteristics of cases identified as profoundly hypokalaemic on admission and response to treatment
Case | Diagnosis | Age (mnths) | Admission
|
Serum potassium change (mmol/l)
|
Base excess at ∼24 hours | Maximum rate potassium infusion (mmol/kg/h) | Total potassium given in ∼24 hours (mmol/kg) | Total fluid administered in 24 hours (ml/kg) | Outcome | ||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Potassium (mmol/l ) | pH | Base excess (mmol) | Early | Late | |||||||||
1 | Gastroenteritis | 8 | 0.7 | 7.09 | −24.5 | 1.1 | 2.7 | −19.2 | 2 | 14.5 | 220 | Alive | |
2 | Gastroenteritis | 14 | 0.9 | 7.06 | −23.3 | 0.9 | 1.8 | — | 0.4 | 6.1 | 210 | Alive | |
3 | Gastroenteritis/ severe protein calorie malnutrition | 12 | 1.1 | 7.23 | −20.2 | — | 1.1 | −20.4 | 1.2 | 4.0 | 200 | Alive | |
4 | Gastroenteritis/ severe protein calorie malnutrition | 25 | 1.2 | — | — | 0.4 | 1.8 | — | 0.5 | 9.0 | 140 | Alive | |
5 | Gastroenteritis/ malaria | 6 | 1.5 | 7.25 | −17.6 | — | 1.3 | −6.5 | 0.2 | 2.7 | 180 | Alive | |
6 | Gastroenteritis | 10 | 1.0 | 7.08 | −13.9 | 0.9 | 3.0* | −14.6 | 1.0 | 7.0* | 220* | Dead* | |
7 | Gastroenteritis | 8 | 1.1 | 7.06 | −24.6 | 0.7 | 2.5 | −26.2 | 0.5 | 5.0 | 195 | Dead† | |
8 | Meningitis/ septicaemia | 4 | 1.5 | 7.12 | −17.5 | — | 2.7 | −21.2 | 1.2 | 7.5 | 150 | Dead‡ | |
Children with inadequate data during admission, including early deaths | |||||||||||||
9 | Gastroenteritis/ severe protein calorie malnutrition | 11 | 0.8 | — | — | — | — | — | — | — | Alive | ||
10 | Gastroenteritis | 6 | 0.9 | 6.97 | −24.9 | — | — | — | — | — | Dead | ||
11 | Gastroenteritis/ severe protein calorie malnutrition | 22 | 1.1 | 7.44 | −6.3 | — | — | — | — | — | Dead | ||
12 | Gastroenteritis/ acute respiratory infection | 7 | 1.3 | 7.38 | −14.7 | — | — | — | — | — | Dead | ||
13 | Gastroenteritis/ severe protein calorie malnutrition | 11 | 1.4 | 7.14 | −19.6 | — | — | — | — | — | Dead |
Died at 18 hours, had 7 mmol/kg potassium before death and fluids at a rate equivalent to 200 ml/kg/24 h. †Died at 48 hours, admitted in coma and did not regain consciousness after admission. ‡Died at about 48 hours as described in the text.
Acknowledgments
We gratefully acknowledge the help of the clinical and laboratory staff of Kilifi District Hospital and the KEMRI/Wellcome Trust Research Laboratories. ME is supported by a Wellcome Trust Career Development Fellowship. This paper is published with the permission of the director of KEMRI.
Footnotes
Funding: ME is supported by a Wellcome Trust career development fellowship.
Competing interests: None declared.
References
- 1.Gennari FJ. Hypokalemia. N Engl J Med. 1998;339:451–458. doi: 10.1056/NEJM199808133390707. [DOI] [PubMed] [Google Scholar]
- 2.World Health Organization. Management of the child with a serious infection or severe malnutrition. Geneva: World Health Organization; 2000. [Google Scholar]
- 3.Heyman S, Nehama H, Horovitz J, Sofer S, Orbach J, Amir Y, et al. Sudden death from fluid resuscitation: lesson from Ruanda. Lancet. 1994;344:1509–1510. doi: 10.1016/s0140-6736(94)90328-x. [DOI] [PubMed] [Google Scholar]