| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
PEDIATRICS Vol. 109 No. 5 May 2002, pp. 985
Glucagon Is Very Unlikely to Have Caused Hyponatremia
To the Editior—Belik et al 1 recommend that glucagon should not be used to treat hypoglycemia in preterm infants. This is because they believe that it caused marked hyponatremia in a 35-week gestation triplet. We believe they have misinterpreted their patient’s pathophysiology, and suggest an alternative explanation for the hyponatremia.
It is difficult to reinterpret their case precisely because their figure has a number of errors, including a nonlinear x axis, and individual clinical laboratory measurements that have been ascribed error bars (and which vary widely between spot values of similar magnitude). Most importantly, the graphs of plasma sodium concentrations and platelet counts must be wrong because the values reported in the text do not correspond with the plotted values. Neither swapping their labels nor their axis scales corrects this. However, the graphed and described values correspond if it is assumed that the upper line represents the platelet count and should have the 0 to 200 scale from the lower axis, and that the lower line is the sodium concentration with a range of 110 to 140 mmol/L.
What is clear is that the infant had been hypoglycemic and had an unexplained convulsion at 93 hours of age. Assuming the graph correction we describe above, the plasma sodium concentration fell to 134 mmol/L 3 hours later, reaching 122 by 12 hours and 116 by the following day, and then recovered over 3 days on 10 mmol of sodium/kg/day. No other interpretation accommodates the description of a plasma sodium nadir of 116 mmol/L.
Belik et al imply that the infant’s convulsion was related to the falling plasma sodium concentration, although their data seem to indicate that the convulsion preceded the fall. We suggest that their infant developed hyponatremia as a result of an uncontrolled urinary sodium loss triggered by the effect of an acute stress on the renal tubules that was insufficient to cause acute tubular necrosis. We have recognized this clinical entity many times, and have documented it precisely in 4 infants in whom we were repeatedly measuring the glomerular filtration rates, urine flow, and sodium losses using the continuous inulin infusion technique. 2 These infants each had an episode of severe stress: birth trauma in 2, necrotizing enterocolitis in 1, and pneumonia in 1. Over the next 2 days each infant had greatly increased urinary sodium losses of between about 8 and 12 mmol/kg/day (Fig 1, top graph) that led to hyponatremia (Fig 1, lower graph) without any change in glomerular filtration rate. These infants also had an increase in the fractional excretion of glucose at the same time, a feature we have seen in a large proportion of infants with milder degrees of stress. 3
The time sequence of changes in the plasma sodium concentrations in the infant described by Belik et al (filled circles, lower graph) are remarkably similar to our cases. We suggest that their infant suffered a stress that led to both the convulsion and the increased urinary sodium losses. It is interesting that their infant had periventricular leukomalacia noted at 7 days that may also reflect that an unrecognized stressful event may have occurred. Unfortunately, Belik et al only measured an untimed urine sodium concentration, and did not measure either timed urinary sodium excretions or spot urinary sodium:creatinine ratios from which these can be deduced. 4 However, if their infant’s hyponatremia was attributable to the same mechanism as ours, it would be expected that a daily sodium intake of 10 mmol/kg would approximately balance the losses for 1 or 2 days, and then lead to a gradual rise to normal, exactly as is seen.
Belik et al blame glucagon for their infant’s hyponatremia despite the fact that this has never been described before and presumably because they were unaware of any other possible mechanism. However, their case closely resembles our infants who suffered an uncontrolled natriuresis for about 2 days after a severe stress without being given any of the drugs mentioned by Belik. Others have reported similar findings. 5 We presume that the renal tubular cells become transiently dysfunctional without becoming necrotic. It would be inappropriate for neonatologists to follow Belik et al’s advice to avoid the use of glucagon in preterm infants on the basis of this single case report.
Malcolm G Coulthard, MD
Edmund N Hey, MD
Royal Victoria Infirmary
Children’s Kidney Unit
Newcastle Upon Tyne, NE1 4LP, United Kingdom
REFERENCES
-
Belik J, Musey J, Trussell KA. Continuous infusion of glucagon induces severe hyponatremia and thrombocytopenia in a premature neonate.
Pediatrics.2001; 107
:595
–597
[Abstract/Free Full Text] - Coulthard MG. A comparison of methods of measuring renal function in preterm babies using inulin. J. Pediatr.1983; 102 :923 –930[CrossRef][Medline]
- Coulthard MG, Hey EN. Renal processing of glucose in well and sick neonates. Arch Dis Child.199981 :F92 –F98
- Coulthard MG, Hey EN, Ruddock V. Creatinine and urea clearances compared to inulin clearance in preterm and mature babies. Early Hum Dev.1985; 11 :11 –19[CrossRef][Medline]
-
Wilkins BH. Renal function in sick very low birthweight infants: 3. Sodium, potassium and water excretion.
Arch Dis Child.1992; 67
:
1154
–1161
[Abstract/Free Full Text]
In Reply—
I thank Drs Coulthard and Hey for their input regarding our recently published case report. 1 Their careful review of our publication disclosed an error in Fig 1 that unfortunately was not noticed before its publication. The serum sodium and platelet count labels were inadvertently swapped.
|
I disagree, however, with Drs Coulthard and Hey’s claim that the hypoglycemia observed in that infant was related to a severe stress. As described, this infant showed good Apgar scores at birth and, aside from a very mild respiratory condition (transient tachypnea) and hypoglycemia, never had any clinically documented stress. Certainly she did not have any of the conditions previously described in Dr Coulthard’s publication. 2 Contrary to their assumption, this infant had documented serum sodium of 121 mmol/L at 90 hours of age, and thus 3 hours before the first clinically noticed generalized seizure.
The evidence that stress-induced hyponatremia can be secondary to increased urinary sodium losses is indisputable. Yet, I don’t believe that the reported infant’s clinical data supports this as the cause of the severe hyponatremia. Contrary to Drs Coulthard and Hey’s letter and as stated in our report, the original description by Carter et al of the use of glucagon infusion in infants disclosed the occurrence of unexplained hyponatremia in 1 of their patients.
Overall, I stand by our recommendation that glucagon infusion be avoided in premature neonates until its effect on the serum sodium levels and platelet count is further clarified.
Jaques Belik, MD
Division of Neonatology
Hospital for Sick Children
Toronto, Ontario, M5G 1X8 Canada
REFERENCES
- Belik J, Musey J, Trussell KA. Continuous infusion of glucagon induces severe hyponatremia and thrombocytopenia in a premature neonate. Pediatrics.2001; 107 :595 –597
- Coulthard MG. Comparison of methods of measuring renal function in preterm babies using inulin. J Pediatr.1983; 102 :923 –930
PEDIATRICS (ISSN 1098-4275). ©2002 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
