Page 77 - 15Diarrhoeaandvomiting

Diarrhoea and vomiting caused by gastroenteritis in children under 5 years

Evidence summary

There was a lack of good-quality studies relevant to these two questions.
Result from one UK study showed that the incidence of biochemical abnormalities was significantly
higher in dehydrated children compared with those who were not dehydrated, while the other
two UK studies did not report separately the distribution of abnormalities among dehydrated
and non-dehydrated children. Although the three studies employed variable definitions of
abnormality for the various biochemical measurements, the incidence of hypernatraemia was
reported to be less than 1% in the two latter studies and in the first study in non-dehydrated
children. High urea levels (>6 or >7 mmol/l) were found in 5–8% of children while variation
was seen regarding the incidence of acidosis. A study from Turkey found a higher incidence of
biochemical abnormalities, but more than half of the children in this study were undernourished.
This study also reported a positive correlation between serum sodium levels and blood glucose
levels in children younger than 2 years with low bicarbonate levels. The last case series from
the USA found 9.4% of children admitted to an emergency department having glucose levels
< 3.3 mmol/l, and the mean duration of vomiting in these children was significantly longer
compared with rest of the children admitted with gastroenteritis.

Evidence from the systematic review suggests that blood levels of bicarbonate > 15 or > 17 mEq/l
(or absence of acidosis) is helpful in decreasing the likelihood of dehydration, while lower levels
are not helpful in increasing the chances of detecting dehydration. High urea/BUN levels,
high serum uric acid and increased anion gap were also found to be unhelpful in detecting
dehydration. Results from another study indicate that urine specific gravity, urine ketone levels
and low urine output are not useful diagnostic tests in detecting dehydration.

GDG translation from evidence to recommendations
The GDG found that there was a lack of satisfactory evidence with regard to the incidence of
clinically important biochemical disturbances in children with gastroenteritis in the UK. One
large study did indicate that hypernatraemia, uraemia and acidosis were more commonly found in
those with clinical dehydration. The available studies did not, however, make clear the frequency
with which such abnormalities occurred in children with varying levels of dehydration. In studies
of large populations of children with gastroenteritis in the UK, the incidence of hypernatraemia
was 1% or less, and those populations included children with severe dehydration. Increased
plasma bicarbonate levels were significantly associated with dehydration but the practical
usefulness of bicarbonate estimation to detect dehydration was unclear. Studies on the potential
value of other blood and urine investigations for the detection of dehydration also failed to
provide evidence in support of their use.
The GDG recognised that children with shock may develop metabolic acidosis and monitoring
of the venous blood acid–base status is clinically important in such cases. Studies of acid–base
status usually include the pH, pCO , bicarbonate, base deficit and lactate. When acidosis is
2
detected, the underlying mechanism should be considered. Diarrhoea is often associated with
substantial bicarbonate loss and this is a common cause of acidosis. In hypovolaemic shock,
tissue blood perfusion is impaired and this can cause lactic acidosis. In those with acidosis due
to bicarbonate loss, spontaneous resolution can be anticipated as the diarrhoea resolves. Patients
with hypovolaemic shock require IV fluid bolus administration (see Chapter 5). Such fluid
therapy would not be indicated in those with acidosis related to bicarbonate loss. To distinguish
+
these two types of acidosis, the GDG suggested that measurement of the anion gap: ([Na ] +
−
−
[K ]) − ([HCO ] + [Cl ]) could be helpful, particularly in those with symptoms or signs of shock.
+
3
If the anion gap was increased, this would suggest impaired tissue perfusion as the underlying
mechanism. If the anion gap was normal, bicarbonate loss was likely to be important .
The GDG considered that routine measurement of plasma glucose was unnecessary in children
who would not otherwise require a blood test. However, if there was reason to clinically
suspect hypoglycaemia, for example in an infant with unexplained drowsiness, then it should
be measured.

52