Page 126 - 15Diarrhoeaandvomiting

Antibiotic therapy

Another retrospective review, 159 conducted in Hong Kong, included 126 children with
salmonella (n = 86), rotavirus (n = 55) or unspecified gastroenteritis (n = 126) who were
admitted to hospital. [EL = 2+]. Demographic, clinical (dehydration, vomiting, fever, diarrhoea,
abdominal pain), stool and medication outcomes were collected from case notes and analysed
according to gastroenteritis type. Patients with salmonella were more likely to have bloody
(OR 6.1; 95% CI 3.2 to 11.7; P < 0.0001) or mucoid stools (OR 4.8; 95% CI 2.6 to 8.9;
P < 0.001) compared with the combined rotavirus/non-specified gastroenteritis groups. They
were statistically significantly younger (median 7.1 months versus 14.6 months; P < 0.0001),
had a longer stay in hospital (median 3.4 days versus 2 days; P < 0.0001), passed more stools
per day (median 6.2 versus 4.2; P < 0.0001) and more of them experienced fever during their
admission (OR 3.6; 95% CI 1.6 to 8.4; P = 0.001). Additionally, patients with salmonella were
statistically significantly more likely to have been given antibiotics than children in the other
two groups (OR 3.6; 95% CI 1.9 to 6.9; P < 0.0001), although administration of antibiotics was
not dependent on age.

Evidence summary

Compared with other enteropathogens, salmonella gastroenteritis has repeatedly been shown
to particularly affect younger children. A retrospective review from Malaysia found that 67% of
all salmonella-infected children were younger than 1 year. Most children developing invasive
salmonellosis (bacteraemia or meningitis) were younger than 6 months. Similarly, a study from
Hong Kong found the median age to be 7.05 months (range 3.9 to 13.6 months). Fever was a
significant characteristic in both studies, compared with other pathogens and as an indication of
invasive salmonellosis.

Dehydration was statistically significantly associated with more severe disease in one study.
Diarrhoea was more frequent in salmonella infection and stools were characteristically bloody
and/or mucoid, although blood in stools was not found to be indicative of invasive salmonellosis.
Vomiting was less frequent than with viral infection. Children with salmonella had longer hospital
stays and were more likely to be treated with antibiotics regardless of age.

GDG translation from evidence to recommendations
The GDG recognised that gastroenteritis in children in the UK is usually a self-limiting illness that
does not require antibiotic treatment. Most cases are due to viral enteric pathogens. Even with
the more common non-viral pathogens, treatment is usually not indicated in the UK (see below).
Antibiotic treatment is associated with a risk of adverse effects, and is a very common cause of
diarrhoea.

In most cases of childhood gastroenteritis, the healthcare professional will not know the causative
agent. Stool microbiological investigations are performed in selected cases only (see Chapter 3).
If a bacterial pathogen is identified by stool culture, the result would not be available at the time
of first presentation.
The recommendations below took into account the experience of the GDG and of the advisers,
and the limitations of the studies identified (not UK-based and with few children in the study
populations).

The GDG was aware of studies conducted in South Africa and Mexico several decades ago in
which antibiotic therapy was administered while awaiting the results of stool microbiological
investigations. There was some evidence that at that time in those studies empirical treatment
had some apparent benefit. However, the GDG did not consider that those studies were relevant
to the current UK setting.
The spectrum of pathogens commonly responsible for gastroenteritis in the UK is such that benefit
from empirical antibiotic treatment would be highly unlikely.

Recommendations regarding the treatment of specific enteric pathogens were also considered.
A series of randomised controlled trials found little evidence of clinical benefit from antibiotic
treatment for children with salmonella gastroenteritis. Indeed, there was evidence to suggest
that treatment might increase the risk of salmonella carriage. The GDG therefore concluded that

101