Page 117 - 15Diarrhoeaandvomiting

Diarrhoea and vomiting caused by gastroenteritis in children under 5 years

100 mg/kg per day ampicillin or placebo suspension in equally divided doses every 6 hours for
5 days). Fifty-seven participants received either IM or oral ampicillin and 53 received either IM
fructose or oral placebo. A random numbers list was used to assign children to treatment groups.
The treatments were prepared in foils for suspension in water before use and allocation was not
known to investigators, patients or outcome assessors. [EL = 1+]

The treatment groups were similar at baseline for age, sex, race, diarrhoeal duration, dehydration
status, clinical examinations and prior treatment with antimicrobial or antidiarrhoeal drugs. No
statistically significant differences were reported between the ampicillin and placebo groups for
the mean number of days until diarrhoea improved or ceased or for the mean number of days
until the patient became afebrile or culture negative. No patient receiving IM ampicillin relapsed
(reversion to positive cultures after a period of negative culture) after the 5 day course of therapy.
Although details of relapse in the placebo group were not presented, the authors asserted that
this finding was statistically significant (P = 0.02) and that fewer patients receiving IM antibiotics
than those receiving either placebo (P = 0.01) or oral ampicillin (P = 0.04) became short-term
salmonella carriers (culture positive any time after completion of therapy).

Evidence summary

The trials included to inform this section were generally of poor quality (except for one RCT)
with small sample sizes, and there was variation in the specific antibiotics used. Nevertheless,
there was consistent evidence from the trials suggesting that antibiotic treatment did not shorten
the duration of diarrhoea or lead to an earlier resolution of clinical symptoms. The good-quality
RCT reported that IM ampicillin protected children against relapse and carriage of salmonella
infection significantly better than placebo or oral ampicillin, while another trial (with EL = 1−)
reported an increase in the carrier rate with oral ampicillin and amoxicillin use compared with
the placebo.

7.2 Campylobacter

Evidence overview

Three RCTs were identified that compared erythromycin treatment with placebo or no treatment
for campylobacter enteritis, and they were conducted in South Africa, Canada and Peru.

The first RCT, from South Africa, 145 examined the effect of erythromycin treatment for
campylobacter-associated enteritis compared with placebo in infants aged 1–24 months. Children
admitted to hospital with diarrhoea of less than 96 hours’ duration and who had not received
any antimicrobial therapy for this illness were included in this study (n = 26). Confirmation of
Campylobacter jejuni and any other infection was from microscopic and culture examination of
stool samples. Results for those children who were infected only with Campylobacter jejuni (n = 8)
are presented here. Exclusion criteria details were not provided. Participants were randomised
to receive an oral suspension of either 40 mg/kg per day erythromycin (n = 4) or placebo (n = 4)
in divided doses for 5 days. Treatments were supplied as granules for reconstitution in pre-
coded containers such that patients, investigators and outcome assessors were blind to treatment
allocation. Follow-up was by daily physical and stool examination for 1 week. [EL = 1−]

Treatment groups were similar at baseline for age, sex, duration of diarrhoea, dehydration
severity and weight. Although the study was well conducted, causative organisms were identified
retrospectively and only eight children with Campylobacter jejuni infection alone were included,
reducing the power of the study for these results. No statistically significant differences were
found between the erythromycin and placebo groups for the mean durations of abnormal stool
frequency and consistency, vomiting, dehydration or fever.
One RCT 146 conducted in Canada recruited children of up to 12 years of age (and their household
contacts) following prospective identification of a positive, erythromycin-sensitive stool culture of
campylobacter. Children with symptoms of enteritis were recalled to hospital and were allocated
to no treatment (n = 12) or to treatment with 40 mg/kg per day erythromycin every 6 hours
for 7 days (n = 15). Exclusion criteria were presence of other enteric pathogens in the stool,
lack of symptoms and antibiotic therapy being given in the 2 weeks prior to recruitment. No

92