Page 118 - 15Diarrhoeaandvomiting

Antibiotic therapy

details of the randomisation process were presented, and assessors were not blinded to treatment
allocation. Children were followed up until the entire household had three consecutive negative
(weekly) stool samples. [EL = 1−]

The trial reported no statistically significant differences in the mean duration of diarrhoea
experienced by participants receiving erythromycin or no treatment. The range in number of days
with diarrhoea was 1–6 days in the erythromycin group and 1–15 days in the group receiving no
treatment. A statistically significant difference was found in the mean number of days until the
first negative culture between those children receiving erythromycin (2.0 ± 1.3 days) and those
receiving no treatment (16.8 ± 12.5 days) (P < 0.001).
The third RCT, from Peru, 147 examined the effects of early erythromycin treatment for
campylobacter-associated enteritis compared with placebo in children aged 3–60 months
brought in as outpatients for treatment of acute diarrhoea (n = 24). Participants had five or more
loose stools a day with gross blood or mucus for no longer than 5 days and had not received
antibiotics for another illness. Children with clinical signs of dehydration or who were under the
third percentile for weight/length ratio (US National Center for Health Statistics standard) or who
had had a separate episode of diarrhoea in the previous fortnight were excluded. Participants
were randomised to receive 50 mg/kg per day erythromycin oral suspension (n = 14) or placebo
oral suspension (n = 10) in four doses for 5 days. Treatments were randomised and pre-coded
by the manufacturers such that patients, investigators and outcome assessors were blind to
treatment allocation. Allocation to treatment groups was prior to stool culture confirmation of
campylobacter. Follow-up by stool culture and parental reporting of symptoms was performed
for 5 days. [EL = 1+]

The mean duration of diarrhoea was significantly lower in the patients receiving erythromycin
(2.4 ± 0.4 days) compared with placebo (4.2 ± 0.3 days), (WMD −1.80 days; 95% CI −2.08 to
−1.52 days). However, the number of patients with normal stools at 5 days in the erythromycin
group (13/14) was not significantly different from those receiving placebo (5/10) (RR 1.86; 95% CI
0.98 to 3.51). The mean number of days until last positive stool culture was significantly lower for
those receiving erythromycin (0.5 ± 0.3 days) compared with the placebo group (2.2 ± 0.6 days)
(WMD −1.70 days; 95% CI −2.10 to −1.30 days), but no statistically significant difference was
found in the proportion of patients in each group with a positive stool culture at 5 days.

Evidence summary

Of the three available small RCTs, one trial had insufficient power to detect any statistically
significant treatment differences in a small subgroup who received erythromycin or placebo
for treatment of Campylobacter jejuni enteritis. There were conflicting results from the two
remaining studies. One trial [EL = 1−] found no difference between the groups in mean duration
of diarrhoea. The second trial [EL = 1+] found that the mean duration of diarrhoea was shorter
with erythromycin treatment, although the ‘diarrhoeal cure’ rate at 5 days was similar between
treatment groups. The difference in mean duration of diarrhoea might be explained by the
second trial’s early recruitment of participants to treatment groups without awaiting stool culture
confirmation of campylobacter. Thus patients would be less likely to have had long episodes of
diarrhoea prior to treatment and might be more uniform in severity of illness during the study.
Both studies found that erythromycin treatment demonstrated antibacterial efficacy by either
reducing the mean number of days until first negative stool culture or the last positive culture.
However, it could not be established in the EL = 1+ trial whether erythromycin treatment caused
fewer patients to excrete campylobacter at day 5 compared with placebo.

7.3 Yersinia

Evidence overview

One RCT was identified. This study 148 was conducted in Canada and examined the treatment
of yersinia enteritis with trimethoprim/sulfamethoxazole compared with placebo (n = 45)
Participants were children younger than 15 years with symptomatic enteritis. Prior to recruitment,
stool samples from participants had been positively cultured for yersinia. Participants and their
household contacts were followed until all had three consecutive negative (weekly) stool

93