Study characteristics

Methods

Study design: randomised controlled trial, double‐blind placebo‐controlled

Study grouping: cross‐over, each arm lasted for 6 months

Carry‐over treatment effect accounted for: yes. Measurements were recorded twice for each 6‐month period. Investigators took an average of the 2 measurements, they then discarded the measurements from the first 3 months, to account for any cumulative effect of the drug. There was no formal washout period

Participants

Inclusion criteria: diagnosis of CF (by sweat test > 60 mmol/L); 1 or more episodes of DIOS in preceding 12 months (based on clinical or radiological evidence and on successful treatment with intestinal lavage, oral N‐acetylcysteine or both)

Exclusion criteria: peptic ulcer disease, inflammatory bowel disease, anatomic intestinal obstruction, serious cardiovascular, neurological, renal or hepatic disease, severe pulmonary dysfunction (maximal mid‐expiration flow rate < 25% normal) pregnancy, regular use of metoclopramide, domperidone or anticholinergic drug

Pre‐treatment: no significant differences in the clinical characteristics of 2 groups (cross‐over)

Baseline characteristics

Age, mean (SD), range: 21.0 (5.9) years, 12.9 to 34.9 years old

Mean (SD) duration of DIOS symptoms: 4.2 (3.1) years

Mean (SD) height percentile: 40.6 (26.3)

Mean (SD) weight percentile: 40.3 (23.2)

Mean (SD) weight for height: 100.3 (9.5) % of ideal

Mean (SD) FVC (% predicted): 80.9 (23.4)%

Mean (SD) FEV₁ (% predicted): 78.2 (28.1)%

Interventions

Cisapride: oral tablets 7.5 mg 3 times daily (before meals) for participants weighing between 40 kg and 50 kg, 10 mg 3 times daily (before meals) for participants weighing over 50 kg

Placebo (identical in appearance to active treatment): oral tablets 7.5 mg 3 times daily (before meals) for participants weighing between 40 kg and 50 kg, 10 mg 3 times daily (before meals) for participants weighing over 50 kg

Outcomes

Gastrointestinal symptoms: a total of 10 symptoms scored 1 to 10 (heartburn, flatulence, regurgitation, fullness, abdominal distension, abdominal pain, diarrhoea, nausea, vomiting, anorexia) and also added to give a total score (20 to 100) where a lower score is better; severity (none, not limiting daily activities, limits daily activities, daily activities not possible) and frequency (never, less than once a week, more than once a week or daily) of symptoms recorded and accounted for in final score for each item

Anthrompometric measurements: mid‐arm circumference; skin fold thickness; change in weight; abdominal circumference

Alteration in global symptoms: measured by both physician and participant, from 'worse' to 'symptom‐free' using scores from ‐1 to +3 (higher is better)

Pulmonary function: FEV₁ % predicted, frequency of pulmonary exacerbations, number of hospital admissions

Nutrient intake: calories (kcal/day), fat (mmol/day)

Stool composition: stool weight (g/day), faecal water content (%), faecal fat (%), calories malabsorbed (%), faecal bile acids (mmol/day), faecal chymotrypsin (10 x 300 units/day)

Supine abdominal radiography

Intestinal lavage therapy

Identification

Sponsorship source: supported by Janssen Pharmaeutica Inc., Canada. No sponsorship mentioned

Country: Canada

Setting: single centre (tertiary centre) at the University of Toronto and the research institute at the Hospital for Sick Children, Toronto

Comments: approved by the human subjects review committee of the Hospital for Sick Children, Toronto. Janssen Pharmaceutica supplied cisapride and placebo tablets

Contact author's name: Sibylle Koletzko

Institution: Children's Hospital, Heinrich Heine University, Dusseldorf, Federal Republic of Germany

Email: [email protected]‐muenchen.de

Address: Lindwurmstraße 4, 80337 München, Germany

The authors did not declare any conflicts of interest

Notes

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Random allocation stated but no specific sequence generation methodology described.

Allocation concealment (selection bias)

Unclear risk

No method described for concealment of allocation.

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

"Double‐blind" trial. Placebo tablets were identical in taste and appearance to cisapride.

Blinding of outcome assessment (detection bias)

Unclear risk

Only 3 outcomes stated that the investigators were blinded.

1. Gastrointestinal symptom scores ‐ the blinded participants acted as their own assessors.

2. Assessment of supine abdominal radiographs ‐ a paediatric radiologist judged these in a blind fashion.

3. Assessment of nutritional intake and stool collection ‐ the blinded participants recorded their own intake and investigators also worked in a blind fashion.

Although blinded participants scored their own global symptoms, physicians assessed them too and we could not make the assumption that the physicians were blinded.

There was no mention of blinding for the other outcomes: anthropometric measurements, number of hospital admissions, pulmonary function and frequency of pulmonary infections, laboratory tests, abdominal circumference and intestinal lavage therapy.

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

No missing data and outcomes fully reported: gastrointestinal symptom scores, global symptom scores and intestinal lavage therapy

Insufficient information to judge whether outcome data are missing: anthropometric measurements, adverse effects, pulmonary function, frequency of pulmonary exacerbations, number of hospital admissions, radiological signs of DIOS and laboratory values

Data missing: for nutritional intake and stool losses, only present data for 10/17 participants, reasons given for missing data but this would have had a big impact on the effect size, since the sample size was small anyway; no mention of ITT analysis, no mention of treating per protocol

Selective reporting (reporting bias)

High risk

No access to trial protocol so not possible to compare planned outcomes to outcomes reported in paper.

Anthropometric measurements: stated in the methods section that "at the end of each treatment period, difference weight and percentage of ideal weight for height, as well as the difference in weight change during the two periods, were compared by means of t tests", however, these changes and results of t tests were not specifically reported in the results, incomplete report means it can not be entered into a meta‐ analysis.

Pulmonary function testing and X‐ray findings: although supposed to be measured "at the end of baseline and 6 month periods" were reported incompletely so cannot be entered into a meta‐analysis.

Cisapride levels: in the methods section, it specified that they would determine cisapride levels "at the end of baseline and 6 month periods" but this was not reported in the results.

Number of hospital admissions: specified as an outcome measure in the methods section, but not reported in the results.

Laboratory test results (blood and urine analysis): incompletely reported so that they could not be entered into a meta‐analysis.

Other bias

Unclear risk

None known, but cannot be sure, there is no known sponsorship source but the trial was described as being supported by Janssen Pharmaceuticals.