Table 2 GRADE summary of findings table

1 (mesalazine)

AC: not discussed

Population: OK? IBS-D population selected based on ROME criteria but paper does not define how they separated IBS-D patients into post-infective and non-post-infective group.

B: not discussed

FU: 0 pts lost to FU

OB: not a concern

L: none

Intervention: Good. Mesalazine is relevant intervention for our purpose.

Question whether outcome measure scales were validated (abdominal pain and distention score)

Outcomes: Good. Outcomes are direct (e.g. abdominal pain, frequency of stool); concern regarding length of follow up (only 30 days)

Indirect comparison: N/A

2 (ondansetron)

AC: randomized; method not discussed

Population: Good. Study population fulfilled ROME II criteria after Giardia infection.

B: subjects and clinical investigators were blinded

Intervention: Ok. Ondansetron is relevant to our study, but the patients were fasted the day before and they were fed a specific meat soup.

FU: 0 pts lost to FU

OB: not a concern

L: none

Outcome: Good. Direct outcomes such as abdominal pain/discomfort, nausea, fullness and satiety assessed; no information on chronicity; some surrogate measures (gastric emptying, max drinking capacity)

Indirect comparison: N/A

3 (prednisolone)

AC: good “A designated phar- macist generated random sequences in blocks of six, five times.”

population: good, representative; Good. Is it problematic that some patients were recruited based on ROME I after gastroenteritis, and some on clinical diagnosis?

B: patients and clinicians blinded; cell count was done blinded

FU: 5 pts lost to FU; analysis performed to take losses into account

intervention: good; Prednisone is relevant to our study.

OB: not a concern

outcomes: surrogate measures (number of serotonin-containing enterochromaffin cells, number of lamina propria T cells) are the primary and secondary outcomes. But further direct outcomes were measured (pain, looseness, urgency and frequency) for up to 3 months; pts kept symptom diaries for 6 weeks of study

L: none

Indirect comparison: N/A

4 (pregestimil)

AC: unclear

population: Ok. Limited study population to infants with lactose intolerance. Cannot be extended to adult population. Diagnosis by chronic diarrhea with gastroenteritis, so not certain if it would fulfill the IBS criteria.

B: Good “The local physician handling the patients was not aware of the formula choice.”

NB: physician decided when and if change in feeding was needed – unblinding?

intervention: good, but specific to infants

FU: 0 pts lost to FU

OB: not a concern

outcomes: good (direct outcome of days to improvement in diarrhea) but no information on longevity of effect.

L: population age, population co-morbidities, outdated?

Indirect comparison: N/A

5 (cholestyramine)

AC: not done (retrospective)

population: Not good. Population mainly screened for bile acid malabsorption (75SeHCAT). Study population is not similar to our patients; some concerns about generalizability

B: not blinded

FU: 7 pts stopped treatment, documented in study, rest of the study deals with only 18 pts who continued/responded

OB: not a concern

intervention: Not good. Cholestyramine would only be used in people who were diagnosed with bile malabsorption. Patients took cholestyramine in different doses and for different lengths of time. Patients were permitted to also take codeine phosphate and loperamide to alleviate symptoms; but this is not taken into consideration during the analysis.

L: Stopping early for benefit: 6 pts who did not improve were not followed further

Measures not validated?

outcomes: OK. Only mentions mean frequency of diarrhea as an outcome relevant to our study.

Indirect comparison: N/A

6 (cholestyramine)

AC: not done (retrospective)

population: Not good. Population mainly screened for bile acid malabsorption (75SeHCAT). Study population is not similar to our patients; some concerns about generalizability

B: not done

FU: 0 pts lost to FU

OB: not a concern

L: not a concern

intervention: Not good. Cholestyramine would only be used in people who were diagnosed with bile malabsorption. The doses of cholestyramine varied between patients.

outcomes: Ok. Only mentions mean stool frequency as an outcome relevant to our study.

Data only up to 2 weeks after start of treatment (longevity?)

Indirect comparison: N/A

7 (metronidazole)

AC: not randomized, sorted based on disease type

population: good, representative (determined by ROME II and gastroenteritis).

B: pts and clinicians not blinded

FU: 9 pts lost to FU; unclear if accounted in analysis

intervention: good

outcomes: Good: validated questionnaire that assesses pain, stool frequency, consistency, etc. and patient follow up 3 weeks post study

OB: validated symptom questionnaire used

L:

Indirect comparison: N/A

8 (mesalamine)

AC: good

population: good, representative (patients referred with clinical diagnosis of PI-IBS).

B: participants and investigators blinded (head pharmacist allocated pts and meds)

intervention: good

FU: 3 pts lost to FU; were accounted for in analysis

outcomes: Good: outcome measures of global improvements, mean change in symptoms, and quality of life.

OB: no concerns (validated measures, clearly states primary and secondary outcome measures)

L:

Indirect comparison: N/A

9 (mesalazine)

AC: not discussed

population: Good: patients diagnosed with ROME II criteria after gastroenteritis.

B: not discussed

FU: 2 pts withdrew due to AE

intervention: Good: mesalazine is relevant to our study.

OB: not a concern

L: outcomes measured not clearly stated; “weekly symptom score/global improvement” used to measure outcomes, unsure of validity

outcomes: Ok. weekly Symptom scores and global improvement at the end of treatment measured, but do not describe what they are.

Indirect comparison: N/A