Table 3.

Summary of meta-analysis results for the individual MedDRA preferred term adverse events pooled across all five trials

Adverse events (AEs)		Breakdown of adverse events reporting
Meta-analysis characteristic	Total	Once in the clinical study report (CSR) and journal publication	CSR	Journal publication
Number of AE terms reported (% of total)	61	30 (49 %)	31 (51 %)	0 (0)
Direction of pooled risk effect in meta-analysis		For all 30 AEs there is agreement in direction of the pooled risk effect between the pairing of documents	• 23 (74 %) showed an increased pooled risk of AE on orlistat • four (13 %) showed no difference • four (13 %) showed increased pooled risk of AE on placebo
AE listings for when there is a change in effect including statistical significance		• Pooled risk effect was greater in journal publication for four AEs; increased defecation, oily spotting, oily evacuation, faecal incontinence • Pooled risk effect was greater in the CSR for two AEs; soft stools, faecal urgency	• two (6 %) of the 23 AEs were statistically significant; faeces discolouration, dry skin a • one (3 %) of the four AEs with increased risk on placebo was statistically significant; haemorrhoids a

Footnote:

^aThese adverse events were mild and unrelated to treatment