It is very uncertain whether comparison 3: anticonvulsant levetiracetam changes dropouts due to adverse events (acute phase) compared with placebo (very low certainty).
It is very uncertain whether comparison 4: antipsychotics changes dropouts due to adverse events (acute phase) compared with placebo (very low certainty).
It is very uncertain whether comparison 5: benzodiazepines changes dropouts due to adverse events (acute phase) compared with placebo (very low certainty).
It is very uncertain whether comparison 1: 5HT1A partial agonists changes dropouts due to adverse events (acute phase) compared with placebo (very low certainty).
It is very uncertain whether comparison 8: NARIs changes dropouts due to adverse events (acute phase) compared with placebo (very low certainty).
It is very uncertain whether comparison 9: NaSSAs changes dropouts due to adverse events (acute phase) compared with placebo (very low certainty).
It is very uncertain whether comparison 10: RIMAs changes dropouts due to adverse events (acute phase) compared with placebo (very low certainty).
Comparison 12: SNRIs probably leads to higher dropouts due to adverse events (acute phase) compared with placebo (moderate certainty).
It is very uncertain whether comparison 2: anticonvulsants/GABAs changes dropouts due to adverse events (acute phase) compared with placebo (very low certainty).
Comparison 13: SSRIs may leads to higher dropouts due to adverse events (acute phase) compared with placebo (low certainty).
Comparison 14: GW876008 showed no clear difference in dropouts due to adverse events (acute phase) compared with placebo (low certainty).
Anticonvulsants showed no clear difference in adverse events—withdrawn for medical reasons compared with placebo (moderate certainty).
It is very uncertain whether leukotriene receptor antagonist changes withdrawals due to adverse events at 26 weeks compared with placebo (very low certainty).
Viaskin milk patch showed no clear difference in withdrawals due to adverse events at 44 weeks compared with placebo (low certainty).
Summary of findings table ‐ Immunoglobulins showed no clear difference in adverse events of any severity or seriousness compared with placebo (low certainty).
Hyperimmune immunoglobulin from human plasma probably leads to higher adverse events at grades 3 to 4 up to 28 days compared with saline placebo (moderate certainty).