Phenytoin (Epilepsy) updated on 04-22-2025

ASD (Autism spectrum disorder): Diagnosis

id	Study   Auto Straub (Controls exposed to LTG) (Phenytoin) (Epilepsy), 2026Straub (Controls unexposed, sick) (Phenytoin) (Epilepsy), 2026Madley-Dowd_SE (Phenytoin) (Controls exposed to LTG), 2024Madley-Dowd_SE (Phenytoin) (Controls unexposed, sick), 2024Madley-Dowd_UK (Phenytoin) (Epilepsy), 2024Bromley (Phenytoin) (Controls exposed to Lamotrigine, sick), 2008Bromley (Phenytoin) (Controls unexposed, disease free), 2008Bromley (Phenytoin) (Controls unexposed, sick), 2008	Lib. in paper	Exposition period  Auto during pregnancy (anytime or not specified)1st trimester2nd trimester3rd trimester2nd and/or 3rd trimesterthroughout pregnancyearly pregnancydays before deliverylate pregnancypreconception-onlypericonceptionalpreconceptional for kinetic reasonat least 1st trimester3 months (or more) before pregnancy or during pregnancy1st and 2nd trimester3 months or more before pregnancy or1st trimesterPaternal exp. genotoxic window only (0–3 months preconception)Paternal exposure: epigenetic window only (3–xxx months preconception)Paternal exposure: genotoxic/epigenetic window (0–xxx months preconception)Paternal exposure: other window	Study type  All studies Case control excluded Only case control	Control type    Auto unexposed (general population or NOS)unexposed, sickunexposed, disease freeexposed to other treatment, sicknot specifiedsiblingwhatever (meta-analysis)	Tags	OR 95%CI	x1/n1 x0/n0	no cases	no exposed	ROB	Ref.
S19463 R82983	Straub (Controls exposed to LTG) (Phenytoin) (Epilepsy), 2026	Autism spectrum disorder - ICD-10 codes: F84.x (except F84.2, F84.3) - Mean follow up (full cohort): 3.4 years	during pregnancy (anytime or not specified)	retrospective cohort (claims database)	exposed to other treatment, sick excluded	Adjustment: No	0.90 [0.47;1.71] C excluded (control group)	13/1,390   34/3,273	47	1,390		ref
S19473 R83106	Straub (Controls unexposed, sick) (Phenytoin) (Epilepsy), 2026	Autism spectrum disorder - ICD-10 codes: F84.x (except F84.2, F84.3) - Mean follow up (full cohort): 3.4 years	during pregnancy (anytime or not specified)	retrospective cohort (claims database)	unexposed, sick	Adjustment: Yes	0.54 [0.29;0.98]	13/1,390   106/8,887	119	1,390		ref
S17393 R72922	Madley-Dowd_SE (Phenytoin) (Controls exposed to LTG), 2024	Autism - ICD-8/9/10 codes - Mean 5-20 years old (according groups)	during pregnancy (anytime or not specified)	population based cohort retrospective	exposed to other treatment, sick excluded	Adjustment: No	2.20 [0.98;4.92] C excluded (control group)	7/147   53/2,383	60	147		ref
S17394 R72926	Madley-Dowd_SE (Phenytoin) (Controls unexposed, sick), 2024	Autism - ICD-8/9/10 codes - Mean 5-20 years old (according groups)	during pregnancy (anytime or not specified)	population based cohort retrospective	unexposed, sick	Adjustment: Yes	0.80 [0.38;1.70]	7/147   425/10,769	432	147		ref
S17253 R72210	Madley-Dowd_UK (Phenytoin) (Epilepsy), 2024	Autism - ICD10 codes - Median follow up 10.75 (7.82-13.62) years old	during pregnancy (anytime or not specified)	retrospective cohort (claims database)	unexposed, sick	Adjustment: Yes	1.19 [0.17;8.52]	-/45   78/4,075	-	45		ref
S6601 R18099	Bromley (Phenytoin) (Controls exposed to Lamotrigine, sick), 2008	Autism spectrum disorder (a majority aged 6 and above)	early pregnancy	prospective cohort	exposed to other treatment, sick excluded	Adjustment: No	5.38 [0.30;95.06] C excluded (control group)	1/9   1/44	2	9		ref
S6594 R18089	Bromley (Phenytoin) (Controls unexposed, disease free), 2008	Autism spectrum disorder (a majority aged 6 and above)	early pregnancy	prospective cohort	unexposed, disease free excluded	Adjustment: No	13.88 [1.30;148.33] C excluded (control group)	1/9   3/336	4	9		ref
S6598 R18094	Bromley (Phenytoin) (Controls unexposed, sick), 2008	Autism spectrum disorder (a majority aged 6 and above)	early pregnancy	prospective cohort	unexposed, sick	Adjustment: No	18.60 [0.69;500.42] C	1/9   0/47	1	9		ref
					Total	4 studies	0.80 [0.39;1.62]		552	1,591
x1: number of endpoints among exposed, n1: number of exposed; x0: number of endpoints among non exposed, n0: number of non exposed; C: calculated odds ratio from numbers of events and effectives

Forest plot

Sensitivity analysis

Publication bias and p-hacking diagnosis

Sub-groups analysis using all included studies

excluded 6601, 6594, 17393, 19463