Table 1.

Overall infertility measure over time in men with epilepsy or bipolar disorder exposed and unexposed to valproate

Outcome	Outcome assessment period after index	Cohort	Patient count before matching	Patient count after matching	Patients excluded because they had the outcome prior to follow-up	Remaining patient denominator analysed	Patients with outcome	Risk	Hazard ratio (95% CI)	Log-Rank p-value
Composite outcome: -male infertilitya -testicular hypofunctionb or atrophyc -low sperm concentration, motility, vitality, normal forms, or semen volumed	Lifetime	Epilepsy or bipolar disorder exposed to valproate	91,917	78,971	617	78,354	942	1.2%	0.932 (0.849–1.024)	0.142
		Epilepsy or bipolar disorder not exposed to valproate	535,803	78,971	675	78,296	838	1.1%
	30 days	Epilepsy or bipolar disorder exposed to valproate	90,571	77,826	621	77,205	30	0.0%	0.741 (0.461–1.189)	0.213
		Epilepsy or bipolar disorder not exposed to valproate	524,095	77,826	638	77,188	40	0.1%
	60 days	Epilepsy not exposed to valproate	89,759	77,099	600	76,499	48	0.1%	0.703 (0.485–1.018)	0.061
		Bipolar disorder exposed to valproate	514,623	77,099	619	76,480	67	0.1%
	90 days	Bipolar disorder not exposed to valproate	88,702	76,278	583	75,695	73	0.1%	0.848 (0.620– 1.160)	0.302
		Epilepsy or bipolar disorder exposed to valproate	513,666	76,278	653	75,625	84	0.1%
	180 days	Epilepsy or bipolar disorder not exposed to valproate	88,687	75,985	590	75,395	131	0.2%	0.951 (0.747– 1.210)	0.681
		Epilepsy exposed to valproate	514,702	75,985	638	75,347	133	0.2%
	360 days	Epilepsy not exposed to valproate	84,068	71,327	511	70,816	191	0.3%	0.910 (0.746, 1.110)	0.353
		Bipolar disorder exposed to valproate	496,609	71,327	518	70,809	199	0.3%
	2 years	Bipolar disorder not exposed to valproate	83,222	70,346	532	70,346	333	0.5%	0.919 (0.790–1.069)	0.272
		Epilepsy or bipolar disorder exposed to valproate	478,058	70,328	550	70,328	338	0.5%
	5 years	Epilepsy or bipolar disorder not exposed to valproate	87,720	75,233	583	74,650	633	0.8%	0.949 (0.848– 1.061)	0.358
		Epilepsy exposed to valproate	503,955	75,233	600	74,633	596	0.8%
	10 years	Epilepsy not exposed to valproate	82,046	70,425	527	69,898	747	1.1%	0.942 (0.849– 1.045)	0.261
		Bipolar disorder exposed to valproate	469,361	70,425	570	69,855	677	1.0%

^a Male infertility (as defined by the World Health Organisation (WHO) through the International Classification of Diseases 10^th Revision Clinical Modification (ICD-10-CM) code N46—which includes N46.0 (Azoospermia), N46.1 (Oligospermia), N46.8 (Other male infertility—capturing where male infertility has been identified but the cause does not fit into any of the other specified (coded) aetiological categories), N46.9 (Male infertility, unspecified—capturing where male infertility that has been identified but the cause is unclear)⁵⁵, and 606 (Infertility, male—which is the ICD-9-CM code equivalent of N46, allowing healthcare systems using older coding to still be represented).

^b Testicular hypofunction (ICD-10-CM code E29.1)—which includes defective biosynthesis of testicular androgen, 5-delta-reductase deficiency (with male pseudohermaphroditism), and testicular hypogonadism.

^c Testicular atrophy (ICD-10-CM code N50.0)—a code which clinicians have the discretion to use when finding evidence of a pathological reduction in size of the testicles, e.g., using an orchidometer or ultrasound).

^d Low sperm concentration (< 16 × 10⁶/ml semen semen), low sperm motility ( < 42% total motility), low sperm vitality ( < 54% viable), low normal forms ( < 4% of sperm have a normal morphology), or low semen volume ( < 1.4 mL). CI = Confidence interval.

N.B: Statistics = Survival analysis using Cox-proportional hazard models with 95% CIs and two-sided Log-Rank p-values with a 0.05 level of significance. Slight variation in overall patient count denominators over time reflects the fact that each analysis was conducted as an independent query on a live, dynamic dataset, where patient inclusion can vary slightly due to real-time clinical updates, such as newly recorded diagnoses, additional data accrual, or changes in data completeness across healthcare network sites. Overall trends and conclusions have not been impacted.