Table 2.

Association between initiation of antidepressants of interest versus selective serotonin reuptake inhibitors (SSRIs) and hospitalization due to serious liver injury (main definition of event, follow-up 6 months)

Antidepressant	Initiators (n = 4,966,825)	Events (n = 382)	Event incidence (per 100,000 person-years)a	Hazard ratio (95% CI)
				Crude	Adjustedb,c
SSRIs	3,543,559	258	19.2	1.00 [Reference]	1.00 [Reference]
Venlafaxine	436,155	36	22.2	1.15 (0.81–1.63)	1.17 (0.83–1.64)
Milnacipran	37,577	0	0.0
Duloxetine	247,250	12	12.6	0.70 (0.39–1.24)	0.54 (0.28–1.02)
Mianserin	293,484	29	21.5	1.43 (0.97–2.10)	0.90 (0.58–1.41)
Mirtazapine	128,593	15	32.8	1.65 (0.98–2.77)	1.17 (0.67–2.02)
Tianeptine	181,289	24	31.6	1.93 (1.27–2.94)	1.35 (0.82–2.23)
Agomelatine	98,918	8	24.6	1.18 (0.58–2.38)	1.07 (0.51–2.23)

CI confidence interval, SSRIs selective serotonin reuptake inhibitors

^aStandardized on sex and age categories (< 50 or ≥ 50 years), SSRI initiators served as the reference group

^bInverse probability of treatment weighting considering the following covariates: inclusion year, sex, age, deprivation index and complementary universal health insurance at inclusion; diabetes, heart failure, chronic renal failure, measurable history of smoking, morbid obesity and measurable history of substance abuse up to 12–36 months before inclusion; reimbursements for drugs potentially associated with hepatotoxicity and immunosuppressant drugs up to 6–12 months before inclusion; aminotransferase testing at inclusion ± 1 month

^cAdditional adjustment on age categories, prescriber specialty at inclusion; psychiatric history in 12 months before inclusion; drugs potentially associated with hepatotoxicity, other antidepressants and aminotransferase testing reimbursed during follow-up