Table 4. Sensitivity analyses–Stratified analyses (with exposure to BZDs studied as binary factor).

			Adjusted HR§ (95% CI)		P Value
		No. of Patients	Unexposed	Exposed to BZDs
Overall
Baseline (fixed) exposure to BZDs		246940	1 (ref)	1.25 (1.19–1.31)	<0.0001
Time-dependent exposure to BZDs		246940	1 (ref)	1.29 (1.23–1.36)	<0.0001
Analyses stratified on
Sex	Male	104217	1 (ref)	1.24 (1.16–1.33)	<0.0001
	Female	142723	1 (ref)	1.25 (1.18–1.33)	<0.0001
Age group	40–59	43748	1 (ref)	1.27 (1.16–1.40)	<0.0001
	60–74	109991	1 (ref)	1.25 (1.17–1.34)	<0.0001
	≥ 75	93201	1 (ref)	1.22 (1.13–1.31)	<0.0001
Exposure to antidepressant	No	201142	1 (ref)	1.24 (1.18–1.31)	<0.0001
	Yes	45798	1 (ref)	1.25 (1.12–1.39)	<0.0001
Hospital sector	Public	83925	1 (ref)	1.26 (1.16–1.35)	<0.0001
	Private	163015	1 (ref)	1.24 (1.18–1.31)	<0.0001
Analyses according to
class of BZDs exposure at baseline
Exposure to Anxiolytic BZDs only		43128	1 (ref)	1.23 (1.16–1.30)	<0.0001
Exposure to Hypnotic BZDs only		36402	1 (ref)	1.28 (1.21–1.36)	<0.0001
Exposure to both BZDs		24163	1 (ref)	1.37 (1.28–1.47)	<0.0001
Exposure to other class of BZDs		16519	1 (ref)	1.28 (1.18–1.39)	<0.0001

§ Hazard Ratios were adjusted for: sex, age at implantation, diabetes mellitus, obesity, Parkinson's disease, immunodeficiency, exposure to antidepressant, oral corticosteroid, antiosteoporotics, psychostimulant, antipsychotic, antiepileptic (non-BZD), anxiolytic or hypnotic (non-BZD), public or private sector, center activity, duration of stay, cement type and bearing surface, in multivariate Cox models.