TABLE 3.

Adjusted Effect of Sedation Type on Post-TAVR Outcomes Using Instrumental Variable Analysis and Propensity Scores

Outcome of Interest	Instrumental Variable Analysis Model				Propensity Score-Based Model
	Conscious Sedation	General Anesthesia	Absolute Risk Difference	p Value	Conscious Sedation	General Anesthesia	Absolute Risk Difference	p Value
Primary endpoint
In-hospital death (%)	1.1	1.3	−0.2 (−0.4 to −0.0)	0.010	0.9	1.5	−0.7 (−0.8 to −0.5)	<0.001
Secondary endpoints
30-day death (%)*	2.0	2.5	−0.5 (−0.7 to −0.1)	<0.001	1.8	2.7	−0.9 (−1.1 to −0.7)	<0.001
Device success (%)	97.3	97.3	−0.0 (−0.2 to 0.3)	0.947	97.5	97.3	0.3 (0.1 to 0.5)	<0.001
Length of hospital stay (days)†	3.5	4.3	−0.7 (−0.8 to −0.7)	<0.001	3.5	4.2	−0.7 (−0.7 to −0.7)	<0.001
Use of inotropic drugs (%)	27.0	31.3	−4.2 (−5.1 to −3.5)	<0.001	26.2	32.9	−6.7 (−7.2 to −6.1)	<0.001
Discharge to home (%)	88.9	86.1	2.8 (2.3 to 3.4)	<0.001	89.4	85.9	3.4 (3.1 to 3.9)	<0.001
Falsification endpoint
In-hospital access site vascular complication (%)	3.0	3.1	−0.1 (0.4 to 0.2)	0.516	2.8	3.4	−0.6 (−0.8 to −0.4)	<0.001

For the instrumental variables model, predicted event rates were calculated on the basis of the average patient in the entire cohort receiving conscious sedation (conscious sedation column) or general anesthesia (general anesthesia column). For the propensity score-based model, predicted event rates were calculated on the basis of the average predicted probability when all patients receive conscious sedation (conscious sedation column) and general anesthesia (general anesthesia column). Absolute risk differences and p values are provided with 1,000 samples used to calculate bootstrap confidence intervals.

^*The percentage of patients with missing 30-day mortality data was 6.8%.

^†Fewer than 1% (0.95%) of patients were censored for length of stay >25 days.

TAVR = transcatheter aortic valve replacement.