Table 4.

Multivariable analysis for multiple complications/death at pelvic exenteration.

Number	Multiple complications§		Perioperative death
	OR (95%CI)	P-value	OR (95%CI)	P-value
Year†	1.100 (1.075–1.126)	<0.001
Race/ethnicity		<0.001*
White	1
Black	0.559 (0.402–0.778)	0.001
Hispanic	1.011 (0.722–1.415)	0.949
Others	1.343 (0.910–1.984)	0.138
Missing	1.347 (1.052–1.724)	0.018
Obesity
No	1
Yes	1.519 (1.162–1.987)	0.002
Charlson Comorbidity Index†	1.197 (1.156–1.240)	<0.001	1.617 (1.434–1.823)	<0.001
Median household income		<0.001*
≤$38,999	1.171 (0.901–1.522)	0.238
$39,000–$47,999	1
$48,000–$62,999	1.572 (1.241–1.992)	<0.001
≥$63,000	1.418 (1.097–1.833)	0.008
Missing	0.598 (0.375–0.955)	0.031
Primary expected payer		<0.001*		<0.001*
Medicare	1		1
Medicaid	0.639 (0.488–0.838)	0.001	na	0.990
Private including HMO	0.769 (0.614–0.962)	0.022	0.155 (0.069–0.351)	<0.001
Others/missing	1.253 (0.885–1.774)	0.204	0.603 (0.215–1.691)	0.336
Hospital bedsize		0.001*		<0.001*
Small	1.484(1.016–2.169)	0.041	1.071 (0.412–2.785)	0.887
Medium	1		1
Large	1.588 (1.245–2.025)	<0.001	0.231 (0.111–0.481)	<0.001
Urinary diversion
No	1
Yes	2.251 (1.799–2.816)	<0.001
Vaginal reconstruction
No	1
Yes	1.390 (1.127–1.713)	0.002
Cancer type		0.005*		0.034*
Cervical	1		1
Uterine	1.123 (0.844–1.496)	0.426	0.949 (0.378–2.284)	0.911
Vaginal	1.092 (0.877–1.359)	0.432	2.642 (1.094–6.379)	0.031
Vulvar	1.697 (1.267–2.273)	<0.001	0.824 (0.246–2.758)	0.753

Binary logistic regression model for analysis. All the covariates with P < 0.05 on univariable analysis were entered in the initial model. Conditional backward method was used to retain covariates with P < 0.05 in the final model. All the covariates in the list were entered in the final model.

^§any two or more complications listed in Table 3.

^†continuous variable.

^*P-value for interaction. Abbreviations: OR, odds ratio; and CI, confidence interval.