Survival after Abdominoperineal and Sphincter-Preserving Resection in Nonmetastatic Rectal Cancer: A Population-Based Time-Trend and Propensity Score-Matched SEER Analysis
Table 3
Bias for abdominoperineal resection.
Logistic regression in raw data
Patient characteristics after exact propensity score matching
OR (95% CI)
Total
APR
CAA
Tumor stage (AJCC 6th ed.)
Stage I
Reference
<0.001
295.2 (28.1%)
194 (28.1%)
101.2 (28.1%)
Stage IIA
1.21 (0.98–1.51)
398.7 (38.0%)
262 (38.0%)
136.7 (38.0%)
Stage IIB
3.53 (1.91–7.31)
1.5 (0.1%)
1 (0.1%)
0.5 (0.1%)
Stage IIIA
0.90 (0.65–1.25)
36.5 (3.5%)
24 (3.5%)
12.5 (3.5%)
Stage IIIB
1.42 (1.12–1.80)
261.7 (24.9%)
172 (24.9%)
89.7 (24.9%)
Stage IIIC
1.52 (1.14–2.04)
56.3 (5.4%)
37 (5.4%)
19.3 (5.4%)
Retrieved regional lymph nodes
<12
Reference
0.155
503.7 (48.0%)
331 (48.0%)
172.7 (48.0%)
12+
0.89 (0.75–1.05)
546.3 (52.0%)
359 (52.0%)
187.3 (52.0%)
Grading
G1
Reference
0.072
22.8 (2.2%)
15 (2.2%)
7.8 (2.2%)
G2
1.22 (0.92–1.61)
969.3 (92.3%)
637 (92.3%)
332.3 (92.3%)
G3/4
1.34 (0.95–1.88)
41.1 (3.9%)
27 (3.9%)
14.1 (3.9%)
Unknown
1.69 (1.12–2.57)
16.7 (1.6%)
11 (1.6%)
5.7 (1.6%)
Radiation
None
Reference
<0.001
228.3 (21.7%)
150 (21.7%)
78.3 (21.7%)
Before surgery
1.75 (1.43–2.13)
768.5 (73.2%)
505 (73.2%)
263.5 (73.2%)
After surgery
1.58 (1.20–2.10)
53.3 (5.1%)
35 (5.1%)
18.3 (5.1%)
Year
2005
Reference
0.155
176.5 (16.8%)
116 (16.8%)
60.5 (16.8%)
2006
1.05 (0.79–1.41)
152.2 (14.5%)
100 (14.5%)
52.2 (14.5%)
2007
0.92 (0.69–1.23)
149.1 (14.2%)
98 (14.2%)
51.1 (14.2%)
2008
1.19 (0.88–1.62)
94.3 (9.0%)
62 (9.0%)
32.3 (9.0%)
2009
0.82 (0.62–1.09)
172 (16.4%)
113 (16.4%)
59 (16.4%)
2010
0.98 (0.73–1.31)
172 (16.4%)
113 (16.4%)
59 (16.4%)
2011
0.81 (0.61–1.08)
133.9 (12.8%)
88 (12.8%)
45.9 (12.8%)
Age
<50
Reference
<0.001
132.4 (12.6%)
87 (12.6%)
45.4 (12.6%)
50–64
1.20 (0.97–1.48)
544.8 (51.9%)
358 (51.9%)
186.8 (51.9%)
65–79
1.86 (1.48–2.33)
343.9 (32.8%)
226 (32.8%)
117.9 (32.8%)
80+
2.76 (1.96–3.94)
28.9 (2.8%)
19 (2.8%)
9.9 (2.8%)
Gender
Male
Reference
0.066
833.9 (79.4%)
548 (79.4%)
285.9 (79.4%)
Female
0.86 (0.73–1.01)
216.1 (20.6%)
142 (20.6%)
74.1 (20.6%)
Ethnicity
Caucasian
Reference
0.008
1033.3 (98.4%)
679 (98.4%)
354.3 (98.4%)
African-American
0.71 (0.55–0.94)
7.6 (0.7%)
5 (0.7%)
2.6 (0.7%)
Other/unknown
1.30 (0.97–1.77)
9.1 (0.9%)
6 (0.9%)
3.1 (0.9%)
Marital status
Married
Reference
0.006
897.8 (85.5%)
590 (85.5%)
307.8 (85.5%)
Single/widowed
1.29 (1.06–1.58)
106.5 (10.1%)
70 (10.1%)
36.5 (10.1%)
Other/unknown
1.33 (1.06–1.69)
45.7 (4.3%)
30 (4.3%)
15.7 (4.3%)
After exclusion of 3,650 patients for the exact propensity score matching, no bias was observed in the remaining 1,050 patients for APR versus CAA (all ). logistic regression with the odds ratio (OR) for APR in the original raw data set (). ratio tests. for comparison of APR versus CAA in weighted Chi-square tests after exact weighted propensity score matching (). Weighted matching causes decimals for the number of patients in the group with CAA.
