Clinical prediction model for inguinal hernia recurrence risk after transabdominal preperitoneal hernia repair surgery in elderly patients

Table 1 Comparison of demographic characteristics and baseline disease data between the two groups, n (%)/mean ± SD

Characteristic	Recurrence group (n = 40)	Nonrecurrence group (n = 260)	χ2/t	P value
Sex			6.909	0.006
Male	38 (95.00)	200 (76.92)
Female	2 (5.00)	60 (23.08)
Age			0.934	0.627
60-70 years	12 (30.00)	61 (23.46)
71-80 years	24 (60.00)	165 (63.46)
> 80 years	4 (10.00)	34 (13.08)
Preoperative BMI (kg/m2)	24.84 ± 2.04	22.07 ± 2.16	7.605	< 0.001
Smoking history	18 (45.00)	26 (10.00)	23.276	< 0.001
Drinking history	2 (5.00)	21 (8.08)	0.389	0.533
Hypertension	17 (42.50)	119 (45.77)	0.150	0.699
Diabetes mellitus	6 (15.00)	32 (12.31)	0.227	0.634
Coronary heart disease	3 (7.50)	20 (7.69)	0.077	0.782
Hyperlipidemia	20 (50.00)	136 (52.31)	0.074	0.786
Unilateral hernia	34 (85.00)	206 (79.23)	0.721	0.396
Bilateral hernia	6 (15.00)	54 (20.77)
Chronic constipation	12 (30.00)	25 (9.62)	11.504	0.001
Hernia type			0.366	0.833
Indirect hernia	31 (77.50)	201 (77.31)
Direct hernia	6 (15.00)	33 (12.69)
Combined hernia	3 (7.50)	26 (10.00)
Hernia sac adhesion			0.752	0.386
Mild	35 (87.50)	213 (81.92)
Moderate-to-severe	5 (12.50)	47 (18.08)
Incarcerated hernia	14 (35.00)	18 (6.92)	25.808	< 0.001

BMI: Body mass index.

Table 2 Comparison of surgical outcomes between the two groups, n (%)/mean ± SD

General data	Recurrence group (n = 40)	Nonrecurrence group (n = 260)	χ2/t	P value
Emergency surgery	18 (45.00)	118 (45.38)	0.002	0.964
Anesthesia method			0.140	0.708
Intraspinal anesthesia	26 (65.00)	161 (61.92)
Local infiltration anesthesia	14 (35.00)	99 (38.08)
Operation time (minutes)	78.81 ± 12.14	80.06 ± 14.75	0.510	0.611
Mesh type			14.535	< 0.001
Polypropylene mesh	11 (27.50)	18 (6.92)
Polytetrafluoroethylene mesh	29 (72.50)	242 (93.08)
Mesh size			0.038	0.845
8 cm × 12 cm	6 (15.00)	36 (13.85)
10 cm × 15 cm	8 (20.00)	58 (22.31)
12 cm × 15 cm	18 (45.00)	121 (46.54)
15 cm × 15 cm	8 (20.00)	45 (17.31)
Chronic postoperative pain	3 (7.50)	28 (10.77)	0.400	0.527
Postoperative hematoma	4 (10.00)	30 (11.54)	0.003	0.986
Postoperative infection	1 (2.50)	3 (1.15)	0.002	0.961

Table 3 Assignment analysis

Variable	Assignment
Sex	Male = 1, female = 0
Preoperative BMI	Continuous variable
Smoking history	Yes = 1, no = 0
Chronic constipation	Yes = 1, no = 0
Incarcerated hernia	Yes = 1, no = 0
Polypropylene mesh patch	Yes = 1, no = 0

In the variable assignment process, “yes” was coded as 1 and “no” was coded as 0 for binary variables included in logistic regression analysis. BMI: Body mass index.

Table 4 Multivariate logistic regression analysis

Risk factor	β	SE	Wald	P value	OR	95%CI
Male	0.953	0.305	9.763	0.001	2.593	1.846-3.641
Preoperative BMI	0.588	0.236	6.208	0.008	1.800	1.147-2.826
Smoking history	0.739	0.282	6.867	0.006	2.093	1.392-3.147
Chronic constipation	1.150	0.406	8.023	0.002	3.159	2.104-4.742
Incarcerated hernia	0.856	0.305	7.877	0.004	2.353	1.748-3.168
Polypropylene mesh patch	0.709	0.281	6.366	0.008	2.031	1.500-2.751

BMI: Body mass index; OR: Odds ratio; CI: Confidence interval.

Table 5 Clinical efficacy of the postoperative recurrence risk prediction model

Item	AUC	95%CI	Sensitivity	Specificity
Training set	0.895	0.831-0.959	75.90%	88.50%
Validation set	0.933	0.828-1.000	81.80%	83.10%

Area under the curve refers to the area under the curve. AUC: Area under the curve; CI: Confidence interval.