Comparison of laparoscopic and open surgery for colorectal malignancy in obese patients: a propensity score-weighted cohort study

Abstract

Background:

Insufficient evidence exists to ascertain the long-term prognosis in patients with obesity undergoing laparoscopic surgery versus open surgery for colorectal cancer.

Methods:

Employing an institutional database from 2009 to 2019, we assessed individuals with a BMI of greater than or equal to 30 kg/m² who underwent surgery for primary stage I–III colorectal adenocarcinoma. The authors used propensity score-weighted analysis to compare short-term and oncologic outcomes between laparoscopic and open surgical approaches.

Results:

This study enrolled 473 patients (open vs. laparoscopic surgery: 220 vs. 253; median follow-up period, 60 months). The laparoscopy group showed a significantly longer operative time (252 vs. 212 min), a higher anastomotic-leakage rate (5.14% vs. 0.91%), and a greater proportion of Clavien–Dindo class greater than III complications (5.93% vs. 1.82%). The open group showed a higher wound infection rate (7.27% vs. 3.16%) and a higher readmission rate (6.36% vs. 2.37%). After propensity score weighting, laparoscopy was inferior to open surgery in terms of long-term overall survival (hazard ratio: 1.43), disease-free survival (1.39), and recurrence rate (21.1% vs. 14.5%). In the subgroup analysis, female patients, older individuals, stage III patients, patients with rectal cancer, and those who underwent surgery after 2014 showed inferior long-term outcomes after laparoscopy.

Conclusions:

Laparoscopic colorectal cancer surgery for patients with obesity requires significant caution. Despite good short-term outcomes, this procedure is associated with hidden risks and poor long-term prognoses. In female patients, older individuals, stage III patients, patients with rectal cancer, and those treated in the late surgery era subgroups, caution is advised when performing laparoscopic surgery.

Introduction

Highlights

• Laparoscopic colorectal (CRC) surgery presents surgical risks in individuals classified as obese (BMI ≥30 kg/m²).

• The propensity score weighting (PSW) method was applied to the most extensive patient cohort from a singular medical center.

• Laparoscopic surgery demonstrated associations with decreased wound infection rates and a shortened overall hospitalization period.

• Laparoscopic procedures displayed suboptimal outcomes when juxtaposed with open surgery in CRC patients characterized as obese.

• Special consideration is warranted for specific subgroups, namely women, elderly individuals, those in stage III, and patients diagnosed with rectal cancer.

The incidence of obesity has been steadily increasing in the general population. Moreover, an increase in BMI by 5 kg/m² is associated with a 30% higher risk of colon cancer in men and a 12% higher risk in women, according to the findings from Kyrgiou et al.¹‘s study. Individuals who are overweight or obese face ~18% and 32% higher risk of colorectal cancer (CRC), respectively, than those of normal weight^2–4.

Laparoscopic surgery for CRC has emerged as a trend only after the findings of four major trials^5–9 confirmed its short-term advantages and equivalent long-term outcomes in CRC. However, the stability of laparoscopic techniques does not reduce surgical challenges in patients with obesity. Numerous reports indicate that patients with obesity undergoing CRC surgery face more short-term risks than those without obesity^10–14. Short-term risks encompass a higher conversion rate, increased postoperative morbidity, and significantly longer operative times in laparoscopic surgery. Therefore, surgeons should exercise extreme caution when dealing with patients with obesity.

Laparoscopy for colorectal surgery has demonstrated short-term benefits for the general patient population, primarily by accelerating postoperative recovery, reducing complications, and minimizing incision size and bleeding. Its oncologic outcomes were comparable to those of open surgeries. However, studies on patients with obesity are relatively scarce. Moreover, while most short-term outcomes are positive in obese patients^15–18, long-term prognostic data are limited^19–22. While the studies suggest a comparable oncologic prognosis between laparoscopy and open surgery, the evidence is considerably limited, partly due to factors like retrospective collecting and an insufficient number of patients. An additional randomized control trial, JCOG404, provides substantial evidence. This study even concluded that a high BMI (>25 kg/m²) was a factor associated with unfavorable long-term outcomes of laparoscopy in patients with colon cancer who underwent curative resection^23,24.

Given the uncertainty regarding the long-term outcomes of laparoscopy in obese patients with CRC, this study aimed to employ propensity score weighting (PSW) to analyze a decade of data comparing laparoscopic and open surgery outcomes in patients with BMI greater than or equal to 30 kg/m² who had undergone treatment for CRC.

Methods

Study design and population

Data were collected from 545 CRC patients with BMI greater than or equal to 30 kg/m² who underwent colorectal surgery at the single-institute academic hospital between January 2009 and December 2019. The work has been reported in line with the STROCSS criteria, Supplemental Digital Content 1, http://links.lww.com/JS9/C700 ²⁵. After the operation, eight patients with non-adenocarcinoma neoplasms or Stage IV CRC were excluded. Finally, 473 patients who underwent radical surgery for CRC were enrolled. (Fig. 1A) Subgroup analyses were also conducted based on sex, age, tumor location, tumor stage, carcinoembryonic antigen (CEA) level, operation year, histology grade, and retrieved lymph node number. (Fig. 1B) The decision to perform open or laparoscopic surgery was based on institutional treatment protocols, the surgeon’s experience, and patient performance.

Figure 1.

(A) Flowchart for patient screening and selection under propensity score weighting. Data were collected from 545 patients with BMI ≥30 kg/m² who underwent colorectal surgery between January 2009 and December 2019. Eight cases that showed non-adenocarcinoma neoplasms and 64 cases with stage IV disease were excluded. Four hundred seventy-three patients (open, 220; laparoscopy, 253) were enrolled in the study and analyzed using propensity score-weighted analysis. (B) The flowchart for subgroup analysis under propensity score weighting. A distinct PSW model was employed for subgroup analyses, encompassing variables such as gender, age, tumor location, tumor stage, CEA level, operation year, histology grade, and the number of retrieved lymph nodes. CRC, colorectal cancer, GIST, gastrointestinal stromal tumor, NET, neuroendocrine tumor; PSW, propensity score weighting.

Data collection

Clinical demographic data, previous health information, preoperative laboratory test results, operative records, pathological features, postoperative care records, and follow-up status data were obtained from the Colorectal Section Tumor Registry of our hospital. Clinical demographic data and previous health information, including age, sex, BMI, comorbidities, and abdominal surgical history, were collected. Preoperative laboratory test results were recorded, such as carcinoembryonic antigen (CEA), hemoglobin (Hb), and albumin levels. Operative records and pathological features were recorded, including operative timing and method, ostomy type, operative time, blood loss, tumor location, tumor size, number of retrieved lymph nodes, histological differentiation, and TNM stage. Postoperative care records were collected, including postoperative mortality and morbidity, Clavien–Dindo classification, length of hospital stay, and incidence of readmission.

Study endpoints

The endpoints were disease-free survival (DFS) and overall survival (OS). The DFS was calculated as the interval from the initial surgical intervention for the primary lesion to the date of first recurrence, death, or last follow-up. Meanwhile, OS was calculated as the interval from the initial surgical intervention for the primary lesion to the date of death or last follow-up. All patients underwent the same follow-up protocols, including outpatient visits every three months for physical examinations, CEA measurements, minimal computed tomography, and a colonoscopy every year until 2022 or death. Recurrence of CRC was confirmed by elevated CEA levels, lesion biopsy, or imaging studies [positron emission tomography (PET), computed tomography (CT), or MRI]. The median follow-up duration was 60 months. All data collection and analyses were supervised and approved by the Institutional Review Board of our hospital in Taiwan.

Statistical analysis

The propensity score is a valuable tool to mimic the randomization in an observational study^26–29. Matching and inverse probability of treatment weighting are popular propensity score methods. The two methods both create a pseudo-population to derive an unbiased estimate of the average treatment effect between exposed and unexposed groups. Like a randomized trial, the baseline characteristics between exposed and unexposed groups would be nearly balanced in the pseudo-population. A previous study compared the matching and the weighting and encouraged both methods to be used in survival analysis^30,31. To avoid a heavy loss of statistical power due to matching, this observational study applied the PSW method to mitigate confounding bias. Propensity scores were calculated using logistic regression analysis. The exposure variable was the surgical approach that patients received. The potential confounding factors were age, sex, BMI, operation year, comorbidities), previous abdominal operation, carcinoembryonic antigen, hemoglobin, albumin, tumor location, tumor stage, histological types, histology grade, retrieved lymph node number, tumor size, and operation timing. Each patient was given a weight: ${\mathrm{w}}_{\mathrm{i}}=\frac{{\mathrm{z}}_{\mathrm{i}}}{{\mathrm{e}}_{\mathrm{i}}}+\frac{1-{\mathrm{z}}_{\mathrm{i}}}{1-{\mathrm{e}}_{\mathrm{i}}}$ , where ${\mathrm{e}}_{\mathrm{i}}$ represents the estimated propensity score of the ith patient based on the fitted logistic regression model, ${\mathrm{z}}_{\mathrm{i}}=1$ denotes laparoscopy, and ${\mathrm{z}}_{\mathrm{i}}=0$ denotes open surgery for the ith patient.

In the real population (before PSW), Pearson’s χ² test and the Wilcoxon rank-sum test were used to compare the laparoscopy and open groups’ baseline characteristics and short-term outcomes. Survival curves were calculated by the Kaplan–Meier (KM) method and compared using the log-rank test. In the pseudo-population (after PSW), the weighted Pearson’s χ² test, the weighted Wilcoxon rank-sum test, and the weighted log-rank test were used instead. Hazard ratios (HRs), 95% CIs, and P values were calculated using the weighted Cox proportional-hazard model. Univariate and multivariate Cox proportional-hazard model were also fitted to compare with the results of the PSW analysis. In addition, subgroup analyses based on sex, age, tumor location, tumor stage, serum CEA level, operation year, histological grade, and number of retrieved lymph nodes were also conducted. All statistical analyses were performed using SAS version 9.4.

Results

This study enrolled 473 CRC patients with obesity who underwent curative resection, of whom 220 and 253 patients underwent open and laparoscopic surgery, respectively. Table 1 lists the baseline characteristics of the two groups before and after PSW. After PSW, only cirrhosis rates in the two groups remained significantly different. For the rectal cancer subgroup, data regarding neoadjuvant therapy and trends in lymph node harvest can be found in Appendix 1, Supplemental Digital Content 2, http://links.lww.com/JS9/C701.

Table 1.

Baseline characteristics of patients treated with open group versus laparoscopy group before and after propensity score weighting.

	Open group	Laparoscopic group	P
	N=220	N=253	Before	After
Age, N (%)			0.83	0.64
$<$ 65 years	123 (55.91)	145 (57.31)
≥65 years	97 (44.09)	108 (42.69)
Sex, N (%)			0.03	0.95
Female	115 (52.27)	106 (41.90)
Male	105 (47.73)	147 (58.10)
BMI, N (%)			0.40	0.68
<35	191 (86.82)	227 (89.72)
≥35	29 (13.18)	26 (10.28)
Operation year, N (%)			<0.01	0.99
<2014	155 (70.45)	74 (29.25)
≥2014	65 (29.55)	179 (70.75)
Previous abdominal operation, N (%)
Appendectomy	17 (7.73)	18 (7.11)	0.94	0.78
Cholecystectomy	11 (5.00)	7 (2.77)	0.31	0.51
Hysterectomy	17 (7.73)	13 (5.14)	0.34	0.52
Oophorectomy	4 (1.82)	6 (2.37)	0.92	0.64
Colon-rectal operation	9 (4.09)	6 (2.37)	0.42	0.72
Comorbidity, N (%)
Hypertension	145 (65.91)	165 (65.22)	0.95	0.43
Cardiac disease	28 (12.73)	40 (15.81)	0.41	0.45
Cerebrovascular accident	12 (5.45)	12 (4.74)	0.89	0.92
Asthma	16 (7.27)	9 (3.56)	0.11	0.80
Diabetes mellitus	60 (27.27)	95 (37.55)	0.02	0.44
Liver cirrhosis	0 (0.00)	7 (2.77)	0.04	<0.01
Carcinoembryonic antigen, N (%)			0.17	0.48
<5 ng/ml	163 (74.09)	202 (79.84)
≥5 ng/ml	57 (25.91)	51 (20.16)
Hemoglobin, N (%)			1.00	0.21
<10 mg/ml	37 (16.82)	43 (17.00)
≥10 mg/ml	183 (83.18)	210 (83.00)
Albumin, N (%)			0.17	0.40
<3.5 mg/dl	16 (7.27)	10 (3.95)
≥3.5 mg/dl	204 (92.73)	243 (96.05)
Tumor location, N (%)			0.52	0.67
Right colon	62 (28.18)	83 (32.81)
Left colon	86 (39.09)	96 (37.94)
Rectum	72 (32.73)	74 (29.25)
Tumor stage (AJCC), N (%)			<0.01	0.79
0&1	45 (20.45)	88 (34.78)
2	78 (35.45)	75 (29.64)
3	97 (44.09)	90 (35.57)
Histological type, N (%)			0.42	0.16
Adenocarcinoma	212 (96.36)	242 (95.65)
Mucinous adenocarcinoma	0 (0.00)	2 (0.79)
Signet ring cell	8 (3.64)	9 (3.56)
Histology grade, N (%)			0.15	0.90
Well-differentiated	28 (12.73)	45 (17.79)
Moderately differentiated	180 (81.82)	188 (74.31)
Poorly differentiated	12 (5.45)	20 (7.91)
Retrieved lymph node (+) number, N (%)			0.09	0.85
<12	19 (8.64)	11 (4.35)
≥12	201 (91.36)	242 (95.65)
Tumor size, N (%)			0.04	0.82
<4 cm	113 (51.36)	155 (61.26)
≥4 cm	107 (48.64)	98 (38.74)
Operation timing, N (%)			0.90	0.65
Elective	218 (99.09)	252 (99.60)
Emergent	2 (0.91)	1 (0.40)

Italic values represent p < 0.05, statistically significant.

Short-term postoperative results

Table 2 presents the postoperative outcomes of open and laparoscopic methods before and after PSW. Operation time was significantly longer in the laparoscopy group than in the open group (252 vs. 212 min, P<0.01). The two groups showed similar postoperative mortality and overall morbidity rates before and after PSW. In the sub-classification of postoperative morbidity, the wound infection rate was significantly higher in the open surgery group (7.27% vs. 3.16%, P=0.02). The anastomotic-leakage rate was higher in the laparoscopy group (5.14% vs. 0.91%, P<0.01), as was the proportion of patients who showed Clavien–Dindo class greater than III complications (5.93% vs. 1.82%, P<0.01). The readmission rate was higher in the open surgery group (6.36% vs. 2.37%, P<0.01). The overall conversion rate for laparoscopic surgery was 2.77%. The two groups showed no significant difference in the length of hospital stay after surgery. The stoma rate after surgery was higher in the open group before PSW and showed no significant difference after PSW.

Table 2.

Postoperative outcomes of patients treated with open group versus laparoscopy group before and after propensity score weighting.

	Open group	Laparoscopic group	P
	N=220	N=253	Before	After
Operation time, medium (IQR), min	212 (162–268)	252 (201–315)	<0.01	<0.01
Blood loss, medium (IQR), ml	50 (50–100)	50 (20–50)	0.15	0.17
Postoperative mortality, N (%)	1 (0.45)	3 (1.19)	0.72	0.07
Abdomen (abscess, peritonitis,…)	1 (0.45)	1 (0.40)	1.00	1.00
Anastomosis (leakage, stenosis,…)	0 (0.00)	2 (0.79)	0.54	0.06
Postoperative morbidity, N (%)	37 (16.82)	39 (15.42)	0.77	0.51
Wound (infection, dehiscence,…)	16 (7.27)	8 (3.16)	0.07	0.02
Lung (atelectasis, pneumonia,…)	3 (1.36)	2 (0.79)	0.88	0.17
Cardiovascular event (MI, CVA, embolism,…)	1 (0.45)	2 (0.79)	1.00	0.14
Bladder dysfunction, N (%)	7 (3.18)	4 (1.58)	0.40	0.27
Gastrointestinal (obstruction, bleeding,…)	10 (4.55)	6 (2.37)	0.29	0.80
Abdomen (abscess, peritonitis,…)	5 (2.27)	5 (1.98)	1.00	0.06
Anastomosis (leakage, stenosis,…)	2 (0.91)	13 (5.14)	0.02	<0.01
Others	1 (0.45)	2 (0.79)	1.00	0.12
Clavien–Dindo classification, N (%)			0.04	<0.01
<III	216 (98.18)	238 (94.07)
≥III	4 (1.82)	15 (5.93)
Readmission, N (%)	14 (6.36)	6 (2.37)	0.05	<0.01
Conversion, N (%)	—	7 (2.77)	—	—
Length of hospital stay, median (IQR), day	9 (8–12)	7 (6–9)	0.69	0.71
Stoma type after operation, N (%)			0.05	0.10
No	181 (82.27)	224 (88.54)
Protective divertive stoma	28 (12.73)	25 (9.88)
End stoma	11 (5.00)	4 (1.58)

Italic values represent p < 0.05, statistically significant.

CVA, cerebrovascular accident; IQR, interquartile range; MI, myocardial infarction.

Long-term oncological outcomes

Table 3 presents the data for long-term oncological outcomes. After propensity score weighting, laparoscopy was inferior to open surgery in terms of long-term OS (HR: 1.43 and 95% CI: 1.10–1.86), DFS (HR, 1.39 and 95% CI: 1.09–1.75), and recurrence rate (21.1% vs. 14.5% and P=0.01). The OS and DFS after PSW in the Kaplan–Meier (KM) analysis (Fig. 2) revealed significant inferiority in the laparoscopy group (OS: P<0.01; DFS: P<0.01). In addition, the result of multivariate analysis (Appendix 3, Supplemental Digital Content 3, http://links.lww.com/JS9/C702) showed that laparoscopy was inferior to open surgery, and the effects were nearly significant (HR: 1.53, 95% CI: 0.97–2.39, and P=0.07 for the OS; HR: 1.45, 95% CI: 0.98–2.15, and P=0.06 for the DFS).

Table 3.

Results of hazard ratio of patients treated with open group versus laparoscopy group before and after propensity score weighting.

	Overall survival				Disease-free survival				Recurrence rate
	Before PSW		After PSW		Before PSW		After PSW		Before PSW		After PSW
	HR [95% CI]	P	HR [95% CI]	P	HR [95% CI]	P	HR [95% CI]	P	Rate (%)	P	Rate (%)	P
Laparoscopy group	1.32 [0.89–1.95]	0.17	1.43 [1.10–1.86]	<0.01	1.22 [0.87–1.72]	0.26	1.39 [1.09–1.75]	<0.01	18.58	0.90	21.12	<0.01
Open group	—	—	—	—	—	—	—	—	17.73	—	14.53	—

HR, hazard ratio; PSW, propensity score weighting.

Figure 2.

Disease-free survival and overall survival of patients treated with open or laparoscopic surgery before and after propensity score weighting. The overall survival (OS) and disease-free survival (DFS) before propensity score weighting (PSW) in the Kaplan–Meier analysis revealed no significant differences in the laparoscopy group (OS, P=0.17; DFS, P=0.26). However, the OS and DFS after PSW in the Kaplan–Meier analysis revealed significant inferiority in the laparoscopy group (OS: P<0.01; DFS: P<0.01).

The results of subgroup analyses based on different characteristics before and after PSW are listed in Table 4. In the female subgroup, laparoscopy was associated with a significantly inferior recurrence rate (RR) and DFS (HR for DFS: 1.49; RR for open vs. laparoscopy: 10.3% vs. 22.2%, P<0.01). The older-age subgroup also showed inferior OS, DFS, and RR for laparoscopy (HR for OS: 1.73; HR for DFS: 1.74; RR for open vs. laparoscopy: 12.5% vs. 20.9%, P=0.03). When considering tumor location, obese patients with rectal cancer showed inferior long-term outcomes after laparoscopy (HR for OS, 1.79; HR for DFS, 1.65; no significant difference in RR, P=0.06). In terms of TNM stage, patients with stage III CRC who underwent laparoscopic surgery had worse long-term outcomes (HR for OS: 1.87; HR for DFS: 1.76; RR for open vs. laparoscopy: 26.0% vs. 45.9%, P<0.01), and stage II patients in the laparoscopy group had a higher recurrence rate (21.1% vs. 9.83%, P<0.01). For surgeries performed in or after 2014, the laparoscopy group showed worse long-term outcomes than the open surgery group (HR for OS: 1.74; HR for DFS: 2.02; RR for open vs. laparoscopy: 9.9% vs. 19.9%, P<0.01), but for surgeries performed before 2014, the two groups showed similar long-term outcomes. In assessments based on histological grade, the laparoscopy group showed a significantly worse DFS (HR: 1.30, P=0.05) and RR (open vs. laparoscopy: 15.0% vs. 26.3%, P<0.01) for moderately differentiated adenocarcinoma. If more than 12 lymph nodes were retrieved in the surgical specimen, the laparoscopy group also showed worse long-term outcomes (HR for OS: 1.40; HR for DFS: 1.39; significance in RR, P<0.01).

Table 4.

Subgroup analysis of hazard ratio of patients treated with open group versus laparoscopy group before and after propensity score weighting.

	Overall survival				Disease-free survival				Recurrence rate
	Before PSW		After PSW		Before PSW		After PSW		Before PSW			After PSW
	HR [95% CI]	P	HR [95% CI]	P	HR [95% CI]	P	HR [95% CI]	P	Rate (%)			Rate (%)
									OG	LG	P	OG	LG	P
Sex
Male	1.21 [0.69–2.14]	0.51	1.24 [0.85–1.80]	0.26	0.99 [0.62–1.58]	0.95	1.12 [0.82–1.54]	0.47	21.90	16.33	0.34	18.49	19.03	0.97
Female	1.46 [0.85–2.53]	0.17	1.31 [0.89–1.93]	0.18	1.54 [0.94–2.53]	0.09	1.49 [1.04–2.14]	0.03	13.91	21.70	0.18	10.34	22.17	<0.01
Age
<65 years	1.25 [0.66–2.36]	0.49	1.15 [0.75–1.78]	0.52	1.07 [0.64–1.77]	0.80	1.16 [0.81–1.66]	0.43	19.51	18.62	0.98	15.07	19.44	0.22
≥65 years	1.34 [0.81–2.21]	0.26	1.73 [1.22–2.44]	<0.01	1.35 [0.85–2.14]	0.21	1.74 [1.26–2.40]	<0.01	15.46	18.52	0.69	12.53	20.86	0.03
Tumor location
Colon	1.33 [0.82–2.17]	0.25	1.22 [0.87–1.71]	0.25	1.23 [0.80–1.91]	0.35	1.22 [0.90–1.64]	0.21	14.86	15.08	1.00	12.96	18.15	0.09
Rectum	1.28 [0.66–2.50]	0.47	1.79 [1.17–2.76]	<0.01	1.24 [0.71–2.14]	0.45	1.65 [1.12–2.42]	0.01	23.61	27.03	0.78	16.65	26.38	0.06
Tumor stage (AJCC)
0&1	1.51 [0.66–3.46]	0.33	1.06 [0.60–1.85]	0.85	1.47 [0.65–3.34]	0.36	1.11 [0.64–1.94]	0.71	4.44	5.68	1.00	2.71	5.67	0.42
2	1.36 [0.66–2.81]	0.40	1.02 [0.60–1.73]	0.94	1.27 [0.69–2.34]	0.44	1.20 [0.78–1.85]	0.40	11.54	20.00	0.22	9.83	21.12	<0.01
3	1.29 [0.72–2.30]	0.39	1.87 [1.29–2.69]	<0.01	1.29 [0.80–2.09]	0.30	1.76 [1.27–2.45]	<0.01	28.87	30.00	0.99	25.99	45.92	<0.01
Carcinoembryonic antigen
$<$ 5 ng/ml	1.29 [0.81–2.06]	0.28	1.21 [0.88–1.66]	0.24	1.18 [0.78–1.79]	0.44	1.17 [0.88–1.57]	0.28	13.50	15.35	0.73	10.94	15.84	0.07
≥5 ng/ml	1.52 [0.72–3.20]	0.27	1.40 [0.81–2.42]	0.23	1.51 [0.83–2.77]	0.18	1.26 [0.81–1.96]	0.32	29.82	31.37	1.00	25.61	29.25	0.67
Operation year
<2014	1.47 [0.90–2.40]	0.12	1.45 [1.05–2.01]	0.03	1.27 [0.80–2.02]	0.31	1.26 [0.92–1.71]	0.15	19.35	21.62	0.82	18.33	20.50	0.64
≥2014	0.99 [0.49–2.00]	0.98	1.74 [1.10–2.75]	0.02	1.17 [0.65–2.11]	0.61	2.02 [1.38–2.96]	<0.01	13.85	17.32	0.65	9.86	19.86	<0.01
Histology grade
Well-differentiated	2.91 [1.07–7.90]	0.04	1.87 [1.06–3.32]	0.03	2.04 [0.78–5.33]	0.15	2.14 [1.20–3.81]	0.01	10.71	6.67	0.86	3.89	8.61	0.40
Moderately differentiated	1.17 [0.74–1.86]	0.51	1.27 [0.93–1.73]	0.13	1.13 [0.77–1.67]	0.54	1.30 [1.00–1.70]	0.05	17.78	19.68	0.74	15.00	23.63	<0.01
Poorly differentiated	0.66 [0.19–2.35]	0.52	0.66 [0.19–2.35]	0.52	1.07 [0.34–3.35]	0.91	1.07 [0.34–3.35]	0.91	33.33	35.00	1.00	33.34	35.00	1.00
Retrieved lymph node (+) number
<12	1.25 [0.39–3.94]	0.71	1.25 [0.39–3.94]	0.71	1.10 [0.36–3.36]	0.87	1.10 [0.36–3.36]	0.87	26.32	36.36	0.87	26.31	36.35	0.87
≥12	1.35 [0.89–2.05]	0.16	1.40 [1.06–1.85]	0.02	1.26 [0.88–1.81]	0.21	1.39 [1.08–1.78]	<0.01	16.92	18.18	0.82	13.70	20.63	<0.01

HR, hazard ratio; LG, laparoscopy group; OG, open group; PSW, propensity score weighting.

We also analyzed the recurrence sites in these patients (Table 5). Brain metastasis (1.58% vs. 0.91%, P<0.01), liver metastasis (9.88% vs. 8.18%, P<0.01), intra-abdominal recurrence (2.77% vs. 1.36%, P=0.01), and other nodes metastasis (2.77% vs. 0.91%, P<0.01) rates were significantly higher after PSW in the laparoscopy group. Additionally, the distribution of recurrences for the rectal cancer subgroup has been analyzed and included in Appendix 2, Supplemental Digital Content 4, http://links.lww.com/JS9/C703.

Table 5.

Recurrence patterns of patients treated with open group versus laparoscopy group before and after propensity score weighting.

	Open group	Laparoscopic group	P
Recurrence	N=220, N (%)	N=253, N (%)	Before	After
Anastomosis	4 (1.82)	4 (1.58)	1.00	1.00
Bone	7 (3.18)	5 (1.98)	0.59	0.78
Brain	2 (0.91)	4 (1.58)	0.81	<0.01
Liver	18 (8.18)	25 (9.88)	0.63	<0.01
Lung	3 (1.36)	2 (0.79)	0.88	0.42
Kidney	2 (0.91)	0	0.42	0.29
Intra-abdominal	3 (1.36)	7 (2.77)	0.46	0.01
Pelvic organs	11 (5.00)	13 (5.14)	1.00	0.15
Perineum	0	1 (0.40)	1.00	0.87
Retroperitoneum	8 (3.64)	6 (2.37)	0.59	0.56
Other sites	6 (2.73)	3 (1.19)	0.38	0.39
Inguinal nodes	0	0	—
Other nodes	2 (0.91)	7 (2.77)	0.26	<0.01

Discussion

This study compared the short- and long-term outcomes of open and laparoscopic CRC surgeries. In comparison with the open method, laparoscopic radical resection for CRC in patients with obesity was associated with lower rates of wound infection and readmission but showed a higher rate of anastomotic leakage after the operation. After PSW, the laparoscopy group showed significantly poorer long-term outcomes regarding OS, DFS, and RR. In the subgroup analysis, female patients, older individuals, stage III patients, patients with rectal cancer, and those who underwent surgery after 2014 showed inferior long-term outcomes in the laparoscopy group.

The data comparing open and laparoscopic surgery in obese patients with CRC are notably limited. Most studies had inadequate sample sizes, statistical shortcomings, and methodological concerns. For example, despite showing no apparent differences in long-term and short-term prognoses, Kazama et al.²⁰ assessed a relatively small sample size of 96 patients, with only 32 in the laparoscopy group. Cai et al.¹⁹ reported that while the two groups showed no apparent difference in DFS and OS in the KM analysis, they showed differences in the proportion of stage III patients in comparison with stage I patients. Despite the higher proportion of stage III patients in the open surgery group, KM survival rates remained comparable to those in the laparoscopy group. Bizzoca et al. utilized propensity score matching (PSM) to create groups of 51 patients, resulting in an asymmetric distribution of age, BMI, tumor location, and tumor stage between the two groups. Notably, stage IV patients constituted 33.3% of the open surgery group but only 9.8% of the laparoscopy group²¹.

Second, scrutiny of these literature reviews indicates a poorer tumor prognosis related to laparoscopy. In the study by Cai et al.¹⁹, despite a higher proportion of stage III patients in the open surgery group, the KM curves showed similar outcomes, indicating a potentially worse prognosis associated with laparoscopic surgery. In the study by Bizzoca et al.²¹, multivariate Cox regression analysis of tumor relapse, which showed a nonsignificant P value for laparoscopy, produced an HR of 3.69. However, when considering a significant staging factor, the HR was reduced to 1.19 with a nonsignificant P value. The findings obtained by Akagi and colleagues closely resemble those of the current study. Although their conclusion suggests that laparoscopy groups have better short-term outcomes and similar long-term outcomes, their data revealed that the 3-year RFS after laparoscopic surgery was not non-inferior to that after open surgery. Additionally, the KM curve for relapse-free survival clearly showed a similar trend of poorer outcomes in the laparoscopy group, consistent with our findings²². The pivotal trial in this context was the JCOG404 trial, which compared open with laparoscopic surgery for CRC. Although laparoscopic surgery was considered a viable alternative for stage II/III colon cancer, the subgroup analysis in that study indicated that a higher BMI (> 25 kg/m²) and pT4 and pN2 (except for recto-sigmoid colon) were associated with less favorable long-term outcomes for laparoscopic surgery in patients with colon cancer who underwent curative resection^23,24. We are utilizing our comprehensive database and design methods to seek answers for the above reasons.

Although laparoscopic surgery for CRC has become a trend, many guidelines emphasize caution when considering its use in patients with obesity³². In patients with high BMI, visceral and abdominal wall fat limits the surgical space and increases technical challenges, increasing operative times for both open and laparoscopic procedures¹⁶. Numerous studies have indicated that laparoscopic surgery in patients with obesity tends to yield fewer lymph nodes than open surgery. The lymph node harvest did not decrease with increasing BMI among early-era patients undergoing surgery (most commonly open surgery)^33–35. However, in laparoscopic procedures, the lymph node harvest tends to decrease as the BMI increases³⁶. Such findings raise concerns regarding laparoscopic surgery in patients with obesity²². The anesthesia risks associated with laparoscopic surgery in older obese patients are also worth considering. Moreover, advanced tumors are more confidently addressed through open surgery, particularly in obese patients with rectal cancer, further adding to the complexity of laparoscopic surgery. Thus, the selection of laparoscopic surgery for patients with obesity often requires consideration of multiple factors by the surgical team.

In our dataset, the advantages and drawbacks of laparoscopic surgery were consistent with the findings of other studies. Furthermore, our laparoscopic surgery group showed a lower rate of cases in which less than 12 lymph nodes were obtained (4.35% vs. 8.64%), indicating that our laparoscopic procedures are safe. Regarding long-term tumor prognosis, our DFS, OS, and RR results were in line with those in the existing literature^{11,12,19,20,22}. However, in specific subgroups, such as women, older individuals, stage III patients, patients with rectal cancer, and those who underwent surgery after 2014, laparoscopic surgery resulted in comparatively poorer outcomes. While proficiency in laparoscopic surgery is expected to improve with the accumulation of surgical experience, our data suggest that the outcomes for laparoscopic procedures worsened after 2014. This led us to suspect a possible correlation with the broader patient selection criteria employed by surgeons.

Considering the concerns mentioned above, our data suggest a reasonable risk of laparoscopic surgery in patients with rectal cancer and stage III CRC. In patients with obesity, surgical assistants should be made aware of the strict limitations on grasping the tissue to achieve optimal surgical visibility. Moreover, surgeons should be mindful of the increased risk of tumor exposure during repeated manipulations and extended periods of pneumoperitoneum. This is identical to the concept that factors such as pT4 and pN2 were associated with less favorable long-term outcomes reported in JCOG404^23,24.

In our department’s previous study comparing laparoscopy and open surgery in older adults³⁷, laparoscopy showed better short-term outcomes and a similar tumor prognosis. However, laparoscopy was a risk factor for older patients with obesity in this study. The findings did not clarify whether these factors were related to compromised immune responses in older obese individuals. Notably, patients who underwent laparoscopic surgery in our department showed a higher probability of developing intestinal leakage, primarily because many surgeons believed that performing a protective stoma in the event of intestinal leakage following laparoscopic surgery was easier than performing the same procedure following open surgery. This reduces the accuracy of assessments determining whether an initial protective stoma is required in laparoscopic cases. Additionally, laparoscopic surgery in obese individuals requires more time, potentially increasing surgical risks in the older population. These factors may contribute to poorer outcomes observed in older individuals.

One possible reason for the poorer prognosis in women and patients with rectal cancer who undergo laparoscopic surgery may be the less common use of concurrent chemoradiotherapy, particularly among women. Furthermore, there could be concerns regarding the larger pelvic area in women and the potential impact of radiation therapy on fertility. Surgeons may lean more toward direct surgery in these cases, potentially reducing the likelihood of favorable laparoscopic outcomes in such patients. Table 5 presents the patient recurrence patterns. Despite a higher proportion of brain, liver, abdominal, and distant lymph node metastases in the laparoscopy group, the distribution of metastatic pathways (hematogenous, lymphatic, or peritoneal) was even, showing no bias towards any one route.

This study had several strengths. First, to our knowledge, the present study included the largest number of patients from a single medical center. Second, the PSW used in the study can preserve sample size to gain greater statistical power than the traditional multivariate analysis by the Cox proportional-hazard model. Third, the prevalence of obesity is relatively lower in Asian countries than in the Western world. This study provides a real-world experience of laparoscopic radical resection in obese patients with CRC, a population in which clinicians may have less experience.

Nevertheless, our study also had some limitations. First, this was a retrospective study conducted at a single institution. Despite advancements in laparoscopic surgical techniques at our medical institution, the choice between laparoscopic and open surgery remains subject to the surgeon’s experience and individual patient characteristics. This selection bias may have influenced our findings. Another limitation was the relatively small size of our patient population, which reduced the statistical power of our study. Third, intergroup differences in confounding factors such as the ASA score, surgical margins, and CME/TME quality were not performed and may have provided additional insights. A randomized controlled trial may be needed to provide more definitive insights into the effectiveness of laparoscopic procedures in this patient population.

Conclusions

Laparoscopic CRC surgery for patients with BMI greater than or equal to 30 kg/m² requires extra caution. Our study found that laparoscopy was associated with lower wound infection rates and a lower overall period of hospitalization but an increased leakage rate. However, laparoscopy performed worse than open surgery regarding OS, DFS, and RR. Furthermore, we observed that in the population of patients with obesity, the long-term prognosis was worse in women, older individuals, stage III patients, patients with rectal cancer, and those treated in the late era after 2014. Laparoscopic surgery in patients with obesity presents several challenges and benefits with hidden risks. Advances in techniques and training can enhance the outcomes of this approach. This study emphasizes the importance of employing safe and personalized surgical approaches.

Ethical approval

We retrieved data regarding our clinicopathologic variables from the Colorectal Section Tumor Registry of Chang Gung Memorial Hospital (CGMH). The Institutional Review Board of CGMH approved this study (IRB No. 202301534B0).

Consent

This retrospective study adheres to ethical guidelines, as it does not involve patient treatment and thus does not violate the requirement for ethics committee approval and fully informed, written consent. The study ensures patient privacy by abstaining from the use of identifiable information, and any essential information is presented with explicit permission obtained in accordance with the necessary consent procedures. The manuscript includes a consent section confirming the acquisition of written informed consent for the publication of relevant details, as stipulated in the ethical guidelines.

Source of funding

The study did not receive any funding, and the authors declare that they do not have any conflict of interest.

Author contribution

Y.J.H.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, visualization, writing—original draft, writing—review and editing. Y.L.Y.: conceptualization, data curation, formal analysis, investigation, supervision, validation, visualization, writing—original draft.J.R.J.: conceptualization, data curation, investigation, methodology, software, writing—original draft.J.F.Y.: data curation, supervision, writing—review and editing. C.K.L.: data curation. W.S.T.: data curation, supervision. Y.P.P.: data curation.Y.J.C.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, visualization, writing—original draft, writing—review and editing. All authors have read and agreed to the published version of the manuscript.

Conflicts of interest disclosure

The authors declare that they have no financial conflict of interest with regards to the content of this report.

Research registration unique identifying number (UIN)

Clinicaltrials.gov

registration ID: NCT06221878.

https://clinicaltrials.gov/study/NCT06221878.

Guarantor

Yu-Jen Hsu and Yih-Jong Chern are the guarantors.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Provenance and peer review

Our paper was not invited. We submitted our research based on our own interests.

Supplementary Material

js9-110-4598-s001.docx ^{(31.2KB, docx)}

js9-110-4598-s002.docx ^{(14.6KB, docx)}

js9-110-4598-s003.docx ^{(45.1KB, docx)}

js9-110-4598-s004.docx ^{(15.5KB, docx)}