The practice of draining body cavities has a long history in medical practice, extending back to antiquity. Draining chest empyema and ascites can be traced back to the time of Hippocrates in historical records [1]. Prophylactic abdominal drainage (ADs) have been used in colorectal surgery to remove intraperitoneal collections such as ascites, chyle, blood, and intestinal contents since the mid-1800s (notably reported by Billroth [2] in 1881). At that time, drainage was performed because these accumulations were believed to have the potential to cause infection. Further, ADs were also useful for early detection of postoperative bleeding or anastomotic leakage [3]. However, it has been widely recognized that there are no clear advantages to the use of prophylactic ADs in open colorectal surgery. Many studies have demonstrated little difference in postoperative complications, such as morbidity or mortality, regardless of AD placement [4-9]. Moreover, the Enhanced Recovery After Surgery (ERAS) guidelines recommend avoiding the routine use of ADs after colorectal surgery [10-12]. Despite these guidelines, a lot of surgeons routinely use this technique. According to an online survey on the adoption of ERAS guidelines among European Society of Coloproctology members in 2019 to 2020, it was found that the “no drainage” guideline was the least well-implemented [13]. This discrepancy underscores the ongoing debate and the lack of unanimous agreement in real-world clinical settings.

Research on ADs, particularly in minimally invasive surgeries, including laparoscopic procedures, has demonstrated a comparatively lower impact than well-established open surgery techniques. Owing to the varying practices implemented by hospitals and countries, more evidence is needed to support the decision not to use ADs in laparoscopic hemicolectomy. Given this clinical setting, to ensure generalizability and objectivity regarding the application of ADs in laparoscopic hemicolectomy, it is necessary to examine the differences in postoperative pain, recovery, complications, and other surgical outcomes between patients who have undergone the procedure with and without ADs (AD vs. no-AD groups). In this study, we attempted to challenge the conventional practice of prophylactic AD placement during laparoscopic hemicolectomy to uncover the potential effects associated with this approach.

Demographics

Data of all adult patients (aged ≥18 years) who underwent colorectal surgery at Seoul National University Bundang Hospital between April 2013 and September 2022 by a single surgeon were retrospectively reviewed using electronic medical records (EMRs). The exclusion criteria included surgeries other than hemicolectomies, presence of benign colonic diseases, requirement for simultaneous surgery for other conditions, and treatment with open hemicolectomies. Data on the baseline patient demographic characteristics, including age, sex, body mass index (BMI), smoking status, alcohol consumption status, presence of comorbidities, American Society of Anesthesiologists physical status (ASA PS) class, and history of abdominal surgery were collected. Information on blood tests, whether emergency surgery was conducted, operation time, estimated blood loss (EBL), type of operation, anastomotic configuration, time to gas passage after surgery, length of hospital stay after surgery, postoperative numerical rating scale (NRS) score for pain, postoperative complications, readmission rates within 30 days after surgery, operation year, and clinical stage of cancer were also reviewed. The final staging followed the American Joint Committee on Cancer 8th edition and is shown in Table 1. Readmissions within 30 days were included, regardless of whether they occurred in the emergency room or outpatient setting. The presence or absence of complications was objectively defined according to the Clavien-Dindo classification [14].

Table 1 . Summary of final staging.

AJCC stage, 8th ed	No-AD group (n = 112)	AD group (n = 92)
O	6	2
I	32	19
II	34	26
III	35	37
IV	5	8

Most were adenocarcinomas..

AJCC, American Joint Committee on Cancer; AD, abdominal drainage..

Abdominal drainage

A Jackson-Pratt drain was used via the port site. There was no indication for inserting a drain, however, the decision to insert the drain was made by the colorectal surgeon after considering various factors, including the patient’s baseline information, bowel condition in the operating room, presence of severe inflammation, and high risk of anastomosis leakage. The drain was typically removed at least 5 days postoperatively, after the patient had consumed soft food and had a bowel movement. In some cases, the drain was kept in place for a longer period.

Outcomes measured

The primary outcome was postoperative NRS pain, specifically measured on the morning of the first, second, and fifth postoperative days (POD). In total, five ports (RUQ, LUQ, umbilicus, RLQ, and LLQ) were utilized. A standardized midline mini-laparotomy was performed with an incision length of 4 cm, ensuring no difference among the patient groups. Patient-controlled analgesia was uniformly used for all patients to control postoperative pain. Additionally, when patients reported severe pain, the nurses notified the physician, who then administered additional analgesics, such as acetaminophen, NSAIDs, or opioids. The time to flatus after surgery, length of hospital stay after surgery, readmission rates within 30 days after surgery, and postoperative complications were assessed as secondary outcomes.

Statistical analyses

Propensity scores were employed to estimate the probability of receiving a particular treatment conditional on baseline characteristics [15,16]. By incorporating these propensity scores into the analysis, the inverse probability of treatment weighting (IPTW) was used to reduce selection bias between the AD and no-AD groups. Weights were obtained by taking the inverse of the propensity scores, representing the probability of receiving an AD on the baseline covariates. A logistic regression model was applied to estimate the propensity scores for each participant, with the group variable as the dependent variable and the baseline covariates as the independent variables. A standardized mean difference (SMD) was calculated to evaluate covariate balance, with an absolute SMD greater than 0.2 indicating an imbalance. Potential confounding variables that could be considered before drain insertion into the operating room were included. Covariates for propensity scores in relation to the baseline characteristics of the patients were as follows: age, sex, BMI, smoking status, alcohol consumption status, presence of comorbidities, ASA PS class, history of abdominal surgery, blood tests, whether emergency surgery was conducted, type of operation, anastomotic configuration, year of operation, and clinical stage of cancer. All baseline characteristics described in Table 2 were used as covariates in regression modeling.

Table 2 . Demographic data of the study participants before and after IPTW.

Variable	Before IPTW (n = 204)				After IPTW (n = 402.8)
	No-AD group	AD group	SMD		No-AD group	AD group	SMD
No. of participants	112	92			227	175.8
Age (yr)	64.0 ± 13.3	66.8 ± 13.7	0.207		66.5 ± 12.4	65.1 ± 13.8	0.107
Male sex	53 (47.3)	49 (53.3)	0.059		121.0 (53.3)	86.9 (49.4)	0.039
BMI (kg/m2)	23.7 ± 3.2	24.7 ± 3.7	0.291		23.9 ± 3.2	24.5 ± 3.4	0.161
Comorbidity	81 (72.3)	66 (71.7)	0.006		171.9 (75.7)	122.3 (69.6)	0.062
ASA PS grade			0.109				0.011
I, II	95 (84.8)	68 (73.9)			174.7 (77.0)	137.2 (78.1)
III, IV	17 (15.2)	24 (26.1)			52.3 (23.0)	38.6 (21.9)
Operation history	30 (26.8)	33 (35.9)	0.091		51.8 (22.8)	55.1 (31.3)	0.085
Smoking			0.178				0.061
Never	82 (73.2)	51 (55.4)			132.2 (58.2)	111.8 (63.6)
Ex	21 (18.8)	28 (30.4)			56.6 (24.9)	45.1 (25.6)
Current	9 (8.0)	13 (14.1)			38.2 (16.8)	18.9 (10.8)
Alcohol consumption			0.015				0.058
Never	75 (67.0)	63 (68.5)			173.8 (76.6)	124.4 (70.7)
Ex	4 (3.6)	3 (3.3)			5.8 (2.6)	4.8 (2.7)
Current	33 (29.5)	26 (28.3)			47.3 (20.9)	46.6 (26.5)
Clinical stage			0.117				0.071
1, 2	63 (56.2)	41 (44.6)			100.9 (44.4)	90.7 (51.6)
3, 4	49 (43.8)	51 (55.4)			126.1 (55.6)	85.1 (48.4)
Hemoglobin (g/dL)	12.5 ± 2.1	11.1 ± 2.4	0.272		12.0 ± 2.2	12.4 ± 2.6	0.189
WBC (×103/μL)	6.8 ± 1.9	7.1 ± 2.5	0.149		7.1 ± 1.9	7.2 ± 2.3	0.029
PLT (×103/μL)	273.2 ± 82.1	278.7±107.8	0.057		282.2 ± 78.6	289.0 ± 100.0	0.070
PT/INR	1.0 ± 0.1	1.0 ± 0.1	0.004		1.0 ± 0.1	1.0 ± 0.1	0.058
Albumin (g/dL)	4.1 ± 0.5	3.9 ± 0.6	0.377		4.0 ± 0.6	4.0 ± 0.6	0.133
Glucose (mg/dL)	112.3 ± 28.0	120.3 ± 33.8	0.259		117.4 ± 28.4	117.4 ± 30.1	0.001
Creatinine (mg/dL)	0.8 ± 0.6	0.9 ± 0.6	0.059		0.8 ± 0.5	0.9 ± 0.6	0.090
Operation year			0.167				0.070
2014–2018	47 (42.0)	54 (58.7)			128.4 (56.6)	87.0 (49.5)
2019–2022	65 (58.0)	38 (41.3)			98.6 (43.4)	88.8 (50.5)
Emergency operation	4 (3.6)	4 (4.3)	0.008		15.9 (7.0)	18.3 (10.4)	0.034
Radicality			0.047				0.001
R0	110 (98.2)	86 (93.5)			216.6 (95.4)	168.0 (95.5)
R2	2 (1.8)	6 (6.5)			10.4 (4.6)	7.8 (4.5)
Operation time (min)	134.7 ± 32.4	164.4 ± 39.9	0.817		152.5 ± 39.5	155.2 ± 36.2	0.075
EBL (mL)	56.3 ± 49.8	81.1 ± 84.8	0.356		65.7 ± 63.4	71.4 ± 72.3	0.081
Operation type			0.205				0.093
LHC	5 (4.5)	23 (25.0)			13.0 (5.7)	26.4 (15.0)
RHC	107 (95.5)	69 (75.0)			213.9 (94.3)	149.4 (85.0)
Co-operationa)	10 (8.9)	9 (9.8)	0.009		25.2 (11.1)	18.1 (10.3)	0.008
Anastomosis type			0.023				0.009
Side-to-side	107 (95.5)	90 (97.8)			221.5 (97.6)	173.2 (98.5)
End-to-side	5 (4.5)	2 (2.2)			5.5 (2.4)	2.6 (1.5)

Values are presented as number only, mean ¡¾ standard deviation, or number (%)..

IPTW, inverse probability of treatment weighting; AD, abdominal drainage; SMD, standardized mean difference; BMI, body mass index; ASA PS, American Society of Anesthesiologists physical status; WBC, white blood cell; PLT, platelet; PT, prothrombin ratio; INR, international normalized ratio; EBL, estimated blood loss; LHC, left hemicolectomy; RHC, right hemicolectomy..

^a)Co-operation with other minor surgery..

Continuous variables are reported as the means or medians for normally distributed variables and compared using a two-sample t-test. The statistical significance of categorical variables was evaluated using the chi-square test or Fisher exact test. Statistical significance was considered at p-values of <0.05. Statistical analyses were performed using the R version 4.1.1 (R Foundation for Statistical Computing).

Study population

Fig. 1 presents the flowchart of patient selection. A total of 1,090 patients who underwent colectomies performed by a single surgeon at Seoul National University Bundang Hospital between April 2013 and September 2022 were retrospectively reviewed based on their EMRs. After excluding 632 patients who underwent other surgeries, such as cecectomy, ileocecectomy, anterior resection, Hartmann operation, and subtotal or total colectomy, only 404 patients who underwent right hemicolectomies and 54 patients who underwent left hemicolectomies (a total of 458 patients) were included. After excluding 131 patients with benign colonic diseases, nine patients who underwent simultaneous surgeries for other cancers, 111 patients who underwent open surgery, 179 patients who underwent laparoscopic right hemicolectomies, and 28 patients who underwent laparoscopic left hemicolectomies were included. Among them, three individuals had missing data, resulting in 204 patients being included in the review (AD group, 92 patients; no-AD group, 112 patients).

Figure 1. Flow chart of patient selection. Three individuals were ruled out for inverse probability of treatment weighting adjustment due to insufficient data.
LHC, left hemicolectomy; RHC, right hemicolectomy.

Demographics

The baseline patient characteristics are shown in Table 2. Initially, variables such as age, BMI, hemoglobin level, albumin level, glucose level, operation time, EBL, and operation type between the two groups had a SMD exceeding 0.2. Therefore, IPTW was applied to adjust and align the SMD to <0.2, ensuring no significant differences in the variables between the two groups. A balanced plot of the potential confounding variables is shown in Fig. 2.

Figure 2. Balance plot of potential confounding variables. Unadjusted means before adjustment with inverse probability of treatment weighting (IPTW) and adjusted means after adjustment with IPTW. Op, operation; EBL, estimated blood loss; BMI, body mass index; WBC, white blood cell; ASA, American Society of Anesthesiologists physical status classification; cStage, clinical stage; Em, emergency; Co-op, co-operation with other minor surgery.

Primary outcomes

The postoperative surgical outcomes are shown in Table 3. After adjustment with IPTW, the NRS on POD 1 was 3.4 ± 1.5 in the no-AD group and 3.3 ± 1.6 in the AD group, with no significant difference (p = 0.666). The NRS on POD 2 was significantly higher in the AD group than in the no-AD group (3.4 ± 0.8 vs. 3.2 ± 0.8, p = 0.043). On POD 5, the NRS was 2.5 ± 0.7 in the no-AD group and 2.4 ± 0.7 in the AD group, with no significant difference (p = 0.204). Postoperative NRS scores for pain on each day are shown in Fig. 3.

Table 3 . Postoperative surgical outcomes before and after IPTW.

Outcomes	Before IPTW (n = 204)				After IPTW (n = 402.8)
	No-AD group (n = 112)	AD group (n = 92)	p-value		No-AD group (n = 227)	AD group (n = 175.8)	p-value
NRS
POD 1	3.7 ± 1.4	3.1 ± 1.6	0.002		3.4 ± 1.5	3.3 ± 1.6	0.666
POD 2	3.3 ± 0.8	3.3 ± 0.9	0.981		3.2 ± 0.8	3.4 ± 0.8	0.043
POD 5	2.4 ± 0.7	2.4 ± 0.8	0.861		2.5 ± 0.7	2.4 ± 0.7	0.204
Postoperative complications	17 (15.2)	13 (14.1)	0.833		28.2 (12.4)	30.4 (17.3)	0.170
Hospital stay (day)	6.9 ± 3.8	7.3 ± 2.9	0.479		6.9 ± 3.0	7.3 ± 2.8	0.298
First time to flatus (day)	2.9 ± 1.0	2.9 ± 1.0	0.566		2.7 ± 0.9	2.9 ± 0.9	0.078
Readmission (%)	2.7	2.2	1.000		1.7	2.3	0.733

Values are presented as mean ± standard deviation, number (%), or percentage only..

IPTW, inverse probability of treatment weighting; AD, abdominal drainage; NRS, numerical rating scale; POD, postoperative day..

Figure 3. Postoperative NRS pain score on each day in the unadjusted (A) and adjusted (B) patient groups. Unadjusted means before adjustment with inverse probability of treatment weighting (IPTW) and adjusted means after adjustment with IPTW.
AD, abdominal drainage.

Secondary outcomes

After adjustment with IPTW, no statistically significant differences in complications were found between the no-AD (28.2 cases, 12.4%) and AD (30.4 cases, 17.3%) groups (p = 0.170). Surgical complications were classified according to the Clavien-Dindo grade, as summarized in Table 4. Grade 1 complications were observed in eight cases in both groups. Grade 2 complications were present in six cases in the no-AD group and three cases in the AD group. Grade 3 complications occurred in two cases in each group. Additionally, one case of a Grade 4 complication was reported in the no-AD group. Notably, in the no-AD group, there were two cases of anastomotic leakage, both of which required reoperation. The length of postoperative hospital stay was 6.8 ± 3.0 days and 7.3 ± 2.8 days in the no-AD and AD groups, respectively, with no statistically significant difference (p = 0.298). The time to flatus was 2.7 ± 0.9 days in the no-AD group and 2.9 ± 0.9 days in the AD group, with no statistically significant difference (p = 0.078). Finally, regarding readmission to the outpatient or emergency departments within 1 month, four cases (1.7%) in the no-AD group and four cases (2.3%) in the AD group were noted, with no significant difference (p = 0.733).

Table 4 . Clavien-Dindo classification of surgical complica­tions.

Clavien-Dindo classification	No AD group (n = 17)	AD group (n = 13)
Grade I	8	8
Grade II	6	3
Grade III	2	2
Grade IV	1	0

AD, abdominal drainage..

To the best of our knowledge, this is the first study to determine the effect of prophylactic AD on postoperative pain in patients with colon cancer undergoing laparoscopic surgery.

The primary objective of this study was to evaluate the effect of prophylactic AD during laparoscopic hemicolectomy. The results showed that the absence of AD did not result in any statistically significant differences in postoperative surgical outcomes, except for postoperative pain, providing crucial evidence for clinical decision-making regarding the use of AD in laparoscopic hemicolectomy. Thus, this provides new insights into the role of AD in laparoscopic hemicolectomy, especially in the context of postoperative pain management.

The placement of surgical drains has long been considered an important aspect of the postoperative management of patients [17]. As mentioned above, AD is commonly performed with the rationale that it may prevent complicated intraabdominal fluid collection, reduce anastomotic leakage, and allow early detection of bleeding, anastomotic leakage, or other complications [18,19]. However, conflicting results have been reported regarding the efficacy and safety of AD. In a meta-analysis by Urbach et al. [6], drains did not effectively detect anastomotic leakage at an early stage. Among the 20 patients with drains who experienced anastomotic leakage, the diagnosis relied on the identification of intestinal content in the effluent in only one case, constituting a 5% detection rate. Drain placement has also been associated with additional adverse events such as increased production of serous fluid, wound infection, and mobility discomfort [20,21]. In addition, AD has been shown to affect the well-being of patients, while indwelling drains have been associated with increased discomfort, which can increase postoperative anxiety [22]. Moreover, many studies do not support the routine use of prophylactic AD after colorectal surgery because of the lack of clinical benefits [23,24]. However, these studies were conducted before the era of minimally invasive surgery, and there is a relative scarcity of research on AD in the context of laparoscopic surgery.

This study found that AD could exacerbate postoperative pain after laparoscopic hemicolectomy. Patients are expected to experience severe pain immediately after surgery, regardless of the presence of AD, which should gradually decrease daily starting from POD 1. In our study, NRS on POD 2 was significantly higher in the AD group than in the no-AD group after IPTW, suggesting that AD had an effect on the degree of postoperative pain. Thus, this study contributes to the limited research on prophylactic AD in hemicolectomy patients.

In our study, we did not observe significant differences between the two groups in terms of length of hospital stay. However, other studies have reported a significantly shorter duration of this parameter. Studies conducted by Hagmüller et al. [25] (with a mean hospital stay of 14.9 days in the AD group vs. 13.3 days in the no-AD group) and Sagar et al. [26] (with a median hospital stay of 12 days in the AD group vs. 13 days in the no-AD group), both published before the implementation of advanced recovery protocols, reported a shorter duration of hospital stay in the no-AD group. Some studies have proposed that refraining from AD implementation could result in improved functional outcomes and a reduced hospital period [27]. This effect may be influenced by differences in postoperative care protocols; however, further research is needed to confirm this hypothesis.

Our study has some limitations. Firstly, it is important to note that although the NRS is a reliable and valid measure of pain intensity and distress, it captures only part of the pain experienced by patients [28]. In addition, the retrospective review of these scores and the small sample size in this study may introduce limitations in terms of reliability. It should be noted that this study was conducted by a single surgeon at a single center. Additionally, given the complexity and difficulty of surgeries, surgeons may have preferred to use AD, which could introduce a potential bias.

In conclusion, the results of this single-center retrospective study suggest that AD can exacerbate postoperative pain in patients without increasing the risk of complications. This study could serve as crucial evidence to support the decision not to insert prophylactic AD in actual clinical practice.

Ethical statements

This study was approved by the Institutional Review Board (IRB) of Seoul National University Bundang Hospital (IRB No. B-2308-844-101). In accordance with the policy of the IRB, the need for informed consent was waived due to the retrospective design and minimal risk to the patients.

Authors’ contributions

Conceptualization: HKO

Data curation: SSH, HP, HHS

Formal analysis: SSH, HKO, EJ

Investigation: SSH, HHS

Methodology: SSH, HKO, EJ, HA, ANS DWK, SBK

Supervision: HKO

Writing–original draft: SSH, HKO

Writing–review & editing: SSH, HKO, HRS, TGL, MJC, MHJ, HA, ANS, DWK, SBK

All authors read and approved the final manuscript.

Conflict of interest

Heung Kwon Oh, serving as the editorial board of Journal of Minimally Invasive Surgery, did not participate in the review process of this article. No other potential conflicts of interest pertinent to this article were reported.

Funding/support

None.

Acknowledgments

The authors would like to acknowledge the presentation of this research at the ACKSS 2023 where it received valuable feedback.

Data availability

The data presented in this study are available upon reasonable request to the corresponding author.