Abstract
Background:
The growing number of premalignant colonic lesions undergoing surgical treatment can lead to increased overtreatment.
Objectives:
We assessed the magnitude of overtreatment by comparing rates of adverse events related to surgical and endoscopic treatment of complex benign polyps.
Design:
This was a single-center retrospective study conducted at a tertiary care hospital.
Methods:
This study included patients with benign colonic lesions treated surgically during 2005–2021 and compared with a cohort with complex lesions (Size, Morphology, Site, and Access (SMSA) ⩾3) treated endoscopically during 2018–2022. Adverse events, need for reintervention, mortality, and length of hospital stay were compared using propensity score (PS)-matching analysis. The cohorts were matched 1:1 with adjustment for sex, age, SMSA, and size as covariates. Surgical and endoscopic adverse events were described using the Clavien–Dindo (surgical group) and AGREE (endoscopic group) classifications.
Results:
Of 240 included patients, PS matching yielded 71 pairs. Adverse events were more frequent with surgical treatment (odds ratio (OR) 3.27; 95% confidence interval (CI) 1.59–6.71), as were severe adverse events (OR 7.5; 95% CI 2.1–27.0), need for reintervention (OR 25.6; 95% CI 3.3–200.0), and mean length of hospital stay (10 vs 0 days, p < 0.001). One (1.4%) patient in the surgical group and none in the endoscopic group died (p = 0.39).
Conclusion:
An excess of severe adverse events with surgical treatment of complex benign polyps reflects overtreatment. Adequate pathways must be established for referral of these lesions for endoscopic treatment.
Plain Language Summary
Background and purpose:
More and more people with certain pre-cancerous colon growths are undergoing surgery, which might not always be necessary. This study looked at how often surgery leads to complications compared to a less invasive procedure called endoscopy.
How the study was done:
Researchers reviewed past medical records from one hospital. They compared patients who had surgery for benign (non-cancerous) colon growths between 2005 and 2021 with those who had endoscopic treatment between 2018 and 2022. They matched patients based on factors like age, sex, and the complexity of the growths to make the comparison fair. The study examined complications, the need for further treatment, hospital stay length, and death rates.
Findings:
Out of 240 patients, 71 pairs were closely matched for comparison. Surgery led to more complications than endoscopy. Serious complications were also much more common with surgery. Patients who had surgery were more likely to need additional procedures and stayed in the hospital much longer (an average of 10 days versus none for the endoscopy group). One person in the surgery group died, while no one in the endoscopy group did.
Conclusion:
Many patients with complex but non-cancerous colon growths may be undergoing unnecessary surgery, leading to more risks and longer recovery times. A better system should be in place to ensure these patients are referred for endoscopic treatment instead.
Keywords
Introduction
Endoscopic polypectomy has led to reduced colorectal cancer (CRC) incidence and mortality.1,2 All endoscopists routinely perform polypectomy, allowing for the resection of most premalignant lesions. However, some lesions are considered highly difficult for endoscopic treatment because of the risk of adverse events such as perforation or bleeding, incomplete resection, or recurrence. These lesions have frequently been referred for surgical treatment, 3 leading to overtreatment.
A precise definition of overtreatment is not available. 4 Some have argued that the treatment of overdiagnosed polyps should be considered overtreatment, even when a specific patient who has been overtreated cannot be clearly identified. 5 Once a pre-malignant colonic polyp is detected, the potential exists for it to become harmful, necessitating treatment of the lesion. This built-in uncertainty may lead to overtreatment, when the proposed treatment results in more adverse events than does the standard option as a comparator. 6
Techniques such as endoscopic mucosal resection or endoscopic submucosal dissection have supported endoscopic treatment of difficult lesions, minimizing the need for surgery. 7 In this context, surgical treatment of benign colonic lesions has emerged as a cause of overtreatment. In the United States, 25% of scheduled colectomies for colonic neoplasia are performed for non-malignant colorectal polyps. 3 Observational studies have shown that compared with endoscopic resection, surgery is associated with higher morbidity and mortality in the management of complex lesions.3,8,9 Only one study has compared adverse events between cohorts undergoing surgery versus endoscopic resection, but that investigation lacked a paired comparison of adverse events and adequate matching of patients in terms of treatment. 10
The principal aim of this study was to describe the magnitude of overtreatment of benign colonic lesions, defining “overtreatment” as an increase in adverse events with surgery compared with endoscopic approaches. We conducted a retrospective study with propensity score (PS) matching to compare 30-day adverse events in patients with benign colonic lesions treated surgically or endoscopically at our center. We also compared the need for reoperation, length of hospital stay, and mortality between the two groups. To address the question of future risk with the more conservative endoscopic approach, a secondary aim was to determine the prevalence of early and late recurrence within the endoscopic treatment group.
Materials and methods
In this observational, single-center study, we included in the first cohort all patients with benign lesions diagnosed by colonoscopy who underwent scheduled surgical treatment during 2005–2021 at the Hospital General Universitario Doctor Balmis in Alicante. The second cohort included all patients with benign colonic lesions classified as complex following the Size, Morphology, Site, and Access (SMSA) classification (SMSA ⩾3) and who received endoscopic treatment during 2018–2022 at the same center, after the introduction of a specific program for endoscopic treatment of difficult lesions. This study was designed in compliance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. 11 A study flow diagram (Figure 1) has been included to illustrate the study process, and the STROBE checklist has been provided as Supplemental Table 1.
The selective flow chart of the study.
Lesion-related definitions
Benign lesions were defined as adenomas with low- or high-grade dysplasia or serrated lesions with or without dysplasia. The left colon location included the descending colon and splenic flexure, and the right colon location included the transverse colon, ascending colon, cecum, and ileocecal valve. Lesion morphology was established using the Paris classification.12,13 The complexity of polyps for endoscopic resection was described using the SMSA classification, which is a scoring system that defines polyp complexity in four levels based on size, morphology, location, and accessibility, and has proved feasible and reliable for planning treatment (Supplemental Table 2). 14 The total score is obtained by summing the points from each domain and is classified into difficulty levels: Level 1: 4–5 points; Level 2: 6–8 points; Level 3: 9–12 points; and Level 4: 13–16 points.
Inclusion and exclusion criteria
Data on patients with benign lesions referred for surgical treatment were obtained from the electronic database of the pathology department, after a search of surgical specimens with a diagnosis of adenoma, serrated polyp, intramucosal carcinoma, or carcinoma in situ. Manual selection of the clinical and pathology reports of these patients was performed. In 2018, a dedicated endoscopy agenda for advanced colonic lesions was established, with two endoscopists trained in the treatment of these lesions (C.M.-S. and J.M.-S.). Data for cases included in this agenda were prospectively collected. Patients with colonic polyps with SMSA ⩾3 during 2018–2022 were selected for comparison with surgical cases and had been treated using endoscopic mucosal resection or endoscopic submucosal dissection. Patients were selected consecutively to ensure a systematic and unbiased inclusion process.
Patients with infiltrating CRC histology (pT1 or higher), a diagnosis of inflammatory bowel disease, familial adenomatous polyposis, or >20 polyps were excluded. We also excluded patients treated with transanal surgery.
Definition of adverse events
We defined adverse events using the Clavien–Dindo classification 15 for surgical adverse events and the AGREE classification 16 for endoscopic adverse events (Supplemental Tables 3 and 4). Severe adverse events were classified as grade 3A or higher in both classifications and included suture dehiscence, infections and collections, ostomy, eventration/evisceration, or paralytic ileus. Severe endoscopic adverse events were defined as 3A or higher using the AGREE classifications and included hemorrhage with hospital admission or perforation.
Statistical analysis
Variables were expressed as counts and frequencies (%) for categorical variables, means (±standard deviations, SDs) for continuous variables, and medians with interquartile ranges (25th–75th) for discrete variables, depending on whether they followed a normal distribution. To compare differences in tumor characteristics, staging, and treatment options between the two time periods, we performed univariable analyses with the Chi-squared test for categorical data. To account for multiple comparisons and control the Type I error rate, the Bonferroni correction was applied. Given that five statistical tests were performed, the significance threshold was adjusted by dividing the original significance level (α = 0.05) by the number of comparisons. Thus, results were considered statistically significant if p < 0.01 to ensure rigorous control of false-positive findings. We also report 99% confidence intervals (CIs) for proportions.
Adverse events, need for reintervention, mortality, and length of hospital stay were compared between the two groups using PS-matching analysis. This method was applied to adjust for significant differences in baseline characteristics between the cohorts and to reduce the influence of possible confounding factors. 17 The two groups were matched 1:1 (71 patients in each group) with adjustment for four covariates (sex, age (±1 year), SMSA, and size). We use nearest neighbor PS matching methods. The caliper width of PS matching was 0.6. All statistical analyses were performed with SPSS 25.0 software (IBM Corp., Armonk, NY, USA) and R software (version 4.4.2; R Foundation for Statical Computing, Vienna, Austria).
Results
We included 240 patients, 136 (56.7%) in the endoscopic group and 104 (43.3%) in the surgical group. Overall, 161 (67.1%) were male, and the mean age was 68 years (SD 10).
Characteristics of benign lesions
Regarding the lesions (Table 1), most were located in the right colon in both cohorts (61.5% surgical group vs 58.8% endoscopic group; p = 0.67). Sessile polyps (0–Is) were the most frequent in the surgical group (55.9%), whereas raised flat (0–IIa) lesions predominated in the endoscopic group (42.3%). The mean lesion size was 43 mm (SD 20) in the surgical group and 36 mm (SD 15) in the endoscopic group (p < 0.001).
Characteristics of patients and colonic lesions by treatment group before and after PS matching.
PS, propensity score; SD, standard deviation; SMSA, size, morphology, site, and access.
The most common histology was adenoma (89.4% in the surgical group vs 88.2% in the endoscopic group; p = 0.77), and the remainder were serrated polyps. Patients in the surgical group had a significantly higher proportion of lesions with high-grade dysplasia (54.8% vs 23.5%, p < 0.001).
In the surgical group, 75% of patients had a SMSA score >12 (SMSA 4), 17.3% had a score of 10–12 (SMSA 3), and 7.7% had a score of 6–9 (SMSA 2). By contrast, the endoscopic group had significantly fewer lesions with a score >12 (51.5%; p < 0.001) and a higher number of lesions with a score of 10–12 (48.5%; p < 0.001).
Comparison of clinical outcomes between the two groups by PS matching
Table 1 shows a comparison of clinical characteristics between the two groups before and after PS matching. Before PS matching, median size, Paris morphology classification, dysplasia, and SMSA score differed significantly between the two groups. PS matching led to 71 matched pairs and averaged the differences for four covariates. Supplemental Table 5 shows the characteristics of patients who were excluded after PS matching.
Adverse events
Table 2 compares adverse events after PS matching between the two groups. In the surgical group, 36 patients (50.7%) had any adverse event in the first 30 days post-surgery. By contrast, adverse events happened in 17 patients (23.9%) in the endoscopic group, for a calculated odds ratio (OR) of 3.3 (95% CI 1.3–8.5).
Adverse events, reintervention rate, mortality, and hospital length of stay by treatment group.
As can be seen in Table 3, the most frequent adverse event in the endoscopic group was immediate bleeding (11 events, 15.5%). In the operated patients, the most frequent adverse events were surgical wound infection (16 events, 22.5%) and suture dehiscence (14 events, 19.4%). Thirteen patients (18.3%) in the surgical group needed an ostomy after reintervention.
Adverse events by treatment group.
Patients can have had more than one adverse event.
When the adverse events were classified according to Clavien–Dindo and AGREE (Table 4), grade I adverse events predominated in the endoscopic group, including needing a stay <24 h and/or only analgesic, diuretic, or antipyretic treatment. By contrast, grade III adverse events predominated in the surgical group, including a need for surgical or interventional radiology treatment, and specifically grade IIIb adverse events, which are those requiring general anesthesia. Severe adverse events classified as grade 3 or higher in the Clavien–Dindo or AGREE classification arose in 21 patients (29.6%) in the surgery group compared with three patients in the endoscopic group (4.2%; OR 7.5, 95% CI 1.4–39.8; Tables 2 and 4).
Adverse events according to the Clavien–Dindo (surgical group, n = 17) and AGREE (endoscopic group, n = 36) classifications.
Only the most advanced adverse event was selected.
As Table 4 shows, 23 patients (32.4%) in the surgical group compared with one patient (1.4%) in the endoscopic group had post-treatment adverse events requiring surgical treatment (reintervention in the surgical cases or surgery in the endoscopic cases: OR 25.6, 95% CI 1.73–378.0; p < 0.001). One patient died in the surgical cohort due to a suture dehiscence that led to septic shock and, ultimately, death; and none in the endoscopic cohort (1.4% vs 0%; p = 0.39). The surgical group had a median hospital stay of 10 (7–21) days. By contrast, the endoscopic group had a median stay of 0 (0–2) days (p < 0.001).
The endoscopic group had eight cases (11.3%) of recurrence, five of them at the first surveillance endoscopy (62.5%), and two at the second (25.0%). All recurrences were successfully treated endoscopically.
Discussion
In this paired case–control study using PS matching, endoscopic treatment of benign complex colonic lesions was associated with significantly fewer adverse events, less need for surgical reoperation, and a shorter hospital stay compared with surgery. Our results quantify overtreatment in several ways in patients with benign colonic polyps treated surgically rather than endoscopically. The risk of severe adverse events was 7 times higher with surgery, and when patients were treated endoscopically, the vast majority of adverse events were resolved endoscopically. In addition, the rate of recurrence of endoscopically removed lesions was low, and they were easily addressed with endoscopy.
Surgery is still a common treatment for complex colonic polyps. An increase in the number of colonic resections for benign lesions has been reported in the United States and Europe,3,18 although recent data suggest a slight downward turn. 19 However, this analysis of a nationwide database still showed that 23.7% of colectomies in the United States were for benign neoplasms, 19 which, based on our findings, likely reflects considerable overtreatment. However, when adequate referral systems are established for treating complex polyps with endoscopy, the rate of surgical treatment of benign lesions clearly declines, 20 as we can see from the data of our center (Supplemental Figure 1).
Screening is a well-defined cause of overdiagnosis and overtreatment. As is well known, screen-detected cancer can be indolent in some cases and never cause symptoms. In such cases, treatment of these tumors will become overtreatment. The magnitude of overdiagnosis and overtreatment in CRC screening has not been clearly elucidated21,22 but the complete natural history of benign colonic lesions is not fully clarified, even for large lesions like those included in our study. The key issue with overtreatment in the context of screening is that it is impossible to know whether the identified abnormality will become a problem. Two influential factors in this prognostic gap are a lack of predictive accuracy at the individual level in terms of the condition evolving into advanced cancer and the fact that even if such predictive accuracy were possible, we would still need to assess what would be acceptable in terms of expected risk or harm. 6
Although surgery remains an adequate alternative when endoscopic treatment is not possible, the high rate of adverse events after colonic surgery is well known. In 2019, De Neree Tot Babberich et al. 23 published a systematic review of 26 articles and showed an increased risk of postoperative morbidity, with a pooled 1-month adverse event rate of 24%, increasing hospital length of stay, and costs. The adverse event rate with endoscopic treatment of benign lesions is clearly lower, with Hassan et al. 24 showing only an 8% adverse event rate in their systematic review. Moreover, most endoscopy adverse events tend to be resolved in the same procedure, without consequences for the patient.
These well-known differences between surgical and endoscopic treatment make a randomized controlled trial an unethical method for comparing these two approaches. As a result, no studies have directly compared surgical versus endoscopic treatment in cases involving polyps with similar characteristics. Here, we sought to overcome these limitations by comparing cases that were carefully matched by age, sex, polyp size, and level of difficulty with endoscopic treatment, as measured by the SMSA scale. 14 Moreover, adverse events were adequately classified using fully comparable scales. With this approach, we could directly estimate the level of overtreatment associated with surgery.
Overtreatment in general refers to medical interventions that are unwarranted because the recipients would be better off without them. 25 In this case, we demonstrate that patients with complex polyps are better treated with endoscopic resection and that efforts should target developing adequate referral processes for these patients. CRC screening, especially fecal immunochemical testing, identifies a number of patients with advanced complex polyps. To avoid overtreatment in these cases, clear and restrictive indications for surgery must be established. Moreover, screening programs must define pathways for adequate referral of patients to specialized endoscopic units that can resolve the case while avoiding surgical overtreatment. Because of the lack of availability of such advanced endoscopic treatment in all centers, referral for additional endoscopic resection occurs in a minority of cases. To increase referrals and a second attempt at endoscopic resection, multidisciplinary committees and effective communication between hospitals are crucial.18,26 Voloyiannis et al. 27 reported that almost 60% of patients in their study who were diagnosed with colonic lesions initially classified as endoscopically unresectable avoided surgical resection after a new therapeutic colonoscopy by an experienced endoscopist, showing the efficacy of this initiative.
Unnecessary treatment negatively affects quality of life. 28 Considering the adverse events that developed in the surgically treated patients, we found a surgical wound infection rate of 27.6% and an anastomotic leak rate of 22.4%. Peery et al. 3 included 12,732 patients undergoing surgical resection and reported a 30-day postoperative adverse event rate of 14%, with significantly lower rates of deep infection (0.7%) and anastomotic leak (2.6%) than we identified in our study. These findings may diverge because most surgeries in the Peery et al. study were laparoscopic, whereas most procedures in our study (72%) were open surgeries. Moreover, the rate of ostomies was 20.4%.
We identified no differences in mortality but did find a significantly lower rate of endoscopy-first patients requiring surgical reintervention (1.4%) versus the surgery-first group (32.4%, p < 0.001). The rate for the surgery group was much higher than reported in other studies because of the greater occurrence of anastomotic dehiscence in our patients. 23 As a consequence of the higher aggressiveness and longer recovery time, as well as the higher rate of reinterventions, our surgical cohort also had a longer hospital stay (10 vs 0 days). Most endoscopic procedures were outpatient, and some patients were admitted for 24-h observation. All of these adverse events from the surgical treatment doubtless negatively affected quality of life.
The main limitations of this study include its single-center design, which limits the external validity of our findings. Also, its retrospective nature may hinder the accurate calculation of certain variables, such as the SMSA score. Patient comorbidities were not considered in the analysis, which may represent a limitation and could affect the external validity of our findings. Moreover, the limitations of this study include the divergence of our findings from some previously published results, which we attribute to heterogeneity in defining adverse events. To address this issue, we used comparable adverse events classifications: the Clavien–Dindo surgical classification and the recently described AGREE classification for endoscopic adverse events.15,16 The different time frames for the cohorts could be a limitation as well because of older surgical cases, but we do not believe that this factor was likely to have affected our results. Most clinical practice guidelines recommend endoscopic resection as the first-line treatment for benign colorectal lesions 29 because endoscopic removal is associated with lower morbidity and mortality, with data based on retrospective studies in which the surgical and endoscopic cohorts differed by more than 10 years due to the shift in the treatment paradigm.10,30 We applied PS-matching analysis to minimize the differences in clinical characteristics between the two groups, but the results still should be interpreted with caution. Nevertheless, despite the favorable results of endoscopic treatment over surgical management for benign colorectal lesions in terms of complications, due to the retrospective nature of this study, further prospective studies with a larger number of participants should be conducted to validate our findings.
DuMontier et al. 28 defined overtreatment as using a treatment for a patient who would gain a greater net benefit from less intensive therapy. Our findings illustrate exactly this pattern. We found that therapeutic endoscopy is a safe intervention and may offer advantages over surgery in terms of morbidity and length of hospital stay. We therefore consider that the endoscopic approach should be the treatment of choice for the vast majority of complex colonic polyps. It is necessary to promote the creation of specialized units to treat these patients and to invest in endoscopist training to provide patients with the best available treatment while avoiding unnecessary adverse events.
Supplemental Material
sj-docx-1-tag-10.1177_17562848251351214 – Supplemental material for Overtreatment in colorectal cancer prevention: comparison between surgical and endoscopic treatment of benign colonic polyps
Supplemental material, sj-docx-1-tag-10.1177_17562848251351214 for Overtreatment in colorectal cancer prevention: comparison between surgical and endoscopic treatment of benign colonic polyps by Noelia Sala-Miquel, Lucía Medina-Prado, Carolina Mangas-Sanjuan, Sandra Baile-Maxía, Cristina Alenda, Lucía Madero-Velázquez, Francisco A. Ruiz-Gómez, Eva Serrano, Enrique Santana, Victor Ausina, María Sáez-Rico, Pedro Zapater, Juan Martínez-Sempere and Rodrigo Jover in Therapeutic Advances in Gastroenterology
Footnotes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
