Short- and Long-term Outcomes of 3-Row Vs 2-Row Circular Staplers for Low Colorectal Anastomosis After Total Mesorectal Excision: A Single-Center, Randomized, Single-Blind, Phase III Trial (THREESTAPLER Study)

Abstract

BACKGROUND:

Anastomotic leakage remains a critical complication in colorectal surgery.

OBJECTIVE:

This study compared the efficacy of 3-row versus 2-row circular staplers in reducing anastomotic leakage rates.

DESIGN:

Prospective, randomized, single-center, noninferiority trial.

SETTINGS:

Clinic of Colorectal and Minimally Invasive Surgery, University Hospital No. 2, Sechenov University.

PATIENTS:

The trial included patients who underwent low anterior resection with total mesorectal excision.

INTERVENTIONS:

Patients were randomly assigned in a 1:1 ratio to the 2-row or 3-row stapler group using centralized, computer-generated randomization. Outcome adjudicators, data collectors, and data analysts were blinded.

MAIN OUTCOME MEASURES:

Anastomotic leakage incidence, intra- and postoperative anastomotic bleeding, 30-day complications, low anterior resection syndrome, anastomotic stenosis, adverse events, time to and permanence of stoma, and 5-year overall survival are the main outcome measures.

RESULTS:

Among 154 patients included in the study, the anastomotic leakage rate was 16.9% (13/77) in the 3-row group compared to 15.6% (12/77) in the 2-row group (p = 0.8). The observed difference between groups was 1.3%, with a 1-sided 95% CI bounded below by –8.47%. The rates of anastomotic stricture were 1.4% (n = 2) in the 2-row group and 0.7% (n = 1) in the 3-row group (p = 1). No patients experienced anastomotic hemorrhage. Stoma closure was successfully performed in 147 patients (95.5%). The 5-year overall survival was 80.2% (95% CI, 69.4%–92.8%) in the 3-row group versus 83.8% (95% CI, 75.6%–92.3%) in the 2-row group (p = 0.80).

LIMITATIONS:

Routine postoperative proctography was not performed, so clinically silent (grade A) leaks may have been missed; the study was underpowered because the assumed 12% anastomotic leakage difference used for sample-size estimation was optimistic; neoadjuvant regimen heterogeneity may have influenced outcomes; and the single-center design may limit generalizability.

CONCLUSIONS:

The 3-row stapler was not noninferior to the 2-row stapler in preventing anastomotic leakage. See Video Abstract.

CLINICAL TRIAL REGISTRATION NUMBER:

NCT03910699

RESULTADOS A CORTO Y LARGO PLAZO DE LAS GRAPADORAS CIRCULARES DE TRES FILAS FRENTE A LAS DE DOS FILAS PARA LA ANASTOMOSIS COLORRECTAL BAJA TRAS LA ESCISIÓN TOTAL DEL MESORRECTO: UN ENSAYO CLÍNICO DE FASE III, ALEATORIZADO, SIMPLE CIEGO Y DE UN SOLO CENTRO (ESTUDIO THREESTAPLER)

ANTECEDENTES:

La fuga anastomótica sigue siendo una complicación crítica en la cirugía colorrectal.

OBJETIVO:

Este estudio comparó la eficacia de las grapadoras circulares de tres filas frente a las de dos filas para reducir las tasas de fuga anastomótica.

DISEÑO:

Ensayo prospectivo, aleatorizado, unicéntrico y de no inferioridad.

LUGAR:

Clínica de Cirugía Colorrectal y Mínimamente Invasiva, Hospital Universitario Nº 2, Universidad Sechenov.

PACIENTES:

El ensayo incluyó pacientes sometidos a resección anterior baja con escisión total del mesorrecto.

INTERVENCIONES:

Los pacientes fueron aleatorizados 1:1 al grupo de grapadoras de dos o tres filas mediante aleatorización centralizada generada por computadora. Los evaluadores de resultados, los recolectores de datos y los analistas de datos estaban cegados.

PRINCIPALES MEDIDAS DE RESULTADO:

Incidencia de fuga anastomótica; sangrado anastomótico intra y postoperatorio; complicaciones a los 30 días; síndrome de resección anterior baja; Estenosis anastomótica; eventos adversos; tiempo hasta la formación y permanencia del estoma; y supervivencia global a 5 años.

RESULTADOS:

Entre los 154 pacientes incluidos en el estudio, la tasa de fuga anastomótica fue del 16,9 % (13/77) en el grupo de 3 filas, en comparación con el 15,6 % (12/77) en el grupo de 2 filas (p = 0,8). La diferencia observada entre los grupos fue del 1,3 %, con un intervalo de confianza unilateral del 95 % inferior a -8,47 %. Las tasas de estenosis anastomótica fueron del 1,4 % (n = 2) en el grupo de 2 filas y del 0,7 % (n = 1) en el grupo de 3 filas (p = 1). Ningún paciente presentó hemorragia anastomótica. El cierre del estoma se realizó con éxito en 147 pacientes (95,5 %). La supervivencia global a los 5 años fue del 80,2 % (intervalo de confianza del 95 %, 69,4-92,8) en el grupo de tres filas frente al 83,8 % (intervalo de confianza del 95 %, 75,6-92,3) en el grupo de dos filas (p = 0,80).

LIMITACIONES:

No se realizó proctografía postoperatoria de rutina; es posible que se hayan pasado por alto fugas clínicamente silentes (grado A); el estudio tuvo una potencia estadística insuficiente, ya que la diferencia del 12 % en la tasa de fugas anastomóticas estimada para el cálculo del tamaño muestral fue optimista; la heterogeneidad del régimen neoadyuvante pudo haber influido en los resultados; y el diseño unicéntrico puede limitar la generalización de los resultados.

CONCLUSIONES:

La grapadora de tres filas no demostró ser no inferior a la de dos filas en la prevención de fugas anastomóticas. (AI-generated translation)

Número de registro del ensayo clínico: NCT03910699

Video Abstract

Video Abstract.

Download video file ^{(131.6MB, mp4)}

Anastomotic leakage (AL) remains a critical complication after colorectal resection, substantially increasing the risks of morbidity and mortality. The consequences of AL can be severe, including peritonitis, sepsis, prolonged hospital stay, reoperation, and an increased risk of local recurrence, which can ultimately lead to life-threatening outcomes.¹

The development of circular staplers has profoundly transformed the landscape of colorectal surgery by providing a reliable method for anastomotic formation. Since its initial implementation, this technology has undergone substantial evolution, with continuous refinements aimed at enhancing anastomotic integrity and reducing postoperative complications. Historically, this pivotal innovation originated in the Soviet Union, where Burtsev et al² invented the first circular stapler at the Maxim Gorky Donetsk Medical Institute (model PKS-25, 1957) in 1957 and they advanced stapler design by introducing a circular stapler with a 2-row staggered staple configuration in 1975.^3,4 The first major clinical outcomes after the introduction of the circular stapler for rectal resection were reported in 1971 by Alexandrov et al.⁵ Subsequent evolution has focused on optimizing stapler design. For instance, matching staple height to tissue thickness is done to achieve optimal compression without causing ischemia. Stapler shapes (eg, B-shape, 3-dimensional) influence wound closure strength, and the multirow configuration aims to enhance hemostasis and evenly distribute compression.⁶ In addition, powered mechanisms have been introduced with the goal of improving tissue compression, hemostasis, and ease of use.⁷

A recent innovation is the introduction of a manual 3-row circular stapler, which offers a theoretical advantage over the established 2-row design through enhanced hemostasis and more uniform tissue compression, which can potentially reduce the risk of AL. In addition, publicly available pricing indicates that the 3-row stapler is approximately 60% to 70% less expensive than comparable 2- or 3-row staplers in some markets, representing a potentially substantial cost saving. However, prospective clinical data validating its safety and efficacy remain limited.

This evidence gap underscores the necessity of conducting rigorous randomized controlled trials with each major technological iteration. Therefore, we designed the THREESTAPLER trial, a randomized controlled trial, to evaluate whether the use of a manual 3-row circular stapler is noninferior to a manual 2-row circular stapler in reducing the incidence of AL after low anterior resection (LAR) with total mesorectal excision (TME).⁸ This investigation seeks to establish this innovation as a safe and viable alternative.

MATERIALS AND METHODS

Study Design

This was a prospective, randomized, controlled, single-blinded, single-center, noninferiority trial with 2 parallel groups and a 1: 1 allocation ratio. The study was approved by the Ethics Committee of Sechenov University (No. 04-19; June 3, 2019). The trial was registered on ClinicalTrials.gov (NCT03910699, https://www.clinicaltrials.gov/study/NCT03910699?term=NCT03910699&rank=1) and the protocol was published in 2019. A detailed description of the trial design has been reported previously.⁸

Study Population

Consecutive adult patients with histologically confirmed adenocarcinoma of the rectum (stage cT0–4a, N0–2, M0–1), located 5 to 15 cm from the anal verge, who were scheduled to undergo elective LAR with TME were systematically screened for eligibility. All patients meeting the inclusion criteria were invited to participate. The inclusion criteria were provision of signed and dated informed consent; stated willingness to comply with all study procedures and availability for the duration of the study; age 18 years or older; histologically confirmed rectal adenocarcinoma located 5 to 15 cm from the anal verge; and planned elective LAR with TME. Exclusion criteria included patient refusal to participate, T4b tumors involving adjacent organs, inability to preserve the left colic artery, inability to perform a double-stapled anastomosis due to a narrow rectal lumen, active pelvic or systemic infection, preexisting anal incontinence (Wexner score 10 or higher), and pregnancy.

Study End Points

The primary end point of this study was the incidence of AL, which was radiologically confirmed by water-soluble contrast enema (proctography). All cases of AL were defined and graded according to the standardized criteria established by the International Study Group of Rectal Cancer.⁹ Key secondary end points included intraoperative and postoperative anastomotic bleeding rates; the overall 30-day postoperative complication rate, graded using the Clavien-Dindo classification; the incidence and severity of LAR syndrome (LARS), assessed with the LARS score; and the incidence of anastomotic stenosis, graded according to the classification system by Cong et al.¹⁰ Additional secondary end points comprised the rate of adverse events, time to stoma closure, rate of permanent stoma, and 5-year overall survival.

Outcome Assessment and Follow-up

For the primary outcome, AL, surveillance extended from the time of initial surgery until the subsequent stoma reversal procedure. Tension was evaluated intraoperatively by confirming that the proximal limb of the bowel, once brought down for a pelvic anastomosis, does not make contact with the coccyx and the presacral fascia in the depths of the pelvis. This observation is inherently a subjective judgment based solely on the surgeon’s personal experience. All postoperative complications occurring during the index hospital admission or within 30 days of surgery were rigorously documented and severity-graded using the Clavien-Dindo classification. Anastomotic stenosis was suspected in patients reporting symptoms of difficult evacuation or narrowed stool. The diagnosis was confirmed either by digital rectal examination revealing a narrowed lumen that did not allow the passage of the examiner’s index finger or radiologically via water-soluble contrast enema (proctography) demonstrating a persistent narrowing of the anastomotic lumen. Confirmed cases were graded according to the classification by Cong et al.¹⁰ The anastomosis was assessed for stenosis from the initial operation until 1 year after stoma reversal, with confirmation via diagnostic imaging. Functional outcomes were measured using the LARS score at 3, 6, and 12 months after stoma reversal. Oncological surveillance for local and distant recurrence was conducted through clinical and radiological evaluation at 3, 6, and 12 months postoperatively, continuing annually thereafter for up to 5 years; this protocol included an annual CT scan of the chest, abdominal MRI, pelvic MRI, and colonoscopy.

Randomization Procedure

An independent statistician generated the allocation sequence using a block-randomization method with varying block sizes ranging from 4 to 6. This sequence was concealed and stored by the central trial office at Sechenov University. Randomization was performed intraoperatively. After the resection of the rectum was completed and immediately before the creation of the anastomosis, the attending surgeon initiated randomization by contacting the central trial office via a dedicated telephone line. A research fellow, who was not involved in the patient’s care or surgical procedure and had no knowledge of the patient’s details beyond a study ID, then revealed the next allocation in the sequence. The research fellow communicated the group assignment directly to the surgical team in the operating room: double stapling technique or triple-stapling technique. The team then proceeded to create the anastomosis using the assigned stapling technique. This process ensured robust allocation concealment until the last possible moment, minimizing selection bias.

Blinding

This trial was conducted in a single-blind manner. Due to the technical nature of the intervention, the operating surgeons were not blinded to the stapler type. However, participants were blinded to the allocation to mitigate performance bias. To minimize assessment bias, outcome adjudicators, data collectors (research fellows), and data analysts were blinded to group assignments throughout the trial.

Data Oversight and Management

A Data Monitoring Committee was established to oversee trial integrity. The committee included independent colorectal surgeons, biostatisticians, ethicists, and senior researchers. The Data Monitoring Committee was responsible for safeguarding randomization integrity, adjudicating postoperative outcomes, and reviewing interim analyses. All data were anonymized and managed by Sechenov University research staff, who were randomly assigned and not involved in patient enrollment or intraoperative allocation, to ensure confidentiality and prevent handling bias.

Interventions

All operations were performed by experienced colorectal surgeons, each with >10 years of specialization and a minimum of 50 patients with rectal cancer per year. Overall, 6 surgeons participated in this study. Patients underwent preoperative mechanical bowel preparation. A D3 lymph node dissection with preservation of the left colic artery was routinely performed.¹¹ Subsequently, TME was conducted with meticulous preservation of the hypogastric nerves and pelvic nerve plexuses. The rectum was transected distal to the tumor using a linear stapler, ensuring a margin of at least 1 cm below the inferior tumor margin. After this, the rectal stump was irrigated. A circular stapler was inserted transanally, and the anvil needle was pierced through 1 edge of the linear staple line. The circular stapler was maintained in position for 15 seconds to ensure adequate tissue compression and hemostasis before complete firing and removal. The end-to-end colorectal anastomosis was performed using either a manual 2-row circular stapler (Ethicon CDH29A, 29 mm) or a manual 3-row circular stapler (Mirus MCS-29R3, 29 mm). The integrity of the 2 donuts of tissue was verified to be complete. An intraoperative air leak test was routinely performed to assess anastomotic integrity. The height of the anastomosis was measured in centimeters from the anal verge by digital examination. Consistent with institutional and prevailing national practice, a protective loop transverse colostomy and a silicone drain were placed in the pelvic cavity and performed in all patients.

Statistical Analysis

The sample-size calculation determined that 154 patients (77 per group) were required to achieve 80% statistical power at a 2-sided α level of 0.05, assuming a noninferiority margin of 5% for the primary end point (AL). This margin was prespecified on the basis of the published range of AL rates after LAR with TME (up to 30%) and the severe clinical consequences of AL,¹² including increased morbidity, mortality, and impaired oncological outcomes. In the context of an expected baseline leakage rate of approximately 16%, consistent with rates reported in contemporary trials, such as GRECCAR 5,¹³ a 5% absolute increase was considered the maximum clinically acceptable deterioration, corresponding to a relative risk increase of ~30%. No deviations from the predefined statistical analysis plan were made after patient enrollment began. Further details of the calculation are provided in the published protocol.⁸ Noninferiority margins were defined using the Newcombe-Wilson score method. Variables included in the multivariable logistic regression model were selected a priori based on clinical relevance and established risk factors for AL reported in the literature, including patient demographics (age, sex, BMI, ASA grade), tumor characteristics (location, stage), neoadjuvant therapy type, operative details (access, operative time, blood loss, anastomotic height, splenic flexure mobilization), and stapler type. Variable selection adhered to the principle of minimizing events per variable to ensure model stability, restricting inclusion to factors with the highest clinical significance. Continuous variables are presented as median and interquartile range and were compared between groups using the Mann-Whitney U test. Categorical variables are summarized as absolute numbers and percentages and were compared using Pearson’s χ² test or Fisher exact probability test, as appropriate. All tables were generated using the gtsummary package (version 2.3.0) in RStudio.¹⁴ Risk factors associated with AL were identified using multivariable logistic regression. Survival analysis was performed using Kaplan-Meier curves, and between-group differences were assessed with the log-rank test. Missing data were handled using a multivariate imputation algorithm implemented via the missRanger package in RStudio except for the overall survival analysis. All p values were 2-sided, and statistical significance was defined at a p value of <0.05. Analyses were conducted using the R Core Team version 4.5.1 (June 13, 2025).

RESULTS

Between May 2019 and October 2021, 171 patients were assessed for eligibility, with 154 meeting inclusion criteria and undergoing randomization (Fig. 1). Baseline demographic and clinical characteristics were well balanced between the 2 study groups, with no significant differences in age, sex distribution, pathological tumor stage, or tumor distance from the anal verge (Table 1). Operative outcomes showed no significant differences in median operative time (260 vs 255 minutes, p = 0.70) or blood loss (100 vs 90 mL, p = 0.20). There were no incidences of anastomotic hemorrhage in either group (Table 2). In addition, no technical difficulties with staplers were observed, except for the need for additional pressure during firing. All air leak tests were negative. In the 2-row stapler group, laparoscopic approach and upfront surgery were more frequently used.

FIGURE 1.

CONSORT 2025 flow diagram. CONSORT = Consolidated Standards of Reporting Trials.

TABLE 1.

Baseline characteristics

Characteristic	Three-row stapler group (N = 77)	Two-row stapler group (N = 77)	p
Sex			>0.99
Female	34 (44.2%)	34 (44.2%)	>0.99
Male	43 (55.8%)	43 (55.8%)	>0.99
Age, median (IQR)	63 (48–68)	60 (53–66)	0.73
BMI, median (IQR)	27 (24–30)	26 (22–29)	0.12
ASA classification			>0.99
I	7 (9.1%)	7 (9.1%)	>0.99
II	48 (62.3%)	47 (61%)	0.87
III	22 (28.6%)	23 (29.9%)	0.73
Tumor location			0.32
Upper rectum	4 (5.2%)	6 (7.8%)	0.51
Middle rectum	55 (71.4%)	60 (77.9%)	0.35
Lower rectum	18 (23.4%)	11 (14.3%)	0.15
Neoadjuvant therapy			0.10
RT	4 (5.2%)	6 (7.8%)	0.51
CRT	27 (35.1%)	17 (22.1%)	0.07
CT	6 (7.8%)	2 (2.6%)	0.28
None	40 (51.9%)	52 (67.5%)	0.049
pStage			0.95
0 (pCR)	2 (2.6%)	1 (1.3%)	>0.99
I	22 (28.6%)	24 (31.2%)	0.73
II	11 (14.3%)	9 (11.7%)	0.81
III	36 (46.8%)	38 (49.4%)	0.87
IV	6 (7.8%)	5 (6.5%)	>0.99

Data reported as n (%) unless otherwise indicated.

CRT = chemoradiotherapy; CT = chemotherapy; IQR = interquartile range; pCR = pathologic complete response; RT = radiotherapy.

TABLE 2.

Intraoperative characteristics

Characteristic	Three-row stapler group (N = 77)	Two-row stapler group (N = 77)	p
Access			0.18
Laparoscopic	34 (44.2%)	47 (61%)	0.04
Robot-assisted	29 (37.7%)	22 (28.6%)	0.23
Open	14 (18.1%)	8 (10.4%)	0.17
Conversion	3 (3.9%)	2 (2.6%)	>0.99
Operative time, min, median (IQR)	260 (215–315)	255 (210–305)	0.74
Blood loss, median (IQR)	100 (50–100)	90 (50–100)	0.25
Splenic flexure mobilization	12 (15.6%)	14 (18.2%)	0.67
Anastomosis location (from the anal verge), cm, median (IQR)	3 (3–4)	3 (3–4)	0.82
Tension of anastomosis	2 (2.6%)	0 (0%)	0.50
Anal tube	2 (2.6%)	4 (5.2%)	0.68

Data reported as n (%) unless otherwise indicated.

IQR = interquartile range.

Overall postoperative complication rates were comparable between the 3-row and 2-row stapler groups (Table 3). The most frequent Clavien-Dindo grade I and II complications observed in our study included fever, pain, wound infection, electrolyte imbalance, and plasma and albumin transfusion, occurring in 23.4% (18/77) and 14.3% (11/77) of patients (p = 0.42), respectively. One patient in the 3-row group experienced a thermal injury to the distal ureter resulting in urinary leakage, which was managed with percutaneous nephrostomy and ureteral stenting. All suspected leaks were investigated primarily with proctography.

TABLE 3.

Postoperative characteristics

Characteristic	Three-row stapler group (N = 77)	Two-row stapler group (N = 77)	p
Anastomotic leakage	13 (16.9%)	12 (15.6%)	0.83
Grade of anastomotic leakage			>0.99
B	9 (11.7%)	9 (11.7%)	>0.99
C	4 (5.2%)	3 (3.9%)	>0.99
Clavien-Dindo complication grade			0.40
I–II	18 (23.4%)	11 (14.3%)	0.10
III–IV	4 (5.2%)	4 (5.2%)	>0.99
Time to first flatus, d, median (IQR)	1 (1–1)	1 (1–1)	0.56
Time to first stool, d, median (IQR)	2 (2–3)	2 (2–3)	0.72
Time to anastomotic leakage, d, median (IQR)	6 (4–7)	6 (4–8)	0.44
Hospital stay, d, median (IQR)	13 (9–17)	11 (9–13)	0.07
Time to stoma closure, d, median (IQR)	39 (20–43)	40 (33–45)	0.24
Stoma closure, d	72 (93.5%)	75 (97.4%)	0.44
Anastomotic stenosis			>0.99
1	1 (1.3%)	1 (1.3%)	>0.99
2a	0 (0%)	1 (1.3%)	>0.99

Data reported as n (%) unless otherwise indicated.

IQR = interquartile range.

The primary end point of AL occurred in 13 patients (16.9%) in the 3-row group compared to 12 patients (15.6%) in the 2-row group (p = 0.83). The observed difference in the primary end point (AL) between the 3-row and 2-row stapler groups was 1.3% (16.9%–15.6%). The 1-sided 95% CI for this difference is defined by its lower bound of −8.47% (Fig. 2).

FIGURE 2.

Noninferiority for anastomotic leakage.

Grade B leaks occurred with equal frequency in both groups (11.7% each), whereas grade C leaks were observed in 5.2% (4/77) of 3-row patients vs 3.9% (3/77) of 2-row patients (p = 1). All cases of grade C AL (n = 7; 4 in the 3-row group and 3 in the 2-row group) presented with peritonitis and required urgent surgical reintervention. The reoperations uniformly consisted of abdominal lavage and placement of additional drains. The place of defect was difficult to determine, as endoscopy was not routinely performed. Anastomotic leaks were managed with washout. No grade A leaks were detected per study design, which mandated radiological confirmation for diagnosis. Tension was subjectively assessed as described. In 2 cases, the tension was noted but required no additional mobilization per surgeon decision, but 1 received a reinforcement suture. Postoperatively, both patients developed grade B anastomotic leaks. No postoperative mortality or Clavien-Dindo grade IV complications were observed in this study.

In a multivariable logistic regression analysis adjusting for patient demographics, tumor characteristics, neoadjuvant therapy, and operative variables, no factor was significantly associated with an increased risk of AL (all p > 0.05; Fig. 3).

FIGURE 3.

Adjusted ORs for risk factors associated with anastomotic leakage.

We performed univariate analysis to investigate the impact of neoadjuvant therapy on the development of AL (see Supplemental Tables 1–3 at https://links.lww.com/DCR/C651). There was no AL during stoma reversal. Anastomotic stenosis developed in 3 patients (3.9%), with 2 cases in the 2-row group and 1 in the 3-row group (p = 1). One of these patients (in the 2-row group) experienced concomitant AL.

At a median follow-up of 24 months, stoma reversal rates were similar between groups (93.5% vs 97.4%, p = 0.44). Stoma-free survival rates at various time points are presented in Figure 4 and Table 4. During the extended 5-year surveillance, 39 patients (25.3%) were lost to follow-up owing to inability to make contact despite repeated attempts (n = 15 for the 2-row group [19.5%]; n = 24 for the 3-row group [31.2%], p = 0.23). The 5-year overall survival was 80.2% (95% CI, 69.4%–92.8%) in the 3-row group versus 83.8% (95% CI, 75.6%–92.3%) in the 2-row group (p = 0.80; Fig. 5). The LARS scores after stoma reversal are presented in Table 5.

FIGURE 4.

Stoma-free survival.

TABLE 4.

Stoma-free survival at 2, 4, 6, 12, and 18 mo (95% CI)

Time point	Three-row stapler group (N = 77)	Two-row stapler group (N = 77)	p
2 mo	46 (33–56)	43 (30–53)	0.16
4 mo	78 (66–86)	64 (51–73)
6 mo	89 (79–94)	76 (64–84)
12 mo	97 (89–99)	95 (86–98)
18 mo	99 (90–100)	99 (91–100)

FIGURE 5.

Five-year overall survival.

TABLE 5.

LARS scores after stoma reversal

Characteristics	Three-row stapler group (N = 77)	Two-row stapler group (N = 77)	p
Time point, median (IQR)
3 mo	35 (28–41)	33 (26–40)	0.51
6 mo	29 (22–35)	28 (20–34)	0.64
12 mo	25 (18–32)	24 (17–30)	0.83
LARS category at 12 mo, n (%)			0.87
No LARS (0–20)	22 (28.6%)	25 (32.5%)	0.60
Minor LARS (21–29)	28 (36.4%)	26 (33.8%)	0.74
Major LARS (30–42)	27 (35.1%)	26 (33.8%)	0.87

IQR = interquartile range; LARS = low anterior resection syndrome.

DISCUSSION

This study represents the first randomized controlled trial comparing the 3-row circular stapler to the established 2-row device for anastomosis after LAR with TME. The primary analysis demonstrated that the lower bound of the CI for the difference in AL rates (−8.47%) exceeded the predefined noninferiority margin of Δ = −5%. Consequently, the noninferiority of the 3-row stapler could not be declared. The observed leakage rates were notably similar between the 2 groups (16.9% vs 15.6%), which contrasted with our initial hypothesis that anticipated a more pronounced reduction in leakage with the 3-row device. The absence of a clinically meaningful difference is the noteworthy finding of this trial. This trial was designed as an IDEAL (Idea, Development, Exploration, Assessment, Long-term Study) stage III assessment, a structured framework for evaluating surgical innovations, in which randomized controlled trials are recommended at this stage to provide rigorous comparative evaluation.¹⁵ Given the complete absence of prospective data, we adopted a deliberate conservative safety assumption of sample-size calculation (30% vs 18% AL rate) to ensure the study was powered to detect a clinically unacceptable increase in risk, prioritizing patient safety and ethical rigor. The 30% rate represents the upper bound of the widely reported range of AL incidence after rectal surgery, as summarized by Tsai et al¹² and Challine et al.¹⁶ The 18% rate was based on the contemporary report of symptomatic AL in similar populations by Challine et al¹⁶ and Matthiessen et al¹⁷ and is consistent with the historical rate at our institution. These figures reflect the substantial baseline leakage risk in this high-risk patient population and were deliberately chosen to ensure the trial could detect a clinically meaningful increase in this risk. Consequently, the trial’s primary strength lies in its ability to definitely rule out a dangerously high complication rate for the new device. The observed near-identical outcomes demonstrate comparable performance of the 3-row circular stapler in clinical practice to that of the established 2-row stapler. Although the conservative design of this study resulted in a wide CI, it also provided the necessary data to rule out clinically important safety risks for future research. It is crucial to interpret the trial’s primary outcome within the appropriate context. In the complete absence of prospective safety data, the fundamental question was whether this device caused unacceptably high complication rates. This foundational safety evaluation was a necessary prerequisite before any cost-effectiveness or superiority comparison could be meaningfully pursued.¹⁵

Despite advancements in surgical techniques and technologies, the incidence of AL remains high, ranging from 3% to 20.3%.^18–23 Only 3 retrospective studies have investigated the tristapling technique in relation to AL after TME.^18–20 The overall incidence of AL observed in this study was 16.3%, which is higher than that reported in other studies. In contrast to previous studies, which found a statistically significant difference between 2-row and 3-row stapler techniques,^18–20 our study found no such difference. The rate of grade B AL in the present study was 11.7% in each group. This finding is consistent with one study reporting a rate of 11.4%, but it contrasts with another that demonstrated a lower incidence of 2.5%.^20,23 Although the direct comparisons should be made with caution due to variations in the definition of LAR, as in our trial, LAR was strictly defined in conjunction with TME as a standardized end point. The incidence of grade C AL was 5% in both the present study and a previous report. The absence of grade A AL may be attributed to the lack of routine proctography in asymptomatic patients, leaving clinically insignificant leaks undetected.

Several intraoperative factors, including operation time, intraoperative blood loss, technical errors, and tension at the anastomotic site, are known to contribute to a higher incidence of AL.²⁴ However, the multivariable analysis in this study showed that none of these factors reached statistical significance as independent predictors. Concerns have been raised that staplers may not always uniformly distribute tension or achieve adequate hemostasis, particularly in cases involving thickened or edematous tissues, which could theoretically increase the risk of AL.²⁵ However, the assessment of these specific factors can be highly subjective. Further ex vivo study may help clarify this. Notably, both cases with intraoperative tension noted occurred in the 3-row group and both resulted in grade B leaks. Although excluding these cases would have reduced the AL rate in the 3-row group, this would introduce selective bias. This observation highlights tension as a critical risk factor and underscores the importance of intraoperative judgment, suggesting that surgeons should have a low threshold for additional mobilization when tension is suspected.²⁴

The 3-row stapler represents a technological evolution, as the additional row of staples is theorized to enhance hemostasis and improve tissue compression.¹⁸ Notably, the present study observed no incidents of anastomotic hemorrhage. This finding is consistent with other studies reporting low hemorrhage rates, ranging from 0% to 2.2% for the 3-row circular stapler group and 0% to 3% for the 2-row circular stapler group.^19,20 This lack of a significant difference is further supported by the meta-analysis conducted by McKechnie et al,²⁶ which found no statistically significant difference in hemorrhage rates between the stapler types.

After colorectal surgery, the incidence of anastomotic stenosis ranges from 3% to 30%. In the present study, anastomotic stenosis was observed in 2 patients (1.4%) in the 2-row group and 1 patient (0.7%) in the 3-row group, a rate that is comparable to that reported in other studies.¹⁹ The presence of a protective stoma may be a contributing factor, potentially due to the lack of fecal stimulation, which can lead to stenosis.²⁷ In addition, tissue ischemia poses a significant risk. In cases of LAR, factors such as excessive tension, edema, and thickening of the intestinal wall can contribute to the development of anastomotic stenosis.²⁸

As for the relationship between neoadjuvant therapy and AL, some studies, such as the multicenter analysis by Park et al,²⁹ have identified preoperative chemoradiation as a significant risk factor for leakage. Similarly, a meta-analysis by Pommergaard et al³⁰ reported that preoperative radiotherapy may be associated with an increased AL. However, more recent studies challenged this association. Brisinda et al³¹ found that long-course chemoradiotherapy did not significantly increase AL rates, whereas Arrezo et al³² concluded that overall neoadjuvant therapy was not linked to AL, but short-course radiotherapy can be associated with leakage. A meta-analysis that included more than 9600 patients demonstrated no significant correlation between neoadjuvant and AL incidence.³³ Our results align with this latter body of evidence, showing comparable leakage rates between stapler groups despite an imbalance in neoadjuvant treatment. This suggests that although neoadjuvant therapy may be one of several contributing factors, it is unlikely to be a primary independent driver of leakage risk.

AL is a significant contributing factor to the conversion of temporary stomas into permanent ones. In this study, the aim was to close patient stomas within 6 weeks after the initial surgical procedure. Existing literature indicates that failure to reverse a stoma within 1 year often results in a permanent stoma.³⁴ In our cohort, stoma reversal was successfully performed in 95.5% of patients. In the remaining cases, stoma closure was precluded either by patient mortality or by cancer recurrence that necessitated a permanent stoma.

This study has several limitations. First, postoperative proctography was not routinely performed; consequently, clinically silent (grade A) anastomotic leaks may have gone undetected. This suggests that the true AL rate may be higher in both groups than reported. Second, the study was underpowered to detect a clinically significant difference between groups, as the assumed 12% difference in AL rates used for the sample-size calculation was overly optimistic. Third, heterogeneity among groups regarding neoadjuvant treatment regimens could have influenced the outcomes. Fourth, although our multivariable analysis controlled for key tumor and treatment variables, we did not systematically collect or analyze established patient-specific risk factors for AL, including nutritional status, specific comorbidities, smoking history, or steroid use. The potential confounding effect cannot be excluded. Finally, the single-center design and investigation of only 2 specific types of circular staplers may limit the generalizability of the findings.

CONCLUSIONS

This study did not demonstrate the noninferiority of the 3-row stapler for preventing AL. However, no clinically significant differences were found between the 3-row and 2-row devices in leakage rates, other complications, or long-term functional and oncological outcomes. Therefore, the 3-row stapler can be considered a safe clinical alternative. The choice of device may depend on surgeon preference, cost, and availability. Further studies with larger sample sizes are warranted.

ACKNOWLEDGMENTS

The data sets generated and/or analyzed during the current study are not publicly available due to confidentiality and institutional policy but are available from the corresponding author on reasonable request.

Supplementary Material

dcr-69-2500-s002.pdf ^{(40.3KB, pdf)}