Laparoscopic management of duodenal atresia: a decade of single surgeon’s perspective

Abstract

Purpose

This study aimed to evaluate the outcomes and learning curve of laparoscopic duodenal atresia repair performed by a single pediatric surgeon over 10 years and compare it with traditional open repair.

Methods

This retrospective study included 28 neonates who underwent surgical repair for congenital duodenal atresia between 2010 and 2024 at a tertiary center. In total, 20 patients underwent laparoscopic repair by a single surgeon between 2013 and 2024, whereas 8 underwent open repair between 2010 and 2014. Clinical characteristics, operative times, feeding milestones, and complications were compared. To assess the effect of surgical experience, the laparoscopic cohort was divided into early and late phases.

Results

The median operative time was longer in the laparoscopic group than in the open group (124 minutes vs. 80 minutes, P < 0.05). However, laparoscopic patients achieved earlier oral intake (5 days vs. 9 days, P = 0.021) and full enteral feeding (8 days vs. 13 days, P = 0.028). No significant differences were observed in complication rates. The operative time improved significantly throughout the laparoscopic series, decreasing from 155 minutes in early cases to 110 minutes in later cases (P = 0.039). No conversions or anastomotic leaks were observed in the late laparoscopic group.

Conclusion

Laparoscopic duodenal atresia repair performed by a single surgeon is safe and effective, allowing for faster postoperative recovery than open repair. With accumulated experience, operative efficiency improves significantly, supporting its broader adoption in neonatal surgery.

INTRODUCTION

Despite its rarity, duodenal atresia (DA) holds significant clinical importance. DA is reported to affect 1 infant per 10,000 births and represents up to 60% of small intestinal atresias [1,2]. A “double-bubble sign” strongly suggests DA, and approximately 73% of cases, particularly complete duodenal obstruction, can be diagnosed prenatally. However, prenatal identification of incomplete obstruction can be challenging [3]. Open duodenoduodenostomy (often in a diamond-shaped configuration) has long been the standard surgical treatment of DA, which historically yields excellent outcomes [3,4,5].

Minimally invasive surgery has increasingly been applied to neonates, and in 2001, Bax et al. [6] reported the first laparoscopic repair of DA, followed by other early adopters in 2002 [7]. Laparoscopic duodenoduodenostomy is technically challenging in neonates, with a steep learning curve, owing to the small working space and delicate tissue. Nevertheless, DA presents a unique scenario where the distal bowel is decompressed (unlike other neonatal obstructions), providing a relatively adequate intraperitoneal workspace despite the patient’s tiny size [8]. Early series and comparative studies have suggested that a laparoscopic approach to the treatment of DA can result in faster postoperative recovery, including earlier return to feeding and shorter hospital stay, while maintaining similar overall complication rates despite the longer operative time [9]. However, laparoscopic DA repair is still regarded as one of the most technically demanding pediatric procedures, and its outcomes are highly operator-dependent [5].

This study presents a single surgeon’s 10-year experience with laparoscopic DA repair, examining surgical outcomes and evolution of the results over time. This retrospective study aimed to evaluate whether the promised benefits of minimally invasive repair were realized in practice and how surgical proficiency affected patient outcomes over the past decade.

METHODS

Ethics statement

The study was approved by the Institutional Review Board of Keimyung University Dongsan Medical Center (No. 2024-05-025-001). This study was conducted in accordance with the Declaration of Helsinki, and the need for written informed consent was waived owing to the retrospective nature of the study.

Study design and setting

A retrospective cohort study was conducted at a single tertiary pediatric hospital. Data of patients who underwent surgical repair of DA between 2010 and 2024 were reviewed. The laparoscopic cohort consisted of all consecutive patients with DA laparoscopically repaired by a single pediatric surgeon between June 2013 and October 2024. The open surgery cohort included patients who underwent conventional open DA repair by an experienced pediatric surgeon between 2010 and 2014 (before the adoption of the laparoscopic approach at our center). This open group was included for the historical comparison of outcomes.

Patients and outcome measures

Infants with congenital DA (intrinsic duodenal obstruction) who received operative treatment (either laparoscopic or open duodenoduodenostomy) were included. Diagnosis of DA was confirmed by neonatal occlusion with a “double-bubble sign” on plain X-ray imaging of the abdomen or upper gastrointestinal findings in the case of duodenal web or stenosis, or by laparoscopic exploration in the case of a double-bubble sign on prenatal ultrasonography. Patients with life-threatening conditions, such as severe cardiac anomalies, which could affect postoperative outcomes, were excluded. The data collected included patient characteristics, such as age, sex, and weight, associated anomalies, operative procedures performed, operative time, postoperative outcomes, and complications. Operative times were defined as the operative duration dedicated to the repair of DA (excluding the time for other concomitant procedures).

Postoperative outcomes were assessed, which included the time to start oral intake, time to reach full enteral feeds, and length of hospital stay. Complications included anastomotic leak, wound complications, and any significant morbidity before discharge. All patients were followed as outpatients, and the duration of postoperative follow-up was recorded (in months). All laparoscopic DA repairs were compared with open repairs before laparoscopy. All open surgery cases were performed by experienced pediatric surgeons between April 2010 and September 2014. The open approach was performed via the traditional transverse abdominal incision.

Laparoscopic technique

General anesthesia is used. The patient is placed in the supine position at the lower end of the operating table, and the surgeon can stand at the end of the table. The camera assistant can be on the surgeon’s right, with the scrub nurse on the other side. The main monitor is placed just over the patient’s head (Fig. 1). The procedure is usually performed using 3 ports. The camera port is placed through an umbilical incision. The abdomen can be inflated using the Hassan technique. A 5-mm cannula is inserted, and a short (20-cm) 5-mm, 30° telescope is used, giving an excellent wide-angle view of the right upper abdomen. A 3-mm port is inserted in the patient’s left mid-abdomen, just above the umbilicus, for the surgeon’s right hand, and a 3-mm port is made in the patient’s right mid-abdomen at the same level as the left port (Fig. 2). The inflating pressure is set to 8–10 mmHg and the flow rate to 2 L/min. Although this range may be relatively high for neonates with a median birth weight of approximately 2 to 3 kg, lower pressures (e.g., 5–7 mmHg) were trialed early in the series but were found to provide suboptimal visualization and working space. Therefore, a standard setting of 8–10 mmHg was adopted throughout the series.

Fig. 1. Team and table setup for laparoscopic duodenal atresia repair. The surgeon stands at the foot of the table, with the assistant to the right, scrub nurse to the left, and the anesthesiologist at the head. The monitor is positioned above the patient’s head.

Fig. 2. Port placement for neonatal laparoscopic surgery. A 5-mm umbilical port is used for the optical scope, and two 3-mm working ports are placed in the right and left upper quadrants. Additional traction sutures are placed through the abdominal wall as needed for optimal exposure.

The dilated proximal duodenum is easily visualized and mobilized from the transverse colon and gallbladder. The 3-mm hook cautery can be used for this step. Then, the smaller distal duodenal segment is identified and sufficiently kocherized to allow a tension-free anastomosis. If the liver obstructs the operative field, the abdominal wall can be sutured to the falciform ligament and gallbladder for better visualization. However, this does not need to be a routine procedure if the surgical field is adequate. A transabdominal stitch for the traction of the duodenal bulb or pylorus is beneficial in this procedure. Frequently, malrotation occurs, and a formal Ladd procedure is performed before creating the duodenal anastomosis. Once the anatomy is clearly defined, enterotomies are performed. A transverse duodenotomy is performed in the proximal, dilated duodenum, and a longitudinal duodenotomy is performed in the distal segment along the antimesenteric side to avoid the ampulla of Vater. A diamond-shaped anastomosis can be performed carefully without injuring the ampulla of Vater. The posterior wall can then be interrupted with 4-0 or 5-0 glyconate monofilament (Monosyn, B. Braun) sutures. The anterior wall is then sutured with a second suture. Anastomotic patency can be checked by insufflating 50 mL of air into the stomach through the previous orogastric tube.

Learning curve segmentation

To evaluate trends over time and the effect of surgical experience on the laparoscopic approach, the 10-year laparoscopic cohort (2013–2024) was divided into 2 chronological phases: an early period (initial 5 years, 2013–2017) and a late period (most recent 6 years, 2018–2024). The early phase included the first 7 laparoscopic cases, and the late phase included the subsequent 13 cases. To assess improvements or changes coincident with the surgeon’s increased experience, patient characteristics and outcomes were compared between these 2 subgroups. In addition, the operative time by case order was plotted to visualize the learning curve (Fig. 3). Violin plots were prepared to compare the time of laparoscopic DA repair between early and late laparoscopic phases (Fig. 4).

Fig. 3. Operative time for 20 consecutive laparoscopic duodenal atresia repairs. A segmented line plot shows reduced operative time in the late phase (cases 8–20), indicating progression along the learning curve.

Fig. 4. Comparison of operative time between early and late phases using a violin plot. Operative time decreased and became more consistent in the late phase, indicating progression along the learning curve.

Statistical analysis

Data analysis was performed using descriptive statistics and appropriate comparative tests. Continuous variables were assessed for normality; if normally distributed, they were compared with an independent-samples t-test, and if not, a nonparametric test (Mann-Whitney U-test) was used. Given the small sample sizes and nonnormal distributions of many measures, median values with ranges were reported, and the Mann-Whitney U-test was employed for the analysis of most continuous outcomes. Categorical variables (such as sex, prematurity, and presence of complications) were compared between the groups using the Fisher exact test. A P-value <0.05 was considered significant. Statistical analyses were performed using IBM SPSS Statistics ver. 20.0 (IBM Corp.). All tests were two-tailed.

RESULTS

Patient characteristics

Overall, 28 infants with DA underwent surgical repair during the study period. Of these, 20 patients were treated with laparoscopic DA repair (minimally invasive duodenoduodenostomy), and 8 patients underwent an open repair via laparotomy. The median gestational age at birth was 36 weeks in both groups (overall range, 29–40 weeks), with approximately 57% of the infants born premature (<37 weeks); however, no significant difference between laparoscopic and open groups was found (P > 0.999). Birth weight ranged from just <1.0 kg to a little >3.1 kg; the median birth weight was approximately 2.3 kg in both cohorts (overall mean, 2.38 ± 0.47 kg). Nine male infants (32.1%) and 19 female infants (67.9%) were included in the study. A prenatal ultrasound diagnosis of duodenal obstruction (polyhydramnios with double-bubble sign) had been made in 64.3% of cases. The laparoscopic group had a higher proportion of prenatal diagnosis (14 of 20, 70.0%) than the open group (4 of 8, 50.0%); however, this difference was not significant (P = 0.400).

Comparison of surgical outcomes: laparoscopy vs. open surgery

Table 1 summarizes the key intraoperative and postoperative outcomes for the laparoscopic and open groups. No significant difference was found in characteristics, such as sex, age, weight, and associated anomalies, between the 2 groups. Similarly, no difference was found in the type, obstruction level, and operative procedure. The median operative time was significantly longer in the laparoscopic group than in the laparotomy group. The laparoscopic group showed better postoperative outcomes in the median time to initiate oral feeding, full feeding, and hospital stay. No significant difference in postoperative complications was observed between the 2 groups. One patient in the laparoscopic group underwent anastomosis revision on postoperative day 3. In the laparotomy group, 1 patient had a catheter-related bloodstream infection, and the other had a wound infection.

Table 1. Comparative study on laparoscopic and open surgery for congenital duodenal atresia.

Values are presented as number only, number (%), or median (interquartile range).

GA, gestational age; POD, postoperative day.

^a)Fisher exact test. ^b)Mann-Whitney U-test. ^c)In cases accompanied by intestinal malrotation, a Ladd’s procedure was performed concurrently during surgery.

Learning curve: early vs. late laparoscopic cohorts

Over the 10-year period, a clear increase in the annual volume of laparoscopic DA repairs was observed, reflecting the increasing experience. The first phase (2013–2017) included 7 laparoscopic cases, and the second phase (2018–2024) included 13 cases. The patient profiles in these 2 subgroups were generally comparable, with no significant differences in baseline characteristics. The median gestational age (37 weeks vs. 36 weeks) and birth weight (2.27 vs. 2.44 kg) did not differ significantly between the early and late laparoscopic periods, nor did the rates of prematurity (42.9% vs. 61.5%, P = 0.642) or prenatally diagnosed cases (57.1% vs. 76.9%, P = 0.613). The sex distribution was comparable (female patients comprised 57.1% of early cases vs. 61.5% of late cases) (Table 2). The operative performance improved markedly over time. Fig. 3 illustrates the operative time plotted by case order for the laparoscopic series, demonstrating a downward trend as experience accrued. In the first 5 years. The median operative time decreased from 155 minutes in early cases to 110 minutes in late cases (P = 0.039). In the latter half of the series, the laparoscopic operative times were much closer to those expected for open surgery. Consistent with this, the single conversion to open operation occurred in the early series; no conversions occurred in any of the last 13 laparoscopic cases. Postoperative recovery in the early vs. late laparoscopic subgroups indicated that the minimally invasive approach was beneficial from the beginning and remained consistent. Feeding milestones were achieved equally quickly in both periods: in both early and late laparoscopic cases, the first oral feeds were typically initiated around postoperative day 5 (median, 5 days in each group), and full feeds were reached by approximately 1 week (median, 8 days in both early and late groups). This indicates that the surgeon achieved the enhanced feeding recovery that laparoscopy can provide, even during the initial learning phase, and this outcome was maintained in later cases. Similarly, hospital stays did not significantly differ between the early and late laparoscopic groups. Early-phase cases had a median postoperative length of stay of 13 days (range, 8–19 days), compared with the 17 days (range, 7–115 days) in the late phase, a difference that was not significant (P = 0.473). A violin plot comparing operative time between early and late phases of laparoscopic DA repair (Fig. 4) shows a significant reduction in operative time in the late group, with a narrower distribution and lower median compared to the early group, reflecting procedural proficiency over time. No significant changes in complication rates were observed over time: the early laparoscopic group had 1 case (14.3%) of early postoperative complications.

Table 2. Comparison of the early and late phases of laparoscopic surgery.

lues are presented as number only, number (%), or median (interquartile range).

GA, gestational age; POD, postoperative day.

^a)Fisher exact test. ^b)Mann-Whitney U-test.

DISCUSSION

The decade-long, single-surgeon experience indicates that laparoscopic DA repair can be successfully implemented with outcomes that are at least equivalent to, and in some aspects better than, the traditional open approach. Despite initial technical hurdles, the minimally invasive repair proved its value in postoperative recovery. In our series, patients who underwent laparoscopy achieved full enteral feeding significantly faster than those who underwent open surgery, a finding that is consistent with the findings of previous studies [9,10]. The early return of bowel function and feeding is a well-recognized benefit of laparoscopy in neonatal abdominal surgery, likely attributable to the reduced handling of the bowel and minimized surgical trauma, leading to a shorter postoperative ileus [11]. In the laparoscopic cohort, the median time to full feeding was 5 days earlier than that in the open cohort, which aligns with the magnitude of difference reported in the literature [11,12]. Early feeding translates to improved nutritional recovery and may facilitate earlier hospital discharge. It is possible that the delayed initiation of oral feeding in the open group was due to more conservative postoperative feeding protocols rather than intrinsic delayed functional recovery. This possibility should be considered when interpreting the observed advantage in feeding milestones among the laparoscopic cohort.

Although this study did not show a significant difference in the length of hospital stay between the laparoscopic and open groups (possibly due to the limited sample size), the laparoscopic group tended to have shorter hospital stays. Many prior series and recent meta-analysis have demonstrated significantly shorter hospital stays with minimally invasive repair of DA [9,13]. In a meta-analysis of 304 laparoscopic vs. 1,044 open DA repairs, Zhang et al. [9] reported that laparoscopy was associated with a meaningfully shorter postoperative hospitalization on average. In the present cohort, the median hospital stay was approximately 5 days less with laparoscopy, and this difference would reach significance with a larger sample. An important point is that improved feeding milestones likely contributed to the reduced length of stay. Once full feed is achieved and weight gain is established, infants are generally discharge-ready, barring other medical issues. No adverse trade-off was observed in safety or morbidity with the laparoscopic approach. No anastomotic leaks were observed in the minimally invasive group, and the overall complication rate was low and comparable with that of the open group. This echoes the findings of prior comparisons that anastomotic complications (leak or stricture) or other major complications did not increase with laparoscopic repair [9,12]. In a recent study, laparoscopy may reduce certain complications, such as wound infections, which were noted to be lower in the laparoscopic cohorts [13]. In our series, only 1 wound infection occurred (in an open case), highlighting the potential benefit of smaller incisions in neonates. The single anastomotic stricture in the laparoscopy group occurred early in our experience and was managed successfully. Such occurrences have also been reported in other series; however, with refinements in technique (e.g., meticulous suturing and ensuring a tension-free anastomosis), they have become quite rare [14].

The longest operative time associated with laparoscopy was the most pronounced difference in intraoperative metrics, particularly in the initial cases. This is not unexpected—a steep learning curve exists for laparoscopic neonatal surgery, and DA repair is one of the more challenging procedures to perform when employing an endoscopic approach. Our data revealed that early laparoscopic cases required substantially longer operative times (median, >3 hours) than open surgery. This finding mirrors the common experience reported in the literature that laparoscopic DA repairs initially take longer operating times due to the delicate intracorporeal suturing and the surgeon’s learning process [10]. However, our series demonstrated that this disadvantage diminished with increased experience. By the latter half of the study period, the median laparoscopic operative duration had dropped to approximately 2.5 hours, approaching the operative times of the open technique. The improvement in the “real” anastomosis time by approximately 45 minutes between the early and late laparoscopic cohorts is a concrete measure of the surgeon’s increased efficiency. After approximately 5–10 cases, the operative workflow became smoother, and the setup and suturing times decreased. The learning curve observed (Fig. 3) indicates that familiarity with neonatal minimally invasive techniques and iterative refinements (such as optimal port placement, suturing method, and use of appropriate instruments) led to meaningful time savings. Other authors have noted that after an initial learning phase, operative times for laparoscopic DA can approach parity with open surgery [14]. Our experience supports this finding, indicating that the operative time penalty can be overcome with practice and should not be a permanent deterrent to adopting the laparoscopic approach.

This study has several limitations. The sample size is modest, particularly for the open surgery group (n = 8), which limits the statistical power to detect differences. The open and laparoscopic groups were not contemporaneous; the open cases were performed in earlier years (2010–2014) by a different surgeon, whereas the laparoscopic cases were more recent. This raises the possibility of temporal confounders. We attempted to mitigate this by focusing on objective surgical-related metrics and the fact that the open surgeon is highly experienced, whereas the laparoscopic surgeon was on a learning curve, which in some ways introduces biases against the laparoscopic group. Another limitation is the retrospective design, as the study relies on the accuracy of medical records for the timing of feeds and complications. However, given that feeding advancement and discharge are protocol-driven in our unit, these endpoints were well-documented. Moreover, pain scores or cosmesis were not formally measured; however, qualitatively, the laparoscopic group had minimal incision pain and virtually invisible scars, which is a subjective benefit for families. Additionally, as this study focused primarily on early postoperative outcomes, we did not investigate long-term biliary complications or hepatobiliary status. This is another inherent limitation of retrospective design. Future studies should consider routine follow-up imaging, such as abdominal ultrasonography, in this patient cohort to evaluate delayed hepatobiliary anomalies, which have been occasionally reported in association with DA. Finally, while a learning curve analysis was performed by splitting time periods, a more sophisticated learning curve analysis (e.g., cumulative sum or risk-adjusted models) was beyond the scope of this study but could further quantify the proficiency gain.

In conclusion, with over a decade of experience, laparoscopic DA repair proved to be a safe and effective technique, yielding clinical outcomes on par with or superior to the traditional open approach. Infants who underwent laparoscopic repair benefited from faster feeding initiation and a trend toward shorter hospital stays without an increase in complications. Operative times for the laparoscopic procedure were initially longer, reflecting the steep learning curve, but improved significantly with surgeon experience, approaching those of open surgery. This single-surgeon series demonstrates that with appropriate expertise, the minimally invasive approach can become the standard of care for DA, offering the advantages of quicker recovery and minimal incision morbidity. We advocate for the continued adoption of laparoscopic repair of neonatal DA by experienced pediatric surgeons, as it is an effective and infant-friendly technique that aligns with the goals of enhanced recovery in this fragile patient population. Ultimately, our decade of outcomes supports the idea that laparoscopic management of DA is feasible and beneficial, validating its role in modern pediatric surgery practice.