Surgical Endoscopy and Other Interventional Techniques

https://doi.org/10.1007/s00464-018-6130-3

Laparoscopic versus open surgery for the repair of congenital

duodenal obstructions in infants and children
Stefan Gfroerer1   · Till‑Martin Theilen1 · Henning Cornelius Fiegel1 · Udo Rolle1

Received: 24 July 2017 / Accepted: 23 February 2018

© Springer Science+Business Media, LLC, part of Springer Nature 2018

Abstract
Background  Laparoscopic repair of congenital duodenal obstruction (LCDO) was described more than 15 years ago. How-
ever, studies comparing outcomes of LCDO with open repair (OCDO) are rare. Standardized assessments of complications
using the Clavien–Dindo classification (CDC) and the comprehensive complication index (CCI) are not available.
Methods  All patients undergoing OCDO or LCDO between 2004 and 2017 were identified from the institutional database
by retrospective analysis. Postoperative outcomes were assessed, including all complications using the CDC and the CCI.
Results  Forty-seven consecutive patients were identified; 27 patients underwent LCDO and 20 patients had OCDO. Both
groups did not differ regarding demographics, associated congenital anomalies, intraoperative pathologic findings, and opera-
tive procedures. LCDO was associated with a longer operative time [mean (SD), 202 (89) vs. 112 (41) min, P < 0.0001],
shorter time to initiation of feeds [median (range), 1 (0–4) vs. 3 (1–12) days, P = 0.0027], and shorter time to full feeds [mean
(SD), 8.2 (4.1) vs. 12.2 (6.4) days, P = 0.0243] compared to OCDO. Shorter length of postoperative hospital stay in LCDO
group was achieved for patients without cardiac anomalies [mean (SD), 9.4 (3.1) days in LCDO group vs. 17.2 (9.4) days
in OCDO, P = 0.0396] and patients without other anomalies [median (range), 12 (3–38) days in LCDO group vs. 21 (7–31)
days in OCDO, P = 0.0460]. LCDO was associated with a lower CCI [median (range) 0 (0–39.7) vs. 4.3 (0–100), P = 0.0270].
Conclusions  Despite a longer operative time for LCDO, a number of advantages of LCDO over OCDO were recognized
comparing both approaches in the repair of congenital duodenal obstruction. Such advantages include a lower morbidity,
reduced time to initiation and completion of full enteral feeds, and shorter length of postoperative hospitalization for patients
without concomitant cardiac anomalies and for patients without other anomalies when operated laparoscopic. In view of the
present results, LCDO, performed in selected patients, appears to represent a viable alternative to OCDO.

Keywords  Congenital duodenal obstruction · Laparoscopic duodenoduodenostomy · Laparoscopic versus open ·

Morbidity · Clavien–Dindo classification · Comprehensive complication index

Congenital duodenal obstruction (CDO) is characterized
by a partial or complete obstruction of the duodenum. The
standard open surgical repair of congenital duodenal obstruc-
tion (OCDO) is performed by a right upper quadrant lapa-
rotomy. Since the first description of laparoscopic Ladd’s
procedure [1] and laparoscopic duodenoduodenostomy
[2] various studies of retrospective case series have shown
that laparoscopic repair of congenital duodenal obstruction
(LCDO) is feasible and efficacious [3–8]. However, to date,
* Stefan Gfroerer
stefan.gfroerer@kgu.de
1
Department of Pediatric Surgery and Pediatric Urology,
University Hospital Frankfurt, Theodor‑Stern‑Kai 7,
60590 Frankfurt am Main, Germany
there are only limited data available, supporting evidence
for either the open or laparoscopic procedure to be superior.
The only systematic review and meta-analysis comparing
laparoscopic and open approach for the correction of CDO
are based on four retrospective case series comprising 180
patients [9]. In each of the included studies, complications
were mentioned, but standardized methodology of reporting
morbidity was not applied [10–13].
A widely used classification of surgical complications is
the Clavien–Dindo classification [14]. The Clavien–Dindo
classification focuses on the therapeutic consequences of the
single most severe complication occurring in a patient in a
given episode. In comparison, the newly proposed compre-
hensive complication index (CCI) combines all complica-
tions after surgery in a single score, and thus reflects the

13
Vol.:(0123456789)

overall morbidity [15]. The experienced working group
around Clavien currently recommends the simultaneous use
of both scoring systems for clinical trials [15].
In our retrospective analysis, we aimed to analyze the
postoperative outcomes of patients with CDO following
LCDO and OCDO including an analysis of postoperative
complications using the Clavien–Dindo classification and
the CCI.
Methods
We performed a retrospective analysis of a series of 47 con-
secutive patients with CDO, as identified through Interna-
tional Classification of Diseases (ICD-10, Version 2017)
codes, that underwent either LCDO or OCDO between
January 2004 and January 2017 at our institution. The study
protocols were approved by the local institutional review
committee (number 85/17). Data were retained in a database,
including demographics, pathologic findings, operative vari-
ables, postoperative outcomes, and a detailed complication
profile. Demographics included age at operation, weight at
operation, gestational age at birth, gender, prenatally sus-
pected diagnosis, and associated congenital anomalies.
Pathologic findings were assigned to either complete or
incomplete duodenal obstruction and additional pathologies.
Pathologies of complete duodenal obstruction were allocated
to atresia type 1 (membrane), atresia type 2 (fibrous cord),
atresia type 3 (duodenal gap), and annular pancreas. Find-
ings of incomplete duodenal obstruction were specified by
diagnoses of web, Ladd’s bands and annular pancreas. Oper-
ative details included operative time (abdominal incision to
close of the abdominal wound; not including time for correc-
tion of extra abdominal pathologies), operative procedure,
and additional operative procedures. Postoperative outcome
variables included time from operation to initiation of feeds,
time from operation to full feeds, length of postoperative
hospital stay (not including the day of operation, but includ-
ing the day of discharge). All complications were recorded
individually in detail and simultaneously scored according
to the Clavien–Dindo classification [14]. This classifica-
tion consists of 7 grades (I, II, IIIa, IIIb, IVa, IVb, and V).
Morbidity minor included complications of grades I and II,
while morbidity major included grades III-V. The compre-
hensive complication index [15], as a sensitive measure of
postoperative complications and their severity, ranges from
0 (no complication) to 100 (death). The CCI was calculated
using the CCI calculator available online (http://www.asses​
surge​ry.com).
The diagnosis of CDO was based on preoperative plain
abdominal X-ray and/or upper gastrointestinal contrast
study and/or gastroduodenoscopy. Computed tomography
13
Surgical Endoscopy
or magnetic resonance imaging was not applied in the diag-
nostic process.
Operative techniques
Operative approach was determined by the attending senior
surgeon. LCDO was performed by the first author in patients
weighing > 1700 g at operation. Laparoscopic approach was
implemented irrespective of preoperative radiologic findings
or associated congenital anomalies. OCDO was performed
via a transverse right upper quadrant incision. LCDO was
performed using an umbilically placed 5 mm 30° camera
and two 3.5-mm working trocars, one placed in the upper
left quadrant, the second in the right lower quadrant. Liver
elevation was achieved through a transabdominal suspension
suture around the Ligamentum teres hepatis or alternatively
by a 3-mm grasper, introduced via an infraxiphoidal stab
incision, directed under the liver, lifted and clamped to the
latero-parietal peritoneum. Subtle preparation of the duode-
num was performed; malrotation corrected as needed, and
Ladd’s bands resected as found. In LCDO, predominantly
diamond-shaped anastomoses were performed; recently also
simple oblique [7] and parallel anastomoses [8] were used
depending on the individual duodenal anatomic situation.
Duodenoduodenostomy in OCDO was performed using clas-
sic diamond-shaped anastomosis [16]. Intraoperative stay
sutures were not applied on a regular basis in OCDO and
LCDO, but were occasionally used when helpful for ana-
tomical reasons. Tapering duodenoplasty was not performed
in either group. Cases that were started laparoscopically and
needed conversion to open procedures were included in the
laparoscopic group. All LCDO were performed by the first
author. All OCDO were performed by, or under the direct
supervision of the first, third, or last author.
Postoperative care and follow‑up
Nasogastric tube was left in situ postoperatively for gastric
decompression. All patients with a duodenal anastomosis
were left nil by mouth on the day of operation and were
allowed to receive enteral feeding on postoperative day 1.
Patients without duodenal anastomosis were allowed to
receive enteral feeding on the day of operation. Day of ini-
tiation of feeds was the day, on which feeding per orally or
via nasogastric tube was initiated. After commencement of
enteral feeding, all patients received nutritional increments,
based on clinical observations, irrespective of the operative
approach. Some greenish gastric fluid postoperatively dur-
ing preprandial routine aspirations of the nasogastric tube
in an otherwise unremarkable clinical course was regarded
as a normal finding, due to the insufficient closure of the
pyloric muscle, and feeding increments were continued. Day
of achievement of full feeds was the day, on which parenteral
Surgical Endoscopy

nutrition was ceased. Upper gastrointestinal (UGI) contrast Table 2  Pathologic findings of 47 patients undergoing laparoscopic
studies were not performed on a routine basis postopera- repair of congenital duodenal obstruction (LCDO) or open repair of
congenital duodenal obstruction (OCDO)
tively. Patients were scheduled for out-patient follow-up
within 4 weeks after discharge. All patients were scheduled LCDO OCDO P
for regular out-patient follow-up, at least once per year. N = 27 N = 20

Complete duodenal obstruction 11 (41%) 14 (70%) 0.0759

Statistical analysis
 Atresia type 1—membrane 2 (7%) 7 (35%)
  Type 2—fibrous cord 0 1 (5%)
Continuous data with a normal distribution are reported as
  Type 3—gap 2 (7%) 1 (5%)
mean (standard deviation, SD) values and compared using
 Annular pancreas 7 (26%) 5 (25%)
the two-sample t test. Continuous data that were not nor-
  additionally to type 3 atresia 1 (4%) 1 (5%)
mally distributed are reported as median (range) values and
Incomplete duodenal obstruction 16 (59%) 6 (30%) resp
compared using the robust Wilcoxon rank sum test. For
 Web 6 (22%) 2 (10%)
categorical variables, comparisons between groups were
 Ladd’s bands 7 (26%) 4 (20%)
performed using the Fisher’s exact test. Median duration of
 Annular pancreas 3 (11%) 0
follow-up was calculated based on survival analysis (reverse
Additional pathologies 23 (85%) 12 (60%)
Kaplan–Meier estimate). Statistical significance was taken
 Second distal stenosis 1 (4%) 0
for two-sided tests at the 5% level. The computer software
 Malrotation 22 (81%) 12 (60%)
R was used for all statistical calculations.
Data are frequency (%)

Results
A total of 47 consecutive patients underwent surgical repair
of CDO; 27 received LCDO and 20 underwent OCDO. A
comparison of demographic data of both groups is shown in
Table 1. Demographic variables and associated congenital
anomalies were similar in both groups.
Table  2 compares intraoperative pathologic findings
in the 2 groups. The proportion of complete/incomplete
duodenal obstructions in each group did not differ (41/59%
in the LCDO vs. 70/30% in the OCDO group, P = 0.0759).
Table 3 compares LCDO and OCDO groups in terms of
operative characteristics. There were no differences with
regard to the applied operative procedures in both groups.
Predominantly a duodenoduodenostomy was performed in
63% of LCDO and 65% of OCDO (P = 1). The operative
time was longer in the LCDO group [mean (SD), 202 (89)

Table 1  Demographic data for 47 patients undergoing laparoscopic repair of congenital duodenal obstruction (LCDO) or open repair of congeni-
tal duodenal obstruction (OCDO)
LCDO OCDO P
N = 27 N = 20

Gender (male:female) 12 (54.5%):15 (60%) 10 (45.5%):10 (40%) 0.7730

Prenatally suspected diagnosis (yes:no) 10 (45.5%):17 (68%) 12 (54.5%):8 (32%) 0.1480
Gestational age at birth (weeks) 38.3 (33.9–40.2) 36.0 (30.1–42.0) 0.0958
Age at operation (days) 9.0 (1–2790) 2.5 (0–190) 0.0824
Weight at operation (kg) 3.01 (1.78–13.89) 2.76 (1.48–6.34) 0.3067
Associated congenital anomalies (patients) 18 (67%) 12 (60%) 0.7610
 Congenital heart disease (patients) 17 (63%) 11 (55%) 0.5500
 Trisomy 21 7 (26%) 7 (35%) 0.5350
 Other anomalies (patients) 10 (37%) 6 (30%) 0.7580
  Details Cornelia de Lange syndrome (1), butterfly Esophageal atresia (1), hypothyreosis (2),
vertebrae (1), hydronephrosis (2), hypo- pes calcaneus (1), hemolytic disease of the
thyreosis (4), funnel trachea (1), ectrodac- newborn (1), biliary duct hypoplasia (1),
tyly (1), polydactyly (1), hypospadia (1), celiac disease (1), atopic eczema (1)
Hirschsprung disease (1), glutaric aciduria
type 1 (1), patent omphalomesenteric duct
(1), sleep apnoea (1)

Data are median with range () or frequency (%)

13
 Surgical Endoscopy

Table 3  Operative variables LCDO OCDO P

and surgical procedures
for 47 patients undergoing N = 27 N = 20
laparoscopic repair of
Operative ­timea (min) 202 (89) 112 (41) < 0.0001
congenital duodenal
obstruction (LCDO) or open Procedure
repair of congenital duodenal  Duodenoduodenostomy 17 (63%) 13 (65%) 1
obstructions (OCDO)  Excision of web and duodenoplasty (Mikulicz) 3 (11%) 3 (15%) 1
 Duodenal liberation from obstructing bands (Ladd) 7 (25%) 4 (20%) 0.6351
Additional procedures
 Jejunoplasty (Mikulicz) for second distal stenosis 1 (4%) 0
 Ladd’s procedure 16 (59%) 7 (35%) 0.1425
 Appendectomy 9 (33%) 8 (40%) 0.7610
 Resection of Meckel’s diverticulum 1 (4%) 3 (15%)
 Transanastomotic nasojejunal tube 3 (11%) 7 (35%) 0.0730
 Intraoperative X-ray 2 (7%) 1 (5%) 1
Conversion to open approach 3 (11%)
Suture size X/0 5 (3–5) 5 (5–6)
CO2 ­flowb (l/min) 1 (0.5–3)
PCO2b (Hg) 8 (8–10)

­ eana or m
Data are m ­ edianb with standard deviation or range () respectively or frequency (%); bold numbers
represent P values < 0.05

vs. 112 (41) min respectively, P < 0.0001]. There were 3/27
(11%) conversions to laparotomy in the LCDO group.
Table 4 compares postoperative outcomes of patients
after LCDO and OCDO. Time from operation to initiation
of feeds was shorter in the LCDO group compared with
OCDO group [median (range), 1 (0–4) vs. 3 (1–12) days,
P = 0.0027]. Time from operation to full enteral feeds was
shorter in the LCDO group [mean (SD), 8.2 (4.1) days vs.
12.2 (6.4) days in the OCDO group, P = 0.0243]. Length
of postoperative hospital stay was similar following LCDO
and OCDO [median (range), 14 (3–38) vs. 21 (7–40) days,
P = 0.1179]; however, length of postoperative hospital stay

Table 4  Postoperative outcomes for 47 patients undergoing laparoscopic repair of congenital duodenal obstruction (LCDO) or open repair of
congenital duodenal obstruction (OCDO)
LCDO OCDO P
N = 27 N = 19

Time from OP to initiation of f­ eedsb (days) 1 (0–4) 3 (1–12) 0.0027

Time from OP to full ­feedsa (days) 8.2 (4.1) 12.2 (6.4) 0.0243
Length of postop. hospital ­stayb (days) 14 (3–38) 21 (7–40) 0.1179
 No cardiac a­ nomalya (days) 9.4 (3.1)—N = 10 17.2 (9.4)—N = 9 0.0396
 No ­trisomyb (days) 9.5 (3–38)—N = 20 16 (7–34)—N = 13 0.0621
 No other a­ nomaliesb (days) 12 (3–38)—N = 17 21 (7–31)—N = 13 0.0460
N = 27 N = 20

Morbidity (Clavien–Dindo grade I–V) 6 (22%) 10 (50%) 0.0650

 Surgical morbidity 2 (7%) 6 (30%) 0.0574
  Anastomotic 0 1 (5%)
  Nonanastomotic 2 (7%) 5 (25%)
 Nonsurgical morbidity 4 (15%) 7 (35%) 0.1645
Morbidity minor (Clavien–Dindo grade I–II) 5 (19%) 7 (35%) 0.3111
Morbidity major (Clavien–Dindo grade III–V) 2 (7%) 5 (25%) 0.1185
Overall reoperation rate 2 (7%) 5 (25%) 0.1185
Comprehensive complication index 0 (0–39.7) 4.3 (0–100) 0.0270

Data are ­meana or m

­ edianb with standard deviation and range () respectively or frequency (%); bold numbers represent P values < 0.05

13
Surgical Endoscopy

was shorter following LCDO for patients without cardiac Table 5  Individual profile of postoperative complications, graded
anomalies [mean (SD), 9.4 (3.1) vs. 17.2 (9.4) in the OCDO according to Clavien–Dindo classification and with calculated com-
prehensive complication index (CCI), for patients following laparo-
group, P = 0.0396] and for patients without other anoma- scopic repair of congenital duodenal obstruction (LCDO) and open
lies [median (range), 12 (3–38) vs. 21 (7–31) in the OCDO repair of congenital duodenal obstruction (OCDO)
group, P = 0.0460]. The hospital stay for patients without
Postoperative complications Clavien– CCI
trisomy did not differ significantly between the groups Dindo
[median (range) 9.5 (3–38) for LCDO vs. 16 (7–34) days grade
in the OCDO group, P = 0.0621]. There were no differences
LCDO
in morbidity between the laparoscopic and open approach
 Patient 16 Transient trocar hernia I 8.7
according to Clavien–Dindo classification; however, assess-
 Patient 25 Colon perforation IIIb 33.7
ing morbidity by means of the CCI, patients undergoing
 Patient 27 Pneumonia II 29.6
LCDO had a lower index [median (range), LCDO 0 (0–39.7)
Pericardial effusion II
vs. OCDO 4.3 (0–100), P = 0.027].
 Patient 37 Subclavian malpuncture, transfusion II 39.7
Table 5 shows the individual postoperative complica-
Hematothorax, pleural drainage IIIb
tion profile following LCDO and OCDO. 2 out of 6 (33%)
 Patient 42 Enterocolitis II 20.9
patients had 2 or more postoperative complications in the
 Patient 45 Postoperative vomiting (> 7 days) I 8.7
LCDO group compared to 5/10 (50%) in the OCDO group.
OCDO
Figure 1 displays absolute frequencies of LCDO and
 Patient 1 Colon perforation, segmental resec- IIIb 47.7
OCDO performed between January 2004 and January 2017. tion
Decrease of operative time for LCDO, started in 2010, is Colostomy closure IIIb
visualized. The mean (SD) operative time of the years  Patient 2 Incisional hernia IIIb 33.7
2010/2011 was 272 (102); over a 5-year period, operative  Patient 3 Central line infection II 20.9
time decreased by 104 min (− 38%) to mean (SD) 168 (66)  Patient 5 Surgical site infection II 22.6
min in 2016/2017 (P = 0.057). Icterus prolongatus I
For the OCDO group, the intervals January 2004–Decem-  Patient 6 Enteritis, dehydration II 29.6
ber 2009 (N = 11) and January 2010–January 2017 (N = 8) Central line infection II
were compared with regard to postoperative outcomes.  Patient 11 Hyperbilirubinemia I 8.7
There were no differences regarding time from operation to  Patient 22 Gastroesophageal reflux II 39.7
initiation of feeds [median (range), 2 (1–12), 2004–2009 vs. Missed Ladd’s bands IIIb
3 (2–6), 2010–2017, P = 0.8645], time from operation to full  Patient 31 Gastroesophageal reflux II 44.9
feeds [median (range), 12 (3–22), 2004–2009 vs. 10 (5–20), Pneumonia II
2010–2017, P = 0.9340] and length of postoperative hospital Missed Ladd’s bands IIIb
stay [median (range), 25 (7–40), 2004–2009 vs. 18,5 (7–34),  Patient 38 Central line infection II 20.9
2010–2017, P = 0.8363].  Patient 39 Gastric bleeding IIIa 100
Median (range) duration of follow-up was 44.9 months Anastomotic leakage IIIb
(0.39–155.01); 34.8 months (0.39–73.7) for the LCDO Cardiac failure, multiorgan dysfunc- IVb
group and 102.4 months (11.3–155.0) for the OCDO group tion
(P = 0.0006). Death V

Discussion comparison with patients undergoing laparotomy. Patients

This study comprises the largest group of laparoscopic
repairs among the few studies comparing open and lapa-
roscopic approaches in CDO surgery (Table  6). To our
knowledge, this is the first study comparing laparoscopic
and open surgical repair of CDO including an assessment
of postoperative complications, using the Clavien–Dindo
classification and the comprehensive complication index.
In this study, laparoscopically operated patients had a lower
CCI compared to patients undergoing laparotomy. Laparo-
scopically operated patients tolerated an earlier commence-
ment and showed an earlier completion of enteral feeding in
without associated cardiac anomalies as well as patients
without other anomalies were discharged home earlier after
laparoscopic procedures for CDO correction.
In order to provide a basis for a comprehensive evalua-
tion of differences between operative techniques, it is crucial
to reveal all emerging complications [17]. A widely used
tool for postoperative morbidity assessment is the Cla-
vien–Dindo classification [14]. This simple and reproduc-
ible classification implements grades I to V to focus on the
individual single most severe postoperative complication.
This complication assessment has recently been augmented
by the CCI, a highly discriminatory index, which includes

13

Fig. 1  Columns represent absolute frequencies of open repair of con-
genital duodenal obstruction (OCDO) and laparoscopic repair of con-
genital duodenal obstruction (LCDO) between January 2004 and Jan-
all complications after a procedure, and thus, reflects the
multidimensional nature of postoperative morbidity [15].
Currently the latter authors recommend using both scoring
systems simultaneously for postoperative outcome assess-
ment. While the CDC in our study did not reveal a differ-
ence, the CCI showed a significant difference between both
groups. This underlines the importance to include the CCI
for postoperative outcome assessment.
Outcome assessment of patients with CDO is complex
due to a high prevalence of comorbidities, predominantly
congenital heart disease (CHD). Roughly 50% of patients
with CDO are diagnosed to have concomitant CHD [18, 19].
The detection rate of CHD in our study was 64%; this may
be attributed to the circumstance that 98% of our patients
received an echocardiography pre- or postoperatively. The
prevalence of CHD in our cohort is in line with findings
recently published by Short et al., who stated a prevalence of
70.1% in their retrospective series [20]. CHD has the poten-
tial to delay transition to full enteral feeding after opera-
tion [21]. In consequence, CHD may prolong the length of
hospital stay after operation and, hence, potentially mask
effects on hospital stays of different surgical procedures. In
our cohort, the subgroup of patients without CHD as well as
the group of patients without other anomalies left hospital
13
Surgical Endoscopy
uary 2017; dashed line shows decrease of operative time for LCDO
between 2010 and 2017
earlier after laparoscopic operation compared to open access.
Spilde et al. previously reported a shorter length of postoper-
ative hospitalization for patients after laparoscopy compared
with those that underwent open repair [10], while other
authors did not find a difference [11–13]. Subgroup analy-
sis was performed in neither of these comparative studies.
Consistent with the finding that laparoscopic operations
are associated with more rapid recovery of intestinal motility
and earlier resumption of normal diet [22], our results show
that LCDO allows for earlier initiation of enteral feeds and
earlier completion of full enteral feeds compared to OCDO.
Comparable results were published by Spilde et al. [10].
However, Spilde et al. suggested an alternative mechanism
for their findings. They stated that all 15 patients undergoing
laparoscopic repair received an upper gastrointestinal (UGI)
contrast study between postoperative days 2 and 11, before
initiating feeding. On the contrary only 4 of the 14 patients
in the open approach group were evaluated by UGI study.
The nasogastric tube was removed after successful contrast
study respectively and enteral feeding was commenced 12 h
later. This almost certainly contributed to the significant dif-
ferences in time to initial feeds, full enteral diet, and hospital
stay in the mentioned study [10]. Two other comparative
studies reported no differences between LCDO and OCDO
Surgical Endoscopy

Table 6  Summary of literature comparing laparoscopic (lap) versus open repair of congenital duodenal obstruction
Spilde et al. [10] Hill et al. [11] Jensen et al. [12] Parmentier et al. [13]

Study design Retrospective case Retrospective case series, Retrospective case series, Retrospective case series,
series, single center single center multicenter single center
Study period 2003–2007 2001–2010 2005–2011 2007–2014
Patients (N) 29 58 64 29
 Laparoscopic (N) 15 22 20 10
 Open (N) 14 36 44 19
Age at operation (days), lap 39.9 ± 28 4 (1–300) 4 (1–22) 2.5 (1–385)
Age at operation (days), 41.8 ± 25 3 (0–150) 6 (0–89) 1 (1–21)
open
Weight at operation (g), lap 3224 ± 1540 2500 (1600–8500) 2560 (1740–3440) Not reported
Weight at operation (g), 3332 ± 1690 2400 (1100–9600) 2570 (1730–4350) Not reported
open
Operative time (min), lap 127 116 (73–164) 145 (91–308) 90 (80–150)
Operative time (min), open 96 103 (71–220) 96 (54–174) Not reported
 P 0.06 < 0.05 0.001 Not reported
Time to initiation of feeds 5.4 7 (0–36) 10 (4–44) 8 (4–24)
(days), lap
Time to initiation of feeds 11.3 9 (0–23) 9 (4–33) 4 (3–13)
(days), open
 P 0.002 > 0.05 0.24 0.009
Time to full enteral feeds 9 Not reported 15 (6–72) 36 (6–70)
(days), lap
Time to full enteral feeds 16.9 Not reported 15 (9–126) 16.5 (8–200)
(days), open
 P 0.007 Not reported 0.69 0.14
Hospital stay (days), lap 12.9 20 (4–149) 20 (13–105) 45.5 (6–204)
Hospital stay (days), open 20.1 19 (6–79) 30 (9–173) 20.5 (13–225)
 P 0.01 > 0.05 0.27 0.09
Complications, lap Stenosis, dilatation Prolonged ileus (4) Late stricture, reoperation Stricture, reoperation (1)
Bacteremia (1) (1) Catheter associated infection
Died of sepsis 5 month Trocar site hernia (1) (3)
postop (1) Missed duodenal web (1) Bacteremia (1)
Complications, open Stenosis, reoperation (1) Leakage, reoperation (1) Incisional hernia (1) Adhesive intestinal obstruc-
Prolonged ileus (10) Surgical site infection (1) tion (1)
Bacteremia (1) Bacteremia (1) Died of GI tract bacterial
Cardiac arrest (1) translocation (1)

groups [11, 12] and one study presented data of earlier initia-
tion of feeds after open repair [13]. These latter authors did
not implement a postoperative UGI contrast study. In our
study, we did not expose operated patients towards a routine
UGI contrast study in either group. Patients of both groups
were started, as early as they clinically tolerated, on enteral
feeding, irrespective of the color of the aspirated gastric
fluid, provided their general condition was appropriate. As
other authors previously mentioned, congenital pyloric dis-
tension persists for some while in the postoperative course,
making bilious reflux into the stomach an expectable phe-
nomenon after corrective operation [10, 13].
The only patient that had suffered anastomotic leakage
among all operated patients had concomitant trisomy 21 and
tetralogy of Fallot. The initial anastomosis was performed
via open surgical access. The duodenal anastomosis was
revised twice by laparotomy and intraoperative assess-
ment of the anastomosis judged the arterial perfusion to be
extremely poor. It was estimated that the repetitive leakage
was a result of a lack of reconstitution due to poor perfusion
rather than a result of purely surgical reasons.
The patients’ third leakage was eventually left with an
abdominal drain and anastomotic leakage closed up spon-
taneously. This patient died 4 weeks postoperatively due to
severe cardiac failure.
There were 2 patients in the OCDO group, where Ladd’s
bands were missed in the initial open duodenoduodenostomy
(patient 22, patient 31; Table 5 respectively). Both patients
underwent open surgical reoperation with Ladd’s bands
resection. One patient (patient 22), who had postnatally

13

undergone simultaneous correction of esophageal atresia,
additionally underwent a fundoplication during reoperation.
Previously documented gastroesophageal reflux symptoms
and recurrent pneumonia in the other patient (patient 31)
subsided following the reoperation. Both cases elucidate the
necessity to watch out for constricting ligaments across the
duodenum. A high visual magnification during laparoscopy
may be helpful, but first and foremost a systematic duodenal
exploration is essential to avoid this complication. To date
none of our patients in follow-up suffers, or is in diagnostic
for, gastroesophageal reflux. However, long-term follow-up
is indicated for all operated patients due to the risk for late
complications [19].
In either OCDO and LCDO groups, one patient suffered
a localized perforation of the transverse colon on postop-
erative day 5 and 7, respectively. Both patients had initially
been operated on the second day of life. Based on examina-
tions of all notes and reports both complications most likely
resulted from thermal lesions by monopolar electrocautery.
These incidents illustrate the vulnerability and susceptibility
of neonatal tissue especially towards heat conduction from
electrocautery. It is advisable to use electrocautery with
utmost caution, preferably bipolar cautery or sealing.
Operative time in the LCDO group was significantly
longer compared to the OCDO group. This is consistent with
published data of comparative studies [10–12]. The decrease
of operative time in the LCDO group reflects a steep learn-
ing curve. Operative time will decrease further when time
consuming intracorporeal knotting is supported or replaced
by the deployment of appropriate instruments for knotting
or stapling in infants.
In this study, all patients, diagnosed with CDO during
a period of 13 years and undergoing surgical repair in our
center, were included. Because there were no exclusions in
this consecutive series, overall results are likely to mirror
pathologies, surgical repairs, comorbidities, and complica-
tions of patients with CDO largely realistic. Demographic
data and associated congenital anomalies appeared balanced
between LCDO and OCDO patient groups. There was no
difference with respect to gestational age at birth, age, and
weight at operation and associated anomalies in both groups.
Bias, related to time-epoch phenomena based on the
different lengths of follow-up of LCDO and OCDO group
appeared unlikely, because there were no differences regard-
ing time to initiation of feeds, time to full feeds, and hospi-
tal stay between patients operated < 2010 and ≥ 2010 in the
OCDO group.
The current study has several limitations, including the
retrospective nature of the study design, small sample size,
single institution experience, and short follow-up period.
A range of diagnoses and surgical procedures are being
subsumed under the umbrella of CDO and the associated
operative corrections. This cluster of congenital anomalies
13
Surgical Endoscopy
is evaluated all together though, because frequently the pre-
cise diagnosis is not established preoperatively but in the
course of surgical exploration. Despite various limitations,
our results are important; based on a consecutive series of
patients the results represent all surgically corrected patients
with a congenital duodenal obstruction over a 13-year period
of time in an experienced specialized department. In view
of the often small case loads of single centers, multicenter
trials are warranted. In an attempt to reduce bias due to con-
founding variables, a propensity score matching should be
implemented.
Conclusions
Consistent with the findings of our study, despite a longer
operative time for LCDO, a number of advantages of LCDO
over OCDO have been recognized when comparing both
approaches in the repair of congenital duodenal obstruction.
Such advantages include a lower morbidity, reduced time to
initiation and completion of full enteral feeds, and shorter
length of postoperative hospitalization for patients without
concomitant cardiac anomalies as well as for patients with-
out other anomalies when operated laparoscopically. In view
of the present results, LCDO, performed in selected patients,
appears to represent a viable alternative to OCDO.
Acknowledgements  The authors would like to acknowledge Rejane
Bager and Emilia Salzmann-Manrique, Biomedical Statisticians, Divi-
sion for Pediatric Stem-Cell Transplantation and Immunology, Univer-
sity Hospital Frankfurt, for the statistical analysis.
Disclosures  Drs. Gfroerer, Theilen, Fiegel, and Rolle have no conflicts
of interest or financial ties to disclose in relation to this manuscript.
References
1. van der Zee DC, Bax NM (1995) Laparoscopic repair of acute
volvulus in a neonate with malrotation. Surg Endosc 9:1123–1124
2. Bax NM, Ure BM, van der Zee DC, van Tuijl I (2001) Laparo-
scopic duodenoduodenostomy for duodenal atresia. Surg Endosc
2:217
3. Rothenberg SS (2002) Laparoscopic duodenoduodenostomy
for duodenal obstruction in infants and children. J Pediatr Surg
7:1088–1089
4. Kay S, Yoder S, Rothenberg S (2009) Laparoscopic duodenoduo-
denostomy in the neonate. J Pediatr Surg 5:906–908
5. Burgmeier C, Schier F (2012) The role of laparoscopy in the treat-
ment of duodenal obstruction in term and preterm infants. Pediatr
Surg Int 10:997–1000
6. Li B, Chen WB, Zhou WY (2013) Laparoscopic methods in the
treatment of congenital duodenal obstruction for neonates. J Lapa-
roendosc Adv Surg Tech 10:881–884
7. Son TN, Liem NT, Kien HH (2014) Laparoscopic simple oblique
duodenoduodenostomy in management of congenital duodenal
obstruction in children. J Laparoendosc Adv Surg Tech 2:163–166
Surgical Endoscopy
8. Oh C, Lee S, Lee SK, Seo JM (2017) Laparoscopic duodenoduo-
denostomy with parallel anastomosis for duodenal atresia. Surg
Endosc 6:2406–2410
9. Chung PH, Wong CW, Ip DK, Tam PK, Wong KK (2017) Is lapa-
roscopic surgery better than open surgery for the repair of con-
genital duodenal obstruction? A review of the current evidences.
J Pediatr Surg 3:498–503
10. Spilde TL, St Peter SD, Keckler SJ, Holcomb GW 3rd, Snyder CL,
Ostlie DJ (2008) Open vs laparoscopic repair of congenital duode-
nal obstructions: a concurrent series. J Pediatr Surg 6:1002–1005
11. Hill S, Koontz CS, Langness SM, Wulkan ML (2011) Laparo-
scopic versus open repair of congenital duodenal obstruction in
infants. J Laparoendosc Adv Surg Tech 10:961–963
12. Jensen AR, Short SS, Anselmo DM, Torres MB, Frykman PK,
Shin CE, Wang K, Nguyen NX (2013) Laparoscopic versus
open treatment of congenital duodenal obstruction: multicenter
short-term outcomes analysis. J Laparoendosc Adv Surg Tech
10:876–880
13. Parmentier B, Peycelon M, Muller CO, El Ghoneimi A, Bon-
nard A (2015) Laparoscopic management of congenital duodenal
atresia or stenosis: a single-center early experience. J Pediatr Surg
11:1833–1836
14. Dindo D, Demartines N, Clavien PA (2004) Classification of sur-
gical complications: a new proposal with evaluation in a cohort
of 6336 patients and results of a survey. Ann Surg 2:205–213
15. Slankamenac K, Graf R, Barkun J, Puhan MA, Clavien PA (2013)
The comprehensive complication index: a novel continuous scale
to measure surgical morbidity. Ann Surg 258:1–7
16. Kimura K, Tsugawa C, Ogawa K, Matsumoto Y, Yamamoto T,
Asada S (1977) Diamond-shaped anastomosis for congenital duo-
denal obstruction. Arch Surg 112:1262–1263
17. Martin RC, Brennan MF, Jaques DP (2002) Quality of complica-
tion reporting in the surgical literature. Ann Surg 6:803–813
18. Grosfeld JL, Rescorla FJ (1993) Duodenal atresia and stenosis:
reassessment of treatment and outcome based on antenatal diag-
nosis, pathologic variance, and long-term follow-up. World J Surg
3:301–309
19. Escobar MA, Ladd AP, Grosfeld JL, West KW, Rescorla FJ,
Scherer LR 3rd, Engum SA, Rouse TM, Billmire DF (2004) Duo-
denal atresia and stenosis: long-term follow-up over 30 years. J
Pediatr Surg 6:867–871
20. Short SS, Pierce JR, Burke RV, Papillon S, Frykman PK, Nguyen
N (2014) Is routine preoperative screening echocardiogram indi-
cated in all children with congenital duodenal obstruction? Pediatr
Surg Int 6:609–614
21. Bairdain S, Yu DC, Lien C, Khan FA, Pathak B, Grabowski MJ,
Zurakowski D, Linden BC (2014) A modern cohort of duodenal
obstruction patients: predictors of delayed transition to full enteral
nutrition. J Nutr Metab. https​://doi.org/10.1155/2014/85082​0
22. Aziz O, Athanasiou T, Tekkis PP, Purkayastha S, Haddow J,
Malinovski V, Paraskeva P, Darzi A (2006) Laparoscopic ver-
sus open appendectomy in children: a meta-analysis. Ann Surg
1:17–27
13