Journal of Pediatric Surgery 49 (2014) 10831086

Contents lists available at ScienceDirect

Journal of Pediatric Surgery

journal homepage: www.elsevier.com/locate/jpedsurg

Risk of incomplete pyloromyotomy and mucosal perforation in open and

laparoscopic pyloromyotomy
Nigel J. Hall a,, Simon Eaton a, Aaron Seims b, Charles M. Leys b, John C. Densmore c, Casey M. Calkins c,
Daniel J. Ostlie d, Shawn D. St Peter d, Richard G. Azizkhan e, Daniel von Allmen e, Jacob C. Langer f,
Eveline Lapidus-Krol f, Sarah Bouchard g, Nelson Pich g, Steven Bruch h, Robert Drongowski h,
Gordon A. MacKinlay i, Claire Clark i, Agostino Pierro a, f
a

UCL Institute of Child Health & Great Ormond Street Hospital for Children, London, UK
Riley Hospital for Children, Indianapolis, IN
c
The Children's Hospital of Wisconsin, Medical College ofWisconsin, Milwaukee, WI
d
Children's Mercy Hospital, Kansas City, MO
e
Cincinnati Children's Hospital Medical Center, Cincinnati, OH
f
Hospital for Sick Children and University of Toronto, Toronto, Canada
g
Hospital Sainte-Justine, Montreal, Canada
h
University of Michigan, Ann Arbor, MI
i
Royal Hospital for Sick Children, Edinburgh, UK
b

a r t i c l e

i n f o

Article history:
Received 4 July 2013
Received in revised form 7 October 2013
Accepted 7 October 2013
Key words:
Pyloric stenosis
Pyloromyotomy
Minimally invasive surgery
Infant

a b s t r a c t
Background: Despite randomized controlled trials and meta-analyses, it remains unclear whether
laparoscopic pyloromyotomy (LP) carries a higher risk of incomplete pyloromyotomy and mucosal
perforation compared with open pyloromyotomy (OP).
Methods: Multicenter study of all pyloromyotomies (May 2007December 2010) at nine high-volume
institutions. The effect of laparoscopy on the procedure-related complications of incomplete pyloromyotomy
and mucosal perforation was determined using binomial logistic regression adjusting for differences among
centers.
Results: Data relating to 2830 pyloromyotomies (1802 [64%] LP) were analyzed. There were 24 cases of
incomplete pyloromyotomy; 3 in the open group (0.29%) and 21 in the laparoscopic group (1.16%). There
were 18 cases of mucosal perforation; 3 in the open group (0.29%) and 15 in the laparoscopic group (0.83%).
The regression model demonstrated that LP was a marginally signicant predictor of incomplete
pyloromyotomy (adjusted difference 0.87% [95% CI 0.0064.083]; P = 0.046) but not of mucosal perforation
(adjusted difference 0.56% [95% CI 0.096 to 3.365]; P = 0.153). Trainees performed a similar proportion of
each procedure (laparoscopic 82.6% vs. open 80.3%; P = 0.2) and grade of primary operator did not affect the
rate of either complication.
Conclusions: This is one of the largest series of pyloromyotomy ever reported. Although laparoscopy is
associated with a statistically signicant increase in the risk of incomplete pyloromyotomy, the effect size is
small and of questionable clinical relevance. Both OP and LP are associated with low rates of mucosal
perforation and incomplete pyloromyotomy in specialist centers, whether trainee or consultant surgeons
perform the procedure.
2014 Elsevier Inc. All rights reserved.

Since its introduction two decades ago, the laparoscopic approach

to pyloromyotomy has gained popularity and has been implemented
by many centers. During this period, a number of groups have
reported comparative outcomes between the traditional open
procedure and the laparoscopic procedure. Such studies initially
took the form of retrospective comparative reports, but there have
since been at least ve prospective randomized controlled trials
Corresponding author at: Surgery Unit, UCL institute of Child Health, 30 Guilford St,
London, WC1N 1EH, UK. Tel.: +44 207 9052 641; fax: +44 207 4046 181.
E-mail address: nigel.hall@ucl.ac.uk (N.J. Hall).
http://dx.doi.org/10.1016/j.jpedsurg.2013.10.014
0022-3468/ 2014 Elsevier Inc. All rights reserved.

(RCTs) [15]. Furthermore, there are three published systematic

reviews with meta-analysis covering this topic [68]. The reported
benets (as dened by the attainment of statistical signicance
[P b 0.05]) in these RCTs and meta-analyses include shorter time to
achieve full feeds [3,6,8], shorter postoperative time in hospital [3,6],
less postoperative vomiting [1], less pain [1,3,4], fewer wound
complications [6], better cosmesis [2] and improved cost-effectiveness [9]. Despite these benets, there remain concerns that the
laparoscopic approach may subject patients to a higher risk of the
specic procedure-related complications of incomplete pyloromyotomy and mucosal perforation [4,6,7,10]. Although a statistically

1084

N.J. Hall et al. / Journal of Pediatric Surgery 49 (2014) 10831086

signicant difference in incidence of these complications has not been

demonstrated in any of the RCTs, none was large enough to detect
such a difference because of the low incidence of these complications.
In meta-analysis there is a trend toward a higher incidence of
incomplete pyloromyotomy using the laparoscopic technique that
approaches statistical signicance (P = 0.06) [6,7] although the
validity of this gure has been questioned owing to concerns over
the statistical techniques used [10,11].
The difculty in determining whether there is really a difference in
the incidence of incomplete pyloromyotomy and mucosal perforation
between laparoscopic and open pyloromyotomy lies in the sample
size required to demonstrate a difference, if one truly exists. Even in
published meta-analyses, the total number of patients available
(owing to the limited number of RCTs included and their individual
sample sizes) remains limited at approximately 500 [68]. We have
previously estimated that a sample size of approximately 1250
patients would be required to demonstrate a statistically signicant
difference in the incidence of incomplete pyloromyotomy between
laparoscopic and open procedures of 2.5% (the unadjusted difference
in incidence obtained from the raw data included in meta-analysis
[10]). Such a large study collecting data from individual patients has
never been reported in the eld of pediatric surgery to our knowledge.
Therefore, we aimed to perform a study large enough to overcome the
sample size limitations of previous studies with adequate power to
detect small differences in the incidence of these complications
between open and laparoscopic approaches.
1. Methods
We performed a retrospective study in nine high-volume specialist
pediatric surgical centers in the UK, USA and Canada. IRB/ethical
approval was obtained for each center. Centers were invited for
inclusion following a mailshot and selected only if they performed at
least 15 laparoscopic pyloromyotomies per year. The study period ran
from 1st May 2007 to 31st December 2010. This start date was chosen
as being the date following which the last patient was randomized
into a previous RCT of open versus laparoscopic pyloromyotomy [3] to
exclude duplication of reporting. Data relating to all pyloromyotomies
for infantile hypertrophic pyloric stenosis performed in this period
were recorded including patient demographics (age, gender), operative approach (open or laparoscopic), conversion rate of laparoscopy,
and grade of operating surgeon (trainee under supervision or
consultant). Grade of operating surgeon was recorded because a
previous RCT reported that a signicantly higher proportion of
laparoscopic procedures were performed by consultants than were
open procedures [3]. Although grade of operating surgeon had no
inuence on the primary outcomes of that RCT, the effect of grade of
primary operator on the primary outcomes of this study remains
unknown. The primary outcomes for this study were the specic
procedure-related complications of (i) incomplete pyloromyotomy
and (ii) mucosal perforation. A mucosal perforation was dened as
breach of the mucosal barrier during pyloromyotomy which was
either identied at the time of the myotomy and repaired immediately, or became apparent in the postoperative period and was
reexplored and repaired. An incomplete pyloromyotomy was dened
as an inadequate myotomy requiring another procedure. Redo
procedures following incomplete pyloromyotomy were excluded
from the analysis.
1.1. Statistical analysis
A binomial multivariate logistic regression model was generated to
calculate incidence of primary outcomes taking into account differences in rate of complications between different centers and grade of
primary operator. Analysis of patient demographics revealed a
statistically signicant difference in age at surgery between open

and laparoscopic groups. Furthermore an initial univariate analysis

also demonstrated that age at surgery, but not gender, was a
signicant determinant of IP. Therefore age at surgery was included
in the nal multivariate model. Non-parametric data were compared
using a MannWhitney U test and proportions were compared using
chi-square test. SPSS version 18 was used for all analyses.
1.2. Sample size
Based on data from previous RCTs and meta-analyses we
calculated that to detect statistically signicant differences in the
incidence of incomplete pyloromyotomy and mucosal perforation
would require a total sample size of approximately 1250 and 390,000
respectively. Given the impossibility of achieving the second of these
sample sizes (and therefore a near-impossible chance that there is a
clinically signicant difference) we designed a study aiming for a
sample of at least 1250 patients, and selected centers and a time scale
in order to achieve this.
2. Results
Overall there were 2830 pyloromyotomies; 1028 were performed
open and 1802 laparoscopically. The distribution of cases between the
nine centers is shown in Table 1. Patients operated on laparoscopically
were slightly older (difference in median age 2 days) and weighed
more (difference in median weight 0.14 kg); these differences were
statistically signicant (Table 2).
Thirty-ve laparoscopic procedures (1.9%) were converted to an
open procedure for reasons including inability to complete the
procedure laparoscopically, suspected mucosal perforation and for
repair of conrmed mucosal perforation. However, not all instances of
mucosal perforation in laparoscopic cases were converted to an open
procedure. All such conversions were analyzed in the laparoscopic
group on an intention-to-treat basis.
In total there were 24 cases of incomplete pyloromyotomy; 3 in
the open group (0.29%) and 21 in the laparoscopic group (1.16%).
There were 18 cases of mucosal perforation; 3 in the open group
(0.29%) and 15 in the laparoscopic group (0.83%). The binomial
logistic regression model demonstrated a signicantly higher incidence of incomplete pyloromyotomy with laparoscopic surgery
compared with open (adjusted difference 0.87% [95% CI 0.006
4.083]; P = 0.046) but no signicant difference in the incidence of
mucosal perforation (Table 3). The distribution of mucosal perforation
and incomplete pyloromyotomy among centers is shown in Table 4.
At least one complication occurred in each center and there was
no signicant relationship between predominant procedure
performed at each center (laparoscopic or open) and distribution
of complications.
The proportion of laparoscopic and open procedures performed by
trainees was similar (laparoscopic 1446/1749 [82.6%] vs. open 465/

Table 1
Distribution and description of cases between centers.
Center

A
B
C
D
E
F
G
H
I
Total
a

414
350
389
502
179
373
108
344
171
2830

Age (days)

Weight (kg)

Male

Median

IQRa

Median

IQRa

36
34
36
34
35
34
38
35
36
35

2948
2745
2747
2745
2747
2646
3049
2749
2947
2747

3.92
3.90
3.96
3.83
4.10
4.00
3.80
4.10
3.76
3.90

3.484.40
3.464.32
3.444.44
3.464.50
3.504.55
3.564.50
3.354.30
3.604.56
3.404.30
3.504.44

350
292
329
424
140
298
83
285
146
2347

84.5
83.4
84.6
84.5
78.2
79.9
76.9
82.8
84.9
82.9

Interquartile range.

N.J. Hall et al. / Journal of Pediatric Surgery 49 (2014) 10831086

Table 2
Demographic details of infants undergoing open or laparoscopic pyloromyotomy.
Operative approach

Age at surgery (days) a

Weight at surgery (kg) a
Male, n (%)
a
b
c

Open

Laparoscopic

34 (2745)
3.85 (3.434.40)
864 (84%)

36 (28-48)
3.99 (3.504.47)
1483 (82%)

0.003b
0.002b
NSc

Data are median (interquartile range).

MannWhitney U test.
Chi-square test.

579 [80.3%]; P = 0.2; data unknown for 502 procedures, 1 center).

Grade of operating surgeon had no effect on the incidence
of either incomplete pyloromyotomy (adjusted difference trainee vs.
consultant 0.438% [95% CI 1.198 to 1.483]) or mucosal
perforation (adjusted difference trainee vs. consultant 0.677%
[95% CI 1.183 to 0.8]).
3. Discussion
Despite high-quality RCTs and meta-analyses, until now it has
remained unknown whether laparoscopic pyloromyotomy carries a
higher incidence of incomplete pyloromyotomy or mucosal perforation than the open procedure. In this multicenter review, comprising
one of the largest series of pyloromyotomy ever reported, we have
identied a small (0.87%) but statistically signicant (P = 0.046)
increase in the incidence of incomplete pyloromyotomy in the
laparoscopic group compared with the open group. Although this is
a statistically signicant difference, we believe that the size of the
clinical effect it describes should be carefully considered. The
difference in incidence of incomplete pyloromyotomy of 0.87%
indicates that the number of children that would need to be treated
laparoscopically in order for one additional episode of incomplete
pyloromyotomy to occur is 115. While there is a higher proportion of
children in the laparoscopic group who will require a repeat
procedure, this marginally higher rate of repeat procedure must be
balanced against previously documented advantages of laparoscopic
pyloromyotomy over the open procedure [1,3,6]. Furthermore, it
should be noted that previous studies reporting shorter recovery
time following laparoscopy have included patients who have had
an incomplete pyloromyotomy [3,6,8]. Thus the low incidence
of an extended postoperative recovery period from incomplete
pyloromyotomy does not offset the advantage of laparoscopy for
the population.
In agreement with previous smaller studies [1,3,4] we have not
identied a statistically signicant difference in incidence of mucosal
perforation between open and laparoscopic pyloromyotomy.
The reason for the higher incidence of incomplete pyloromyotomy
with laparoscopic procedure is not immediately apparent. While it is
possible that the lack of tactile feedback during laparoscopy results in
a less precise pyloromyotomy, St Peter and colleagues [1] have
impressively demonstrated in an RCT that by ensuring a pyloromyotomy of adequate length, this complication can be avoided altogether.
By ensuring a pyloromyotomy of minimum length 2 cm (measured
Table 3
Rates of procedural complications of pyloromyotomy.
Operative approach

Mucosal perforation (%)

Incomplete
pyloromyotomy (%)

Open

Laparoscopic

0.29
0.29

0.83
1.16

* Binomial logistic regression analysis.

a
Adjusted for center and age.

Adjusted differencea,
(%) (95% CI)

P*

0.56% (0.096 to 3.365)

0.87% (0.006 to 4.083)

0.153
0.046

1085

Table 4
Distribution and description of complications between centers.
Center

A
B
C
D
E
F
G
H
I
Total

Open

Laparoscopic

Total, n (%)

MP

IP

MP

IP

MP

IP

86
265
0
449
2
118
15
89
4
1028

0
1

1
0
1
0
0
0
3

0
3

0
0
0
0
0
0
3

328
85
389
53
177
255
93
255
167
1802

2
1
1
0
1
4
4
1
1
15

9
1
1
0
0
1
3
5
1
21

2 (0.48%)
2 (0.57%)
1 (0.26%)
1 (0.19%)
1 (0.56%)
5 (1.34%)
4 (3.70%)
1 (0.29%)
1 (0.58%)
18 (0.64%)

9 (2.17%)
4 (1.14%)
1 (0.26%)
0 ()
0 ()
1 (0.27%)
3 (2.78%)
5 (1.45%)
1 (0.58%)
24 (0.84%)

MP, mucosal perforation; IP, incomplete pyloromyotomy.

intraoperatively with an intracorporally placed length of string)

they achieved a zero incidence of incomplete pyloromyotomy
during laparoscopic procedure. These results were available to all
centers prior to the rst patient in this study being treated, yet still
overall a higher incidence with laparoscopy was encountered. Other
factors that cannot be controlled for outside the strict connes of an
RCT in a real-world setting may be contributory. Our results
therefore describe the rates of complication applicable to patients
outside an RCT.
We have also demonstrated that grade of primary operator
(trainee under supervision or consultant/attending) has no signicant
effect on the incidence of either incomplete pyloromyotomy or
mucosal perforation. A previous RCT comparing open and laparoscopic pyloromyotomy also reported no signicant effect on primary
outcomes attributable to grade of primary operator [3]. However, the
majority of laparoscopic procedures were performed by a consultant
surgeon while the majority of open procedures were performed by a
supervised trainee. In comparison, the majority of both open and
laparoscopic procedures in this series were performed by a trainee
under supervision. The fact that these results are valid for procedures
performed predominantly by trainees under supervision clearly has
important implications for service delivery as well as for surgical
training. These results demonstrate that equally good outcomes can
be achieved from open and minimally invasive surgery within
pediatric surgical training institutions where the majority of procedures are performed by trainees.
The main strength of this study is its size. This is one of the largest
series of pyloromyotomies ever reported. This high number of cases,
combined with the multicenter nature of the study, allows us to draw
reliable conclusions applicable to a large number of similar centers
worldwide. This study demonstrates that to detect small differences
in treatment outcomes large, multicenter collaborations are essential.
This study has demonstrated a statistically signicant difference in
incidence of incomplete pyloromyotomy when smaller RCTs have
failed to do so because of sample size limitations. These data also
come from a real-life setting; that is all patients were included
and inclusion in the study was not limited by the eligibility criteria of
an RCT.
A potential limitation to this study is that data were collected
retrospectively. From the outset, we considered whether to perform a
second, larger, multicenter randomized study as a follow-on from our
previous trials [1,3] or a retrospective review. In general terms,
prospective studies are preferred as they reduce the inuence of
biases on the results of a study. A study in which any of the outcomes
are subjective or consist of a continuous variable may be particularly
prone to the inuence of bias in this way. However, a study such as the
current one that uses well-dened, objective, categorical endpoints is
highly unlikely to be inuenced by the sources of bias that exist in a
typical retrospective study. In addition, the procedure-related
complications analyzed in this study are diagnosed either during the

1086

N.J. Hall et al. / Journal of Pediatric Surgery 49 (2014) 10831086

operation or soon after it, and patients are treated in the same center.
This reduces the chance of these complications being missed from the
analysis and not reported. We therefore justify this retrospective
study which has the added advantage of providing data to the surgical
community more rapidly than a prospective study. Interestingly, age
at surgery and weight at surgery were signicantly higher (by 2 days
and 140 g respectively) in the group undergoing laparoscopic
pyloromyotomy than the group undergoing open pyloromyotomy in
this study. Although this is a potential confounding factor (and source
of bias), we have controlled for age at surgery in the analyses by
including age in our regression model. We did not additionally control
for weight in the regression analysis, as age and weight are highly
correlated as expected and controlling for both age and weight would
yield an overadjusted model.
A further observation is that these data were obtained from highvolume, specialist pediatric surgical centers. Whether the results of
this study are generalizable and applicable to smaller or nonspecialist
units is not known. There is a recognized association between patient
outcomes and both high volume and specialization [12,13]. Overall
the procedural complication rates reported in this study are among
the lowest reported for either open or laparoscopic procedures, a
factor that we suspect is at least partly caused by volume and
specialization. While Langer and To [14] have shown improved
outcomes after open pyloromyotomy done by a pediatric Surgeon
compared with that by a general surgeon and Haricharan and
colleagues [15] have shown an increased rate of complications
when laparoscopic pyloromyotomy is performed by a general surgical
resident rather than a pediatric surgery resident, it is also true that Ali
and colleagues [16] have demonstrated that laparoscopic pyloromyotomy can be safely and effectively performed by a non-specialist
general surgeon in a non-specialist center.
These data may be interpreted differently by different groups of
surgeons. Laparoscopic enthusiasts may regard the difference in rate
of incomplete pyloromyotomy between laparoscopic and open
approaches to be so small as to be of questionable clinical signicance.
This, combined with the previously reported benets of laparoscopy,
may be viewed as continued justication in support of the use of the
laparoscopic approach. Conversely, surgeons who favor open pyloromyotomy may focus on the statistically higher incidence of
incomplete pyloromyotomy with laparoscopy. We believe that this
study demonstrates that the rates of incomplete pyloromyotomy and
mucosal perforation at high-volume pediatric surgical centers are low
during both open and laparoscopic procedures and justify the
continued use of both procedures in such institutions. However, we
would recommend that particular attention be paid to the length of

the pyloromyotomy during laparoscopy especially in young or

small infants.
4. Summary
Open and laparoscopic pyloromyotomies have similar rates of
mucosal perforation and incomplete pyloromyotomy. Both approaches to pyloromyotomy are safe and carry an acceptable rate of
these complications in high-volume, specialist centers.
References
[1] St Peter SD, Holcomb III GW, Calkins CM, et al. Open versus laparoscopic
pyloromyotomy for pyloric stenosis: a prospective, randomized trial. Ann Surg
2006;244:36370.
[2] Siddiqui S, Heidel RE, Angel CA, et al. Pyloromyotomy: randomized control trial of
laparoscopic vs open technique. J Pediatr Surg 2012;47:938.
[3] Hall NJ, Pacilli M, Eaton S, et al. Recovery after open versus laparoscopic
pyloromyotomy for pyloric stenosis: a double-blind multicentre randomised
controlled trial. Lancet 2009;373:3908.
[4] Leclair MD, Plattner V, Mirallie E, et al. Laparoscopic pyloromyotomy for
hypertrophic pyloric stenosis: a prospective, randomized controlled trial. J Pediatr
Surg 2007;42:6928.
[5] Greason KL, Allshouse MJ, Thompson WR, et al. A prospective, randomized
evaluation of laparoscopic versus open pyloromyotomy in the treatment of
infantile hypertrophic pyloric stenosis. Pediatr Endosurgery Innov Techn 1997;1:
1759.
[6] Sola JE, Neville HL. Laparoscopic vs open pyloromyotomy: a systematic review and
metaanalysis. J Pediatr Surg 2009;44:16317.
[7] Jia WQ, Tian JH, Yang KH, et al. Open versus laparoscopic pyloromyotomy for
pyloric stenosis: a meta-analysis of randomized controlled trials. Eur J Pediatr Surg
2011;21:7781.
[8] Oomen MW, Hoekstra LT, Bakx R, et al. Open versus laparoscopic pyloromyotomy
for hypertrophic pyloric stenosis: a systematic review and meta-analysis focusing
on major complications. Surg Endosc 2012;26:210410.
[9] Carrington EV, Hall NJ, Pacilli M, et al. Cost-effectiveness of laparoscopic versus
open pyloromyotomy. J Surg Res 2012;178:31520.
[10] Hall NJ, Eaton S, Pierro A. Editorial on Open versus laparoscopic pyloromyotomy
for pyloric stenosis: a meta-analysis of randomized controlled trials by Jia et al.
Eur J Pediatr Surg 2011;21:756.
[11] Eaton S, Hall NJ, Pierro A. Zero-total event trials and incomplete pyloromyotomy.
J Pediatr Surg 2009;44:24345.
[12] Safford SD, Pietrobon R, Safford KM, et al. A study of 11,003 patients with
hypertrophic pyloric stenosis and the association between surgeon and hospital
volume and outcomes. J Pediatr Surg 2005;40:96772.
[13] Chowdhury MM, Dagash H, Pierro A. A systematic review of the impact of
volume of surgery and specialization on patient outcome. Br J Surg 2007;94:
14561.
[14] Langer JC, To T. Does pediatric surgical specialty training affect outcome after
Ramstedt pyloromyotomy? A population-based study. Pediatrics 2004;113:
13427.
[15] Haricharan RN, Aprahamian CJ, Celik A, et al. Laparoscopic pyloromyotomy: effect
of resident training on complications. J Pediatr Surg 2008;43:97101.
[16] Ali A, Tripuraneni G, Velmurugan S, et al. Laparoscopic pyloromyotomy is both
safe and effective in a district hospital. Surg Endosc 2008;22:1513.