Original article
Survival of children following Nissen fundoplication
R. Wockenforth1, C. S. Gillespie2 and B. Jaffray3
1Department of Paediatric Surgery, The Royal Victoria Infirmary, and 2School of Mathematics & Statistics and 3Clinical Medical Sciences, University of
Newcastle, Newcastle upon Tyne, UK
Correspondence to: Mr B. Jaffray, Department of Paediatric Surgery, The Royal Victoria Infirmary, Queen Victoria Road, Newcastle upon Tyne NE2 2PY,
UK (e-mail: bruce.jaffray@ncl.ac.uk)
Background: Analyses of survival after fundoplication in childhood are often restricted to 30-day
mortality, or to the neurologically impaired. The objective of this study was to report actuarial survival
and variables associated with mortality for all children undergoing fundoplication.
Methods: This was a prospective observational study of fundoplication surgery by one surgeon; the
endpoint was survival. Using a Cox proportional hazards model, gastrostomy, neurological status,
tracheostomy, congenital cardiac disease, syndromic status, presence of congenital anomaly, other
chronic disease, weight z-score at time of surgery, need for revisional fundoplication, use of laparoscopic
surgery, gastric drainage procedures, age and sex were assessed for their influence on survival.
Results: Two-hundred and thirty children underwent 255 fundoplications at a median age of 3·6 years.
Forty-six children (20·0 per cent) died during amedian follow-up of 2·8 (range 0·5–11·2) years. Statistical
modelling showed gastrostomy (relative risk of death 11·04, P < 0·001), cerebral palsy (relative risk 6·58,
P = 0·021) and female sex (relative risk 2·12, P = 0·015) to be associated with reduced survival. Revisional
fundoplication was associated with improved survival (relative risk of death 0·37, P = 0·037). Survivors
had significantly higher weight z-scores (−1·4 versus −2·9 for those who died; P = 0·001). The 5-year
survival rate after fundoplication for children with cerebral palsy and gastrostomy was 59 per cent.
Conclusion: Survival of children following fundoplication is related principally to the presence of a
gastrostomy and neurological status. Estimates of children’s life expectancy should take account of the
poorer survival of neurologically impaired children who undergo fundoplication, presumably due to the
related co-morbidities that lead to a gastrostomy.
Paper accepted 1 December 2010
Published online in Wiley Online Library (www.bjs.co.uk). DOI: 10.1002/bjs.7415
Introduction
In children, common complications of gastro-oesophageal
reflux disease (GORD) include oesophagitis, vomiting
with failure to thrive, and aspiration. Children with
these complications frequently undergo fundoplication1,
most often with the Nissen technique2. The efficacy of
this procedure is usually measured by indices of reflux
control3,4. There are few analyses of survival, with reports
either restricted to deaths within 1 month of surgery2,
or failing to provide actuarial analysis of survival5,6.
One recent study restricted to children with neurological
impairment suggested that approximately 90 per cent were
alive 5 years after fundoplication7.
Many children undergoing fundoplication suffer from
cerebral palsy, which is associated with a reduced life
expectancy. Although some authors relate the short life of
these children to the severity of the neurodisability8,9,
others have suggested that functional abilities such as
feeding are more important10. The degree to which life
expectancy is reduced is controversial11, but the use of a
surgical intervention such as fundoplication as a predictive
variable in assessing life expectancy among such children
has not been described previously.
This studywas an actuarial analysis of survival of children
following fundoplication. The aim was to identify variables
independently associated with mortality.
Methods
Details for all children undergoing fundoplication by a
single surgeon between January 1998 and December 2009
 2011 British Journal of Surgery Society Ltd British Journal of Surgery
Published by John Wiley & Sons Ltd

R. Wockenforth, C. S. Gillespie and B. Jaffray
were recorded prospectively in a customized database at
the time of surgery. Details recorded included indications
for surgery, age at first operation, whether the child had
a gastrostomy or a tracheostomy, whether a pyloroplasty
was performed, whether the child subsequently underwent
a revisional fundoplication, and whether the procedure was
performed using laparoscopic or conventional surgery.
Case-note review allowed identification of the further
following variables: weight at operation, neurological
impairment, congenital cardiac disease and any surgery
performed to correct this, syndromic status, congenital
anomalies, and coexisting chronic disease. The variables
were examined for influence on survival.
The diagnosis of cerebral palsy was confirmed by cross-
referencing the database with the records of the North
of England Collaborative Cerebral Palsy Study. Data
were also cross-referenced with the Northern Congenital
Anomaly Survey data and with the regional cardiac surgery
database.
The regional ethics committee confirmed that this study
design did not require ethics committee approval.
Statistical analysis
Statistical analysis was performed using the programRwith
the survival package12. A full Cox proportional hazards
model was fitted to the data, and the residuals were
checked for deviations away from the model assumptions.
The residuals indicated that there was strong evidence for
non-proportional hazards for cerebral palsy. However, the
inclusion of a gastrostomy/cerebral palsy interaction term
corrected this problem. Interaction effects involving sex
and age were also examined. Co-variables were removed
according to the Wald statistic. Model validity was checked
by starting with an empty model and progressively adding
significant terms. The final model contained five co-
variables including the non-significant interaction term.
In a data set of this type, many variables are likely to
be correlated. For example, patients with cerebral palsy
were more likely to have had a gastrostomy. Rather
than modelling each co-variable individually, which would
suggest many spurious relationships, the risk factors
were jointly modelled simultaneously, thereby significantly
reducing the chance of type 1 errors.
Kaplan–Meier survival curves are presented, fromwhich
5 year survival rates and median survival were calculated.
Graphs were generated using the ggplot2 package13.
Mortality rates were calculated for the following
risk factors, singly and in combination: cerebral palsy,
gastrostomy, tracheostomy and congenital cardiac disease.
A child was classed as normal by the absence of any of these
variables or any syndrome or congenital anomaly.
Weights on the day of surgery were converted to z-
scores using the data from the Child Growth Foundation
for weight distribution among British children14 and were
compared using Student’s t test. P < 0·050 was considered
statistically significant.
Results
Two-hundred and thirty children (105 girls) underwent
255 fundoplication operations; their median age at first
fundoplication was 3·6 (range 0·1–18·9) years. Median
follow-up was 2·8 (range 0·5–11·2) years. Forty-six
children (20·0 per cent) died during follow-up. Indications
for surgery are listed in Table 1. Incidence of death
according to the child’s risk factors is detailed in Table 2.
Seventy-six children had cerebral palsy, 114 had a
gastrostomy, and 18 a tracheostomy. Forty-two children
had more than one fundoplication, 20 of whom had
undergone the original fundoplication by a surgeon other
than the senior author andwere referred for their revisional
Table 1 Indications for fundoplication
No. of patients (n = 230)
GORD 68 (29·6)
GORD and vomiting 40 (17·4)
Vomiting 37 (16·1)
Acute life-threatening incident 30 (13·0)
Failure to thrive 22 (9·6)
Aspiration 19 (8·3)
Combination 9 (3·9)
Other 5 (2·2)
Values in parentheses are percentages. GORD, gastro-oesophageal reflux
disease.
Table 2 Deaths according to risk factor status
Status*
No. of
patients
No. of
deaths
Normal (no risk factors) 83 0
Cerebral palsy only 15 2
Cerebral palsy + gastrostomy 47 17
Cerebral palsy + gastrostomy +
tracheostomy
5 2
Cerebral palsy + gastrostomy +
congenital cardiac disease
5 3
Gastrostomy only 20 5
Tracheostomy only 1 0
Gastrostomy + tracheostomy 3 1
Congenital cardiac disease only 6 2
Congenital cardiac disease +
gastrostomy
8 2
All patients 230 46
*Not all possible combinations of risk factors are presented.
 2011 British Journal of Surgery Society Ltd www.bjs.co.uk British Journal of Surgery
Published by John Wiley & Sons Ltd

Survival of children following Nissen fundoplication
surgery. Of the remaining 22, 18 underwent one revisional
procedure and one child had two revisional procedures;
two children had three revisional procedures, and one
of these children’s original operation was performed by
another surgeon. Thus a total of 255 fundoplications
were performed by the senior author in this series.
Two laparoscopic fundoplications were converted to open
surgery; both were analysed as a laparoscopic procedure.
Forty-one children were diagnosed with a named
syndrome, 30 had other congenital anomalies and 22 had
undergone surgery for congenital cardiac disease.
Pyloroplasty was performed on nine children early in
the series and was abandoned after May 2002 because of
the severity of dumping symptoms.
The mean(s.d.) weight z-score at the time of surgery was
−1·8(2·2).
Survival analysis
Variables that significantly influenced survival were
gastrostomy (P < 0·001), cerebral palsy (P = 0·021), sex
(P = 0·015) and need for revisional surgery (P = 0·037).
The strongest risk factor for death in this series was the
presence of a gastrostomy. For childrenwith a gastrostomy,
the relative risk of death (compared with other children
without a gastrostomy) was 11·04 (95 per cent c.i. 3·16
to 38·63). When patients had both cerebral palsy and
a gastrostomy, there was a pronounced shortening in
lifespan (Fig. 1). For children without cerebral palsy, the
5-year survival rate for a child with a gastrostomy following
fundoplication was 71 per cent, compared with 97 per cent
for a child without a gastrostomy. For children with a
gastrostomy and cerebral palsy, the 5-year survival rate was
59 per cent.
Cerebral palsy was confirmed as a risk factor for
mortality. The relative risk of death in children with
cerebral palsy was 6·58 (95 per cent c.i. 1·32 to 32·78)
compared with that in children without cerebral palsy.
The 30-day mortality rate for children with cerebral palsy
undergoing fundoplication was 11 per cent (8 of 76). By
age 20 years, it was calculated that less than 30 per cent
of children with cerebral palsy and gastrostomy would be
expected to be alive (data not shown).
Revision of the fundoplication was associated with
reduced mortality. The relative risk of death for children
who did not undergo revision was 2·73 (95 per cent c.i.
1·06 to 7·04).
Survival analysis also identified sex as a significant risk
factor (Fig. 2). Relative risk of death for female compared
with male children was 2·12 (1·16 to 3·89).
None of the other possible explanatory variables,
including congenital cardiac disease, tracheostomy, weight
99
44
13
45
98
40
13
40
98
37
12
35
No. at risk
No CP, no
gastrostomy
101
53
15
61
No CP, with
gastrostomy
CP, no
gastrostomy
CP, with
gastrostomy
No CP, no gastrostomy
No CP, with gastrostomy
CP, no gastrostomy
CP, with gastrostomy
2 4 6
Time after surgery (years)
100
75
50
25
0
O
ve
ra
ll s
ur
viv
al
(%
)
Fig. 1 Effect of gastrostomy and cerebral palsy (CP) on the
survival of 230 children undergoing Nissen fundoplication
69
81
30
33
68
80
29
29
68
79
27
27
No. at risk
No CP, male 70
84
34
42
No CP, female
CP, male
CP, female
2 4 6
Time after surgery (years)
100
75
50
25
0
O
ve
ra
ll s
ur
viv
al
(%
)
No CP, male
No CP, female
CP, male
CP, female
Fig. 2 Effect of sex and cerebral palsy (CP) on the survival of 230
children undergoing Nissen fundoplication
 2011 British Journal of Surgery Society Ltd www.bjs.co.uk British Journal of Surgery
Published by John Wiley & Sons Ltd

R. Wockenforth, C. S. Gillespie and B. Jaffray
z-scores and syndromic status, had a significant effect on
mortality rates.
Effect of weight z-scores
Children who had a gastrostomy were significantly lighter
than those who did not (mean(s.d.) z-score −2·9(2·0)
versus −0·6(1·7) respectively; P < 0·001, 95 per cent c.i.
of difference 1·6 to 2·8).
Children with cerebral palsy were significantly lighter
than children without cerebral palsy at the time
of fundoplication (mean(s.d.) z-score −2·8(2·2) versus
−1·1(2·0) respectively; P < 0·001, 95 per cent c.i. of
difference 0·9 to 2·3).
Children who died were also significantly lighter than
survivors at the time of their fundoplication (mean(s.d.)
z-score −2·9(2·3) versus −1·4(2·1); P = 0·001, 95 per cent
c.i. of difference 0·6 to 2·4).
Discussion
The two principal findings in this study of the
survival of children following fundoplication are the
association between gastrostomy and death, independent
of neurological status, and the extremely poor survival of
neurologically impaired children, which was significantly
worse than previous estimates9–11.
Other variables affecting survival were sex and revisional
fundoplication. Male sex appeared to be protective. Other
studies have suggested no effect of sex on survival among
children with neurodisability10, and the present authors
are unable to explain this observation. An effect of sex
on survival has, however, been reported for other chronic
diseases15.
Revisional fundoplication improved survival. The likely
explanation is that this group must survive long enough for
their fundoplication to fail and are therefore selected for
survival at least until their second operation. Alternatively,
restoration of protection from reflux by the revisional
procedure may improve survival.
This was a single-surgeon series, thereby removing
the systemic error present in multisurgeon or multi-
institutional series. The only changes in practice were the
adoption of laparoscopic surgery and the abandonment of
gastric drainage procedures, which had no effect on survival
The integrity of the data was ensured by being recorded
prospectively by the surgeon at the time of operation.
Among limitations of the present study is the possibility
that patients were somehow a selected group with more
severe disabilities. Referrals for fundoplication were taken
from across the North of England, with no attempt to
select or screen children. It is possible that this cohort had
poorer survival by chance. There has been no evidence of
increasing severity of neurodisability fromUKorEuropean
registries (A. F. Colver, personal communication), and
studies of life expectancy among children with cerebral
palsy suggest that survival is improving16.
Another limitation is that the cause of death could not
be ascertained. Many children did not have a post-mortem
examination. Although aspiration pneumonia could be
suspected as the commonest cause of death, there was
no proof for this.
A systematic review of gastrostomy feeding amongst
neurologically impaired children suggested increased
mortality17, but there was heterogeneous use of antireflux
procedures. The present study supports the excess
mortality associated with gastrostomy tube placement.
The 5-year survival rate of 59 per cent for children with
cerebral palsy and gastrostomy undergoing fundoplication
in the present series contrasts with the 61 per cent reported
by an Australian study, with heterogeneous use of an
antireflux procedure18. Indirect support for the contention
that gastrostomy is associated with poorer survival is the
mortality rate of 21 per cent for neurologically impaired
children treated with a gastrojejunal tube, compared with
12 per cent in children treated with fundoplication7.
The reasons why gastrostomy has such effects on survival
may be questioned when a tracheostomy, which intuitively
would seem a more hazardous procedure, confers no excess
risk. It seems likely that the underlying disease process that
necessitates gastrostomy leads to early death.
Previous studies of survival following fundoplication
are unsatisfactory, because of either a lack of actuarial
analysis, or a failure to assess the effects of other variables
such as gastrostomy. Mortality rates seem unrealistic.
The 20·0 per cent mortality rate in the present study can
be compared with the published values of 0 per cent19,
0·8 per cent2, 6·7 per cent6 and 9 per cent20. One study
describing poorer survival than in the present cohort
reported a 51 per centmortality rate5, but had a high rate of
operative complications leading to death; nine (26 per cent)
of 35 children had died within a month of surgery. The
corresponding value in the present study for death within
30 days of surgery among children with cerebral palsy was
11 per cent (8 of 76).
The 59 per cent 5-year survival rate of children with
cerebral palsy and gastrostomy undergoing fundoplication
is notably worse than the approximately 90 per cent
survival recently reported from a multi-institutional
American study utilizing data from an administrative
database7, or the 83 per cent survival rate reported
 2011 British Journal of Surgery Society Ltd www.bjs.co.uk British Journal of Surgery
Published by John Wiley & Sons Ltd

Survival of children following Nissen fundoplication
from a Canadian study21, possibly reflecting incomplete
ascertainment of deaths.
Survival of children with cerebral palsy in this region
has been described by Hutton and colleagues9; the survival
of such children undergoing fundoplication was much
worse than that of children with severe disability recorded
on a whole-population register of children with cerebral
palsy. Whereas Hutton and co-workers found that at least
74 per cent of childrenwith severe disability survived to the
age of 20 years, in the present study less than 30 per cent
were alive by this time point. A study from Mersey region
suggested that only 51 per cent of children with severe
disability are alive by age 20 years8; although closer to the
results of the present investigation, this still represents a
major difference. The present study most closely resembles
the 1998 report of Strauss et al.10, who noted the severe
effects of feeding problems on the survival of children
with cerebral palsy in California. Californian children
with a gastrostomy had a median survival of only 7 years,
similar to the life expectancy found in the present study
for children with cerebral palsy undergoing fundoplication
and receiving a gastrostomy.
Mortality after this common operation is entirely related
the presence of risk factors, principally gastrostomy and
cerebral palsy; female sex was also influential. There were
no deaths in the 83 children who had no risk factors;
normal children do not die after fundoplication.
The short survival described may have implications for
legal settlements based on the predicted life expectancy of
neurologically impaired children.
The principal variables predicting early mortality are
gastrostomy and cerebral palsy. Nutrition is clearly part
of the problem. Children who died weighed significantly
less than survivors, although both were below average in
weight. Whether more attention to nutritional issues in
such children can alter survival should be the subject of
research.
The common practice of gastrostomy tube insertion
in neurologically impaired children who are failing to
thrive seems to mark out the group with the poorest
survival. Whether to insert a gastrostomy tube at the time
of fundoplication in a neurologically impaired child may
be a difficult decision. The present study did not address
the alternative possibility of avoiding fundoplication by
insertion of a gastrojejunal tube7, but it can be speculated
that a gastrojejunal tube would have the same effect on
mortality as a gastrostomy tube, in that it does not alter
underlying disease processes.
This study has shown that some children with
cerebral palsy and profound feeding difficulties undergoing
gastrostomy tube insertion and fundoplication have a
severely limited life expectancy that may not be extended
by antireflux surgery. Medical professionals and parents
need to appreciate the severe consequences of the child’s
disabilities. Surgeons should be aware that some children
may be close to their predicted time of death when
approaching surgery, and are likely to die within 30 days
of operative intervention.
Acknowledgements
The authors thank Professor A. F. Colver for assistance in
cross-referencing children with the diagnosis of cerebral
palsy with the North of England Collaborative Cerebral
Palsy Study, and for comments on the manuscript; Mr J.
L. Hamilton and Dr C. Wren for facilitating use of the
regional database of children undergoing cardiac surgery;
Mary Bythell of the Northern Congenital Anomaly
Survey for cross-referencing the database; and Dr M. G.
Coulthard, Professor S. M. Griffin, Dr S. K. Bunn and Mr
M. Gatt for helpful suggestions. The authors declare no
conflict of interest.
References
1 Goldin AB, Garrison M, Christakis D. Variations between
hospitals in antireflux procedures in children. Arch Pediatr
Adolesc Med 2009; 163: 658–663.
2 Fonkalsrud EW, Ashcraft KW, Coran AG, Ellis DG,
Grosfeld JL, Tunell WP et al. Surgical treatment of
gastroesophageal reflux in children: a combined hospital
study of 7467 patients. Pediatrics 1998; 101: 419–422.
3 Srivastava R, Berry JG, Hall M, Downey EC, O’Gorman M,
Dean JM et al. Reflux related hospital admissions after
fundoplication in children with neurological impairment:
retrospective cohort study. Br Med J 2009; 339: b4411.
4 Lee SL, Sydorak RM, Chiu VY, Hsu JW, Applebaum H,
Haigh PI. Long-term antireflux medication use following
pediatric Nissen fundoplication. Arch Surg 2008; 143:
873–876.
5 Smith CD, Othersen HB Jr, Gogan NJ, Walker JD. Nissen
fundoplication in children with profound neurologic
disability. High risks and unmet goals. Ann Surg 1991; 215:
654–658.
6 Tovar JA, Luis AL, Encinas JL, Burgos L, Pederiva F,
Martinez L et al. Pediatric surgeons and gastroesophageal
reflux. J Pediatr Surg 2007; 42: 277–283.
7 Srivastava R, Downey EC, O’Gorman M, Feola P,
Samore M, Holubkov R et al. Impact of fundoplication versus
gastrojejunal feeding tubes on mortality and in preventing
aspiration pneumonia in young children with neurologic
impairment who have gastroesophageal reflux disease.
Pediatrics 2009; 123: 338–345.
 2011 British Journal of Surgery Society Ltd www.bjs.co.uk British Journal of Surgery
Published by John Wiley & Sons Ltd

R. Wockenforth, C. S. Gillespie and B. Jaffray
8 Hutton JL, Pharoah PO. Effects of cognitive, motor, and
sensory disabilities on survival in cerebral palsy. Arch Dis
Child 2002; 86: 84–89.
9 Hutton JL, Colver AF, Mackie PC. Effect of severity of
disability on survival in north east England cerebral palsy
cohort. Arch Dis Child 2000; 83: 468–474.
10 Strauss DJ, Shavelle RM, Anderson TW. Life expectancy of
children with cerebral palsy. Pediatr Neurol 1998; 18:
143–149.
11 Strauss D, Shavelle R. Life expectancy in cerebral palsy. Arch
Dis Child 2001; 85: 442–442.
12 R Core Development Team. R: A Language and Environment
for Statistical Computing. http://www.R-project.org [accessed
10 January 2009].
13 Wickham H. ggplot2: Elegant Graphs for Data Analysis.
Springer: New York, 2009.
14 Child Growth Foundation UK. The British 1990 Growth
Reference. Revised September 1996. Child Growth Foundation:
London, 1996.
15 Rosenfeld M, Davis R, FitzSimmons S, Pepe M, Ramsey B.
Gender gap in cystic fibrosis mortality. Am J Epidemiol 1997;
145: 794–803.
16 Strauss D, Shavelle R, Reynolds R, Rosenbloom L,
Day S. Survival in cerebral palsy in the last 20 years:
signs of improvement? Dev Med Child Neurol 2007; 49:
86–92.
17 Sleigh G, Brocklehurst P. Gastrostomy feeding in cerebral
palsy: a systematic review. Arch Dis Child 2004; 89: 534–539.
18 Catto-Smith AG, Jimenez S. Morbidity and mortality after
percutaneous endoscopic gastrostomy in children with
neurological disability. J Gastroenterol Hepatol 2006; 21:
734–738.
19 Albanese CT, Towbin RB, Ulman I, Lewis J, Smith SD.
Percutaneous gastrojejunostomy versus Nissen fundoplication
for enteral feeding of the neurologically impaired child with
gastroesophageal reflux. J Pediatr 1993; 123: 371–375.
20 Pearl RH, Robie DK, Ein SH, Shandling B, Wesson DE,
Superina R et al. Complications of gastroesophageal
antireflux surgery in neurologically impaired versus
neurologically normal children. J Pediatr Surg 1990; 25:
1169–1173.
21 Wales PW, Diamond IR, Dutta S, Muraca S, Chait P,
Connolly B et al. Fundoplication and gastrostomy versus
image-guided gastrojejunal tube for enteral feeding in
neurologically impaired children with gastroesophageal
reflux. J Pediatr Surg 2002; 37: 407–411.
 2011 British Journal of Surgery Society Ltd www.bjs.co.uk British Journal of Surgery
Published by John Wiley & Sons Ltd