Pediatr Surg Int DOI 10.1007/s00383-015-3707-y
ORIGINAL ARTICLE
Percutaneous endoscopic gastrostomy (PEG) does not worsen vomiting in children Madhavi Kakade1,2 • David Coyle1,2 • Dermot T. McDowell1,2 • John Gillick1,2
Accepted: 7 April 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose We aimed to evaluate the rate and examine potential predictors of subsequent anti-reflux procedures in a population undergoing percutaneous endoscopic gastrostomy (PEG) insertion. Materials We retrospectively reviewed the pre- and postoperative clinical course of patients undergoing PEG insertion over a 10-year period with respect to indication, underlying co-morbidity, and GER investigation and management. Results We reviewed data on 170 patients. Neurological disability (e.g., cerebral palsy) was the most common underlying condition in those undergoing PEG insertion (n = 104) followed by cystic fibrosis (n = 29). Oropharyngeal dysphagia and failure to thrive were the commonest indications for PEG. Eight patients (4.7 %) reported increased frequency of vomiting after PEG, 6 (75 %) of whom had a pre-operative diagnosis of GER. Two (25 %) patients from this sub-group subsequently required anti-reflux surgery. Patient’s with neurological disease were not at increased risk of new-onset GER or increased vomiting following PEG insertion compared to those with non-neurological conditions (p = 0.259). In total, 8 (4.7 %) and 7 (4.1 %) patients underwent fundoplication and gastrojejunal tube insertion, respectively. Conclusions PEG insertion does not appear to induce symptomatic gastro-oesophageal reflux in the majority of & Madhavi Kakade [email protected] 1
Department of Pediatric Surgery, Temple Street Children’s University Hospital, Dublin 1, Ireland
2
Department of Pediatric Surgery, Our Lady’s Children’s Hospital, Crumlin, Dublin 12, Ireland
children, suggesting that in the majority of cases, a concurrent anti-reflux procedure is unnecessary. Parents should be counseled accordingly. Keywords Percutaneous endoscopic gastrostomy Vomiting Anti-reflux surgery Gastro-esophageal reflux Fundoplication
Introduction Percutaneous endoscopic gastrostomy (PEG) tube placement is commonly undertaken in children for a variety of indications, most commonly oropharyngeal dysphagia and failure to thrive in those with medical co-morbidity [1, 2]. Where feasible, it has been widely accepted as a safe and acceptable minimally invasive alternative to open gastrostomy [3, 4]. Children undergoing gastrostomy placement frequently suffer from medical conditions already associated with a high incidence of gastro-esophageal reflux (GER), such as neurological impairment and cystic fibrosis. The induction or exacerbation of GER has historically been a primary concern with gastrostomy placement. Therefore, a simultaneous anti-reflux procedure such as a Nissen’s or Thal’s fundoplication has traditionally been performed at the time of gastrostomy placement in those considered to be at risk for problematic GER, necessitating either an open or laparoscopic approach [5, 6]. Evidence supporting this practice has been contentious. It has been demonstrated that the decision to perform simultaneous fundoplication was rarely informed by definitive medical testing or medical treatment response, and was more frequently influenced by physician and surgeon preference [5]. Patients with neurological impairment, who comprise the majority of candidates for PEG placement, have been
123
Pediatr Surg Int
purported to be the group most at risk of requiring subsequent fundoplication [7]. Recurrence rates of GER after fundoplication, however, are high—up to 30 % at 12 months—and the addition of a fundoplication at the time of gastrostomy placement does not appear to reduce the rate of hospitalization secondary to reflux-associated illness [8, 9]. Conversion of a gastrostomy into a gastro-jejunal tube for feeding is an alternative anti-reflux measure to fundoplication. Its effectiveness is unclear and it is associated with frequent hospital visits and potentially with jejunal perforation [10]. It follows that a useful predictor of those most likely to benefit from anti-reflux surgery in association with gastrostomy placement would be beneficial. The use of both oesophageal pH studies and an upper GI contrast series preoperatively has been described but may not prove especially useful in determining the need for a concomitant anti-reflux procedure [11, 12]. We aimed to evaluate the patient characteristics pre-disposing to the need for a subsequent anti-reflux procedure fundoplication for GER after PEG placement and to determine the impact, if any, pre-operative upper GI contrast studies had on predicting outcome after PEG placement.
Methods We performed a retrospective review of the medical records of 176 patients, selected on the basis of consecutive hospital medical record number, who underwent gastrostomy insertion in two pediatric tertiary referral hospitals in Dublin over a period from 2000 to 2010. We excluded all those whose gastrostomy insertions were not carried out endoscopically (n = 6). We recorded basic patient demographics (gender, age at PEG insertion), the underlying diagnosis and the clinical indication for PEG feeding (e.g. oropharyngeal
Table 1 Primary underlying medical conditions and clinical indications for PEG insertion for study population
dysphagia). Where data was available, we also recorded the presence of documented pre-existing clinical or radiological evidence of GER, including the findings of pre-operative upper GI contrast series, where performed. We examined the post-operative follow-up medical records with reference to post-operative complications, worsening clinical signs of GER and vomiting, the necessity for an anti-reflux procedure or gastro-jejunal (GJ) tube and the impact of any such procedure on symptom control. Statistical analysis was carried out using a standard statistical software package (Statistical Package for Social Sciences [SPSS] v21.0).
Results Population characteristics A total of 170 patients were included in this study, of which 90 were males and 80 were females. The median age at the time of PEG insertion was 3.01 years (range 0–19.3 years). The commonest indications for insertion of PEG were inadequate caloric intake causing failure to thrive in 87 cases (51.2 %), oropharyngeal dysphagia in 26 (15.3 %) or a combination of the two (25 cases, 14.7 %). Other indications included delivery of essential medications, oro-aversion, and gastric decompression. Forty-eight (28.2 %) children had more than one indication for insertion of PEG. Children with primary neurological conditions, in particular cerebral palsy, were the most common group to undergo PEG insertion by diagnosis, followed by those with cystic fibrosis and those with metabolic disorders. The underlying diagnosis for all children in this study is listed in Table 1.
Underlying diagnosis
n = 170 (%)
Neurological conditions (incl. cerebral palsy)
104 (61.2 %)
Cystic Fibrosis (CF)
29 (17.1 %)
Metabolic disorders
13 (7.6 %)
Congenital cardiac disease Craniofacial and airway disorder Miscellaneous causes
5 (2.9 %) 4 (2.4 %) 15 (8.8 %)
Clinical indication Insufficient caloric intake with failure to thrive (FTT) Oropharyngeal dysphagia (unsafe swallow)
87 (51.2 %) 26 (15.3 %)
Combination of FTT and oropharyngeal dysphagia
25 (14.7 %)
Oral aversion
2 (1.2 %)
Others
2 (1.2 %)
Combination of above indications
123
4 (2.4 %)
Delivery of essential medication
24 (14.1 %)
Pediatr Surg Int Table 2 Representation of anti-reflux measures in place prior to PEG insertion
Vomiting and GER after PEG insertion
Anti-reflux treatment
n = 170 (%)
Proton pump inhibitors
35 (20.6 %)
There was incomplete follow-up regarding the pattern of vomiting following PEG insertion in 15 cases. Regarding those with pre-operative GER, the vomiting pattern of 70 (92.1 %) patients was noted to be unchanged or improved at follow-up, while 6 (7.9 %) noted a worse pattern of vomiting (Fig. 1). In those with no evidence of GER preoperatively, symptoms developed following PEG insertion in 2 cases (2.6 %), with the remainder noting no change or an improvement in the frequency and pattern of vomiting observed. Patients with neurological disorders were not found to be at higher risk of deteriorating symptoms of GER after PEG insertion compared to a pooled cohort of patients who did not have neurological disorders (OR: 0.232, p = 0.259). Of the 8 patients whose vomiting and symptoms of GER worsened post-operatively, seven had neurological disorders, one of whom had no pre-operative GOR, and one had a metabolic disorder and also had no pre-operative history of GOR. All required PEG due to either inadequate caloric intake causing failure to thrive, oropharyngeal dysphagia causing unsafe swallow, or a combination of both. Four were males and four were females. Two of this sub-group of patients proceeded to undergo an additional anti-reflux procedure (one Nissen’s fundoplication, one gastro-jejunal tube insertion).
H2 receptor antagonist
24 (14.1 %)
Combination anti-acid with anti-reflux suspension
20 (11.8 %)
Combination anti-acids
20 (11.8 %)
Anti-reflux suspension alone (e.g. GavisconÒ)
4 (2.4 %)
Conservative (e.g., positioning) or no treatment
57 (33.5 %)
Incomplete records
10 (5.9 %)
Anti-acids include H2 receptor antagonists and proton pump inhibitors
Seventy-five patients (41 %) had a pre-operative upper gastrointestinal contrast series during their medical workup in advance of surgery. Clinical or radiological evidence of gastro-esophageal reflux (GER) was present in 80 patients (48.2 %). GER was more common pre-operatively in patients with metabolic disorders (8 patients [61.5 %]) and in patients with neurological disorders (55 patients [52.9 %]) compared to those with cystic fibrosis (5 patients [17.2 %], p = 0.01 and p = 0.001, respectively). Aspiration was a common clinical feature in our cohort of patients pre-operatively, being evident in 73 patients (42.9 %), 55 (75.3 %) of whom had a neurological disorder. Most children were on some form of anti-reflux or antiacid measures pre-operatively, such as anti-reflux suspension (e.g., GavisconÒ), H2 receptor antagonists, Proton pump inhibitors or a combination of the three, while 57 children were not on any anti-reflux or anti-acid medical treatment, including 7, who were managed with positioning alone. Further information regarding the medical management of these patients is detailed in Table 2. Post-operative complications Forty-one patients were recorded to have experienced an early post-operative complication. The most common was surgical site infection (n = 15), followed by PEG site over-granulation (n = 11), tube dislodgement (n = 3), gastrostomy site leak (n = 3) and pulmonary complications of surgery (n = 2) and a variety of other causes including dental injury, unspecified febrile illnesses and diarrhea. One patient developed peritonitis in the setting of a peritoneal dialysis catheter post-operatively. Late complications occurred in 19 patients. Tube dislodgement was the most common (n = 8), followed by tube fracture (n = 3), PEG site infection (n = 3), and PEG site over-granulation (n = 3). One patient each developed a gastro-colic and gastro-cutaneous fistulas as a result of PEG.
Anti-reflux surgery In total, eight patients underwent Nissen’s fundoplication at follow-up. This group consisted of four males and four females. Four patients had an underlying neurological disorder, 1 had cystic fibrosis, 1 had an underlying cardiac disorder, 1 was post-repair of a congenital diaphragmatic hernia, and one had trisomy 21 and was known to have a peptic oesophageal stricture. Fundoplication was carried out a median of 9 months following PEG insertion. Gastro-jejunal tubes were inserted in seven patients (four males, three females) a median of 9 months after PEG insertion. The primary diagnosis in five of these cases was a neurological disorder, one had a metabolic disorder and one had an oesophageal stricture. None of these patients had undergone a previous fundoplication.
Discussion Since the first description of the percutaneous endoscopic technique of gastrostomy placement in 1980, PEG has become the standard way to provide long-term enteral nutrition in children, particularly those with neurological
123
Pediatr Surg Int
Fig. 1 Flow chart of vomiting pattern & need for Anti reflux procedure
disability. For the two decades after its inception, PEG was documented in multiple studies as being associated with abnormal pH studies and clinically significant GER in large proportions of patients, especially those with neurological disability [13–16]. In the adult population, the lower esophageal sphincter relaxation triggered by gastric distension during PEG feeding was identified as the probable cause for GER after PEG insertion [17]. Prior to this Seekri et al. [18] demonstrated, using scintigraphy, upper GI contrast studies and pH studies, that PEG placement along the lesser curvature was associated with a lower incidence of GER. In addition to this, correctly adjusting feeding volumes and good medical management of GER have been suggested as useful measures in obviating the need for an anti-reflux procedure [19]. Cystic fibrosis cohort of patients figured particularly strongly in our study, this does not frequently feature in many other studies on this subject. It has previously been demonstrated by Sheikh et al. [20] that controlling reflux in this population improves pulmonary and nutritional outcomes in these patients. In addition, they demonstrated that patients who were fed by gastrostomy tube for 6 months
123
prior to undergoing anti-reflux surgery benefited the most from their procedure with regards to these outcomes compared to those who were gastrostomy fed for less than 6 months [20]. This group fared notably well in our study, with only one going onto require a fundoplication. In recent years, there has been a paradigm shift in the management of GER in the context of gastrostomy placement. A systematic review by Noble et al. [21] in 2012 found that although the evidence concerning this area is inconsistent, the effect of PEG insertion on GER does not appear to be causal in nature. A Cochrane Collaboration systematic review failed to identify a single trial satisfying its criteria for analysis [22]. A large retrospective analysis of 684 patients treated with gastrostomy insertion, of whom 278 underwent PEG, concluded that, with a fundoplication rate of 9.1 % (twice that seen in our series), fundoplication should not be performed empirically in the absence of complicated reflux [7]. Of note, they also identified those with neurological disability, namely cerebral palsy and anoxic brain injury, as the most at risk group that may require an anti-reflux procedure [7]. In our series, patients with neurological disability accounted for 50 % of those
Pediatr Surg Int
undergoing subsequent fundoplication and approximately 70 % of those undergoing GJ tube placement, which is actually representative of the proportion of the population they comprised. The heterogeneous nature of our study population compared to other studies may account for this. The use of an upper GI contrast series in pre-operative planning for PEG insertion is controversial. As with any medical investigation, the question of what impact an upper GI contrast series has on pre-operative planning needs to be asked. Its use to identify GER is questionable. The identification of GER in children with or without neurological impairment does not necessarily preclude a good clinical response to PEG feeding. This has been previously examined with multiple intra-luminal impedance monitoring/PH studies for a 24-h period following PEG placement. The authors did not identify any worsening of GER following PEG placement [23]. In our cohort, upper GI contrast studies were undertaken on 41 % of patients. Of the eight patients who subsequently required a Nissen’s fundoplication, only one did not have a pre-operative contrast study. Thus, the findings of the contrast study alone did not aid decision-making in determining the need for fundoplication. The utility of a pre-operative contrast study may be in determining gastric emptying and assessing for the presence of a downstream pathology such as an asymptomatic midgut malrotation, which would alter the operative plan. In general, we found that the pattern of vomiting in patients after PEG insertion was largely unchanged in those with a pre-operative history of GER, although in the minority (*8 %) who did complain of worsened vomiting, medical management was effective in controlling symptoms in all but one. Conversely, those who proceeded to fundoplication were largely those, whose GER either improved or remained unchanged after PEG insertion. Newonset GER was only found in two cases following PEG insertion, with medical management sufficing in both cases to control symptoms. In summary, pre-operative upper gastrointestinal contrast studies are not useful in determining the need for a concomitant anti-reflux operation in children undergoing PEG insertion. Only a small proportion of patients with preoperative GER undergoing PEG insertion will require a fundoplication at follow-up. While patients with neurological disability comprise the majority of such patients, they are not disproportionately represented in this subgroup. We conclude that the concomitant use of an antireflux operation at the time of PEG insertion is rarely necessary, and then only in those with a pre-operative history of GER. A trial period on PEG feeds will allow better patient selection for anti-reflux surgery and improve the nutritional status of those patients prior to a more major procedure.
Acknowledgments The authors declare no conflict of interest. Due to the retrospective nature of the study ethical approval was not sought on the basis of local hospital guidelines. No sources of funding were utilized in the carrying out of this research.
References 1. Van Biervliet S, Van Renterghem K, Vande Putte D, Vande Velde S, De Bruyne R, Van Winckel M (2014) Gastrostomy use in children: a 3-year single centre experience. Acta Gastroenterol Belg 77(1):8–12 2. Minar P, Garland J, Martinez A, Werlin S (2011) Safety of percutaneous endoscopic gastrostomy in medically complicated infants. J Pediatr Gastroenterol Nutr 53(3):293–295. doi:10.1097/ MPG.0b013e318215c41b 3. Saitua F, Acuna R, Herrera P (2003) Percutaneous endoscopic gastrostomy: the technique of choice? J Pediatr Surg 38(10):1512–1515 4. Akay B, Capizzani TR, Lee AM, Drongowski RA, Geiger JD, Hirschl RB, Mychaliska GB (2010) Gastrostomy tube placement in infants and children: is there a preferred technique? J Pediatr Surg 45(6):1147–1152. doi:10.1016/j.jpedsurg.2010.02.079 5. Fox D, Barnard J, Campagna EJ, Dickinson LM, Bruny J, Kempe A (2012) Fundoplication and the pediatric surgeon: implications for shared decision-making and the medical home. Acad Pediatr 12(6):558–566. doi:10.1016/j.acap.2012.07.006 6. Toporowska-Kowalska E, Gebora-Kowalska B, Fendler W, Popinska K, Szlagatys-Sidorkiewicz A, Grzybowska-Chlebowczyk U, Wiernicka A, Borkowska A, Sibilska M, Wiecek S, Hapyn E, Kierkus J (2013) Diagnosis of gastroesophageal reflux and anti-reflux procedures among Polish children with gastrostomies: a 10-year nationwide analysis. Eur J Clin Nutr 67(11):1169–1174. doi:10.1038/ejcn.2013.164 7. Ponsky TA, Gasior AC, Parry J, Sharp SW, Boulanger S, Parry R, Ostlie DJ, St Peter SD (2013) Need for subsequent fundoplication after gastrostomy based on patient characteristics. J Surg Res 179(1):1–4. doi:10.1016/j.jss.2012.03.064 8. Ferluga ED, Sathe NA, Krishnaswami S, McPheeters ML (2014) Surgical intervention for feeding and nutrition difficulties in cerebral palsy: a systematic review. Dev Med Child Neurol 56(1):31–43. doi:10.1111/dmcn.12170 9. Barnhart DC, Hall M, Mahant S, Goldin AB, Berry JG, Faix RG, Dean JM, Srivastava R (2013) Effectiveness of fundoplication at the time of gastrostomy in infants with neurological impairment. JAMA Pediatr 167(10):911–918. doi:10.1001/jamapediatrics. 2013.334 10. King M, Barnhart DC, O’Gorman M, Downey EC, Jackson D, Mundorff M, Holubkov R, Feola P, Srivastava R (2014) Effect of gastrojejunal feedings on visits and costs in children with neurologic impairment. J Pediatr Gastroenterol Nutr 58(4):518–524. doi:10.1097/MPG.0000000000000227 11. Wilson GJ, van der Zee DC, Bax NM (2006) Endoscopic gastrostomy placement in the child with gastroesophageal reflux: is concomitant antireflux surgery indicated? J Pediatr Surg 41(8):1441–1445. doi:10.1016/j.jpedsurg.2006.04.021 12. Gauderer MWL (2006) Gastrostomy. In: Puri P, Hollwarth M (eds) Pediatric Surgery. Springer, Berlin, Heidelberg, pp 181–196 13. Grunow JE, al-Hafidh A, Tunell WP (1989) Gastroesophageal reflux following percutaneous endoscopic gastrostomy in children. J Pediatr Surg 24(1):42–44 Discussion 44–45 14. Berezin S, Schwarz SM, Halata MS, Newman LJ (1986) Gastroesophageal reflux secondary to gastrostomy tube placement. Am J Dis Child 140(7):699–701
123
Pediatr Surg Int 15. Jolley SG, Smith EI, Tunell WP (1985) Protective antireflux operation with feeding gastrostomy. Experience with children. Ann Surg 201(6):736–740 16. Mollitt DL, Golladay ES, Seibert JJ (1985) Symptomatic gastroesophageal reflux following gastrostomy in neurologically impaired patients. Pediatrics 75(6):1124–1126 17. Coben RM, Weintraub A, DiMarino AJ Jr, Cohen S (1994) Gastroesophageal reflux during gastrostomy feeding. Gastroenterology 106(1):13–18 18. Seekri IK, Rescorla FJ, Canal DF, Zollinger TW, Saywell R Jr, Grosfeld JL (1991) Lesser curvature gastrostomy reduces the incidence of postoperative gastroesophageal reflux. J Pediatr Surg 26(8):982–984 discussion 984–985 19. Gauderer MW (1988) Feeding gastrostomy or feeding gastrostomy plus antireflux procedure? J Pediatr Gastroenterol Nutr 7(6):795–796 20. Sheikh SI, Ryan-Wenger NA, McCoy KS (2013) Outcomes of surgical management of severe GERD in patients with cystic
123
fibrosis. Pediatr Pulmonol 48(6):556–562. doi:10.1002/ppul. 22630 21. Noble LJ, Dalzell AM, El-Matary W (2012) The relationship between percutaneous endoscopic gastrostomy and gastro-oesophageal reflux disease in children: a systematic review. Surg Endosc 26(9):2504–2512. doi:10.1007/s00464-012-2221-8 22. Vernon-Roberts A, Sullivan PB (2013) Fundoplication versus postoperative medication for gastro-oesophageal reflux in children with neurological impairment undergoing gastrostomy. Cochrane Database Syst Rev 8:CD006151. doi:10.1002/ 14651858.CD006151.pub3 23. Toporowska-Kowalska E, Gebora-Kowalska B, Jablonski J, Fendler W, Wasowska-Krolikowska K (2011) Influence of percutaneous endoscopic gastrostomy on gastro-oesophageal reflux evaluated by multiple intraluminal impedance in children with neurological impairment. Dev Med Child Neurol 53(10):938–943. doi:10.1111/j.1469-8749.2011.04031.x
ORIGINAL ARTICLE
Percutaneous endoscopic gastrostomy (PEG) does not worsen vomiting in children Madhavi Kakade1,2 • David Coyle1,2 • Dermot T. McDowell1,2 • John Gillick1,2
Accepted: 7 April 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose We aimed to evaluate the rate and examine potential predictors of subsequent anti-reflux procedures in a population undergoing percutaneous endoscopic gastrostomy (PEG) insertion. Materials We retrospectively reviewed the pre- and postoperative clinical course of patients undergoing PEG insertion over a 10-year period with respect to indication, underlying co-morbidity, and GER investigation and management. Results We reviewed data on 170 patients. Neurological disability (e.g., cerebral palsy) was the most common underlying condition in those undergoing PEG insertion (n = 104) followed by cystic fibrosis (n = 29). Oropharyngeal dysphagia and failure to thrive were the commonest indications for PEG. Eight patients (4.7 %) reported increased frequency of vomiting after PEG, 6 (75 %) of whom had a pre-operative diagnosis of GER. Two (25 %) patients from this sub-group subsequently required anti-reflux surgery. Patient’s with neurological disease were not at increased risk of new-onset GER or increased vomiting following PEG insertion compared to those with non-neurological conditions (p = 0.259). In total, 8 (4.7 %) and 7 (4.1 %) patients underwent fundoplication and gastrojejunal tube insertion, respectively. Conclusions PEG insertion does not appear to induce symptomatic gastro-oesophageal reflux in the majority of & Madhavi Kakade [email protected] 1
Department of Pediatric Surgery, Temple Street Children’s University Hospital, Dublin 1, Ireland
2
Department of Pediatric Surgery, Our Lady’s Children’s Hospital, Crumlin, Dublin 12, Ireland
children, suggesting that in the majority of cases, a concurrent anti-reflux procedure is unnecessary. Parents should be counseled accordingly. Keywords Percutaneous endoscopic gastrostomy Vomiting Anti-reflux surgery Gastro-esophageal reflux Fundoplication
Introduction Percutaneous endoscopic gastrostomy (PEG) tube placement is commonly undertaken in children for a variety of indications, most commonly oropharyngeal dysphagia and failure to thrive in those with medical co-morbidity [1, 2]. Where feasible, it has been widely accepted as a safe and acceptable minimally invasive alternative to open gastrostomy [3, 4]. Children undergoing gastrostomy placement frequently suffer from medical conditions already associated with a high incidence of gastro-esophageal reflux (GER), such as neurological impairment and cystic fibrosis. The induction or exacerbation of GER has historically been a primary concern with gastrostomy placement. Therefore, a simultaneous anti-reflux procedure such as a Nissen’s or Thal’s fundoplication has traditionally been performed at the time of gastrostomy placement in those considered to be at risk for problematic GER, necessitating either an open or laparoscopic approach [5, 6]. Evidence supporting this practice has been contentious. It has been demonstrated that the decision to perform simultaneous fundoplication was rarely informed by definitive medical testing or medical treatment response, and was more frequently influenced by physician and surgeon preference [5]. Patients with neurological impairment, who comprise the majority of candidates for PEG placement, have been
123
Pediatr Surg Int
purported to be the group most at risk of requiring subsequent fundoplication [7]. Recurrence rates of GER after fundoplication, however, are high—up to 30 % at 12 months—and the addition of a fundoplication at the time of gastrostomy placement does not appear to reduce the rate of hospitalization secondary to reflux-associated illness [8, 9]. Conversion of a gastrostomy into a gastro-jejunal tube for feeding is an alternative anti-reflux measure to fundoplication. Its effectiveness is unclear and it is associated with frequent hospital visits and potentially with jejunal perforation [10]. It follows that a useful predictor of those most likely to benefit from anti-reflux surgery in association with gastrostomy placement would be beneficial. The use of both oesophageal pH studies and an upper GI contrast series preoperatively has been described but may not prove especially useful in determining the need for a concomitant anti-reflux procedure [11, 12]. We aimed to evaluate the patient characteristics pre-disposing to the need for a subsequent anti-reflux procedure fundoplication for GER after PEG placement and to determine the impact, if any, pre-operative upper GI contrast studies had on predicting outcome after PEG placement.
Methods We performed a retrospective review of the medical records of 176 patients, selected on the basis of consecutive hospital medical record number, who underwent gastrostomy insertion in two pediatric tertiary referral hospitals in Dublin over a period from 2000 to 2010. We excluded all those whose gastrostomy insertions were not carried out endoscopically (n = 6). We recorded basic patient demographics (gender, age at PEG insertion), the underlying diagnosis and the clinical indication for PEG feeding (e.g. oropharyngeal
Table 1 Primary underlying medical conditions and clinical indications for PEG insertion for study population
dysphagia). Where data was available, we also recorded the presence of documented pre-existing clinical or radiological evidence of GER, including the findings of pre-operative upper GI contrast series, where performed. We examined the post-operative follow-up medical records with reference to post-operative complications, worsening clinical signs of GER and vomiting, the necessity for an anti-reflux procedure or gastro-jejunal (GJ) tube and the impact of any such procedure on symptom control. Statistical analysis was carried out using a standard statistical software package (Statistical Package for Social Sciences [SPSS] v21.0).
Results Population characteristics A total of 170 patients were included in this study, of which 90 were males and 80 were females. The median age at the time of PEG insertion was 3.01 years (range 0–19.3 years). The commonest indications for insertion of PEG were inadequate caloric intake causing failure to thrive in 87 cases (51.2 %), oropharyngeal dysphagia in 26 (15.3 %) or a combination of the two (25 cases, 14.7 %). Other indications included delivery of essential medications, oro-aversion, and gastric decompression. Forty-eight (28.2 %) children had more than one indication for insertion of PEG. Children with primary neurological conditions, in particular cerebral palsy, were the most common group to undergo PEG insertion by diagnosis, followed by those with cystic fibrosis and those with metabolic disorders. The underlying diagnosis for all children in this study is listed in Table 1.
Underlying diagnosis
n = 170 (%)
Neurological conditions (incl. cerebral palsy)
104 (61.2 %)
Cystic Fibrosis (CF)
29 (17.1 %)
Metabolic disorders
13 (7.6 %)
Congenital cardiac disease Craniofacial and airway disorder Miscellaneous causes
5 (2.9 %) 4 (2.4 %) 15 (8.8 %)
Clinical indication Insufficient caloric intake with failure to thrive (FTT) Oropharyngeal dysphagia (unsafe swallow)
87 (51.2 %) 26 (15.3 %)
Combination of FTT and oropharyngeal dysphagia
25 (14.7 %)
Oral aversion
2 (1.2 %)
Others
2 (1.2 %)
Combination of above indications
123
4 (2.4 %)
Delivery of essential medication
24 (14.1 %)
Pediatr Surg Int Table 2 Representation of anti-reflux measures in place prior to PEG insertion
Vomiting and GER after PEG insertion
Anti-reflux treatment
n = 170 (%)
Proton pump inhibitors
35 (20.6 %)
There was incomplete follow-up regarding the pattern of vomiting following PEG insertion in 15 cases. Regarding those with pre-operative GER, the vomiting pattern of 70 (92.1 %) patients was noted to be unchanged or improved at follow-up, while 6 (7.9 %) noted a worse pattern of vomiting (Fig. 1). In those with no evidence of GER preoperatively, symptoms developed following PEG insertion in 2 cases (2.6 %), with the remainder noting no change or an improvement in the frequency and pattern of vomiting observed. Patients with neurological disorders were not found to be at higher risk of deteriorating symptoms of GER after PEG insertion compared to a pooled cohort of patients who did not have neurological disorders (OR: 0.232, p = 0.259). Of the 8 patients whose vomiting and symptoms of GER worsened post-operatively, seven had neurological disorders, one of whom had no pre-operative GOR, and one had a metabolic disorder and also had no pre-operative history of GOR. All required PEG due to either inadequate caloric intake causing failure to thrive, oropharyngeal dysphagia causing unsafe swallow, or a combination of both. Four were males and four were females. Two of this sub-group of patients proceeded to undergo an additional anti-reflux procedure (one Nissen’s fundoplication, one gastro-jejunal tube insertion).
H2 receptor antagonist
24 (14.1 %)
Combination anti-acid with anti-reflux suspension
20 (11.8 %)
Combination anti-acids
20 (11.8 %)
Anti-reflux suspension alone (e.g. GavisconÒ)
4 (2.4 %)
Conservative (e.g., positioning) or no treatment
57 (33.5 %)
Incomplete records
10 (5.9 %)
Anti-acids include H2 receptor antagonists and proton pump inhibitors
Seventy-five patients (41 %) had a pre-operative upper gastrointestinal contrast series during their medical workup in advance of surgery. Clinical or radiological evidence of gastro-esophageal reflux (GER) was present in 80 patients (48.2 %). GER was more common pre-operatively in patients with metabolic disorders (8 patients [61.5 %]) and in patients with neurological disorders (55 patients [52.9 %]) compared to those with cystic fibrosis (5 patients [17.2 %], p = 0.01 and p = 0.001, respectively). Aspiration was a common clinical feature in our cohort of patients pre-operatively, being evident in 73 patients (42.9 %), 55 (75.3 %) of whom had a neurological disorder. Most children were on some form of anti-reflux or antiacid measures pre-operatively, such as anti-reflux suspension (e.g., GavisconÒ), H2 receptor antagonists, Proton pump inhibitors or a combination of the three, while 57 children were not on any anti-reflux or anti-acid medical treatment, including 7, who were managed with positioning alone. Further information regarding the medical management of these patients is detailed in Table 2. Post-operative complications Forty-one patients were recorded to have experienced an early post-operative complication. The most common was surgical site infection (n = 15), followed by PEG site over-granulation (n = 11), tube dislodgement (n = 3), gastrostomy site leak (n = 3) and pulmonary complications of surgery (n = 2) and a variety of other causes including dental injury, unspecified febrile illnesses and diarrhea. One patient developed peritonitis in the setting of a peritoneal dialysis catheter post-operatively. Late complications occurred in 19 patients. Tube dislodgement was the most common (n = 8), followed by tube fracture (n = 3), PEG site infection (n = 3), and PEG site over-granulation (n = 3). One patient each developed a gastro-colic and gastro-cutaneous fistulas as a result of PEG.
Anti-reflux surgery In total, eight patients underwent Nissen’s fundoplication at follow-up. This group consisted of four males and four females. Four patients had an underlying neurological disorder, 1 had cystic fibrosis, 1 had an underlying cardiac disorder, 1 was post-repair of a congenital diaphragmatic hernia, and one had trisomy 21 and was known to have a peptic oesophageal stricture. Fundoplication was carried out a median of 9 months following PEG insertion. Gastro-jejunal tubes were inserted in seven patients (four males, three females) a median of 9 months after PEG insertion. The primary diagnosis in five of these cases was a neurological disorder, one had a metabolic disorder and one had an oesophageal stricture. None of these patients had undergone a previous fundoplication.
Discussion Since the first description of the percutaneous endoscopic technique of gastrostomy placement in 1980, PEG has become the standard way to provide long-term enteral nutrition in children, particularly those with neurological
123
Pediatr Surg Int
Fig. 1 Flow chart of vomiting pattern & need for Anti reflux procedure
disability. For the two decades after its inception, PEG was documented in multiple studies as being associated with abnormal pH studies and clinically significant GER in large proportions of patients, especially those with neurological disability [13–16]. In the adult population, the lower esophageal sphincter relaxation triggered by gastric distension during PEG feeding was identified as the probable cause for GER after PEG insertion [17]. Prior to this Seekri et al. [18] demonstrated, using scintigraphy, upper GI contrast studies and pH studies, that PEG placement along the lesser curvature was associated with a lower incidence of GER. In addition to this, correctly adjusting feeding volumes and good medical management of GER have been suggested as useful measures in obviating the need for an anti-reflux procedure [19]. Cystic fibrosis cohort of patients figured particularly strongly in our study, this does not frequently feature in many other studies on this subject. It has previously been demonstrated by Sheikh et al. [20] that controlling reflux in this population improves pulmonary and nutritional outcomes in these patients. In addition, they demonstrated that patients who were fed by gastrostomy tube for 6 months
123
prior to undergoing anti-reflux surgery benefited the most from their procedure with regards to these outcomes compared to those who were gastrostomy fed for less than 6 months [20]. This group fared notably well in our study, with only one going onto require a fundoplication. In recent years, there has been a paradigm shift in the management of GER in the context of gastrostomy placement. A systematic review by Noble et al. [21] in 2012 found that although the evidence concerning this area is inconsistent, the effect of PEG insertion on GER does not appear to be causal in nature. A Cochrane Collaboration systematic review failed to identify a single trial satisfying its criteria for analysis [22]. A large retrospective analysis of 684 patients treated with gastrostomy insertion, of whom 278 underwent PEG, concluded that, with a fundoplication rate of 9.1 % (twice that seen in our series), fundoplication should not be performed empirically in the absence of complicated reflux [7]. Of note, they also identified those with neurological disability, namely cerebral palsy and anoxic brain injury, as the most at risk group that may require an anti-reflux procedure [7]. In our series, patients with neurological disability accounted for 50 % of those
Pediatr Surg Int
undergoing subsequent fundoplication and approximately 70 % of those undergoing GJ tube placement, which is actually representative of the proportion of the population they comprised. The heterogeneous nature of our study population compared to other studies may account for this. The use of an upper GI contrast series in pre-operative planning for PEG insertion is controversial. As with any medical investigation, the question of what impact an upper GI contrast series has on pre-operative planning needs to be asked. Its use to identify GER is questionable. The identification of GER in children with or without neurological impairment does not necessarily preclude a good clinical response to PEG feeding. This has been previously examined with multiple intra-luminal impedance monitoring/PH studies for a 24-h period following PEG placement. The authors did not identify any worsening of GER following PEG placement [23]. In our cohort, upper GI contrast studies were undertaken on 41 % of patients. Of the eight patients who subsequently required a Nissen’s fundoplication, only one did not have a pre-operative contrast study. Thus, the findings of the contrast study alone did not aid decision-making in determining the need for fundoplication. The utility of a pre-operative contrast study may be in determining gastric emptying and assessing for the presence of a downstream pathology such as an asymptomatic midgut malrotation, which would alter the operative plan. In general, we found that the pattern of vomiting in patients after PEG insertion was largely unchanged in those with a pre-operative history of GER, although in the minority (*8 %) who did complain of worsened vomiting, medical management was effective in controlling symptoms in all but one. Conversely, those who proceeded to fundoplication were largely those, whose GER either improved or remained unchanged after PEG insertion. Newonset GER was only found in two cases following PEG insertion, with medical management sufficing in both cases to control symptoms. In summary, pre-operative upper gastrointestinal contrast studies are not useful in determining the need for a concomitant anti-reflux operation in children undergoing PEG insertion. Only a small proportion of patients with preoperative GER undergoing PEG insertion will require a fundoplication at follow-up. While patients with neurological disability comprise the majority of such patients, they are not disproportionately represented in this subgroup. We conclude that the concomitant use of an antireflux operation at the time of PEG insertion is rarely necessary, and then only in those with a pre-operative history of GER. A trial period on PEG feeds will allow better patient selection for anti-reflux surgery and improve the nutritional status of those patients prior to a more major procedure.
Acknowledgments The authors declare no conflict of interest. Due to the retrospective nature of the study ethical approval was not sought on the basis of local hospital guidelines. No sources of funding were utilized in the carrying out of this research.
References 1. Van Biervliet S, Van Renterghem K, Vande Putte D, Vande Velde S, De Bruyne R, Van Winckel M (2014) Gastrostomy use in children: a 3-year single centre experience. Acta Gastroenterol Belg 77(1):8–12 2. Minar P, Garland J, Martinez A, Werlin S (2011) Safety of percutaneous endoscopic gastrostomy in medically complicated infants. J Pediatr Gastroenterol Nutr 53(3):293–295. doi:10.1097/ MPG.0b013e318215c41b 3. Saitua F, Acuna R, Herrera P (2003) Percutaneous endoscopic gastrostomy: the technique of choice? J Pediatr Surg 38(10):1512–1515 4. Akay B, Capizzani TR, Lee AM, Drongowski RA, Geiger JD, Hirschl RB, Mychaliska GB (2010) Gastrostomy tube placement in infants and children: is there a preferred technique? J Pediatr Surg 45(6):1147–1152. doi:10.1016/j.jpedsurg.2010.02.079 5. Fox D, Barnard J, Campagna EJ, Dickinson LM, Bruny J, Kempe A (2012) Fundoplication and the pediatric surgeon: implications for shared decision-making and the medical home. Acad Pediatr 12(6):558–566. doi:10.1016/j.acap.2012.07.006 6. Toporowska-Kowalska E, Gebora-Kowalska B, Fendler W, Popinska K, Szlagatys-Sidorkiewicz A, Grzybowska-Chlebowczyk U, Wiernicka A, Borkowska A, Sibilska M, Wiecek S, Hapyn E, Kierkus J (2013) Diagnosis of gastroesophageal reflux and anti-reflux procedures among Polish children with gastrostomies: a 10-year nationwide analysis. Eur J Clin Nutr 67(11):1169–1174. doi:10.1038/ejcn.2013.164 7. Ponsky TA, Gasior AC, Parry J, Sharp SW, Boulanger S, Parry R, Ostlie DJ, St Peter SD (2013) Need for subsequent fundoplication after gastrostomy based on patient characteristics. J Surg Res 179(1):1–4. doi:10.1016/j.jss.2012.03.064 8. Ferluga ED, Sathe NA, Krishnaswami S, McPheeters ML (2014) Surgical intervention for feeding and nutrition difficulties in cerebral palsy: a systematic review. Dev Med Child Neurol 56(1):31–43. doi:10.1111/dmcn.12170 9. Barnhart DC, Hall M, Mahant S, Goldin AB, Berry JG, Faix RG, Dean JM, Srivastava R (2013) Effectiveness of fundoplication at the time of gastrostomy in infants with neurological impairment. JAMA Pediatr 167(10):911–918. doi:10.1001/jamapediatrics. 2013.334 10. King M, Barnhart DC, O’Gorman M, Downey EC, Jackson D, Mundorff M, Holubkov R, Feola P, Srivastava R (2014) Effect of gastrojejunal feedings on visits and costs in children with neurologic impairment. J Pediatr Gastroenterol Nutr 58(4):518–524. doi:10.1097/MPG.0000000000000227 11. Wilson GJ, van der Zee DC, Bax NM (2006) Endoscopic gastrostomy placement in the child with gastroesophageal reflux: is concomitant antireflux surgery indicated? J Pediatr Surg 41(8):1441–1445. doi:10.1016/j.jpedsurg.2006.04.021 12. Gauderer MWL (2006) Gastrostomy. In: Puri P, Hollwarth M (eds) Pediatric Surgery. Springer, Berlin, Heidelberg, pp 181–196 13. Grunow JE, al-Hafidh A, Tunell WP (1989) Gastroesophageal reflux following percutaneous endoscopic gastrostomy in children. J Pediatr Surg 24(1):42–44 Discussion 44–45 14. Berezin S, Schwarz SM, Halata MS, Newman LJ (1986) Gastroesophageal reflux secondary to gastrostomy tube placement. Am J Dis Child 140(7):699–701
123
Pediatr Surg Int 15. Jolley SG, Smith EI, Tunell WP (1985) Protective antireflux operation with feeding gastrostomy. Experience with children. Ann Surg 201(6):736–740 16. Mollitt DL, Golladay ES, Seibert JJ (1985) Symptomatic gastroesophageal reflux following gastrostomy in neurologically impaired patients. Pediatrics 75(6):1124–1126 17. Coben RM, Weintraub A, DiMarino AJ Jr, Cohen S (1994) Gastroesophageal reflux during gastrostomy feeding. Gastroenterology 106(1):13–18 18. Seekri IK, Rescorla FJ, Canal DF, Zollinger TW, Saywell R Jr, Grosfeld JL (1991) Lesser curvature gastrostomy reduces the incidence of postoperative gastroesophageal reflux. J Pediatr Surg 26(8):982–984 discussion 984–985 19. Gauderer MW (1988) Feeding gastrostomy or feeding gastrostomy plus antireflux procedure? J Pediatr Gastroenterol Nutr 7(6):795–796 20. Sheikh SI, Ryan-Wenger NA, McCoy KS (2013) Outcomes of surgical management of severe GERD in patients with cystic
123
fibrosis. Pediatr Pulmonol 48(6):556–562. doi:10.1002/ppul. 22630 21. Noble LJ, Dalzell AM, El-Matary W (2012) The relationship between percutaneous endoscopic gastrostomy and gastro-oesophageal reflux disease in children: a systematic review. Surg Endosc 26(9):2504–2512. doi:10.1007/s00464-012-2221-8 22. Vernon-Roberts A, Sullivan PB (2013) Fundoplication versus postoperative medication for gastro-oesophageal reflux in children with neurological impairment undergoing gastrostomy. Cochrane Database Syst Rev 8:CD006151. doi:10.1002/ 14651858.CD006151.pub3 23. Toporowska-Kowalska E, Gebora-Kowalska B, Jablonski J, Fendler W, Wasowska-Krolikowska K (2011) Influence of percutaneous endoscopic gastrostomy on gastro-oesophageal reflux evaluated by multiple intraluminal impedance in children with neurological impairment. Dev Med Child Neurol 53(10):938–943. doi:10.1111/j.1469-8749.2011.04031.x
![[Percutaneous Endoscopic Gastrostomy (PEG)].](/assets/img/hold.png)
