J Gastrointest Canc (2014) 45:452–459 DOI 10.1007/s12029-014-9630-y
ORIGINAL RESEARCH
Complications of Percutaneous Endoscopic Gastrostomy Tube Insertion in Cancer Patients: A Retrospective Study Hala Mansoor & Muhammad Adnan Masood & Muhammed Aasim Yusuf
Published online: 26 June 2014 # Springer Science+Business Media New York 2014
Abstract Aim Dysphagia, with associated malnutrition is a common feature in up to 64 % of patients with cancers of the head and neck region, the oesophagus and the gastro-oesophageal junction (Donaldson and Lenon Cancer 43(Suppl 5):2036–52, 1979). These patients usually require alternate routes of feeding during treatment. However, these methods of feeding are not without associated complications. This study was carried out to assess the complications associated with percutaneous endoscopic gastrostomy tube insertion, the commonest means of providing nutrition in this patient group. Methods Two hundred and sixty patients with successful percutaneous endoscopic gastrostomy (PEG) tube insertion were retrospectively reviewed to assess the complications of the procedure. Results The overall complication rate was 25 % (65 patients). Fifty-five patients (21 %) patients had minor complications. Ten patients (3.8 %) suffered a major complication, including peritonitis and major peri-PEG infection each in three patients (1.2 %), while oesophageal perforation, collection around the PEG insertion site requiring operative drainage, bleeding from the PEG insertion site and tumour seeding at the gastrostomy site were each seen in one (0.4 %) patient. Conclusion Percutaneous endoscopic gastrostomy tube insertion is a useful means of providing enteral nutrition to patients H. Mansoor (*) : M. A. Masood : M. A. Yusuf Shaukat Khanum Memorial Cancer Hospital & Research Centre, Lahore, Pakistan e-mail: [email protected] H. Mansoor e-mail: [email protected] M. A. Masood e-mail: [email protected] M. A. Yusuf e-mail: [email protected]
with swallowing problems. Patients with head and neck cancer who are likely to develop dysphagia should be assessed promptly for PEG tube insertion before the start of radiation. Early recognition and prompt treatment of serious complications can avoid potential fatality associated with PEG tubes. Keywords Percutaneous endoscopic gastrostomy (PEG) . Head and neck . Cancer of the oesophagus and gastro-oesophageal junction . Nutrition
Introduction Since the introduction of percutaneous endoscopic gastrostomy (PEG) by Gauderer and Ponsky in 1980 [1], it has become the most commonly employed technique of prolonged enteral feeding in patients with impaired swallowing. This includes patients with neurological causes of dysphagia as well as those with neoplasms of the oral cavity, oropharynx, hypopharynx, nasopharynx, larynx, oesophagus and gastro-oesophageal junction [2]. Most patients with these malignancies already have poor nutritional status at the time of diagnosis due to dysphagia, and as many as 64 % are malnourished [3]. Radical surgery, involving either partial or total glossectomy or mandibulectomy, may further impair swallowing [4]. Chemotherapy may lead to loss of appetite while radiotherapy causes a local radiation reaction that leads to poor oral intake [3, 5, 6]. Studies on cancer patients have suggested improved outcomes and fewer interruptions of treatment in patients who received nutritional supplementation [7–9]. Lees performed a prospective trial, comparing nasogastric feeding with PEG feeding in head and neck cancer patients undergoing radiotherapy and concluded that the final outcome of these patients depended mainly on the nutritional status achieved, regardless of the method of feeding [10]. While PEG feeding is generally preferred, as it is thought to
J Gastrointest Canc (2014) 45:452–459
cause less discomfort as well as for cosmetic reasons, PEG insertion is not entirely safe and carries its own risks, including a high rate of insertion-site infection in up to 41 % of patients and other less frequent, but more serious side-effects, including septicemia, gut perforation, peritonitis and aspiration pneumonia, among others [2, 11, 12]. However, PEG insertion remains a common indication for upper gastrointestinal tract endoscopy [13]. The various techniques used for PEG insertion include the pull (Ponsky), introducer (Russell) and push (Sachs-Vine) and the direct method [2]. Of these, Ponsky’s pull technique is the most frequently used method [14]. At our hospital, all patients who present with dysphagia due to cancer of the head and neck region, the pharynx, the oesophagus or the gastro-oesophageal junction, or in whom swallowing problems are anticipated during chemoradiotherapy, are assessed for PEG tube insertion. The aim of this study was to retrospectively analyse the data of our hospital patients to determine the outcome of PEG tube insertion during the year 2012 for all such patients at our institution.
453
standard upper GI endoscopy or oesophago-gastroduodenoscopy to evaluate the anatomy of the upper GI tract. The anterior wall of the stomach is identified, and an effort is made to ensure that there are no intervening viscerae between the stomach and the anterior abdominal wall. This is done by applying digital pressure on the anterior abdominal wall, which can be seen as an indentation of the stomach endoscopically as well as by trans-illumination of the abdominal wall, through the stomach. A failure to trans-illuminate is an absolute contraindication in attempting PEG tube insertion. Local anaesthetic is then injected at the site selected for gastric puncture, a trocar and cannula is used to puncture the stomach and a guide wire is inserted through the cannula into the gastric lumen. This is grasped endoscopically and pulled out through the mouth. A 20-Fr gastrostomy tube is then attached to the wire and pulled back down the oesophagus and into the stomach, from where it is pulled into place through the anterior abdominal wall. It is our practice to administer a single dose of prophylactic intravenous antibiotic immediately prior to the procedure. Data Collection
Materials and Methods Patients This retrospective study was conducted at Shaukat Khanum Memorial Cancer Hospital & Research Centre, Lahore, Pakistan, after obtaining approval from the Shaukat Khanum Memorial Cancer Hospital & Research Centre Institutional Review Board, in accordance with the principles of the Helsinki Declaration. The data of 260 consecutive patients (159 males, age range 18–81 years; mean age 50 years) who underwent PEG tube insertion from January 2012 till December 2012 was reviewed retrospectively. The majority of these patients had dysphagia secondary to cancers of the head and neck region, the oesophagus and/or the gastro-oesophageal junction. The patients were usually referred from the head and neck or gastroenterology clinics. Patients who were not suitable for PEG insertion, either because of limited mouth opening or due to failure of trans-illumination during the procedure, were referred for radiological gastrostomy tube insertion or for surgical gastrostomy. Patients with gastro-oesophageal junction tumours with extensive gastric involvement were judged unsuitable for PEG tubes and had either nasojejunal feeding tube insertion or oesophageal stenting, in accordance with decisions made in multidisciplinary team (MDT) meetings. Technique of PEG Tube Insertion PEG tubes were inserted by Ponsky’s ‘pull’ technique, after obtaining informed consent. It is our practice to insert a 20-Fr PEG tube in all adult patients. The procedure begins with
The charts of all patients who had a PEG tube inserted at our institution during 2012 were reviewed retrospectively till June 2013, to assess clinical outcomes and complications. Data was collected by reviewing clinic and emergency room visits, inpatient admission notes, from radiology and endoscopy reports and from patients and their families by telephonic survey. A data set was compiled, consisting of the patients’ demographic profile, the date of the procedure, diagnosis, indication for the procedure and the resultant complications. Complications Complications were considered to be either procedurerelated, and occurring during endoscopy, or PEG-related, and occurring later, at a variable time post-procedure. Complications were classed as fatal, major or minor. Major complications were those that resulted in major morbidity for the patient and included repeat procedures, bowel perforation, septicaemia, gastrointestinal haemorrhage, peri-stomal (peri-PEG) or intra-abdominal collection or abscess formation, peritonitis requiring surgery, aspiration pneumonia, tumour seeding at the PEG site, gastro-cutaneous fistula and major leakage caused by the PEG tube. Complications were classed as minor if they required minimal intervention or were self-limited, and included minor discharge around the PEG insertion site, requiring simple wound care or oral antibiotics (peri-PEG infection), mild peri-stomal leak (peri-PEG leak), mild pain at the PEG site, dislodged tubes, asymptomatic pneumo-peritoneum and tube blockade.
454
J Gastrointest Canc (2014) 45:452–459
Statistical Analysis The rate of complications after PEG tube insertion was analysed using SPSS version 19.
Results Patient Characteristics A total of 260 patients (159 males, age range 18–81 ears; mean 50 years) underwent PEG tube insertion at our institution in 2012. One hundred and thirty-two patients had head and neck cancers, 94 patients had carcinoma of the oesophagus, while 31 had carcinoma of the gastro-oesophageal junction (Table 1). Complications A total of 65 patients (25 %) had PEG-related complications. Fifty-five patients (21 %) had minor complications while 10 patients (3.8 %) suffered major complications. Of the 10 patients with major complications, 3 (1.2 %) patients had peritonitis, while another 3 (1.2 %) had severe peri-PEG infection requiring hospitalization. One of these
Table 1 Demographics Parameters Number of patients Males Females Cancer types Head and neck Carcinoma of the tongue Carcinoma of the hypopharynx Carcinoma of the nasopharynx Carcinoma of the oropharynx Carcinoma of the larynx Carcinoma of the alveolus Carcinoma of the thyroid Carcinoma of the lower lip Osteosarcoma Metastatic cervical node with unknown primary Oesophagus Gastro-oesophageal junction Anaplastic large cell lymphoma Diffuse large B cell lymphoma Post-lobectomy oesophageal fistula Total
Numbers (n)
159 101 132 34 31 31 15 10 6 2 1 1 1 94 31 1 1 1 260
three patients had three episodes of severe peri-PEG infection. This was a 75-year-old non-diabetic female with carcinoma of the oesophagus who had a PEG tube in situ for 6 months. One patient each (0.4 %) had a peri-PEG collection, bleeding from around the PEG insertion site (peri-PEG bleed), tumour seeding at the PEG site and oesophageal perforation during access dilatation of an oesophageal cancer. Seven patients with oesophageal cancer underwent access dilatation prior to PEG tube insertion, out of which one developed an oesophageal perforation, as discussed above. In this patient, the perforation was diagnosed within 2 h of the procedure and the patient underwent emergency oesophagectomy later the same day—had a smooth post-operative course and was discharged home on day 14. The patient with a peri-PEG bleed developed this on the sixth day post-PEG insertion, presenting with bleeding from the PEG site and melaena, with a 2.7 g/dl drop in haemoglobin. Upper GI endoscopy showed a large clot in the stomach; however, no active bleeding was seen. Bleeding stopped after applying a suture on the skin at the PEG site. In the patient with tumour implantation at the PEG site, this was proven histologically and required local radiotherapy. Among the minor complications seen, 36 (13.8 %) patients experienced mild peri-PEG infection requiring treatment with oral antibiotics, while 12 patients (4.6 %) complained of mild peri-PEG discomfort up to 1 week post-insertion. Four patients (1.5 %) had peri-PEG leak requiring change to a larger calibre PEG tube, from 20 to 24 Fr, and three (1.2 %) patients had tube dislodgement. There were no fatalities related to the procedure of PEG tube insertion nor were there any subsequently attributable to the PEG tube (Table 2).
Table 2 Complications of PEG tube insertion Type of complications
Number (n)
Major complications Peritonitis Severe peri-PEG infection Peri-PEG collection Peri-PEG bleed Tumour seeding at the PEG site Oesophageal perforation Minor complications Mild peri-PEG infection Mild peri-PEG discomfort Peri-PEG leak Tube dislodgement Total
10 3 3 1 1 1 1 55 36 12 4 3 65
PEG percutaneous endoscopic gastrostomy
J Gastrointest Canc (2014) 45:452–459
Discussion Nutritional support is commonly required for patients with head and neck cancer as well as those with carcinoma of the oesophagus and gastro-oesophageal junction. These patients frequently present with dysphagia, depending upon the site and extent of the cancer and often require neo-adjuvant treatment, followed by surgery, including glossectomy, mandibulectomy or oesophagectomy. In many cases, subsequent adjuvant treatment, either radiotherapy or chemotherapy, alone or in combination, may also be needed. All of these treatment modalities lead to further problems with feeding and can cause worsening of nutritional status in these already undernourished patients [4, 6]. While chemotherapy does contribute towards weight loss by causing loss of appetite [5], it is mainly the radiotherapy that leads to mucositis, stomatitis, xerostomia, alteration in taste sensation and alteration of viscosity of saliva, all contributing towards poor oral intake [3, 6]. These patients require an alternate source of fluid and nutritional supplementation if their treatment is to be continued uninterrupted and, as discussed, in order to improve the final outcome of treatment. To achieve this, they require a source of enteral feeding that is both safe and effective. In the past, this was achieved with naso-gastric tubes, but these have largely been supplanted by gastrostomy tube insertion and, since the late 1980s, these are usually inserted by the percutaneous endoscopic method. Interestingly, a randomized controlled trial of nasogastric tubes versus PEG tubes in such patients has never been carried out. While upper gastrointestinal endoscopy is generally considered to be a safe procedure in carefully selected patients, rare but serious complications include cardiopulmonary compromise, haemorrhage, perforation and aspiration [15]. However, most of this data comes from healthy patients and not from patients with cancer with poor nutritional status [16]. The reported rate of complications as a result of PEG tube insertion varies in different studies from 4 to 41 % [2, 11, 12, 17–23]. While there is no consensus on how these complications are defined, major complications are generally regarded as those that are either fatal or cause major patient discomfort or lead to a repeat procedure, to surgical intervention or to hospitalization. The incidence of such major complications varies from 3 to 4 % in different studies [20, 22, 24]. Minor complications are generally regarded as those that are selflimiting or require minimal intervention and vary in incidence in the literature from 7.4 to 41 % [2, 11, 12, 20, 22]. In our study, a total of 65 out of 260 patients experienced PEG-related complications. Ten patients (3.8 %) suffered major complications while 55 patients (21 %) had minor complications. Of the 10 patients who suffered major complications, 3 (1.2 %) patients had peritonitis, and a similar percentage had severe peri-PEG infection requiring hospitalization. There was no mortality related to the endoscopy or
455
attributable to having a PEG tube in place, in our cohort. Among the minor complications, 36 (13.8 %) patients experienced mild peri-PEG infection, and 12 (4.6 %) patients had mild peri-PEG pain, while 4 (1.5 %) had peri-PEG leak requiring tube change. We are unsure whether any of our patients with local site breakdown, infection or tube dislodgement had an over-tight flange contributing to this, since ours is a retrospective review, but our results are largely consistent with earlier studies. Ponsky and Gauderer conducted a prospective study on 150 patients including 50 children and 100 adults. There were 18 patients with head and neck cancer, two had gastric outlet obstruction due to peritoneal carcinomatosis and one had radiation enteritis. The rest were non-cancer patients. The incidence of complications was 10 %, much lower than our complication rate, and there was no procedure related mortality. Seven patients had peri-PEG infection, three had tube dislodgements and two developed gastro-colic fistulae—one in a child and the other in an adult. There were no cases of peritonitis, haemorrhage, leakage or gastric outlet obstruction [25]. Since the introduction of the pull technique of PEG tube insertion, by Ponsky, various modifications in gastrostomy tube insertion procedure have been tried including the push, simplified push and introducer techniques as well as the direct method. Each technique has its own advantages as well as disadvantages. The push technique is quite similar to the pull technique, with the difference that after the guide wire has been brought out through the oral cavity, the PEG tube is pushed down through the oesophagus and out through the stoma site in the anterior abdominal wall. The simplified push technique avoids pulling the gastrostomy tube through the upper digestive tract by inserting it directly, through the anterior abdominal wall, using a modified Seldinger technique. This technique has been found to be particularly useful for paediatric patients. Russell’s introducer technique utilizes Tfasteners to secure the stomach to the abdominal wall to avoid gastric displacement during tube insertion and inadvertent injury to other abdominal viscera, following which a guide wire with serial dilators is used to insert the PEG through the anterior abdominal wall using a modified Seldinger technique. The direct method is a modification of the introducer method, wherein the initial steps of insertion are the same as for the introducer technique, but in which, following dilation of the tract, the PEG tube is inserted over the guide wire using an obturator. With the pull technique, which we use, large bore gastrostomy tubes can be placed and the bumper at one end of the tube prevents tube displacement. The potential disadvantage of the pull/push technique is considered to be tube contamination with oropharyngeal/oesophageal pathogens leading to increased rates of peri-PEG infection as compared with the introducer/direct method. Tumour implantation at the gastrostomy site is also thought to be due to direct tumour seeding of the PEG tube as it passes through the tumour.
456
However, tumour implantation at the gastrostomy site has also been reported after open gastrostomies in which there was no contamination of the tubes with tumour. There is an increased risk of bleeding with the introducer technique, especially if a large trocar is used as well as of tube displacement [2, 17, 26–30]. In a large prospective study by Ehrsson et al., evaluating 156 head and neck cancer patients, the incidence of complications was higher than the figure we report. Sixty-five (42 %) patients had complications, including seven (5 %) deaths, 33 (21 %) major and 25 (16 %) minor complications. Of the seven patients who died, three had procedure-related mortality, including fatal peritonitis in two and pneumoperitoneum leading to renal failure and death in one patient. Four patients had fatal complications which were reported to be related to the PEG tube, including a case of fatal necrotizing fasciitis around the PEG; two patients had gastro-intestinal bleeding and one had severe diarrhoea with paralytic ileus leading to death. It is unclear from this paper how diarrhoea could have been caused by PEG tube insertion, however. In five of these seven deaths, the PEG tube was inserted after the completion of radiation therapy, while in another patient, the tube was inserted during radiation. The seventh patient died before definitive cancer treatment could be started. Severe complications involved wound infection in 26 patients, major leakage in 5, septicemia in two and pneumonia, subcutaneous emphysema and ileus in one patient each. Minor complications included pain around the PEG site in 22 patients, minor leakage in 12, granulation tissue in 11, undefined problems with ‘PEG material’ in 10, feed regurgitation and minor bleeding in 3, accidental tube dislodgement in 2 and a blocked PEG tube and an undefined wound caused by the PEG tube in one patient each [2]. In another large prospective multi-centre study on head and neck cancer patients by Grant et al., a total of 172 gastrostomy procedures were performed over a year. One hundred and twenty-one patients had PEG tubes inserted, while 51 patients considered unsuitable for endoscopic gastrostomy tube insertion had radiological insertion of a gastrostomy tube. The overall complication rate in patients with PEG tube insertion was 31.4 %, including 2.5 % major and 28.9 % minor. One patient developed severe gastrointestinal bleeding within the first week of PEG insertion leading to death. There were two major bleeds in the PEG group and 35 minor complications, including 11 patients with minor infections, 6 with abdominal pain, 5 with peri-PEG leaks, 4 each with an over-tight flange requiring adjustment and local inflammation, 2 with a minor bleed and one each with flange upside down and broken gastrostomy tube [12]. Most of the studies reviewed had a follow-up period of up to 6 months, but Larson et al. followed patients for up to 2 years post-PEG tube insertion, and we feel that this study is therefore still relevant today, despite the fact that it was performed more than 25 years ago. In their cohort, procedure-
J Gastrointest Canc (2014) 45:452–459
related mortality was 1 %, with major complications in 3 % and minor complications in 13 % [20]. Most of the studies discussed reported the complications associated with PEG-tube insertion but did not report an increased risk with any particular technique. Tucker et al. carried out a retrospective study comparing ‘push’ versus pull methods of PEG tube insertion, in patients with head and neck cancer, and reported significantly higher rates of complications with pull (30 %; 15/50) versus push (0 %; 0/29) methods [31]. However, in a prospective study of 100 patients by Akkersdijk et al., there was no significant difference in complications between the pull and push methods. The use of preprocedure antibiotics significantly reduced the rate of periPEG infection [32]. The initial studies using the introducer technique had low success rate of tube insertion; however, with the use of T-fasteners, the success rate has increased and complications has decreased. Many recent studies have complication rates of less than 11 and 0 % mortality [33, 34]. In contrast, the study by Dyck et al., which compared pull versus ‘introducer’ method of PEG tube insertion in head and neck cancer patients, found higher rates of complications with the introducer method, including a 48 % (11/24) risk of short-term complications, accidental removal 17 % (4/24), wound infection 12 % (3/24) and perforation 12 % (3/24) with the introducer method, compared to 12 % (4/33), 0 % (0/33), 1 % (3/33) and 3 % (3/33) with the pull method, respectively. This increased rate of complications in this retrospective study with the introducer technique could be related to selection bias, as all patients with high-grade stenosis underwent PEG tube insertion by the introducer method [35]. Bankhead et al. compared the complication rate of different methods of gastrostomy tube insertion including PEG versus open surgical gastrostomy versus laparoscopic gastrostomy and found the incidence of complications was much lower with PEG and the open group compared to laparoscopic technique and concluded that PEG should be the preferred method of gastrostomy tube insertion [36]. Rustom et al. compared the complications of percutaneous endoscopic, radiologically inserted gastrostomy (RIG) and surgical gastrostomy tubes in 78 patients with head and neck cancer. The incidence of minor complications was the same in these groups, but tube blockage and dislodgment was lowest in the PEG group. Thirty-day mortality was 4 %, including two patients in the RIG group and one patient in the surgical gastrostomy group. There were no deaths in the PEG group, and the author concluded that PEG should be the first choice for gastrostomy tube insertion in head and neck cancer patients [37]. In the study by Deurloo et al. in 130 attempted RIG insertions, the complication rate was 30 %, including 14 % minor and 8.4 % major complications, with one death (0.7 %) [38]. In another study of RIG by Cantwell et al., the overall complication rate was 19 %, including 7 % minor and 12 % major complications. There was no fatality [39].
J Gastrointest Canc (2014) 45:452–459
457
Considering the complications associated with PEG tube insertion, including the low but real risk of mortality associated with this procedure, some studies recommend the use of nasogastric feeding tubes for enteral feeding if this is likely to be required only for a short period of up to 4 weeks, with others suggesting that NG feeding might be prolonged for up to 12 weeks [10, 40]. Nugent et al. compared PEG versus NG feeding in patients with head and neck cancer patients with regard to weight and treatment interruptions. Both groups were comparable with regard to radiotherapy delays. The duration of enteral feeding was prolonged with PEG as compared to NGT; with patients fed by naso-gastric tube resumed oral feeding earlier, but there was no statistically significant difference in weight in both groups [41]. Similar results were obtained in a study on oropharyngeal cancer by William et al., which again raised concern regarding the enteral feeding dependence with prophylactic PEG feeding and encouraged earlier resumption of oral feeding. There was no significant impact on survival with either feeding methods (PEG versus NG) [42]. Sadasivan et al. performed a prospective study of 100 head and neck cancer patients comparing NG with PEG feeding (50 patients each in both arms). PEG feeding was found to be more efficacious for nutrition with regard to haemoglobin level, weight gain and mid arm circumference after 1 week, 6 weeks and 6 months post-tube insertion (p
ORIGINAL RESEARCH
Complications of Percutaneous Endoscopic Gastrostomy Tube Insertion in Cancer Patients: A Retrospective Study Hala Mansoor & Muhammad Adnan Masood & Muhammed Aasim Yusuf
Published online: 26 June 2014 # Springer Science+Business Media New York 2014
Abstract Aim Dysphagia, with associated malnutrition is a common feature in up to 64 % of patients with cancers of the head and neck region, the oesophagus and the gastro-oesophageal junction (Donaldson and Lenon Cancer 43(Suppl 5):2036–52, 1979). These patients usually require alternate routes of feeding during treatment. However, these methods of feeding are not without associated complications. This study was carried out to assess the complications associated with percutaneous endoscopic gastrostomy tube insertion, the commonest means of providing nutrition in this patient group. Methods Two hundred and sixty patients with successful percutaneous endoscopic gastrostomy (PEG) tube insertion were retrospectively reviewed to assess the complications of the procedure. Results The overall complication rate was 25 % (65 patients). Fifty-five patients (21 %) patients had minor complications. Ten patients (3.8 %) suffered a major complication, including peritonitis and major peri-PEG infection each in three patients (1.2 %), while oesophageal perforation, collection around the PEG insertion site requiring operative drainage, bleeding from the PEG insertion site and tumour seeding at the gastrostomy site were each seen in one (0.4 %) patient. Conclusion Percutaneous endoscopic gastrostomy tube insertion is a useful means of providing enteral nutrition to patients H. Mansoor (*) : M. A. Masood : M. A. Yusuf Shaukat Khanum Memorial Cancer Hospital & Research Centre, Lahore, Pakistan e-mail: [email protected] H. Mansoor e-mail: [email protected] M. A. Masood e-mail: [email protected] M. A. Yusuf e-mail: [email protected]
with swallowing problems. Patients with head and neck cancer who are likely to develop dysphagia should be assessed promptly for PEG tube insertion before the start of radiation. Early recognition and prompt treatment of serious complications can avoid potential fatality associated with PEG tubes. Keywords Percutaneous endoscopic gastrostomy (PEG) . Head and neck . Cancer of the oesophagus and gastro-oesophageal junction . Nutrition
Introduction Since the introduction of percutaneous endoscopic gastrostomy (PEG) by Gauderer and Ponsky in 1980 [1], it has become the most commonly employed technique of prolonged enteral feeding in patients with impaired swallowing. This includes patients with neurological causes of dysphagia as well as those with neoplasms of the oral cavity, oropharynx, hypopharynx, nasopharynx, larynx, oesophagus and gastro-oesophageal junction [2]. Most patients with these malignancies already have poor nutritional status at the time of diagnosis due to dysphagia, and as many as 64 % are malnourished [3]. Radical surgery, involving either partial or total glossectomy or mandibulectomy, may further impair swallowing [4]. Chemotherapy may lead to loss of appetite while radiotherapy causes a local radiation reaction that leads to poor oral intake [3, 5, 6]. Studies on cancer patients have suggested improved outcomes and fewer interruptions of treatment in patients who received nutritional supplementation [7–9]. Lees performed a prospective trial, comparing nasogastric feeding with PEG feeding in head and neck cancer patients undergoing radiotherapy and concluded that the final outcome of these patients depended mainly on the nutritional status achieved, regardless of the method of feeding [10]. While PEG feeding is generally preferred, as it is thought to
J Gastrointest Canc (2014) 45:452–459
cause less discomfort as well as for cosmetic reasons, PEG insertion is not entirely safe and carries its own risks, including a high rate of insertion-site infection in up to 41 % of patients and other less frequent, but more serious side-effects, including septicemia, gut perforation, peritonitis and aspiration pneumonia, among others [2, 11, 12]. However, PEG insertion remains a common indication for upper gastrointestinal tract endoscopy [13]. The various techniques used for PEG insertion include the pull (Ponsky), introducer (Russell) and push (Sachs-Vine) and the direct method [2]. Of these, Ponsky’s pull technique is the most frequently used method [14]. At our hospital, all patients who present with dysphagia due to cancer of the head and neck region, the pharynx, the oesophagus or the gastro-oesophageal junction, or in whom swallowing problems are anticipated during chemoradiotherapy, are assessed for PEG tube insertion. The aim of this study was to retrospectively analyse the data of our hospital patients to determine the outcome of PEG tube insertion during the year 2012 for all such patients at our institution.
453
standard upper GI endoscopy or oesophago-gastroduodenoscopy to evaluate the anatomy of the upper GI tract. The anterior wall of the stomach is identified, and an effort is made to ensure that there are no intervening viscerae between the stomach and the anterior abdominal wall. This is done by applying digital pressure on the anterior abdominal wall, which can be seen as an indentation of the stomach endoscopically as well as by trans-illumination of the abdominal wall, through the stomach. A failure to trans-illuminate is an absolute contraindication in attempting PEG tube insertion. Local anaesthetic is then injected at the site selected for gastric puncture, a trocar and cannula is used to puncture the stomach and a guide wire is inserted through the cannula into the gastric lumen. This is grasped endoscopically and pulled out through the mouth. A 20-Fr gastrostomy tube is then attached to the wire and pulled back down the oesophagus and into the stomach, from where it is pulled into place through the anterior abdominal wall. It is our practice to administer a single dose of prophylactic intravenous antibiotic immediately prior to the procedure. Data Collection
Materials and Methods Patients This retrospective study was conducted at Shaukat Khanum Memorial Cancer Hospital & Research Centre, Lahore, Pakistan, after obtaining approval from the Shaukat Khanum Memorial Cancer Hospital & Research Centre Institutional Review Board, in accordance with the principles of the Helsinki Declaration. The data of 260 consecutive patients (159 males, age range 18–81 years; mean age 50 years) who underwent PEG tube insertion from January 2012 till December 2012 was reviewed retrospectively. The majority of these patients had dysphagia secondary to cancers of the head and neck region, the oesophagus and/or the gastro-oesophageal junction. The patients were usually referred from the head and neck or gastroenterology clinics. Patients who were not suitable for PEG insertion, either because of limited mouth opening or due to failure of trans-illumination during the procedure, were referred for radiological gastrostomy tube insertion or for surgical gastrostomy. Patients with gastro-oesophageal junction tumours with extensive gastric involvement were judged unsuitable for PEG tubes and had either nasojejunal feeding tube insertion or oesophageal stenting, in accordance with decisions made in multidisciplinary team (MDT) meetings. Technique of PEG Tube Insertion PEG tubes were inserted by Ponsky’s ‘pull’ technique, after obtaining informed consent. It is our practice to insert a 20-Fr PEG tube in all adult patients. The procedure begins with
The charts of all patients who had a PEG tube inserted at our institution during 2012 were reviewed retrospectively till June 2013, to assess clinical outcomes and complications. Data was collected by reviewing clinic and emergency room visits, inpatient admission notes, from radiology and endoscopy reports and from patients and their families by telephonic survey. A data set was compiled, consisting of the patients’ demographic profile, the date of the procedure, diagnosis, indication for the procedure and the resultant complications. Complications Complications were considered to be either procedurerelated, and occurring during endoscopy, or PEG-related, and occurring later, at a variable time post-procedure. Complications were classed as fatal, major or minor. Major complications were those that resulted in major morbidity for the patient and included repeat procedures, bowel perforation, septicaemia, gastrointestinal haemorrhage, peri-stomal (peri-PEG) or intra-abdominal collection or abscess formation, peritonitis requiring surgery, aspiration pneumonia, tumour seeding at the PEG site, gastro-cutaneous fistula and major leakage caused by the PEG tube. Complications were classed as minor if they required minimal intervention or were self-limited, and included minor discharge around the PEG insertion site, requiring simple wound care or oral antibiotics (peri-PEG infection), mild peri-stomal leak (peri-PEG leak), mild pain at the PEG site, dislodged tubes, asymptomatic pneumo-peritoneum and tube blockade.
454
J Gastrointest Canc (2014) 45:452–459
Statistical Analysis The rate of complications after PEG tube insertion was analysed using SPSS version 19.
Results Patient Characteristics A total of 260 patients (159 males, age range 18–81 ears; mean 50 years) underwent PEG tube insertion at our institution in 2012. One hundred and thirty-two patients had head and neck cancers, 94 patients had carcinoma of the oesophagus, while 31 had carcinoma of the gastro-oesophageal junction (Table 1). Complications A total of 65 patients (25 %) had PEG-related complications. Fifty-five patients (21 %) had minor complications while 10 patients (3.8 %) suffered major complications. Of the 10 patients with major complications, 3 (1.2 %) patients had peritonitis, while another 3 (1.2 %) had severe peri-PEG infection requiring hospitalization. One of these
Table 1 Demographics Parameters Number of patients Males Females Cancer types Head and neck Carcinoma of the tongue Carcinoma of the hypopharynx Carcinoma of the nasopharynx Carcinoma of the oropharynx Carcinoma of the larynx Carcinoma of the alveolus Carcinoma of the thyroid Carcinoma of the lower lip Osteosarcoma Metastatic cervical node with unknown primary Oesophagus Gastro-oesophageal junction Anaplastic large cell lymphoma Diffuse large B cell lymphoma Post-lobectomy oesophageal fistula Total
Numbers (n)
159 101 132 34 31 31 15 10 6 2 1 1 1 94 31 1 1 1 260
three patients had three episodes of severe peri-PEG infection. This was a 75-year-old non-diabetic female with carcinoma of the oesophagus who had a PEG tube in situ for 6 months. One patient each (0.4 %) had a peri-PEG collection, bleeding from around the PEG insertion site (peri-PEG bleed), tumour seeding at the PEG site and oesophageal perforation during access dilatation of an oesophageal cancer. Seven patients with oesophageal cancer underwent access dilatation prior to PEG tube insertion, out of which one developed an oesophageal perforation, as discussed above. In this patient, the perforation was diagnosed within 2 h of the procedure and the patient underwent emergency oesophagectomy later the same day—had a smooth post-operative course and was discharged home on day 14. The patient with a peri-PEG bleed developed this on the sixth day post-PEG insertion, presenting with bleeding from the PEG site and melaena, with a 2.7 g/dl drop in haemoglobin. Upper GI endoscopy showed a large clot in the stomach; however, no active bleeding was seen. Bleeding stopped after applying a suture on the skin at the PEG site. In the patient with tumour implantation at the PEG site, this was proven histologically and required local radiotherapy. Among the minor complications seen, 36 (13.8 %) patients experienced mild peri-PEG infection requiring treatment with oral antibiotics, while 12 patients (4.6 %) complained of mild peri-PEG discomfort up to 1 week post-insertion. Four patients (1.5 %) had peri-PEG leak requiring change to a larger calibre PEG tube, from 20 to 24 Fr, and three (1.2 %) patients had tube dislodgement. There were no fatalities related to the procedure of PEG tube insertion nor were there any subsequently attributable to the PEG tube (Table 2).
Table 2 Complications of PEG tube insertion Type of complications
Number (n)
Major complications Peritonitis Severe peri-PEG infection Peri-PEG collection Peri-PEG bleed Tumour seeding at the PEG site Oesophageal perforation Minor complications Mild peri-PEG infection Mild peri-PEG discomfort Peri-PEG leak Tube dislodgement Total
10 3 3 1 1 1 1 55 36 12 4 3 65
PEG percutaneous endoscopic gastrostomy
J Gastrointest Canc (2014) 45:452–459
Discussion Nutritional support is commonly required for patients with head and neck cancer as well as those with carcinoma of the oesophagus and gastro-oesophageal junction. These patients frequently present with dysphagia, depending upon the site and extent of the cancer and often require neo-adjuvant treatment, followed by surgery, including glossectomy, mandibulectomy or oesophagectomy. In many cases, subsequent adjuvant treatment, either radiotherapy or chemotherapy, alone or in combination, may also be needed. All of these treatment modalities lead to further problems with feeding and can cause worsening of nutritional status in these already undernourished patients [4, 6]. While chemotherapy does contribute towards weight loss by causing loss of appetite [5], it is mainly the radiotherapy that leads to mucositis, stomatitis, xerostomia, alteration in taste sensation and alteration of viscosity of saliva, all contributing towards poor oral intake [3, 6]. These patients require an alternate source of fluid and nutritional supplementation if their treatment is to be continued uninterrupted and, as discussed, in order to improve the final outcome of treatment. To achieve this, they require a source of enteral feeding that is both safe and effective. In the past, this was achieved with naso-gastric tubes, but these have largely been supplanted by gastrostomy tube insertion and, since the late 1980s, these are usually inserted by the percutaneous endoscopic method. Interestingly, a randomized controlled trial of nasogastric tubes versus PEG tubes in such patients has never been carried out. While upper gastrointestinal endoscopy is generally considered to be a safe procedure in carefully selected patients, rare but serious complications include cardiopulmonary compromise, haemorrhage, perforation and aspiration [15]. However, most of this data comes from healthy patients and not from patients with cancer with poor nutritional status [16]. The reported rate of complications as a result of PEG tube insertion varies in different studies from 4 to 41 % [2, 11, 12, 17–23]. While there is no consensus on how these complications are defined, major complications are generally regarded as those that are either fatal or cause major patient discomfort or lead to a repeat procedure, to surgical intervention or to hospitalization. The incidence of such major complications varies from 3 to 4 % in different studies [20, 22, 24]. Minor complications are generally regarded as those that are selflimiting or require minimal intervention and vary in incidence in the literature from 7.4 to 41 % [2, 11, 12, 20, 22]. In our study, a total of 65 out of 260 patients experienced PEG-related complications. Ten patients (3.8 %) suffered major complications while 55 patients (21 %) had minor complications. Of the 10 patients who suffered major complications, 3 (1.2 %) patients had peritonitis, and a similar percentage had severe peri-PEG infection requiring hospitalization. There was no mortality related to the endoscopy or
455
attributable to having a PEG tube in place, in our cohort. Among the minor complications, 36 (13.8 %) patients experienced mild peri-PEG infection, and 12 (4.6 %) patients had mild peri-PEG pain, while 4 (1.5 %) had peri-PEG leak requiring tube change. We are unsure whether any of our patients with local site breakdown, infection or tube dislodgement had an over-tight flange contributing to this, since ours is a retrospective review, but our results are largely consistent with earlier studies. Ponsky and Gauderer conducted a prospective study on 150 patients including 50 children and 100 adults. There were 18 patients with head and neck cancer, two had gastric outlet obstruction due to peritoneal carcinomatosis and one had radiation enteritis. The rest were non-cancer patients. The incidence of complications was 10 %, much lower than our complication rate, and there was no procedure related mortality. Seven patients had peri-PEG infection, three had tube dislodgements and two developed gastro-colic fistulae—one in a child and the other in an adult. There were no cases of peritonitis, haemorrhage, leakage or gastric outlet obstruction [25]. Since the introduction of the pull technique of PEG tube insertion, by Ponsky, various modifications in gastrostomy tube insertion procedure have been tried including the push, simplified push and introducer techniques as well as the direct method. Each technique has its own advantages as well as disadvantages. The push technique is quite similar to the pull technique, with the difference that after the guide wire has been brought out through the oral cavity, the PEG tube is pushed down through the oesophagus and out through the stoma site in the anterior abdominal wall. The simplified push technique avoids pulling the gastrostomy tube through the upper digestive tract by inserting it directly, through the anterior abdominal wall, using a modified Seldinger technique. This technique has been found to be particularly useful for paediatric patients. Russell’s introducer technique utilizes Tfasteners to secure the stomach to the abdominal wall to avoid gastric displacement during tube insertion and inadvertent injury to other abdominal viscera, following which a guide wire with serial dilators is used to insert the PEG through the anterior abdominal wall using a modified Seldinger technique. The direct method is a modification of the introducer method, wherein the initial steps of insertion are the same as for the introducer technique, but in which, following dilation of the tract, the PEG tube is inserted over the guide wire using an obturator. With the pull technique, which we use, large bore gastrostomy tubes can be placed and the bumper at one end of the tube prevents tube displacement. The potential disadvantage of the pull/push technique is considered to be tube contamination with oropharyngeal/oesophageal pathogens leading to increased rates of peri-PEG infection as compared with the introducer/direct method. Tumour implantation at the gastrostomy site is also thought to be due to direct tumour seeding of the PEG tube as it passes through the tumour.
456
However, tumour implantation at the gastrostomy site has also been reported after open gastrostomies in which there was no contamination of the tubes with tumour. There is an increased risk of bleeding with the introducer technique, especially if a large trocar is used as well as of tube displacement [2, 17, 26–30]. In a large prospective study by Ehrsson et al., evaluating 156 head and neck cancer patients, the incidence of complications was higher than the figure we report. Sixty-five (42 %) patients had complications, including seven (5 %) deaths, 33 (21 %) major and 25 (16 %) minor complications. Of the seven patients who died, three had procedure-related mortality, including fatal peritonitis in two and pneumoperitoneum leading to renal failure and death in one patient. Four patients had fatal complications which were reported to be related to the PEG tube, including a case of fatal necrotizing fasciitis around the PEG; two patients had gastro-intestinal bleeding and one had severe diarrhoea with paralytic ileus leading to death. It is unclear from this paper how diarrhoea could have been caused by PEG tube insertion, however. In five of these seven deaths, the PEG tube was inserted after the completion of radiation therapy, while in another patient, the tube was inserted during radiation. The seventh patient died before definitive cancer treatment could be started. Severe complications involved wound infection in 26 patients, major leakage in 5, septicemia in two and pneumonia, subcutaneous emphysema and ileus in one patient each. Minor complications included pain around the PEG site in 22 patients, minor leakage in 12, granulation tissue in 11, undefined problems with ‘PEG material’ in 10, feed regurgitation and minor bleeding in 3, accidental tube dislodgement in 2 and a blocked PEG tube and an undefined wound caused by the PEG tube in one patient each [2]. In another large prospective multi-centre study on head and neck cancer patients by Grant et al., a total of 172 gastrostomy procedures were performed over a year. One hundred and twenty-one patients had PEG tubes inserted, while 51 patients considered unsuitable for endoscopic gastrostomy tube insertion had radiological insertion of a gastrostomy tube. The overall complication rate in patients with PEG tube insertion was 31.4 %, including 2.5 % major and 28.9 % minor. One patient developed severe gastrointestinal bleeding within the first week of PEG insertion leading to death. There were two major bleeds in the PEG group and 35 minor complications, including 11 patients with minor infections, 6 with abdominal pain, 5 with peri-PEG leaks, 4 each with an over-tight flange requiring adjustment and local inflammation, 2 with a minor bleed and one each with flange upside down and broken gastrostomy tube [12]. Most of the studies reviewed had a follow-up period of up to 6 months, but Larson et al. followed patients for up to 2 years post-PEG tube insertion, and we feel that this study is therefore still relevant today, despite the fact that it was performed more than 25 years ago. In their cohort, procedure-
J Gastrointest Canc (2014) 45:452–459
related mortality was 1 %, with major complications in 3 % and minor complications in 13 % [20]. Most of the studies discussed reported the complications associated with PEG-tube insertion but did not report an increased risk with any particular technique. Tucker et al. carried out a retrospective study comparing ‘push’ versus pull methods of PEG tube insertion, in patients with head and neck cancer, and reported significantly higher rates of complications with pull (30 %; 15/50) versus push (0 %; 0/29) methods [31]. However, in a prospective study of 100 patients by Akkersdijk et al., there was no significant difference in complications between the pull and push methods. The use of preprocedure antibiotics significantly reduced the rate of periPEG infection [32]. The initial studies using the introducer technique had low success rate of tube insertion; however, with the use of T-fasteners, the success rate has increased and complications has decreased. Many recent studies have complication rates of less than 11 and 0 % mortality [33, 34]. In contrast, the study by Dyck et al., which compared pull versus ‘introducer’ method of PEG tube insertion in head and neck cancer patients, found higher rates of complications with the introducer method, including a 48 % (11/24) risk of short-term complications, accidental removal 17 % (4/24), wound infection 12 % (3/24) and perforation 12 % (3/24) with the introducer method, compared to 12 % (4/33), 0 % (0/33), 1 % (3/33) and 3 % (3/33) with the pull method, respectively. This increased rate of complications in this retrospective study with the introducer technique could be related to selection bias, as all patients with high-grade stenosis underwent PEG tube insertion by the introducer method [35]. Bankhead et al. compared the complication rate of different methods of gastrostomy tube insertion including PEG versus open surgical gastrostomy versus laparoscopic gastrostomy and found the incidence of complications was much lower with PEG and the open group compared to laparoscopic technique and concluded that PEG should be the preferred method of gastrostomy tube insertion [36]. Rustom et al. compared the complications of percutaneous endoscopic, radiologically inserted gastrostomy (RIG) and surgical gastrostomy tubes in 78 patients with head and neck cancer. The incidence of minor complications was the same in these groups, but tube blockage and dislodgment was lowest in the PEG group. Thirty-day mortality was 4 %, including two patients in the RIG group and one patient in the surgical gastrostomy group. There were no deaths in the PEG group, and the author concluded that PEG should be the first choice for gastrostomy tube insertion in head and neck cancer patients [37]. In the study by Deurloo et al. in 130 attempted RIG insertions, the complication rate was 30 %, including 14 % minor and 8.4 % major complications, with one death (0.7 %) [38]. In another study of RIG by Cantwell et al., the overall complication rate was 19 %, including 7 % minor and 12 % major complications. There was no fatality [39].
J Gastrointest Canc (2014) 45:452–459
457
Considering the complications associated with PEG tube insertion, including the low but real risk of mortality associated with this procedure, some studies recommend the use of nasogastric feeding tubes for enteral feeding if this is likely to be required only for a short period of up to 4 weeks, with others suggesting that NG feeding might be prolonged for up to 12 weeks [10, 40]. Nugent et al. compared PEG versus NG feeding in patients with head and neck cancer patients with regard to weight and treatment interruptions. Both groups were comparable with regard to radiotherapy delays. The duration of enteral feeding was prolonged with PEG as compared to NGT; with patients fed by naso-gastric tube resumed oral feeding earlier, but there was no statistically significant difference in weight in both groups [41]. Similar results were obtained in a study on oropharyngeal cancer by William et al., which again raised concern regarding the enteral feeding dependence with prophylactic PEG feeding and encouraged earlier resumption of oral feeding. There was no significant impact on survival with either feeding methods (PEG versus NG) [42]. Sadasivan et al. performed a prospective study of 100 head and neck cancer patients comparing NG with PEG feeding (50 patients each in both arms). PEG feeding was found to be more efficacious for nutrition with regard to haemoglobin level, weight gain and mid arm circumference after 1 week, 6 weeks and 6 months post-tube insertion (p
![[Complications of percutaneous endoscopic gastrostomy tube insertion in cancer patients: a retrospective study].](/assets/img/hold.png)
