Complications and Mortality After Percutaneous Endoscopic Gastrostomy in Geriatrics: A Prospective Multicenter Observational Trial

Complications and Mortality After Percutaneous Endoscopic Gastrostomy in Geriatrics: A Prospective Multicenter Observational Trial

JAMDA 13 (2012) 228e233 JAMDA journal homepage: Original Study Complications and Mortality After Percutaneous Endoscopic Gastrostomy ...

258KB Sizes 0 Downloads 0 Views

JAMDA 13 (2012) 228e233

JAMDA journal homepage:

Original Study

Complications and Mortality After Percutaneous Endoscopic Gastrostomy in Geriatrics: A Prospective Multicenter Observational Trial Rainer Wirth MD a, b, *, Christian Voss MD a, Christine Smoliner MSc a, Cornel C. Sieber MD b, c, Juergen M. Bauer MD b, c, Dorothee Volkert PhD b a

Department of Internal Medicine and Geriatrics, St. Marien-Hospital Borken, Borken, Germany Friedrich-Alexander-University Erlangen-Nuremberg, Chair of Geriatric Medicine, Erlangen, Germany c Department of Internal Medicine II, Nuremberg Hospital, Nuremberg, Germany b

a b s t r a c t Keywords: Complications elderly mortality percutaneous endoscopic gastrostomy tube feeding

Background: Generally, the high short-term mortality after percutaneous endoscopic gastrostomy (PEG) in geriatric patients is attributed to the severity of their underlying diseases. However, the procedurerelated mortality in this group is unknown. Methods: This prospective multicenter observational study gathered information about 197 geriatric patients treated with PEG insertion, including the indication for PEG insertion and the prevalence of postprocedure complications and analyzed how these factors related to mortality. Results: Dysphagia (64%) and insufficient food intake (76%) were the most frequent indications for PEG insertion. Severe complications after PEG insertion occurred in 9.6% of patients. Mortality was 9.6% in hospital, as well as 18.4% at 1 month. Six months after PEG placement, with 81 patients lost to follow-up, mortality was 51.9%. Hospital mortality was significantly higher in patients with severe complications caused by PEG insertion (47.4% vs 5.6%; P < .001). A regression analysis that corrected for confounding factors revealed that severe complications in general (HR 6.9; 95% CI: 2.6e18.1; P < .001), peritonitis (HR 33.1; 95% CI: 3.7e293.2; P ¼ .002), and severe wound infections (HR 6.9; 95% CI: 1.9e24.9; P ¼ .003) were each independently associated with hospital mortality. Considering the prevalence of procedure-related complications and their association with early mortality after PEG insertion, the procedure-related mortality rate in geriatric patients was at least 2% in this study. Conclusion: Although the prevalence of complications after PEG in this study of multimorbid geriatric patients is within the expected range, the procedure-related mortality is higher than expected. Published by Elsevier Inc. on behalf of the American Medical Directors Association, Inc.

Because of its high prevalence and serious consequences in the geriatric population, malnutrition is a major health concern.1 Therapeutic interventions include dietary modification, careful feeding by hand, food fortification and supplementation with tube feeding, and parenteral nutrition.2 Enteral nutrition via percutaneous endoscopic gastrostomy (PEG) is a widely accepted method for long-term tube feeding. The European Society for Clinical Nutrition and Nutrition guidelines on PEG recommend enteral

This work was supported by a research grant from Fresenius-Kabi. The Sponsor had no role in study design, data collection, data interpretation or the writing of the manuscript. Parts of the paper were presented at the Congress of the German Society for Internal Medicine (DGIM), April 11, 2010, Wiesbaden, Germany. * Address correspondence to Rainer Wirth, MD, Department of Internal Medicine and Geriatrics, St. Marien-Hospital Borken, Am Boltenhof 7, D-46325 Borken, Germany. E-mail address: [email protected] (R. Wirth).

nutrition via PEG if oral nutrition is expected to be insufficient for longer than 3 weeks.3 PEG was first introduced by Gauderer et al4 in pediatrics. Over time, PEG came to be used more for other patients and indications. Presently, PEG is used predominantly for elderly patients with various indications, even though its impact on survival and quality of life in this population remains unclear.5e7 Although it might seem intuitive that providing nutrition would improve a patient’s prognosis, one has to take into account the potential complications involved in PEG placement and tube feeding, as well as the consideration that improved nutrition may not affect a poor prognosis because of the severity of the underlying disease. Surprisingly, there are limited data about PEG complications and mortality in the multimorbid geriatric population, even though such complications are considerable, even potentially fatal,8 and short-term mortality is high.6 Among clinicians and general practitioners, the complications of PEG insertion are believed to be few and minor and the procedure-related mortality is reported to be well below 1%.7,9e12 However, older subjects face a considerably

1525-8610/$ - see front matter Published by Elsevier Inc. on behalf of the American Medical Directors Association, Inc. doi:10.1016/j.jamda.2011.07.010

R. Wirth et al. / JAMDA 13 (2012) 228e233

high short-term mortality rate of about 19% in the 1 month after PEG placement,6 which might be explained by an unfortunate selection of patients or a previously unrecognized procedurerelated mortality. To our knowledge, no study has investigated the connection between postprocedure complications and mortality after PEG insertion in the vulnerable multimorbid geriatric population. The following prospective multicenter observational trial was conducted to gather information about patients receiving PEG and their indications for PEG insertion, to measure the prevalence of complications and mortality after percutaneous endoscopic gastrostomy in geriatric patients and to evaluate the relationship between the two. Methods Study Design and Ethics Seventeen acute care geriatric hospital units, located throughout Germany, participated in this prospective observational trial. Participating departments included primary care hospitals as well as teaching and university hospitals. Inclusion criteria were any PEG insertion for the purpose of tube feeding in patients of the geriatric department during the first 6 months of 2006. Exclusion criteria were patients younger than 60 years and those without written informed consent. All PEG insertions were performed endoscopically using the pull method by gastroenterologists or internists with experience in endoscopy. The indication for every PEG placement was identified by the responsible geriatrician. The study was performed according to the Helsinki Declaration for Human Studies in accordance with the regulations of German ethical committees. The study was approved by the ethical committee of the University of Munster, Germany, and the medical council of Munster, Germany. Documentation The attending physician documented all items (see Tables 1 and 2). Activities of daily living were measured by the Barthel Index, which was determined by the responsible nurse using a scale from 0 to 100, with low values indicating greater impairment. The Mini-Mental State Examination was administered by trained occupational therapists during the routine geriatric assessment within the first 3 days after admission. All other parameters, such as the patient’s nutritional and cognitive state, were clinically judged by the attending physician. A weight loss of more than 5% in 3 months or more than 10% in the 6 months before PEG insertion was considered to be significant. The primary and secondary diagnoses were defined as those contributing to the PEG indication. The questions about indications for PEG placement, and primary and secondary diagnoses allowed multiple answers within specified categories. Antibiotic prophylaxis was defined as antimicrobial therapy for the sole purpose of preventing PEG-site infections. Complications were defined as early or late, depending on whether symptoms developed within the first 3 days after PEG insertion or later than that. Minor wound infections, diarrhea, vomiting, and abdominal pain were classified as minor complications. Peritonitis, cardiorespiratory insufficiency, gastric bleeding, aspiration, and severe wound infections were classified as severe complications. Wound infections requiring systemic antibiotics were classified as severe. Furthermore, this form documented death in the hospital, PEG removal, oral nutrition at discharge, and the amount of tube feed given. A 6-month follow-up was performed and included questions about mortality, continuing presence of the PEG, and whether tube


Table 1 Characteristics of the Participants, Hospital Stay, and PEG Insertion, given as n (%) or Mean  SD (range) Number of participants 197 Female 111 Male 86 Age, y 80.6  7.7 Barthel Index at admission, points 23.5  33.7 Clinical impression of cognition No disorientation 33 Mild disorientation 39 Severe disorientation 88 No data 37 Length of hospital stay, d 32.1  18.7 Body weight, kg 64.0  15.2 Body height, m 1.66  0.10 2 23.2  5.2 BMI, kg/m Weight course before PEG Loss (> 5% body weight) 52 No loss (> 5% body weight) 62 No data 83 Clinical impression of nutritional state Undernourished 88 Normal 82 Obese 23 No data 4 Dependency with meal intake at admission (according to Barthel Completely dependent 116 In need of help 52 Independent 28 No data 1 PEG after days of hospital stay, d 17.5  18.8 Antibiotic prophylaxis before PEG insertion Yes 102 No 95 No data 0 Additional oral food intake at hospital discharge Yes 86 No 93 No data 18 Additional oral fluid intake at hospital discharge Yes 80 No 99 No data 18

(100) (56.3) (43.7) (60.0e99.9) (0e100) (16.8) (19.8) (44.7) (18.8) (0e197) (32.0e123.7) (1.36e1.93) (9.9e46.9) (26.4) (31.5) (42.1) (44.7) (41.6) (11.7) (2.0) Index) (58.9) (26.4) (14.2) (0.5) (0e192) (51.8) (48.2) (0.0) (43.7) (47.2) (9.1) (40.6) (50.3) (9.1)

BMI, body mass index; PEG, percutaneous endoscopic gastrostomy.

feeding and oral nutrition were being given. Some departments stated before study initiation that they could not do the 6-month follow-up. These departments provided data only on the index hospitalization.

Statistics Data were analyzed using SPSS version 17 (SPSS Inc, Chicago, IL). Descriptive statistics were used to calculate the item frequency, mean, standard deviation, median, and range. Correlation analysis was performed using Spearman’s rank correlation coefficient. The Mann-Whitney test was used to compare 2 groups of metric data. Group comparisons of ordinal and nominal data were performed using the chi-square test. Kruskal-Wallis analysis was used to compare multiple groups of metric data. Survival curves were calculated by the Kaplan-Meier method. The log-rank test was used to compare the survival rates between two different groups. Cox proportional hazards regression was used to assess the simultaneous effects of different factors on hospital, and 1-month and 6-month mortality. All factors with significant group differences of P less than .10 were included in the Cox proportional hazards regression analysis using the stepwise inclusion method. A P value of less than .05 was considered statistically significant.


R. Wirth et al. / JAMDA 13 (2012) 228e233

Table 2 Indications and Diagnoses Contributing to PEG Insertion, given as n (%) Indication for tube feeding* Dysphagia Aspiration Esophageal obstruction Insufficient food intake Insufficient fluid intake Application of medication No data Primary diagnosis* Dementia Depression Stroke Parkinson’s disease Cancer Coma Other No data Secondary diagnosis* Dementia Depression Stroke Parkinson’s disease Cancer Coma Other No data

126 48 4 150 145 84 1

(64.0) (24.4) (2.0) (76.1) (73.6) (42.6) (0.5)

37 9 96 20 8 2 14 11

(18.8) (4.6) (48.7) (10.2) (4.1) (1.0) (7.1) (5.6)

43 18 25 9 7 1 13 90

(21.8) (9.1) (12.7) (4.6) (3.6) (0.5) (6.6) (45.7)

PEG, percutaneous endoscopic gastrostomy. * Multiple answers were possible.

an additional clinical rating of the cognitive state. The main characteristics of the patients and hospital stay, as well as indications for PEG placement and contributing diagnoses, are shown in Tables 1 and 2. The prevalence of complications and the mortality of patients with complications are listed in Table 3. In this study, 50 (25.4%) patients had only minor complications, 10 (5.1%) had only severe complications, and 9 (4.8%) had both minor and severe complications. There were no significant differences in the rates of complications after PEG placement among the participating hospital units. Of all patients, 51.8% received an antibiotic prophylaxis before PEG insertion. There was no significant association between antimicrobial prophylaxis and the prevalence of complications and mortality at any time. Antibiotic prophylaxis appeared to have no significant influence on the rate of aspiration pneumonia and minor wound infections. Patients receiving antibiotic prophylaxis showed a nonsignificant trend toward fewer severe wound infections (1.0% vs 4.2%; P ¼ .150) and an increased prevalence of diarrhea (15.7% vs 9.5%; P ¼ .191). Among the 56 patients documented to have survived after 6 months, 11 (19.6%) patients had their PEG removed and three (5.3%) were no longer using their PEGs. In other words, 75% of those surviving after 6 months were still being fed via PEG, with onethird of patients receiving additional oral intake of food and beverages (see also Figure 1). Three patients (1.5 %) died within 1 day after PEG placement. Because none of these patients were classified as having

Results Of the 197 PEG placements, a median of 7.0 and a mean of 11.6 ( 8.9) PEG placements per study unit were documented. Body height was measured in 83 (42.1%) patients and estimated in 102 (51.8%) patients, with no information in 12 (6.1%) patients. Body weight was measured in 119 (60.4%) patients, estimated in 66 (33.5%) patients, and was unavailable in 12 (6.1%) patients. Therefore, the body mass index was available for 185 patients, with approximately half of these measurements being estimates. The Barthel Index was documented for all patients. Because of dysphasia after stroke or other reasons, the MMSE could be performed in only 42 (21.3%) patients with a mean of 17.3  8.5 (range 0e30) points. Having anticipated this problem, the study asked for Table 3 Complications and Hospital Mortality after PEG Insertion (n ¼ 197) Complication

Prevalence, n (%)

Severe complications Peritonitis Cardiorespiratory failure Gastric bleeding Early aspiration pneumonia Late aspiration pneumonia Severe wound infection Total number of patients Minor complications Early abdominal pain Late abdominal pain Minor wound infection Diarrhea Vomiting Total number of patients Total number of complications Patients with any complication Patients without complications

19 1 0 2 4 7 5 19 77 6 5 29 25 12 59 96 69 128

Hospital Mortality, n (%)

(0.5) (0.0) (1.0) (2.0) (3.6) (2.5) (9.6)

1 0 0 2 3 3 9

(100.0) (0.0) (0.0) (50.0) (42.9) (60.0) (47.4)

(3.0) (2.5) (14.7) (12.7) (6.1) (29.9)*

1 2 4 4 1 3

(16.7) (40.0) (13.8) (16.0) (8.3) (6.0)y

(35.0)* (65.0)

12 (6.1) 7 (5.5)

PEG, percutaneous endoscopic gastrostomy. * The total number of patients is less than the sum of complications because some patients had multiple complications. y Mortality of those with only minor complications (n ¼ 50).

Fig. 1. Course and outcomes of the study participants.

R. Wirth et al. / JAMDA 13 (2012) 228e233

complications, the cause of death remains unknown. Overall hospital mortality was 9.6%. The 19 patients who died during their hospital stay did so after a mean period of 13.7 ( 10.4) days after PEG insertion, with a median of 11.5 (range: 1e34) days. The 1-month mortality rate was 18.4%. Of the 178 participants surviving the hospital stay, 81 (45.5%) were lost to follow-up at the optional 6-month follow-up. Patients with data available from the 6-month follow-up showed no significant differences in activities of daily living, cognitive state, indications, and complications compared with the 81 patients lost at the follow-up. Of the 97 (54.5%) participants with 6-month follow-up data, an additional 41 (42.3%) patients died within 6 months after PEG placement. Thus, in addition to the hospital mortality rate of 9.6%, the total mortality rate after 6 months was 51.9 % (Figure 1).

Factors possibly influencing mortality There were no significant differences in hospital mortality after PEG placement among the participating hospital units. No significant correlation was found between the number of documented PEGs per unit and hospital mortality. Furthermore, there were no significant differences between patients who died in the hospital and those who survived in any of the following factors: clinical impression of the cognitive state, MMSE, Barthel Index at the time of admission, age, sex, body mass index, clinical impression of the nutritional state, prior weight loss, indication for PEG placement, primary and secondary diagnoses, antibiotic prophylaxis before PEG placement, and minor complications after PEG placement. Thus, only a few factors were identified as possibly influencing hospital mortality after PEG insertion. A Cox proportional hazards regression analysis revealed peritonitis, aspiration pneumonia after more than 3 days, severe wound infection, and severe complications in general, as well as a primary diagnosis of coma, as the most significant risk factors for hospital mortality (Table 4). Of the 178 patients without severe complications, 10 (5.6%) died during their hospital stay. In the 19 patients with severe complications, hospital mortality was significantly higher (47.4%, n ¼ 9; P < .001). When comparing the patients who survived after 1 month with those patients who did not, a Cox proportional hazards analysis revealed coma as the primary diagnosis, and persistent or lateonset abdominal pain, severe wound infection, and severe

complications in general as being predictive for 1-month mortality after PEG insertion. In 97 (54.4%) of the 178 patients surviving the hospital stay, 6-month follow-up data were reported. There were no significant differences among the 6-month mortalities of the participating hospitals. No correlation was found between the 6-month mortality and the number of PEGs per participating hospital. Cox proportional hazard regression analysis revealed dementia as a secondary diagnosis, no oral nutrition at hospital discharge and age as significant risk factors for 6-month mortality in geriatric patients. Dependency for food intake at the time of hospital admission was related to a low rate of mortality 6 months after PEG insertion (Table 4). Of the patients with no severe complications with available follow-up data, 46 (46.5%) died within 6 months. In the 17 patients who experienced severe complications and had available follow-up data (2 of 19 were lost at follow-up), the 6-month mortality rate was significantly higher at 82.4% (n ¼ 14; P ¼ .006; Figure 2). Discussion Although it may seem intuitive that tube feeding improves survival, there is very little evidence for this in the literature.6 There is some evidence to suggest that PEG feeding may decrease mortality in specific subgroups, such as dysphagic patients with stroke13 or amyotrophic lateral sclerosis.14 However, PEG tubes may also introduce additional morbidity leading to adverse events, such as peritonitis, severe wound infections, and other complications. For instance, a higher prevalence of aspiration has been reported among tube-fed patients compared with those without tubes.15e17 Dysphagia (64%) and stroke (49%) appear to be the main indication for PEG insertion in this cohort of geriatric patients in Germany, which is consistent with the findings of a previous retrospective study.18 Dementia was the primary indication for PEG insertion in only 18.8% of the patients, which reflects the fact that dementia is not a frequent indication for PEG insertion in geriatric hospital departments and is in agreement with guidelines on enteral feeding in geriatric patients.19 The fact that the mean length of hospital stay was 32 days shows the high resource requirements of geriatric patients receiving PEG. In Germany, the mean length of hospital stay is 6 days in general and 16 days in acute care geriatric hospital departments, like the

Table 4 Risk Factors for Mortality after PEG Insertion (Cox Proportional Hazard Ratio)

Risk factors for hospital mortality (n ¼ 197) Eating independently at admission Coma Peritonitis Severe wound infection Late aspiration pneumonia Severe complications Risk factors for 1-month mortality (n ¼ 113) Coma Severe wound infection Late or persistent abdominal pain Any abdominal pain Severe complications Risk factors for 6-month mortality (n ¼ 85) Dementia as secondary diagnosis Age No oral nutrition at hospital discharge Dependent with meal intake at admission PEG, percutaneous endoscopic gastrostomy.


95% CI

P Value

1.09 19.32 33.06 6.92 5.09 6.88

0.32e3.69 2.15e173.62 3.73e293.25 1.92e24.93 1.42e18.25 2.62e18.06

.895 .008 .002 .003 .013 <.001

36.93 3.58 7.40 4.70 4.28

7.10e192.05 1.01e12.66 2.03e27.03 1.34e16.43 1.65e11.09

<.001 .048 .002 .016 .003

4.81 1.09 2.43 0.18

2.19e10.56 1.03e1.15 1.23e4.81 0.07e0.49

<.001 .004 .011 .001


Fig. 2. Survival rates with and without severe complications.


R. Wirth et al. / JAMDA 13 (2012) 228e233

participating units. This is because of the possibility of attending a rehabilitation procedure contemporaneous with acute care, which generates an extra payment for the hospital. The observation that the documented PEGs were inserted after a mean period of 17.5 days after hospital admission may be interpreted in two different ways. On one hand, these patients may have partly suffered complications during their hospital stay that necessitated PEG insertion. On the other hand, it may instead reflect the fact that geriatricians consider the indication for PEG insertion carefully and often begin with noninvasive treatment, such as food supplementation or nasogastric tube feeding, before conducting PEG insertion for long-term feeding. The 1- and 6-month mortality rates of 18.4% and 51.9%, respectively, in patients with PEG placement are to a great extent in concordance with the data provided by a meta-analysis on PEG placement in older people.6 Although the hospital mortality of 9.6% is within the expected range for multimorbid geriatric patients whose overall hospital mortality is about 5.0%,18 the 6-month mortality of 51.9% is strikingly high, probably mainly reflecting the severity of the disease of these patients. However, in contradiction to this hypothesis, we found no association between mortality and the patients’ functional, cognitive, and nutritional states, which would be expected if disease severity is the link to mortality. The prevalence of minor and severe complications was comparable with other studies and with a recent meta-analysis.7 There were no significant differences in the prevalence of complications or mortality among the participating centers, suggesting that patient selection and the quality of the endoscopic procedure were not relevantly different. Whereas short- and long-term mortality rates were not particularly high in patients with special indications or diagnoses, patients with severe complications comprised a large share of all patients with early mortality, ie, hospital and 1-month mortality. Patients with severe complications after PEG insertion showed high rates of hospital mortality and 6-month mortality (47.4% and 82.4%, respectively). Regarding the survival curve of patients with and without severe complications, it becomes clear that severe complications correspond to mortality within the first 30 days (Figure 2). After that period, the widely spaced survival curves run parallel to each other again. This association between complications and mortality gives rise to the suspicion that procedure-related mortality rates after PEG insertion in multimorbid geriatric patients are considerably higher than those reported in mixed populations and it suggests that complications after PEG insertion may have a significant, previously unrecognized influence on the early mortality of geriatric patients receiving PEG. For instance, some studies classified mortality to be procedure-related only if death occurred within a 48-hour period or if mortality was a result of complications beginning less than 48 hours after the procedure.11 However, wound infections often occur after that period and may be fatal, as reported in detail for one of our study participants.8 Five (2.5%) of the participants of this study developed severe wound infections without any other complication and 3 of them (60.0%) died in the hospital after a mean period of 16 days after PEG placement, demonstrating that severe wound infections are a significant risk factor for short-term mortality. In addition, 3 patients (1.5%) died unexpectedly within 1 day after PEG insertion without reported complications. Dying within 1 day after PEG insertion suggests that the death may be partly attributable to the procedure itself, even though complications were not reported. According to the regression analysis, peritonitis, aspiration pneumonia, and severe wound infections were the main risk factors for hospital mortality after PEG insertion. A pessimistic estimation of procedure-related mortality in this cohort has to include the 1 patient who died of peritonitis, the three patients who died after severe wound infections, the five patients who died after aspiration

pneumonia, and the three patients who died within 1 day after PEG insertion with no reported complications. This approach leads to an unexpectedly high procedure-related mortality rate of 6.1%. An optimistic estimation would include only patients with reported complications and exclude cases of aspiration pneumonia, which may be a complication of the underlying disease itself. Yet, even this approach should consider the one patient dying of peritonitis and the 3 patients dying after severe wound infections as procedure-related deaths, which would result in an estimated procedure-related mortality rate of 2%. This demonstrates that, although the overall mortality and the prevalence of complications in this study is comparable to that found in other studies, we found that the fatality rate in geriatric patients is increased, similar to what is observed for other vulnerable patient groups. For example, in patients with head and neck cancer, a 2.2% rate of procedurerelated mortality after PEG was reported.20 In addition, a national enquiry from the United Kingdom reported that 43% of all patients who died within 30 days after PEG insertion died within the first week, suggesting a relevant number of futile PEG insertions and significant morbidity being introduced by the procedure.21 Thus, the potential risks of PEG placement and feeding should be taken into account when patients and proxies are given medical advice about PEG insertion. This study has several limitations. One limitation is the incomplete follow-up data. As this study had only very minor financial support, the design was kept as simple as possible, and therefore included only an optional follow-up. Therefore, 45% of patients discharged from the hospital were lost at the 6-month follow-up, because some centers were not able to perform the follow-up, which may have biased the follow-up data. Second, there is no documentation about the cause of death, so that the supposed link between complications and mortality is based only on statistics. In addition, some of the baseline data, such as the cognitive and nutritional states, were obtained without validated measures, using only the clinical judgment of the attending physician, which partly limits the validity of the data.

Conclusion Although the prevalence of serious and minor complications is not increased in multimorbid geriatric patients, the procedurerelated mortality rate following PEG is at least 2%, which is higher than that of mixed patient populations. Consequently, the serious consequences of complications after PEG insertion in the vulnerable group of geriatric patients require careful and experienced consideration by the referring physician. Further studies should investigate whether or not alternative methods, such as nasogastric feeding and energy-dense supplements, lead to less complications and lower procedure-related mortality for certain indications.

References 1. Pirlich M, Schutz T, Norman K, et al. The German hospital malnutrition study. Clin Nutr 2006;25:563e572. 2. Bauer JM, Wirth R, Volkert D, et al. Malnutrition, sarcopenia and cachexia in the elderly: From pathophysiology to treatment. Conclusions of an international meeting of experts, sponsored by the BANSS Foundation. Dtsch Med Wochenschr 2008;133:305e310. 3. Loser C, Aschl G, Hebuterne X, et al. ESPEN guidelines on artificial enteral nutritiondpercutaneous endoscopic gastrostomy (PEG). Clin Nutr 2005;24: 848e861. 4. Gauderer MW, Ponsky JL, Izant RJ. Gastrostomy without laparotomy: A percutaneous endoscopic technique. J Pediatr Surg 1980;15:872e875. 5. Gauderer MW. Percutaneous endoscopic gastrostomy and the evolution of contemporary long-term enteral access. Clin Nutr 2002;21:103e110. 6. Mitchell SL, Tetroe JM. Survival after percutaneous endoscopic gastrostomy placement in older persons. J Gerontol A Biol Sci Med Sci 2000;55:735e739.

R. Wirth et al. / JAMDA 13 (2012) 228e233 7. Potack JZ, Chokhavatia S. Complications of and controversies associated with percutaneous endoscopic gastrostomy: Report of a case and literature review. Medscape J Med 2008;10:142. 8. Wirth R, Bauer J, Sieber C. Necrotizing Candida infection after percutaneous endoscopic gastrostomy: A fatal and rare complication. JPEN J Parenter Enteral Nutr 2008;32:285e287. 9. Schrag SP, Sharma R, Jaik NP, et al. Complications related to percutaneous endoscopic gastrostomy (PEG) tubes. A comprehensive clinical review. J Gastrointestin Liver Dis 2007;16:407e418. 10. Figueiredo FA, da Costa MC, Pelosi AD, et al. Predicting outcomes and complications of percutaneous endoscopic gastrostomy. Endoscopy 2007;39: 333e338. 11. Rimon E, Kagansky N, Levy S. Percutaneous endoscopic gastrostomy; evidence of different prognosis in various patient subgroups. Age Ageing 2005;34: 353e357. 12. Janes SE, Price CS, Khan S. Percutaneous endoscopic gastrostomy: 30-day mortality trends and risk factors. J Postgrad Med 2005;51:23e28. 13. Dennis MS, Lewis SC, Warlow C. Effect of timing and method of enteral tube feeding for dysphagic stroke patients (FOOD): A multicentre randomised controlled trial. Lancet 2005;365:764e772.


14. Mazzini L, Corra T, Zaccala M, et al. Percutaneous endoscopic gastrostomy and enteral nutrition in amyotrophic lateral sclerosis. J Neurol 1995;242:695e698. 15. Kadakia SC, Sullivan HO, Starnes E. Percutaneous endoscopic gastrostomy or jejunostomy and the incidence of aspiration in 79 patients. Am J Surg 1992; 164:114e118. 16. Peck A, Cohen CE, Mulvihill MN. Long-term enteral feeding of aged demented nursing home patients. J Am Geriatr Soc 1990;38:1195e1198. 17. Finucane TE, Bynum JP. Use of tube feeding to prevent aspiration pneumonia. Lancet 1996;348:1421e1424. 18. Wirth R, Volkert D, Bauer JM, et al. PEG tube placement in German geriatric wardsda retrospective data-base analysis. Z Gerontol Geriatr 2007;40:21e30. 19. Volkert D, Berner YN, Berry E, et al. ESPEN guidelines on enteral nutrition: Geriatrics. Clin Nutr 2006;25:330e360. 20. Grant DG, Bradley PT, Pothier DD, et al. Complications following gastrostomy tube insertion in patients with head and neck cancer: A prospective multiinstitution study, systematic review and meta-analysis. Clin Otolaryngol 2009;34:103e112. 21. Johnston SD, Tham TC, Mason M. Death after PEG: Results of the National Confidential Enquiry into Patient Outcome and Death. Gastrointest Endosc 2008;68:223e227.