Maxim S Petrov1, M Isabel TD Correia2, John A Windsor3
1Department of Surgery, Nizhny Novgorod State Medical Academy. Nizhny Novgorod, Russia
2Department of Surgery, Federal University of Minas Gerais. Belo Horizonte, Brazil.
3Department of Surgery, University of Auckland. Auckland, New Zealand
Received January 28th, 2008 - Accepted April 15th, 2008
Context Nasogastric tube feeding is safe and well tolerated in most critically ill patients. However, its safety and tolerance in the setting of severe acute pancreatitis is debatable. Objective We aimed to review all available studies on nasogastric feeding in patients with severe acute pancreatitis to determine the safety and tolerance of this approach. A further aim was to perform a meta-analysis of the available randomized controlled trials regarding nasogastric versus nasojejunal feeding. Methods Three electronic databases (Cochrane Central Register of Controlled Trials, EMBASE and MEDLINE) and the abstracts of major gastroenterological meetings were reviewed. Meta-analysis was performed using the random effects model. Main outcome measures The summary estimates were reported as risk ratio (RR) with 95% confidence interval (95% CI). Results A total of four studies on nasogastric tube feeding in 92 patients with predicted severe acute pancreatitis were identified. Documented infected pancreatic necrosis developed in 11 patients (16.9%) and multiple organ failure in 10 (15.4%) out of 65 patients with available data. Overall, there were 15 deaths (16.3%). An exacerbation of pain after initiation of feeding occurred in 3 (4.3%) out of 69 patients with available data. Full tolerance was achieved in 73 (79.3%) patients who did not require temporary reduction, stoppage or withdrawal of nasogastric feeding. The results of nasogastric feeding as compared to nasojejunal feeding, were no worse in terms of mortality (RR=0.77; 95% CI: 0.37 to 1.62; P=0.50) or intolerance of feeding (RR=1.09; 95% CI: 0.46 to 2.59; P=0.84). Conclusion Nasogastric feeding appears safe and well tolerated in patients with predicted severe acute pancreatitis. An adequately powered randomized trial on nasogastric versus nasojejunal feeding is required to support this approach as routine clinical management.
Enteral Nutrition; Pancreatitis, Acute Necrotizing /complications
RCT: randomized controlled trials
Nutritional support plays an important role in the management of patients with severe acute pancreatitis [1, 2, 3]. It has been convincingly demonstrated in numerous studies that enteral nutrition is preferable to parenteral nutrition as it leads to significantly better glycemic control and decreases infectious complications and mortality [4, 5, 6, 7]. With these apparent benefits, the question has been to determine the most optimal site of tube feeding administration. The alternatives include nasojejunal and nasogastric tube placement. The former requires the assistance of an endoscopist or a radiologist which may result in a delay in commencing enteral nutrition. This delay may have an impact on the clinical outcome as it is now believed that enteral nutrition should commence as soon as possible after hospital admission in order to maximize clinical benefits . In contrast, a nasogastric feeding tube can be inserted immediately and with ease so that pre-pyloric feeding can be started without delay.
A number of randomized controlled trials (RCTs) and the latest meta-analysis  have demonstrated the equivalence of nasogastric and nasojejunal tube feeding in terms of safety and tolerance in critically ill patients. While this may be true for this group of patients in general, it is recognized that patients with severe acute pancreatitis are particularly prone to gastric ileus because of the subjacent inflamed pancreas . This has been given as a reason for providing enteral nutrition into the jejunum [1, 11]. Another reason given to avoid the provision of enteral nutrition proximal to the jejunum has been the concern that pancreatic exocrine stimulation will result in an increased severity of the acute pancreatitis . Most studies reveal that the majority of patients receiving enteral nutrition for severe acute pancreatitis have received nasojejunal tube feeding, but there are some studies of successful nasogastric feeding [13, 14, 15].
The only review on nasogastric feeding in acute pancreatitis  attempted to define the feasibility of this route of nutrition by metaanalyzing the data from RCTs on nasogastric versus ‘conventional’ nutrition. The pooled estimates and variance of the treatment effect were based on the statistical aggregation of the results from studies with essentially different comparators (i.e., total parenteral and nasojejunal tube feeding). Such an approach might be misleading as nasojejunal, but not parenteral, nutrition is now considered the therapy of choice in severe acute pancreatitis. Moreover, there was a marked heterogeneity in baseline risk among the studies involved in that meta-analysis, particularly in regards to age and gender ratio [15, 17], and incorrect pooled estimates were presented due to inaccurate data input [14, 15]. Furthermore, the previous review  did not determine the efficacy of nasogastric tube feeding alone.
The aim of this systematic review was to assess the relative efficacy of nasogastric versus nasojejunal feeding in severe acute pancreatitis, and to determine the safety and tolerance of nasogastric tube feeding. This was done by analyzing all of the literature (randomized and non-randomized studies) relating to acute pancreatitis and nasogastric tube feeding.
A computerized literature search of the Cochrane Central Register of Controlled Trials, EMBASE and MEDLINE through December 1st, 2007 was conducted. The key words for Cochrane Central Register of Controlled Trials were “acute pancreatitis” and “nutrition”. The key words for EMBASE included the terms “acute pancreatitis” and “enteral nutrition” or “enteral feeding”. The same key words used for EMBASE were used for MEDLINE. No language restrictions were applied. The bibliographies of all selected articles found which included information on nasogastric tube feeding in acute pancreatitis were reviewed in an attempt to find other relevant articles. The abstracts of major pancreatology meetings (Digestive Disease Week (DDW), United European Gastroenterology Week (UEGW), International Hepato-Pancreato-Biliary Association (IHPBA), and European Pancreatic Club (EPC)) until 2007 were also manually screened.
The following selection criteria were used to identify published studies for inclusion in this systematic review: study design: cohort study or RCT; population: patients with predicted severe acute pancreatitis; intervention: nasogastric tube feeding; outcome: at least one of the following: tolerance, organ failure, infectious complications, and mortality.
The records extracted by the initial search were scanned to exclude obviously irrelevant studies. Full-text articles were retrieved and reviewed by two authors (MSP, MITDC) independently with the aim of applying inclusion criteria. All differences found between the two reviewers were resolved by discussion among the three authors of this paper.
The methodological quality of the randomized studies included was estimated using a modification of the previously published quality score . It consists of 8 parameters with a quality score range from 0 to 16 points (Table 1).
Chosen a priori, a meta-analysis (Review Manager - RevMan (Computer program); Version 5.0. Copenhagen, The Nordic Cochrane Centre, The Cochrane Collaboration, 2008) was carried out on the data from the RCTs of nasogastric versus nasojejunal feeding. The outcome data were combined to determine the risk ratio (RR), with its 95% confidence interval (95% CI). The presence of heterogeneity was assessed using the I2 measure, with I2 greater than 25% indicating significant heterogeneity . Irrespective of the degree of heterogeneity of effect among the trials included, a random effects model was used.
A total of 396 publications were identified using the above-mentioned search strategy. Of these, 392 articles did not meet the inclusion criteria and were subsequently excluded. Figure 1 details the selection process. A total of 4 studies were included in this review [13, 14, 15, 17]. One study was a cohort study  whereas the other three studies were RCTs [14, 15, 17]. The control groups for the RCTs were nasojejunal feeding in two [14, 15] and parenteral feeding in the third one . Table 2 demonstrates the characteristics of the studies included in this systematic review, including the assessment of study quality. Overall, 93 patients with predicted severe acute pancreatitis were enrolled in these trials. The severity of the patients at admission was comparable in all 4 cohorts, based on APACHE II scoring. Table 3 presents the baseline characteristics of the patients who received nasogastric tube feeding. One patient had a protocol violation; therefore, the results were available in 92 patients (the characteristics of the protocol violator are unknown).
The main clinical outcomes of the studies are summarized in Table 4. Thirty-five of the 92 patients (38.0%) required ventilatory support. There was no evidence of aspiration pneumonia in any of the patients. The other outcomes were available in 3 studies only (65 cases): infected pancreatic necrosis was revealed in 11 patients (16.9%), multiple organ failure developed in 10 patients (15.4%), and 12 patients underwent surgery (18.5%). The mortality rate was 16.3% (15 cases).
Nasogastric feeding-related outcomes, including safety and tolerance, are presented in Table 5. Full tolerance was achieved in 73 of the 92 (79.3%) patients who did not require temporary reduction, stoppage or withdrawal of nasogastric feeding. The 19 patients who had a modification of the nasogastric tube feeding regimen presented signs of gastric ileus (n=6), diarrhea (n=10), or repeatedly removed their feeding tube (n=3). Three patients out of 69 (4.3%) experienced an exacerbation of pain after commencement of nutrition, although it was not a reason for the discontinuation of tube feeding in any of the cases.
The meta-analysis was restricted to the randomized studies of nasogastric versus nasojejunal feeding. In two eligible trials [14, 15], a total of 43 patients received enteral nutrition via the nasogastric route and 36 patients via the nasojejunal route. The use of nasogastric feeding resulted in a nonsignificant reduction in the risk of death (RR=0.77; 95% CI: 0.37 to 1.62; P=0.50) (Figure 2). The number of nutritionassociated adverse events was similar between the two groups. As a consequence, nasogastric feeding was associated with nonsignificant increases in the risk of diarrhea (RR=1.42; 95% CI: 0.46 to 4.38; P=0.54) (Figure 3) and the exacerbation of pain following feeding (RR=1.74; 95% CI: 0.24 to 12.79; P=0.58) (Figure 4). Overall, patients in both groups did not differ significantly in terms of an intolerance to feeding (RR=1.09; 95% CI: 0.46 to 2.59; P=0.84) (Figure 5). There was no heterogeneity between the study results for all comparisons (I2=0%).
This systematic review of the literature demonstrates the safety and tolerance of nasogastric tube feeding in 4 out of 5 patients with predicted severe acute pancreatitis. The clinical outcomes were within the expected range. Nasogastric tube feeding-related problems occurred in about 20% of patients, but they were relatively minor. The metaanalysis also demonstrated that there was no difference between nasogastric and nasojejunal tube feeding with respect to safety and tolerance in the two available RCTs. However, there were limitations in both the RCTs. Some data on certain essential clinical outcomes were not presented in the RCT from the United Kingdom . In addition, it was noted that it is likely that jejunal tube feeding in this trial was probably duodenal (because true jejunal placement would have been difficult with the types of feeding tubes and placement techniques used) meaning that both feeding arms may have been equally proinflammatory . The defect of the RCT from India  was that there was a considerable delay (7.8±6.5 and 5.7±4.7 days after symptom onset in the nasogastric and nasojejunal groups, respectively) in commencing enteral nutrition. Apart from these concerns, both trials were insufficiently powered to detect any difference or to prove equivalence between the groups studied in terms of any clinical outcome. For example, an adequately powered RCT would need to enrol 153 patients per arm in order to show a decrease in mortality from 16% (average rate in the nasogastric group in the present review) to 6% (best results in the nasojejunal group of RCTs on enteral versus parenteral nutrition [21, 22]) with 80% power and alpha value equal to 0.05 (two-sided).
One of the most important issues in considering tube feeding in acute pancreatitis is the effect of nutrition on pancreatic exocrine function. It was shown by O’Keefe et al.  that all forms of enteral nutrition stimulate pancreatic secretion. In particular, when compared to placebo-saline, an oral liquid polymeric diet resulted in a significantly higher level of amylase (P<0.01) and lipase (P<0.01); a duodenal polymeric enteral formula led to an increased level of amylase (P<0.01), lipase (P<0.01) and trypsin (P<0.01); a duodenal elemental feeding formula resulted in an elevated level of lipase (P<0.05). The same research group also compared pancreatic secretory response to tube feeding delivered into the duodenum, mid jejunum (40-60 cm distal to the ligament of Treitz), and distal jejunum (100-120 cm distal to the ligament of Treitz) . Even though the authors did not find a direct relationship between the decrease in enzyme secretion and the distance to the mid-distal jejunum, they demonstrated a significantly lower secretion of trypsin (P<0.01) and lipase (P<0.05) in response to the elemental formula delivered into the jejunum (40 cm or more distal to the ligament of Treitz) in comparison with the duodenum. Moreover, the trypsin and lipase secretory response in the mid-distal jejunum group was as low as in the control group (fasting).
It should be noted, however, that these studies [23, 24] on the effects of enteral feeding on pancreatic exocrine function were in healthy subjects. There is now convincing evidence  that patients with acute pancreatitis have significantly lower rates of pancreatic enzyme secretion into the duodenum as compared to healthy subjects. Furthermore, when patients with mild/moderate acute pancreatitis were compared to those with severe acute pancreatitis, a lower secretion of trypsin (6- fold), amylase (22-fold) and lipase (42-fold) was found in the latter group, suggesting that duodenal secretion of pancreatic enzymes is inversely related to the severity of acute pancreatitis . In line with this finding, another study  showed 86% rate of pancreatic exocrine insufficiency (measured by fecal pancreatic elastase-1) in patients recovering from severe attacks of acute pancreatitis. Moreover, the severity of pancreatic exocrine insufficiency correlated with the extent of pancreatic necrosis. These data suggest that injured acinar cells are not able to fully respond to the physiological stimuli of secretion into the duodenum and this may be of some help in explaining the findings of this review, namely, that nasogastric tube feeding does not appear to aggravate the severity of acute pancreatitis.
In conclusion, the present review has appraised the current evidence regarding nasogastric tube feeding to patients with predicted severe acute pancreatitis. The evidence base is small but does show that enteral nutrition administered by means of the nasogastric route is safe and well tolerated in 79% of patients with predicted severe acute pancreatitis. The statistically aggregated data from two randomized trials on the direct comparison of nasogastric and nasojejunal nutrition are encouraging as they demonstrate no difference in safety and tolerance between these routes. An adequately powered randomized trial on nasogastric versus nasojejunal feeding is required to support this approach before early nasogastric tube feeding can be established as the standard of care.
The authors have no potential conflicts of interest