When should nutritional support be started in patients with acute pancreatitis?

Nutritional support in acute necrotising pancreatitis

The provision of nutritional support for the patient with ANP is an essential component of supportive therapy, especially since many patients with pancreatitis are nutritionally depleted prior to their illness and face increased metabolic demands throughout the course of their disorder. Failure to reverse or prevent malnutrition, and a prolonged negative nitrogen balance, increases mortality rates. The route by which nutrition is administered is, none the less, still debated. In the last two decades there has been a trend away from the use of total parenteral nutrition (TPN) in favour of enteral nutrition (EN) in supporting the critically ill. Studies suggest that TPN started within 24 hours of admission to ICU, compared with a watch-and-wait policy and/or EN where tolerated is associated with increased infective complications and hospital length of stay.27

TREATMENT

Management of pancreatitis is generally supportive, although identification and control of any underlying treatable causes should be pursued. Bowel rest, fluid replacement and management, nutritional support, and pain control are central to supportive care.

Eliminating oral or gastric intake is usually sufficient for bowel rest. However, bowel decompression via nasogastric suction may be needed for patients with persistent vomiting. Fluid replacement and correction of electrolyte imbalances are key in the initial phase. In severe or necrotizing pancreatitis, tremendous amounts of fluid can accumulate in and around the inflamed pancreas and lead to requirements for large amounts of intravenous fluid. Careful monitoring of hemodynamic status, urine output, and serum electrolytes is warranted.

Total parenteral nutrition (TPN) should be initiated if enteral feeding is expected to be withheld for more than 3 to 5 days. Studies suggest that intrajejunal feeding is as safe and as well tolerated as TPN in acute pancreatitis. Such enteral feeding may moderate the acute phase response and thereby decrease the risk for septic complications or the systemic inflammatory response syndrome. Intestinal barrier integrity is better maintained, which results in decreased bacterial translocation when the gut is fed. In addition, placement of a central venous catheter can often be avoided. A low-fat elemental formula (e.g., Vivonex TEN) should be used when feeding is directly into the jejunum.

Pain control can be provided with hydromorphone hydrochloride (Dilaudid), which is preferred over morphine. Morphine is believed to produce contraction of the sphincter of Oddi and therefore has the potential to interfere with drainage of biliary and pancreatic secretions.

Treatable causes of pancreatitis should be identified and addressed. This could include the discontinuation of potentially toxic drugs such as valproic acid or other antiepileptic drugs, chemotherapeutic agents, estrogens, corticosteroids, or diuretics (Table 100-2). Metabolic causes of pancreatitis, such as hypertriglyceridemia, hyperlipidemia, and hypercalcemia, should be addressed and treated.

A sudden rise in pancreatic enzymes, persistent elevation, or worsening/persistence of symptoms suggests the development of a pseudocyst. Pseudocysts complicate the course in approximately 10% of patients unless associated with abdominal trauma, in which case they occur in more than 50% of patients. Serial ultrasound often reveals evidence of pancreatitis not identified at presentation or earlier in the disease course. In all but the mildest self-limited cases, it is advisable to repeat an ultrasound after approximately 4 weeks to evaluate for possible pseudocyst formation because they usually become evident 2 to 3 weeks after the acute attack. Pseudocysts are most often (85%) located in the body or tail of the pancreas. Most frequently associated with trauma, pseudocysts generally resolve spontaneously, and medical management with several weeks of TPN or jejunal feeding may be necessary. Surgical, endoscopic, or image-guided percutaneous drainage should be undertaken if the pseudocyst is increasing in size or continuing to produce pain.

Surgical intervention should be avoided in the acute setting unless intervention for obstruction of biliary or pancreatic ducts is deemed necessary. ERCP with sphincterotomy, balloon dilation, or stent placement may be an option in this setting. Usually, correction of anatomic defects should be scheduled electively after recovery from the acute episode. Partial pancreatectomy or pancreaticojejunostomy may be indicated in patients with intrapancreatic ductal strictures.12,13

Bacterial superinfection is a serious complication of pancreatitis and is more likely to occur in the setting of necrosis of pancreatic tissue and more commonly later in the clinical course. Strategies to prevent pancreatic infection include enteral administration of nonabsorbable antibiotics to decontaminate the gut lumen and prophylactic systemic antibiotics. Although promising, studies using these modalities in adults have yielded conflicting results. However, some experts recommend administration of prophylactic intravenous antibiotic therapy to patients with proven necrotizing pancreatitis. Because coverage of enteric gram-negative organisms is the goal, piperacillin-tazobactam and imipenem are good choices.

The timing of reintroducing oral (or gastric) feeding is not absolute, and it is not necessary to wait for normalization of pancreatic enzymes. Reintroduction of oral feeding may be considered when pain and ileus have resolved and the patient regains an appetite. A clear liquid diet can be started and advanced with reintroduction of fat from 10 to 20 g/kg daily as tolerated. A low-fat diet may be prescribed until complete recovery of acute pancreatitis, but this restriction does not appear to affect the progression of chronic pancreatitis and may contribute to malnutrition if used chronically. In chronic pancreatitis, it is advisable to administer pancreatic enzyme supplementation with each meal to reduce digestive stress on the pancreas. In children, the amount of replacement enzyme should not exceed 2000 U lipase/kg to avoid complications such as fibrotic strictures of the colon, particularly in chronic pancreatitis. Complete correction is not readily achieved. Proton pump inhibitors should be administered as long as pancreatic enzyme supplements are given to prevent acid-induced inactivation of lipase in the stomach and duodenum.2

The Ranson criteria (Table 100-3) and other systems, such as the APACHE III (Acute Physiology and Chronic Health Evaluation) scoring system, have been used to predict the prognosis in adults. They have not been evaluated in pediatric patients, and pediatric risk stratification systems need further study.14

Enteral nutrition – of value in AP?

The use of early enteral nutrition in AP has been studied in a number of publications and summarized findings point to its feasibility, cost reduction, a decrease in septic complications, hospital stay, modulation of the inflammatory response and facilitation of gut function. Weaknesses, though, have been that most studies have included only a limited number of patients, with in general a substantial delay before the initiation of enteral nutrition, using a non-defined enteral formula and a mix of patients with varying severity of disease [54–59]. By using the addition of probiotics (Lb. Plantarum 299) to enteral nutrition in patients with prognosticated severe AP, the incidence of sepsis and need for surgical intervention could be significantly decreased, as evaluated in 45 patients [60]. By tradition, a more distal positioning of the enteral tube has been advocated, aiming at reaching a position of the tip at about the level of the ligament of Treitz. Enteral nutrition provided through a nasogastric tube has also been successful in severe AP in 22 of 26 patients [61]. The use of nasogastric early enteral nutrition has also been demonstrated to be feasible and without side effects as compared with traditional treatment of ‘pancreatic rest’, and this regime provided better blood glucose control in patients with prognosticated severe AP [62]. In patients with mild AP, a recent randomized clinical study has demonstrated that by allowing patients to eat directly, hospital stay decreased by one-third without any notable side effects or increase in, for example, refeeding pancreatitis [63].

Nutritional Support

The most recent Cochrane meta-analysis of enteral versus parenteral nutrition in acute pancreatitis, published in 2010, included eight RCTs.24 Overall, there was a reduction in mortality, systemic infection, and multiorgan dysfunction in pooled analyses for the patients receiving enteral nutrition; therefore, enteral nutrition is strongly preferred over parenteral nutrition for patients with severe acute pancreatitis. Several trials, including one non-inferiority study, have failed to detect a difference between nasogastric compared to nasojejunal route of enteral nutrition,25–27 indicating that early enteral nutrition through nasogastric feeding may be sufficient. A multicenter study from the University of Pittsburgh (Study of Enteral Nutrition in Acute Pancreatitis) is underway to address the optimal form of enteral nutrition in acute pancreatitis. In addition, a multicenter prospective RCT is underway in the Netherlands to address the impact that timing of enteral nutrition has on outcome in patients with severe acute pancreatitis. Although initially demonstrating promising results in small nonrandomized clinical trials, the addition of probiotics to enteral nutrition has been shown to be potentially detrimental in a large-scale RCT among patients with severe acute pancreatitis28; therefore, use of probitics is not recommended.

Results

With the search strategy 254 publications were retrieved from which 13 RCTs comparing EN versus PN in acute pancreatitis were identified. Seven trials43–49 were excluded because the principal outcome variables (incidence of hyperglycemia and/or need for insulin therapy) were not reported. Eventually, six RCTs13,21–25 satisfied all of the inclusion criteria. Figure 1 shows the details of the search, the selection process, and the reasons for exclusion. Four of six studies included only patients with severe acute pancreatitis,13,22,24,25 while the study population of two other trials consisted of both mild and severe patients.21,23 The funnel plots for the different outcomes show no clear pattern of bias (data not presented). Table 1 presents the study characteristics for the included trials. A total of 264 non-diabetic patients with acute pancreatitis were included in the systematic review, of whom 126 received EN and 138 received PN. Four of six RCTs reached the Jadad quality score grade 3 (Table 2). Table 3 displays the demographic data of patients. The outcome data for the six studies are presented in Table 4.

Table 1. Summary of study characteristics for the included trials.

Table 2. Quality assessment of trials included in the systematic review.

Table 3. Baseline characteristics of patients.

Table 4. Summary of outcomes data of trials included in the systematic review.

All six trials registered the need for insulin therapy. Fifty-six patients who were not previously diabetics (21.2%) developed hyperglycemia requiring insulin; 15 of 126 (11.9%) with EN and 41 of 138 (29.7%) with PN. In absence of marked heterogeneity across studies, EN resulted in a statistically significant reduction in risk for insulin therapy (RR 0.41; 95% CI 0.24–0.70; p=0.001) (Fig. 2). This risk reduction was also significant when only patients from RCTs on severe acute pancreatitis were considered (Table 5).

Table 5. Results of meta-analysis.

Three of six RCTs presented hyperglycemia data. All three studies included only patients with severe acute pancreatitis. Ninety-four patients without prior diabetes mellitus (60.3%) developed hyperglycemia: 31 of 76 (40.8%) in the EN group and 63 of 80 (78.8%) in the PN group. With a marked heterogeneity across studies, EN resulted in a statistically significant reduction in risk for hyperglycemia (RR 0.53; 95% CI 0.29–0.98; p=0.04) (Fig. 3).

Total infectious complications were reported in all six trials. In the absence of marked heterogeneity across studies, EN resulted in a statistically significant reduction in risk for total infectious complications (Table 5). The same applies for the risk reduction outcome when only patients from RCTs on severe acute pancreatitis were considered (Table 5).

Mortality was reported in six RCTs. With a marked heterogeneity across studies, EN was not associated with significant reduction in mortality risk (Table 5). However, when only patients from RCTs on severe acute pancreatitis were considered, there was no marked heterogeneity across studies and EN resulted in a statistically significant reduction in risk for mortality (Table 5).

Mild and moderate pancreatitis

Mild and moderate pancreatitis is usually self-limiting, responds to short periods of bowel rest, and resolves before any significant malnutrition occurs. In a study by Louie and colleagues [3] comparing the route of EN to PN in acute pancreatitis, 184 patients had a Ranson's severity score of 3 or greater; of these, 120 patients tolerated oral intake within 5 days of admission. Sax and colleagues [12] reported that 80% of patients who had Ranson's criteria less than or equal to 2 tolerated an oral diet by hospital day 7. In addition, 54% of patients who had Ranson's criteria greater than or equal to 2 tolerated a diet by hospital day 7. Abou-Assi and colleagues [4] treated patients with bowel rest, intravenous fluids, and analgesics for 48 hours after establishing a diagnosis of acute pancreatitis. Seventy-five percent of these patients improved within this time period without the need for specialized nutrition support. Furthermore, 87% of patients resumed an oral diet within the next 2 days. These studies all support withholding specialized nutrition support in mild and moderate pancreatitis, since an oral diet can be instituted within a short period of time after the initial diagnosis.

Composition of enteral nutrition and immunonutrition

There is a wide range of EN formulations with varying purported benefits in AP and, for simplicity, is divided into three categories: (1) oligomeric feeds, (2) polymeric feeds, and (3) “immunonutrition.” Oligomeric, also known as semielemental, formulations contain small peptides, medium-chain fatty acids, and simple polysaccharides that do not require digestion by pancreatic enzymes and theoretically offer greater pancreatic rest than more complex polymeric formulations that contain full proteins, complex lipids, and carbohydrates.33,34 Two meta-analyses, however, comparing oligomeric with polymeric formulations found no difference in terms of feeding intolerance, mortality, or LOS between the two formulations.35 There is no apparent clinical advantage to the use of oligomeric formulations and more inexpensive polymeric formulations should be readily used.

Immunonutrition broadly refers to specialized formulations containing immunomodulatory supplements that are thought to offer benefit by modifying the immune response associated with AP. The most well studied of these include formulations supplemented with one of either glutamine, arginine, omega-3 fatty acids, nucleotides, and fiber enrichment. Trials in other clinical settings involving critically ill patients given immunonutrition-supplemented EN, specifically with glutamine and arginine, have described trends toward lower infectious complications and mortality compared with standard EN.36–38 The benefit of immunonutrition-supplemented EN in AP is less established and based on low-quality studies. A Cochrane systematic review by Poropat and colleagues39 of 15 trials investigating EN formulations specifically containing immunonutrition components given to patients with AP found no difference in all-cause mortality or occurrence of SIRS when compared with other EN formulations. Nearly all the trials in this review were noted to be of low quality with a high risk for bias. A separate meta-analysis by Petrov and colleagues40 similarly did not demonstrate a clinical benefit with immunonutrition formulations in regard to LOS or infectious complications.

There are trials, largely from China, showing clinical benefit of PN supplemented with glutamine or glutamine administered intravenously in SAP with regards to lower infectious complications, LOS, and resolution of inflammatory markers.41–44 Glutamine is postulated to exert an immunomodulatory effect via increasing lymphocyte mitogenic function, whereas decreasing production of proinflammatory cytokines, such as interleukin-6 and tumor necrosis factor-α and antioxidant properties. Glutamine also supports the growth of other rapidly dividing cells, such as enterocytes.45 These benefits have not been demonstrated in trials with glutamine-supplemented EN,43,46 currently limiting the clinical relevance of glutamine in AP until further investigation with high-quality trials can be performed.

Patients with AP are known to have gut dysbiosis or unfavorable imbalance of gut microbiota that may contribute to associated inflammation.3,47 The use of probiotics, substances containing live microorganisms of healthy gut flora, however, has been shown to be detrimental in AP. The PROPATRIA trial, a multicenter, double-blind, placebo-controlled RCT of nearly 300 patients with predicted SAP conducted in the Netherlands aimed to reduce infectious complications in patients with SAP through the use of enteral probiotic preparations. The study compared the use a multispecies mixture of two different Bifidobacterium species, three different Lactobacillus species, and one Lactococcus species with placebo. Findings from the trial demonstrated no significant difference in the primary end point of infectious complications and a two- to three-fold increase in mortality in patients who received probiotics.48 Until more studies can establish an acceptable safety margin and consistent dosing for probiotic administration, their use should be avoided in AP.

The benefit of immunonutrition in AP is currently unclear and, as recommended by the AGA, warrants further investigation with high-quality RCTs to support routine use. Current evidence is lacking in supporting the use of immunonutrition-supplemented EN but there may be some benefit of glutamine-supplemented PN in patients with SAP requiring PN. A Cochrane systematic review and network meta-analysis by DiMartino and colleagues49 is ongoing to further clarify the benefit of immunonutrition supplementation in EN and PN in AP.

AP-V.3. Nutrition

The nutrition of children with AP is carried out according to recommendations based on clinical experience in adults and the limited studies in pediatrics. No RCTs have been done in children to address the role of nutrition in AP [18, 22, 23, 29, 30].

Influence of Disease Severity on Gut Integrity and Likelihood for Enteral Nutrition to Change Outcome

As has been shown in animal models [31,32] and in clinical studies involving other disease processes (eg, trauma) [33], pancreatitis patients with greater disease severity have greater degrees of intestinal permeability and are more likely to sustain a favorable response to early EN. In a study of patients with acute pancreatitis using urinary excretion of polyethylene glycol as a marker of intestinal permeability, Ammori and coworkers [34] showed that patients with mild pancreatitis had a degree of permeability that was no greater than that seen in controls without pancreatitis. Patients with severe but uncomplicated pancreatitis had a fourfold increase in permeability as measured by urinary polyethylene glycol levels, whereas patients with severe pancreatitis complicated by multiple organ failure had a fourfold higher increase in permeability [34]. Clearly, greater degrees of disease severity are associated with greater degrees to which the channels between the intestinal epithelial cells open and gut integrity is lost. Over the first four PRCTs evaluating EN versus PN in acute pancreatitis, the percentage of patients with severe pancreatitis seemed to be a key factor in whether the EN significantly altered clinical outcome [21–23,25]. In our Louisville study, only 19% of the study patients had severe pancreatitis, and no differences in clinical outcome were seen between the two routes of nutrition support [21]. In the studies by Abou-Assi and coworkers [25] and Windsor and coworkers [23], 35% to 38% of patients had severe pancreatitis and results indicated that early use of EN hastened resolution of SIRS and decreased the time to resolution of the disease process compared with PN [23,25]. In a Greek study by Kalfarentzos and coworkers [22], where 100% of patients had severe pancreatitis with necrosis on CT scan, early EN was shown to reduce the number of patients with complications significantly from 75% to 44%, and percent of patients with septic morbidity from 50% to 28% compared with use of PN.

Patients with severe pancreatitis (designated by APACHE II score ≥10 and ≥3 Ranson's criteria) had a high rate of complications (38%) and mortality (19%), and close to 0% chance of advancing to oral diet within 7 days [35–38]. For these patients, EN may be expected to change clinical outcome. Patients with lower scores have mild to moderate pancreatitis; have a low rate of complications (6%); negligible mortality; and a greater than 80% chance of advancing to an oral diet within the first week of hospitalization [35–38]. The former group of patients needs specialized nutrition support and are most likely to experience an improved outcome with early enteral feeding. The latter group, in contrast, does not need specialized nutrition support, is not expected to experience improved outcome by the enteral or parenteral route of feeding, and instead simply needs supportive care with intravenous analgesia and fluid resuscitation.