Early nasogastric feeding versus parenteral nutrition in severe acute pancreatitis: A retrospective study

Objective: To compare the efficacy and safety of early nasogastric enteral nutrition (EN) with total parenteral nutrition (TPN) in patients with severe acute pancreatitis (SAP). Methods: From July 2008 to July 2014,185 patients with SAP admitted to our centre were enrolled in this retrospective study. They were divided into EN group (n=89) and TPN group (n=96) based on the nutrition support modes. Patients in EN group received nasogastric EN support, while patients in TPN group received TPN support within 72 hours of disease onset. The medical records were reviewed and clinical factors were retrospectively analyzed. Results: There were no significant differences in baseline characteristics between two groups. EN group had significantly lower incidence of pancreatic infections (P=0.0333) and extrapancreatic infections (P=0.0431). Significantly shorter hospital stay (P=0.0355) and intensive-care stay (P=0.0313) were found in EN group. TPN group was found to have significantly greater incidence of multiple organ dysfunction syndrome (MODS) (P=0.0338) and mortality (P=0.0382). Moreover, the incidence of hyperglycemia was significantly higher in TPN group (P=0.0454). Conclusions: Early nasogastric EN was feasible and significantly decreased the incidence of infectious complications as well as the frequency of MODS and mortality caused by SAP.

(TPN) with no beneficial effects on intestinal permeability or the inflammatory response. 7 On the basis of aforementioned facts, we conducted this retrospective study to evaluate the efficacy and safety of early nasogastric enteral nutrition as compared with TPN in patients with SAP.

Patients:
Between June 2008 and June 2014, 185 patients admitted to First People's Hospital affiliated to Huzhou University Medical College (Zhejiang Province, China) with a clinical diagnosis of SAP within 48h of disease onset were enrolled in this retrospective study. According to the nutrition support modes, they were divided into two groups: 96 in TPN group and 89 in EN group.
The diagnostic and classification criteria for SAP included at least two of the following features: 8 1. Abdominal pain consistent with AP; 2. Amylase activity at least four folds of the upper limit of normal; and 3. Characteristic findings of AP on contrastenhanced computed tomography (CT) and, less commonly, on magnetic resonance imaging or transabdominal ultrasonography. The exclusion criteria included the following: 1. Previously with cardiac failure or pulmonary edema; 2. Pregnant; 3. Underwent early surgical treatment; 4. Known allergy to any of the ingredients; 5. With poorly controlled diabetes This study was conducted in accordance with the principles of the Declaration of Helsinki and "Good Clinical Practice" guidelines. Written informed consent was obtained from all patients and study protocol was approved by the Institutional Review Board. Treatment: Patients were monitored and treated according to clinical routine including pain control, symptomatic and organ supportive treatment. Broad-spectrum antibiotic therapy was used before microbiological examination. 9 Intervention was indicated when infection of pancreatic necrosis was proven by the fine needle aspiration or when sepsis persisted despite maximal support on the intensive care unit. Patients with infected pancreatic necrosis were operated upon (necrosectomy and open packing or closed continuous lavage). Patients with pancreatic abscess were an indication for percutaneous drainage.
Nutrition support was begun within 72 hours of disease onset for a minimum of seven days in both groups, supplying daily 30 kcal/kg and 1.5 g/kg of protein, based on admission weight. In both groups, standard formulas without specific immunomodulating nutrients were used. TPN (Kabiven PI, Sino-Swed LTD., Wuxi, China) was delivered through a central venous catheter, initially infused at half of the calculated energy requirements; then increased over 48 hours to achieve 100% of the target energy rate. EN (Fresubin original, Sino-Swed LTD.) was administrated through a nasogastric feeding tube (size 8FG Flocare polyurethane feeding tubes, Nutricia LTD., Trowbridge, UK), at a rate of 25 ml/hour and increased by 10 ml/hour every 6 hours, until the desired caloric intake was reached. The NG feeding tubes were inserted on the ward by the medical staff. The position of the feeding tube was checked by aspiration and pH measurement. A chest radiograph was performed if necessary. The aim was to reach full nutrition within 48 hours after initiation. Oral feeding was reintroduced when amylase and CRP levels had decreased and abdominal pain had resolved. Regular hospital diet was introduced gradually, in general initially starting with liquid and then solid food. Outcome Measurements: The medical documents were reviewed for data collection. The incidence of pancreatic infectious complications (i.e. infected pancreatic necrosis and pancreatic abscess), extrapancreatic infectious complications, feedingassociated complications (diarrhea, abdominal bloating, hyperglycemia, dislodged nasogastric tubes, etc.) and organ failure during hospital admission were retrospectively compared. Diagnosis of pancreatic infectious complications was based on microbiological examination (Gram stain and culture for aerobic and anaerobic germs and fungi) of samples obtained by ultrasoundguided fine-needle aspiration of (peri-) pancreatic collections and confirmed in pancreatic tissue specimens obtained either during surgery or autopsy in patients operated upon or who died, respectively. 10 Also, nutrition support outcomes, total hospital, intensive-care stay and mortality rate were compared. Hyperglycemia during feeding was defined as blood glucose >10 mmol/L. Local and systemic complications of acute pancreatitis were defined as previously reported. 11 To assess organ failure, the Marshall score was used. 12 Statistical Analysis: All measurements were expressed as mean ± SD. The statistical analyses were performed using the two-sample t-test and adjusted Chi-square test for the two groups.
The Exact Chi-square test was also used if individual cell size was less than five counts. P value <0.05 was considered statistically significant.

Baseline Characteristics of Patients:
No significant difference was found in terms of gender, age, BMI, body weight, underlying comorbidity, etiological factor, APACHE II score and CRP level at admission (Table-I Table-II). Despite our efforts, tubes were displaced in five cases and a replacing maneuver was required. Eight patients in EN group developed abdominal bloating. In these patients, the volume of their enteral nutrition was temporarily reduced for between one and two days. A troublesome diarrhea occurred in nine patients in EN group. In these patients, the rate of feeding was temporarily reduced and four of the patients received loperamide as well. In all patients resumption of full enteral support was achieved for the remainder of the feeding period. In TPN group, three patient developed diarrhea. The incidence of hyperglycemia during feeding was significantly higher in TPN group (32/96 vs. 18/89, P = 0.0454) ( Table-II Table-III). No patient interrupted their oral feeding because of pain relapse. Nutrition support was maintained until death in dead cases.

DISCUSSION
Mortality of SAP has two peaks, i.e., "early" during the first week and "late" after 1 to 3 weeks. 13 The second peak in mortality usually involves MODS together with infections, which are frequently caused by gram-negative bacteria. The predominance of gram-negative bacteria isolated from infected pancreatic tissue resemble common gastrointestinal flora, suggesting that they reach the pancreas by translocation from the gut. Thereby, preservation or restoration of the intestinal barrier function may have a beneficial impact on infectious morbidity from severe acute pancreatitis and may reduce mortality. Early enteral feeding is recommended to preserve the intestinal barrier to prevent bacterial translocation along with avoiding the complications associated with parenteral nutrition in SAP. 14 Proper timing is probably crucial for achieving success with therapeutic interventions, including modulation of inflammatory mediator production and release. In SAP, a therapeutic window exists up to about 48 and 72 hours from pain onset, i.e., in the time phase usually required for the development of remote organ dysfunction. 15 In present study, we initialed TPN and EN in 61.4±6.4 hours and 62.3±7.1 hours after disease onset respectively (P=0.3658). In present study, mortality rate of EN group was significant lower. In some reports, the majority of deaths in SAP are attributed to infected necrosis and secondary multi-organ failure, which means late mortality. [16][17][18] Notably, six patients in TPN group and none in EN group died in the second week after disease onset. This would suggest that early enteral feeding could efficiently preserve the intestinal barrier to prevent bacterial translocation and reduce 'late' mortality in patients with SAP.
Pancreatic infections and subsequent septic complications have emerged as the main risk factor for late death in severe acute pancreatitis. It is widely accepted that early enteral feeding is critical to improving outcomes in patients with SAP since the first demonstrated the beneficial effect of an elemental enteral diet in 6 patients with acute pancreatitis in 1973. 14 We found that more patients developed pancreatic infectious complications (P=0.0333) and extrapancreatic infections (P=0.0431) in TPN group. A significant decrease in the incidence of infectious complications in EN group may lead to the reduction of surgical interventions in the enter ally fed patients. Consequently, the length of hospital stay and intensive-care stay in EN group was significantly shorter. Similar to that mentioned above regarding the incidence of infectious complications, the incidence of MODS (P=0.0338) was significantly lower in EN group.
In present study, full rate of nutrition was achieved in all patients and there was no significant difference in the time to full rate of nutrition support after initiation. Total feeding-associated complications did not statistically difference between two groups (P=0.8147). Hyperglycemia is common in SAP, and in the present study incidence of hyperglycemia during feeding was significantly lower in EN group (P=0.0454). The effects of Yulong Tao et al. normoglycemia in SAP have not yet been studied, but potentially this concept might further improve outcome also in patients with SAP.
In conclusion, this retrospective study has confirmed that nasogastric EN may be safely delivered to patients with SAP without causing feeding-associated complications. Based on the improved outcomes, nasogastric EN should be started within 72 hours after disease onset in all patients with severe acute pancreatitis. Because of the retrospective nature and the small sample size, further prospective study with large sample size is needed to confirm the results of our study.

Grant Support & Financial Disclosures:
None.