Acute pancreatitis at the Aga Khan University Hospital, Nairobi: a two year audit
M. Mutebi, MBChB, A. Abdallah, MBChB, FRCS, Department of Surgery, Aga Khan University Hospital,
Nairobi, P.O. Box 30270- 00100, Nairobi, Kenya and H. Saidi, BSc., MBChB, MMed, FCS, Department of Surgery, Aga Khan University Hospital and Department of Human Anatomy, College of Health Sciences, University of Nairobi, P.O. Box 30197-00100, Nairobi, Kenya
Dr. M. Mutebi, Department of Surgery, Aga Khan University Hospital, Nairobi, P.O. Box 30270- 00100, Nairobi, Kenya,
BACKGROUND: Acute pancreatitis ranges in severity from a mild, self-limiting to a fulminant disease with systemic decompensation. The treatment of the severe form of the disease may be difficult with mortality rates of up to 30%. There are published, evidence-based guidelines for optimizing outcome of the disease (1,2). We performed this audit to determine whether our clinical practice conformed to these guidelines.
AIM: To audit the management of patients admitted with acute pancreatitis at the Aga Khan University Hospital (AKUH) and compare the current practice with accepted international guidelines (1,2) with respect to diagnostic modalities, severity stratification, critical care unit (CCU) management for severe disease, timing of treatment intervention strategies and mortality.
DESIGN: Retrospective audit of hospital records.
METHODS: The medical records of all patients admitted to the AKUH with a diagnosis of acute pancreatitis from 1st January 2005 to 31st January 2007 were reviewed.
RESULTS: Thirty five patients were admitted with a confirmed diagnosis of acute pancreatitis in the two year period. Twenty six patients were males (74%). The median age was 46 years (range 29-82 years). The aetiology of the acute pancreatitis was alcohol in 51%, gallstones in 11%, other causes in 9% and idiopathic in 29% of patients. The median length of hospital stay was 11 days (range 1–38). The diagnosis of acute pancreatitis was confirmed by amylase and lipase assays or with CT scan evidence of pancreatitis. Only three patients (8%) had formal severity stratification (Ranson’s score). Eight patients (23%) had severe disease as defined by pancreatic necrosis and need for critical care. Twenty seven patients with mild disease also underwent abdominal CT scans and only 54% of all patients had an initial ultrasound to exclude gallstones. The timing of these investigations was arbitrary. Ten patients with mild disease received unnecessary prophylactic antibiotics including metronidazole, cefuroxime, and tazobactam/ piperacillin for a median period of 7 days. In severe disease where antibiotic use is possibly justifiable, a carbapenem based antibiotic was prescribed for four patients. Nasojejunal feeding was instituted early in six patients with severe disease and parenteral nutrition was also used exclusively in one patient. The overall mortality was 2.9% with the only death occurring in the severe subgroup thereby making the mortality rate in those patients with severe acute pancreatitis in this audit 12.5%.
CONCLUSIONS: The current management of acute pancreatitis at AKUH is physician dependant and not in conformity with the established and recommended guidelines. The CT scans were over-prescribed, their timing inappropriate and efforts to exclude the cause of pancreatitis moderate. The mortality rate is acceptable by international standards despite uniform application of diagnostic and risk stratification tools.
Pancreatitis was first characterised by Lord Monihyan in 1925, as the ‘most terrible of calamities to affect the viscera’ and its attendant mortality as, ‘the most formidable of catastrophes’ (3). Despite considerable evolution in its management, acute pancreatitis continues to present therapeutic challenges due to its variable clinical presentation and natural history, ranging from mild innocuous self-limiting disease to frank necrosis of the gland and multiple organ failure (4). Markers of severe disease as defined by the Atlanta symposium include a Ranson’s score >3, Apache score of >8 or presence of one or more organ failures or local complications. The overall mortality in hospitalised patients is approximately 10%, but may increase to 30% in those with severe disease (5). A multiple organ dysfunction syndrome (MODS) usually presents within days of the severe form of the disease with secondary complications and further clinical deterioration due to infected necrosis and systemic sepsis developing in the second week.
Institution of prompt intensive care has decreased the morbidity (6) over the years and improved the overall quality of life in survivors of severe disease (7). Early recognition of patients at risk for severe disease using any of a variety of risk stratification methods (Ranson’s, Modified Glasgow and Apache II scores) has also modified outcomes (2) by dictating the pace and setting of care. Several evidence-based guidelines (British, International Association of Pancreatology, Japanese etc.) have been designed to provide recommendations for managing acute pancreatitis. These guidelines include methods for early diagnosis, critical care management for severe acute pancreatitis and the timing of radiological investigations and surgical interventions. The concordance of clinical practice with the guidelines for therapy is however variable. Even if they know the current recommended guidelines, clinicians do not always follow these recommendations (12). This audit compares the management of acute pancreatitis at the AKUH with these recommended international guidelines.
The medical records of all patients admitted with a diagnosis of acute pancreatitis between 1st January 2005 and 31st January 2007 were reviewed. All patients presented with epigastric pain. Patients who had a presumptive diagnosis of pancreatitis, who subsequently had an alternative definitive diagnosis made (gastritis or peptic ulcer disease (PUD) confirmed on endoscopy), were excluded from this audit.
The points audited were as follows:
Method of diagnosis/diagnostic modality
Risk stratification of patients
Other therapeutic measures such as early enteral feeding and antibiotic usage
The aetiology of acute pancreatitis was classified as gallstones (demonstrated by ultrasound), alcohol (when average daily intake exceeded 50g in the absence of other causes), other causes e.g. secondary to anti-retroviral use, post ERCP or as idiopathic (when no cause could be readily identified). Records were obtained on the mode of diagnosis, the duration of symptoms prior to admission to AKUH, length of stay in AKUH, predictors of severity utilised, timing of CT scan, surgical interventions and complications.
Organ failures were defined as: respiratory failure requiring ventilatory support, circulatory failure unresponsive to fluid therapy or altered mentation (low GCS). Local complications included pancreatic necrosis, pseudocyst formation, fluid collection or infected necrosis. Respiratory complications included pleural effusions, lung atelectasis or pneumonia.
Thirty five records were reviewed. Twenty six (74%) were males and nine were females (26%) The median age was 46 years (range 29–82 years).
All diagnoses were made on the basis of symptoms (epigastric pain lasting 1-2 days ±back pain ± constitutional symptoms of nausea and vomiting), signs (epigastric tenderness) and elevated amylase (three times the normal) or elevated lipase (>60 micromoles/l). Thirty patients (86%) underwent biochemical diagnostic tests with 30 patients having an initial amylase and 29 having an initial lipase performed. Twenty four patients had both amylase and lipase measured. The remaining fi ve had the diagnosis made radiologically. Twenty patients(57%)underwent serial amylase and lipase level measurement at a median interval of 24 hours. Twenty six patients (74%) had a CT scan done. This demonstrated pancreatic necrosis in 5 patients (14.3%), acute fluid collections in 11 patients (31%), oedema and fat infiltration in 8 patients (23%), pleural effusions in 5 patients (14%) and normal findings 10 patients (27%). The average timing before CT scan was performed was 48 hours (range 24-72 hours). Nineteen patients (54%) had an abdominal ultrasound with a similar percentage undergoing both CT and ultrasound investigations.
The etiology was attributed to gallstones in four patients (11%), alcohol in 18 (51%) and other causes in three patients (9%). The latter group included patients who developed pancreatitis secondary to anti-retroviral drugs or as a complication of ERCP. There was no clearly established cause in 10 patients (29%) and the records revealed minimal attempts to determine the etiology in this group (Table 1).
Risk stratification of patients and level of care
Only three patients (8%) had formal severity stratification based on the Ranson’s criteria (Table 1) and all these had mild disease. Nine patients (25.7%) were managed in critical care units (four in Intensive Care Unit and five in High Dependency Unit; one was admitted due to significant co-morbidity and not the perceived severity of pancreatitis). The admission to a critical care unit was dictated by the perceived severity of disease by the attending physician, with no reproducible objective admission criteria aiding these decisions. Two patients were presumed to have infected necrosis on the basis of an initial scan confirming >50% necrosis and an initial improvement in symptoms on antibiotic therapy, then a clinical decline during the 2nd and 3rd week of management. One of these patients underwent percutaneous drainage of their pancreatic fluid collections which cultured E. coli.
Antibiotic therapy, nutritional support and medical care
The spectrum of antibiotics utilized included metronidazole, ciprofloxacin, meropenem, tazobactam/piperacillin and amikacin. These were started at varying times, most within 24 hours of admission, and were given for a duration of five days to one week. Eighteen patients (51.4%) received antibiotics, 10 (22.6%) of whom had mild disease.
All patients were initially nil by mouth and subsequently had feeding reinstituted when it was assessed that they would tolerate enteral intake. Patients with mild disease were recommenced on normal diet after an average duration of 48 hours. Of the CCU patients, six (75%) had nasojejunal feeds instituted between 48-72 hours. One patient received nasogastric feeds and another with severe disease received exclusive parenteral nutrition. One of the 7 patients on enteral feeds also received parenteral feeds. The patient who was admitted with mild disease to the CCU was also commenced on enteral feeding.
Respiratory failure occurred in two patients (5.7%) and the duration of mechanical ventilation was 4 days and 11 days respectively. One patient had circulatory failure requiring ionotropic support. The median length of hospital stay was 11 days (range 1-38 days). The overall mortality was 2.9 % (one patient). This patient had severe acute pancreatitis and was one of the two patients who developed presumed infected necrosis. Thirty three patients were eventually discharged home. One out of the four patients with biliary cause to their pancreatitis was readmitted a month later for an elective laparoscopic cholecystectomy.
This audit demonstrates the management practice for acute appendicitis at the AKUH over a two year period. The results of the audit indicate that both amylase and lipase levels were used fr diagnosis. In addition, 20 patients (57%) underwent serial lipase and amylase level evaluations presumed to be an indicator of progression of disease. This practice is at variance with recommendations. Although amylase provides acceptable accuracy of diagnosis, where available, lipase alone is preferred at times for diagnosis (1). Given the natural evolution and peaks and troughs of these enzymes, they have no prognosticating value and there is no added benefit to doing them serially other than adding expense.
The results have further indicated that patients at our institution had an index CT scan within 24-27 hours of their admission on the basis of a diagnosis of acute pancreatitis, regardless of severity stratification. Of the 26 patients who had a CT scan performed, only 8 had severe disease. Given an average presentation time of 1-2 days, the duration of time may have been insufficient to appreciate radiological changes in the pancreas indicating severe disease. Only 5 out of the 8 patients with severe disease who required CCU care had demonstrated necrosis on CT. This may have been due to inappropriate timing of the imaging. Indeed, several patients (both severe and mild) underwent a second CT because of a deteriorating clinical picture or persistent pain. The recommendation is that patients with persistent organ failure, signs of sepsis or clinical deterioration 6-10 days after admission will require CT scan.
The proportion of patients with an intermediate etiology to their pancreatitis was 29% and is higher than the acceptable for idiopathic disease. Possibly, an ultrasound performed in the index pancreatitis could have proved more beneficial and effective in determining the possible etiology. A more comprehensive approach in excluding the less common causes of acute pancreatitis may be indicated in the future to decrease the number of patients labelled as having idiopathic pancreatitis.
Only three patients were stratified according to risk using the Ranson’s criteria and all three had mild disease. Eight patients who were managed in the critical care unit were deemed to have severe disease at admission by criteria other than published guidelines. Presumably the widespread use of CT scan served the purpose of isolating those with local complications. It is recommended that severity be defined by the Atlanta criteria (Ranson’s score>3, Apache score >8 or presence of one or more organ failures or local complications) and that patients with acute severe pancreatitis be managed in a critical care unit.
In the current series, antibiotics were used in an arbitrary manner. Eighteen patients received antibiotics against eight patients with severe disease. Thus, 10 patients received unnecessary antibiotics. Of the patients with severe disease, only four received a carbapenem-based antibiotic. In their systematic review, Henrich et al demonstrated benefits with antibiotic therapy only in patients with severe pancreatitis. Nordback (10), showed benefit of imipenem over placebo treatment in managing patients with severe pancreatitis. Bassi et al (11) demonstrated superiority of imipenem over quinolone combinations. Prophylactic antibiotic use in severe disease minimizes the risk of developing secondary infection of the necrotic of the gland due to bacterial translocation. Issenman et al (4) showed no benefit to the use of antibiotics in severe disease. However, their study used ciprofloxacin and metronidazole, which have been shown to have no proven benefit in outcome with severe disease. This perhaps further consolidates the case for a carbapenem-based antibiotic. The results of the available randomized trials and meta-analyses are conflicting and the use of prophylactic antibiotics remains a controversial issue at present.
Early enteral feeding is advocated in patients with severe acute pancreatitis. This preserves the gut mucosal integrity which prevents bacterial translocation and possibly lowers the chances of developing infected necrosis (2). The overall management in the current audit showed a cognizance of nutritional requirements and most severe patients had their feeds instituted early. However, one patient received exclusive parenteral nutrition which may have further complicated his hospital stay as it is associated with a greater central line infections and sepsis as suggested by Kalfarentzos (12).
None of our patients with demonstrated biliary pancreatitis had a cholecystectomy performed at their index admission as per the recommendations for mild disease proposed by Heinrich et al (2). One patient had a cholecystectomy a month later. There was no active follow up for the remaining three patients. An early cholecystectomy removes the ‘seat of disease’ in biliary pancreatitis, and prevents recurrent bouts of disease which are associated with considerable morbidity.
Our audit is limited by its retrospective nature. Some of the records were incomplete and thus could not be included in our study and therefore the results could be biased. It is difficult to categorize the true aetiology of the pancreatitis except where cholelithiasis was demonstrated or where some attempt at quantifying alcohol had been made. There was frequently a report of heavy alcohol use and no quantification. In the absence of any other obvious risk factors or other causes, we attributed the aetiology to alcohol and this may have skewed our results. The mortality rate was 2.9% overall and 12.5% for the group with severe disease. Although this might suggest that patients with acute pancreatitis can be managed without guidelines with a low associated mortality, the sample is small to make such a conclusion as a number of patients could have been missed due to inappropriate diagnosis. Our patients were over-investigated and over-treated. There were more ICU admissions, more antibiotic use, more CT scans and laboratory investigations than were possibly necessary. This is not a cost-effective method of managing patients with acute appendicitis in a resource-limited environment. In the majority of our patients, in whom favorable outcome was anticipated, these tests only increase the financial burden to the patients.
The audit illustrates a need for a paradigm shift in the current management of patients with acute appendicitis. The routine use of an accepted scoring system is necessary to better categorize our patients.
There is need for establishment of clear protocols which will have considerable cost implications particularly in a resource strained region like our own. Our recommendations are:
Use of a single serum assay (preferably serum lipase where available.)
Risk stratification of patients using the Ranson score (all parameters easily available)
Abdominal ultrasound during admission with the first episode of pancreatitis and a CT scan in patients with severe disease at the appropriate interval
The authors acknowledge the assistance of the Department of Surgery and the Medical Records Department at the Aga Khan University Hospital, who were able to facilitate this study.
UK guidelines for the management of acute pancreatitis: UK working party on acute pancreatitis. 2003
Heinrich S., Schaffer M., Rousson V., et al. Edidence-based treatment of acute pancreatitis. Ann. Surg. 2006; 243: 154-168.
Monihyan B. Acute pancreatitis. Ann. Surg. 1925; 81: 132-142.
Isennmann R., Ran B. and Berger H.G. Early severe acute pancreatitis: characteristics of a new subgroup. Pancreas. 2001; 22: 274.
Sarles H. Revised classification of pancreatitis. Marseilles. Dig. Dis. Sci. 1985; 30: 573.
Bank S. Clinical course of acute pancreatitis: What has changed in recent years? In: Malfertheiner P. (ed.) Acute Pancreatitis Novel Concepts in Biology and Therapy. Berlin: Blackwell Science, 1999; pp. 164.
Fenton-Lee D. and Irene C. Pancreatic necrosis: assessment of outcome related to quality of life and cost management. Brit. J. Surg. 1993; 80: 1579-82.
Foitzik T. and Klar E. (Non) Compliance with guidelines for the management of severe acute pancreatitis among German surgeons. Pancreatology. 2007; 7: 80-5.
Bradley E.L. A clinical based classification system for acute pancreatitis. Summary of the international symposium on AP., Atlanta, GA. September 11-13., 1992. Arch. Surg. 1993; 128: 586-90.
Nordback I. Early treatment with antibiotics reduces the need for surgery in acute necrotizing pancreatitis: a single centre randomized study. J. Gastrointestinal Surg. 2001; 5: 113-8.
Bassi C., Falconi M., Talamini G., et al. Controlled clinical trial of perfloxacin versus imipenem in severe acute pancreatitis. Gastroenterology. 1998; 115: 1513-7.
Kalfarentzos F. Enteral nutrition is superior to parenteral nutrition in severe acute pancreatitis: Results of a randomized prospective trial. Brit. J. Surg. 1997; 84: 1665-1669.