Audit

M.G. DUBE, D.N. LOBO, B.J. ROWLANDS and I.J. BECKINGHAM
Division of Gastrointestinal Surgery, University Hospital, Queen’s Medical Centre, Nottingham, UK

Background: A case note audit of existing practices in the management of acute pancreatitis (AP) at a district general hospital (DGH) and a teaching hospital was undertaken to determine if practices were in accordance with published UK guidelines. Methods: Casenotes of all adults admitted with AP over a period of one year at the two hospitals were reviewed. Results: Ninety-five patients were treated for AP at the teaching hospital and 52 at the DGH. The age, sex and aetiological distributions at the two hospitals were similar. Fifteen (15.8%) patients at the teaching hospital and eight (15.3%) at the DGH had severe AP. Four patients died at each hospital. Prognostic Glasgow criteria tests (excluding LDH) were completed within 48 hours in 43% patients at the teaching hospital and 48% at the DGH. Five of the twenty-five cholecystectomies at the teaching hospital and 4/18 at the DGH were performed within four weeks after admission with AP. Conclusion: Audit of current practice has highlighted deficiencies at many levels compared with current evidence-based guidelines, although this has not resulted in unexpected mortality. It remains to be seen whether new measures to aid compliance with guidelines will result in improvement in morbidity and mortality.

Acute pancreatitis (AP) accounts for about 3% of hospital admissions with abdominal pain.1 It is often a difficult condition to manage and various regimens have been tried over the years. Multicentre audits have revealed deficiencies in the management of the disease and a lack of standardised protocols both within and between institutions.^2,3 Overall mortality for the condition has remained unchanged over two decades. ^4,5 The British Society of Gastroenterology (BSG), therefore, commissioned and published national guidelines for the management of AP.6 These were endorsed by other specialist associations and became widely available in June1998. The guidelines include a summary of audit goals. In the light of these, we audited existing practices in the management of AP in a district general hospital (DGH) and a teaching hospital in the mid-Trent region.

The aim was to discover if the management of AP in the two hospitals was in accordance with the guidelines and to identify areas where the recommended standards were not being met.

The period of study was from 1 April 1997 to 31 March 1998, which was just prior to the availability of the BSG guidelines. In both hospitals, case notes of all adult patients whose admissions during the study period were coded for any form of pancreatitis, or who had a serum amylase concentration of over three times the normal recorded during the study period were requisitioned. Of the 84 records obtained at the DGH and 197 at the teaching hospital, 52 patients at the former and 95 at the latter actually had acute pancreatitis on review of the notes. Results of investigations, if not found in the notes, were traced through the laboratory computer systems. Demographic information and data regarding aetiology, prognostic investigations, imaging, and various aspects of management (including use of antibiotics, intensive care unit (ICU) stay and nutritional support) were collected. Final outcome, dates and results of endoscopic retrograde cholangiopancreatography (ERCP), if performed, complications, and date of cholecystectomy, if indicated, were also recorded. All data were then analysed using SPSS^Ž for Windows^TM Release 9.0 software (SPSS Inc., Chicago, USA) and the findings compared with the stated goals in the national guidelines.

The demographic profile (Table 1) of the patients and aetiological factors (Figure 1) in the two hospitals were not significantly different. As there were no ERCPs carried out at the DGH, there were no cases of post-ERCP AP at that hospital. The correct diagnosis was made in all patients in the DGH and in 99% of patients in the teaching hospital within 24 hours of admission, as evidenced in the daily progress notes.

Based on the Glasgow criteria, prognostic blood tests that should be carried out on admission and/or within 48 hours are white cell count, glucose, urea, calcium, albumin, lactate dehydrogenase (LDH) and arterial blood gas analysis.7 Serum C-reactive protein (CRP) concentrations have independent prognostic value and are also recommended. No patient at the DGH and very few at the teaching hospital had LDH or CRP assayed (Figure 2). Excluding LDH, only 48% of patients at the DGH and 43% of patients at the teaching hospital had the complete set of prognostic investigations.

Seventy-one per cent of patients at the DGH had abdominal ultrasound (US) within 72 hours of admission. At the teaching hospital, 65% of the patients, excluding those with post-ERCP AP, had an US within 72 hours. Only five patients (9.6%) at the DGH had contrast-enhanced CT scans and these were all performed between 15 and 35 days after admission for diagnosis/follow-up of peripancreatic fluid collections. At the teaching hospital 16 patients (16.7%) had a staging CT scan, three of them on the first day post-admission, and the rest between 3 and 16 days later. Four of these scans were for patients with severe AP between 3 and 10 days post-admission.

Table 2 shows the data, including mortality, for severe AP, as defined in the Atlanta classification.⁸ There were no deaths in patients with mild AP.

Of the four deaths at the DGH, one patient was an 80-year-old nursing home resident who was admitted in a critically-ill state. He also had significant co-morbidity and after an initial period of resuscitation, a decision was made not to treat actively. Two other patients died 4 and 15 days after admission respectively, of multi-organ failure. Despite having three positive Glasgow criteria on admission, both were only transferred to ICU after significant deterioration on the general wards, on the 2nd and 3rd day after admission, respectively. Finally, the last patient was not fully investigated on admission and was diagnosed after 10 hours when his condition was deteriorating. He was then transferred to ICU but died 12 hours later.

Figure 1: Aetiology of acute pancreatitis. There was no statistically significant difference between the two hospitals (P=0.092). DGH = district general hospital; ERCP = endoscopic retrograde cholangiopancreatography

Figure 2: Proportion of patients in whom prognostic tests were not performed within 48 hours of admission. * P=0.008, Fisher’s exact test; all other differences not statistically significant. ABG = arterial blood gases; LDH = lactate dehydrogenase; CRP = C-reactive protein; DGH = district general hospital

Figure 3: Use of prophylactic antibiotics in mild and severe acute pancreatitis. * • 2 test; ** Fisher’s exact test. AP = acute pancreatitis; DGH = district general hospital; TH = teaching hospital.

At the teaching hospital, two deaths occurred within 24 hours in patients over the age of 80 years, who were very unwell on admission and did not respond to treatment. One death occurred in a patient who was admitted in a shocked state three days after the onset of symptoms. She was treated aggressively but died of a pulmonary embolism two days later. The last death occurred in a patient who was formally recorded on admission as having predicted severe AP, based on the Glasgow criteria. However, he was initially managed on the general ward and was transferred to ICU after 2 days. He had deranged liver function tests, and a dilated common bile duct on an ultrasound scan performed soon after admission. ERCP was attempted three days after admission but failed due to inability to pass the endoscope through a very oedematous duodenum. He died after 19 days of conservative management on the ICU.

	Number of patients at DGH	Number of patients at teaching hospital	P value	Test
Severe AP	8 (15.3%)	15 (15.8%)	0.948	X² test
Organ failure	4	7
Local complications	3	5
Both	1	3
Mortality* Overall	4 (7.7%)	4 (4.2%)	0.454	Fisher’s exact test
In severe AP	4 (50%)	4 (26.7%)	0.371	Fisher’s exact test

*All deaths were due to multiple organ dysfunction syndrome. AP = acute pancreatitis; DGH = district general hospital.

Antibiotics were commenced on admission for 30 (58%) patients at the DGH and 28 (29.5%) at the teaching hospital. Figure 3 shows the patterns of antibiotic prescription on admission in patients with mild and severe AP in the two hospitals. The most commonly used antibiotic was intravenous cefuroxime (750mg, 8-hourly).

Neither hospital had a surgical high dependency unit at the time of the study. At the DGH, four patients were transferred to the ICU, one of whom survived. At the teaching hospital, there were nine referrals and six transfers to the ICU, the remaining three being managed on the general wards with ICU advice. One patient, who was formally scored as having severe AP on admission, but was transferred to ICU two days later when deteriorating, died.

Only two patients with severe AP at the DGH had any nutritional support - one parenterally, started at 21 days and one enterally at four days. At the teaching hospital, four patients had nutritional support - one parenteral, one enteral and two both (serially), all started three days after admission.

There were no facilities for ERCP at the DGH, and patients were referred to other hospitals for late ERCP, when indicated. At the teaching hospital, two of the six patients with gallstone-associated severe AP had an ERCP attempted in the acute phase. One of these failed, and ductal stones were cleared in the other. Twenty-one patients had ERCP later and of these 15 were normal, three failed, two had ductal stones cleared and one had the common bile duct stented as stones could not be cleared.

Table 3 shows the time-scale over which cholecystectomies were performed in patients with gallstone-associated AP, and the outcome in others who did not have cholecystectomy. At the DGH, nine patients had 12 re-admissions between them while on the waiting list for cholecystectomies. Of these, one died of severe recurrent AP five months after the initial admission. Five patients at the teaching hospital, had six readmissions, one of them being for severe AP with considerable morbidity. Table 4 summarises the audit findings in the two hospitals with respect to the audit goals set out in the guidelines.

The period of study immediately precedes the wide availability of the BSG guidelines, hence patients were not expected to have been managed expressly according to them. However, the evidence-base in the literature that was used to formulate the guidelines did exist and Table 4 clearly shows that the guidelines were not being met in a large proportion of patients. Very few case notes had a formal record of the Glasgow score, although possibly the scores may have been computed on ward rounds.

The use of imaging differed in the two hospitals. While 71% of patients at the DGH had US within 72 hours, no CT scans were obtained between three and ten days after admission. At the teaching hospital, on the other hand, CT scans were obtained too soon in three patients. The guidelines suggest that dynamic contrast-enhanced CT scanning should be performed in all patients with severe AP between three and ten days of admission. The early CT scans at the teaching hospital were not intended to be diagnostic, as the diagnosis had already been made on clinical and biochemical grounds. Moreover, three of the scans were performed without contrast enhancement which limits their value in assessing disease severity. Thus, it would appear that practice relating to CT scanning was haphazard at both hospitals. Incomplete prognostic scores may, to some extent, have been responsible for inappropriate use of CT scans. Clearly, there is a need for rationalisation of the use of CT scanning in both hospitals. Radiologists, as well as treating surgeons and physicians, should be made aware of this, as there are often constraints in the imaging departments which have a bearing on the choice of modality, timing of imaging and use of expensive radiological contrast agents.

There was no standard protocol for starting antibiotics on admission in patients with AP at either hospital. Figure 3 shows that there appears to have been inappropriate use (or lack of use) of antibiotic prophylaxis in mild and severe AP at both the hospitals. There is no evidence that antibiotics are of benefit in mild AP, but in severe AP prophylactic antibiotics may help prevent septic complications and reduce mortality, although cefuroxime in the dose of 750mg 8-hourly may be inadequate for this purpose. ^9,10

Neither hospital had a surgical high dependency unit; both have ICUs, with heavy demand on bed usage. Hence, it is not surprising that patients with severe AP were not routinely admitted to the ICU, at either hospital. ICU help was sought when patients began to deteriorate, but even then transfer was often delayed. Ideally, provision must be made for a high dependency bed to be available for every patient who is predicted to have severe AP, on the basis of a full set of prognostic investigations.

There is no reference to nutritional support in the BSG guidelines and this audit has shown that the prescription of nutritional support is rather haphazard.6 Published literature has not shown that nutritional support affects the underlying disease process in patients with AP. ¹¹ However, patients with severe AP have high metabolic and nutrient demands and artificial nutritional support may help prevent the additional consequences of undernutrition and starvation. Intra-jejunal feeding may be preferred when practical, although parenteral nutrition may be necessary in some patients.

The guidelines state that mortality from AP should be lower than 10% overall and less than 30% in the severe group. There were four deaths at the DGH, which represented 7.7% of all admissions, but 50% of the patients with severe AP. The four deaths at the teaching hospital represent 4.2% of all admissions and 26.7 % of the admissions with severe AP. Despite the small numbers, there may be room for improvement at the DGH in this regard, particularly, as two of the deaths occurred late, in patients who were referred to ICU on the second and third day after admission, respectively. These statistics, however, do not take into consideration deaths that may have occurred in patients with undiagnosed AP, as post-mortem data were not available at either hospital.

Fifteen of 21 late ERCPs in patients with gallstone-associated AP at the teaching hospital were normal and another three failed. As shown in Figure 1, post-ERCP AP accounted for 14% of patients from this hospital. We must clarify here that this represents about 2.5% of all the ERCPs performed in the hospital over the year. With a low positive bile duct stone rate the use of pre-operative magnetic resonance cholangiopancreatography or operative cholangiography with single stage laparoscopic cholecystectomy and ductal exploration may be the way forward. ¹²

At both hospitals, only a small proportion of cholecystectomies were performed within the recommended time period. The delay in cholecystectomy resulted in 17 readmissions (Table 3), one death from severe recurrent AP at the DGH, and a large pseudocyst secondary to recurrent AP in a patient at the teaching hospital. This pseudocyst failed to resolve and the patient ultimately had a cholecystostomy and a cystogastrostomy.

	Number of patients at DGH	Number of patients at Teaching Hospital
Total gallstone pancreatitis	27	45
Total cholecystectomies	18	25
Performed within 2 weeks	2	1
Performed between 2 and 4 weeks	2	4
Performed over the next 16 months	14	20
Deaths	3*	0
Re-admissions with AP while on waiting list for cholecystectomy	12 re-admissions in 8 patients	6 re-admissions in 5 patients
ERCP + ES as definitive treatment	1	4
Medical or other reasons for not operating	3	11
Refused operation	2	3
Lost to follow-up	0	2

* Two patients died of severe gallstone AP during first admission and one died from recurrent severe AP while on the waiting list for cholecystectomy. AP = acute pancreatitis; DGH = district general hospital; ERCP = endoscopic retrograde cholangiopancreatography; ES = endoscopic sphincterotomy

This audit has demonstrated deficiencies in the management of AP in both hospitals. Existing practices were sub-optimal and lacked uniformity in a number of areas. Deficiencies were noted in the completion of prognostic investigations and early severity stratification, and in performance of early cholecystectomy where indicated. Correction of the former may lead to earlier recognition of severe AP and to reduction of mortality in this group. Improvement in the early cholecystectomy rate may reduce the morbidity of readmission. Policy on the appropriate use of antibiotics, CT scanning, ERCP and nutritional support needs to be defined and implemented, so as to derive maximum benefit while minimising costs.

Specific recommendations have been made in both hospitals taking into account the findings of these audits. These include the use of a specifically designed form for every patient admitted with AP to ensure a uniform plan of investigation, severity stratification and management. Patients with mild gallstone-associated AP are not to be put on the routine waiting list for cholecystectomy on discharge but are to be given a specific early date for operation. We understand that the guidelines are themselves under revision, and a future audit against those would show if the changes introduced have made a difference.

Audit Goals	Findings at teaching hospital	Findings at DGH
Mortality <10% overall, < 30% for severe AP	Mortality 4.2% overall, 26.7% for severe AP	Mortality 7.7% overall, 50% for severe AP
Correct diagnosis within 48 hours	Achieved in 99% (as evidenced in daily progress notes)	(as evidenced in daily Achieved in 100% progress notes)
Severity stratification within 48 hours	Formal record in 6.3%; complete set of tests (except LDH) in 43%	No formal record; complete set of tests (except LDH) in 48%
Aetiology assessment - idiopathic group not more than 20-25%	Idiopathic group 19%	Idiopathic group 21%
HDU/ICU initially for severe cases	40% of severe cases transferred to ICU	50% of severe cases transferred to ICU
Full radiological and ERCP facilities	Available	ERCP not available
Dynamic CT scanning for severe AP between 3 and 10 days of admission	Performed for 26.7% of patients with severe AP	Not performed for any patient severe AP
Cholecystectomy within 2-4 weeks for mild gallstone pancreatitis	22.2% of all cholecystectomies performed within 4 weeks	20% of all cholecystectomies performed within 4 weeks

AP = acute pancreatitis; DGH = district general hospital; LDH = lactate dehydrogenase; HDU = high dependency unit; ICU = intensive care unit; ERCP = endoscopic retrograde cholangiopancreatography.

This paper has been presented in part to the Association of Surgeons of Great Britain and Ireland, Cardiff, May 2000, and published in abstract form. Br J S urg 2000; 87(S1): 62

1. de Dombal F (ed). The acute abdomen: definitions, diseases and decisions. In: Diagnosis of Acute Abdominal Pain. 2nd edn. London: Churchill Livingston, 1991: 19-30
2. Mann DV, Hershman MJ, Hittinger R, Glazer G. Multicentre audit of death from acute pancreatitis. Br J Surg 1994; 81: 890-3
3. Toh SKC, Phillips S, Johnson CD. A prospective audit against national standards of the presentation and management of acute pancreatitis in the South of England. Gut 2000; 46: 239-43
4. Corfield AP, Cooper MJ, Williamson RC. Acute pancreatitis: a lethal disease of increasing incidence. Gut 1985; 26: 724-9
5. Wilson C, Imrie CW, Carter DC. Fatal acute pancreatitis. Gut 1988; 29: 782-8
6. Glazer G, Mann DV. United Kingdom guidelines for the management of acute pancreatitis. Gut 1998; 42(Suppl 2): S1-S13
7. Blamey SL, Imrie CW, O’Neill J, Gilmour WH, Carter DC. Prognostic factors in acute pancreatitis. Gut 1984; 25: 1340-6
8. Bradley EL III. A clinically based classification system for acute pancreatitis. Summary of the International Symposium on Acute Pancreatitis, Atlanta, Ga, September 11 through 13, 1992. Arch Surg 1993; 128: 586-90
9. Powell JJ, Miles R, Siriwardena AK. Antibiotic prophylaxis in the initial management of severe acute pancreatitis. Br J Surg 1998; 85: 582-7
10. Golub R, Siddiqi F, Pohl D. Role of antibiotics in acute pancreatitis: A meta-analysis. J Gastrointest Surg 1998: 2: 496-503
11. Lobo DN, Memon MA, Allison SP, Rowlands BJ. The evolution of nutritional support in acute pancreatitis. Br J Surg 2000; 87: 694-705
12. Cuschieri A, Lezoche E, Morino M, Croce E, Lacy A, Toouli J, et al. EAES multicenter prospective randomised trial comparing two-stage vs single-stage management of patients with gallstone disease and ductal calculi. Surg Endosc 1999; 13: 952-7

Copyright date: 10th August 2001
Correspondence: M.G. Dube, Section of Surgery, E Floor, West Block, University Hospital, Queen’s Medical Centre, Nottingham NG7 2UH, UK

DGH	Teaching	P value hospital	Test
Total number of patients	52	95	-	-
Median age in years	55	60	0.862	MW*
(range)	(17-90)	(19-95)
Male:Female	25:28	44:51	0.985	X²test

AUDIT SECTION