Article Text
PDF
Abstract
BACKGROUND The pharmacological inhibition of exocrine pancreatic secretion with the somatostatin analogue octreotide has been advocated as a specific treatment of acute pancreatitis.
AIM To investigate the efficacy of octreotide in acute pancreatitis in a randomised, placebo controlled trial.
METHODS 302 patients from 32 hospitals, fulfilling the criteria for moderate to severe acute pancreatitis within 96 hours of the onset of symptoms, were randomly assigned to one of three treatment groups: group P (n=103) received placebo, while groups O1 (n=98) and O2 (n=101) received 100 and 200 μg of octreotide, respectively, by subcutaneous injection three times daily for seven days. The primary outcome variable was a score composed of mortality and 15 typical complications of acute pancreatitis.
RESULTS The three groups were well matched with respect to pretreatment characteristics. An intent to treat analysis of all 302 patients revealed no significant differences among treatment groups with respect to mortality (P: 16%; O1: 15%; O2: 12%), the rate of newly developed complications, the duration of pain, surgical interventions, or the length of the hospital stay. A valid for efficacy analysis (251 patients) also revealed no significant differences.
CONCLUSIONS This trial shows no benefit of octreotide in the treatment of acute pancreatitis.
- acute pancreatitis
- somatostatin
- octreotide
- randomised controlled multicentre trial
Abbreviations used in this paper
- ARDS
- adult respiratory distress syndrome
- CRF
- case report form
- CT
- contrast enhanced computed tomography
- ERCP
- endoscopic retrograde cholangiopancreatography
- ITT
- intent to treat
- O
- octreotide
- P
- placebo
- RCT
- randomised controlled trial
- SIRS
- systemic inflammatory response syndrome
- VFE
- valid for efficacy
Statistics from Altmetric.com
Acute pancreatitis is clinically classified into mild and severe forms.1 The majority of patients suffer from mild acute pancreatitis, a self limiting disease which responds well to conservative treatment. In up to 20% of patients with acute pancreatitis, however, the disease progresses to a severe form involving both pancreatic and extrapancreatic necrosis. While the morbidity and mortality of mild acute pancreatitis are low, patients with severe acute pancreatitis commonly develop systemic complications and have a mortality of 10–50%, despite all currently available treatment strategies.2-8
Although the initial steps in the pathogenesis of acute pancreatitis are not fully understood, the term “autodigestion”, introduced by Chiari in 1896, summarises the prevailing concept that the acute inflammation is triggered by the activation of digestive enzymes.9 ,10 The earliest trial of somatostatin, a hormone which inhibits exocrine pancreatic secretion, as a rational therapeutic approach in human acute pancreatitis, yielded “an impressive clinical improvement”.11 The findings of later, randomised trials of somatostatin in acute pancreatitis were controversial.12-17 However, a meta-analysis of all adequately designed studies concluded that somatostatin treatment reduced mortality from 14% to 6.2%.18 In contrast to experimental findings,19 clinical studies showed that basal exocrine secretion is essentially normal during the acute inflammatory phase of acute pancreatitis.20 ,21 These results provoked interest in the synthetic somatostatin analogue octreotide as a potentially useful specific treatment for acute pancreatitis. In an open phase I/II study of patients with moderate and severe acute pancreatitis treated with three different doses of octreotide (100, 200, or 500 μg three times daily), an apparent reduction in the rate of newly developed complications was found which was particularly pronounced in the 200 μg group.22 We were thus encouraged to evaluate the efficacy of this drug in a randomised, controlled, multicentre trial.
Patients and methods
This report was prepared in accordance with the Consolidated Standards of Reporting Trials (the “CONSORT statement”).23
STUDY DESIGN AND PROTOCOL
The study protocol was approved by the Ethics Committees of the Universities of Bern (Switzerland) and Ulm and Bremen (Germany). The two treatment groups received octreotide in different doses, 100 μg three times daily (O1) and 200 μg three times daily (O2); the control group received placebo (P). It was intended to enrol 100 patients in each group, as suggested by a power analysis (see below).
Patients with moderate or severe acute pancreatitis were the target group of the study; mild cases were excluded. To be enrolled, patients were required to fulfil four necessary criteria and four of 10 supporting criteria, as defined in a previous multicentre trial (table1).8 The necessary and supporting criteria were waived in cases where the diagnosis of severe acute pancreatitis was confirmed by laparotomy within four days of the onset of symptoms. The exclusion criteria of the previous trial were also used in this trial (table2).8
Inclusion criteria
Exclusion criteria
BLINDING AND RANDOMISATION
Each of the participating centres received packages of vials containing either octreotide or placebo. A multiple of six packages (6, 12, 18, etc.) was sent to each centre. Each package contained equal numbers of vials for groups P (n=2), O1 (n=2), and O2 (n=2) in a blinded fashion. The packages were used sequentially as the patients were enrolled in the study. There was thus a strict randomisation in three parallel groups in each hospital, and a strict double blind procedure was guaranteed. The unbroken codes of the individual packages were deposited with the chief investigator in each hospital. A provision was made that the code could be broken in case of severe adverse events after notification of both the principal investigator and trial coordinator.
MEDICATION
The study medication was stored at 2–8°C. The first subcutaneous injection of the study medication was given immediately after enrolment of the patient. Further injections were given at eight hour intervals over the following seven days.
BASAL TREATMENT
The standardised basal treatment was according to generally accepted principles of supportive care in acute pancreatitis,8 including nothing by mouth diet, nasogastric tube (obligatory in severe acute pancreatitis, optional in moderate acute pancreatitis), central venous catheter, input/output monitoring, and adequate parenteral replacement of fluid and electrolytes. Lipid infusions were permitted, but parenteral hyperalimentation was not. Table 3 summarises further treatments used in some cases; these had to be recorded in the case report forms (CRF).
Optional treatments
PRIMARY AND SECONDARY OUTCOME VARIABLES
The efficacy of the treatment was assessed by a score composed of mortality and the occurrence of 15 typical complications (tables 4 and5). Death was considered to be worse than all other complications together. Therefore, early deaths without complications were assessed adequately. The chief investigator in each hospital was advised to look for these complications each day during the treatment phase (day 0 to day 7) and in the follow up period on days 8–11, 12–14, and 15–21, and on the day of discharge. The severity of acute pancreatitis during the study period was graded according to a scoring system defined in a previous study on the basis of these complications.8 The clinical change in each patient was calculated as the score given on enrolment in the trial, minus the total score for newly developed complications during hospitalisation. A zero or negative calculated value thus indicates clinical deterioration, while a positive value indicates clinical improvement.
Definitions of local, organ, and metabolic complications of severe acute pancreatitis
Complication score as primary outcome variable
The duration of pain and the Apache II scores over the study period were used as secondary outcome variables. Other parameters were monitored for safety reasons (table 6) daily from day 0 to day 7, and thereafter at the same intervals as those used to assess efficacy.
Secondary outcome variables and clinical and laboratory parameters monitored for safety reasons
The preparation, randomisation, and delivery of the study medication, as well as the monitoring of the study centres by checking the information in the CRFs, were carried out by Novartis (formerly Sandoz), Nuremberg (Germany). The study headquarters was located in the Department of Visceral and Transplantation Surgery of the University of Bern (Switzerland) and was responsible for medical problems, data checking, and data entry. Checking of the data for plausibility, and all statistical analyses, were performed by the Department of Biometry and Medical Documentation of the University of Ulm (Germany), after the code was broken.
POWER ANALYSIS AND STATISTICS
In the multicentre study with the use of gabexate mesilate in moderate to severe acute pancreatitis, the percentage of patients with one or more newly developed complications in the placebo group was 63%.8 In the present octreotide study, conducted in patients with disease of the same degree of severity, we hypothesised that octreotide treatment would reduce the complication rate from 63% to 35%. A power analysis indicated that 100 patients would have to be enrolled in each of the three groups to detect such a reduction at the one sided 5% significance level with a power of 98.8%. If the power is reduced to 80% a decrease in the complication rate from 63% to 45% can be detected under the same conditions.
Two orthogonal tests were applied to the primary and secondary outcome variables: the two octreotide groups, taken together, were tested against placebo (O1, O2 versus P), and high dose octreotide was tested against low dose octreotide (O2 versus O1). A χ2 or Fisher’s exact test was applied for qualitative variables, and a Wilcoxon U or Kruskal-Wallis H test for quantitative variables.
The study was designed for one single outcome variable. However, the restriction to one outcome variable is only relevant if the study delivers a significant result. If several outcome variables are tested, but none of these tests is significant, this gives even more evidence that octreotide is not helpful in moderate to severe acute pancreatitis than one test on the primary outcome variable. In the case of non-significance the problem is not to control type 1 error, but to minimise type 2 error. The latter is true if no significance is achieved despite testing several outcome variables.
Results
PATIENT CHARACTERISTICS
Thirty one centres in Germany and one in Switzerland participated in this trial; 302 patients with moderate to severe acute pancreatitis were enrolled from November 1993 to March 1996. The mean age was 50 years (range 18–93); there were 198 men and 104 women (table 7). The aetiology of acute pancreatitis was alcohol abuse in 126 patients (42%), gallstones in 111 (37%), alcohol abuse and gallstones in four, alcohol and other causes in two, and other aetiologies in 59 (20%). The median time elapsing from the onset of symptoms to entry into the study was 44 hours (range 0–145), and the patients fulfilled a median of five supporting criteria (range 3–10) on enrolment. The median Ranson, Apache II, and complication scores were 3 (range 0–9), 8 (range 0–37), and 6 (range 1–30) points, respectively. Contrast enhanced computed tomography (CT) was performed within 96 hours in 264 (87%) of the 302 patients; pancreatic oedema with exudate, and varying extents of pancreatic and peripancreatic necrosis, were found in 133 (44%) and 108 (36%) patients, respectively.
Characteristics of the 302 patients in the intent to treat analysis
The intent to treat (ITT) analysis included all 302 patients enrolled in the study. A total of 103 patients were treated with placebo (P), 98 with octreotide 100 μg three times daily (O1), and 101 with octreotide 200 μg three times daily (O2). Table 7 gives the patient characteristics in the three treatment groups. Fifty one patients were excluded from the valid for efficacy (VFE) analysis for various reasons: unfulfilled inclusion criteria (n=17), discontinuation or incorrect use of study medications (n=14), withdrawal of consent (n=10), and previous laparotomy with necrosectomy (n=19) (fig 1). Table8 summarises the characteristics of the remaining 251 patients considered in the VFE evaluation (90 in P, 78 in O1, and 83 in O2).
Flow chart of the allocation of patients to the study groups. ITT, intent to treat analysis; VFE, valid for efficacy analysis. Some patients were excluded for more than one reason.
Characteristics of the 251 patients in the valid for efficacy analysis
COMPARABILITY OF GROUPS
There were no demographic differences between treatment groups in the categorical variables sex, aetiology, severity of disease, and presence of pre-existing and accompanying diseases (tables 7 and 8). Twenty three variables were used to assess the comparability of groups in the ITT and VFE analyses; there was therefore an expectation of 1.15 falsely significant tests at the 5% significance level. The only indication of a non-comparability of groups was a nominally significant difference in the frequency of necrosectomy at enrolment in the study (p=0.023); however, even though this p value rose above the 5% significance level after adjustment for multiple testing with the Bonferroni-Holm correction, all patients with initial laparotomy and necrosectomy were excluded from the VFE analysis. Additionally, the influence of the centres on mortality was estimated with a logistic regression calculation. Neither the centres nor the treatment groups were significantly different in mortality according to either the ITT (p=0.22 and p=0.85, respectively) or VFE (both p=0.54) analyses.
PRIMARY OUTCOME VARIABLES
In the ITT analysis, 14% (43/302) of the patients died, with a mortality of 16% (16/103), 15% (15/98), and 12% (12/101), in group P, O1, and O2, respectively (p=0.71). The three groups also did not differ in early and late mortality (table 9). Nearly three quarters of the patients developed new complications in the course of the study (P: n=73, 71%, O1: n=74, 76%; O2: n=73, 72%); there was no significant difference among the three groups in the overall morbidity or in the frequency of each of the 15 specified complications (table 10). The median difference between the complication scores at inclusion and over the course of treatment (newly developed complications only) was 3 in the placebo group and 3 in the two octreotide groups combined (p=0.12). The comparison of O2 with O1 revealed a median difference of complication scores of 4 in the high dose octreotide group, as opposed to 3 in the low dose octreotide group (p=0.059; fig 2). The VFE analysis yielded a mortality of 11% (27/251), distributed as 14% (13/90) in group P, 9% (7/78) in O1, and 8% (7/83) in O2 (p=0.37). The other figures were similar to those found in the ITT analysis (table11).
Primary and secondary outcome variables in the intent to treat analysis
Complications at inclusion of the patients and newly developed complications in the intent to treat analysis
Scores for complications on enrolment, newly developed complications, and their difference in the intent to treat patient group. Results expressed as medians and ranges. The difference in complication scores (enrolment minus new complications) was computed for each patient. The median of the individual differences was then calculated.
Primary and secondary outcome variables in the valid for efficacy analysis
SECONDARY OUTCOME VARIABLES
The median duration of pain was six days in both P and O1, and seven days in O2, according to either the ITT analysis or the VFE analysis (tables 9 and 11). There were no differences in the Apache II score (fig 3) or in the laboratory parameters listed in table6.
Course of the Apache II score in the intent to treat patient group. Results are expressed as medians and ranges.
Surgical intervention for necrosectomy was necessary in 46 patients during the course of the study, 19 (18%) in P, 13 (13%) in O1, and 14 (14%) in O2. Other pancreatitis related operations (for example, relaparotomy for abscess, symptomatic postacute pseudocysts, etc., and cholecystectomy) were performed in 34 patients (33%) in P, 34 (35%) in O1, and 42 (42%) in O2. The rates of surgical intervention were similar in the VFE patient population.
The median length of hospital stay was 21 days (range 0–210) in both the ITT and the VFE analyses and did not differ significantly among treatment groups (tables 9 and 11).
Discussion
The idea that the pharmacological inhibition of exocrine pancreatic secretion might serve as a specific treatment for acute pancreatitis has been a subject of controversy for more than three decades. Activated pancreatic enzymes seem to be involved in the generation and progression of this “autodigestive” acute inflammatory disease, and it was therefore hypothesised that the administration of somatostatin or its analogue octreotide might help prevent progression of the disease and its complications. Nonetheless, the present randomised, controlled, multicentre study, whose size and design give it the power to detect a major clinical benefit if present, found no evidence for any positive effect of octreotide treatment on the course and outcome of patients with moderate to severe acute pancreatitis.
The morbidity and mortality of severe acute pancreatitis are still distressingly high.2-5 Its natural course has two phases. The first phase (during the first two weeks after onset) is characterised by a systemic inflammatory response syndrome (SIRS) due to the liberation of vasoactive and toxic mediators by the necrotising process.24 ,25 In the second phase (during the third week and beyond), septic complications occur as a result of infection of pancreatic necrosis.26
Although the initiating pathophysiological processes are not known, the disease is thought to result from destruction of the pancreas and peripancreatic fatty tissue by digestive enzymes which are produced in the acinar cells and activated intracellularly or extracellularly.9 ,10 Two treatment options might be of value to limit the “autodigestive” local damage and to interfere with SIRS in the early phase of acute pancreatitis: inhibition of the activated enzymes with protease inhibitors; or inhibition of exocrine secretion to “put the gland to rest” and prevent the further release and activation of digestive enzymes. Protease inhibitors (such as aprotinin and gabexate mesilate), hormones (such as calcitonin and glucagon), and atropine have all been clinically tested as possible specific treatments for acute pancreatitis, so far without success.6-8
Somatostatin and its analogue octreotide have been tested in several experimental models of acute pancreatitis, with contradictory results, sometimes negative,27-31 and sometimes positive.32-37 The first clinical study of somatostatin treatment, published in 1980, revealed “an impressive improvement” in all 14 patients treated with a continuous intravenous infusion of somatostatin.11 This encouraging result provided the motivation for further trials, none of which showed a significant benefit on the prognosis of patients with acute pancreatitis.12-17 In 1991, Carballo and colleagues18 published a meta-analysis of six randomised clinical trials involving somatostatin. The individual trials had not revealed any statistically significant benefit; in the meta-analysis, however, the 207 patients in the somatostatin group had a significantly lower mortality than the 207 patients in the placebo group (6.2% versus 14%, p<0.01). A possible explanation for the apparent lack of effect in the individual clinical studies was suggested by the finding that, in various animal models of acute pancreatitis, the basal and stimulated pancreatic exocrine secretion are reduced or even nearly abolished.19 Yet this has been found not to be the case in human acute pancreatitis, where the initial interdigestive exocrine pancreatic secretion remains in the normal range or is even increased.20 ,21
Octreotide, a long acting octapeptide analogue of the native hormone somatostatin, became clinically available in 1982.38 It has proved to be a powerful inhibitor of basal and stimulated exocrine pancreatic secretion in healthy volunteers, and, unlike somatostatin, can be given subcutaneously.39 The results of a phase I/II study with three different octreotide dosages suggested a possible benefit in acute pancreatitis.22 The patients receiving 200 μg of octreotide three times daily had fewer complications than those receiving 100 μg three times daily, or than historical controls.
While our trial was being conducted, other clinical studies of somatostatin or octreotide treatment in acute pancreatitis were published, with conflicting results (table 12).40-46These studies, however, are open to criticism: they either were not randomised, controlled trials,42 ,44 ,45 included mainly patients with mild acute pancreatitis,40-43 or contained only a small number of patients.40-42 44-46 The present multicentre trial sought to overcome all these problems by including a sufficiently large number of patients with an adequate level of disease severity, and randomising them to either octreotide or placebo treatment within 96 hours of the onset of symptoms. This period was chosen so that the patients in the study would all be in the early phase of the disease. From early surgical interventions in severely ill patients, and from serial measurements of C-reactive protein, it has been shown that pancreatic necrosis develops early in the course of the disease (within 96 hours), regardless of the underlying aetiology.47 ,48 It therefore seemed reasonable to evaluate the possible benefit of octreotide within this early, “vulnerable” period. The severity of acute pancreatitis among the patients in our study can be seen from the median Ranson and Apache II scores (3 and 8, respectively), from the overall mortality rate (14%), and from the fact that 80% of the patients had initial CT findings consistent with severe acute pancreatitis of Ranson grade C or worse.49 However, octreotide was not found to be beneficial to our patients, either in the intent to treat (n=302) or in the valid for efficacy (n=251) analysis.
Prospective clinical studies using somatostatin or its analogue octreotide in human acute pancreatitis
In other clinical circumstances, octreotide has been found beneficial if given prophylactically before invasive pancreatic procedures are performed, such as ERCP,50 elective pancreatic surgery,51-54 and pancreatic transplantation.55 However, for the great majority of patients with acute pancreatitis, prophylactic treatment is not possible; octreotide therapy may well come too late, because of the unavoidable interval between the onset of symptoms and hospitalisation. This problem is likely to impair the usefulness of any drug, not just octreotide, used to interfere with the early phase of acute pancreatitis.
In conclusion, octreotide treatment for acute pancreatitis cannot be recommended on the basis of our results. The early phase of severe acute pancreatitis with SIRS can be managed successfully today with maximal intensive care support, without any specific medication directed against the pathophysiological process of acute pancreatitis. In the late phase, however, septic complications frequently occur and are the most important risk factor for death. Preventing infection of pancreatic necrosis thus seems to be the most promising approach to the treatment of this disease at present.26 ,56-59
Acknowledgments
The German Pancreatitis Study Group: M W Büchler, W Uhl, R Vogel, and S Schmid, Department of Surgery, University Hospital of Bern; H G Beger and B Rau, Department of Surgery, University Hospital of Ulm; H P Bruch and G Kolbert, Department of Surgery, University of Lübeck; P G Lankisch and R Mahlke, Department of Internal Medicine, Städt. Krankenhaus Lüneburg; G Adler and B Glasbrenner, Department of Internal Medicine, University of Ulm; D Lorenz, Department of Surgery, University of Greifswald; H Bünte and U Sulkowski, Department of Surgery, University of Münster; H Goebel and M Rünzi, Department of Internal Medicine, University of Essen; W Kozuschek, Department of Surgery, Knappschaftskrankenhaus Langendreer Bochum; H Troidl and E Eypasch, Department of Surgery, University of Köln; I Klempa, Department of Surgery, Krankenhaus der Freien Hansestadt Bremen; G Allmendinger, Department of Internal Medicine, Klinik am Eichert Göppingen; S Liebe and M Löhr, Department of Internal Medicine, University of Rostock; J Witte, M Wienbeck and N Zügel, Departments of Gastroenterology and Surgery, Zentralklinikum Augsburg; H Zirngibl, Department of Surgery, University of Regensburg; A Hornung, Department of Surgery, Krankenhaus Bad Cannstatt; G Ramadori, Department of Gastroenterology and Endocrinology, University of Göttingen; E L Zurmeyer, Department of Surgery, Klinikum Ludwigshafen; W F Caspary, Department of Gastroenterology, University of Frankfurt; W Hartig, Department of Surgery, Städt. Klinikum St Georg Leipzig; R Bähr, Department of Surgery, Städt. Klinikum Karlsruhe; B Kremer, Department of Surgery, University of Kiel; P Czygan, Department of Internal Medicine, Lukaskrankenhaus Neuss; P Malfertheiner and U von Arnim, Department of Internal Medicine, University of Magdeburg; J Mössner, Department of Internal Medicine, University of Leipzig; D Kraus, Department of Surgery, Klinikum Nürnberg; D Häussinger and C Niederau, Department of Internal Medicine, University of Düsseldorf; T Sauerbruch, Department of Internal Medicine, University of Bonn; H Lippert and H U Schulz, Department of Surgery, University of Magdeburg; M Rothmund, Department of Surgery, University of Marburg; E Farthmann, Department of Surgery, University of Freiburg; W Schmiegel, Department of Internal Medicine, Knappschaftskrankenhaus Langendreer Bochum. We are indebted to the patients, to the chief investigators and physicians at the participating hospitals, and to Dr Ethan Taub for editorial assistance.
Abbreviations used in this paper
- ARDS
- adult respiratory distress syndrome
- CRF
- case report form
- CT
- contrast enhanced computed tomography
- ERCP
- endoscopic retrograde cholangiopancreatography
- ITT
- intent to treat
- O
- octreotide
- P
- placebo
- RCT
- randomised controlled trial
- SIRS
- systemic inflammatory response syndrome
- VFE
- valid for efficacy