The remarkable renewal of interest in the use of LIVER, PANCREAS, AND BILIARY TRACT

Transcription

1 GASTROENTEROLOGY 1998;115: LIVER, PANCREAS, AND BILIARY TRACT Controlled Clinical Trial of Pefloxacin Versus Imipenem in Severe Acute Pancreatitis CLAUDIO BASSI,* MASSIMO FALCONI,* GIORGIO TALAMINI, GENEROSO UOMO, GUIDO PAPACCIO, CHRISTOS DERVENIS, ROBERTO SALVIA,* ELISA BERTAZZONI MINELLI, # and PAOLO PEDERZOLI* *Surgical, Medical, and # Pharmacology Departments, Borgo Roma University Hospital, Verona, Italy; Pancreatic Disease Center, Cardarelli Hospital, Naples, Italy; 1st Surgical Division, Umberto I Hospital, Venice, Italy; and Surgical Department, Agia Holga Hospital, Athens, Greece Background & Aims: Antibiotic prophylaxis in severe pancreatitis has recently yielded promising clinical results, with imipenem significantly reducing the incidence of infected necrosis compared with an untreated control group. On the bases of pefloxacin s spectrum of action and pancreatic penetration, we investigated whether such drugs represent a valid alternative to imipenem. Methods: In a multicenter study, 60 patients with severe acute pancreatitis with necrosis affecting at least 50% of the pancreas were randomly allocated to receive intravenous treatment for 2 weeks with pefloxacin, 400 mg twice daily (30 patients), or imipenem, 500 mg three times daily (30 patients), within 120 hours of onset of symptoms. Age, sex, body weight, Ranson and Apache II scores, C-reactive protein, etiology, and time from onset of symptoms to treatment were well matched in the two groups. Results: The incidences of infected necrosis and extrapancreatic infections were 34% and 44%, respectively, in the pefloxacin group and 10% and 20% in the imipenem group. Imipenem proved significantly more effective in prevention of pancreatic infections (P I 0.05). Mortality was not significantly different in the two groups. Conclusions: Despite its theoretical potential, pefloxacin is inferior to imipenem in the prevention of infections associated with severe pancreatitis. The remarkable renewal of interest in the use of antibiotics in severe pancreatitis 1 19 is attributable to two factors: first, there is a widespread conviction that infected pancreatic necrosis is the main cause of morbidity and mortality associated with the disease 5 ; second, prophylaxis on a sounder scientific basis than in the past 4,5,12,13 19 has yielded promising clinical results In 1993, the Verona group 7 was the first to show that an antibiotic such as imipenem was capable of significantly reducing the incidence of infected necrosis in treated patients compared with that in controls. Over the following years, various controlled trials conducted with different drugs and treatment regimens 8 11 confirmed a positive trend in favor of the systematic use of antibiotics A number of uncontrolled clinical experiences, 6,20,21 on the strength not only of the abovementioned results reported in humans but also of data obtained in the experimental setting, 22 have been conducted with imipenem, achieving a general reduction in the incidence of pancreatic infections compared with historical control subjects. Theoretically, the main alternative to imipenem should be a drug belonging to the fluoroquinolone class. 4,5,13 15,23 25 In light of these considerations, we designed a prospective, randomized, multicenter trial using pefloxacin, which was the only fluoroquinolone commercially available in the injectable form when the trial protocol was drawn up. The aim of the trial was to assess the role of pefloxacin in the prevention of infected necrosis in patients suffering from pancreatitis with a severe necrotic component ( 50% of the pancreatic volume) compared with a control population that, for obvious ethical reasons, had to be a well-matched patient group treated with imipenem. Patients and Methods Sixty patients progressively observed and screened out of a total of 246 with acute necrotizing pancreatitis in four centers (Surgical Department, University of Verona, 24 patients; Pancreatic Disease Center, Cardarelli Hospital, Naples, 13 patients; Surgical Department, Agia Holga Hospital, Athens, 13 patients; 1st Surgical Division, Mestre Hospital, 10 patients) over the period from July 1991 to July 1997 were recruited into the trial after it was ascertained that they fulfilled the following inclusion criteria: no history of pancreatic disease; definite diagnosis of severe pancreatitis of any Abbreviations used in this paper: CT, computed tomography; CRP, C-reactive protein; US, ultrasonography; FNA, fine-needle aspiration; MOF, multiorgan failure by the American Gastroenterological Association /98/$3.00

2 1514 BASSI ET AL. GASTROENTEROLOGY Vol. 115, No. 6 etiology with onset of pain symptoms occurring not more than 5 days before admission; definite evidence of pancreatic necrosis as detected by computed tomography (CT) and intravenous contrast medium and confirmed by C-reactive protein (CRP) values above 100 mg/l and extending to at least 50% of the volume of the gland; and no antibiotic intake during the hours immediately before admission. After informed consent was obtained, the patients were randomly allocated (random numbers table) to 2 weeks intravenous treatment with either imipenem, 500 mg 3 times daily (group I), or pefloxacin, 400 mg 2 times daily (group P). All patients received the usual support therapy and total parenteral nutrition. Patients with pancreatitis of biliary etiology underwent endoscopic sphincterotomy within 72 hours of admission. The clinical course was assessed using routine laboratory tests and markers for suspected infection of pancreatic origin. 5 The morphological picture was evaluated at weekly intervals by means of ultrasonography (US), CT, or both. In the presence of suspected infection of the necrosis, patients underwent fineneedle aspiration (FNA) and culture examinations. 6 Evidence of infection prompted recourse to surgical treatment. 5 The primary end points evaluated were differences in incidence of pancreatic and extrapancreatic infections, the latter comprising pulmonary (clinical pneumonia confirmed by radiographs), urinary tract, and intravenous catheter infections confirmed by positive cultures. Secondarily, the 2 treatment groups were also compared in terms of need for surgery, length of hospital stay, and mortality. Statistical Analysis The goal of the study was to test the null hypothesis that (1) the proportion of pancreatic infections and (2) the incidence of mortality were identical in the 2 groups. The criterion for significance ( ) was set at 0.05 (2-tailed test). To achieve a statistical power of 80%, the proposed sample size was 50 patients per group. Data were analyzed using the Fisher exact, 2, and Mann Whitney U tests. Results Table 1 shows the main demographic data, etiology of pancreatitis, and severity parameter values recorded in the patients analyzed; there were no significant differences in these values between the 2 treatment groups. Two patients (1 in group I and 1 in group P) had CT findings that, in an overall review a posteriori, were not entirely consistent with necrosis affecting at least 50% of the pancreatic volume. Another 2 patients (both in group P) did not complete the scheduled antibiotic treatment because they died on days 4 and 8, respectively, as a result of multiorgan failure (MOF) with bacteriologically sterile necrosis (fulminant acute pancreatitis). We considered all these patients in the final analysis on an intention-totreat basis; in any event, their inclusion in the analysis population had no decisive bearing on the final outcome of the trial. The 17% overall mortality rate, in conjunction with the mean Ranson score at 24 hours (4.6) and the mean Apache II score at admission to hospital (11.5), clearly bears witness to the effective severity of disease in the sample studied. Also well-matched in the 2 groups was the mean time that elapsed between onset of disease symptoms and antibiotic treatment (52.8 vs hours). Given the trial entry criterion regarding extent of necrosis, the risk of developing infection was evenly matched in the 2 treatment groups. 26 Because of the broad spectrum of activity of the drugs used, both of which are capable of covering common urinary tract and respiratory pathogens, no changes in antibiotic administration were necessary during the scheduled 14-day treatment period. The clinical course 5 called for execution of an FNA in Table 1. Epidemiological Data, Severity, and Etiology of Severe Acute Pancreatitis in the Two Treatment Groups Imipenem Pefloxacin Total P Sex (M/F) 16/14 18/12 34/26 NS Age (yr ) 50 (34 67) 54 (36 70) 51 (34 70) NS Body wt (kg ) 66 (51 95) 66 (52 93) 66 (51 95) NS Ranson score a 4.4 (3 6) 4.7 (3 6) 4.6 (3 6) NS Apache II score a 12 (10 21) 11 (9 22) 11.5 (9 22) NS CRP (mg/l) 301 ( ) 314 ( ) 307 ( ) NS Etiology Biliary (66.1%) NS Biliary alcoholic (14%) NS Alcoholic 4 5 NS Post-ERCP 2 0 NS Idiopathic 2 2 NS Days from abdominal pain to admission 2.1 (1.5 5) 1.9 (0.5 5) 2.0 (0.5 5) NS NOTE. n 60 patients. Data represent means (ranges). ERCP, endoscopic retrograde cholangiopancreatography. a At admission.

3 December 1998 PEFLOXACIN VS. IMIPENEM IN ACUTE PANCREATITIS of 60 patients, 14 in group I and 19 in group P; the examination was repeated in 12 patients. Eighteen patients (10 at the first FNA and 8 at the second) had positive results, and 20 had negative. Three patients (in group P) had positive results during the second week of therapy; after undergoing surgery, they continued the antibiotic treatment up to the end of the scheduled course of therapy. Six patients in group P had positive results at FNA in the third and fourth weeks after the onset of pancreatitis, and 1 had positive results in the fifth week. All 3 infected patients in group I had positive results during the fourth week. Ten of 30 patients in group P developed infected necrosis (34%), compared with 3 of 30 patients in group I (10.0%); the difference in favor of imipenem was statistically significant (P 0.034). All patients with infected necrosis were treated surgically; thus, statistically the need for surgery for this indication was significantly greater in the pefloxacin group. The incidence of extrapancreatic infections was also greater in group P (44% vs. 20.0%), though not significantly (P 0.059). Seven group P patients and 2 group I patients had concomitant pancreatic and extrapancreatic infections. The extrapancreatic infections detected in a total of 13 group P patients consisted of 9 bronchopulmonary infections, 5 urinary tract infections, and 2 infections of the central venous catheter; 2 patients had concomitant lung and urinary tract infections, and 1 had a combination of bronchopulmonary infection and intravascular catheter infection. All 6 extrapancreatic infections in group I affected the respiratory tract. Eight of the 10 group P patients with infected necrosis had positive blood culture results (4 deaths), as did all 3 group I patients (3 deaths). The mean hospital stay in the group I survivors was 29 days (range, years) and 31 days in group P (range, years). Despite the favorable trend registered in the imipenem group, mortality was not significantly different in the 2 treatment groups (10% vs. 24%; P 0.18). All patients who developed infected necrosis despite imipenem prophylaxis died after surgical treatment (1 of MOF caused by Escherichia coli infection and 2 of MOF caused by polymicrobial infection with Staphylococcus aureus associated with E. coli and Candida glabrata, respectively). There were 5 of 10 postoperative deaths in the infected patients in group P (MOF caused by monomicrobial infections in 4 patients 3 C. glabrata and 1 Xanthomonas malthophilia and 1 death caused by polymicrobial infection with E. coli associated with Bacteroides fragilis). Table 2 lists the microbiological isolates in the patients Table 2. Microbiological Isolates in 13 Patients With Infected Necrosis in the Two Treatment Groups with infected necrosis. These microorganisms proved resistant to the antibiotics used or were not covered by their antimicrobial activity spectra (Candida). Discussion Pefloxacin group a Imipenem group b Staphylococcus aureus (methicillin-resistant) 3 2 Candida glabrata 3 1 Pseudomonas aeruginosa 2 0 Escherichia coli 1 2 Bacteroides fragilis 1 0 Xanthomonas malthophilia 1 0 Acinetobacter 1 0 Enterococcus 1 0 Klebsiella 1 0 a Polymicrobial (4 patients): 2 P. aeruginosa S. aureus, 1 E. coli B. fragilis, 1 Klebsiella Acinetobacter, 8 positive blood cultures. b Polymicrobial (2 patients): 1 E. coli S. aureus, 1 C. glabrata S. aureus, 3 positive blood cultures. Although infectious complications are still regarded as the primary cause of mortality in severe pancreatitis, 5,27 the most recent studies 6,20,21 appear to provide evidence of a reduction in the overall infection rate. On the whole, the data obtained in the present study relating to the low incidence of pancreatic infections (13 of 60 patients, 22%) confirm the trend reported in the above-mentioned uncontrolled series compared with historical data from the same centers. Since the introduction of antibiotic prophylaxis protocols, Banks et al. 6 report that the incidence of infected necrosis is down from 67% to 32%, and Ho and Frey 21 report a decrease from 75% to 20%. The selection of a restricted patient sample, moreover, explains the lengthy period taken to conduct this trial (6 years) and the limited number of patients recruited, the latter unquestionably being a factor with a major bearing on the lack of statistical significance of the difference in mortality rates. On the other hand, recruitment of 50 patients per group would have further and unacceptably prolonged the study period by at least another 2 3 years with no appreciable increase in the statistical power of the study. The choice of a maximum-risk patient sample exclusively composed of patients with severe pancreatitis (extent 50%) in the present trial at least partly offsets the drawback presented by the fairly limited number of patients recruited. An adequate spectrum of action, together with effective, long-lasting pancreatic concentrations and low resistance, would appear to be the essential properties

4 1516 BASSI ET AL. GASTROENTEROLOGY Vol. 115, No. 6 required of an antibiotic for it to be useful in the prophylaxis of infection in severe pancreatitis. 4,5,13 19 Theoretically, pefloxacin possesses these properties; the decision to test the drug compared with a control group also receiving prophylactic antibiotic treatment was based on the strength of the evidence reported in the literature that there can now be little doubt as to the effective soundness of a policy involving the routine use of suitable antibiotics in severe pancreatitis. 6 11,18 21,28 Our study, with its 2 antibiotic treatment groups, differs materially from other trials that included non antibiotic-treated control subjects, 8,11 in that it also included patients with biliary pancreatitis requiring endoscopic retrograde cholangiopancreatography and notoriously at risk of concomitant cholangitis. Comparison of pefloxacin and imipenem showed a 34% incidence of infected necrosis in the sample treated with pefloxacin compared with a significantly lower figure of 10% in the imipenem-treated group. The percentage of infections in group P is comparable to the incidences reported for many untreated control groups in similar trials. Pederzoli et al., 7 for instance, report an infection rate of 30%, Sainio et al. 8 report a rate of 40%, Luiten et al. 10 report a rate of 38%, and Delcenserie et al. 11 report a rate of 33%. Pefloxacin, despite its theoretical potential, as indicated by the data on its spectrum of action and its bioavailability at the pancreatic level even in the course of the disease, 4,5,13 15,23 25 presented such limited efficacy in our series that it can hardly be considered a valid alternative to imipenem in prevention of infected pancreatic necrosis. In our conditions, all deaths were caused by septic shock syndrome. The pathogens responsible in group P were either resistant microorganisms such as Staphylococcus or Pseudomonas aeruginosa or nonsusceptible microorganisms such as fungi or multiresistant bacteria that are very difficult to treat (Xanthomonas malthophilia, Acinetobacter). The serum and pancreatic concentrations of pefloxacin were substantially higher 15 than the mean inhibitory concentrations (MIC 90 s) of the bacteria isolated from infected tissue. 29 These results probably represent a selection phenomenon because pefloxacin (like other fluoroquinolones) has been shown to provide incomplete cover against this group of bacteria. Pefloxacin, both alone and in combination with other agents, is effective in the treatment of neutropenic sepsis and gram-negative bacteremia, including nosocomially acquired infections. It is also considered effective in selective decontamination of the gastrointestinal tract. 29 Therapeutic failures are generally caused by methicillinresistant staphylococci and P. aeruginosa. Resistance to fluoroquinolones and other unrelated antibiotics is now more frequent among nosocomial S. aureus and is rapidly increasing. 30 Despite its broad spectrum and substantial intrinsic antimicrobial activity, imipenem can select resistant microorganisms. 29 In group I, the deaths were attributable to infections caused by E. coli in 1 patient and 2 polymicrobial infections sustained by S. aureus and E. coli and by S. aureus and C. glabrata, respectively, i.e., by resistant or nonsusceptible microorganisms. Our experience with pefloxacin in acute necrotizing pancreatitis reflects the infection pattern and emergent bacterial resistance trend described in other hospital departments. Recent data have shown increasing staphylococcal and enterococcal colonization in intensive care unit patients. 31 The incidence of fungemia, and primarily candidemia, has also been rising substantially over recent years. There is a discernible shift toward non-albicans Candida species as new emergent pathogens: C. glabrata fungemia, for instance, has been associated with a higher mortality rate than infections with other species. 32 Infection may occur despite appropriate antibiotic therapy, and in many cases the prognosis is poor. The infection and the underlying disease probably join forces to worsen the patient s condition. There does not appear to be any very appreciable difference between grampositive and gram-negative bacteria in terms of hemodynamic parameters. Host factors and the underlying disease are major determinants in the development of septic shock in bacteremic patients. 33 Infections in the group I patients set in some time after the end of antibiotic prophylaxis, whereas in group P the necrosis became infected during the treatment. In addition to antibiotic use and other measures, the clinical outcome in such patients also appears to be related above all to patient characteristics and most notably to the extent of necrosis and to high scores for clinicobiochemical indicators. The latter are representative of the patient s general condition, including altered host defenses, cell reactivity, and endothelial damage with production of naturally occurring substances such as cytokines and nitric oxide in response to inflammatory and proinflammatory stimuli In conclusion, the prophylaxis of infected necrosis and pancreatic abscesses continues to intrigue clinical pancreatologists. To the best of our knowledge, pefloxacin is the first quinolone to be tested in acute pancreatitis and has yielded disappointing results, leaving imipenem as the drug of choice for single-drug prophylaxis. Other -lactam antibiotics (mezlocillin and cefuroxime 8 ) used in the clinical setting have shown only limited efficacy, prob-

5 December 1998 PEFLOXACIN VS. IMIPENEM IN ACUTE PANCREATITIS 1517 ably because of their restricted spectrum of activity against the organisms responsible. The evolution of the bacterial flora responsible for the pancreatic infections; the onset of drug resistance; the possibility of adopting other measures, 5,16,18 particularly intestinal decontamination 10 ; and the rational prospect of an effective drug combination 4,5,11,16 all constitute formidable stimuli to those working in the fields of basic and clinical research aimed at improving prevention and treatment of infected pancreatic necrosis. References 1. Craig RM, Dordal E, Myles L. The use of ampicillin in acute pancreatitis. Ann Intern Med 1975;83: Howes R, Zuidema GD, Cameron JL. Evolution of prophylactic antibiotics in acute pancreatitis. J Surg Res 1975;18: Finch WT, Sawyers JL, Schenker S. A prospective study to determine the efficacy of antibiotics in acute pancreatitis. Ann Surg 1976;183: Bradley EL III. Antibiotics in acute pancreatitis. Current status and future directions. Am J Surg 1989;158: Bassi C. Infected pancreatic necrosis. Int J Pancreatol 1994;16: Banks PA, Gerzof SG, Langevin RE, Silverman SG, Sica GT, Hughes MD. CT-guided aspiration of suspected pancreatic infection. Int J Pancreatol 1995;18: Pederzoli P, Bassi C, Vesentini S, Campedelli A. A randomized multicenter clinical trial of antibiotic prophylaxis of septic complications in acute necrotizing pancreatitis with imipenem. Surg Gynecol Obstet 1993;176: Sainio V, Kemppainen E, Puolakkainen P, Taavitsainen M, Kivisaari L, Valtonen V, Happianen R, Schröder T, Kivilaakso E. Early antibiotic treatment in acute necrotizing pancreatitis. Lancet 1995;346: Takeda K, Matsuno S, Sunamura M, Kakugawa Y. Continuous regional arterial infusion of protease inhibitor and antibiotics in acute necrotizing pancreatitis. Am J Surg 1996;171: Luiten EJT, Hop WCJ, Lange JF, Bruning HA. Controlled clinical trial of selective decontamination for the treatment of severe acute pancreatitis. Ann Surg 1995;222: Delcenserie R, Yzet T, Ducroix JP. Prophylactic antibiotics in treatment of severe acute alcoholic pancreatitis. Pancreas 1996; 13: Bassi C, Falconi M, Girelli R, Nifosi F, Elio A, Pederzoli P. Microbiological findings in severe pancreatitis. Surg Res Commun 1989;5: Büchler M, Malfertheiner P, Friess H, Isenmann R, Vaneck E, Grimm H, Schlegel P, Friess T, Beger H. Human pancreatic tissue concentration of bactericidal antibiotics. Gastroenterology 1992; 103: Bassi C, Pederzoli P, Vesentini S, Falconi M, Bonora A, Abbas H, Benini A, Pederroci P. Behavior of antibiotics during human necrotizing pancreatitis. Antimicrob Agents Chemother 1994;38: Bertazzoni Minelli E, Benini A, Muner A, Bassi C, et al. Pefloxacin penetration into human necrotic pancreatic tissue. J Antimicrob Chemother 1996;38: Barie PS. A critical review of antibiotic prophylaxis in severe acute pancreatitis. Am J Surg 1996;172: Oldach D. Antibiotic prophylaxis for necrotizing pancreatitis. Lancet 1995;346: Bassi C, Falconi M, Caldiron E, Salvia R, Bertazzoni EM, Pederzoli P. Use of antibiotics in necrotizing pancreatitis. Probl Gen Surg 1997;13: Steinberg W, Tenner S. Medical progress: acute pancreatitis. N Engl J Med 1994;330: Uomo G, Visconti M, Manes G, Calise F, Laccetti M, Rabitti PG. Nonsurgical treatment of acute necrotizing pancreatitis. Pancreas 1996;12: Ho HS, Frey CF. The role of antibiotic prophylaxis in severe acute pancreatitis. Arch Surg 1997;132: Mithöfer K, Fernandez-del Castillo C, Ferraro MJ, Lewandrowski K, Rattner DW, Warshaw AL. Antibiotic treatment improves survival in experimental acute necrotizing pancreatitis. Gastroenterology 1996;110: Tydén G, Brattström C, Malmborg AS. Early antibiotic treatment in acute necrotizing pancreatitis. Lancet 1995;346: Brattström C, Malmborg AS, Tydén G. Penetration of ciprofloxacin and ofloxacin into human allograft pancreatic juice. J Antimicrob Chemother 1988;22: Brattström C, Malmborg AS, Tydén G. Penetration of pefloxacin into human allograft pancreatic juice. J Antimicrob Chemother 1990;25: Vesentini S, Bassi C, Talamini G, Cavallini G, Campedelli A, Pederzolli P. Prospective comparison of C-reactive protein level, Ranson score and contrast-enhanced computed tomography in the prediction of septic complications of acute pancreatitis. Br J Surg 1993;80: Renner IG, Savage WT, Pantja JW, Ranner VJ. Death due to acute pancreatitis: a retrospective analysis of 405 autopsy cases. Dig Dis Sci 1985;30: Powell JJ, Miles R, Siriwardena AK. Antibiotic prophylaxis in the initial management of severe acute pancreatitis. Br J Surg 1998;85: Kucers A, Crowe S, Grayson ML, Hoy J. The use of antibiotics. 5th ed. Oxford, England: Butterworth-Heinemann, 1997: Coronado VG, Edwards JR, Culver DH, et al. Ciprofloxacin resistance among nosocomial Pseudomonas aeruginosa and Staphylococcus aureus in the United States. Infect Control Hosp Epidemiol 1995;16: Dellamonica P, Bernard E. Fluoroquinolones and surgical prophylaxis. Drugs 1993;45(suppl 3): Pfaller MA. Epidemiology and control of fungal infections. Clin Infect Dis 1994;19(suppl 1):S8 S Leibovici L, Drucker M, Konigsberger H, Samra Z, Harrari S, Ashkenazi S, Pitlik SD. Septic shock in bacteremic patients: risk factors, features and prognosis. Scand J Infect Dis 1997;29: Glauser MP, Zanetti G, Baumgartner JD, Cohen J. Septic shock: pathogenesis. Lancet 1991;338: Bone RC. Why sepsis trials fail. JAMA 1996;276: Kaboré AF, Denis M, Bergeron MG. Association of nitric oxide production by kidney proximal tubular cells in response to lipopolysaccharide and cytokines with cellular damage. Antimicrob Agents Chemother 1997;41: Received March 10, Accepted September 14, Address requests for reprints to: Claudio Bassi, M.D., Dipartimento di Scienze Chirurgiche, Servizio di Chirurgia Endocrina e Pancreatica, Policlinico B.go Roma, Via delle Menegone, Verona, Italy. Fax: (39) Supported by Italian Ministry of the University (MURST 60%) grant The authors thank Anthony Steele, M.A. Oxon., Lecturer in Medical English, University of Verona, for the translation of this article.