Imipenem: Antimicrobial Activity, Susceptibility, Administration and Dosage, Toxicity, Clinical Uses etc.

The carbapenems are a group of bicyclic beta-lactam compounds with a common carbapenem nucleus. Imipenem (N-formimidoyl-thienamycin, C12H17N3O4S; average molecular weight 299.3460) is one of these and it has substantial activity against both aerobic and anaerobic bacteria, including many (but not all) beta-lactamase-producing strains.
In the 1970s, Beecham Research Laboratories identified a carbapenem group, called olivanic acids, which were beta-lactamase inhibitors and broad-spectrum antibiotics (Butterworth et al., 1979).

At about the same time, Merck, Sharp and Dohme Research Laboratories independently identified thienamycin, derived from Streptomyces cattleya. This had the same carbapenem nucleus, but different side chains, which increased its antibacterial potency (Cassidy et al., 1981). Its chemical instability in concentrated solution was overcome by crystallization of the N-formimidoyl derivative of thienamycin, called imipenem. Urinary recovery in vivo was less than expected owing to extensive renal tubular metabolism by a brush border dipeptidase enzyme, dehydropeptidase I. A selective competitive antagonist of this enzyme was then developed and named cilastatin, which has similar pharmacokinetics to imipenem. Imipenem and cilastatin are marketed in combination in a 1:1 ratio (Kahan et al., 1983). Cilastatin not only prevents the degradation of imipenem, but may also protect the kidneys against potential toxic effects exerted by higher doses of imipenem (Rodloff et al., 2006).

The chemical structure of imipenem is shown in Figure 36.1.

ANTIMICROBIAL ACTIVITY

a. Routine susceptibility

Imipenem is active against a broad range of Gram-positive and Gramnegative bacteria, including both aerobes and anaerobes. Imipenem may have activity against some Mycobacteria spp., but Mycoplasma, Chlamydia, Legionella, Stenotrophomonas, Burkholderia, Clostridium difficile, and methicillin-resistant Staphylococcus aureus (MRSA) are not within its antimicrobial spectrum (Rodloff et al., 2006).

Gram-positive cocci

From the clinical standpoint, the activity of imipenem against Gram-positive cocci is generally far less emphasized than against Gramnegative bacteria, because antibiotics that have activity purely against Gram-positive bacteria are available. However, in terms of in vitro activity, imipenem has potent activity against many of the common Gram-positive human pathogens.

Imipenem has good inhibitory activity against Enterococcus faecalis, but, similar to penicillin G, it is not bactericidal (Eliopoulos and Moellering, 1981). Beta-lactamase-producing strains of E. faecalis are imipenem sensitive, but as most of these strains also exhibit highlevel gentamicin resistance, a synergistic imipenem–aminoglycoside combination usually cannot be obtained for the purpose of treating E. faecalis endocarditis (Hindes et al., 1989; Markowitz et al., 1991).

Gram-positive bacilli

Listeria monocytogenes is sensitive to imipenem (Kropp et al., 1985; Safdar et al., 2002; Safdar and Armstrong, 2003), but imipenem alone or combined with gentamicin is less effective than ampicillin or ampicillin–gentamicin against L. monocytogenes infections in experimental animals (Kim, 1986). Also L. monocytogenes mutants, resistant to imipenem and penicillin G can be selected in vitro. In these, penicillin-binding protein (PBP) 3 was altered and had decreased affinity for imipenem and penicillin G (Pierre et al., 1990).

Gram-positive anaerobes

Anaerobes such as Peptococcus, Peptostreptococcus, Actinomyces, Clostridium, Propionibacterium and Lactobacillus spp. are nearly always imipenem sensitive (Table 36.2) (Aldridge, 2002; Goldstein et al., 2004; Koeth et al., 2004; Edmiston et al., 2005; Wexler et al., 2005; Roberts et al., 2006; Tanaka et al., 2006). Clostridium difficile is less sensitive, most strains needing 8–16 mg/ml for inhibition (Goldstein et al., 2004; Tanaka et al., 2006).

Gram-negative cocci

Both Neisseria meningitidis and N. gonorrhoeae are very sensitive to imipenem. Beta-lactamase-producing strains are equally susceptible (Jones, 1985; Kropp et al., 1985). The in vitro activity of imipenem against common Gram-negative pathogens is summarized in Table 36.2.

Gram-negative bacilli

Haemophilus influenzae and H. parainfluenzae are also imipenem sensitive, including beta-lactamase-producing strains (Kropp et al., 1985; Cerami and Shungu, 1986). Ampicillin-resistant H. influenzae strains, which do not produce beta-lactamase (‘‘BLNAR’’), but whose resistance is intrinsic and chromosomally mediated, usually have reduced susceptibility to other penicillins and cephalosporins.

Gram-negative anaerobes

Bacteroides fragilis and other members of the B. fragilis group, such as B. thetaiotaomicron, B. ovatus, B. vulgatus, B. distasonis, B. uniformis, and B. caccae are nearly always imipenem sensitive. According to a national survey in the USA from 1997 to 2004 using data for 5225 isolates referred by ten medical centers, the susceptibility rates of the Bacteroides fragilis group against imipenem were higher than 99.5%. The susceptibility trends showed an overall significant decrease in the geometric mean MIC for imipenem against all the Bacteroides fragilis group (Snydman et al., 2007). Although imipenem resistance rates among Bacteroides spp. strains remain very low, strains which produce a zinc-dependent metallo-beta-lactamase encoded by the cfiA gene have been reported.

Mycobacteria

Rapidly growing mycobacteria including Mycobacterium fortuitum, M. chelonae, and M. abscessus are variably susceptible to imipenem (Brown-Elliott and Wallace, 2002; Uslan et al., 2006; Huang et al., 2008). Typically, greater than 90% of M. fortuitum strains are susceptible to imipenem, whereas 40–60% of M. chelonae and M. abscessus strains are susceptible (Brown-Elliott and Wallace, 2002). Mycobacterium marinum strains, which causes ‘‘swimming pool granuloma’’ or ‘‘fish tank granuloma,’’ is relatively sensitive to imipenem (MIC50, 2 mg/ml, MIC90, 8 mg/ml, and range of MIC 0.5– 16 mg/ml) (Aubry et al., 2000). Resistance of M. tuberculosis to betalactam antibiotics (such as imipenem) is thought to be mediated by a class A beta-lactamase.

b. Emerging resistance and cross-resistance

Enterobacteriaceae

Clinical isolates of imipenem-resistant Enterobacteriaceae have been reported recently. Klebsiella pneumoniae is the most frequently reported imipemem-resistant member of the Enterobacteriaceae. However, several other genera (including Escherichia and Enterobacter) have been reported to become imipenem resistant. Resistance of the Enterobacteriaceae to imipenem can occur by a number of mechanisms (Paterson, 2006).

The first mechanism is the combination of porin loss and the expression of AmpC-type enzymes (ACT-1 and CMY-4) or ESBLs (SHV- and TEM-type). The second mechanism is the presence of a carbapenemase, a b-lactamase capable of hydrolysis of carbapenems, such as imipenem. Two types of carbapenemases have been detected in K. pneumoniae so far, namely class B metallo-b-lactamases (IMP-type and VIM-type) and class A carbapenemase [K. pneumoniae carbapenemasetype (KPC)]. These carbapenemases are mostly plasmid mediated and, to make matters worse, their genes tend to be harbored on a plasmid with other genes related to resistance (not only to other b-lactams, but also to fluoroquinolones and aminoglycosides).

Pseudomonas aeruginosa and Acinetobacter baumannii

Pseudomonas aeruginosa and A. baumannii are frequently responsible for hospital-acquired infections and treated with carbapenems, such as imipenem. Thus, emergence and spread of carbapenem-resistant P. aeruginosa and A. baumannii is of great concern. P. aeruginosa and A. baumannii become imipenem resistant by two principal mechanisms, either the loss of or reduced expression of the outer membrane porin (OMP) or the production of b-lactamases that hydrolyze imipenem. Efflux pumps may also contribute to the resistance of P. aeruginosa to various antibiotics. The most common system, MexAB-OprM, causes resistance to meropenem, but not imipenem (Rodloff et al., 2006).

MECHANISM OF DRUG ACTION

Imipenem, like all other b-lactam agents, inhibits bacterial cell wall synthesis by binding to and inactivating relevant transpeptidases, known as penicillin-binding proteins (PBPs) (Rodloff et al., 2006). Imipenem binds preferentially to PBP2 and PBP1, the transpeptidases implicated in elongation of the bacterial cell wall, and has weak affinity for PBP3, the primary target of aminopenicillins and cephalosporins. In general, a wide range of beta-lactam antibiotics which are highly bactericidal for rapidly growing bacteria become bacteriostatic for slowly growing bacteria (Cozens et al., 1986). However, imipenem seems to trigger autolysins and to produce a rapid bactericidal effect to both rapidly and slowly growing bacteria (Cozens et al., 1989).

MODE OF DRUG ADMINISTRATION AND DOSAGE

a. Adults

Imipenem is most commonly administered in a dosage of 500 mg every 6 hours or 1 g every 8 hours, intravenously (Rodloff et al., 2006). Both dosing regimens result in approximately the same time above MIC (Mouton et al., 2000). The maximum recommended dose of imipenem is 50 mg/kg/day.

b. Newborn infants and children

For pediatric patients Z3 months of age, the recommended dose is 15–25 mg/kg per dose every 6 hours (60–100 mg/kg/day, maximum daily dose, 2 g) (Ayalew et al., 2003; Rodloff et al., 2006). Imipenem is not recommended in pediatric patients o30 kg with impaired renal function, as no information is available on appropriate dosing (Merck, 2007).

c. Altered dosages

Impaired renal function

The product information of Primaxint gives a complicated scheme for dose adjustment based on impaired renal function. In brief, a reduction in dose is recommended when a patient has a creatinine clearance of r70 ml min/1.73 m2 and/or a body weight less than 70 kg. According to this scheme, the final adjusted dosage regimen also depends on the type and severity of infection and whether or not the infection is caused by a fully susceptible strain (as opposed to a ‘‘moderately susceptible’’ organism, such as some strains of P. aeruginosa) (Merck, 2007). A simplified dose adjustment schedule is to give the normal dose when creatinine clearance is W50 ml/min, 50% of the dose when creatinine clearance is 10–50 ml/min, and 25% of the dose when the creatinine clearance is less than 10 ml/min (Aronoff et al., 2007).

Impaired hepatic function

Cilastatin accumulates in patients with hepatic dysfunction. However, the adverse effects of this are unknown (Trotman et al., 2005). There is no recommendation for dose adjustment of imipenem–cilastatin in patients with hepatic dysfunction.

The elderly

In general, dosing modification for the elderly with normal renal function is not considered necessary. In a small trial of patients aged between 68 and 83 years who received i.v. imipenem 0.5 g 6-hourly, dosage reduction was not necessary if their glomerular filtration rate (GFR) exceeded 30 ml/min (Finch et al., 1986).

PHARMACOKINETICS AND PHARMACODYNAMICS

Imipenem is widely distributed in the body. After intravenous administration, it can be detected in sputum, pus, pleural fluid, synovial fluid, bone, aqueous humor, interstitial fluid, and in peritoneal fluid in patients undergoing elective abdominal surgery. CSF concentrations in patients with uninflamed meninges were relatively low (0.8 mg/ml) and the level in saliva was also low (0.38 mg/ml) (Jacobs et al., 1986; MacGregor et al., 1986; Wise et al., 1986; Rolando et al., 1994). In patients undergoing thoracotomy, the imipenem concentration was 10.5 mg/ml in pericardial fluid, but only 0.28 mg/g in lung tissue, 2.25 hours after administration of 1 g imipenem–cilastatin i.v. (Benoni et al., 1987).

c. Clinically important pharmacokinetic and pharmacodynamic features

As with other b-lactam antibiotics, imipenem shows time-dependent antibiotic killing and its bactericidal activity relates most to the time that serum drug concentrations remain above the MIC (TWMIC) for a given organism (Craig, 1998; Turnidge, 1998; Sun et al., 2005). For effective bactericidal activity of imipenem, drug concentrations should exceed the MIC value for at least 40% of the dosing interval against a given organism (Sun et al., 2005; Rodloff et al., 2006). The Optimizing Pharmacodynamic Target Attainment using the MYSTIC (Meropenem Yearly Susceptibility Test Imformation Collection) Antibiogram (OPTAMA) program combines MIC information from a global surveillance study (MYSTIC program) with information derived from Monte Carlo simulation.

d. Excretion

Imipenem is predominantly excreted by glomerular filtration. Renal tubular secretion accounts for only a very small fraction of renal elimination (Norrby et al., 1983a; Kropp et al., 1985). Co-administration of probenecid or dehydropeptidase inhibitors has only a slight effect on imipenem pharmacokinetics and its serum half-life is unaltered (Norrby et al., 1983a; Norrby et al., 1983b).

Urinary recovery of imipenem administered alone is low and variable (6–38% of the dose), because a renal tubular dipeptidase enzyme metabolizes the drug by opening its lactam ring. A ratio of imipenem to cilastatin of at least 1:1 is necessary to maintain effective antibacterial urinary levels of imipenem, and therefore this fixed ratio has been adopted for clinical use of imipenem–cilastatin. Urinary excretion of each entity then accounts for about 70% of the dose (Rogers et al., 1985).

e. Drug interactions

Imipenem seems to be less prone to lowering concentrations of valproic acid than meropenem or panipenem. There has been only one case report that showed decreased concentrations of valproic acid during treatment by imipenem (Perea Falomir et al., 2006).

TOXICITY

a. Seizures

Imipenem–cilastatin has been considered to have proconvulsive activity and it has generally been considered that the drug is more prone to cause seizures than other carbapenem class antibiotics. It is noteworthy, however, that US product labeling indicates that the seizure rate for imipenem is 0.4%, compared with 0.5–0.7% for meropenem (Rodloff et al., 2006).

b. Nausea and vomiting

Nausea and/or vomiting has occurred in some 3–4% of patients receiving the drug. In a few, persistent vomiting necessitated stopping the drug. Slowing the rate of i.v. infusion appeared to lessen this side-effect in some patients (Calandra et al., 1985; Zajac et al., 1985). When high doses of imipenem were given to patients with cystic fibrosis, nausea and vomiting were more common and more severe (Pedersen et al., 1987).

c. Diarrhea and C. difficile colitis

Diarrhea has been observed in 3% of patients treated by imipenem– cilastatin (Leyland et al., 1992; Norrby and Gildon, 1999). In a study more than 20 years ago, C. difficile colitis during treatment by imipenem was reported to be rare (0.1% of all treated patients) (Calandra et al., 1986). However, the epidemiology and nature of C. difficile has dramatically changed recently. A recent systematic review and meta-analysis comparing various monotherapies of b-lactam agents in patients with neutropenenic fever demonstrated that imipenem was associated with significantly more frequent pseudomembranous colitis than cephalosporins (RR 2.07, 95% CI 1.28–3.34) (Paul et al., 2005b).

d. Hypersensitivity reactions

Some 2–3% of patients treated with imipenem–cilastatin have developed a rash, pruritus, or urticaria. It has been said that imipenem may be cross-allergenic with penicillins and cephalosporins and it should be avoided in patients with previous allergic reactions to these drugs, especially if the previous reaction was severe (Barza, 1985; Wang et al., 1985; Calandra et al., 1986; Boguniewicz and Leung, 1995). On the basis of positive skin tests, a 47.4% (9/19) rate of crossreactivity was found in a study performed more than 20 years ago(Saxon et al., 1988).

e. Hematologic side-effects

Neutropenia, a known side-effect of beta-lactam antibiotics, occurs in a small number of patients treated with imipenem–cilastatin. This was usually reversible on ceasing the drug (Gentry, 1985; Calandra et al., 1986). Eosinophilia was more common, but it was not associated with clinical abnormalities. A small number of patients developed a positive Coombs’ test, but there were no cases of hemolytic anemia. Changes in platelets and abnormal prothrombin times were rare (Calandra et al., 1986).

f. Hepatotoxicity

Abnormalities in liver function tests have been seen during imipenem– cilastatin therapy. These have usually been transient and without clinical signs of disease. Three patients during the early trials with imipenem–cilastatin developed jaundice, necessitating the discontinuation of the drug (Calandra et al., 1986).

g. Nephrotoxicity

In animals, concomitant administration of cilastatin eliminates the nephrotoxicity associated with high doses of imipenem–cilastatin (Kahan et al., 1983; Norrby, 1985). Daily doses of 4 g imipenem– cilastatin for up to 4 weeks have not been associated with nephrotoxicity in humans.

h. Risks in pregnancy

Imipenem–cilastatin is a Food and Drug Administration pregnancy category C drug (Merck, 2007) and an Australian Drug Evaluation Committee category B3 agent (Therapeutic Goods Administration, 1999). Imipenem crosses the placenta in considerable quantity. Heikkila et al. (1992) demonstrated that the concentration in amniotic fluid was 47%739% of simultaneous maternal concentration of plasma when sampled 3 hours after the infusion in early pregnancy, and 16%725% when sampled 30 minutes after the infusion in late pregnancy, and that mean concentration in umbilical venous and arterial blood was 33%712% and 31%713% of that in the maternal blood, respectively. The manufacturer of the drug states that there is no evidence of embryotoxicity or teratogenicity in animal experiments (Merck, 2007). There are, however, no adequate and well-controlled studies in pregnant women.

i. Breastfeeding

No reports describing the use of imipenem–cilastatin during human lactation are available and the effects on the nursing infant from exposure to the drug in milk are unknown. Until more data are available, we suggest using caution when considering the use of imipenem–cilastatin in lactating women.

CLINICAL USES OF THE DRUG

a. Hospital-acquired and healthcareassociated pneumonia, including ventilator-associated pneumonia

Initial empiric therapy

Imipenem–cilastatin is recommended in guidelines of the Infectious Diseases Society of America (IDSA) and the American Thoracic Society (ATS) as one option for the initial empirical treatment of hospital-acquired pneumonia (HAP), healthcare-associated pneumonia (HCAP), and ventilator-associated pneumonia (VAP). In particular, patients with duration of hospitalization Z5 days, antimicrobial therapy in the preceding 90 days, immunosuppressive disease/therapy, and presence in a hospital with a high frequency of antimicrobial resistance have a high risk of infection with a multiresistant Gram-negative bacillus (for example, P. aeruginosa, Enterobacter spp., K. pneumoniae, Acinetobacter spp., E. coli, and S. marcescens), and therefore may require imipenem (ATS/IDSA, 2005).

Definitive treatment

Pseudomonas aeruginosa

In 20–50% of patients with P. aeruginosa pneumonia, resistance to imipenem develops during therapy (Cometta et al., 1994; Fink et al., 1994; Jaccard et al., 1998; Carmeli et al., 1999; Zanetti et al., 2003; Nseir et al., 2008). This is most likely because of emergence of strains which have lost the OprD porin (Bonomo and Szabo, 2006). It is important to note that other antibiotics may also be associated with a substantial rate of emergence of resistance (Fink et al., 1994). Some have questioned the efficacy of imipenem–cilastatin in the treatment of P. aeruginosa pneumonia.

b. Intra-abdominal infections

Intra-abdominal infections are nearly always polymicrobial. Both aerobes and anaerobes may be involved, including Bacteroides spp. More resistant strains, requiring use of imipenem–cilastatin, may be related to nosocomial cases (for example, P. aeruginosa, Acinetobacter spp., ESBL-producing Enterobacteriaceae, Enterococcus spp). Vancomycin-resistant enterococci, MRSA, and Candida spp. may also occur in this setting, but will not be covered by imipenem (Dupont, 2007).

c. Neutropenic fever

Imipenem–cilastatin is one of the alternatives of initial antimicrobial therapy for neutropenic fever in high-risk patients. Imipenem has activity againt most of the causative pathogens in neutropenic fever including P. aeruginosa, Enterobacteriaceae (including ESBL producers), Streptococcus spp., and E. faecalis. As noted above under Grampositive cocci, imipenem will not reliably cover MRSA, MRSE, E. faecium, and nonbacterial pathogens. Imipenem–cilastatin therapy for neutropenic fever has been supported by substantial clinical experience.

d. Complicated urinary tract infections

Complicated urinary tract infection (UTI) may have different etiologies than uncomplicated UTI. Organisms isolated in complicated UTI, such as E. coli, Klebsiella spp., E. cloacae, S. marcescens, P. mirabilis, P. aeruginosa, and enterococci, may be more likely to be multiresistant and therefore need carbapenem therapy. Two randomized controlled trials demonstrated that imipenem–cilastatin was as effective as meropenem and piperacillin–tazobactam in the treatment of complicated UTI (see Table 36.12) (Cox et al., 1995; Naber et al., 2002).

e. Polymicrobial necrotizing fasciitis

Polymicrobial necrotizing fasciitis is an uncommon but life-threatening disease, associated with diabetes mellitus, morbid obesity, alcoholism, parenteral drug use, abdominal surgery, decubitus ulcer, perianal abscess, or vulvovaginal infection (DiNubile and Lipsky, 2004; Stevens et al., 2005). It is caused mainly by nongroupable streptococci, Enterobacteriaceae, Bacteroides spp., Peptostreptococcus spp., or P. aeruginosa (DiNubile and Lipsky, 2004).

f. Severe diabetic foot infections

There has been only one randomized controlled trial of imipenem– cilastatin for the treatment of diabetic foot infection, which demonstrated that ampicillin–sulbactam was as effective as imipenem–cilastatin for limb-threatening foot infection in diabetic patients (Grayson et al., 1994).

g. Prophylactic use in acute necrotizing pancreatitis

Secondary pancreatic or peripancreatic infection of acute necrotizing pancreatitis is a very important and controversial issue. Infection occurs in 40–70% of patients with necrotizing pancreatitis in the second or third week after onset of pancreatitis, and is the leading cause of the mortality and morbidity in this condition. Thus, early prevention of infectious complication of acute necrotizing pancreatitis has been advocated (Frossard et al., 2008). The causative organisms of pancreatic or peripancreatic infection secondary to acute necrotizing pancreatitis include E. coli, Klebsiella spp., Enterobacter spp., Proteus spp., P. aeruginosa, Bacteroides spp., Clostridium spp., and Enterococcus spp., all of which should be covered by carbapenems, such as imipenem (Dellinger et al., 2007).

h. Osteomyelitis

Imipenem–cilastatin has been used for the treatment of chronic osteomyelitis (Gentry, 1985; Gentry, 1988; MacGregor and Gentry, 1985). It is potentially useful in polymicrobial infections caused by nosocomial bacteria (such as methicillin-susceptible S. aureus and Gram-negative bacteria). Such osteomyelitis often follows trauma or orthopedic surgical procedures and often there are predisposing host factors, such as diabetes mellitus or peripheral vascular disease. In some parts of the world, ESBL-producing organisms are an increasingly frequent component of the flora causing diabetic foot infections (Gadepalli et al., 2006). In this context, carbapenems may be particularly useful for osteomyelitis.

References

Abdel-Haq NM, Papadopol R, Asmar BI et al. (2006). Antibiotic susceptibilities of Yersinia enterocolitica recovered from children over a 12-year period. Int J Antimicrob Agents 27: 449.
Acar JF, Goldstein FW, Kitzis MD et al. (1983). Activity of imipenem on aerobic bacteria. J Antimicrob Chemother 12 (Suppl D): 37.
Alarabi AA, Cars O, Danielson BG et al. (1990). Pharmacokinetics of intravenous imipenem/cilastatin during intermittent haemofiltration. J Antimicrob Chemother 26: 91.
Banks PA, Freeman ML (2006). Practice guidelines in acute pancreatitis. Am J Gastroenterol 101: 2379.
Banks PA (2006). Practice guidelines in acute pancreatitis. Am J Gastroenterol 101: 2379.
Cozens RM, Markiewicz Z, Tuomanen E (1989). Role of autolysins in the activities of imipenem and CGP 31608, a novel penem, against slowly growing bacteria. Antimicrob Agents Chemother 33: 1819.
Craig WA (1998). Pharmacokinetic/pharmacodynamic parameters: rationale for antibacterial dosing of mice and men. Clin Infect Dis 26: 1 quiz 11. Dambrauskas Z, Gulbinas A, Pundzius J et al. (2007). Meta-analysis of prophylactic parenteral antibiotic use in acute necrotizing pancreatitis. Medicina (Kaunas) 43: 291.
EUCAST (2008). Clinical breakpoints. Accessed February 5, 2009. Available from: www.escmid.org/research_projects/eucast/clinical_breakpoints/.
Farrugia DC, Eykyn SJ, Smyth EG (1994). Campylobacter fetus endocarditis: two case reports and review. Clin Infect Dis 18: 443.
Gurusamy KS, Farouk M, Tweedie JH (2005). UK guidelines for management of acute pancreatitis: is it time to change? Gut 54: 1344.
Haney JC, Pappas TN (2007). Necrotizing pancreatitis: diagnosis and management. Surg Clin North Am 87: 1431.
Hedberg M, Nord CE, the ESCMID Study Group on Antimicrobial Resistance in Anaerobic Bacteria (2003). Antimicrobial susceptibility of Bacteroides fragilis group isolates in Europe. Clin Microbiol Infect 9: 475.