Effect of adding metronidazole to meropenem on clinical efficacy in complicated intra-abdominal infections: a retrospective study Authors: Hajime Nakashima, Motoyasu Miyazaki, Naoaki Nishimura, Ryo Maeyama, Kenji Yoshikuni, Osamu Imakyure Published: 2026 ABSTRACT BackgroundComplicated intra-abdominal infections (cIAI), often caused by a diverse range of organisms, are commonly managed with appropriate empiric antimicrobial therapy, particularly broad-spectrum agents such as meropenem (MEPM). Anaerobes contribute to 30%–80% of the causative pathogens, and Bacteroides fragilis, a typical anaerobe, has been reported to exhibit increased resistance to antimicrobial agents. Metronidazole (MNZ) has shown significant antibacterial activity against anaerobes and a different mechanism of action; therefore, its addition to MEPM may improve its clinical efficacy. This study aimed to evaluate the impact of adding MNZ to MEPM on the clinical efficacy of empiric antimicrobial therapy for cIAI.MethodsThis retrospective study spanned 9 years and included 114 patients who underwent empiric antimicrobial therapy with MEPM monotherapy or MEPM plus MNZ combination therapy for cIAI caused by lower gastrointestinal perforation in conjunction with surgery or percutaneous drainage. Propensity score matching was applied to adjust for background factors. The primary outcome was the cure rate, while secondary outcomes included 28-day mortality and the duration of hospitalization following surgical intervention.ResultsThere were no significant differences in the cure rates, 28-day mortality or hospitalization durations between the MEPM monotherapy group and the MEPM plus MNZ group [cure rates: 91.2% (83/91) vs. 91.3% (21/23), respectively, p = 0.999; 28-day mortality: 2.2% (2/91) vs. 4.3% (1/23), respectively, p = 0.495; hospitalization durations, median (interquartile range): 18 days (11.5–26.5) vs. 20 days (15.5–32.0), respectively; p = 0.214]. Twenty patients were selected for each group using propensity score matching. No significant differences were observed in cure rates, 28-day mortality, or hospitalization durations between the MEPM monotherapy group and the MEPM plus MNZ group [cure rates: 85.0% (17/20) vs. 95.0% (19/20), respectively, p = 0.605; 28-day mortality: 10% (2/20) vs. 5% (1/20), respectively, p = 0.999; hospitalization durations, median (interquartile range): 16 days (11.8–24) vs. 20.5 days (16.3–38.3), respectively; p = 0.184].ConclusionIn this single-center retrospective cohort, there was no evidence that adding MNZ to MEPM improved clinical efficacy in empiric antimicrobial therapy with infection-source control in cIAI.Clinical trial numberNot applicable. Background Perforation involving the lower gastrointestinal tract below the small bowel can develop into complicated intra-abdominal infections (cIAI) caused by intra-abdominal bacterial leaks, leading to more serious conditions [1]. cIAI is defined as an infection involving multiple abdominal organs, causing localized or diffuse peritonitis and requiring control of the infection source and appropriate antibiotic treatment [2]. It is often a polymicrobial infection, and the causative bacteria include aerobic Gram-negative rods, such as Escherichia coli and Klebsiella pneumoniae, which are representative bacteria of the intestinal flora; Gram-positive cocci, such as Streptococcus species and Enterococcus species; and anaerobic bacteria among other bacterial species [1, 3–6]. Immediate empiric antibiotic therapy is often needed to cover a wide range of causative pathogens before culture and susceptibility results can dictate the choice of antimicrobials [2, 7]. The revised guidelines of the Surgical Infection Society [7] for the management of intra-abdominal infections recommend meropenem (MEPM), doripenem, imipenem–cilastatin, and cefepime plus metronidazole (MNZ) as options for empiric antimicrobial therapy for patients with sepsis, those with an acute physiology and chronic health evaluation II (APACHE II) score of ≥ 10, and those with diffuse peritonitis. In cIAI, the rate of isolation of anaerobes has been reported to be approximately 30% to 80%, with various bacterial species detected, including Bacteroides species, Clostridium species, Peptostreptococcus species, Prevotella species, and Fusobacterium species [4, 8, 9]. Among them, Bacteroides species, the most common obligate anaerobic Gram-negative rods, are normal flora in the gastrointestinal tract and cause infections in the intra-abdominal cavity, pelvic cavity, skin soft tissues, and other areas [1, 10, 11]. The most pathogenic species is the representative bacterial species B. fragilis [12, 13]. The mortality rate for Bacteroides species with bacteremia has been reported to be 14%–43% [14–18]. Therefore, it is crucial to select antibiotics that are effective against anaerobes, such as Bacteroides species, which are commonly detected in cIAI treatment. MEPM, a broad-spectrum carbapenem antibiotic, exhibits antimicrobial activity against Gram-positive bacteria; Gram-negative bacteria; and obligate anaerobic bacteria, including Bacteroides species, and is used to treat various bacterial infections [19]. However, the susceptibility rate of B. fragilis to MEPM has been declining in the recent years [20, 21]. On the other hand, MNZ exhibits high activity against anaerobes and is frequently used in combination with cephalosporins and fluoroquinolones, which have weak antibacterial activity against anaerobes, for managing intra-abdominal infections [22–25]. The rate of antibiotic resistance in B. fragilis ranges from 3.9% to 17% for MEPM, whereas it is 0%–4% for MNZ, indicating that B. fragilis is highly susceptible to MNZ [20, 21, 26–29]. Regarding the mechanism of action, MEPM exhibits high affinity for penicillin-binding proteins and inhibits bacterial cell-wall synthesis [30], whereas MNZ is absorbed into bacteria under anaerobic conditions, where it undergoes reduction and is converted into nitroso compounds; these nitroso compounds bind to deoxyribonucleic acid (DNA) and inhibit DNA synthesis, thereby exhibiting antibacterial activity [31]. Hence, clinical efficacy may improve when combining two drugs with different mechanisms of action for treating infections, such as cIAI, which have a high isolated frequency of anaerobic bacteria. Furthermore, the addition of MNZ may be a useful option for treating B. fragilis strains that show resistance to MEPM. However, no reported studies have assessed the clinical efficacy of the MEPM plus MNZ combination therapy as the empiric antimicrobials for cIAI. The present study aimed to evaluate the impact of adding MNZ to MEPM on the clinical efficacy of empiric antimicrobial therapy for cIAI. We retrospectively compared the cure rates, 28-day mortality and the duration of hospitalization, between the MEPM monotherapy and MEPM plus MNZ groups. Methods Study population This retrospective study included patients with colorectal perforation, small bowel perforation, or perforated appendicitis who were admitted to Japan Community Health Care Organization Kyushu Hospital between January 2016 and December 2024. The choice between MEPM monotherapy and MEPM plus MNZ was left to the discretion of the attending physician, as no standardized institutional protocol was in place. Patients who underwent surgery or percutaneous drainage for generalized peritonitis and were administered MEPM or MEPM combined with MNZ as the empiric antimicrobials for cIAI were considered eligible. The exclusion criteria were as follows: patients who were treated with conservative therapy using only antimicrobials, were not treated with MEPM or MEPM plus MNZ as initial empirical therapy, were < 18 years of age, were treated with the target antibiotics < 3 days, or had been transferred before treatment completion. Clinical backgrounds Patient information, treatment course, and outcomes were retrospectively retrieved from the electronic medical records, while microbial data from the submitted cultures were obtained from the microbiology laboratory database. The extracted data included age, sex, body weight, modified APACHE II score [32], intensive care unit (ICU) admission, serum creatinine, duration and dosage of target antibiotic administration, switch to oral antibiotics, perforation site, treatment details, concomitant bacteremia, and comorbidities. Creatinine clearance was calculated using the Cockcroft–Gault equation [33]. The perforation site and treatment details were obtained from the primary care physician’s diagnostic or surgical records. Patient comorbidities were considered present if documented in the admission history and physical examination. Definitions The primary outcome was the cure rate. Evaluation of clinical cure and treatment failure as outcome measures was based on prior studies of cIAI [8, 34]. Cure was defined as the complete resolution or marked improvement of infection-related signs and symptoms, completion of antibiotic therapy, and no requirement for additional surgical intervention or percutaneous drainage. Treatment failure was defined as death during cIAI treatment, or persistent or recurrent intra-abdominal infection leading to poor clinical outcomes, necessitating a change from the target antibiotics or requiring additional surgical intervention or percutaneous drainage. Secondary outcomes included 28-day mortality and the duration of hospitalization following surgical intervention. The duration of hospitalization was considered from the first day of surgery or percutaneous drainage. Antibiotic treatment For antibiotic treatment, either MEPM or MEPM plus MNZ (the target antibiotics) were continued until the end of treatment or when empirical treatment was changed to definitive treatment based on culture and susceptibility reports. If Enterococcus species resistant to the target antibiotics were isolated, vancomycin was additionally administered by the attending physician as deemed necessary. MEPM dosing regimens were categorized as 1 g every 8 h, 1 g every 12 h, 500 mg every 8 h, 500 mg every 12 h, or 500 mg every 24 h. MEPM was administered intravenously at doses determined by the attending physician in accordance with the hospital’s renal function–based protocol, infused over 30 min to 1 h via a venous catheter. MNZ was given intravenously at a dose of 500 mg every 8 h, also infused over 30 min to 1 h through a catheter. Statistical analysis The MEPM and MEPM plus MNZ groups were compared using the Mann–Whitney U test for continuous variables and the Fisher’s exact test for categorical variables, with statistical significance set at p < 0.05. This study was a single-center retrospective cohort. Patient allocation to treatment groups was non-random, and antibiotic selection was left to the discretion of the attending physician, as no standardized institutional protocol existed. To adjust for potential confounding factors, propensity score matching was performed. Logistic regression analysis was conducted with the target antibiotics as the independent variable, and age, modified APACHE II score, ICU admission, creatinine clearance, daily MEPM dose, and primary infection— clinically relevant variables associated with infection severity and prognosis —as explanatory variables to calculate propensity scores. Based on these scores, one-to-one nearest neighbor matching without replacement was applied, using a caliper width set at 20% of the standard deviation. Statistical analysis was performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan) [35]. Results Patient baseline characteristics Between January 2016 to December 2024, among the 361 patients who were admitted with colorectal perforation, small bowel perforation, or perforated appendicitis, 114 patients were included in this retrospective study, excluding 247 patients who did not meet the inclusion criteria (Fig. 1). Fig. 1Study flow chart. MEPM: meropenem; MNZ: metronidazole Study flow chart. MEPM: meropenem; MNZ: metronidazole The baseline characteristics of the sample are presented in Table 1. The median (interquartile range [IQR]) age of the patients was 69 (62.25–78), 49.1% were male, and the median (IQR) modified APACHE II score at the time of surgical intervention or percutaneous drainage was 11 (8–15). The median (IQR) creatinine clearance was 58 (33.9–83.8) mL/min, and the median duration (IQR) of targeted antibiotic administration was 7 (6–9) days. Approximately half of the patients were administered 3,000 mg/day of MEPM. Colorectal perforation was the most common site of gastrointestinal perforation (56.1%, 64/114). Multiple bacteria were isolated from the cultured samples collected in 62.3% (71/114) of all patients, whereas no significant bacteria were isolated in 23.5% of the patients. Twenty-one patients (18.4%) were switched from the target antibiotics to oral agents, of whom 15 (71.4%) were transitioned to levofloxacin. The most common comorbidity was hypertension, present in approximately 40% of the patients. Solid tumors were present in 28 patients (24.6%), while hematological malignancies were present in only 2 patients (1.8%). Table 1Characteristics of the patients with complicated intra-abdominal infectionsCharacteristicBefore propensity score matchingAfter propensity score matchingTotal cohort(n = 114)Treated with MEPM(n = 91)Treated with MEPM + MNZ(n = 23)p valueTreated with MEPM(n = 20)Treated with MEPM + MNZ(n = 20)p valueSMDAge in yr, median (IQR)69 (62.3–78)69 (63–78.5)69 (57–76)0.37169.5 (63.5–75)69.5 (60.3–76)0.9990.059Male, n (%)56 (49.1)45 (49.5)11 (47.8)0.9997 (35.0)9 (45.0)0.7480.205Wt (kg), median (IQR)55 (48–63.2)54.9 (48–62)56 (48.3–67.5)0.39357.0 (51.1–62.8)55.9 (48.4–67.4)0.8180.043Creatinine clearance on day of surgical intervention (ml/min), median (IQR)58 (33.9–83.8)56.7 (33.5–81.4)65.3 (43.9–98.0)0.18978.2 (58.5–93.1)63.0 (44.7–96.8)0.4610.251Modified APACHEⅡ score on day of surgical intervention, median (IQR)11 (8–15)11 (8.5–15)11 (7.5–14.5)0.95810.5 (7.8–12.8)11.5 (6.5–14.3)0.7450.100Intensive care unit admission,n (%)39 (34.2)35 (38.5)4 (17.4)0.0844 (20.0)4 (20.0)0.999< 0.001Duration of target antibiotic administration (day),median (IQR)7 (6–9)7 (6–9)8 (6–9)0.6006.5 (5–7)8 (6–9)0.0200.446Daily dose of MEPM, n (%) 3,000 mg60 (52.6)44 (48.4)16 (69.6)0.10114 (70.0)14 (70.0)0.999< 0.001 2,000 mg15 (13.2)14 (15.4)1 (4.3)0.2983 (15.0)1 (5.0)0.6050.338 1,500 mg13 (11.4)10 (11.0)3 (13.0)0.7241 (5.0)3 (15.0)0.6050.338 1,000 mg17 (14.9)14 (15.4)3 (13.0)0.9991 (5.0)2 (10.0)0.9990.191 500 mg9 (7.9)9 (9.9)0 (0)0.2001 (5.0)0 (0)0.9990.324Primary diagnosis, n (%) Colorectal perforation64 (56.1)55 (60.4)9 (39.1)0.09914 (70.0)9 (45.0)0.2000.523 Small bowel perforation17 (14.9)12 (13.2)5 (21.7)0.3310 (0)4 (20.0)0.1060.707 Perforated appendicitis33 (28.9)24 (26.4)9 (39.1)0.3036 (30.0)7 (35.0)0.9990.107Complicated with bacteremia, n (%)17 (14.9)13 (14.3)4 (17.4)0.7466 (30.0)3 (15.0)0.4510.365Number of pathogens from culture, n (%) Polymicrobial71 (62.3)58 (63.7)13 (56.5)0.6319 (45.0)11 (55.0)0.7520.201 Monomicrobial16 (14.0)10 (11.0)6 (26.1)0.0892 (10.0)6 (30.0)0.2350.516 No detection27 (23.7)23 (25.3)4 (17.4)0.5859 (45.0)3 (15.0)0.0820.693Switched to an oral agent, n (%)21 (18.4)15 (16.5)6 (26.1)0.3664 (20.0)5 (25.0)0.9990.120 Trimethoprim/sulfamethoxazole2 (1.8)0 (0)2 (8.7)0.0390 (0)2 (10.0)0.4870.471 Levofloxacin15 (13.2)12 (13.2)3 (13.0)0.9994 (20.0)2 (10.0)0.6610.283 Metronidazole4 (3.5)3 (3.3)1 (4.3)0.9990 (0)1 (5.0)0.9990.324Patients with the following medical history, n (%) Hypertension45 (39.5)37 (40.7)8 (34.8)0.6428 (40.0)7 (35.0)0.9990.103 Type 2 diabetes15 (13.2)10 (11.0)5 (21.7)0.1803 (15.0)4 (20.0)0.9990.132 Ischemic heart disease5 (4.4)3 (3.3)2 (8.7)0.2641 (5.0)1 (5.0)0.999< 0.001 Heart failure10 (8.8)7 (7.7)3 (13.0)0.4190 (0)2 (10.0)0.4870.471 Cerebrovascular disease10 (8.8)10 (11.0)0 (0)0.2093 (15.0)0 (0)0.2310.594 Chronic kidney disease14 (12.3)11 (12.1)3 (13.0)0.9991 (5.0)2 (10.0)0.9990.191 Solid tumors28 (24.6)21 (23.1)7 (30.4)0.5886 (30.0)6 (30.0)0.999< 0.001 Hematological malignancies2 (1.8)2 (2.2)0 (0)0.9990 (0)0 (0)0.999< 0.001MEPM: meropenem; MNZ: metronidazole; IQR: interquartile range; APACHE II: Acute Physiology and Chronic Health Evaluation II; SMD: standardized mean difference Table 1Characteristics of the patients with complicated intra-abdominal infectionsCharacteristicBefore propensity score matchingAfter propensity score matchingTotal cohort(n = 114)Treated with MEPM(n = 91)Treated with MEPM + MNZ(n = 23)p valueTreated with MEPM(n = 20)Treated with MEPM + MNZ(n = 20)p valueSMDAge in yr, median (IQR)69 (62.3–78)69 (63–78.5)69 (57–76)0.37169.5 (63.5–75)69.5 (60.3–76)0.9990.059Male, n (%)56 (49.1)45 (49.5)11 (47.8)0.9997 (35.0)9 (45.0)0.7480.205Wt (kg), median (IQR)55 (48–63.2)54.9 (48–62)56 (48.3–67.5)0.39357.0 (51.1–62.8)55.9 (48.4–67.4)0.8180.043Creatinine clearance on day of surgical intervention (ml/min), median (IQR)58 (33.9–83.8)56.7 (33.5–81.4)65.3 (43.9–98.0)0.18978.2 (58.5–93.1)63.0 (44.7–96.8)0.4610.251Modified APACHEⅡ score on day of surgical intervention, median (IQR)11 (8–15)11 (8.5–15)11 (7.5–14.5)0.95810.5 (7.8–12.8)11.5 (6.5–14.3)0.7450.100Intensive care unit admission,n (%)39 (34.2)35 (38.5)4 (17.4)0.0844 (20.0)4 (20.0)0.999< 0.001Duration of target antibiotic administration (day),median (IQR)7 (6–9)7 (6–9)8 (6–9)0.6006.5 (5–7)8 (6–9)0.0200.446Daily dose of MEPM, n (%) 3,000 mg60 (52.6)44 (48.4)16 (69.6)0.10114 (70.0)14 (70.0)0.999< 0.001 2,000 mg15 (13.2)14 (15.4)1 (4.3)0.2983 (15.0)1 (5.0)0.6050.338 1,500 mg13 (11.4)10 (11.0)3 (13.0)0.7241 (5.0)3 (15.0)0.6050.338 1,000 mg17 (14.9)14 (15.4)3 (13.0)0.9991 (5.0)2 (10.0)0.9990.191 500 mg9 (7.9)9 (9.9)0 (0)0.2001 (5.0)0 (0)0.9990.324Primary diagnosis, n (%) Colorectal perforation64 (56.1)55 (60.4)9 (39.1)0.09914 (70.0)9 (45.0)0.2000.523 Small bowel perforation17 (14.9)12 (13.2)5 (21.7)0.3310 (0)4 (20.0)0.1060.707 Perforated appendicitis33 (28.9)24 (26.4)9 (39.1)0.3036 (30.0)7 (35.0)0.9990.107Complicated with bacteremia, n (%)17 (14.9)13 (14.3)4 (17.4)0.7466 (30.0)3 (15.0)0.4510.365Number of pathogens from culture, n (%) Polymicrobial71 (62.3)58 (63.7)13 (56.5)0.6319 (45.0)11 (55.0)0.7520.201 Monomicrobial16 (14.0)10 (11.0)6 (26.1)0.0892 (10.0)6 (30.0)0.2350.516 No detection27 (23.7)23 (25.3)4 (17.4)0.5859 (45.0)3 (15.0)0.0820.693Switched to an oral agent, n (%)21 (18.4)15 (16.5)6 (26.1)0.3664 (20.0)5 (25.0)0.9990.120 Trimethoprim/sulfamethoxazole2 (1.8)0 (0)2 (8.7)0.0390 (0)2 (10.0)0.4870.471 Levofloxacin15 (13.2)12 (13.2)3 (13.0)0.9994 (20.0)2 (10.0)0.6610.283 Metronidazole4 (3.5)3 (3.3)1 (4.3)0.9990 (0)1 (5.0)0.9990.324Patients with the following medical history, n (%) Hypertension45 (39.5)37 (40.7)8 (34.8)0.6428 (40.0)7 (35.0)0.9990.103 Type 2 diabetes15 (13.2)10 (11.0)5 (21.7)0.1803 (15.0)4 (20.0)0.9990.132 Ischemic heart disease5 (4.4)3 (3.3)2 (8.7)0.2641 (5.0)1 (5.0)0.999< 0.001 Heart failure10 (8.8)7 (7.7)3 (13.0)0.4190 (0)2 (10.0)0.4870.471 Cerebrovascular disease10 (8.8)10 (11.0)0 (0)0.2093 (15.0)0 (0)0.2310.594 Chronic kidney disease14 (12.3)11 (12.1)3 (13.0)0.9991 (5.0)2 (10.0)0.9990.191 Solid tumors28 (24.6)21 (23.1)7 (30.4)0.5886 (30.0)6 (30.0)0.999< 0.001 Hematological malignancies2 (1.8)2 (2.2)0 (0)0.9990 (0)0 (0)0.999< 0.001MEPM: meropenem; MNZ: metronidazole; IQR: interquartile range; APACHE II: Acute Physiology and Chronic Health Evaluation II; SMD: standardized mean difference Characteristics of the patients with complicated intra-abdominal infections Intensive care unit admission, n (%) Duration of target antibiotic administration (day), median (IQR) MEPM: meropenem; MNZ: metronidazole; IQR: interquartile range; APACHE II: Acute Physiology and Chronic Health Evaluation II; SMD: standardized mean difference There were 91 patients in the MEPM monotherapy group and 23 in the MEPM plus MNZ group, with no significant difference in the duration of target antibiotic administration (median [IQR]: 7 [6–9] days and 8 [6–9] days, respectively; p = 0.600). Furthermore, no significant difference was observed between the two groups in terms of daily MEPM dose or the proportion of patients switched to oral antibiotics. In addition, there were no significant differences in perforation site or comorbidities between the two groups. The incidence of bacteremia was 14.3% (13/91) in the MEPM group and 17.4% (4/23) in the MEPM plus MNZ group (p = 0.746). After propensity score matching, 20 patients were included in the MEPM monotherapy group and 20 in the MEPM plus MNZ group. A significant difference was observed in the duration of administration of the target antibiotics (median [IQR]: 6.5 [5–7] days vs. 8 [6–9] days, p = 0.020); however, no differences were found in other variables. Isolated pathogens The major bacteria isolated from cultures at the time of surgical intervention are shown in Table 2. In the MEPM group, the most common bacteria isolated was Escherichia coli (51.6%, 47/91), followed by Enterococcus species, Streptococcus species, and Bacteroides non-fragilis, which were isolated from approximately 25% of the patients. In contrast, Escherichia coli and Streptococcus species were the most common pathogens isolated in the MEPM plus MNZ group, both of which were isolated from 39.1% (9/23) of the patients, followed by Enterococcus species and Prevotella species, which were isolated from 26.1% (6/23) of the patients. In addition, the isolation rates of anaerobic bacteria were 42.9% (39/91) in the MEPM group and 52.2% (12/23) in the MEPM plus MNZ group. Table 2Isolated pathogens from the samples obtained at the time of the surgical treatmentPathogen category organism/group, n (%)Total cohort(n = 114)Treated with MEPM(n = 91)Treated with MEPM + MNZ(n = 23)Aerobic Gram-negative75 (65.8)60 (65.9)15 (65.2) Escherichia coli56 (49.1)47 (51.6)9 (39.1) Escherichia coli (ESBL-positive)15 (13.2)13 (14.3)2 (8.7) Klebsiella pneumoniae7 (6.1)5 (5.5)2 (8.7) Klebsiella pneumoniae (ESBL-positive)1 (0.9)1 (1.1)0 (0) Enterobacter cloacae9 (7.9)8 (8.8)1 (4.3) Pseudomonas aeruginosa14 (12.3)13 (14.3)1 (4.3)Aerobic Gram-positive56 (49.1)44 (48.4)12 (52.2) Enterococcus species29 (25.4)23 (25.3)6 (26.1) Enterococcus faecium9 (7.9)6 (6.6)3 (13.0) Streptococcus species30 (26.3)21 (23.1)9 (39.1) Staphylococcus species5 (4.4)3 (3.3)2 (8.7) Staphylococcus aureus (MRSA)1 (0.9)0 (0)1 (4.3)Anaerobes51 (44.7)39 (42.9)12 (52.2) Bacteroides fragilis12 (10.5)10 (11.0)2 (8.7) Non-fragilis Bacteroides species24 (21.1)22 (24.2)2 (8.7) Peptostreptococcus species5 (4.4)3 (3.3)2 (8.7) Prevotella species22 (19.3)16 (17.6)6 (26.1) Fusobacterium species5 (4.4)1 (1.1)4 (17.4) Bifidobacterium species4 (3.5)3 (3.3)1 (4.3) Clostridium species3 (2.6)3 (3.3)0 (0)MEPM: meropenem; MNZ: metronidazole; MRSA, methicillin-resistant Staphylococcus aureus Table 2Isolated pathogens from the samples obtained at the time of the surgical treatmentPathogen category organism/group, n (%)Total cohort(n = 114)Treated with MEPM(n = 91)Treated with MEPM + MNZ(n = 23)Aerobic Gram-negative75 (65.8)60 (65.9)15 (65.2) Escherichia coli56 (49.1)47 (51.6)9 (39.1) Escherichia coli (ESBL-positive)15 (13.2)13 (14.3)2 (8.7) Klebsiella pneumoniae7 (6.1)5 (5.5)2 (8.7) Klebsiella pneumoniae (ESBL-positive)1 (0.9)1 (1.1)0 (0) Enterobacter cloacae9 (7.9)8 (8.8)1 (4.3) Pseudomonas aeruginosa14 (12.3)13 (14.3)1 (4.3)Aerobic Gram-positive56 (49.1)44 (48.4)12 (52.2) Enterococcus species29 (25.4)23 (25.3)6 (26.1) Enterococcus faecium9 (7.9)6 (6.6)3 (13.0) Streptococcus species30 (26.3)21 (23.1)9 (39.1) Staphylococcus species5 (4.4)3 (3.3)2 (8.7) Staphylococcus aureus (MRSA)1 (0.9)0 (0)1 (4.3)Anaerobes51 (44.7)39 (42.9)12 (52.2) Bacteroides fragilis12 (10.5)10 (11.0)2 (8.7) Non-fragilis Bacteroides species24 (21.1)22 (24.2)2 (8.7) Peptostreptococcus species5 (4.4)3 (3.3)2 (8.7) Prevotella species22 (19.3)16 (17.6)6 (26.1) Fusobacterium species5 (4.4)1 (1.1)4 (17.4) Bifidobacterium species4 (3.5)3 (3.3)1 (4.3) Clostridium species3 (2.6)3 (3.3)0 (0)MEPM: meropenem; MNZ: metronidazole; MRSA, methicillin-resistant Staphylococcus aureus Isolated pathogens from the samples obtained at the time of the surgical treatment MEPM: meropenem; MNZ: metronidazole; MRSA, methicillin-resistant Staphylococcus aureus Cure rates, 28-day mortality and hospitalization durations Comparison of the primary outcome and secondary outcomes between the MEPM and MEPM plus MNZ groups are shown in Table 3. No significant differences were observed between the two groups in terms of cure rates, 28-day mortality rates, or length of hospital stay following surgical intervention, either before or after propensity score matching. Among the 8 patients who had treatment failure in the MEPM group, 2 patients died, 4 required additional laparotomy or percutaneous drainage due to postoperative abscess exacerbation, and 2 required a change of antibiotics. In contrast, in the MEPM plus MNZ group, 1 patient died, and 1 patient required a change of antibiotics due to insufficient clinical efficacy. In the results after propensity score matching, among the 3 patients who had treatment failure in the MEPM group, 2 patients died and 1 patient required a percutaneous drainage due to postoperative abscess exacerbation. In contrast, in the MEPM + MNZ group 1 patient died. Table 3Primary and secondary outcomesBefore propensity score matchingAfter propensity score matchingTreated with MEPM(n = 91)Treated with MEPM + MNZ(n = 23)Odds ratio(95%CI)Difference*(95%CI)p valueTreated with MEPM(n = 20)Treated with MEPM + MNZ(n = 20)Odds ratio(95%CI)Difference*(95%CI)p valueCure, n (%)83 (91.2)21 (91.3)0.988(0.095–5.487)0.99917 (85.0)19 (95.0)0.307(0.005–4.243)0.60528-day mortality2 (2.2)1 (4.3)2.008(0.033–40.221)0.4952 (10.0)1 (5.0)0.482(0.008–10.024)0.999Duration of hospitalization aftersurgical intervention (day), median (IQR)18 (11.5–26.5)20 (15.5–32.0)-3(-8.0–1.025)0.21416 (11.8–24)20.5 (16.3–38.3)-4(-12.0–2.0)0.184MEPM: meropenem; MNZ: metronidazole; IQR: interquartile range; *Hodges–Lehmann median difference Table 3Primary and secondary outcomesBefore propensity score matchingAfter propensity score matchingTreated with MEPM(n = 91)Treated with MEPM + MNZ(n = 23)Odds ratio(95%CI)Difference*(95%CI)p valueTreated with MEPM(n = 20)Treated with MEPM + MNZ(n = 20)Odds ratio(95%CI)Difference*(95%CI)p valueCure, n (%)83 (91.2)21 (91.3)0.988(0.095–5.487)0.99917 (85.0)19 (95.0)0.307(0.005–4.243)0.60528-day mortality2 (2.2)1 (4.3)2.008(0.033–40.221)0.4952 (10.0)1 (5.0)0.482(0.008–10.024)0.999Duration of hospitalization aftersurgical intervention (day), median (IQR)18 (11.5–26.5)20 (15.5–32.0)-3(-8.0–1.025)0.21416 (11.8–24)20.5 (16.3–38.3)-4(-12.0–2.0)0.184MEPM: meropenem; MNZ: metronidazole; IQR: interquartile range; *Hodges–Lehmann median difference Primary and secondary outcomes 0.988 (0.095–5.487) 0.307 (0.005–4.243) 2.008 (0.033–40.221) 0.482 (0.008–10.024) Duration of hospitalization after surgical intervention (day), median (IQR) -3 (-8.0–1.025) -4 (-12.0–2.0) MEPM: meropenem; MNZ: metronidazole; IQR: interquartile range; *Hodges–Lehmann median difference Discussion In this study, there was no evidence that adding MNZ to MEPM improved clinical efficacy in empiric antimicrobial therapy with infection-source control in cIAI. The cure rates were 91.2% (83/91) in the MEPM monotherapy group and 91.3% (21/23) in the MEPM plus MNZ group. The clinical trial by Mazuski et al. compared the efficacy of MEPM with that of ceftazidime–avibactam plus MNZ in the treatment of cIAI and observed that the cure rates ranged from 81.6% to 92.5%, a finding that is similar to that reported in the present study [8]. This study enrolled patients with lower gastrointestinal perforation complicated by generalized peritonitis, and the median (IQR) modified APACHE II score was 11 (8–15). These patients are classified as either high risk for community-acquired infections or low or high risk for hospital-acquired or healthcare-associated infections according to the Surgical Infection Society Guidelines for the Management of Intra-abdominal Infections [7]. In this case, carbapenems are recommended because they cover not only aerobic Gram-negative rods, Gram-positive cocci, and obligate anaerobic bacteria, but also Enterobacterales that produce extended-spectrum β-lactamase (ESBL) and AmpC-β-lactamase. Therefore, treatment of the eligible patients with MEPM was considered appropriate. Previous studies have reported that polymicrobial infections documented in 40.0%–84.6% of cases with cIAI [8, 9, 34, 36]. In addition, multiple bacteria were isolated in 62.3% (71/114) of the patients in this study. Anaerobes were isolated in 44.7% (51/114) of the patients, and polymicrobial infections with anaerobes and Gram-negative rods were documented in 39.5% (45/114) of the patients. These results are similar to those in previous studies [4, 8, 37]. In addition, previous studies have reported that the isolation rate of B. fragilis, the most common obligate anaerobic bacteria, ranged from 10.8% to 35% [2, 34, 36, 38], and in this study, it was isolated in 10.5% (12/114) of the patients. However, the isolation rate of B. fragilis was low in the MEPM plus MNZ group, at 8.7% (2/23). The small sample size in the MEPM plus MNZ group and the non-randomized design may have affected this result. Although drug susceptibility testing for anaerobic bacteria was not performed in this study, Japanese surveillance data show that while there are few reports of B. fragilis resistance to MNZ, resistance rates to MEPM are increasing [39]. Therefore, it is possible that the MEPM monotherapy group, which had a high isolation rate of B. fragilis, included more patients with lower antimicrobial susceptibility than the MEPM plus MNZ therapy group. In contrast, the isolation rate of Prevotella species was 26.1% (6/23) in the MEPM plus MNZ group, which was higher than the 17.6% (16/91) observed in the MEPM group. Nationwide surveillance in Japan has reported that MEPM resistance among Prevotella species has remained very low over an extended period, ranging from 0% to 2.3% [40]. Therefore, the difference in isolation rates of Prevotella species may have had limited influence on clinical outcomes. Among the patients in whom Enterococcus faecium resistant to the target antibiotics was isolated, vancomycin was additionally administered to 2 out of 6 patients in the MEPM group and 2 out of 3 patients in the MEPM plus MNZ group. There were no failures in treatment among these 9 cases. Therefore, it is considered that the 5 cases in which E. faecium was isolated and additional antibiotics were deemed unnecessary did not affect the cure rate. Methicillin-resistant Staphylococcus aureus also was isolated from one patient in the MEPM plus MNZ group, but treatment was judged to be unnecessary. All detected ESBL-producing Escherichia coli and Klebsiella pneumoniae were susceptible to MEPM; therefore, these organisms likely did not affect study outcomes. Carbapenem-resistant Enterobacterales were not isolated in this study. There were no significant differences between the two groups in the proportion of patients switched from target antibiotics to oral agents or in the types of oral antibiotics used. However, 20 of the 21 patients who transitioned to oral therapy had achieved clinical cure. This suggests that switching to oral agents may have contributed to maintaining clinical improvement and reducing hospital stay. The Infectious Diseases Society of America guidelines for the management of cIAI recommend 4–7 days of antibiotic treatment for patients with cIAI, provided that adequate source control has been achieved [2]. The median (IQR) duration of administration of the target antibiotics in this study was 7 (6–9) days, which was relatively long. Antibiotics administered in conjunction with surgical intervention for intra-abdominal infections are reported to be effective with durations of 4 days or less [41–44]. However, there are limited reports on critically ill patients because most studies target patients with mild to moderate illness. In a review and meta-analysis of 155 patients with bacteremia, including 40 cases with intra-abdominal infection, Havey et al. reported no difference in the cure rate or survival rate between 5 and 7 days and 7–21 days of antimicrobial administration [45]. In contrast, Montravers et al. compared the short-term administration of antibiotics for 8 days with long-term administration for ≤ 15 days in critically ill patients with intra-abdominal infections, approximately 30% of whom were on mechanical ventilation, with a median Simplified Acute Physiology Score II of 45 [46]. They reported no difference between the two groups in terms of reoperation or mortality; however, additional drainage and bacteremia were more common in the short-term administration group. Therefore, although the source of infection has been reported to be adequately controlled in critically ill patients, the available data are insufficient to support a fixed short-term treatment period. In the present study, 32 patients (28.9%) had a modified APACHE II score of ≥ 15, and 17 patients (14.9%) had bacteremia complications. In these critically ill patients, there is currently no clear evidence to support the long-term administration of antibiotics, and the duration of antibiotic administration had to be adjusted individually based on the clinical course. This study has several limitations. First, as a single center retrospective study, it is difficult to adjust for all confounding factors influencing clinical outcomes, and the possibility of unmeasured confounding remains. Although no significant differences were observed between the two groups in terms of major comorbidities, 2 patients had a high risk for infection due to continuing immunosuppression associated with underlying diseases, such as hematological malignancies and their treatment; however, their nutritional status and activities of daily living were not evaluated. In addition, even after propensity score matching, differences in duration of antibiotic administration remained between the groups, which may have influenced outcomes. Second, relapse after hospital discharge could not be systematically evaluated in this retrospective study. Third, regarding the sample size, on the basis of previous studies, assuming that there are differences in efficacy among patients with cIAI who achieve cure rates of 85% for MEPM monotherapy and 95% for MEPM plus MNZ combination therapy, a sample size of 161 patients per group was calculated to achieve 80% power with a two-sided alpha level of 5%. However, this study was conducted at a single-center, and the number of enrolled patients was lower than this threshold. Therefore, the results of this study do not demonstrate equivalence or non-inferiority, and findings showing no significant difference may reflect a potential Type II error. Fourth, because the antimicrobial susceptibility of obligate anaerobic bacteria, including B. fragilis, was not assessed, we were unable to confirm the presence or absence of resistant strains and thus could not verify the microbiological rationale for adding MNZ to MEPM. In addition, local susceptibility data for obligate anaerobes were not available, preventing evaluation of whether changes in anaerobic susceptibility over the 9-year study period influenced outcomes. In recent years, antimicrobial resistance in the Bacteroides species has become a major issue [20], and the MEPM susceptibility has also been reported to be declining [20, 21]; therefore, surveillance of the susceptibility of anaerobes is crucial. Multi-center data collection and analysis are needed to be addressed in the future. Despite these constraints, this is the first report examining the effects of adding MNZ to MEPM in the empirical antimicrobial treatment for cIAI. Although MNZ has strong antimicrobial activity against anaerobes and MEPM has a broad antimicrobial spectrum covering Gram-positive and -negative pathogens and anaerobes, including the Bacteroides species, no additional clinical benefit was observed in this cohort. Conclusions In this single center retrospective cohort, we found no evidence that adding MNZ to MEPM improved clinical efficacy in empiric antimicrobial therapy with infection source control for cIAI. The excessive use of antibiotics not only increases the risk of side effects but is also associated with the outbreak of antimicrobial-resistant bacteria and increased medical expenses. Combination antibiotic therapy needs to be administered with careful consideration of the antimicrobial spectrum, tissue penetration of each antimicrobial agent, and the prevalence of antimicrobial-resistant bacteria. We hope that the results obtained in this study will serve as the foundation for appropriate use of antibiotics in cIAI.