Efficacy of De-Escalation to Cefmetazole in Patients with Bacteremic Urinary Tract Infections Caused by Extended-Spectrum β-Lactamase-Producing Escherichia coli

Abstract

This study aimed to clarify the optimal value for the unbound cefmetazole concentration to remain above the minimum inhibitory concentration (MIC) (fT ≥ MIC) for efficacy of de-escalation to cefmetazole in patients with bacteremic urinary tract infection by extended-spectrum β-lactamase-producing Escherichia coli. This double-center retrospective observational study was conducted at Tokyo Bay Urayasu Ichikawa Medical Center and Keio University Hospital from January 2012 to October 2022. Efficacy was determined via clinical evaluation (mortality rate, recurrence rate, vital changes) and bacteriological evaluation, and the optimal fT ≥ MIC was calculated via receiver operating characteristic curve analysis. As a result, the number of patients evaluated were 40 (35 and 5 in the treatment success and treatment failure groups, respectively). Univariate analysis showed that fT ≥ MIC, recurrence rate, and MIC for cefmetazole against bacteria were significantly different for the two groups (p < 0.05). Receiver operating characteristic curve analysis showed that the optimal fT ≥ MIC indicating efficacy was 57% (area under the curve: 0.94, 95% confidence interval: 0.86–1.00, p = 0.002). All patients with fT ≥ MIC ≥ 57% had successful treatment, whereas the frequency of treatment failure was high among those with fT ≥ MIC <57%. The optimal fT ≥ MIC for the clinical efficacy of de-escalation to cefmetazole in patients with bacteremic urinary tract infection by extended-spectrum β-lactamase-producing E. coli was fT ≥ MIC ≥ 57%. This finding would be useful for optimal dosing of cefmetazole.

INTRODUCTION

In recent years, there has been a global increase in the number of patients infected with antimicrobial-resistant bacteria, especially extended-spectrum β-lactamase-producing Escherichia coli (ESBL-E), which is associated with increased morbidity and mortality and classified as a Serious Threat.^1,2) ESBL-E strains are typically treated with carbapenems for a variety of diseases.³⁾ However, increasing the use of carbapenem antibiotics can create selective pressure and lead to the emergence of drug-resistant bacteria.^4,5) Because carbapenem antibiotics have a broad spectrum, they can be a last resort against drug resistance^6,7); therefore, their use needs to be reduced.

Among the cephalosporin antibiotics, cefmetazole (CMZ) is the only cephamycin drug available in Japan that is stable against hydrolysis by ESBL-E.⁸⁾ Retrospective studies have revealed that carbapenems are not superior to CMZ with regard to the 30-d mortality associated with ESBL-E bacteremia.^9,10) In a multicenter observational study, the clinical and bacteriological efficacy of CMZ was comparable to that of meropenem for invasive urinary tract infection (UTI) caused by ESBL-E.¹¹⁾ Therefore, CMZ has recently attracted considerable attention as an alternative to carbapenems. However, the Infectious Diseases Society of America 2024 Guidance maintains that cephamycins are not recommended for the treatment of ESBL-E infections until more clinical outcome data are available and optimal dosing is determined.¹²⁾ Furthermore, in the case of UTIs in hospitalized elderly or male patients, considering the possibility of Pseudomonas aeruginosa and using drugs that are active against it is common.^13,14) Because CMZ does not have P. aeruginosa activity, drugs with Pseudomonas aeruginosa activity (such as meropenem and piperacillin-tazobactam) are often used for empiric therapy. These drugs often exhibit activity against ESBL-E. Therefore, in clinical practice, it is commonly used to de-escalate from broad-spectrum antibiotics. CMZ is mainly excreted by the kidneys, partially excreted in bile,¹⁵⁾ and has a high protein-binding rate of approximately 80%.¹⁶⁾ An efficacy index for the design of antimicrobial drug dosing in humans has been established using pharmacokinetic/pharmacodynamic (PK/PD) parameters from a mouse thigh infection model.¹⁷⁾ The PK/PD parameter of the exposure time that the unbound drug concentration remains above the minimum inhibitory concentration (MIC) for a bacterium (fT ≥ MIC) is used in establishing optimal dosing regimens.¹⁷⁾ In a mouse thigh-infection model, treatment of ESBL-E infections with CMZ required 69.6% fT ≥ MIC.¹⁸⁾ For some cephalosporins, such as cefotaxime (CTX) and cefepime, bacteriostatic effects were observed in mice infected with Enterobacteriaceae bacteria when fT ≥ MIC exceeded approximately 35% of the dosing interval.¹⁹⁾ The maximum bactericidal effect of cephalosporins in mice infected with Enterobacteriaceae bacteria occurs when fT ≥ MIC exceeds 60–70% of the dosing interval.²⁰⁾ However, the optimal fT ≥ MIC for de-escalation to CMZ in patients with bacteremic UTI by ESBL-E has not yet been clarified in clinical practice. Therefore, in the current study, we clarify the optimal fT ≥ MIC to establish the clinical effectiveness of de-escalation to CMZ in patients with bacteremic UTI by ESBL-E.

PATIENTS AND METHODS

Patient Information

This retrospective observational study was conducted on patients treated with CMZ against bacteremic UTI caused by ESBL-E at the Tokyo Bay Urayasu Ichikawa Medical Center and Keio University Hospital between January 2012 and October 2022. This study was reviewed and approved by the Keio University School of Medicine, Ethics Committee (Approval No.: 20221159), the Tokyo Bay Urayasu Ichikawa Medical Center Ethics Review Committee (Approval No.: 819), and Keio University Faculty of Pharmacy, Ethics Committee (Approval No.: 230920-3). Informed consent was obtained using an opt-out document on the hospital website.

The inclusion criteria were as follows: 1) ESBL-E was detected in the urinary tract and blood; 2) diagnosis of bacteremic UTI by ESBL-E; 3) the bacterial strain was isolated and stored; and 4) CMZ was used for at least 2 d after de-escalation. The exclusion criteria were as follows: 1) diseases other than UTI were suspected; 2) antibiotics susceptible to ESBL-E were used within 28 d before hospitalization (a washout period was provided to allow sufficient time for the antibiotics to become ineffective)¹¹⁾; 3) patients who have been taking antibiotics for more than 30 d and who have been using them inappropriately (patients who were given the antibiotics for purposes other than the treatment of infections, such as best supportive care); 4) hemodialysis; 5) CMZ was used in empiric therapy; and 6) bacteria other than E. coli were detected.

Data Collection

The following patient information was collected from the electronic medical records: age, sex, body weight (BW), Cockcroft–Gault renal function (CL_CR), albumin (ALB), body temperature, white blood cell (WBC) counts, and C-reactive protein (CRP) levels before the start of treatment; body temperature, WBC counts, and CRP levels during treatment; severity before the start of treatment (Charlson comorbidity score, SOFA, and quick SOFA); ICU admission; cerebrovascular disease; cardiovascular disease; liver disease; renal disease; diabetes; malignant tumor; immunosuppression; urological complications; history of hospitalization within 3 months prior to starting treatment; history of antibiotic use; death within 30 d of starting treatment; recurrence within 30 d of the end of treatment; change from CMZ to another drug; bacteriological evaluation; improvement in clinical findings within 3 d of starting treatment; improvement in clinical findings after 4 d of treatment; antibiotics used in empiric therapy; efficacy of empiric therapy; MIC of CMZ against bacteria; time from culture collection to initiation of effective antibiotics; duration of empiric therapy; duration of injectable antibiotics; number of patients switching to oral medication after CMZ use; and the total duration of antibiotics. CL_CR was calculated using the Cockcroft–Gault formula and BW. The area under the curve (AUC) for blood concentration was calculated by dividing the dose by unbound CMZ clearance (CL). If the dose was changed, the AUC was recalculated. CL was calculated from CL_CR and ALB based on population pharmacokinetic (PPK) analysis using the unbound CMZ concentration. CL = 16.2 × (CL_CR/4.36)^0.781 × (ALB/28)^1.2 L/h.²¹⁾ Distribution volume (V) was calculated using V = 47.1 × (BW/62.8)^0.751 L.²¹⁾ fT ≥ MIC after using CMZ was calculated using Phoenix WinNonlin™ (Version 8.3.5; Certara, NJ, U.S.A.). Each patient’s data were inputted into WinNonlin using Excel. Excel also includes all changes in dosage, administration interval, and administration time from the start to the end of CMZ administration for each patient, as well as changes in test value data (CL_CR, ALB and BW). The PPK parameters (CL and V) for each patient were calculated based on the previously reported formula described in the method above after incorporating each patient’s data, and the blood concentration curve was drawn in WinNonlin. If the dose was changed, the fT ≥ MIC was recalculated. As the primary endpoint, treatment success was defined as survival within 30 d of the start of treatment, no recurrence within 30 d of the end of treatment, and treatment success as judged by the clinician. Clinical efficacy was determined to be present if the primary endpoint was observed. Treatment failure was defined as death within 30 d of the start of treatment, recurrence within 30 d of the end of treatment, and treatment failure as judged by the clinician. The clinician’s criteria for determining treatment success and failure were confirmed by checking the doctor’s records after the start of treatment. Treatment failure was defined as a change from CMZ to another drug when symptoms worsened or when the treatment was not expected to be effective. Regarding recurrence, ESBL-E was detected again in the urinary tract, and the patient was diagnosed with UTI. The following information regarding vital changes, immunosuppression, and urinary complications was obtained from previous reports.¹¹⁾ Secondary endpoints were the presence or absence of normalization of body temperature, white blood cells, and CRP within 6 d of the start of treatment, and the presence or absence of bacterial disappearance (only if confirmed). A body temperature of 37.5°C or higher was considered fever. An elevated WBC count was considered if the count exceeded 12000 cells/μL. An increased CRP level was considered when it exceeded 10 mg/dL. The criteria for immunosuppression were as follows: 1) steroid/glucocorticoid use (equivalent to at least 20 mg of prednisone per day for at least 1 month); 2) immunosuppressant (such as selective T cell costimulatory blocker, methotrexate); 3) anticancer drug; 4) myelodysplastic syndrome; 5) febrile neutropenia (<500 cells/μL); 6) received an organ transplant within the past 3 months; and 7) human immunodeficiency virus infection.¹¹⁾ Urological complications included obstruction or stricture of the urinary tract (e.g., bladder and prostate cancer, ureteral stricture, and prostatic hypertrophy), presence of foreign bodies (e.g., ureteral stents and urinary tract stones), ileal conduit, cystostomy, and urinary retention due to neurological disease.

Bacterial Characterization

Enterobacterales in the samples were identified and subjected to antimicrobial susceptibility testing using the Microscan WalkAway® system (Beckman Coulter, Brea, CA, U.S.A.) with accompanying panels of the Microscan Neg® series (Neg Combo EN 4 J) (Beckman Coulter). To confirm that the identified bacteria were ESBL-producers, ESBL-E screening test (ceftriaxone (CRO), ceftazidime (CAZ), cefotaxime, and aztreonam (AZT) ≥ 2 mg/L) was performed, and the inhibition zone of clavulanic acid was confirmed according to the Clinical and Laboratory Standards Institute (CLSI) criteria.²²⁾ The MIC distribution and susceptibility of the bacteria were determined. The MIC of CMZ against ESBL-E was determined via the broth microdilution method using CLSI standards in the MIC range of 0.125–256 mg/L.^22,23) Microbroth dilutions were as per a previously reported method.²¹⁾

Calculation of Optimal fT ≥ MIC

Using the presence or absence of efficacy and the fT ≥ MIC, the optimal fT ≥ MIC for the clinical efficacy of de-escalation to CMZ was calculated by receiver operating characteristic (ROC) analysis and precision-recall (PR) curve. PR curve is an analysis where the vertical axis is the recall rate (recall) and the horizontal axis is the precision rate (precision).²⁴⁾ In the case of imbalanced data, reliability may not be guaranteed by ROC analysis alone; therefore, the PR curve is also performed.^24,25) The results of ROC analysis using imbalanced data are complemented by calculating the AUC using the PR curve.

Efficacy Evaluation and Consideration of Optimal Administration Method

We confirmed the number of patients with successful and failure treatments, the proportion of patients with successful treatment, and the fT ≥ MIC and compared the differences in efficacy depending on whether the optimal fT ≥ MIC was achieved. Concurrently, we have shown the distribution of fT ≥ MIC depending on whether or not optimal fT ≥ MIC was achieved. Regarding the optimal dosing regimen to achieve the optimal fT ≥ MIC for patients with treatment failure, we used Phoenix WinNonlin™ to perform a simulation analysis of a dosing regimen that increases the number of doses or extends the dosing time.

Statistical Analysis

SPSS version 28.0 (SPSS Inc., Chicago, IL, U.S.A.) and R (Version 4.4.2) were used for data analysis. Patient information collected from electronic medical records was compared between the treatment success and treatment failure groups. Continuous variables were checked for normality, and in the absence of normality, the Mann–Whitney U test was performed. If normality was present, an equal variance was considered. If there was equal variability, a t-test was performed; if there was no equal variability, Welch’s test was performed. The Fisher’s exact test was used to analyze categorical variables. p-Value <0.05 was considered a significant difference.

RESULTS

Patient Data

A total of 40 patients were enrolled: 35 in the treatment success group and 5 in the treatment failure group. The median age of the patients was 79 (interquartile range [IQR]: 73–83), 79 (IQR: 73–83), and 78 (IQR: 74–90) years in the overall population, treatment success, and treatment failure groups, respectively. The proportions of men were 55.0% overall, 57.1% in the treatment success group and 40.0% in the treatment failure group. There was a significant difference in fT ≥ MIC, relapses within 30 d at the end of treatment, and MIC of CMZ against bacteria between the treatment success and failure groups (p < 0.05) (Table 1).

Table 1. Patients Enrolled in This Study

	All patients (n = 40)	Treatment success (n = 35)	Treatment failure (n = 5)	p-Value
Age (years)	79 (73–83)	79 (73–83)	78 (74–90)	0.538
Sex (male)	22 (55.0)	20 (57.1)	2 (40.0)	0.402
Body weight (kg)	58.4 (46.6–66.0)	58.5 (47.6–65.3)	58.3 (43.9–66.0)	0.907
CLCR (mL/min)a)	37.9 (22.3–66.3)	33.1 (22.0–62.9)	56.8 (38.6–74.5)	0.379
Albumin (g/L)	32 (27–38)	32 (27–38)	34 (28–39)	0.809
AUC (mg·h/L)b)	1680 (1066–3571)	1690 (1210–3588)	1077 (461–1964)	0.298
fT ≥ MIC (%)c)	84 (60–100)	87.5 (72–100)	50 (40–56)	<0.001
Fever over 37.5 °Cd)	25 (62.5)	23 (65.7)	2 (40.0)	0.264
Leukocytosis (WBC > 12,000 cells/μL)d,e)	18 (50.0)	16 (50.0)	2 (50.0)	0.699
CRP >10 mg/dLd,e)	14 (41.2)	13 (41.9)	1 (25.0)	0.449
Charlson comorbidity indexd)	8 (6–10)	8 (6–10)	6 (6–9)	0.843
SOFA scored)	3 (1–4)	2 (1–4)	3 (2–3)	1.000
Quick SOFA scored)	0 (0–1)	0 (0–1)	0 (1–1)	0.473
Admission to ICU	0	0	0
Cerebrovascular disease	11 (27.5)	9 (25.7)	2 (40.0)	0.422
Cardiovascular disease	23 (57.5)	21 (60.0)	2 (40.0)	0.354
Liver disease	2 (5.0)	2 (5.7)	0	0.763
Kidney disease	16 (40.0)	15 (42.9)	1 (20.0)	0.323
Diabetes mellitus	14 (35.0)	13 (37.1)	1 (20.0)	0.418
Cancer	15 (37.5)	12 (34.3)	3 (60.0)	0.264
Immunosuppression	4 (10.0)	2 (5.7)	2 (40.0)	0.069
Urological complications	18 (45.0)	15 (42.9)	3 (60.0)	0.402
Hospitalization historyf)	12 (30.0)	9 (25.7)	3 (60.0)	0.149
Antibiotic use historyf)	15 (37.5)	12 (34.3)	3 (60.0)	0.264
Outcome
30-d mortality	0	0	0
Recurrence within 30 d	4 (10.0)	0	4 (80)	<0.001
Switching from CMZ to other antibiotics	1 (2.5)	0	1 (20.0)	0.125
Microbiologically effective	1 (100)	1 (100)	0
Early defervescenceg)	22 (88.0)	21 (91.3)	1 (50.0)	0.230
Early improvement of leukocytosisg)	14 (77.8)	13 (81.3)	1 (50.0)	0.405
Early improvement of CRPg)	22 (88.0)	21 (91.3)	1 (100)	0.786
Late defervescenceh)	3 (100)	2 (100)	1 (100)
Late improvement of leukocytosish)	4 (100)	3 (100)	1 (100)
Late improvement of CRPh)	3 (100)	3 (100)	0
Types of Urinary Tract Infections
Pyelonephritis	29 (72.5)	25 (71.4)	4 (80.0)	0.578
Prostatitis	5 (12.5)	5 (14.3)	0	0.493
Device (catheter or stent)-related
Removed	5 (12.5)	5 (14.3)	0	0.493
Not removed	1 (2.5)	0	1 (20.0)	0.125
Empiric therapy
Ampicillin-sulbactam	1 (2.5)	1 (2.9)	0	0.875
Piperacillin-tazobactam	15 (37.5)	15 (42.9)	0	0.081
Cefazolin	2 (5.0)	1 (2.9)	1 (20.0)	0.237
Ceftriaxone	17 (42.5)	13 (37.1)	4 (80.0)	0.093
Cefepime	1 (2.5)	1 (2.9)	0	0.875
Cefozopran	1 (2.5)	1 (2.9)	0	0.875
Meropenem	21 (52.5)	19 (54.3)	2 (40.0)	0.451
Levofloxacin	2 (5.0)	1 (2.9)	1 (20.0)	0.237
Faropenem	1 (2.5)	1 (2.9)	0	0.875
Empiric treatment effectiveness	30 (75.0)	28 (80.0)	2 (40.0)	0.089
MIC for CMZ against bacteria (mg/L)	2 (1–4)	2 (1–4)	4 (4–4)	0.030
Time to start effective antibiotics from culture collection (h)	22 (3–59)	18 (3–59)	45 (16–46)	0.634
Empiric therapy duration (d)	5 (3–5)	4 (4–5)	5 (3–6)	0.891
Intravenous antibiotic duration (d)	13 (9–15)	13 (9–15)	12 (7–15)	0.858
Oral switch after CMZ treatment	9 (22.5)	8 (22.9)	1 (20)	0.689
Total antibiotic administration duration (d)	15 (11–17)	15 (12–18)	15 (8–16)	0.634

Data are presented as medians (interquartile range) for continuous variables and numbers (%) for categorical variables. Abbreviations: AUC, area under the curve; CL_CR, creatinine clearance; CMZ, cefmetazole; CRP, C-reactive protein; MIC, minimum inhibitory concentration; SOFA, Sequential Organ Failure Assessment; WBC, white blood cell. a) CL_CR was estimated using the Cockcroft–Gault equation with the measured body weight. b) AUC was calculated by dividing the total dose by the unbound CMZ clearance (CL). CL was calculated using albumin, and CL_CR using a model formula constructed from the population pharmacokinetic analysis of unbound cefmetazole concentrations. c) fT ≥ MIC was the exposure time that the unbound drug concentration remains above the MIC for the bacterium. d) Data before the antibiotic start date. e) Available data included 36 patients with WBC counts and 34 patients with CRP levels. f) Within 3 months before starting antibiotics. g) Improvement within 3 d of treatment initiation. h) Improvement 4 d after treatment initiation.

Bacterial Characterization

Based on the measurement results of 40 strains of causative bacteria, the MICs of CRO, CAZ, CTX, and AZT were ≥2 mg/L, and all strains were found to be ESBL-E strains (Table 2). All the strains were susceptible to piperacillin/tazobactam, CMZ, and meropenem (Table 2). Figure 1 shows the results of the MIC measurements of CMZ against the 40 strains of causative bacteria. The MIC₅₀ of CMZ was 2 mg/L and the MIC₉₀ was 4 mg/L.

Table 2. Minimum Inhibitory Concentration (MIC) and Susceptibility of the Isolated Bacteria

	SAM	TZP	CFZ	CMZ	CRO	CTX	CAZ	FEP	AZT	MEM	LVX
MIC50	16/8	≤16/4	≤16	≤8	>2	>2	8	>16	8	≤1	>4
MIC90	≤16/8	≤16/4	≤16	≤8	>2	>2	>8	>16	>8	≤1	>4
Susceptible rate (%)	35%	100%	0%	100%	0%	0%	0%	0%	0%	100%	30%

AZT, aztreonam; CAZ, Ceftazidime; CFZ, cefazolin; CMZ, cefmetazole; CRO, ceftriaxone, CTX, cefotaxime; FEP, cefepime; LVX, levofloxacin; MEM, meropenem; SAM, ampicillin-sulbactam; TZP, piperacillin-tazobactam.

Fig. 1. Minimum Inhibitory Concentration (MIC) Distribution for Cefmetazole against Extended-Spectrum β-Lactamase-Producing Escherichia coli (40 Clinical Isolates)

Gray bar indicates the number of patients that showed treatment failure; black bar indicates the number of patients that showed successful treatment; and the right Y-axis indicates the cumulative rate.

Calculation of Optimal fT ≥ MIC

From the ROC analysis conducted using efficacy status and fT ≥ MIC, the optimal fT ≥ MIC for the clinical efficacy of de-escalation to CMZ was 56.6% (AUC 0.94, 95% confidence interval 0.86–1.00, p = 0.002) (Fig. 2). In addition, the AUC calculated by PR curve was 0.99 (recall 0.97, Precision 0.87) (Fig. 3).

Fig. 2. Confirmation of the Optimal fT ≥ MIC to Ensure the Effectiveness of De-Escalation to Cefmetazole against Extended-Spectrum β-Lactamase-Producing Escherichia coli by the Receiver Operating Characteristic (ROC) Curve

Area under the curve: 0.94 (95% confidence interval, 0.86–1); p = 0.002; sensitivity, 91.4%; specificity, 100%.

Fig. 3. Confirmation of the Optimal fT ≥ MIC to Ensure the Effectiveness of De-Escalation to Cefmetazole against Extended-Spectrum β-Lactamase-Producing Escherichia coli by the Precision-Recall Curve

Area under the curve: 0.99 (recall 0.97, Precision 0.87).

Efficacy Evaluation and Consideration of Optimal Administration Method

The optimal fT ≥ MIC was defined as fT ≥ MIC ≥57%; Table 3 shows the difference in fT ≥ MIC when optimal fT ≥ MIC was achieved and when it was not. Of the 35 successfully treated patients, 91.4% had fT ≥ MIC ≥57%, whereas only 8.6% had fT ≥ MIC <57%. The distribution of fT ≥ MIC in the treatment successful and treatment failure groups and the efficacy depending on whether optimal fT ≥ MIC was achieved are shown in Fig. 4. The success rate of treatment was 100% when fT ≥ MIC ≥57% but dropped significantly to 37.5% when fT ≥ MIC <57%. Table 4 shows the results of an investigation using Phoenix WinNonlin™ regarding the optimal dosing method to achieve fT ≥ MIC ≥57% for patients in whom the treatment failed. A recommended dosing regimen was shown to achieve fT ≥ MIC ≥57% by increasing the number of doses and extending the infusion time (24-h infusion).

Table 3. Relationship between Therapeutic Efficacy and fT ≥ MIC by Achieving Optimal fT ≥ MIC

	Treatment success rate (%)	fT ≥ MIC (%)a)	Treatment failure rate (%)	fT ≥ MIC (%)a)	p-Valueb)
fT ≥ MIC ≥57%	91.4 (32/35)	94 (72–100)	0 (0/0)	n.d	<0.001
fT ≥ MIC <57%	8.6 (3/35)	48 (40–48)	100 (5/5)	50 (40–56)

The fT ≥ MIC are presented as medians (interquartile range). MIC, minimum inhibitory concentration. a) fT ≥ MIC was the exposure time that the unbound drug concentration remains above the MIC for the bacterium. b) Significant difference in treatment success with or without achieving optimal fT ≥ MIC.

Fig. 4. Distribution of fT ≥ MIC Is Shown for Each Optimal fT ≥ MIC

Gray squares indicate treatment failure, black circles indicate treatment success, and dashed lines indicate fT ≥ MIC = 57%.

Table 4. Recommended Dosing Regimens for Successful Treatment of Patients with Bacteremic Urinary Tract Infections Who Showed Failed Treatment

No.	CLCR (mL/min)a)	ALB (g/L)	BW (kg)	CL (L/h)b)	MIC (mg/L)	Regimens for treatment failure (1-h infusion)	fT ≥ MIC (%)c)	Recommended dosing regimens (1-h infusion)	fT ≥ MIC (%)c)	Recommended dosing regimens (24-h infusion)	fT ≥ MIC (%)c)
1	111	27	58.3	21.8	4	2 g every 12 h	40	2 g every 8 h	60	4 g every 24 h	96
2	59.4	34	67.7	17.3	4	1 g every 12 h	40	1 g every 8 h	60	2 g every 24 h	100
3	46.0	28	43.9	11.3	8	1 g every 8 h	50	1 g every 6 h	64	3 g every 24 h	100
4	42.8	27	43.1	10.2	4	1 g every 12 h	56	1 g every 8 h	84	2 g every 24 h	100
5	83.4	34	66.0	22.4	2	2 g every 12 h	56	2 g every 8 h	84	4 g every 24 h	100

ALB, albumin; BW, body weight; CMZ, cefmetazole; CL, clearance; CL_CR, creatinine clearance; MIC, minimum inhibitory concentration.

a) CL_CR was estimated using the Cockcroft–Gault equation with the measured body weight. b) CL was calculated from albumin and CL_CR using a model formula constructed from the population pharmacokinetic analysis of unbound cefmetazole concentrations. c) fT ≥ MIC was the exposure time that the unbound drug concentration remains above the MIC for the bacterium.

DISCUSSION

In this study, we developed an administration design based on PK/PD in patients with bacteremic UTI caused by ESBL-E and evaluated its effectiveness. As a result, the optimal fT ≥ MIC for the clinical efficacy of de-escalation to CMZ for patients with ESBL-E bacteremic UTI was newly clarified as fT ≥ MIC ≥57%. There has been no report that has calculated the optimal fT ≥ MIC for the clinical efficacy of cephems, including CMZ, in humans; hence, this report is considered to be of high clinical significance.

There were no differences in immunosuppression, severity, or urological complications between the treatment success and treatment failure groups (Table 1). By contrast, there were significant differences in fT ≥ MIC and MIC for CMZ against bacteria. Lower doses or less frequent dosing in patients with high CL of CMZ or high MIC of CMZ against bacteria may lead to low fT ≥ MIC and treatment failure (Table 1). In this study, most patients had their urinary catheters and ureteral stents replaced or removed, but one patient who showed failed treatment still had a stent inserted. We believe that fT ≥ MIC should be more strictly controlled, especially in patients with inadequate source control.

Because the number of patients in this study was small and there were concerns about the validity of the optimal fT ≥ MIC determined using ROC analysis, we also used PR curve to complement the results of the ROC analysis.^24,25) The AUC of the ROC curve analysis was 0.94 (Fig. 2), and the AUC of the PR curve was 0.99 (Fig. 3), both of which are very high AUCs. The PR curve is an analysis that focuses on positivity, and in this study, we wanted to calculate the optimal fT ≥ MIC considering the overall balance; hence, we used the results of ROC analysis.

The PK/PD parameter for efficacy of β-lactams has been shown to be fT ≥ MIC, and the same is true for cephalosporins. The optimal fT ≥ MIC of cefotaxime against Klebsiella pneumoniae in the lungs of neutropenic mice is 30–40% bacteriostatic effect and 60–70% maximum bactericidal effect.²⁰⁾ The optimal fT ≥ MIC in this study (fT ≥ MIC ≥ 57%) was equivalent to the previously reported static effect fT ≥ MIC = 57.6% of CMZ against ESBL-E using a mouse thigh infection model.¹⁸⁾ In a previous report, a 1 log₁₀ kill reduction of fT ≥ MIC = 69.6% was achieved.²¹⁾ In our PK/PD analysis for ESBL-E, the optimal fT ≥ MIC was set at fT ≥ MIC ≥ 70% based on the 69.6% from the above report.²¹⁾ The PK/PD parameters obtained from human studies were similar to those obtained from animal studies using a mouse thigh infection model.¹⁷⁾ We believe that the reason why the optimal fT ≥ MIC (fT ≥ MIC ≥ 57%) was lower than fT ≥ MIC = 69.6% in the previous report is related to the pathological condition of UTI.¹⁸⁾ CMZ is a renally excreted drug, and its concentration in the urine is higher than that in the blood.¹⁵⁾ Therefore, it is conceivable that the optimal fT ≥ MIC was low because the CMZ concentration in urine was higher than that in the blood, and a sufficient bactericidal effect was achieved.

Confirming the relationship between the efficacy and fT ≥ MIC, treatment was successful in all patients with UTI when fT ≥ MIC was ≥57% (Fig. 4). If bacteremia is caused by a UTI, we believe that it can be treated with fT ≥ MIC ≥ 57%, which is a PK/PD parameter of bacteriostatic efficacy. However, as it has not been investigated in other diseases, we believe that by using the nomogram fT ≥ MIC ≥ 70%, it may be possible to determine the optimal dosing regimen according to the PK/PD parameter.²¹⁾ Therefore, CMZ can be used more appropriately by correctly selecting the optimal fT ≥ MIC for each disease.

In this simulation analysis, we did not use the blood concentration of each patient, so it may be less accurate than a simulation analysis based on concentration. However, CL and V were estimated to use each patient’s CL_CR, albumin and BW, and concentration simulation was performed using their individual PPK parameters. The PPK parameters used in this analysis were calculated via PPK analysis, and because the patient background in previous research was similar to the patient background of this study, the effectiveness judgment by simulation using PPK parameters for individual patients is considered to be appropriate.²¹⁾

Based on the results of this simulation analysis, continuous infusion is considered to be an effective strategy for increasing fT ≥ MIC (Table 4). Designing an administration where fT ≥ MIC = 100% is possible, therefore, it may become a treatment option even for patients with severe infections.

In this study, there were no patients in either group who discontinued treatment owing to side effects caused by CMZ. There is currently no evidence about the relationship between cefmetazole’s fT ≥ MIC and side effects, and the related details are unknown.

One limitation of this study is that the simulation analysis was performed using CL and V calculated from population pharmacokinetic analysis, rather than measuring the blood concentration of patients. Therefore, detailed accuracy and bias could not be calculated. Another limitation of the current study is the small number of patients. To further validate our results, we believe that it is necessary to increase the number of cases and conduct further analyses.

CONCLUSION

The optimal fT ≥ MIC for the clinical efficacy of de-escalation to CMZ in patients with bacteremic UTI caused by ESBL-E was shown to be fT ≥ MIC ≥57%. The success rate of treatment was 100% when fT ≥ MIC ≥57%. These results may lead to optimal dosing regimens when using CMZ for patients with bacteremic UTI caused by ESBL-E. Large-scale prospective clinical studies are needed to make these results more robust.

Acknowledgments

The authors thank the staff at the Division of Clinical Laboratory, Tokyo Bay Urayasu Ichikawa Medical Center, Japan. They also thank the Department of Hospital Information Systems, Keio University Hospital, as well as the staff in the Department of Clinical Laboratory Technology Office. This research received no specific grants from any funding agency in the public, commercial, or not-for-profit sectors.

Author Contributions

Takaya Namiki, Yuta Yokoyama, Yoshifumi Uwamino, and Rentaro Oda contributed to the study conception and design. Data collections were performed by Takaya Namiki. Analysis and interpretation of data were performed by Takaya Namiki, Yuta Yokoyama, Hitoshi Kawazoe, Hisakazu Ohtani, Kazuaki Matsumoto, Hideki Hashi, Sayo Suzuki, and Tomonori Nakamura. The first draft of the manuscript was written by Takaya Namiki and was revised critically by Yuta Yokoyama, Motonori Kimura, Shogo Fukuda, Shoji Seyama, Osamu Iketani, Yoshifumi Uwamino, Aya Jibiki, Hitoshi Kawazoe, Hisakazu Ohtani, Naoki Hasegawa, Kazuaki Matsumoto, Rentaro Oda, Hideki Hashi, Sayo Suzuki, and Tomonori Nakamura. All authors read and approved the final manuscript.

Conflict of Interest

HK received research funding from Eli Lilly outside the submitted work. Dr. KM received Grants from Meiji Seika Pharma, Sumitomo Pharma, and Shionogi outside of the submitted work. TN received research funding from Chugai, Daiichi Sankyo, Kyowa Kirin, Otsuka Pharmaceutical, Sanofi, and Shionogi outside the submitted work. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

REFERENCES

• 1)  Septimus EJ. Antimicrobial resistance: an antimicrobial/diagnostic stewardship and infection prevention approach. Med. Clin. North Am., 102, 819–829 (2018).
• 2)  Tamma PD, Aitken SL, Bonomo RA, Mathers AJ, van Duin D, Clancy CJ. Spectrum β-lactamase Producing Enterobacterales (ESBL-E), Carbapenem-Resistant Enterobacterales (CRE), and Pseudomonas aeruginosa with Difficult-to-Treat Resistance (DTR-P. aeruginosa) Version 1.1 2. Running Title: AMR Treatment Guidance. IDSA, Arlington Virginia United States (2022).
• 3)  CDC, Atlanta, Georgia, “Antibiotic resistance threats in the United States.”: <https://www.cdc.gov/antimicrobial-resistance/media/pdfs/2019-ar-threats-report-508.pdf>, accessed 20 December, 2023.
• 4)  Hussein K, Sprecher H, Mashiach T, Oren I, Kassis I, Finkelstein R. Carbapenem resistance among Klebsiella pneumoniae isolates risk factors, molecular characteristics, and susceptibility patterns. Infect. Control Hosp. Epidemiol., 30, 666–671 (2009).
• 5)  Paiboonvong T, Tedtaisong P, Montakantikul P, Gorsanan S, Tantisiriwat W. Correlation between carbapenem consumption and carbapenems susceptibility profiles of Acinetobacter baumannii and Pseudomonas aeruginosa in an academic medical center in Thailand. Antibiotics (Basel), 11, 143 (2022).
• 6)  Exner M, Bhattacharya S, Christiansen B, Gebel J, Goroncy-Bermes P, Hartemann P, Heeg P, Ilschner C, Kramer A, Larson E, Merkens W, Mielke M, Oltmanns P, Ross B, Rotter M, Schmithausen RM, Sonntag HG, Trautmann M. Antibiotic resistance: what is so special about multidrug-resistant Gram-negative bacteria? GMS Hyg. Infect. Control., 12, Doc05 (2017).
• 7)  Armstrong T, Fenn SJ, Hardie KR. JMM Profile: carbapenems: a broad-spectrum antibiotic. J. Med. Microbiol., 70, 001462 (2021).
• 8)  Pitout JDD, Laupland KB. Extended-spectrum β-lactamase-producing Enterobacteriaceae: an emerging public-health concern. Lancet Infect. Dis., 8, 159–166 (2008).
• 9)  Matsumura Y, Yamamoto M, Nagao M, Komori T, Fujita N, Hayashi A, Shimizu T, Watanabe H, Doi S, Tanaka M, Takakura S, Ichiyama. Multicenter retrospective study of cefmetazole and flomoxef for treatment of extended-spectrum-β-lactamase-producing Escherichia coli bacteremia. Antimicrob. Agents Chemother, 59, 5107–5113 (2015).
• 10)  Fukuchi T, Iwata K, Kobayashi S, Nakamura T, Ohji G. Cefmetazole for bacteremia caused by ESBL-producing enterobacteriaceae comparing with carbapenems. BMC Infect. Dis., 16, 427 (2016).
• 11)  Hayakawa K, Matsumura Y, Uemura K, Tsuzuki S, Sakurai A, Tanizaki R, Shinohara K, Hashimoto T, Hase R, Matono T, Kato H, Mawatari M, Hara H, Hamada Y, Saito S, Ohmagari N, Doi Y. Effectiveness of cefmetazole versus meropenem for invasive urinary tract infections caused by extended-spectrum β-lactamase-producing Escherichia coli. Antimicrob. Agents Chemother., 67, e0051023 (2023).
• 12)  Tamma PD, Heil EL, Justo JA, Mathers AJ, Satlin MJ, Bonomo RA. Infectious Diseases Society of America 2024 guidance on the treatment of antimicrobial-resistant gram-negative infections. Clin. Infect. Dis., ciae403 (2024).
• 13)  Lamas Ferreiro JL, Álvarez Otero J, González González L, Novoa Lamazares L, Arca Blanco A, Bermúdez Sanjurjo JR, Rodríguez Conde I, Fernández Soneira M, de la Fuente Aguado J. Pseudomonas aeruginosa urinary tract infections in hospitalized patients: mortality and prognostic factors. PLOS ONE, 12, e0178178 (2017).
• 14)  Tabibian JH, Gornbein J, Heidari A, Dien SL, Lau VH, Chahal P, Churchill BM, Haake DA. Uropathogens and host characteristics. J. Clin. Microbiol., 46, 3980–3986 (2008).
• 15)  Ohkawa M, Orito M, Sugata T, Shimamura M, Sawaki M, Nakashita E, Kuroda K, Sasahara K. Pharmacokinetics of cefmetazole in normal subjects and in patients with impaired renal function. Antimicrob. Agents Chemother., 18, 386–389 (1980).
• 16)  Tan JS, Salstrom S-J, Signs SA, Hoffman HE, File TM. Pharmacokinetics of intravenous cefmetazole with emphasis on comparison between predicted theoretical levels in tissue and actual skin window fluid levels. Antimicrob. Agents Chemother., 33, 924–927 (1989).
• 17)  Craig WA. Pharmacokinetic/pharmacodynamic parameters: rationale for antibacterial dosing of mice and men. Clin. Infect. Dis., 26, 1–10 (1998).
• 18)  Takemura W, Tashiro S, Hayashi M, Igarashi Y, Liu X, Mizukami Y, Kojima N, Morita T, Enoki Y, Taguchi K, Yokoyama Y, Nakamura T, Matsumoto K. Cefmetazole as an Alternative to Carbapenems Against Extended-Spectrum Beta-Lactamase-Producing Escherichia coli Infections Based on In Vitro and In Vivo Pharmacokinetics/Pharmacodynamics Experiments. Pharm. Res., 38, 1839–1846 (2021).
• 19)  Ambrose PG, Owens RC Jr, Garvey MJ, Jones RN. Pharmacodynamic considerations in the treatment of moderate to severe pseudomonal infections with cefepime. J. Antimicrob. Chemother., 49, 445–453 (2002).
• 20)  Craig WA. Interrelationship between pharmacokinetics and pharmacodynamics in determining dosage regimens for broad-spectrum cephalosporins. Diagn. Microbiol. Infect. Dis., 22, 89–96 (1995).
• 21)  Namiki T, Yokoyama Y, Hashi H, Oda R, Jibiki A, Kawazoe H, Matsumoto K, Suzuki S, Nakamura T. Pharmacokinetics/pharmacodynamics analysis and establishment of optimal dosing regimens using unbound cefmetazole concentration for patients infected with Extended-Spectrum β-lactamase producing Enterobacterales (ESBL-E). Pharmacotherapy, 44, 149–162 (2024).
• 22)  Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing: 32nd edition. CLSI supplement M100. Clinical and Laboratory Standards Institute, Berwyn, PA, 2022.
• 23)  Clinical and Laboratory Standards Institute. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically; approved tandard 10th edition. CLSI document M07–A10. Clinical and Laboratory Standards Institute, Berwyn, PA, 2015.
• 24)  Davis J, Goadrich M. “The Relationship Between Precision-Recall and ROC Curves.”: <https://ftp.cs.wisc.edu/machine-learning/shavlik-group/davis.icml06.pdf>, accessed 25 November, 2024.
• 25)  Movahedi F, Padman R, Antaki JF. “Limitations of ROC on Imbalanced Data: Evaluation of LVAD Mortality Risk Scores. (2020).”: <https://arxiv.org/abs/2010.16253>, accessed 13 January, 2025.