Ertapenem-Containing Double-Carbapenem Therapy for Treatment of Infections Caused by Carbapenem-Resistant Klebsiella pneumoniae

Abstract

We describe outcomes of patients with infections with carbapenem-resistant Klebsiella pneumoniae (CRKP) who received ertapenem-containing double-carbapenem therapy (ECDCT). Clinical success was observed in 7/18 (39%) patients overall: bloodstream infections, 3/7 (43%); pneumonia, 1/5 (20%); intraabdominal infections, 0/2 (0%); urinary tract infections, 2/3 (67%); and a skin and skin structure infection, 1/1 (100%). Microbiologic success was observed in 11/14 (79%) evaluable patients; 5/18 (28%) patients died. ECDCT may be effective for CRKP infections with limited treatment options.

TEXT

Since the first identified Klebsiella pneumoniae carbapenemase (KPC) enzymes discovered in K. pneumoniae in 1996, there has been an increasing incidence of this multidrug-resistant pathogen globally (1–3). The drugs most often active against carbapenem-resistant K. pneumoniae (CRKP) are tigecycline, aminoglycosides, and polymyxins, all of which have significant downsides (1, 4–6). Limited recent data have evaluated ertapenem-containing double-carbapenem therapy (ECDCT) as an alternative regimen to treat KPC-producing CRKP infections. Of the available carbapenems, ertapenem is the most readily hydrolyzed by KPC enzymes (7). Synergistic bactericidal activity has been shown in vitro when ertapenem is given in combination with meropenem or doripenem. The proposed mechanism of synergy with ECDCT is that by administering ertapenem with a second carbapenem, beta-lactamase enzymes are consumed by the interaction with ertapenem, allowing the more active carbapenem to affect the organism (8). In vitro studies and case reports suggest a possible benefit of ECDCT for the treatment of KPC-producing Klebsiella infections, though the published data are limited (8–11).

This study was performed to describe the outcomes of patients with infections due to CRKP who received ECDCT in our institution. It was a noninterventional, retrospective chart review of patients with cultures for CRKP who were treated with ECDCT from October 2013 to November 2014. ECDCT was recommended to be given as 1 g of ertapenem administered daily 1 h prior to the first dose of meropenem or doripenem. Recommended doses were 2 g of meropenem every 8 h and 500 mg of doripenem every 8 h or equivalent renally adjusted doses. Included patients were ≥18 years of age with a positive culture for CRKP who received ECDCT for ≥48 h for the treatment of infections meeting the definitions of the Centers for Disease Control and Prevention (12). Our automated susceptibility system, the BD Phoenix Automated System (BD Diagnostics, Sparks, MD), determined antibiotic susceptibilities based on the 2009 Clinical and Laboratory Standards Institute (CLSI) M100-S19 breakpoints (13). On manual review, we considered anything with a MIC of >1 μg/ml as resistant on the basis of the 2010 revised CLSI M100-S10 breakpoints for carbapenems (14). If an imipenem MIC of ≥4 μg/ml was detected by the automated system, it was verified by Etest (bioMérieux USA, Durham, NC). Patients with susceptibility to any carbapenems were excluded from analysis. If a patient had more than one episode of CRKP-associated infection, only the first episode treated with ECDCT was evaluated. The Temple University Institutional Review Board approved this study.

The patient data collected included information on comorbidities, demographics, immune status, previous hospital admissions and antibiotic use, and length of stay. The need for mechanical ventilation, intensive care unit (ICU) admission, and APACHE II score were assessed on the day the index culture was collected (15). The Charlson comorbidity index (CI) determined the extent of comorbid illnesses (16). Patients were considered immunocompromised if they had a bone marrow transplant, a solid organ transplant, active cancer, a splenectomy, a gamma globulin deficiency, or an HIV infection with a CD4 cell count of <250 cells/mm³ or received at least 20 mg of prednisone/day for at least 30 days. Clinical success at the end of therapy was the primary outcome of interest and was defined as the resolution of signs and symptoms of infection, including fever, leukocytosis, and local symptoms without relapse of infection within 30 days. Secondary outcomes included microbiologic success, death within 30 days, time to a negative culture, and adverse effects. Microbiological success was defined as a culture negative for CRKP at the end of therapy. Descriptive statistics were used to analyze the data.

Nineteen patients treated with ECDCT were identified, and 18 of them met the inclusion criteria. One patient treated for a urinary tract infection (UTI) did not meet the criteria for that diagnosis. The characteristics of the patients included are described in Table 1. There was a high degree of acute illness, with a median APACHE II score of 21 (interquartile range [IQR], 13.25 to 25.75). Notable characteristics included a high percentage of patients with an ICU admission, recent hospitalizations and antibiotic use, pulmonary disease, cardiovascular disease, and a requirement for mechanical ventilation. Two patients had isolates with an imipenem MIC of 4 mg/liter initially (Table 2, cases 1 and 16), though for both patients, before the initiation of ECDCT, later isolates tested had an MIC of >4 mg/liter. Doripenem was administered to only one patient (Table 2, case 16).

TABLE 1.

Patient characteristics^a

Characteristic	BSI (n = 7)	PNA (n = 5)	UTI (n = 3)	SSSI (n = 1)	IAI (n = 2)	All infections (n = 18)
Age (yr)	56 (47.5–63.5)	62 (51–79)	65 (64–68)	50	67	62.5 (51–67)
No. (%) of males	3 (43)	2 (40)	2 (67)	1	2 (100)	10 (56)
Charlson CI	3 (2.5–3.5)	3 (3–4)	3 (1.5–4)	6	5	3 (2.25–4.75)
No. (%) with:
Cardiovascular disease	4 (57)	2 (40)	2 (67)	1	1 (50)	10 (56)
Pulmonary disease	3 (43)	5 (100)	2 (67)	0	0 (0)	10 (56)
Diabetes	2 (29)	2 (40)	2 (67)	1	1 (50)	8 (44)
Immunocompromised state	4 (57)	1 (20)	0 (0)	0	2 (100)	7 (39)
Hospitalization within 90 days	3 (43)	4 (80)	2 (67)	1	1 (50)	11 (61)
Antibiotics within 90 days	7 (100)	4 (80)	2 (67)	1	2 (100)	16 (89)
Mechanical ventilation	3 (43)	4 (80)	1 (33)	0	2 (100)	10 (56)
Dialysis	2 (29)	1 (20)	0 (0)	0	1 (50)	4 (22)
Length of stay (days)	35 (27–69.5)	21 (13–26)	33 (32.5–38)	14	34	31.5 (22.25–37.25)
APACHE II score	22 (21–24)	21 (19–28)	9 (7–19.5)	14	18.5	21 (13.25–25.75)
No. (%) in ICU	7 (100)	4 (80)	1 (33)	1	2 (100)	15 (83)

^aMedians (IQRs) are shown for continuous variables.

TABLE 2.

Summary of cases and outcomes

Clinical category and case no.	Type of infection (source)	Time (days)			Time (h) to ECDCTb	ECDCT duration (days)	Concurrent Gram-negative coverageg	Comment(s)
		Of prior Gram-negative therapy (drug)a	To negative culture	To negative culture on ECDCT
Success
1	BSI (IAI)	4 (TZP), 5 (IPM)	27	19	186	72	DOX	Liver abscess aspiration, long course of therapy for source control until eventual healing of hepatic abscess and CTc negative for abscess/fluid collection
2	BSI (CVC)d	10 (TZP)	1	0	80	8	None	CVC removed
3	BSI (urine)	5 (DOX), 2 (TZP)	4	0	126	16	DOX
4	PNA	3 (SXT)	Unknown	Unknown	95	9	None
5	SSSI	3 (FEP)	Unknown	Unknown	56	36	None	Debridement, VAC,e and flap placement; multiple admissions for debridement; VAC eventual removed, only local wound care required
6	UTI	7 (DOX)	9	1	188	15	None	Second episode of CRKP UTI, initial episode treated with AMK
7	UTI	2 (CIP)	5	2	66	14	None
Failure
8	BSI (IAI)	2 (TZP)	8	6	57	16	None	Common bile duct injury with repair, developed perihepatic abscess, subsequent drain placement, relapse of BSI within 30 days
9	BSI (PNA)	7 (FEP)	3	3	7	22	GEN	Polymicrobial BSI with Candida albicans (micafungin, success), relapse of BSI within 30 days
10	BSI (SSSI)	2 (TZP)	4	2	57	4	AMK	Developed seizure and completed 43-day course of PMB + TGC
11	BSI (unclear)	2 (MEM), 4 (FEP), 5 (CIP), 2 (CST)	1	0	106	2	CIP	Polymicrobial BSI with Acinetobacter baumannii, relapse of BSI within 30 days, developed seizure, completed 15-day course of ATM + CST + TGC, death within 30 days
12	PNA	16 (FEP), 1 (TZP)	Unknown	Unknown	74.5	10	CIP	Polymicrobial PNA with P. aeruginosa, initial treatment with FEP for P. aeruginosa, ECDCT initiated after clinical worsening, death within 30 days
13	PNA	2 (TZP)	Cultures did not clear	Cultures did not clear	49	2	None	Switched to PMB + MEM, treatment for 2 additional days before death within 30 days
14	PNA	4 (TOB), 2 (PMB), 3 (CIP), 2 (MEM)	Cultures did not clear	Cultures did not clear	111	10	TGC, PMB	Polymicrobial PNA with P. aeruginosa, death within 30 days
15	PNA	2 (IPM), 4 (TZP)	Cultures did not clear	Cultures did not clear	72.5	21	CIP, TGC	Developed BSI with Pseudomonas putida while on ECDCT, death within 30 days
16	IAI	10 (IPM)	Unknown	Unknown	261	20	None	Polymicrobial infection with C. albicans, rectal perforation, underwent abdominal washout, found to have pelvic abscess, performed sigmoid colostomy and pelvic drain placed, readmission within 30 days of therapy completion with abdominal pain and drainage, abdominal wound culture positive for CRKP
17	IAI	5 (FEP)	8	3	131	7	None	Polymicrobial infection with Proteus mirabilis and Enterococcus faecium, found to have anastomotic leak, underwent exploratory laparotomy, small bowel resection, rectal stump closure, and end ileostomy; CRKP but not E. faecium cleared
18	UTI	2 (FEP), 2 (CIP)	13	11	60	20	GEN, DOX	Also treated for MRSAf endocarditis (daptomycin, success), aortic and mitral valve replacement, relapse of infection within 30 days

^aPrior Gram-negative therapy was defined as the most recent course of antibiotics. TZP, piperacillin-tazobactam; IPM, imipenem; DOX, doxycycline; SXT, trimethoprim-sulfamethoxazole; FEP, cefepime; CIP, ciprofloxacin; CST, colistin; ATM, aztreonam; TGC, tigecycline; PMB, polymyxin B; MEM, meropenem; TOB, tobramycin.

^bTime to ECDCT defined as the time from the index culture to administration of ECDCT.

^cCT, computed tomography.

^dCVC, central venous catheter.

^eVAC, vacuum-assisted closure.

^fMRSA, methicillin-resistant Staphylococcus aureus.

^gGEN, gentamicin; AMK, amikacin.

Patient cases are described in Table 2. Clinical success was observed in 7/18 (39%) patients. Clinical success was observed in 1/1 (100%) patient treated for a skin and skin structure infection (SSSI), 2/3 (67%) patients treated for UTIs, 3/7 (43%) patients treated for bloodstream infections (BSIs), 1/5 (20%) patients treated for pneumonia (PNA), and 0/2 (0%) patients treated for intraabdominal infections (IAIs). Microbiologic success was observed in 11/14 (79%) evaluable patients. The microbiological outcomes of four patients could not be assessed. Two of the four patients were treated for PNA; one achieved clinical success, and the other was initially treated for PNA due to Pseudomonas aeruginosa. Upon clinical worsening, ECDCT was added to treat the CRKP that was subsequently isolated; the patient was treated for 10 days before dying. One of the four patients was treated for an SSSI and improved clinically after being discharged and readmitted for multiple procedures to debride the wound. The final patient whose microbiologic success could not be assessed was treated for an IAI and improved clinically after ECDCT but was readmitted within 30 days of completing ECDCT for drainage of an abdominal wound. All of the patients treated for BSIs or UTIs achieved microbiologic success (7/7 and 3/3, respectively). Three patients treated for BSIs had negative blood cultures prior to the initiation of ECDCT. One of the patients treated for IAIs achieved microbiologic success but not clinical success. Microbiological success did not occur in any of the three evaluable patients treated for PNA (0%).

Death within 30 days occurred in 5/18 (28%) patients with BSIs 1/7 (14%) or PNA 4/5 (80%). Two patients (11%) developed seizures during ECDCT, one of which was a new-onset seizure.

A previous study by Bulik and Nicolau evaluated ECDCT for KPC-producing CRKP in vitro and in an in vivo murine thigh infection model. When observing bacterial densities in vitro, the use of doripenem and ertapenem dual therapy maintained a 3-log reduction in bacterial activity for 16 h, as opposed to 6 h with either doripenem or ertapenem monotherapy at a doripenem MIC of 4 mg/liter and an ertapenem MIC of 64 mg/liter. The in vivo murine thigh infection model showed statistically significant bacterial burden reductions with the combination compared to doripenem alone (8).

Karaiskos et al. described the outcomes of patients with infections due to KPC-producing CRKP, i.e., five patients with BSIs and nine patients with UTIs. All of the patients received ECDCT and achieved clinical and microbiologic success at 14 days. At the 1-month follow-up, all of the patients had apparent clinical and microbiologic success, except that 4 of 12 patients colonized with CRKP in their stool had recurrent UTIs by day 28 (11). A case series of three different patients treated with ECDCT was also recently published by the same authors. The case series described clinical improvement in three patients treated for KPC-producing infections, i.e., two BSIs and a UTI (9). An additional case report of BSI due to CRKP was published that showed success with ECDCT after therapies with colistin, meropenem and rifampin, and then colistin plus fosfomycin had all failed. That paper also included an in vitro analysis of dual-carbapenem therapy that showed after 4 h of treatment, dual-carbapenem therapy demonstrated a 3-log kill (10).

This study is limited by several factors. The analysis was retrospective in nature, and we were limited to an analysis of existing data. It was a small, single-center study without a comparator group, which limits its generalizability. The primary outcome of clinical success is subject to different interpretations. Since our laboratory does not perform genetic testing of CRKP isolates, we chose to treat patients with ECDCT on the basis of phenotypic carbapenem resistance instead of proven KPC enzyme production, which may not be as reliable. Our laboratory did not report the actual MICs for the organism when they were >4 μg/ml. The lack of knowledge of the precise MIC may limit whether or not ertapenem added any benefit to meropenem or doripenem, since it is possible to achieve useful concentrations above minimally elevated carbapenem MICs with aggressive dosing (17, 18). Though some patients had concurrent and deep-seated infections and we believe the concurrent infections were adequately treated, it is difficult to evaluate which resulted in clinical failure. This, however, is characteristic of the types of patients and severity of illness that are associated with multidrug-resistant infections.

Our data suggest that ECDCT may be effective for the treatment of infections due to CRKP with limited treatment options. Clinical success was observed in 7/18 (39%) patients, and death occurred in 5/18 (28%) patients. We believe that ECDCT is worthy of future study for the treatment of infections due to CRKP.