Treatment of Infections by OXA-48-Producing Enterobacteriaceae

Carbapenemase-producing Enterobacteriaceae (CPE) contribute significantly to the global public health threat of antimicrobial resistance. OXA-48 and its variants are unique carbapenemases with low-level hydrolytic activity toward carbapenems but no intrinsic activity against expanded-spectrum cephalosporins.

ABSTRACT

Carbapenemase-producing Enterobacteriaceae (CPE) contribute significantly to the global public health threat of antimicrobial resistance. OXA-48 and its variants are unique carbapenemases with low-level hydrolytic activity toward carbapenems but no intrinsic activity against expanded-spectrum cephalosporins. bla_OXA-48 is typically located on a plasmid but may also be integrated chromosomally, and this gene has progressively disseminated throughout Europe and the Middle East. Despite the inability of OXA-48-like carbapenemases to hydrolyze expanded-spectrum cephalosporins, pooled isolates demonstrate high variable resistance to ceftazidime and cefepime, likely representing high rates of extended-spectrum beta-lactamase (ESBL) coproduction. In vitro data from pooled studies suggest that avibactam is the most potent beta-lactamase inhibitor when combined with ceftazidime, cefepime, aztreonam, meropenem, or imipenem. Resistance to novel avibactam combinations such as imipenem-avibactam or aztreonam-avibactam has not yet been reported in OXA-48 producers, although only a few clinical isolates have been tested. Although combination therapy is thought to improve the chances of clinical cure and survival in CPE infection, successful outcomes were seen in ∼70% of patients with infections caused by OXA-48-producing Enterobacteriaceae treated with ceftazidime-avibactam monotherapy. A carbapenem in combination with either amikacin or colistin has achieved treatment success in a few case reports. Uncertainty remains regarding the best treatment options and strategies for managing these infections. Newly available antibiotics such as ceftazidime-avibactam show promise; however, recent reports of resistance are concerning. Newer choices of antimicrobial agents will likely be required to combat this problem.

INTRODUCTION

Infection due to carbapenemase-producing Enterobacteriaceae (CPE) carries a significantly high attributable mortality (1). CPE infections are steadily increasing worldwide and spreading to several countries (2). Moreover, they are increasing as a proportion of the total Enterobacteriaceae infections, with Klebsiella species being the predominant type (3). OXA-48 is a carbapenemase produced by a growing number of CPE isolates (4). Plasmid-mediated spread has led it to become a growing problem in the Middle East, North Africa, and Europe, especially in the Mediterranean (2, 4). Turkey, the Middle East, and North African countries (mainly Morocco, Tunisia, Egypt, and Libya) are considered the principal reservoirs (5). OXA-181 appears to be unique in having India as its main reservoir (6). There have been OXA-48-like producing isolates reported in Lebanon, the Sultanate of Oman, Saudi Arabia, Kuwait, and recently in European countries, including France, Germany, The Netherlands, Italy, Belgium, the United Kingdom, Ireland, Slovenia, Switzerland, and Spain (5). Although thought to be almost absent from the Americas, reports have revealed their slow emergence (3). Australia and New Zealand have both reported sporadic cases (7). A recent community-acquired urinary tract infection in Algeria has raised concerns regarding community reservoirs of bla_OXA-48 (8). Hospital outbreaks involving OXA-48-producing Enterobacteriaceae have been reported in France, Germany, Switzerland, the Netherlands, and the United Kingdom (9). The clonal expansion of Escherichia coli ST38 carrying a chromosomally integrated bla_OXA-48 has also been documented (10). The OXA-48 enzyme is an Ambler class D beta-lactamase that hydrolyzes carbapenems but shows very weak activity against extended-spectrum cephalosporins such as cefepime and ceftazidime (2), although isolates are frequently multidrug resistant as they combine multiple resistance mechanisms (11). The gene encoding OXA-48, bla_OXA-48, appears to have originated from Shewanella oneidensis that naturally harbors bla_OXA-54, which has a 92% amino acid sequence homology with bla_OXA-48 (12).

The first identification of OXA-48-producing Enterobacteriaceae was from an isolate in Turkey in 2001 (13). Shortly thereafter, there was an outbreak of OXA-48-producing Klebsiella pneumoniae isolates reported in Istanbul in 2006 (14). Since its recognition, there have been increasing numbers of reports of OXA-48-producing organisms worldwide, particularly in Europe (2). OXA-48 carbapenemases appear to be uncommon in the United States, with the first reported OXA-48-producing isolate identified in 2009 (5, 15). Of 52 identified isolates producing OXA-48-like carbapenemases in the United States between 2010 and 2015, two-thirds were acquired from outside the United States (most frequently India), and the majority of these involved patients hospitalized while overseas (3). OXA-48-like carbapenemases have also been subsequently described, which include OXA-181 and OXA-163 (2). The OXA-181 variant, which differs from OXA-48 by four amino acid substitutions, is prominent in India and has been associated with other carbapenemase genes, including bla_NDM-1 and bla_VIM-5 (16). OXA-181 has a similar pattern of resistance to OXA-48; however, OXA-163 hydrolyzes extended-spectrum cephalosporins but has only weak activity against carbapenems (2). Reviewing the treatment efficacy of different antimicrobials is somewhat challenging due to a paucity of clinical data, often related to the inherent difficulties in laboratory detection of OXA-48 producers due to many of the isolates falling within the susceptible range for carbapenems. The challenge of detecting OXA-48 producers is outside the scope of this review and the authors direct the readers to other reviews on this topic (6). Currently, there is a lack of evidence and data on treating infections due to OXA-48 producers, as recent CPE literature has focused more on Klebsiella pneumoniae carbapenemase (KPC)-producers.

IN VITRO ACTIVITY AGAINST OXA-48 PRODUCERS

OXA-48 producers display a unique in vitro susceptibility pattern, often making detection difficult with phenotypic methods alone (17). Initial early descriptions of the dissemination of OXA-48 enzymes were linked to a single IncL/M-type plasmid not containing any other resistance genes (5). However, more recently, epidemiologic studies have demonstrated a strong association with other beta-lactamase genes, in particular, extended-spectrum beta-lactamases (ESBLs) (18). OXA-48 is known to hydrolyze penicillins at a high level and carbapenems at a low level, sparing expanded-spectrum cephalosporins (5). Most current available beta-lactamase inhibitors (BLI) act only against class A beta-lactamases, remaining ineffective against OXA-48 producers. This has led to a recent surge in research evaluating the in vitro effectiveness of new BLI and beta-lactam–BLI combinations. Ceftazidime, cefepime, and aztreonam are not substrates for the OXA-48 beta-lactamase (19, 20). Numerous studies have demonstrated ceftazidime in vitro efficacy against OXA-48 producers, in the absence of ESBL coproduction (18, 21–23). Aztreonam, a member of the monobactam class of antibiotics, is stable from hydrolysis by OXA-48 enzymes and also demonstrates activity against metallo-beta-lactamase (MBL)-producing organisms (20, 24). There appears to be variability in the levels of hydrolysis among the different carbapenem antimicrobials. Carbapenems have been evaluated both in vitro and in a murine peritonitis model (20, 25). Doripenem and imipenem, despite testing susceptible in vitro, were not efficacious in an in vivo murine model; ertapenem displayed little activity in vivo and in vitro with an MIC of 3 mg/liter (20). Moreover, all 30 OXA-48 isolates characterized in a recent study from the United States were ertapenem resistant (26). There have been a number of case reports and series describing treatment failures with carbapenem-containing regimens in the treatment of OXA-48-producing bacterial infections (4, 27–29). Dual carbapenem combinations, particularly those with imipenem, have shown synergy in vitro against both OXA-48 and KPC producers (30, 31). Although, only some study isolates (K. pneumoniae but not recombinant E. coli) showed this result and produced values close to the cutoff value for defining synergy (30). Interestingly, levofloxacin exhibited activity against isogenic OXA-48 strains, but demonstrated little activity against clinical isolates and failed to suppress infection in an in vivo mouse model. Temocillin, a carboxypenicillin used to treat multiresistant Gram-negative organisms, is hydrolyzed efficiently by OXA-48 and failed to demonstrate efficacy in a mouse model (20). Moreover, high-level temocillin resistance has been included in algorithms to aid in OXA-48 recognition (32). Amdinocillin, an old extended-spectrum penicillin, had in vitro activity in 48% of OXA-48 producers in one study when using an extrapolated urinary breakpoint MIC of 8 mg/liter (33). Polymyxins, namely, colistin, have shown inconsistent and variable results in in vitro studies, with many reporting resistance in K. pneumoniae, E. coli, and Enterobacter species clinical isolates found to produce OXA-48 (11, 34). Colistin resistance, likely mediated by alterations in the mgrB gene, was probably acquired after clonal dissemination of an OXA-48-producing isolate identified during an outbreak in France in 2014 (34). Moreover, the addition of fosfomycin to colistin was found to be antagonistic against 12 OXA-48 strains (11). This finding was not statistically significant and it is unclear whether it is specific to OXA-48-producing K. pneumoniae isolates, as there were no controls used in the study (11). Inconsistencies with regard to in vitro susceptibility have been observed with the glycylcycline antibiotic tigecycline, with a broad range of percentages for susceptible isolates identified across different studies (35–38). Table 1 lists MICs for a variety of these drugs.

TABLE 1.

In vitro activity against OXA-48 producers

Drug	Reference(s)	Total no. of isolates tested across all studies	MIC (mg/liter)			% susceptible
			50%	90%	Range
Ceftazidime	18, 21, 23, 26, 33, 35, 36, 37, 62, 70	674	0.12 to 256	128 to 512	0.06 to 512	8 to 40
Cefepime	18, 21, 22, 26, 46, 50, 70	242	0.25 to 32	16 to 512	0.12 to >512	0 to 46
Aztreonam	18, 33, 35, 36, 37, 70	599	128	128 to >512	<0.015 to >512	12.2 to 27
Meropenem	18, 26, 35, 36–40, 50, 62, 71	734	0.06 to 128	0.25 to 256	2 to 256	0 to 87.8
Doripenem	18, 26, 70	170	>4	0.5 to >4	0.25 to >4	10 to 83.3
Ertapenem	18, 26, 35, 36	373	>1	1 to >8	>1	0 to 4.1
Imipenem	18, 23, 26, 35, 36, 72	781	0.12 to 32	8 to 256	1 to 256	0 to 74.8
Levofloxacin	18, 70	140	>4	>4	0.03 to >4	0 to 20.9
Amdinocillin	33	14	—a	—	—	43
Plazomicin (ACHN-490)	73, 74	19	<0.12 to 64	—	<0.12 to 64	—
Amikacin	18, 26, 35, 36, 37, 73	628	1 to 8	2 to >64	0.5 to >64	36.7 to 100
Gentamicin	73	19	0.25 to 128	—	0.25 to >128	68
Colistin	18, 26, 34–36, 38	520	<0.12 to 0.25	1 to 128	0.015 to 128	75.4 to 100
Tigecycline	18, 26, 35, 36–38	649	0.25 to 1	2	0.12 to 8	36 to 91
Fosfomycin	35, 36	281	—	—	—	44.8 to 82
Piperacillin-tazobactam	18, 62, 70	155	128	128 to 256	32 to 256	0
Ceftazidime-avibactam	18, 19, 21, 23, 33, 37, 62	705	0.03 to 1	0.5 to 256	<0.008 to 256	80 to 100
Cefepime-avibactam	21	26	0.016 to 0.12	0.5	<0.008 to 0.5	100
Cefepime-zidebactam	22	15	<0.03	—	—	100
Cefepime-tazobactam	46	25	1 to 8	—	1 to 8	80
Imipenem-avibactam	21	26	0.016 to 0.5	0.5	<0.008 to 2	100
Imipenem-relebactam	23	4	1 to 8	—	1 to 8	25
Aztreonam-avibactam	24, 33	71	0.12 to 0.25	0.25 to 1	0.12 to 1	100
Meropenem-vaborbactam	39, 71	62	16	≥32	0.12 to >32	0 to 3.8
Ceftolozane-tazobactam	19, 47	21	1 to >256	>256	0.5 to >256	13 to 50
Cefiderocol (S-649266)	50	6	0.25	2	0.125 to 2	—
Meropenem–CB-618	40	2	0.25 to 0.5	—	0.25 to 0.5	—

^a—, data not available.

Novel beta-lactamase inhibitors.

New and established beta-lactamase inhibitors have been extensively studied for their activity against OXA-48-producing Enterobacteriaceae. OXA-48 carbapenemases are not effectively inhibited by commonly used BLI such as clavulanic acid, tazobactam, and sulbactam, which mainly inhibit class A beta-lactamases (21). Avibactam is a diazabicyclooctane (DBO), a non-beta-lactam BLI which may have the ability to inhibit OXA-48 beta-lactamase by forming a stable covalent complex (21). In one study, it improved the efficacy of imipenem against OXA-48-producing K. pneumoniae isolates with an MIC₅₀ of 0.016 mg/liter, equating to a 500-fold reduction in the MIC₅₀ value. It also demonstrates inhibition of class A and C beta-lactamases, making its combination with cefepime and ceftazidime effective against OXA-48–ESBL-coproducing organisms, resulting in 266- and 1,000-fold reductions in MIC₅₀ value, respectively (21, 23). Avibactam has also been combined with aztreonam, exhibiting a stability from hydrolysis with OXA-48, MBLs, ESBLs, AmpC, and KPC (33). Vaborbactam, a cyclic boronic acid BLI, exhibited no activity against isolates harboring bla_OXA-48 when combined with meropenem (39). Moreover, relebactam, another novel BLI, also showed no activity when combined with imipenem (23). It is unclear whether the novel DBO BLI zidebactam is effective at inhibiting OXA-48 due to it being combined in one study with cefepime, which is known to be stable against OXA-48 (22). Compound CB-618 is another novel DBO BLI that has shown promise, with in vitro susceptibility of OXA-48 producers when combined with meropenem (40). Compound ETX2514 is yet another broad-spectrum DBO with activity against class D beta-lactamases and is currently in clinical development in combination with sulbactam for the treatment of Acinetobacter baumannii infections (41). ETX2514 also potently inhibits penicillin-binding proteins, resulting in intrinsic antibacterial activity that has confounded its analysis as a novel BLI (42). WCK 234 is a DBO that displayed uniquely high activity against strains with class D carbapenemases when paired with imipenem and meropenem in one study. It is a pure BLI and is not known to display direct antibacterial activity (43). LN-1-255 is a penicillin sulfone derivative highly effective at inhibiting OXA-48, as demonstrated in clinical isolates and in porin-deficient transformed K. pneumoniae. It has a 33-fold higher inhibition efficiency than tazobactam and binds OXA-48 very effectively (44). A focused fragment library has been set up to examine the in vitro inhibitory activity of 49 benzoic acid derivative fragments against OXA-48 as well as to observe their binding pockets and conformations (45). This approach shows promise in identifying new molecules for building novel BLIs.

Novel antibiotics.

The MIC values for ceftazidime-resistant OXA-48 producers, presumably due to the coproduction of ESBLs, to the new antibiotic ceftolozane-tazobactam are mostly higher than for cefepime-tazobactam (46). In a study of 353 OXA-48-producing isolates, 81.9% of those with a ceftazidime MIC of ≤4 mg/liter were susceptible (EUCAST and CLSI breakpoint, ≤1 mg/liter) to ceftolozane-tazobactam (C/T), with only 8.1% of those with an MIC of ≥4 mg/liter to ceftazidime reported as susceptible to C/T (47). The mechanism for this observation is unknown and needs further investigation; however, explanations include the carriage of additional porin mutations or the possibility of OXA-48 overwhelming tazobactam, rendering it unable to protect ceftolozane against AmpC or ESBL enzymes. The next-generation aminoglycoside plazomicin inhibited all carbapenemase-producing K. pneumoniae isolates (MIC < 1 mg/liter) (48). As with other aminoglycosides, its in vitro activity is compromised by the presence of ribosomal methylases, including ArmA and RmtC. The frequency of these resistance determinants in OXA-48 producers is uncertain but understood to be significant (49). Plazomicin showed activity against all CPE isolates irrespective of species, carbapenemase production, or level of resistance to antibiotics. In addition, synergy was observed at physiologically attainable concentrations when plazomicin was combined with colistin, meropenem, and fosfomycin (48). A new siderophore cephalosporin, cefiderocol (S-649266), demonstrated activity against OXA-48-producing strains of K. pneumoniae and Enterobacter cloacae, with MIC values between <0.125 and 2 mg/liter (50). The novel fluorocycline antibiotic eravacycline demonstrated in vitro efficacy against CPE isolates and carbapenem-nonsusceptible Acinetobacter baumannii (51, 52). It has demonstrated noninferiority to ertapenem in the treatment of complicated intraabdominal infection (cIAI) and is awaiting planned clinical trials assessing its utility in complicated urinary tract infections (cUTIs) (53). Although efficacy in OXA-48-like producers has not been specifically evaluated, eravacycline remains a possible candidate for future treatment options for nosocomial and infections due to multidrug-resistant (MDR) pathogens.

RISK FACTORS FOR OXA-48-PRODUCING ORGANISM COLONIZATION AND INFECTION

Identifying patient risk factors for colonization and infection is of great importance in relation to the timely institution of effective therapy. In areas where OXA-48-producing organisms are rarely encountered, recent travel to an area with a high prevalence provides a significant risk for acquisition (3). The identified risk factors for colonization of OXA-48-producing organisms include previous intraabdominal surgery, gastrointestinal or biliary endoscopy, hospitalization and intensive care unit (ICU) admission, and antibiotic intake (54). Male sex and advanced age were identified as factors in another study (55). A multivariate analysis identified the prior administration of a third- or fourth-generation cephalosporin (P < 0.001) and beta-lactam–BLI combination (P < 0.001) as strong risk factors for colonization (54). An estimated 10% to 30% of patients who are colonized with OXA-48-producing CPE will develop infections (56). Another study identified a modest <10% infection rate in colonized patients (57). The risk factors for infection are previous interhospital transfers, admission to an ICU within the previous 3 months, invasive procedures during hospitalization (including the insertion of a central venous line), and mechanical ventilation (56). In addition, the cumulative infection risk was found to progressively increase during the first 100 days in the hospital. Overall, higher rates of colonization and/or infection were seen with other multidrug-resistant Gram-negative organisms than with OXA-48-producing CPE.

TREATMENT OPTIONS

An ideal antimicrobial therapy for OXA-48-producing Enterobacteriaceae infection has not been established. A scarcity of clinical data, the heterogeneity in pathogen genotypes assessed, and the nonrandomization and retrospective nature of most studies have contributed to our inability to answer critical questions. Extended-spectrum cephalosporin monotherapy has been used successfully in two case reports for presumed ESBL-negative OXA-48-positive infection, although this is uncommon (58, 59). Carbapenem monotherapy has been used, with various reported rates of success. Meropenem and imipenem were the only carbapenems used as monotherapies in confirmed infections with OXA-48-producing organisms, with clinical cure and survival rates ranging between 0% and 66% (4, 27–29). Most cases were bloodstream infections, often complicating intravenous catheter placement. The treatment failure of carbapenem monotherapy is likely higher, as many treatment successes involved the removal of the central venous catheters. Amikacin has also been used with moderate success, with two successful outcomes in catheter-related bloodstream infections; however, a death was observed in a case of complicated urinary tract infection (27). Its success is likely also related to the prevalence of alternative mechanisms of resistance in local isolates. Colistin monotherapy resulted in a clinical cure in one case of OXA-48-producing K. pneumoniae bacteremia (27). Tigecycline has been used as a single-agent therapy in two cases of bacteremia secondary to a complicated intraabdominal and soft tissue infection (27). Although surgical intervention was a key component of their management, both patients had successful outcomes with tigecycline (27). Ceftazidime-avibactam as a salvage therapy for CPE infection has shown promise in a case series. In hematology patients with CPE bacteremia, 28/38 patients (73.7%) achieved clinical and/or microbiological cure with ceftazidime-avibactam as a monotherapy or in combination with another antimicrobial, with a delay in therapy being associated with worse outcomes (60). For infections caused by OXA-48 producers, salvage ceftazidime-avibactam produced clinical cure, microbiological cure, and survival to discharge in 8/13 (61.5%), 6/13 (46%), and 5/13 (38.5%) cases, respectively (60). In comparison, in those infected with KPC-producing Enterobacteriaceae, the treatment yielded rates of 17/23 (74%), 18/23 (78%), and 17/23 (74%), respectively. Of all CPE organisms, clinical cure occurred among 68% and 32% of KPC and OXA-48 producers, respectively, suggesting an advantage of ceftazidime-avibactam against this type of carbapenemase (60). Castón et al. reported a higher probability of clinical cure with ceftazidime-avibactam in CPE bacteremia; however, no significant difference in mortality was noted compared to that with other antimicrobial regimens (61). Although the majority of CPE isolates included in that study produced an OXA-48 carbapenemase (61.3%), a significant number produced a Klebsiella pneumoniae carbapenemase (KPC) (61). Surveillance data have suggested that resistance to ceftazidime-avibactam among CPE isolates (excluding those producing Ambler class B beta-lactamases) is rare. A recent case report documented the genetic evidence for MIC elevations resulting in full nonsusceptibility emerging in vivo through the evolution of bla_CTX-M-14 in a OXA-48-producing K. pneumoniae strain (62).

Combination therapy is often used for CPE infections and is proposed to offer reduced mortality and fewer treatment failures (63). A large international prospective study noted an improved survival within combination therapy, but only in those with high mortality scores (64). In infections involving OXA-48 carbapenemases, dual- and triple-therapy regimes containing various combinations of a carbapenem, aminoglycoside, colistin, tigecycline, and fosfomycin have been used with moderate success. A carbapenem, either imipenem or meropenem, plus colistin has been used in seven cases involving catheter-related bloodstream infections, pneumonia, and intraabdominal infection (4, 27, 29, 65). Only 3/7 (42.8%) patients treated with this regimen achieved a clinical cure and survived. However, the addition of amikacin to a carbapenem, when there is susceptibility to amikacin, achieved better outcomes, with 3/4 (75%) patients with OXA-48 bloodstream infections treated with this regimen surviving (27, 66). Colistin has been used in combination with both tigecycline and fosfomycin in six and four patients, respectively, in a large retrospective case series by Navarro-San Francisco et al. which documented survival rates of 5/6 (83%) and 3/4 (75%) (27). Non-carbapenem- and non-colistin-based dual therapies with tigecycline and amikacin were used in four patients in the same case series, which reported survival in 50% (27). Treatment failure and mortality were lower among those who received colistin, carbapenem, and/or an aminoglycoside in a combination regimen, although the combination of colistin and an aminoglycoside was noted to be significantly nephrotoxic (63). An increase in survival was noted with carbapenem-containing regimens when meropenem MIC values were <4 mg/liter (63).

Conversely, another study documenting an outbreak of OXA-48-producing CPE reported an increase in mortality associated with imipenem-containing therapy (4). A lack of adequate detection of OXA-48-producing strains was noted in this case series, attributed to low-level resistance to carbapenems and a lack of additional ESBLs, which may have contributed to the higher mortality rate. In bloodstream infections due to OXA-48-producing Enterobacteriaceae, a non-colistin-based combination regimen was found to be an independent risk factor for death in one study (67). When colistin was used, the survival rates were not significantly different when it was combined with either a carbapenem, a aminoglycoside, or a broad-spectrum cephalosporin in the setting of bacteremia (27). Interestingly, colistin has been reported as the least effective when used as a monotherapy (67).

Ceftazidime-avibactam combined with aztreonam in vitro demonstrated a synergistic effect against MBL-coproducing Enterobacteriaceae in two reports, being particularly effective against K. pneumoniae isolates, again highlighting the need to assess additional mechanisms of carbapenem resistance in OXA-48-carrying strains (68). Lastly, nonantibiotic options have been explored with an OXA-48-type carbapenemase-producing K. pneumoniae diabetic foot osteomyelitis acquired in Tunisia, which was successfully treated with hyperbaric oxygen therapy alone (69). Overall, these observations highlight the variability and unpredictability of appropriate therapeutic options for this type of infection. Table 2 lists the treatment options for OXA-48 producers.

TABLE 2.

Treatment options for OXA-48 producers

Drug regimen	Reference(s)	No. of patients	Type of infection (n)	Microbiological cure (%)	Survival at 14–30 d (%)
Ceftriaxone	27, 59	2	Endometritis (1)	—a	100
			Bloodstream infection (1)
Carbapenem	4, 27–29	10	Pneumonia (1)	—	0–66
			Bloodstream infection (9)
Amikacin	27	3	Bloodstream infection (3)	—	66
Colistin	27	1	Bloodstream infection (1)	—	100
Tigecycline	27	2	Bloodstream infection (2)	—	100
Ceftazidime-avibactam	60, 61	18	Bloodstream infection (5)	25	22.7–70.6
Ceftazidime plus colistin	58	1	Catheter-related bloodstream infection (1)	—	100
Carbapenem plus amikacin	27, 66	4	Bloodstream infection (4)	—	66–100
Carbapenem plus colistin	4, 27, 29, 65	7	Catheter-related bloodstream infection (1)	—	0–100
			Bloodstream infection (4)
			Pneumonia (1)
			Intraabdominal infection (1)
Colistin plus fosfomycin	27	4	Bloodstream infection (4)	—	75
Colistin plus amikacin	27	1	Bloodstream infection (1)	—	100
Tigecycline plus amikacin	27	4	Bloodstream infection (4)	—	50
Tigecycline plus fosfomycin	27	1	Bloodstream infection (1)	—	100
Tigecycline plus colistin	27	6	Bloodstream infection (6)	—	83
Carbapenem plus amikacin plus colistin	4	2	Catheter-related bloodstream infection (1)	—	50
			Pneumonia (1)
Colistin plus amikacin plus tigecycline	27	2	Bloodstream infection (2)	—	50

^a—, data not available.

TREATMENT STRATEGIES

An approach to treating infections due to OXA-48-producing Enterobacteriaceae is undefined and yet to be rigorously studied. Given the unavailability of a single effective antibiotic, multiple strategies have been proposed and advocated for in various retrospective studies involving low patient numbers. Combination therapies using two or three agents with demonstrated activities against an OXA-48-producing isolate have demonstrated a mortality benefit compared with that for monotherapy, in a systematic review (63). No difference was noted between dual and triple-therapy regimens. Another study reported greater survival and microbiological cure of patients with OXA-48-producing Enterobacteriaceae bacteremia than in those treated with a single agent; however, this was attributed to greater source control and catheter removal (67). Despite a theoretical benefit, there are no clinical data to support the notion that combination therapy reduces the emergence of resistance (63).

CONCLUSION

Patients with infection due to OXA-48-producing Enterobacteriaceae experience high mortality, as evidenced by a 30-day mortality rate of 50% in patients with bacteremia (27). The progressive spread of bla_OXA-48-containing organisms has been clearly documented in Africa, the Middle East, and Europe, particularly in the Mediterranean, and is an active threat to the rest of the world (2, 3). OXA-48 has low-level carbapenem hydrolytic activity and resistance, making it difficult to detect by standard laboratory methods and making treatment failure common. A number of antibiotics have shown in vitro activity against OXA-48-producing Enterobacteriaceae; however, the results have been largely inconsistent. In addition, colistin-based regimens are toxic and not well tolerated (63). Ceftazidime-avibactam perhaps has proven to be most successful according to available in vitro and clinical data; however, reports of resistance are emerging. New BLIs, paired with appropriate beta-lactams, are being developed which may provide a solution in the future. New antibiotic choices and randomized controlled trials are needed to inform future decision making and increase the armamentarium used to treat these difficult infections.