To the Editor: Infections caused by Acinetobacter spp., especially A. baumannii, have been increasingly documented in recent years. Carbapenems tend to be empirically prescribed as first-choice drugs for severe invasive infections caused by Acinetobacter spp. other than A. baumannii because these microbes are usually susceptible to carbapenems. However, infections with carbapenem-resistant Acinetobacter spp. have been increasingly reported during the past 15 years. In A. baumannii, carbapenems are usually inactivated by intrinsic oxacillinase (OXA)-51–like, acquired OXA-23–like, or OXA-58–like carbapenemases. Moreover, production of acquired metallo-β-lactamases (MBLs) of the Verona integron (VIM), imipenemase (IMP), or New Delhi (NDM) types has been detected among carbapenem-resistant Acinetobacter species, including A. baumannii, A. junii, A. bereziniae, A. nosocomialis, and A. pittii (1). We report a case of infection with carbapenem-resistant A. soli producing another MBL type, Tripoli MBL 2 (TMB-2), in a man in Japan.

A man in his 60s who had mesenteric injury, pelvic fracture, and intestinal perforation from a traffic accident was admitted to Okazaki City Hospital in Aichi, Japan, on May 3, 2013. After surgery, cefmetazole was prescribed on May 6 (1 g 2×/d for 7 d). On May 12, symptoms of infection developed in the patient, and 2 sets of blood samples were drawn from different vessels for bacterial culture. The following day, cefmetazole was discontinued, and ciprofloxacin (0.3 g 2×/d) and piperacillin/tazobactam (4.5 g 2×/d) were started. Acinetobacter isolates resistant to piperacillin/tazobactam and carbapenems were then recovered from the blood samples, so piperacillin/tazobactam was discontinued on May 14. After that, ceftriaxone (2 g 2×/d) and gentamicin (0.04 g 2×/d) were successively prescribed, in addition to ciprofloxacin; the symptoms of infection improved, and all antimicrobial drugs were discontinued by May 26. Additional blood cultures performed on May 17, 21, and 28 yielded negative results for Acinetobacter spp. However, the patient’s condition worsened on June 5. Meropenem (0.5 g 4×/d) was then given, but the patient died of multiorgan failure on June 7.

The bacterial isolates from the initial blood cultures were identified as A. soli by nucleotide sequencing of the rpoB and gyrB genes and assigned identification no. HK001. MICs of β-lactams, measured by the agar dilution method in accordance with the guideline M07-A9 of the Clinical and Laboratory Standards Institute (http://clsi.org), were as follows: sulbactam/ampicillin, >128 mg/L; piperacillin, >128 mg/L; tazobactam/piperacillin, >128 mg/L; cefotaxime, >64 mg/L; ceftazidime, >64 mg/L; aztreonam, 64 mg/L; cefmetazole, >128 mg/L; imipenem, 8 mg/L; meropenem, 32 mg/L; and doripenem, 32 mg/L. However, MICs of gentamicin, amikacin, levofloxacin, ciprofloxacin, colistin, and tigecycline were below the breakpoints of susceptibility as listed in Clinical and Laboratory Standards Institute document M100-S23. Carbapenem resistance was not transferred from A. soli HK001 to Escherichia coli strain CSH-2 (metB F^– NA^r Rif^r) by conjugation. A double-disk synergy test was initially performed by using sodium mercaptoacetic acid (SMA) (2) and ceftazidime and meropenem disks (Eiken Chemical Co., Ltd, Tokyo, Japan), and results suggested MBL production. The modified Hodge test was then performed, and ertapenem and meropenem disks gave clear positive results (data not shown). PCR was performed to detect bla_OXA-23–like, bla_OXA-24/40–like, bla_OXA-51–like, bla_OXA-58–like, bla_IMP-1, bla_IMP-2, bla_VIM-1, bla_VIM-2, bla_NDM-1, bla_SMB-1, and bla_TMB-1 genes. Nucleotide sequence analyses showed that the A. soli isolate harbored bla_TMB-2 and bla_OXA-58. The modified SMA-disk method (3) was reevaluated to determine whether it could successfully detect TMB-2 production in A. soli HK001. Apparent positive results were obtained when disks containing imipenem, meropenem, or ertapenem were used, particularly when the edge-to-edge distance between 2 disks containing SMA and a carbapenem, respectively, was kept at 5 mm (Figure, top row). However, when the distance between the ertapenem and SMA disks was >10 mm, MBL production was more difficult to detect (Figure, lower 2 rows). This finding may be the result of co-production of OXA-58 by the isolate.

More than 30 Acinetobacter species had been registered by January 2012 (4); A. soli was initially isolated from the soil of a mountain forest in South Korea in 2007 (5) and has been recovered from blood cultures of 5 neonates in Brazil (6). Carbapenem-resistant A. soli co-harboring bla_IMP-1 and bla_OXA-58–like genes was identified in April 2011 in Japan and is frequently recovered from bacteremia patients (7). TMB-1 was reported in 2012 in an Achromobacter xylosoxidans isolate from a hospital in Tripoli, Libya (8); TMB-2 was later reported in Japan (9). The TMB-2–producing A. soli strain that we isolated came from a blood culture, indicating that A. soli is a potential cause of bloodstream infections or bacteremia. A. soli has also been detected in lice and keds of domestic animals (10), indicating that A. soli may inhabit natural environments and that injuries and bites by arthropods might present a risk for invasive infections. Isolates of Acinetobacter species, particularly those recovered from blood culture, should be identified to species type to enable further evaluation of the clinical significance of carbapenem-resistant A. soli strains.