Dissemination of a Clone of Unusual Phenotype of Pandrug-Resistant Acinetobacter baumannii at a University Hospital in Taiwan

Abstract

From December 2002 to February 2003, 15 isolates of pandrug-resistant unidentified Acinetobacter species were recovered from seven patients treated on different wards or intensive care units. Both 16S-23S rRNA intergenic spacer PCR-restriction fragment length polymorphism profiles and sequence analysis of these isolates identified them as Acinetobacter baumannii. This pandrug-resistant A. baumannii strain with an unusual phenotype could persist in humans for long periods and was widely disseminated throughout the hospital.

Bacteria that constitute the genus Acinetobacter were originally identified in the early 20th century, but it was recognized as a ubiquitous pathogen only in the last decade (2). Acinetobacter baumannii (one member of the Acinetobacter calcoaceticus-A. baumannii complex) makes up 73% of all Acinetobacter clinical isolates and is the Acinetobacter species that is the most commonly involved in clinical infections (3, 25). Nosocomial infections caused by multidrug-resistant A. baumannii have been reported in recent years (4, 5, 12, 13, 24). The emergence of carbapenem-resistant A. baumannii (CRAB) was reported in the United States in 1991 (8). Since then, CRAB infections and hospital-wide outbreaks have been reported worldwide (1, 5). Isolates of pandrug-resistant A. baumannii (PDRAB), which are resistant to all antibiotics routinely tested (i.e., ampicillin-sulbactam, ceftazidime, piperacillin-tazobactam, cefepime, aztreonam, ciprofloxacin, trovafloxacin, moxifloxacin, garenoxacin, amikacin, imipenem, and meropenem), were first recovered in May 1998 at the National Taiwan University Hospital (NTUH) (14, 17). Since then, clusters of PDRAB infections and nosocomial outbreaks have persisted, although the incidence of nosocomial infections caused by PDRAB has declined in the past 2 years (12, 13, 15).

In December 2002, an isolate of pandrug-resistant Acinetobacter species was recovered from the respiratory secretions of a hospitalized patient. This isolate was presumptively identified to be an unusual phenotype of pandrug-resistant A. baumannii (PDRABup) because of its negative reaction to 10% lactose. In the following 3 months a total of 15 isolates of this unusual phenotype of pandrug-resistant Acinetobacter were recovered from various clinical specimens from seven patients at the hospital.

The clinical characteristics of the seven patients are shown in Table 1. The mean age of the patients was 64 years (age range, 14 to 90 years). One patient had an underlying malignancy (cholangiocarcinoma), but none of the patients had hematological malignancies or immunodeficiency. All except one patient (patient 5) had fever as the clinical presentation of the infection. Six patients were bedridden due to stroke or head injury. All patients except patients 2 and 5 had received endotracheal mechanical ventilation. The numbers of cultures of clinical samples from these patients prior to the isolation of PDRABup ranged from 3 to 15. The PDRABup isolates were first recovered from these patients 9 to 38 days after admission. All infections or colonizations due to PDRABup were hospital acquired.

TABLE 1.

Clinical characteristics of seven patients with samples positive by culture for pandrug-resistant A. baumannii with an unusual phenotype who were treated at NTUH^a

Patient no.	Sex/age (yr)	Underlying diseases or conditions	Clinical syndrome	Isolation of A. baumannii			Coisolates	Earlier antibiotic treatment	Antibiotic treatment	Outcome
				Site	Isolate designation	Date (day/mo/yr)
1	F/68	Ischemic stroke, RHD, CHF, DM, UTI, pneumonia	Fever	Sputum	A1	2/12/2002	S. maltsophilia, MRSA	Amoxicillin-clavulanate	Ceftazidime, cefepime	Survival
				Sputum	A2	18/12/2002
				Sputum	A3	23/12/2002
				Sputum	A4	2/1/2003
2	M/81	Previous stroke, BPH, recurrent UTI	Fever	Urine	B	5/12/2002	None	Ceftazidime, cefazolin	None	Survival
3	F/76	Cerebellar hemorrhage, post-tracheostomy, hypertension, pneumonia, UTI	Fever	Sputum	C1	9/12/2002	S. maltophilia, Proteus mirabilis, P. aeruginosa, Klebsiella pneumoniae, Candida albicans	Cefuroxime, vancomycin	Flomoxef, ciprofloxacin, meropenem	Survival
				Sputum	C2	30/12/2002
				Sputum	C3	10/1/2002
				Sputum	C4	20/1/2003
				Urine	C5	27/1/2003
4	F/14	Encephalitis, status epilepticus, respiratory failure, UTI	Fever	Sputum, pressure sore	D	12/12/2002	MRSA	Ciprofloxacin, clindamycin, cefepime, vancomycin, meropenem	Meropenem	Survival
5	M/60	Previous stroke, DM, CHF, renal insufficiency, hypoxic encephalopathy, UTI	Shock (no fever)	Wound (amputation site)	E	21/1/2003	MRSA	Imipenem, vancomycin, cefepime	None	Survival
6	F/90	Head injury, intracranial hemorrhage, CHF, cardiogenic shock, cellulites, acute renal failure, pneumonia	Fever	Sputum	F	2/2/2003	P. aeruginosa	Cefepime, vancomycin	Ciprofloxacin, cefepime	Death
7	M/65	Chronic renal failure, UTI, cholangiocarcinoma with obstructive jaundice and BTI, surgery	Fever, shock	Sputum, urine, PTCD	G	20/2/2003	MRSA, Chryseobacterium menigosepticum, S. maltophilia	Cefazolin, cefoxitin, metronidazole, ceftazidime, imipenem, ciprofloxacin	Ciprofloxacin	Death

^aAbbreviations: F, female; M, male; BPH, benign prostatic hypertrophy; BTI, biliary tract infection; CHF, congestive heart failure; DM, diabetes mellitus; PTCD, percutaneous transhepatic cholangial drainage; RHD, rheumatic heart disease; UTI, urinary tract infection.

The most common site of infection was the respiratory tract. PDRABup was isolated from multiple sites in three patients (patients 3, 4, and 7). Additional isolates were identified from the same specimens in six patients, including methicillin-resistant Staphylococcus aureus (MRSA) (four patients), Stenotrophomonas maltophilia (three patients), and Pseudomonas aeruginosa (two patients). All but one (patient 1) received various antimicrobial agents for 7 to 14 days before acquisition of PDRABup. After notification of a culture positive for PDRABup, two patients (patients 2 and 5) did not receive any parenteral antimicrobial agents; and five patients were treated with meropenem, ciprofloxacin, or extended-spectrum cephalosporins, according to the susceptibility testing results for the additional isolates. Five patients survived.

All 15 isolates were gram-negative bacilli, oxidase negative, nonhemolytic, lactose (10%) negative, and glucose nonacidifiers. When they were grown on Trypticase soy agar supplemented with 5% sheep's blood, the colonies were mucoid and slightly pink. They were identified as Acinetobacter lwoffii by the Vitek GNI system (97% presumptive identification; bioMerieux, Marcy l'Etoile, France), A. baumannii-A. calcoaceticus complex by the API 20NE system (76.8 to 83% presumptive identification; bioMerieux), and Acinetobacter species by the Phoenix system (90% confidence value; Becton-Dickinson, Sparks, Md.). A control strain, A. baumannii ATCC 19606, was 10% lactose positive and a glucose acidifier and was identified as A. baumannii-A. calcoaceticus complex by the Vitek GNI system (99% presumptive identification) and the API 20NE system (99% presumptive identification) and as Acinetobacter species by the Phoenix system (92% confidence value) (Table 2).

TABLE 2.

Biochemical profiles of A. baumannii ATCC 19606 and the PDRABup clone

Reaction	Biochemical profile
	A. baumannii ATCC 19606	PDRABup clone
Growth at:
37°C	+	+
41°C	+	+
Acid production from:
Glucose	+	−
Lactose	+	−
Sucrose	−	−
Galactose	−	−
Fructose	−	−
Mannitol	−	−
α-Ketoglutaric acid	+	+
Tiglic acid	+	+
Utilization of:
Acetate	+	+
Citrate	+	+
Malonate	+	+
Glycine	+	+
Leucine	+	+
Presence of:
Arginine dihydrolase	+	−
Orinithine decarboxylase	−	−
Urease	−	−
Esculin hydrolysis	−	−

PCR amplification of the complete 16S rRNA gene and direct sequencing of the amplification product were performed as described previously (16). PCR amplification of the 16S-23S intergenic spacer sequences (ITSs) was also performed, and the amplicons were digested with AluI and then subjected to restriction fragment length polymorphism (RFLP) analysis and direct sequencing based on a previously described protocol (6). The 16S rRNA sequencing data for these isolates (650 nucleotides) were identical, and the sequences of the isolates were comparable to those of the A. baumannii-A. calcoaceticus complex or Acinetobacter junii. The results of 16S-23S ITS PCR-RFLP analysis and the following sequencing analysis (883 nucleotides) were identical and confirmed the identification of the isolates as A. baumannii. We detected only one nucleotide difference between the 16S-23S ITS sequences of our isolates and that of A. baumannii ATCC 19606. The percent matches between the sequences of our isolates and the sequences of isolates with GenBank accession numbers U60279 (A. baumannii), U60280 (Acinetobacter ATCC 19004, genomospecies 3), U60281 (Acinetobacter ATCC 17903, genomospecies 13), and U60278 (A. calcoaceticus, genomospecies 1) were 99.0, 95.7, 95.7, and 95.7%, respectively.

The MICs of 14 antimicrobial agents for the 15 PDRABup isolates were determined by the agar dilution method according to the guidelines established by the National Committee for Clinical Laboratory Standards (21, 22). All these isolates were uniformly resistant to ampicillin-sulbactam (MICs, 128 to >128 μg/ml), ceftazidime (MICs, 64 to >128 μg/ml), piperacillin-tazobactam (MICs, 128 to >128 μg/ml), cefepime (MICs, 16 to 32 μg/ml), aztreonam (MICs, 64 to 128 μg/ml), ciprofloxacin (MICs, 64 to 128 μg/ml), trovafloxacin (MICs, 8 to 16 μg/ml), moxifloxacin (MICs, 4 μg/ml), garenoxacin (MICs, 8 to 32 μg/ml), amikacin (MICs, >128 μg/ml), imipenem (MICs, 8 to 16 μg/ml), and meropenem (MICs, 16 to >128 μg/ml).

For synergy analysis, five pairs of antimicrobial disks (ceftazidime and amikacin, cefepime and amikacin, imipenem and amikacin, imipenem and ampicillin-clavulanate, and imipenem-ciprofloxacin) were applied onto unsupplemented Mueller-Hinton agar with a distance of 15 mm (center to center) between two disks. Synergy between two antimicrobial agents was identified as the presence of an enhanced zone of inhibition between two antimicrobial disks. Among the five pairs of antimicrobial agents tested for synergy, only imipenem and ampicillin-sulbactam showed an enhanced zone of inhibition between the two disks for all 15 isolates (Fig. 1).

FIG. 1.

Enhanced zone of inhibition between imipenem (IPM) and ampicillin-sulbactam (SAM) disks for the PDRABup isolates.

The genotypes of the 15 isolates of PDRABup and five PDRAB isolates, one isolate each of clones 1 to 5 reported previously (14), were determined by pulsed-field gel electrophoresis (PFGE) (14). The DNA was digested with the restriction enzyme SfiI, and the restriction fragments were separated in a CHEF-DRIII unit (Bio-Rad, Hercules, Calif.). All 15 PDRABup isolates had identical PFGE profiles, and this profile was different from that for the five PDRAB isolates (Fig. 2) (the PFGE profile for one of the five PDRAB isolates is shown in Fig. 2). Four isolates from sputum specimens recovered from patient 1 at an interval of 1 month had identical PFGE profiles. Five isolates recovered from patient 3 (four from sputum specimens and one from a urine sample) at an interval of 2 months also had identical PFGE profiles.

FIG. 2.

Profiles obtained by PFGE for A. baumannii after digestion with SfiI. Lane M, molecular size marker; lanes 1 to 11, PDRABup isolates A1, A3, A4, B, C1, C4, C5, D, E, F, and G, respectively (see Table 1 for isolate designations); lane 12, a PDRAB isolate belonging to clone 5, the major clone of PDRAB shown in a previous study (14).

This report describes a PDRAB clone with a phenotype different from that of the PDRAB clones previously found at NTUH and characterizes a nosocomial outbreak due to these organisms over a 3-month period. Our results demonstrate three important facets. First, although glucose-nonoxidizing A. baumannii isolates account for 5% of all clinical A. baumannii isolates (3, 25), isolates of A. baumannii (including PDRAB isolates) with negative reactions for both glucose and lactose have never been seen at NTUH, until now. Because of problems with the identification of Acinetobacter species in routine clinical microbiology laboratories, a phenotypic scheme for the identification of genomospecies 1 to 12 has been described previously (3); however, by using this system, discrepancies with the identities obtained by DNA-DNA hybridization, 16S rRNA sequencing, and 16S-23S ITS PCR-RFLP analysis and sequencing have been found (6, 10, 16, 23). In the present study, the biochemical profiles of these asaccharolytic isolates (glucose-, lactose-, xylose-, and mannitol-nonoxidizing strains) of Acinetobacter species generated by three commercial biochemical identification kits did not allow categorization of the isolates as any particular genomospecies of Acinetobacter (3, 7, 25). Although only the 16S rRNA sequencing data suggested the identification of A. baumannii-A. calcoaceticus complex or A. junii, the 16S-23S ITS PCR-RFLP analysis and the subsequent sequencing analysis confirmed the identities of the isolates as A. baumannii.

Second, the PDRABup isolates had PFGE profiles different from those of 10 PDRAB clones recognized in a previous study (14), indicating that these PDRABup isolates belong to a newly emerging clone. As seen previously for other PDRAB isolates and other nonfermentative gram-negative bacteria, isolates of this clone could also persist in humans for long periods (infection or colonization for weeks and months) (11, 14).

Finally, all isolates were highly resistant to extended-spectrum cephalosporins, carbapenems, fluoroquinolones, aminoglycosides, and ampicillin-sulbactam. Recent studies have demonstrated that treatment with sulbactam alone at higher doses or treatment with sulbactam in combination with other agents is efficacious against nosocomial infections caused by multiresistant A. baumannii (9, 14, 18, 19, 20). Although synergy was detected by the disk method only for the combination of imipenem plus ampicillin-sulbactam, more studies including time-kill and in vivo animal studies should be performed to establish the treatment options.

The environmental source and mode of spread of the PDRABup isolates described here is obscure. During the period of this outbreak (December 2002 to February 2003), several small clusters of PDRAB infections still occurred in intensive care units, and environmental surveillance (samples from air humidifiers, the hands of medical staff, mattresses, stock solutions, sinks, taps, and a portable X-ray machine) failed to find the organism. Fortunately, this clone circulated in the hospital for 3 months and disappeared spontaneously; however, the classic PDRAB isolates persisted.

Clinically, it is difficult to determine the pathogenic role of this organism because of the poor underlying medical conditions of the patients, the polymicrobial growth in specimens from infected sites, and the absence of concurrent bacteremia due to this organism. Two well-known pathogens (P. aeruginosa and MRSA) were also identified in the two patients who died (patients 6 and 7).

In summary, we report on a nosocomial outbreak due to a novel PDRAB clone that occurred in seven patients at NTUH over a 3-month period. Because of the lack of sufficient phenotypic discriminating criteria for the identification of Acinetobacter species, molecular methods should be conducted, particularly with isolates with unusual phenotypes.