Molecular Epidemiological Investigation Using a Randomly Amplified Polymorphic DNA Assay of Burkholderia cepacia Isolates from Nosocomial Outbreaks

Abstract

We experienced two Burkholderia cepacia outbreaks over a 1-year period. During this period, 28 B. cepacia isolates were obtained from clinical specimens, and 2 were obtained from environmental specimens (i.e., from a nebulizer solution and a nebulizer tube). These 30 isolates were subjected to the PCR-based randomly amplified polymorphic DNA (RAPD) assay as well as to pulsed-field gel electrophoresis (PFGE). In the first outbreak, in which eight patients hospitalized in the Trauma and Critical Care Center were involved, the RAPD assay revealed that all 20 isolates obtained from clinical specimens and the 2 isolates from environmental specimens had identical DNA profiles. These RAPD data enabled us to pinpoint a possible source and to take countermeasures to prevent further spread of the epidemic-causing strain. In the second outbreak, two consecutive B. cepacia infection/colonization cases were seen in the surgery ward. The RAPD profiles of four isolates obtained were again identical, but they were distinct from those seen in the first outbreak, clearly indicating that the second outbreak was not related to the first. Thus, our experience demonstrated that the RAPD assay is a useful and reliable tool for epidemiological studies of B. cepacia isolates from nosocomial outbreaks. Since the RAPD assay could provide discriminatory potential and reproducibility comparable to those of the widely used PFGE assay with less complexity and in a shorter time, the introduction of the RAPD assay into hospital microbiology laboratories as a routine technique may help prevent nosocomial outbreaks.

Burkholderia cepacia, a ubiquitous bacterial species in the natural environment, is an important opportunistic pathogen, causing respiratory-tract infections in patients with cystic fibrosis and infections in various sites in immunocompromised hosts (1, 6, 18, 20). B. cepacia has been recovered from hospital environments, medical devices, and a variety of solutions used in clinical practice (13). In fact, several nosocomial outbreaks due to this organism have been reported (6, 15, 17). It may be transmitted via patient-to-patient contact, environmental contamination, and/or contact with health care workers (9). In order to prevent the nosocomial spread of epidemic-causing strains, accurate and prompt determination of the sources and routes of transmission is essential.

Recently, genomic typing of B. cepacia by means of DNA fingerprinting using pulsed-field gel electrophoresis (PFGE) and ribotyping based on the genomic characterization of strains have been reported (3, 7, 8, 11). Although PFGE analysis for strain identification is highly reproducible and discriminative, it has not been adopted as a routine examination procedure in clinical microbiology laboratories because it is time-consuming and technically complicated (12, 21). In contrast to the PFGE method, the randomly amplified polymorphic DNA (RAPD) assay, originally referred to as the arbitrarily primed PCR (AP-PCR), appears to yield results of comparable significance with a less complex procedure and in a shorter time (2). The aim of this study is to examine the applicability of the RAPD assay for determining the epidemiological relationships of B. cepacia isolates in a nosocomial outbreak.

MATERIALS AND METHODS

Patients and clinical samples.

From November 1995 to September 1996, 28 B. cepacia isolates were obtained from the clinical specimens of 13 patients in the Trauma and Critical Care Center (TCC) and other wards at Kyorin University Hospital (Fig. 1). Of the 13 patients, 9 (patients B through I and patient L) stayed in the TCC from 3 to 88 days (mean duration, 23 days). The sputum specimens from patients B through I were submitted to the microbiology laboratory between January and March 1996, and those from patient L were submitted between July and August 1996. A retrospective review of their medical records revealed that these nine patients received identical nebulized medication, i.e., a mixture of bromhexine hydrochloride, tyloxapol, isotonic sodium chloride solution, and distilled water, during their hospitalization in the TCC. SONICLIZER 305 (ATOM Co., Tokyo, Japan) machines (Fig. 2) were used for nebulizing of medication for patients in the TCC. The remaining four patients stayed in the wards indicated in parentheses and received the following nebulized medication: patient A (Respiratory Disease Department), methylprednisolone sodium succinate and isotonic sodium chloride solution; patient J (Thoracic Surgery Department), tyloxapol, isotonic sodium chloride solution, and distilled water; patient K (Thoracic Surgery Department), no nebulized medication; and patient M (Respiratory Disease Department), ambroxol hydrochloride, procaterol hydrochloride, and isotonic sodium chloride solution. Nebulizer machines used in these wards were of the same type as those used in the TCC. Sputum specimens from patients A, J, and M were submitted to the microbiology laboratory, while a blood sample from patient K was submitted (Table 1). The durations of stay for these 13 patients in their respective wards are shown in Fig. 1. Among these 13 patients, one (patient M) developed pneumonia and another (patient K) developed bacteremia due to B. cepacia, while the remaining 11 patients had B. cepacia colonization in the respiratory tract. Before March 1996, each solution for nebulizing was mixed and stored in the mixing bottle in each ward, including the TCC, and 0.1% benzalkonium chloride was used to sanitize the nebulizer machines, including their tubings.

FIG. 1.

Temporal profile of hospitalization of patients and the time of B. cepacia isolation from each patient between November 1995 and September 1996. Hospitalization periods are indicated by solid lines (TCC) and dotted lines (other wards). The following symbols indicate the RAPD profiles and times of B. cepacia isolation: ◊, R1 (isolate from patient A); ▾, R2 (isolates from patient B through I); , R3 (isolates from patients J and K); ▿, R4 (isolate from patient L); □, R5 (isolate from patient M). The short vertical arrow indicates B. cepacia isolation from environmental samplings in the TCC. The long vertical arrow indicates when new measures against the B. cepacia nosocomial outbreak were initiated. Solid circles indicate the first nosocomial outbreak, which occurred in the TCC, and open triangles indicate the second outbreak, which occurred in the surgical ward.

FIG. 2.

Ultrasonic nebulizer (SONICLIZER 305) used for nebulizing of medication for patients in the TCC and in other wards. The epidemic-causing strain in the first outbreak was obtained from the tube connected to the machine.

TABLE 1.

Patients and B. cepacia isolates

Type and no. of isolate	Patient	Patient location	Culture source	Admission diagnosis	Date of isolation (day.mo.yr)	API profilea	RAPD profile	PFGE profile
Clinical
1	A	Respiratory	Sputum	Systemic lupus erythmatodes	27.11.95	A1	R1	P1
2	B	TCC	Sputum	Acute myelocytic leukemia	31.01.96	A2	R2	P2
3	B		Sputum		01.02.96	A2	R2	P2
4	B		CVCTb		05.02.96	A1	R2	P2
5	Cc	Cardiac medicine	Sputum	Pneumonia	01.02.96	A2	R2	P2
6	C		Sputum		05.02.96	A1	R2	P2
7	C		Sputum		06.02.96	A1	R2	P2
8	C		Sputum		13.02.96	A1	R2	P2
9	C		Sputum		24.02.96	A1	R2	P2
10	C		Sputum		08.03.96	A1	R2	P2
11	D	TCC	Sputum	Inferior myocardial infarction	03.02.96	A1	R2	P2
12	E	TCC	Sputum	Burns	05.02.96	A1	R2	P2
13	Fc	Neurosurgery	Sputum	Subarachnoidal hemorrhage	05.02.96	A1	R2	P2
14	G	TCC	Sputum	Multiple trauma	13.02.96	A1	R2	P2
15	H	TCC	Sputum	Dissecting aortic aneurysm	19.02.96	A2	R2	P2
16	H		Sputum		22.02.96	A2	R2	P2
17	H		Sputum		24.02.96	A2	R2	P2
18	H		Sputum		27.02.96	A2	R2	P2
19	H		Sputum		28.02.96	A2	R2	P2
20	H		Sputum		29.02.96	A2	R2	P2
21	I	TCC	Sputum	Head injury	27.02.96	A2	R2	P2
22	J	Thoracic Surgery	Sputum	Cancer of the esophagus	16.07.96	A1	R3	P3
23	J		Pus		16.07.96	A2	R3	P3
24	J		Pharynx secretion		20.08.96	A2	R3	P3
25	K	Thoracic Surgery	Blood	Cancer of the rectum	23.07.96	A1	R3	P3
26	L	TCC	Sputum	Subarachnoidal hemorrhage	31.07.96	A2	R4	P4
27	L		Sputum		02.08.96	A1	R4	P4
28	M	Respiratory	Sputum	Tuberculosis	18.09.96	A1	R5	P5
Environmental
29		TCC	Nebulizer solution		29.02.96	A1	R2	P2
30		TCC	Nebulizer tube		29.02.96	A1	R2	P2

^aA1, 0467577; A2, 0477577. 

^bCVCT, central venous catheter tip. 

^cInitially patients C and F were hospitalized in the TCC. 

Sampling from respiratory-therapy equipment.

The high incidence of B. cepacia isolation from sputum specimens between the period from 31 January to 8 March 1996 prompted us to examine whether the respiratory-therapy machines were the environmental source of B. cepacia. The following swab samples were collected for culture: six samples from the nebulizer tubes, six from the nebulizing chambers, six from the medication cups, and six from the working water chambers which were connected to the main unit of each of the six SONICLIZER 305 machines in the TCC. In one case, the nebulizer solution was also sampled from a mixing bottle used for both patients H and I. In order to carry this out, the solution was centrifuged for 10 min at 2,000 × g and the resultant pellets were processed for culture. All 25 samples were inoculated onto 5% sheep blood agar (Oriental Yeast Co., Tokyo, Japan) and incubated for 48 h at 35°C in a humidified atmosphere. Each component of the nebulizer solution, including bromhexine hydrochloride, tyloxapol, isotonic sodium chloride solution, and distilled water, was also sampled separately.

API 20 NE test.

B. cepacia isolates were identified by the analytical profile index procedure by using the API 20NE system (API-BioMerieux, La Balme les Grottes, France).

RAPD analysis.

Total B. cepacia DNA was prepared as described previously (19) and analyzed by random PCR using two PCR primers synthesized in house: RPKHM1 (5′-AAGCCGGTGAGTTATCTGGCC-3′) and RPKHM2 (5′-CGTAACCGGACTGGGGCGTGT-3′) (14). Each 50 μl reaction mixture was composed of 100 mM Tris-HCl buffer (pH 8.3) containing 500 mM KCl, 15 mM MgCl₂, 200 μM deoxyribonucleoside triphosphate, 2.5 μM each primer, 50 ng of DNA, and 1 U of Taq DNA polymerase (Life Technologies, Inc., Rockville, Md.). Amplification was performed by using a DNA thermal cycler (Zaimoreactor II; ATTO, Tokyo, Japan) programmed for 5 min at 94°C followed by 35 cycles, each consisting of 30 s at 94°C, 30 s at 42°C, and 30 s at 72°C, and a final extension period of 5 min at 72°C. The amplification products were subjected to electrophoresis on 2.4% agarose gels and detected by staining with ethidium bromide, and the gels were photographed under UV illumination. The DNA from each isolate was subjected to the RAPD assay at least three times.

PFGE analysis.

B. cepacia isolates were cultivated at 35°C in heart infusion broth for 18 h with shaking (200 rpm). After culture, the medium was centrifuged for 10 min at 6,000 × g. The resulting pellets were washed twice with Pett IV (PIV) buffer (1 M NaCl–10 mM EDTA [pH 8.0]). After addition of an equal volume of PIV buffer containing 2% agarose (In Cert Agar; FMC BioProducts, Rockland, Maine), each mixture was poured into a plug mold and allowed to cool. Each plug was placed in 500 μl of cell lysis buffer (4% NaCl, 100 mM EDTA [pH 8.0], 10 mM Tris-HCl [pH 8.0], 0.2% deoxycholates, 0.5% Sarkosyl) containing proteinase K (final concentration, 1 mg/ml; Boehringer, Mannheim, Germany) and incubated overnight at 52°C. Then the plugs were washed four times with 1× TE buffer (10 mM Tris-HCl [pH 8.0]–1 mM EDTA [pH 8.0]) containing 1 mg of phenylmethylsulfonyl fluoride/ml and were digested overnight at 35°C by each being placed in 20 μl of an appropriate restriction buffer containing 20 U of XbaI (TaKaRa Shuzo Co., Kyoto, Japan). The plugs were then loaded into the wells of 1% agarose gels (Pulsed-Field Certified Agarose, Ultrapure DNA Grade Agarose; Bio-Rad, Richmond, Calif.) containing 0.5× TBE buffer (44.5 mM Tris, 44.5 mM boric acid, 1.25 mM EDTA [pH 8.0]), and the gels were processed with a CHEF-DR (Bio-Rad) with an initial time of 5 s and a final time of 20 s at 6.0 V/cm and a buffer temperature of 14°C, by using a ramp with an included angle of 120°. The λ phage marker was used as the standard. The gels were stained with ethidium bromide and photographed under UV illumination. The restriction enzyme DNA fragment patterns were inspected visually and were determined to be genetically similar when the electrophoretic mobility profiles of the DNA fragments were completely concordant and genetically dissimilar when the patterns differed by one or more DNA bands.

RESULTS

Bacterial strains.

A total of 30 B. cepacia isolates were obtained. Of these, 28 were obtained from clinical samples: 24 from sputum specimens from 12 patients and 1 each from a blood specimen, a central venous catheter tip specimen, a pus specimen, and a pharynx secretion specimen. From environmental samplings, one isolate was recovered from the nebulizer solution taken from a mixing bottle used for both patients H and I, and another was obtained from a tube of a nebulizer machine in the TCC (Table 1). However, B. cepacia was not isolated from the components of the solutions used in the TCC when samples from these components were separately cultured.

Phenotypic analysis.

The 28 isolates were classified into two biochemical groups, A1 (0467577) and A2 (0477577), according to their API profiles, and the two isolates obtained from environmental samples were classified as A1 (Table 1).

Genotypic analysis.

By RAPD analysis, two fingerprint patterns, R1 and R2, were observed among the 21 isolates taken from patients A to I between November 1995 and March 1996. All of these isolates, except for that from patient A (R1), showed the same pattern (R2) as the environmental isolates obtained from the nebulizer solution and a tube of a nebulizer machine in the TCC. Therefore, the R2 pattern was suspected to be characteristic of the epidemic-causing strain (R2) in the first outbreak. On the other hand, the RAPD profile of isolates from patient L, who stayed in the TCC a few months later, showed another pattern (R4). Four isolates from patients J and K all showed an identical RAPD pattern, R3, while an isolate from patient M showed another pattern (R5) (Table 1; Fig. 3 and 4). The DNA from each isolate was subjected to the RAPD assay at least three times, and the results were highly reproducible for these isolates.

FIG. 3.

DNA fingerprints of 30 B. cepacia isolates determined by the RAPD assay. Isolates were taken from patients A (lane 1), B (lanes 2 to 4), C (lanes 5 to 10), D (lane 11), E (lane 12), F (lane 13), G (lane 14), H (lanes 15 to 20), I (lane 21), J (lanes 22 to 24), K (lane 25), L (lanes 26 to 27), and M (lane 28) and from environmental sources (lanes 29 and 30). Leftmost lane, X174/HinfI digest used as a DNA size marker. DNA fingerprints of lanes 2 through 21, 29, and 30 indicate that the strain responsible for the nosocomial outbreak in the TCC has an R2 RAPD profile, while those of lanes 22 through 25 indicate that the strain responsible for the outbreak in the surgical ward has an R3 RAPD profile.

FIG. 4.

DNA fingerprints of B. cepacia isolates determined by PFGE assay. The lane arrangement of isolates is identical to that in Fig. 3. Lane M, phage marker. DNA fingerprints of lanes 2 through 21, 29, and 30 indicate that the strain responsible for the outbreak in the TCC has a P2 PFGE profile, while those of lanes 22 through 25 indicate that the strain responsible for the outbreak in the surgical ward has a P3 PFGE profile.

PFGE analysis of all 30 isolates revealed five different patterns (P1 through P5), and the profiles of the individual isolates corresponded precisely to the RAPD profiles (Table 1; Fig. 3 and 4).

DISCUSSION

Precise characterization of suspected strains from nosocomial outbreaks is essential to the determination and implementation of correct and prompt measures. Although several techniques based on serological and/or biochemical profiles have been widely used for phenotypic characterization of B. cepacia (4, 5, 10), these conventional methods often fail to provide confirmative data (16). In our study, for example, the API procedure enabled us to classify 28 clinical isolates into two biochemically distinct groups. However, the API patterns were not discriminative enough to unambiguously distinguish epidemic-causing strains from environmental strains and to also elucidate the routes of transmission among several wards.

In contrast to the API data, the RAPD assay divided the 28 isolates into five groups, each with a distinct RAPD pattern. Moreover, the RAPD assay revealed that all the B. cepacia isolates obtained between January and March 1996 exhibited identical RAPD profiles (i.e., R2) which were distinct from those of B. cepacia isolates obtained in November 1995 and also in July and September 1996. These RAPD data led us to suspect that isolates obtained between January and March 1996 were derived from a single source and enabled us to pinpoint the possible source of the nosocomial outbreak. Upon a review of medical records, it became clear that the patients from whom B. cepacia isolates with an R2 RAPD profile were obtained had overlapping hospitalization periods in the TCC. Furthermore, it became evident that these patients were under nebulized medication while they were cared for in the TCC. Therefore, environmental surveys of the nebulizer equipment as well as the nebulizer solution were carried out, revealing the presence of B. cepacia with RAPD profiles identical to those obtained from clinical samples. Since all the patients involved in the first outbreak were staying in the TCC, environmental surveys were performed in this area. Based on these clinical and environmental survey results, countermeasures to prevent further spread of the epidemic-causing strain were taken from 8 March onwards; these include the daily replacement of the mixing bottle together with the solution in it and the thorough disinfection of nebulizer devices, including the tubing, by formalin fumigation at least twice a week. Consequently, no B. cepacia isolates were detected for the subsequent 3 months, and as a result, no further environmental surveys involving other wards were carried out at that time. This indicates that nebulized medication was the most likely route and/or source of B. cepacia infection/colonization, although how the nebulizer tubing became contaminated at the beginning of this incident, i.e., whether the epidemic-causing strain was originally derived from one of the eight patients or, alternatively, was derived from an exogenous source, remains unknown.

Three months later, however, two consecutive B. cepacia infection/colonization cases were seen in the surgery ward. This time, another series of nosocomial B. cepacia isolates was suspected to be responsible for this second outbreak. Actually, two B. cepacia isolates obtained from these two patients demonstrated an identical RAPD profile R3, which differed from the profile of the epidemic-causing strain (R2). As a possible route of B. cepacia transmission in the second outbreak, indirect transmission, presumably via medical personnel, was mainly suspected, since no medical instruments, including respiratory-therapy machines, were shared by these two patients. Moreover, they stayed apart from each other within the same ward, although their hospitalization periods overlapped. We cannot exclude the possibility, however, that these two patients became infected with B. cepacia from a common, but unknown hospital source. Fortunately, there was no further isolation of B. cepacia. Thus, careful observation of the routine precautionary measures against the spread of infections successfully prevented further spread of the R3-type epidemic-causing strain, and accordingly, no further measures were taken to pinpoint the original source of B. cepacia in this second outbreak.

Based on our experience of two B. cepacia outbreaks in our hospital, it was clearly proven that the RAPD assay is a simple, time-saving technique useful in epidemiological investigations by clinical microbiology laboratories of nosocomial outbreaks (9). RAPD analysis of B. cepacia isolates, in fact, could provide bacterial genomic typing results within a single working day, whereas PFGE takes at least 4 days to produce data of almost identical clinical significance. Moreover, the RAPD assay was highly reproducible and sensitive enough to discriminate the clonal diversity of B. cepacia in tracing the source and the routes of the nosocomical outbreak. The RAPD assay also enabled us to process a large number of samples simultaneously, which may be essential in the event of a large outbreak. Although there have been some concerns about the discriminatory ability of the RAPD data compared with the PFGE data and also about the reproducibility and reliability of the RAPD assay in evaluating the clonal diversity of B. cepacia, we conclude that the RAPD assay is a potentially useful method for genotypic determination in clinical microbiology laboratories.