Quantitative Real-Time PCR for Detection of Acinetobacter baumannii Colonization in the Hospital Environment

Abstract

A real-time PCR assay was developed for detecting the presence of Acinetobacter baumannii on hospital equipment and compared to conventional bacterial culture using 100 hospital environmental samples. The real-time PCR detected contaminated surfaces in 4 h with high sensitivity (100%) compared to conventional culture. Thirty-eight percent of samples were positive by real-time PCR and negative by bacterial culture (false positives), possibly indicating the widespread presence of bacterial DNA that is not associated with viable bacteria.

TEXT

Nosocomial infections caused by drug-resistant bacteria represent an important clinical challenge. Acinetobacter baumannii has become one of the most problematic causative agents of nosocomial infections due to its remarkable ability to survive on hospital surfaces and acquire antibiotic resistance, resulting in the global emergence of multidrug-resistant strains with resistance to multiple antibiotic classes (5). A. baumannii has been especially problematic in critically ill patients in the intensive care setting, as it is an important cause of ventilator-associated pneumonia and bacteremia. In this context, patients are exposed to A. baumannii via contact with contaminated hospital equipment or by contact with hospital personnel carrying the bacteria. A number of studies have demonstrated widespread contamination with A. baumannii on hospital environmental surfaces, most notably in intensive care units (ICUs) (1, 4, 8, 9).

Environmental surveillance protocols have been employed for the identification of hospital equipment colonized by A. baumannii so that appropriate decontamination procedures can be carried out (1, 4, 8, 9). Since these surveillance methods employ conventional bacterial culture to determine the presence of A. baumannii, definitive species identification can require between 24 and 48 h. Nucleic acid-based tests, such as real-time PCR, have been employed for the identification of numerous bacterial pathogens (2); however, to our knowledge this technique has not been applied to identifying contaminated hospital equipment. The objective of the present study was to develop a real-time PCR for identifying hospital surfaces colonized by A. baumannii.

A real-time PCR assay was developed using TaqMan chemistry for the amplification of nucleotides 774 to 859 of the outer membrane protein A gene (ompA; accession number AY485227). The ompA gene was chosen because it is present in all sequenced genomes of A. baumannii available in the public domain (as of March 2010), and the sequences chosen for the primers and probe correspond to regions highly conserved between published A. baumannii ompA sequences (100% sequence identity). The primers OmpA Forward (5′-TCTTGGTGGTCACTTGAAGC-3′) and Ompa Reverse (5′-ACTCTTGTGGTTGTGGAGCA-3′) and the probe (5′-AAGTTGCTCCAGTTGAACCAACTCCA-3′), 5′ labeled with 6-carboxyfluorescein and the 3′ labeled with 6-carboxytetramethylrhodamine, were used. A quantification standard, pGEM-ompA, was constructed by inserting the ompA gene from the ATCC 19606 strain of A. baumannii into the pGEM-T Easy vector (Promega) after amplification with the primers 5′-ACAGGATCCATGAAATTGAGTCGTATT-3′ and 5′-ACAGGGCCCTTATTGAGCTGCTGCA-3′. Each 50-μl reaction mix consisted of 25 μl of the 2× TaqMan Universal PCR master mix (Applied Biosystems), 10 μl of DNA (sample or quantification control), OmpA Forward and OmpA Reverse primers at a concentration of 300 nM each, and the probe at a concentration of 100 nM. PCR parameters were 50°C for 2 min, 95°C for 10 min, and then 38 cycles at 95°C for 30 s and 62°C for 1 min. All assays were carried out on a Stratagene Mx3005P thermal cycler. Assay characteristics, including reaction efficiency, dynamic range, intra- and interassay variability, and limit of detection, were determined as described previously (7). The sensitivity of the assay for detecting genomic DNA from diverse A. baumannii strains was determined using purified genomic DNA (QIAamp DNA minikit; Qiagen) from 20 clonally distinct clinical isolates, as determined by pulsed-field gel electrophoresis or repetitive element PCR (REP-PCR). The specificity of the assay for A. baumannii was determined using genomic DNA from a clinical isolate of each of the following species: Pseudomonas aeruginosa, Klebsiella pneumoniae, Moraxella catarrhalis, and Escherichia coli. In all assays, three concentrations of the quantification standard were used (6.2 × 10⁶, 6.2 × 10³, and 6.2 ×10¹) to determine the number of genome copies present in unknown samples.

A total of 100 environmental samples were collected from the general ICU (n = 50) and the trauma/neurosurgical ICU (n = 50) at the Virgen del Rocío University Hospital on separate days in June 2011. Each surface was sampled in duplicate using sterile swabs moistened with physiologic saline. One swab from each surface was used to detect A. baumannii using conventional culture methods previously described by our group (6). Briefly, the swab was placed in 1 ml of Luria-Bertani medium and incubated for 24 h at 37°C to enrich bacteria present in the sample. One hundred microliters from each enrichment culture was plated on Leeds Acinetobacter medium (LAM; Hardy Diagnostics) (3) and incubated at 37°C for 24 h to select for A. baumannii. The definitive identification of bacteria that grew on LAM plates was made by matrix-assisted laser desorption ionization–time-of-flight (MALDI-TOF) analysis using a MALDI biotyper (Bruker Daltonics). To detect and quantify A. baumannii genomic DNA using real-time PCR, the second swab from each surface was placed in 1 ml physiologic saline and vortexed vigorously, and the DNA from 200 μl of the sample was extracted using the QIAamp DNA minikit and eluted in 200 μl water. Ten microliters of the extracted DNA was used in the real-time assay as described above. The number of genome copies present in the sample was determined by extrapolation from the quantification standards performed in parallel. Negative controls consisted of testing the eluant from unused swabs treated using the DNA isolation procedure described above. Samples were considered positive if a threshold cycle was reached during the 38 cycles.

The analytical characteristics of the ompA real-time PCR were determined using purified genomic DNA from the ATCC 19606 strain and the pGEM-ompA plasmid. Amplification was linear over 9 log dilutions of the plasmid pGEM-ompa (r² = 0.991; slope = −3.58), and the amplification efficiency was 0.90. Intra-assay and interassay variabilities using 7,200 copies of genomic DNA were 0.8 and 1.32%, respectively. The limit of detection of the assay was 6.8 copies of genomic DNA, as this quantity could reproducibly be amplified. The assay was able to amplify 20 clonally distinct clinical isolates of A. baumannii, indicating that diverse A. baumannii strains could be detected. The real-time assay was negative when genomic DNA from clinical isolates of Pseudomonas aeruginosa, Klebsiella pneumoniae, Moraxella catarrhalis, and Escherichia coli was used in the assay. Taken together, these results indicate that the ompA real-time assay is sensitive and specific for detecting and quantifying A. baumannii genomic DNA.

Results from environmental samples in the ICUs demonstrated that A. baumannii was identified on 39% of surfaces using bacterial culture (Table 1), a prevalence similar to those of previous reports describing the presence of A. baumannii on environmental surfaces in intensive care settings in which this species is endemic (1, 8). Duplicate samples tested using real-time PCR showed the presence of A. baumannii on 77% of surfaces, a significantly higher level than the results obtained with bacterial culture (P < 0.0001 by chi-squared test). Data from the general ICU demonstrated 27 culture-positive and 43 PCR-positive samples, whereas data from the trauma/neurosurgical ICU demonstrated 11 culture-positive and 34 PCR-positive samples. The quantification of the number of genome copies present in samples taken from environmental surfaces demonstrated a wide range, with between 744 and 189,131 copies present and a median of 19,696. Samples from surfaces positive for A. baumannii (n = 39) by bacterial culture had significantly more genome copies (median [interquartile range], 27,851 [12,950 to 63,324]) than samples from the 38 negative surfaces (13,798 [2,185 to 25,092]; P = 0.002 by Mann-Whitney U test). Importantly, definitive results could be obtained in 4 h using the real-time PCR assay versus 48 h using bacterial culture. Compared to conventional culture, the real-time PCR assay demonstrated a sensitivity of 100%, a specificity of 37.7%, a positive predictive value of 50.6%, and a negative predictive value of 100%.

Table 1.

Presence of Acinetobacter baumannii on intensive care unit environmental surfaces using bacterial culture and real-time PCR

Surface	No. of samples	No. (%) with growth on LAMa	No. (%) positive by real-time PCR	Copies of genomic DNA (median [range])
Bed rail	13	9 (69.2)	12 (92.3)	29,110 (9,794–123,531)
Bedside table	13	4 (30.8)	9 (69.2)	19,583 (10,172–98,010)
AHRDb	12	6 (50.0)	10 (83.3)	20,031 (6,351–66,375)
Intravenous (i.v.) pole/i.v. valves	12	3 (25.0)	9 (75.0)	32,157 (1,616–55,969)
Bedside chairs	7	5 (71.4)	7 (100.0)	30,137 (1,176–126,423)
Equipment carts	6	1 (16.7)	4 (66.7)	41,267 (816–48,314)
Infusion pumps	5	3 (60.0)	4 (80.0)	27,890 (3,346–65,835)
Folders	5	2 (40.0)	3 (60.0)	1,842 (1,781–10,814)
Doorknobs	5	1 (20.0)	4 (80.0)	18,377 (6,847–24,300)
Keyboards	4	2 (50.0)	4 (100.0)	7,249 (2,191–189,131)
Storage cabinets	3	1 (33.3)	2 (66.7)	7,314 (3,005–11,622)
Nurses' station	3	1 (33.3)	1 (33.3)	17,775
Sink	3	0 (0.0)	2 (66.7)	2,372 (1,792–2,951)
Light switch	2	0 (0.0)	1 (50.0)	1,261
Air conditioning grate	2	0 (0.0)	2 (100.0)	3,965 (1,858–6,071)
Ambubag	2	0 (0.0)	0 (0.0)
Dialysis module	1	1 (100.0)	1 (100.0)	93,920
Phone	1	0 (0.0)	1 (100.0)	744
Ultrasound equipment	1	0 (0.0)	1 (100.0)	17,597
Total	100	39 (39.0)	77 (77.0)	19,696 (744–189,131)

^aLeeds Acinetobacter medium.

^bAlcohol-based hand rub dispensers.

Our results indicate that surfaces colonized by A. baumannii can be rapidly identified using real-time PCR with high sensitivity (100%). However, we observed a high frequency of samples that are negative by conventional bacterial culture but positive by real-time PCR. One possibility is that some Acinetobacter strains that are detected by real-time PCR are not able to grow on Leeds Acinetobacter medium. A second possibility explaining this difference is that bacterial culture detects viable bacteria, whereas the real-time PCR measures the presence of genomic DNA in the sample. We show that 38% of surfaces contained A. baumannii DNA in the absence of detectable viable A. baumannii, suggesting the widespread presence of genomic DNA that is not associated with viable bacteria. We hypothesize that the presence of this DNA results from decontamination procedures that effectively kill viable bacteria but do not completely remove bacterial remains from the decontaminated surface. Interestingly, although a number of studies have detected the presence of viable A. baumannii in the hospital setting (1, 8, 9), no study has characterized the presence of free bacterial genomic DNA in this environment. Further study is required to determine if the presence of free bacterial genomic DNA in the hospital environment is of clinical importance.