Acinetobacter baumannii has emerged worldwide as an important nosocomial pathogen.1 The carbapenems are highly active against A. baumannii, but their activity has been threatened by the widespread of Class D carbapenemases, mainly OXA-23, which is highly prevalent in Brazil.2

Treatment of patients infected by carbapenem-resistant A. baumannii (CRAB) has been very difficult and therapy is usually based on polymyxins.3 The use of poymyxins has been guided by the isolation of the organism and detection of carbapenem resistance. However, final results by conventional microbiological methods usually take two to three days, which consequently results in a delay in commencing appropriate therapy for CRAB, which may potentially contribute to adverse outcomes in patients with severe infections such as ventilator-associated pneumonia (VAP).4

In this study, we aimed to develop a SYBR green-based real time PCR (qPCR) to detect the blaOXA-23 gene directly from respiratory tract specimens in patients under mechanical ventilation, potentially serving as a surrogate marker for the presence of CRAB in those patients.

Twenty samples of endotracheal aspirates (ETAs), one specimen per patient, obtained from August to December 2010 from patients under mechanical ventilation at Hospital São Lucas, Porto Alegre, Brazil, were analyzed. All samples were routinely cultivated and only the growth of ≥105CFU/mL of A. baumannii was considered positive in ETA culture.

DNA extraction from 200μL of the bacterial suspension (0.5 McFarland standard) and from ∼400μL of ETA samples were carried out using the Invitek RTP DNA/RNA Virus Mini Kit according to the protocol suggested by the manufacturer.

Two μL of the extracted DNA were used in a total volume of 25μL of PCR mix containing SYBR Green II (Invitrogen). For primer design, multiple nucleotide sequences were aligned to search for conserved regions within the blaOXA-23 gene. The primers used to amplify the blaOXA-23 gene were (F: 5′-AAA GAA GTA AAA CGT ATT GGT TTC G-3′ and R: 5′-CCC AAC CAG TCT TTC CAA AA-3′), which correspond to nucleotides 518–442 (sense) and 724–743 from A. baumannii (accession number AY795964.1). qPCR was performed using the LightCycler platform (Roche, Germany) and confirmation of the amplicon was done by melting curve analysis (TM∼80°C). The specificities of the primers for the detection of blaOXA-23 gene were evaluated by the BLAST search program. No matches to the primers were found other than blaOXA-23. Negative (0.9% sodium chloride) and positive controls (an OXA-23 producing A. baumannii strain confirmed by direct sequencing analysis) were used in each qPCR.

Of 20 ETAs, CRAB isolates were recovered in 10, carbapenem-susceptible A. baumannii (CSAB) in three, and seven cultures were negative to A. baumannii. The results of blaOXA-23 qPCR in both colony and ETA are shown in Table 1.

The number of samples analyzed did not allow for a thorough sensitivity and specificity evaluation of the qPCR in ETA, but concordant result rates between both culture and qPCR of ETA and qPCR of the colony and qPCR of ETA were considered satisfactory. Nonetheless, a few disagreements between results were observed. One culture positive specimen for CRAB presented negative results for blaOXA-23 in both qPCR of the colony and ETA. It is possible that carbapenem-resistance in such isolate was due to other mechanisms (other carbapenemases or by a combination of non-carbapenemases enzymes plus porin modification and/or efflux pump hyperexpression). A thorough investigation of other resistance mechanisms to carbapenems was beyond the scope of our study. One CRAB and one CSAB presented negative result for blaOXA-23 in qPCR of the colony, but showed positive qPCR in ETA. Although these may be considered false-positive results, we cannot rule out the real presence of a subpopulation of A. baumannii harboring the blaOXA-23 gene in the samples, which could not be detected by the phenotypic method used to isolate this pathogen from ETA. Although the presence of CRAB in lower numbers may be simply considered colonization, not considering growth <105CFU/mL in our study impaired the interpretation of these discordant results.

Overall, we believe that our findings support the potential use of qPCR to detect the presence of blaOXA-23 gene directly from ETAs, and possibly the detection of other carbapenemase-encoding genes, such as blaOXA-143 gene for example, which has been shown to be highly prevalent in some Brazilian Southeastern cities,5 as surrogate marker for the presence or absence of CRAB in patients under mechanical ventilation. The final results can be available in a much shorter period (one working day) when compared to conventional bacterial culture and susceptibility tests results (may take up to 72h or more), consequently saving time to identify CRAB in patients with proved or suspected VAP, and potentially shortening the time for initiating appropriate therapy for this difficult to treat infections.