Carbapenem resistant Pseudomonas aeruginosa and Acinetobacter baumannii at Mulago Hospital in Kampala, Uganda (2007–2009)

Background Multidrug resistant Pseudomonas aeruginosa and Acinetobacter baumannii are common causes of health care associated infections worldwide. Carbapenems are effective against infections caused by multidrug resistant Gram-negative bacteria including Pseudomonas and Acinetobacter species. However, their use is threatened by the emergence of carbapenemase-producing strains. The aim of this study was to determine the prevalence of carbapenem-resistant P. aeruginosa and A. baumannii at Mulago Hospital in Kampala Uganda, and to establish whether the hospital environment harbors carbapenem-resistant Gram-negative rods. Results Between February 2007 and September 2009, a total of 869 clinical specimens were processed for culture and sensitivity testing yielding 42 (5 %) P. aeruginosa and 29 (3 %) A. baumannii isolates, of which 24 % (10/42) P. aeruginosa and 31 % (9/29) A. baumannii were carbapenem-resistant. Additionally, 80 samples from the hospital environment were randomly collected and similarly processed yielding 58 % (46/80) P. aeruginosa and 14 % (11/80) A. baumannii, of which 33 % (15/46) P. aeruginosa and 55 % (6/11) A. baumannii were carbapenem-resistant. The total number of isolates studied was 128. Carbapenemase genes detected were blaIMP-like (36 %, 9/25), blaVIM-like (32 %, 8/25), blaSPM-like (16 %, 4/25); blaNDM-1-like (4 %, 1/25) in carbapenem-resistant P. aeruginosa, and blaOXA-23-like (60 %, 9/15), blaOXA-24-like (7 %, 1/15), blaOXA-58-like (13 %, 2/15), and blaVIM-like (13 %, 2/15) in carbapenem-resistant A. baumannii. Furthermore, class 1 integrons were detected in 38 % (48/128) of P. aeruginosa and Acinetobacter, 37 % (26/71) of which were in clinical isolates and 39 % (22/57) in environment isolates. Gene cassettes were found in 25 % (12/48) of integron-positive isolates. These were aminoglycoside adenylyltransferase ant(4′)-IIb (3 isolates); trimethoprim-resistant dihydrofolate reductase dfrA (2 isolates); adenyltransferase aadAB (3 isolates); QacE delta1 multidrug exporter (2 isolates); quinolone resistance pentapeptide repeat protein qnr (1 isolate); and metallo-β-lactamase genes blaVIM-4-like, blaIMP-19-like, and blaIMP-26-like (1 isolate each). Gene cassettes were missing in 75 % (36/48) of the integron-positive isolates. Conclusions The prevalence of carbapenem-resistant P. aeruginosa and Acinetobacter among hospitalized patients at Mulago Hospital is low compared to rates from South-East Asia. However, it is high among isolates and in the environment, which is of concern given that the hospital environment is a potential source of infection for hospitalized patients and health care workers. Electronic supplementary material The online version of this article (doi:10.1186/s40064-016-2986-7) contains supplementary material, which is available to authorized users.


Background
Multidrug resistant Pseudomonas aeruginosa and Acinetobacter baumannii are common causes of nosocomial infections worldwide (Falagas et al. 2005;Giamarellos-Bourboulis et al. 2003;Gootz and Marra 2008;Tam et al. 2010;Turton et al. 2005). Carbapenems are the most effective drugs against infections caused by multidrug resistant Gram-negative bacteria including Pseudomonas and Acinetobacter species (Papp-Wallace et al. 2011;Manenzhe et al. 2015). However, their use in the management of these infections is threatened by the emergence of carbapenemase producing strains. Carbapenemases are β-lactamase enzymes with capacity to hydrolyze carbapenems, penicillins, and cephalosporins; they were first described in the 1990s and have continued to be reported every year with increasing rates (Papp-Wallace et al. 2011;Tenover 2006;Queenan and Bush 2007).
Carbapenemases are assigned to three of four classes of β-lactamases; Ambler classes A, B, and D that are differentiated based on the hydrolytic mechanisms at their active sites (Manenzhe et al. 2015;Queenan and Bush 2007). Class A and D carbapenemases are referred to as Serine carbapenemases because they have Serine (amino acid) at the active site (Serine-dependent), whereas class B carbapenemases have zinc (zinc-dependent) and are referred to as metallo-β-lactamases. Ambler class A carbapenemases can be plasmid encoded or chromosomal and are inhibited by clavulanic acid, a β-lactamase inhibitor; SME, IMI, NMC, GES, and KPC families are the most frequently identified class A carbapenemases mostly in Klebsiella pneumoniae. Class B metallo-β-lactamases are plasmid-encoded (in some cases chromosomal) and the most common enzymes include the VIM, IMP, SPM, GIM, SIM and NDM families. Metallo-β-lactamases have been detected primarily in P. aeruginosa but they are also increasingly being detected in Acinetobacter species (Tsakris et al. 2006) and the Enterobacteriaceae (Turton et al. 2005(Turton et al. , 2006Okoche et al. 2015); NDM producing Enterobacteriaceae are currently a major concern due to their rapid spread worldwide (Manenzhe et al. 2015). Furthermore, class B enzymes are able to hydrolyze β-lactams except aztreonam (a monobactam) and their hydrolytic activity is inhibited by EDTA (ethylenediammine tetra acetic acid), but not clavulanic acid. On the other hand, class D enzymes, also referred to as the OXA-type carbapenemases, are subdivided into five families; OXA-23, OXA-24/40, OXA-48 and OXA-58 families that are mainly plasmid-encoded (Queenan and Bush 2007), and OXA-51 that is chromosomally encoded and intrinsic (naturally found) in A. baumannii (Manenzhe et al. 2015;Turton et al. 2006). Though, OXA-51 confers resistance or reduced susceptibility to carbapenems only when its expression is up-regulated by genetic reorganization (SMI P 8 2014). Class D enzymes are not inhibited by clavulanic acid or EDTA.
Whereas carbapenemase-producing bacteria are well characterized in high-income countries, little is known about them in Uganda and Africa at large (Manenzhe et al. 2015). Metallo-β-lactamase producing bacteria from a tertiary-care center in Nairobi, Kenya, was characterized in 2008, in a first report of VIM-2-producing P. aeruginosa in East Africa (Pitout et al. 2008). Furthermore, in a systematic review of 83 surveillance studies on carbapenemases in Africa (Manenzhe et al. 2015), it was revealed that most studies were from 15 of the 54 countries in Africa, mainly in Northern and Southern Africa with no report of negative results in studies that screened for carbapenemase-producing bacteria in humans in hospitals (Manenzhe et al. 2015). Indeed, prior to 2010 there were only seven reports of carbapenemase-producing bacteria in Africa with OXA-58, OXA-48, OXA-23, VIM-2 and VIM-4 documented as prevalent in carbapenemase-producing bacteria in outbreaks (Manenzhe et al. 2015). Carbapenemase-producing bacteria elaborating OXA-23, OXA-24, OXA-58, VIM-2 and IMP-1 were also isolated from hospital environments (Manenzhe et al. 2015). Recently in Uganda, we characterized carbapenem-resistant Enterobacteriaceae isolated from patients admitted at Mulago National Referral Hospital in Kampala (Okoche et al. 2015).
The aim of this study was to determine the prevalence of carbapenem resistant P. aeruginosa and A. baumannii at Mulago National Referral Hospital, and to establish whether the Mulago Hospital environment harbors carbapenem resistant Gram-negative rods. Herein we describe the susceptibility patterns and carbapenemase genes harbored by the isolates, as it might help in the global comparisons of resistance mechanisms of multidrug resistant Gram-negative bacteria.

Study setting and design
This study was conducted at Mulago National Referral Hospital in Kampala, Uganda. Mulago is a 1500-bed tertiary hospital belonging to the Ministry of Health, Uganda. With its free medical care, the hospital is highly Keywords: Carbapenemase genes, Metallo-beta-lactamases, OXA-carbapenemases, Class 1 integrons, Hospital environment attractive for the peri-urban low-income population around the capital Kampala where the infectious disease burden is high. The laboratory procedures were performed in Clinical Microbiology and Molecular Biology Laboratories of the Department of Medical Microbiology, College of Health Sciences, Makerere University.
Between February 2007 and September 2009, carbapenem resistant P. aeruginosa and A. baumannii were isolated from hospitalized patients at Mulago Hospital, in a laboratory surveillance study that aimed to identify carbapenem resistant Gram-negative rods at Mulago Hospital. Within this period, 869 clinical specimens from hospitalized patients were processed by the Clinical Microbiology Laboratory for culture and sensitivity testing (one specimen per patient). Specimens processed were blood (51), cerebral spinal fluid (49), tracheal aspirates (163), ear swabs (197), sputum (204), urine catheters (98), and pus (107). Following detection of P. aeruginosa and Acinetobacter species in the specimens, 80 samples were randomly collected from the hospital environment (surgical/ medical wards including the intensive care units, ICUs); they included water (13), disinfectants like chlorhexidine gluconate (15), cleaning materials like mops and squeezers (15), sink swabs (22), and floor swabs (15). Water and disinfectants were sampled using sterile syringes with needles while swabs were used to sample sinks and wet floors.

Identification of P. aeruginosa and A. baumannii
All clinical and environment samples were processed within 2 h of collection for identification of Gram-negative bacteria. Isolates were recovered on blood agar after incubating at 35-37 °C for 24-48 h. Then, single colonies were sub-cultured on MacConkey agar and incubated at 35-37 °C for 24-48 h. Isolates were presumptively identified based on colony morphology, Gram-staining properties and biochemical characteristics [oxidase, triple sugar iron (TSI), sulphur indole and motility (SIM), citrate, and urease tests]. Colonial morphological features (i.e. colonies with characteristic spreading pattern and serrated edges, fruity sweet-grape smell, and bright green color) were used to identify Pseudomonas isolates. Positive catalase and oxidase tests, negative TSI and glucose fermentation tests and growth at 42 °C were used to distinguish P. aeruginosa from other lactose non-fermenting Gramnegative rods. Acinetobacter was presumptively identified based on negative motility and catalase tests, negative oxidase and glucose fermentation tests, and inability to grow under anaerobic conditions.

Identification of isolates to species level and drug susceptibility testing
To confirm A. baumannii to species level, PCR-amplification followed by DNA sequencing of the species-specific region of the bla OXA-51 -like gene intrinsic to A. baumannii (Manenzhe et al. 2015;Turton et al. 2006) was performed, using chromosomal DNA extracted from presumptuously identified isolates as template. To confirm P. aeruginosa to species level, and to determine the antimicrobial susceptibility profiles of both P. aeruginosa and A. baumannii, minimum inhibitory concentrations (MICs) were performed using the 'Phoenix Automated Microbiology System' from Becton and Dickson, Franklin Lakes, NJ, USA. This system has combination testing panels that include (a) an identification (ID) side with dried substrates for bacterial identification; the ID portion of the Phoenix panels utilizes a series of conventional, chromogenic, and fluorogenic biochemical tests to identify the organism, (b) an antimicrobial susceptibility testing (AST) side with varying concentrations of antimicrobial agents, and (c) growth and fluorescent controls at appropriate well locations. Specimen processing and Gram staining procedure was performed according to the manufacturer's guidelines.
Phoenix panels were inoculated with standardized inoculum according to the manufacturer's guidelines; occasionally, minor modifications were performed as described elsewhere (Carroll et al. 2006). Briefly, after determining the Gram staining properties of the isolates, nonselective medium (blood agar) was used to prepare fresh pure cultures for isolate identification (ID) and antimicrobial susceptibility testing (AST). Isolates were inoculated into appropriate ID/AST combination panels for Gram-negative isolates that were loaded into the instrument and incubated at 35 °C. The ID broth was inoculated with bacterial colonies adjusted to a 0.5 McFarland standard, and the suspension poured into the ID side of the Phoenix panel after a 30 µl aliquot was removed and saved for AST. For AST, the Phoenix AST Indicator Solution was added to the AST broth tubes and mixed by inversion. The AST side of the combination panel contains 84 wells with dried antimicrobial panels and one growth control well. One free-falling drop of the AST indicator was added to the AST broth tube, and 30 µl of the standardized ID broth suspension transferred to the AST broth and incubated for 16 h at 35 °C. Samples were read automatically at the instrument's set parameters.
Following isolate identification and AST, carbapenemsusceptible isolates were retested with the disc diffusion susceptibility method (10 µg imipenem or 10 µg meropenem, BiolabZrt, Budapest, Hungary) to detect isolates with inhibition zone diameters of ≤25 mm as Clinical and Laboratory Standards Institute (CLSI) recommends screening them for carbapenemase production (Wikler 2006).

Carbapenemase assays
To detect carbapenemase activity in carbapenem-resistant isolates, carbapenemase assays were performed with the modified Hodge test (MHT) and the imipenem-EDTA test using K. pneumoniae ATCC 700603 and E. coli ATCC 25922 as indicator strains (Okoche et al. 2015;Miriagou et al. 2010;Asthana et al. 2014). In the MHT assay, a 1:10 dilution of the indicator strains was made by diluting 0.5 ml of culture (at 0.5 McFarland) to 5 ml with sterile saline, which was streaked all over the Mueller-Hinton Agar (MHA) plate using a sterile swab. Then, 10 µg meropenem disk (BiolabZrt, Budapest, Hungary) was placed at the center of the MHA plate. Each test isolate was streaked in a straight line from the disk to the edge of the plate. K. pneumoniae ATCC BAA-1705 and K. pneumoniae ATCC BAA-1706 served as positive and negative controls, respectively. Strain BAA-1705 possesses a K. pneumoniae carbapenemase KPC-2 that is highly active against cephamycins, carbapenems, and to several extended spectrum beta-lactamases (ESBLs) (Broberg et al. 2013). Positive or negative results were interpreted according to the guidelines of CLSI and the UK Standards for Microbiology Investigations (SMI P 8 2014;Wikler 2006;Asthana et al. 2014).
To detect metallo-β-lactamase activity, the imipenem-EDTA double-disk synergy test was performed using an overnight liquid culture of the test isolate adjusted to a turbidity of 0.5 McFarland standard, and spread on the surface of MHA plates. Then, two discs with 10 µg imipenem each were placed on the agar 15 mm apart (centerto-center); 10 µl of 0.5 M EDTA was added to one of the imipenem disc to achieve a disc content of 1.5 mg. After incubating at 37 °C overnight, an increase in inhibition zone diameter of ≥5 mm in the EDTA-supplemented disc was interpreted as positive for metallo-β-lactamase production (SMI P 8 2014; Asthana et al. 2014).
Furthermore, carbapenem-susceptible isolates with disc inhibition zone diameters of ≤25 mm were also tested for carbapenemase activity (Wikler 2006); isolates with positive results were screened by PCR for carbapenemase genes.
Chromosomal DNA used as templates in PCRs was extracted by the cetyltrimethyl ammonium bromide (CTAB) method (Andreou 2013;William and Feil 2012) and dissolved in 100 µl of sterile Tris-EDTA (TE) buffer. PCRs were performed in 10 µl volumes with 100 ng DNA template, custom Master-mix (1×), 0.5 µM each of forward and reverse primer, and 1.25 U Taq DNA polymerase, in a Techne TC-412 thermal cycler (Techne, UK). 5 µl of the amplified PCR product was analyzed by electrophoresis on 1 % agarose gels at 120 constant voltage for 1 h. PCR products were cleaned with the QIAquick PCR-purification kit (Qiagen, Hilden, Germany) and shipped to the United States for sequencing (ACGT Inc., Wheeling IL). Although our focus on PCR was screening mainly carbapenem-resistant isolates, carbapenemsusceptible isolates with disc inhibition zone diameters of ≤25 mm were also screened; for class 1 integron gene cassettes, all isolates were screened by PCR.

Quality control
Negative controls for the PCR-amplified carbapenemase genes included reactions with only water (no DNA), and DNA template extracted from carbapenemase-negative strains K. pneumoniae DSMZ 9377, E. coli ATCC 25922 and P. aeruginosa ATCC 27853. Positive control reactions included template DNA extracted from carbapenemaseproducing strains (K. pneumoniae Nr.8 for bla NDM-1 , K. pneumoniae 714 for bla  ) and a previously characterized P. aeruginosa clinical strain from Giessen for bla IMP that was obtained from the Institute of Microbiology, Giessen, Germany [see Mushi et al. (2014)]. For bla VIM , the positive control strain was obtained from the RESET research collaboration, Germany [see Fischer et al. (2012)]. Additionally, targeted DNA sequencing of PCR-products and confirmation of sequenced amplicons through BLAST-searching at National Center for Biotechnology Information (NCBI) was performed. Isolates with sequenced amplicons that did not match sequences for the genes being studied were excluded.
Other genes detected in integrons were the putative glucose dehydrogenase precursor and hypothetical genes common to A. baumannii class 1 integrons (3 isolates) and non-ribosomal peptide synthetase (pyoverdine sidechain peptide synthetase) (1 isolate) in P. aeruginosa, Additional file 1: Table S1.

Discussion
In this study, we have described carbapenem-resistant P. aeruginosa and A. baumannii isolated from hospitalized patients and the environment at Mulago Hospital in Kampala, Uganda. As species identification is highly desirable to allow proper interpretation of the results (SMI P 8 2014), the isolates were successfully identified to species level using a rigorous methodology. While the isolates from hospitalized patients were from clinically relevant specimens referred to the diagnostic laboratory for culture and sensitivity testing, we could not rule-out colonization implying that some of the isolates might have been not clinically relevant, given the high rate of  and Acinetobacter species at Mulago National Referral Hospital in Kampala, 2008-2009   The prevalence of carbapenem resistant strains was relatively low among hospitalized patients; however, it was comparatively high in isolates particularly in the environment. This is of concern as the hospital environment is known to be a potential source of infection for hospitalized patients and health care workers. Furthermore, most isolates in this study were multidrug resistant with rates (i.e. 65 % overall; 73 % hospitalized patients, 54 % environment) comparable to those reported by Pitout et al. in Kenya (Pitout et al. 2008), Lee et al. in Korea (Lee et al. 2007) and Gu et al. in China (Gu et al. 2007), but lower than rates reported from Latin America (Labarca et al. 2016) and India (Uma Karthika et al. 2009). This could be due to differences in antibiotic usage between Uganda and the countries where these studies were done (Manikal et al. 2000).
It is important to note that carbapenemases are not the only mechanisms of acquired resistance to carbapenems; other resistance mechanisms in P. aeruginosa include upregulated efflux pumps and loss of the outermembrane protein encoding gene oprD (SMI P 8 2014). As such, when screening for carbapenemases, two confounders have to be ruled-out (SMI P 8 2014); (a) not all carbapenem-resistant isolates produce a carbapenemase, (b) not all carbapenemase producers are resistant to carbapenems. In this study we did not screen for carbapenemases in carbapenem-susceptible isolates with disc inhibition zone diameters of >25 mm. Furthermore, while the automated systems efficiently detect carbapenem resistance, the inbuilt software in these systems is not always accurate at correctly inferring the presence of carbapenemases. The Phoenix BD expert system that we used is noted for its high sensitivity (ability to detect carbapenem resistance) but on the other hand, low specificity (ability to distinguish true carbapenemase producers) (SMI P 8 2014; Woodford et al. 2010). Hence, carbapenem-resistant isolates in this study were screened for carbapenemase production with the MHT and imipenem/EDTA test that have been effectively used to validate carbapenemase producers. These tests also can distinguish carbepenem-resistance mediated by carbapenemases from the one mediated by other mechanisms (SMI P 8 2014; Asthana et al. 2014). Our data showed that carbapeneme-resistance in this study particularly in P. aeruginosa, was mainly mediated by metallo-βlactamases. Nevertheless, the reported low rates of carbapenemase activity (particularly with MHT) among carbapenem-resistant isolates could be due to the fact that the detection of carbapenemase activity in clinical isolates is challenging (Queenan and Bush 2007); the MHT assay suffers from low sensitivity, and interpretation of its results can be subjective (i.e. the identification of the clover-leaf indentation).
Furthermore, carbapenem resistant isolates (both P. aeruginosa and A. baumannii with carbapenemase genes), which lacked carbapenemase activity were also detected. As outlined above, this might reflect the difficulty in detecting carbapenemase production particularly in Acinetobacter (SMI P 8 2014); OXA-carbapenemases often have poor enzymatic activity leading to sub-optimal activity in some strains. Furthermore, carbapenem resistant isolates without carbapenemase genes and carbapenemase activity detected in this study alludes to the occurrence of additional non-carbapenemase resistance mechanisms in these isolates (outlined earlier) particularly in P. aeruginosa [e.g. upregulated efflux pumps, oprD loss (SMI P 8 2014)]. Non-carbapenemase mediated mechanisms need further study as we did not extensively characterize them in this study.