High prevalence of methicillin resistant Staphylococcus aureus in the surgical units of Mulago hospital in Kampala, Uganda

Background There is limited data on Methicillin resistant Staphylococcus aureus (MRSA) in Uganda where, as in most low income countries, the routine use of chromogenic agar for MRSA detection is not affordable. We aimed to determine MRSA prevalence among patients, healthcare workers (HCW) and the environment in the burns units at Mulago hospital, and compare the performance of CHROMagar with oxacillin for detection of MRSA. Results One hundred samples (from 25 patients; 36 HCW; and 39 from the environment, one sample per person/item) were cultured for the isolation of Staphylococcus aureus. Forty one S. aureus isolates were recovered from 13 patients, 13 HCW and 15 from the environment, all of which were oxacillin resistant and mecA/femA/nuc-positive. MRSA prevalence was 46% (41/89) among patients, HCW and the environment, and 100% (41/41) among the isolates. For CHROMagar, MRSA prevalence was 29% (26/89) among patients, HCW and the environment, and 63% (26/41) among the isolates. There was high prevalence of multidrug resistant isolates, which concomitantly possessed virulence and antimicrobial resistance determinants, notably biofilms, hemolysins, toxin and ica genes. One isolate positive for all determinants possessed the bhp homologue which encodes the biofilm associated protein (BAP), a rare finding in human isolates. SCCmec type I was the most common at 54% prevalence (22/41), followed by SCCmec type V (15%, 6/41) and SCCmec type IV (7%, 3/41). SCCmec types II and III were not detected and 10 isolates (24%) were non-typeable. Conclusions Hyper-virulent methicillin resistant Staphylococcus aureus is prevalent in the burns unit of Mulago hospital.


Background
There is a global outbreak of Methicillin resistant Staphylococcus aureus (MRSA) infections, particularly in industrialized countries [1]. However, there is limited data on MRSA prevalence in Uganda, where~10% of the surgical procedures become septic with S. aureus being the most frequent pathogen isolated [2][3][4]. Ojulong et al. 2009 determined the prevalence of MRSA in patients with post-operative surgical wound infections in the surgical wards of Mulago hospital in Kampala, Uganda [5], but the distribution of isolates among healthcare workers (HCW) and the environment was not determined. Here, we aimed to determine MRSA prevalence among patients, HCW and the environment in the burns units at Mulago hospital, and compare the performance of CHROMagar with oxacillin for detection of MRSA (since the routine use of chromogenic agar is usually not affordable in most low income settings).

Study setting and sampling
This cross-sectional study was performed during November, 2009 , beds and table surfaces) were swabbed, as well as air samples on settle plates (with blood agar). For patients, wounds or nostrils were swabbed while for HCW, hands or nostrils were swabbed. One hundred samples (one sample per person or item) from 25 patients, 36 HCW and 39 items (including air settle plates) were obtained using sterile swabs and transported to the laboratory for culture in tubes containing Amies transport medium (Biolab, Budapest, Hungary).

Cultures and drug susceptibility testing (DST)
Swabs were cultured on blood agar and Mannitol salt agar at 37°C for 24 hours, ensuring growth of distinct colonies. The predominant colony type per sample was selected, and S. aureus was identified to species level microbiologically as previously described [6]. Drug susceptibility testing (erythromycin, 15 μg; vancomycin, 30 μg; gentamicin, 10 μg; oxacillin, 1 μg; tetracycline, 30 μg; chloramphenicol, 30 μg; penicillin, 10 units; and sulphamethoxazole-trimethoprim, 1.25/23.15 μg) was performed with the disc diffusion method (Biolab inc, Budapest, Hungary). Isolates resistant to erythromycin were screened for the macrolide-lincosamide and streptrogramin B (MLS B ) phenotype. MRSA was identified based on oxacillin resistance following standard procedures [7], and confirmed by PCR detection of the mecA gene. For comparison, oxacillin resistant-mecA positive isolates were further screened for growth on CHROMagar (BD diagnostics, Sparks, USA).

SCCmec genotyping
Staphylococcus Cassette Chromosomal mec (SCCmec) genotyping was performed in a multiplex PCR protocol described by Boye et al, 2007 (see additional file 1 for primers and PCR conditions) [8]. The SCCmec types were determined based on the banding patterns upon agarose gel electrophoresis of the amplicons [8]. Isolates with no visible bands were classified as non-typeable [8].

Detection of virulence determinants Biofilms
Biofilms were detected with the microtiter plate method [9] and the biofilm unit calculated according to Amaral et al. [10]. Briefly, assays were performed in triplicates in tryptic soya broth (TSB) with 1% glucose in 96-well polystyrene flat-bottom tissue culture plates. Isolates were incubated at 37°C overnight with gentle shaking and standardized to OD 600 = 0.005 with normal saline. Then, 50 μl of standardized cells mixed with 150 μl TSB/1% glucose were incubated at 37°C for 17 hr. After washing three times with sterile water and staining with crystal violet for 15 minutes, cells were washed again and incubated at room temperature for 1 hr in 95% ethanol. Then, biofilms were measured with a spectrophotometer at OD 570 . The biofilm unit (BU) was calculated using negative control values with the formula A 1 / A 2 , where A 1 is the test value while A 2 is the negative control value. Isolates with BU > 2× the negative control value were considered biofilm producers and were classified as: weak, 0.182 < BU < 0.364; moderate, 0.364 < BU < 0.728; strong, BU > 0.728 [10]. The biofilm forming S. epidermidis RP62A and its non-biofilm forming variant (ATCC 12228) were used as controls.

Discussion
In this study, high prevalence of MDR-MRSA was found in the burns unit of Mulago hospital, predisposing patients to infection with intractable isolates and underscoring the need for improved infection control practices in this setting. Ojulong et al 2009, reported a relatively lower prevalence (31.6%) from the general surgery ward, possibly because this earlier study determined MRSA infections in only patients with post-operative surgical wound infections [5]. Although data are still limited, there are emerging reports of prevalent MRSA infections in sub-Saharan Africa [19].
CHROMagar, generally considered more efficient at detecting MRSA than oxacillin discs [20] was inefficient in this setting (i.e. 46% vs. 29% prevalence, respectively). Since all isolates were oxacillin resistant, mecA-positive (mecA encodes the penicillin binding protein 2a, the molecular determinant for methicillin resistance [21]), nucand femA-positive, they were confirmed as MRSA. Prior to use, the CHROMagar batch passed quality control screening with known MRSA and MSSA strains; yet only 26 isolates grew mauve colonies (which is indicative of MRSA); the use CHROMagar in this setting may need further investigations.
A high prevalence of biofilm/ica-positive isolates correlated with that of MDR isolates. Although biofilms/ica genes are debatable as virulence markers [22], in this study, the biofilm/ica-positive isolates were concomitantly positive for other virulence genes. Notably was the absolute prevalence for staphylococcal hemolysins: hla and hld. The hla gene encodes a dermanecrotic and neurotoxic toxin that is also responsible for abscess formation; hld producing S. aureus can cause severe enteritis, while hlg lyses mammalian red blood cells and together with tst-1 (toxic shock syndrome toxin-1), can be involved in the pathogenesis of toxic shock syndrome (TSS) [15]. PVL, also prevalent in this study, causes severe disease in children and young adults with no known exposure to healthcare establishment, and is used as a stable marker for community acquired MRSA [15]. Furthermore, the staphylococcal super-antigenic toxins, sea and tst-1 were also detected (85% and 10% prevalence respectively). sea producing strains are responsible for staphylococcal food intoxications, while tst strains produce antigens that are responsible for TSS.

Conclusion
Hypervirulent methicillin resistant S. aureus is prevalent in the burns unit of Mulago hospital.