Antimicrobial susceptibility pattern of Staphylococcus aureus isolates from clinical specimens at Kenyatta National Hospital

Objective To determine antibiotic susceptibility pattern of S. aureus isolates from clinical specimens collected from patients at Kenyatta National Hospital from March 2014–February 2016, and to determine the prevalence and quarterly trends of MRSA throughout the study period. Results A total of 944 S. aureus isolates were analyzed. High sensitivity of S. aureus was observed for quinupristin/dalfopristin (100%), tigecycline (98.2), imipenem (98%), nitrofurantoin (97.6%), linezolid (97.3%), teicoplanin (97.1%) and vancomycin (95.1%). High resistance was recorded against penicillin G (91.9%), trimethoprim/sulfamethoxazole (56.9%) and tetracycline (33.2%). MRSA prevalence among the patients at KNH was 27.8%. Highest proportion (80%) of MRSA was in burns unit. Both MRSA and MSSA were highly susceptible to quinupristin/dalfopristin, tigecycline, linezolid, nitrofurantoin, ampicillin/sulbactam and vancomycin and showed high resistance to commonly used antibiotics such as gentamycin, erythromycin, levofloxacin and tetracycline. A majority of isolates were from pus specimen (68%).


Introduction
Staphylococcus aureus is a Gram-positive bacterium living as a commensal on the skin, mouth and upper respiratory system, making it a risk factor for opportunistic and nosocomial infections. Resistance to commonly used antimicrobial drugs is frequently encountered with S. aureus. Some of the mechanisms in resistance include; inactivation of antibiotics by the enzymes, decreased affinity for the antibiotics caused by alteration of the target, efflux pumps, and trapping of the antibiotic [1].
Staphylococcus aureus causes skin, bone, soft tissue infections, urinary tract infections, pneumonia, healthcare-associated bacteremia in community and hospital settings and other invasive infections. Multidrug-resistant strains particularly Multidrug Resistant Staphylococcus aureus (MRSA) strains are common causes of nosocomial infections and are associated with increased morbidity and mortality [2,3].
Report from the National Nosocomial Infections Surveillance System of the Centers for Disease Control andPrevention, (2013-2015) showed that MRSA in India and USA accounts for > 60% of S. aureus isolates causing nosocomial infection in ICUs [4].
There is no conclusive local data on the magnitude of multidrug resistance MRSA infection burden in Kenyan hospitals. In study done from patients in select hospitals in Nairobi, MRSA infections were mostly isolated at public healthcare facilities serving economically disadvantaged Nairobi's population, like those living in urban informal settlements [5].
Another study done in Kenya showed there was high number of genetically indistinguishable isolates, which suggested there was local transmission of MSSA and MRSA [6].
Multidrug-resistant strains limit the therapeutic options, creating an economic and social burden to the healthcare system. Horizontal gene transfer in the hospital setting is responsible for disseminating antibiotic resistant determinants. Chromosomal mutation antibiotics selection is also responsible for antibiotics resistance.
The aim of this study was to evaluate the antimicrobial susceptibility pattern of Staphylococcus aureus at KNH from April 2014 to February 2016. Previous studies done in Kenya were targeting specific locations in the hospital and were within short periods of time. This study gives a broader picture of the susceptibility pattern of S aureus in different locations in 3 years.

Methodology
This was a retrospective study based on electronic laboratory records of Staphylococcus aureus isolates from clinical specimens analyzed at KNH microbiology laboratory.
Data was retrieved from VITEK-2 Antibiotic Susceptibility Testing System, imported to WHONET software through BACLINK (World Health Organization). Analysis was done using WHONET and IBM SPSS V.21. VITEK 2 Gram Positive identification card (bioMerieux) was used to identify S. aureus sub-species aureus. All the isolates had an ID confidence of excellent identification with an average percent probability of 96%.
Methicillin resistance was determined using cefoxitin screening.
All isolates were from patients' clinical specimens (mainly pus, urine, blood and tracheal aspirates) and were analyzed according to the 2015 Clinical & Laboratory Standards Institute (CLSI M100-S25) standards.
The sample size was calculated using Fishers formula [7] Univariate analysis was done using frequency distributions and proportions for categorical variance such as antimicrobial resistance and gender.
Bivariate analysis was done by use of Chi square to assess for association between categorical variables such as susceptibility of antibiotics and infections location. Data was presented in tables and graphs.
As shown in Table 1, isolates from all the specimens showed 100% sensitivity to quinupristin/dalfopristin. Isolates from HVS showed high resistance to tetracycline (41%), erythromycin (34%) and clindamycin (19%), no resistance was recorded from isolates from CSF to clindamycin.

MRSA distribution by year
The overall prevalence of MRSA was 27.8%. MRSA prevalence for 2014 was 34.9% (103/298) and 2015 recorded 25.8% (124/481). 2016 had MRSA prevalence of 21.7% representing samples for 2 months (January and February). There was no significant difference between MRSA isolation and the year of isolation (P = 0.159).

MRSA Distribution by specimen
The highest prevalence of MRSA was observed in pus specimen 153/232 (66%).

Discussion
The aim of this study was to evaluate the antimicrobial susceptibility pattern of S. aureus from clinical specimens at Kenyatta National Hospital. In our study, most S. aureus strains (68%) were isolated from pus specimen. This is consistent with a previous study done at KNH [8]. In a study done to determine the antimicrobial susceptibility pattern of S. aureus strains isolated from hospitalized patients in Iran, most of the isolates were from blood specimens (29%) [9]. Another study done on prevalence and antibiotic susceptibility pattern of S. aureus from clinical isolates in Nigeria showed a majority of the isolates were from urine specimens (76%) [10].
The high number of S. aureus isolated in pus may be attributed to exposure of wounds which makes them more prone to infections and poor hygiene.
Staphylococcus aureus isolates showed high sensitivity to quinupristin/dalfopristin, imipenem, nitrofurantoin, tigecycline, ampicillin/sulbactam and linezolid. This is consistent with a similar study done in Iran [11]. Research done on antibiotics currently used in the treatment of infections caused by S. aureus in Australia indicates quinupristin/dalfopristin and linezolid have good antistaphylococcal activity but are very expensive [12].
In our study, high resistance of S. aureus isolates was observed against penicillin G (91.9%), and trimethoprim/ sulfamethoxazole (56.9%). This finding is similar to those of studies done at KNH [13], in Eritrea [14], in Nigeria [15], and Namibia [16]. This resistance could be attributed to the mechanism of resistance like the permeability barrier, efflux pumps, mutational or recombinational changes in the target enzymes and acquired resistance by drug-resistant target enzymes in trimethoprim/sulfamethoxazole and alteration of the target with decreased affinity for the antibiotic in penicillin [1].
In this study, MRSA was tested using cefoxitin screening. Overall MRSA prevalence was 27.8%. This prevalence was lower than in previous studies that reported  [13], 31.5% in Uganda [17] and 46.3% in Iran [18]. This difference could be attributable to interventions that may have been effected during the study period such as infection control and improved antimicrobial stewardship and/or appropriate antibiotic use, this difference in MRSA prevalence could also be attributed to the different laboratory techniques used to correctly identify MRSA. KNH started using automated VITEK-2 system in mid 2013. VITEK @ 2 Gram Positive identification card (vitek @ 2 gp card bioMerieux) used at KNH is sensitive and specific to subspecies S. aureus aureus [18,19].
In contrast, MRSA prevalence was lower in studies done in two private hospitals in Nairobi, Kenya, which showed a 3.8% prevalence [20]; a study in Eritrea that recorded 9% prevalence and 0.03% prevalence in Dutch hospitals [13,21]. The low prevalence of MRSA in private hospitals could be attributed to better infection controls. This shows there is high variance of MRSA prevalence from different countries. Majority of MRSA was isolated from pus specimens, 154/232 (66%). Our finding concurs with studies done in two private hospitals in Nairobi, a Namibian institute of pathology and a tertiary health institution in Nigeria [14,15,22]. In contrast, other studies done in Nigeria, Iran and Jamaica showed different specimens were predominant [9,23,24]. The high number of MRSA from pus in our study could be due to exposure of wounds and abscesses to S. aureus. Carriage of S. aureus on the skin makes wounds more prone to MRSA infections.
In this study, 5% of MRSA isolates were resistant to vancomycin. This finding is similar to a study done in Iran which showed 5% of the MRSA isolates were resistant to vancomycin [11]. This contrasts similar studies done in a tertiary care hospital in India and pediatrics and neonatal intensive care patients at KNH which, respectively showed 3.5 and 1% resistance to vancomycin among MRSA [12,23]. Studies done on antimicrobial susceptibility of MRSA in hospitalized patients in Iran, two hospitals in India and two private hospitals in Kenya showed 100% susceptibility to vancomycin [8,22,25]. Our study showed S. aureus isolates were highly susceptible to newer drugs. These drugs include; quinupristin/dalfopristin, tigecycline, imipenem, teicoplanin, vancomycin and linezolid. Similar studies done in Kenya and USA have shown S. aureus to be highly susceptible to ceftobiprole, tigecycline, linezolid, teicoplanin, vancomycin and daptomycin [22,26].
This finding differs from a study done by Arianpoor et al. on antimicrobial susceptibility pattern of S. aureus isolates against newly marketed antibiotics in Iran which showed 5.5% of MRSA isolates were resistant to linezolid, 5.9% of to quinupristin-dalfopristin and 18.9% to tigecycline [11].

Conclusion and recommendations
Our study showed that burns unit had the highest (80%) proportion of MRSA isolates. Both MRSA and MSSA were highly susceptible to quinupristin/dalfopristin, tigecycline, linezolid, nitrofurantoin, ampicillin/sulbactam, and vancomycin, but showed high resistance to commonly used antibiotics such as gentamycin, erythromycin, levofloxacin, SMX-TMP and tetracycline. Infection control measures should be enhanced in burns unit. Information from this study may be used in future as a baseline for follow-up to the susceptibility trend of various drugs to be used for the treatment of S. aureus infections. Routine screening of MRSA and regular studies should be conducted to find out the sources of MRSA. It is important to do culture and sensitivity of relevant specimens when S aureus infection is suspected. There is need for further research on molecular studies evaluating the resistance genes and monitoring the epidemiology of multiple drug resistant S. aureus and MRSA.

Limitations
This was a retrospective study, where some information such as specimen type, collection date, age of the patient, clinical information, previous antibiotics use, duration of patient stay in the hospital and outcome of the therapy were missing.