Bacteria from infected surgical wounds and their antimicrobial resistance in Hawassa University Referral Teaching Hospital , Southern Ethiopia

A study was carried out from November 2010 to June 2011 in Hawassa University Referral Teaching Hospital to identify bacterial species involved in post-operative wound infections and to determine their antimicrobial resistance pattern. The study involves 100 surgical patients with post-surgical wound infections. Swab samples of wound discharge were collected for bacteriological examination and inoculated on appropriate culture media. Isolates were identified and characterized by standard methods and antibiotic resistance was determined using the Kirby-Bauer disk diffusion method. A total of 177 bacterial isolates were identified in the study. The most dominant isolates were Staphylococcus aureus, Klebsiella spp., Escherichia coli and coagulase negative staphylococci (CoNS) accounting for 45 (25.4%), 32 (18.1%), 30 (16.9%) and 26 (14.7%) of the isolates respectively. Other bacteria isolated include Pseudomonas aeruginosa (9.0%), Proteus spp. (6.8%), Streptococci (5.1%), Citrobacter spp. (2.3%) and Enterobacter spp. (1.7%). Of the 177 isolates, 173 (97.7%) were resistant to at least 1 antimicrobial, while 164 (92.7%) were resistant to ≥2 antimicrobials. Resistance of isolated organisms was 76.3% to amoxicillin, 71.2% to penicillin, 56.9% to vancomycin, 39.5% to ceftriaxone and norfloxacin and 31.1% to gentamicin. The susceptibility of S. aureus was 64.4% to gentamicin but it was 100% resistant to amoxicillin. All isolates of P. aeruginosa were resistant to penicillin and amoxicillin. The rate of resistance of S. aureus to 2 or more antimicrobials was 97.8% and that of P. aeruginosa was 100%. This study confirms that the bacteria commonly implicated in post-operative wound infections: S. aureus, Klebsiella, E. coli, CoNS, and P. aeruginosa, continued to dominate and have developed high level of drug resistance to some important antibiotics. Periodic surveillance of the species of bacteria involved in post-operative wound infection and determination of their antimicrobial resistance is recommended for empirical treatment.


INTRODUCTION
Despite advances in infection control, surgical site infections (SSIs), formerly called surgical wound infections, remain a substantial cause of morbidity and mortality among hospitalized patients (Mangram et al., 1999).They were the 3 rd most frequently reported nosocomial infection, accounting for 14 to 16% of all nosocomial infections among hospitalized patients in the US (Emori and Gaynes, 1993).Surgical wound infection increases hospital stay (Poulsen et al., 1994;Mitt et al., 2005) and thereby the cost of medical treatment (Boyce et al., 1990;Kirkland et al., 1999).It doubles the patient's risk of death after surgery (Kirkland et al., 1999) and causes patient discomfort.It also places a significant burden on the health system (Poulsen et al., 1994), especially in Africa where resources are limited.
Post-operative wound infection risk depends on a number of factors.Those most frequently cited in the literature include length of surgical procedure, surgical procedure category, obesity, use of pre-operative prophylactic antibiotics, colonization with microorganisms, age, sex, anemia, diabetes mellitus, malnutrition, smoking and length of pre-operative hospitalization (Barber et al., 1995;Anvikar et al., 1999;Ahmed et al., 2007).The virulence and invasive capability of the organisms influence the risk of infection, but the physiological state of the tissue in the wound and immunological integrity of the host also have importance in determining whether infection occurs (Fry, 2003).
Wound infection is most commonly characterized by the classic signs of redness (rubor), pain (dolor), swelling (tumor), elevated incisional tissue temperature (color) and systemic fever.Ultimately, the wound is filled with necrotic tissue, neutrophils, bacteria and proteinaceous fluid that together constitute pus (Fry, 2003).
A high level of post-operative wound infection (17.9%) was reported from one of the largest hospitals in Ethiopia, with significant proportion of isolates showing drug resistance (Tekie, 2008).However, there is no adequate data on the problem in the country in general and in the study area in particular, in spite of the necessity of periodic review of isolates and their drug resistance patterns for effective prevention and treatment of surgical wound infections.
This study, therefore, aimed to find out bacterial species involved in post-operative wound infections and to determine their antimicrobial resistance pattern in Hawassa University Referral Teaching Hospital.

The study area
The study was conducted at Hawassa University Referral Teaching Hospital (HURTH) from November 2010 to June 2011.Hawassa town is located at 7°5' Latitude N and 38° 29 ' longitude E, 275 km Guta et al. 1119 south of Addis Ababa, the capital of Ethiopia.It is the capital of Southern Nations Nationalities and Peoples Region.HURTH serves as a main referral center for the Southern part of Ethiopia serving roughly 10 million people in the region and surrounding areas.The hospital went operational in 2003 and is affiliated to the Hawassa University.There are a total of 400 beds in the hospital.However currently only 250 are in use; of which 58 are surgical.

Patients' profile
The study involved 100 patients, 37 males and 63 females, who underwent surgical operations for various reasons and developed post-operative surgical wound infection in the study period.Eighty one of the patients underwent emergency surgery while 19 were operated on elective bases.The study included surgical patients from general surgery, pediatrics and obstetrics and gynecology wards.The age of the patients ranged from 1 to 77 years.

Sample collection and transportation
Samples for bacteriology, from apparently infected surgical wounds, were collected aseptically with sterile cotton tipped swabs.The wounds were examined for suggestive signs of infection: redness, pain, swelling, elevated tissue temperature, discharge and systemic fever (Fry, 2003) before the patients were discharged from the hospital.
Examination for suggestive signs and symptoms of infection and specimen collection was carried out by experienced nurses working in the respective wards.When infection was clinically suspected, the area around the surgical wound was cleaned with 70% ethyl alchohol.The exudates were collected from the depth of the wound using sterile cotton tipped swabs.All the specimens collected were immediately transported to the microbiology laboratory of the hospital for bacteriological examination.Samples were inoculated onto appropriate culture media as soon as they arrived at the laboratory.

Isolation and identification
The samples were inoculated on blood agar, mannitol salt agar (selective medium for S. aureus) and MacConkey agar (Oxoid, Basingstoke, England).The plates were incubated at 37°C for 24 -48 h aerobically.All positive cultures were identified by their characteristic appearance on their respective media, Gram staining reactions and confirmed by the pattern of biochemical reactions using the standard methods (Cheesbrough, 2006).Members of the family Enterobacteriaceae and other Gram-negative rods were identified by indole and H2S production, gas formation, citrate utilization, motility and oxidase and urease test.For Gram-positive bacteria coagulase and catalase tests were used (Cheesbrough, 2006).
Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License diffusion method (NCCLS, 2003) for six antimicrobials currently used in the hospital setting and available on the market.Morphologically identical 4-6 colonies from overnight culture were touched with sterile loop and inoculated in 5 ml nutrient broth and incubated for 4 h at 37°C.Turbidity of the broth culture was matched to 0.5 McFarland standards.The surface of Mueller-Hinton agar plate was evenly inoculated with the culture using a sterile cotton swab.The antibiotic discs were applied to the surface of the inoculated agar using sterile forceps.After 18-24 h of incubation, the diameter of growth inhibition around the discs were measured using a millimeter scale to the nearest millimeter and interpreted as sensitive, intermediate or resistant according to National Committee for Clinical Laboratory Standards (NCCLS) (2003).The antimicrobial agents tested were: amoxicillin (Aml), norfloxacin (Nor), vancomycin (Van), gentamicin (Gen), penicillin (Pen) and ceftriaxone (Cro) (Oxoid, Basingstoke, England).

Data analysis
The collected data were entered in MS Excel and analyzed using Stata version 9 for windows (Stata Corp. College Station, TX).Sex and age of the patient, type of surgery (elective or emergency), and ward (surgical, obstetrics and gynecology and pediatrics) were considered as potential factors which could affect the prevalence of a specific bacteria in post-operative wound infection.Association of these factors with the frequency of bacterial type (species/genus) isolated was analyzed using univariable logistic regression analysis.

Ethical consideration
The study obtained ethical clearance from the Institutional Review Board of Hawassa Collage of Medicine and Health Sciences.Patients, or parents in case of children, were told about the objectives of the study and those who gave their informed consent were included in the study.

Bacterial isolates
Of the 100 samples collected from infected post-opera- A total of 177 bacterial isolates were recovered from the 100 samples examined.Out of these 177 isolates, 97 (54.8%) were Gram negative bacteria while the rest 80 (45.2%) were Gram positive.The proportion of each bacterial isolate to the total number of isolates is presented in Table 1.The most dominant isolates were S. aureus, Klebseilla spp., E. coli and CoNS accounting for 25.4,18.1, 16.9 and 14.7% of the total isolates, respectively.

Effects of sex, age, type of surgery and ward on the types and frequency of bacterial isolates
The results of univariable analyses of frequency of isolation of bacteria from infected post-operative wounds with possible risk factors are summarized in Table 2. Frequency of isolation of Klebsiella and Pseudomonas were significantly associated (P <0.05) with gender of the surgical patient.Klebsiella was most frequently isolated from female patients (41.3 vs. 16.2%)while Pseudomonas predominated in male patients (29.3% Vs 7.9%).Staphylococcus aureus tended to be more prevalent in females (52.4% Vs 32.4%) but the difference just failed to be significant (P =0.053).There was, however, no significant association (P >0.05) between sex of the patient and the frequency of other isolates recovered in the study.
Age of the patient and type of surgery (elective or emergency) appeared to have no significant association (P >0.05) with the frequency of isolation of any type of bacteria identified in the study.
Samples from patients with post-operative wound infection in gynecology and obstetrics ward resulted in a significantly (P <0.05) high S. aureus isolation than samples from pediatrics and surgery wards.Similarly, samples from gynecology and obstetrics ward yielded a significantly (P <0.05) more Klebsiella as compared to samples from surgical ward.However, post-operative infected wounds from surgical ward (42.5%) had a higher prevalence of CoNS as compared to wounds in obstetrics and gynecology (13.5%) and pediatric (17.4%) wards (P <0.05).

Antimicrobial resistance
Antimicrobial susceptibility of all isolates recovered in this study is summarized in Table 3.All the 177 bacterial isolates obtained from infected surgical wounds were tested for 6 antimicrobials.Of the 177 isolates belonging to 9 species/genera, 173 (97.7%) were resistant to one or more antimicrobials, while 164 (92.7%) were resistant to 2 -6 antimicrobials.
The overall resistance of isolated organisms, irrespective of species/genus, was 76.3% to amoxicillin, 71.2% to penicillin, 56.9% to vancomycin, 39.5% to ce-ftriaxone and norfloxacin each and 31.1% to gentamicin.
All of the 45 S. aureus isolates were resistant to 1 or more antimicrobials and 44 (97.8%) were resistant to ≥2 antimicrobials.Resistance of S. aureus was detected for all the 6 antibiotics tested in this study.All the 45 isolates (100%) were resistant to amoxicillin, while 30 (66.7%) were resistant to vancomycin.However, 29 (64.4%) were susceptible to gentamicin.
Resistance was also high in Klebsiella isolates: 30 (93.8%) being resistant to 1 or more antibiotics and 23 (71.9%) to 2 or more antimicrobials.The highest resistance of Klebsiella (78.1%) was recorded against amoxicillin.All the 30 E. coli isolated in this study were resistant to 2 or more antimicrobials and all of them were resistant to penicillin.On the other hand, all of them were susceptible to norfloxacin and gentamicin.Of the 26 CoNS 24 (92.3%) were resistant to ≥2 antimicrobials.Resistance to penicillin among CoNS isolates was 92.3%.All the 16 Pseudomonas isolates had resistance Cro (ceftriaxone), Nor (norfloxacin), Gen (gentamicin), Van (vancomycin), Pen (penicillin), Aml (amoxicillin).
to ≥2 antimicrobials and all of them were resistant to penicillin and amoxicillin.Only 4 (25.0%)isolates of P. aeruginosa were found susceptible to at least 1 of the antibiotics tested and it was only to gentamicin.
Resistance to 2 or more antimicrobials was 100% in Proteus spp., and resistance to penicillin, amoxicillin and vancomycin was 91.7, 91.7 and 83.3% respectively.Eight (88.9%) of the 9 streptococci were resistant to ≥2 antimicrobials and 7 (77.8%)isolates were resistant to amoxicillin and 5 (56.6%) to gentamicin.All Citrobacter and Enterobacter spp.isolated in this study were resistant to penicillin and amoxicillin.However, no Citrobacter was resistant to gentamicin.

Bacterial isolates
Out of the 100 samples collected from post-operative wound infections, 92 (92%) were culture positive yielding a total of 177 bacterial isolates.Similarly, Rao and associates (2013) recovered bacterial isolates from 96 of 100 samples collected from clinically suspected postoperative wound infections in India.A Nigerian study reported isolation rate of 96.2% from infected wounds (Sule et al., 2002).The absence of bacterial growth in samples collected from apparently infected surgical wounds could be due to the effect of antimicrobials which are used routinely in surgery, it can also be due to antiseptics used for cleaning the wounds or it may even be due to the body's defense mechanism overcoming the infection.It is also possible that some organisms were unable to grow under the aerobic condition the samples were cultured.Anaerobic bacteria (Mangram et al., 1999;Rao et al., 2013) and Candida albicans (Mangram et al., 1999;Hidron et al., 2008) were implicated in some wound infections in the literature.
The bacteria isolated in our study were commonly associated with surgical wound infections in Ethiopia and elsewhere (Ahmed et al., 2007;Tekie, 2008;Shanthi et al., 2012;Rao et al., 2013).Similar to our observation (25.4%) S. aureus was reported to be the most dominant isolate from surgical wound infections representing 25-45.1% of the total isolates (Ahmed et al., 2007;Anguzu and Olila, 2007;Khorvash et al., 2008;Tekie, 2008;Shanthi et al., 2012).The high prevalence of S. aureus infection may be partly explained by the presence of these bacteria in the nose, skin and intestinal tract of human beings (Vandepitte et al., 2003).With the disruption of natural skin barrier, S. aureus which are common bacteria on surfaces, easily find their way into surgical sites.
Klebsiella spp.(18.1%) and E. coli (16.9%) were the 2 nd and 3 rd most frequent isolates in our study which is consistent with the report of Khorvash et al. (2008) from Iran.A Nigerian study indicated Klebsiella spp.(25.3%) as the most dominant isolate from wound infections (Sule et al., 2002).Anvikar and associates (1999) reported K. pneumoniae as the most common isolate from clean surgical wounds.A study conducted in India reported E. coli as the most common isolate (20.8%) followed by S. aureus and Pseudomonas species (Rao et al., 2013).
In our study Pseudomonas was high in males than females.This finding agrees with the report from India by Ruhil et al. (2009) and Ranjan et al. (2010) where the prevalence of P. aeruginosa was higher in male patients, as compared to the females.A similar result was reported from Addis Ababa, Ethiopia, (Tekie, 2008) where the rate of P. aeruginosa infection was relatively higher in males than in females.
In this study, the highest proportions of S. aureus and Klebsiella spp.were isolated from gynecology and obstetrics wards.Our result was in agreement with a report from Nigeria by Sule et al. (2002) where Klebsiella species were found to be the most common in wounds from obstetrics and gynecology units.
In our study, highest proportion of CoNS was isolated from general surgery ward as compared to other wards.A similar result has been reported from a teaching hospital in Addis Ababa, Ethiopia, (Tekie, 2008) where CoNS were recovered with high rate from surgical emergency ward.

Antimicrobial resistance
All the S. aureus isolates were resistant to amoxicillin while the majority of S. aureus (64.4%) were sensitive to gentamicin.This finding was similar to results of Anguzu and Olila (2007) where the majority of the S. aureus isolates (87.5%) were sensitive to gentamicin.The sensitivity of S. aureus to gentamicin recorded in a Nigerian study (60.0%) was also in agreement with our result (Nwachukwu et al., 2009).In the same Nigerian study, 10 out of 20 (50%) isolates of S. aureus were sensitive to norfloxacin, supporting our finding of 51.1% resistance.The resistance of P. aeruginosa to gentamicin was 50.0% in exact agreement with the resistance level recorded in a study performed by Nwachukwu and associates (2009) in Nigeria.
In conclusion this study indicated that S. aureus, Klebsiella, E. coli, CoNS and P. aeruginosa were the most important bacteria responsible for post-operative wound infection in Hawassa University Referral Teaching Hospital.High rates of drug resistance to some commonly used antibiotics were observed in this study and this warrants attention to the problem.Resistance was especially high in S. aureus, P. aeruginosa and Proteus spp.
Periodic surveillance of bacterial species involved in post-operative wound infections and their antimicrobial resistance is recommended to minimize the incidence of post-operative wound infections, shorten duration of recovery and reduce cost of medical treatment through empirical treatment which takes into account the epidemiological information on the species of bacteria involved and their antimicrobial resistance.

Table 1 .
Frequency of isolation and percentage of the total number of isolates of bacteria from post-operative wound infections (n=100).
a Coagulase negative staphylococci.
All polymicrobial infections in our study involved Gram positive and Gram negative bacteria with S. aureus and Klebsiella species being the most common association in 16 cases.Other frequent associations were: S. aureus and E. coli (10 cases), S. aureus and P. aeruginosa (9 cases), CoNS and Klebsiella spp.(8 cases) and CoNS and E. coli (7 cases).

Table 2 .
Association of frequency of bacterial isolates, from infected post-operative wounds, with sex, age, type of surgery and ward.
a Gynecology and obstetrics;b Coagulase negative staphylococci.

Table 3 .
Number (percentage)of antimicrobial resistant isolates of bacteria from surgical wound infections.