Introduction

Bacterial infections of wounds are among the leading causes of morbidity and mortality throughout the world and are regarded as one of the most common nosocomial infections. Wound infections have been reported to vary between 3 and 11% in developed countries and estimated to be as high as 40% in developing countries [1,2,3]. Wound infections increase with the degree of wound contamination, and it is estimated that 50% of wounds contaminated by bacteria become clinically infected [4].

Drug resistance impinges on the quality of patient care through its associated mortality, morbidity and significant economic consequences [5]. In hospital practice, 30–50% of antibiotics are prescribed for surgical prophylaxis and 30–90% of these prophylaxes are inappropriate [6]. Inappropriate use of antibiotics increases selection pressure favouring the emergence of pathogenic drug-resistant bacteria which makes the choice of empirical antimicrobial agents more complicated [7, 8].

Extended spectrum beta-lactamase (ESBL) producing organisms are another type of common bacteria resistant to antibiotics. ESBL producing Gram-negative rods (GNRs) have spread all over the world [8, 9]. The prevalence of ESBL producing GNR varies across the world from 50 to 80% [8, 10, 11]. About 33% of infections by ESBL producers are deadly. In Tanzania, the death rate due to ESBL producing GNR is as high as 13.9% [12].

Comparing to Gram-negative, Gram-positives bacteria have been reported to be less prevalent causing wound infections [8, 11, 13,14,15]. Staphylococcus aureus (S. aureus) has been reported to be the most common isolated bacteria from different wound types. Pseudomonas aeruginosa are commonly isolated in infected wounds following surgeries and burns whereas Enterococcus species and Enterobacteriaceae are commonly isolated from wounds in immune-compromised patients and abdominal surgeries [4, 8, 16,17,18].

The majority of the isolates from infected wounds are known to be resistant to ampicillin and amoxicillin. Large numbers of S. aureus are methicillin-resistant S. aureus (MRSA) and most bacteria isolated are sensitive to quinolones, aminoglycosides and monobactam [10, 11, 19,20,21].

Infection in a wound delays healing, prolongs hospital stay, increases trauma, poses risk for disarticulation and amputation, increases need for medical care and increases treatment costs [22]. This makes infection of wounds a matter of concern and makes it necessary to study the causative agents of these infections and their antibiogram.

Main text

Characteristics of participants and enrolment procedures

Patients with Surgical Sites Infections (SSI), infected diabetic wounds, infected wounds due to trauma, and patients with other infected wounds admitted in surgical ward at Kilimanjaro Christian Medical Centre (KCMC) from July 2013 to June 2014 were included in this study. Prior to enrolment in the study, patients were examined by a physician for a suspected or actual wound sepsis using the following criteria; ‘cellulitis’, ‘maladour’, ‘pain’, ‘delayed healing’, ‘deterioration in the wound’ or ‘wound breakdown’ and ‘increase in exudate volume’. Patients presenting with at least three of these clinical signs were enrolled in the study. Chronic wound was differentiated from acute wound if it failed to heal within 4 weeks and showed no sign of improvement within 8 weeks.

Data collection

Pus swabs collection and culture

Wound swabs were collected from patients with infected diabetic wounds, surgical sites, trauma and other wounds by the research nurse. To avoid contaminating the swab with skin flora, pus or necrotic tissue, the wound was thoroughly cleansed with 60–120 mL sterile normal saline prior to taking the sample. Sterile gauze was used to remove excess saline from the wound surface and the pus swabs were collected using sterile swab by swabbing at the middle of the wound. When there were two or more wounds in the same location, separate swabs were used for each wound. A swab moistened with sterile normal saline was rolled deep in the wounds and inserted immediately into a tube containing Stuart’s transport media for preservation of microbes and then transported to the laboratory [8, 16]. Pus swabs were streaked on Blood Agar (BA) and MacConkey Agar (MCA) plates and incubated aerobically for 18–24 h at 37 °C. They were then observed for bacterial growth. Plates with no growth and with growth were re-incubated for another 18–48 h for isolation of bacteria that require extended incubation (slow growers) [8, 16, 17].

Identification of bacterial pathogens

Standard techniques were used for identification of pathogenic bacteria isolated in pure cultures. Characteristic morphological appearances of colonies on media, Gram stains and standard biochemical tests including catalase, coagulase, oxidase, Voges Proskauer, hydrogen sulphide production, urease, methyl red, indole, citrate, CAMP test and sugar utilisation were used to characterise bacteria and identify them [17].

Antibiotics susceptibility testing

Drug susceptibility tests were performed using the Kirby–Bauer disk diffusion method according to Clinical and Laboratory Standards Institute (CLSI) guidelines. A sterile swab was dipped into the suspension of the isolate in normal saline, squeezed free from excess fluid against the side of tube and spread over the Mueller–Hinton agar plate. The density of suspension to be inoculated was determined by comparison with the optical density of McFarland 0.5 Barium sulfate solution. Sensitivity discs of appropriate antibiotics were placed onto the media and incubated at 37 °C for 16–18 h except for coagulase-negative staphylococci which was incubated for 24 h and methicillin-resistant staphylococci at 35 °C [23]. Zones of inhibition were read and, incubation and resistance rates to respective antibiotics were determined.

ESBL production screening

ESBLs production was tested by the disc diffusion method on Mueller–Hinton Agar according to the CLSI guidelines and confirmed by the double disc approximation method [23].

Statistical analysis

Clinical, demographic and laboratory data were entered and linked for each patient using Statistical Package for Social Science software version 20 (IBM Corp, Chicago IL). Thereafter, data were cleaned and analysed using Stata software (Version 13, StataCorp, College Station, Texas). Numeric variables were summarised using measures of central tendency with their respective measures of dispersion while frequency and percentages were used to summarise categorical data.

Results

General characteristics of the study participants

In total, 93 patients diagnosed with infected wounds were enrolled in this study. Male patients numbered 63 (67.7%). The median (range) age at recruitment was 45 (1–80) years. Most of participants, i.e. 65 (71.4%) had acute wounds. The majority of patients, i.e. 82 (90.1%) indicated to have used antibiotics either as a prophylaxis or treatment previously. Additional file 1: Table S1 shows these results.

Bacteria isolated

A total of 93 wound swabs from 93 patients were cultured and 146 bacteria were isolated. Of them 91.4% had bacterial growth. Gram stains of pure cultures showed 91 (62.3%) of the isolates were gram-negative rods. A total of 144 pathogenic bacteria were isolated from 83 cultures. Staphylococcus aureus was the most common isolate (16.0%) followed by other Coliforms and Enterococcus spp. (12.5% each). Figure 1 shows these results.

Fig. 1
figure 1

Species of bacteria isolated

Full size image

More than one-third of the wound infections were caused by single isolate, i.e. 38 (44.7%). According to the type of wounds, Staphylococcus aureus was the most isolated bacteria in acute wounds (29.1%) followed by Pseudomonas aeruginosa (18.2%) and other Coliforms (23%). Whereas in chronic wounds, Proteus mirabilis (26.9%) followed by Enterococcus species and Escherichia coli (23.1%) were the most common isolated bacteria. Additional file 2: Table S2 and Additional file 3: Figure S1 depict these results.

Antibiogram of the isolated bacteria

Staphylococcus aureus showed high resistance to amoxicillin (61.9%). Most Gram-negative rods isolated were very resistant to amoxicillin–clavulanate and cotrimoxazole (66.7–100%) respectively. Tables 1 and 2 show these findings.

Table 1 Drug resistance patterns of Gram positive isolates
Full size table
Table 2 Antibiotic resistance patterns for Gram negative rods isolated (n = 91)
Full size table

ESBL producing Gram-negative rods

The 44 Gram-negative rods isolated, including Klebsiella pneumoniae, Escherichia coli and Proteus species, were phenotypically tested for ESBL production. Half (50%) of these isolates were ESBL producers. All ESBL-producing Gram-negative rods showed 100% resistance rates to ceftriaxone, cefotaxime and cotrimoxazole. These bacteria showed resistance rates of 60–100% to amoxClav, ceftazidime and gentamycin. All ESBL producing GNR showed no resistance to amikacin.

Discussion

In this study, Gram-negative rods were the predominant and leading cause of wound infections. These findings are in line with those of previous studies in Asia and other African settings [8, 11, 13, 14]. This might be due to high resistances to antibiotics showed by Gram-negative bacteria compared to Gram-positive isolates, and therefore their persistence in infected wounds. Furthermore, chronic wounds were infected by multiple Gram-negative rods. The multiple bacterial infections in this case might be due to impaired immune responses associated with diabetes. These results are in accord with recent studies in Tanzania where polymicrobial wound infections were reported to be common in diabetic wounds [8, 24].

In general, Staphylococcus aureus was the most common bacteria isolated in this study. This is a finding consistent with most studies done across the world [10, 17, 21]. S. aureus are normal flora of the skin and anterior nares, therefore they can easily contaminate wounds and cause infections. Moreover, S. aureus are known to have a vast number of virulence factors that increase their ability to cause infections when compared to other bacteria. Our findings are contrary to the study conducted in a similar setting where Pseudomonas aeruginosa was the common isolate in SSI. These variations could be attributed to several factors including the nature of the surgical site itself, the wound site, the type of prophylactic antibiotics used for infections prevention, the level of nursing care given and the measures taken to prevent nosocomial infections [8, 10].

Enterococcus was the most common bacteria isolated in diabetic wounds, perhaps due to their opportunistic pathogen behaviour since lowered immune responses are associated with diabetes. Other common isolates from IDFUs were Proteus and Klebsiella, which are known to be common isolates in chronic wounds. These bacteria had high rates of ESBL production and showed high multiple drug resistance (MDR) rates. Studies elsewhere have reported similar findings [13,14,15, 20, 24].

In this study, high drug resistance was observed for amoxicillin–clavulanate and cotrimoxazole. These antibiotics are relatively cheap and readily available. These, together with policies that do not restrict antibiotics accessibility to patients, might have caused the irrational overuse of these drugs which might have led to bacterial resistance. Cephalosporins were ineffective against most Gram-negative rods. This might be due to mutational emergence and the spread of ESBL-producing Gram-negative rods and the extensive use of these antibiotics in both treatment and prophylaxis. In this setting, good responses were seen for ciprofloxacin and amikacin. In our setting, the use of these antibiotics is highly restricted due to their adverse side effects. Ciprofloxacin has been recommended only in certain bacterial infections. Furthermore, amikacin is very expensive in our community and the majority of patients could not afford to use it. The different levels of resistance to cefotaxime and ceftriaxone are surprising, although this may be influenced by the fact that not all isolates were tested against all antibiotics. Other studies showed findings in accordance with these [8, 10, 19].

Limitation

Despite being a commonly-used, non-invasive and cost-effective method, swabs might provide a poor specimen as compared to needle aspiration or tissue biopsy if not collected appropriately. Improper specimen collection affects the results obtained, often by reflecting normal skin flora and colonizing organisms, making it difficult to decide which organisms are the true pathogens. However, our results are not likely to be affected by this since the wound was cleansed thoroughly prior to swab collection. Moreover, the small sample size did not allow us to conduct advanced statistical analyses which could have potentially strengthened this study.