MATERIALS AND METHODS

The details of the materials and methods used in the isolation and characterization of the bacteria have been previously described by Mthembu et al.²⁰. Briefly wound specimens of 18 (22% male and 78% female) randomly selected hospitalized diabetic patients were collected from three rural-based Northern KwaZulu-Natal hospitals (named X, Y and Z) in South Africa between 2015 and 2016. To collect the wound specimen medical doctors firstly cleaned the wound with sterile saline and sterile cotton pads, after which they then introduced sterile swabs at the base of the wound and subsequently inserted the swab into Amies transport media to maintain the specimen during transportation to the University of Zululand’s Biochemistry laboratory.

Antibiotics: All the antibiotic discs were supplied by Polychem (OXOID). Nine antibiotics representing different antibiotic classes were selected and are as shown in Table 1.

Antibiotic susceptibility test: Antibiotic susceptibility testing of the bacteria isolates was done according to the Bauer et al.²¹ method as recommended by the Clinical Laboratory Standard Institute (CLSI)²². The test isolates were grown on Mueller-Hinton broth overnight for 18-24 h at 37°C after which turbidity of the suspension was adjusted to match 0.5 McFarland’s standard. A volume of 100 μL of the diluted suspension was inoculated into Mueller-Hinton agar containing plates and streaked with sterile swabs. Antibiotic disks were tested against different bacterial groups such as Staphylococci, Enterococci, Gram-negative bacilli from Enterobacteriaceae, non-Enterobacteriaceae Gram-negative bacilli and Gram-positive rods. The antibiotic disks were evenly placed on plates aseptically and incubated for 18-24 h overnight at 37°C. The choices and interpretive standards of antibiotics were based on the CLSI²³ guideline. The zones of inhibition of the bacteria were measured and classified into three groups: resistant, intermediate and susceptible according to the CLSI²³ guideline. Multidrug-resistant patterns were assessed whereby, bacteria identities resistant to two or more classes of antibiotics were recorded as multi-drug resistant¹⁸.

Table 1:	Antibiotics used in the study and their respective classes

Statistical analysis: Variants were analyzed using one-way ANOVA and two-way ANOVA through Graphpad prism version 6, whereby graphs and figures were constructed at a 95% confidence level unless stated otherwise.

RESULTS

Antibiotic susceptibility test: The 5 groups of bacteria showed varying susceptibility patterns as shown in Table 2. Multidrug resistance patterns were noted among the different groups, Enterococci (83%), Enterobacteriaceae (55%), Non-Enterobacteriaceae (50%), Staphylococci (50%) and the Gram-positive rods (33%). Table 3 shows the percentage resistance observed for each bacteria group against an individual antibiotic. Resistance percentages of 100% were noted among the Staphylococci and Enterococci group against the antibiotic penicillin, while non-Enterobacteriaceae and the Gram-positive rods exhibited 100% resistance against ampicillin. The highest resistance was observed against β-lactams in all groups. Table 4 shows the overall susceptibility patterns of each individual antibiotic against all the bacteria across the different groups. It is of interest that a 100 and 91% resistance against penicillin and ampicillin was noted, respectively. Antibiotic resistance was further evaluated in each of the hospitals under study to obtain the percentage resistance of the bacteria recovered against each individual antibiotic and the results are presented in Fig. 1.

Fig. 1:	Antibiotic resistance percentages (for each antibiotic) against bacteria from the different hospitals. AMP: Ampicillin, CIP: Ciprofloxacin, CAZ: Ceftazidime, VA: Vancomycin, P: Penicillin, IMP: Imipenem, CN: Gentamicin, E: Erythromycin, F-EP: Cefepime

Table 2:	Antibiotic sensitivity profiles of the bacteria
-: Means not tested, S: Susceptible, I: Intermediate, R: Resistant, AMP: Ampicillin, CIP: Ciprofloxacin, CAZ: Ceftazidime, VA: Vancomycin, P: Penicillin, IMP: Imipenem, CN: Gentamicin, E: Erythromycin, F-EP: Cefepime

Bacteria from Hospital X showed the highest resistance patterns against erythromycin (80%) and ciprofloxacin (70%), while bacteria from hospital Z showed the highest resistance at 50% against vancomycin and penicillin. Hospital Y bacteria was most resistant to imipenem (45%). The data shown to be significantly different (p<0.0001).

DISCUSSION

This study revealed very high levels of antibiotic resistance of up to a 100% among the different bacterial species isolated from the wounds of diabetic patients, this was noted mainly within the β-lactam class of antibiotics (Table 3) of which, some are among the first line of treatment given in the event of infection. This therefore limits the treatment options, further compromising the health and quality of life of the patients as diabetic wounds are generally slow healing in nature. This also corresponds to Zubair et al.¹² and Shahi et al.¹³ findings which allude that the burden of antibiotic-resistant bacteria has become common in diabetic patients with wounds. Multidrug resistance patterns were also noted in the study with the highest percentage being against the Enterococci group (Table 2). The Enterococci spp. have been reported to be resistant to multiple antibiotics in clinical settings, with various resistant mechanisms being reported against all antimicrobial agents in clinical practice²⁴.

Table 3:	Resistance of different bacteria groups against antibiotics

This indicates the pressure imposed on diabetics who are contaminated by multidrug resistant bacteria and become immuno-compromised¹. The other clinical challenge of antibiotic resistance is the decline in efficacy of antibiotic treatment, with the dry pipeline of developing new antibiotic agents which threatens the provision of prompt treatment⁴.

The bacteria were mostly resistant to β-lactam (penicillin, ampicillin, cefepime and ceftazidime) and aminoglycoside (gentamicin) (Table 3, 4). The observed resistance may be exhibited by bacterial cell walls (β-lactam) and ribosomal proteins (gentamicin). β-lactams are bactericidal and inhibit the cell wall synthesis of different organisms²⁵. Penicillin is most active against Gram-positives while ampicillin is an excellent broad-spectrum antibiotic²⁶. Cephalosporins and carbapenems are active against Gram-negatives²⁷ however, in this study most bacteria were resistant to these antibiotics. Bacteria can protect themselves against β-lactams through alteration in penicillin binding proteins (PBPs) that reduce the affinity of β-lactams to the sites of action and the production of β-lactamases that are able to hydrolyze the β-lactam ring²⁸. The frequent use of β-lactams as the first line of treatment has played a major role in the development of antibiotic resistant bacteria²⁹. As a result, infections caused by β-lactam resistant organisms have increased over the years²⁷. To make matters worse, penicillin is still being prescribed without considering its effects on resistance³⁰.

The declining effectiveness of antibiotics is believed to be also driven by mismanagement of antibiotics³¹ in hospitals where most infection causing bacteria are isolated¹. In the study, hospital X was implicated in resistance against ciprofloxacin, a broad spectrum bactericidal antibiotic that is a last line of defense in antibiotic treatment^26,32 and erythromycin, which is considered bacteriostatic and targets gram-positives ^26,33. Resistance to quinolones can be a severe clinical problem due to their importance in health³⁴. Therefore, the use of ciprofloxacin needs to be well monitored.

Table 4:	Overall antibiotic susceptibility patterns
n: number of organisms tested, Mode of action25

Bacteria resistant to ciprofloxacin can mediate resistance through mutations that alter the drug targets or reduce drug accumulation and develop plasmids that protect bacterial cells from the lethal effect of quinolones^32,34 while erythromycin and vancomycin can be resisted through methylation of bacterial ribosomes altered cell-wall precursors, respectively³³, which all compromise the available treatment and increase the burden of antibiotic resistance.

Enterobacteriaceae showed varying susceptibility patterns, with 55% resistance to antibiotics. It has been reported to be resistant to a number of antibiotics through the production of β-lactamase^10,27. All Proteus mirabilis (P. mirabilis) (predominant) isolates were shown to be susceptible to ciprofloxacin while 86% was resistant to Imipenem. Imipenem resistance has been reported to be natural in P. mirabilis³⁵. Imipenem is normally the last line of treatment in extended-spectrum β-lactamase producing organisms^10,35. Some multidrug resistant strains of P. mirabilis produce these enzymes^9,29. Among the Klebsiella pneumoniae (K. pneumoniae) and Escherichia coli species, multidrug resistant patterns were observed against β-lactams, ciprofloxacin and gentamicin in which resistance has been previously reported by Iredell et al.³⁶. The ability of Klebsiella species to rapidly spread in the hospital environment contributes to their resistance which has been closely associated with the production of extended-spectrum β-lactamases (ESBL), mostly mediated by plasmids³⁷. In 2011 K. pneumoniae was documented with the most extreme cases of multidrug resistant bacterial infection in South Africa¹. This emphasizes the importance of routine susceptibility testing before treatment is prescribed¹⁴.

The resistance observed on skin commensals, Staphylococcus epidermidis and Corynebacterium striatum, is normally experienced in patients receiving treatment through antibiotics, especially the broad-spectrum³⁸ such as ciprofloxacin and gentamicin. Enterococcus faecalis, also regarded as normal skin flora³⁹, showed multidrug resistant patterns with one species (Enterococcus faecalis (A4)), showing resistance to all test antibiotics. The resistance shown by the normal skin flora poses a threat to other sensitive bacteria and patients³⁸. Kiser et al.²⁶, reported that susceptibility testing of bacteria from an anatomic site for which they are found should be avoided because results may encourage the physician to treat normal conditions. The problem is that skin microflora has become resistant to the current treatment⁴⁰, as evidenced by their multidrug-resistant patterns recorded in the study. Treatment in immune compromised patients such as diabetics where bacteria are most likely to become pathogenic is now a challenge.

CONCLUSION

These findings show that it is important to evaluate possible resistance patterns of bacteria since some bacterial strains have become resistant to current treatment. The resistance conferred by Gram-negative bacilli and some skin commensals pose a threat to antibiotics used in the treatment of wound infection and may lead to a public health crisis.

SIGNIFICANCE STATEMENT

This study discovers the antibiotic resistance patterns of some bacteria isolated from wounds of diabetic patients in some Northern KwaZulu-Natal hospitals in South Africa, which can provide beneficial information for policy makers in the health care system and for public health awareness programmes. This study will help the researcher to uncover the critical areas of antibiotic resistance in rural based hospitals in South Africa that many researchers were not able to explore. Thus evidence is now available in literature revealing the presence of multidrug resistant bacteria inhabiting the wounds of hospitalized diabetic patients in these hospitals.

ACKNOWLEDGMENT

Authors would like to acknowledge the Research Committee of the University of Zululand for the funding under project number UZREC 171110-030PGM 2015/195, the participating hospitals, patients and Dr Bonginkosi Shoba for their support.