Diabetic Foot Infection in Morocco: Microbiological Profile

Objective. The objective of this work is to describe the microbiology of diabetic foot infections (DFIs). Materials and Methods. The authors included all samples of infected diabetic foot ulcers between January 2009 and June 2014 at the Mohammed Vth Military Teaching Hospital of Rabat, Morocco. Results. The researchers collected 199 samples corresponding to 157 patients. The mean age of the patients was 59 years ± 12 years. Of the collected samples, deep samples represented 41% and swab samples 59%. Direct examination indicated anaerobic infection in 32% of the cases. There were 307 bacteria isolates from both deep and swab samples. There was no statistically significant association between the sampling method and isolate species (P = 0.237). Enterobacteriaceae, Staphylococcus aureus, Streptococcus sp, nonfermenting gram-negative bacilli (NFGNB), and Enterococcus sp represented 31.8%, 12.6 %, 12.3%, 11.7%, and 8.7% of the isolates, respectively. Methicillin-resistant S. aureus represented 4.7% of S. aureus isolates. Enterobacteriaceae and NFGNB-producing extended spectrum β-lactamases represented 14.1% and 5.1%, respectively, with isolates producing carabapenemase representing 3.8% and 38.5%. Piperacillin-tazobactam, imipenem, and ciprofloxacin resistance concerned 7.5%, 4.7%, and 25.5%, respectively, of isolated Enterobacteriaceae, and 35.9%, 30.7%, and 35.9% of NFGNB. Low susceptibility to β-lactams was found in 4.9% of Streptococcus sp isolates and 4.9% of Streptococcus sp isolates were resistant to moxifloxacin. Conclusion. Gram-negative bacilli are responsible for 43% of DFIs, and multidrug-resistant GNB is a challenging issue in DFI management. The sampling method doesn’t seem to impact the bacteriological profile; however, this finding must be confirmed with further study.

Diabetes affects 285 million people worldwide. In Morocco 6.6% (1.5 million) of the population lives with diabetes.¹ Diabetic foot is a complication of 5%-10% of diabetes cases and a multidisciplinary approach is required for managing it.² Antibiotic therapy is indicated in case of infection proved by consensual criteria such as those defined by Infectious Diseases Society of America guidelines (eg, signs of inflammation, purulent secretions, or secondary signs such as other type of secretions, friable tissue, or odor).³ Diabetic foot infections can range from moderate to severe; as severity increases, the risk of amputation grows by as much as 90%, especially in developing countries.^4-7

Diabetic foot infections (DFIs) are the first cause of hospitalization among individuals with diabetes.⁸ When choosing an antibiotic therapy, clinicians must consider the main suspected bacteria, bone infection, and ecologic risk of multiresistant isolate emergence. Thus, probabilistic antibiotic therapy must be regularly evaluated and adapted to microbiological data.

 A microbiological profile of a DFI is a variable dependent on the acute or chronic character of the wound, hospitalization occurrence and duration, prior antibiotic therapy, and sampling technique.⁸ Acute DFIs implicate generally monomicrobial flora, while chronic DFIs are polymicrobial. Bacteriological documentation is based on deep sampling techniques obtained from infected tissue. In fact, antibiotic therapy based on microbiological data of wound swabs seems to adequately cover only 62%-93% of the pathogens isolated from tissue and deep sampling.^9-12 

The study objective was to describe microbiological aspects of DFIs in Mohammed Vth Military Teaching Hospital, Rabat, Morocco, and the antibiotic therapy susceptibility profile of isolates comparing wound swab samples and deep sampling techniques. 

The study included all DFI samples collected at the Department of Bacteriology between January 2009 and June 2014 at the Mohammed Vth Military Teaching Hospital, Rabat, Morocco. The samples were classified in 2 groups according to sampling method: swab samples and deep samples. Deep samples included tissue and bone samples, which were pretreated with sonication, and pus aspiration samples. All specimens were manipulated using a class II biological safety cabinet (Bio II Advance, Telstar Life Science Solutions, Bristol, PA). 

Cultures were performed on Columbia agar base oxoid with 5% horse blood and boiled horse blood agar supplemented with vitamins and mannitol salt agar (Chapman medium). All culture media used in this study were oxoid dehydrated media prepared according to manufacturer instructions. The cultures were incubated at 37°C aerobic conditions with 5% of CO₂. 

The deep samples were also plated in Shaedler agar and blood agar supplemented with colistin and nalidixic acid in anaerobic conditions. The cultures were incubated at 37°C for 48 hours. Aerobic and anaerobic enrichment broths were plated for deep samples and reinoculated 48 hours later.

A microscopic examination after Gram staining looked for the presence of a cellular reaction—whether or not there are polynuclears and whether or not they are altered—and bacterial flora to describe its density, morphology, and monomicrobial or polymicrobial aspect. Isolate identification was done by standard bacteriological techniques. 

Antimicrobial susceptibility testing and extended spectrum β-lactamase (ESBL) testing were performed by a disk diffusion method, and interpretation was done in accordance with  the Comité d’Antibiogramme de la Société Française de Microbiologie (CASFM) 2013 recommendations.¹³ Carbapenemase screening was based on a modified Hodge test result.¹³

The age, gender, and hospitalization departments of patients were recorded from laboratory prescriptions and the hospital’s database laboratory information system. 

Investigators compared bacteriological results of deep and swab sampling methods. 

Statistical analysis used SPSS version 13.0 software (SPSS Inc, Chicago, IL). Quantitative variables were expressed by mean ± standard deviation, and qualitative variables were expressed by effectives and percentages.  Statistical association was researched by chi-square test or Student’s t test, according to variable type; P < 0.05 was considered significant.

Demographic data. During the study period, 199 specimens were collected corresponding to 157 patients (31 women and 126 men). The mean age was 59 years ± 12 years, and 80% of patients were more than 50 years old (Table 1). 

Sampling techniques. Swab samples represented 57.3% (114/199), while deep samples represented 42.7% (85/199) comprised of bone samples 2% (4/199), tissue samples 9.1% (18/199), and pus aspirates 31.6% (63/199) (Table 1).

Microbiological data. Direct examination of gram-stained smears showed monomorphic aspect in 40.7% of cases (35 deep samples and 44 swab samples), and polymorphous samples represented 57.7% of cases (49 deep samples and 63 swab samples). 

Direct examination of deep samples showed an anaerobic infection in 25.2% of cases (21.2% deep samples vs 28.4% of swab samples).

Sterile cultures represented 10.6% of deep samples vs 12.3% of swab samples. Polymicrobial cultures represented 58.8% (50/85) of deep samples vs 50.0% (57/114) of swab samples. There was no statistical association between polymicrobial or monomicrobial nature of culture and the sampling method (P = 0.259) (Table 2). 

The positive culture identified 307 bacterial isolates and 3 yeast isolates (140 deep samples vs 167 swab samples). Gram-negative bacilli (GNB), gram-positive cocci (GPC), and gram-positive bacilli (GPB) accounted for 48.8% (150/307), 45% (138/307), and 6.2% (19/307) of isolates, respectively. Their distribution depended on the sampling method (Table 2).

The predominant species in both deep and swab samples was Staphylococcus aureus (11.8% vs 13.3%). Enterobacteriaceae were isolated in 29.8% vs 34.3% of cases, Enterococcus sp in 11.8% vs 6.1%, Streptococcus sp in 16.4% vs 9.9%, nonfermenting GNB (NFGNB) in 8.6% vs 14.4%, coagulase-negative Staphylococcus sp in 7.9% vs 7.7%, and Corynebacterium sp in 4.6% vs 3.9%, respectively.

Escherichia coli (9.3%) was the most frequent among Enterobacteriaceae isolates from deep samples followed by Proteus sp (7.1%) and Enterobacter sp (6.4%). Enterobacter sp was predominant (11.4%) among Enterobacteriaceae isolates in surface swabs followed by Proteus sp (7.22%) and Klebsiella sp (6.0%). Isolates species distribution is shown in Table 2. There was no statistical association between the isolate species and the sampling method (P = 0.237).

Antibiotic susceptibility. Enterobacteriaceae were producing ESBL in 14.1% (15/106), and carbapenemase in 5.6% (6/106) of all samples (swab and deep samples included). 

The resistance rate of Enterobacteriaceae to amoxicillin-clavulanic acid, piperacillin-tazobactam, ertapenem, and ciprofloxacin were 60.4% (64/106), 8.1% (8/98), 5.0% (5/99), and 26.2% (27/103), respectively  (Figure 1). 

Figure 1. Antibiotic susceptibility of Enterobacteriaceae. R: resistant; I: intermediate; S: susceptible

Nonfermenting gram-negative bacilli produced ESBL in 5.1% (2/39) and a carbapenemase in 38.5% (15/39) of samples. The resistance rates of NFGNB to ticarcillin-clavulanic acid was 35.4% (11/31), Piperacillin-tazobactam 36.8% (14/38), ceftazidime 46.8% (15/32), imipenem 32.4% (12/37), gentamicin 36.4% (12/33), and ciprofloxacin 36.8% (14/38) (Figure 2). Methicillin-resistant S. aureus (MRSA) represented 4.7% (2/42) of S. aureus isolates. Resistance rates to other antibiotics are reported in Figure 3. 

Figure 2. Antibiotic susceptibility of nonfermenting gram-negative bacilli.R: resistant; I: intermediate; S: susceptible

Figure 3. Antibiotic susceptibility of Staphylococusaureus isolates. R: resistant; I: intermediate; S: susceptible

Methicillin-resistant coagulase-negative Staphylococcus represented 43.5% (10/23) of isolates. Resistance rates to other antibiotics are reported in Figure 4. The resistance rates of Enterococcus sp to ampicillin and moxifloxacin were 0.0% (0/29) and 7.1% (2/28), respectively (Figure 5). The resistance rates of Streptococcus sp to penicillin and moxifloxacin were 11.4% (4/35) and 5.4 % (2/37), respectively (Figure 6). No Enterococcus sp, Staphylococcus sp, or Streptococcus sp isolates were glycopeptide resistant. 

Figure 4. Antibiotic susceptibility of coagulase negative Staphylococcussp isolates. R: resistant; I: intermediate; S: susceptible

Figure 5. Antibiotic susceptibility of Enterococcus sp isolates. R: resistant; I: intermediate; S: susceptible

Figure 6. Antibiotic susceptibility of Streptococcus sp isolates. R: resistant; I: intermediate; S: susceptible

The slight clinical presentation that can delay diagnosis,^14,15 ischemia,¹⁵ multidrug-resistant bacteria issues, and bone infections (osteomyelitis) are all factors that make the management of DFIs difficult.¹⁴ The male predominance found in this series is described in other studies.^14,16 The average age of patients with a DFI is between 50 years and 60 years, but neither age nor gender appear to influence the isolated pathogen species or the polymicrobial character of the sample.² 

The recommended sampling method for DFI microbiological diagnosis is tissue samples.¹⁷ Even if commonly used, the value of swab sampling is controversial. Some studies report the equivalence of the results obtained in the surface and deep samples,¹⁸ while others argue that only the tissue samples allow the distinction between colonization and true infection.⁹ No statistically significant difference was found between the 2 sampling methods in this study. However, comparison was biased by the retrospective nature of the study; no patient could benefit from both of the types of samples. The isolation of strict anaerobes is also an aspect for which the authors have not been able to establish the effectiveness of both techniques. 

According to the literature data, strict anaerobes are isolated in 6%-79% of DFIs.^14,17-20 This disparity in reported prevalences is attributed to the quality of the samples and the difference in the performance and sophistication of the available techniques.¹⁸ The prevalence is associated with the Wagner Classification System grade, chronicity of the lesion, and prior antibiotic therapy. Diabetic foot infections caused by obligate anaerobes are slower to heal.^18-20 The discrepancy between the direct examination and culture isolates obtained in the authors’ study indicated an anaerobic infection in 28.4% of cases. Technical constraints have not allowed the strict anaerobes isolation in this study. 

Bacteria cultures of DFIs are mostly polymicrobial.^{3,8,14,17,19-26} Whatever the sampling method, more than half of the samples allowed the isolation of 2 or more pathogens. The polymicrobial nature would be more common in chronic or recurrent DFIs with prior antibiotic therapy. The patients with first-time DFIs would have monomicrobial samples with predominance of gram-positive cocci.²⁶

This predominance of gram-positive cocci, including S aureus and β-hemolytic Streptococci, affects all DFIs observed in developed countries. All DFI studies, as far as the authors know, in industrialized countries report gram-positive cocci predominance.^8,14,23 The geographic origin, sociocultural parameters, and level of hygiene of each patient seem to influence the prevalence of the germs.¹⁴ A case series from India and Malaysia reported a high incidence of gram-negative bacteria, especially Enterobacteriaceae and Pseudomonas sp.^19,20,27 

In the authors’ study, GPB were predominant on deep samples, while GNB were most common on swab samples. However, no statistically significant association was found between the sampling method and isolated bacteria genus or species. 

S. aureus was most prevalent regardless of sampling method, and represented 12.6% of all isolated bacteria. Only 4.7% of S. aureus were resistant to methicillin, and it is the species most frequently reported in DFIs in developed countries.^8,25,26,28 Its prevalence was 42.1% in Lavery et al¹⁶ who focused on MRSA which was found in 29.8% of DFIs in the same study sample. S. aureus was also reported in 29% in the cohort of Yates et al,²³ 23% of which were MRSA. In 3 Indian studies,^19,20,27S. aureus accounts for 13%-19% of isolates. In a study by Shanmugam and colleagues,²⁷ MRSA isolates represented 9%. Malecki et al¹⁴ and Al Benwan et al²⁹ reported prevalences close to the current study, with 15% and 11% S aureus. Methicillin-resistant S. aureus represented 6% and 7%, respectively, of S. aureus isolates in Malecki and colleagues’ and Al Benwan and coauthors’ series are MRSA. However, Hannat and coauthors³⁰ reported higher prevalences (56% S aureus and 43% MRSA) in their study, similar to those described in western studies such as Lavery et al.¹⁶ Resistance to vancomycin is rarely reported. 

Enterobacteriaceae represented 31.8% of all isolates (29.3% in deep samples vs 34.3% in surface charges) in the authors’ study. The prevalence of enteric bacteria may be associated with lack of control of fecal matter, precarious hygiene conditions, and elderly patients, due to their higher incidence of chronic medical problems and residence in long-term care facilities, both of which are risk factors for enteric bacterial infections. Proteus sp, followed by E. coli and Klebsiella sp are the most frequently reported species.^3,14,27 In the authors’ study, Enterobacter sp (8.4%) was predominant among Enterobacteriaceae after S aureus. The same finding is reported in the series of Ji et al³¹ with a 16.7% prevalence. The saprophytic character of Enterobacter sp would be a favorable element in the colonization of diabetic feet and also of the contamination of swab samples.

Bacterial epidemiology of DFIs is characterized by the high rate of multidrug resistant organisms. Indeed, in the present study the rate of ESBL among Enterobacteriaceae was 14%. This rate varies from 6%-36%, according to other studies in the literature.^14,23,27,32 Carbapenemase-producing isolates (5.6% in this study) were found in 20% of Klebsiella sp and 40% of Proteus sp.^27,33 Ciprofloxacin-resistant Enterobacteriaceae represented 26.2%, which is less than the reported rate of resistance of up to 90% of the Enterobacteriaceae.^27,33

Pseudomonas sp represented 8.8% in the authors’ study, similar to a 10% prevalence found by Pellizer et al.³⁴Pseudomonas sp are frequently isolated in Indian studies,^19,20,27 and their isolation is associated with chronic DFIs.^8,23,27

The prevalence of Acinetobacter baumannii isolates was low (3.9%). All Acinetobacter baumannii isolates were carbapenemase-producing (eg, 11.1% of Pseudomonas aeruginosa isolates). Resistance to third-generation cephalosporins affected almost half of the NFGNB isolates, and resistance to gentamicin and quinolones was found in one-third of them. The reported antibiotic resistance profile of NFGNB is alarming; Shanmugam et al²⁷ reported 50% resistance to quinolones and third-generation cephalosporins among Acinetobacter baumannii isolates, 50% resistance to quinolones, and 60% resistance to third-generation cephalosporins among P. aeruginosa.²⁷

Initial antibiotic therapy is often empirical, based on epidemiological and clinical data. The physician must consider the clinical severity of the DFI, recent antibiotic use, DFI history, data from local bacterial epidemiology, and factors related to the patient.⁴ Early optimal antibiotic therapy, even before receiving laboratory results, plays an important role in the DFI prognosis as well as in rational insulin therapy and local measures such as revascularization, debridement, decompression of the foot, or other surgical measures.⁸

Many studies have attempted to determine the best combination of antibiotics to treat DFIs. Malecki et al¹⁴ suggests the superiority of associations compared to monotherapy and recommends the combination of amoxicillin-clavulanic acid or piperacillin-tazobactam with an aminoglycoside.¹² Other studies recommend other associations including imipenem, piperacillin-tazobactam, and vancomycin.²⁹ Amoxicillin-clavulanic acid,³⁵ piperacillin-tazobactam,³⁶ moxifloxacin,⁴ and ertapenem³⁷ have been proposed for empirical monotherapy.

In a systematic review, no suggested therapeutic protocol seems to be superior to others.³⁸ 

Considering the microbiological data from the authors’ study and the regional epidemiological and clinical context, empiric antibiotic therapy for DFIs must cover either GNB or GPC. The authors suggest, in order of preference, the use of piperacillin-tazobactam, ticarcillin-clavulanic acid, fourth-generation cephalosporin, ertapenem, and an association of fluoroquinolones with one of the aforementioned antibiotics. (Taking into consideration the high level of resistance to third-generation cephalosporins among GNB that are predominant in this study, the authors do not recommend them for use in treatment.) 

The limitations of this study are those of any retrospective study, including bias due to rigorous clinical definition and sampling method, as well as the lack of data on antibiotic treatments received by patients and about the patients’ clinical courses. 

Although S aureus is the predominant species in DFIs, resistance to methicillin is rare. Isolated GNB prevalence rates appear similar to those reported in developing countries with an alarming rate of multidrug-resistant bacteria. The contribution of surface samples in DFI diagnosis seems comparable to deep samples in the present study. However, considering the small sample size and retrospective study design, this finding should first be confirmed through a rigorous, larger, multicenter prospective comparison study.

From the Bacteriology Department, Mohammed Vth Military Teaching Hospital of Rabat; Traumalogy Department, Mohammed Vth Military Teaching Hospital of Rabat; and Pharmacy Department, Mohammed Vth Military Teaching Hospital of Rabat, Rabat, Morocco

Address correspondence to:
Bouchra Belefquih, MD
Bacteriology Department
Mohammed Vth Military Teaching Hospital
Rabat, Morocco
bbelefquih@yahoo.fr 

Disclosure: The authors disclose no financial or other conflicts of interest.