The Microflora of Chronic Diabetic Foot Ulcers Based on Culture and Molecular Examination: A Descriptive Study

Infection of chronic diabetic foot ulcers (DFUs) is a major concern in patients with diabetes mellitus. Purpose: This prospective, descriptive study was conducted to evaluate clinical wound parameters and to determine the aerobic and anaerobic microflora of DFUs with 1 or more clinical signs of infection using culture and molecular methods. Methods: Patients with a DFU and clinical signs of infection receiving care at a tertiary care hospital in Varanasi, India, were consecutively enrolled. Patient and wound characteristics were assessed, and the cultures obtained were analyzed quantitatively to detect aerobes/facultative anaerobes and by polymerase chain reaction for common anaerobes. If no organisms were found using these methods, sequence analysis of bacterial 16S ribosomal RNA (16SrRNA) was used. Clinical, demographic, and microbial flora variables were compared using the chi-squared test, and predictors of culture results were ascertained using multiple logistic regressions. Results: Forty (40) patients participated. Of those, 30 (75%) had positive culture results with a total of 64 isolates (2.13 isolates/ulcer). The ratio of aerobes to anaerobes was 1.24:1 (35/29); Peptococcus spp was the most frequent isolate (15, 23.4%). Proteobacteria and Firmicutes were detected by 16SrRNA in 6 of the 10 samples (60%), their presence not detected by other methods. Ulcer size <11.84 cm² (OR: 4.71; 95% CI: 0.93-23.68) and ulcer on the dorsum of the foot (OR: 0.92; 95% CI: 0.05-16.42) were significantly associated (P <.05) with monomicrobial microflora. Conclusion: Gram-negative aerobic/facultative anaerobes predominated in the 30 DFUs that exhibited clinical signs of infection. Further experimental studies are required to understand the diverse microorganisms present in DFUs and their potential role in wound infections. 

Diabetes is a rapidly increasing epidemic, especially in developing countries such as India. The World Health Organization¹ estimates a 78% prevalence of diabetes, representing a cause for global concern. Presently, diabetic foot infections (DFIs) are the most common cause of hospitalization among people with diabetes.² With the overwhelming global burden of the disease and its ever increasing prevalence, DFIs remain a major complication in these patients.³ The vast spectrum of DFI may vary from superficial infections to extensive disseminated ulcers and often leads to lower extremity amputation. In one of the first systematic cross-sectional analysis⁴ of diabetic foot ulcers (DFUs), which involved 52 cases of neuropathic, nonischemic, noninfected foot wounds, ulcer depth and duration were associated with an array of microorganisms inhabiting the wound. The study also revealed the heterogeneity of the microbial flora inhabiting the ulcers; the authors concluded that analysis of the entire microbial community, instead of culturing only predominant pathogens, is the best approach for appropriate clinical treatment. A hospital-based, case-controlled study⁵ among 47 patients with a DFU noted risk factors for lower extremity amputations included poor glycemic control (adjusted odds ratio [OR]: 20.47), hypertension (adjusted OR: 3.67), hypertriglyceridemia (adjusted OR: 5.58), and infection (adjusted OR: 12.97). Because the treatment of foot infection often commences by empirically targeting a broader spectrum of organisms, frequent updates on the most probable organisms leading to such infections might provide a better clinical response.⁶

Despite a considerable number of studies on DFI, a review⁷ has shown that consensus is lacking on certain issues such as colonization status of wounds, standard laboratory methods for quantitative bacterial counts, and association of microbiological evidence with clinical outcomes.  In this context, critical colonization (bacterial count >10⁵ colony forming units [CFU]/g of tissue) with clinical signs of infection in DFUs often is cited by specialists in wound management as a reason for delayed wound healing.⁷ However, this concept is not evidence-based, and studies (including a cross-sectional study⁸ of 64 patients with a DFU) have found no definite relationship between quantitative bacterial counts, symptoms, and rate of wound healing. Because the majority of the studies are culture-based, the entire microflora often is not established in ordinary laboratory media, especially in cases of fastidious organisms that require complex nutrition. Recent data using molecular technique⁹ and data from tropical countries¹⁰ revealed a changing spectrum of etiologies responsible for chronic infections in DFU.

The microflora (ie, the aggregative community of genomes present in the site of infection, also known as microbiomes) of DFUs usually consist of a heterogenous mixture of colonizers and pathogens.⁴ Because it has been rightly said that microflora of DFI varies based on “environmental, hygienic, and cultural issues,”⁷ knowing the prevalence and types of microorganisms associated with DFI based on regional variances would be helpful in treatment outcomes.

The purpose of this prospective, descriptive study was to evaluate clinical wound parameters and to determine the aerobic and anaerobic flora of 1 or more clinical signs of infection, using culture and molecular methods, in patients with DFUs attending a tertiary care hospital in Varanasi, India. 

Setting and patients. The study was conducted in the Departments of Microbiology and General Surgery associated with a tertiary care university hospital in North India. All patients with a DFU who were enrolled at the hospital’s wound clinic/completed the registration form and treated were eligible to participate if they met the following inclusion criteria: age 18 to 80 years, clinical diagnosis of infection at the time of presentation, DFUs of >6 weeks’ duration (suggesting chronic nature), and not on antibiotics for at least 2 weeks (antibiotics significantly alter and modify bacterial flora). Clinical suspicion of infection was determined by the presence of 1 or more factors: exudate, redness, discolored ulcer edges, pain, local edema, and foul odor. Patients who were pregnant and patients diagnosed with malignancy or osteomyelitis, sinus leading to bone, an ulcer involving the toe or gangrenous area, and systemic sepsis relating to conditions with severity or risk of increased infection were excluded. Written informed consent was obtained from all participants and 1 family member. The study was approved by the institution’s ethics committee.

Data collection and clinical assessment of the ulcers. Patient demographic data (including age, gender) and clinical history (chief complaints, past medical history, current medication) of diabetes were collected from every participant based on a pretested, predesigned questionnaire. Details such as name and contact information were avoided to maintain patient anonymity; data were recorded using codes. Information regarding the ulcers included anatomical location, surface area calculated by multiplying the greatest length with the perpendicular greatest width as measured by ruler, ulcer duration, and duration of diabetes; digital images were taken at the first visit for assessment and documentation. Relevant laboratory investigations to ascertain glycemic control, anemia, liver function (measurement of levels of serum glutamic-oxaloacetic transaminase, serum glutamic pyruvic transaminase, alkaline phosphatise, and serum bilirubin), and renal function (serum urea, creatinine and uric acid) were noted. The data were accessible only to the research team of this study.

Sample collection and wound culture. Samples were collected both by Levine’s technique¹¹ and needle aspiration. Although tissue biopsy could have provided both qualitative and quantitative data, this invasive technique was not applied due to the added risk for wound infection. Instead, the swab technique was employed, which has been reported in a systematic review¹² to provide results comparable to biopsy without being invasive. 

Before obtaining the sample, the wound was cleansed with sterile gauze moistened with sterile normal saline to remove surface contamination. Subsequently, using Levine’s technique, a sterile cotton swab was rotated over a 1 cm² area for 5 seconds with sufficient pressure to cause tissue fluid to be expressed. The tip of the swab was broken off in a sterile transport media (phosphate-buffered saline [PBS]) and transported immediately to the hospital’s microbiology laboratory. Tissue fluid also was collected by sterile syringe in sterile vials and sent for culture. The samples were divided into 2 parts — one for quantitative aerobic bacterial culture and the other for detection of common anaerobes by polymerase chain reaction (PCR).

The laboratory processed the samples per protocol. All samples were gram stained for direct examination. For culture, cotton swabs in PBS were thoroughly vortexed before quantitative culture and then serially diluted in normal saline and plated on MacConkey’s agar media and blood agar media (Hi Media Pvt Labs, Mumbai, India). Tissue fluid was processed in the same manner. After overnight incubation, the plates were visually inspected and colonies of each type of bacteria were counted. Plates with 50 to 500 colonies were considered readable and the exact number of colonies was counted.¹³ CFUs then were calculated to determine the total bacterial count on each plate using the formula: 

CFU/mL = (number of colonies x dilution factor)/volume of culture plate.

CFU >1 x 10⁵ was considered critical colonization/infection and counts below were considered benign colonization.¹¹

Bacterial colonies that grew aerobically were identified and reported using standard biochemical tests.¹⁴ Gram stain findings and swab culture of tissue fluids were corroborated for uniformity.

PCR for detection of anaerobes. PCRs were performed to detect common anaerobes (ie, Bacteroides, Clostridium, Peptostreptococcus, Peptococcus, and Ruminococcus) because an anaerobic culture facility was not available. These anaerobes were selected for study based on their predominance in DFUs from previous research.^4,6,7,10 DNA was isolated from the tissue fluid samples using QIAamp DNA mini kit (QIAGEN, Hilden, Germany) according to the manufacturer’s instructions. The quality of the DNA was measured by NanoDrop analysis (Thermo Fisher Scientific, Waltham, MA). PCR amplifications were performed by using primer sets from a previous study with slight modifications.¹⁵ The primer sequences and the reaction conditions are described in Table 1.^14,15

16SrRNA gene sequencing. Samples that did not yield/demonstrate any aerobic or anaerobic flora were subjected to 16S ribosomal RNA (16SrRNA) PCR and sequencing¹⁶ to confirm whether they were bacteriologically sterile (see Table 1).^14,15 The sequences from 16SrRNA PCR positive samples were analyzed using the mothur package version 1.23.¹⁷ Based on instructions, the sequences were screened and aligned to the SILVA reference set of NAST aligner of mothur for identification of the bacteria. Bacterial identification was done based on designated operational taxonomic units of the sequences. 

Antibiotic susceptibility testing. The modified Kirby-Bauer disc diffusion method was used to test antibiotic susceptibility of the aerobic isolates per the 2017 Clinical and Laboratory Standards Institute guidelines.¹⁸ The antibiotics used in this study are listed in Table 2. Isolates showing resistance to ≥1 antimicrobial agent in ≥3 antimicrobial categories were considered multidrug resistant (MDR), and resistance to ≥1 antimicrobial agent in all but ≤2 antimicrobial categories were included as extensively drug resistant (XDR).¹⁹

Data and statistical analysis. All patients with clinical signs of infections and positive for microorganisms were divided into 2 groups based on culture and PCR results. One group comprised cases with 1 type of microflora (monomicrobial) and the other group comprised of cases with more than 1 type of microflora (polymicrobial). Quantitative variables were denoted by mean ± standard deviation, and qualitative variables were shown as percentages. The clinical parameters of the DFU were correlated with the monomicrobial and polymicrobial culture results by chi-squared test using SPSS, version 19 (IBM Corporation, Armonk, NY). Multiple logistic regressions were used to identify independent predictors of monomicrobial and polymicrobial infections. Comorbid factors (eg, anemia and abnormalities in liver and renal function) were not considered for analysis. The critical value of P, indicating the probability of significant difference, was taken as <.05 for comparison.

Demographic and clinical characteristics. Participants included 40 consecutive, nonduplicate patients (22 men, 18 women; mean/median age 55 [range 25 – 70] years). Mean ulcer size was 18.34 ± 16.25 cm² (range 1.1 cm² – 145 cm²), and mean ulcer duration was 18.16 ± 19.8 weeks (range 6 weeks to more than 2 years). Mean duration of diabetes mellitus was 60 months (range 6 months to 15 years). Mean HbA1C value (an indicator of glycemic control) was 8.49 ± 1.39. Patient details, including comorbidities and signs of infection, are summarized in Table 3.

Detection of microbial flora. A total of 30 (75%) samples showed evidence of microbial presence either by aerobic culture or PCR for anaerobes (see Figure 1) or both. Among these, 64 isolates were identified, giving a mean of 2.13 species per ulcer. Although 15 samples (37.5%) were positive only for aerobic organisms, 1 (2.5%) showed only anaerobic flora and 14 (35%) were positive for both aerobic and anaerobic organisms. The ratio of aerobes to anaerobes was 1.24:1 (35/29). Peptococcus spp was the most frequent isolate (15, 23.4%). All organisms isolated by aerobic culture method had a mean colony count 14.09 x 10⁵ ± 12.16 above 10⁵ CFU/mL, with counts ranging from 5 x 10⁵ to >10⁶ CFU/mL. The bacteriological analysis of the wounds is summarized in Table 3.

Ten (10, 25%) samples did not indicate the presence of organisms either on aerobic media or by PCR. However, 6 of these samples showed the presence of bacteria using 16SrRNA PCR. Using sequencing and alignment of the sequences, organisms could be classified into 2 phyla: Proteobacteria and Firmicutes. Although 1 sample contained unclassified organisms, 1 showed the presence of Streptococcus spp and 1 showed Alcaligenes faecalis. The remaining samples contained organisms closely related to Sphingomonas spp, Streptococcus spp, and A faecalis. 

Antibiotic susceptibility testing. The isolates demonstrated resistance to a number of commonly used antibiotics. Of the 28 aerobic isolates, 21 (75%) were found to be MDR. One XDR each was reported in Pseudomonas aeruginosa and Klebsiella pneumoniae, and 2 isolates of Escherichia coli also were found to be XDR. The antibiotic resistance pattern of the isolates is shown in Table 4.

Microbial flora and clinical factors. Significant differences (P <.05) in the presence of monomicrobial flora were observed for smaller (<11.84 cm²) versus larger ulcers (OR: 4.71; 95% CI: 0.93-23.68) and for wounds located on the dorsum of the foot (OR: 0.92; 95% CI: 0.05-16.42)  (see Table 5). 

According to a cross-sectional study²⁰ among 32 patients with neuropathic ulcers, the burden of microbial flora in DFI is one of the most important factors related to the pathophysiology of DFUs. Owing to the fact that DFUs are a growing public health concern and that microorganisms play a major role in their delayed healing, a complete knowledge of the prevailing microflora in these DFUs is essential. This study revealed a wide array of aerobic and anaerobic colonizers and pathogens in infected ulcers of patients with diabetes.

A quantitative culture was performed to assess the load of aerobes and facultative anaerobes in DFUs with clinical signs of infection to delineate problematic pathogens from benign colonizers, because these organisms also are isolated as the most common contaminants. The presence of these organisms in significant number in these ulcers suggests their probable pathogenic role. A large amount of the literature considers culture-based methods the way to establish the flora of ulcers with suspected infection. Although such methods are cost effective, they have the serious disadvantage of underreporting organisms, especially those that are fastidious.¹³ On the other hand, community profiling based on the 16SrRNA bacterial gene provides a comprehensive picture of the microflora of DFU and may reveal the complex events that occur sequentially from colonization to differentiation of niche and formation of heterogenous biofilms.²¹ However, such methods are more than 10 times as costly as culture-based methods. 

This study employed a cost-effective approach to characterize the microflora of DFUs. The predominant aerobic/facultative anaerobic pathogens were isolated and quantitatively identified by conventional culture. Because capability for anaerobic culture is limited in many laboratories, including the authors’, conventional PCR targeting major anaerobes present in DFUs were sought. Gene profiling using 16SrRNA was performed for culture and PCR- negative but clinically infected wound samples, revealing the constraints of a culture-based reporting system. The majority (60%) of the presumed sterile samples were found to harbor complex microbiota of unclassified and unusual microorganisms that were not detectable by routine laboratory practices (only culturable and viable organisms are easily identified by this method). Although the pathogenic role of these microcommunities could not be definitely ascertained, which is in turn a major limitation with microflora study, their presence in wounds with clinical signs of infection demonstrated the diverse microbial nature of such type of ulcers along with their complex community structure. The study showed the presence of anaerobes as mixed flora along with aerobes/facultative anaerobes. The presence of other anaerobes could not be ruled out in this study, because few anaerobes were targeted through molecular methods and culture capability was not available. Antibiograms were performed only for aerobes, but the widespread presence of mixed anaerobic flora in the clinically infected wounds warrants their probable pathogenic role and need for presumptive treatment.

A cross-sectional study⁴ involving 52 cases of neuropathic DFU showed depth and duration of DFUs are directly linked with the type of microbial flora inhabiting the ulcer. This study⁴ found ulcers that were deep and of longer duration were associated with higher levels of anaerobes and Proteobacteria, while on the other hand, superficial and short-duration ulcers were associated with predominance of gram-positive cocci. In the current study, gram-negative organisms E coli and Klebsiella spp were the most common aerobic flora, while Peptococcus spp was the most common anaerobic flora. As experimental studies²² involving in depth characterization of microbial flora using several molecular methods have shown, in the presence of pathogenic bacteria in the deepest part of the ulcers, a meticulous sample collection is important to delineate DFU microflora. Studies conducted in India have shown the changing epidemiology of microflora of DFUs with predominant gram-negative flora^23-25 instead of gram-positive pathogens such as Staphylococcus and Streptococcus spp as noted in samples obtained in Western countries.²⁶ In one of the first comprehensive surveys²⁶ on microbial flora of hospitalized patients of DFU in India, gram-negative aerobic bacteria were found to be the most frequent isolates. A prospective study²⁴ on 105 patients with DFI reported gram-negative bacteria as the causative organism in 70.8% cases. Similarly, a hospital-based prospective study²⁵ documented the abundance of these organisms in the infected DFUs of 75 patients. However, most of this research is based on the results of aerobic culture; therefore, no data on anaerobic organisms is provided. In this context, it is interesting to note that several factors have been studied that affect microflora, including antibiotic use.⁹ Systemic and topical use of antibiotics alter the structure of the bacterial community in chronic wounds by reducing gram-positive flora such as Streptococcaceae and increase gram-negative flora such as Pseudomonadaceae.²⁷ This factor is reflected in the majority of the Indian studies, including the present study, reporting predominantly gram-negative flora; this might be due to unrestricted antibiotic use in India as compared to more developed countries (in the west). Additionally, the majority (75%) of the cultured isolates were MDR, revealing their capacity to persist. Such infections also have been previously implicated in the DFUs of hospitalized patients in India, where 72% of the infected DFU cases were MDR.²³

In the current study, risk for monomicrobial and polymicrobial infections was assessed based on clinical parameters. In this study, smaller ulcers contained significantly less diverse microflora than larger ulcers. As the usual trend for treating DFUs is based on the assumption of the dominant flora, it may be advisable to manage small-sized ulcers with antibiotics targeting less diverse flora as compared to bigger ulcers where a broad-spectrum antibiotic may be required. 

In the current study, anaerobic culture could not be performed and only a few anaerobes were targeted by PCR. Bacterial profile based on the 16SrRNA gene, which could have better predicted the community structure, could not be performed for all the samples due to financial constraints. 

A prospective, descriptive study of patients with DFU and clinical signs of infection was conducted in a tertiary care hospital in India to evaluate clinical wound parameters and to determine the aerobic and anaerobic flora. Gram-negative bacteria were found to be the predominant bacterial strain by wound culture (E coli [30%] and Klebsiella [30%]). Among anaerobes, Peptococcus was most common (23.4%). This knowledge is important in guiding clinicians for treatment of such infections. Smaller ulcers (<11.84 cm²) were significantly more likely to contain monomicrobial and/or less diverse flora. The study also revealed that DFUs with clinical signs of infection might show false negative culture results due to presence of uncommon organisms that can be detected only by molecular methods. Further experimental studies on larger sample sizes and by advanced molecular methods would help in determining the wide spectrum of emerging microorganisms responsible for such infections and aid in their effective management. 

Dr. Banerjee is an Associate Professor, Dr. Das is a senior resident, and Ms. Singh is a research scholar, Department of Microbiology; Dr. Bansal is a senior resident and Dr. Basu is a Professor, Department of General Surgery, Institute of Medical Sciences, Banaras Hindu University, Varanasi, India. Please address correspondence to: Tuhina Banerjee, MD, PhD, Associate Professor, Department of Microbiology, Institute of Medical Sciences, Banaras Hindu University, Varanasi-221005, India; email: drtuhina@yahoo.com.