Pathogens Causing Diabetic Foot Infection and the Reliability of the Superficial Culture

Abstract

Background:

Diabetic foot infection (DFI) is a severe complication of diabetes and a leading cause of hospitalization in the population with diabetes. Empirical intravenous antibiotic agents are initiated according to guidelines based on previously published data on typical pathogens. Therefore, regular evaluation of the pathogens in DFI and their resistance is important to validate current therapies. We evaluated the most current data on bacterial cultures in patients treated at our hospital for DFI and the resistance to the most common antibiotic agents, as well as the reliability of superficial cultures compared with deep tissue cultures.

Patients and Methods:

This retrospective study was performed at the University Hospital of Tampere and comprised 325 patients with 405 hospitalizations for DFI during the years 2010–2014.

Results:

The most frequent pathogens in superficial and deep samples were Staphylococcus aureus (36.9%), gram-negative bacilli (24.6%), and β-hemolytic streptococci (BHS, 19.5%). Septicemia was caused most often by Staphylococcus aureus and BHS (34.6% each). The specificity of superficial culture was 91.8%–92.8% and sensitivity 66.7%–87.5%.

Conclusions:

This study indicates the need to cover Staphylococcus aureus, BHS, and gram-negative bacilli when treating DFI. The reliability of superficial culture was surprisingly good.

Diabetes and its complications are a substantial burden for the healthcare system, and diabetic foot infections (DFIs) are the major cause of costs [1]. The prevalence of diabetes has increased over the last few decades and is expected to continue increasing [2]. Treatment of DFIs is time-consuming because of long intravenous antimicrobial therapies and often multiple surgical procedures. The length of hospital stay is the most crucial factor considering the cost of treatment, and the second greatest contributor to cost is the antimicrobial treatment [3].

Pathogens causing DFIs are usually identified by obtaining bacterial cultures. However, antimicrobial therapy is often initiated before the culture results are available. Therefore, up-to-date knowledge of the usual pathogens is important for choosing the best empirical antimicrobial therapy.

Deep tissue biopsy culture is considered to be the gold standard for identifying chronic wound pathogens [4]. There are different methods of obtaining cultures, including sterile swabs or needle aspirations from purulent wound drainage and tissue biopsy with curette or knife. In earlier studies, superficial cultures were estimated to have mediocre sensitivity and specificity [5]. However, some other studies indicate that superficial cultures can provide clinically relevant data [4,6,7].

This retrospective cohort study was performed to evaluate the contemporary pathogens of DFIs and their bacterial resistance. Our secondary intention was to evaluate the reliability of superficial culture compared with deep tissue or bone culture in detecting these pathogens in patients with DFIs.

Patients and Methods

The study was conducted at Tampere University Hospital after Institutional Review Board approval (ETL-code R14545S). The study cohort consisted of all patients with DFI who required hospitalization between January 1, 2010, and December 31, 2014, with a minimum of five years of follow-up for studying long-term survival and costs after DFI to be reported later. In this article we report microbiologic findings of this data. No changes in microbiologic guidelines or swab culture techniques have been made within the last five years.

Patient identification

Electronic medical records were searched for patients treated for DFI with at least one of the following criteria: diabetes diagnosis and C-reactive protein (CRP) ≥100 at any point during the hospitalization; any level of lower extremity amputation; or any lower extremity wound debridement in the operating room with at least one bacterial culture within 30 days after the operation. Identified cases (defined as a period of hospitalization) were reviewed manually to include only confirmed DFIs in the present study. Patients with ischemic or neuropathic wound without acute bone or soft-tissue infection and patients with soft tissue infection not located in the foot (e.g., erysipelas or cellulitis above the ankle) were excluded (Fig. 1). After the computer search and manual revision, identified cases were cross-checked with the existing registry of DFIs (years 2012–2014) to ensure that the computer algorithm had found all known cases.

FIG. 1.

Flowchart of case identification.

The following data were collected from patient records: patient demographics, registered diagnoses, length of hospital stay, and microbiologic and clinical chemistry data from one week prior to hospitalization to the end of the hospital stay. Wound status was defined manually from patient records using the University of Texas Staging System for Diabetic Foot Ulcers (UT scale). One case was excluded because staging the wound retrospectively from the records was not possible (Fig. 1).

Materials and measures

During the study period, the following methods were used routinely in our hospital to obtain bacterial cultures. Superficial culture was collected with a standard sterile cotton swab after cleaning the wound. Swab specimens were collected according to the local guidelines by a specialist nurse. To avoid contaminating bacterial flora, wounds were debrided from necrotic or non-viable tissue or slough and rinsed with tap water or saline before swabbing. Deep tissue culture was collected after wound debridement using a dermal curette or by excisional biopsy. Bone culture was obtained in the operating room after wound debridement using a separate set of instruments not used for the debridement. Blood cultures were collected as a venous blood sample and placed immediately into a blood culture flask. Cultures from superficial swabs are always more or less an interpretation of the sample because a specialist in the laboratory is choosing to report only clinically important bacterial strains, and others are classified as normal bacterial flora. Superficial swabs also identify only aerobic bacteria, not anaerobic bacteria.

We evaluated total resistance to common antibiotics based on antibiotic sensitivity testing and the inherent antibiotic resistance. An assumption of resistance was not possible in only a few cases, and the resistance was classified as undefined.

Statistical analysis

For statistical analysis, deep tissue and bone cultures were classified as deep culture. We also categorized bacteria into groups based on their role in clinical settings. The frequencies of different pathogen groups were recorded according to the sample type (superficial, deep, blood). The frequencies of resistant bacteria were recorded separately for every sample type. We used cross-tabulation to compare superficial and deep cultures and evaluated the significance between groups using Pearson χ² test or Fisher exact test as appropriate. All statistical analyses were performed by a statistician using the SPSS software package (IBM SPSS Statistics, version 24.0. Armonk, NY).

Results

A total of 325 patients (238 [73.2%] males and 87 [26.8%] females) with 405 periods of hospitalization fulfilled the inclusion criteria and were included in the analysis. A case was defined as one period of hospitalization. The median age of the patients was 67 years (IQR 17.4). The comorbidities are presented in Table 1. Bacterial cultures were obtained during 261 of 405 (64%) of hospitalization periods with a total of 456 individual samples collected (Table 2).

Table 1.

Comorbidities in the Study Cases

Comorbidities (n = 405)	n (%)
Hypertension	169 (41.7%)
Atherosclerosis	267 (65.9%)
Ischemic heart disease	113 (27.9%)
Chronic obstructive chronic pulmonary disease	21 (5.2%)
Renal failure	83 (20.5%)
Heart failure	84 (20.7%)
Dyslipidemia	25 (6.2%)

Table 2.

Number of Bacterial Cultures by Sample Type

Sample type	Cultures obtained, n (% of cases)
Superficial	160 (39.5%)
Deep tissue	113 (27.9%)
Bone	46 (11.4%)
Blood	137 (33.8%)
Deep (deep tissue and bone combined)	120 (29.6%)
Superficial and deep sample obtained	66 (16.3%)

Staphylococcus aureus was the most frequent pathogen in superficial and deep samples (36.9%), followed by gram-negative bacilli (24.6%) and β-hemolytic streptococci (BHS; 19.5%, Table 3). Staphylococcus aureus and BHS were the most common cause of septicemia (34.6% each). Pseudomonas is included in the gram-negative bacilli group in our analyses. Pseudomonas was found in nine (6.6%) superficial and 10 (9.0%) deep cultures. Anaerobes were present in 10.1% of deep and 10.7% of blood cultures, respectively.

Table 3.

Frequencies of the Most Common Pathogens According to Different Sample Types

	Superficial	Deep	Combined^a	Blood
Pathogen	n = 136 (%)	n = 111 (%)	n = 195 (%)	n = 26 (%)	p ^b
Staphylococcus aureus	62 (45.6%)	36 (32.4%)	72 (36.9%)	9 (34.6%)	0.0375
BHS	37 (27.2%)	25 (22.5%)	38 (19.5%)	9 (34.6%)	0.4615
Gram-negative bacilli	38 (27.9%)	38 (34.2%)	48 (24.6%)	5 (19.2%)	0.3324
CoNS	5 (3.7%)	19 (17.1%)	20 (10.3%)	1 (3.8%)	0.0004
Other	27 (19.9%)	41 (36.9%)	47 (24.1%)	4 (15.4%)	0.0040

BHS = β-hemolytic streptococci; CoNS, coagulase-negative staphylococci.

Superficial and deep sample combined.

Superficial vs deep samples, χ² test.

Table 4 presents the total antibiotic resistance in the deep and blood cultures based on antibiotic sensitivity tests and known inherent antibiotic resistance. Methicillin-resistant Staphylococcus aureus (MRSA) was identified in nine (6.6%) superficial and two (1.8%) deep tissue cultures. Methicillin-resistant Staphylococcus aureus was not found in any bone or blood culture.

Table 4.

Total Resistance within the Deep Culture and Blood

	Sensitive, n (%)	Resistant, n (%)	Undefined, n (%)
Deep sample (n = 137)
Cefuroxime	93 (67.9%)	43 (31.4%)	1 (0.7%)
Ceftriaxone	95 (69.3%)	41 (29.9%)	1 (0.7%)
Clindamycin	74 (54.0%)	63 (46.0%)	0
Levofloxacin	94 (68.6%)	33 (24.1%)	10 (7.3%)
Piperacillin-tazobactam	93 (67.9%)	38 (27.7%)	6 (4.4%)
Oxacillin	69 (50.4%)	63 (46.0%)	5 (3.6%)
Blood sample (n = 28)
Cefuroxime	21 (75.0%)	7 (25.0%)	0
Ceftriaxone	22 (78.6%)	6 (21.4%)	0
Clindamycin	21 (75%)	7 (25.0%)	0
Levofloxacin	24 (85.7%)	4 (14.3%)	0
Piperacillin-tazobactam	20 (71.4%)	8 (28.6%)	0
Oxacillin	16 (57.1%)	9 (32.1%)	3 (10.7%)

Deep tissue and superficial cultures were available in 66 cases. In deep tissue cultures, BHS were present in 15 cases, Staphylococcus aureus in 24 cases, and gram-negative bacilli in 17 cases. To evaluate the reliability of superficial culture, deep culture was considered a gold standard. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of superficial culture were calculated for BHS, Staphylococcus aureus, and gram-negative bacilli (Table 5).

Table 5.

Sensitivity, Specificity, and Predictive Values of Superficial Samples Compared with Deep Samples

Pathogen	Sensitivity	Specificity	PPV	NPV
BHS	66.7	92.2	71.4	90.4
Staphylococcus aureus	87.5	92.9	87.5	92.9
Gram-negative bacilli	76.5	91.8	76.5	91.8

PPV = positive predictive value; NPV = negative predictive value; BHS = β-hemolytic streptococci.

Cases were divided into two groups based on the UT scale evaluation of the wound. Group B consisted of wound infections without ischemia, and wounds in group D were infected and ischemic. Gram-negative bacilli were more prevalent in ischemic wounds (p < 0.005). Table 6 presents the frequencies of both groups considering superficial samples, deep samples, and both combined.

Table 6.

Bacterial Frequencies in the Non-Ischemic (B, n = 165) and Ischemic (D, n = 239) Groups

	Superficial, n (%)			Deep, n (%)			Combined, n (%)
	B	D	p^a	B	D	p^a	B	D	p^a
Staphylococcus aureus	29 (46.0)	33 (45.8)	1.000	21 (35.6)	15 (28.8)	0.543	50 (41.0)	48 (38.7)	0.795
BHS	22 (34.9)	15 (20.8)	0.083	12 (20.3)	13 (25.0)	0.651	34 (27.9)	28 (22.6)	0.380
Gram-negative bacilli	10 (15.9)	28 (38.9)	0.004	16 (27.1)	22 (42.3)	0.111	26 (21.3)	50 (40.3)	0.001
CoNS	4 (6.3)	1 (1.3)	0.184	13 (22.0)	6 (11.5)	0.207	17 (13.9)	7 (5.6)	0.032

BHS = β-hemolytic streptococci; CoNS = coagulase-negative staphylococci.

Fisher's exact test

Discussion

In this study, we compared bacterial findings from superficial and deep culture samples of DFIs. The study had three major findings. First, the sensitivity and specificity of superficial culture were good compared with deep tissue samples. Second, Staphylococcus aureus, BHS, and gram-negative bacilli are the most common bacteria isolated from cultures in our hospital, and most of the bacteria were sensitive to the most commonly used empirical antimicrobial regimens for DFIs.

Staphylococcus aureus was the most prevalent finding in concordance with numerous previous studies [8 –10]. It was also responsible for every third bacteremia in this study. Beta-hemolytic streptococci was the most common cause of bacteremia (34.6%) along with Staphylococcus aureus, even though the frequency of BHS was lower in superficial (21.9%) and deep tissue (15.7%) cultures. Diabetes has been associated with a higher risk of bacteremia because of hemolytic streptococci [11]. The recommendation of the International Working Group on the Diabetic Foot (IWGDF) is always to cover Staphylococcus aureus and BHS when choosing empirical antibiotic treatment for DFIs [12].

In previous studies, gram-negative bacilli have accounted for 57%–60% of all bacterial isolates in DFIs [8,10,13]. In a recent Scandinavian study, gram-negative bacteria were present in 44% of DFI cases [14]. In our data, the frequency of gram-negative bacilli was 24.6%. We compared the bacterial flora between ischemic and non-ischemic infections and the frequency of gram-negative bacilli was higher in ischemic wounds. Ischemic wounds are prone to necrosis, which may predispose to mixed infection and more frequent presence of gram negative bacilli. Long duration of the wound and previous antibiotic treatments have been associated with mixed infections in diabetic wounds [15]. These factors were not, however, assessed in the present study. Both ischemic foot and the presence of gram-negative bacilli are predictors of amputation [16]. Although the clinical importance of vascular status when treating DFI is commonly known, not too many studies have compared bacterial incidence between ischemic and non-ischemic DFIs. Our data recommend covering gram-negative bacilli when choosing empirical antibiotic therapy for ischemic DFIs, as the IWGDF guidelines also suggest [12].

In our data, coagulase-negative staphylococci (CoNS) were rare in superficial cultures but were present in >10% of deep cultures (Table 3). This is most probably because in superficial cultures, CoNS is not specifically reported. Coagulase-negative staphylococci have been considered merely as colonizing flora and have little clinical importance in diabetic foot ulcers. However, according to a recent study, CoNS (specifically Staphylococcus epidermidis) may be the sole pathogen in diabetic osteomyelitis [17]. Considering patients with osteomyelitis, superficial swabs have been found to be unreliable compared with percutaneous bone biopsies [18]. Therefore, obtaining deep cultures is advisable, especially if osteomyelitis is suspected.

The overall resistance to cephalosporins, levofloxacin, and piperacillin-tazobactam was relatively low, with a 25%–30% resistance rate, whereas resistance to oxacillin and clindamycin (46% for both) was more common. This reflects the low number of MRSA infections and the relatively high proportion of gram-negative bacilli and enterococci found in our study. Current IWGDF guidance [12] recommends selecting an antibiotic based on the severity of infection and targeting empiric therapy at Staphylococcus aureus and BHS in mild infections. The cases in the present study comprise more severe infections, supporting the IWGDF recommendation of β-lactams with β-lactamase inhibitors or cephalosporins in moderate and severe cases.

Clindamycin resistance in Staphylococcus aureus and BHS isolates was low in our study (6.5% and 12.0%, respectively). Narrow-spectrum antibiotic treatment covering gram-positive cocci is recommended in mild DFIs [12,19]. Based on our data, clindamycin may be used as an alternative to β-lactam antibiotic agents in mild DFIs. However, gram-negative bacilli and enterococci were relatively frequent, especially in ischemic wounds. Because these bacteria are inherently resistant to clindamycin, it may not be suitable as monotherapy in more severe DFIs.

In this study, the sensitivity of superficial culture was good for Staphylococcus aureus, BHS, and gram-negative bacilli. There is some perception that superficial cultures are reliable compared with deep tissue cultures in cases of DFI without bone infection [6]. However, the sensitivity of superficial swab culture in detecting the pathogens causing bone infection accurately is <40% [20]. In a systematic review, the estimated sensitivity and specificity of superficial culture was 49% and 62%, respectively [5]. However, this review assessed the overall predictive values for all organisms together, although they may be different for every organism. Our data indicate that the sensitivity of a superficial culture in detecting Staphylococcus aureus (87.5%) is fairly good, but is clearly poorer for BHS (66.7%). A recent study reported similar overall sensitivity of 62.5% for streptococci, but only 75% for Staphylococcus aureus [21].

In most DFI cases, antibiotic therapy has to be initiated empirically before the bacterial culture results are available. The resistance data may be useful in cases in which the clinical course seems to be unfavorable. However, whether clinically efficient antibiotic therapy should be changed according to the culture results is an important aspect to ponder. In a multicenter study in the United Kingdom, the blinded clinician panel recommended a change in the antibiotic therapy between the wound swab and tissue sample in the majority of cases [22]. However, the question remains of whether this change would lead to a better outcome for the patient or just an unnecessary use of broad spectrum antibiotics. Streamlining antibiotic therapy according to the culture results in the case of a favorable clinical response could be considered, especially when switching from parenteral to oral route [19].

The strength of this study was that our data include all of the hospitalized patients with DFIs treated in our hospital's catchment area, which has a population base of approximately 525,000. All hospitalized patients with DFIs were included in the study, regardless of the department in which they were treated. Our data also include all patients with DFIs over five years, which is a large number of patients. Both superficial and deep cultures were obtained in 66 cases. The cohort size in earlier studies comparing deep and superficial cultures has varied between 26 and 100 [5].

A limitation of this study is its retrospective nature. Some cases had no bacterial cultures obtained during hospitalization. In addition, many cases did not include both superficial and deep cultures. This study also includes only cases treated in a tertiary hospital department, whereas a large proportion of DFIs are also treated in primary, secondary, or outpatient care. However, as the most severe cases of DFI are usually treated in tertiary hospital departments, our study provides targeted data about that population.

In conclusion, our study shows that Staphylococcus aureus, BHS, and gram-negative bacilli are the most frequent causes of DFI in our hospital. International guidelines recommend covering Staphylococcus aureus and BHS in antibiotic treatment of DFIs, which is supported by the present study. In addition, in most cases, covering gram-negative bacilli is also recommended. Furthermore, our data indicate that superficial culture is quite reliable when obtained correctly and osteomyelitis is absent. However, obtaining deep cultures is recommended in most cases when possible.

Footnotes

Funding Information

This work was funded by Competitive State Research Financing of the Expert Responsibility area of the Tampere University Hospital, grant 9R025 and 9S025. None of the authors have any conflict of interest regarding the content of this article.

Author Disclosure Statement

No competing financial interests exist.

References

Hicks

, Selvarajah

, Mathioudakis

, et al. Burden of infected diabetic foot ulcers on hospital admissions and costs. Ann Vasc Surg, 2016; 33:149–158.

International Diabetes

Federation

. IDF Diabetes

Atlas

, 8th ed. Brussels, Belgium: International Diabetes

Federation

, 2017. www.diabetesatlas.org (Last accessed July 16, 2020).

Tan

, Hong

, Shen

, et al. Costs of Patients admitted for diabetic foot problems. Ann Acad Med Singapore, 2015; 44:567–570.

Haalboom

, Blokhuis-Arkes

MHE

, Beuk

, et al. Wound swab and wound biopsy yield similar culture results. Wound Repair Regen, 2018; 26:192–199.

Chakraborti

, Le

, Yanofsky

. Sensitivity of superficial cultures in lower extremity wounds. J Hosp Med, 2010; 5:415–420.

Slater

, Lazarovitch

, Boldur

, et al. Swab cultures accurately identify bacterial pathogens in diabetic foot wounds not involving bone. Diabet Med, 2004; 21:705–709.

Stallard

. When and how to perform cultures on chronic wounds?. J Wound Ostomy Cont Nurs, 2018; 45:179–186.

Commons

, Robinson

, Gawler

, et al. High burden of diabetic foot infections in the top end of Australia: An emerging health crisis (DEFINE study). Diabetes Res Clin Pract, 2015; 110:147–157.

Akhi

, Ghotaslou

, Asgharzadeh

, et al. Bacterial etiology and antibiotic susceptibility pattern of diabetic foot infections in Tabriz, Iran. GMS Hyg Infect Control, 2015; 10:Doc02.

10.

Henig

, Pogue

, Cha

, et al. Epidemiology of diabetic foot infection in the metro-Detroit area with a focus on independent predictors for pathogens resistant to recommended empiric antimicrobial therapy. Open Forum Infect Dis, 2018; 5:ofy245.

11.

Hatipoglu

, Mutluoglu

, Turhan

, et al. Causative pathogens and antibiotic resistance in diabetic foot infections: A prospective multi-center study. J Diabetes Complications, 2016; 30:910–916.

12.

IWGDF Guidelines. Infection guideline. May 2019. https://iwgdfguidelines.org/infection-guideline/ (Last assessed October 2, 2019).

13.

, Qi

, Yuan

, et al. Microbiological profile and clinical characteristics of diabetic foot infection in northern China: A retrospective multicentre survey in the Beijing area. J Med Microbiol, 2018; 67:160–168.

14.

Jääskeläinen

, Hagberg

, Forsblom

, Järvinen

. Microbiological etiology and treatment of complicated skin and skin structure infections in diabetic and nondiabetic patients in a population-based study. Open Forum Infect Dis, 2017; 4:ofx044.

15.

Uçkay

, Aragón-Sánchez

, Lew

, Lipsky

. Diabetic foot infections: What have we learned in the last 30 years?. Int J Infect Dis, 2015; 40:81–91.

16.

Saltoglu

, Yemisen

, Ergonul

, et al. Predictors for limb loss among patient with diabetic foot infections: An observational retrospective multicentric study in Turkey. Clin Microbiol Infect, 2015; 21:659–664.

17.

Aragón-Sánchez

, Lázaro-Martínez

, Hernández-Herrero

, et al. Clinical significance of the isolation of Staphylococcus epidermidis from bone biopsy in diabetic foot osteomyelitis. August 27, 2010. Diabet Foot Ankle [Epub ahead of print; DOI: 10.3402/dfa.v1i0.5418.]

18.

Senneville

, Melliez

, Beltrand

, et al. Culture of percutaneous bone biopsy specimens for diagnosis of diabetic foot osteomyelitis: Concordance with ulcer swab cultures. Clin Infect Dis, 2006; 42:57–62.

19.

Lipsky

, Berendt

, Cornia

, et al. 2012 Infectious Diseases Society of America Clinical Practice Guideline for the Diagnosis and Treatment of Diabetic Foot Infections. J Am Podiatr Med Assoc, 2013; 103:2–7.

20.

Elamurugan

, Jagdish

, Kate

, Chandra Parija

. Role of bone biopsy specimen culture in the management of diabetic foot osteomyelitis. Int J Surg, 2011; 9:214–216.

21.

Huang

, Cao

, Zou

, et al. A comparison of tissue versus swab culturing of infected diabetic foot wounds. Int J Endocrinol, 2016; 2016:8198714.

22.

Nelson

, Wright-Hughes

, Backhouse

, et al. CODIFI (Concordance in Diabetic Foot Ulcer Infection): A cross-sectional study of wound swab versus tissue sampling in infected diabetic foot ulcers in England. BMJ Open, 2018; 8:e019437.