Microbiological profile and antimicrobial resistance among diabetic foot infections in Lebanon

Abstract

Diabetic Foot Infection (DFI) is a challenging complication of diabetes mellitus with a high burden in the Middle East where there is a marked increase in diabetes prevalence and complications. Early detection of DFI and the infectious organisms could result in the early initiation of appropriate antibiotic therapy and improved outcomes. DFI microbiological profiles differ between countries. In our region, Western guidelines are used when initiating treatment for DFI in the absence of local guidance. The purpose of our study was to determine the microbiologic profile and antimicrobial susceptibility of the DFI admissions at a large tertiary referral centre in Beirut and review other reported series in Lebanon and our region. This is a retrospective observational study of patients with DFI admitted to the American University of Beirut Medical Centre from January 2008 to June 2017. The bacteriologic isolation and antimicrobial susceptibility tests were performed according to standard microbiological methods. Between 2008 and 2017, 319 diabetic patients with DFU were admitted to AUBMC, and deep‐tissue cultures were taken for 179 patients. From 179 deep tissue cultures, 314 bacterial isolates were obtained. Fifty‐four percent of patients had the polymicrobial infection. Aerobic gram‐negative rods (GNR) were more prevalent than gram‐positive cocci (GPC) (55%, 39%, respectively). The most common isolate was Escherichia coli (15%) followed by Enterococcus (14%) and Pseudomonas aeruginosa (11%). Staphylococcus aureus isolates accounted for 9% with 50% of them being methicillin‐resistant (MRSA). Among Enterobacteriaceae, 37% of isolates were fluoroquinolone‐resistant, 25% were ESBL producers, and 2% were carbapenem‐resistant. Antibiotic resistance was significantly associated with prior usage of antibiotics. Anaerobes were isolated in 1% and Candida species in 5% of isolates. The sensitivity, specificity, PPV, and NPV of swab culture recovery of pathogens compared with deep tissue culture were (76%, 72%, 76%, 72%) and (94%, 81%, 91%, 86%) for gram‐positive and gram‐negative organisms, respectively. The microbiological profile of DFI in Lebanon is comparable to other countries in the MENA region with big differences compared with the West. Therefore, it is imperative to develop local guidelines for antimicrobial treatment. The high prevalence of GNR in DFI and the high fluoroquinolone resistance should be taken into consideration when choosing empiric antibiotics. Empiric treatment for MRSA or Pseudomonas does not appear necessary except for patients with specific risk factors.

1. INTRODUCTION

Diabetes is an important cause of morbidity around the world, and its burden has markedly increased in the Eastern Mediterranean Region over the last 25 years. ¹ The prevalence of diabetes in the Middle East and North Africa (MENA) region is the highest compared with other regions, ranging from 8% to 24% with expected increase. ²

Diabetes has a wide range of complications with diabetic foot ulcer (DFU) being the most costly yet preventable. ³ DFU affects 25% of diabetic patients during their lifetimes and is considered one of their major causes of hospitalisation where 20% to 25% of diabetic patients hospitalisations are related to foot problems. ³ , ⁴ In Lebanon, the prevalence of foot ulcers in diabetic patients is approximately 8.6%, with a high prevalence of predisposing factors, such as peripheral neuropathy and peripheral arterial occlusive disease in 40% and 18% of patients, respectively. ² , ⁵

The aetiology of DFU is multifactorial. It includes peripheral arterial disease, peripheral neuropathy, foot deformity, trauma, and impaired resistance to infection. ⁶ The management of DFU depends on identifying the vascular and infectious status of the wound. The mainstay of treatment includes foot care and hygiene, diabetes control, antibiotics, revascularisation, and surgical debridement or amputation when indicated. ⁷

Infection is a serious consequence of DFU, as 58% of ulcers are infected upon presentation and the number increases to 82% in hospitalised patients. ⁸ Diabetic foot infection (DFI) can result in poor outcomes causing more than 50% of lower limb amputations. ⁸ , ⁹ The recent guidelines for the treatment of DFI recommend selecting empirical antibiotic treatment for clinically infected wounds depending on the severity of the infection with reliance on appropriately obtained cultures and sensitivity and patients' response to treatment in determining the definitive therapy. ¹⁰

There is a wide variation in the microbiological profile of DFIs as reported in the literature. The most common organisms isolated from DFUs in North America and Europe are aerobic gram‐positive cocci, mainly Staphylococcus aureus. In contrast, studies from countries in Asia and Africa show that aerobic gram‐negative organisms are more common. ¹¹ In addition, the emergence of antimicrobial resistance is a compounding problem, associated with treatment failure and poor outcomes in many clinical settings. The natural evolution of bacteria along with the uncontrolled widespread use of antibiotics are the predisposing factors. ¹² , ¹³

Several studies were conducted in the MENA region describing the microbiological profile in DFI. ¹¹ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ In Lebanon, two studies were conducted during the past decade with different findings yet no local guidelines were established. Our study aims to determine the microbiological profile, antimicrobial resistance in a tertiary hospital in Lebanon, and review the findings with existing data to provide a useful resource for creating management guidelines in the MENA region.

2. METHODS

We conducted a single centre retrospective study of 356 diabetic patients admitted to the American University of Beirut Medical Centre (AUBMC) with the diagnosis of DFU between January 2008 and June 2017. After obtaining ethical approval from AUBMC institutional review board, we reviewed the medical records and collected demographic data, including age, gender, origin, and comorbidities (hypertension, dyslipidemia, chronic kidney disease, coronary artery disease, congestive heart failure). Diabetes history, including type, duration, and treatment by oral medications only or with insulin as well as diabetes complications, such as sensory neuropathy and peripheral artery disease were also gathered.

Wound infection was classified according to the International Working Group for Diabetic Foot (IWGDF) and Infectious Diseases Society of America (IDSA) criteria. ¹⁹ Infected wounds were classified as mild, moderate, and severe based on the size (of cellulitis) and depth (or level of tissue involved) of the infection and the presence of systemic manifestations of infection or metabolic instability. Ulcers were also classified according to the Wagner‐Meggit system. ²⁰

We collected Microbiology results of swab culture and deep tissue or bone biopsy. We used the tissue culture results to compare our results to other national and international studies. Identification of culture growth was done at the AUBMC bacteriology laboratory as per standard procedures and guidelines. ²¹

We defined osteomyelitis by the presence of one of four determinants; plain X‐ray findings indicative of osteomyelitis, MRI findings when available, positive bone culture, or positive pathology results.

All data were collected and entered into SPSS statistical software. Descriptive analysis was conducted; demographics and clinical data were summarised using percentages, mean ± standard deviation. A chi‐square test was used in comparing patients according to antimicrobial resistance. Statistical significance was determined as P ≤ .05.

3. RESULTS

3.1. Demographics

This study included 356 patients who were admitted with the diagnosis of DFU. The mean age of the patients was 66.9 ± 12.2 years, with 246(69%) males. Most patients were Lebanese (71%) and the rest were from neighbouring countries mainly Iraq (24%). The majority had type 2 diabetes (98%) with a mean duration of diabetes of 18.4 ± 9.1 years. The median length of stay was 8 days, range (1‐112) days. The general characteristics of the study population are shown in (Table 1).

TABLE 1.

General characteristics of the DFU patients

Characteristic		N = 356
Mean age (y)		66.9 ± 12.2
Origin	Lebanese	254 (71%)
	Non‐Lebanese	102 (29%)
Gender	Male	246 (69%)
	Female	110 (31%)
Residence	Beirut	146 (41%)
	Out of Beirut	114 (32%)
	Iraq	84 (24%)
	Others	12 (3%)
DM	Type 1	8 (2%)
	Type 2	348 (98%)
DM treatment	Oral medications	163 (46%)
	Insulin	193 (54%)
Mean DM duration (y)		18.4 ± 9.1
DM complications	PAD	272 (76%)
	CAD	171 (48%)
	CVA	34 (10%)
	Neuropathy	219 (62%)
	Nephropathy	121 (34%)
Comorbidities	HTN	300 (84%)
	DL	255 (72%)
	CHF	94 (26%)

3.2. Infectious status at presentation

Upon admission, and according to the IWGDF/IDSA classification, 37(10%) patients did not have an infection, 131(37%) had a mild infection, 164(46%) had a moderate infection, and 24 (7%) had a severe infection. Only 83 (23%) had fever upon presentation, and 65% had received antibiotics before admission. C‐reactive protein (CRP) was measured at baseline in 128 (36%) of the patients with a median value of 46.6 mg/L (range 0‐441 mg/L). Osteomyelitis was diagnosed in 179(50%) of patients. Patients' characteristics at presentation are described in (Table 2).

TABLE 2.

Infectious status upon presentation

Characteristics upon presentation		N = 356
Fever (%yes)		83 (23%)
Mean temperature (celsius)		36.9 ± 0.6
Osteomyelitis (%yes)		179 (50%)
IWGDF/IDSA	No infection	37 (10%)
	Mild	131 (37%)
	Moderate	164 (46%)
	severe	24 (7%)
Wagner classification	0	3 (1%)
	1	26 (7%)
	2	86 (24%)
	3	69 (19%)
	4	144 (40%)
	5	28 (8%)
Treatment before presentation	PO antibiotics	121 (34%)
	IV antibiotics	105 (30%)
	Topical Antibiotics	7 (2%)
	None	120 (34%)

3.3. Management post‐clinical presentation

Out of the 356 hospitalised patients, 332 (93%) received intravenous antibiotics, 4 (1%) received per oral antibiotics, 3 (1%) topical antibiotics, and 17 (5%) did not receive any antibiotics. The mean total duration of antibiotics, including in hospital and after discharge was 22 ± 16 days. Some patients underwent various vascular interventions according to the surgeons' assessment. Vascular interventions included balloon angioplasty in 48 (13%), patch angioplasty in 8 (2%), and bypass surgery in 34 (9%). 276 (72%) of all patients underwent different surgical procedures ranging from debridement to amputation at different levels. (Table 3).

TABLE 3.

Management of DFU patients

Management strategy			N = 356
Antibiotics	No ABx		17 (5%)
	IV ABx		332 (93%)
	PO ABx only		4 (1%)
	Topical		3 (1%)
Surgery (n = 276)	Debridement		151
	Skin grafting		23
	Osteotomy		47
	Angiography		78
	Vascular intervention (n = 73)	Balloon angioplasty	48
		Patch angioplasty	8
		Bypass surgery	34
	Amputation	Total	152
	Minor amputation (n = 121)	Digits	65
		Ray	20
		Transmetatarsal	32
		Syme	2
	Major amputation (n = 31)	BKA	24
		AKA	7

3.4. Tissue culture results

A total of 179 hospitalised patients underwent deep tissue bacterial cultures, of those 172 were positive with 92 (54%) being polymicrobial and seven negative cultures.

The total number of isolates was 314, 174 (58%) of the isolates were aerobic gram‐negative bacteria and 122 (41%) were aerobic gram‐positive bacteria.

The most common isolated bacteria were E. coli (15%), followed by Enterococcus (14%) and P. aeruginosa (11%). (Table 4)

TABLE 4.

Tissue enterobacteriaceae

Enterobacteriaceae	n = 314
E. coli	47 (15%)
P. mirabilis	29 (9%)
M. morganii	19 (6%)
K.Pneumoniae	9 (3%)
E. cloacae	7 (2%)
C. freundii	6 (2%)
P. vulgaris	6 (2%)
Serratia	5 (2%)
K. oxytoca	2 (1%)
E. aerogenes	2 (1%)

S. aureus was isolated in 9%, and 50% of S. aureus isolates were methicillin‐resistant (MRSA). Staph‐coagulase negative bacteria were isolated in 10% of cases of which 48% were oxacillin resistant. Enterobacteriaceae family represented the highest number of isolates (42%). Among this family species, 37% were fluoroquinolones resistant, 25% were extended spectrum beta‐lactamase (ESBL) producers, 2% were carbapenem‐resistant, and 52% were resistant to trimethoprim/sulfamethoxazole. P. aeruginosa was resistant to fluoroquinolones in 18% of cases, to carbapenems in 3%, and to cephalosporins in 8% Table 5. Acinetobacter baumannii comprised 2% of the isolates and 37% of them were multidrug‐resistant. Fungi were present in 15 cultures representing 5% of the total isolates. C. albicans was isolated from five samples, non‐albicans from nine samples, and both from one patient. (Table 5).

TABLE 5.

The results of tissue culture and antimicrobial resistance

Bacteria type		Number of isolates = 314
Gram‐positive bacteria		122 (39%)
Staph‐coagulase negative	Total	31 (10%)
	Oxacillin resistant	15 (5%)
S. aureus	Total	28 (9%)
	MRSA	14 (4%)
Enterococcus	Total	44 (14%)
	Ampicillin resistant	8 (3%)
	Vancomycin resistant	1 (0.3%)
Streptococcus	Total	12 (4%)
	Penicillin resistant	1 (0.3%)
	Clindamycin resistant	5 (2%)
GPR (Diphtheroides)		7 (2%)
Gram‐negative bacteria		174 (55%)
Enterobacteriaceae	Total	132 (42%)
	Fluoroquinolones resistant	49 (16%)
	ESBL	23 (7%)
	CRE	2 (1%)
	TMP/SMX resistant	69 (22%)
P. aeruginosa	Total	34 (11%)
	Fluoroquinolone resistant	6 (2%)
	MDR	0 (0%)
	Carbapenem resistant	1 (0.3%)
	Cephalosporin resistant	3 (1%)
Acinetobacter	Total	8 (2%)
	MDR	3 (1%)
Anaerobes		3 (1%)
Fungi	Total	15 (5%)
	C. albicans	5 (2%)
	Candida non‐albicans	9 (3%)
	both	1 (0.3%)

3.5. Swab culture validity

Out of 179 patients who had tissue culture, 146 patients had also a wound swab culture taken simultaneously. Both swab and tissue culture had identical isolates in 44% of the patients. By comparing the results of swab and tissue culture, we found that swab culture had 76% sensitivity, 72% specificity, 76% positive predictive value (PPV), and 72% negative predictive value (NPV) in detecting aerobic gram‐positive cocci. Regarding gram‐negative rods, the swab culture had a sensitivity, specificity, PPV, NPV of 94%, 81%, 91%, and 86%, respectively (Table 6).

TABLE 6.

Sensitivity and specificity of swab culture

	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
Gram‐positive cocci	76	72	76	72
Gram‐negative rods	94	81	91	86

Abbreviations: NPV, negative predictive value; PPV, positive predictive value.

3.6. Antimicrobial resistance

3.6.1. Methicillin‐resistant S. aureus (MRSA)

In order to assess the risk factors for MRSA, we compared methicillin sensitive S. aureus (MSSA) in the tissue cultures (14 isolates) to MRSA in tissue culture (14 isolates).

There was a significant difference in the patient origin of MRSA infection, with 71% of MRSA being detected in Iraqi patients, (P = .007). No significant difference was found in the history of previous antibiotics treatment pre‐admission. By comparing the rate of MRSA in the last 5 years (2013‐2017) to that of the previous years (2008‐2012), no significant difference was detected (Table 7).

TABLE 7.

Antimicrobial resistance among different bacterial species

Methicillin resistance		MSSA (n = 14)	MRSA (n = 14)	P
S. aureus (n = 28)
Origin	Lebanese	11 (79)	3 (21)	.007
	Iraqi	2 (14)	10 (71)
	Others	1 (7)	1 (7)
Time	2008 to 2012	3 (21)	3 (21)	1.000
	2013 to 2017	11 (79)	11 (79)
Treatment pre‐admission	None	4 (29)	4 (29)	.904
	IV	6 (43)	5 (36)
	PO	4 (29)	5 (36)

Fluoroquinolones resistance		Resistant (n = 49)	Sensitive (n = 51)	P
Enterobacteriaceae (n = 100)
Origin	Lebanese	42 (82)	28 (57)	.023
	Iraqi	7 (14)	16 (33)
	Others	2 (4)	5 (10)
Time	2008 to 2012	21 (41)	14 (29)	.186
	2013 to 2017	30 (59)	35 (71)
AB pre‐admission	None	19 (37)	9 (18)	.027
	IV	14 (28)	24 (49)
	PO	15 (29)	16 (33)
	Topical	3 (6)	0 (0)

ESBL		ESBL (n = 33)	Non‐ESBL (n = 67)	P
Enterobacteriaceae (n = 100)
Origin	Lebanese	19 (76)	51 (76)	.159
	Iraqi	11 (33)	12 (18)
	Others	3 (9)	4 (6)
Time	2008 to 2012	9 (27)	26 (39)	.256
	2013 to 2017	24 (73)	41 (61)
AB pre‐admission	None	3 (9)	25 (37)	.010
	IV	16 (49)	22 (33)
	PO	14 (42)	17 (25)
	Topical	0 (0)	3 (5)

3.6.2. Fluoroquinolones resistance Enterobacteriaceae

Out of 132 Enterobacteriaceae isolates in tissue culture, 49 were fluoroquinolones resistant species.

By comparing those with fluoroquinolone‐resistant species to those with sensitive ones, we found a significant difference in the patient origin as 82% of fluoroquinolone‐resistant Enterobacteriaceae was detected in Lebanese patients compared with 18% in other nationalities (P = .023). Treatment before admission also affected the resistance as 82% of patients with resistant species received antibiotics pre‐presentation compared with 57% of patients with sensitive species (P = .027). The prevalence of fluoroquinolone resistance was shown to be increasing over the last 10 years. (Figure 1).

FIGURE 1.

Changes in incidence of fluoroquinolones resistant Enterobacteriaceae infection over time

3.6.3. ESBL producing Enterobacteriaceae

Out of 132 Enterobacteriaceae isolates in tissue culture, 33 were ESBL producing species. No significant difference was found in the origin between patients with ESBL producing species and patients with non‐producing species; however, ESBL producing species was significantly higher among patients who had antibiotic treatment before admission 91% compared with 58% of patients with non‐ESBL producer species (P = .010). (Table 7).

3.6.4. Tissue culture results according to patient's origin

Out of 172 positive tissue cultures, 113 were taken from Lebanese and 59 from non‐Lebanese. The only significant difference was the higher prevalence of GNR and Pseudomonas in Lebanese patients compared with non‐Lebanese (P = .015, .022, respectively). (Table 8).

TABLE 8.

Bacteriology in Lebanese and non‐Lebanese patients

Total number of positive tissue culture (172)		Lebanese	Non‐Lebanese	P
		n = 113	n = 59
Poly‐microbial		62 (55)	30 (51)	.616
GPC		66 (58)	40 (68)	.229
	Staph‐coagulase negative	24 (21)	7 (12)	.129
	S. aureus	14 (12)	14 (24)	.056
	Enterococcus	27 (24)	17 (29)	.483
	Streptococcus	9 (8)	3 (5)	.482
	GPR	3 (3)	5 (9)	.085
GNR		87 (77)	35 (60)	.015
	Enterobacteriaceae	70 (61)	30 (51)	.161
	Pseudomonas	28 (25)	6 (10)	.022
	Acinetobacter	4 (4)	4 (7)	.335
Anaerobes		1 (1)	1 (2)	.638

4. DISCUSSION

DFU is a major complication of diabetes affecting approximately 8% of diabetic patients in Lebanon. ² The proper treatment of foot infection is essential as the failure to timely diagnose and initiate appropriate antibiotics can lead to limb‐threatening infection and eventually limb loss. ²² This study examined the microbiological profile and antimicrobial resistance if DFI in Lebanon with a special focus on potential management gaps. In Lebanon, data regarding DFI are still scarce and comparing our data with current local studies showed mild variation in infection profile among different centres as outlined in Table 9. ²³ , ²⁴

TABLE 9.

Comparison of DFU infections among different centres in Lebanon

	Current study (179)	LAU 23 (128)	St. Joseph 24 (167)
Time period	2008 to 2017	2015 to 2016	2000 to 2011
Gram‐positive isolates	39%	33%	—
Gram‐negative isolates	55%	66%	—
Enterobacteriaceae of which susceptible to ciprofloxacin susceptible to quinolones	42% 37%	42% 48.6%	39%
E. coli	15%	12.1%	11.9%
Proteus	—	12.6%	8.5%
Enterococcus	14%	18%	11%
Pseudomonas of which susceptible to ciprofloxacin susceptible to imipenem susceptible to pipe/tazo susceptible to cefta/clavulanate quinolone	11% 37% — — — 18%	18.6% 69% 96% 91% 70% —	19.1% — 26% 22% 31% 37%
S. aureus of which Staph coag neg	9% 10%	18.6%	11.0% 1.3%
MSSA		15.3%
MRSA	50%	21.4%	29%
Poly‐microbial infections	54%	38%	54.2%

In concordance with a well‐known high prevalence of poly‐microbial infection in DFUs, our study showed that 54% of patients have polymicrobial infection. The rate of polymicrobial infection in other studies ranged from 50% to 84% and in Lebanese studies 38% to 54%. ⁹ , ²⁵

In developed countries, most DFIs are caused by gram‐positive cocci mainly S. aureus. Gram‐positive aerobes are reported in 66% to 84% of DFIs in those settings. ⁹ , ²⁶ , ²⁷ On the contrary, studies from Asian and African countries show a higher prevalence of gram‐negative organisms in DFI: aerobic gram‐negative organisms were found in 51.2% of DFI in Kuwait and in 65.1% in India. ¹⁴ , ²⁵

In our study, tissue culture grew aerobic gram‐negative organisms in 55% of isolates, aerobic gram‐positive organisms in 39%, anaerobes in 1%, and fungi in 5%. This high prevalence of gram‐negative organisms in developing countries might be mainly due to the severity of infection upon presentation, 50% of our patients presented initially with moderate or severe infection and more than 60% had grade three or more Wagner classification. Several investigators reported a high prevalence of gram‐negative bacteria in higher grade ulcers compared with milder ones. ²⁸ , ²⁹

S. aureus is the most common bacteria isolated from DFI in many studies, our results show that E. coli was the most common isolate in 15%. ¹⁵ , ¹⁷ , ²⁷ , ³⁰ Acinetobacter sp. isolates were comparable to rates reported by others. ¹⁵ , ¹⁷ Pseudomonas was cultured in 11% of the isolates in our study, which is lower than the percentages reported by other studies from Lebanon and neighbouring countries such as in Kuwait (17%) and Turkey (15%), but higher than developed countries where the rates are less than 10%. ¹⁰ , ¹¹ , ¹⁴ , ¹⁶ However, despite Pseudomonas growing from wound cultures, many studies show that patients improve without targeted antimicrobial coverage. Therefore, empirical anti‐pseudomonal antibiotic treatment could be reserved for severe infections and in countries with high Pseudomonas prevalence. ¹⁰

Anaerobes were isolated from 1% of our samples. In the literature, there is a large discrepancy in the rate of anaerobic bacteria in DFI, both over time and between countries. The rate of anaerobic isolation from DFI reported in the literature differed from 0% to 45%, which may be due to the unstandardised culture methods especially the type of sample, transportation, and laboratory techniques. ³¹

Yeast infection represented 5% of the total isolates in our study, which is higher than rates reported by other studies. ⁹ , ¹⁴ Studies suggest that chronic wounds wrapped with dressing and topical antibiotics create the ideal environment and favour the growth of yeasts. ³²

The frequency of MRSA in DFI has increased dramatically over the last years worldwide. ¹⁰ Our data shows that 14 (50%) of S. aureus isolates were methicillin‐resistant. The rates of MRSA are high as reported in different countries reaching 85% in Algeria. ¹⁴ , ¹⁵ , ¹⁷ , ²⁷ , ³³ This high rate of resistance in S. aureus in our study could be due to increased MRSA prevalence in the local community, in addition to the absence of strict guidelines for antibiotics prescription and lack of adherence to infection control measures in the community hospital setting. We also found that MRSA was significantly more common in Iraqi patients; as they usually present with more severe infections and have received several courses of antibiotics before presentation to our centre. However, MRSA made up only 4% of the total isolates in this study and 3.1% to 3.9% of the other two Lebanese studies. As such, we question the need for MRSA coverage in empirical antibiotic treatment. International guidelines recommended MRSA coverage if the local prevalence of MRSA exceeded 50% of all isolates in mild infection and 30% in severe infection. ¹⁰

This study showed that the previous antibiotic treatment was associated with significantly higher rates of both ESBL and fluoroquinolone resistance. Kandemir et al also reported that antibiotic treatment and its duration were the significant risk factors for multi‐drug resistant organisms. ³⁴ A previous study in Lebanon reported the increasing rates of ESBL producing organisms and attributed it to recent antibiotics use and prolonged hospital stay. ³⁵

Despite the general agreement that deep tissue culture is more accurate than superficial swabs, swab cultures are still used routinely by health care providers to guide antimicrobial treatment. This may be due to the ease in obtaining a superficial culture, while deep tissue culture requires careful surgical debridement in the operating room. According to the literature, the sensitivity and specificity of swab cultures varied between 26% to 93% and 32% to 96%, respectively. ³⁶ In this study swab culture sensitivity in detecting GPC was 76% and in GNR was 72%, specificity in detecting GPC was 94% and in GNR was 81%. We concluded that swab culture can be used as a reasonable initial option in determining diabetic foot microbiology and guiding antibiotic treatment when deep tissue culture is not available.

Out of 270 patients who underwent surgical intervention, deep tissue culture was taken in 179 patients only. Being the most accurate method of detecting the bacteria and guiding the management, the surgeons should be encouraged to obtain deep tissue culture in all clinically infected diabetic foot admissions.

In this study, 80% of the patients with mild or no infection received intravenous antibiotics. International guidelines recommended intravenous antibiotics for only severe infections and some cases of moderate infection. The unnecessary use of antibiotics may affect the microbiologic profile and increase rates of resistant organisms and affect negatively ongoing antimicrobial stewardship efforts.

Patients with osteomyelitis received antibiotics for 28 ± 16 days, and those without osteomyelitis received antibiotics for 14 ± 12 days; this is in alignment with international guidelines. Those guidelines recommended stopping antibiotics once infection signs have resolved and not until the total healing of the wound. Skin and soft tissue infection need 1 to 2 weeks of antibiotics. In osteomyelitis, antibiotics are recommended for >4 weeks for persistent infected or necrotic bone and for short duration 2 to 5 days after radical resection.

The international guidelines recommended plain radiograph for all patients with DFI to look for bone abnormalities or foreign bodies and MRI if osteomyelitis is suspected. In this study, a plain radiograph was done in 60% of patients and only 29% of patients with osteomyelitis had MRI.

5. CONCLUSION

The microbiological profile of DFI in Lebanon is comparable to other countries in the MENA region with big differences compared with the Western countries. This fact stresses the need for the creation of local guidelines for antimicrobial treatment. The high prevalence of GNR in DFI and the high fluoroquinolone resistance should be taken into consideration when choosing empiric antibiotic therapy. Empiric treatment for MRSA or Pseudomonas does not appear necessary except for patients with risk factors for severe infection. Swab culture may be a reasonable initial option in determining DFU bacteriology and guiding the antimicrobial treatment. More studies are needed to assess the prevalence and epidemiology locally in the countries of the MENA in addition to the standardised classification scales. Based on those, adapted treatment guidelines could be developed by regional experts. It is imperative also to spread awareness regarding abuse and unnecessary use of antibiotics in an area plagued by antimicrobial resistance.

AUTHORS CONTRIBUTION

Lamia Jouhar, Rola F. Jaafar, Rakan Nasreddine: conception and design of the work; ensuring data integrity; statistical analysis; write‐up; revision; final approval of the manuscript. Lamia Jouhar, Omar Itani: data collection; analysis. Rola F. Jaafar, Fady Haddad, Nisrine Rizk: write‐up; revision. Jamal J. Hoballah: conception of the work; revision; final approval of the manuscript.

Jouhar L, Jaafar RF, Nasreddine R, et al. Microbiological profile and antimicrobial resistance among diabetic foot infections in Lebanon. Int Wound J. 2020;17:1764–1773. 10.1111/iwj.13465