Foot infections account for about 20% of all hospitalisations in people with diabetes and at least 50% of all non-traumatic lower limb amputations in the US (Cunha, 2000). They are usually preceded by foot ulcers and have a worse outcome than foot infections in people without diabetes, owing to the impaired microvascular circulation, neuropathy, anatomical alterations and various metabolic disturbances associated with diabetes (Cunha, 2000).
Most moderate-to-severe diabetic foot infections are polymicrobial. Empirical therapy should therefore include broad-spectrum antibiotics; these are often given parenterally, necessitating hospitalisation of the patient (Lipsky, 1999).
Aim of the study
The aim of this study was to investigate the prevalence of pathogenic organisms in severe foot infections and their resistance to the empirical antibiotic regimens used for initial treatment. It reflects the experience of the multidisciplinary team working in the diabetic foot clinic of the Endocrinology Department of Hospital Geral de Santo António, Porto, Portugal.
Patients and methods
Patients
A retrospective analysis of 185 infected foot ulcers was conducted between January 2000 and November 2002. Bacteriological examinations were performed in the microbiology laboratory. The examinations comprised Gram staining and culture on appropriate media to identify aerobic and anaerobic bacteria. Sensitivity tests were performed on the organisms isolated.
Eighty-one per cent of the patients studied had type 2 diabetes and 19% had type 1 diabetes. The mean age of patients was 63 years (range 23–89 years) and 59% were male. The mean duration of diabetes since diagnosis was 18 years (range 1–42 years).
A neuropathic ulcer was present in 86.5% of patients, and a vascular ulcer in 13.5%. All patients had clinical evidence of severe infection, defined as ulcer grade III/IV on the Wagner scale (Wagner et al, 2001).
All patients had been started on empirical intravenous antibiotic therapy. The combinations used were:
- imipenem/cilastatin (500mg qds)
- ciprofloxacin (400mg bd) + clindamycin (600mg tds)
- vancomycin (1g bd)
- amoxicillin/clavulanate (1.2g tds).
Methods
Samples for staining and culture were obtained from deep tissues, and by needle aspiration when pus was present. Evidence of osteomyelitis was sought by radiology and/or probing to the bone.
Statistical analysis was performed using the Fisher–Irwin or chi-squared test. Statistical significance was set at p<0.05.
Results
Pathogenic organisms isolated
Bacterial growth was detected in all the samples. Fifty-three per cent of the wound cultures were polymicrobial.
The most prevalent organisms isolated were Gram-positive aerobes (55%), followed by Gram-negative aerobes (45%).
Among the Gram-positive aerobes, Staphylococcus aureus was the most common (60%), and 53% of these were methicillin-resistant Staphylococcus aureus (MRSA) (Figure 1). MRSA was isolated only from patients who had been treated at least once before, and with different antibiotics.
The most prevalent Gram-negative aerobe isolated was Pseudomonas aeruginosa (29%).
Anaerobes were present in 2.7% of the isolates (Figure 2).
Resistance to the antibiotic regimens
All patients had received empirical treatment initially. The distribution of antibiotic regimens used (Figure 3) were:
- amoxicillin/clavulanate (51%)
- imipenem/cilastatin (30%)
- ciprofloxacin/clindamycin (17%)
- vancomycin (2%).
Organisms resistant to the antibiotic regimens selected were found in the isolates of 44 (23.8%) patients. The organisms responsible were: MRSA (31), Pseudomonas aeruginosa (7), Serratia marcescens (4) and Morganella morganii (2) (Figure 4).
Thirty per cent of isolates were resistant to the amoxicillin/clavulanate combination, 38% to ciprofloxacin/clindamycin and 5.4% to imipenem/cilastatin. The difference was statistically significant only when imipenem/ cilastatin was compared with either ciprofloxacin/clindamycin (p=0.0003) or amoxicillin/clavulanate (p=0.0005) (Figure 5).
Outcomes
Antibiotic therapy led to improvement in most patients (127; 68.7%) and cure in five (2.7%) patients (cure being defined as disappearance of all signs and symptoms of infection at the end of treatment and absence of relapse during follow-up).
Three patients (1.6%) underwent a major amputation and 23 (12.4%) a minor amputation. Twenty-seven patients (14.6%) were lost to follow-up.
Relationships with outcomes
Resistant bacteria were isolated from 34.6% (9/26) of patients who underwent amputation, compared with 22.0% (35/159) of patients not undergoing amputation, p=0.24 (Figure 6).
Analysing the same relationship, but looking only at MRSA, we found MRSA in the isolates of 38.5% (10/26) of patients who underwent amputation, compared with 13.2% (21/159) of patients not undergoing amputation, p=0.0036 (Figure 7).
Outcomes involving osteomyelitis
Evidence of osteomyelitis was found in 90 (48.6%) patients. MRSA was present in 20% of the isolates from patients with osteomyelitis, compared with 13.7% of those from patients with no osteomyelitis, p=0.44 (Figure 8).
Osteomyelitis was present in 57.6% of patients who underwent amputation compared with 47.2% of patients not submitted to amputation, p=0.24 (Figure 9).
Conclusion
The prevalence of aerobic bacteria found in isolates in this study is similar to that described in other international studies (Boutoille et al, 2000; Temple and Nahatna, 2000). The high percentage of MRSA found in our isolates (16.7%) is also similar to that reported in many previous studies (of 20–40%: e.g. Tentolouris et al, 1999; Day and Armstrong, 1997).
The infections subjected to analysis in our study were serious, complex and had been submitted to previous treatment, so the number of wounds found with polymicrobial infection is not surprising.
The prevalence of anaerobes isolated is much lower than that usually reported (74–95%) (Gerding, 1995; Wagner et al, 2001). This difference could be due to insufficient care in the handling of the samples in these studies. MRSA accounted for a significant percentage of resistant isolates and amputations, contributing to a worse outcome. This organism is an increasing problem in diabetic foot clinics.
Differences in resistance rates between the antibiotic combinations were only significant when imipenem/cilastatin was compared with the other regimens. There was no statistically significant difference between ciprofloxacin/clindamycin and amoxicillin/clavulanate. This result can be attributed to the broader spectrum of these two combinations.
Osteomyelitis, usually reported as having a poor outcome, was not associated with a greater number of amputations in our study. It is likely that other factors, which were not analysed individually, may also influence to the final outcome.
Treatment of a diabetic foot infection requires a multidisciplinary approach. General measures must include elimination of the focus of infection, good blood glucose control, anticoagulation, etc. Antibiotic therapy is not the most important component of management of the diabetic foot; however, it is a vital part of management in the presence of extensive deep involvement or systemic signs of infection (Hartemann-Heurtier et al, 2000).
For patients at high risk of infection from nosocomially acquired pathogens, the initial antibiotic therapy must cover MRSA and resistant Pseudomonas aeruginosa and Enterobacteriae. In all cases, when definitive reliable cultures are reported, initial antibiotic regimens should be revised to narrow the coverage to specific pathogens and ensure better outcomes. In most cases, the duration of antibiotic treatment will determine clinical outcome; in the presence of osteomyelitis, antibiotic therapy should be continued for at least two months.
A carefully planned rehabilitation programme using adapted soles and/or orthopaedic shoes can considerably reduce the frequency of recurrence of infection.
Prevention, treatment and rehabilitation programmes are best accomplished by a multidisciplinary approach involving the endocrinologist and the vascular and orthopaedic surgery teams (Vaccaro et al, 2002; Gulam and Abbas, 2002; O’Rourke et al, 2002). One such project was begun in our outpatient diabetic foot clinic in 1987, and has shown promising results, especially with regard to amputations: since 1995 the percentage of major amputations performed has fallen from 8.2% to 5.2%.
