Concordance in diabetic foot ulceration: a cross-sectional study of agreement between wound swabbing and tissue sampling in infected ulcers.

Background: There is inadequate evidence to advise clinicians on the relative merits of swabbing versus tissue sampling of infected diabetic foot ulcers (DFUs).

Objectives: To determine (1) concordance between culture results from wound swabs and tissue samples from the same ulcer; (2) whether or not differences in bacterial profiles from swabs and tissue samples are clinically relevant; (3) concordance between results from conventional culture versus polymerase chain reaction (PCR); and (4) prognosis for patients with an infected DFU at 12 months' follow-up.

Methods: This was a cross-sectional, multicentre study involving patients with diabetes and a foot ulcer that was deemed to be infected by their clinician. Microbiology specimens for culture were taken contemporaneously by swab and by tissue sampling from the same wound. In a substudy, specimens were also processed by PCR. A virtual 'blinded' clinical review compared the appropriateness of patients' initial antibiotic regimens based on the results of swab and tissue specimens. Patients' case notes were reviewed at 12 months to assess prognosis.

Results: The main study recruited 400 patients, with 247 patients in the clinical review. There were 12 patients in the PCR study and 299 patients in the prognosis study. Patients' median age was 63 years (range 26-99 years), their diabetes duration was 15 years (range 2 weeks-57 years), and their index ulcer duration was 1.8 months (range 3 days-12 years). Half of the ulcers were neuropathic and the remainder were ischaemic/neuroischaemic. Tissue results reported more than one pathogen in significantly more specimens than swabs {86.1% vs. 70.1% of patients, 15.9% difference [95% confidence interval (CI) 11.8% to 20.1%], McNemar's -value < 0.0001}. The two sampling techniques reported a difference in the identity of pathogens for 58% of patients. The number of pathogens differed in 50.4% of patients. In the clinical review study, clinicians agreed on the need for a change in therapy for 73.3% of patients (considering swab and tissue results separately), but significantly more tissue than swab samples required a change in therapy. Compared with traditional culture, the PCR technique reported additional pathogens for both swab and tissue samples in six (50%) patients and reported the same pathogens in four (33.3%) patients and different pathogens in two (16.7%) patients. The estimated healing rate was 44.5% (95% CI 38.9% to 50.1%). At 12 months post sampling, 45 (15.1%) patients had died, 52 (17.4%) patients had a lower-extremity ipsilateral amputation and 18 (6.0%) patients had revascularisation surgery.

Limitations: We did not investigate the potential impact of microbiological information on care. We cannot determine if the improved information yield from tissue sampling is attributable to sample collection, sample handling, processing or reporting.

Conclusions: Tissue sampling reported both more pathogens and more organisms overall than swabbing. Both techniques missed some organisms, with tissue sampling missing fewer than swabbing. Results from tissue sampling more frequently led to a (virtual) recommended change in therapy. Long-term prognosis for patients with an infected foot ulcer was poor.

Future Work: Research is needed to determine the effect of sampling/processing techniques on clinical outcomes and antibiotic stewardship.

Funding: The National Institute for Health Research Health Technology Assessment programme.