Aims/hypothesis
CODIFI2 compared wound swabbing and tissue sampling in people with infected diabetic foot ulcers (DFU) to determine effects on clinical outcomes.
Methods
Multicentre, Phase III, prospective, non-blind, 2-arm parallel group, randomised controlled trial comparing time to ulcer healing (primary outcome), proportions healed, antimicrobial regimen, ulcer area reduction, hospitalisation duration, and time to death, for swab compared to tissue sampling. Allocation was via central and independent randomisation system, with minimisation by DFU site, number, type, size, location, and duration.
Follow-up was 52-104 weeks, with healing confirmed by a blinded assessor. Sample-size target was 730 participants for 90% power to detect 12.5% difference in healing at 52-weeks.
Results
Between May 2019 and May 2022 149 participants were recruited (75 Swab, 74 Tissue) from 21 UK sites. The 52-week cumulative incidence of confirmed healing as first event was 45.3% (33.5% to 56.4%) and 44.6% (33.0 to 55.6%) for swab vs tissue. Hazard ratio (HR) for healing for tissue vs swab was 1.01 (95% CI 0.65 to 1.55). Median (IQR) days in hospital was 17 (12 to 39) for swab and 16 (10 to 32) for tissue. Seventeen swab and 7 tissue participants died during follow-up and 18.7% and 24.3% participants in the swab and tissue groups respectively had an amputation.
Conclusions/interpretation
This trial was underpowered to determine whether swab or tissue sampling impacted rate of healing or time to healing. Clinical prescribing and patient outcomes differed slightly between groups, hence the clinical benefit of tissue sampling is not established.
ORCID: 
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)