Design: Multicentre randomised controlled trial, 25 US sites
Population: 170 adults with type 1 or 2 diabetes and foot ulcers of 1–25 cm² with under 40% area reduction after 4 weeks of standard care
Primary outcome: Complete wound closure within 12 weeks, confirmed 2 weeks later
Effect: 42% vs 27%; absolute difference 14.6% (95% CI 0.15% to 27.9%), p=0.047

PREPARE randomised 170 patients at 25 US sites whose diabetic foot ulcers had shrunk by less than 40% over a four-week screening period of standard care. Standard care meant sharp debridement, saline cleansing, moisture-balancing dressings and offloading. One group also received a purified native type 1 collagen matrix containing polyhexamethylene biguanide (PHMB), an antimicrobial.

Complete closure at 12 weeks, confirmed two weeks later, was 42% with the matrix against 27% with standard care alone, an absolute difference of about 15 percentage points. The lower end of the confidence interval was 0.15 points, barely above no effect. Mean area reduction was 75% against 58%. No adverse events were attributed to the product.

The groups were unequal in size (96 against 74), the trial was open to the obvious difficulty of blinding a dressing, and the result is fragile. It adds to an evidence base for advanced wound products that is broad but generally thin. The basics — debridement, offloading, infection control and revascularisation where needed — still do most of the work.

Confirm a stalled ulcer has had four weeks of genuine offloading and debridement before reaching for an advanced dressing.
Less than 40% area reduction at four weeks was the threshold used here to define a hard-to-heal ulcer; measuring area makes this decision possible.
Check perfusion and exclude osteomyelitis before attributing poor healing to the wound bed.
Consider an antimicrobial collagen matrix as an adjunct where offloading is in place and the ulcer is still not closing, if it is available and affordable.
The result was statistically significant only just; do not let an adjunctive dressing replace offloading.

Why it matters

It gives a measurable trigger — under 40% shrinkage at four weeks — for when standard care is failing.

The statistics, in plain English

The 95% confidence interval for the extra closure rate runs from 0.15 to 27.9 percentage points. The true benefit could be anywhere from almost nothing to a large effect; a trial of 170 cannot narrow it further. A p-value of 0.047 sits right at the conventional cut-off, so a slightly different sample could have tipped it the other way.