Every one of us has quoted a healing rate. In clinic, in a grant, in a talk, in a conversation with a patient sitting on the exam table asking the only question that matters: will this close? Most of those quoted numbers came from the control arms of industry trials, which is a bit like estimating the height of the population by measuring the people who fit through a particular door.
The NIH-funded Diabetic Foot Consortium has now done the harder thing. In a new paper in Wound Repair and Regeneration, Xu C, Schmidt BM, Ye W, Roy S, Sen CK, Pop-Busui R, Gaynanova I and colleagues report longitudinal healing and amputation trajectories from two prospective DFC cohorts — the Open Wound Master Protocol (n = 419) and the completed c-Myc biomarker study (n = 140). Real patients, real standard of care, broad inclusion criteria, across community clinics and tertiary centres.
The headline numbers
In the Master Protocol, roughly 26% of diabetic foot ulcers had healed by week 12 (95% CI 21–31%). By week 32, that reached about 49% (95% CI 43–56%). Read that second one again: at eight months, half of these wounds were still open.
Amputation rates ran 4.3% by week 12 (95% CI 2.2–6.5%) and 10% by week 32 (95% CI 6.3–14%). The c-Myc cohort, with tighter inclusion criteria and a protocolised care pathway including mandated total contact casting and sharp debridement, did better at 12 weeks — about 38% healed and 2.5% amputated.
Crucially, the DFC team treated amputation as a competing risk using Aalen–Johansen estimation rather than quietly censoring those patients out. That is a methodological choice with moral weight. Amputations are not missing data. They are the outcome we exist to prevent.
Triangulating against Fife, Lavery, and Margolis
What makes this paper genuinely useful is how well it converges with the other serious attempts to answer the same question from completely different angles.
- Trial control arms. Coye TL, Bargas Ochoa M, Zulbaran-Rojas A, Tarricone A, Chung J, Najafi B and Lavery LA pooled 32 randomised controlled trials published between 1996 and 2023 and found a standard-of-care healing rate of 33.15% (95% CI 31.18–35.11%), with a mean healing time around 50 days. Their conclusion was blunt: standard of care shows limited effectiveness, with only about one third of patients achieving closure.
- Real-world registry. Fife CE, Carter MJ and colleagues, working from the US Wound Registry across 96 clinics, 11,784 patients and 25,114 DFUs, documented 39.4% healing in total-contact-cast-treated ulcers versus 37.2% in those not casted — and one-year amputation of 2.2% versus 5.2%. The sting in that paper is elsewhere: offloading was documented at only 2.2% of 221,192 visits, and among clinics that owned casting materials, 96.3% of the ulcers that did not get a TCC were eligible for one.
- Historical cohorts. The DFC estimates sit squarely on top of the Margolis benchmarks from 1999, 2003 and 2022, which is the whole point. Twenty-plus years, and the curve has not moved.
Four datasets, four methodologies, one answer: about a third of diabetic foot ulcers close at twelve weeks. That is our real denominator.
Why the trial number and the clinic number drift apart
Here is where Caroline Fife’s earlier work becomes indispensable. Carter MJ, Fife CE, Walker D and Thomson B took 17 wound-care RCTs and asked a deceptively simple question: what fraction of an actual wound clinic population would have been thrown out by these exclusion criteria? In 15 of the 17 trials, more than half. Even after stripping out the exclusions that were arguably not clinically meaningful, 14 of 17 still excluded a quarter to a half of real patients.
The DFC paper quantifies the mechanism from the other direction. Predicted healing rates differed by 30 to 40 percentage points between small, short-duration ulcers and large, long-duration ones. Which means enrolment criteria alone can move a trial’s headline result by more than any dressing in the study. Choose your inclusion criteria and you have largely chosen your answer before the first patient is randomised.
A ruler and one question
The predictive modelling is my favourite part, because it is so unglamorous. Baseline wound area and baseline wound duration — log-transformed, two variables — produced an AUC at or above 0.72 from week 4 through week 24, peaking near 0.78. Both were significant at every timepoint out to 32 weeks.
Wound depth added nothing. Once area and duration were in the model, depth was not a significant predictor, which the authors attribute in part to how notoriously badly we measure it.
So the single best available prognostic instrument in the diabetic foot remains a disposable ruler and the question “how long have you had this?” We have spent a great deal of money looking for something better. That is not an argument against biomarkers — it is precisely why the DFC exists, and precisely the bar any candidate biomarker now has to clear. Beat area and duration, or go home.
The authors also flag that the most informative window sits between weeks 12 and 20, which is exactly where most trials have already stopped looking.
What this means for how we design trials
The sample-size work is immediately actionable. For a 20 percentage-point absolute effect, roughly 180 participants total gets you 80% power. Chase a 10-point effect and you need 650 to 780. Chase 5 points and you are looking at 3,000-plus, which is to say you are not running that trial.
And the twelve-week endpoint, that arbitrary convention we have all inherited and none of us chose, comes out looking indefensible. Shorter trials need smaller samples because outcome variance is lower — but as the authors put it, they risk failing to capture the full treatment effect. A large ulcer may be responding beautifully at week 12 and simply not be closed yet. We score it a failure and move on.
The remission frame
Regular readers know where I am going. If half of these wounds are still open at eight months, and one in ten patients has lost part of a foot by then, the language of cure is not merely optimistic — it is misleading. This is a chronic, relapsing, remitting disease of a barrier organ, and it deserves the same longitudinal surveillance architecture oncology built decades ago. Ulcer-free, hospital-free, activity-rich days. Not a checkbox at week 12.
The DFC Master Protocol is heading toward 5,000 participants. That is the closest thing our field has to a properly instrumented natural-history study, and it is being built by people who chose to report the uncomfortable number rather than the flattering one.
Congratulations to Charlotte Xu, Brian Schmidt, Wen Ye, Sashwati Roy, Chandan Sen, Rodica Pop-Busui, Irina Gaynanova, Cathie Spino, Teresa Jones, Crystal Holmes, Kellen Chen and the whole consortium. This is the kind of paper that quietly resets a field’s assumptions. Now let us go put the biostatistician, the vascular surgeon, the podiatrist, the endocrinologist and the biomarker chemist in the same room and beat 26%.
According to PubMed:
Xu C, Schmidt BM, Krambrink A, Park J, Song PXK, Jones TLZ, Holmes CM, Chen K, Ye W, Kolenic G, Roy S, Sen CK, Pop-Busui R, Spino C, Gaynanova I. Longitudinal Healing and Amputation Trajectories in Diabetic Foot Ulcers: Predictive Power of Wound Area and Duration and Sample-Size Implications From the Diabetic Foot Consortium. Wound Repair Regen. 2026;34(4):e70189. https://doi.org/10.1111/wrr.70189
Coye TL, Bargas Ochoa M, Zulbaran-Rojas A, Martinez Leal B, Quattas A, Tarricone A, Chung J, Najafi B, Lavery LA. Healing of diabetic neuropathic foot ulcers receiving standard treatment in randomised controlled trials: A random effects meta-analysis. Wound Repair Regen. 2025;33(1):e13237. https://doi.org/10.1111/wrr.13237
Fife CE, Carter MJ, Walker D, Thomson B, Eckert KA. Diabetic foot ulcer off-loading: The gap between evidence and practice. Data from the US Wound Registry. Adv Skin Wound Care. 2014;27(7):310-6. https://doi.org/10.1097/01.ASW.0000450831.65667.89
Carter MJ, Fife CE, Walker D, Thomson B. Estimating the applicability of wound care randomized controlled trials to general wound-care populations by estimating the percentage of individuals excluded from a typical wound-care population in such trials. Adv Skin Wound Care. 2009;22(7):316-24. https://doi.org/10.1097/01.ASW.0000305486.06358.e0
The Diabetic Foot Consortium is supported by NIDDK grants U01DK119100, U01DK119083, U01DK119094, U01DK119099 and U24DK122927.
#DiabeticFoot #DiabeticFootUlcer #DFU #DiabeticFootConsortium #WoundHealing #WoundCare #AmputationPrevention #ActAgainstAmputation #ALPSLimb #LimbPreservation #ClinicalTrials #TrialDesign #RealWorldEvidence #Offloading #TotalContactCast #DiabeticFootInRemission #MeasureWhatWeManage #ToeAndFlow #Biomarkers #NIDDK #diabeticfootonline
Leave a Reply