Your bedside vascular tests are missing most PAD in diabetes

This is the paper most likely to change what you do this fortnight, and it is not a comfortable one.

The DMPAD study set out to answer a question most of us have quietly assumed was settled: how accurate are the bedside vascular tests we use to screen for peripheral arterial disease in people with diabetes? It ran prospectively across 16 UK sites, recruiting from both primary and secondary care, and enrolled 604 people with diabetes between April 2022 and November 2023. Of those, 536 completed the study with valid reference imaging.

Five index tests were assessed: ankle brachial pressure index, exercise ABPI, toe brachial pressure index, visual Doppler and audible Doppler. Each was compared against CT or MR angiography as the reference standard, scored with the ANGIO system, with 50% or greater stenosis defining PAD. The radiologists were blinded to the bedside results. The primary outcome was sensitivity, which is the right question for a screening test: of the people who genuinely have disease, how many does the test actually pick up? PAD was present in 37% of participants, so this was not a population where disease was rare.

It was published in Diabetes Care and funded by the NIHR.

Sensitivity was poor across the board. Not borderline, not modest. Poor.

Read those numbers the way a patient would experience them. Of the people who actually had significant arterial disease on angiography, a handheld audible Doppler assessment identified around a third. ABPI found about four in ten. The toe pressure index, the test many of us reach for precisely because we know ABPI struggles with calcified vessels, still missed nearly half.

The authors' conclusion is blunt: these five tests demonstrated poor diagnostic performance and should not be considered accurate diagnostic tools in this population.

A normal result is not a clearance. This is the single change worth making. If you have been treating a normal ABPI or a triphasic-sounding Doppler as evidence that the vasculature is fine, that inference does not hold in a person with diabetes. On these numbers, a negative bedside test moves the odds far less than most of us have assumed. The test can raise your suspicion when it is abnormal. It cannot lower it much when it is normal.

That reframes the assessment rather than abolishing it. Keep doing the tests, because an abnormal result is still informative and still triggers referral. What changes is what you do with a reassuring one.

Let the clinical picture outrank the number. If the story in front of you suggests arterial disease, a wound that will not progress despite good offloading and debridement, rest pain, absent hair, a foot that looks dusky when dependent, claudication the patient has explained away as age, then a normal ABPI should not be the thing that stops you referring. Document the discrepancy and escalate on the clinical picture. That is a defensible position and now you have a large prospective study behind it.

Say it out loud in the notes. There is a medicolegal dimension here worth being deliberate about. If you record "ABPI 1.0, vascular status satisfactory" and that patient later presents with critical limb ischaemia, the note reads as a clearance you were never entitled to give. Recording something closer to "ABPI 1.0, noting limited sensitivity in diabetes; referring on clinical grounds" reflects the evidence and protects you.

Know what the study did and did not test. Two honest caveats. The reference standard was anatomical, 50% or greater stenosis on angiography, so this is a question about detecting arterial narrowing rather than about predicting whether a given wound will heal. Perfusion and healing are not the same thing as stenosis, and a pressure-based test is measuring something genuinely different from an angiogram. Second, this was a diagnostic accuracy study, not a trial of a new pathway, so it tells you the tests underperform without telling you exactly what to replace them with. The practical answer for most of us in private practice is a lower threshold for vascular referral and imaging, not a new gadget.

The change to make this fortnight. Pick your diabetic foot assessment template and change one line. Wherever it currently lets a normal ABPI or Doppler close out the vascular section, rewrite it so the vascular decision rests on the clinical picture with the bedside test as supporting information. Then, for the next high-risk patient whose bedside tests look fine but whose foot does not, refer anyway.

In 536 people with diabetes assessed against angiography, ABPI, exercise ABPI, toe pressures, visual Doppler and audible Doppler all had sensitivity between 36% and 55%. Every one of them missed more disease than it caught. Treat an abnormal result as useful and a normal result as weak reassurance, and let the clinical picture drive referral in the high-risk foot.