A Score Built for Mortality, Applied to a Man Who Is Losing the Use of Both Feet
A single patient with polyarteritis nodosa and a Five-Factor Score of zero. The disagreement is over whether a scoring tool built to predict death actually has anything useful to say about the nerve damage that's actually happening to him.
Carl B., 55, has owned the same auto-repair shop for twenty-two years and hasn't taken a sick day in longer than that, which is part of why the pain that started in both his shins two months ago took him this long to bring in — pain he first blamed on standing all day, then on age, until his left foot started dragging and he couldn't tell where the ground was under his right one either. Nerve conduction studies confirmed mononeuritis multiplex in both legs, an asymmetric-onset but now genuinely bilateral pattern of nerve injury that's left him using a cane for the first time in his life. Alongside the nerve findings: a lacy, mottled rash across both calves (livedo reticularis), fifteen pounds of unintentional weight loss, and new hypertension nobody had seen before. Angiography found the classic picture — microaneurysms scattered through his mesenteric and hepatic circulation — confirming polyarteritis nodosa, with his hepatitis serologies both negative, ruling out the hepatitis-B-associated form that would have changed the entire treatment approach toward antivirals rather than immunosuppression. What his workup did NOT find matters as much as what it did: no renal artery involvement, no renal insufficiency, no cardiomyopathy, no GI bleeding, no CNS disease. By the Five-Factor Score, the tool used to predict mortality risk and guide how aggressively to treat PAN, Carl scores a flat zero.
An FFS of zero has historically meant glucocorticoids alone, with cyclophosphamide reserved for relapse or treatment failure — a real, evidence-supported approach from French Vasculitis Study Group data showing acceptable outcomes in exactly that low-risk group. Carl scores zero on either version of the tool: the original 1996 score (Guillevin), built on proteinuria, renal insufficiency, cardiomyopathy, gastrointestinal involvement and CNS signs, and the 2009 revision (Guillevin, Medicine 2011), which dropped CNS disease and proteinuria and added age over 65. He is 55, with none of the rest.
But the Five-Factor Score was built and validated against mortality — renal, cardiac, gastrointestinal and central nervous system failure, the organ systems that actually kill people with PAN. Neither version contains a peripheral-nerve item. It was never designed to predict, and doesn't capture, what happens to a peripheral nerve under active vasculitic attack. Axonal injury from mononeuritis multiplex, once severe enough, doesn't reliably regenerate even after the underlying vasculitis is controlled — meaning the actual clock on Carl's disability may be running faster than the actual clock on any of the five factors his score is built to measure. A tool that correctly says he isn't at meaningfully elevated risk of dying from this disease may be saying nothing at all about whether he's going to walk normally again.
What the score was actually built to predict
His Five-Factor Score is zero, and that score is real, validated evidence — patients at this level have historically done well on glucocorticoids alone, with cyclophosphamide reserved for relapse. Cyclophosphamide's cumulative toxicity, including gonadal and malignancy risk, is a real cost I wouldn't front-load for a fifty-five-year-old with an otherwise normal life expectancy who doesn't meet the organ-threatening criteria the score was built to flag.
I'd push back on what that score is actually telling us here. The FFS was validated against mortality from renal, cardiac, GI and CNS failure — and when it was revised in 2009 the CNS item came out and age came in. Peripheral nerve outcomes were never in either version. It was never built or tested against them. Bilateral mononeuritis multiplex this severe is active, ongoing axonal injury, and once axonal loss is severe enough, nerve regeneration doesn't reliably happen even after the vasculitis itself is brought under control. Waiting for a relapse to add cyclophosphamide risks waiting past the point where his nerves can still recover.
I understand the toxicity concern, and it's a real cost — but it's being weighed against a mortality-risk score that was never designed to measure the specific harm actually happening to him.
I think the neurologist is naming a real blind spot rather than disputing the score's validity for what it was built to measure. An FFS of zero correctly tells us Carl's 5-year mortality risk from this disease is low — around 9% on the 2009 score, 12% on the 1996 one — it says nothing about whether his bilateral foot drop resolves. Those are different predictions, and treating a low-mortality score as equivalent to 'treat conservatively regardless of what else is happening' misreads what the tool actually validates. I'd add cyclophosphamide now, specifically because of the neurologic severity and bilaterality — not because his mesenteric or renal findings changed, they haven't, but because the actual disability at stake here is being driven by a manifestation the score was never built to weigh.
Agreed: start cyclophosphamide induction alongside glucocorticoids now, explicitly on neurologic grounds rather than waiting for a Five-Factor-Score-defined relapse or organ-threatening event to justify it.
The rheumatologist's original caution about cyclophosphamide's cumulative toxicity was not withdrawn — it was accepted as a real cost being knowingly paid, given the specific, disabling, and potentially irreversible nature of his current nerve injury.