The visit runs to a script. A patient describes the tingling that started in the toes and crept up to the arches, the numbness that turns a gravel driveway into a minefield after dark, the burning that arrives the moment the lights go off. The clinician nods, because this is familiar, names it peripheral neuropathy, and reaches for the pad. The patient leaves with gabapentin, or its pricier cousin pregabalin, or the antidepressant duloxetine, and a quiet fact goes unspoken in the exchange: none of those drugs was built to repair a nerve, and none has been shown to. No one checks whether the diabetes pill in the same patient’s other prescription is quietly starving the nerve everyone is so busy medicating.
The distance between what the pills do and what the patient assumes they do is the subject of a piece published this week on TrialSite by Ronald Kostoff, a retired Georgia Tech researcher who spent a career mining the medical literature for patterns the specialists inside any one field rarely step back far enough to see. He treats peripheral neuropathy and “heavy legs” as a combined neurological and vascular problem, and his prescription is the one the pharmaceutical model has no way to write: remove what is causing the damage, and do it before reaching for anything that only quiets the symptom.
What the drugs actually deliver
Start with the drug everyone trusts. Gabapentin is the sixth most prescribed medication in the United States, with nearly 50 million prescriptions filled a year, the bulk of them written off-label and much of that for nerve pain. The franchise is enormous. The evidence underneath it is thinner than the volume implies.
The largest synthesis, a Cochrane review pooling 37 trials and nearly 6,000 patients, looked hardest at the two conditions gabapentin is actually approved to treat. In painful diabetic neuropathy, 38 percent of patients on the drug reported at least a 50 percent drop in pain, against 21 percent on placebo. Read that the way a clinician should: roughly six patients treated for every one who gets meaningful relief, and, by the review authors’ own accounting, more than half who get no worthwhile relief at all while still carrying the side effects. That is the best case, the on-label case.
Step outside the approved indications and the floor drops. When the American Academy of Family Physicians reviewed the off-label sprawl, it warned that guidelines and review articles “do physicians and patients a disservice” when they take the modest benefit seen in diabetic neuropathy and shingles pain and stretch it across every other kind of nerve pain. The drug gets handed out for far more than it was ever shown to touch.
And that is only the benefit column. The cost column carries dizziness, sedation, unsteadiness, weight gain, and fluid retention, plus a withdrawal syndrome that, after sustained use, can include seizures if the drug is stopped abruptly. A retrospective analysis of matched patients across dozens of health systems added another line to the ledger: higher five-year rates of stroke, heart failure, and venous clots among people taking gabapentin or pregabalin, with stroke running about 31 percent higher in the gabapentin group. The drugs quiet the nerve. They do not heal it.
That is not an argument against symptom relief, which a person in pain has every right to want. It is a description of the ceiling. Nothing in the standard regimen reverses the underlying damage, and nothing in it asks why the nerve is dying.
Kostoff’s inversion
Kostoff’s answer to that “why” is the part the pharmacy aisle has no reason to raise. His protocol for chronic disease, built on a body of work that catalogs roughly 8,000 causes spread across some 4,000 diseases, rests on a principle conventional medicine recites and then ignores: removal of cause is a necessary condition for restorative treatment to work. The method is five plain steps, a real exposure and habit history, then behavioral and clinical testing, then lab work, then identifying and stripping out the contributing factors before any disease-specific treatment goes in, with the low-risk options, diet and exercise and removing toxic exposures, coming ahead of the prescriptions.
One instruction sets it apart from the usual lip service. The protocol tells clinicians to eliminate medicinal drugs that themselves turn up on the list of contributing factors, “unless these drugs are absolutely necessary.” For peripheral neuropathy that stops being abstract fast. The condition has a long, well-documented list of removable causes: poor glycemic control, alcohol, vitamin deficiencies, heavy-metal and solvent exposure, and a roster of drugs that injure nerves directly, from certain chemotherapies to some antibiotics. The symptom-first model medicates the alarm. Kostoff’s model goes looking for the fire.
The diabetes drug that feeds the disease
Metformin is the first-line drug for type 2 diabetes, taken by the exact population most likely to develop diabetic neuropathy. Long-term metformin use also depletes vitamin B12, in up to roughly 30 percent of chronic users depending on where the cutoff is drawn. And neuropathy from B12 deficiency is, in the words of the clinical literature, “clinically indistinguishable” from diabetic peripheral neuropathy: the same numbness, the same burning feet, the same exam findings. So a patient can arrive with nerve damage partly manufactured by the drug treating their diabetes, have it labeled diabetic neuropathy, and walk out with a gabapentin script that does nothing about the missing vitamin. Caught early, the deficiency can be arrested with repletion. Caught late, after a few years under a prescription that masked the warning, the damage is permanent. The American Diabetes Association now recommends periodic B12 testing in metformin-treated patients, especially those with neuropathy, which is about as close to an institutional admission that the cause was being missed as you will find in a guideline.
That is one cause, one drug, one vitamin. Kostoff’s argument is that the same shape, a removable cause hiding under a symptomatic prescription, repeats across the whole disease.
Why no one ran the trial
Ask the obvious question and the funding map answers it. If cause-first care for neuropathy is this plausible, where is the trial that puts it head to head against standard drug treatment? It does not exist. No one has run it, and the reason is not scientific.
There is no patent on telling a patient to check their B12, change their diet, and stop a nerve-toxic medication. There is an enormous revenue stream attached to writing the sixth-most-prescribed drug in the country one more time, nearly 50 million times a year. The research dollars follow the molecule, the cause-first approach stays under-studied because almost no one is paid to study it, and then the thinness of that evidence base gets cited as the reason to keep reaching for the molecule. The gap funds itself.
Honesty cuts both ways here. Kostoff’s paradigm is a literature synthesis, not a regimen put through a randomized trial, and its pieces are uneven. Glycemic control and B12 repletion sit on firm ground. Pulling nerve-toxic drugs is ordinary clinical sense. The bolder claim, that diet can reverse established neuropathy, rests so far on preclinical work, including dietary-reversal experiments in diabetic mice that no one has reproduced at scale in people. That limit is worth stating plainly. It indicts the funding, not the idea.
What the evidence does establish is narrower and harder to wave off. The standard drugs help a minority modestly, carry harms that are now turning up in five-year outcome data, and reverse nothing. The causes of neuropathy are frequently findable and often removable. And in the most common case of all, the drug prescribed alongside the disease may be feeding it.
Kostoff is asking clinicians to do the unglamorous thing and find the cause before they medicate the symptom. The patient in the exam room, leaving with a gabapentin script and no B12 test, is the standing evidence that most of medicine still works the other way around.
Sources
- TrialSite – Ronald Kostoff, “Unified Drug-Free Healing Paradigm for Peripheral Neuropathy and Heavy Legs” (2026)
- Cochrane Database of Systematic Reviews – Wiffen et al., gabapentin for chronic neuropathic pain in adults (2017)
- American Academy of Family Physicians – Gabapentinoids for Pain: Potential Unintended Consequences (2019)
- Kostoff – Prevention and Reversal of Chronic Diseases: A Protocol (Public Health and Toxicology)
- PMC – Peripheral Neuropathy as a Consequence of Metformin-Induced Vitamin B12 Deficiency in T2DM (2017)
- PubMed – Metformin-induced vitamin B12 deficiency presenting as a peripheral neuropathy (2010)
- PubMed – Cardiovascular risk of gabapentin and pregabalin in patients with diabetic neuropathy (2022)
- ClinCalc DrugStats – Gabapentin, U.S. prescription volume and ranking
- Diabetes – Amelioration of Peripheral Neuropathy in Mouse Models of Diabetes by Dietary Reversal (2018)