Two published trials have tested laser photobiomodulation in people who already have peripheral neuropathy, and both were run in neuropathy caused by chemotherapy. One compared a very low-power laser against a sham device and found the laser group’s response rate holding at twelve weeks while the control group’s fell.1 The other used a class 4 laser, reported improvement in nerve score and walking distance, and had no control group at all.2 Neither reports nerve regeneration, reversal, or a cure, and neither enrolled patients whose neuropathy comes from diabetes or metabolic disease.
That is a thin evidence base, and the honest thing to do with a thin evidence base is not to shrink from it but to read it carefully: what each trial measured, who it let in, who it kept out, and what its design was never built to answer. That is what the rest of this page does.
The short version
- Two trials, two very different lasers. The only sham-controlled trial used a class 2M laser rated in milliwatts. The only trial that used a class 4 laser had no control group.
- Dose is the variable, not the word “laser”. Wavelength, power density, energy density and treated area differ enormously between the two trials, and light therapy has a documented biphasic dose response.3
- Order of events matters. In the sham-controlled trial the control arm responded slightly more often at six weeks. The arms separated only at twelve weeks, which raises a question about how long a course needs to be.1
- Both trials studied cancer survivors. Neither enrolled diabetic, nutritional, autoimmune or idiopathic neuropathy.
- The registered diabetic neuropathy laser trials screened out the harder patients — peripheral vascular disease, open wounds on the treatment area, recent diabetic ketoacidosis, and, in one protocol, anyone who had taken gabapentin, pregabalin, tramadol or an opioid in the previous month.4,5
- Coverage is a payer decision, not a clinical finding. The 2006 Medicare determination that gets quoted here is about infrared diode pads, not about laser photobiomodulation.6
- Map the driver first. Which mechanism is active decides whether light-based therapy belongs in the plan at all.
What the question actually means
When people ask whether class 4 laser helps neuropathy, they usually mean whether laser light improves neuropathy symptoms in humans. The evidence base is real, but it is small, and it does not travel evenly across neuropathy types.
Neuropathy is a category, not a diagnosis
The same burning-feet complaint can come from metabolic injury, a vitamin deficiency, a toxic exposure, chemotherapy, an autoimmune process, or a structural cause. Those are different diseases that happen to share a symptom.
A trial run in one of those populations says very little about the others. That is not a criticism of the trials. It is a statement about what a trial in a defined population can and cannot be extended to cover.
Laser class describes hazard, not dose
Laser safety classes run from 1 up to 4 and describe how hazardous a beam is to the eye and skin. They are a regulatory hazard rating. They do not tell you the wavelength, the power density at the tissue, the energy delivered, or the area treated — and those are the things that determine what happens biologically.
This matters more here than it usually does, because the two published trials sit at opposite ends of that scale, and the one with a control arm is the one at the bottom.
Dose is the variable that separates these trials
Both trials are described as photobiomodulation. In terms of the light actually delivered, they have almost nothing in common.
The parameters that define a photobiomodulation dose
A photobiomodulation dose is defined by wavelength, power, power density at the tissue surface, energy density (fluence), the size of the treated area, whether the beam is continuous or pulsed, how long each session lasts, and how many sessions are given over what period. Change any one of those and you have changed the treatment.
The sham-controlled trial used a class 2M continuous-wave diode laser delivering 8 milliwatts through a 3.2 millimeter aperture, applied point by point — sixteen points in the interdigital spaces of the hands and feet, plus ten points over the cutaneous landmarks for the C6–T1 and L5–S1 nerve roots on both sides. The dose was prescribed as 1 joule per point, escalating to 2 joules per point in later sessions as tolerated, twice a week for six weeks.1
The class 4 trial used a device combining two diodes at 905 nanometers and 808 nanometers, one pulsed and one continuous, with peak powers of 25 watts and 1.1 watts, a power density of 0.168 watts per square centimeter, and a beam spot of 150 square centimeters covering the plantar surfaces of the hands and feet bilaterally. The dose was prescribed as an energy density — 6 or 8 joules per square centimeter — twice weekly for three weeks.2
Notice that the two doses are not even reported in the same units. One is joules per point at twenty-six discrete points; the other is joules per square centimeter across a large field. There is no clean conversion between them, which means the two trials cannot be pooled, compared, or averaged in any meaningful way.
The biphasic dose response
This is not a bookkeeping problem. It is the central reason results in this literature diverge.
A 2009 review in the journal Dose-Response, from the Wellman Center for Photomedicine at Massachusetts General Hospital, described a biphasic dose response that has been observed repeatedly in low-level light therapy: low levels of light stimulate and repair tissue more effectively than higher levels do, a pattern the authors describe using the Arndt-Schulz curve.3 More is not better, and past a point more is worse.
The same review makes the point directly relevant to this page: the difficulty of choosing rationally among a large number of illumination parameters — wavelength, fluence, power density, pulse structure and treatment timing — has produced a body of literature containing both negative and positive studies.3 A negative photobiomodulation trial can mean the therapy does not work, or it can mean that particular dose did not work, and from outside the trial those two possibilities look identical.
So when a reader encounters mixed results in light therapy, the appropriate response is not to conclude the field is unserious. It is to ask what dose was tested, in whom, and for how long.
The sham-controlled trial: 44 cancer survivors, 2022
This is the only randomized, sham-controlled trial of photobiomodulation in established neuropathy discussed here, and it deserves to be read closely rather than quoted selectively.
Who was studied
Teng and colleagues, publishing in Supportive Care in Cancer, enrolled 44 cancer survivors who still had neuropathy symptoms at least three months after finishing potentially neurotoxic chemotherapy. Participants were randomized 2:1 — 29 to laser, 15 to sham — and treated twice weekly for twelve sessions over six weeks, with assessments at baseline, six weeks and twelve weeks.1
The group was 61 percent female with a mean age of about 62. Just under half had colorectal cancer and about a third had breast cancer; a mean of roughly fifteen months had passed since chemotherapy. Neuropathy was grade 1 in 64 percent and grade 2 in 36 percent — that is, mild to moderate.1
Who was screened out
The trial listed six specific exclusions, and they are worth reading as a description of who was not in the room.1
- Anyone unable to lie supine for a thirty-minute period.
- Anyone with an open wound or ulcer over the treatment area.
- Anyone with a clinical diagnosis of peripheral neuropathy from another cause. The paper notes that diabetes without neuropathy was not itself an exclusion — but diabetes with neuropathy was, because that is a second cause.
- Anyone with uncontrolled psychiatric illness or cognitive dysfunction that could interfere with completing assessments.
- Anyone with a life expectancy under three months.
- Anyone using a concurrent complementary therapy for neuropathy during the study.
Read together, this describes a relatively well, relatively mobile group with a single identified cause of neuropathy, intact skin on the feet, no competing treatments running alongside, and mild to moderate severity. That is a clean trial population. It is not the population that walks into a neuropathy clinic with type 2 diabetes, an old foot ulcer, three prescriptions for nerve pain and fifteen years of symptoms.
What the response rates showed, and in what order
Response rates were 48 percent in the laser group and 53 percent in the sham group at six weeks, and 45 percent and 33 percent respectively at twelve weeks.1 Stated plainly: the sham group responded slightly more often at the end of treatment, and the two arms separated only at follow-up, when the laser group held its gains and the control group’s declined.
That finding must stay on the page. It also has to be read with three facts attached.
First, this was a non-comparative trial by design. Each arm was tested against a fixed statistical threshold, not against the other arm. The published null hypothesis was that the true response rate is 5 percent, and the sample size was chosen to give 80 percent power if the true response rate on treatment were 20 percent.1 The authors state explicitly that the design did not permit statistical comparison between the treatment and control arms.
Second, the control arm had fifteen people in it. In an arm of fifteen, one participant is worth about seven percentage points. The five-point gap at six weeks is therefore smaller than a single person’s response.
Third, the laser arm’s response rate did exceed the trial’s prespecified threshold at both time points, and reported side effects were low grade.1 The authors’ own conclusion is measured: the result warrants evaluation in an appropriately powered phase III trial.
Why a sham arm is hard to build in a light-therapy trial
The paper describes its sham honestly. Sham participants got the identical visit schedule, the same eye mask, the same equipment and the same application points, with the laser aperture blocked by an opaque aluminum cover, so that the sensory experience would match the active treatment.1 That is a careful design, and with a device delivering 8 milliwatts there is little heat or sensation to give the difference away.
Blinding was still incomplete, and the authors say so. The trial was single-blinded. The clinical neuropathy assessment was performed by an unblinded physician investigator, and the paper’s own limitations list names the unblinded clinical assessor and therapist as a weakness.1 An unblinded assessor can shift a clinician-rated score in either direction.
None of this is a reason to discount the trial. It is a reason to hold both arms’ numbers loosely, including the six-week ones.
What this design could not capture
The authors list the limitations themselves: small sample size, a short intervention period, the unblinded assessor and therapist, and unavoidable delays to study visits caused by COVID-19 restrictions. They also note that no neurophysiological testing was performed because of resource constraints, and suggest it as an objective measure for a future trial.1
Three further things follow from the design rather than from any fault in it. The trial tested laser alone, so it cannot say anything about laser used alongside metabolic correction, nutritional repletion or exercise. Follow-up stopped at twelve weeks, which is six weeks after the last session, so nothing in it speaks to durability beyond that. And with no nerve conduction testing, the trial measured symptoms and function rather than the nerve itself.
The NEUROLIGHT pilot trial: 60 patients, a class 4 laser, 2026
The second trial is the one that actually used a class 4 device, and it is a pilot without a control group.
Who was studied
Claes and colleagues, publishing in Lasers in Medical Science, treated 60 cancer patients with chemotherapy-induced peripheral neuropathy at Jessa Hospital in Hasselt, Belgium, between February 2022 and November 2023. Twenty-eight received six sessions at 6 joules per square centimeter and thirty-two received six sessions at 8 joules per square centimeter, twice weekly over three weeks.2
Most participants were women, with mean ages around 61 to 64. Breast cancer was the most common primary, followed by colorectal, head and neck, ovarian, endometrial, bladder, lung and prostate.2
Who was screened out
The eligibility rules were narrower than they first appear. To be included, a patient had to have a physician diagnosis of chemotherapy-induced neuropathy, be at least 18, have had one of a specific list of neurotoxic agents — paclitaxel, docetaxel, oxaliplatin, cisplatin, thalidomide, bortezomib or vincristine — with the last dose at least two weeks earlier, and have a Fitzpatrick skin type between I and V.2
That last criterion excluded Fitzpatrick type VI, the most deeply pigmented skin. Pigment absorbs light, so this is a real limit on who the reported doses have been tested in.
The stated exclusions were two: interrupting more than two consecutive treatment sessions, and having had neuropathy symptoms before receiving chemotherapy.2 That second one is the important one. It removes exactly the patient who arrives at a neuropathy clinic with pre-existing diabetic or idiopathic neuropathy that chemotherapy then made worse — a common and difficult presentation, and one this trial cannot speak to.
What moved and what did not
The modified Total Neuropathy Score improved significantly over time (P = 0.048), as did the six-minute walk test (P < 0.001), with walking distance improving about 13 percent in the lower-dose group and about 18 percent in the higher-dose group. Pain scores on the numeric rating scale improved over time (P < 0.001), with better scores in the lower-dose group (P = 0.034). Patient satisfaction improved (P = 0.003). No adverse events were reported.2
Not everything moved. There was no significant change in the FACT/GOG-NTX total score or in its neurotoxicity subscale, meaning the broader quality-of-life instrument did not register the same signal.2
That split between a walking test and a questionnaire is common in neuropathy research, and it is a practical argument for tracking what you can do — how far you can walk, how steady you are — rather than a single global score.
It is also worth noting that the lower dose, not the higher one, produced the better pain scores, and that the authors decline to name an optimal fluence on that basis.2 That is the biphasic dose response showing up inside a single trial.
What this design could not capture
The central limitation is structural: both groups received photobiomodulation. There was no sham arm, so improvement over time cannot be separated from natural recovery, from regression to the mean, or from a placebo effect. The authors say this plainly and accept that a placebo effect could partly explain the results.2
They also report that the sample was too small to analyze results by chemotherapy type, number of cycles, time since chemotherapy ended, prior lines of neurotoxic treatment, or neuropathy-inducing comorbidities such as diabetes — and that the effect of concurrent drugs like duloxetine and pregabalin was not assessed.2 In other words, the trial could not tell you whether a diabetic patient in the group did better or worse than a non-diabetic one.
Twenty of the sixty patients were lost to follow-up, twelve in the lower-dose group and eight in the higher-dose group.2 That is a third of the sample, and it is a further reason to read the later time points cautiously.
How long does a course need to be?
The trajectory in the sham-controlled trial is the most practically useful thing on this page. The laser arm did not lead at six weeks. It led at twelve.1
Two readings are available and the trial cannot distinguish between them. Either the laser effect takes longer to appear than a treatment course lasts, or the control group’s early response was a short-lived placebo response that faded on schedule while the laser group’s did not. Both are consistent with the numbers.
Either way, the practical implication is the same: judging a course of photobiomodulation at the moment the last session ends is judging it too early. It also means follow-up length is a fair question to ask of any trial in this field. The sham-controlled trial stopped at twelve weeks. The class 4 pilot collected data at baseline, after the six sessions, three weeks later, and again at six months and one year — but with no control arm, its long-term numbers cannot be attributed to the laser.1,2
Diabetic peripheral neuropathy: the gap, and who the trials there screened out
Diabetic peripheral neuropathy is the clearest gap in this evidence. Neither trial above enrolled that population, and the mechanisms differ: metabolic and microvascular injury varies between patients in a way that a single fixed protocol is unlikely to address uniformly. Several of those drivers are covered in our review of the hidden drivers behind peripheral neuropathy.
Two registered sham-controlled trials, and what they excluded
Laser trials in diabetic neuropathy have been registered and run. Two sham-controlled trials of the same 635-nanometer, three-diode, 17-milliwatt device were registered on ClinicalTrials.gov, both applying the laser fifteen minutes per foot, twice weekly for six weeks.4,5
The first completed with 30 participants.4 The second was terminated because it could not recruit, after enrolling 18; summary results are posted on its registry record.5 Neither is a peer-reviewed publication, and neither is large enough to settle anything. What they do document precisely is who was allowed in.
Both protocols required foot pain that was chronic, bilateral, and roughly symmetrical between the two feet, rated at least 50 on a 0–100 visual analog scale, on a stable diabetes medication regimen for the preceding 30 days.4,5
The exclusions are the striking part. Across the two records, participants were excluded for any of the following:4,5
- Foot pain that was unilateral or noticeably different between the two feet.
- Pain rated below 50 on the visual analog scale — that is, milder cases.
- Foot pain not definitively attributed to diabetes, or attributed to diabetes plus something else.
- Serious organ disease or other serious primary disease.
- Diabetic ketosis, ketoacidosis or severe infection within the previous two weeks.
- An active chronic pain condition — the second record names chronic fatigue syndrome, fibromyalgia, endometriosis, inflammatory bowel disease, interstitial cystitis and peripheral vascular disease.
- Cancer, or treatment for cancer, in the previous six months.
- An active infection, wound or other external trauma anywhere on the treatment area.
- Previous surgery or an implanted device for diabetic neuropathy foot pain.
- Analgesics or NSAIDs within seven days; antidepressants within 30 days; gabapentin, pregabalin, tramadol or opioid medicines within 30 days; local anesthetic injections within 30 days.
- Ongoing litigation or disability benefits related to the study parameters.
Why that matters for a patient with metabolic disease
Line those criteria up against a real clinic waiting room. A patient with peripheral vascular disease is out. A patient with a healing ulcer is out. A patient whose neuropathy is part metabolic and part something else is out. A patient with fibromyalgia alongside their neuropathy is out. A patient on gabapentin or pregabalin — which is a large share of people who have been treated for painful neuropathy — is out unless they stop for a month. A patient whose pain is bad in one foot and mild in the other is out.
This is not sharp practice by the investigators. Trials are built this way on purpose, to reduce noise and isolate a single effect. But it has a consequence that gets skipped over constantly: a modest or null result in a population selected this tightly does not establish that the treatment fails in the patients who were screened out. It establishes what happened in the patients who were let in.
The patients who were screened out here — vascular disease, unstable glycemic control, open wounds, multiple pain conditions, polypharmacy, mixed causes — are precisely the patients a metabolic neuropathy practice sees most days. The evidence is genuinely silent about them. Silence is not a negative finding, and it is not a positive one either.
Coverage is a payer decision, not a clinical finding
A national coverage determination in effect since October 24, 2006 states that infrared therapy devices are not covered by Medicare for treating diabetic or non-diabetic peripheral sensory neuropathy, or wounds or ulcers of the skin and subcutaneous tissues, including for pain arising from those conditions.6,7
The determination is specific about what it is describing. It defines an infrared therapy device as an array of juxtaposed infrared diodes affixed to a flexible pad held in contact with the skin, which can also produce local warming as a secondary effect.6 That is a different piece of equipment from a laser, and the determination was written about the evidence for those pad devices.
It is a payer decision about a device category, not a clinical conclusion about laser photobiomodulation, and it should not be read as either an endorsement or a verdict. Because it dates from 2006 it also predates both trials on this page. Local coverage rules may differ, and nothing here is a benefits determination for any individual.
Where a service is not covered, patients sign an Advance Beneficiary Notice before it is provided, so that the financial side is settled in writing in advance rather than discovered afterward. It is fair to ask any clinic which device it is using and which coverage language it is relying on.
How light-based therapy fits into an evaluation-first plan
Class 4 photobiomodulation, class 3B cold laser, and whole-body infrared are offered here as part of a plan built around whichever driver the evaluation identifies. They are not offered as nerve regeneration or reversal, and no success rate is published for them.
Before any of it is discussed, the evaluation looks for which mechanism is active. Electrodiagnostic testing (EMG/NCS) is performed on site, and VNG (videonystagmography) is available as a diagnostic test of inner-ear balance function when dizziness or unsteadiness is part of the picture — it is a test, never a treatment.
The published trials tested light on its own, which is the only way to isolate its effect but not the way neuropathy is usually treated. Here, light-based therapy sits alongside metabolic and nutritional care aimed at the driver itself. That combination has not been tested in a randomized trial, and saying so is more useful than implying it has.
Orthobiologic injections (PRP and BMAC) are offered here as well.
What all of this adds up to is a conversation rather than a protocol. The risks of photobiomodulation reported in these trials were low — mild, low-grade events in one and none in the other — and the potential benefit in a patient like yours is genuinely uncertain. Both halves of that sentence belong in the discussion before anyone starts a course of treatment.
What to ask before you commit to a course of treatment
- Which neuropathy driver do you believe is dominant in my case, and what did you base that on?
- Is the plan supported by trial evidence in a population like mine, or extrapolated from a different one?
- If it is extrapolated, what specifically about my case makes you think it transfers?
- How many sessions, over what period, at what wavelength and what energy density?
- When will we judge whether it is working, and why then rather than at the last session?
- What function will we measure, not just what pain score?
- Which device is this, what class is it, and what coverage language applies to it?
Where we are and how to start
Regenerve is located at 4477 Woodson Rd #104, St. Louis, MO 63134, minutes from St. Louis Lambert International Airport, and we see patients from across the St. Louis region, Missouri and Illinois. You can call or text (314) 886-5902, or email info@regenerve.com.
The free Nerve Damage Score is a five-question starting point for the driver question, which is the question that decides whether light-based therapy belongs in your plan at all.
Frequently asked questions
Is the evidence for photobiomodulation in neuropathy strong enough to rely on?
On its own, no — there are two published trials in established neuropathy, both in chemotherapy-induced cases, one small and sham-controlled and one with no control group at all.1,2 That is a reason to treat it as one component of a plan aimed at an identified driver, rather than as the plan. It is not a reason to call it disproven, because the trials that exist have not studied most of the people asking the question. See the treatments offered here.
Does that evidence transfer to diabetic neuropathy?
Not automatically. Diabetic peripheral neuropathy involves metabolic and microvascular drivers that vary between patients, and neither published trial on this page enrolled that population. The registered diabetic neuropathy laser trials that do exist excluded peripheral vascular disease, open wounds on the feet, recent ketoacidosis and patients taking common nerve pain medications, so they say little about a complex metabolic patient either.4,5 See how diabetic peripheral neuropathy is approached here.
The sham group did better at six weeks. Does that mean the laser did nothing?
It means the trial did not show a laser advantage at the end of treatment, which is worth knowing. It does not mean nothing happened, for three reasons: the trial was designed as non-comparative and its authors state it could not statistically compare the arms; the control arm had only fifteen people, so a five-point gap is smaller than one participant’s response; and by twelve weeks the laser group had held its response while the control group’s had fallen.1 See which specialist to see when neuropathy has not improved.
How is photobiomodulation different from Qutenza?
They are unrelated mechanisms with very different evidence. Qutenza (capsaicin 8% patch) is FDA-approved in adults only for neuropathic pain from postherpetic neuralgia and from diabetic peripheral neuropathy of the feet, and any other use is off-label. It is applied in clinic by a clinician, never dispensed for home use, and repeated no more often than every three months. See how Qutenza is used in clinic.
Will Medicare cover light therapy for neuropathy?
A national coverage determination in effect since October 2006 makes infrared therapy devices non-covered for peripheral sensory neuropathy, wounds and ulcers.6 That determination describes arrays of infrared diodes on a flexible pad, not laser photobiomodulation, and coverage is a payer decision rather than a statement about whether something works. Where a service is not covered, an Advance Beneficiary Notice is signed before treatment. See contact us with coverage questions.
Sources
- Teng C, Egger S, Blinman PL, Vardy JL. Evaluating laser photobiomodulation for chemotherapy-induced peripheral neuropathy: a randomised phase II trial. Supportive Care in Cancer. 2022;31(1):52. https://pmc.ncbi.nlm.nih.gov/articles/PMC9758032/ (44 cancer survivors at least 3 months after neurotoxic chemotherapy; randomized 2:1, 29 laser and 15 sham; class 2M continuous diode laser, 8 mW through a 3.2 mm aperture, 1–2 J per point at 26 points, twice weekly for 12 sessions; response rates 48% laser and 53% sham at 6 weeks, 45% and 33% at 12 weeks; non-comparative design; null hypothesis of a 5% true response rate; sham delivered with the aperture occluded by an opaque aluminium cover; limitations include small sample, short intervention period and unblinded clinical assessor and therapist).
- Claes M, Lodewijckx J, Robijns J, Hermans S, Peeters P, Mebis J. Evaluating the efficacy of photobiomodulation therapy in the management of chemotherapy-induced peripheral neuropathy: a pilot trial (NEUROLIGHT trial). Lasers in Medical Science. 2026;41(1):57. https://pmc.ncbi.nlm.nih.gov/articles/PMC12988979/ (60 cancer patients with chemotherapy-induced peripheral neuropathy at Jessa Hospital, Hasselt, Belgium, February 2022 to November 2023; class IV MLS M6 laser, 905 nm and 808 nm, peak power 25 W and 1.1 W, power density 0.168 W/cm², 150 cm² beam spot; 6 sessions at 6 J/cm² (n = 28) or 8 J/cm² (n = 32); Fitzpatrick skin types I–V only; exclusions were interruption of more than two consecutive treatments and neuropathy symptoms preceding chemotherapy; mTNS P = 0.048, 6MWT P < 0.001, NRS pain P < 0.001 favoring the 6 J/cm² group P = 0.034, no significant difference in FACT/GOG-NTX total score or neurotoxicity subscale; no control group; 20 of 60 lost to follow-up).
- Huang YY, Chen AC, Carroll JD, Hamblin MR. Biphasic dose response in low level light therapy. Dose-Response. 2009;7(4):358–383. https://pmc.ncbi.nlm.nih.gov/articles/PMC2790317/ (low levels of light stimulate and repair tissue more effectively than higher levels, described by the Arndt-Schulz curve; the difficulty of choosing among wavelength, fluence, power density, pulse structure and treatment timing has produced both negative and positive studies).
- ClinicalTrials.gov. Study of Low Level Laser Therapy to Treat Diabetic Peripheral Neuropathy Foot Pain (NCT02461225). Erchonia Corporation. https://clinicaltrials.gov/study/NCT02461225 (double-blind, placebo-controlled randomized evaluation of the Erchonia FX-635; completed with 30 participants; three 17 mW 635 nm diodes, 15 minutes per foot, twice weekly for 6 weeks; full inclusion and exclusion criteria as summarized above).
- ClinicalTrials.gov. An Evaluation of the Effect of Low Level Laser Therapy on Diabetic Peripheral Neuropathy Pain (NCT04006392). Erchonia Corporation. https://clinicaltrials.gov/study/NCT04006392 (double-blind, placebo-controlled randomized evaluation of the Erchonia FX-635; terminated for inability to recruit after enrolling 18 participants; results posted March 14, 2023; exclusions include peripheral vascular disease, active infection or wound on the treatment area, serious organ disease, and recent diabetic ketosis or ketoacidosis).
- Centers for Medicare & Medicaid Services. National Coverage Determination 270.6: Infrared Therapy Devices. Effective October 24, 2006. https://www.cms.gov/medicare-coverage-database/view/ncd.aspx?ncdid=315&ncdver=1 (infrared and near-infrared light and heat, including monochromatic infrared energy, non-covered for diabetic and non-diabetic peripheral sensory neuropathy, wounds and ulcers of the skin and subcutaneous tissues, including pain arising from those conditions; the determination describes an array of juxtaposed infrared diodes affixed to a flexible pad held against the skin).
- Centers for Medicare & Medicaid Services. Decision Memo for Infrared Therapy Devices (CAG-00291N). 2006. https://www.cms.gov/medicare-coverage-database/view/ncacal-decision-memo.aspx?proposed=N&ncaid=176 (the national coverage analysis underlying NCD 270.6).