When no controlled trial of a tinnitus treatment exists — as with cupping, reflexology and reiki — that absence is the answer, and three questions about any study you are shown will tell you most of the rest.
This site has articles on ginkgo, acupuncture, PEMF, low-level laser, hyperbaric oxygen, CBD, magnesium and zinc, melatonin, branded tinnitus supplements such as Lipo-Flavonoid and chiropractic care. It has none on cupping, reflexology, reiki or crystal healing.
That is not an oversight, and leaving it unexplained would be its own kind of dishonesty. This page is the explanation, and more usefully the method, because the same three questions work on whatever you have just been offered.
The search, stated so you can rerun it
In September 2026, on PubMed, searching the treatment name together with "tinnitus":
- Cupping — six records. None is a controlled trial. Two are single-patient case reports, one for Ménière's disease and one for sudden hearing loss. The rest study cupping as one component of a combined acupuncture, moxibustion and cupping protocol, which cannot tell you what the cupping contributed.
- Reiki — nothing. Not a weak literature; an empty one.
- Reflexology — dozens of records, and this is where it gets interesting. See below.
- Homeopathy — a small number, including one useful trial. Also below.
You can run each of these yourself in about a minute. That is the point of putting the search terms here rather than only the conclusion.
Why the number of results tells you almost nothing
Reflexology looks, by count, like a researched topic. Open the records and they are studies of manual therapy for temporomandibular and cervical disorders, physiotherapy for neck pain, osteopathic correction of jaw dysfunction, auricular acupressure. Real research on real interventions — the site covers that territory in TMJ and tinnitus and somatic tinnitus — but not one of them is a study of foot reflexology.
Filtering to randomised controlled trials is meant to fix this, and it introduces its own trap. Filter the homeopathy search that way and one of the trials returned is a study of a homeopathic remedy for the side effects of quinine, in pregnant women being treated for malaria in Benin. Tinnitus is a quinine side effect, which is why the record matches. It is not a trial of homeopathy for tinnitus.
A marketing page can truthfully say "dozens of studies" about either of these. The count is not the evidence. The papers are, and reading titles is enough to sort most of it.
The one that was tested properly
Homeopathy for tinnitus is worth following through, because it went further than most and the result is instructive.
In 1998 the British Journal of Audiology published a double-blind, placebo-controlled trial of a homeopathic preparation at D60 potency against a matched placebo, in 28 people with tinnitus. Outcomes were assessed at four points using visual analogue scales, questionnaires, and a battery of audiological measurements.
Two things came out of it, and the gap between them is the whole lesson. Fourteen of the 28 participants said on a questionnaire that they preferred the remedy to the placebo — half the group, which as a testimonial would be persuasive. And the analysis of variance found no significant improvement on either the visual analogue scores or the audiological measures. The authors concluded that the preparation could not be shown to be more effective than the matched placebo.
Set against it, a 2007 paper in the International Tinnitus Journal on a homeopathic product for tinnitus relief: 15 patients, no control group, no placebo, no randomisation, a descriptive analysis, and a positive conclusion.
That is the same shape as much of the laser therapy literature, and once you have seen it you keep seeing it: the well-designed study finds nothing, the uncontrolled one finds benefit, and both get cited as "studies".
The three questions that do the work
You do not need to appraise a paper properly to get most of the value. For anything you are offered:
- Was there a placebo group, and were people randomly assigned to it? Tinnitus fluctuates on its own, and expectation moves how bothersome it feels. Without a control group a study cannot separate the treatment from either. This one question disqualifies most of what gets cited.
- How many people? Fifteen is a pilot at best. The tinnitus literature is full of small studies, and small positive results are the ones least likely to survive a larger trial.
- Was this treatment tested on its own? The cupping records are the clean example — cupping given alongside acupuncture and moxibustion, in a study that reports the package. Whatever happened, it cannot be attributed to the cupping.
The how this site reads evidence page describes the standard applied here, and finding and reading trials goes further into registry entries and what a trial record tells you before any results exist.
What absence of evidence is and is not
It is not proof that something does nothing. Plenty of effective treatments were unstudied before they were studied, and a treatment that has never been trialled is in a different position from one that has been trialled and failed.
What it does mean is concrete. Nobody has shown that it works; you are being asked to pay for it anyway; and the cost is not only money but the attention and hope spent on it, which are finite in a condition where the approaches with real evidence ask for patience.
This page is about the case where nothing has been published. The harder case is where something has been published and the result is weaker than it sounds — the patterns that make a weak tinnitus result look strong sets out nine of them, each with a cited example from this site.
If you want to try something harmless while keeping that in view, that is a reasonable adult decision and this site is not going to lecture you about it. What is worth refusing is the framing that it has been shown to work, when a minute on PubMed shows otherwise. One treatment does have a page here despite failing every test above, and for a reason worth stating: ear candling is not merely unsupported, it can cause the blockage it claims to relieve. Absence of evidence and evidence of harm are different findings.