Combining two tinnitus treatments worked slightly better on average in a 461-patient randomised trial, about 3 points on a 100-point scale, but only because a stronger treatment offset a weaker one, not because treatments add up. "Try several things at once" is reasonable-sounding advice, and until recently there was no good trial behind it. Now there is one, it is large, and its result is more interesting than a yes.
What the trial did
An international, multicentre, parallel-arm, superiority randomised controlled trial, published in Nature Communications in 2025, set out to test whether combination therapies beat single interventions for chronic subjective tinnitus. It is the UNITI trial — registered as NCT04663828, the same identifier carried by its published statistical analysis plan — the EU study this site has been describing as pending for some time.
- 674 people screened, 461 enrolled (190 female), aged 18 to 80, all with chronic subjective tinnitus and at least mild tinnitus handicap.
- Recruited from five clinical sites across the EU, randomised by a web-based system and stratified by hearing and distress level so the arms started comparable.
- 230 assigned to a single treatment, 231 to a combination of two.
- Four treatments, alone or paired, producing ten treatment arms: cognitive-behavioural therapy, hearing aids, app-based structured counselling, and app-based sound therapy.
- 12 weeks. Primary outcome: change in total Tinnitus Handicap Inventory score from baseline to week 12, intention-to-treat. All statistical analysis was performed blinded to treatment allocation.
Two design details are worth flagging before the result. The app-based arms were delivered without any direct contact or guidance from clinicians — fully unguided, so the result speaks to apps used on their own rather than with a clinician's support. And the analysis was blinded even though the participants could not be, which is the right mitigation for a trial where concealing the treatment is impossible.
What it found
Every arm improved. The Tinnitus Handicap Inventory runs from 0 to 100, so these are changes on a 100-point scale. Least-squares mean changes from baseline to week 12:
- Single treatment: −11.7 (95% CI −14.4 to −9.0)
- Combination treatment: −14.9 (95% CI −17.7 to −12.1)
- p = 0.034 for the difference
So combination treatment won its superiority test. No serious adverse events occurred.
Now size it. The gap between the two strategies is about 3.2 THI points. The responder threshold this field commonly applies to the THI — the one the bimodal device trial used to define who counted as improved — is 7 points.
The advantage of combining two treatments over one was real, statistically significant, and roughly half the size of the smallest change a person tends to notice. Both strategies produced improvements well above that threshold on their own. It is the difference between them that is small.
The finding that matters most
Here is where this trial earns its place, and it is in the mechanism rather than the headline.
The authors looked for a synergistic effect — two treatments producing more together than either could alone — and did not find one. What they found instead they call a compensatory effect: a more effective treatment offsets the clinical effects of a less effective treatment.
Read that carefully, because it inverts the usual reading of "combination therapy is better".
It means combinations did not amplify anything. They averaged. Pair a strong treatment with a weak one and the strong one carries the pair — which lifts the combination group's average above the single group's, because the single group contains people who received only the weak options with nothing to offset them.
The trial states the consequence explicitly: cognitive-behavioural therapy and hearing aids alone had large effect sizes, which could not be further increased by combination treatment.
What that means if you are the patient
If CBT or properly fitted hearing aids are already in your plan, this trial does not support adding a second treatment to improve the result. Those were the two strong components, and combining them with something else did not make them work better. The gain from stacking went to people whose single treatment was one of the weaker ones.
If neither is available to you, that is the gap worth closing — before adding a second thing. A second modality does not substitute for access to an effective first one, and the honest reading of this trial is that which treatment you get matters more than how many.
Where combination genuinely helps is as insurance. If nobody can tell in advance which treatment will suit you — and mostly nobody can — then a combination reduces the chance of spending twelve weeks on the one that does not. That is a real benefit. It is just a different benefit from "two treatments work better than one", and it argues for pairing a strong treatment with a plausible second rather than for accumulating treatments indefinitely.
What it does not settle
It tested four treatments, not all of them. Bimodal neuromodulation, sound therapy as ordinarily delivered, and everything else on the comparison table were not in these arms. A compensatory pattern among these four is suggestive about combining treatments generally; it is not a measurement of it.
Twelve weeks is twelve weeks. Whether the gap between strategies widens, closes or reverses over a year is not something this trial can say.
The app arms were unguided by design, so their weaker showing is evidence about unguided apps specifically, not about app-delivered therapy with a clinician involved — which is a different intervention with its own trial evidence, and a considerably better track record.