96.8% of people who got steroid injections for Meniere's said their vertigo improved. In the placebo group it was 96.6%. That is from a 2023 Cochrane review: two small trials, 60 people in total.
The same questions keep coming up in the Meniere's community: does salt matter, does betahistine work, are the injections worth it. So I read the guideline US ENTs work from (AAO-HNS, 2020) and the Cochrane reviews behind it. This is a summary of what those documents say.
Steroid injections: almost everyone improved
Nearly everyone in both groups said their vertigo was better. With so few people studied and such a high improvement rate on placebo, the trials cannot tell a treatment that works from one that does not. The same review found that steroid injections may slightly reduce how often attacks happen (3 trials, 472 people), but that evidence is low-certainty too.
Salt: barely tested
The guideline lists one small randomized trial of salt restriction, and it found no effect. Cochrane looked for placebo-controlled trials of salt or caffeine restriction and found none. So it is not proven and not disproven. It has barely been tested.
The guideline still recommends talking to patients about diet changes, because the risk is low. If you want to see where your own intake sits, the sodium calculator and the low-sodium diet guide are a good start.
Betahistine: placebo did as well
In the best trial so far (BEMED), attacks went down over 9 months in the betahistine groups and in the placebo group, with no significant difference between them.
The line that explains a lot
The guideline says up to 60% vertigo control has been seen in placebo groups. Meniere's comes in waves and often quiets down on its own. People also tend to start something new when their symptoms are at their worst, and symptoms often settle afterwards anyway. That makes it very easy to credit, or blame, the wrong thing.
What this does not mean
None of this means stop what you are doing. Something with weak trial evidence can still be what works for you, and that is a conversation for your ENT. "Not well tested" and "does not work" are different statements, and most of what is listed here is the first one.
Testing one trigger at a time
If trials cannot settle it, a careful personal experiment is the next best thing: change one thing, keep everything else steady, log it, and give it long enough to see past a lucky month. The caffeine and alcohol guide walks through a four-week version.
For those who tested one trigger at a time: did the result match what you believed before?