This post was amended based on the comment from the NNT team below.
Written by Clay Smith
Spoon Feed
Supplementation with vitamin D and calcium, in addition to standard canalith repositioning maneuvers, reduced the number of episodes of benign paroxysmal peripheral vertigo (BPPV) per year in adult patients.
Why does this matter?
Otoconia are calcium crystals that make the otolithic membrane heavier than the surrounding fluid. Calcium is required for otoconia crystallization. Low calcium may lead to greater otoconia turnover and release of crystalline debris, which might precipitate an attack of BPPV. Anecdotally, some patients have improved with vitamin D and calcium supplementation. Would this prove true in a RCT?
Epley + calcium = relief? Looks like it may…
This was a multicenter randomized trial that enrolled 1050 patients with BPPV who all had canalith repositioning maneuvers. About half received vitamin D and calcium supplementation, and half did not. Specifically, the treatment group got vitamin D 400 IU and 500 mg of calcium carbonate twice a day for one year if their serum vitamin D level was lower than 20 ng/ml. There was no placebo, and the non-treatment patients were not to take vitamin D or calcium supplements. For the primary outcome, the annual recurrence rate per person-year using intention to treat was 0.83 in the treatment arm vs. 1.10 in the non-treatment group. Two patients in the treatment arm developed hypercalcemia and had to stop medication.
The trial was retrospectively registered on ClinicalTrials.gov due to translation delays from Korean to English. That’s not ideal. They also changed the primary outcome from proportion of patients with recurrence to the absolute number of recurrences per patient (annual recurrence rate of BPPV per one-person year). Turns out, the overall proportion of patients with recurrence was also lower with treatment: 37.8 vs. 46.7%, p=0.005 (NNT = 11). There was large loss to follow up, with only 80% of the intervention group making it a year and 92% of the non-treatment group. Despite these challenges, on intention to treat, per protocol, and Kaplan-Meier time to event curve analysis, all showed benefit with treatment. I think this treatment effect is probably real. But I wonder if expecting to test vitamin D levels, treat with vitamin D and calcium, and recheck calcium and vitamin D levels for a year is realistic? Something more pragmatic might be a better approach in future studies.
Source
Prevention of Benign Paroxysmal Positional Vertigo with Vit D Supplementation: A Randomized Trial. Neurology. 2020 Aug 5:10.1212/WNL.0000000000010343. doi: 10.1212/WNL.0000000000010343. Online ahead of print.
Open in Read by QxMD
The researchers of this open-label, outpatient, telephone follow-up, Vitamin D trial report an erroneous NNT roughly three times lower (better) than the correct value. Their ‘NNT’ of 3.7 is an inversion of the absolute risk difference (0.27) between groups in the primary outcome. However their primary outcome is a difference between two continuous variables, and thus a continuous variable itself, and cannot be used to calculate an NNT. The correct NNT is roughly 11, reflecting an 8.9% difference in their secondary outcome (originally their primary outcome, changed after the trial began) comparing proportions of patients with recurrence.
It would be helpful if you could point this out to readers, as the researchers (and editorial staff of the journal Neurology), among others around the world, are apparently unfamiliar with the NNT concept. This trial’s abstract and paper are quite misleading for those who use the NNT as a comparative metric.
Thanks,
The NNT Team
You are absolutely right. I have amended the post. Apologies for missing that.