So many people with long COVID were already taking B vitamins that it was the top reason for turning them away from this trial. Then, over 28 days, the vitamins did no better than going without

Screening for this trial turned up a telling detail. Of 2,498 people with long COVID assessed at three German university sites, only 323 were randomised, and the main reason people were ruled out was recent use of one of the trial's treatments. B vitamins were the most common cause. Those who did join had been ill for a median of 79 weeks and had already tried an average of 2.8 treatments for it.

The PreVitaCOV trial, published in eClinicalMedicine on 24 October, is the first randomised, double-blind, placebo-controlled test of these treatments in long COVID, according to its authors. Everyone took two things each morning for 28 days. One was prednisolone, a steroid (20 mg for five days, then 5 mg for 23 days), or a dummy. The other was a B vitamin tablet (100 mg of B1, 50 mg of B6 and 500 micrograms of B12), or a dummy. That made four groups, and it let the team judge each treatment across the two groups that got it and the two that didn't.

The B vitamins didn't move the main measure, a combined score for fatigue, breathlessness, thinking problems, anxiety and depression, where lower is better. After adjusting for starting scores, people on the vitamins differed from those not on them by -0.1 points (95% CI -1.2 to +1.0). On a separate fatigue questionnaire, the estimate leaned slightly the wrong way, at +0.8 points, and that wasn't significant either. Quality of life, pain, mood and a one-minute sit-to-stand test showed no significant effect of the vitamins.

Yet people on the vitamins did feel better. Their score fell by 3.2 points over the month, against 2.9 on placebo. This is the part worth remembering if you've ever taken something and improved: both changes were statistically significant within their own groups, and only the comparison shows the vitamins weren't the reason. The authors say the placebo effect, the natural ups and downs of the illness, and the attention that comes with being in a trial could all have contributed. Some participants, they add, seemed to see taking part as a way to win recognition for the condition, though nobody measured what that did for them.

Prednisolone did better, but only modestly. Its groups improved by 4.3 and 4.5 points, an adjusted difference of -1.4 (95% CI -2.5 to -0.4). The team had set 3 points as the change that would count as clinically meaningful, so this result is statistically significant and still falls short. They worked out that about 7.5 people would need treating for one extra person to get a meaningful result, but that was a post-hoc figure and its interval runs from 4 to 55. The differences between groups were no longer significant on days 60 and 180, after the capsules had stopped.

Safety looked unremarkable over this short stretch. Adverse events were reported by 78% to 86% of each group, placebo included, and no serious event was judged related to the treatment. But a month tells you little about the steroid risks the authors themselves list, including high blood sugar, high blood pressure and mood changes. People with diabetes or severe comorbid conditions weren't allowed in, so the trial says nothing about them. This isn't a case for trying prednisolone at home: the authors say any benefit must be weighed against harms, and that the result needs confirming.

There are limits on the vitamin finding too. The main paper reports no blood levels of B1, B6 or B12, so we can't tell whether anyone was short of them to begin with. A real deficiency is a separate question, and one a blood test and a doctor can answer. The trial ran for 28 days in one country and left out people who were bedridden or needed nursing care. Results from its computer-based attention tests will be reported separately, so that part of the thinking story isn't in yet.

What it does offer is a clear answer for this one combination: at these doses, for a month, a high-strength B vitamin added nothing measurable to what a dummy capsule did. And the B6 dose raises a question of its own for anyone taking it for much longer. At 50 mg a day, it's more than four times the 12 mg a day the European Food Safety Authority set in 2023 as the adult upper level, because too much B6 can damage nerves.