Vitamin D improves mood

People with low vitamin D report low mood more often. That link is real and it shows up consistently across large observational populations. What has not followed is the next step: when trials give people vitamin D and wait, mood mostly does not move. The exception is short, mostly small trials in people who already have depression, which often report improvement; longer trials in the same kind of people mostly do not, and the two largest recent ones split.

The three biggest trials found nothing. VITAL-DEP followed 18,353 US adults over 50 for a median of 5.3 years on 2,000 IU a day. Depression risk was the same in both groups, and the average change in mood score was 0.01 points on a scale where 0.5 was set in advance as the smallest difference that would matter. D-Health followed 21,315 older Australians for five years on monthly high-dose vitamin D and found a 0.02 point difference and no change in how many people started an antidepressant. The Finnish FIND trial gave 2,434 adults over 60 either 1,600 or 3,200 IU a day or placebo for five years and found no difference in new diagnoses of major depression, at 14, 11 and 8 cases. In a longer follow-up that continued for about three years after the tablets stopped, the higher-dose group had fewer diagnoses, a difference at the edge of statistical significance that rests on 63 cases in all.

The obvious response is that these were people who mostly were not deficient, so there was nothing to fix. That response has been tested. D-Vitaal recruited 155 older adults who were both low in vitamin D and already had clinically relevant depressive symptoms, and gave them 1,200 IU a day for a year. Their blood levels roughly doubled. Their depressive symptoms did not change, and neither did anything else the trial measured.

Two newer trials in adults already being treated for depression point in opposite directions. In Finland, DepFuD gave 281 outpatients aged 18 to 65 either 4,000 or 400 IU a day for six months alongside their usual treatment, and depression scores did not differ between the two doses. Because both groups took some vitamin D, it tests whether more helps rather than whether any does. In Brazil, a trial of 224 adults with recurrent major depression, all continuing their usual treatment, gave 50,000 IU a week or placebo for six months and reported clinician-rated depression scores falling from about 27 to about 8 on vitamin D, against about 25 to about 22 on placebo. That trial assigned people by coin toss, its vitamin D group started with more severe depression, and the recruitment and follow-up dates it reports span about five months, shorter than its six-month treatment period.

Meta-analyses pooling all the trials do report a benefit, and reading why they disagree with the large trials is the most useful thing here. The broadest pooled analysis cited here, 31 trials and 24,189 participants searched to the end of 2022, finds a small improvement overall and a larger one in people who already had depressive symptoms. Its own subgroup breakdown then shows where that comes from: trials lasting eight weeks or less show a clear effect, trials running 24 to 52 weeks show almost nothing, and trials longer than a year point very slightly the wrong way. The difference between those groups was statistically significant. VITAL-DEP is inside that pool and contributes about three quarters of its participants, which is possible because pooling weights trials by the precision of their result rather than by how many people they enrolled. That analysis, and the 29-trial one described below, also include a small 2016 trial in people with depression that its journal retracted in 2021. Two pooled analyses have tried to pin the effect on starting vitamin D level and reached opposite answers. One found benefit only in people above 50 nmol/L and none below it. The other found benefit concentrated in people below 50 nmol/L. The second pooled 29 trials totalling 4,504 participants, meaning no large trial is inside it at all. Where a moderator flips direction depending on which trials are pooled, the moderator is not doing the work the trial size and duration are.

The biology is not the weak part. Vitamin D receptors are genuinely present in the adult brain, confirmed by proteomics after years of contested antibody work, and mapped in mice to the cortex, amygdala and hypothalamus, with the receptor's natural ligand activating those specific neurons. The same work also reports that receptor expression in brain is lower than in gut or kidney, and that two antibodies previously published as brain-specific are not. A plausible pathway exists. It has not produced a reproducible effect on mood in people. Roughly 60 to 70 randomized trials have now tested vitamin D against a depression or mood outcome, with doses from 400 IU a day to 50,000 IU a week over durations from five days to five years, in people whose starting levels, ages and depression status all differed. NIH's Office of Dietary Supplements summarises the same split: low vitamin D is associated with higher depression risk, and supplementation has not been shown to prevent depression or treat depressive symptoms in trials.

Vitamin D has a documented harm at high intakes, which is too much calcium in the blood. LiverTox, the US government's drug-induced liver injury database, reports that vitamin D intoxication generally arises with intakes above 50,000 IU a day, and that lower doses can cause it in susceptible people such as those with renal osteodystrophy, a bone disorder of kidney disease. It lists symptoms including thirst, nausea, vomiting, constipation, fatigue, bone pain and muscle cramps, with possible kidney damage, and names mislabeled supplements, over-fortified milk or foods, and prescription or dispensing errors as common causes. It rates vitamin D an unlikely cause of liver injury. The Office of Dietary Supplements lists 4,000 IU a day as the upper intake level for adults. The Brazilian trial's 50,000 IU a week averages about 7,100 IU a day; participants in that trial had blood calcium measured and regular medical follow-up.