BSEM Members' Focus
See here for letters / articles from BSEM members --> Margaret Moss
Margeret Moss (BSEM)
Comment on:
Castro M, King TS et al. Effect of Vitamin D3 on Asthma Treatment Failures in Adults With Symptomatic Asthma and Lower Vitamin D Levels, The VIDA Randomised Clinical Trial. JAMA online 18.5.14. See full text here
BSEM Member Margaret Moss MA UCTD DipION MBANT CBiol MSB:
Nutrients work as a team. To treat asthma I should use a range of nutrients relevant to the individual, perhaps magnesium, manganese, molybdenum, vitamin B2, vitamin B5, vitamin C, vitamin D, omega 3 fatty acids, and magnesium sulphate baths, together with a change of diet and lifestyle. That way I should expect success.
If you pick out difficult cases, give them one nutrient alone, make your trial brief, and start listing what goes wrong from the start, before the level of the nutrient has built up, then you should not expect statistically significant success. If in addition, you prescribe a corticosteroid drug, known to interfere with the activation of the nutrient, in this case vitamin D3, your hope of success must be further reduced.
I wonder why only 9 out of 14 pre-specified secondary outcomes were analysed. What caused this change of mind?
Out of 1523 assessed for eligibility 455 were excluded initially, mainly as 25-hydroxyvitamin D levels were within the reference range. The researchers entered 1068 into the run-in, and then excluded 660 more, mainly because these no longer met the inclusion criteria, presumably because their asthma was no longer bad enough. To remain in the trial, a participant had to have failed in treatment before the trial started and then in the run-in as well. So the researchers ended up with only 408 to randomise, presumably the worst cases. Maybe those excluded would have picked up more quickly. In any case, reducing the number of participants makes it harder to reach statistical significance.
The groups were not the same to start with:
There was a median proportion of 0 days with controlled asthma in the vitamin D group, versus a median of 7 in the placebo group, in the last 14 days of the run-in. 37% of the vitamin D group versus 31% of the placebo group had had emergency hospital or doctor visits in the previous year. 34 % of the vitamin D group versus 30% of the placebo group were on systemic steroids. 30% of the vitamin D group had missed work or school, or were unable to do housework, compared with 24% of the placebo group. 33% versus 29% were on oral steroids and 47% v 40% on inhaled steroids. 23% versus 18% were enrolled in the fall, so that their 28 weeks’ treatment was over the winter. However 4% versus 6% had been hospitalised, 24% versus 26% had been on a leukotriene receptor antagonist or 5-lipoxygenase inhibitor, 60% versus 64% had been on an inhaled steroid plus a long-acting bronchodilator.
Exacerbations were at a rate of 28 per person per year in the vitamin D group versus 45 in the placebo group (p value 0.06), adjusted hazard ratio 0.63. So a few more participants, or a slightly longer trial, and it would probably have been statistically significant. Treatment failure was 0.58 events per person per year in the vitamin D group versus 0.74 in the placebo group. 96% in the vitamin D group versus 91% in the placebo group achieved a 50% reduction in Ciclesonide. 89% versus 80% achieved a 75% reduction in Ciclesonide. The vitamin D group received an average of 111mcg per day compared with the placebo group’s 126mcg/day (p=0.02). Non-black participants demonstrated a significant reduction in the rate of 1st exacerbation. Black participants could be expected to need longer to bring up their blood levels of vitamin D. The rate of 1st exacerbation was 13% in the vitamin D group versus 19% in the placebo group. It was 11% in those whose 25-hydroxyvitamin D reached 30ng/mL, giving a hazard ratio of 0.6. Each 10 ng/ml increase in 25-hydroxyvitamin D was associated with a reduction in the overall rate of treatment failures (HR 0.9, p=0.04) and the rate of exacerbation (HR 0.8. p=0.02). So, if the researchers had continued the trial long enough to bring more people up to the acceptable range of 25-hydroxyvitamin D, then the results could be expected to have been better.
There was a long list of conflicts of interest, which may or may not have contributed to the design of the study, or the interpretation of the figures.
The conclusion was that the figures did not support the use of vitamin D3 in asthma. It should have been that the study was promising, but did not go on long enough to reach significance in all measures. Moreover, success might be achieved more rapidly in a population with milder disease.