Dr. Daniel Riemer has very kindly supplied me with the full text of the paper by Rejnmark et al I referenced in the last post.
Thank you Dr. Riemer!
The study is of women who were studied to see the effects of hormone replacement therapy (estrogen) on a variety of paramers, including temporal changes in levels of Vitamin D binding protein and on levels of 25 (OH) D3, which I will continue to abbreviate as 25-D.
It turns out there were no significant effects of estrogen replacement on D status. Of interest to us, they measured 25-D at 0, 1, 2 and 5 years. The assay used at time zero was different from the one used subsequently, so analysis of 25-D variability was only assessed between times 1 and 2 and 2 and 5. The methods section relates that they drew these levels at the same time of the year for each subject to attempt to minimize the effect of seasonal variation. That's good. They also say that about 40% of the subjects were taking D supplements. No comment about how static the levels of supplementation were.
So the shortest interval was a one year interval from time 1yr to time 2 yr. The intra-individual variability here averaged 13% with a range of 6% to 26%.
The three-year interval from time 2 yr to time 5yr had average variabilty of 16% with range of 7% to 29%.
The maximum interval of 4 years from time 1yr to time 5yr was 19% with a range of 12% to 27%.
What can we conclude and what is the relevance to those of us testing ourselves while supplementing?
I think we can say that up to 20% variation in two values obtained a year apart in someone who is presumed to be at equilibrium in their dosing regime could be basically "physiologic drift" - a real change in 25-D levels but not unexpected and not pathologic. Yet another reason to etiher test frequently (like twice a year) or to bias your level to the high side in case you drift lower a year later.
For those with variations over shorter periods of time, like one month, it seems reasonable to expect smaller amounts of variability due to this "physiologic drift" and if large differences are seen and you think you have been at equilibrium for a while (not changed your dosing regime for many months) one should probably look at lab error as the most likely cause, as miscalibration seems to be give the largest "delta" among the causes for variation at presumed equilibrium we have seen.
Who really knows, though? It is still possible that there could be high physiologic variability over short periods like one month. It simply has not been studied enough to say for sure to the best of my knowledge.