A recent study from Denmark reported that there was a “reverse J-shaped association” between serum vitamin D levels [25(OH)D] and all-cause mortality rate [Durup et al., 2012]. They took snapshots of serum 25(OH)D concentrations from 247,574 individuals, then followed them for a median time of about 3 years.
The graph of the results showed that the risk for all-cause mortality was 2.5 times greater for those with 25(OH)D concentrations of about 2 ng/ml, decreased in a nearly linear manned to 12 ng/ml, then slowly to a value of 1.0, the minimum, at 20 ng/ml, then rose gradually from 30 ng/ml to a value of 1.5 at 64 ng/ml. Thus, those with very low serum 25(OH)D concentrations had 2.5 times the mortality rate of those with serum 25(OH)D concentrations between 20 and 30 ng/ml.
This finding flies in the face of numerous studies reporting benefits of higher serum 25(OH)D concentrations for the diseases contributing the most to mortality rates in Europe and the United States. Diseases like many types of cancer, cardiovascular disease, diabetes mellitus, and infectious diseases. Thus, the study should be considered in detail and in comparison to other papers in the peer-reviewed journal literature.
Perhaps the most important shortcoming was that no information was obtained about health conditions at the time of enrollment or death. A possible explanation for the higher mortality rate for those with higher 25(OH)D concentrations at time of enrollment is that they had been diagnosed with osteoporosis or other conditions and were advised by their physician to take vitamin D supplements.
Since high vitamin D doses are generally unavailable over the counter in Denmark, prescriptions by physicians were very likely for those with high 25(OH)D concentrations in this study. There were 2,527 subjects with serum 25(OH)D concentrations higher than 140 nmol/l (56 ng/ml). Taking high vitamin D doses after a lifetime of low serum 25(OH)D concentrations may not correct conditions that take years to develop such as calcified arteries, cancer tumors, and osteoporosis.
Support for this hypothesis is given in two recent studies in the United States. In one for women, frailty status found a U-shaped relation between serum 25(OH)D concentration and frailty status [Ensrud et al., 2010], while the one for men found a monotonic decrease in frailty status with increasing serum 25(OH)D concentration [Ensrud et al., 2011]. Women in the United States are much more likely to be diagnosed with osteoporosis and advised to take vitamin D than are men.
Frailty was defined in these studies as:
There are many journal papers reporting health benefits associated with vitamin D. Most of the serum 25(OH)D-health outcome relations as discussed in a recent review [Grant, 2011]. Thus, the Durup et al.  study should be considered interesting but perhaps biased at higher serum 25(OH)D concentrations by subjects only recently taking high doses of vitamin D.
Meanwhile, keeping one’s serum 25(OH)D concentration in the 40-60 ng/ml or higher range should not pose any problems, especially in light of a recent study of thirty-five pastoral Maasai and twenty-five Hadzabe hunter-gatherers living in Tanzania where “the mean serum 25(OH)D concentrations of Maasai and Hadzabe were 119 (range 58-167) and 109 (range 71-171) nmol/l, respectively” (47.6 and 43.6 ng/ml).