Distribution analysis of vitamin D highlights differences in population subgroups: preliminary observations from a pilot study in UK adults.

Abstract

There is no consensus between Authors on the definition of a replete or deficient vitamin D state. Our aim was to describe a suitable method that could be used to compare vitamin D data in subject groups with small or large numbers. Two hundred and forty indigenous asymptomatic, non-pregnant adult subjects recruited from a single-consultation outpatient attendance with normal biochemistry, represented a sample of our inner city district population. 25-hydroxyvitamin D (25,OHD3) levels were measured to illustrate the effects of season, sex and ethnic group on vitamin D levels and subjected to distribution analysis. This method quantifies as a percentage the distribution of 25,OHD3 concentrations (observed concentration, OC) in pooled group data. The data can be expressed as distribution frequency domains or cumulative frequency ogives (0-100%) or transformed into discrete linear probits, amenable to regression analysis. An estimate of the OC50 (mid-point) and upper (either OC75 or OC95) or lower (either OC25 or OC5) range or at any other frequency between subject groups can be compared. A marked difference in 25,OHD3 levels between Asian and non-Asian asymptomatic adult subjects was seen during both seasons. 25,OHD3 deficiency was defined as at or below the OC25 for the non-Asian group (for both sexes: winter < 13.36 ng/ml, summer <13.38 ng/ml). The majority of Asians of both sexes were 25,OHD3 deficient (winter 94%, summer 82%). The distribution analysis provides an easy technique to compare 25,OHD3 status of different subject groups, allowing the description of populations using either longitudinal or cross-sectional data. This method may offer a way of describing 25,OHD3 deficiency between observers worldwide.