Participants and intervention
A cross-sectional study was carried out in 71 residents of dementia care units of the nursing home Topaz Overduin in Katwijk (the Netherlands), using cholecalciferol capsules, once a week 5600 IE or cholecalciferol drops (50,000 IU/ml), 3 drops a week (7500 IU). The capsules contained cholecalciferol 100 IE/mg, cellulose microcrystalline PH102, magnesium stearate en lactose monohydrate (180). Drops were a watery mixture composed of cholecalciferol concentrate in oil, citric acid monohydrate, star anise oil, potassium sorbates, polysorbatum 80 (polyoxyethylene sorbitan monooleat), sugar syrup, and purified water.
The only exclusion criterion was the use of vitamin D for less than 3 months.
All blood samples were drawn on the same day.
The best parameter for vitamin D status, 25-hydroxyvitamin D [25(OH)D], was measured by a radioimmunoassay (25-OH-vitamin D RIA, Diasorin, Stillwater, MN, USA). The assay has 100% cross reactivity with 25(OH)D2 and 25(OH)D3. Total imprecision (Interassay coefficient of variation) is 9.4% at 22 nmol/l. SCAL Medical Diagnostics (Foundation central primary care laboratory) is certified by the Dutch Board for Accreditation and participates in external quality assessment schemes organized by the Foundation Quality Control Medical Laboratory Diagnostics.
Ethical approval
Ethical approval was not necessary under Dutch regulations since this study was a retrospective patient file study. Among all elderly care physicians in the participating nursing home, drawing blood samples, in order to check the effects of vitamin D supplementation on individual serum 25(OH)D levels, is considered to be a quality of care standard and therefore part of good medical practice. Therefore, the drawing of blood samples was for clinical purposes and complied with The Dutch Law of Agreement to Medical Treatment (WGBO), and not as research that has to comply with the Dutch law on Medical Research in Humans (WMO), for which ethical approval is required. Informed consent however is also necessary for all diagnostic and clinical work under the WGBO, so patients who did not want their blood samples taken, could refuse.
Potential factors that influence 25(OH)D concentration
From the patient files, data were collected on factors possibly influencing a rise in serum 25(OH)D: age, comorbidity, number and sort of medication use (anticonvulsant medication and corticosteroids were taken into account because these increase catabolism of 25(OH)D [13]), body mass index (BMI), sun exposure, Modification of Diet in Renal Disease (MDRD) as an estimate of the renal function, and Functional Ambulation Classification (FAC) scores as an assessment of the mobility, performed by physiotherapists and physicians. The FAC consists of five items. Scores range from 0–5 (0 = Nonfunctional ambulation; 1 = Ambulator-dependent for physical assistance level II; 2 = Ambulator-dependent for physical assistance level I; 3 = Ambulator-dependent for supervision; 4 = Ambulator- independent level surfaces only; and 5 = Ambulator independent) [16].
Comorbidity was expressed as the number of chronic diseases, taking into account 7 major conditions [17], i.e. chronic obstructive pulmonary disease, cardiac disease, peripheral arterial disease, diabetes mellitus, stroke, cancer and rheumatoid arthritis/osteoarthritis, and measured by review of the patients’ medical files. BMI was calculated as body weight in kg divided by height in m2 and subsequently categorized into three groups: underweight (BMI < 20 kg/m2), normal weight (BMI: 20-25kg/m2) and overweight (BMI > 25 kg/m2).
Adequate compliance was defined to exist when more than 80% of the vitamin D medication was ingested.
Statistical analysis
Main outcome measure was serum 25(OH)D level after supplementation.
We express continuous variables as means ± standard deviation and categorical variables as percentages. We used Student’s T for comparisons, as well as Chi-squared test.
A logistic regression analysis was used to assess the possible predictors of 25(OH)D insufficiency. P-values were considered significant at a p-value < 0.05. Odds ratios (OR) and 95% confidence intervals (95% CI) were calculated to estimate the strength of the association when the p-value was significant. All analyses were conducted with SPSS 21 software.