(1990). Johnson et al. (1980) conducted a double-blind clinical trial of men (all over 65 years) and women (all over 60 years) treated with 50 µg (2,000 IU)/day of vitamin D for approximately 6 months. Assuming a normally distributed population, these data as presented would appear to suggest that the risk of hypercalcemia in a population exposed to intakes of 50 µg (2,000 IU)/day ranges from 5/1,000 to less than 1/10,000. In addition, Honkanen et al. (1990) reported that 45 µg (1,800 IU)/day of supplementary vitamin D administered to Finnish women aged 65 to 72 years for 3 months produced no ill effects.

Data Selection. Data from several studies in infants (Fomon et al., 1966; Jeans and Stearns, 1938; Stearns, 1968) were judged appropriate for use in deriving a UL for infants up to 1 year of age since the data document the duration and magnitude of intake, and as an aggregate, they define a dose-response relationship. Available data from animal studies were judged inappropriate due to their greater uncertainty.

Identification of a NOAEL (or LOAEL) and Critical Endpoint. Jeans and Stearns (1938) found retarded linear growth in 35 infants up to 1 year of age who received 45 to 112.5 µg (1,800 to 4,500 IU)/day of vitamin D as supplements (without regard to sunlight exposure, which was potentially considerable during the summer months) when compared with infants receiving supplemental doses of 8.5 µg (340 IU)/day or less for a minimum of 6 months. At 45 weeks of age, infants were found to have a linear growth rate 7 cm lower than the controls.

Fomon et al. (1966), in a similar study, explored the effects on linear growth in infants (n = 13) ingesting 34.5 to 54.3 µg (1,380 to 2,170 IU)/day of dietary vitamin D (mean = 44.4 µg or 1,775 IU/day) from fortified evaporated milk formulas as the only source of vitamin D compared with infants who were receiving 8.8 to 13.8 µg (350 to 550 IU)/day (n = 11) from another batch of formula. No effect was found in infants who were enrolled in the study during the first 9 days after birth up to 6 months of age. Given the small sample size used in this study, it was deemed appropriate to deviate from the model for the development of ULs (see Chapter 3 which