were judged inappropriate for use in deriving a tolerable upper intake level (UL) since the study subjects had hypoparathyroidism, which possibly increased their susceptibility to vitamin D toxicity. Irnell (1969) also reported a case of nephrocalcinosis and renal insufficiency in a thyroidectomized patient taking 1,125 µg (45,000 IU)/day of vitamin D for 6 years.

Animal data in monkeys (Peng and Taylor, 1980; Peng et al., 1978), rabbits (Lehner et al., 1967), and pigs (Kummerow et al., 1976) suggest that calcification may also occur in nonrenal tissue. Human studies of cardiovascular effects are largely negative or equivocal. Although Linden (1974) observed that myocardial infarct patients in Tromso, Norway, were more likely to consume vitamin D in excess of 30 µg (1,200 IU)/day than were matched controls, two subsequent studies (Schmidt-Gayk et al., 1977; Vik et al., 1979) failed to confirm these results.

Although the data for nephrocalcinosis and arteriosclerosis are insufficient for determination of a UL, they point to the necessity for conservatism. There is a large uncertainty about progressive health effects, particularly on cardiovascular tissue and the kidney, with regular ingestion of even moderately high amounts of vitamin D over several decades.

Data Selection. The most appropriate data available for the derivation of a UL for adults are provided by several studies evaluating the effect of vitamin D intake on serum calcium in humans (Honkanen et al., 1990; Johnson et al., 1980; Narang et al., 1984). The available animal data were not used to derive a UL for adults because the data were judged to have greater associated uncertainty than the human data.

Identification of a NOAEL (or LOAEL) and a Critical Endpoint. Narang et al. (1984) studied serum calcium levels in humans, with and without tuberculosis, where diet was supplemented with daily vitamin D doses of 10, 20, 30, 60, and 95 µg (400, 800, 1,200, 2,400, and 3,800 IU) for 3 months. Thirty healthy males and females rang-