months of life. These earlier estimates of the chromium concentrations, however, cannot be used to accurately predict the additional needs of chromium during pregnancy.

Because of the lack of data to estimate the additional chromium requirement during pregnancy, the AI is determined by extrapolating up from adolescent girls and adult women, as described in Chapter 2. Carmichael and coworkers (1997) reported that the median weight gain of 7,002 women who had good pregnancy outcomes was 16 kg. In six studies of U.S. women, no consistent relationship between maternal age and weight gain was observed (IOM, 1990). Therefore, 16 kg is added to the reference weight for adolescent girls and adult women for extrapolation.

The AI for lactation is estimated on the basis of the chromium intake necessary to replace chromium secreted in human milk plus the AI for women. The amount that must be absorbed to replace the chromium secreted in milk is 0.252 μg/L × 0.78 L/day, or 200 ng/day. If absorption is estimated at 1 percent, 20 μg/day of chromium must be consumed beyond the usual intake to compensate for the milk losses. If absorption is only 0.5 percent, an additional 40 μg/day would be required. In the one study available on dietary intakes of lactating women, chromium intake was 41 μg/day (Anderson et al., 1993a).

Women do not appear to reduce urinary chromium excretion during lactation to compensate for increased needs (Mohamedshah et al., 1998). To calculate an AI for chromium during lactation, it is assumed that 1 percent of chromium is absorbed and 0.2 μg/day is secreted in human milk. Therefore 20 μg is added to the AI for adolescent girls and adult women, and the AI is rounded.