Compared with nonusers, women in our study who took calcium supplements had a modest but statistically significant increase in the risk for kidney stones (relative risk, 1.20 [95% CI, 1.02 to 1.41]). This increase in risk is nearly identical to our findings among men [1]; in men who had calcium supplement intake greater than 500 mg/d, the relative risk compared with nonusers was 1.23 (CI, 0.84 to 1.79). The point estimates in all three dosage categories were similar; in the highest intake group, the relative risk was 1.21 (CI, 0.96 to 1.52) compared with nonusers. Although this apparent 21% increase in risk is not statistically significant, it would be a mistake to characterize it as "no increase at all." The most common strength of calcium supplement is 500 mg of calcium carbonate, which contains only 200 mg of elemental calcium. Most multivitamins contain less than 100 mg of elemental calcium.

Dr. Heaney believes that we have overinterpreted our data because of the lack of a dose-response relation between supplemental calcium and the risk for incident kidney stones. However, such a relation is not necessary to be consistent with our conclusions. On the basis of the mechanism of the treatment of absorptive hyperoxaluria (discussed in our paper and described by Dr. Heaney), a dose-response relation may not be expected. For calcium to bind to oxalate in the gastrointestinal tract, the calcium must be ingested when the oxalate is consumed. If calcium is not taken at the same time as oxalate, then calcium is absorbed. This leads to an increase in urinary calcium excretion with no change in urinary oxalate, thereby resulting in an increase in urinary calcium oxalate supersaturation. It is well known that it is critical for patients with absorptive hyperoxaluria to consume the calcium supplements with food. As we stated in our article, most supplement users probably did not consume their calcium supplement with oxalate. Perhaps the lack of a dose-response relation is related to the fact that persons in the higher-dose groups took more than one tablet per day, which would increase the likelihood of having had calcium in the gastrointestinal tract when oxalate was consumed. We cannot exclude the possibility that the finding for supplemental calcium is the result of chance, but the likelihood is low (P = 0.03).

The point of the analysis of supplemental calcium was not to discourage use of supplements but to quantify the risk, which had not been available previously. The underlying risk to a woman who has never had a kidney stone and uses calcium supplements is still very low (1.2 cases/1000 women per year).