Vitamin D_{3} | ||||||
---|---|---|---|---|---|---|

Placebo | 1,000 IU/day | 2,000 IU/day | 3,000 IU/day | Linear trend in change in percent density^{d} | ||

Comparisons of percent density | (n = 96) | (n = 96) | (n = 99) | (n = 100) | ß (95% CI) | P |

Percent density at baseline (mean ± SD) | 40.0 ± 16.5 | 38.2 ± 14.9 | 37.3 ± 15.1 | 38.5 ± 16.6 | ||

Percent density at end of intervention (mean ± SD)^{a} | 34.2 ± 14.9 | 32.8 ± 14.2 | 31.5 ± 14.2 | 34.7 ± 15.7 | ||

Change in percent density^{b} (mean ± SE) | –5.7 ± 0.5 | –5.5 ± 0.5 | –5.9 ± 0.5 | –3.8 ± 0.5 | ||

Difference from placebo^{c} | 0.0 | 0.3 | –0.1 | 1.9 | 0.53 (0.07–0.99) | 0.02 |

(95% CI)^{c} | (–1.5–2.1) | (–1.9–1.7) | (0.1–3.7) | |||

P^{c} | 1.0 | 1.0 | 0.03 |

↵

^{a}The Pearson correlation coefficient between percent mammographic density at baseline and at the end of the one-year intervention in each of the four study arms ranges between 0.94 and 0.95.↵

^{b}Mean ± SE changes between percent mammographic breast density at the end of the one-year intervention and that at baseline are estimated from ANOVA and adjusted for sites.↵

^{c}Differences in the mean of change in percent mammographic breast density comparing study arms receiving vitamin D_{3}with placebo, 95% confidence intervals and*P*values, are estimated from ANOVA and adjusted for sites; 95% confidence intervals and*P*values are adjusted for multiple comparisons (Bonferroni correction).↵

^{d}ß (95% confidence interval) and*P*value for linear trend are estimated from regression models and adjusted for sites. Study arm is treated as a continuous variable, and ß represents mean difference in change in percent mammographic breast density for increments of 1,000 IU/day of vitamin D_{3}.