Selection biases were controlled by this being a nested study of cases and controls on a cohort of pregnant patients, taking the women who began prenatal control before week 12, both in cases as in controls, guaranteeing their having the same probability of being exposed to the protective measure (calcium during weeks 12 to 16) and without still manifesting the disease. The detection bias of the case or non-case was minimized with the of primigravidae women who received calcium during this period. Classification bias was controlled by detecting and discarding clinical histories of poor quality.