unpublished observations

for Diabetes Four independent studies, three in theU.S. and one in Norway, have re-cently examined 254,481 persons to find a relationship between each individ-ual’s own birth weight and his or her risk of subsequently developing diabetes (1– 4). The first two studies used type 2 dia-betes as the sought end point; the second two focused on gestational diabetes. The occurrence of diabetes in a pregnant woman is in itself a marker for a height-ened risk of subsequently developing frank diabetes, usually type 2. Table 1 shows the relative risk (RR) of subse-quently developing diabetes at or below each designated birth weight; data are shown as RR (95 % CI). Birth weights that were originally published in Imperial units have been converted to SI units. While none of these observations can in themselves answer questions of causa-tion, nonetheless the RR ratios found in the four studies are remarkably uniform and highly predictive. None of these studies has identified a statistically significant increase in risk as-sociated with birth weight greater than expected, but together they establish be-yond all reasonable doubt that being small for gestational age at birth increases one’s risk of developing diabetes later in life. Moreover, they suggest that the smaller one begins life, the more likely one will be diabetic at its end