Previous studies have demonstrated an association between low weight at birth and risk of later development of non-insulin-dependent diabetes mellitus (NIDDM). It is not known whether this association is due to an impact of intrauterine malnutrition per se, or whether it is due to a coincidence between the putative "NIDDM susceptibility genotype" and a genetically determined low weight at birth. It is also unclear whether differences in gestational age, maternal height, birth order and/or sex could explain the association. Twins are born of the same mother and have similar gestational ages. Furthermore, monozygotic (MZ) twins have identical genotypes. Original midwife birth weight record determinations were traced in MZ and dizygotic (DZ) twins discordant for NIDDM. Birth weights were lower in the NIDDM twins (n = 2 x 14) compared with both their identical (MZ; n = 14) and non-identical (DZ; n = 14) non-diabetic co-twins, respectively (MZ: mean +/- SEM 2634 +/- 135 vs 2829 +/- 131 g, p < 0.02; DZ: 2509 +/- 135 vs 2854 +/- 168 g, p < 0.02). Using a similar approach in 39 MZ and DZ twin pairs discordant for impaired glucose tolerance (IGT), no significantly lower birth weights were detected in the IGT twins compared with their normal glucose tolerant co-twins. However, when a larger group of twins with different glucose tolerance were considered, birth weights were lower in the twins with abnormal glucose tolerance (NIDDM + IGT; n = 106; 2622 +/- 45 g) and IGT (n = 62: 2613 +/- 55 g) compared with twins with normal glucose tolerance (n = 112: 2800 +/- 51 g; p = 0.01 and p = 0.03, respectively). Furthermore, the twins with the lowest birth weights among the two co-twins had the highest plasma glucose concentrations 120 min after the 75-g oral glucose load (n = 86 pairs: 9.6 +/- 0.6 vs 8.0 +/- 0.4 mmol/l, p = 0.03). In conclusion, the association between low birth weight and NIDDM in twins is at least partly independent of genotype and may be due to intrauterine malnutrition. IGT was also associated with low birth weight in twins. However, the possibility cannot be excluded that the association between low birth weight and IGT could be due to a coincidence with a certain genotype causing both low birth weight and IGT in some subjects.