A detailed longitudinal examination of plasma lipoprotein subfraction concentrations and compositions in pregnancy was performed with the objective of discovering the pattern of change in lipoprotein subfractions. Plasma triglyceride, cholesterol, very low density lipoprotein (VLDL1), very low density lipoprotein (VLDL2), intermediate density lipoprotein (IDL), low density lipoprotein (LDL) and its subfractions (LDL-I, LDL-II, LDL-III), and high density lipoprotein-cholesterol (HDL cholesterol) were quantified in 10 normal pregnant women from 10 weeks of gestation and at 5 weekly intervals thereafter, until 35 weeks of gestation, together with circulating hepatic lipase (at 10 and 35 weeks) and serum estradiol concentration. Median concentrations of VLDL1 (19-109 mg/dL), VLDL2 (17-103 mg/dL) and IDL (26-124 mg/dL) increased in parallel (maximum increase around 5-fold) as plasma triglyceride increased with advancing gestation. This contrasts with observations in the normal non-pregnant female, where higher concentrations of plasma triglyceride are associated with preferentially higher VLDL1 concentrations. The rise in IDL was also remarkable as this does not normally accompany changes in plasma triglyceride. LDL mass increased by 70% (200-353 mg/dL) between 10 and 35 weeks, and in 6 of the 10 women studied, the LDL subfraction pattern was modified towards a smaller denser pattern in a manner suggestive of a "threshold" transition, with the proportion of LDL-III increasing at the expense of LDL-II, whereas in the other 4 women, LDL subfraction profile remained unchanged throughout pregnancy. Interestingly, this "threshold" transition, if it occurred, did so at varying gestational ages and triglyceride concentrations for different women. The likelihood of LDL subfraction change and the final concentration of small, dense. LDL-III were related to the 10-week triglyceride concentration (R2 = 36.7%, P = 0.063) and to the rate of change in triglyceride for a given increment in estrogen (R2 = 48.6%, P = 0.025). In addition, VLDL1 mass exceeded 100 mg/dL during pregnancy only in those individuals in whom LDL profile perturbation was evident (chi 2, P < 0.001). LDL profile change was evident at the lowest triglyceride concentrations in the 2 individuals with the highest increments in triglyceride corrected for estrogen. On the basis of these longitudinal observations, we conclude the following: 1) as plasma triglyceride increases in pregnancy, there are parallel rises in median concentrations of VLDL1, VLDL2 and IDL, around 5-fold; 2) as a result of this progressive increase in plasma triglyceride, in particular in VLDL1, the LDL profile is altered in some individuals towards smaller, dense particles; 3) in general, the higher the initial (booking) fasting plasma triglyceride concentration or the larger the rate of change in triglyceride for a given increment in estradiol, the greater the probability of change in LDL profile towards smaller denser species; 4) significantly, LDL subclass perturbation towards smaller denser species occurs not in a gradual and progressive manner but exhibits "threshold" behavior; and finally, 5) this threshold is achieved at differing gestational ages and triglyceride concentrations for different women.