An epidemiologic investigation of California twins

This dissertation investigates demographic trends in twinning using data from the California vital statistics birth and fetal death records (fetal deaths at more than 20 weeks gestation) for the years 1983-2003. The probability of twins increased by 10.1% from 1983-1992 and increased by 20.1% from 1993-2003, nearly doubling the previous increase. All maternal race-ethnicity groups showed increases in probability of twins with increasing maternal age. Parous women compared to nulliparous women had larger increases in the probability of twins. The probability of males in twins decreased from 1983-1992 and increased from 1993-2003; while in singletons the probability appeared stable during this time period.;Where twin zygosity has not been established through biological means, accurate estimation of zygosity is critical. Differences in concordance between monozygotic (MZ) and dizygotic (DZ) twins are a means of elucidating the contributions of genetic factors to the development of a given condition. The primary methods used for zygosity estimation are the maximum likelihood procedure and Weinberg's differential rule. Chapter three of this dissertation uses twin sex pairing data from twenty-one twin data sources where the data were collected under voluntary and active ascertainment methods. Maximum likelihood estimation of the probability of dizygotic twins ranges from 0.083 (95% approximate CI: 0.082, 0.085) to 0.750 (95% approximate CI: 0.749, 0.752) for voluntary ascertainment data sources and from 0.374 (95% approximate CI: 0.373, 0.375) to 0.987 (95% approximate CI: 0.959, 1.016) for active ascertainment data sources. In seventeen of the twenty-one twin data sources differences of 0.005 or less occur between maximum likelihood and Weinberg zygosity estimation methods. The Weinberg and maximum likelihood estimates are negligibly different in most applications however; the probability of a dizygotic twin may vary under twin data ascertainment methods.;The study of twins may be a powerful approach for identifying clues about cardiovascular defects and is explored in Chapter Four; for example, elevated risks of cardiac defects in MZ relative to DZ twins may indicate that genetics is more important than environment. Chapter four analyses revealed increased prevalences in twins compared to singletons in all 16 cardiovascular categories. The largest prevalence ratio was observed in atresia and stenosis of the pulmonary artery (cardiovascular category 12) (PR=4.01, 95% CI: 3.44, 4.68) and seven of the cardiovascular categories have at least double the prevalence in twins compared to singletons. Like-sex twins had increased prevalence of cardiovascular defects compared to unlike sex twins. Probabilities of concordance for aortic defects were higher among MZ than DZ twins, which is in agreement with findings from familial recurrence studies of cardiovascular defects.