The object of this work is to try to elucidate the role of genetic and environmental factors in the aetiology of diabetes by studying a series of identical twins.Concor&ince in identical (monozygotic) twins does not necessarily mean that a disease is genetic in origin. Twins usually live together in early life and thus share the same environment. Concordance could therefore be the result of genetic or environmental similarity. However, in later life most twins live apart and then concordance does suggest a genetic disease. Discordance, on the other hand, must indicate that a disease is due, at least in part, to nongenetic factors. We are therefore particularly interested in discordance in younger and concordance in older twins.
Previous Twin StudiesThere have been five large studies of diabetic twins [1-5] but all have defects; none has categorised the twins as insulin dependent diabetics (IDDs) or noninsulin dependent diabetics (NIDDs) and in 0nly two were the unaffected twins examined by glucose tolerance.Then Berg [1] reported 47 identical twin pairs of whom 35 were over 43 years old. All of these were concordant for diabetes on history or glucose tolerance testing, but of the 12 younger twin pairs only 6 were concordant. White [2] found 16 of 33 pairs to be concordant but did not categorise them by age or type of diabetes. Gottlieb and Root [3] who, like White, worked at the Joslin Clinic but, we assume, were reporting on different patients, found seven out of 10 pairs in whom diabetes was diagnosed over the age of 40 were concordant, compared to only two out of 20 younger twins. Harvald and Hauge [4] ascertained twins from the Danish twin register. Of 47 "maturity-onset" pairs 26 were concordant, of 36 younger onset pairs only 12 were concordant but unaffected twins were not tested. Pollin et al [5] studied 53 identical twin pairs among US ex-service men aged 43 to 53 and found only three pairs were concordant but again unaffected twins were not tested.All but one of these studies have also reported concordance rates in non-identical twins. Then Berg [1] found nine out of 50 pairs of non-identical twins to be concordant, White [2] two out of 63, Gottlieb and Root [3] two out of 70 and Harvald and Hauge [4] 22 out of 158. The overall figure for concordance in nonidentical twin pairs in these four studies is 35 out of 341 (10%).In spite of the limitations of these studies three broad conclusions can be drawn. 1) Identical twins always show a higher concordance rate than non-identical twins irrespective of their age at diagnosis.2) Younger onset pairs of identical twins are often discordant for diabetes.3) Older onset pairs, on the other hand, are usually concordant for diabetes.Our own previous results [6] confirm these general conclusions. Of 96 pairs of identical twins 59 index twins became diabetic before the age of 40, 31 were concordant and 28 discordant; of the 37 pairs in which the index twin was diagnosed after the age of 40 all but three were concordant. Similar proportions were found as the study has...