Twin study using mortality data: a new sampling method

Abstract

Background: To evaluate the success of a novel approach to twin studies using death discordant twin pairs in a disease of low prevalence.

Method: A population study based on all registered deaths at the Office of Population Censuses and Surveys of England and Wales, classified under the ICD code 335.2 (motor neuron disease [MND]) for the period 1979-1989 inclusive. From the above database of 10,872 people, individuals born after 31 December 1899 were traced in the Birth Indices for England & Wales to enable identification of possible twins. In all 131 twin pairs were found and the co-twin details were sent to the National Health Service Central Register (NHS-CR) to enable location of the relevant Family Health Services Authority and thence the co-twin's general practitioner (GP). A letter requesting access was sent to the family practitioners. If given, the co-twin can be approached and interviewed.

Results: The search produced: 54 living co-twins; 31 who died as adults; 29 infant deaths; 5 emigrated; 3 incorrectly diagnosed index twins; and 9 untraceable co-twins. Among the adult deaths two concordant pairs were identified. This has created the largest twin population sample worldwide for MND.

Conclusion: This new twin study method is clearly viable, and has produced a large unbiased sample compared to that possible using traditional methods. It relies heavily on the accuracy of death certificates and zygosity reporting by living co-twins, but is possibly the only way of collecting twins in rare conditions.