R D Farmer1, R B Newson, K MacRae, R A Lawrenson, F Tyrer. 1. Department of Public Health and Primary Care, Charing Cross and Westminster Medical School, (University of London), Chelsea and Westminster Hospital.
Abstract
STUDY OBJECTIVE: To investigate whether there has been an increase of venous thromboembolism (VTE) mortality in European countries, concurrent with the replacement of second generation by third generation combined oral contraceptives (COCs). Such an increase has been predicted, and reportedly detected, because published studies have detected an increased incidence of VTE associated with third generation rather than second generation COC use. DESIGN: Data were collected on population and annual VTE mortality in women 15-34 and 35-49 years old, and on second and third generation COC sales, from 1981 to 1994 in 13 European countries. Data from the seven most populous countries were analysed by linear regression of annual VTE mortality, in the 15-34 and 15-49 age groups, with respect to calculated total and third generation COC use rates, and the regression coefficients used to estimate mortality differences between second generation users and non-users and between third and second generation users, respectively. MAIN RESULTS: The estimated mortality differences in all seven countries had confidence intervals wide enough to contain both zero and the excess mortalities expected from the results of published studies. This was true both for the mortality difference between third and second generation COC users and for that between second generation users and COC non-users. CONCLUSIONS: Mortality differences of the size expected from the published studies cannot be measured using annual national VTE mortality and COC sales data alone, because of residual interannual variation in VTE mortality, and possibly confounding between rising third generation market share and total COC use.
STUDY OBJECTIVE: To investigate whether there has been an increase of venous thromboembolism (VTE) mortality in European countries, concurrent with the replacement of second generation by third generation combined oral contraceptives (COCs). Such an increase has been predicted, and reportedly detected, because published studies have detected an increased incidence of VTE associated with third generation rather than second generation COC use. DESIGN: Data were collected on population and annual VTE mortality in women 15-34 and 35-49 years old, and on second and third generation COC sales, from 1981 to 1994 in 13 European countries. Data from the seven most populous countries were analysed by linear regression of annual VTE mortality, in the 15-34 and 15-49 age groups, with respect to calculated total and third generation COC use rates, and the regression coefficients used to estimate mortality differences between second generation users and non-users and between third and second generation users, respectively. MAIN RESULTS: The estimated mortality differences in all seven countries had confidence intervals wide enough to contain both zero and the excess mortalities expected from the results of published studies. This was true both for the mortality difference between third and second generation COC users and for that between second generation users and COC non-users. CONCLUSIONS: Mortality differences of the size expected from the published studies cannot be measured using annual national VTE mortality and COC sales data alone, because of residual interannual variation in VTE mortality, and possibly confounding between rising third generation market share and total COC use.