Effects of manufacturing oral contraceptives on blood clotting. [Letter]

The three objections to our paper (30 June p 1761) raised by Drs T W Meade and S G Thompson in their letter (25 August p 493) on the effects of oral contraceptive manufacture on blood clotting can be refuted. These refer to the small numbers the possible effect of diabetes in one subject and the statistical analysis. A point which may not have been sufficiently explicit in our paper was that although the number of subjects who suffered high exposure was small the many replicate tests performed allowed valid statistical analysis and produced highly significant results. The clotting changes in this highly exposed group were marked and typical of the pattern which has now been established following long-term studies with oestrogen-progestogen combinations given by mouth as oral contraceptives. Such changes have not been found in the long-term assessment of controls or patients with uncomplicated diabetes. The typical "pill pattern" changes found in this group in factors VII and X fibrinogen and thrombelastography leave little doubt of the valid comparison between exposure during manufacture and oral ingestion of oestrogen-progestogen contraceptives. The only difference in the present study is the rise in platelet count which we believe was due to the postmenopausal bleeding from which the women workers suffered. Finally the criticism of the inappropriate use of the unpaired t test in the statistical analysis is groundless because firstly there is no obvious violation of the assumptions of the t test and secondly the robustness of the t statistic secures its use in comparing two groups. We have no doubt therefore about the significance of our findings. (Full-Text)