Response to Letter Regarding Article, “Hormone Therapy and Venous Thromboembolism Among Postmenopausal Women: Impact of the Route of Estrogen Administration and Progestogens: The ESTHER Study”
We thank Drs Micheletti and Chevallier for their interest in our report.1 First, we believe that odds ratios (ORs) and 95% confidence intervals (CIs) estimated from logistic regressions provide adequate information about significance and the size and direction of the effect of norpregnane derivatives. Because elevation in venous thromboembolism (VTE) risk is substantial (OR: 4) and significantly different from 1 (with the 95% CI not crossing 1), our results suggest a thrombogenic effect of norpregnanes, and the probability value (P<0.006) indicates that the probability of the result being due to chance is very small.
Second, only the main effects of the route of estrogen administration and type of progestogens were estimated with a joint model (Table 2 of the original article1). Stratified analyses by route of estrogen administration and type of progestogens have also been performed. Among transdermal estrogen users, women received estrogen alone (10 cases and 35 controls; OR: 0.8, 95% CI: 0.4 to 1.8 after adjustment for obesity, family history of VTE, and varicose veins) or combined with either micronized progesterone (13 cases and 63 controls; OR: 0.6, 95% CI: 0.3 to 1.2), pregnane derivatives (16 cases and 51 controls; OR: 0.8, 95% CI: 0.4 to 1.6), or norpregnane derivatives (28 cases and 31 controls; OR: 3.1, 95% CI: 1.7 to 5.9). Among oral estrogen users, women received estrogen alone (4 cases and 5 controls) or combined with either micronized progesterone (6 cases and no controls), pregnane derivatives (23 cases and 28 controls), norpregnane derivatives (12 cases and 6 controls), or nortestosterone derivatives (12 cases and 7 controls). There was no significant difference in VTE risk between any of the progestogen subgroups among current users of oral estrogen (overall OR: 4.5, 95% CI: 2.6 to 7.5).
Third, to allow for adequate numbers of subjects within subgroups, stratified analysis by time of exposure used the median of the distribution (5 years) as a cutoff point. Unlike oral estrogens, there was no significant interaction between the time of exposure to either transdermal estrogens or norpregnane derivatives and VTE risk. Therefore, differences in exposure time to hormone therapy cannot explain our results.
Finally, although our results may be clinically relevant, we acknowledge that interpretation of data may have been biased by the inclusion of women with hyperestrogenic symptoms who were prescribed norpregnane derivatives. This prescription bias was emphasized in the Discussion section. Regarding the absence of thrombogenic mechanism underlying our results, Micheletti and Chevallier quote an inconclusive small trial2 that failed to also show the well-known activation of blood coagulation among women using oral estrogens. In addition, relevant hemostatic tests such as plasma-activated protein C sensitivity were not included as end points in this trial. Because relevant data are lacking, we are presently investigating the impact of norpregnanes on hemostasis among users of hormone therapy in the Study of NorpregnAnes on Coagulation (SNAC study).
Dr Scarabin has received research grants from Inserm, Fondation pour la Recherche Médicale, Fondation de France, Aventis, Besins International, Sanofi, and Servier Institute. The remaining authors report no conflicts.
Canonico M, Oger E, Plu-Bureau G, Conard J, Meyer G, Levesque H, Trillot N, Barrellier MT, Wahl D, Emmerich J, Scarabin PY; Estrogen and Thromboembolism Risk (ESTHER) Study Group. Hormone therapy and venous thromboembolism among postmenopausal women: impact of the route of estrogen administration and progestogens: the ESTHER study. Circulation. 2007; 115: 840–845.