Group-Based Trajectory Modeling of the Estradiol Trajectories During the Menopausal Transition Among Women in the Study of Women's Health Across the Nation (SWAN)

• Activity Number: 550
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Ping G. Tepper*+ and Bobby Jones and Sybil Crawford and Huiyong Zheng and Kristine Ruppert and Howard Kravitz and Yinjuan Lian and Maria M. Brooks
• Companies: University of Pittsburgh and University of Pittsburgh Medical Center and University of Massachusetts Medical School and University of Michigan and University of Pittsburgh and Rush University Medical Center and University of Pittsburgh and University of Pittsburgh
• Keywords: group-based trajectory modeling ; estradiol ; menopausal transition ; trajectory clustering

Abstract

Estradiol (E2) declines over the menopausal transition, but its trajectory variability is undefined. We studied annual serum E2 levels relative to the final menstrual period (FMP) in SWAN. Group-based trajectory modeling (GBTM) was used to model E2 trajectory variability with a polynomial relationship between E2 and time to FMP. Time-stable and time-varying factors were adjusted in the model. A total of 1585 women with up to 13 annual visits were analyzed. We found 4 distinct E2 trajectories: low and slow (SL, 24.5%), flat (23.8%), steep (ST, 11.0%), and rise/steep decline (RST, 40.7%). Compared to the RST group, women in the flat group were more likely to be African American (AA) (log-odds (ß)=1.05), overweight (ß=1.47) or obese (ß=3.84), but less likely to be Chinese (ß =-3.43), women in the SL group were more likely to be Chinese (ß =1.50), and women in the ST group were more likely to be AA (ß =0.95) but less likely to be overweight (ß =-1.66) (all p< .05). This is the first time that distinct E2 trajectories were identified and risk factors for these patterns documented. GBTM enabled the discovery of meaningful biological trajectory clusters.