Not all menopause is created equal: impact of surgical menopause on heart failure risk in women

Introduction A longer reproductive lifespan has been suggested to be protective against heart failure (HF), likely reflecting the cardioprotective influence of endogenous estrogen over time. However, the impact of gynecologic surgery and the marked abrupt disruption in endogenous estrogen exposure on HF risk is underexplored. Purpose To determine if surgical menopause, oophorectomy, hysterectomy, and age at procedure is associated with HF risk. Methods Using NHLBI’s BIOLINCC resource, data from the Atherosclerosis Risk in Communities (ARIC), Cardiovascular Health Study (CHS), Framingham Heart Study (FHS), Jackson Heart Study (JHS), and Multiethnic Study of Atherosclerosis (MESA) were aggregated and harmonized to create a cohort of 25,518 [ES1] women. Primary outcomes were HF incidence as adjudicated by each study and age at onset of HF. Using Google Big Query and R, we evaluated associations between primary outcomes and variables describing surgical menopause, oophorectomy, and hysterectomy. Results Shown in Figure 1, exposure to surgical menopause, oophorectomy, and hysterectomy were all associated with a higher odds of incident HF (OR 1.47 [1.37–1.56], p<0.001; OR 1.41 [1.32–1.51], p<0.001; OR 1.45 [1.36–1.55], p<0.001, respectively). Undergoing both hysterectomy and oophorectomy was associated with a 54% higher odds of incident HF (OR 1.54 [1.42–1.66], p<0.001). The association between surgery and incident HF varied across age (p-interaction<0.001). The relationship also differed by surgery type (oophorectomy vs hysterectomy p-interaction[ES2] <0.001). However, there was no difference in incident HF risk based on number of ovaries removed (OR 0.99 [0.86–1.13], p=0.84; 1 vs. 2 removed, p-interaction= 0.86). Shown in Figure 2, exposure to surgical menopause (-1.28±0.35 years, p<0.001), oophorectomy (-1.49±0.38 years, p<0.001), and hysterectomy (-1.47±0.35 years, p<0.001) were all associated with an earlier age at HF onset. Later age at oophorectomy years was associated with later HF onset (p<0.001), whereas age at hysterectomy per 5 years was not associated with age at HF onset (p=0.245). The association between surgery and age of HF onset varied by age (p-interaction=0.033) and surgery type (p-interaction=0.033), but not by the number of ovaries removed (p-interaction=0.238). Conclusions Risk of incident HF and age of diagnosis following gynecologic surgery appears to be driven more by surgical type and timing than by the number of ovaries removed, suggesting that any disruption of ovarian function may be sufficient to influence HF outcomes. Future research into the potential role of hormone replacement therapy in modifying this risk.