Working and episodic memory trajectories over time following bilateral oophorectomy in early midlife women

Abstract

AbstractBackgroundEarly ovarian hormone deprivation due to bilateral salpingo‐oophorectomy (BSO; ovarian and fallopian tube removal) is a risk for late‐life AD, but estradiol replacement therapy (ERT) may be protective (Rocca et al., 2007). Working and episodic memory decrements have been observed without ERT soon after surgery (Sherwin et al., 1988) and on average 5 years post‐BSO in midlife compared to premenopausal controls (Gervais et al., 2020). Unknown is whether memory continues to decrease over time in early midlife, a critical period for understanding cognitive trajectories that presage AD. We investigated memory changes in early midlife over time post‐BSO, and whether those taking ERT had different trajectories than those without.MethodParticipants with BSO taking ERT (BSO+ERT: M age = 44.5y; n = 71) were compared to those not taking ERT (BSO: M age = 43.2y; n = 55), recruited from Toronto and Montreal, Canada, and Linköping, Sweden, and completed up to four yearly visits. Episodic memory was evaluated using Logical Memory Task (LM) percent forgetting and Rey Auditory Verbal Learning Test (RAVLT) total learning. Working memory was assessed using Spatial Working Memory Task (SPWM) trial 1 errors, Digit Ordering Task (DOT) errors, and Digit Span Backward Task (DSB) span length. Multilevel models were used to examine whether ERT influenced memory over time post‐BSO and disaggregate within‐ and between‐person effects. Estimated marginal means (Aiken and West, 1991) were calculated to delineate significant interactions.ResultThere were no significant changes over time post‐BSO on LM or RAVLT. The DSB showed significant within‐ and between‐person effects; the BSO group showed decreases in span length over time post‐BSO, while the BSO+ERT group showed increases and had significantly longer span lengths at seven years post‐BSO and beyond (Fig.1A&B). The SPWM Task showed a significant within‐person effect; the BSO group had stable errors over time post‐BSO, while the BSO+ERT group had working memory improvements and made fewer errors (Fig.1C).ConclusionFindings suggest that immediate verbal episodic memory declines we observed previously may stabilize. They also reveal that without ERT, most women have working memory declines post‐BSO. These data support ERT administration post‐BSO for working memory preservation. If ERT is contraindicated, cognitive training or lifestyle interventions may mitigate working memory declines.