Cumulative Loneliness and Memory Aging Among Middle-Aged and Older Men and Women in the United States

As the subjective experience of social isolation, loneliness has been associated with poorcognitive health outcomes in later life, including increased risk of Alzheimer’s disease andrelated dementias (ADRD). At least three knowledge gaps exist in the prior literature that need tobe filled for loneliness to become a practical intervention target to reduce ADRD risk at thepopulation-level. First, due to the dynamic, time-varying nature of loneliness, existing studiesthat have measured loneliness at a single point in time or over short periods of time may besubject to reverse causation, whereby rather than being a risk factor, loneliness could be apreclinical syndrome of ADRD. Second, there exist conflicting findings regarding sex/genderspecific associations between loneliness and ADRD, potentially due to information bias thatresults from differential loneliness under-reporting between men and women. This possibilityremains understudied, potentially because there is no gold standard data with which to validateself-reported psychosocial measures in research interviews. Third, although loneliness istheorized as a chronic psychosocial stressor that may persist over time to induce systemicinflammation, it remains unclear whether systemic inflammation-related biological mechanismsmay mediate the association between cumulative loneliness and ADRD.Using data from middle-aged and older adults in the US Health and Retirement Study(HRS), this dissertation applied rigorous study designs and complex statistical methods toaddress the three knowledge gaps described above. Chapter 2 constructed a prospective cohortstudy and fitted mixed-effects linear regression models to examine the association betweenxiiicumulative loneliness from 1996-2004 and subsequent memory function and rate of decline from2004-2016. Chapter 3 applied a quantitative probabilistic bias analysis to investigate thedirection and magnitude of the potential information bias arising from differential lonelinessunder-reporting between men and women in the effect modification by sex/gender of theassociation between cumulative loneliness and rate of memory decline that was observed inChapter 2. Chapter 4 used causal mediation analysis under a counterfactual framework toexamine the mediating effect of latent systemic inflammation factor scores in the associationbetween cumulative loneliness and memory function.Results suggest that a longer duration of loneliness from 1996-2004 was associated withlower memory scores in 2004 and a faster rate of memory decline from 2004-2016. Theassociation was stronger among adults aged ≥65 than those aged <65 and was stronger amongwomen than men. The observed effect modification by gender was not impacted by anydifferential loneliness under-reporting between men and women, as estimated by the quantitativebias analysis. Despite the imprecise estimates, cumulative loneliness was associated with higherlevel systemic inflammation. However, systemic inflammation did not mediate the associationbetween cumulative loneliness and memory function.Findings from this dissertation indicate that cumulative loneliness in mid-to-later life maybe a salient risk factor for accelerated memory aging, especially among women aged ≥65 in theUnited States. Further research is warranted to investigate the potential biological mechanismslinking loneliness and memory aging.