Impaired glucose tolerance in Mild Cognitive Impairment is associated with lower brain amyloid on PET

Background Among community‐dwelling older adults with mild cognitive impairment (MCI), there exists little knowledge regarding how metabolic function impacts brain amyloid pathology (measured using PET) and cognitive status. Method This study explored, in participants diagnosed with MCI enrolled in the Wake Forest Alzheimer’s Disease Research Center (ADRC) Clinical Core, markers of baseline metabolic risk (Oral Glucose Tolerance Testing of blood glucose 120‐minute post challenge; OGTT‐2h), amyloid deposition by global PiB PET SUVr and disease subtype (MCI with and without amnestic symptoms) at a subsequent visit, as well as genetic ( APOE ‐ε4 carrier status) and demographic variables (age, sex, education). Consensus diagnosis of MCI was adjudicated by a panel of experts close in time to PET. OGTT‐2h>140 mg/dL indicates impaired glucose tolerance (IGT). Result 59 MCI participants underwent OGTT at initial visit (39 normoglycemic, 20 IGT), and amyloid PET at a later visit (mean delay=362days). N=55 also had APOE genotype determined. PiB SUVR was not statistically different ( p =.15) between MCI‐NG (1.51) and MCI‐IGT (1.33). However, in models controlling for age, sex, years education, and time between OGTT and PET, glycemic status was significantly associated ( p =.007) with amyloid, such that MCI‐IGT had lower SUVR. APOE ‐ε4 status did not attenuate this effect (glycemic status p =.011). A similar significant association was present for continuous OGTT‐2h values before ( p =.006) and after ( p =.02) controlling for APOE ‐ε4 status, with higher OGTT‐2h associated with lower SUVR. Scatterplots of the negative association between OGTT‐2h and PiB SUVR are presented in Figure 1 (before [a] and after stratifying by APOE‐ ε4 status [b]). Associations between glycemic status and PiB SUVR were similar when restricting to participants with amnestic symptoms (i.e., MCI‐A or MCI‐MDA; n=52), before ( p =.008) and after controlling for APOE ‐ε4 status ( p =.012). In none of these analyses was an interaction term of APOE ‐ε4 status and the glycemic variable significant ( p ’s>.23). Conclusion Among community‐dwelling participants with MCI, elevated OGTT‐2h and IGT are associated with lower PiB SUVR. Participants with IGT, at the same level of cognition (MCI), have lower brain amyloid. In these MCI‐IGT participants, factors other than amyloid may be driving MCI status.