Background: We reported that preclinical AD as detected by the amyloid-imaging agent PIB is associated with
progression to symptomatic AD in cognitively normal persons. Here, we update the predictive power of PET PIB
imaging for symptomatic AD and cognitive impairment and explore whether PIB uptake and cerebrospinal fluid
(CSF) biomarker measures concomitantly predict time to cognitive impairment.
Methods: Participants (n=195; mean age 71.6 years) were CDR=0 at time of baseline PET PIB scan and were
assessed annually for a mean of 2.9 years. A subsample (n=114) had CSF within 1 year of PET/PIB imaging. Cox
proportional hazards models were used to examine time to cognitive impairment (CDR>0) in both samples and
to DAT in the larger sample as a function of the mean cortical binding potential (MCBP) for PIB at baseline scan.
Results: In the entire sample, 28 participants developed cognitive impairment over the follow-up period, 11 of
whom were diagnosed with DAT. Higher baseline MCBP (p=0.02) predicted a faster time to a DAT diagnosis
(p=0.02) and also predicted time to cognitive impairment generally (p=0.001). In the subsample with CSF data,
12 participants developed cognitive impairment. When considered alone, MCBP (p=0.0257), Aβ42 (p=0.0163), and
the ratios of tau/Aβ42 (p=0.0026) and ptau181/Aβ42 (p=0.0016) predicted time to cognitive impairment. When MCBP
was paired with each of the CSF variables in a series of Cox proportional hazards models, either MCBP or the CSF
variable, but not both, significantly predicted incident cognitive impairment.
Conclusions: Preclinical AD, as detected either by PET PIB or by CSF biomarkers, is a harbinger of both cognitive
impairment generally and symptomatic AD specifically. Incorporating both MCBP and CSF biomarkers in the
models does not add to the predictive power of using either biomarker alone.