• Alzheimer's disease;
  • amyloid;
  • disease progression;
  • meta-analysis;
  • PET imaging

Our aim was to conduct a meta-analysis of longitudinal studies assessing the association between Pittsburgh compound B (PiB) retention and the progression of cognitive status in healthy elderly, in patients with mild cognitive impairment (MCI) and in patients with Alzheimer's disease (AD). PubMed, MEDLINE, Embase and the Cochrane Library up to April 2013 were searched for studies reporting PiB retention at baseline and conversion of clinical status at follow-up, with manual searches of bibliographies of key retrieved articles and relevant reviews. Two independent reviewers extracted data on individual numbers with PiB positive or negative status at baseline and corresponding numbers of patients with cognitive decline at follow-up (conversion from healthy elderly to MCI or AD, or from MCI to AD, or a Mini-Mental State Examination decline >3 for AD patients). Relative risks were pooled using fixed-effects or random-effects models as appropriate. Associations were tested in subgroups representing three different phases of AD. Publication bias was evaluated with funnel plots. Twelve cohort studies including 1275 participants were included with a follow-up period ranging from 1 to 3.8 years. The pooled adjusted relative risks were 3.75 (95% confidence interval 2.76–5.09; P for heterogeneity 0.16; fixed-effects model) for disease progression, 1.73 (0.63–4.75; P for heterogeneity 0.27; fixed-effects model) for AD patients (four studies), 4.03 (2.68–6.07; P for heterogeneity 0.49; fixed-effects model) for MCI patients (eight studies) and 3.67 (2.25–5.99; P for heterogeneity 0.26; fixed-effects model) for disease progression in healthy elderly (six studies). Baseline PiB positive status is associated with a significantly increased risk of cognitive progression in healthy elderly and MCI patients.