While APOE ɛ4 is the major genetic risk factor for Alzheimer’s disease (AD), amyloid dysmetabolism is an initial or early event predicting clinical disease and is an important focus for secondary intervention trials. To improve identification of cases with increased AD risk, we evaluated recruitment procedures using pathological CSF concentrations of Aβ42 (pAβ) and APOE ɛ4 as risk markers in a multi-center study in Norway. In total, 490 subjects aged 40–80 y were included after response to advertisements and media coverage or memory clinics referrals. Controls (n = 164) were classified as normal controls without first-degree relatives with dementia (NC), normal controls with first-degree relatives with dementia (NCFD), or controls scoring below norms on cognitive screening. Patients (n = 301) were classified as subjective cognitive decline or mild cognitive impairment. Subjects underwent a clinical and cognitive examination and MRI according to standardized protocols. Core biomarkers in CSF from 411 and APOE genotype from 445 subjects were obtained. Cases (both self-referrals (n = 180) and memory clinics referrals (n = 87)) had increased fractions of pAβ and APOE ɛ4 frequency compared to NC. Also, NCFD had higher APOE ɛ4 frequencies without increased fraction of pAβ compared to NC, and cases recruited from memory clinics had higher fractions of pAβ and APOE ɛ4 frequency than self-referred. This study shows that memory clinic referrals are pAβ enriched, whereas self-referred and NCFD cases more frequently are pAβ negative but at risk (APOE ɛ4 positive), suitable for primary intervention.