Background Clusterin (ApoJ) is an amyloid-associated protein and plays an important role in Alzheimer’s disease (AD) pathology. Recent genome-wide association studies have indicated that certain genetic variants increase the risk of developing AD. To determine if the expression of clusterin is different in AD patients, both systemically and locally in the brain, differs between (subgroups of) AD patients and non-AD cases, an assay available that detects clusterin in both plasma and cerebrospinal fluid (CSF) with equal sensitivity would be helpful. Methods We compared four different commercially available antibodies in their ability to detect recombinant clusterin and immune-purified human clusterin. Specificity was tested on western blot and in ELISA systems, and selection was based on the ability to detect clusterin in CSF and plasma. A sandwich ELISA was developed and validated with monoclonal antibody G7 as capture, and rabbit polyclonal (Alexis) antibodies for detection. Results Our ELISA measured clusterin concentrations in plasma and CSF with dynamic ranges of 2–70 mg/L and 0.5–40 mg/L, respectively. The assays showed 99.8% recovery in CSF and 97% recovery in plasma. Intra-assay coefficient of variation was 1.4% and inter-assay 8.8%. The assay shows no cross-reactivity with related apolipoproteins. Clusterin quantification is dependent on the type of storage for plasma samples. A single freeze/thaw cycle caused fluctuations of clusterin concentrations in plasma, while clusterin in CSF is stable for up to five cycles. Conclusions We have successfully developed a clusterin ELISA that reliably measures CSF and plasma clusterin concentrations. In a pilot study, all samples gave results that were well within the dynamic range of the assay, with low variations. Freshly stored plasma samples are crucial for accurate clusterin quantification.