Updated 22 November 2009

20 November 2009. Part 1 in this series laid out the rationale for a new initiative called the Alzheimer’s Association QC Program for CSF Biomarkers, and Part 2 detailed the possible reasons why centers come up with such different results when testing how much Aβ42, tau, or phospho-tau float in a given person’s cerebrospinal fluid. Here, now, is a summary of the nuts and bolts of the QC program itself. The program combines a standardized protocol for sample collection that participants are encouraged to adhere to with periodic measurement of Swedish surplus CSF pool samples.

The protocol is freely available for download. The pooled CSF is available from Kaj Blennow and Henrik Zetterberg, at Sahlgrenska University Hospital in Molndal near Göteborg, Sweden.

This group has established pools yielding several thousand samples of human CSF with defined concentrations of these proteins. This is possible because, unlike in the U.S. where fear of lumbar puncture is quite common, in Sweden this procedure has long been part of routine neurologic practice. It is frequently performed in psychiatric, geriatric, and even some general medical practice, as well. Spinal taps have proven to be safe if performed correctly. The one reported side effect is headache; its frequency ranged from 0.9 to 4.1 percent in several published studies of consecutive taps (Blennow et al., 1993; Andreasen et al., 2001; Peskind et al., 2005). In Sweden, infection is routinely ruled out with spinal taps, and every medical student learns how to perform them. Swedish medical practice, therefore, generates a large amount of surplus CSF that used to be discarded, Blennow said, but now is captured for use in reference pools.

Three times a year, the Sahlgrenska group will ship three quality control samples to each participating site. Two samples are different, one is the same. The identical one will be stored on site and analyzed later to track longitudinal deviation over time. Each site analyzes the pool samples with the same assay they use for their research projects and clinical practice, and then sends the pool samples’ results back to the Sahlgrenska group. At each timepoint, three designated reference labs—Les Shaw’s at the University of Pennsylvania, Philadelphia; Rien Blankenstein’s at VU Medical University in Amsterdam, the Netherlands; and Piotr Lewczuk’s at University of Erlangen, Germany, receive and analyze an additional six copies of each sample. The Sahlgrenska site serves as a reference site, too. The results of these combined runs will serve to establish a reference range for each sample.

Blennow’s group will summarize the participating sites’ results in a report, where each lab can compare how it fares. “Then they know if they are within or outside acceptable limits, and can track down the problem if indeed there is one,” Zetterberg said. These problems are all fixable, he added. Sometimes it is as easy as replacing a dying lamp in the ELISA plate reader. The key is knowing when the readout begins to go astray, Zetterberg said.

The Alzheimer’s Association so far has funded this initiative for 3 years with the possibility for extension. Blennow hopes that the QC program will continue indefinitely. External quality control accompanies most established testing in other medical disciplines; in heart disease, for example, QC programs have been running for 30 years. “We will need external QC as long as we use CSF biomarkers clinically and in drug trials,” Blennow said.

Medical areas such as liver and kidney disease have had similar early problems with site variation and have brought them under control by analyzing the same aliquots of control sample several times a year and comparing results between labs in a broad-based effort. Likewise, tests for prolactin and thyroid-stimulating hormone have overcome variability problems and now test with around 10 percent variance among hospitals.

Last but not least, this kind of shared effort can lead to universal cutoffs, scientists agreed. At present, site and assay variations require that each Alzheimer disease center calculates its own cutoff. For blood glucose, for example, a worldwide cutoff exists to delineate normal from abnormal. “That was possible only because a QC program has been in place for many years, and it is why we intend to keep this QC program going,” Blennow said. Readers interested in participating in the program can contact the QC Programme Coordinator at neurokemi.su@vgregion.se or Henrik Zetterberg at henrik.zetterberg@clinchem.gu.se.—Gabrielle Strobel.

This concludes a three-part series. See also Parts 1 and 2. Read the entire series [.pdf].