Received on September 29, 2008
Accepted on February 18, 2009

General Clinical Chemistry

Toward Standardization of Insulin Immunoassays

¹ Department of Pathology, Virginia Commonwealth University, Richmond, VA
² Laboratory for Analytical Chemistry, Faculty of Pharmaceutical Sciences, Gent University, Gent, Belgium
³ The Regional Endocrine Laboratory, University Hospital Birmingham, University of Birmingham, Birmingham, UK
⁴ Mercodia, Uppsala, Sweden
⁵ Nilsson Measurement Quality, Uppsala, Sweden
⁶ Department of Laboratory Medicine and Pathology, Medical School, University of Minnesota, Minneapolis, MN
⁷ Department of Pathology, Virginia Commonwealth University, Richmond, VA; Laboratory for Analytical Chemistry, Faculty of Pharmaceutical Sciences, Gent University, Gent, Belgium; The Regional Endocrine Laboratory, University Hospital Birmingham, University of Birmingham, Birmingham, UK; Mercodia, Uppsala, Sweden; Nilsson Measurement Quality, Uppsala, Sweden, and Department of Laboratory Medicine and Pathology, Medical School, University of Minnesota, Minneapolis, MN

^* To whom correspondence should be addressed. E-mail: gmiller{at}vcu.edu.

BACKGROUND: Measurement of circulating insulin may improve the classification and management of diabetes mellitus and assist in treating people with insulin resistance.

METHODS: A work group convened by the American Diabetes Association evaluated results for a panel of 39 single donor sera measured by 10 commercial insulin methods from 9 manufacturers against an isotope dilution–liquid chromatography/tandem mass spectrometry (IDMS) measurement procedure calibrated using purified recombinant insulin. We used a candidate primary (pure substance) reference material, pooled serum, and single donor sera to evaluate approaches to achieve improved agreement of results between the routine and reference measurement procedures.

RESULTS: Four of 10 methods had 95% of individual serum results within 32% of the IDMS concentrations. However, the bias vs IDMS was more than 15.5% for 7 of 10 methods in 36%–100% of individual samples. A purified recombinant insulin preparation used as a common calibrator did not improve harmonization of results among routine methods but was not used as instructed by all participants. Calibration using serum pools achieved bias <15.5% for nearly all results in the concentration range covered by the pools (>60 pmol/L). Calibration using a panel of individual sera was the most effective to improve harmonization of results over the full measuring range.

CONCLUSIONS: Agreement among methods can be improved by establishing traceability to the IDMS procedure using a panel of native sera. Pooled sera may be useful as trueness control materials. The usefulness of the pure insulin primary reference material [candidate reference material for insulin (cRMI)] requires clarification of protocols used by manufacturers.