Measurement of islet cell antibodies in the Type 1 Diabetes Genetics Consortium: efforts to harmonize procedures among the laboratories

Abstract

Background: and

Purpose: Three network laboratories measured antibodies to islet autoantigens. Antibodies to glutamic acid decarboxylase (GAD65 [GADA]) and the intracellular portion of protein tyrosine phosphatase (IA-2(ic) [IA-2A]) were measured by similar, but not identical, methods in samples from participants in the Type 1 Diabetes Genetics Consortium (T1DGC).

Methods: All laboratories used radiobinding assays to detect antibodies to in vitro transcribed and translated antigen, but with different local standards, calibrated against the World Health Organization (WHO) reference reagent. Using a common method to calculate WHO units/mL, we compared results reported on samples included in the Diabetes Autoantibody Standardization Program (DASP), and developed standard methods for reporting in WHO units/mL. We evaluated intra-assay and inter-assay coefficient of variation (CV) in blind duplicate samples and assay comparability in four DASP workshops.

Results: Values were linearly related in the three laboratories for both GADA and IA-2A, and intra-assay technical errors for values within the standard curve were below 13% for GADA and below 8.5% for IA-2A. Correlations in samples tested 1-2 years apart were >97%. Over the course of the study, internal CVs were 10-20% with one exception, and the laboratories concordantly called samples GADA or IA-2A positive or negative in 96.7% and 99.6% of duplicates within the standard curve. Despite acceptable CVs and general concordance in ranking samples, the laboratories differed markedly in absolute values for GADA and IA-2A reported in WHO units/mL in DASP over a large range of values.

Limitations: With three laboratories using different assay methods (including calibrators), consistent values among them could not be attained.

Conclusions: Modifications in the assays are needed to improve comparability of results expressed as WHO units/mL across laboratories. It will be essential to retain high intra- and inter-assay precision, sensitivity and specificity and to confirm the accuracy of harmonized methods.

Figure 1

GADA inter-assay comparisons of blind duplicates among the T1DGC autoantibody laboratories. Mean values of the original and repeat assays are plotted. Results above the highest standard have been extrapolated from the standard curve.

Figure 2

IA-2A inter-assay comparisons of blind duplicates among the T1DGC autoantibody laboratories. Mean values of the original and repeat assays are plotted. Results above the highest standard have been extrapolated from the standard curve.

Figure 3

Sensitivity and specificity for GADA (left) and IA-2A (right) in four DASP proficiency evaluations among the three T1DGC autoantibody laboratories. Results are shown for assays using the IA-2_ic clone with the exception of the North American DASP 2002 results that used the IA-2_bdc clone. A common set of 100 control sera were used for DASP 2002-2005, but 50 were substituted for the 2007 workshop.

Figure 4

Comparisons of (a) GADA (upper panel) and (b) IA-2A (lower panel) results for cases reported by the three T1DGC autoantibody laboratories in the DASP 2005 workshop. Samples are ordered according to the median antibody level reported by all laboratories participating in DASP 2005.