Сахарный диабет (May 2018)

Differential diagnostic utilities of combined testing for islet cell antibody, glutamic acid decarboxylase antibody, and tyrosine phosphatase antibody

  • Alexei V. Timofeev,
  • Igor E. Koltunov,
  • Elena E. Petriaikina,
  • Irina G. Rybkina,
  • Lubov N. Samsonova,
  • Anatoly N. Tiulpakov,
  • Natalia A. Zubkova,
  • Irina G. Kolomina,
  • Evgenia A. Evsjukova,
  • Sergey S. Bukin,
  • Alexey C. Khrushchev

DOI
https://doi.org/10.14341/DM9364
Journal volume & issue
Vol. 21, no. 2
pp. 74 – 83

Abstract

Read online

Background. Beta-cell antibody tests are used for the differential diagnosis of diabetes mellitus. They permit to discriminate between the type 1 diabetes (T1D) and non-autoimmune diabetes types. To choose an appropriate test for ruling in or ruling out the T1D a physician needs to know how conclusive test results are. The most powerful estimate of test conclusiveness is its likelihood ratio (LHR). The higher LHR of a positive result (LHR+), the more posttest probability of T1D; the lower LHR of a negative result (LHR−), the less posttest probability of T1D. Aims. To compare conclusiveness of single and combined tests for antibodies to islet cells (ICA), glutamate decarboxylase (GADA), and tyrosine phosphatase IA-2 (IA-2A), and to evaluate posttest probabilities of T1D at various pretest probabilities. Methods. All antibodies were tested in parallel in 169 children and adolescents with a new-onset T1D, and in 169 persons without this disease. ICA, GADA, and IA-2A were determined by indirect immunofluorescence, radioimmune assay, and ELISA, respectively. LHR+ and LHR− were calculated with the MedCalc Statistical Software. Posttest T1D probabilities were calculated from Bayes theorem-based equation. Results. Among single tests, an ICA test had the greatest LHR+ and the smallest LHR−, and consequently was the most reliable either for ruling in or ruling out the T1D. Among test combinations, an ICA&GADA combination had the greatest LHR+ and was the most suitable for T1D confirmation. The triple combination ICA&GADA&IA-2A had the smallest LHR− and was the most suitable for T1D exclusion. Conclusions. In the differential diagnosis of diabetes, the most appropriate test for ruling in the T1D is the double combination ICA&GADA. With both antibodies positive, this combination provides the highest posttest T1D probabilities at any pretest probability. The most appropriate test for ruling out the T1D is the triple combination ICA&GADA&IA-2A. With all three antibodies negative, this combination provides the lowest posttest T1D probabilities.

Keywords