Department of Public Health and Community Medicine/Primary Health Care, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden. Electronic address: margareta.leonardsson-hellgren@vgregion.se.
To explore and compare sensitivity and specificity for HbA1c =48mmol/mol as a predictor for type 2 diabetes mellitus (T2DM) in two populations with different ethnicity and to examine the predictive value of two levels of HbA1c (=42mmol/mol, =39mmol/mol) for prediabetes in these populations.
Four cohorts were examined with an oral glucose tolerance test. (1) The MEDIM Study (n=1991 individuals of Swedish and Iraqi ancestry); (2) The Skaraborg Project (n=1327 individuals of Swedish ancestry); (3) The 4-D study (n=424 individuals of Swedish, Iraqi and Turkish ancestry); (4) The Flemingsberg study (n=212 participants of Turkish ancestry).
HbA1c =48mmol/mol had a sensitivity for T2DM of 31% and 25% respectively in individuals of Middle-East and Swedish ancestry. The positive and negative predictive value was high in both populations (70.3, 96.4 and 96.2, 97.6 respectively). Using HbA1c =42mmol/mol and =39mmol/mol as a predictor for prediabetes gave a sensitivity of 17% and 36% in individuals of Middle-East and 15% and 34% in individuals of Swedish ancestry.
Even if HbA1c =48mmol/mol is a valuable diagnostic tool, it is a blunt and insensitive tool for screening and would exclude most people with T2DM, independent of ancestry and age. HbA1c is an inefficient way to detect individuals with prediabetes.