Can HbA1c combined with fasting plasma glucose help to assess priority for GCK-MODY vs HNF1A-MODY genetic testing?

Excerpt

GCK-MODY and HNF1A-MODY are the most common subtypes of Maturity Onset Diabetes of the Young (MODY) [1‐3]. HNF1A-MODY mutation carriers may respond to sulfonylureas, while GCK-MODY does not necessitate therapy; therefore, molecular diagnosis is instrumental to guide therapeutic decision. In many laboratories next-generation sequencing is not available yet, and genetic testing of common MODY genes is performed using Sanger DNA sequencing. Thus, establishing which gene has to be screened first is mainly based on clinician expertise. Though HNF1A mutation carriers are frequently diagnosed with overt diabetes, both GCK-MODY and HNF1A-MODY can present with impaired fasting glucose (IFG). In the latter case, oral glucose tolerance test (OGTT) may help to distinguish between the two types [4]; however, a less expensive and time-consuming test would be preferable. GCK-MODY is usually considered in pediatric patients negative to type 1 diabetes (T1D)-related autoantibodies and stable, mild fasting hyperglycaemia (100–150 mg/dl). In addition, an HbA1c ≥ 6% (42 mmol/mol) [5] or ≤ 7.6% (60 mmol/mol) [2] has been utilized as criterium to select candidates for GCK genetic testing. In this study, we wanted to evaluate the usefulness of HbA1c combined with fasting glucose (FPG) < 150 mg/dl to prioritize sequencing between the two genes. Ninety-nine cases with GCK-MODY and 45 with HNF1A-MODY confirmed by genetic testing were selected among patients attending Italian pediatric diabetes centers. The following data were collected: age at referral, sex, 2 FPG values at presentation (before therapeutic intervention) and HbA1c. Statistical analysis was run with SPSS version 20.0. GCK-MODY patients were younger (9.0 ± 4.1 vs 12.9 ± 2.8 years p < 0.001) and had lower HbA1c [6.3 ± 0.4% (45.5 ± 4.7 mmol/mol) vs 7.8 ± 2.3% (61.5 ± 24.7 mmol/mol), p < 0.001] than HNF1A-MODY. Of note, only seven GCK-MODY subjects (7/99, or 7%) had one or both fasting glucose values > 150 mg/dl and an HbA1c value between 6 and 7.1% (42–54 mmol/mol). The other 92 (92%) GCK-MODY patients and 23/45 (51%) HNF1A-MODY patients had two values < 150 mg/dl. Further analysis of these two subgroups showed that an HbA1c cut-off of 7.3% (56 mmol/mol) could identify 91 out of 92 GCK mutations carriers (ROC analysis) with high sensitivity (100%) and specificity (80.7%) (AUC 0.904). Of twenty-three patients with HNF1A gene mutations and 2 FPG values below 150 mg/dl, 17 (73.9%) had an HbA1c above 7.6% (60 mmol/mol), and would have been correctly prioritized according to criteria previously used by others [2], while 6 (26.1%) had HbA1c between > 7.3% (56 mmol/mol) and ≤ 7.6% (60 mmol/mol) and would have been inappropriately screened for GCK if selected by FPG criteria alone. When we considered age, HbA1c and gender in the regression analysis of patients with two FPG values below 150 mg/dl, these three variables predicted the risk of HNF1A-MODY (Cox and Snell R² = 0.211, Table 1). To date, selection of individuals with IFG (or mild diabetes), and negative to type 1 diabetes antibodies seems the best strategy to identify carriers of GCK mutations: utilizing this approach we established that GCK-MODY accounts for 4.7% out 3781 patients referred to 15 Italian pediatric diabetes clinics between 2007 and 2012 [1]. In that paper, we suggested to screen GCK gene first in patients with FPG below 150 mg/dl, and HNF1A gene in those with FPG above 150 mg/dl [1]. Here we propose to combine a cut-off of FPG ≤ 150 mg/dl and > 7.3% (56 mmol/mol) for HbA1c to screen HNF1A first, irrespective of other clinical information. In conclusion, the prediction model to prioritize MODY gene analysis by Sanger sequencing in the pediatric setting should take into consideration primarily fasting plasma glucose and HbA1c. …