Insulin secretion in patients with latent autoimmune diabetes (LADA): half way between type 1 and type 2 diabetes: action LADA 9

In this cross-sectional study, we enrolled 32 subjects with LADA, 33 subjects with type 1 and 30 subjects with type 2 diabetes mellitus. A total of 11 study subjects, 4 with LADA and 7 with type 1 diabetes, were recruited in Hospital de Sant Pau. All other subjects were included at Hospital Arnau de Vilanova. All study subjects were 30 – 70 years old at the time of diagnosis. Type 1 and type 2 diabetes mellitus were diagnosed according to the standard criteria as described previously[8], and LADA was defined as patients aged 30 – 70 years at the diagnosis of diabetes who did not require insulin for at least 6 months after diagnosis, with glutamic acid decarboxylase (GAD) autoantibody (GADAb) or tyrosine phosphatase autoantibody (IA-2Ab) positivity. The study was initially part of the Action LADA project that was extended locally for an additional period of 3 years[10]. The study protocol was designed to recruit 10 subjects in the following disease duration categories for each group: 6 – 18 months, 19 months – 5 years, and 5 – 10 years from diagnosis. This number of subjects was established based on previous studies that yielded clear differences when this sample size was used to compare LADA and type 2 diabetes[1, 2]. Thus, at least 30 patients of each diabetes type were included. None of the subjects had impaired kidney function, liver disease or any other condition known to interfere with the C-peptide response. Patients were included consecutively after they signed a written informed consent document. The study protocol was approved by the local Ethics Committee of the two centers where patients were recruited, i.e. Hospital de Sant Pau (Barcelona) and Hospital Arnau de Vilanova (Lleida).

For each subject, age, sex, weight, height, body mass index, and disease duration were obtained as recently described[8]. The time to start of insulin was calculated as the time between the date of diagnosis and the date of the first insulin treatment. In addition, glucose and HbA1c levels were tested using standard laboratory assays. A detailed description of the methods was recently published[8].

Antibodies to islet cell antigens (GADA65 and IA-2) were measured with two commercially available ELISA kits: GAD65 Autoantibody ELISA and IA-2 Autoantibody ELISA (both from DRG Diagnostic, Marburg, Germany), as previously described[11, 12]. Optimal cut-off values for positivity were set at 5 and 15 U/mL for GADAb and IA-2Ab, respectively. Both assays showed good performance when tested in the Diabetes Antibody Standardization Program Workshops. In the Diabetes Antibody Standardization Program Workshop 2007, sensitivities and specificities were 94% and 97% for GADAb, respectively, and 66% and 95% for IA-2Ab; in the Diabetes Antibody Standardization Program Workshop 2009, the figures were 82% and 95% for GADAb, and 60% and 100% for IA-2Ab, respectively.

The test was performed following the same procedures as those described in the European C-Peptide Trial in which our group participated[13]. Briefly, after an overnight fasting period, all study subjects were scheduled to start the test before 10 a.m. the next morning. Subjects were asked to withhold any short-acting insulin for at least 6 hours before the test. The standard meal consisted of 6 ml/kg of a Boost HP drink (Nestle Nutrition, Vevey, Switzerland); the subjects were given up to 360 ml for up to 5 minutes. C-peptide samples were collected at times -5 and 0 minutes before (the mean of these two values was taken as the fasting concentration), and 10, 20, 30, 60, 90 and 120 minutes after the final intake of the mixed-meal.

C-peptide was measured by a chemiluminescence assay (Immulite 2000, Siemens Medical Solutions Diagnostics, LA, USA), with reference values of 0.30 – 2.4 nmol/l (intra and interassay coefficients of variation were less than 5%). The lower detection limit was 0.03 nmol/l. Any value below the lower detection limit was arbitrarily assigned a concentration of 0.029 nmol/l.

Continuous variables are expressed as the mean (± SD) or the median (interquartile range ([IQR]) where appropriate. Categorical variables are described using the absolute and relative frequencies. Variables with a skewed distribution were log-transformed before analysis. Comparisons between groups were performed using the non-parametric Mann–Whitney U test for numerical variables and the χ² test for percentages. The area under the curve of C-peptide (AUCCP) was computed from all timed collections using the trapezoidal rule[9]. The analyses of the AUCCP and peak C-peptide were adjusted for sex, body mass index, age, duration of diabetes, fasting glucose, HbA1c and basal C-peptide. The results were considered to be statistically significant if the two-sided p-value was less than 0.05. All analyses were performed with SPSS for Windows version 15.0 (Chicago, USA).