We read with great interest the recent article by Hosszufalusi et al. (1) reporting that latent autoimmune diabetes in adults (LADA) manifests clinical features similar to those of adult-onset type 1 diabetes with rapid progression, and that prevalences of predisposing alleles and haplotypes did not differ between these two forms of autoimmune diabetes. Patients with LADA have been found to be heterogeneous in their clinical attributes (2). Some patients develop insulin deficiency within a few years (LADA-type 1), while others show clinical and metabolic markers similar to patterns in antibody-negative type 2 diabetes (LADA-type 2). As the authors noted, selection bias may have been present among the patients with LADA in their study because they tested for antibodies against islet cell cytoplasma (ICA) only in patients clinically suspected to have LADA. Therefore, patients with LADA in their study were not representative of all LADA patients, but represented only a subgroup, namely LADA-type 1. As a result, patient characteristics and prevalences of predisposing alleles and haplotypes were similar to those seen in adult-onset type 1 diabetes. We previously reported that HLA-DRB1 alleles influenced the prognosis of Japanese patients with diabetes who were positive for antibodies to GAD (3). Patients with LADA who later developed insulin deficiency showed increased prevalence of one of the predisposing alleles. Notably, patients with LADA who did not develop insulin deficiency were more likely to have protective alleles and less likely to have predisposing alleles than patients with type 1 diabetes showing rapid progression.
In addition, the authors diagnosed type 1 diabetes when patients had typical diabetes symptoms, were ketosis-prone, and required prompt insulin treatment at the time of diagnosis. The median and upper interquartile fasting concentrations of C-peptide (0.46 and 1.05 nmol/l, respectively) in the subjects with type 1 diabetes were somewhat higher than we expected, given their tendency toward ketosis and immediate need for insulin treatment. One suspects that patients who required insulin within 6 months of diagnosis as opposed to the time of diagnosis may have been included in the type 1 diabetes category. We propose that those patients should be placed in another subgroup, given their clinical differences from typical type 1 diabetic patients.
Demonstrating that ICA detected at diagnosis disappeared in six LADA patients with a relatively long disease course, the authors speculated that the tendency of ICA to disappear with increasing disease duration was similar in LADA and type 1 diabetes. In fact, the meaning of this disappearance may differ between LADA and type 1 diabetes. ICA positivity has been reported to persist in LADA (4), in contrast to the disappearance during the course of type 1 diabetes. Low levels of antibodies to GAD in patients with LADA declined to undetectable levels in our study (5), which suggests the possibility of pseudopositivity of antibodies to GAD.
In conclusion, patients with LADA constitute a clinically and genetically heterogeneous group. A more precise classification of autoimmune diabetes in adults is needed to define differences between forms of autoimmune diabetes in adults.
