A Case of Fulminant Type 1 Diabetes with Transient Production of Anti-Glutamic Acid Decarboxylase Antibody

Introduction

Fulminant type 1 diabetes is characterized by the abrupt onset of ketoacidosis and the low frequency of islet-specific autoantibodies [1]. The rapid elevation of blood glucose is typically revealed by a disproportionally low HbA1c level compared with the degree of hyperglycemia. However, it is rarely possible to trace the time course of the decline in insulin secretory capacity accompanied by the elevation of blood glucose in patients with fulminant type 1 diabetes.

Fulminant type 1 diabetes is not rare in Japan, and accounts for approximately 20% of adult-onset ketosis-prone type 1 diabetes, with no sex differences [2], although the incidence of type 1 diabetes is low in East Asian countries when compared with European countries and the US. More than 90% of cases of the disease were adult-onset [2].

The disease was originally described as an idiopathic nonautoimmune subtype of type 1 diabetes [1]. Actually, the prevalence of anti-glutamic acid decarboxylase (GAD) antibody was shown to be ∼5% in patients with fulminant type 1 diabetes [2,3], whereas anti-IA-2 and anti-ZnT8 antibodies are almost always negative. The absence of autoantibodies is one of the reference findings in the current diagnostic criteria of fulminant type 1 diabetes [4]. However, the low frequency of anti-GAD antibodies does not exclude the autoimmune mechanism in the pathogenesis of fulminant type 1 diabetes, because GAD-reactive Th1 cells are frequently detected in peripheral blood [5]. Furthermore, severe insulitis with the infiltration of dendritic cells, macrophages and CD8+ cells was reported in the pancreata of patients who died shortly after the onset of fulminant type 1 diabetes [6].

Thus, the measurement of anti-GAD antibody may provide valuable information not only for the diagnosis of patients but also for elucidating the pathogenic mechanism of the disease. However, the frequency and titer of anti-GAD antibodies may be underestimated, because anti-GAD antibodies may be transient and disappear early in some patients with fulminant type 1 diabetes. Herein, we report a case in which we observed the process of the extremely rapid onset of diabetes and the early decline in anti-GAD antibody titers during the inpatient stay.

Case Report

A 61-year-old man was brought to the emergency department of Chikugo City Hospital with acute confusion. He had suffered from significant thirst, polyuria, and fatigue for 2 days. His body height, weight, and body mass index were 163 cm, 61.2 kg, and 23.0 kg/m2, respectively. In addition, he had lost 4 kg of body weight during the previous week. A physical examination revealed disorientation, dry skin, rapid deep breathing (Kussmaul breathing), normal blood pressure (131/90 mmHg), and tachycardia (98 beats/min). Blood tests demonstrated marked hyperglycemia (1327 mg/dL), ketonemia (acetoacetate: 3023 µmol/L; 3-hydroxybutyrate: 14 012 µmol/L), and moderately elevated HbA1c (7.2%) and glycoalbumin (22.3%) (Table 1). Arterial blood gas analysis showed partially compensated metabolic acidosis (pH 7.192, PaCO2 18.6 mmHg, and HCO3 6.9 mmol/L) (Table 1). Blood tests also showed elevated levels of serum amylase, elastase, phospholipase A2, potassium, creatinine, creatine kinase, transaminases, and uric acid (Table 1). Serum C-peptide was not detectable (<0.03 ng/mL) (Table 1). He was immediately admitted to the hospital.

One month before admission, he had developed sudden sensorineural hearing loss and received betamethasone for 8 days at a starting dose of 4 mg/day at an otorhinolaryngology clinic. Six days before admission, he visited the city hospital with a sore throat, cough, and fever. Laboratory tests showed leukopenia (3200/µL) and a slightly increased level of C-reactive protein (0.84 mg/dL). His random blood glucose level was not high (117 mg/dL) at that time. He was prescribed acetaminophen and kampo maoto. Several parameters were later measured in a serum stocked in a freezer at the first visit, and the results showed 13.0% glycoalbumin, 5.07 ng/mL C-peptide, 24.4 µIU/ mL immunological insulin, and 332 U/L amylase. Despite the elevation of serum amylase, he did not have abdominal pain.

After admission, he was treated with intravenous fluid administration and continuous intravenous infusion of regular insulin. On the 3rd hospital day, the treatment was switched to multiple daily injections of insulin, which was continued thereafter. Serum C-peptide was undetectable, even after a glucagon injection. The glycoalbumin level increased to 25.2% on the 5th hospital day, whereas serum amylase gradually decreased and became normal on the 25th hospital day. An abdominal CT-scan performed on the 3rd hospital day showed hepatic steatosis with no obvious pancreatic abnormality. Although he exhibited clinical characteristics of fulminant type 1 diabetes, the serum anti-GAD antibody titer determined by enzyme-linked immunosorbent assay (ELISA) using a full-length recombinant protein (Cosmic, Tokyo, Japan) [7] (normal, <5.0 U/mL) was elevated (111 U/mL) at admission (Table 1). The antibody titer decreased to 22.8 U/mL after 2 weeks, and became negative 1 year later. He was negative for the anti-IA-2 antibody and thyroid-associated autoantibodies (Table 1). No significant elevations were observed in titers of anti-virus antibodies that were determined in this patient (Table 1). The genotypes of HLA class II were DRB1*08: 02-DQB1*03: 02 and DRB1*09: 01-DQB1*03: 03. He was treated with multiple daily insulin injections using intermittently scanned continuous glucose monitoring (Abbott Japan). One year after the onset of hyperglycemia, HbA1c and glycated albumin were 7.1% and 25.3%, respectively. Serum C-peptide remained undetectable.

Discussion

Fulminant type 1 diabetes is a subtype of type 1 diabetes in which the process of beta cell destruction is extremely rapid. The disease manifests as sudden onset of diabetic ketoacidosis, often accompanied by gastrointestinal symptoms and higher serum pancreatic enzyme concentrations. Flu-like symptoms frequently precede disease onset [2,4].

Although the abrupt onset of disease is a remarkable feature of fulminant type 1 diabetes, as well as the fact that the mean duration from the onset of hyperglycemic symptoms to the start of insulin therapy was reportedly 4.4 days [2], it is usually impossible to identify when insulin deficiency and hyperglycemia begin. However, in the present case, clinical data collected 6 days before the onset of ketoacidosis were available. Neither the elevation of plasma glucose or glycoalbumin, nor the decrease in C-peptide or immunoreactive insulin were observed at that time, thus proving that the critical damage to pancreatic beta cells occurred over a short period of time. Moreover, the interval between the start of the symptoms of hyperglycemia and the development of ketoacidosis was only 2 days.

Viral infection and the subsequent immune reaction have been proposed as the pathogenic mechanism of fulminant type 1 diabetes [8–10]. The common cold-like symptoms preceding the onset of ketoacidosis in this patient suggest that viral infection of pancreatic beta cells may be involved in the pathogenesis of the disease, although no significant elevation was observed in viral antibody titers (according to what we surveyed). It is unknown whether betamethasone administered for the treatment of sudden sensorineural hearing loss is associated with viral infections.

Notably, the anti-GAD antibody titer rapidly decreased from 111 U/mL, when the patient was brought to the emergency department, to 22.8 U/mL at 14 days later. The autoantibody test was negative 1 year later. It is well known that both the prevalence of anti-GAD antibodies and their titer in type 1 diabetes tend to decrease over time. In a nationwide survey of the HLA genotype in fulminant type 1 diabetes [11], anti-GAD antibodies were assessed within 1 week after the onset of diabetes. This time range is likely appropriate for the determination of anti-GAD antibody positivity. However, our observations suggest that, if the purpose is to quantify the antibody titer, the first measurement of anti-GAD antibody should be performed as early as possible. It should be noted that because the ELISA uses full-length recombinant GAD protein, there may be some discrepancy of results between the ELISA and previously used radioimmunoassay [7].

This patient had a susceptible haplotype DRB1*09: 01-DQB1*03: 03, which was shown to be significantly more common in anti-GAD antibody-positive patients with fulminant type 1 diabetes [11]. In addition, he suffered a serious hyperglycemic emergency only 2 days after the onset of symptoms associated with hyperglycemia. These features were in accordance with a previous report showing that the frequency of anti-GAD antibody was higher in patients who had marked hyperglycemia and severe ketoacidosis, and that the antibody titers were inversely correlated with the time period from the development of any prodromal symptoms to the diagnosis of fulminant type 1 diabetes [12]. It can be assumed that viral-induced massive beta cell destruction can result in the generation of anti-GAD antibodies on the genetic background of the susceptible HLA class II haplo-type. The transient elevation of the anti-GAD antibody titer was likely attributable to abrupt and complete beta cell destruction.

Conclusions

These observations prove the extremely rapid decline in insulin secretory capacity and underline the importance of the timing of anti-GAD antibody measurement in fulminant type 1 diabetes.