Fulminant Type 1 Diabetes Mellitus: A Study of Nine Cases

Abstract

Objective:

The study aims to explore the prevalence and clinical features of fulminant type 1 diabetes mellitus (F1DM).

Subjects and Methods:

We analyzed 11,202 patients who were diagnosed having diabetes and admitted to the Department of Endocrinology, Nanjin First Hospital, Nanjin, China, from September 2005 to April 2011. Among 198 patients classified having type 1 diabetes mellitus (T1DM), nine patients were diagnosed having F1DM. Clinical features of F1DM were analyzed and compared with typical T1DM.

Results:

The prevalence of F1DM was 4.5% (nine of 198 T1DM patients) in this study. F1DM patients had significantly high levels of random blood glucose (40±9 mmol/L). However, glycated hemoglobin (7.1±1.0%) and C-reactive peptide (0.21±0.15 μg/L) at 2 h postprandial were significantly lower in F1DM patients compared with typical T1DM patients. Six of them had flu-like symptoms. Five of them showed gastrointestinal symptoms. Three patients experienced disturbance of consciousness. However, none of these symptoms was found in patients with typical T1DM. A continuous intravenous insulin infusion therapy using a portable pump was given during the acute phase to control hyperglycemia and ketoacidosis. All F1DM patients survived after treatment.

Conclusion:

In this single hospital-based study, nine Chinese patients met the criteria for diagnosis of F1DM. Although F1DM may be fatal, prompt treatment is required and useful to rescue these patients.

Introduction

F ulminant type 1 diabetes mellitus (F1DM) is a subtype of type 1 diabetes mellitus (T1DM) that was found and defined by Imagawa et al.¹ in 2000. It is characterized by markedly rapid progression of hyperglycemia and ketoacidosis, which is due to complete destruction of pancreatic β-cells.¹ Generally, a rapid onset of diabetic ketoacidosis will occur within a short period after hyperglycemia with normal or near-normal glycated hemoglobin (HbA1c) level at onset and complete β-cell destruction. The duration of the disease from a normal β-cell mass and normoglycemia to almost total destruction of β-cells and ketoacidosis will be only a few days and rarely exceeds 1 week. Without proper treatment, ketoacidosis is fatal. Studies have suggested that pathogenesis of this disease involves both genetic background and viral infection.^1,2 Accumulated evidence has shown that human leukocyte antigen (HLA) class II is associated with this disease.³

A nationwide survey in Japan found that fulminant diabetes mellitus accounted for 15–20% of Japanese T1DM with ketosis or ketoacidosis at onset.⁴ In Korea, the prevalence of F1DM was 7.1% among all patients newly diagnosed with T1DM and 30.4% among patients with adult-onset diabetes.⁵

The prevalence of F1DM in China has also been reported. Zheng et al.⁶ found that the prevalence of F1DM was 9.1% (eight patients) among all newly diagnosed T1DM patients in a single hospital-based study. Here we reported nine cases of F1DM from the Nanjing First Hospital, Nanjing Medical University, Nanjing, China, along with a comparison of typical T1DM.

Subjects and Methods

Patients

Patients diagnosed with diabetes mellitus (n=11,202) were admitted to the Department of Endocrinology, Nanjing First Hospital, from September 2005 to April 2011. Of these patients, 10,017 were classified as having type 2 diabetes mellitus, and 198 were classified as having T1DM; 987 cases were found to be unclassified diabetes mellitus. Ketoacidosis was seen at onset for 33 newly diagnosed T1DM patients. Nine of the 198 T1DM patient (4.8%) were diagnosed having F1DM, based on the following criteria: (1) ketoacidosis within 7 days after hyperglycemia; (2) random blood glucose ≥16 mmol/L with HbA1c<8.5%; (3) urinary C-reactive peptide (CRP)<10 μg/day or serum CRP during fasting <0.3 μg/L, with 2-h postprandial CRP<0.5 μg/L.² The exclusion criteria were acute myocardial infarction, acute pancreatitis, and liver or kidney diseases. The characteristics and clinical symptom were investigated and compared with those of the remaining 24 typical T1DM patients.

Laboratory measurements

HbA1c was measured by the Bio-Rad (Hercules, CA) Variant™ II automated HbA1c analyzer, and CRP was measured with the Roche Diagnostics (Indianapolis, IN) E170 analyzer. Ketone bodies were measured using a semiquantification method and expressed as 0=normal, 1=mild, 2=moderate, and 3=severe. Blood glucose, glutamate-pyruvate transaminase (GPT), glutamic-oxaloacetic transaminase (GOT), γ-glutamyl transpeptidase (GGT), creatine kinase, creatine, alkaline phosphatase (ALP), blood urea nitrogen (BUN), blood potassium, sodium, chloride, cholesterol, triglyceride, high-density lipoprotein-cholesterol, albumin, and leukocyte count were determined with a Kodak blood biochemistry analysis machine (Johnson-Johnson, New Brunswick, NJ) and Mindray (Shen Zhen, China) blood analysis machine. Anti-insulin antibody was determined with a radioimmunoassay kit purchased from Beijing North Institute of Biological Technology (Beijing, China). Anti-glutamate decarboxylase antibody was determined with an enzyme-linked immunosorbent assay kit purchased from Beijing LiMi Biotechnology Co. (Beijing).

Statistical analysis

All data are expressed as mean±SD values. All statistical analyses were performed using SPSS version 15 (SPSS, Inc., Chicago, IL). Paired t test or χ ² test was used to determine the significance of within-group differences. All tests were performed with a statistical significance level of 5%.

Results

General characteristics of patients

There were significant differences in age of onset, body weight index, and course of disease (Table 1). The onset age of F1DM patient was much older (46±7 years) compared with typical T1DM patients (19±6 years) (t=10.79, P<0.01). The body mass index of F1DM patients was much higher (22.3±1.9 kg/m²) compared with typical T1DM patients (16.8±2.7 kg/m²) (t=5.49, P<0.01). Most of the F1DM patients did not have a family history (eight of nine, 88.9%), whereas 70.8% (14 of 24) of the typical T1DM patients also did not have a family history. The mean hospital stay was 24.9±5.2 days for F1DM patients versus 12.6±6.3 days for typical T1DM patients (t=9.33, P<0.05).

Table 1.

Patient Demographics

	F1DM	Typical T1DM	t value	P value
Sex distribution (M/F)	5/4	6/18
Age at onset (years)	46±7	19±6	10.79	<0.01
Family history (yes/no)	1/8	7/17	χ ²=1.16	>0.05
BMI (kg/m²)	22.3±1.9	16.8±2.7	5.49	<0.01
Hemoglobin (g/L)	12.83±1.72	12.86±1.01	0.43	>0.05
Hospital stay (days)	24.9±5.2	12.6±6.3	9.33	<0.05

BMI, body mass index; F1DM, fulminant type 1 diabetes mellitus; T1DM, type 1 diabetes mellitus.

The disease course was a few days in patients with F1DM, which was much shorter than that of typical T1DM (Table 2). None of the F1DM patients died after treatment.

Table 2.

Clinical Symptoms of the Patients

	F1DM	Typical T1DM	t value	P value
Flu-like symptoms	6	0	χ ²=19.56	<0.01
Gastrointestinal symptoms (n)	5	0	χ ²=15.17	<0.01
Course of disease (days)	3.3±2.8	52.5±32.6	4.47	<0.01
Disturbance of consciousness (n)	3	0	χ ²=8.8	<0.01
Breath rate (breaths/min)	25±4.7	19.2±0.8	5.87	<0.01
Heart rate (beats/min)	91±11	85±4	5.87	<0.01
Temperature (°C)
Day 1	36.27±1.01	36.89±0.63	1.91	>0.05
Day 2	37.15±0.35	36.79±0.44	1.86	>0.05

F1DM, fulminant type 1 diabetes mellitus; T1DM, type 1 diabetes mellitus.

Clinical symptoms

Six of the nine (66.7%) F1DM patients had flu-like symptoms, including sore throat, cough, and fever (Table 2). Five of the nine patients showed gastrointestinal symptoms, including nausea, vomiting, and abdominal pain. One-third of them experienced disturbance of consciousness. However, none of these symptoms was found in patients with typical T1DM (P<0.01 for all).

The breath rate of F1DM patients (25±4.7/min) was much faster than that of typical T1DM patients (19.2±0.8/min) (t=5.87, P<0.01). Heart rate was also faster in patients with fulminant T1DM (91±11 beats/min) compared with typical T1DM patients (85±4 beats/min) (t=5.78, P<0.01). However, no significant difference was found regarding blood pressure. Ultrasound and computed tomography did not shown any pancreatic abnormality.

Laboratory measurements

A significantly high level of random blood glucose (40±9 mmol/L) was found in F1DM patients compared with typical T1DM patients (23±4 mmol/L) (t=9.22, P<0.01) (Table 3). However, the HbA1c (7.1±1.0%) and CRP (0.21±0.15 μg/L) levels at 2 h postprandial were significantly lower in F1DM patients compared with typical T1DM patients (14.4±2.2% [t=7.66, P<0.01] and 0.58±0.39 μg/L [t=2.74, P=0.01], respectively), whereas no significant difference was found in fasting and urine CRP levels between the two groups. Increased leukocyte (18.2±7.8×10⁹/L) and neutrophil (81±4%) numbers were found in F1DM patients compared with typical T1DM patients (t=7.33 and t=6.27, respectively; P<0.05 for both). No significant differences in anti-glutamate decarboxylase and anti-insulin antibodies were found between two groups.

Table 3.

Laboratory Measurements

	F1DM	Typical T1DM	t value	P value
Random blood glucose (mmol/L)	40±9	23±4	9.22	<0.01
HbA1c (%)	7.1+1	14.4	7.66	<0.01
CRP at 2 h postprandial (μ/L)	0.21±0.15	0.58±0.39	2.74	0.0001
Urine ketone body (+)	1.78±0.67	2.17±0.82	1.28	>0.05
Blood pH	7.18±0.17	7.31±0.08	3.07	<0.01
Access base (mmol/L)	−15±7.7	−8±7	2.23	<0.05
Glutamate-pyruvate transaminase (U/L)	145±133	10±2	5.1	0.01
Glutamic-oxaloacetic transaminase (U/L)	140±116	18±6	5.3	0.01
γ-Glutamyl transpeptidase (U/L)	105±49	12±3	9.62	<0.01
Creatine kinase (U/L)	407±246	79±10	6.73	<0.01
Creatine (U/L)	157±45	55±16	9.73	<0.01
Alkaline phosphatase (U/L)	146±36	111±38	2.27	<0.05
BUN (mmol/L)	13.9±5.4	3.6±0.9	9.3	<0.01
Potassium (mmol/L)	5.4±0.6	3.4±0.8	6.51	<0.01
Sodium (mmol/L)	129±13	136±7	1.82	>0.05
Chloride (mmol/L)	96±8	99±6	1.46	>0.05
Cholesterol (mmol/L)	3.6±0.5	4.4±1.0	2.28	<0.05
Triglyceride (mmol/L)	1.3±0.6	1.2±1.1	0.38	>0.05
HDL-cholesterol (mmol/L)	0.62±0.13	0.89±0.12	5.35	<0.01
Leukocytes (×10⁹/L)	18.2±7.8	6±1.9	7.33	<0.01
Neutrophils (%)	81±4	58±10	6.27	<0.01
Albumin (g/L)	35.4±2.8	37.2±2.3	1.8	>0.05
Anti-insulin antibody (+/−)	0/9	0/24	1.45	>0.05
Anti-glutamate decarboxylase antibody (+/−)	1/8	9/15	χ ²=2.16	<0.05

BUN, blood urea nitrogen; CRP, C-reactive peptide; F1DM, fulminant type 1 diabetes mellitus; HbA1c, glycated hemoglobin; HDL, high-density lipoprotein; T1DM, type 1 diabetes mellitus.

Creatine kinase, creatine, GPT, GOT, GGT, ALP, BUN, and blood potassium levels of F1DM patients were significantly higher than those of typical T1DM patients (P<0.05 for all). Cholesterol (3.6±0.5 mmol/L), high-density lipoprotein-cholesterol (0.62±0.13 mmol/L), blood pH values (7.18±0.17), and base access in F1DM were significantly lower than those in typical T1DM (P<0.05 for all). No differences in sodium, chloride, triglycerides, and albumin were found between the two groups.

Treatment

When the diagnosis of F1DM is strongly suspected and diagnosed, treatment of the patient should be started immediately. Continuous intravenous insulin infusion therapy using a portable pump to deliver a biosynthetic short-acting human insulin, Humulin R (Eli Lilly and Co., Indianapolis, IN), was given for 24–72 h to control hyperglycemia and ketoacidosis. After the acute phase, Humulin R will be continuously injected four times a day or continuously administered through an insulin pump. This treatment strategy was followed until the patient was discharged. Only one patient with F1DM was still using an insulin pump at discharge. The remaining eight F1DM and 24 typical T1DM patients switched to insulin injection at discharge: three preprandial injections of Humulin R at breakfast, lunch, and dinner and one injection before bedtime with insulin glargine or Humulin N.

All the patients in this study survived and were discharged after treatment. The insulin dosage in F1DM patients was 0.48±0.13 U/kg, whereas it was 0.68±0.33 U/kg for typical T1DM patients (t=1.77, P>0.05).

Discussion

In this study we have shown that the prevalence of F1DM was 4.5% among all patients newly diagnosed with T1DM in Nanjing, China. The clinical course of the disease and its biological parameters are similar to those reported in Japan⁴ or Korea.⁵

In this single hospital-based study lasting for 5.5 years, the prevalence of F1DM was 4.5% (nine of 198 T1DM patients). This percentage is smaller than that reported by Zhen et al.,⁶ who found the prevalence of F1DM was 9.1% of newly diagnosed T1DM patients, and that in Japan, with 20% of acute-onset T1DM cases,⁴ and in Korea, with 7.1%.⁵

Although the age at onset of F1DM could range from 1 to 80 years, the age at onset of F1DM is significantly older than that for typical T1DM. In the current study, the age at onset was 46±7 years, which was significantly older than for typical T1DM (19±6 years). This is similar to the report of Imagawa et al.¹ that the mean age at onset was 39±16 years in the Japanese population. In comparison, the age in the current study was older than in the other two studies in China: Zheng et al.⁶ reported the average age of F1DM patients was 31±7 years, and Lu et al.⁷ reported patients as young as 22.4±6.7 years. This difference could be due to regional variation.

Table 2 shows the symptoms observed at or immediately before onset of the F1DM. In this type of diabetes, the duration from the start of symptoms associated with hyperglycemia, such as thirst, polyuria, or polydipsia, is mostly within 7 days.^1,2 It was 3.3±2.8 days in the current study, which is much shorter than that (52.5±32.6 days) seen in typical T1DM. This is consistent with the studies of Imagawa et al.⁴ (4 days), Zheng et al.⁶ (2 days), and Lu et al.⁷ (3.4 days). This marked difference suggests that F1DM has a striking fast disease progression. The most remarkable symptoms observed immediately before the onset of the disease are common cold–like symptoms (66.7%), followed by abdominal symptoms (55.6%) and disturbance of consciousness (33%) in this study. Although leukocyte count and neutrophil number also increased, no fever was found in these patients. This suggests viral infection might be partially involved in the development of F1DM.

The mechanism underlying the development of F1DM is unclear. It might be related to genetic susceptibility,³ autoimmunity,⁸ and viral infection.⁹ Evidence ahs demonstrated that different genotypes of HLA plays important roles in conferring susceptibility or resistance to the development of T1DM, especially HLADR4-DQ4, which was found in 41.8% of F1DM patients.¹⁰ The study of Nakanishi and Saitoh¹¹ found that simultaneous appearance of HLA-24, DQA1*03, and DR9 could cause F1DM and destroy β-cells.

Autoimmunity may also be involved in the development of F1DM.^8,10 In the current study, anti-insulin antibody levels of F1DM and typical T1DM patients were all negative. However, one F1DM patient was positive for anti-glutamate decarboxylase antibody, suggesting an autoimmune mechanism may also be involved in the development of F1DM in this patient. The acute and strong destruction of β-cells and ketoacidosis may cause multiple organ injury, including liver, kidney, etc. This results in increased levels of GPT, GOT, GGT, ALP, BUN, creatine kinase, creatine, etc.

Most of the patients had flu-like and gastrointestinal symptoms. This suggests that F1DM patients might have a viral infection. This might be a key step in the development of F1DM. The infection caused an autoimmune reaction that is so strong that it caused damage to the pancreas and other organs. Therefore, both β- and α-cells are greatly reduced in number in F1DM.

Ketoacidosis is usually seen at the onset of this type of diabetes. Plasma ketone bodies are elevated (ketosis) at the onset of F1DM with arterial blood gas pH below 7.35 (acidosis) in most patients.^1,2 Among the serum electrolytes, only potassium was significantly increased in this study; sodium and chloride levels were insignificantly decreased. These changes are more severe than those in autoimmune T1DM, indicating the more severe metabolic derangement in F1DM. In the current study, the ketone body level was increased with pH at 7.18±0.17; the base excess was −15±7.7 mmol/L. Serum sodium and chloride levels were insignificantly decreased, whereas potassium content was significantly increased to 5.4±0.5 mmol/L (t=6.51, P<0.05). These are consistent with the findings of Imagawa et al.¹

The window of opportunity for treatment of F1DM is limited. When the diagnosis of F1DM is strongly suspected, treatment of the patient must be started immediately. Otherwise, the disease will quickly deteriorate, leading to the death of the patient within 24 h. The treatment of F1DM in the nine patients described here is similar to that of typical T1DM. At onset, when ketoacidosis is present, intravenous infusion of saline and insulin will be started. A portable insulin pump is used to continuously infuse Humulin R for 24–72 h to control hyperglycemia and ketoacidosis. When ketoacidosis is over, we either used portable insulin pumps or switch to insulin injection. The use of an insulin pump is convenient for dosage adjustment and provides a fast and continuous control of glucose.

In this one hospital-based study, the nine Chinese patients met the criteria for diagnosis of F1DM. Although F1DM may be fatal, prompt treatment is a must to rescue these patients.

Footnotes

Author Disclosure Statement

The authors have declared that no competing interests exist.

References

Imagawa

, Hanafusa

, Miyagawa

, Matsuzawa

. A novel subtype of type 1 diabetes mellitus characterized by a rapid onset and an absence of diabetes-related antibodies. N Engl J Med, 2000; 342:301–307.

Hanafusa

, Imagawa

. Fulminant type 1 diabetes: a novel clinical entity requiring special attention by all medical practitioners. Nat Clin Pract Endocrinol Metab, 2007; 3:36–45.

Imagawa

, Hanafusa

, Uchigata

, Kanatsuka

, Kawasaki

, Kobayashi

, Shimada

, Shimizu

, Maruyama

, Makino

. Different contribution of class II HLA in fulminant and typical autoimmune type 1 diabetes mellitus. Diabetologia, 2005; 48:294–300.

Imagawa

, Hanafusa

, Uchigata

, Kanatsuka

, Kawasaki

, Kobayashi

, Shimada

, Shimizu

, Toyoda

, Maruyama

, Makino

. Fulminant type 1 diabetes: a nationwide survey in Japan. Diabetes Care, 2003; 26:2345–2352.

Cho

, Kim

, Ko

, Koo

, Yang

, Park

, Lee

, Park

. Fulminant type 1 diabetes in Korea: high prevalence among patients with adult-onset type 1 diabetes. Diabetologia, 2007; 50:2276–2279.

Zheng

, Lin

, Yang

, Zhang

, Tan

, Yang

, Xiao

, Che

, Zhou

. Prevalence and clinical features of fulminant type 1 diabetes [in Chinese] Chin J Endocrinol Metab, 2010; 26:188–191.

, Liu

, Shao

, Lai

, Zou

, Yan

. Clinical analysis of five cases of fulminant type I diabetes mellitus [in chinese] Chin J Endocrinol Metab, 2010; 26:192–194.

Shimada

, Maruyama

. Encephalomyocarditis-virus-induced diabetes model resembles “fulminant” type 1 diabetes in humans. Diabetologia, 2004; 47:1854–1855.

Imagawa

, Hanafusa

, Makino

, Miyagawa

, Juto

. High titres of IgA antibodies to enterovirus in fulminant type-1 diabetes. Diabetologia, 2005; 48:290–293.

10.

Imagawa

, Hanafusa

. Pathogenesis of fulminant type 1 diabetes. Rev Diabet Stud, 2006; 3:169–177.

11.

Nakanishi

, Saitoh

. Clinical and genetic characteristics of patients with type 1 diabetes associated with interferon therapy. Diabetes Care, 2011; 34:471–473.