Endocrinol.Japon. 1982, 29 (1),87-94 Various Phenotypes of Diabetes Mellitus as Ultimate Outcome of Acutely Developed Diabetic State Induced by Viral Infection KYOHEI NONAKA, HIROYUKI TOYOSHIMA*, MITSUYOSHI NAMBA AND SEIICHIRO TARUI The Second Department of Internal Medicine, Osaka University Medical School, Osaka 553 *Minoo City Hospital , Osaka 562 Abstract For 4 to 8years we followed up 3 diabetic patients in whom the onset of diabetes seemed to be closelyrelated to the well-documented Epstein-Barr virus infection(Case 1) or Coxsackie B4 virus infection(Case 2, 3). Although all developed acute ketosis-prone diabetes in the convalescent stage of the viral infections,the subsequent clinicalcourses were quite differentfrom each other. Case 1 has remained consistentlyinsulin-dependent and associatedwith positiveislet cell antibody, gastricparietal cell antibody, thyroglobulin hemoagglutinating antibody and thyroidal microsomal hemoagglutinating antibody. Case 2 restored normal glucose tolerance. Case 3 has become noninsulin-dependent diabetes mel- litusafter a 6year interval.Thus, it is reasonably presumed that viruses could be respons- iblefor the occurrence of differentphenotypes of diabetes. In at leastsome cases of acute-onsetinsu- logicalattack on the isletstriggered by a viral lin-dependentdiabetes (IDDM), an etiologicalinfection. Therefore, it is particularlyim- role of viralinfections has recentlybeen sug- portant at presentto accumulate data on long- gested by epidemiological,clinical, and ex- term clinicalobservations on acute diabetes perimental research. We have, however, as preceded by or associated with well-docu- yet, littleinformation on the ultimate clinical mented viralinfections. outcome of the virus-inducedketosis-prone For 4 to 8years we have been able to ob- diabeticstate after the acute phase in man. As serve and follow up 3 diabetic patients,in for encephalomyocarditis(EMC) virus-induced whom well-documented viralinfections seemed diabetesin mice, it was reported that a large to be the caustiveagents. In thispaper we give number of the animals showed only a transient detaileddata obtainedin our long-termclinical hyperglycemia (Boucher and Notkins, 1973). observationsof thesecases, including 1 case of Taking such reports into consideration,it Epstein-Barr(E-B) virus-inducedIDDM. So seems too simple to assume that the acute far as we know, there have been only 2 case infectiousinjury to the islettissue always reports in literatureanywhere in the world resultsin permanent IDDM. Moreover, it (Burgess et al.,1974; Ganda et al.,1977) of stillremains to be clarifiedwhether long- acute diabetesfollowing an attack of infec- standing isletcell dysfunctionin IDDM has tious mononucleosis. Here we could add a close relationshipto the sustainedimmuno- another case report of E-B virus-induced IDDM. In addition,on the basis of the re- Received October 5, 1981. sultsof these studies,we feelthat virusesare Endocrinol. Japon. 88 NONAKA et al. February 1982 responsible for the occurrence of different method of Bottazzo et al. (1974). Serum antibody to a microsomal component of the thyroid (MCHA) or phenotypes of diabetes. antibody to a thyroglobulin (TGHA) was measured using a kit from Fujizoki Co., Tokyo. Methods Case Reports Since 1973 we have continued to try to select, among patients with newly-diagnosed acute-onset Case 1 ketosis-prone diabetes, definite cases in which viral infections were considered to be the cause. A 7-yr-old Japanese boy was admitted be- In order to discriminate diabetes associated with cause of marked hyperglycemia and ketonemia. insulinopenia due to specific viral injury from diabetes Fifty days prior to admission he had had precipitated or aggravated by non-specific stress of vomiting and diarrhea followed by a fever up infections, we have adopted the following principles for the selection of "virus-induced" diabetes: to 39℃ for 3weeks. One week after these 1) In "virus-induced diabetes" -neither diabetes nor symptoms were relieved, he had marked glucose intolerance could be detected in the pa- polydipsia, polyuria and polyphagia, losing tient's history. If possible, it is desirable that direct evidence of negative urine glucose or normal 3kg in body weight. On admission, he had glucose tolerance in the past be available. systemic lymphnode enlargements and hepato- 2) Ketotic diabetes appears in the convalescent stage megaly. His laboratory data revealed; a leuko- of a viral infection rather than at the peak of it. cytosis of 9700/cmm with 3% atypical lympho- 3) Definite serological data are obtained for the di- agnosis of viral infections. cytes, PG 600mg/dl and β-hydroxybutyrate 4) Marked hypoinsulinism associated with hyper- 44.7mg/dl. Serological analyses showed the glycemia is demonstrated in the acute stage of Paul-Bunnell heterophile antibody test to be diabetes. Although some cases of virus-induced diabetes positive (1: 224) on admission and negative a month later. The immunofluorescence test would be probably excluded by these principles, it seems to be rather necessary at present to select definite for antibody to E-B VCA was at a titer of cases as strictly as possible. 1:160 on. admission, rose to as high as 1:640 Neutralizing antibodies to coxsackie viruses be- after 3weeks and remained at the same level longing to group B were measured as follows (Toyo- shima et al., 1976): mixtures of various types of viruses for an additional 3months. The titer of IgM in a dose of 100 tissue culture infective dose (TCID50) for E-B virus was 1:640 on hospitalization, and aliquots of serially diluted serum were incubated decreased to 1:40 a month later and re- for 1hour at 37℃, then transferred onto a monolayer mained at the same level for the next 3months of cynomolgus monkey kidney cell line (MIK) cell. After another 1h incubation at 37℃, antibody titers (Figure 1). After successful treatment of acute were determined by the cytopathic effect of a virus on diabetes with insulin for a month, an oral MIK cell. An antibody titerto E-B viralcapsid antigen (VCA) glucose tolerance test (O-GTT, 1.75g/kg of was measured by the methods of Hinuma et al. (1967) body weight) was performed and demonstrated and Naito et al. (1970),IgM antibody specificto E-B an almost total absence of C-peptide immuno- virus by the method of Schmitz and Scherer (1972), reactivity (CPR) response. Since then insulin an antibody titerto E-B virus-induced early antigen therapy with an ever increasing dose has been (EA) by the method of Henle et al.(1970). Paul-Bun- nell heterophile antibody (PB) was determined by the essential up to now for more than 4years and original method (Paul and Bunnell, 1932) with slight a half (Figure 2). He had no family history of modification (Kumagai, 1951). Plasma glucose (PG) diabetes and had not been obese. Urinalysis was measured with a Beckman's auto-analyzer using performed 1year prior to his admission was glucose oxidase. Plasma immunoreactive insulin (IRI) was assayed with a Phadebas kit based on the method found to be negative for glucose according to of Wide and Porath (1966). Plasma immunoreactive his physical record at elementary school. His glucagon (IRG) was measured by the method by histocompatibility antigens were HLA-AW26, Nonaka and Foa (1969). Serum isletcell antibody AW24, BW40, B7 and DR1, DRW9. MCHA (ICA) and gastricparietal cell antibody (GPCA) were detected with the help of Dr. G. F. Bottazzo with the has appeared at the 7th month of diabetes, Vol.29, No.1 OUTCOME OF VIRUS-INDUCED DIABETES 89 Fig. 1. Sequential changes in levelsof viralantibody titerin the acute phase. Anti-VCA: antibody to E-B viral capsid antigen, anti-EA: antibody to E-B virus induced early antigen, EVB-IgM: IgM antibody specificto E-B virus, P-B: heterophileantibody determined by modified Paul-Bunnel test,(C) Bn: co- xsackie virus belonging to group Bn. Panel A: Case 1, Panel B: Case 2 and Panel C: Case 3. Fig. 2. Clinicalcourse of Case 1 (N.S. 7yrs. Male). Diabetes is permanent and consistentlyinsulin-dependent. ICA: isletcell antibody, GPCA: gastricparietal cell antibody, TGHA: Thyroglobulin hemoagglutinating antibody, MCHA: thyroid microsomal hemoagglutinating antibody. Endocrinol.Japon. 90 NONAKA et al. February 1982 and TGHA has also appeared at the 14th than 7years untilnow (Figure3). His diabetes month. Even at the end of the 3rd diabetic was thereforetransient and reversible.Cox- year, ICA and GPCA which were determined sackie virus B4 (CB4) neutralizingantibody for the first time were found to be weakly titerwas as high as 1:128 on admission, positive. increasedto more than 1:256 after 37days, and returnedto normal after6months (Figure Case 2 1). The family historywas negativefor dia- A 26-yr-old Japanese male complained of betes. ICA was negativeon 1 occasion. His symptoms of common cold followed by histocompatibilityantigens were HLA-A2, thirst and a marked loss of 10kg in body All and BW35,-. weight. Two weeks later, on admission, he was precomatous with PG 588mg/dl, plasma Case 3 IRI 12μU/ml, plasma IRG 440pg/ml (as- A 44-year-oldJapanese male was admitted sayed with Unger's 30K), arterial pH 7.05 with acute diabetes following a sore throat. with Pco2 11.2mmHg, base excess -28.5 He lost7.5kg in body weight in a week. On mEq/l, β-hydroxybutyrate 316mg/dl and admission, his FPG was 286mg/dl and IRI aceto-acetate 7.2mg/dl. Serum sodium was responded very poorly to O-GTT. After 40 130mEq/l, potassium 5.4mEq/l, chloride 95 days of insulintreatment his metabolic con- mEq/l, urea nitrogen 45mg/dl, cholesterol trol was well achieved and no drug therapy 476mg/dl. After 40days of insulin therapy, he was required afterwards. For 5years since underwent complete remission with the resto- then, O-GTTs have been either normal or ration of fasting PG to normal level and borderlinemost of the time. CB4 antibody normal patterns of PG and IRI on 100g O- titerwas 1:128 on admission and was normal- GTT which have been maintained for more ized in 3months (Figure 1).