X/00/$03.00/0 Vol. 85, No. 12 The Journal of Clinical Endocrinology & Metabolism Copyright 2000 by The Endocrine Society

Transcription

1 X/00/$03.00/0 Vol. 85, No. 12 The Journal of Clinical Endocrinology & Metabolism Printed in U.S.A. Copyright 2000 by The Endocrine Society Prognostic Factors for the Course of Cell Function in Autoimmune Diabetes* C. TÖRN, M. LANDIN-OLSSON, Å. LERNMARK, J. P. PALMER, H. J. ARNQVIST, G. BLOHMÉ, F. LITHNER, B. LITTORIN, L. NYSTRÖM, B. SCHERSTÉN, G. SUNDKVIST, L. WIBELL, AND J. ÖSTMAN Department of Medicine (C.T., M.L.-O.), University Hospital, Lund, Sweden; Department of Medicine, R. H. Williams Laboratory (Å.L.) and Division of Endocrinology, Metabolism and Nutrition, Diabetes Endocrinology Research Center and Diabetes Care Center, DVA Puget Sound Health Care System (J.P.P.), University of Washington, Seattle, Washington; Department of Medicine and Care (H.J.A.), Linköping University, Linköping, Sweden; Department of Internal Medicine (G.B.), Stockholm Söder Hospital, Stockholm, Sweden; Department of Medicine (F.L.), Umeå University Hospital, Umeå, Sweden; Department of Community Health Sciences (B.L.), University of Malmö/Lund, Malmö, Sweden; Department of Public Health and Clinical Medicine (L.N.), University of Umeå, Umeå, Sweden; Department of Community Health Sciences Dalby/Lund (B.S.), Lund, Sweden; Department of Endocrinology (G.S.), Malmö University Hospital, Malmö, Sweden; Department of Medicine (L.W.), University Hospital, Uppsala, Sweden; and Department of Medicine (J.Ö.), Huddinge Hospital, Stockholm, Sweden ABSTRACT This study presents a 2-yr follow-up of 281 patients, aged yr, diagnosed with diabetes between 1992 and At diagnosis, 224 (80%) patients were positive for at least one of the following autoantibodies: islet cell antibodies (ICAs), glutamic acid decarboxylase antibodies (GADAs), or tyrosine phosphatase antibodies (IA-2As); the remaining 57 (20%) patients were negative for all three autoantibodies. At diagnosis, C-peptide levels were lower (0.27; nmol/l) in autoantibody-positive patients compared with autoantibodynegative patients (0.51; nmol/l; P 0.001). After 2 yr, C-peptide levels had decreased significantly in patients with autoimmune diabetes (0.20; nmol/l; P ), but not in autoantibody-negative patients. In patients with autoimmune diabetes, a low initial level of C-peptide (odds ratio, 2.6; 95% confidence interval, ) and a high level of GADAs (odds ratio, 2.5; 95% IN TYPE 1 diabetes autoantibodies can be detected many years before the clinical presentation (1 4). At diagnosis, about 80% of type 1 diabetic patients have pancreatic islet autoantibodies, islet cell antibodies (ICAs), glutamic acid decarboxylase antibodies (GADAs), tyrosine phosphatase antibodies (IA-2As), or insulin autoantibodies (IAAs) (5 9). These autoantibodies are thought to indicate a progressive autoimmune disease in the cells associated with a gradual decrease in insulin secretion. At clinical onset of type 1 diabetes, an endogenous insulin production can still be measured as C-peptide (10 13), and the C-peptide levels are higher in adults than in children at diagnosis (14 16). It is likely that the -cell destruction continues after diagnosis, Received October 25, Revision received September 7, Accepted September 9, Address correspondence and requests for reprints to: Carina Törn, B 11, BMC, Lund, Sweden. Carina.Torn@med.lu.se. * Supported by NIH Grant DK-42654, the Swedish Medical Council (Grant B96-27X ), the Novo Nordisk Foundation, the Swedish Diabetes Association, the Malmö Diabetes Association, the Lundströms Foundation, and the Stig Alméns Foundation. confidence interval, ) were risk factors for a C-peptide level below the reference level of 0.25 nmol/l 2 yr after diagnosis. Body mass index had a significant effect in the multivariate analysis only when initial C-peptide was not considered. Factors such as age, gender, levels of ICA or IA-2A or insulin autoantibodies (analyzed in a subset of 180 patients) had no effect on the decrease in -cell function. It is concluded that the absence of pancreatic islet autoantibodies at diagnosis were highly predictive for a maintained -cell function during the 2 yr after diagnosis, whereas high levels of GADA indicated a course of decreased -cell function with low levels of C-peptide. In autoimmune diabetes, an initial low level of C-peptide was a strong risk factor for a decrease in -cell function and conversely high C- peptide levels were protective. Other factors such as age, gender, body mass index, levels of ICA, IA-2A or IAA had no prognostic importance. (J Clin Endocrinol Metab 85: , 2000) with varying rate until the cells are depleted. Because the pathogenesis of type 1 diabetes is unknown, it is difficult to define factors of importance for future decrease in C-peptide levels. C-peptide levels at diagnosis, age, gender, presence, or absence of ICA; degree of obesity and puberty; and levels of glycosylated hemoglobin are factors described to influence the -cell function after diagnosis in adolescents (17, 18). There have been contradictory reports about the importance of ICA levels for the remaining C-peptide (14, 19), GADA has been reported to be associated to lower levels of C-peptide during follow-up (20 22), a high C-peptide level at diagnosis has been associated with a higher C-peptide level during the first year of follow-up (23, 24), and male gender has been reported to be associated with a longer remission period after clinical onset (25). The population of young adult diabetic patients consists of a mixture of patients with type 1 or type 2 diabetes who are difficult to separate on a clinical basis. Many adult patients with type 1 diabetes have, contrary to children, a well preserved -cell function at diagnosis. It is, therefore, of importance to improve the clinical classification and to find prog- 4619

2 4620 TÖRN ET AL. JCE&M 2000 Vol. 85 No. 12 nostic factors for the -cell function in the autoimmune type of diabetes. In this study, we have determined pancreatic islet autoantibodies at diagnosis and used this to separate patients into autoimmune and nonautoimmune diabetes, and we hypothesized that autoantibodies at diagnosis could give an indication of initial and future -cell function. The main aim of this study was to identify prognostic factors among those variables [age, body mass index (BMI), C-peptide, gender, and autoantibodies] that are measured at the time of clinical diagnosis in most patients with tentative type 1 diabetes. Specifically, we wanted to test whether any of the four islet autoantibodies (ICA, GADA, IA-2A, or IAA) or the other previously mentioned clinical variables were more important than others in determining the course of -cell function after diagnosis in autoantibody-positive patients. Subjects Materials and Methods In this nationwide population-based study, patients were registered in the Diabetes Incidence Study in Sweden (DISS). Blood samples were obtained from consecutive patients, yr of age, newly diagnosed with diabetes (except gestational diabetes) between 1992 and The diagnosis of diabetes mellitus and reporting to DISS was done by the patients treating physician, and diagnosis was based on blood glucose levels according to criteria from WHO (26). All information was reported on a standardized form to DISS, including duration of symptoms, presence of coma, blood glucose, degree of ketonuria, body weight and height, and presence of ketoacidosis (bicarbonate 15 mmol/l and/or ph 7.3), together with clinical classification of diabetes. Blood samples and clinical classification were obtained from 764 patients at diagnosis. Of patients included in the present report, 67% (187 of 281) donated the first blood sample within 7 days from diagnosis and 78% (220 of 281) within 14 days after diagnosis. During follow-up, all patients were contacted annually by mail and asked to donate a new blood sample. Patients who did not respond to the first letter were reminded with an additional letter. Both fasting and random samples were accepted, to maximize the number of samples and also to get the first sample as close to diagnosis as possible. There is no marked difference between fasting and nonfasting C-peptide in autoantibodypositive patients (27). The samples were taken at the local hospitals and sent by mail to our laboratory for analyses. In the present report, only the 281 patients who donated samples at diagnosis and at the two yearly sampling occasions were considered. There were no differences in gender, age, C-peptide levels, BMI, or frequency of autoantibody-positive patients or levels of autoantibodies, between patients who donated yearly samples (responders; n 281) and patients who donated samples only at diagnosis (nonresponders; n 246) (Table 1). The 281 patients were divided into two groups, based on autoantibody status at diagnosis and not on clinical classification. One group consisted of 224 (80%) patients positive for at least one of ICAs, GADAs, or IA-2As at diagnosis, and the other group consisted of 57 (20%) patients who were negative for all three autoantibodies. The study was approved by the Ethical Committees at all regional centers for DISS (Stockholm, Göteborg, Linköping, Lund, Umeå, and Uppsala). Assays C-peptide was analyzed by RIA using a commercial kit (Euro-Diagnostica AB, Malmö, Sweden). The reference range for the C-peptide assay was nmol/l, and the detection limit was 0.13 nmol/l. The intra-assay variation was 5% ( nmol/l), and the total variation (sum of inter- and intra-assay variation) was 7% in the same measurement interval. ICAs were analyzed using a two-color immunofluorescence method, as described previously (28). The samples were diluted until negative. Thereafter, the highest positive titer was converted into Juvenile Diabetes Foundation Units (JDF-U), in accordance with a standard curve for the specific pancreas used. The lower detection limit was 6 JDF-U and was considered positive. The sensitivity was 100% and specificity 88% for the pancreas used in this study when tested in the International Diabetes Workshop (IDW) for standardization (29). GADAs were analyzed with a radioimmunoprecipitation assay, as described in detail (30, 31). The reference range was defined using 833 blood samples from controls matched for age and sex. An index below 0.08 (97.5 percentile) was considered negative. The sensitivity was 81% and specificity 95% when tested in IDW for standardization (29). IA-2As were analyzed by radioimmunoprecipitation assay using the ICA512 complementary DNA (32). It is likely that IA-2 and ICA512 represents the same protein and epitope because the overlapping region of their complementary DNA is identical except for one nucleotide (33). In the IDW (29), both the longer and the shorter versions of the protein used in RIAs gave similar sensitivity and specificity. An index below 0.05 (97.5 percentile) was considered as negative. The reference index was defined using the same controls as for GADA. IAAs were analyzed in a radiobinding assay with displacement of cold insulin (9). A level above 0.7% of binding was considered positive. This threshold was based on previous results from healthy individuals. IAAs were only analyzed in samples taken within 7 days (180 of 281 in the responder group and 140 of 246 in the nonresponder group) after insulin treatment was initiated to avoid interference with antibodies formed against exogenous insulin. Statistical analyses Because levels of autoantibodies and C-peptide were not normally distributed, results are given as median and interquartile range. A C- peptide level of 0.25 nmol/l was used as a cut-off value for a low level after 2 yr because this is the lower level of the reference interval. The McNemar s test was used to test whether frequencies of patients with a C-peptide above 0.25 nmol/l had changed over time. The Spearman-rank correlation test (r s ) was used to test whether age, BMI, initial levels of C-peptide, ICA, GADA, IA-2A or IAA correlated TABLE 1. Characteristics at diagnosis of patients who donated yearly blood samples (responders, n 281) and patients who donated samples only at diagnosis (nonresponders, n 246) Variables Responders (n 281) Median; interquartile range No. Nonresponders (n 246) Median; interquartile range Age (yr) 26; ; BMI (kg/m 2 ) 21.8; ; C-peptide (nmol/l) 0.29; ; Male gender (%, 95% CI) 60; ; ICA (JDF-U) 131; (57%) 131; (47%) GADA (index) 0.47; (63%) 0.41; (55%) IA-2A (index) 0.89; (41%) a 0.79; (38%) IAA ( % binding) 1.6; (23%) b 2.4; (21%) b a IA-2A was missing in one subject at diagnosis. b IAA was analyzed in a subpopulation (with samples taken within 7 days of diagnosis) of 180 patients in the responder group and 140 in the nonresponder group. No.

3 C-PEPTIDE IN AUTOIMMUNE DIABETES 4621 to levels of C-peptide during the 2 yr of follow-up. The Mann-Whitney U test was used to compare differences in C-peptide levels between groups. Friedman s test was used to test for differences in levels of C-peptide over time, and differences were further tested with the Wilcoxon signed rank test. A multiple logistic regression analysis was used to identify risk factors (age, BMI, initial level of C-peptide, ICA, GADA, IA-2A, IAA, or gender) for a C-peptide level below 0.25 nmol/l 2 yr after diagnosis. In this analysis, the initial level of C-peptide was categorized into three groups (below 0.25 nmol/l, nmol/l and above 0.50 nmol/l), as well as BMI (below 21 kg/m 2, kg/m 2 and above 25 kg/m 2 ). A P less than, 0.05 was considered significant. The Statistical Package for Social Sciences (version 6.1 for Macintosh; SPSS, Inc., Chicago, IL ) was used for the statistical analyses. Results Natural course of C-peptide At diagnosis, the autoantibody-negative patients had significantly higher C-peptide levels (n 57; 0.51; nmol/l) compared with autoantibody-positive patients (n 224; 0.27; nmol/l; P 0.001) (Fig. 1). In autoantibody-positive patients, C-peptide levels were unchanged during the first year after diagnosis, but declined during the second year of follow-up (0.20; nmol/l); P ). Autoantibody-negative patients (n 57) had no significant changes during these 2 yr (Fig. 1). The only difference in C-peptide levels between genders was observed after 1 yr in autoantibody-positive patients; C-peptide levels were higher in men (n 136; 0.30; nmol/l) compared with levels in women (n 87; 0.22; nmol/l; P 0.022). At diagnosis, 54.9% (123/224) of autoantibodypositive patients had a C-peptide level above the lower reference value of 0.25 nmol/l, and after 2 yr the frequency had decreased to 41.5% (93 of 224; P 0.01). In autoantibodynegative patients there was no significant difference in frequency of patients with a C-peptide above 0.25 nmol/l; it was 80.7% (46 of 57) at diagnosis and 66.7% (38 of 57) after 2 yr. FIG. 1. C-peptide levels were lower in autoantibody-positive patients (n 224; P 0.001) compared with autoantibody-negative patients (n 57) at diagnosis and during the 2 yr of follow-up. C-peptide levels decreased during the second year in autoantibody-positive patients (P )., Medians of the separate groups; lines, interquartile ranges. The time axis has been displaced to the right for the autoantibody negative patients. Relation between initial levels of autoantibodies and C-peptide In the group of autoimmune diabetes (n 224), high initial levels of C-peptide were significantly correlated to high levels of C-peptide during follow-up, and high levels of GADA were correlated to low levels of C-peptide at diagnosis and during follow-up (Table 2). BMIs were positively correlated to initial levels of C-peptide (r s 0.32; P 0.001) and also to age (r s 0.28; P 0.001). Predictive factors for C-peptide level after 2 years in autoantibody-positive patients A low C-peptide level and high GADA at diagnosis were found to be significant risk factors for a low (below the reference limit of 0.25 nmol/l) remaining C-peptide after 2 yr in a logistic regression model. The OR for C-peptide was 2.6 (95%CI, ) and for GADA 2.5 (95%CI, ). In the next regression model, only patients with an initial C-peptide above 0.25 nmol/l (n 121) were considered. Also in this model, the lower initial C-peptide in the interval was the strongest risk factor (OR 3.7; 95%CI, ) for a low C-peptide after 2 yr, whereas GADA did not reach significance as a risk factor in this analysis. In the third regression model, where initial C-peptide was not included, high GADA at diagnosis could be demonstrated to be a significant risk factor (OR 2.9; 95%CI, ) and also low BMI (OR %CI, ) (Table 3). Factors as age, gender, or levels of other pancreatic islet autoantibodies (ICA or IA-2A or IAA) were not significant. Discussion In this study, we have followed the natural course of -cell function in newly diagnosed young adult diabetic patients of both type 1 and type 2. Because the clinical classification of diabetes into type 1 and type 2 is difficult in this age group of patients, we have used autoantibody status at diagnosis as a base for classification into autoimmune and nonautoimmune diabetes. As many as 224 of 281 (80%) of these diabetic patients aged yr old were positive for one or several autoantibodies at diagnosis, and only 57 (20%) were autoantibody negative. The presence or absence of autoantibodies at diagnosis were of major importance for C-peptide levels at diagnosis and during follow-up. We found that the presence of autoantibodies predicted a lower final C-peptide level after 2 yr compared with the initial level, whereas absence of autoantibodies resulted in a preserved -cell function without significant change in C-peptide level. This was more or less expected because patients with autoimmune diabetes have a continuous process of destruction of cells, probably caused by a combination of humoral and a cellular activity (34, 35). In autoantibody-negative patients, the mechanism of disease is different since the immune system may not be involved and these patients are more likely to be conventional type 2. In autoimmune diabetes, a low level of C-peptide at diagnosis was an indicator for a low -cell function (C-peptide below 0.25 nmol/l) after 2 yr. A high initial level of GADA was a significant risk factor for a C- peptide level below 0.25 nmol/l, within 2 yr after diagnosis,

4 4622 TÖRN ET AL. JCE&M 2000 Vol. 85 No. 12 TABLE 2. High C-peptide levels at diagnosis correlated to high C-peptide levels during follow-up, and high levels of GADA at diagnosis showed a significant correlation to low levels of C-peptide at diagnosis and during follow-up in autoantibody-positive patients (n 224) Variables at diagnosis whereas levels of other pancreatic autoantibodies (ICA or IA-2A or IAA), age, or gender were of no importance when all these variables were compared in a multivariate logistic regression analysis. Because 46% of the patients with autoimmune diabetes had a C-peptide level below 0.25 nmol/l already at diagnosis and these patients could not be expected to improve their -cell function during follow-up, the same analysis was repeated without these patients with no or very low -cell function. Despite this, a moderate C-peptide level at diagnosis was still a strong risk factor, whereas GADA was no longer significant, probably because of the decreased number of observations. In a third regression analysis, without inclusion of initial C-peptide, high GADA and low BMI were significant risk factors. BMI was positively correlated to C-peptide at diagnosis C-peptide after 1 yr C-peptide after 2 yr r s P r s P r s P Age NS 0.13 NS NS BMI C-peptide NA NA ICA NS NS GADA IA-2A NS NS NS IAA a NS NS NS a IAA was analyzed in a subpopulation of 180 patients (with samples taken within 7 days of diagnosis), and 152 of these patients were positive for at least one of the four autoantibodies (ICA, GADA, IA-2A, or IAA). The correlation analysis for IAA was done in these 152 patients. NS, Not significant. TABLE 3. A low C-peptide and high GADA at diagnosis were found to be significant risk factors for a decreased cell function (final C-peptide level below 0.25 nmol/l) within 2 yr of diagnosis in autoantibody-positive patients (n 222) in a multiple logistic regression analysis (left). The same multiple logistic regression analysis was repeated in the subgroup with samples donated within 7 days in which IAA was also analyzed, and at least one of the four autoantibodies were found in 151 of these patients (right) Independent variable OR 95% CI OR 95% CI Including IAA Patients with autoimmune diabetes (n 222) Patients with autoimmune diabetes (n 151) C-peptide a GADA Gender IA-2A BMI b Age ICA IAA NA NA Initial C-peptide 0.25 nmol/l (n 121) Initial C-peptide 0.25 nmol/l (n 75) C-peptide a GADA Gender IA-2A BMI b Age ICA IAA NA NA Initial C-peptide omitted from analysis (n 222) Initial C-peptide omitted from analysis (n 151) GADA Gender IA-2A BMI b Age ICA IAA NA NA a C-peptide was categorized into three groups: below 0.25 nmol/l (3), nmol/l (2), and above 0.50 nmol/l (1). b BMI was categorized into three groups: below 21 kg/m 2 (3), kg/m 2 (2), and above 25 kg/m 2 (1). NA, Not applicable. age and initial C-peptide in the univariate analysis but had a significant effect of its own in the multivariate analysis only when initial C-peptide was omitted. A nationwide study of this size, with several years of follow-up, is difficult to perform with complete sampling. To avoid a bias in the interpretation of the results of the followed group, we did a careful comparison of responders and nonresponders for the data available at diagnosis. Because no differences in sex ratio or proportion of autoantibody positives or any other important variables were found, we conclude that the 281 patients followed for 2 yr are representative for the entire cohort. From the patients point of view, it is of highest priority to find prognostic factors for the course of -cell function. Even

5 C-PEPTIDE IN AUTOIMMUNE DIABETES 4623 if the endogenous insulin production is insufficient to keep normal blood glucose levels, this insulin production facilitates a good metabolic control with only small or moderate doses of exogenous insulin. The finding that more than half of the young adult patients with autoimmune diabetes had a C-peptide level above the lower reference range suggests that part of the -cell function is left at the time of clinical diagnosis. In addition to this, we have showed that no significant decrease of -cell function occurs during the first year of disease, but during the second year. During the first year after diagnosis, the -cell function was in a steady state, and when future intervention become available, the degenerating process might be stopped to prolong this steady-state period. The finding that patients diagnosed with low levels of C-peptide and high levels of GADA are in particular risk to decrease in -cell function could be taken into account for future intervention studies. In conclusion, the presence of autoantibodies were predictive for a decline of -cell function, measured as C- peptide, and conversely the absence of autoantibodies predicted a course of preserved -cell function during the first 2 yr after diagnosis. In autoimmune diabetes, low C-peptide level and also high GADA at diagnosis were risk factors for a decrease in -cell function. The levels of other autoantibodies (ICA or IA-2A or IAA) or factors such as age, BMI, or gender were of no prognostic importance for the course of -cell function. Acknowledgments Birgitta Persson, Gunnel Jonsson, Berit Persson, Eine Valtersson, and Regina Park are acknowledged for expert technical assistance. We thank our patients for participation and Kerstin Lantz and Madeleine Morein, who kept records of the patients. Jonas Ranstam is thanked for precious advice on statistical methods. Dr. Anders Isaksson is thanked for providing C-peptide analyses at the Department of Clinical Chemistry, University Hospital (Lund, Sweden). Dr. George Eisenbarth (Denver, CO) is thanked for kindly donating plasmid pica512 bdc. References 1. Gorsuch AN, Spencer KM, Lister J, et al Evidence for a long prediabetic period in type 1 (insulin-dependent) diabetes mellitus. Lancet. 2: Bärmeier H, McCulloch DK, Neifing JL, et al Risk for developing type 1 (insulin-dependent) diabetes mellitus and the presence of islet 64K antibodies. Diabetologia. 34: Veijola R, Vähäsalo P, Tuomilehto-Wolf E, et al Human leukocyte antigen identity and DQ risk alleles in autoantibody-positive siblings of children with IDDM are associated with reduced early insulin response. Diabetes. 44: Whittingham S, Byron SL, Tuomilehto J, et al Autoantibodies associated with presymptomatic insulin-dependent diabetes mellitus in women. Diabetes Med. 14: Landin-Olsson M, Karlsson A, Dahlquist G, Blom L, Lernmark Å, Sundkvist G Islet cell and other organ-specific autoantibodies in all children developing type 1 (insulin-independent) diabetes mellitus in Sweden during one year and in matched controls. Diabetologia. 32: Landin-Olsson M, Karlsson FA, Lernmark Å, Sundkvist G Islet cell and thyrogastric antibodies in 633 consecutive years old patients in the Diabetes Incidence Study in Sweden (DISS). Diabetes. 41: Vandewalle CL, Falorni A, Svanholm S, et al High diagnostic sensitivity of glutamate decarboxylase autoantibodies in IDDM with clinical onset between age 20 and 40 years. J Clin Endocrinol Metab. 80: Savola K, Bonifacio E, Sabbah E, et al IA-2 antibodies - a sensitive marker of IDDM with clinical onset in childhood and adolescence. Diabetologia. 41: Palmer JP, Asplin CM, Clemons P, et al Insulin antibodies in insulindependent diabetics before insulin treatment. Science. 222: Agner T, Damm P, Binder C Remission in IDDM: prospective study of basal C-peptide and insulin dose in 268 consecutive patients. Diabetes Care. 10: Landin-Olsson M, Nilsson KO, Lernmark Å, Sundkvist G Islet cell antibodies and fasting C-peptide predict insulin requirement at diagnosis of diabetes mellitus. Diabetologia. 33: Arnqvist HJ, Littorin B, Nyström L, et al Difficulties in classifying diabetes at presentation in the young adult. Diabetes Med. 10: Wroblewski M, Gottsäter A, Lindgarde F, Fernlund P, Sundkvist G Gender, autoantibodies, and obesity in newly diagnosed diabetic patients aged years. Diabetes Care. 21: Wallensteen M, Dahlquist G, Persson B, et al Factors influencing the magnitude, duration, and rate of fall of B-cell function in type 1 (insulindependent) diabetes children followed for two years from their clinical diagnosis. Diabetologia. 31: Karjalainen J, Salmela P, Ilonen J, Surcel H-M, Knip M A comparison of childhood and adult type 1 diabetes mellitus. N Engl J Med. 320: Snorgaard O, Lassen LH, Binder C Homogeneity in pattern of decline of -cell function in IDDM. Diabetes Care. 15: Schiffrin A, Suissa S, Weitzner G, Poussier P, Lalla D Factors predicting course of -cell function in IDDM. Diabetes Care. 15: Klipper-Aurbach Y, Wasserman M, Braunspiegel-Weintrob N, et al Mathematical formulae for the prediction of the residual beta cell function during the first two years of disease in children and adolescents with insulindependent diabetes mellitus. Med Hypotheses. 45: Gottsäter A, Landin-Olsson M, Fernlund P, Lernmark Å, Sundkvist G cell function in relation to islet cell antibodies (ICA) during the first three years after the clinical diagnosis of diabetes in non-insulin-dependent (type II) diabetic patients. Diabetes Care. 16: Petersen JS, Dyrberg T, Karlsen AE, et al Glutamic acid decarboxylase (GAD 65 ) autoantibodies in prediction of -cell function and remission in recent-onset IDDM after cyclosporin treatment. Diabetes. 43: Gottsäter A, Landin-Olsson M, Lernmark Å, Fernlund P, Sundkvist G, Hagopian WA Glutamate decarboxylase antibody levels predict rate of betacell decline in adult onset diabetes. Diabetes Res Clin Pract. 27: Bonfanti R, Bazzigaluppi E, Calori G, et al Parameters associated with residual insulin secretion during the first year of disease in children and adolescents with type 1 diabetes mellitus. Diabetes Med. 15: Marner B, Agner T, Binder C, et al Increased reduction in fasting C-peptide is associated with islet cell antibodies in type I (insulin-dependent) diabetic patients. Diabetologia. 28: Martin S, Pawlowski B, Greulich B, Ziegler A, Mandroup-Poulsen T, Mahon J Natural course of remission in IDDM during 1 st yr after diagnosis. Diabetes Care. 15: Schölin A, Berne C, Schvarcz E, Karlsson FA, Björk E Factors predicting clinical remission in adult patients with type 1 diabetes. J Int Med. 245: World Health Organization World Health Organization technical report series no. 727: diabetes mellitus. Report of a WHO Study Group. Geneva: WHO. 27. Törn C, Landin-Olsson M, Scherstén B. Predictability of C-peptide for autoimmune diabetes in young adult diabetic patients. Pract Diabetes Int. In press. 28. Landin-Olsson M, Sundkvist G, Lernmark Å Prolonged incubation in the two-colour immunofluorescence test increases the prevalence and titres of islet cell antibodies in type 1 (insulin-dependent) diabetes mellitus. Diabetologia. 30: Verge CF, Stenger D, Bonifacio E, et al Combined use of autoantibodies (IA-2 autoantibody, GAD autoantibody, insulin autoantibody, cytoplasmic islet cell antibodies) in type 1 diabetes. Diabetes. 47: Grubin CE, Daniels T, Toivola B, et al A novel radioligand binding assay to determine diagnostic accuracy of isoform-specific glutamic acid decarboxylase antibodies in childhood IDDM. Diabetologia. 37: Falorni A, Örtqvist E, Persson B, Lernmark Å Radioimmunoassays for glutamic acid decarboxylase (GAD65) and GAD65 autoantibodies using 35 Sor 3 H recombinant human ligands. J Immunol Methods. 186: Gianani R, Rabin DU, Verge CF, et al ICA512 autoantibody radioassay. Diabetes. 44: Bonifacio E, Lampasona V, Genovese S, Ferrari M, Bosi E Identification of protein tyrosine phosphatase-like IA-2 (Islet cell antigen 512) as the insulindependent diabetes-related 37/40 autoantigen and a target of islet-cell antibodies. J Immunol. 155: Duranovic-Bellò I, Hummel M, Ziegler AG Cellular immune response to diverse islet cell antigens in IDDM. Diabetes. 45: Hummel M, Durinovic-Bello I, Ziegler A-G Relation between cellular and humoral immunity to islet cell antigens in type 1 diabetes. J Autoimmunity. 9: