The longitudinal loss of islet autoantibody responses from diagnosis of type 1 diabetes occurs progressively over follow-up and is determined by low autoantibody titres, early-onset, and genetic variants.

The clinical usefulness of post-diagnosis islet autoantibody levels is unclear and factors that drive autoantibody persistence are poorly defined in type 1 diabetes (T1D). Our aim was to characterise the longitudinal loss of islet autoantibody responses after diagnosis in a large, prospectively sampled UK cohort. Participants with T1D [n = 577] providing a diagnosis sample [range -1.0 to 2.0 years] and at least one post-diagnosis sample (<32.0 years) were tested for autoantibodies to glutamate decarboxylase 65 (GADA), islet antigen-2 (IA-2A), and zinc transporter 8 (ZnT8A). Select HLA and non-HLA SNPs were considered. Non-genetic and genetic factors were assessed by multivariable logistic regression models for autoantibody positivity at initial sampling and autoantibody loss at final sampling. For GADA, IA-2A, and ZnT8A, 70.8%, 76.8%, and 40.1%, respectively, remained positive at the final sampling. Non-genetic predictors of autoantibody loss were low baseline autoantibody titres (P < 0.0001), longer diabetes duration (P < 0.0001), and age-at-onset under 8 years (P < 0.01--0.05). Adjusting for non-genetic covariates, GADA loss was associated with low-risk HLA class II genotypes (P = 0.005), and SNPs associated with autoimmunity RELA/11q13 (P = 0.017), LPP/3q28 (P = 0.004), and negatively with IFIH1/2q24 (P = 0.018). IA-2A loss was not associated with genetic factors independent of other covariates, while ZnT8A loss was associated with the presence of HLA A*24 (P = 0.019) and weakly negatively with RELA/11q13 (P = 0.049). The largest longitudinal study of islet autoantibody responses from diagnosis of T1D shows that autoantibody loss is heterogeneous and influenced by low titres at onset, longer duration, earlier age-at-onset, and genetic variants. These data may inform clinical trials where post-diagnosis participants are recruited.

Islet autoantibody profiles associated with higher diabetes risk in Lithuanian compared with English schoolchildren.

During a 15-year period, the incidence of type 1 diabetes has doubled in Lithuania, while increasing by a third in England; however, England still has a higher incidence. Analysis of sera collected from non-diabetic schoolchildren from Lithuania and England more than 20 years ago showed a similar number of multiple autoantibody-positive schoolchildren between the populations, but a higher prevalence of islet antigen-2 autoantibodies (IA-2A) in English schoolchildren. We aimed to use recently developed, more specific islet autoantibody tests to characterize differences in humoral autoimmunity between these two general population cohorts in greater detail. Samples from 88 Lithuanian and 133 English schoolchildren previously found islet autoantibody-positive were selected for measurement of additional islet autoantibodies by radioimmunoassay. Samples were tested for autoantibodies to zinc transporter 8 (ZnT8A), GAD (96-585), the protein tyrosine phosphatase region of islet antigen-2 (PTPA) and the related IA-2ßA, while autoantibodies to IA-2A were reassayed using the current harmonized method. IA-2-related autoantibodies PTPA (0·13 versus 0·45%, P = 0·027) and IA-2ßA (0 versus 0·35%, P < 0·001), but not IA-2A measured using the harmonized method, were less common in Lithuanian compared to English schoolchildren. Lithuanian schoolchildren who were islet autoantibody-positive were positive for fewer biochemical autoantibodies compared with English schoolchildren (P = 0·043). Background rates of islet autoimmunity in childhood differ subtly between countries, which have different incidences of type 1 diabetes. The optimal screening strategy (age and combination of markers) for detection of islet autoimmunity may vary between countries, dependent upon the pattern of autoantibodies found in the general population.

The first 142 amino acids of glutamate decarboxylase do not contribute to epitopes recognized by autoantibodies associated with Type 1 diabetes.

AIMS: Glutamate decarboxylase (GAD) antibodies are the most widely used predictive marker for Type 1 diabetes, but many individuals currently found to be GAD antibody-positive are unlikely to develop diabetes. We have shown previously that radioimmunoassays using N-terminally truncated 35 S-GAD65 (96-585) offer better disease specificity with similar sensitivity to full-length 35 S-GAD65 (1-585). To determine whether assay performance could be improved further, we evaluated a more radically truncated 35 S-GAD65 (143-585) radiolabel. METHODS: Samples from people with recent-onset Type 1 diabetes (n = 157) and their first-degree relatives (n = 745) from the Bart's-Oxford family study of childhood diabetes were measured for GAD antibodies using 35 S-labelled GAD65 (143-585). These were screened previously using a local radioimmunoassay with 35 S-GAD65 (1-585). A subset was also tested by enzyme-linked immunosorbent assay (ELISA), which performs well in international workshops, but requires 10 times more serum. Results were compared with GAD antibody measurements using 35 S-GAD65 (1-585) and 35 S-GAD65 (96-585). RESULTS: Sensitivity of GAD antibody measurement was maintained using 35 S-GAD65 (143-585) compared with 35 S-GAD65 (1-585) and 35 S-GAD65 (96-585). Specificity for Type 1 diabetes was improved compared with 35 S-GAD65 (1-585), but was similar to 35 S-GAD65 (96-585). Relatives found to be GAD antibody-positive using these truncated labels were at increased risk of diabetes progression within 15 years, compared with those positive for GAD(1-585) antibody only, and at similar risk to those found GAD antibody-positive by ELISA. CONCLUSIONS: The first 142 amino acids of GAD65 do not contribute to epitopes recognized by Type 1 diabetes-associated GAD antibodies. Low-volume radioimmunoassays using N-terminally truncated 35 S-GAD65 are more specific than those using full-length GAD65 and offer practical alternatives to the GAD antibody ELISA for identifying children at increased risk of Type 1 diabetes.

A quarter of patients with type 1 diabetes have co-existing non-islet autoimmunity: the findings of a UK population-based family study.

Individuals with type 1 diabetes (T1D) are at increased risk of coeliac disease (CD), autoimmune thyroiditis and autoimmune gastritis, but the absolute risks are unclear. The aim of this study was to investigate the prevalence of autoantibodies to tissue transglutaminase (TGA), thyroid peroxidase (TPOA) and gastric H+ /K+ -ATPase (ATPA) and their genetic associations in a well-characterized population-based cohort of individuals with T1D from the Bart's-Oxford family study for whom islet autoantibody prevalence data were already available. Autoantibodies in sera from 1072 patients (males/females 604/468; median age 11·8 years, median T1D duration 2·7 months) were measured by radioimmunoassays; HLA class II risk genotype was analysed in 973 (91%) using polymerase chain reaction with sequence specific primers (PCR-SSP). The prevalence of TGA (and/or history of CD), TPOA and ATPA in patients was 9·0, 9·6 and 8·2%, respectively; 3·1% had two or more autoantibodies. Females were at higher risk of multiple autoimmunity; TGA/CD were associated with younger age and TPOA with older age. ATPA were uncommon in patients under 5 years, and more common in older patients. Anti-glutamate decarboxylase autoantibodies were predictive of co-existing TPOA/ATPA. TGA/CD were associated with human leucocyte antigen (HLA) DR3-DQ2, with the DR3-DQ2/DR3-DQ2 genotype conferring the highest risk, followed by DR4-DQ8/DR4-DQ8. ATPA were associated with DR3-DQ2, DRB1*0404 (in males) and the DR3-DQ2/DR4-DQ8 genotype. TPOA were associated with the DR3-DQ2/DR3-DQ2 genotype. Almost one-quarter of patients diagnosed with T1D aged under 21 years have at least one other organ-specific autoantibody. HLA class II genetic profiling may be useful in identifying those at risk of multiple autoimmunity.

Worth the wait: type 1 diabetes prospective birth cohort studies enter adolescence.

Autoantibodies to islet cell proteins currently provide the only reliable indication that the process leading to type 1 diabetes has started. The period from the first detection of islet autoantibodies to clinical onset of diabetes can last months or years. Longitudinal birth cohort family studies give crucial information concerning the natural history of islet autoimmunity and have already shown that islet autoantibodies, which precede diabetes development, often appear in early infancy. In this issue of Diabetologia, Ziegler et al (DOI: 10.1007/s00125-012-2472-x ) and Parikka et al (DOI: 10.1007/s00125-012-2523-3 ) report findings from their birth cohort studies after numerous children have entered adolescence, allowing a more complete picture of islet autoimmunity in childhood to be revealed. Both groups are in accord that, between 6 months and 3 years of age, there is an explosion of islet autoimmunity in susceptible children and that the great majority (approximately 80%) of genetically at-risk children who present with diabetes before adolescence develop islet autoimmunity at this young age. These findings emphasise the importance of early life events in disease pathogenesis and have major implications for efforts aimed at preventing type 1 diabetes.

'Insulin autoantibody affinity measurement using a single concentration of unlabelled insulin competitor discriminates risk in relatives of patients with type 1 diabetes.

Development of high-risk combinations of multiple islet autoantibodies and type 1 diabetes is associated with high-affinity insulin autoantibodies (IAA), but IAA affinity measurements require large serum volumes. We therefore investigated whether a simplified method of IAA affinity measurement using a low concentration of unlabelled insulin (ULI) competitor discriminated between moderate-high- and low-affinity IAA and identified individuals at highest risk of disease. Samples were assayed by radiobinding microassay using high (4·0 × 10(-5) mol/l) and low (7 × 10(-9) mol/l) ULI concentrations for competitive displacement in three cohorts of IAA-positive individuals; (1) 68 patients with newly-diagnosed type 1 diabetes; (2) 40 healthy schoolchildren; and (3) 114 relatives of patients with type 1 diabetes followed prospectively for disease development (median follow-up 13 years). IAA results obtained with low ULI were expressed as a percentage of those obtained with high ULI and this was used to classify samples as low or moderate-high affinity (0-50% and >50%, respectively). Sixty-eight patient samples were positive with high and 67 (99%) with low ULI. Forty schoolchildren were IAA-positive with high and 22 (55%) with low ULI (P < 0·001). Of the relatives, 113 were positive with high and 83 (73%) with low ULI (P < 0·001). In relatives, moderate-high affinity IAA were associated with multiple islet antibodies (P < 0·001) and greater diabetes risk than low affinity IAA (P < 0·001). A single low concentration of ULI competitor can act as a surrogate for complex IAA affinity measurements and identifies those IAA-positive relatives at highest risk of disease progression.

Immunological and C-peptide studies of patients with diabetes in northern Ethiopia: existence of an unusual subgroup possibly related to malnutrition.

AIMS/HYPOTHESIS: Surveys in northern Ethiopia have demonstrated that apparent type 1 diabetes occurs more frequently than elsewhere in Africa and, indeed, in other parts of the world. We therefore investigated in detail a cohort of diabetic patients from this region to clarify the nature of this type of diabetes. METHODS: All patients attending the diabetic clinic at Mekelle Hospital in the Tigray region of northern Ethiopia were investigated over a 6 week period. Clinical, demographic and anthropometric data were collected, as well as measurements of HbA(1c), fasting lipid profile, fasting serum C-peptide and serum markers of beta cell autoimmunity, i.e. islet antigen-2 and GAD antibodies (GADA). RESULTS: Of 105 patients seen, 69 (66%) were on insulin treatment and had been from or close to diagnosis. Their median age and diabetes duration were 30 and 5 years, respectively, with a male excess of 2:1. Median BMI was 20.6 kg/m². Despite these clinical characteristics suggestive of type 1 diabetes, only 42 of 69 (61%) patients were C-peptide-negative and 35% GADA-positive. Overall, 38 (36%) of the total group (n = 105) had immunological or C-peptide characteristics inconsistent with typical type 1 or type 2 diabetes. The clinical characteristics, local prevalence of undernutrition, and GADA and C-peptide heterogeneity suggest a malnutrition-related form of diabetes. CONCLUSIONS/INTERPRETATION: Not all patients in northern Ethiopia with apparent type 1 diabetes appear to have the form of disease seen in Europids; their disease may, in fact, be related to malnutrition.

Autoantibodies to islet antigen-2 are associated with HLA-DRB1*07 and DRB1*09 haplotypes as well as DRB1*04 at onset of type 1 diabetes: the possible role of HLA-DQA in autoimmunity to IA-2.

AIMS/HYPOTHESIS: To further our understanding of antigen presentation by HLA class II molecules, we have examined the influence of HLA class II genotype on expression of autoantibodies to islet antigen-2 (IA-2A). METHODS: HLA class II genotype and IA-2A were determined within 3 months of diagnosis in 618 patients with type 1 diabetes (median age 11 years [range 0.7-20.9]). Antibodies to the juxtamembrane region of IA-2 were measured by a radiobinding assay in 481 of 484 IA-2A-positive patients. RESULTS: IA-2A prevalence was highest in patients carrying at least one HLA-DRB1*04-DQA1*0301 (385 of 450; 86%), DRB1*07-DQA1*(0201 or 0301) (58 of 64; 91%) or DRB1*09-DQA1*0301 haplotype (18 of 19; 95%). Multiple regression showed that IA-2A were strongly associated with the number of these haplotypes carried; only 69 of 132 (52%) patients carrying none of these haplotypes had IA-2A, compared with 322 of 391 (82%) patients with one and 93 of 95 (98%) with two of these haplotypes (p < 0.001). IA-2 juxtamembrane antibodies were less frequent in IA-2A-positive patients with one (35%) or two (36%) DRB1*03-DQB1*02 or DRB1*07-DQB1*02 haplotypes than in those negative for these haplotypes (52%) (p = 0.002), but showed an independent positive association with IA-2A level (p < 0.001). CONCLUSIONS/INTERPRETATION: HLA class II alleles strongly influence the prevalence of IA-2A. The high IA-2A prevalence in patients carrying DRB1*04, DRB1*07 and DRB1*09 alleles in linkage disequilibrium with DQA1*0301 or the closely related DQA1*0201 suggests the humoral response to IA-2 may be driven by HLA-DQA1 genes.

Autoantibodies to IA-2beta improve diabetes risk assessment in high-risk relatives.

AIMS/HYPOTHESIS: The aim of this study was to evaluate the prognostic significance of autoantibodies to IA-2beta (IA2betaA) in a large, well-characterised population of islet cell antibody (ICA)-positive relatives followed for 5 years in the European Nicotinamide Diabetes Intervention Trial. METHODS: Autoantibodies to insulin (IAA), glutamate decarboxylase (GADA) and IA-2 (IA2A) were measured in 549 participants at study entry, and IA2A-positive samples tested for IA2betaA. First-phase insulin response (FPIR) and oral glucose tolerance were determined at baseline. RESULTS: Of 212 ICA/IA2A-positive participants (median age 12.1 years; 57% male), 113 developed diabetes (5 year cumulative risk 56%), and 148 were also GADA-positive and IAA-positive (4Ab-positive). IA2betaA were detected in 137 (65%) ICA/IA2A-positive participants and were associated with an increased 5 year diabetes risk (IA2betaA-positive 65 vs 39% in IA2betaA-negative, p=0.0002). The effect was most marked in 4Ab-positive relatives (72% vs 52%, p=0.003). Metabolic testing further refined risk assessment. Among 101 4Ab-positive relatives with IA2betaA, the 5 year risk was 94% in those with a low FPIR (vs 50% in those with a normal FPIR, p<0.0001), and 95% in those with impaired glucose tolerance (IGT) (vs 66% in those with normal glucose tolerance, p<0.0001). The median time to diagnosis of 4Ab/IA2betaA-positive participants with a low FPIR was 1.5 years. Multivariate analysis confirmed IA2betaA status, antibody number, young age, FPIR and IGT as independent determinants of risk. CONCLUSIONS/INTERPRETATION: IA2betaA are associated with a very high risk of diabetes in ICA/IA2A-positive relatives. Testing for IA2A/IA2betaA compares favourably with the IVGTT in identifying a subgroup of autoantibody-positive relatives at increased risk. IA2betaA determination should be added to screening protocols of future intervention trials.

Magnetic resonance imaging: a reliable method for measuring pancreatic volume in Type 1 diabetes.

AIM: To validate magnetic resonance imaging (MRI) for monitoring pancreatic atrophy in Type 1 diabetes. METHODS: Twelve male patients with Type 1 diabetes of duration >or= 10 years (median age 28, range 19-32 years) and 12 healthy controls (median age 30, range 22-36 years) were invited for two abdominal MRI scans, 14 days apart. Four sequences were used: standard T1-weighted; standard T2-weighted; volumetric interpolated breath-hold examination (VIBE); and T1-weighted breath hold with fat suppression (T1BHFS). The pancreas was identified on coded images by one observer and volumes estimated by interpolation. RESULTS: Eleven patients and all controls were scanned twice. Visualization of the pancreas was best with VIBE and T1BHFS, allowing volume estimation from 47 and 46 scans, respectively. The pancreatic volume of patients estimated from these sequences were half those of controls (52.4 ml, +/- 17.1 ml, mean +/- sd) vs. (101 ml, +/- 19.5 ml, P < 0.001) and estimates showed little bias between visits; mean difference 1.1 ml (95% CI; -3.1 to 5.3 ml, P = 0.61) using VIBE and -2.6 ml (-5.8 to 0.6 ml, P = 0.03) using T1BHFS. Both sequences gave similar precision; the standard deviation of the differences in volume estimates between visits was 9.7 ml for VIBE and 7.3 ml for T1BHFS, although mean volumes estimated from T1BHFS were 4.9 ml lower (-8.2 to -1.7 ml, P = 0.005). CONCLUSIONS: Pancreatic volume can be measured reliably using MRI and shows a 48% reduction in long-standing Type 1 diabetes as compared with age-matched normal subjects. MRI should prove useful in determining the natural history of pancreatic atrophy in diabetes.

Type 1 diabetes risk assessment: improvement by follow-up measurements in young islet autoantibody-positive relatives.

AIMS/HYPOTHESIS: Combinations of autoantibody characteristics, including antibody number, titre, subclass and epitope have been shown to stratify type 1 diabetes risk in islet autoantibody-positive relatives. The aim of this study was to determine whether autoantibody characteristics change over time, the nature of such changes, and their implications for the development of diabetes. METHODS: Five-hundred and thirteen follow-up samples from 141 islet autoantibody-positive first-degree relatives were tested for islet autoantibody titre, IgG subclass, and GAD and IA-2 antibody epitope. All samples were categorised according to four risk stratification models. Relatives had a median follow-up of 6.8 years and 48 developed diabetes during follow-up. Survival analysis was used to determine the probability of change in risk category and of progression to diabetes. RESULTS: For each stratification model, the majority of relatives (71-81%) remained in the same risk category throughout follow-up. In the remainder, changes occurred both from lower to higher and from higher to lower risk categories. For all four models, relatives aged < 15 years were more likely to change risk category than those aged >15 years (0.001 < p < 0.03). Relatives whose autoantibody status changed from low- to high-risk categories had a higher risk of diabetes than relatives who remained in low-risk categories, and inclusion of autoantibody status during follow-up improved diabetes risk stratification in Cox proportional hazards models (p < 0.001). CONCLUSIONS/INTERPRETATION: Changes in islet autoantibodies are relevant to pathogenesis, and are likely to signal alterations in the disease process. Detection of changes through follow-up measurement will improve diabetes risk stratification, particularly in young individuals.

GAD antibodies in probands and their relatives in a cohort clinically selected for Type 2 diabetes.

AIMS: A subset of patients who present as if they have Type 2 diabetes have positive pancreatic autoantibodies, and have been referred to as having latent autoimmune diabetes in adults (LADA). We assessed the prevalence and clinical characteristics of patients with glutamic acid decarboxylase antibodies (GADA) in a cohort clinically selected for Type 2 diabetes and determined the presence of diabetes and GADA in their first-degree relatives. METHODS: GADA were measured in 2059 subjects, not known to be related, and clinically selected as having Type 2 diabetes for genetic studies. Clinical characteristics were compared in GADA positive and GADA negative subjects. Diabetes and GAD antibody status were compared in 208 first-degree relatives of GADA positive and GADA negative probands. RESULTS: Of the subjects, 136 (7%) were GADA positive. Compared with the GADA negative subjects, they were slimmer (P < 0.001), diagnosed at a younger age (P = 0.011) and progressed to insulin faster (P < 0.001). Thirty-three per cent of GADA positive subjects had a first-degree relative with diabetes compared with 42% of GADA negative subjects (P = 0.034). The overall prevalence of GADA was similar in the first-degree relatives of GADA positive and GADA negative probands (4 v 5%), and 19 of 22 (86%) diabetic relatives of GADA positive probands were GADA negative. CONCLUSION: Despite clinically selecting a Type 2 diabetes cohort, 7% were GADA positive with an altered phenotype. These GADA positive patients had a strong family history of non-autoimmune diabetes. This suggests that, in this subgroup of patients, autoimmune pancreatic beta-cell destruction occurs on a background of Type 2 diabetes genetic susceptibility.

The prevalence of insulin autoantibodies at the onset of Type 1 diabetes is higher in males than females during adolescence.

AIMS/HYPOTHESIS: The incidence of Type 1 diabetes shows little sex bias up to age 15 years, but more males are diagnosed in early adult life. Humoral responses to the beta cell antigen insulin could help to reveal the mechanism underlying this difference. We therefore determined the influence of sex on the prevalence of insulin autoantibodies (IAA) at diagnosis. METHODS: IAA were measured by radiobinding assay in 598 patients with newly diagnosed Type 1 diabetes (aged 10.5, range 0.8-20.7 years, 333 male), and analysed according to age, sex and HLA class II genotype. RESULTS: Overall, 74% of males and 65% of females had IAA above the 97.5(th) centile of 2860 schoolchildren ( p=0.028). IAA prevalence was similar in males and females under the age of 15 (0-4 yr, 95% vs 88%; 5-9 yr, 76% vs 73%; 10-14 yr, 67% vs 58%), but male excess was seen between 15 and 21 years (66% vs. 32%, p(corr)=0.016). HLA class II genotype was available for 426 patients. IAA prevalence in DR4 homozygous patients was 87%, in DR4 heterozygous patients 72% and in DR4 negative patients 55% ( p<0.001). Multivariate analysis showed independent association of IAA with age ( p<0.001), number of DR4 alleles ( p<0.001) and male sex ( p=0.002). CONCLUSIONS/INTERPRETATION: The prevalence of IAA in patients with newly diagnosed Type 1 diabetes is higher in males than females between 15 and 21 years of age. The lower prevalence of IAA in adolescent females implies sex-specific modulation of the autoimmune process during puberty.

Assessment of faecal occult blood loss by qualitative and quantitative methods.

BACKGROUND: Various methods exist for the assessment of faecal occult blood loss in a patient with suspected gastrointestinal blood loss. METHODS: The present study examined the effectiveness and financial implications of a qualitative guaiac-based method (Haemoccult) of faecal occult blood detection and a quantitative measure of haeme-derived porphyrins (Hemoquant) in 184 patients who underwent assessment of faecal blood loss by both methods over a three year period during assessment of iron deficiency anaemia. MAIN FINDINGS: At least one Haemoccult test was positive in 72.2% of patients while Hemoquant was suggestive of significant blood loss (> 2mg haemoglobin/g faeces) in 29.9%. Patients underwent a total of 324 further endoscopic or radiological investigations of which 76.5% demonstrated no abnormality. A diagnosis was reached in 60 patients (32.6%). A significant potential source of gastrointestinal bleeding was found in 48 patients (26.1%). Hemoquant achieved a sensitivity of 62.5% and a specificity of 81.6% while with Haemoccult it was 85.4% and 32.4%, respectively. Hemoquant was normal in 18 patients with significant gastrointestinal conditions including peptic ulcers and colonic polyps. While Haemoccult only missed 7 lesions, two of these were colonic cancers. The quantitative nature of the Hemoquant test gave little clue as to diagnosis. CONCLUSION: Neither of the tests examined was ideal but Hemoquant had an overall better performance and further investigation of patients with evidence of blood loss from this test should be mandatory.

Islet autoantibodies, nationality and gender: a multinational screening study in first-degree relatives of patients with Type I diabetes.

AIMS/HYPOTHESIS: First-degree relatives of patients with Type I (insulin-dependent) diabetes mellitus diagnosed at 20 years of age or under were screened for islet cell antibodies (ICA) in the course of recruitment to an international diabetes prevention trial. Our aim was to evaluate the influence of age, gender, proband characteristics and nationality on the prevalence of ICA and co-existence of autoantibodies to GAD, IA-2 and insulin. METHODS: A central laboratory screened samples from 10 326 non-diabetic relatives who were aged less than 40 years, from eight European countries for ICA. Antibodies to GAD and IA-2 were measured in all samples with ICA of 10 JDF units or more. RESULTS: Overall, 8.9 % of relatives had ICA of 10 JDF units or more, 3.8 % with ICA of 20 JDF units or more. Of 921 relatives with ICA of 10 JDF units or more, 29 % had co-existing antibodies to GAD or IA-2 or both. ICA of 10 JDF units or more were more prevalent in males (10.8 %) than females (7.3 %). ICA with GAD or IA-2 antibodies or both were also more common in males (3.4 %) than females (1.9 %) and in relatives under 20 years of age (3.5 % vs 1.5 %). Multiple regression analysis showed nationality to be a determinant of ICA of 10 JDF units or more but not of ICA of 20 JDF units or more or of ICA with co-existing islet antibodies, and confirmed the importance of age and gender as determinants of islet autoimmunity. CONCLUSIONS/INTERPRETATION: Relatives from different European countries have similar rates of islet autoimmunity despite wide variation in the background incidence of childhood diabetes, and male excess is equally evident in all populations. The male excess of ICA and islet autoimmunity over 10 years of age reflects the higher male incidence of Type I diabetes in this age group, and suggests that boys may be more likely than girls to develop islet autoimmunity during adolescence.