Remnant cholesterol predicts progression of diabetic nephropathy and retinopathy in type 1 diabetes.

BACKGROUND: We aimed to assess whether remnant cholesterol concentration and variability predict the progression of diabetic nephropathy (DN) and severe diabetic retinopathy (SDR) in type 1 diabetes. METHODS: This observational prospective study covered 5150 FinnDiane Study participants. Remnant cholesterol was calculated as total cholesterol - LDL cholesterol - HDL cholesterol and variability as the coefficient of variation. DN category was based on consensus albuminuria reference limits and the progression status was confirmed from medical files. SDR was defined as retinal laser treatment. For 1338 individuals, the severity of diabetic retinopathy (DR) was graded using the ETDRS classification protocol. Median (IQR) follow-up time was 8.0 (4.9-13.7) years for DN and 14.3 (10.4-16.3) for SDR. RESULTS: Remnant cholesterol (mmol L-1 ) was higher with increasing baseline DN category (P < 0.001). A difference was also seen comparing non-progressors (0.41 [0.32-0.55]) with progressors (0.55 [0.40-0.85]), P < 0.001. In a Cox regression analysis, remnant cholesterol predicted DN progression, independently of diabetes duration, sex, HbA1c , systolic blood pressure, smoking, BMI, estimated glucose disposal rate and estimated glomerular filtration rate (HR: 1.51 [1.27-1.79]). Remnant cholesterol was also higher in those who developed SDR (0.47 [0.36-0.66]) than those who did not (0.40 [0.32-0.53]), P < 0.001, and the concentration increased stepwise with increasing DR severity (P < 0.001). Regarding SDR, the HR for remnant cholesterol was 1.52 (1.26-1.83) with the most stringent adjustment. However, remnant cholesterol variability was not independently associated with the outcomes. CONCLUSIONS: Remnant cholesterol concentration, but not variability, predicts DN progression and development of SDR. However, it remains to be elucidated whether the associations are causal or not.

Genome-wide search for genes affecting the age at diagnosis of type 1 diabetes.

BACKGROUND: Type 1 diabetes (T1D) is an autoimmune disease affecting individuals in the early years of life. Although previous studies have identified genetic loci influencing T1D diagnosis age, these studies did not investigate the genome with high resolution. OBJECTIVE AND METHODS: We performed a genome-wide meta-analysis for age at diagnosis with cohorts from Finland (Finnish Diabetic Nephropathy Study), the United Kingdom (UK Genetic Resource Investigating Diabetes) and Sardinia. Through SNP associations, transcriptome-wide association analysis linked T1D diagnosis age and gene expression. RESULTS: We identified two chromosomal regions associated with T1D diagnosis age: multiple independent variants in the HLA region on chromosome 6 and a locus on chromosome 17q12. We performed gene-level association tests with transcriptome prediction models from two whole blood datasets, lymphocyte cell line, spleen, pancreas and small intestine tissues. Of the non-HLA genes, lower PNMT expression in whole blood, and higher IKZF3 and ZPBP2, and lower ORMDL3 and GSDMB transcription levels in multiple tissues were associated with lower T1D diagnosis age (FDR = 0.05). These genes lie on chr17q12 which is associated with T1D, other autoimmune diseases, and childhood asthma. Additionally, higher expression of PHF20L1, a gene not previously implicated in T1D, was associated with lower diagnosis age in lymphocytes, pancreas, and spleen. Altogether, the non-HLA associations were enriched in open chromatin in various blood cells, blood vessel tissues and foetal thymus tissue. CONCLUSION: Multiple genes on chr17q12 and PHF20L1 on chr8 were associated with T1D diagnosis age and only further studies may elucidate the role of these genes for immunity and T1D onset.

The association between bacterial infections and the risk of coronary heart disease in type 1 diabetes.

BACKGROUND: Diabetes increases the risk of infections and coronary heart disease (CHD). Whether infections increase the risk of CHD and how this applies to individuals with diabetes is unclear. OBJECTIVES: To investigate the association between bacterial infections and the risk of CHD in type 1 diabetes. METHODS: Individuals with type 1 diabetes (n = 3781) were recruited from the Finnish Diabetic Nephropathy Study (FinnDiane), a prospective follow-up study. CHD was defined as incident events: fatal or nonfatal myocardial infarction, coronary artery bypass surgery or percutaneous coronary intervention, identified through national hospital discharge register data. Infections were identified through national register data on all antibiotic purchases from outpatient care. Register data were available from 1 January 1995 to 31 December 2015. Bacterial lipopolysaccharide (LPS) activity was measured from serum samples at baseline. Data on traditional risk factors for CHD were collected during baseline and consecutive visits. RESULTS: Individuals with an incident CHD event (n = 370) had a higher mean number of antibiotic purchases per follow-up year compared to those without incident CHD (1.34 [95% CI: 1.16-1.52], versus 0.79 [0.76-0.82], P < 0.001), as well as higher levels of LPS activity (0.64 [0.60-0.67], versus 0.58 EU mL-1 [0.57-0.59], P < 0.001). In multivariable-adjusted Cox proportional hazards models, the mean number of antibiotic purchases per follow-up year was an independent risk factor for incident CHD (HR 1.21, 95% CI: 1.14-1.29, P < 0.0001). High LPS activity was a risk factor for incident CHD (HR 1.93 [1.34-2.78], P < 0.001) after adjusting for static confounders. CONCLUSION: Bacterial infections are associated with an increased risk of incident CHD in individuals with type 1 diabetes.

Associations of dietary macronutrient and fibre intake with glycaemia in individuals with Type 1 diabetes.

AIMS: To study the association between dietary intake and glycaemia in Type 1 diabetes. METHODS: Data on energy and nutrient intakes, and the mean and coefficient of variation of self-monitored blood glucose measurements were obtained from records completed by 1000 adults with Type 1 diabetes. Associations between these measures of glycaemia and dietary intake were investigated using generalized linear regression, with and without macronutrient substitution. RESULTS: In the first set of analyses, fibre intake was associated with lower mean self-monitored blood glucose values (ß = -0.428, 95% CI -0.624 to -0.231; P<0.001). In these same analyses, carbohydrate (ß = 0.011, 95% CI 0.002 to 0.020; P=0.014), alcohol (ß = 0.013, 95% CI 0.003 to 0.023; P=0.009) and monounsaturated fatty acid intakes (ß=0.012, 95% CI 0.001 to 0.023; P=0.029) were associated with higher variability in blood glucose measurements. In the macronutrient substitution analyses, substituting proteins for either carbohydrates (ß = -0.026, 95% CI -0.040 to -0.013; P<0.001), fats (ß = -0.018, 95% CI -0.033 to -0.004; P=0.014), or alcohol (ß = -0.026, 95% CI -0.045 to -0.006; P=0.010), or fats for carbohydrates (ß=-0.009, 95% CI -0.017 to -0.001; P=0.030), were all associated with lower variability in the measured blood glucose values. After adjusting for fibre intake, no significant results were observed in analyses of mean self-monitored blood glucose. CONCLUSIONS: This observational, cross-sectional study indicates that dietary fibre is associated with lower mean blood glucose concentrations in people with Type 1 diabetes. Glycaemic excursions were reduced when protein was substituted for other macronutrients and when fat replaced carbohydrate, after adjusting for fibre intake.

Association between habitual coffee consumption and metabolic syndrome in type 1 diabetes.

BACKGROUND AND AIMS: In the general population, habitual coffee consumption is inversely associated with the metabolic syndrome, a syndrome that is rather common also in patients with type 1 diabetes. However, whether coffee intake is beneficially related to the metabolic syndrome also in type 1 diabetes, is not known. We, therefore, studied the potential association between coffee consumption and the metabolic syndrome in a large population of individuals with type 1 diabetes. Furthermore, we investigated whether coffee consumption is associated with insulin resistance (estimated glucose disposal rate, eGDR), kidney function (estimated glomerular filtration rate, eGFR), and low-grade chronic inflammation (high-sensitivity C-reactive protein, hsCRP). METHODS AND RESULTS: Data from 1040 participants in the Finnish Diabetic Nephropathy Study were included in these cross-sectional analyses. Metabolic syndrome was assumed if at least 3 of the following cardiovascular risk factors were present: central obesity, high blood pressure, low HDL-cholesterol concentration, high triglyceride concentration, and hyperglycaemia. Subjects were categorized based on self-reported daily coffee intake: non-consumers (<1 cup/d), low (≥1 cups/d < 3), moderate (≥3 cups/d < 5), and high coffee consumption (≥5 cups/d). In multivariable logistic regression analysis, moderate and high coffee consumption was associated with increased odds of the metabolic syndrome. Moreover, any level of coffee consumption was associated with increased risk of the blood pressure-component. An increasing trend was observed in the eGFR with increasing coffee consumption. CONCLUSIONS: In type 1 diabetes, high coffee intake is associated with the metabolic syndrome, and especially its blood pressure-component.

Association of serum 25-hydroxyvitamin D concentration with HLA-B, -DRB1 and -DQB1 genetic polymorphisms.

BACKGROUND/OBJECTIVES: The human leukocyte antigen (HLA) gene region associates with the risk for several autoimmune diseases, including type 1 diabetes. An association between vitamin D deficiency and several autoimmune diseases has been suggested. We tested the association between serum 25-hydroxyvitamin D (25OHD) concentrations and HLA alleles in pregnant Finnish women. SUBJECTS/METHODS: HLA-B (n=395), HLA-DRB1 (n=501) and HLA-DQB1 (n=475) alleles were genotyped in pregnant women (mothers of children who later developed type 1 diabetes and mothers of non-diabetic children). HLA-B alleles were divided into supertypes that share similar peptide-binding specificity. Serum 25OHD concentration had been previously measured in these women from sera collected during the first trimester of pregnancy. Multiple testing was controlled for using the false discovery rate method. RESULTS: An association was found between 25OHD concentration and HLA-B44 supertype (P=0.009); women with HLA-B44 supertype (B*18, B*37, B*40 and B*44 alleles) had lower 25OHD concentrations. No association was found between HLA-DRB1 or -DQB1 alleles and 25OHD concentration. CONCLUSIONS: In this study we found for the first time an association between HLA genetic polymorphisms and 25OHD concentration. In future studies, the mechanistic background of this association and the role of vitamin D in the regulation of HLA gene expression should be investigated.

Patients with Type 1 diabetes consuming alcoholic spirits have an increased risk of microvascular complications.

AIMS: Our aim was to evaluate the effect of the amount of alcohol consumption and the type of beverage on the risk of diabetic nephropathy and severe diabetic retinopathy. METHODS: The alcohol consumption data were available from 3608 patients with Type 1 diabetes participating in the Finnish Diabetic Nephropathy Study (FinnDiane). We assessed the cross-sectional association between alcohol consumption and diabetic nephropathy as well as retinopathy. Patients were divided into different groups according to the amount of alcohol and the type of beverage they were consuming. RESULTS: In the multivariate analysis, the odds ratio for nephropathy was 1.39 (95% CI 1.05-1.84) for abstainers and 2.44 (95% CI 1.49-3.99) for former users compared with light consumers. The results were similar in retinopathy, with an odds ratio of 1.42 (95% CI 1.11-1.82) for abstainers and 1.73 (95% CI 1.07-2.79) for former users. No difference between light consumers and moderate or heavy consumers was observed. Compared with wine drinkers, men consuming mostly alcoholic spirits had a higher risk of nephropathy with an odds ratio of 2.80 (95% CI 1.15-6.81). In women, there was no difference in the risk of nephropathy between the different beverage types. Alcoholic spirit consumers had a higher risk of retinopathy with an odds ratio of 2.32 (95% CI 1.35-4.00). There was no difference between wine and beer consumers. CONCLUSIONS: Alcoholic spirit drinkers carry a higher risk of nephropathy and severe retinopathy compared with wine drinkers. Lifelong abstainers and former users of alcohol have a higher risk of nephropathy and severe retinopathy compared with light consumers.

Associations and interactions between lipid profiles, retinopathy and nephropathy in patients with type 1 diabetes: the FinnDiane Study.

OBJECTIVES: The aim of this study was to investigate the associations between lipid profiles and retinopathy in the large nationwide FinnDiane Study and to examine interactions and correlations between retinopathy, nephropathy and lipid variables. DESIGN AND SUBJECTS: A total of 1465 patients with type 1 diabetes, available lipid profiles, ophthalmic records and fundus photographs were included in the study. The Early Treatment of Diabetic Retinopathy Study scale was used to assess the severity of retinopathy. In an independent cohort of 1100 patients, laser treatment was used to define severe diabetic retinopathy. RESULTS: HDL cholesterol was associated with proliferative retinopathy (PDR), and triglycerides were associated with mild nonproliferative retinopathy (NPDR) independently of nephropathy and other conventional risk factors (P < 0.01). Significant interactions were seen between albumin excretion rate (AER), retinopathy status and lipid parameters (including triglycerides, non-HDL cholesterol and apolipoprotein B; P < 0.001). Highly different correlations between AER and lipid variables were observed in patients without retinopathy or with mild NPDR compared with patients with moderate to severe NPDR or PDR. Similar interactions and correlations were observed in an independent cohort stratified by laser treatment. In patients without retinopathy or with mild NPDR, AER was low despite HDL cholesterol in the lowest or triglycerides, total cholesterol or LDL cholesterol in the highest quartiles. CONCLUSIONS: Nephropathy had a strong effect on the associations between lipid variables and retinopathy, whilst dyslipidaemia was associated with nephropathy only in the presence of retinopathy. This finding suggests the existence of shared pathogenic mechanisms between retinopathy and nephropathy which could be targeted to prevent complications in patients with metabolic risk factors.

HbA1c variability is associated with an increased risk of retinopathy requiring laser treatment in type 1 diabetes.

AIMS/HYPOTHESIS: This study aimed to investigate whether variation in long-term glycaemia in type 1 diabetes as measured by HbA1c variability is associated with the cumulative incidence and risk of retinopathy requiring laser treatment. METHODS: The effect of HbA1c variability was assessed in 2,019 Finnish Diabetic Nephropathy (FinnDiane) study patients. The patients were studied in two partially overlapping subcohorts with either verified first laser treatment (n = 1,459) or retinopathy severity graded from ophthalmic records with the Early Treatment of Diabetic Retinopathy Study (ETDRS) scale (n = 1,346). The ratio of intrapersonal SD and mean of serially measured HbA1c was considered an estimate of HbA1c variability. RESULTS: A subcohort of 1,459 patients did not have laser treatment prior to the first FinnDiane visit and 174 of these patients were treated during a mean follow-up period of 5.2 ± 2.2 years. The 5 year cumulative incidence of laser treatment was 19% (95% CI 15, 24) in the highest quartile of HbA1c variability and 10% (95% CI 7, 12) in the lowest quartile (p < 0.001, Gray's test) with a corresponding HR of 1.6 (95% CI 1.1, 2.5; p = 0.02) adjusted for renal status, diabetes duration, mean HbA1c, blood pressure, sex and number of HbA1c measurements. In a subcohort of 1,346 patients, 434 patients had proliferative diabetic retinopathy (PDR). Patients in the highest quartile of HbA1c variability had an increased risk of PDR compared with the lowest quartile (HR 1.7 [95% CI 1.3, 2.2]; p < 0.001]). CONCLUSIONS/INTERPRETATION: HbA1c variability was associated with an increased cumulative incidence and risk of retinopathy requiring laser treatment in type 1 diabetes.

Hyperfiltration in type 1 diabetes: does it exist and does it matter for nephropathy?

AIMS/HYPOTHESIS: Hyperfiltration is widely regarded as a contributing factor to the development of microalbuminuria and progressive nephropathy in type 1 diabetes. However, recent studies have questioned this conclusion. METHODS: To address this conflicting evidence, we examined the association between hyperfiltration and progression to microalbuminuria in 2,318 adults with type 1 diabetes. We also compared the estimated GFR in our diabetic patients with rates observed in 6,247 adults from the Finnish general population, using age- and sex-specific z scores. RESULTS: The distribution of estimated GFR in adults with type 1 diabetes and normoalbuminuria was not significantly different from that expected in the general population (p = 0.51, Mann-Whitney test). Type 1 diabetic patients with a higher estimated GFR were also no more likely to develop microalbuminuria over a median of 5.2 years of follow-up than those with normal estimated GFR. This was the case regardless of whether hyperfiltration was defined by an absolute threshold, deciles of estimated GFR or a z score, using creatinine- or cystatin-based clearance formulas in men or in women. CONCLUSIONS/INTERPRETATION: Together with other studies, these data suggest that creatinine- or cystatin-based estimates of GFR do not predict the development of microalbuminuria in patients with type 1 diabetes. Moreover, in the absence of incipient or overt nephropathy, conventionally determined renal function in patients with type 1 diabetes appears no different from that in the general population. This is hardly surprising, given that these individuals, by all definitions, do not have kidney disease.

Serum 25-hydroxyvitamin D level during early pregnancy and type 1 diabetes risk in the offspring.

AIMS/HYPOTHESIS: Vitamin D deficiency during the fetal period or infancy is one of the suggested environmental factors for type 1 diabetes and for its increasing incidence. To test this hypothesis we compared serum 25-hydroxyvitamin D (25(OH)D) levels during early pregnancy in mothers of children who subsequently developed type 1 diabetes (case mothers) with mothers of non-diabetic healthy children (control mothers) of the same age. METHODS: Children with type 1 diabetes were identified from the nationwide prescription register. 25(OH)D concentration was measured from serum samples collected during the first trimester of pregnancy from all Finnish women (Finnish Maternity Cohort). A total of 343 case mothers and 343 control mothers were included in the study. Samples were collected throughout the year. Samples from case and control mothers were matched on the day of collection. RESULTS: Mean 25(OH)D levels in case mothers (43.9 nmol/l) and control mothers (43.7 nmol/l) were not different. Of all mothers, 481 (70.1%) were vitamin D-deficient or -insufficient. CONCLUSIONS/INTERPRETATION: No difference was found in serum 25(OH)D concentrations during first trimester of pregnancy between mothers whose children later on developed type 1 diabetes, and mothers of non-diabetic ' healthy' children of the same age. It is difficult to detect possible effects of mothers' vitamin D deficiency during early pregnancy on the development of type 1 diabetes in the offspring in this population, as such a large proportion of mothers were vitamin D-deficient or -insufficient.

Purchase of antidepressant agents by patients with type 1 diabetes is associated with increased mortality rates in women but not in men.

AIMS/HYPOTHESIS: Individuals with diabetes have increased mortality rates compared with the general population. In patients with type 2 diabetes depression further contributes to the increased mortality. Depression and mortality rates in patients with type 1 diabetes are an understudied phenomenon. We therefore studied their association in a prospective setting. METHODS: We followed 4,174 participants (51% men, age 39 ± 12 years, diabetes duration 22 ± 12 years [mean ± SD]) in the Finnish Diabetic Nephropathy Study (FinnDiane) for an average of 9 years. Depression was defined as purchase of antidepressant agents at baseline and during follow-up. These data were obtained from the Finnish Drug Prescription Register. Data on all-cause mortality and cause of death were obtained from the Finnish Cause of Death Register. RESULTS: At baseline, 313 (7.5%) patients had purchased antidepressant agents. During follow-up 758 (18.2%) additional cases were observed. Purchasers of antidepressant agents at baseline had the highest 10-year cumulative mortality rate (22.5% [95% CI 18.1, 26.6]), followed by those with such purchases during follow-up (18.0% [15.4, 20.5]) and those with no purchases (10.1% [9.0, 11.2], p < 0.001). In the adjusted Cox regression models (age, diabetes duration, diastolic blood pressure, smoking, HbA(1c) and nephropathy), the purchase of antidepressant agents at baseline was associated with mortality in women, but not in men. Cardiovascular diseases were the major cause of death in non-purchasers of antidepressant agents. In antidepressant purchasers, chronic diabetic complications were the most frequent underlying cause of death. CONCLUSIONS/INTERPRETATION: In a population of patients with type 1 diabetes, purchase of antidepressant agents was associated with increased mortality rates in women, but not in men.

Sex-related differences in the long-term risk of microvascular complications by age at onset of type 1 diabetes.

AIMS/HYPOTHESIS: This study examined sex-related differences in the cumulative risk of proliferative retinopathy (PR) and end-stage renal disease (ESRD) over 40 years of duration of type 1 diabetes according to age at diabetes onset. METHODS: We assessed 4,416 patients from the Finnish Diabetic Nephropathy Study population. Kaplan-Meier analysis was used to provide cumulative incidence rates and Cox regression analyses for HRs. RESULTS: There were no sex-related differences in the cumulative incidence of ESRD in patients diagnosed with type 1 diabetes between 0 to 4 and 5 to 9 years. Thereafter the risk started to diverge. The cumulative incidence of ESRD in patients diagnosed between 10 to 14 and ≥15 years was 17.4% (95% CI 13.4-21.2) and 13.0% (9.6-16.2) respectively in women, while in men it was 32.2% (28.0-36.1) and 24.6% (20.8-28.1) respectively. The respective HRs were (onset at 10 to 14 years) 1.9 (p < 0.0001) and (onset at ≥15 years) 1.8 (p < 0.001), respectively. There was no difference in the risk of PR between men and women diagnosed between 0 and 4 years of age, but progressive sex-related differences in the cumulative incidence of PR were observed with increasing age at onset. The HRs for men in the age-at-onset groups 5 to 9, 10 to 14 and ≥15 years of age were 1.3 (95% CI 1.0-1.6), 1.3 (1.1-1.6) and 2.1 (1.6-2.6) compared with women in these groups, respectively. CONCLUSIONS/INTERPRETATION: The difference between the sexes with regard to risk of diabetic microvascular complications is highly dependent on the age at onset of diabetes. The risk of ESRD and PR risk doubled in men compared with women when age at onset was ≥15 years.

Cumulative cost of prescription medication in outpatients with type 1 diabetes in Finland.

AIMS/HYPOTHESIS: We examined the 11-year cumulative outpatient cost of prescription medication in patients with type 1 diabetes by subgroups according to the presence of complications and duration of diabetes. METHODS: This longitudinal study included a nationally representative cohort of patients with type 1 diabetes (N = 3,717) from the Finnish Diabetic Nephropathy Study (FinnDiane). The data were linked to the Drug Prescription Register. The cumulative cost was calculated between 1998 and 2008. Information on complications was updated until 2008. Patients were divided into 10-year groups according to the duration of diabetes in 1998. Generalised linear mixed models under gamma distribution were used to evaluate the costs. RESULTS: Approximately 25% of the patients had macrovascular disease (MVD) and/or end-stage renal disease (ESRD). The adjusted cumulative cost of medications increased 56% when MVD was present compared with those without complications. In patients with ESRD or with both complications present the cost increased fourfold or 15-fold, respectively, when diabetes medications were excluded. The proportion of renal failure related medications (immunosuppressants, peritoneal dialytics and erythropoietin) accounted for more than 70% of these costs. The cost of diabetes medication was rather stable, irrespective of complication status or duration of diabetes. However, when complications were present these costs were markedly lower in all 10-year duration groups. CONCLUSIONS/INTERPRETATION: This study shows that ESRD has a great impact on outpatient prescription medication costs. Since no considerable differences were observed in the cost of diabetes medication, the increase was completely due to the cost of medications related to comorbidity.

A population-specific diabetogenic haplotype HLA-A2,Cw1,B56,DR4,DQ8 is associated with high birthweight in Finnish diabetic families.

Children with type 1 diabetes (T1D) susceptibility HLA genotypes are shown to have an increased birthweight. We investigated to what extent T1D-predisposing HLA haplotypes were associated with increased birthweight. A total of 1255 Finnish children comprising those with T1D and their non-diabetic siblings were investigated. A total of 342 children and their non-diabetic parents were HLA genotyped. Birthweight data were obtained from the national Medical Birth Registry. The population-specific diabetogenic haplotype HLA-A2,Cw1,B56,DR4,DQ8 was associated with high birthweight (P=0.0280) in families with a diabetic offspring. Other T1D-predisposing HLA haplotypes showed nonsignificant tendency with high birthweight. More infants with a birthweight >or=4000 g were born in families with a T1D offspring than in the general Finnish population (P=0.0139). The previously observed direct association between birthweight and T1D risk may be mediated through the modulating effects that T1D susceptibility HLA genes have on weight. High birthweight and subsequent weight gain may accelerate the ongoing pancreatic autoimmune process in genetically susceptible individuals. The high proportion of infants having a birthweight >or=4000 g in families with a diabetic offspring raises a concern of potential adverse health outcomes that high birthweight can have.

Early mortality in EURODIAB population-based cohorts of type 1 diabetes diagnosed in childhood since 1989.

AIMS/HYPOTHESIS: The aims of this study were to provide a contemporary picture of mortality and causes of death in Europe following a diagnosis of type 1 diabetes made before the 15th birthday, and to examine excess mortality by country for possible links to incidence level or national prosperity. METHODS: Thirteen population-based EURODIAB registers in 12 countries followed-up 28,887 children diagnosed since 1989, either by record linkage to population registers or through contact with doctors providing care. RESULTS: There were 141 deaths in the cohort during 219,061 person-years of follow-up compared with 69.1 deaths expected from national mortality rates, a standardised mortality ratio (SMR) of 2.0 (95% CI 1.7-2.4). The SMR varied from 0 to 4.7 between countries, but showed little relationship with the country's incidence rate or gross domestic product (US$ per capita). The SMR did not change significantly with attained age, calendar period or time since diagnosis. The female SMR (2.7; 95% CI 2.0-3.5) was greater than the male SMR (1.8; 95% CI 1.4-2.2), although absolute numbers of excess deaths were similar in the two sexes. One-third of deaths were classified as directly attributable to diabetes (many with mention of ketoacidosis) and half were unrelated to diabetes. There was a non-significant excess of accidental/violent deaths (48 observed vs 40.7 expected; SMR 1.2; 95% CI 0.9-1.6) but little excess in suicides (11 observed, 10.2 expected; SMR 1.1; 95% CI 0.5-1.9). CONCLUSIONS/INTERPRETATION: Before the onset of late complications, significant excess mortality existed following the diagnosis of type 1 diabetes in childhood, even in recent years. Variation between countries in this excess could not be explained.