From www.medscape.com "scientific papers are the ones that challenge the prevailing wisdom, or dogma. Here are 15 such articles from 2018, in no particular order."

 

I fulltext pdf utan lösenord

https://www.thelancet.com/journals/landia/article/PIIS2213-8587(17)30362-5/fulltext

VOLUME 6, ISSUE 2, P122-129, FEBRUARY 01, 2018

Frequency and phenotype of type 1 diabetes in the first six decades of life: a cross-sectional, genetically stratified survival analysis from UK Biobank

Abstract 

Background

Type 1 diabetes is typically considered a disease of children and young adults. Genetic susceptibility to young-onset type 1 diabetes is well defined and does not predispose to type 2 diabetes. It is not known how frequently genetic susceptibility to type 1 diabetes leads to a diagnosis of diabetes after age 30 years. We aimed to investigate the frequency and phenotype of type 1 diabetes resulting from high genetic susceptibility in the first six decades of life.

Methods

In this cross-sectional analysis, we used a type 1 diabetes genetic risk score based on 29 common variants to identify individuals of white European descent in UK Biobank in the half of the population with high or low genetic susceptibility to type 1 diabetes. We used Kaplan-Meier analysis to evaluate the number of cases of diabetes in both groups in the first six decades of life. We genetically defined type 1 diabetes as the additional cases of diabetes that occurred in the high genetic susceptibility group compared with the low genetic susceptibility group. All remaining cases were defined as type 2 diabetes. We assessed the clinical characteristics of the groups with genetically defined type 1 or type 2 diabetes.

Findings

13 250 (3·5%) of 379 511 white European individuals in UK Biobank had developed diabetes in the first six decades of life. 1286 more cases of diabetes were in the half of the population with high genetic susceptibility to type 1 diabetes than in the half of the population with low genetic susceptibility.

These genetically defined cases of type 1 diabetes were distributed across all ages of diagnosis;

537 (42%) were in individuals diagnosed when aged 31–60 years, representing 4% (537/12 233) of all diabetes cases diagnosed after age 30 years.

The clinical characteristics of the group diagnosed with type 1 diabetes when aged 31–60 years were similar to the clinical characteristics of the group diagnosed with type 1 diabetes when aged 30 years or younger.

For individuals diagnosed with diabetes when aged 31–60 years, the clinical characteristics of type 1 diabetes differed from those of type 2 diabetes: they had a lower BMI (27·4 kg/m2 [95% CI 26·7–28·0] vs 32·4 kg/m2 [32·2–32·5]; p<0·0001), were more likely to use insulin in the first year after diagnosis (89% [476/537] vs 6% [648/11 696]; p<0·0001), and were more likely to have diabetic ketoacidosis (11% [61/537] vs 0·3% [30/11 696]; p<0·0001).

Interpretation

Genetic susceptibility to type 1 diabetes results in non-obesity-related, insulin-dependent diabetes, which presents throughout the first six decades of life.

Our results highlight the difficulty of identifying type 1 diabetes after age 30 years because of the increasing background prevalence of type 2 diabetes. Failure to diagnose late-onset type 1 diabetes can have serious consequences because these patients rapidly develop insulin dependency.

From the article

Introduction

Type 1 diabetes is caused by autoimmune destruction of pancreatic β cells in genetically predisposed individuals and results in severe insulin deficiency with a requirement for treatment with insulin. It is typically considered a disease of childhood and adolescence, but can occur at any age. Type 2 diabetes is predominantly a disease of adulthood and is associated with obesity, insulin resistance, and relative but not absolute insulin deficiency. Type 2 diabetes is initially treated with lifestyle measures and oral glucose-lowering drugs.

Correct diagnosis of type 1 diabetes in young people (<20 years) is usually straightforward because it accounts for most (≥85%) cases of diabetes in that population.

 By contrast, identification of type 1 diabetes in adults older than 30 years is challenging because of the much higher prevalence of type 2 diabetes than type 1 diabetes in older adults (type 1 diabetes accounts for <5% of all cases).

Errors are often made when diagnosing type 1 diabetes later in life. For example, more than 50% of patients diagnosed with type 1 diabetes after age 35 years were shown to have type 2 diabetes in long-term follow-up.

Conversely, many older patients initially believed to have type 2 diabetes because of their age at diagnosis deteriorate rapidly and are subsequently found to have type 1 diabetes.

Few studies have investigated how frequently type 1 diabetes presents in later life.

 

Discussion

We used a novel genetically stratified survival analysis to identify cases of type 1 diabetes in a cross-sectional population of 379 511 people in UK Biobank. Genetically defined type 1 diabetes represented 9·7% (1286 of 13 250) of all diabetes cases diagnosed in the first six decades of life, with 58% (749 cases, 74% of all cases of diabetes in this age range) of type 1 diabetes cases diagnosed when patients were aged 30 years or younger and 42% (537 cases, 4% of all cases) of cases diagnosed when individuals were aged 31–60 years. The phenotype of type 1 diabetes was similar for the group diagnosed when aged 31–60 years and the group diagnosed when aged 30 years or younger; individuals in both groups had a low BMI, rapidly progressed to insulin treatment, and were at increased risk of diabetic ketoacidosis compared with participants with assumed type 2 diabetes. Our findings suggest that type 1 diabetes presents throughout the first six decades of life. Although type 1 diabetes is easy to identify and study in individuals aged 30 years or younger, identification of type 1 diabetes in individuals who present when older than 30 years is more challenging because of the much greater background incidence of type 2 diabetes in older populations.

We have shown that genetic predisposition to type 1 diabetes, defined in children, predisposes to diabetes at all ages. Our subtraction method allowed a genetic definition of disease. Although monogenic disease is usually defined by its genetic aetiology, to our knowledge, this study is the first to use a subtraction method with genetic predisposition to define cases of a common complex polygenic disease.

Numerous studies have shown that type 1 diabetes can occur in individuals older than 30 years, but researchers have been unable to quantify to what degree it occurs. Approaches to this dilemma have broadly categorised type 1 diabetes either by clinical definitions, C-peptide negativity, or autoantibody positivity. In clinical practice, further investigations are rarely done and, when they are, interpretation is challenging because of the high prevalence of type 2 diabetes in the population older than 30 years. Clinical diagnosis is often based on poorly discriminatory clinical characteristics and, as a result, misdiagnosis is common. Among individuals diagnosed with diabetes when older than 35 years, 56% of those who commence insulin treatment immediately do not progress to absolute insulin deficiency, and 7% of those not initially treated with insulin do progress to absolute insulin deficiency.

The clinical characteristics resulting from a genetic definition of type 1 diabetes can be compared with the clinical characteristics of type 1 diabetes defined by the presence of islet autoantibodies. In children and young adults with diabetes, who are typically non-obese and rapidly insulin dependent, both genetically defined type 1 diabetes and islet-antibody-positive diabetes define almost all patients. We identified a similar phenotype when genetically defined type 1 diabetes was diagnosed in individuals aged 31–60 years: 89% of patients were treated with insulin within 1 year of diagnosis and about one in nine were admitted to hospital with documented diabetic ketoacidosis. By contrast, most patients with diabetes diagnosed in middle age who have islet antibodies do not usually proceed rapidly to insulin therapy. In latent autoimmune diabetes of adults (LADA)—defined as patients diagnosed when aged 35–70 years who have antibodies to GAD and who are not initially treated with insulin—about 30–40% of patients are treated with insulin within the first 3 years after diagnosis, and the BMI, lipid profiles, and genetic susceptibility of these patients are intermediate between young-onset type 1 diabetes and type 2 diabetes.

Rather than a single intermediate phenotype, a possible explanation is that patients who are antibody positive when the prevalence of type 1 diabetes is low represent a mixture of true positives (young-onset type 1 diabetes phenotype) and false positives (type 2 diabetes). The phenotypic heterogeneity seen in LADA, depending on the population studied and the titre and number of antibodies used, supports this concept.

Further research is needed to investigate genetically defined type 1 diabetes, autoantibodies, and C-peptide within a single cohort.

Our genetically stratified survival analysis approach to identifying type 1 diabetes in a population has several advantages. First, the permanence of genetic risk from birth allows survival analysis to be done in a cross-sectional study. Second, the survival curve for type 2 diabetes in the low genetic risk score group allows use of a sensitive, but not specific, cutoff to identify almost all type 1 diabetes cases and then remove type 2 diabetes cases by subtraction. Third, a genetic definition of type 1 diabetes at a population level identifies cases without making assumptions about associated clinical features or requiring autoantibody measurement. This approach meant that we were able to assess clinical features in our study because they were not used to define type 1 diabetes. This strategy of using type 1 genetic susceptibility is independent of and complementary to methods based on clinical features, autoantibodies, or C-peptide measurement. Notably, this method could potentially be applied to other polygenic diseases with strong genetic susceptibility to provide an unbiased population assessment of their contribution to a common phenotype with multiple alternative causes.

This study has clear clinical implications. Our findings alert clinicians that type 1 diabetes occurs often after age 30 years, but that it is difficult to detect because of the predominance of type 2 diabetes in older adults. A high index of suspicion for type 1 diabetes in later life is important because it has a rapidly progressive phenotype with a substantial risk of diabetic ketoacidosis. Diagnosis of type 1 diabetes should be considered in any middle-aged patient with type 2 diabetes who does not show good glycaemic control on rapidly escalating therapy, especially if they are slim. At the individual level, a high type 1 diabetes genetic risk score (above the median of the population) cannot be used to diagnose type 1 diabetes in a patient with diabetes diagnosed in middle age because most (>90%) individuals with a high score will have type 2 diabetes. By contrast, a low genetic risk score (below the median of the population) can be used to say that a patient is unlikely to have type 1 diabetes. This genetic exclusion method is also used in existing clinical guidelines for diagnosis of coeliac disease and ankylosing spondylitis.

Calculating the probability of a type 1 diabetes diagnosis for an individual will be helped in the future by development of statistical models that incorporate quantitative clinical features, such as age of diagnosis and BMI, with the type 1 diabetes genetic risk score.

This study has several limitations. UK Biobank is a voluntary study recruiting from 22 sites across the UK; these volunteers are typically healthier and of higher socioeconomic status than the general UK population.

All participants have reached the age of 37 years, introducing a selection bias against people who are unhealthy or have not survived to this age. This selection bias might lead to an underestimate of type 1 diabetes at all ages and particularly in those diagnosed as children because they would be more likely to have serious morbidity or mortality from complications that could reduce likelihood of entry into the study. In our study, the prevalence of diabetes diagnosed in children younger than 10 years was 35 cases per 100 000 population, which is lower than expected from previous studies (around 100 cases per 100 000 population).

A further contributing factor is that UK Biobank participants were born between 1934 and 1971, and type 1 diabetes has been presenting at younger ages over time.

Underestimation of type 1 diabetes diagnosed in the youngest decade would have reduced the proportion of cases diagnosed when aged 30 years or younger in our study. However, this underestimation will not alter the phenotypic similarity in the clinical characteristics of patients diagnosed in the two age groups or the finding that presentation after age 30 years is common.

The age distribution of people in UK Biobank meant that we only had power to assess diabetes in individuals aged 60 years or younger. Therefore, we cannot predict whether the frequency of type 1 diabetes diminishes after age 60 years. We limited the study to individuals of white European descent because the type 1 diabetes genetic risk score has only been validated in white people. This approach means the results are specific to this ethnic group, and other studies are needed to validate the type 1 diabetes genetic risk score in other ethnic groups and to investigate older-onset type 1 diabetes in these populations. We used a type 1 diabetes genetic risk score developed in a reference population who were all diagnosed with type 1 diabetes when younger than 17 years.

 Although the type 1 diabetes genetic risk score has not been tested in an adult-onset type 1 diabetes population, genetic susceptibility to type 1 diabetes in later life almost completely overlaps with that of childhood diabetes.

Large studies of well defined type 1 and type 2 diabetes populations have shown that almost no overlap exists between risk of type 1 diabetes and risk of type 2 diabetes. Additionally, in the Wellcome Trust Case Control Consortium cohort, type 2 diabetes had an identical distribution of type 1 diabetes genetic risk scores to a control population without diabetes and a population with monogenic diabetes. These data support our assumption that the excess of cases in the high genetic risk score group are type 1 diabetes.

In conclusion, use of a novel genetic approach to define type 1 diabetes has shown that it presents across the first six decades of life and should not be considered a disease of children and young adults. Whatever age it presents, type 1 diabetes is associated with rapid requirement for insulin and risk of ketoacidosis, suggesting that it is not a milder phenotype if diagnosed later in life. A key area for both clinical practice and research in the future is to improve recognition of late-onset type 1 diabetes.

Nyhetsinfo

www red DiabetologNytt