What is the role of puberty in the development of islet autoimmunity and progression to type 1 diabetes?

doi:10.21203/rs.3.rs-1983795/v1

In many populations, the peak period of incidence of type 1 diabetes (T1D) has been observed to be around 10–14 years of age, coinciding with puberty, but direct evidence of the role of puberty in the development of T1D is limited. We therefore aimed to investigate whether puberty and the timing of its onset are associated with the development of islet autoimmunity (IA) and subsequent progression to T1D. A Finnish population-based cohort of children with HLA-DQB1-conferred susceptibility to T1D was followed from 7 years of age until 15 years of age or until a diagnosis of T1D (n = 6920). T1D-associated autoantibodies and growth were measured at 3- to 12-month intervals, and pubertal onset timing was assessed based on growth. The analyses used a three-state survival model. IA was defined as being either positive for islet cell antibodies plus at least one biochemical autoantibody (ICA + 1) or as being repeatedly positive for at least one biochemical autoantibody (BC1). During follow-up, 211 children (3.2% of those who were ICA + 1-defined IA-negative at 7 years) or 198 children (5.3%, BC1) developed IA, and 169 (31.9% or 25.6% of children observed to be ICA + 1 or BC1-defined IA-positive during follow-up, respectively) progressed to T1D. Puberty was associated with an increase in the risk of progression to T1D, but only from ICA + 1-defined IA (hazard ratio: 1.57; 95% confidence interval: 1.14, 2.16), and the timing of pubertal onset did not affect the association. No association between puberty and the risk of IA was detected. In conclusion, puberty may affect the risk of progression but is not a risk factor for IA.

Puberty

Type 1 diabetes

Islet autoimmunity

Cohort study

Multi-state model

Type 1 diabetes (T1D) is an immune-mediated chronic disease in which insulin-producing β cells perish. The clinical disease is preceded by a preclinical phase of islet autoimmunity in which T1D-associated autoantibodies can be detected in the blood. [1] The majority of children diagnosed with T1D test positive for multiple autoantibodies during the preclinical phase, and the risk of subsequent progression to the clinical disease increases with increasing number of detected autoantibodies [2].

Over the last few decades, the incidence of T1D in children aged <15 years has increased in many countries, although in more recent years, the incidence rates in some high-risk European countries, including Finland, have plateaued or even decreased [3,4]. However, Finland still has the highest incidence globally with a rate of at least 52.2 per 100 000 person-years during 2015–2018 [4]. While genetics plays an important role in the development of the disease, the remarkable differences in incidence rates within countries and between neighboring countries with similar genetic backgrounds support the importance of environmental factors in the disease process [1].

Puberty and/or insulin resistance have been suggested as potential factors that may play a role in the development of T1D [5]. The incidence of islet autoimmunity peaks around 1–2 years, although it can occur at any age [6,7], and a second peak of islet autoimmunity around puberty has been suggested, but the evidence is inconsistent [8,9]. Nevertheless, the incidence of T1D has been observed to reach a peak around 10–14 years, coinciding with the pubertal period in many, although not all, populations and time periods [10–12].

Insulin sensitivity decreases at the onset of puberty, is at its lowest around mid-puberty, and recovers to pre-pubertal levels once puberty has fully passed [13,14]. This decrease in insulin sensitivity may contribute to the progression of T1D, especially in people already at high risk of developing the disease [15–19], but there are indications that the role of insulin resistance as an accelerator of the disease process is minor [20]. Insulin resistance has also been linked to other phenomena, including aging, physical inactivity, and obesity [19]. No association with the risk of islet autoimmunity has been detected [21], and the direct evidence of the role of puberty in the development of T1D, including intermediate islet autoimmunity, is limited.

We therefore aimed to study whether puberty and the age at onset of it are associated with the development of islet autoimmunity and subsequent progression from islet autoimmunity to T1D in the large, population-based Finnish Type 1 Diabetes Prediction and Prevention (DIPP) cohort of young children with HLA-DQB1 -conferred susceptibility to T1D, who were followed up intensively from birth through puberty.

2.1 Participants

The data originated from the Finnish prospective population-based DIPP Study [22]. Participants in the DIPP birth cohort were born at the Turku (1994–), Oulu (1995–), and Tampere (1997–) University Hospitals. Newborns were screened for HLA-conferred genetic susceptibility to T1D using cord blood samples. Originally, children carrying the genotypes HLA-DQB1*02/*03:02 and HLA-DQB1*03:02/x (x≠*02, *03:01, or *06:02/3 until March 1997, and x≠*02, *03:01, or *06:02 thereafter) were selected for follow-up [22]. Later on, more DQB1 and DQA1 alleles and subtypes of DRB1*04 alleles were included in the screening to increase the specificity and sensitivity of the procedure [23]. Siblings of the participants were also tested for HLA-conferred susceptibility and invited for follow-up if an increased risk was detected.

The original data included all DIPP participants born before 2020 and their genetically susceptible siblings who participated in the follow-up. Data from 7 years of age were used in the analyses, and participants who had at least four height measurements were included in the assessment of pubertal onset (n_girls=3118, n_boys=3803). Of these, all who had not developed T1D before 7 years of age were included in the final analyses (n=6920, Fig. 1). Children who had already tested positive for islet autoimmunity at 7 years of age contributed only to the analysis of the progression hazards.

2.2 Background characteristics

Information on sex was collected as part of a structured questionnaire completed in the hospital after delivery. The child’s body mass index (BMI) at different ages was calculated using the available height and weight measurements obtained during follow-up. To assess BMI at 7 years of age, a polynomial spline mixed-effects model with random intercept and random slopes was fitted to BMI values by age, and each child’s BMI at 7 years was obtained from the individual BMI–age trajectory. For the analyses, the participants were classified into two categories based on the derived BMI value as follows: overweight (ISO-BMI > 25, including overweight and obese) and other (including normal weight and underweight) children, based on the Finnish ISO-BMI classification for children [24]. The ISO-BMI classification defines age-specific BMI cut-off points for underweight, overweight, and obesity for ages 2 to 18 years among Finnish children.

2.3 Follow-up characteristics

Participants were monitored for T1D-associated autoantibodies, growth, and weight up to the age of 15 years or until diagnosed with T1D. If autoantibodies were detected before the age of 15 years but the clinical disease was not diagnosed by that point, monitoring would continue. For children born until 2010, antibodies were analyzed at the ages of 3, 6, 12, 18, and 24 months and then annually at the Tampere and Oulu centers and at 3-month intervals until age of 2 years and then semiannually at the Turku center. For children born since 2010, antibodies were analyzed at the ages of 3, 6, 9, 12, and 24 months and then annually at all centers. After the possible detection of autoantibodies, the participants were followed up at 3-month intervals. Data for T1D diagnosis were obtained from the Finnish Pediatric Diabetes Registry [4] or from the study clinics.

2.4 Immunological methods

Four T1D -associated autoantibodies were analyzed, with ICAs used as the primary screening tool for children born before the end of 2002. If a child tested positive for ICA, all the child’s preceding (starting from birth) and subsequent samples were analyzed for three biochemical autoantibodies: insulin autoantibodies (IAA), glutamic acid decarboxylase antibodies (GADA), and islet antigen 2 antibodies (IA-2A). For children born from 2003 onwards, all four autoantibodies from all their samples were regularly analyzed. ICAs were quantified using a standard indirect immunofluorescence method and IAAs, GADAs, and IA-2As using specific radio-binding assays, as described previously [25]. Transplacentally transferred autoantibodies were monitored from the maternal samples and from the children’s early samples to exclude those from the analyses.

2.5 Definition of outcomes

Islet autoimmunity and clinical T1D diagnosis observed after age of 7 years were used as the outcomes of the study. Because islet autoimmunity is not an unambiguous state, we used two different definitions for it: 1) repeated positivity for ICA together with at least one biochemical autoantibody (ICA+1) and 2) repeated positivity for at least one biochemical autoantibody (BC1). As the risk of progression increases significantly with increasing numbers of detected autoantibodies [2], BC1 can be considered a broader definition for islet autoimmunity than ICA+1.

2.6 Determination of age at pubertal onset

Age at pubertal onset was assessed based on individual growth characteristics—peak height velocity (cm per year) and age at peak height velocity (years )—and on overweight status (based on ISO-BMI [24]) before the pubertal growth spurt. The assessment was carried out using a previously constructed and validated time-to-pubertal onset model [26]. The children were also divided (at the 33rd and 67th percentiles) into three equally sized groups based on age at pubertal onset. We labeled these group as the early, middle, and late pubertal onset groups; the age cutoff points for the groups were 10.0 and 11.0 years for girls and 11.3 and 11.8 years for boys.

2.7 Ethics

Parents gave their written informed consent for genetic testing of their newborn infant and for participation in the follow-up. The study adheres to the Declaration of Helsinki, and the ethics committee of Northern Ostrobothnia Hospital District approved the study protocol. Data were pseudonymized before the corresponding author accessed them for analysis.

2.8 Statistical analysis

2.8.1 The three-state survival model

Progressive three-state survival models [27] for continuous time data were used to study the association between puberty and the development of islet autoimmunity and progression to T1D. The hazards were assumed to follow a Weibull distribution. We defined state 1 as being negative for islet autoimmunity without T1D, state 2 as having islet autoimmunity without T1D, and state 3 as being diagnosed with clinical T1D. Time to islet autoimmunity was considered to be interval censored between the last negative islet autoimmunity measurement and the first positive islet autoimmunity measurement, meaning that the true transition time was unknown, although it was known to lie within a specific time interval [28]. T1D was the absorbing state—backward transition from this state was not considered possible, and entry into this state ended the follow-up. Three irreversible transitions were therefore possible: 1 → 2 (islet autoimmunity), 1 → 3 (direct transition to T1D), and 2 → 3 (progression). A few 1 → 3 transitions were considered to be unobserved 1 → 2 → 3 paths based on the rarity of the T1D diagnosis without detected autoantibodies [6]. This three-state model is illustrated in Supplementary Fig. 1.

In addition to the observed transitions, the possibility of unobserved transitions was accounted for in the models. Unobserved transitions for islet autoimmunity consisted of children under follow-up who had no measurements to assess their status at 7 years (potential islet autoimmunity at 7 years) followed by an observed islet autoimmunity (ICA+1: n=16; BC1: n=27) together with those with a T1D diagnosis but without initial observation of islet autoimmunity (n=3). The latter were also considered to be unobserved transitions from islet autoimmunity to T1D (progression).

The models were adjusted for the sex of the child (girl or boy) and overweight status (underweight/normal weight or overweight/obese) at 7 years. The potential modifying effect of the actual timing of pubertal onset (early, middle, or late) on the association between puberty and the transition-specific risks was investigated by adding the corresponding interaction term to the model. Type I errors due to multiple testing were controlled for using the false discovery rate (FDR) method [29] for the adjusted results. The FDR was set at 0.05, and a result was considered a discovery if the p-value was smaller than the corresponding FDR critical value.

2.8.2 The pubertal effect function

Puberty was added to the three-state model as a novel time-dependent function. Different origins, durations, and shapes for the effect were considered as alternative forms of pubertal influence. In the function definition, origin was defined as the initiation of the effect—whether the pubertal effect starts at the estimated onset or before it; duration was defined as the time period of the pubertal effect; and the different shapes of the pubertal effect consisted of ramping-up, flat, and fading-off periods, each accounting for different proportions of the overall duration of the effect but with equal durations for ramping-up and fading-off. The functional form of the pubertal effect is illustrated in detail in the Supplementary Methods.

We considered a range of alternative functional forms. The origin was assumed to occur at onset or 1 year before; the duration ranged from 1 to 4 years, at 1-year intervals; and the shape was expressed as a percentage of the entire duration at 10% intervals (i.e., 11 options), with ramping-up and fading-off periods being equally long. This led to close to one hundred presumed possible forms of the pubertal effect. It was assumed that the same origin, duration, and shape applied to both transitions in each model, but their regression coefficients could be different.

The pubertal effect function used was chosen in a data-driven manner based on the Akaike information criterion (AIC) by choosing the smallest average AIC over the different islet autoimmunity definitions. The final multi-state models included a pubertal effect with an origin at 1 year before the estimated onset, a duration of 3 years, and a shape comprising a 0.3-year (10%) ramping-up, a 2.4-year (80%) flat period, and a 0.3-year (10%) fading-off. We verified the performance of the data-driven model selection by implementing a small simulation study, which demonstrated probability of approximately 80% for detecting a pubertal effect as well as the ability to correctly estimate the relative risk even under a slight misspecification of the functional form (see Supplementary Methods).

2.8.3 Sensitivity analysis

We investigated the sensitivity of the results in several ways. The possibility of an error in the growth-based estimation of pubertal onset was evaluated by fitting the same model to ten datasets, with the timing of onset randomly drawn from a normal distribution corresponding to the individual’s pubertal onset prediction interval obtained from the time-to-pubertal onset model [26]. Results from the models were combined using Rubin’s Rules [30].

To verify that the significant findings were not specific to the choice of a particular functional form, we tested a range of models with different origins, durations, and/or shapes and compared them to the results using the chosen functional form.

Finally, additional analyses defining islet autoimmunity as repeated positivity for at least two biochemical autoantibodies (BC2) were performed to determine whether the definition of islet autoimmunity affected the results.

Of the 6920 participants, 3117 (45.0%) were girls and 1572 (22.7%) were overweight or obese at the age of 7 years. The mean estimated age at pubertal onset was 10.6 (sd: 0.98) years for girls and 11.6 (sd: 0.62) years for boys.

Depending on the islet autoimmunity definition, either 530 (ICA + 1, 7.7%) or 660 (BC1, 9.5%) participants were observed to be positive for islet autoimmunity either at the start (at 7 years) or during the follow-up. Of these, either 211 (ICA + 1) or 198 (BC1) developed islet autoimmunity during the follow-up period (Table 1). Among all the participants, 172 (2.5%) developed T1D during the follow-up, at a median age of 11.3 (interquartile range: 9.3–13.3) years, 169 of whom were previously positive for islet autoimmunity (Table 1). Neither the sex of the child nor overweight status at 7 years of age were found to be associated with the risk of islet autoimmunity or the risk of progression to T1D (Table 2).

Despite trending in the direction of elevated risk, puberty was not significantly associated with the risk of islet autoimmunity, but it was associated with the risk of progression from ICA + 1-defined islet autoimmunity to T1D. The hazard for a pubertal child was 1.57 (95% CI: 1.14, 2.16) times that of a non-pubertal child (Table 3, Supplementary Fig. 2), and this association is illustrated in Fig. 2, which shows that at a given age, the currently pubertal children tended to have higher progression probabilities than those who were pre- or post-pubertal. No association between puberty and the risk of progression was detected with the BC1 definition.

The timing of pubertal onset did not affect the association between puberty and the risk of progression (Table 3). All the results remained similar in the sensitivity analyses after considering uncertainty in the estimation of puberty onset (data not shown). The statistically significant finding of an association between puberty and ICA + 1-defined progression to T1D was consistent under different forms of pubertal effect functions, which did not markedly weaken the model fit (AIC; Supplementary Table 1). The results with the additional BC2 islet autoimmunity definition were similar to the results using the BC1 definition (data not shown).

We investigated the role of puberty in the development of islet autoimmunity and progression to T1D. No association was found between puberty and the risk of islet autoimmunity, but puberty was associated with the risk of progression to T1D using the ICA + 1 definition for islet autoimmunity. The hazard ratio of progression to T1D from BC1-defined islet autoimmunity was in the same direction, but did not reach statistical significance. The timing of pubertal onset did not modify the association.

This study is among the first assessments of the role of puberty in T1D development. Insulin resistance is a normal feature of puberty, and previous studies have implied that it may contribute to the progression of T1D [15–19]. From that perspective, the potential association found between puberty and the higher risk of progression to T1D in the present study is consistent with the existing literature. In addition, the production of growth hormone and insulin-like growth factor-1 increases during puberty, and both have been linked to insulin resistance during puberty [31, 32], pointing to the possibility that they may contribute to the association found.

Based on a large Biobank study from the UK, there are associations between the timing of puberty as either early or late and higher risks of numerous adverse outcomes, including gynecological, cardio-metabolic, gastrointestinal, musculoskeletal, and neuro-cognitive, as well as a range of cancers in both men and women [33]. Since pubertal timing seems to have a broad influence on later health, we wanted to assess the effect of the timing on the association between puberty and the risk of T1D progression. However, our analyses did not find any notable differences in the particular association between the pubertal timing groups of early, middle and late.

The results were somewhat sensitive to the specification of the pubertal effect function. The AIC suggested models in which an association between puberty and progression was present, and we demonstrated through sensitivity analyses that a misspecification of the effect could lead to a loss of significance, although not a change in the direction of the effect. This underlines the need to model this effect carefully. The identified duration for the pubertal effect closely agreed with the previously observed durations of puberty, which were 4.1 years for girls and 3.8 years for boys [34].

The major strength of the current study is the large and unique longitudinal cohort with its long follow-up for autoantibodies, growth, and T1D. The use of the multi-state model enabled a state representation that is well suited to the T1D disease process because autoantibodies usually precede the clinical disease. Modeling enabled the risk factors for progression to be studied while considering the child’s preceding states, and the flexible inclusion of the pubertal effect as a function for the multi-state model enabled a comparison of different potential forms of the effect.

Limitations of the study include the lack of systematic biochemical autoantibody (IAA, GADA, and IA-2A) measurements. Among children born before 2003, ICA was used as a primary screening tool, meaning that biochemical antibodies were detected only if ICA was positive. This may have affected the results regarding the biochemical endpoint because some autoantibody-positive measurements may have been missed. However, since a positive test for biochemical autoantibodies without the appearance of ICA is uncommon [6, 35]^, we believe that it is unlikely that this would have had a major effect on the results. In addition, the previous simulation study in the same data showed that the effect of selective sampling on the regression coefficients for the secondary autoantibody outcomes is minor [36]. Antigen specificity of ICA is not known, and it is of interest that sensitivity analysis with a BC2 definition gave similar results to those of BC1, suggesting possible contribution of other autoantibodies than IAA, GADA, or IA-2A in the ICA + BC1 group, e.g., zinc transporter 8 autoantibodies. Another limitation is that the study did not involve an actual pubertal follow-up, and the pubertal onset timings were estimated on the basis of a validated growth assessment method [26]. The generalizability of the results might also be limited because the current study was conducted in subjects with increased HLA-conferred genetic risk for T1D.

In conclusion, this study suggests that puberty could have an accelerating role in progression to T1D but has no role in the development of islet autoimmunity. The study is novel in its investigation of the direct role of puberty—not via insulin sensitivity—in the development of T1D, together with the ability to model disease progression and puberty together. Based on our results, the previously observed peak in T1D incidence during adolescence [1] could be associated with pubertal development.

Funding:

The study was supported by the Academy of Finland (grants 63672, 68292, 79685, 79686, 80846, 114666, 126813, 129492, 139391, 201988, 210632, 276475, 307996, 308065, 308066, and Center of Excellence in Molecular Systems Immunology and Physiology Research 2012-2017, Decision No. 250114), URL: https://www.aka.fi/en/; European Foundation for the Study of Diabetes (EFSD) award supported by EFSD/JDRF/Lilly, URL: http://www.europeandiabetesfoundation.org/; the Finnish Diabetes Association, URL: https://www.diabetes.fi/en/finnish_diabetes_association; the Finnish Diabetes Research Foundation, URL: https://www.diabetestutkimus.fi/en; the Juho Vainio Foundation, URL: https://juhovainionsaatio.fi/en/juho-vainio-foundation/; the Juvenile Diabetes Research Foundation International (grants 4-1998-274, 4-1999-731, 4-2001-435, 1-SRA-2016-342-M-R, 1-SRA-2019-732-M-B, and 3-SRA-2020-955-S-B), URL: https://www.jdrf.org/; the Competitive State Research Financing of the Expert Responsibility area of Tampere University Hospital (grants 9E082, 9F089, 9G087, 9H092, 9J147, 9K149, 9L042, 9L117, 9M114, 9N086, 9P057, 9R055, 9S074, 9U065, and 9V072); Oulu University Hospital Research Funds; Turku University Hospital (state research funding VTR); the European Comission (grant BMH4-CT98-3314), URL: https://ec.europa.eu/info/index_en; the Novo Nordisk Foundation, URL: https://novonordiskfonden.dk/en/; Special Research Funds for University Hospitals in Finland; the Sigrid Jusélius Foundation, URL: https://www.sigridjuselius.fi/en/; and author E.J.P. has received research support from the Alfred Kordelin Foundation (200356), URL: https://kordelin.fi/en/frontpage.

The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have no relevant financial or non-financial interests to disclose.

CRediT authorship contribution statement:

Essi J. Peltonen: Methodology, Validation, Formal analysis, Writing - Original Draft, Visualization. Riitta Veijola: Conceptualization, Investigation, Data Curation, Supervision, Writing - Review & Editing, Funding acquisition. Jorma Ilonen: Investigation, Writing - Review & Editing. Mikael Knip: Conceptualization, Investigation, Data Curation, Writing - Review & Editing, Funding acquisition. Harri Niinikoski: Conceptualization, Investigation, Data Curation, Writing - Review & Editing, Funding acquisition. Jorma Toppari: Conceptualization, Investigation, Data Curation, Writing - Review & Editing, Funding acquisition. Helena E. Virtanen: Conceptualization, Writing - Review & Editing, Funding acquisition. Suvi M. Virtanen: Conceptualization, Investigation, Supervision, Writing - Review & Editing, Funding acquisition. Jaakko Peltonen: Conceptualization, Methodology, Supervision, Writing - Review & Editing, Visualization. Jaakko Nevalainen: Conceptualization, Methodology, Supervision, Writing - Review & Editing, Visualization.

Ethics approval: The study adheres to the Declaration of Helsinki, and the ethics committee of Northern Ostrobothnia Hospital District has approved the study protocol. Data were pseudonymized before the corresponding author accessed them for the analyses.

Consent to participate: Parents gave their written informed consent for genetic testing of their newborn infant and for participation in the follow-up.

Norris JM, Johnson RK, Stene LC. Type 1 diabetes-early life origins and changing epidemiology. Lancet Diabetes Endocrinol 2020;8:226–38. https://doi.org/https://doi.org/10.1016/S2213-8587(19)30412-7.
Ziegler A-G, Rewers M, Simell O, Simell T, Lempainen J, Steck A, et al. Seroconversion to multiple islet autoantibodies and risk of progression to diabetes in children. JAMA 2013;309:2473–9. https://doi.org/https://doi.org/10.1001/jama.2013.6285.
Patterson CC, Harjutsalo V, Rosenbauer J, Neu A, Cinek O, Skrivarhaug T, et al. Trends and cyclical variation in the incidence of childhood type 1 diabetes in 26 European centres in the 25 year period 1989–2013: a multicentre prospective registration study. Diabetologia 2019;62:408–17. https://doi.org/https://doi.org/10.1007/s00125-018-4763-3.
Parviainen A, But A, Siljander H, Knip M, The Finnish Pediatric Diabetes Register. Decreased Incidence of Type 1 Diabetes in Young Finnish Children. Diabetes Care 2020;43:2953–8. https://doi.org/https://doi.org/10.2337/dc20-0604.
Craig ME, Kim KW, Isaacs SR, Penno MA, Hamilton-Williams EE, Couper JJ, et al. Early-life factors contributing to type 1 diabetes. Diabetologia 2019;62:1823–34. https://doi.org/https://doi.org/10.1007/s00125-019-4942-x.
Parikka V, Näntö-Salonen K, Saarinen M, Simell T, Ilonen J, Hyöty H, et al. Early seroconversion and rapidly increasing autoantibody concentrations predict prepubertal manifestation of type 1 diabetes in children at genetic risk. Diabetologia 2012;55:1926–36. https://doi.org/https://doi.org/10.1007/s00125-012-2523-3.
Ziegler A-G, Bonifacio E, the BABYDIAB-BABYDIET Study Group. Age-related islet autoantibody incidence in offspring of patients with type 1 diabetes. Diabetologia 2012;55:1937–43. https://doi.org/https://doi.org/10.1007/s00125-012-2472-x.
Ziegler A-G, Nepom GT. Prediction and pathogenesis in type 1 diabetes. Immunity 2010;32:468–78. https://doi.org/https://doi.org/10.1016/j.immuni.2010.03.018.
Chiu SW-Y, Leslie RDG, Beyan H. Non-genetic factors in the pathogenesis of type 1 diabetes. In: Taylor K, Hyöty H, Toniolo A, Zuckerman AJ, editors. Diabetes and Viruses, New York: Springer; 2013, p. 25–34. https://doi.org/https://doi.org/10.1007/978-1-4614-4051-2_4.
Karvonen M, Viik-Kajander M, Moltchanova E, Libman I, LaPorte R, Tuomilehto J. Incidence of childhood type 1 diabetes worldwide. Diabetes Mondiale (DiaMond) Project Group. Diabetes Care 2000;23:1516–26. https://doi.org/https://doi.org/10.2337/diacare.23.10.1516.
Onda Y, Sugihara S, Ogata T, Yokoya S, Yokoyama T, Tajima N, et al. Incidence and prevalence of childhood‐onset Type 1 diabetes in Japan: the T1D study. Diabet Med 2017;34:909–15. https://doi.org/https://doi.org/10.1111/dme.13295.
Weng J, Zhou Z, Guo L, Zhu D, Ji L, Luo X, et al. Incidence of type 1 diabetes in China, 2010-13: population based study. BMJ 2018;360:j5295. https://doi.org/https://doi.org/10.1136/bmj.j5295.
Moran A, Jacobs DR, Steinberger J, Hong C-P, Prineas R, Luepker R, et al. Insulin resistance during puberty: results from clamp studies in 357 children. Diabetes 1999;48:2039–44. https://doi.org/https://doi.org/10.2337/diabetes.48.10.2039.
Kelsey MM, Zeitler PS. Insulin resistance of puberty. Curr Diab Rep 2016;16:64. https://doi.org/https://doi.org/10.1007/s11892-016-0751-5.
Fourlanos S, Narendran P, Byrnes GB, Colman PG, Harrison LC. Insulin resistance is a risk factor for progression to type 1 diabetes. Diabetologia 2004;47:1661–7. https://doi.org/https://doi.org/10.1007/s00125-004-1507-3.
Mrena S, Virtanen SM, Laippala P, Kulmala P, Hannila M-L, Åkerblom HK, et al. Models for predicting type 1 diabetes in siblings of affected children. Diabetes Care 2006;29:662–7. https://doi.org/https://doi.org/10.2337/diacare.29.03.06.dc05-0774.
Xu P, Cuthbertson D, Greenbaum C, Palmer JP, Krischer JP. Role of insulin resistance in predicting progression to type 1 diabetes. Diabetes Care 2007;30:2314–20. https://doi.org/https://doi.org/10.2337/dc06-2389.
Bingley PJ, Mahon JL, Gale EAM. Insulin resistance and progression to type 1 diabetes in the European Nicotinamide Diabetes Intervention Trial (ENDIT). Diabetes Care 2008;31:146–50. https://doi.org/https://doi.org/10.2337/dc07-0103.
Pang TTL, Narendran P. Addressing insulin resistance in Type 1 diabetes. Diabet Med 2008;25:1015–24. https://doi.org/https://doi.org/10.1111/j.1464-5491.2008.02493.x.
Siljander HT, Hermann R, Hekkala A, Lähde J, Tanner L, Keskinen P, et al. Insulin secretion and sensitivity in the prediction of type 1 diabetes in children with advanced β-cell autoimmunity. Eur J Endocrinol 2013;169:479–85. https://doi.org/https://doi.org/10.1530/EJE-13-0206.
Winkler C, Marienfeld S, Zwilling M, Bonifacio E, Ziegler A-G. Is islet autoimmunity related to insulin sensitivity or body weight in children of parents with type 1 diabetes? Diabetologia 2009;52:2072–8. https://doi.org/https://doi.org/10.1007/s00125-009-1461-1.
Kupila A, Muona P, Simell T, Arvilommi P, Savolainen H, Hmlinen A-M, et al. Feasibility of genetic and immunological prediction of type I diabetes in a population-based birth cohort. Diabetologia 2001;44:290–7. https://doi.org/https://doi.org/10.1007/s001250051616.
Ilonen J, Hammais A, Laine A-P, Lempainen J, Vaarala O, Veijola R, et al. Patterns of β-Cell Autoantibody Appearance and Genetic Associations During the First Years of Life. Diabetes 2013;62:3636–40. https://doi.org/https://doi.org/10.2337/db13-0300.
Saari A, Sankilampi U, Hannila M-L, Kiviniemi V, Kesseli K, Dunkel L. New Finnish growth references for children and adolescents aged 0 to 20 years: length/height-for-age, weight-for-length/height, and body mass index-for-age. Ann Med 2011;43:235–48. https://doi.org/https://doi.org/10.3109/07853890.2010.515603.
Virtanen SM, Kenward MG, Erkkola M, Kautiainen S, Kronberg-Kippilä C, Hakulinen T, et al. Age at introduction of new foods and advanced beta cell autoimmunity in young children with HLA-conferred susceptibility to type 1 diabetes. Diabetologia 2006;49:1512–21. https://doi.org/https://doi.org/10.1007/s00125-006-0236-1.
Syrjälä E, Niinikoski H, Virtanen HE, Ilonen J, Knip M, Hutri-Kähönen N, et al. Determining the timing of pubertal onset via a multicohort analysis of growth. PLoS One 2021;16:e0260137. https://doi.org/https://doi.org/10.1371/journal.pone.0260137.
Van Den Hout A. Multi-state survival models for interval-censored data. CRC Press; 2016.
Collett D. Modelling survival data in medical research. 3rd ed. Bristol, UK: CRC Press; 2015.
Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J R Stat Soc Ser B 1995;57:289–300. https://doi.org/https://doi.org/10.1111/j.2517-6161.1995.tb02031.x.
Rubin DB. Multiple imputation for nonresponse in surveys. vol. 81. John Wiley & Sons; 2004.
Moran A, Jacobs Jr DR, Steinberger J, Cohen P, Hong C-P, Prineas R, et al. Association between the insulin resistance of puberty and the insulin-like growth factor-I/growth hormone axis. J Clin Endocrinol Metab 2002;87:4817–20. https://doi.org/https://doi.org/10.1210/jc.2002-020517.
Vijayakumar A, Yakar S, LeRoith D. The intricate role of growth hormone in metabolism. Front Endocrinol (Lausanne) 2011;2:32. https://doi.org/https://doi.org/10.3389/fendo.2011.00032.
Day FR, Elks CE, Murray A, Ong KK, Perry JRB. Puberty timing associated with diabetes, cardiovascular disease and also diverse health outcomes in men and women: the UK Biobank study. Sci Rep 2015;5:1–12. https://doi.org/https://doi.org/10.1038/srep11208.
Cheng HL, Harris SR, Sritharan M, Behan MJ, Medlow SD, Steinbeck KS. The tempo of puberty and its relationship to adolescent health and well-being: A systematic review. Acta Paediatr 2020;109:900–13. https://doi.org/https://doi.org/10.1111/apa.15092.
Pöllänen PM, Lempainen J, Laine A-P, Toppari J, Veijola R, Vähäsalo P, et al. Characterisation of rapid progressors to type 1 diabetes among children with HLA-conferred disease susceptibility. Diabetologia 2017;60:1284–93. https://doi.org/https://doi.org/10.1007/s00125-017-4258-7.
Nevalainen J, Datta S, Toppari J, Ilonen J, Hyöty H, Veijola R, et al. Frailty modeling under a selective sampling protocol: an application to type 1 diabetes related autoantibodies. Stat Med 2021;40:6410–20. https://doi.org/https://doi.org/10.1002/sim.9190.

Table 1

Distribution of type 1 diabetes development status at 7 years of age (the beginning of follow-up) for different outcomes and the numbers and proportions of transitions
	Status at 7 years of age, n (% of N = 6920)				Number of transitions^a
	No IA	IA	Potential IA^b	Potential IA^b followed by IA	IA^c	Progression^d
ICA + 1	6599 (95.4%)	303 (4.4%)	2 (0.03%)	16 (0.2%)	211 (3.2%)	169 (31.9%)
BC1	3724 (53.8%)	435 (6.3%)	2734 (39.5%)	27 (0.4%)	198 (5.3%)	169 (25.6%)
IA: islet autoimmunity; T1D: type 1 diabetes; Progression: progression from IA to T1D
^a In addition to observed transitions, the possibility of unobserved transitions was accounted for in the model.
^b Children being followed up (T1D-free) who had no measurements to assess IA status at 7 years of age.
^c Observed transitions with the proportions of children observed to be IA-negative at 7 years of age.
^d Observed transitions with the proportions of children observed to be IA-positive during follow-up (at 7 years of age or with transition during follow-up).

Table 2

Distribution of background characteristics and their association with the risk of islet autoimmunity and with the risk of progression to type 1 diabetes for different outcomes; numbers include both unobserved and observed transitions
		Sex					Overweight status at 7y
		Girls^a n = 3117		Boys n = 3803			Underweight/ normal weight^a n = 5348		Overweight/ obese n = 1572
	Transition	n (%)	Total N	n (%)	Total N	HR (95% CI)	n (%)	Total N	n (%)	Total N	HR (95% CI)
ICA + 1	IA^b	92 (3.1)	2994	138 (3.8)	3623	1.26 (0.95, 1.67)	169 (3.3)	5102	61 (4.0)	1515	1.18 (0.87, 1.61)
ICA + 1	Progression^c	76 (35.3)	215	96 (30.2)	318	0.80 (0.59, 1.09)	142 (34.2)	415	30 (25.4)	118	0.82 (0.56, 1.22)
BC1	IA^b	92 (3.1)	2935	136 (3.8)	3550	1.29 (0.97, 1.71)	172 (3.4)	5002	56 (3.8)	1483	1.10 (0.80, 1.52)
BC1	Progression^c	76 (27.7)	274	96 (24.7)	389	0.99 (0.73, 1.35)	142 (27.4)	518	30 (20.7)	145	0.74 (0.49, 1.10)
IA: islet autoimmunity; T1D: type 1 diabetes; Progression: progression from IA to T1D
^a Reference group.
^b Total N = those not observed to be IA-positive at 7 years of age; n = those who developed IA or were diagnosed with T1D with unobserved IA.
^c Total N = those observed to be IA-positive during follow-up (at 7 years of age or with transition during follow-up) or diagnosed with T1D with unobserved IA; n = those diagnosed with T1D.

Table 3

Number of transitions for different outcomes and hazard ratios for, both overall and by timing of pubertal onset; numbers include both unobserved and observed transitions
	Transition	Timing of pubertal onset	n (%)^a	Unadjusted model HR (95% CI)	P-value^b	Adjusted model^c HR (95% CI)
ICA + 1	IA	Early	80 (3.7)	1.33 (0.87, 2.03)	0.542
		Middle	87 (3.7)	1.48 (0.94, 2.32)
		Late	63 (3.0)	0.88 (0.49, 1.57)
		Overall	230 (3.5)	1.30 (0.94, 1.81)		1.25 (0.90, 1.74)
	Progression	Early	47 (27.6)	1.41 (0.88, 2.26)	0.659
		Middle	83 (37.1)	1.53 (0.95, 2.47)
		Late	42 (29.0)	1.96 (1.24, 3.11)
		Overall	172 (32.3)	1.63 (1.19, 2.24)		1.57 (1.14, 2.16)^d
BC1	IA	Early	68 (3.2)	1.05 (0.64, 1.70)	0.296
		Middle	98 (4.3)	1.72 (1.10, 2.68)
		Late	62 (3.0)	0.76 (0.40, 1.47)
		Overall	228 (3.5)	1.18 (0.83, 1.68)		1.24 (0.87, 1.77)
	Progression	Early	47 (22.2)	1.23 (0.77, 1.98)	0.954
		Middle	83 (30.0)	1.28 (0.79, 2.08)
		Late	42 (23.3)	1.37 (0.86, 2.19)
		Overall	172 (25.9)	1.30 (0.95, 1.79)		1.31 (0.95, 1.80)
IA: islet autoimmunity; T1D: type 1 diabetes; Progression: progression from IA to T1D
^a Proportion of (for IA) those not observed to be IA-positive at 7 years of age or (for progression) those observed to be IA-positive during follow-up (at 7 years of age or with transition during follow-up) or those diagnosed with T1D with unobserved IA.
^b Wald test for equality of timing-specific hazard ratios.
^c Adjusted for sex and overweight status at 7 years of age.
^d Statistically significant after correction for multiple testing, with p-value of 0.0056 and FDR critical value of 0.0063.

What is the role of puberty in the development of islet autoimmunity and progression to type 1 diabetes?

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Material and Methods

3. Results

4. Discussion

Declarations

References

Tables

Supplementary Files

Status:

Journal Publication

Version 1