High prevalence of depressive symptoms among people with pediatric‐onset and adolescent‐onset type 1 diabetes: A cross‐sectional analysis of the Diabetes Study from the Center of Tokyo Women's Medical University

ABSTRACT Aims/Introduction To investigate the prevalence of depressive symptoms by the age of onset of type 1 diabetes, and its association with the condition of individuals with pediatric‐ and adolescent‐onset type 1 diabetes. Materials and Methods This single‐center cross‐sectional study enrolled Japanese participants with type 1 diabetes. All participants completed a questionnaire about their diabetes‐related condition and the Patient Health Questionnaire‐9, which was used to evaluate depression. Individuals with a Patient Health Questionnaire‐9 score of ≥10 points were defined as having moderate depressive symptoms. Results A total of 1,267 participants (mean age 40 years; mean duration of type 1 diabetes 21 years; 68% female; mean glycated hemoglobin 7.8%) were included and classified according to the age of onset of type 1 diabetes to identify the proportion of moderate depressive symptoms in each group: 21% (0–12 years), 18% (13–19 years) and 13% (20–40 years). The prevalence of moderate depressive symptoms was significantly higher among participants with pediatric‐onset type 1 diabetes (P < 0.05). Moderate depressive symptoms were associated with increased glycated hemoglobin, neuropathy and hypoglycemia unawareness. Conclusions Regular screening for depressive symptoms and hypoglycemia awareness is important. Healthcare professionals should provide appropriate psychosocial care for people with pediatric‐onset and adolescent‐onset type 1 diabetes from childhood through to adulthood.


INTRODUCTION
Type 1 diabetes is a chronic disease that frequently manifests during childhood and requires daily intensive self-management. For a smooth transition from the pediatric to adult medical care departments, patients with type 1 diabetes must be mentally independent and psychosocially mature 1 . However, 15% of children aged 13-19 years report a mental burden associated with type 1 diabetes 2,3 . Compared with their counterparts without diabetes, adolescent children with type 1 diabetes are more likely to have mental health problems, such as depression, eating disorders and adjustment disorders [4][5][6] . The SEARCH for Diabetes in Youth Study carried out in the USA showed that 14% of young people with type 1 diabetes had mild depression, whereas 8.6% had moderate-to-severe depression 7 . Therefore, we hypothesized that individuals with pediatric-onset type 1 diabetes are more likely to develop depressive symptoms than those with adult-onset type 1 diabetes.
The complications of depression negatively affect self-care behaviors, such as compliance with insulin administration and self-measurement of blood glucose levels 4,8 , which result in poor glycemic control, diabetic ketoacidosis 9 , chronic complications in early adulthood and deterioration of the quality of life [10][11][12] . Therefore, psychiatric problems in childhood and adolescence often hinder smooth medical care transition, and these problems continue even after the transfer to the adult care department. It is, therefore, important to evaluate the long-term perspective of the lives of individuals with type 1 diabetes.
In a recent cross-sectional analysis of the Diabetes Study from the Center of Tokyo Women's Medical University (DIA-CET), we reported that 24.7% and 14.5% of participants with type 1 diabetes had mild and moderate-to-severe depression, respectively 13 . However, we did not study the prevalence of depressive symptoms with regard to the age of onset of type 1 diabetes. Therefore, in the present study, we aimed to investigate the impact of age of onset of type 1 diabetes on the prevalence of depressive symptoms to highlight the burden of depressive symptoms in pediatric and adolescent-onset type 1 diabetes in adulthood, and clarify the association between depressive symptoms and diabetes-related complications. The insights from this analysis are expected to provide support for necessary care, including the definition of a transition period, to ameliorate depression in individuals with type 1 diabetes.

Participants
This single-center cross-sectional study enrolled Japanese participants who had type 1 diabetes with an age of onset of ≤40 years, were hospitalized at or regularly visited the Diabetes Center of Tokyo Women's Medical University, had participated in the DIACET survey initiated in October 2012 and had complete data for all nine items of the Patient Health Questionnaire-9 (PHQ-9).

DIACET and PHQ-9
The protocol of the DIACET study has been described previously 14 . The PHQ-9 is a self-administered questionnaire on health status, including glycemic control, medications and diabetes-related complications (e.g., presence of diabetic retinopathy, hemodialysis treatment, diabetic neuropathy and hypoglycemia unawareness). Neuropathy was diagnosed from the presence of numbness or pain in the hands and feet. Furthermore, the use of oral antihypertensive and therapeutic agents for dyslipidemia was investigated. The PHQ-9 questionnaires were distributed to eligible participants during regular outpatient visits or during hospitalizations between October 2012 and February 2013, and the completed PHQ-9 forms were returned by postal mail or collected during the subsequent outpatient visit or hospitalizations. The PHQ-9 assesses the presence, frequency and severity of depressive symptoms 15 . For each item, respondents are instructed to choose one of the following scores according to the frequency of occurrence within a 2-week period: 0, not at all; 1, several days; 2, more than half of the days; or 3, nearly every day. Based on the total score of the responses to all questions (0-27), participants were categorized into three groups according to the severity of depression as: no (0-4), mild [5][6][7][8][9] or moderate depressive symptoms (≥10). The Japanese version of the PHQ-9 showed adequate validity and reliability in the evaluation by Muramatsu et al. 16 Data on glycated hemoglobin (HbA1c) levels were obtained for the 90 days before and after the distribution of the questionnaire, and were measured using high-performance liquid chromatography. Hypoglycemia unawareness was defined as hypoglycemia requiring another individual's assistance or a lack of symptoms when blood glucose was <70 mg/dL.

Statistical analysis
Continuous data are expressed as the arithmetic meanstandard deviation, and categorical data are expressed as the frequency with percentage. Among the three groups classified according to the age of onset, the Jonckheere-Terpstra test was used to compare continuous data, whereas the Cochran-Armitage test was used for binary data. Data from the two groups that were classified by the severity of depression according to the PHQ-9 scores (<10 and ≥10) underwent one-way ANOVA testing to compare continuous data, and the v 2 -test to compare binary data. For multivariate analysis using JMP 15 (SAS Institute, Cary, NC, USA), the following candidate variables were used to identify the risk of moderate depressive symptoms: age, duration of diabetes, sex, HbA1c, hypoglycemia unawareness, retinopathy, neuropathy and hemodialysis.

RESULTS
In the present study, we included 1,279 people who had been diagnosed with incident type 1 diabetes at ≤40 years-of-age. In this cohort, the mean age was 40 -13 years, 68% of the participants were female and the mean duration of type 1 diabetes was 21 -11 years. When classified by the age of onset, 413, 259 and 607 participants had developed type 1 diabetes in childhood (0-12 years, pediatric-onset type 1 diabetes), adolescence (13-19 years, adolescent-onset type 1 diabetes) and adulthood (20-40 years, adult-onset type 1 diabetes), respectively. Table 1 shows the differences in the clinical characteristics among the three groups. Participants in the pediatric-onset type 1 diabetes group were significantly younger (P < 0.0001), had a significantly longer duration of type 1 diabetes (P < 0.0001), and had a significantly higher prevalence of hypoglycemia unawareness and retinopathy (P < 0.0005 and P < 0.0001, respectively). The prevalence of oral administration of therapeutic agents for hypertension and dyslipidemia was higher in the adult-onset group (P < 0.05 and P < 0.0005, respectively). The frequency and severity of depressive symptoms according to the age at onset of type 1 diabetes are shown in Figure 1. Moderate depressive symptoms were noted in 21%, 18% and 13% of participants with pediatric-onset, adolescentonset and adult-onset type 1 diabetes, respectively. The prevalence of moderate depressive symptoms significantly increased with lower age of onset (P < 0.05, Cochran-Armitage propensity test). In contrast, the prevalence of mild depressive symptoms was nearly similar for all the groups classified by the age of onset.
To examine the clinical characteristics of participants with pediatric-onset and adolescent-onset type 1 diabetes with moderate depressive symptoms, we categorized the participants into two subgroups according to their total PHQ-9 scores: <10 points and ≥10 points (Table 2). Consequently, we observed significantly higher HbA1c levels (P < 0.0001), and higher rates of neuropathy (P < 0.0001), retinopathy (P < 0.0001) and hemodialysis (P < 0.05) in participants with moderate depressive symptoms.
The results of multivariate logistic regression analysis for the likelihood of moderate depressive symptoms are shown in Table 3. A PHQ-9 score ≥10 points was treated as a categorical outcome variable. Age, duration of diabetes, sex, HbA1c level, hypoglycemia unawareness, retinopathy, neuropathy and hemodialysis were selected as explanatory variables. In pediatric-onset and adolescent-onset type 1 diabetes, the likelihood of moderate depressive symptoms increased by 70% (P < 0.05) with the presence of hypoglycemia unawareness, whereas no correlation was observed between depression and hypoglycemia unawareness in participants with adult-onset type 1 diabetes. High HbA1c levels and the presence of neuropathy were related to moderate depressive symptoms at all ages of onset. Among participants with adult-onset type 1 diabetes, women had more depressive symptoms than men.

DISCUSSION
In the present single-center cross-sectional cohort analysis, we clarified the difference in the prevalence of depressive symptoms by the age of onset of type 1 diabetes in participants with the disease. The prevalence of moderate depressive symptoms was significantly higher in participants with pediatriconset and adolescent-onset type 1 diabetes than in those with adult-onset type 1 diabetes. Furthermore, moderate depressive symptoms were associated with high HbA1c, presence of neuropathy and hypoglycemia unawareness among participants who were aged <20 years at the onset of type 1 diabetes.
The prevalence of depression in type 1 diabetes was 2.4-to 3.8-fold higher than that of people without type 1 diabetes 17 . In a meta-analysis, 30% of participants with type 1 diabetes diagnosed at age <20 years had depressive symptoms 18 . Recently, it was reported that 9-20% of adult patients with type 1 diabetes showed higher scores on the PHQ-9 (≥10 points) 19,20 . The results of the present study are very similar to those of previous studies. However, when 77 children with type 1 diabetes in Japan aged 10-18 years were evaluated using the Children's Depression Inventory, none showed depressive symptoms 21 . Compared with the present result, reasons that could explain the difference in the proportion of depressive symptoms in other studies are the sample size and the advances in the treatment of type 1 diabetes. In the present study, we evaluated depressive symptoms in people living with type 1 diabetes for an average of 21 years. The number of participants who persistently experienced depressive symptoms from childhood to adulthood is unclear. Furthermore, the environment around children with type 1 diabetes has changed remarkably in the past few decades (e.g., diet therapy and the choice of freedom in foods have greatly improved with the dissemination of carbohydrate-counting therapy). Currently, individuals can choose their insulin requirements according to their lifestyle. However, most pediatric-onset and adolescent-onset type 1 diabetes participants in the present study received strict dietary and exercise therapy guidance at the time of type 1 diabetes onset, and it is possible that the burden of diabetes has continued until now. The present study is the first to identify that depressive symptoms are more frequent in individuals with pediatric-onset and adolescent-onset type 1 diabetes than in those with adultonset type 1 diabetes. Children with type 1 diabetes go through various life events, such as dreaming of the future, deciding their life course, entrance exams, romance, employment, marriage and having a child. During these processes, if they feel discriminated against due to type 1 diabetes, their selfaffirmation and self-esteem will decrease, and the stigma caused by type 1 diabetes will be internalized, which might result in depressive symptoms 22 . Stigma experienced during adolescence might be more prominently associated with depression due to immaturity. Furthermore, there might be gaps between pediatric and adult diabetes healthcare providers 23 . Diabetes healthcare providers for adults might think that patients with pediatric-onset type 1 diabetes are accustomed to living with type 1 diabetes, and thus might not consider the psychosocial issues of their patients. The present results showed that information about continuous mental support was important to avoid depressive symptoms in individuals, regardless of the duration of type 1 diabetes.
Furthermore, depression is associated with fear and anxiety of hypoglycemia 24 . In addition, depression increases the risk of severe hypoglycemia and diabetic ketoacidosis 20 . Conversely, a history of these conditions increases the risk of depression, and depression and severe hypoglycemia have an interactive relationship 25 . In the present study, moderate depressive symptoms were associated with a history of hypoglycemia unawareness in pediatric-onset and adolescent-onset type 1 diabetes. We could not explain why hypoglycemia unawareness was associated with depressive symptoms in participants with pediatric-onset and adolescent-onset type 1 diabetes but was not associated with depressive symptoms in those with adult-onset type 1 diabetes. Among individuals who experienced severe hypoglycemic events that resulted in an emergency room visit or hospitalization, 75% had an elevated risk of depression compared with individuals who did not experience such events 25 Figure 1 | Frequency and severity of depressive symptoms according to the age at onset of type 1 diabetes. Participants were categorized into three groups based on the severity of depressive symptoms on the total Patient Health Questionnaire-9 (PHQ-9) score. The prevalence of moderate depressive symptoms significantly increased with a lower age of onset of type 1 diabetes (P < 0.05). were considered as stressful life events. In the present study, we did not confirm the treatment for hypoglycemia unawareness. People with pediatric-onset and adolescent-onset type 1 diabetes could visit the emergency room more frequently than people with adult-onset type 1 diabetes because of the worry of their parents or caregivers. Additionally, other factors (such as working environment, relationships with family and friends, income, and underlying health conditions besides diabetes) might have a greater influence on depressive symptoms for people with adult-onset type 1 diabetes. Yet, little research has been carried out on the association between depression and hypoglycemia unawareness, and further study, including prospective research, is necessary. Thus, avoiding hypoglycemia unawareness was necessary to decrease depressive symptoms. Currently available technologies, such as continuous glucose monitoring and sensor-augmented insulin therapy, have shown a positive impact on reducing hypoglycemia unawareness 26 . Higher PHQ-9 scores were associated with increased odds ratios for retinopathy, neuropathy, and end-stage renal disease in both type 1 diabetes and type 2 diabetes 9, 14 . Furthermore, a longitudinal study showed an association between depression and diabetic retinopathy 27 and nephropathy 19,28 . Therefore, it is important to consider strategies to reduce depressive symptoms to prevent the onset and progression of type 1 diabetesrelated complications. Collaborative care interventions with diabetologists and non-physician care coordinators resulted in statistically significant improvements in depressive symptoms and cardiometabolic indices, including HbA1c, blood pressure and low-density lipoprotein cholesterol in individuals with type 2 diabetes 29 . As the duration of type 1 diabetes in pediatric and adolescent-onset cases is longer than the type 1 diabetes duration in adult-onset cases, people with pediatric and adolescentonset type 1 diabetes are likely to have diabetes-related complications when they are younger. Therefore, the transition period from pediatric to adult diabetes healthcare providers was appropriate for psychosocial screening 30,31 . Earlier detection of depressive symptoms and collaboration with mental health specialists could prevent the onset and progression of diabetesrelated complications.
The present study had several limitations. First, this study was carried out at a single university hospital; therefore, the findings are not generalizable to all individuals with type 1 diabetes. Second, from the perspective of precise intervention for depression, depressive symptoms according to the PHQ-9 should be supplemented by a clinical diagnosis of depression from a psychiatrist. Third, selection bias cannot be ruled out, because participation in the DIACET Study was voluntary, and fewer persons with depression might have participated in the study. Fourth, other conditions, such as family relations, incomes, educations, medical insurance, body mass index, exercise, smoking, alcohol, eating disorders and 25 (OH)D3 levels, could be associated with depressive symptoms. A future prospective study could study these associations in detail. Finally, this was a cross-sectional survey, which limited the identification of a causal relationship between depressive symptoms, the age of onset of type 1 diabetes and complications. Therefore, prospective surveys are necessary to clarify the longitudinal changes in depression symptoms, and the acute and chronic complications in people with type 1 diabetes.
The present study highlighted the association between depressive symptoms and the age of onset of type 1 diabetes among participants with the disease. With the evidence of the elevated prevalence of moderate depressive symptoms in adulthood among individuals with pediatric-onset and adolescentonset type 1 diabetes, it is important to avoid hypoglycemia unawareness and consider screening for depression regularly, including in the transitional period from childhood to adulthood in individuals with type 1 diabetes. Early detection of depression can facilitate effective treatment and help minimize