Psychoneuroendocrinology (2015) 55, 48—58
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/psyneuen
Associations between HbA1c and depressive symptoms in young adults with early-onset type 1 diabetes夽 Christina Bächle a,b,∗, Karin Lange c, Anna Stahl-Pehe a,b, Katty Castillo a,b, Reinhard W. Holl b,d, Guido Giani a,b, Joachim Rosenbauer a,b a
German Diabetes Center, Institute for Biometrics and Epidemiology, Auf’m Hennekamp 65, D-40225 Düsseldorf, Germany b German Center for Diabetes Research (DZD), Germany c Hannover Medical School, Department of Medical Psychology, Carl-Neuberg-Str. 1, OE 5430, D-30626 Hannover, Germany d University of Ulm, Institute of Epidemiology and Medical Biometry, Albert-Einstein-Allee 41, D-89081 Ulm, Germany Received 9 December 2014; received in revised form 15 January 2015; accepted 30 January 2015
KEYWORDS Type 1 diabetes; Duration; Depression; Depressive disorder; Metabolic control; Hemoglobin A1c
Summary Objective: This study sought to evaluate the associations between metabolic control and each DSM-5 (Diagnostic and Statistical Manual, fifth edition) symptom of depression among young women and men with early-onset long-duration type 1 diabetes. Methods: The data of 202 18—21-year-old patients with type 1 diabetes from a populationbased, nationwide survey (40.1% male) with a mean age of 19.4 (standard deviation 0.9) years, a mean HbA1c level of 8.3% (1.6%) (i.e., 67 [17.5] mmol/mol), and a mean diabetes duration of 15.7 (1.0) years were included. The German version of the Patient Health Questionnaire (PHQ-9) was used to assess depression symptoms. For each PHQ-9 depressive symptom, the mean HbA1c values of screening-positive and screening-negative patients were compared via t-test. The associations between HbA1c levels and depressive symptoms were analyzed using multiple linear regression analyses and stepwise adjustments for individual, socioeconomic and health-related covariates.
夽
In cooperation with the German Pediatric Surveillance Unit (ESPED) and the DPV-Science initiative, supported by the Competence Network Diabetes Mellitus (support codes 01GI0802, 01GI1109A) and the German Center for Diabetes Research (DZD). ∗ Corresponding author at: Institute for Biometrics and Epidemiology, German Diabetes Center, Leibniz Institute for Diabetes Research at Heinrich Heine University, Auf’m Hennekamp 65, D-40225 Düsseldorf, Germany. Tel.: +49 211 3382 411; fax: +49 211 3382 677. E-mail address: [email protected] (C. Bächle). http://dx.doi.org/10.1016/j.psyneuen.2015.01.026 0306-4530/© 2015 Elsevier Ltd. All rights reserved.
Associations between HbA1c and depressive symptoms in T1D
49
Results: Exactly 43.0% and 33.3% of female and male participants reported at least one depressive symptom, and 5.0% and 2.5% met the DSM-5 criteria for major depressive syndrome. HbA1c levels increased with psychomotor agitation/retardation (women), overeating/poor appetite (men/women), lethargy (men), and sleep difficulty (men). Overeating/poor appetite, lethargy, and total PHQ-9 score (per score increase by one) were associated with increased HbA1c levels of 1.10, 0.96 and 0.09 units (%), respectively. Conclusions: The associations between depressive symptoms and HbA1c levels vary by symptom and sex. Differentiating the symptoms of depression and targeted interventions might help to improve metabolic outcomes in young adults with early-onset type 1 diabetes and depression. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Depression and diabetes have serious individual, societal and economic effects, and they often co-occur (Roy and Lloyd, 2012). According to Manarte et al. (2010), the prevalence of depression among patients with diabetes is rapidly increasing; furthermore, diabetes and depression will become the most prevalent health problems in the 21st century. According to the results of three systematic reviews (Anderson et al., 2001; Groot et al., 2001; Barnard et al., 2006), the prevalence of clinical depression among adults with type 1 diabetes was approximately four times greater than that of a control group without diabetes (12.0% vs. 3.2%) (Roy and Lloyd, 2012). Compared with type 2 diabetes, however, the data regarding the prevalence of comorbid depression among patients with type 1 diabetes remain scarce, thereby limiting the reliability of the reported findings (Barnard et al., 2006; Roy and Lloyd, 2012). The coexistence of depression or depressive symptoms and type 1 diabetes has been associated with insufficient diabetes outcomes and increased admission to hospitals, emergency units and outpatient consultations, resulting in higher total healthcare costs and increased mortality rates (Ciechanowski et al., 2000; Groot et al., 2001; Lawrence et al., 2006; Roy and Lloyd, 2012; van Dooren et al., 2013; Plener et al., 2014). Depression has been frequently associated with elevated hemoglobin A1c (HbA1c) levels (Lawrence et al., 2006; Sacco and Bykowski, 2010; Corathers et al., 2013; Melin et al., 2013; Plener et al., 2014). In turn, HbA1c levels have been associated with acute and chronic complications (Bryden et al., 2003; Peters and Laffel, 2011). Therefore, HbA1c is likely a central mediator of the association between depression and long-term outcomes (Bot et al., 2013). In this context, young adults with type 1 diabetes are of special interest. During development, they are challenged in various ways; like their peers without diabetes, they must overcome educational, economic and social challenges. In addition, they must take full responsibility for their disease management (Arnett, 2000; Hamilton and Daneman, 2002; Anderson, 2010; Johnson et al., 2013). Specific challenges arise during the transition from pediatric to adult diabetes care, which is recommended to occur between the ages of 18 and 21 years (Cooley and Sagerman, 2011). In a subsample of the SEARCH for Diabetes in Youth Study, leaving pediatric diabetes care was associated with a 2.5-fold increase in the odds for poor diabetes control; however, the factors that
mediate this effect have not yet been identified (Lotstein et al., 2013). Previous research suggested that young adults with type 1 diabetes are at an increased risk for depressive symptoms, which result in insufficient metabolic control. However, the association between these conditions has rarely been analyzed. Of the 23 studies included in a recently published systematic review on the prevalence and metabolic implications of depression among patients with type 1 diabetes up to the age of 25 years (Johnson et al., 2013), only one study included the data of patients older than 18 years (de Wit and Snoek, 2011). The wide participants’ age range (age 9—19 years) and the small number of adults in that study limit the validity of the results with regard to young adults. Another recent study assessed the prevalence and associations between the symptoms of depression/antidepressant medication and the metabolic outcomes among patients with type 1 diabetes who were younger than 25 years. However, this study was based on a clinical diagnosis of depressive symptoms rather than standardized measures, and specific age groups were not targeted (Plener et al., 2014). Former research on the association between depression (using the DSM-IV criteria) and glycemic control primarily concentrated on depression in its entirety without differentiating between its distinct symptoms, as was previously recommended (Lux and Kendler, 2010; Bot et al., 2013). Additional knowledge concerning the associations between each depressive symptom and specific glycemic outcomes might provide deeper insight into the etiology of comorbid depression. This research might also help to develop individualized interventions for young adults (Bot et al., 2013). To the best of our knowledge, only two studies have thoroughly analyzed the associations between the specific symptoms of depression and metabolic control among patients with type 1 diabetes (McGrady and Hood, 2010; Bot et al., 2013). Although female and male participants are differentially affected by depression (Grey et al., 2002; Manarte et al., 2010), those studies did not report sex differences. The age range of the participants was wide, and none of the studies focused on age groups with potentially increased risks for both depressive symptoms and poor metabolic control. Patients with long disease durations are predisposed to late diabetes sequelae (Dabelea, 2009; Downie et al., 2011). Like in other groups, the comorbid depression in emerging adulthood associated with poor metabolic control (Lawrence et al., 2006; Sacco and Bykowski, 2010; Corathers et al., 2013; Melin et al., 2013; Plener et al.,
50 2014) might particularly worsen future health perspectives. Therefore, the aim of the current study was to analyze the associations between metabolic control and each of the nine DSM-5 symptoms of depression. Metabolic control was operationalized as the recent HbA1c level in young adults with early type 1 diabetes onset and a disease duration of at least ten years. We hypothesized that individual symptoms of depression would be differently associated with metabolic control and that the results would vary by sex.
2. Material and methods 2.1. Data sources A nationwide, population-based cohort study called the ‘‘Clinical Course of Type 1 Diabetes in Children, Adolescents and Young Adults with Disease Onset in Preschool Age’’ was initiated as part of the German Competence Network on Diabetes Mellitus. Details concerning the study design have been described previously (Stahl et al., 2012). In brief, children, adolescents and young adults were invited to participate in the questionnaire survey if (1) type 1 diabetes was diagnosed between 0 and 4 years old (i.e., early diabetes onset), and (2) the duration of type 1 diabetes was at least 10 years (i.e., the disease onset occurred from 1993 to 1999). Between September 2009 and December 2010, exactly 2241 11—21-year-old patients and their parents (in the case of younger participants) were asked to complete one of three standardized questionnaires (specifically arranged questionnaires for children and adolescents, young adults, or parents of children with type 1 diabetes) and to send them back to the study center with their written informed consent. The ethical review board at Heinrich Heine University Düsseldorf approved this study. In total, 629 children/adolescents and their parents as well as 211 young adults with type 1 diabetes returned the comprehensive questionnaires, and additional 280 patients/parents answered short questionnaires, resulting in an overall response rate of 50%.
2.2. Sample Of the 18—21-year-old participants, 85 men and 126 women with early-onset type 1 diabetes returned the comprehensive questionnaire, resulting in a response rate of 30% (Stahl-Pehe et al., 2014). Comparisons of the 211 participants with the non-participants showed that men were underrepresented in the sample (40.3% male participants vs. 53.5% male non-participants, p < 0.001). The mean age of the participants was significantly lower (0.18 years, p = 0.017), and the disease duration was 0.20 years briefer (p = 0.010) than those of the non-participants. Age at diabetes onset did not significantly differ between the two groups. The current analyses were based on 202 participants with complete data for the brief version of the Patient Health Questionnaire (PHQ-9).
C. Bächle et al.
2.3. Variable assessment To assess depressive symptoms, the young adult participants completed the German version of the PHQ-9, which was jointly developed with the authors of the original PHQ (Löwe et al., 2002) and validated in particular among diabetes patients (van Steenbergen-Weijenburg et al., 2010). The nine items of the PHQ-9 were based on the DSM-IV criteria (Diagnostic and Statistical Manual, fourth edition) for major depressive disorder. When introducing the DSM-5 system, the American Psychiatry Association (APA) recently recommended the PHQ-9 as an emerging severity measure for depression among adults (American Psychiatry Association, 2014). The participants answered nine questions regarding whether, over the previous 2 weeks, they were affected by any of the core symptoms of depression using four response categories ranging from ‘‘not at all’’ (corresponding to 0) to ‘‘nearly every day’’ (corresponding to 3) (Kroenke et al., 2010; Sacco and Bykowski, 2010). The screening for the key depressive symptoms anhedonia and depressed mood and further depressive symptoms sleep difficulty, lethargy, overeating/poor appetite, feelings of worthlessness/low self-esteem, difficulties in concentration, and psychomotor retardation/agitation (i.e., questions 1—8) was evaluated as positive when the respective depressive symptom was present for at least 7 days over the previous 2 weeks. Question 9 (suicidal ideation) was valued as positive whenever it was confirmed. The PHQ-9 enables both categorical and continuous data analyses (Kroenke et al., 2001). Categorically, the PHQ-9 serves as a screening instrument for a major depressive syndrome (MDS) with a positive result when five or more of the depressive symptoms were screening-positive, necessarily including one of the key depressive symptoms dysphoria or anhedonia. When only two to four of the depressive items were present (again, including one of the key depressive symptoms dysphoria or anhedonia), an other depressive syndrome (ODS) was indicated. Summing the points of each answer resulted in a total PHQ-9 score (range 0—27) that was used to assess the severity of the depressive symptoms. The score was categorized as mild (5—9), moderate (10—14), moderately severe (15—19), or severe (≥20) (Kroenke et al., 2001, 2010; Löwe et al., 2002). The sociodemographic variables included age, sex, family structure/residence, and socioeconomic status. The composite socioeconomic status index of each participant was determined based on his or her educational, professional and income data. Given that many of the patients were still receiving vocational training, the socioeconomic status levels of the participants who stated that they were not the principal earner in the household were approximated using those of the principal earners in the household based on their education levels, professional statuses and household incomes. Composite socioeconomic status levels were then classified into three categories (low, middle, and high), following the usual procedures applied for Germany (Stahl et al., 2012). As health-related covariates, smoking and body mass index (BMI, kg/m2 ) were included. The latter was calculated using self-reported height and weight data, and
Associations between HbA1c and depressive symptoms in T1D participants were classified as underweight (BMI
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/psyneuen
Associations between HbA1c and depressive symptoms in young adults with early-onset type 1 diabetes夽 Christina Bächle a,b,∗, Karin Lange c, Anna Stahl-Pehe a,b, Katty Castillo a,b, Reinhard W. Holl b,d, Guido Giani a,b, Joachim Rosenbauer a,b a
German Diabetes Center, Institute for Biometrics and Epidemiology, Auf’m Hennekamp 65, D-40225 Düsseldorf, Germany b German Center for Diabetes Research (DZD), Germany c Hannover Medical School, Department of Medical Psychology, Carl-Neuberg-Str. 1, OE 5430, D-30626 Hannover, Germany d University of Ulm, Institute of Epidemiology and Medical Biometry, Albert-Einstein-Allee 41, D-89081 Ulm, Germany Received 9 December 2014; received in revised form 15 January 2015; accepted 30 January 2015
KEYWORDS Type 1 diabetes; Duration; Depression; Depressive disorder; Metabolic control; Hemoglobin A1c
Summary Objective: This study sought to evaluate the associations between metabolic control and each DSM-5 (Diagnostic and Statistical Manual, fifth edition) symptom of depression among young women and men with early-onset long-duration type 1 diabetes. Methods: The data of 202 18—21-year-old patients with type 1 diabetes from a populationbased, nationwide survey (40.1% male) with a mean age of 19.4 (standard deviation 0.9) years, a mean HbA1c level of 8.3% (1.6%) (i.e., 67 [17.5] mmol/mol), and a mean diabetes duration of 15.7 (1.0) years were included. The German version of the Patient Health Questionnaire (PHQ-9) was used to assess depression symptoms. For each PHQ-9 depressive symptom, the mean HbA1c values of screening-positive and screening-negative patients were compared via t-test. The associations between HbA1c levels and depressive symptoms were analyzed using multiple linear regression analyses and stepwise adjustments for individual, socioeconomic and health-related covariates.
夽
In cooperation with the German Pediatric Surveillance Unit (ESPED) and the DPV-Science initiative, supported by the Competence Network Diabetes Mellitus (support codes 01GI0802, 01GI1109A) and the German Center for Diabetes Research (DZD). ∗ Corresponding author at: Institute for Biometrics and Epidemiology, German Diabetes Center, Leibniz Institute for Diabetes Research at Heinrich Heine University, Auf’m Hennekamp 65, D-40225 Düsseldorf, Germany. Tel.: +49 211 3382 411; fax: +49 211 3382 677. E-mail address: [email protected] (C. Bächle). http://dx.doi.org/10.1016/j.psyneuen.2015.01.026 0306-4530/© 2015 Elsevier Ltd. All rights reserved.
Associations between HbA1c and depressive symptoms in T1D
49
Results: Exactly 43.0% and 33.3% of female and male participants reported at least one depressive symptom, and 5.0% and 2.5% met the DSM-5 criteria for major depressive syndrome. HbA1c levels increased with psychomotor agitation/retardation (women), overeating/poor appetite (men/women), lethargy (men), and sleep difficulty (men). Overeating/poor appetite, lethargy, and total PHQ-9 score (per score increase by one) were associated with increased HbA1c levels of 1.10, 0.96 and 0.09 units (%), respectively. Conclusions: The associations between depressive symptoms and HbA1c levels vary by symptom and sex. Differentiating the symptoms of depression and targeted interventions might help to improve metabolic outcomes in young adults with early-onset type 1 diabetes and depression. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Depression and diabetes have serious individual, societal and economic effects, and they often co-occur (Roy and Lloyd, 2012). According to Manarte et al. (2010), the prevalence of depression among patients with diabetes is rapidly increasing; furthermore, diabetes and depression will become the most prevalent health problems in the 21st century. According to the results of three systematic reviews (Anderson et al., 2001; Groot et al., 2001; Barnard et al., 2006), the prevalence of clinical depression among adults with type 1 diabetes was approximately four times greater than that of a control group without diabetes (12.0% vs. 3.2%) (Roy and Lloyd, 2012). Compared with type 2 diabetes, however, the data regarding the prevalence of comorbid depression among patients with type 1 diabetes remain scarce, thereby limiting the reliability of the reported findings (Barnard et al., 2006; Roy and Lloyd, 2012). The coexistence of depression or depressive symptoms and type 1 diabetes has been associated with insufficient diabetes outcomes and increased admission to hospitals, emergency units and outpatient consultations, resulting in higher total healthcare costs and increased mortality rates (Ciechanowski et al., 2000; Groot et al., 2001; Lawrence et al., 2006; Roy and Lloyd, 2012; van Dooren et al., 2013; Plener et al., 2014). Depression has been frequently associated with elevated hemoglobin A1c (HbA1c) levels (Lawrence et al., 2006; Sacco and Bykowski, 2010; Corathers et al., 2013; Melin et al., 2013; Plener et al., 2014). In turn, HbA1c levels have been associated with acute and chronic complications (Bryden et al., 2003; Peters and Laffel, 2011). Therefore, HbA1c is likely a central mediator of the association between depression and long-term outcomes (Bot et al., 2013). In this context, young adults with type 1 diabetes are of special interest. During development, they are challenged in various ways; like their peers without diabetes, they must overcome educational, economic and social challenges. In addition, they must take full responsibility for their disease management (Arnett, 2000; Hamilton and Daneman, 2002; Anderson, 2010; Johnson et al., 2013). Specific challenges arise during the transition from pediatric to adult diabetes care, which is recommended to occur between the ages of 18 and 21 years (Cooley and Sagerman, 2011). In a subsample of the SEARCH for Diabetes in Youth Study, leaving pediatric diabetes care was associated with a 2.5-fold increase in the odds for poor diabetes control; however, the factors that
mediate this effect have not yet been identified (Lotstein et al., 2013). Previous research suggested that young adults with type 1 diabetes are at an increased risk for depressive symptoms, which result in insufficient metabolic control. However, the association between these conditions has rarely been analyzed. Of the 23 studies included in a recently published systematic review on the prevalence and metabolic implications of depression among patients with type 1 diabetes up to the age of 25 years (Johnson et al., 2013), only one study included the data of patients older than 18 years (de Wit and Snoek, 2011). The wide participants’ age range (age 9—19 years) and the small number of adults in that study limit the validity of the results with regard to young adults. Another recent study assessed the prevalence and associations between the symptoms of depression/antidepressant medication and the metabolic outcomes among patients with type 1 diabetes who were younger than 25 years. However, this study was based on a clinical diagnosis of depressive symptoms rather than standardized measures, and specific age groups were not targeted (Plener et al., 2014). Former research on the association between depression (using the DSM-IV criteria) and glycemic control primarily concentrated on depression in its entirety without differentiating between its distinct symptoms, as was previously recommended (Lux and Kendler, 2010; Bot et al., 2013). Additional knowledge concerning the associations between each depressive symptom and specific glycemic outcomes might provide deeper insight into the etiology of comorbid depression. This research might also help to develop individualized interventions for young adults (Bot et al., 2013). To the best of our knowledge, only two studies have thoroughly analyzed the associations between the specific symptoms of depression and metabolic control among patients with type 1 diabetes (McGrady and Hood, 2010; Bot et al., 2013). Although female and male participants are differentially affected by depression (Grey et al., 2002; Manarte et al., 2010), those studies did not report sex differences. The age range of the participants was wide, and none of the studies focused on age groups with potentially increased risks for both depressive symptoms and poor metabolic control. Patients with long disease durations are predisposed to late diabetes sequelae (Dabelea, 2009; Downie et al., 2011). Like in other groups, the comorbid depression in emerging adulthood associated with poor metabolic control (Lawrence et al., 2006; Sacco and Bykowski, 2010; Corathers et al., 2013; Melin et al., 2013; Plener et al.,
50 2014) might particularly worsen future health perspectives. Therefore, the aim of the current study was to analyze the associations between metabolic control and each of the nine DSM-5 symptoms of depression. Metabolic control was operationalized as the recent HbA1c level in young adults with early type 1 diabetes onset and a disease duration of at least ten years. We hypothesized that individual symptoms of depression would be differently associated with metabolic control and that the results would vary by sex.
2. Material and methods 2.1. Data sources A nationwide, population-based cohort study called the ‘‘Clinical Course of Type 1 Diabetes in Children, Adolescents and Young Adults with Disease Onset in Preschool Age’’ was initiated as part of the German Competence Network on Diabetes Mellitus. Details concerning the study design have been described previously (Stahl et al., 2012). In brief, children, adolescents and young adults were invited to participate in the questionnaire survey if (1) type 1 diabetes was diagnosed between 0 and 4 years old (i.e., early diabetes onset), and (2) the duration of type 1 diabetes was at least 10 years (i.e., the disease onset occurred from 1993 to 1999). Between September 2009 and December 2010, exactly 2241 11—21-year-old patients and their parents (in the case of younger participants) were asked to complete one of three standardized questionnaires (specifically arranged questionnaires for children and adolescents, young adults, or parents of children with type 1 diabetes) and to send them back to the study center with their written informed consent. The ethical review board at Heinrich Heine University Düsseldorf approved this study. In total, 629 children/adolescents and their parents as well as 211 young adults with type 1 diabetes returned the comprehensive questionnaires, and additional 280 patients/parents answered short questionnaires, resulting in an overall response rate of 50%.
2.2. Sample Of the 18—21-year-old participants, 85 men and 126 women with early-onset type 1 diabetes returned the comprehensive questionnaire, resulting in a response rate of 30% (Stahl-Pehe et al., 2014). Comparisons of the 211 participants with the non-participants showed that men were underrepresented in the sample (40.3% male participants vs. 53.5% male non-participants, p < 0.001). The mean age of the participants was significantly lower (0.18 years, p = 0.017), and the disease duration was 0.20 years briefer (p = 0.010) than those of the non-participants. Age at diabetes onset did not significantly differ between the two groups. The current analyses were based on 202 participants with complete data for the brief version of the Patient Health Questionnaire (PHQ-9).
C. Bächle et al.
2.3. Variable assessment To assess depressive symptoms, the young adult participants completed the German version of the PHQ-9, which was jointly developed with the authors of the original PHQ (Löwe et al., 2002) and validated in particular among diabetes patients (van Steenbergen-Weijenburg et al., 2010). The nine items of the PHQ-9 were based on the DSM-IV criteria (Diagnostic and Statistical Manual, fourth edition) for major depressive disorder. When introducing the DSM-5 system, the American Psychiatry Association (APA) recently recommended the PHQ-9 as an emerging severity measure for depression among adults (American Psychiatry Association, 2014). The participants answered nine questions regarding whether, over the previous 2 weeks, they were affected by any of the core symptoms of depression using four response categories ranging from ‘‘not at all’’ (corresponding to 0) to ‘‘nearly every day’’ (corresponding to 3) (Kroenke et al., 2010; Sacco and Bykowski, 2010). The screening for the key depressive symptoms anhedonia and depressed mood and further depressive symptoms sleep difficulty, lethargy, overeating/poor appetite, feelings of worthlessness/low self-esteem, difficulties in concentration, and psychomotor retardation/agitation (i.e., questions 1—8) was evaluated as positive when the respective depressive symptom was present for at least 7 days over the previous 2 weeks. Question 9 (suicidal ideation) was valued as positive whenever it was confirmed. The PHQ-9 enables both categorical and continuous data analyses (Kroenke et al., 2001). Categorically, the PHQ-9 serves as a screening instrument for a major depressive syndrome (MDS) with a positive result when five or more of the depressive symptoms were screening-positive, necessarily including one of the key depressive symptoms dysphoria or anhedonia. When only two to four of the depressive items were present (again, including one of the key depressive symptoms dysphoria or anhedonia), an other depressive syndrome (ODS) was indicated. Summing the points of each answer resulted in a total PHQ-9 score (range 0—27) that was used to assess the severity of the depressive symptoms. The score was categorized as mild (5—9), moderate (10—14), moderately severe (15—19), or severe (≥20) (Kroenke et al., 2001, 2010; Löwe et al., 2002). The sociodemographic variables included age, sex, family structure/residence, and socioeconomic status. The composite socioeconomic status index of each participant was determined based on his or her educational, professional and income data. Given that many of the patients were still receiving vocational training, the socioeconomic status levels of the participants who stated that they were not the principal earner in the household were approximated using those of the principal earners in the household based on their education levels, professional statuses and household incomes. Composite socioeconomic status levels were then classified into three categories (low, middle, and high), following the usual procedures applied for Germany (Stahl et al., 2012). As health-related covariates, smoking and body mass index (BMI, kg/m2 ) were included. The latter was calculated using self-reported height and weight data, and
Associations between HbA1c and depressive symptoms in T1D participants were classified as underweight (BMI
