GRAPHICAL ABSTRACT

Highlights

• Using the Korean NHIS cohort, PA change after new T2DM diagnosis was analyzed.

• PA level by MET-min/wk was assessed from two consecutive health screenings.

• PA change after diabetes diagnosis was linked to incident depression.

• Increasing PA lowered depression risk, while decreased PA raised it.

• Increasing PA may be helpful for mental health in patients with newly diagnosed T2DM.

INTRODUCTION

Type 2 diabetes mellitus (T2DM) has numerous comorbidities including physical ones such as cardiovascular diseases as well as psychiatric disorders [1,2]. Depression is the most common psychiatric comorbidity of diabetes affecting 16.5% of the patients with T2DM [3], which are prone to mood disorders due to the emotional burden of managing a lifelong condition and persistent social stigma on metabolic disorders [4,5]. Depression of patients with T2DM leads to negative health outcomes such as poor adherence to treatment and inadequate glycemic control [6-8]. In addition, depression decreases quality of life (QoL) in patients with T2DM, which is independent of the lower QoL caused by the diabetes itself [9,10]. Therefore, efforts are needed to treat and prevent depression in people with T2DM, and prevention activities include identifying modifiable risk factors.

Physical inactivity is one of the key risk factors of depression, and numerous studies on the general population have shown that engaging in regular physical activity (PA) has a protective effect against depression, regardless of age, gender, or global region [11,12]. Increasing PA compared to baseline levels also lowers the risk of incident depression in the general population [13,14]. Considering that 54% of people with diabetes have no or insufficient physical activities [15], a significant number of patients with T2DM would benefit from increasing the amount of exercise. A few studies have reported that PA is associated with lower depressive symptoms in the diabetic population [16,17]. Similarly, a meta-analysis by Narita et al. [18] revealed that interventions on PA are effective in treating diabetes-related depression. However, there is only limited evidence on whether PA is also effective in preventing depression in patients with T2DM. To our knowledge, there has been no prior research suggesting that changes in PA, not the baseline activity, are associated with the risk of incident depression in patients with T2DM and without a previous history of depression.

Thus, this nationwide population-based cohort study aimed to evaluate the association of changes in physical activities and the risk of incident depression in participants with newly diagnosed T2DM.

METHODS

Participants

Our study utilized a nationwide, population-based dataset from the National Health Insurance Service (NHIS) of Korea, comprising randomly sampled data from approximately 2.2 million patients diagnosed with T2DM. From this database, we identified 319,125 individuals who were newly diagnosed with T2DM between 2009 and 2015 and completed two consecutive health examinations: 1 within 2 years prior to their T2DM diagnosis and the other within 2 years following their diagnosis. Among these, we first excluded 2,766 individuals due to missing data related to PA or covariates. We subsequently excluded an additional 61,740 participants who had been diagnosed with depression prior to the index date. Ultimately, as shown in Fig. 1, our final population included 254,619 patients who met the inclusion criteria, having undergone both health checkups with complete data on PA and no prior history of depression. Approval for this study was obtained from the Seoul National University Hospital Institutional Review Board (E-2204-038-1314). Additionally, the NHIS Big Data Steering Department provided approval with research management number NHIS-2023-1-695. The requirement for informed consent from individual participants was waived because the NHIS database consists of strictly anonymized clinical data, with access limited to authorized individuals in compliance with the Personal Information Protection Act guidelines.

Assessment of the PA change

PA levels were assessed using self-administered questionnaires based on a 7-day recall method, a validated approach for population-level PA monitoring [19]. In addition, this questionnaire has been used in previous population-based studies and its reliability and validity have been evaluated in Korean settings. For example, the Korean version of the Global Physical Activity Questionnaire (K-GPAQ) demonstrated test–retest reliability (κ or correlation coefficients approximately 0.60 to 0.70 for recreational PA) and a modest but significant correlation with accelerometer-based measurements (r=0.30–0.40) [20,21]. The PA questionnaires were self-administered by participants as part of the standardized NHIS health screening program and collected under the supervision of trained medical staff at examination centers. Questionnaires were administered at the time of routine national health examinations conducted within 2 years before and 2 years after the diagnosis of T2DM. Utilizing data from the two health examinations closest to the T2DM diagnosis date one within 2 years prior and the other within 2 years following diagnosis we calculated PA levels. Participants indicated how often they engaged each week in light (≥30 minutes/day; e.g., casual walking, light exercise), moderate (≥30 minutes/day; e.g., fast walking, easy biking), and vigorous (≥20 minutes/day; e.g., jogging, aerobic workouts, intense cycling, hiking) PA. Light, moderate, and vigorous PA were assigned metabolic equivalent of task (MET) scores of 2.9, 4.0, and 7.0, respectively, according to prior literature [22,23]. Total energy spent on PA per week (in MET-min/wk) was calculated by multiplying frequency, duration, and MET value for each PA level and summing the results. Based on total MET-min/wk, individuals were grouped into four categories: 0, 1–499, 500–999 (aligning with recommended PA levels), and ≥1,000. PA change was determined by subtracting the pre-diagnosis MET-min/wk value from the post-diagnosis figure. Additionally, we conducted comparative analyses across 16 subgroups defined by MET categories at both health examinations. Four separate analyses were performed, each stratified by baseline PA levels measured during the first examination. The first analysis included individuals with 0 MET-min/wk at baseline, comparing those who maintained 0 MET-min/wk at the second examination with those whose MET-min/wk increased. The second analysis involved participants initially categorized as 1–499 MET-min/wk, comparing those who maintained their MET category with those who experienced MET changes at the second examination. Similarly, the third analysis assessed participants initially within the 500–999 MET-min/wk range, and the fourth analysis included individuals initially classified with ≥1,000 MET-min/wk, focusing on comparative changes, particularly MET reductions, during the second examination period.

Follow-up for new-onset depression

When defining the onset of depression, the International Classification of Diseases, 10th Revision (ICD-10) code and antidepressant drug prescriptions were used. Among the ICD-10 codes, patients diagnosed with F32 and F33 and prescribed antidepressants at the same time with at least two hospital visits were defined as a diagnosis of depression [24,25]. As F33 is a code for recurrent depressive disorders, so for some subjects it is difficult to determine when the first depressive episode occurred, but it was used to extract additional patients due to the limitation of the data accessibility range. To find out the effect of exercise on the onset of depression, patients with depression before new diabetes diagnosis were removed, and through this, only new depression diagnoses were defined after the new diabetes diagnosis date.

Statistical analysis

In this study, we employed a moving index date approach to define the follow-up period, beginning from the second health examination conducted after the initial diagnosis of T2DM, rather than a fixed baseline date. Participants were followed up for an average duration of 7.765 years. Follow-up was concluded on the earliest occurrence among the date of newly diagnosed depression, death, or December 31, 2022. Participants who did not experience any event during the study period were followed until December 31, 2022. Baseline characteristics were described according to sex, incorporating socioeconomic data from the year of T2DM diagnosis and health examination results from the second assessment post-diagnosis. Cox proportional hazards models were utilized for survival analyses. The first Cox regression model adjusted only for age and sex. The second model further adjusted for socioeconomic status variables, including age, sex, household income, and additional health screening variables including factors like body mass index (BMI), fasting blood glucose (FBG), total cholesterol (TCHO), systolic blood pressure (SBP), alcohol intake, smoking status, and the Charlson comorbidity index (CCI). Variables such as sex, PA, household income, alcohol intake, smoking status, and CCI were categorized, while BMI, FBG, TCHO, SBP, and age were treated as continuous variables. In the primary analysis, the association between continuous changes in MET (MET-min/wk) and the risk of depression was evaluated for the entire cohort using restricted cubic spline curves, with knots placed at 200 units of MET-min/wk changes. In the secondary analysis, participants were stratified based on baseline MET-min/wk levels obtained from the initial health examination, and the risk of depression associated with changes in MET-min/wk by the second examination was assessed through survival analysis. All analyses were performed using SAS software version 9.4 (SAS Institute, Cary, NC, USA), and statistical significance was determined using a two-sided P value <0.05.

Additional analyses

Two additional analyses were conducted: firstly, depression risk was evaluated based solely on MET-min/wk status at the initial pre-diagnosis health examination; secondly, depression risk was analyzed based solely on MET-min/wk status from the post-diagnosis health examination.

Subgroup analyses

To explore potential heterogeneity in the association between changes in PA and incident depression, subgroup analyses were performed by age group (<65 years vs. ≥65 years), sex, BMI (<25 kg/m² vs. ≥25 kg/m²), smoking status (current vs. non-current), alcohol intake (none vs. mild-to-heavy), and CCI (<2 vs. ≥2). Participants were stratified by baseline PA level (0 or ≥1,000 MET-min/wk) at the pre-diagnosis health examination. Within each stratum, the association between changes in PA and incident depression was re-estimated. Interaction was tested by including an interaction term between PA change group and the subgroup variable in the Cox proportional hazards models with the same adjustment variables as in the main analysis. A two-sided P value for trend and interaction <0.05 was considered statistically significant. These subgroup analyses were exploratory in nature.

RESULTS

Baseline characteristics

A total of 254,619 participants were included in this study, with a mean±standard deviation age of 56.6±12.0 years. Of these participants, 130,309 were men, accounting for approximately 51.2% of the total cohort. Prior to the diagnosis of T2DM, 22.9% of participants reported no PA (0 MET-min/wk); this proportion was higher among women (26.5%) than men (19.5%). Conversely, 19.7% of participants had higher PA levels (≥1,000 MET-min/wk), with a higher proportion observed among men (22.5%) compared to women (16.8%). Similar patterns were observed at the second health screening following the diagnosis of T2DM. Detailed baseline characteristics by sex are presented in Table 1.

Changes in PA (MET-min/wk) after newly diagnosed T2DM and the risk of depression

Fig. 2 illustrates the association between changes in MET-min/wk, treated as a continuous variable, and the risk of new-onset depression using restricted cubic spline curves. In the overall cohort, greater increases in PA from the pre-diagnosis to post-diagnosis period were significantly associated with lower risk of depression, whereas greater decreases in PA were significantly associated with higher depression risk. These trends were statistically significant when considering the confidence intervals (CIs).

Increased PA and depression

Table 2 shows the results of subgroup analyses based on PA levels at the initial health screening. Compared to participants who remained physically inactive (0 MET-min/wk), patients increasing their PA to 500–999 MET-min/wk had a significantly reduced depression incidence (incidence rate per 10,000 person-years [PY], 12.95; hazard ratio [HR], 0.77; 95% CI, 0.62 to 0.95). Similarly, patients increasing PA to ≥1,000 MET-min/wk also showed a significantly lower incidence of depression (10.92 per 10,000 PY; HR, 0.75; 95% CI, 0.57 to 0.997). A clear gradient in risk corresponding to increasing exposure levels was identified (P for trend=0.0056), indicating that higher increases in PA correlated with greater reductions in depression risk. Table 2 also shows that among participants initially categorized as moderately active (1–499 or 500–999 MET-min/wk), reductions in PA were significantly associated with increased depression risk. Both groups displayed statistically significant dose-response trends (P for trend=0.0173 and 0.0232, respectively), supporting the relationship between PA changes and depression risk.

Decreased PA and depression

As shown in Table 2, participants initially engaging in ≥1,000 MET-min/wk who decreased MET-min/wk demonstrated a non-significant increase in depression risk compared to those maintaining higher PA levels. Specifically, the incidence rates and HRs for those decreasing to 500–999 MET-min/wk (incidence rate per 10,000 PY, 10.59; HR, 1.12; 95% CI, 0.89 to 1.43), 1–499 MET-min/wk (12.48 per 10,000 PY; HR, 1.26; 95% CI, 0.96 to 1.66), and becoming inactive (14.13 per 10,000 PY; HR, 1.33; 95% CI, 0.97 to 1.81) indicated a trend towards increased risk of depression with reduced PA. Despite non-significant CIs, a meaningful trend indicating a progressive association with rising exposure levels was detected (P for trend=0.0342). Among participants initially categorized as moderately active (1–499 or 500–999 MET-min/wk), increases in PA initially showed marginal increases in depression risk, followed by a non-significant trend towards risk reduction with larger activity increases.

Additional analyses

As shown in Supplementary Table 1, among initially inactive participants (0 MET-min/wk), higher PA levels prior to diabetes diagnosis significantly reduced depression risk. Specifically, participants with ≥1,000 MET-min/wk had significantly lower depression incidence (10.66 per 10,000 PY; HR, 0.79; 95% CI, 0.63 to 0.98), confirming a significant dose-response relationship (P for trend=0.0274). Additionally, as shown in Supplementary Table 2, analyses based on post-diagnosis PA showed significantly lower depression risk for all groups with increased PA compared to inactive participants (1–499 MET-min/wk: 11.49 per 10,000 PY; HR, 0.88; 95% CI, 0.79 to 0.98; 500–999 MET-min/wk: 10.10 per 10,000 PY; HR, 0.79; 95% CI, 0.71 to 0.89; ≥1,000 MET-min/wk: 9.07 per 10,000 PY; HR, 0.74; 95% CI, 0.65 to 0.84). All groups demonstrated significant dose-response relationships (P for trend <0.0001).

Subgroup analyses

To further explore potential heterogeneity of our findings, additional stratified analyses were conducted according to baseline PA levels prior to diabetes diagnosis. Supplementary Tables 3 and 4 report both the P values for trend and the P values for interaction. Among participants who reported no PA (0 MET-min/wk) at the pre-diagnosis examination, an increase in MET-min/wk at the second screening was consistently associated with a lower risk of incident depression, showing a graded inverse relationship (P for trend <0.05 in several subgroups). The protective association of increasing PA was generally observed across most strata including age, sex, BMI, smoking status, alcohol intake, and comorbidity level while the overall direction and magnitude of the associations were concordant with the main analysis. Conversely, among individuals who were physically active (≥1,000 MET-min/wk) before diabetes diagnosis, a reduction in PA was associated with a higher risk of developing depression, again consistent with the primary findings. Although a few subgroup-specific estimates (e.g., older adults and non-smokers) showed borderline statistical significance, the overall trend suggested that maintaining or increasing PA was associated with a lower depression risk among patients with newly diagnosed T2DM. Collectively, these supplementary analyses reaffirmed the robustness of the main results and supported the inverse dose-response relationship between changes in PA and subsequent depression risk. Although subgroup-specific point estimates for the trend differed, the formal test for interaction was not statistically significant (P for interaction >0.05), indicating that we did not find evidence that subgroup characteristics modified the effect of PA change on depression risk. To further evaluate whether the impact of decreased PA differed according to baseline PA levels, we additionally conducted subgroup analyses restricted to participants with low (1–499 MET-min/wk) and moderate (500–999 MET-min/wk) baseline activity. In both subgroups, decreases in PA were associated with an increased risk of incident depression, consistent with the direction of the main findings (Supplementary Tables 5 and 6).

DISCUSSION

This nationwide population-based cohort study demonstrated that changes in PA before and after T2DM diagnosis had a significant effect on the risk of new-onset depression. The association was evident in both directions of the change; an increase in MET was linked to a decrease in depression, whereas a decrease in MET led to an increased risk of depression. Furthermore, this trend was consistent regardless of whether the individual had no, moderate, or high levels of previous exercise. In addition, engaging in PA, even at mild levels, has a protective effect against developing depression compared to not being active at all, whether before or after a T2DM diagnosis.

A few studies have shown that changes in metabolic risk factors can help prevent depression in the diabetic population, although none of them have focused on changes in the level of PA alone. A cohort study by Kim et al. [26] displayed that both increases and decreases in body weight were associated with a higher risk of incident depression in patients with T2DM. Another cohort study by An et al. [27] showed that variability in FBG, BMI, blood pressure, and TCHO affects the risk of incident depression in participants with T2DM. Our study demonstrated results in line with these prior research by focusing on changes in PA.

As PA is a factor that is amenable to structured intervention in healthcare settings [28,29], some previous studies have suggested that exercise programs can improve mental health indicators such as depressive symptom scores in patients with T2DM [30-32]. Our findings are consistent with these studies, but we focused more on the preventive intervention rather than the therapeutic effects of PA by completely excluding depression before the time of T2DM diagnosis.

Several mechanisms support that exercise is protective against depression in individuals newly diagnosed with T2DM. One such mechanism involves apelin, which regulates glucose and lipid metabolism and has been suggested as a novel therapeutic agent for diabetes [33,34]. Apelin can also explain the antidepressant effects of exercise because it inhibits neuroinflammation and is also associated with hippocampal neuroplasticity [35,36]. Another mechanism involves hypothalamic-pituitary-adrenal (HPA) axis, which affects cognition and behavior in general by regulating the stress response. Dysregulation of the HPA axis due to chronic stress contributes to the development of depression [37]. Patients with T2DM have subclinical hypercortisolism and abnormalities in the diurnal rhythm of cortisol, which may explain their vulnerability to depression [38,39]. Exercise stabilizes the HPA axis by reducing stress reactivity in both acute and habitual settings, which would be helpful for mental health in the diabetic population [40,41]. Glucose metabolism and inflammation may also serve as mechanisms linking diabetes, exercise, and depression [42-45]. Psychosocial stress is also relevant, given that diabetes has a detrimental effect on the body image of affected individuals [46], and it can be alleviated by engaging in PA [47].

The strength of this study is that it obtained a large study population from a nationalized database, which increases the generalizability of the results; the definitions of T2DM and depression were based on diagnostic codes, which reflect real-world clinical practice; and both the direction and amplitude of MET change were considered. We classified participants into four PA groups and demonstrated a dose-response relationship between changes in PA and depression.

Clinical implications

It should also be noted that PA was assessed only at two time points, before and after diabetes diagnosis, which limited our ability to capture long-term trajectories of behavior. Nevertheless, this approach was intentionally chosen to focus on the early behavioral changes occurring around the time of diagnosis, a period when patients are most likely to modify their lifestyle. By leveraging both pre- and post-diagnosis examinations, our study provides unique insights into the dynamic impact of PA changes on mental health in newly diagnosed patients. This study has implications for both clinical practice and public health policy. Given that changes in exercise behavior before and after a T2DM diagnosis have been shown to affect psychological well-being, it would be helpful for clinicians to suggest changes in exercise level for newly diagnosed patients with T2DM. Individuals who have not been exercising regularly can be encouraged to start some activity, and those who have been mildly physically active can be encouraged to increase the amount of exercise. Our results also suggest that it is also crucial to maintain the existing exercise levels, as decreased PA was associated with adverse mental health outcomes. Furthermore, our findings indicate that the greater the increase in MET among previously inactive individuals, the larger the reduction in the HR for depression, which could provide a rationale for health policy interventions targeting population with newly diagnosed T2DM.

Limitations

This study has several limitations. First, due to its retrospective design, the possibility of reverse causation cannot be fully excluded. However, this issue was minimized by clearly separating the exposure assessment period (newly diagnosed T2DM) from the outcome follow-up period (incident depression) and by excluding participants with prior depression. Second, the definition of depression relied on ICD-10 codes F32 and F33, which may not have captured all clinically relevant cases, such as those treated under different diagnostic codes. Nevertheless, the use of claim-based codes enabled identification of patients with physician-diagnosed depression while minimizing inclusion of subclinical symptoms [48]. Third, we did not differentiate between those requiring insulin and those treated with oral agents, nor assess the effect of this factor on glycemic control or comorbidity reduction. We attempted to mitigate this limitation by adjusting for key demographic, clinical, and comorbidity variables. Fourth, this study focused solely on depression as a mental health outcome. Other psychiatric disorders such as anxiety, bipolar disorder, or schizophrenia were not separately identified or excluded. Because depression is the most common psychiatric comorbidity among diabetic patients, we prioritized it for analytic clarity. Nevertheless, unmeasured comorbid psychiatric conditions may have influenced our findings, and this potential residual confounding should be considered in interpretation. Fifth, PA was measured only twice, within 2 years before and after diabetes diagnosis, limiting evaluation of long-term or time-varying patterns during the average 7.76-year follow-up. This may have introduced some exposure misclassification. However, focusing on early behavioral changes after diagnosis remains clinically meaningful, and the questionnaire used has demonstrated acceptable reliability and validity in Korean populations, supporting the robustness of this measurement despite temporal limitations. Sixth, subgroup analyses were exploratory in nature and should be interpreted with caution. Nevertheless, they provide additional context for understanding how the proposed associations may vary across different patient characteristics.

Conclusions

This research suggests that changes in PA among patients with newly diagnosed T2DM are associated with the risk of incident depression, regardless of the levels of previous exercise. Even mild exercise is protective for mental health in T2DM population, and increasing the level of PA is also helpful, whereas decreased activity elevates the risk of depression. These findings indicate that intervention for increasing or at least maintaining the level of PA would benefit the mental health of individuals with newly diagnosed T2DM.

SUPPLEMENTARY MATERIALS

NOTES

CONFLICTS OF INTEREST

No potential conflict of interest relevant to this article was reported.

AUTHOR CONTRIBUTIONS

Conception or design: S.P., B.K., S.J., S.M.P., D.H.L., S.J.C.

Acquisition, analysis, or interpretation of data: S.P., B.K., H. J.K., S.J.P., J.S., J.C.

Project administration and supervision: S.J., S.M.P., D.H.L., S.J.C.

Writing the original draft: all authors.

FUNDING

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (RS-2024-00440371 to Seogsong Jeong and NRF-2021R1A5A 2030333, to Dae Ho Lee). Also, this work was supported by the Gachon University Gil Medical Center (FRD2023-20 to Soo Jung Choi).

ACKNOWLEDGMENTS

Sangwoo Park, Hye Jun Kim, Sun Jae Park, Jihun Song, and Jina Chung received a BK21 FOUR education program scholarship. It was offered by the National Research Foundation of Korea.

Fig. 1.

Study flow for the study. T2DM, type 2 diabetes mellitus.

Fig. 2.

Association between changes in physical activity (metabolic equivalent of task [MET]-min/wk) after newly diagnosed type 2 diabetes mellitus and the risk of depression. Restricted cubic spline curve was constructed to examine the non-linear association between changes in physical activity (defined as the difference in MET-min/wk after versus before type 2 diabetes mellitus diagnosis) and the risk of depression. The solid line represents the adjusted hazard ratio, and the shaded area indicates the 95% confidence intervals derived from Cox proportional hazards regression models. Changes in physical activity were modeled as a continuous variable. A change of 0 MET-min/wk represents no difference in physical activity levels before and after type 2 diabetes mellitus diagnosis. Restricted Cubic Spline curve was generated using knots placed at intervals of 200 MET-min/wk. The model was adjusted for age, sex, household income, body mass index, systolic and diastolic blood pressure, fasting serum glucose, total cholesterol, alcohol intake, and Charlson comorbidity index.

REFERENCES

• 1. Iglay K, Hannachi H, Joseph Howie P, Xu J, Li X, Engel SS, et al. Prevalence and co-prevalence of comorbidities among patients with type 2 diabetes mellitus. Curr Med Res Opin 2016;32:1243-52.ArticlePubMed
• 2. Semenkovich K, Brown ME, Svrakic DM, Lustman PJ. Depression in type 2 diabetes mellitus: prevalence, impact, and treatment. Drugs 2015;75:577-87.ArticlePubMedPDF
• 3. Anderson RJ, Freedland KE, Clouse RE, Lustman PJ. The prevalence of comorbid depression in adults with diabetes: a meta-analysis. Diabetes Care 2001;24:1069-78.PubMed
• 4. Akyirem S, Ekpor E, Namumbejja Abwoye D, Batten J, Nelson LE. Type 2 diabetes stigma and its association with clinical, psychological, and behavioral outcomes: a systematic review and meta-analysis. Diabetes Res Clin Pract 2023;202:110774.ArticlePubMed
• 5. Chapman DP, Perry GS, Strine TW. The vital link between chronic disease and depressive disorders. Prev Chronic Dis 2005;2:A14.
• 6. Papelbaum M, Moreira RO, Coutinho W, Kupfer R, Zagury L, Freitas S, et al. Depression, glycemic control and type 2 diabetes. Diabetol Metab Syndr 2011;3:26.ArticlePubMedPMCPDF
• 7. Lustman PJ, Clouse RE. Depression in diabetic patients: the relationship between mood and glycemic control. J Diabetes Complications 2005;19:113-22.ArticlePubMed
• 8. Ciechanowski PS, Katon WJ, Russo JE. Depression and diabetes: impact of depressive symptoms on adherence, function, and costs. Arch Intern Med 2000;160:3278-85.ArticlePubMed
• 9. Schram MT, Baan CA, Pouwer F. Depression and quality of life in patients with diabetes: a systematic review from the European depression in diabetes (EDID) research consortium. Curr Diabetes Rev 2009;5:112-9.ArticlePubMedPMC
• 10. Eren I, Erdi O, Sahin M. The effect of depression on quality of life of patients with type II diabetes mellitus. Depress Anxiety 2008;25:98-106.ArticlePubMed
• 11. Schuch FB, Vancampfort D, Firth J, Rosenbaum S, Ward PB, Silva ES, et al. Physical activity and incident depression: a meta-analysis of prospective cohort studies. Am J Psychiatry 2018;175:631-48.ArticlePubMed
• 12. Pearce M, Garcia L, Abbas A, Strain T, Schuch FB, Golubic R, et al. Association between physical activity and risk of depression: a systematic review and meta-analysis. JAMA Psychiatry 2022;79:550-9.ArticlePubMedPMC
• 13. Kim SY, Park JH, Lee MY, Oh KS, Shin DW, Shin YC. Physical activity and the prevention of depression: a cohort study. Gen Hosp Psychiatry 2019;60:90-7.ArticlePubMed
• 14. Dishman RK, McDowell CP, Herring MP. Customary physical activity and odds of depression: a systematic review and meta-analysis of 111 prospective cohort studies. Br J Sports Med 2021;55:926-34.ArticlePubMed
• 15. Shin HY, Park S, Park SM. Gaps in health behaviours and use of preventive services between patients with diabetes and the general population: a population-based cross-sectional study. BMJ Open 2018;8:e017937.ArticlePubMedPMC
• 16. Lee J, Callaghan T, Ory M, Zhao H, Bolin J. Difference in the risk of depressive symptoms associated with physical activity in persons with diabetes: across age, gender, and race/ethnicity. J Affect Disord 2020;269:108-16.ArticlePubMed
• 17. Lee J, Kim J, Chow A, Piatt JA. Different levels of physical activity, physical health, happiness, and depression among older adults with diabetes. Gerontol Geriatr Med 2021;7:2333721421995623.ArticlePubMedPMCPDF
• 18. Narita Z, Inagawa T, Stickley A, Sugawara N. Physical activity for diabetes-related depression: a systematic review and meta-analysis. J Psychiatr Res 2019;113:100-7.ArticlePubMed
• 19. Kim K, Choi S, Hwang SE, Son JS, Lee JK, Oh J, et al. Changes in exercise frequency and cardiovascular outcomes in older adults. Eur Heart J 2020;41:1490-9.ArticlePubMedPDF
• 20. Lee J, Lee C, Min J, Kang DW, Kim JY, Yang HI, et al. Development of the Korean Global Physical Activity Questionnaire: reliability and validity study. Glob Health Promot 2020;27:44-55.ArticlePubMedPDF
• 21. Bae WK, Cho B, Son KY. Validity and reliability of the Korean version of neighborhood physical activity questionnaire. Korean J Fam Med 2015;36:135-40.ArticlePubMedPMC
• 22. Jeong SW, Kim SH, Kang SH, Kim HJ, Yoon CH, Youn TJ, et al. Mortality reduction with physical activity in patients with and without cardiovascular disease. Eur Heart J 2019;40:3547-55.ArticlePubMedPMCPDF
• 23. Kim SR, Choi S, Kim K, Chang J, Kim SM, Cho Y, et al. Association of the combined effects of air pollution and changes in physical activity with cardiovascular disease in young adults. Eur Heart J 2021;42:2487-97.ArticlePubMedPDF
• 24. Cho S, Park S, Lee SK, Oh SN, Kim KH, Ko A, et al. Associations of changes in alcohol consumption on the risk of depression/suicide among initial nondrinkers. Depress Anxiety 2024;2024:7560390.ArticlePubMedPMC
• 25. Yu J, Park S, Jeong S, Ko A, Lee J, Han S, et al. Association of cholecystectomy with short-term and long-term risks of depression and suicide. Sci Rep 2025;15:6557.ArticlePubMedPMCPDF
• 26. Kim EY, Lee YB, Lee GN, Han K, Jeong SM. Associations between body weight change and incidence of major depressive disorder in patients with type 2 diabetes mellitus: a nationwide longitudinal follow-up cohort study of 1.1 million. Psychol Med 2024;54:2380-8.ArticlePubMed
• 27. An JH, Han KD, Jeon HJ. Higher metabolic variability increases the risk of depressive disorder in type 2 diabetes mellitus: a longitudinal nationwide cohort study. Front Psychiatry 2023;14:1217104.ArticlePubMedPMC
• 28. Joseph MS, Tincopa MA, Walden P, Jackson E, Conte ML, Rubenfire M. The impact of structured exercise programs on metabolic syndrome and its components: a systematic review. Diabetes Metab Syndr Obes 2019;12:2395-404.PubMedPMC
• 29. Health Quality Ontario. Structured education and neuromuscular exercise program for hip and/or knee osteoarthritis: a health technology assessment. Ont Health Technol Assess Ser 2018;18:1-110.
• 30. Lincoln AK, Shepherd A, Johnson PL, Castaneda-Sceppa C. The impact of resistance exercise training on the mental health of older Puerto Rican adults with type 2 diabetes. J Gerontol B Psychol Sci Soc Sci 2011;66:567-70.ArticlePubMedPMC
• 31. Saiiari A, Moslehi M. Effects of crawl swimming on depression in type 2 diabetic patients. Procedia Soc Behav Sci 2011;30:2156-60.Article
• 32. Pibernik-Okanovic M, Hermanns N, Ajdukovic D, Kos J, Prasek M, Sekerija M, et al. Does treatment of subsyndromal depression improve depression-related and diabetes-related outcomes?: a randomised controlled comparison of psychoeducation, physical exercise and enhanced treatment as usual. Trials 2015;16:305.PubMedPMC
• 33. Castan-Laurell I, Dray C, Attane C, Duparc T, Knauf C, Valet P. Apelin, diabetes, and obesity. Endocrine 2011;40:1-9.ArticlePubMedPDF
• 34. Castan-Laurell I, Dray C, Knauf C, Kunduzova O, Valet P. Apelin, a promising target for type 2 diabetes treatment? Trends Endocrinol Metab 2012;23:234-41.ArticlePubMed
• 35. Yu J, Yau SY. Apelin as a novel mediator of antidepressant effects of physical exercise. Int J Neuropsychopharmacol 2025;28:pi205.ArticlePDF
• 36. Tian Y, Wang R, Liu L, Zhang W, Liu H, Jiang L, et al. The regulatory effects of the apelin/APJ system on depression: a prospective therapeutic target. Neuropeptides 2023;102:102382.ArticlePubMed
• 37. Packard AE, Egan AE, Ulrich-Lai YM. HPA axis interactions with behavioral systems. Compr Physiol 2016;6:1897-934.ArticlePubMedPMCPDF
• 38. Champaneri S, Wand GS, Malhotra SS, Casagrande SS, Golden SH. Biological basis of depression in adults with diabetes. Curr Diab Rep 2010;10:396-405.ArticlePubMedPDF
• 39. Roy M, Collier B, Roy A. Hypothalamic-pituitary-adrenal axis dysregulation among diabetic outpatients. Psychiatry Res 1990;31:31-7.ArticlePubMed
• 40. Zschucke E, Renneberg B, Dimeo F, Wustenberg T, Strohle A. The stress-buffering effect of acute exercise: evidence for HPA axis negative feedback. Psychoneuroendocrinology 2015;51:414-25.ArticlePubMed
• 41. Puterman E, O’Donovan A, Adler NE, Tomiyama AJ, Kemeny M, Wolkowitz OM, et al. Physical activity moderates effects of stressor-induced rumination on cortisol reactivity. Psychosom Med 2011;73:604-11.ArticlePubMedPMC
• 42. Bouwman V, Adriaanse MC, van ‘t Riet E, Snoek FJ, Dekker JM, Nijpels G. Depression, anxiety and glucose metabolism in the general dutch population: the new Hoorn study. PLoS One 2010;5:e9971.ArticlePubMedPMC
• 43. Dunstan DW, Salmon J, Owen N, Armstrong T, Zimmet PZ, Welborn TA, et al. Physical activity and television viewing in relation to risk of undiagnosed abnormal glucose metabolism in adults. Diabetes Care 2004;27:2603-9.ArticlePubMedPDF
• 44. Laake JP, Stahl D, Amiel SA, Petrak F, Sherwood RA, Pickup JC, et al. The association between depressive symptoms and systemic inflammation in people with type 2 diabetes: findings from the South London Diabetes Study. Diabetes Care 2014;37:2186-92.ArticlePubMedPDF
• 45. Frank P, Kaushal A, Poole L, Lawes S, Chalder T, Cadar D. Systemic low-grade inflammation and subsequent depressive symptoms: is there a mediating role of physical activity? Brain Behav Immun 2019;80:688-96.ArticlePubMed
• 46. Chao H, Lao I, Hao L, Lin C. Association of body image and health beliefs with health behaviors in patients with diabetes: a cross-sectional study. Diabetes Educ 2012;38:705-14.ArticlePubMedPDF
• 47. Guicciardi M, Lecis R, Anziani C, Corgiolu L, Porru A, Pusceddu M, et al. Type 2 diabetes mellitus, physical activity, exercise self-efficacy, and body satisfaction: an application of the transtheoretical model in older adults. Health Psychol Behav Med 2014;2:748-58.ArticlePubMed
• 48. Fiest KM, Jette N, Quan H, St Germaine-Smith C, Metcalfe A, Patten SB, et al. Systematic review and assessment of validated case definitions for depression in administrative data. BMC Psychiatry 2014;14:289.ArticlePubMedPMCPDF

Figure & Data

References

Citations

Citations to this article as recorded by  

About this article

Park S, Kim B, Kim HJ, Park SJ, Song J, Chung J, Jeong S, Park SM, Lee DH, Choi SJ. Association between Changes in Physical Activity and Incident Depression among Patients with Newly Diagnosed Type 2 Diabetes Mellitus. Diabetes Metab J. 2026 Feb 23. doi: 10.4093/dmj.2025.0766. Epub ahead of print.

Received: Aug 17, 2025; Accepted: Dec 18, 2025

DOI: https://doi.org/10.4093/dmj.2025.0766.

Download citation ↓