GRAPHICAL ABSTRACT

Highlights

• Higher CRF linked to lower diabetic nephropathy risk in T2DM patients.

• No significant associations were found between CRF and retinopathy or neuropathy.

• Triglycerides and white blood cell count mediated about 20% of the observed association.

• Genetic predisposition did not modify the protective effect of CRF on nephropathy.

INTRODUCTION

Diabetes represents one of the most rapidly growing public health challenges worldwide. In 2024, approximately 589 million adults worldwide were living with diabetes, and this number is projected to surge to 853 million by 2050, with type 2 diabetes mellitus (T2DM) accounting for the majority of cases [1]. Diabetic microvascular complications, including diabetic neuropathy, diabetic retinopathy, and diabetic kidney disease, affect more than 50% of patients with T2DM [2,3]. These complications frequently lead to severe outcomes, such as end-stage renal disease, vision loss, neuropathic pain, and paresthesia, imposing a significant health and economic burdens on individuals, families, and health systems [4]. Given the high prevalence and serious consequences, there is an urgent public health need for identification of modifiable risk factors to prevent or postpone microvascular complication development in patients with T2DM.

Cardiorespiratory fitness (CRF), defined as the ability of the circulatory and respiratory systems to deliver oxygen to skeletal muscles during sustained physical activity, is an integrated physiological marker [5]. Low CRF is a well-established, strong, and independent predictor of diabetes, cardiovascular diseases, a range of cancers, and mortality [6-8]. Moreover, improving CRF is correlated with a substantial enhancement in survival rates [9]. In patients with T2DM, low CRF is a well-recognized risk factor for macrovascular disease, such as cardiovascular disease and stroke [10-13]. Higher CRF is associated with better endothelial function, lower inflammation, and more favorable lipid profiles, which are protective factors for microvascular health [14,15]. However, the relationship between CRF and the incidence of microvascular complications in patients with T2DM remains insufficiently investigated. Moreover, while studies have linked CRF with a range of blood biomarkers, including lipid profiles, liver function biomarkers, and systemic inflammatory factors, whether and the extent of their mediation effect in the association between CRF and diabetic microvascular complications is unclear [7,16,17].

To address this knowledge gap, we aimed to evaluate the association between CRF and the incidence of specific microvascular complications in T2DM patients using a large prospective cohort study of the UK Biobank. We also examined whether genetic predisposition to diabetic complications modifies this relationship. Furthermore, we also investigated the potential mediating roles of key circulating biomarkers.

METHODS

Study design and population

This prospective analysis utilized data from the UK Biobank, a large community-based cohort that recruited over 500,000 participants aged 37 to 73 years between March 2006 and October 2010 from across the United Kingdom to investigate common diseases in middle-aged and older adults. Participants attended one of 22 assessment centers across England, Scotland, and Wales, where they completed nurse-led touchscreen questionnaires, physical measurements, and biological sample collections at the baseline. The specific methods of data collection have been described previously [18,19].

This study utilized a subsample of 65,421 participants who completed a submaximal cycle ergometer test at baseline. Prevalent T2DM at baseline was identified using UK Biobank algorithms described by Eastwood et al. [20] which incorporated hospital inpatient records, self-reported medical history, and medication use, achieving an accuracy of 96%. Glycosylated hemoglobin (HbA1c) was also considered as part of the algorithm, with a cutoff of ≥48 mmol/mol (6.5%) used to define diabetes. Among the 65,421 participants with available data on CRF, we excluded those without diabetes at baseline (n=61,681), those diagnosed with type 1 diabetes mellitus (n=32), and those with prevalent microvascular disease at baseline (n=606). Finally, 3,102 patients were included in the analysis (Supplementary Fig. 1).

Ethics statement

All participants provided informed consent, and the study was approved by the North West–Haydock Research Ethics Committee (16/NW/0274). Data for the present study were retrieved from the UK Biobank under application ID 121022.

Measurements of CRF

CRF was assessed as maximal oxygen uptake (VO_2max, mL/min/kg) using a submaximal cycle ergometer test without direct gas analysis. Before the test, participants completed a risk stratification questionnaire, after which an individualized test protocol was assigned. The test consisted of four phases: a 15-second pretest phase, a 2-minute constant workload phase, a 4-minute incremental phase during which workload increased up to 35% of the estimated maximum for low-risk individuals and 50% for minimal-risk individuals, and finally, a 1-minute recovery phase. Throughout the cycling phases, participants maintained a cadence of 60 revolutions per minute, and electrocardiograms were continuously recorded. More details can be found at https://biobank.ndph.ox.ac.uk/showcase/label.cgi?id=267.

VO_2max was estimated using the prediction equation developed by Gonzales et al. [21], which integrates heart rate response features obtained during both the flat and ramped phases of the test, adjusts for the individualized ramp rate, and extrapolates the steady-state HR-work-rate relationship to agepredicted maximal heart rate [21]. This equation has been validated against directly measured VO₂max in UK Biobank participants, showing Pearson coefficients ranging from 0.68 to 0.74, with no significant mean bias. CRF was categorized into a three-level categorical variable: low (<20th percentile), moderate (20th to 60th percentiles), and high (>60th percentile) within sex and 10-year age groups (e.g., 50–60 years), in accordance with standard approaches to the analysis of CRF [22]. The VO_2max values were transformed into maximal metabolic equivalent of task (MET, 1 MET=3.5 mL/kg/min) when CRF was evaluated as a continuous variable.

Assessment of the circulating biomarkers

Blood samples were collected from consenting participants at recruitment, separated by components and stored at UK Biobank (–80°C and liquid nitrogen) until analysis. Blood biomarkers were externally validated with stringent quality control in the UK Biobank. We selected a panel of potential mediating biomarkers based on their biological pathways previously implicated in linking CRF to microvascular diseases, including lipid profile (total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglycerides, apolipoprotein A, apolipoprotein B, and lipoprotein A), liver function (alanine aminotransferase, alkaline phosphatase, aspartate aminotransferase, gamma-glutamyl transferase [GGT], total bilirubin, total protein, and albumin), and inflammation (C-reactive protein, and white blood cell [WBC] count).

Outcome ascertainment

Outcomes of interest were diabetic nephropathy, diabetic neuropathy, and diabetic retinopathy, which were identified using cumulative hospital inpatient records and death record linkage to national death registries. The definition was described according to the 10th revisions of the International Classification of Diseases (ICD-10) and self-reported data fields (Supplementary Table 1). We compared the date of the first diagnosis with the baseline date to distinguish between baseline and incident complications. Participants with a first recorded diagnosis after the baseline visit date were considered incident cases. At the time of analysis, hospital admission data were available until 31 October 2022 for England, 31 August 2022 for Scotland, and 31 May 2022 for Wales.

Calculation of polygenic risk score for diabetic complications

To assess genetic predisposition to diabetic complications, we employed the multi-polygenic risk score (multiPRS) model previously developed and validated by Tremblay et al. [23]. This model integrates 10 weighted polygenic risk scores (wPRS) that capture 598 independent single nucleotide polymorphisms (SNPs) associated with major risk factors and complications related to T2DM, including, obesity, blood pressure, renal function, lipids, and cardiovascular events. Detailed information of these SNPs is listed in Supplementary Table 2. In the Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation (ADVANCE) derivation, the 10 wPRS, along with the first principal component of ethnicity, sex, age at onset and diabetes duration were included into one logistic regression model to develop the multiPRS model, which achieved an area under the receiver operating characteristic curve (AUC) of 0.67 for predicting combined microvascular or macrovascular complications of T2DM; when applied to UK Biobank, the AUC for predicting incident low estimated glomerular filtration rate (eGFR) was 0.67 to 0.70 [23]. To apply this multiPRS model in the present study, we extracted genotype dosages for 598 SNPs from the UK Biobank imputed genetic data [24]. Each wPRS was calculated by summing the number of effect alleles (coded 0, 1, or 2) weighted by their respective effect sizes derived from genome-wide association study. Then, we applied the logistic regression from the ADVANCE-derived multiPRS model to our UK Biobank data, which generated a continuous multiPRS value for each participant. Consistent with Tremblay et al. [23], we defined a high-risk group as those in the top 30% of the multiPRS model, and a low-risk group as those in the remaining 70%.

Covariates

Potential confounders were included in the analysis, encompassing sociodemographic data, lifestyle behaviors, physical assessments, and medical history. Sociodemographic variables included age (years, continuous), sex (female/male), ethnicity (White/others), education level (university or college degree/others), and the Townsend deprivation index (continuous), which was used to quantify socioeconomic status. Lifestyle behaviors covered smoking status (never, former, or current), alcohol intake (grams per day, continuous), and physical activity level (total moderate-to-vigorous activity in MET-minutes/week, continuous) and dietary habits. Diet habits were assessed using a healthy diet score (continuous), where higher scores indicated dietary patterns more aligned with cardiovascular and metabolic health priorities [25]. Trained nurses measured height and weight at the assessment center, and body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared (continuous). Systolic and diastolic blood pressure were measured using automated devices during the baseline assessment (continuous). Medical history included diabetes duration (<1, 1–4, 5–9, ≥10 years), HbA1c level (<53, ≥53 mmol/mol [7%]), eGFR (calculated using the Chronic Kidney Disease Epidemiology Collaboration 2021 creatinine equation, continuous), diabetes medication use (none, only oral medication, only insulin, and oral medication plus insulin), lipid-lowering treatment (yes or no), antihypertensive medication treatment (yes or no), and use of aspirin (yes or no). Missing data for covariates (missing rate range, 0.13% to 19.47%) (Supplementary Table 3) were handled using multiple imputation by chained equations to minimize potential bias.

Statistical analysis

Categorical variables were presented as percentages, and normal continuous variables as mean and standard deviation, while non-normal variables as medians and interquartile ranges (IQRs). Continuous variables were assessed by one-way analysis of variance (ANOVA) or the Kruskal-Wallis test. Categorical data were assessed using the chi-square test.

Cox proportional hazards regression models were used to evaluate the associations of CRF with incident diabetic microvascular complications, and the results were reported as hazard ratios (HRs) and 95% confidence intervals (CIs). Three adjusted models were built. Model 1 was unadjusted. Model 2 was adjusted for age, sex, ethnicity, education level, Townsend deprivation index, smoking status, daily alcohol intake, healthy diet score, physical activity level, and BMI. Model 3 was further adjusted for diabetes duration, HbA1c level, diabetes medication use, lipid-lowering treatment, antihypertensive medication treatment, and use of aspirin. The proportional hazards assumption was assessed using the Schoenfeld residuals method, and no violation was detected (P=0.20 for global testing). Restricted cubic splines (RCS) with four knots (5th, 35th, 65th, and 95th percentiles) analysis for smooth curve fitting was used to visually illustrate the dose-response relationship. The association between the genetic risk and the incidence of microvascular disease was evaluated using Cox proportional hazards models. Then, we tested for effect modification by genetic predisposition by introducing a product term between CRF and the genetic risk group into the fully adjusted Cox model, with significance determined by the likelihood ratio test. To assess the joint associations of CRF and genetic risk, we created a combined variable defining four exposure groups. For this analysis, participants were categorized into four groups: moderate/high CRF+low risk; moderate/high CRF+high risk; low CRF+low risk; and low CRF+high risk. HRs across these groups were then estimated using Cox proportional hazards models, with fit+low risk group as the reference.

According to predefined mediation principles, biomarkers associated with both CRF and the incident microvascular complications were selected for the mediation analysis [26]. These criteria were tested with multivariable-adjusted linear regression and Cox regression. For the mediation analyses, we used the CMAverse R package. For each mediator, total effect (TE), indirect effect (IE), and direct effect were calculated with a combination of the mediator and outcome models adjusting for all the covariates included in model 3. The proportion mediated was calculated as IE divided by TE, and 95% CIs were derived from nonparametric bootstrap resamples.

Stratified analyses were conducted by age (≤60, >60 years), sex (female, male), BMI (<30, ≥30 kg/m²), diabetes duration (≤4, >4 years), use of diabetes medication (yes, no), and HbA1c (<53, ≥53 mmol/mol), and an interaction term with sleep duration category was tested to investigate the potential effect modification. Sensitivity analyses were performed to test the robustness of the results. First, we performed the analysis after excluding cases occurring within the first 3 years of follow-up to minimize the potential reverse causation. Second, Fine and Gray proportional subdistribution hazards regression models were constructed to account for the possible competing risk of death. Third, we performed the analysis by excluding patients with missing values for any covariates included in model 3. Fourth, we further adjusted for baseline eGFR, systolic blood pressure, and diastolic blood pressure based on model 3. To assess the potential influence of lipid-lowering therapy on the mediating role of triglycerides, we conducted sensitivity mediation analyses stratified by baseline use of lipid-lowering medications.

Statistical analyses were performed in the R version 4.4.3 (R Foundation for Statistical Computing, Vienna, Austria). A two-tailed P<0.05 was considered statistically significant (two-sided tests).

RESULTS

Baseline characteristics of the study population

This study included 3,102 participants with T2DM, of whom 61.6% were male and the mean age was 59.2 years. 1,335 (43.0%) participants were in the low-CRF group, 1,237 (39.9%) in the moderate-CRF group, and 530 (17.1%) in the high-CRF group, respectively. The baseline characteristics of these participants are presented in Table 1. Compared to participants in the low-CRF group, those with moderate or high CRF were more likely to be less deprived, highly educated, with a lower BMI, more alcohol intake, better diet quality, and higher levels of physical activity. They were also more prone to have lower HbA1c levels, less use of diabetes medication, antihypertensive medication, lipid-lowering treatment, and aspirin at baseline.

Association between CRF and risk of microvascular complications

During a median follow-up period of 12.47 years (IQR, 12.35 to 12.63) years, 331 cases of diabetic nephropathy, 268 cases of diabetic retinopathy, and 88 cases of diabetic neuropathy were documented. The associations between CRF and microvascular disease are presented in Table 2. Cox regression analyses revealed that higher CRF was significantly associated with a lower risk of diabetic nephropathy. After adjusting for the confounding factors, compared to participants with low CRF, those with moderate CRF had a 22% lower risk of diabetic nephropathy (HR, 0.78; 95% CI, 0.61 to 0.99), while those with high CRF had a 45% lower risk (HR, 0.55; 95% CI, 0.36 to 0.85). When analyzed as a continuous variable, each 1-MET increment in CRF was associated with an 11% lower risk of diabetic nephropathy (HR, 0.89; 95% CI, 0.81 to 0.98), and RCS analysis revealed a significant linear relationship (P non-linear= 0.107), as presented in Fig. 1. However, no statistically significant associations were observed between CRF and the risks of diabetic retinopathy or neuropathy.

Interaction and joint association of CRF and genetic risk

As shown in Supplementary Table 4, individuals with a high genetic risk showed a 1.57-fold higher risk of diabetic nephropathy incidence compared to those with a low genetic risk (HR, 1.57; 95% CI, 1.21 to 2.02). No significant interaction between CRF and genetic susceptibility for diabetic nephropathy risk was found (P for interaction=0.543) (Supplementary Table 5). The joint association of CRF group and genetic risk group with diabetic nephropathy is shown in Table 3. Compared with the moderate/high CRF+low risk group, all other groups showed markedly higher risks. The highest risk was observed in the low CRF+high risk group (HR, 2.03; 95% CI, 1.42 to 2.89), followed by moderate/high+high-risk group (HR, 1.73; 95% CI, 1.22 to 2.46), and low CRF+low risk group (HR, 1.55; 95% CI, 1.12 to 2.14).

Mediation analysis

All biomarkers were significantly associated with CRF except for albumin, apolipoprotein A, and lipoprotein A (Supplementary Table 6). Among these, three biomarkers were significantly associated with the risk of diabetic nephropathy, namely GGT, triglycerides and WBC (Supplementary Table 7). Mediation analysis suggested that triglyceride and WBC explained 7.46% and 12.88%, respectively, of the association between higher CRF and lower risk of diabetic nephropathy (Fig. 2).

Secondary analysis and sensitivity analysis

Consistent results were observed when the analyses were stratified by age, sex, BMI, diabetes duration, use of diabetes medication, and HbA1c level. No significant interaction was observed between CRF and the stratified factors on the outcomes (Supplementary Table 8). In the sensitivity analyses, the results were generally robust when excluding individuals who developed diabetic microvascular complications within the first 3 years of follow-up, using competing risk models accounting for death, excluding patients with missing covariate data, or additionally adjusting for baseline eGFR, systolic blood pressure, and diastolic blood pressure (Supplementary Table 9). In sensitivity analyses stratified by lipid-lowering medication use, the proportion of the association mediated by triglycerides was 6.92% among participants using lipid-lowering medications and 7.91% among non-users, both similar to the overall estimate, although the 95% CI for the non-user group included 1 (Supplementary Fig. 2).

DISCUSSION

In this large prospective cohort study of 3,102 patients with T2DM from the UK Biobank, we found that higher CRF was associated with a lower risk of incident diabetic nephropathy. Specifically, compared to participants with low CRF, those with moderate and high CRF exhibited 22% and 45% lower risks of diabetic nephropathy, and a clear dose-response relationship was observed, with each 1-MET increment in CRF associated with an 11% lower risk. However, the relationships with retinopathy and neuropathy did not reach significance. Furthermore, we found that genetic predisposition to complications did not modify the association between CRF and diabetic nephropathy, and mediation analysis identified that triglyceride levels and WBC explained approximately 20% of the effect of CRF on diabetic nephropathy.

Previous studies have predominantly focused on the association between CRF and macrovascular outcomes in patients with T2DM, while the evidence on CRF and microvascular complications is limited and inconclusive. The Cooper Center Longitudinal Study showed that higher CRF was associated with a lower risk of chronic kidney disease (CKD) in generally healthy adults, including those who developed diabetes later in life, but these findings were constrained by the predominant inclusion of adults over 65 years old, and the specific number of T2DM cases was not reported [27]. However, another cohort study involving only 74 men with T2DM found no significant relationship between CRF and CKD, but the limited sample size and exclusive focus on male participants likely resulted in low statistical power [28]. To our knowledge, the present study represents the largest longitudinal study investigating the association between CRF and microvascular disease among patients with T2DM. Our results provide strong evidence that CRF might be an independent predictor of diabetic nephropathy, thereby extending its prognostic value beyond macrovascular endpoints.

We further explored whether genetic predisposition to diabetic complications modifies the relationship between CRF and microvascular outcomes. Although individuals with a high genetic risk had a 1.57-fold higher risk of diabetic nephropathy compared to those with low risk, no significant interaction was observed between CRF and genetic susceptibility. This suggests that the protective effect of CRF is consistent across different genetic risk profiles. In the joint analysis, participants with both low CRF and high genetic risk had the highest incidence of nephropathy, reinforcing that improving CRF may benefit even those at high genetic risk.

Mediation analysis provided mechanistic insights by identifying triglycerides and WBC as potential mediators in the association between CRF and diabetic nephropathy, accounting for 7.46% and 12.88% of the observed association, respectively. These findings align with the known roles of dyslipidemia and chronic inflammation in the pathogenesis of diabetic kidney disease, suggesting that CRF may exert part of its protective effect through improving lipid profiles and reducing systemic inflammation [29,30].

The absence of statistically significant associations for retinopathy and neuropathy warrants careful consideration. The observed incidence rates of retinopathy and neuropathy in our cohort were lower than expected based on general epidemiologic data, which might reduce the statistical power to detect potentially modest associations, particularly for neuropathy which had the fewest events [31,32]. Besides, pathophysiological differences among microvascular complications may explain the discrepant findings [33,34]. Nonetheless, the direction of the point estimates for retinopathy and neuropathy was consistently inverse, suggesting that a modest protective effect of CRF remains plausible. Future studies with larger sample sizes or more sensitive outcome ascertainment methods are needed to clarify these relationships.

Currently, professional organizations such as the American Heart Association advocates for the formal assessment of CRF as a ‘clinical vital sign’ and recommend its integration into clinical decision-making for a wide range of chronic conditions [5]. Our study provides further evidence that CRF offers valuable prognostic information on diabetic nephropathy risk in T2DM. Though about half of the variation in CRF is heritable and it is also influenced by non-modifiable factors such as age, sex and underlying disease states, CRF largely remains a modifiable risk factor [35,36]. Increased physical activity and structured exercise training are the primary evidence-based methods to improve CRF [37,38]. These interventions effectively target hyperglycemia, insulin resistance, dyslipidemia, endothelial dysfunction, oxidative stress, and inflammation, all of which are involved in the pathophysiology of microvascular injury [39,40]. Emerging evidence supports the clinical benefit of exercise on improving kidney function. A 6-month aerobic-exercise program in overweight/obese men with T2DM reduced the prevalence of microalbuminuria and the level of serum N-acetyl-β-D-glycosaminidase, an early marker of tubular injury [41]. Similarly, a 12-week supervised exercise intervention in patients with T2DM and stage 2–3 CKD yielded a 6%-12% increase in eGFR [42].

The strengths of this study include its prospective design, large sample size, long follow-up period, comprehensive consideration of covariates, and integration of genetic and biomarker data. Several limitations should be acknowledged. First, although the submaximal cycle test has been validated against cardiopulmonary exercise testing, it remains an estimate rather than a direct measurement of VO_2max, potentially introducing non-differential misclassification that would bias associations toward the null. Second, reliance on hospital inpatient records and ICD-10 codes may underestimate the incidence of outcomes, particularly milder or earlier stages managed in outpatient settings; although adding urine albumin-to-creatinine ratio could have improved sensitivity, it was available for about 50% of the diabetes sub-cohort and lacked repeated measures. Third, residual confounding by unmeasured factors, such as genetic predisposition, cannot be entirely excluded due to the inherent limitations of observational study designs. Fourth, while the mediation effects of triglycerides were similar across subgroups defined by lipid-lowering medication use, the smaller sample size in the non-user subgroup limited the statistical power, which requires validation in larger studies. Finally, since participants in the UK Biobank are predominantly of European descent, the generalizability of our results to other populations may be limited, underscoring the need for further research in diverse ethnic and racial groups.

In conclusion, this prospective cohort study demonstrated that higher CRF is independently associated with a significantly reduced risk of developing diabetic nephropathy in individuals with T2DM. The association was not modified by genetic risk, and was partially mediated by triglycerides and WBC. These findings highlight CRF as a potent and modifiable risk factor for microvascular disease, and support the integration of CRF assessment into clinical practice for risk stratification and development of personalized exercise prescriptions, ultimately helping to reduce the burden of diabetic kidney disease in this high-risk population.

SUPPLEMENTARY MATERIALS

NOTES

CONFLICTS OF INTEREST

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

AUTHOR CONTRIBUTIONS

Conception or design: H.C., H.G., Y.C.

Acquisition, analysis, or interpretation of data: all authors.

Drafting the work or revising: A.X., H.Z., Y.C.

Final approval of the manuscript: all authors.

FUNDING

The study was financially supported by the grants from National Natural Science Foundation of China (82270333; 81600260; 82300956, 82200148, and 82304910), the Natural Science Foundation of Guangdong Province, China (2024A1515013067; 20 22A1515012358), Guangzhou Science and Technology Program (2023A04J1087, 2024B03J1344), the Natural Science Foundation of Hubei Province, China (Grant No. 2024AFB925), Highlevel Talents Introduction Plan of Guangdong Provincial People’s Hospital (KY012023007), the Guangdong Provincial People’s Hospital Supporting Fund for Talent Program (No. KJ012020629), Clinical Research Special Fund of Guangdong Medical Association (2024HY-A6002), and National Science and Technology Innovation Major Project-Research Project on Prevention and Treatment of Cancer, Cardiovascular, Respiratory and Metabolic Diseases (2023ZD0504202; 2023ZD0504204).

ACKNOWLEDGMENTS

The author expresses gratitude to the UK Biobank participants for their involvement and contributions to the research. This study utilized resources from the UK Biobank under application number 644424520.

Fig. 1.

Dose-response associations between cardiorespiratory fitness (CRF) and risk of diabetic nephropathy (A), retinopathy (B), and neuropathy (C). HR, hazard ratio; CI, confidence interval.

Fig. 2.

Association of cardiorespiratory fitness with diabetic nephropathy by triglycerides (A) and white blood cell count (B). HR, hazard ratio; CI, confidence interval.

REFERENCES

• 1. International Diabetes Federation. IDF diabetes atlas 2025. Available from: https://diabetesatlas.org/resources/idf-diabetesatlas-2025 (cited 2026 Mar 27).
• 2. Chatterjee S, Khunti K, Davies MJ. Type 2 diabetes. Lancet 2017;389:2239-51.ArticlePubMed
• 3. Litwak L, Goh SY, Hussein Z, Malek R, Prusty V, Khamseh ME. Prevalence of diabetes complications in people with type 2 diabetes mellitus and its association with baseline characteristics in the multinational A1chieve study. Diabetol Metab Syndr 2013;5:57.ArticlePubMedPMCPDF
• 4. Chen HY, Kuo S, Su PF, Wu JS, Ou HT. Health care costs associated with macrovascular, microvascular, and metabolic complications of type 2 diabetes across time: estimates from a population-based cohort of more than 0.8 million individuals with up to 15 years of follow-up. Diabetes Care 2020;43:1732-40.ArticlePubMedPMCPDF
• 5. Ross R, Blair SN, Arena R, Church TS, Despres JP, Franklin BA, et al. Importance of assessing cardiorespiratory fitness in clinical practice: a case for fitness as a clinical vital sign: a scientific statement from the American Heart Association. Circulation 2016;134:e653-99.ArticlePubMed
• 6. Chen Y, Yang H, Li D, Zhou L, Lin J, Yin X, et al. Association of cardiorespiratory fitness with the incidence and progression trajectory of cardiometabolic multimorbidity. Br J Sports Med 2025;59:306-15.ArticlePubMed
• 7. Kunutsor SK, Kaminsky LA, Lehoczki A, Laukkanen JA. Unraveling the link between cardiorespiratory fitness and cancer: a state-of-the-art review. Geroscience 2024;46:5559-85.ArticlePubMedPMCPDF
• 8. Zhou H, Wang Y, Song X, Xu T, Xia C, Guo Y, et al. Independent and joint association of fat-to-muscle mass ratio and cardiorespiratory fitness with type 2 diabetes mellitus incidence: a prospective cohort study. Diabetes Obes Metab 2026;28:463-71.PubMed
• 9. Clausen JS, Marott JL, Holtermann A, Gyntelberg F, Jensen MT. Midlife cardiorespiratory fitness and the long-term risk of mortality: 46 years of follow-up. J Am Coll Cardiol 2018;72:987-95.PubMed
• 10. Wills AC, Vazquez Arreola E, Olaiya MT, Curtis JM, Hellgren MI, Hanson RL, et al. Cardiorespiratory fitness, BMI, mortality, and cardiovascular disease in adults with overweight/obesity and type 2 diabetes. Med Sci Sports Exerc 2022;54:994-1001.ArticlePubMedPMC
• 11. Pierre-Louis B, Aronow WS, Yoon JH, Ahn C, DeLuca AJ, Weiss MB, et al. Incidence of myocardial infarction or stroke or death at 47-month follow-up in patients with diabetes and a predicted exercise capacity 85% during an exercise treadmill sestamibi stress test. Prev Cardiol 2010;13:14-7.PubMed
• 12. Church TS, LaMonte MJ, Barlow CE, Blair SN. Cardiorespiratory fitness and body mass index as predictors of cardiovascular disease mortality among men with diabetes. Arch Intern Med 2005;165:2114-20.ArticlePubMed
• 13. Kokkinos P, Myers J, Nylen E, Panagiotakos DB, Manolis A, Pittaras A, et al. Exercise capacity and all-cause mortality in African American and Caucasian men with type 2 diabetes. Diabetes Care 2009;32:623-8.ArticlePubMedPMCPDF
• 14. Montero D. The association of cardiorespiratory fitness with endothelial or smooth muscle vasodilator function. Eur J Prev Cardiol 2015;22:1200-11.ArticlePubMedPDF
• 15. Lin X, Zhang X, Guo J, Roberts CK, McKenzie S, Wu WC, et al. Effects of exercise training on cardiorespiratory fitness and biomarkers of cardiometabolic health: a systematic review and meta-analysis of randomized controlled trials. J Am Heart Assoc 2015;4:e002014.ArticlePubMedPMC
• 16. Parto P, Lavie CJ, Swift D, Sui X. The role of cardiorespiratory fitness on plasma lipid levels. Expert Rev Cardiovasc Ther 2015;13:1177-83.ArticlePubMed
• 17. Florea VV, Gajjar P, Huang S, Tang J, Zhao S, Davenport M, et al. Hepatic steatosis and fibrosis, cardiorespiratory fitness, and metabolic mediators in the community. Liver Int 2025;45:e16147.ArticlePubMedPMC
• 18. Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, et al. UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med 2015;12:e1001779.ArticlePubMedPMC
• 19. Caleyachetty R, Littlejohns T, Lacey B, Besevic J, Conroy M, Collins R, et al. United Kingdom Biobank (UK Biobank): JACC Focus Seminar 6/8. J Am Coll Cardiol 2021;78:56-65.PubMed
• 20. Eastwood SV, Mathur R, Atkinson M, Brophy S, Sudlow C, Flaig R, et al. Algorithms for the capture and adjudication of prevalent and incident diabetes in UK Biobank. PLoS One 2016;11:e0162388.ArticlePubMedPMC
• 21. Gonzales TI, Westgate K, Strain T, Hollidge S, Jeon J, Christensen DL, et al. Cardiorespiratory fitness assessment using risk-stratified exercise testing and dose-response relationships with disease outcomes. Sci Rep 2021;11:15315.ArticlePubMedPMCPDF
• 22. Blair SN, Kohl HW, Paffenbarger RS, Clark DG, Cooper KH, Gibbons LW. Physical fitness and all-cause mortality: a prospective study of healthy men and women. JAMA 1989;262:2395-401.ArticlePubMed
• 23. Tremblay J, Haloui M, Attaoua R, Tahir R, Hishmih C, Harvey F, et al. Polygenic risk scores predict diabetes complications and their response to intensive blood pressure and glucose control. Diabetologia 2021;64:2012-25.ArticlePubMedPMCPDF
• 24. Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, et al. The UK Biobank resource with deep phenotyping and genomic data. Nature 2018;562:203-9.ArticlePubMedPMCPDF
• 25. Mozaffarian D. Dietary and policy priorities for cardiovascular disease, diabetes, and obesity: a comprehensive review. Circulation 2016;133:187-225.ArticlePubMedPMC
• 26. MacKinnon DP, Fairchild AJ, Fritz MS. Mediation analysis. Annu Rev Psychol 2007;58:593-614.ArticlePubMedPMC
• 27. DeFina LF, Barlow CE, Radford NB, Leonard D, Willis BL. The association between midlife cardiorespiratory fitness and later life chronic kidney disease: the Cooper Center Longitudinal Study. Prev Med 2016;89:178-83.ArticlePubMed
• 28. Kunutsor SK, Isiozor NM, Myers J, Seidu S, Khunti K, Laukkanen JA. Baseline and usual cardiorespiratory fitness and the risk of chronic kidney disease: a prospective study and meta-analysis of published observational cohort studies. Geroscience 2023;45:1761-74.ArticlePubMedPMCPDF
• 29. Tu QM, Jin HM, Yang XH. Lipid abnormality in diabetic kidney disease and potential treatment advancements. Front Endocrinol (Lausanne) 2025;16:1503711.ArticlePubMedPMC
• 30. Jung CY, Yoo TH. Pathophysiologic mechanisms and potential biomarkers in diabetic kidney disease. Diabetes Metab J 2022;46:181-97.ArticlePubMedPMCPDF
• 31. Savelieff MG, Elafros MA, Viswanathan V, Jensen TS, Bennett DL, Feldman EL. The global and regional burden of diabetic peripheral neuropathy. Nat Rev Neurol 2025;21:17-31.ArticlePubMedPMCPDF
• 32. Ting DS, Cheung GC, Wong TY. Diabetic retinopathy: global prevalence, major risk factors, screening practices and public health challenges: a review. Clin Exp Ophthalmol 2016;44:260-77.ArticlePubMedPMC
• 33. Yu MG, Gordin D, Fu J, Park K, Li Q, King GL. Protective factors and the pathogenesis of complications in diabetes. Endocr Rev 2024;45:227-52.ArticlePubMedPMCPDF
• 34. Lyssenko V, Vaag A. Genetics of diabetes-associated microvascular complications. Diabetologia 2023;66:1601-13.ArticlePubMedPMCPDF
• 35. Bouchard C. Genomic predictors of trainability. Exp Physiol 2012;97:347-52.ArticlePubMedPDF
• 36. Fletcher GF, Ades PA, Kligfield P, Arena R, Balady GJ, Bittner VA, et al. Exercise standards for testing and training: a scientific statement from the American Heart Association. Circulation 2013;128:873-934.ArticlePubMed
• 37. Al-Mhanna SB, Batrakoulis A, Wan Ghazali WS, Mohamed M, Aldayel A, Alhussain MH, et al. Effects of combined aerobic and resistance training on glycemic control, blood pressure, inflammation, cardiorespiratory fitness and quality of life in patients with type 2 diabetes and overweight/obesity: a systematic review and meta-analysis. PeerJ 2024;12:e17525.ArticlePubMedPMCPDF
• 38. Franklin BA, Eijsvogels TM, Pandey A, Quindry J, Toth PP. Physical activity, cardiorespiratory fitness, and cardiovascular health: a clinical practice statement of the American Society for Preventive Cardiology Part II: Physical activity, cardiorespiratory fitness, minimum and goal intensities for exercise training, prescriptive methods, and special patient populations. Am J Prev Cardiol 2022;12:100425.ArticlePubMedPMC
• 39. Faselis C, Katsimardou A, Imprialos K, Deligkaris P, Kallistratos M, Dimitriadis K. Microvascular complications of type 2 diabetes mellitus. Curr Vasc Pharmacol 2020;18:117-24.ArticlePubMed
• 40. Barrett EJ, Liu Z, Khamaisi M, King GL, Klein R, Klein BE, et al. Diabetic microvascular disease: an Endocrine Society scientific statement. J Clin Endocrinol Metab 2017;102:4343-410.ArticlePubMedPMC
• 41. Lazarevic G, Antic S, Vlahovic P, Djordjevic V, Zvezdanovic L, Stefanovic V. Effects of aerobic exercise on microalbuminuria and enzymuria in type 2 diabetic patients. Ren Fail 2007;29:199-205.ArticlePubMed
• 42. Nylen ES, Gandhi SM, Kheirbek R, Kokkinos P. Enhanced fitness and renal function in type 2 diabetes. Diabet Med 2015;32:1342-5.ArticlePubMed

Figure & Data

References

Citations

Citations to this article as recorded by  

About this article

Xu A, Zhou H, Wang C, He Q, Wu P, Wu W, Wu H, Kong AP, Cao H, Guan H, Cheng Y. Cardiorespiratory Fitness and Risk of Microvascular Complications in Patients with Type 2 Diabetes Mellitus. Diabetes Metab J. 2026 May 18. doi: 10.4093/dmj.2025.1109. Epub ahead of print.

Received: Nov 05, 2025; Accepted: Feb 10, 2026

DOI: https://doi.org/10.4093/dmj.2025.1109.

Download citation ↓