February 2025
Volume 14, Issue 2
Open Access
Public Health  |   February 2025
Associations Between Diabetic Retinopathy and Frailty: Insights From the National Health and Nutrition Examination Survey and Mendelian Randomization
Author Affiliations & Notes
  • Jianqi Chen
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China
  • Xuhao Chen
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China
  • Xu Cao
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China
  • Xiaohua Zhuo
    Department of Pathophysiology, School of Medicine, Sun Yat-Sen University, Shenzhen, China
    Shenzhen Eye Institute, Shenzhen Eye Hospital, Jinan University, Shenzhen, China
  • Yuwen Wen
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China
  • Guitong Ye
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China
  • Yuan Zhang
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China
  • Jinan Zhan
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China
  • Hongmei Tan
    Department of Pathophysiology, School of Medicine, Sun Yat-Sen University, Shenzhen, China
  • Yingting Zhu
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China
  • Yehong Zhuo
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China
  • Correspondence: Yingting Zhu, State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, 54 Xianlie Rd., Guangzhou 510060, China. e-mail: [email protected] 
  • Yehong Zhuo, State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, 54 Xianlie Rd., Guangzhou 510060, China. e-mail: [email protected] 
  • Footnotes
     JC and XC contributed equally to this work as co-first authors.
Translational Vision Science & Technology February 2025, Vol.14, 2. doi:https://doi.org/10.1167/tvst.14.2.2
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Jianqi Chen, Xuhao Chen, Xu Cao, Xiaohua Zhuo, Yuwen Wen, Guitong Ye, Yuan Zhang, Jinan Zhan, Hongmei Tan, Yingting Zhu, Yehong Zhuo; Associations Between Diabetic Retinopathy and Frailty: Insights From the National Health and Nutrition Examination Survey and Mendelian Randomization. Trans. Vis. Sci. Tech. 2025;14(2):2. https://doi.org/10.1167/tvst.14.2.2.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: To elucidate the relationship between diabetic retinopathy (DR) and frailty and investigate genetic correlations and causality.

Methods: We analyzed data from the US National Health and Nutrition Examination Survey, which included 1003 individuals with diabetes. DR was evaluated via nonmydriatic retinal imaging, and frailty was measured using a 49-item frailty index. The association between DR and frailty was assessed using survey-weighted logistics regression adjusted for multiple covariates, including age, sex, race, education level, family income-to-poverty ratio, marital status, and Healthy Eating Index. Genetic correlations and causal relationships were investigated through linkage disequilibrium score regression and bidirectional Mendelian randomization.

Results: DR was significantly associated with higher odds of frailty after full adjustment (odd ratio [OR] = 4.25; 95% confidence interval [CI], 1.08−16.67; P = 0.040). The association was robust and did not significantly differ across age (P interaction = 0.080), sex (P interaction = 0.216), or race (P interaction = 0.749) groups. DR exhibited a moderate but significant genetic correlation with frailty (rg = 0.27, standard error = 0.04; P = 2.43 × 10−10). Genetically inferred DR was significantly associated with a greater frailty index (β = 0.03; 95% CI, 0.01−0.05; P < 0.001), whereas frailty was not associated with DR risk (OR = 1.20; 95% CI, 0.80−1.81; P = 0.376).

Conclusions: Our findings suggest that DR is associated with an increased risk of frailty, indicating that DR not only impairs vision but also accelerates physical decline.

Translational Relevance: This study highlights the critical need for integrated care approaches that incorporate frailty screening and proactive management in individuals with DR to prevent further health deterioration and improve both quality of life and long-term outcomes.

Introduction
Diabetic retinopathy (DR) is a leading cause of visual impairment and blindness worldwide, and its incidence is projected to increase due to the rising prevalence of diabetes and population aging.13 This chronic progressive disease typically begins asymptomatically, progressing from nonproliferative to proliferative DR and macular edema without timely intervention. Proliferative DR is characterized by new blood vessel formation in the retina and optic disc, accompanied by preretinal and vitreous hemorrhages.4 DR significantly increases the risk of blindness, imposing considerable challenges for affected individuals, families, communities, and healthcare systems globally. 
As an aging-related syndrome characterized by reduced resilience and increased susceptibility to stressors,5 frailty is also considered a major global public health concern that is linked to adverse outcomes in old age, such as lower life expectancy and higher risk of disability.6 A thorough analysis encompassing 62 nations and areas found that 12% of older adults are affected by frailty.7 Over the last 20 years, biological and epidemiological studies have significantly enhanced our understanding of the physiological causes of frailty. Despite these advances, specific interventions for treating frailty remain limited. From a public health perspective, further research is needed to understand the interactions between disorders and frailty, potentially paving the way for innovative long-term care strategies for middle-aged and older adults. 
Because DR and frailty are both common conditions in the aging population with diabetes, it is feasible that they may potentially be interrelated. Both conditions share common pathological mechanisms, such as chronic inflammation, oxidative stress, and vascular dysfunction, all of which contribute to systemic physiological decline.8 The vision loss caused by DR directly exacerbates physical inactivity, increases fall risk, and accelerates mobility decline, which are key contributors to frailty development.9 Frailty imposes additional risks of falls, disability, and mortality, which may worsen outcomes in individuals with DR. Despite these plausible links, the relationship between DR and frailty remains underexplored, and it remains unclear whether monitoring and preventing frailty are necessary in middle-aged and older adults with DR. Addressing this gap is crucial, as understanding this association could inform targeted interventions to improve patient outcomes and long-term health management and mitigate the dual burden of these conditions. 
In this study, we utilized two analytical methods to explore the relationship between DR and frailty (Fig. 1). The primary objective was to assess the association between DR and frailty by analyzing large-scale, nationally representative data from US populations using the National Health and Nutrition Examination Survey (NHANES). Additionally, we investigated genetic correlations and causal relationships through linkage disequilibrium score regression (LDSC) and Mendelian randomization (MR) based on a genome-wide association study (GWAS) in European populations.10 LDSC quantifies the genetic correlation between DR and frailty by analyzing GWAS statistics, providing insights into the shared genetic architecture underlying these conditions.10 In contrast, MR leverages genetic variants as instrumental variables to infer causality, minimizing confounding and reverse causation.11 These methods provide a robust framework to explore the genetic basis of the DR–frailty association while validating observational findings. 
Figure 1.
 
Visual representation of the study design.
Figure 1.
 
Visual representation of the study design.
Methods
Cross-Sectional Investigation
Data and Participants
The data were obtained from the NHANES, an ongoing cross-sectional survey that samples noninstitutionalized US individuals.12 Ethical approval for the NHANES was granted by the Ethics Review Board of the National Center for Health Statistics (Protocol #2005-06 and its continuation). All participants provided prior informed consent. We used NHANES data from the 2005−2006 and 2007−2008 cycles, which included participants with diabetes who were ≥40 years old. 
Diagnosis of Diabetes and DR
Based on the algorithm of Haddad et al.,13 participants were considered to have diabetes if they met the following criteria: self-reported diabetes, a glycosylated hemoglobin A1c level ≥6.5%, or use of antidiabetic medications. DR was identified from fundus images obtained via nonmydriatic retinal imaging using a Canon CR6-45NM non-mydriatic retinal camera and Canon EOS 10D camera (Canon USA, Huntington, New York). Anonymous graders reviewed fundus images at the University of Wisconsin Ocular Epidemiologic Reading Center. Disagreements were resolved through consensus by a senior grader or senior ophthalmologist. DR severity was determined following the Early Treatment Diabetic Retinopathy Study protocol.14 Participants were classified into five categories based on the worse-affected eye: (1) no retinopathy, (2) mild nonproliferative DR (NPDR), (3) moderate NPDR, (4) severe NPDR, or (5) proliferative DR. Considering the relatively small sample size, we subsequently grouped participants into two categories for further analysis: with DR and without DR. 
Frailty Index Measurement
The frailty index deficit accumulation method is widely used to measure frailty in middle-aged and older adults. Our frailty index was constructed based on previous studies.15,16 Frailty indices include traits that are indicators of health decline, exhibit increased risk with aging, are prevalent among the population, span across various bodily systems, and have data available for ≥80% of participants. Every health deficiency was measured using a 0 to 1 scale, which allowed us to incorporate both continuous and categorical health indicators. The frailty index included 49 health deficits spanning over a wide range of systems and a variety of medical conditions. We also considered physical capabilities and measurements, critical laboratory markers, overall wellness, and involvement with health care (Supplementary Table S1). The frailty index value was calculated by dividing the score of identified deficits for an individual by the total number of applicable deficits. A frailty index value ≤ 0.10 was considered to indicate non-frail, with reference to a previous study.17 
Covariates
Data on sociodemographic characteristics, including age, ethnicity, sex, education, marital status, and family income-to-poverty ratio, were gathered through interviews. These variables were selected based on their potential to influence metabolic control and health outcomes in populations with diabetes.18,19 We also employed the Healthy Eating Index (HEI) as a covariate because healthy eating was shown to be relevant to both frailty and diabetic complications.20,21 The HEI reflects adherence to dietary guidelines, with total values derived by summing scores from each dietary component, resulting in a range from 0 (indicating poor diet quality) to 100 (indicating excellent diet quality).20 
Statistical Analysis
To investigate the relationship between DR and frailty, we used survey-weighted logistics regression analysis, including three models, progressively adjusting for additional variables. The initial model was unadjusted; the second model accounted for age, sex, and ethnicity; and the third model further adjusted for income, education, marital status, and HEI. Additionally, a sensitivity analysis using an interaction approach was conducted to explore potential differences in the association between DR and frailty across age, sex, and ethnicity groups. When an interaction effect was identified, hierarchical survey-weighted generalized linear models were applied. 
To address the missing data for the family income-to-poverty ratio (7.58% missingness), we employed multiple imputation with chained equations using the mice package in R (R Foundation for Statistical Computing, Vienna, Austria). Predictive mean matching was used as the imputation method, which is robust for imputing continuous variables. The imputation model included demographic variables (age, sex, race, education, and marital status), health-related variables (HEI score), and the research variables (frailty score and DR category in the worse eye). A total of 50 iterations were performed to ensure the stability and convergence of the imputation model. 
We integrated the NHANES framework into our analysis by including the appropriate sampling weight, strata, and primary sampling unit. All statistical tests were conducted as two-tailed, with a significance level set at P < 0.05. 
GWAS-Based Investigation
Data Source
Single-nucleotide polymorphisms (SNPs) were used as instrumental variables. Frailty-associated SNPs were sourced from a comprehensive GWAS meta-analysis, including European participants from the UK Biobank (164,610 individuals) and TwinGen (10,616 individuals).22 The frailty index, encompassing 49 health deficits, is well validated and has been extensively applied clinically.22 DR summary statistics were sourced from the GWAS European population data from the FinnGen consortium database.23 These data encompassed 10,413 cases and 308,633 controls. DR diagnosis within the FinnGen dataset was confirmed according to the International Classification of Disease code. 
The Medical Ethics Committee of the Zhongshan Ophthalmic Center at Sun Yat-Sen University in Guangzhou, China, approved the GWAS analysis for this study and waived the requirement for informed consent (approval number 2023KYPJ350). This study adhered to the tenets of the Declaration of Helsinki. 
Instrumental Variable Selection
We initially screened for SNPs with strong associations with exposure using a genome-wide significance threshold (P < 5 × 10−8). Subsequently, we applied linkage disequilibrium (LD) clumping to these SNPs using PLINK 1.9,24 setting a stringent clumping threshold of R2 = 0.001 within a 10,000-kB window.25 Only the SNP with the lowest P value was retained in cases of LD. To determine the risk of weak instrument bias among the retained SNPs, we excluded SNPs with an F statistic below 10.26 Additionally, SNPs directly associated with the outcome (P < 5 × 10−5) were also eliminated, and potential reverse causation instruments were excluded using the Steiger filtering process.27 When SNPs for the exposure were absent in the outcome datasets, we substituted them with appropriate proxy SNPs, ensuring a minimum LD R2 of 0.8, where applicable. 
Statistical Analysis
Linkage Disequilibrium Score Regression
We assessed SNP-based genetic heritability and the correlation between DR and frailty using the LDSC method.10 LDSC assesses the relationship between test statistics and LD, allowing for differentiation between inflation from true polygenic signals and statistical biases. This approach enables genetic correlation (rg) estimation using GWAS summary statistics without overlapping sample interference.10 We estimated genetic covariance by multiplying the DR-associated variant and frailty index z-scores and regressing the product against the LD score. Normalizing the genetic covariance by SNP heritability revealed the genetic correlation between DR and frailty. 
Bidirectional Two-Sample MR
We performed bidirectional two-sample MR analyses to explore potential causal relationships. We performed MR analysis bidirectionally, first considering frailty as the exposure to determine whether a higher frailty index increases susceptibility to DR and subsequently as the outcome to investigate whether DR contributes to increased frailty index. 
The primary analytical approach used in this study was inverse variance weighting (IVW).11 A fixed-effects IVW model was first used to calculate Cochran's Q statistic, utilized for detecting heterogeneity among the genetic variations. Heterogeneity was deemed present if the P value was below 0.05. When heterogeneity was identified, the effects were re-evaluated with a multiplicative random-effects IVW model. We employed multiple robust methods for comparison, including MR‒Egger, weighted median, weighted mode, simple mode, and MR–PRESSO. The MR‒Egger regression intercept was used to assess the presence of horizontal pleiotropy, with a P value less than 0.05 indicating potential directional pleiotropy.11 Sensitivity analysis was also performed using a leave-one-out approach. The effects were expressed as β coefficients with 95% confidence intervals (CIs) for frailty index outcomes, and these were transformed into odds ratios (ORs) for DR-related outcomes. Power calculations were conducted using a previously established MR framework.28 Statistical significance was determined using two-tailed tests with an alpha level of 0.05. All analyses were performed using R 4.2.2. 
Results
Associations Between DR and Frailty
In the cross-sectional analysis, we included 1003 NHANES participants with diabetes: 490 men (46.66%) and 513 women (53.34%). Individuals with DR tended to have lower incomes (Table 1). DR was not correlated with frailty in the raw model (OR = 3.22; 95% CI, 0.96–10.76; P = 0.057). However, DR was correlated with increasing odds of frailty after controlling for age, sex, and ethnicity (OR = 3.71; 95% CI, 1.06–13.00; P = 0.041) and after further adjustment for all variables (OR = 4.25; 95% CI, 1.08–16.67; P = 0.040). The variation in the association did not reach significance across age (P interaction = 0.080), sex (P interaction = 0.216), or race (P interaction = 0.749) groups, indicating no notable interaction effects (Table 2). 
Table 1.
 
Survey-Weighted Sample Characteristics According to the Presence of Diabetic Retinopathy
Table 1.
 
Survey-Weighted Sample Characteristics According to the Presence of Diabetic Retinopathy
Table 2.
 
Association Between DR and Frailty in the NHANES
Table 2.
 
Association Between DR and Frailty in the NHANES
Genetic Correlation Between DR and Frailty
The LDSC analysis demonstrated a modest but significant genetic association between DR and frailty (rg = 0.27; SE = 0.04; P = 2.43 × 10−10) (Fig. 2A). 
Figure 2.
 
Genetic correlations and causal relationships between DR and frailty. (A) Results of the LDSC analysis. (B) Results of the bidirectional MR analysis. The scatterplots illustrate the potential effects of SNPs, with the slope of each plot indicating the evaluated effect size for each method.
Figure 2.
 
Genetic correlations and causal relationships between DR and frailty. (A) Results of the LDSC analysis. (B) Results of the bidirectional MR analysis. The scatterplots illustrate the potential effects of SNPs, with the slope of each plot indicating the evaluated effect size for each method.
Causal Relationship Between DR and Frailty
Causal Effect of Frailty on DR
The MR–Egger regression intercept showed no significant indication of directional pleiotropy (P = 0.661) in the examination of the causative relationship between frailty and DR. Furthermore, there was no notable heterogeneity across the genetic variants (Cochran's Q = 9.67, P = 0.645) (Supplementary Table S2). Consequently, we performed a fixed-effects IVW analysis, which was corroborated by the sensitivity analyses. The analysis indicated no significant causal connection between the genetically predicted frailty index and DR prevalence (OR = 1.20; 95% CI, 0.80−1.81; P = 0.376) (Fig. 2B). The power of the analysis was 14.4%. None of the 13 SNPs linked with the frailty index had a significant correlation with either an elevated or reduced risk of DR, and the leave-one-out analysis corroborated that no single SNP exerted a disproportionate impact on the total findings (Supplementary Fig. S1). 
Causal Effect of DR on Frailty
The MR–Egger regression intercept revealed no significant indication of directional pleiotropy (P = 0.592) in the MR analysis regarding the causal effect of DR on frailty. The genetic variants exhibited considerable heterogeneity (Cochran's Q = 51.53; P < 0.001) (Supplementary Table S2). Employing multiplicative random-effects IVW, we established that an increased likelihood of a higher frailty index was associated with genetically inferred DR (β = 0.03; 95% CI, 0.01−0.05; P < 0.001). The power of the analysis was 100%. The link was substantiated by the weighted median (β = 0.04; 95% CI, 0.02−0.06; P < 0.001), weighted mode (β = 0.05; 95% CI, 0.02−0.08; P = 0.006), simple mode (β = 0.04; 95% CI, 0−0.08; P = 0.044), and MR–PRESSO analyses (β = 0.04; 95% CI, 0.02−0.06; P < 0.001) (Fig. 2B). When examining the individual and collective influences of DR-related SNPs on frailty, eight of the 23 SNPs were significantly associated with a risk of higher frailty index (rs73061095, rs141935215, rs481887, rs6820509, rs6679677, rs34872471, rs2057726, and rs55869430). In contrast, the remaining SNPs showed no significant associations. Moreover, the leave-one-out analysis revealed that no single SNP drove the overall effect (Supplementary Fig. S2). 
Discussion
Our cross-sectional analysis suggested a potential association between DR and increased odds of frailty after adjusting for covariates. The GWAS-based analyses revealed a moderate but significant genetic association between these two conditions. MR confirmed a causal effect of DR on a higher frailty index but did not provide evidence supporting a causal influence of frailty on DR. 
Research indicates that frailty is prevalent among individuals with diabetes, with studies showing that older adults with diabetes are more likely to experience frailty compared with those without diabetes, with common frailty components including exhaustion, low physical activity, and weak handgrip strength.29 This increased frailty not only affects the quality of life but also correlates with higher rates of hospitalization and mortality.30 Our study further suggests that DR may contribute to worsening physical function and accelerate frailty progression, potentially adversely impacting long-term outcomes in diabetes management. 
As a common microvascular complication of diabetes, DR significantly impairs visual acuity, leading to profound vision loss or blindness in severe cases. This deterioration in vision limits mobility, disrupts spatial awareness, and deters patients from engaging in regular exercise or basic daily activities.9,14 The resulting sedentary lifestyle accelerates decline in physical fitness, muscle strength, and overall health, contributing to frailty development. Our cross-sectional analysis confirmed this connection, demonstrating greater odds of frailty in individuals with DR than in those without DR. Physiologically, declining wellness and capability precipitate numerous adverse outcomes, including reduced adherence to medical advice, greater dependence on familial and social support systems, and impaired diabetes management.31 This promotes DR progression and traps patients in a cycle of deteriorating health and increased vulnerability, negatively impacting disease management, healthcare outcomes, associated costs, and patient well-being.31 Psychologically, DR significantly contributes to distress, with depression, anxiety, and general distress becoming more pronounced in advanced stages or when visual acuity substantially declines.32 Socially, DR leads to social interaction disruption, social isolation, and increased dependency. Furthermore, as the primary cause of blindness among working-age adults, it undermines employment opportunities and productivity, increasing poverty rates.33,34 
Increasing evidence also supports a causal link between DR and frailty, partly attributable to overlapping risk factors and pathophysiological pathways, such as inflammation, oxidative stress, and disruptions in neuronal and vascular functions. These shared mechanisms indicate a complex interplay at the cellular and molecular levels, predisposing individuals to the co-occurrence of DR and frailty.8 According to previous studies, frailty and DR share inflammatory factors, such as interleukin-6 (IL-6).35,36 IL-6 is a multifunctional inflammatory cytokine produced by various cell types that has been implicated in the development of age-related conditions because its levels naturally increase with age.35 Elevated IL-6 levels have also been linked to reduced muscle strength and physical disability, both of which are central to frailty.37 Recent findings confirm a strong association between elevated IL-6 levels and frailty or pre-frailty status in older adults.35 In a recent meta-analysis, IL-6 levels were significantly higher in the DR group than in the control group, suggesting a correlation between IL-6 levels and DR.36 Our LDSC analysis revealed a moderate but significant genetic correlation between DR and frailty, suggesting a potential comorbidity relationship. Further research into the mechanisms underlying the link between DR and frailty will enhance our understanding and facilitate the development of targeted interventions to mitigate their impact. 
Our findings highlight the considerable frailty burden placed on individuals with diabetes and its vision-related complications, emphasizing the need for holistic care strategies that encompass both the medical and psychosocial dimensions of DR. Specialized services, such as family counseling, job support, financial aid, and advice on social media engagement can alleviate the social consequences of DR. These services are meticulously designed to address the specific needs of individuals at their most vulnerable moments. This insight is pivotal for policymakers allocating healthcare budgets and rehabilitation professionals optimizing patient support. 
Specifically, effective DR management is vital for reducing the risk of frailty and mitigating adverse long-term outcomes. Effective management of blood glucose, blood pressure, and potentially blood lipids is essential for reducing the risk of developing and worsening DR. Laser therapy can help preserve vision in patients with proliferative DR and macular edema, although its ability to reverse visual loss is limited. In the advanced stages of DR, vitrectomy surgery may be needed. Novel treatments, such as steroid injections and anti-vascular endothelial growth factor agents, offer potential benefits by reducing retinal damage and may be particularly advantageous for patients who are unresponsive to traditional methods. Innovative approaches for DR treatment should be further developed, with emerging approaches such as inhibiting other angiogenic factors, regenerative therapy, and topical treatments showing promise.38 
This study has several strengths. First, we used a 49-item frailty index, offering a holistic view of the frailty model. Second, we accounted for the cross-sectional NHANES data by adopting a bidirectional MR approach to address confounding and reverse causality biases. Finally, to confirm the consistency and robustness of the findings, we performed various sensitivity analyses, including interaction analysis in NHANES and weighted median, weighted mode, simple mode, MR–PRESSO, and leave-one-out analyses in MR. 
Our study also has some limitations. In the NHANES sample, the use of imputation in the income-to-poverty ratio may have introduced bias. The interaction analyses may be limited by the relatively small subgroup sample sizes, which could have reduced the statistical power and led to nonsignificant results. Furthermore, the diabetes status was mainly determined through self-report, and the cross-sectional design of NHANES limits the ability to infer causality between DR and frailty. However, to mitigate biases from self-reported diabetes in NHANES, glycosylated hemoglobin A1c levels were included as an objective biomarker in the diagnosis criteria, and MR was applied to address confounding and reverse causation. In the GWAS analysis, due to data availability constraints, all data pertained to European populations. Thus, extending the identified associations to other populations should be done with caution. Additionally, the MR analysis with DR as the exposure and frailty as the outcome had a low statistical power, which may be insufficient to detect positive associations. Moreover, MR may not precisely quantify the association between DR and frailty due to the genetic variance explained ratio, but the qualitative findings are still reliable.39 The combined analysis with NHANES data could offer a more comprehensive quantitative assessment of the relationship. 
In conclusion, DR was associated with a greater risk of frailty, indicating that DR not only impairs vision but also accelerates physical decline, highlighting the critical need for integrated care approaches that incorporate frailty screening and proactive management in patients with DR. Early identification of frailty through targeted screening could enable timely interventions, such as resistance training, multicomponent exercise programs, and nutritional support, to preserve physical function and slow frailty progression. These strategies not only can improve patient outcomes but may also be cost effective by reducing hospitalization rates and delaying severe complications. Future research should explore the cost effectiveness of such integrated care models, optimizing health outcomes and quality of life for these individuals. 
Acknowledgments
The authors thank J. L. Atkins and her study group for providing the frailty GWAS datasets, and the participants and investigators of the FinnGen study. We also thank the use of BioRender to create Figure 1. 
Supported by grants from the National Key R&D Project of China (2020YFA0112701), Guangdong Basic and Applied Basic Research Foundation (2024A1515013058), and Science and Technology Program of Guangzhou, China (202206080005). The sponsors or funding organizations had no role in the design or conduct of this research. 
Disclosure: J. Chen, None; X. Chen, None; X. Cao, None; X. Zhuo, None; Y. Wen, None; G. Ye, None; Y. Zhang, None; J. Zhan, None; H. Tan, None; Y. Zhu, None; Y. Zhuo, None 
References
Sabanayagam C, Banu R, Chee ML, et al. Incidence and progression of diabetic retinopathy: a systematic review. Lancet Diabetes Endocrinol. 2019; 7(2): 140–149. [CrossRef] [PubMed]
Chen J, Zhu Y, Li Z, et al. Global impact of population aging on vision loss prevalence: a population-based study. Glob Transit. 2024; 6(suppl 7): 28–36.
Chen J, Zhu Y, Li Z, et al. Trends in prevalence rates of blindness among patients with diabetic retinopathy in high-income countries from 1990 to 2019: a joinpoint regression analysis. Diabetes Res Clin Pract. 2023; 202: 110823. [CrossRef] [PubMed]
Wu L, Fernandez-Loaiza P, Sauma J, Hernandez-Bogantes E, Masis M. Classification of diabetic retinopathy and diabetic macular edema. World J Diabetes. 2013; 4(6): 290–294. [CrossRef] [PubMed]
Cesari M, Prince M, Thiyagarajan JA, et al. Frailty: an emerging public health priority. J Am Med Dir Assoc. 2016; 17(3): 188–192. [CrossRef] [PubMed]
Kojima G, Iliffe S, Walters K. Frailty index as a predictor of mortality: a systematic review and meta-analysis. Age Ageing. 2018; 47(2): 193–200. [CrossRef] [PubMed]
O'Caoimh R, Sezgin D, O'Donovan MR, et al. Prevalence of frailty in 62 countries across the world: a systematic review and meta-analysis of population-level studies. Age Ageing. 2021; 50(1): 96–104. [CrossRef] [PubMed]
Antonetti DA, Silva PS, Stitt AW. Current understanding of the molecular and cellular pathology of diabetic retinopathy. Nat Rev Endocrinol. 2021; 17(4): 195–206. [CrossRef] [PubMed]
Cooper O, Taylor DJ, Crabb DP, Sim DA, McBain H. Psychological, social and everyday visual impact of diabetic macular oedema and diabetic retinopathy: a systematic review. Diabet Med. 2020; 37(6): 924–933. [CrossRef] [PubMed]
Bulik-Sullivan B, Finucane HK, Anttila V, et al. An atlas of genetic correlations across human diseases and traits. Nat Genet. 2015; 47(11): 1236–1241. [CrossRef] [PubMed]
Emdin CA, Khera AV, Kathiresan S. Mendelian randomization. JAMA. 2017; 318(19): 1925–1926. [CrossRef] [PubMed]
Centers for Disease Control and Prevention. National Health and Nutrition Examination Survey. Available at: https://wwwn.cdc.gov/nchs/nhanes/analyticguidelines.aspx. Accessed February 22, 2024.
Haddad MB, Lash TL, Castro KG, et al. Tuberculosis infection among people with diabetes: United States population differences by race/ethnicity. Am J Prev Med. 2020; 58(6): 858–863. [CrossRef] [PubMed]
Willis JR, Doan QV, Gleeson M, et al. Vision-related functional burden of diabetic retinopathy across severity levels in the United States. JAMA Ophthalmol. 2017; 135(9): 926–932. [CrossRef] [PubMed]
Hakeem FF, Bernabé E, Sabbah W. Association between oral health and frailty among American older adults. J Am Med Dir Assoc. 2021; 22(3): 559–563.e2. [CrossRef] [PubMed]
Chen J, Cao X, Zhuo X, et al. Relationships between frailty and the risk of glaucoma in middle-aged and older adults [published online ahead of print October 5, 2024]. Ophthalmol Glaucoma, https://doi.org/10.1016/j.ogla.2024.09.006.
Guan Z, Ma L, Wu C. Association between serum Klotho and physical frailty in middle-aged and older adults: finding from the National Health and Nutrition Examination Survey. J Am Med Dir Assoc. 2023; 24(8): 1173–1178.e2. [CrossRef] [PubMed]
Li B, Cheng X, Huang Y, et al. The differences of metabolic profiles, socioeconomic status and diabetic retinopathy in U.S. working-age and elderly adults with diabetes: results from NHANES 1999-2018 [published online ahead of print August 5, 2024]. Acta Diabetol, https://doi.org/10.1007/s00592-024-02328-8.
Bonilla-Escobar FJ, Eibel MR, Le L, Gallagher DS, Waxman EL. Follow-up in a point-of-care diabetic retinopathy program in Pittsburgh: a non-concurrent retrospective cohort study. BMC Ophthalmol. 2024; 24(1): 356. [CrossRef] [PubMed]
Goshen A, Goldbourt U, Benyamini Y, Shimony T, Keinan-Boker L, Gerber Y. Diet quality and incident frailty in adults 65 years or older: the Israeli Longitudinal Study on Aging. Mayo Clin Proc. 2023; 98(12): 1774–1784. [CrossRef] [PubMed]
Liu G, Li Y, Pan A, et al. Adherence to a healthy lifestyle in association with microvascular complications among adults with type 2 diabetes. JAMA Netw Open. 2023; 6(1): e2252239. [CrossRef] [PubMed]
Atkins JL, Jylhävä J, Pedersen NL, et al. A genome-wide association study of the frailty index highlights brain pathways in ageing. Aging Cell. 2021; 20(9): e13459. [CrossRef] [PubMed]
Kurki MI, Karjalainen J, Palta P, et al. FinnGen provides genetic insights from a well-phenotyped isolated population. Nature. 2023; 613(7944): 508–518. [CrossRef] [PubMed]
Purcell S, Neale B, Todd-Brown K, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007; 81(3): 559–575. [CrossRef] [PubMed]
Auton A, Brooks LD, Durbin RM, et al. A global reference for human genetic variation. Nature. 2015; 526(7571): 68–74. [PubMed]
Burgess S, Thompson SG. Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol. 2011; 40(3): 755–764. [CrossRef] [PubMed]
Hemani G, Tilling K, Davey Smith G. Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLoS Genet. 2017; 13(11): e1007081. [CrossRef] [PubMed]
Pierce BL, Ahsan H, Vanderweele TJ. Power and instrument strength requirements for Mendelian randomization studies using multiple genetic variants. Int J Epidemiol. 2011; 40(3): 740–752. [CrossRef] [PubMed]
Yanase T, Yanagita I, Muta K, Nawata H. Frailty in elderly diabetes patients. Endocr J. 2018; 65(1): 1–11. [CrossRef] [PubMed]
Miao Z, Zhang Q, Yin J, Li L, Feng Y. Impact of frailty on mortality, hospitalization, cardiovascular events, and complications in patients with diabetes mellitus: a systematic review and meta-analysis. Diabetol Metab Syndr. 2024; 16(1): 116. [CrossRef] [PubMed]
Fenwick E, Rees G, Pesudovs K, et al. Social and emotional impact of diabetic retinopathy: a review. Clin Exp Ophthalmol. 2012; 40(1): 27–38. [CrossRef] [PubMed]
Nguyen TT, Wong TY, Islam FM, et al. Evidence of early retinal microvascular changes in patients with type 2 diabetes and depression. Psychosom Med. 2010; 72(6): 535–538. [CrossRef] [PubMed]
Kang EY, Chen TH, Garg SJ, et al. Association of statin therapy with prevention of vision-threatening diabetic retinopathy. JAMA Ophthalmol. 2019; 137(4): 363–371. [CrossRef] [PubMed]
Chen J, Yang C, Zheng W, et al. Global, regional, and national epidemiology of visual impairment in working-age individuals, 1990-2019. JAMA Ophthalmol. 2024; 142(1): 25–32. [CrossRef] [PubMed]
Marcos-Pérez D, Sánchez-Flores M, Proietti S, et al. Association of inflammatory mediators with frailty status in older adults: results from a systematic review and meta-analysis. Geroscience. 2020; 42(6): 1451–1473. [CrossRef] [PubMed]
Yao Y, Li R, Du J, Long L, Li X, Luo N. Interleukin-6 and diabetic retinopathy: a systematic review and meta-analysis. Curr Eye Res. 2019; 44(5): 564–574. [CrossRef] [PubMed]
Ferrucci L, Penninx BW, Volpato S, et al. Change in muscle strength explains accelerated decline of physical function in older women with high interleukin-6 serum levels. J Am Geriatr Soc. 2002; 50(12): 1947–1954. [CrossRef] [PubMed]
Cheung N, Mitchell P, Wong TY. Diabetic retinopathy. Lancet. 2010; 376(9735): 124–136. [CrossRef] [PubMed]
Chen J, Cao X, Chen X, et al. Causal relationship between central corneal thickness and open-angle glaucoma: evidence from Mendelian randomization. Exp Eye Res. 2024; 246: 110000. [CrossRef] [PubMed]
Figure 1.
 
Visual representation of the study design.
Figure 1.
 
Visual representation of the study design.
Figure 2.
 
Genetic correlations and causal relationships between DR and frailty. (A) Results of the LDSC analysis. (B) Results of the bidirectional MR analysis. The scatterplots illustrate the potential effects of SNPs, with the slope of each plot indicating the evaluated effect size for each method.
Figure 2.
 
Genetic correlations and causal relationships between DR and frailty. (A) Results of the LDSC analysis. (B) Results of the bidirectional MR analysis. The scatterplots illustrate the potential effects of SNPs, with the slope of each plot indicating the evaluated effect size for each method.
Table 1.
 
Survey-Weighted Sample Characteristics According to the Presence of Diabetic Retinopathy
Table 1.
 
Survey-Weighted Sample Characteristics According to the Presence of Diabetic Retinopathy
Table 2.
 
Association Between DR and Frailty in the NHANES
Table 2.
 
Association Between DR and Frailty in the NHANES
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×