Introduction

Type 2 diabetes mellitus (T2DM) is a global health issue. In 2015, an estimated 415 million people worldwide had T2DM, and this number is expected to rise to 642 million by 20401. As the prevalence of diabetes increases and the lifespan of patients extends, diabetic retinopathy (DR) and its associated visual impairments are also rapidly increasing worldwide2,3,4. DR is the most common microvascular complication of T2DM and is recognized as a leading cause of blindness and visual impairment in the working-age population globally3,5. Although numerous studies have focused on the pathogenesis and influencing factors of DR, including epigenetic mechanisms such as DNA methylation, histone modifications, as well as miRNAs and long non-coding RNA (lnc-RNA) regulation6,7,8, the exploration of the relationship between serum albumin (ALB) and DR remains inadequate. Serum albumin (ALB) is the most abundant protein in circulation and exhibits significant antioxidant activity, effectively scavenging free radicals9,10. Given its physiological properties, low ALB may be an underappreciated risk factor associated with DR11,12.

While some studies have investigated the relationship between ALB and the pathogenesis and complications of T2DM, the conclusions drawn have been diverse and contradictory13,14. Therefore, this study aims to delve deeper into the association between ALB levels and DR in patients with T2DM through a cross-sectional study.

Methods

Research setting and population

A single-center, cross-sectional analysis was conducted in the Endocrinology Department of the Affiliated Hospital of Qingdao University between January 1, 2021, and December 1, 2022. This period was chosen based on the availability of continuous data from our hospital’s internal database. It provided a complete and unbiased set of patient records, ensuring the consistency and reliability of our research. Furthermore, during this timeframe, the epidemiological trends of diabetic retinopathy and the treatment strategies for type 2 diabetes remained stable, helping to minimize the impact of external confounding factors. We included inpatients aged 18 and above with a T2DM diagnosis. Patient data were extracted from the hospital’s electronic medical records system.

Inclusion and exclusion criteria

Inclusion Criteria:

T2DM diagnosis based on American Diabetes Association15.

Comprehensive eye examination records, particularly retinal status.

Ability and willingness to provide a full set of medical and laboratory data, including ALB levels.

Signed informed consent for study participation.

Exclusion Criteria:

Diagnosis of Type 1 Diabetes or gestational diabetes.

Severe systemic or ocular diseases, such as uveitis or glaucoma.

Treatments or surgeries, e.g., intravitreal injections, within the last three months.

Incomplete or inconsistent clinical and laboratory data.

Participant selection is detailed in Fig. 1.

Figure 1
figure 1

Flowchart of study population.

Full size image

Measurement of serum albumin

ALB levels were determined using the bromocresol green method, per manufacturer’s guidelines. The reference range is 40–55 g/L.

DR status

DR assessment involved a slit lamp microscope and optical coherence tomography. An ophthalmologist made the DR diagnosis based on observed retinopathy changes from funduscopic examinations16.

Ethics statement

This study adhered to the Declaration of Helsinki principles, procured participant informed consent, and received ethical approval from the Ethics Committee of the Affiliated Hospital of Qingdao University (QYFY WZLL 28169).

Covariates

Demographics and Clinical Data: Information included gender, age, diabetes duration, and complications.

Anthropometrics: Data included height, weight, and BMI. BMI = weight (kg)/height2 (m2).

Lifestyle Factors: Smoking and alcohol consumption statuses were determined.

Biochemical Analyses: After 8–10 h of fasting, blood samples were collected for analysis, including lipid profiles, Cr, FPG, liver function, HbA1c, and UAER. Except for HbA1c, tests were conducted using the Hitachi 7600 analyzer. eGFR was calculated using the formula by Andrew S. Levey et al.17. Fatty liver was evaluated based on ultrasound. Hypertension is defined as a systolic blood pressure (SBP) of 140 mmHg or greater and/or a diastolic blood pressure (DBP) of 90 mmHg or higher, based on measurements taken in a clinical setting on three separate occasions, without the influence of antihypertensive medications18. Diabetic nephropathy (DKD) is characterized by patients with a documented history of type 2 diabetes mellitus (T2DM) presenting with either albuminuria (urinary albumin-to-creatinine ratio, UACR, exceeding 30 mg/g) or an estimated glomerular filtration rate (eGFR) below 60 ml/min/1.73 m219. The diagnosis of Diabetic Peripheral Neuropathy (DPN) was established upon the analysis of clinical symptoms (such as sensory loss, pain, and muscular weakness), neurological examinations, and the outcomes of nerve conduction studies20.

Statistical methods

Continuous variables were either mean [Standard Deviation (SD)] or a median [Interquartile Range (IQR)]. DR differences were evaluated using the χ2 test, student T-test, or Man Whitney U test. Multivariate binary logistic regression assessed the ALB-DR correlation. Nonlinearity was examined using the Generalized Additive Model. Stratified binary logistic regression was the method of choice for subgroup analyses. Statistical analyses were conducted using R software, version 4.2.0, and EmpowerStats software. P values < 0.05 were considered significant.

Ethical approval

This research was conducted in accordance with the principles set forth in the Declaration of Helsinki. Informed consent was obtained from all participants. The study received the endorsement of the Ethics Committee of the Affiliated Hospital of Qingdao University (Approval No. QYFY WZLL 28169).

Results

Participant characterization description

In Table 1, participants were categorized based on the presence or absence of Diabetic Retinopathy (DR). Of the 1949 participants, 1144 (58.76%) did not have DR, while 803 (41.24%) did. There were significant differences between the two groups in terms of age (P < 0.001), duration of diabetes (P < 0.001), blood glucose levels (HbA1c, P < 0.001), and lipid levels (TG, TC, both P < 0.05). Furthermore, distinct differences were noted in AST (P < 0.001), ALT (P < 0.001), and the presence of comorbid conditions such as fatty liver (P < 0.001), diabetic nephropathy (P < 0.001), and diabetic peripheral neuropathy (P < 0.001).

Table 1 baseline characteristics of participants:
Full size table

Correlation between Albumin and Diabetic Retinopathy:

Regarding serum albumin (ALB) levels presented in Table 1, the mean ALB for participants with DR was 39.35 ± 3.43 g/l, while for those without DR it was 40.22 ± 3.10 g/l. A significant association was observed between ALB levels and the presence of DR (P < 0.001).

As illustrated in Table 2, we evaluated the relationship between serum albumin (ALB) and diabetic retinopathy (DR) using logistic regression analysis after adjusting for potential confounders. The non-adjusted model revealed that for each 10 g/L increment in ALB, the risk of DR was significantly reduced by 56% (OR = 0.44, 95% CI 0.33, 0.58, P < 0.0001). This association persisted across all models. In Adjusted Model I, every 10 g/L increase in ALB was associated with a 51% decrease in DR risk (OR = 0.49, 95% CI 0.37–0.66, P < 0.0001). In Adjusted Model II, the corresponding decrease was 34% (OR = 0.66, 95% CI 0.48–0.94, P = 0.0123). In Model III, it was 33% (OR = 0.67, 95% CI 0.48–0.94, P = 0.0209).

Table 2 Relationship between Serum Albumin Levels and DR in different models.
Full size table

Image Annotation: The generalized additive model plot delineates the association between ALB levels (X-axis) and predicted DR risk (Y-axis).

To elucidate the association between ALB levels and DR risk, curve fitting analyses, as illustrated in Fig. 2, were conducted, adjusting for confounders such as age, sex, BMI, duration of diabetes, smoking status, FBG, HbA1c, lipid profile (total cholesterol and triglycerides), liver enzymes (AST and ALT), and comorbidities like fatty liver disease, diabetic nephropathy, and diabetic peripheral neuropathy. As depicted in Fig. 2, the analyses revealed a significant inverse relationship between ALB levels and DR risk. Importantly, as shown in Fig. 2, no inflection point was observed, denoting a consistent decrease in DR risk with rising ALB levels across the observed range.

Figure 2
figure 2

Curve Fitting: Evaluating the relationship between serum albumin levels (ALB) and diabetic retinopathy risk via generalized additive models.

Full size image

As shown in Table 3, subsequent stratified analyses were conducted considering variables such as age, gender, BMI, diabetes duration, smoking status, FBG, HbA1c, fatty liver disease, diabetic nephropathy, and diabetic peripheral neuropathy. The interaction P-values for all categories were less than 0.05 in these subgroups. This consistency underscores the significant association between ALB and DR across varied conditions and demographics. Notably, the direction of association varied among some subgroups, yet the robustness of the ALB-DR relationship was unwavering, further reinforcing the pivotal role of ALB in DR pathogenesis.

Table 3 Stratified analyses of ALB levels and diabetic retinopathy risk.
Full size table

Discussion

In this cross-sectional study, we analyzed data from 1947 type 2 diabetic patients and found a significant negative correlation between serum albumin (ALB) levels and diabetic retinopathy (DR). This association remained after adjusting for major confounders such as age and gender. Although our findings are consistent with those of Gao-Xiang Wang’s 2022 study21, it is important to note that our study is based on data from Chinese hospitalized patients, providing further evidence for the consistency of the relationship between ALB and DR across different populations. It is worth mentioning that Wang et al. chose to conduct a quartile analysis based on ALB levels, while we opted for ALB grouping in units of 10, a method that helps reveal the strength of the association across different ranges. Furthermore, to gain a deeper understanding of this association, we performed curve fitting analysis to explore potential nonlinear relationships.

Serum albumin (ALB) is considered to have significant antioxidant activity, particularly its 34th cysteine residue, which is the most abundant source of free thiol in plasma22,23. This is particularly important for type 2 diabetic patients, who commonly experience oxidative stress and chronic inflammation. In these patients, the antioxidant properties of ALB may play a key role in defending against oxidative damage24,25.

In addition, ALB can bind to advanced glycation end products (AGEs), which are produced through non-enzymatic glycation reactions under hyperglycemic conditions 26,27. AGEs are closely associated with the development of diabetic retinopathy28,29, and ALB may mitigate their damaging effects on the retina by binding to them30.

Moreover, serum albumin plays a key role in maintaining vascular permeability and stability, which helps reduce vascular leakage and edema, maintaining the integrity of blood vessel walls and preventing the formation and progression of retinopathy24,31.

Overall, serum albumin’s antioxidant effects, binding to AGEs, and maintenance of vascular integrity provide a biological basis for its protective role in type 2 diabetic retinopathy, further explaining the negative correlation between ALB and DR observed in this study.

However, there are several key limitations in this study. Firstly, due to the cross-sectional design, we were only able to collect baseline serum albumin measurements, making it difficult to establish causation. Future studies could consider using a longitudinal cohort design to more accurately explore the relationship between serum albumin levels and diabetic retinopathy.

Secondly, although we adjusted for multiple variables, there is still the possibility of unaccounted confounding factors. Future studies could consider using matched study designs or propensity score matching to reduce the potential for confounding.

Lastly, our sample was sourced from the endocrinology department of a single hospital, which may have resulted in selection bias. To enhance the generalizability and representativeness of the study results, we recommend future research be conducted in multiple hospitals or different regions. Additionally, it’s important to note that the sample size, determined by data from 1947 patients, was not based on a priori power calculations. Nevertheless, it offered adequate statistical power (81.8%) for detecting an OR of 0.67. We acknowledge that future studies could benefit from predefined sample size estimations to enhance study design and findings. Despite these considerations, we believe the current sample size lends credibility to our results within the context of this study.

Despite these limitations, our study provides valuable insights into the relationship between serum albumin levels and type 2 diabetic retinopathy. We hope that future researchers will further explore this field, taking into account the limitations mentioned in our study.

Conclusions

In conclusion, our research unveils a marked negative correlation between serum albumin levels and the presence of diabetic retinopathy among type 2 diabetic patients. This association holds even after accounting for confounding factors, emphasizing the potential protective role of serum albumin. Our findings not only enrich the current body of evidence but also suggest that maintaining optimal serum albumin levels could be a therapeutic target for mitigating the risk of diabetic retinopathy. As we delve deeper into this association, the insights gained hold promise for practical clinical implications, directing future research pathways and interventions in diabetes care.