Effect of Ocular Perfusion Pressure on Incidence of Diabetic Retinopathy in Type 2 Diabetes: A Two-Year Prospective Study

Abstract

Purpose: To investigate the association between mean ocular perfusion pressure (MOPP), estimated cerebrospinal fluid pressure (CSFP), and changes in diabetic retinopathy (DR) in a Southern Chinese population with type 2 diabetes (T2DM).

Methods: A total of 1224 subjects from the Guangzhou Diabetic Eye Study were enrolled. Systolic blood pressure (SBP), diastolic blood pressure (DBP), and intraocular pressure (IOP) were measured. MOPP was calculated with the formula: MOPP = 2/3 [DBP + 1/3 (SBP − DBP)] − IOP. CSFP was calculated using the formula: CSFP = 0.44 × body mass index (kg/m²) + 0.16 × DBP − 0.18 × age (years) − 1.91. Incidence, progression, and regression of DR were graded based on seven-field 45° conventional fundus photographs at baseline and during two-year follow-up examinations according to the United Kingdom National Diabetic Eye Screening Program guidelines.

Results: Higher MOPP was associated with DR incidence in the multivariate model (per 1 mm Hg increase: relative risk, 1.05; 95% confidence interval, 1.01–1.09; P = 0.02) and was not associated with DR development and DR regression in two-year follow-up of T2DM patients. However, CSFP was not associated with DR changes (incidence, progression, or regression).

Conclusions: The higher MOPP is an independent risk factor for DR incidence among T2DM patients in a Southern Chinese cohort. Monitoring MOPP and managing blood pressure can be part of a comprehensive approach to prevent or delay the onset of DR in T2DM patients.

Translational Relevance: MOPP might be an indicator for the detection of DR incidence.

Introduction

Diabetic retinopathy (DR) is one of the most common microvascular complications of diabetes and has become a global public health issue.^1–3 Identifying risk factors for the incidence and progression of DR is crucial for diabetic patients. Blood pressure is a well-known risk factor for DR incidence and progression.^4–7 Furthermore, the diabetic population exhibits dysfunctional retinal perfusion.^8,9 Mean ocular perfusion pressure (MOPP) and cerebrospinal fluid pressure (CSFP) influence retinal blood flow, subsequently increasing retinal capillary hydrostatic pressure and capillary leakage.^10,11 Prior research has reported associations between MOPP, CSFP, and DR,^12–15 but conclusions have been inconsistent. Moreover, cohort studies exploring these associations in Asian patients are rare. The Guangzhou Diabetic Eye Study (GDES) represents the first large-scale community-based study among type 2 diabetes mellitus (T2DM) patients in southern China. We conducted a prospective cohort study in GDES, aimed to investigate the impacts of MOPP and CSFP on DR changes using two years of longitudinal data. 

Methods

Study Participants

The GDES is a prospective ongoing cohort study that has enrolled T2DM patients from communities in Guangzhou (ISRCTN registry no. 15853192). Before enrollment, participants were diagnosed with T2DM at comprehensive hospitals and subsequently followed up in community health centers. They were then referred to the Zhongshan Ophthalmic Center, where they underwent baseline ophthalmic and physical examinations, with annual follow-ups thereafter. The study adheres to the Declaration of Helsinki guidelines. Written informed consent was obtained from all participants, and the study was approved by the ethics committee of the Zhongshan Ophthalmic Center at Sun Yat-Sen University, Guangzhou, China (2017KYPJ094). The exclusion criteria included (1) severe systemic diseases other than T2DM and hypertension, such as stroke, heart disease, cancer, or kidney disease; (2) a history of major systemic surgeries, such as renal transplantation, cardiac bypass, or thrombolysis; (3) any other ocular disease, including glaucoma, retinal or choroidal disease, and ocular trauma; (4) a history of ocular interventions, including refractive surgery, intravitreal injections, retinal laser procedures, or intraocular surgery; (5) inability to cooperate with survey questionnaires or examinations; (6) inability to undergo examination with dilated pupils because of corneal ulcers, shallow anterior chambers, or severe refractive media opacity. 

Between November 2017 and December 2019, a total of 1768 T2DM patients with two-year follow-up were enrolled. After excluding 544 participants, 1224 participants were ultimately included in the analysis. 

Assessment and Grading of DR

After pupil dilation, DR was assessed using seven-field 45° conventional color photographs captured with a digital retinal camera (Canon CX-1, Tokyo, Japan), adhering to the standardized Early Treatment Diabetic Retinopathy Study protocol. Diagnosis and grading of DR were performed according to the United Kingdom National Diabetic Eye Screening Program (UK NDESP) guidelines by two experienced ophthalmologists who were blinded to the research protocol. The severity of DR was classified as R0, R1, R2, or R3.¹⁶ 

Vision-threatening DR is defined by the presence of proliferative DR, clinically significant diabetic macular edema, or both.¹⁷ DR incidence was defined as newly developed of any DR during the follow-up period in eyes with no DR at baseline. DR progression was characterized by an increase of one or more steps in DR severity during follow-up in eyes with existing DR at baseline. Conversely, DR regression was defined as a DR level decrease of at least one step during the follow-up in eyes with DR at baseline.¹⁸ For each participant, only the data from the worse eye were used. 

MOPP and CSFP

Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured with the patient in a seated position using an electronic sphygmomanometer (Hem-907; Omron, Kyoto, Japan) and recorded twice at intervals of at least three minutes. The average SBP and DBP values were calculated from the two readings that were closest. Intraocular pressure (IOP) was determined using a non-contact tonometer (CT-1; Topcon, Tokyo, Japan). After exclusion of the highest and lowest values, the average IOP was used for further analysis. Body mass index (BMI) was calculated as weight (kg) divided by the square of height (m). MOPP was calculated with the formula: MOPP = 2/3 [DBP + 1/3 (SBP − DBP)] − IOP. CSFP was calculated using the formula: CSFP = 0.44 × BMI (kg/m²) + 0.16 × DBP − 0.18 × age (years) − 1.91.¹⁹ 

Assessment of Systemic Risk Factors

Demographic information and medical histories, including age, gender, duration of diabetes, insulin usage, and history of systemic diseases, were collected via a standardized questionnaire. Hemoglobin A1c (HbA1c), serum uric acid (SUA), serum creatinine (Scr), total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) were measured following the standardized procedures of accredited laboratories in China. 

Statistical Analysis

In this study, all statistical analyses were conducted using SPSS software (version 27.0, IBM, Armonk, NY, USA). Continuous data were expressed as mean ± standard deviation (SD), whereas categorical variables are presented as numbers (%). The t tests were used to evaluate the statistical differences in demographic and ocular parameters. The χ² tests were used for categorical variable analysis. Two logistic regression models were used to investigate the associations between MOPP, CSFP, DBP, SBP, mean arterial pressure (MAP), and PP with DR changes (incidence, progression, or regression) at the two-year follow-up. Initially, the associations were adjusted for age and sex in the first model. Subsequently, adjustments were made for selected risk factors (P < 0.05 in the first model and universally accepted risk factors) in the multivariable-adjusted logistic regression models. To control the Type I error rate and account for multiple comparisons, we applied a Bonferroni correction to the P values. Relative risks (RRs) with 95% confidence intervals (CIs) were reported. All tests were two-sided, and a P value < 0.05 was deemed statistically significant. 

Results

Baseline Characteristics of Subjects

A total of 1224 participants with T2DM were included in the analysis (281 with DR and 943 with No DR). Table 1 presents the baseline characteristics of these patients. The DR patients had a higher proportion of males (46.6% vs. 41.7%, P = 0.02), a higher proportion of insulin use (43.4% vs. 17.7%, P < 0.001) and a longer diabetes duration (12.62 ± 8.10 vs. 8.16 ± 6.38 years, P < 0.001) compared to No DR patients. The DR patients had higher SBP (138.63 ± 20.42 vs. 133.40 ± 18.29 mm Hg, P = 0.04), higher HbA1c (7.84% ± 1.79% vs. 6.73% ± 1.24%, P < 0.001), higher TG levels (2.29 ± 1.62 vs. 1.86 ± 0.88 mg/d, P < 0.001), higher SUA levels (368.34 ± 106.04 vs. 358.97 ± 91.69 mg/dL, P = 0.02), and higher Scr levels (81.00 ± 33.80 vs. 71.16 ± 17.49 mg/dL, P < 0.001) compared with No DR patients. There were no significant difference in MOPP (45.65 ± 8.42 vs. 44.82 ± 7.61 mm Hg, P = 0.06) and CSFP (8.55 ± 3.24 vs. 8.44 ± 2.95 mm Hg, P = 0.06) between the DR patients and No DR patients. 

Table 1.

 

Characteristics of Patients With or Without DR at Baseline

 

Association of MOPP, CSFP, DBP, SBP, MAP, and PP With DR Incidence

Table 2 shows the association of MOPP, CSFP, DBP, SBP, MAP, and PP with DR incidence. Among 943 without DR at baseline, 41 (4.35%) developed any DR. After adjustment for age and sex, DR incidence was associated with higher DBP (per 1 mm Hg increase: RR = 1.03; 95% CI, 1.00–1.05; P = 0.04), MAP (per 1 mm Hg increase: RR = 1.03; 95% CI, 1.00–1.05; P = 0.02), and MOPP (per 1 mm Hg increase: RR = 1.05; 95% CI, 1.01–1.09; P = 0.01). However, in the multivariable-adjusted model, only DBP (per 1 mm Hg increase: RR = 1.03; 95% CI, 1.00–1.05; P = 0.04) and MOPP (per 1 mm Hg increase: RR = 1.05; 95% CI, 1.01–1.09; P = 0.02) remained significant. (Fig.) 

Table 2.

 

Association Between MOPP, DBP, SBP, MAP, PP, and CSFP With DR Incidence in Two-Years Follow-Up

 

Figure.

 

Forest plots for multivariate logistic association between MOPP, DBP, SBP, MAP, PP, and CSFP with DR incidence in two-year follow-up of T2DM. In the multivariable-adjusted model, only DBP (per 1 mm Hg increase: RR = 1.03; 95% CI, 1.00–1.05; P = 0.04) and MOPP (per 1 mm Hg increase: RR = 1.05; 95% CI, 1.01–1.09; P = 0.02) remained significant.

Association of MOPP, CSFP, DBP, SBP, MAP, and PP With DR Progression

Table 3 shows the association of MOPP, CSFP, DBP, SBP, MAP, and PP with DR progression. Among the 281 with DR at baseline, 20 (7.11%) had DR progression of ≥1 step. After adjustment for age and sex, DR progression was only associated with PP (per 1 mm Hg increase: RR = 1.03; 95% CI, 1.00–1.06; P = 0.04). However, in the multivariable-adjusted model, the difference was not significant. 

Table 3.

 

Association Between MOPP, DBP, SBP, MAP, PP, and CSFP With DR Progression in Two-Years Follow-Up

 

Association of MOPP, CSFP, DBP, SBP, MAP, and PP With DR Regression

Table 4 shows the association of MOPP, CSFP, DBP, SBP, MAP, and PP with DR regression. Among the 281 with DR at baseline, 30 (10.68%) had DR regression of ≥1 step. After adjustment for age and sex, DR regression was only associated with SBP (per 1 mm Hg increase: RR = 1.02; 95% CI, 1.00–1.04; P = 0.07). However, this association was not statistically significant in the multivariable-adjusted model. The logistic regression models revealed no association between DR changes (incidence, progression, or regression) and CSFP at the two-year follow-up. 

Table 4.

 

Association Between MOPP, DBP, SBP, MAP, PP, and CSFP With DR Regression in Two-Years Follow-Up

 

Discussion

In this prospective observational cohort study, our results suggest that, after adjusting for these universally accepted risk factors, we identified the higher MOPP was associated with DR incidence, not associated with DR development and DR regression in two-year follow-up of T2DM patients. However, we found no associations between CSFP and DR changes. Additionally, our results suggest that the higher MOPP may be an independent risk factor in DR incidence in our real-world study, in which T2DM patients predominantly had no or mild DR. The bloodstream in any organization is generated through perfusion pressure; retinal perfusion pressure has attracted the attention of investigators. In the early 1990s, hemodynamic changes in diabetes patients sparked scholarly interest, and the result showed that patients with DR had hyperperfusion of retinal circulation compared with patients without DR.¹² High retinal perfusion can explain a few aspects of DR pathophysiology.⁸ First, the higher retinal perfusion pressure can increase the circumferential stress and injure the retinal vessel wall, resulting in dilation with subsequent hyperperfusion. Second, elevated perfusion pressure can also reduce retinal perfusion by causing retinal vasculature autoregulation.¹² Third, elevated perfusion pressure and the resultant stress changes can also increase the net pressure gradient from vessels to tissue, resulting in retinal capillary leakage (Starling's forces) and increased risk of rupture (Laplace's law).²⁰ All of these result in retinal edema, hemorrhage, and capillary dropout, which are clinically manifested as DR. 

A higher MOPP increases retinal blood flow, thereby elevating retinal capillary hydrostatic pressure and enhancing capillary leakage.¹⁰ Furthermore, an increased estimated CSFP suggests elevated retinal vein pressure. Consequently, MOPP and CSFP are presumed to affect capillary pressure, subsequently leading to capillary leakage,¹¹ which is presumably associated with the incidence and progression of DR. The impact of MOPP and CSFP on the incidence and progression of DR holds clinical significance, because these ocular parameters can be easily measured and modified in clinic. Additionally, MOPP can be reduced by systemic antihypertension medications, and CSFP can be decreased through the use of systemic carbonic anhydrase inhibitors. 

However, the association between the two ocular parameters and DR remains inconsistent.^{3,11–15,21–23} In Asia, baseline data from the Fushun Diabetic Retinopathy Cohort Study (FS-DIRECT), a community-based prospective cohort study, indicated that increased MOPP is associated with the presence of any type of DR in northeastern Chinese patients with T2DM.²² Longitudinal data from FS-DIRECT, within a multivariate model, demonstrated that increasing MOPP was linked to incidence of DR, but not to its progression or regression.¹⁸ Additionally, CSFP showed no association with the incidence, progression, or regression of DR. Meanwhile, the Beijing Eye Study (BES) revealed that higher CSFP was associated with the cumulative 10-year incidence of DR in northern Chinese patients.²¹ A population-based, cross-sectional evaluation from the Sankara Nethralaya Diabetic Retinopathy Epidemiology and Molecular Genetic Study (SN-DREAMS), using multivariate analysis, found no association between MOPP and DR in a South Indian subpopulation with diabetes.¹⁵ Unexpectedly, the Singapore Indian Eye Study revealed that higher CSFP was positively associated with DR progression yet negatively associated with its incidence among migrant Indians residing in Singapore.³ Conflicting evidence exists in the literature concerning the relationship between the two ocular parameters and DR in Asia due to several inconsistencies among studies. First, the types of studies varied and included cross-sectional, cohort, and population-based studies, with cohort studies being notably scarce. Second, there were discrepancies in the baseline data of the subjects, including participants' age, diabetes duration, prior retinal status, and previous glycemic control status. For instance, in FS-DIRECT, 506 eyes with DR were reported among 1322 T2DM subjects (38.28%) at baseline.¹⁸ In BES, the study was divided into urban and countryside part, excluding all individuals diagnosed with DR at baseline (n = 87).²¹ In the Singapore Indian Eye Study, 204 subjects with DR were noted among 705 T2DM subjects (28.94%) at baseline.³ Third, follow-up durations varied: in FS-DIRECT, the mean interval was 21.1 months, BES had a 10-year follow-up from 2001 to 2011, and the follow-up period in FS-DIRECT spanned six years. 

The GDES is the first large-scale cohort study of T2DM patients in southern China. In this study, there were 281 subjects with DR out of 1224 T2DM subjects (22.96%) at baseline, a proportion lower than the 38.28% found in FS-DIRECT. GDES, an ongoing study, enrolled T2DM patients from communities in Guangzhou between November 2017 and December 2019. Conversely, FS-DIRECT recruited T2DM subjects from Fushun city in northeastern China between July 2012 and May 2013. With increased public awareness and advancements in diabetes mellitus (DM) diagnostic technologies, diagnosis and treatment of DM now ensue earlier than in the past. The sample in our study represents the current situation of the T2DM population in Guangzhou district. Our study is a prospective observational cohort study where most participants were T2DM patients with no or mild DR, which represents a real-world situation. Our results, after adjusting for universally accepted risk factors, suggest that a higher MOPP was associated with an increased incidence of DR over a two-year follow-up of T2DM patients, indicating that higher MOPP might be an independent risk factor for DR incidence. However, no associations were found between CSFP and DR changes. 

The MOPP is calculated as two thirds of the difference between the MAP and the IOP, with blood pressure being a significant risk factor for DR. In our study, a higher DBP was linked to an increased incidence of DR. When incorporated into the regression model, both the interactions between the formula components and MOPP, and the actual impact of MOPP, which serves as a surrogate for blood pressure, should be considered. 

The strength of this study is a real-world study based on T2DM patients, which focuses on T2DM patients who are frequently excluded from RCTs, thereby enhancing the generalizability of the findings. Additionally, this research characterized a large cohort through the use of standard seven-field fundus photography to confirm DR diagnoses. Other key strengths include a large sample size and adjustments for confounding factors. 

Despite its strengths, this study has some limitations. First and most importantly, the two ocular parameters, the MOPP and CSFP, were not directly measured but estimated from derived formulas. Second, the study included only T2DM patients from China. Therefore caution should be exercised when generalizing the outcomes to other ethnicities and individuals with T1DM. 

In conclusion, our findings indicate that higher MOPP is an independent risk factor for DR incidence among T2DM patients in China. Conversely, CSFP did not correlate with the incidence, progression, or regression of DR. Monitoring MOPP and managing blood pressure can be part of a comprehensive approach that includes regular eye examinations, glycemic control, and lifestyle modifications to prevent or delay the onset of DR inT2DM patients Nonetheless, further research is required to determine whether interventions to reduce MOPP can decelerate the progression of DR. 

Acknowledgments

The authors express their sincere gratitude to everyone who contributed to the completion of this study. We are grateful to the GDES and staff members, who generously provided their time, knowledge, and resources for this study. 

Supported by research grants from the Science and Technology Program of Guangzhou of China (grants 2024A03J0333) and National Natural Science Foundation of China (grants 81970807 and 82271099). 

Disclosure: Y. Yan, None; X. Wu, None; Y. Li, None; Y. Fan, None; L. Li, None; C.-K. Tsui, None; K. Liu, None; W. Huang, None; X. Liang, None; A. Hu, None