Association of stage of diabetic retinopathy on intraocular pressure

Abstract

Purpose

To determine whether stage of diabetic retinopathy (DR) is associated with increased intraocular pressure (IOP).

Methods

The medical records of patients seen in the outpatient ophthalmology clinic at Sidney & Lois Eskenazi Community Hospital in Indianapolis, Indiana, from January 2022 to August 2023 were reviewed retrospectively. The highest IOP of those presenting without type 2 diabetes mellitus (T2DM), T2DM without DR, and varying severity of DR was recorded. Main outcome measures included median and average IOPs recorded from patients who presented without diabetes, diabetes without retinopathy, and mild, moderate, severe nonproliferative and proliferative diabetic retinopathy. Comparisons between groups generated P values, with ≤0.05 indicating statistical significance.

Results

A total of 389 patients (759 eyes) were included. Mean patient age was 58. Male patients accounted for 286 eyes (37.7%); Hispanic patients, for 270 eyes (35.6%); patients identifying as black, for 261 eyes (34.4%); white patients, for 195 eyes (25.7%); and Asian patients, for 33 eyes (4.3%). Overall, analysis showed IOP increased between eyes without T2DM compared with mild nonproliferative DR (P = 0.0373), mild DR compared to moderate nonproliferative DR (P = 0.0117), and moderate DR compared with severe nonproliferative DR (P = 0.0001). The IOP observed between severe nonproliferative DR to proliferative DR (P = 0.5351) was not statistically significant.

Conclusions

In our study cohort, there was a positive correlation between elevated IOP and more advanced stage of DR, indicating that patients with diabetes but no findings of DR on retinal examination seem to be at greater risk of elevated IOP than those with mild-to-moderate DR. These results suggest that IOP in patients with T2DM, with or without DR, should be closely monitored over time to prevent future IOP-related complications.

Primary open-angle glaucoma (POAG) is the leading cause of irreversible blindness worldwide, affecting approximately 60 million people, with the incidence of new cases increasing each year.1 The leading risk factor in the development of POAG is increased intraocular pressure (IOP).2 IOP is maintained by the balance between continuous generation and drainage of aqueous humor, which provides nutrients and removes waste from the eye.3,4 The aqueous humor is generated by the ciliary body and drained via two pathways: the conventional and uveoscleral pathways.3,4 The conventional pathway includes the trabecular meshwork and Schlemm’s canal, through which approximately 85% of aqueous humor is drained.3,4 In POAG, the conventional pathway is obstructed through mechanisms not yet fully understood. This obstruction impedes the drainage of aqueous humor, leading to increased IOP and a higher risk of glaucomatous damage to the optic nerve.3,4 Although there is currently no cure for glaucoma, treatments aim to prevent further damage by lowering IOP.5 Consequently, identifying and managing risk factors for elevated IOP is essential for preventing the development of this disease.

One of the proposed risk factors for developing glaucoma is type 2 diabetes mellitus (T2DM),2 which affects more than 500 million people worldwide and is increasing in prevalence.6 T2DM is known to cause many microvascular complications, including the vessels that supply the eye.7 Some studies have suggested that T2DM significantly increases the risk of developing glaucoma, whereas others believe this may be a coincidental finding, with factors such as hyperlipidemia, race, and age confounding the results.2,8–10

Diabetic retinopathy (DR) is a known complication in patients with poorly controlled blood sugar where damage to the retinal vasculature occurs.11,12 Symptoms include increased floaters, blurriness, loss of color perception, and areas of lost vision.11,12 There are four stages of DR that are used to stratify severity in clinic.11,12 Early or mild nonproliferative DR (NPDR) is described by the presence of at least one microaneurysm on the retinal examination. Moderate NPDR consists of multiple microaneurysms, dot-and-blot hemorrhages, cotton wool spots, or venous beading. Severe NPDR is diagnosed by the 4-2-1 rule: if the patient has diffuse intraretinal hemorrhages and microaneurysms in four quadrants, venous beading in at least two quadrants, or at least one intraretinal microvascular abnormality, a diagnosis of severe NPDR can be assigned. Approximately 50%–75% of patients in the severe NPDR stage will transition to the final stage, proliferative diabetic retinopathy (PDR), within 1 year.13 In PDR, disorganized neovascularization is seen on fundus examination and may increase the risk of acute glaucoma due to obstruction of aqueous humor outflow through bleeding or growth of a neovascular membrane over the trabecular meshwork.11–13

Some studies indicate that diabetes is associated with POAG; however, DR severity is currently not well understood in the development of POAG.14,15 Although DR is associated with poorly controlled blood glucose levels, research on the correlation between DR stage and its effect on IOP is scarce. The current study compared IOP levels in different stages of DR to determine whether the severity of DR is a risk factor in developing POAG. This could contribute to our understanding of changes in IOP and glaucoma in the diabetic population. Due to the increasing number of patients with ocular findings associated with T2DM and glaucoma each year, understanding the correlation between these two disease states can be helpful in preventing irreversible vision loss in the growing diabetic population.

Subjects and Methods

Ethical review and study protocol, including exemption of written informed consent, was reviewed and approved by the Indiana University and Eskenazi Health Institutional Review Boards. This study followed the regulations of US Health Insurance Portability and Accountability Act of 1996 and adhered to the tenets of the Declaration of Helsinki. The medical records of patients seen at the Sidney & Lois Eskenazi Hospital by the Indiana University School of Medicine Department of Ophthalmology from January 2022 to August 2023 were reviewed retrospectively. Age, race, sex, central corneal thickness (CCT), presence of T2DM and stage of DR corresponding with the patient’s highest IOP were extracted from the record for all visits. IOP measurements were made using handheld TonoPen (AMETEK Reichert, Depew, NY) or by Goldmann applanation devices at the slit lamp biomicroscope, with DR stage evaluated by practitioner on dilated fundus examination. Patients with known factors associated with increased IOP, such as active vitreous hemorrhage, current treatment with intravitreal injections, steroid use, or family history of glaucoma/ocular hypertension were excluded. Those with a history of trauma to the eye, prior IOP-lowering surgery and procedures, or signs of active inflammation were not included. Given the possibility of prior intervention, patient eyes with any previous diagnosis of glaucoma were excluded in this study as well. Patients on IOP lowering medications were not included. The no-T2DM group was determined by patients who received routine primary care examinations without a prior diagnosis. Once data was collected, it was analyzed with one-way ANOVA with significance determined by P < 0.05 (*), P < 0.01 (**), P < 0.001 (***), and P < 0.0001 (****). Men and women within each stage were compared using an unpaired t test, with P values of ≤0.05 considered statistically significant. In this study, mild NPDR is represented as stage 1, moderate NPDR as stage 2, severe NPDR by stage 3, and PDR by stage 4.

Results

A total of 759 eyes of 389 patients were included. Of these 759 eyes, 62% of which were in female patients. The range of available CCT for this patient population was 540–575 μm. The average patient age was 58 years, and most eyes belonged to Hispanic, black, and white patients (Table 1). Figure 1 demonstrates a significantly elevated IOP in the T2DM without DR to those without T2DM (P < 0.0001). The median and average for the highest recorded IOP in the no-T2DM group was 4 mm Hg lower compared with the T2DM without DR condition (Table 2). The IOP of the no-T2DM group was significantly lower compared to stage 1. IOP significantly increased with increasing stage until comparing stage 3 to stage 4, which trended in a similar fashion (Table 3). The T2DM without DR group was nonsignificant compared with stage 3 DR. Overall, the median and average IOP measurements increased with the increasing stage of DR (Table 2).

Table 1.

Eye demographics: patient eyes examined classified by condition

Figure 1.

Distribution of highest recorded intraocular pressure (IOP) readings by presence of diabetes mellitus (DM) and stage of diabetic retinopathy (DR). The violin plot shows the distribution of IOP data points in each condition. The red dashed line indicates the median; the top and lower black dashed lines, the cutoff for the top 75% and lower 25% of IOP readings.

Table 2.

TOTAL	DR stage 1	DR stage 2	DR stage 3	DR stage 4	DM without DR	No DM
Number of values	105	99	103	96	164	192
Minimum, maximum	9, 23	10, 24	12, 28	13, 36	12, 25	7, 21
Median (IQR)	15 (13-16.5)	16 (15-18)	18 (16-21)	19 (17-20)	18 (16-20)	14(13-15)
Mean ± SD	15.01 ± 2.581	16.27 ± 2.762	18.65 ±2.899	19.3 ±3.645	18.04 ± 2.468	14.04 ±2.257
SEM	0.2519	0.2776	0.2857	0.372	0.1927	0.1629
Male	DR stage 1	DR stage 2	DR stage 3	DR stage 4	DM without DR	No DM
Number of values	53	42	40	41	46	64
Minimum, maximum	9, 19	12, 20	12, 23	14, 26	13, 24	10, 18
Median (IQR)	15 (13-16.5)	17 (15-18)	18 (16-19)	19 (17-19)	18 (16-20)	14 (12-15)
75% percentile	16.5	18	19	20	20	15
SEM	0.3719	0.3201	0.3922	0.3714	0.3617	0.2544
Female	DR stage 1	DR stage 2	DR stage 3	DR stage 4	DM without DR	No DM
Number of values	52	57	63	55	118	128
Minimum, maximum	12, 23	10, 24	13, 28	13, 36	12, 25	7,21
Median (IQR)	15 (13.25-16.75)	16 (14-18)	19(17-21)	19 (17-21)	18 (16-20)	14(13-15.75)
Mean ± SD	15.33 ± 2.431	16.12 ± 3.185	19.35 ±2.947	19.87 ±4.291	17.96 ±2.479	14.16 ±2.357
SEM	0.3371	0.4219	0.3712	0.5785	0.2282	0.2083
Hispanic	DR stage 1	DR stage 2	DR stage 3	DR stage 4	DM without DR	No DM
Number of values	29	44	50	45	30	72
Minimum, maximum	9, 23	10, 21	12, 25	13, 24	15, 25	7, 17
Median (IQR)	14 (13-17.5)	16.5 (15-17)	18 (16-20)	19 (16.5-20)	18.5 (17-20)	14 (12-15)
Mean ± SD	14.62 ±3.509	16 ±2.178	18.32 ±3.087	18.47 ±2.341	18.7 ±2.168	13.65 ±2.15
SEM	0.6517	0.3284	0.4365	0.349	0.3958	0.2534
Black	DR stage 1	DR stage 2	DR stage 3	DR stage 4	DM without DR	No DM
Number of values	45	32	26	34	76	48
Minimum, maximum	12, 20	12, 24	14, 28	16, 36	12, 24	9, 21
Median (IQR)	15 (13.5-17)	15.5 (14-18.75)	18 (16-19.25)	20 (17.75-22.25)	18 (16-18)	14 (13-16.75)
Mean ± SD	15.47 ±2.302	16.28 ± 3.476	18.31 ±2.923	20.62 ±4.499	17.76 ±2.668	14.71 ± 2.76
SEM	0.3432	0.6144	0.5732	0.7716	0.306	0.3983
White	DR stage 1	DR stage 2	DR stage 3	DR stage 4	DM without DR	No DM
Number of values	27	21	26	15	46	60
Minimum, maximum	11, 18	12, 20	15, 24	14, 32	13, 22	8, 18
Median (IQR)	14 (14-16)	18 (15-19.5)	20 (17.75-21)	18 (16-19)	18 (16.75-20)	14 (13.25-15)
Mean ± SD	14.44 ± 1.739	17.19 ± 2.442	19.58 ±2.403	18.87 ±4.291	18.07 ±2.265	14.18 ± 1.918
SEM	0.3348	0.5328	0.4712	1.108	0.3339	0.2476

SD, standard deviation; SEM, standard error of mean.

Table 3.

Comparison by Sex

To understand whether sex affects the correlation between IOP levels and stage of DR, we analyzed the IOP levels with increasing stage of DR in both male and female patients. In the male patients (Figure 2A), the no-T2DM group was not significant compared with DR stage 1 but was significantly lower compared with the other stages. Stage 2 compared with stage 3 and stage 3 to stage 4 trended toward the overall data but did not reach statistical significance (Table 3). The IOP from the T2DM without DR was comparable to stages 3 and 4 in this subanalysis.

Figure 2.

Presence of diabetes mellitus (DM) and diabetic retinopathy (DR) stages on intraocular pressure (IOP) readings by sex. The violin plot shows the distribution of IOP data points in each condition. The red dashed line indicates the median; the top and lower black dashed lines, the cutoff for the top 75% and lower 25% of IOP readings. A, Highest recorded IOP in male patient eyes for each condition. B, Highest recorded IOP in female patient eyes for each condition. C, Comparison between male and female eyes by condition.

The eyes of female patients (shown in Figure 2B) showed significantly lower IOP for the no-T2DM group compared with T2DM without DR. Similar to the male analysis, the no-T2DM group IOP was not significant compared to stage 1, but significantly lower compared with stages 2, 3, and 4 (Table 3). The T2DM without DR group was significantly elevated compared to no-T2DM, stage 1, and stage 2 and significantly lower than stage 4. The IOP from stage 3 was comparable to T2DM without DR.

Comparing male to female in each condition (shown in Figure 2C), no-T2DM (P = 0.2690), stage 1 (P = 0.2136), stage 2 (P = 0.5320), stage 4 (P = 0.0755), and T2DM without DR (P = 0.4813) were similar. Female patients in this study showed a significantly higher IOP in stage 3 (P = 0.0018) than men. Overall, these results indicate that sex is not a significant determining factor for the tendency of IOP levels in increasing stage of DR.

Comparison by Race

To examine whether race affects the tendency of IOP levels in stage of DR, a subanalysis of the Hispanic, black, and white patients in our study population was performed (shown in Figure 3A–D). Due to the low sample size, Asian patients were included in the overall studies but subgroup analysis was not performed.

Figure 3.

Diabetes mellitus (DM) and diabetic retinopathy (DR) stages on intraocular pressure (IOP) readings by race. The red dashed line indicates the median; the top and lower black dashed lines, the cutoff for the top 75% and lower 25% of IOP readings. A, Highest recorded IOP in Hispanic patient eyes for each condition. B, Highest recorded IOP in black patient eyes for each condition. C, Highest recorded IOP in white patient eyes for each condition. D, Comparison between Hispanic, black, white eyes by condition.

The IOP of the eyes belonging to the Hispanic (Figure 3A), black (Figure 3B), and white (Figure 3C) all increased with increasing stage. Of note, all groups shared a similar lack of significance when no-T2DM was compared with stage 1 DR. The IOP in T2DM without DR was comparable to stage 3 DR in all groups.

In comparing races within stages (Figure 3D), the eyes of black patients had a significantly higher IOP than eyes of Hispanic patients (P = 0.0347) in the no-T2DM group, but there was no difference between Hispanic and white (P = 0.3730) and black and white patients (P = 0.4547). There was no significance between different races in stage 1, Hispanic versus black (P = 0.3593), Hispanic versus white (P = 0.9649), and black versus white (P = 0.2412). In stage 2, there were no significant differences in IOP between Hispanic and black patients (P = 0.8970), Hispanic and white patients (P = 0.2310), or black and white patients (P = 0.4630). Similarly, in Stage 3, IOP did not differ significantly between Hispanic and black patients (P = 0.9998), Hispanic and white patients (P = 0.1745), or black and white patients (P = 0.2567). In stage 4, black patients had significantly higher IOP compared to Hispanic patients (P = 0.0259), while no significant differences were found between Hispanic and white patients (P = 0.9256) or between black and white patients (P = 0.2606). In the T2DM without DR group, no significant differences in IOP were observed between Hispanic and black patients (P = 0.0347), Hispanic and white patients (P = 0.5155), or black and white patients (P = 0.7883). These findings suggest that race is not a primary factor influencing IOP differences across stages of DR.

Discussion

Elevated IOP is a known risk factor for the development of glaucoma, and the incidence of both T2DM and POAG continues to rise.16,17 Previous studies have suggested that the presence of T2DM significantly increases the risk of developing glaucoma.9,18–21 Studies have also shown that T2DM can raise IOP, although the underlying mechanism remains unclear.14,15 Sato and Roy22 demonstrated that under hyperglycemic conditions, trabecular meshwork cells lining Schlemm’s canal increase the expression of fibronectin and fibrosis mRNA, which impairs aqueous humor drainage and increases the risk of developing POAG.22 Other potential mechanisms include the generation of advanced glycated end products that accelerate trabecular meshwork senescence and apoptosis through nitric oxide and superoxide substances.23 Additionally, the observed activation of protein kinase C disturbs the metalloproteinases of the trabecular meshwork, leading to disruption of function.24,25

Little research has been done on the relationship between stages of DR and IOP. Although a prolonged and more elevated HbA1c is usually correlated with a more severe stage of DR, this raises the question as to whether higher-stage patients are at greater risk for developing POAG.26 Current management of DR does not include a close examination of trending IOP and assessment for glaucoma.

Our results are concordant with previous population studies that show that IOP is elevated in T2DM patients compared with those who do not have T2DM.14,15 Our analysis showed a significant difference between the no-T2DM group and stage 1 DR, although this finding was not consistent across all subgroups. We attribute this to stage 1 being the mildest form, with the possibility that some patients were normoglycemic recently. The significance may be attributed to patients with a history of prolonged and nonprogressing stage 1 T2DM. The gradual rise in the highest IOP readings with advancing stages suggests increasing IOP instability as the stage progresses. The lack of significance between stage 3 and stage 4 may be due to the small sample size, or it could indicate that once a patient progresses beyond moderate NPDR, IOP variations stabilize. However, the increased range seen in the eyes belonging to black patients and white patients in stage 4 indicates this stabilization may not apply to all patients (Table 2). In patients with T2DM but no DR, the median and average IOP values are lower than those observed in stages 3 and 4, but they are still within a comparable range. This could be due to variations in HbA1c levels and the duration of T2DM diagnoses. Additionally, patients without DR may be less likely to actively manage their T2DM if they are not experiencing symptoms.

Of note, compared to patients in the T2DM without DR group, patients in the stages 1 and 2 DR groups had significantly lower IOP, indicating a potential lower risk in the development of POAG in the future (Figure 1). On the other hand, stage 4 has a significantly higher IOP than T2DM without DR and may represent an increased risk in patients with severe retinal vascular changes. Our data suggests stage 3 DR patients have similar IOP to the T2DM without DR group. Current management does not consider mild or moderate NPDR having lower IOP and risk of POAG than patients with T2DM without retinal vasculature changes and warrants further investigation.

Limitations of this study include the lack of HbA1c measurements and the inability to determine the timing of T2DM in relation to DR stages, as many patients did not regularly follow up with a primary care physician or undergo routine lab work. With HbA1c data or a history of blood glucose management, additional correlations between DR stage and IOP could be explored, highlighting the need for future investigation. For this reason, alternate metabolic comorbidities, such as the presence of hyperlipidemia, cannot be evaluated. Some studies have shown that concurrent metabolic diseases, such as those relating to cholesterol, triglycerides, and insulin resistance, may alter aqueous flow dynamics.27 It is also difficult to assess whether the patients in our population who are prescribed a statin medication have well controlled cholesterol. Additionally, a small subset of patients who had received intravitreal injections in their past may experience a sustained IOP elevation. This may contribute to some patients with PDR having elevated pressure.28,29 Also, not all patients evaluated for DR underwent pachymetry at the time of IOP measurements, and as a result, the relationship between CCT and IOP was not assessed in this study. Previous research has demonstrated that diabetes can impact the corneal endothelium and CCT, which in turn may affect IOP measurements.30,31 However, the magnitude of these changes is thought to be relatively minor; it has been estimated that higher blood glucose levels in T2DM patients are associated with an average 12.8 μm increase in CCT compared with normoglycemic patients.30 This small variation is unlikely to contribute to significant clinical measurement errors. Furthermore, the absence of additional analyses, such as retinal nerve fiber layer thickness measurements and optic nerve photography, for some patients at the time of IOP measurement limited their inclusion in the study. Incorporating these tools in future research could provide valuable insights into how the stage of DR influences retinal changes and the risk of glaucoma, further advancing our understanding of this important relationship.

There was no significance difference between the male and female eyes in our study population. This coincides with previous studies that have shown no substantial difference in baseline IOP based on sex.32 Some studies suggest females may have a slightly higher IOP at baseline than men, similar to the data represented here.32 DR does not appear to significantly affect the IOP in one sex more than the other when compared by stage. Previous studies proposed that patients of African descent could have an elevated IOP compared with the non-Hispanic white population, with Hispanic population IOP between these two.33 Although there is a significant difference between the stage 4 and no-T2DM groups, we did not observe large racial differences in IOP when comparing stages and presence of T2DM.

In this study, we investigated the effect of DR stages on IOP. DR is a common complication of T2DM and is classified into stages based on severity. Although changes in IOP across the DR stages have not been thoroughly explored, fluctuations in IOP are critical for the prevention of glaucoma. Notably, a recent prospective study by Sperling et al34 found no significant correlation between DR stage and the diagnosis of ocular hypertension or glaucoma. What sets this study apart is our focus on examining IOP variations across the different stages of DR, rather than concentrating solely on new diagnoses. This research is pivotal to our understanding of why glaucoma appears to be more prevalent in diabetic patients and in advancing our understanding of the underlying mechanisms involved.

Our findings suggest that more advanced stages of DR are associated with a higher risk of uncontrolled IOP. However, a key observation in our study is that T2DM patients without DR have an increased risk of elevated IOP compared with those in DR stages 1 and 2, yet a decreased risk compared to those with PDR. These associations were consistent across sex and race. This is an underexplored area, and future research investigating IOP fluctuations in this population will be crucial for more effective management.

Given these findings, it may be beneficial to closely monitor and track IOP in T2DM patients, including those without DR. Clinics should consider routine HbA1c testing, especially for patients who lack regular healthcare, because not all T2DM cases present with DR. Additionally, routine optical coherence tomography (OCT) could help assess changes in the cup-to-disc ratio and optic nerve health. As the incidence of glaucoma-induced irreversible blindness continues to rise, optimizing the management of T2DM and DR patients is essential for preserving vision.