Introduction

Diabetic retinopathy (DR), a common microvascular complication of type 2 diabetes (T2DM), stands as a leading cause of new-onset visual impairment among the working-age population in developed countries. As the prevalence of T2DM and the lifespan of patients with T2DM increase, the prevalence of DR is also rising [1,2,3,4]. DR is caused by the activation of the polyol and hexosamine pathologic pathways, de novo synthesis of diacylglycerol-protein kinase C, and the production of free radicals and advanced glycation end-products, all of which are induced by hyperglycemia [5, 6]. Vision loss due to DR can occur through several mechanisms, such as diabetic macular edema, vitreous hemorrhages, tractional retinal detachment, neovascular glaucoma, and more [7]. Like T2DM, hypertension (HTN) is another significant systemic disease that can result in visual impairment. High blood pressure can lead to visual loss via subfoveal hemorrhages, subretinal fluids, macular exudates, macular edema, and optic nerve swelling [8, 9].

The advent of optical coherence tomography angiography (OCTA), which allows for the visualization of the fine retinal microvasculature across various layers, has fostered numerous studies that have examined retinal microvasculature damage in patients with DR and HTN. For instance, Zihan et al. [10]. found that the progression of DR was associated with an expansion of the foveal avascular zone (FAZ) area and a decrease in the vessel density (VD) of the deep capillary plexus (DCP). Durbin et al. [11]. reported a negative correlation between the VD of the superficial retinal layer and best-corrected visual acuity (BCVA), severity of DR, and age. In the context of HTN, Sargues et al. [12]. reported reduced parafoveal VD in DCP and choriocapillaris (CC) in patients with HTN, and Chua et al. [13]. found that patients with severe HTN, who exhibited changes in cardiac muscle structures, had lower VD of superficial capillary plexus (SCP) than normal subjects.

Given that both DR and HTN can damage retinal microvasculature, and considering the doubling of HTN prevalence in the presence of T2DM, it is expected that concomitant DR and HTN would lead to greater microvascular damage [14,15,16,17]. However, to date, no studies have explored this possibility in detail. Therefore, we investigated the effect of HTN on the inner retinal layer and macular microvasculature in patients with DR by comparison of inner retinal thickness using optical coherence tomography (OCT) and VDs using OCTA.

Methods

Patients

This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of Konyang University Hospital, Daejeon, Republic of Korea (No.2023-05-042). Due to the retrospective design of the study, the requirement for informed consent was waived. We conducted a retrospective review of the charts of 227 patients diagnosed with T2DM or DR who had undergone checkups for retinal abnormalities between January 2018 and January 2023.

Subjects were categorized into three groups: subjects with T2DM without clinical DR (T2DM group), subjects with clinical DR (DR-HTN group), and subjects with both DR and HTN (DR + HTN group). In subjects with DR, the severity of DR was graded according to the International Clinical Early Treatment Diabetic Retinopathy Study classification. Patients with mild, moderate, and severe nonproliferative DR (NPDR) were included in the study. Patients with very severe NPDR and proliferative DR, who required laser treatment or had already received laser photocoagulation, which could potentially cause retinal atrophy, were excluded. In addition, patients who had received intravitreal anti-VEGF injections to reduce DME were also excluded. All HTN patients had well-controlled blood pressure. The exclusion criteria also encompassed a history of any systemic disease other than T2DM and HTN, uncontrolled blood pressure > 140/90 mmHg even after treatment, intraocular pressure (IOP) ≥ 21 mmHg, BCVA < 20/400, central macular thickness ≥ 300 μm, any ophthalmic condition that could affect the retinal layer thickness such as glaucoma, retinal and neuro-ophthalmic diseases, and history of intraocular surgery except for cataract extraction. If both eyes met the inclusion criteria, the eye with the higher image quality was selected for analysis.

Ganglion cell complex thickness measurements using OCT

Retinal layer thicknesses were measured by an experienced examiner utilizing spectral domain OCT (SD-OCT, Spectralis; Heidelberg Engineering, Heidelberg, Germany). Built-in segmentation systems of the Spectralis device, specifically the Heidelberg Eye Explorer (version 6.9a) were employed for each layer thickness measurement from the SD-OCT. The retinal layer thickness in the parafoveal area was determined by averaging the values from the superior, nasal, inferior, and temporal segments within the 1–3 mm range. The thicknesses of segmented layers, including the retinal nerve fiber layer (NFL), ganglion cell layer (GCL), and inner plexiform layer (IPL), were measured. The ganglion cell complex (GCC) thickness, representing the thickness of the inner retinal layer, was derived by adding the thicknesses of the NFL, GCL, and IPL.

Any scans exhibiting motion artifacts, misalignment, segmentation system errors, incomplete image taking, or low resolution were excluded from the analysis.

VD measurement using OCTA

OCTA was performed using a Spectralis OCT2 device (Spectralis; Heidelberg Engineering, Heidelberg, Germany). Images of the SCP (defined as the layer from the ILM to IPL), the DCP (defined as the layer from IPL outer border OPL), and the CC (20 μm thick layer from Bruch’s membrane) were visualized automatically by the device embedded software. A 6.0 × 4.5 mm OCTA scan, centered on the fovea, was acquired. The VD was calculated using open-source processing software (ImageJ, National Institutes of Health, Bethesda, MD, USA). In particular, 8-bit images were selected and adjusted using default threshold settings. This automatic adjustment distinguishes features of interest from the background in segmented grayscale images. Using this binarised image, the VD was subsequently calculated by dividing the area of white pixels by the total number of pixels. OCTA quality was automatically assessed by the device’s embedded software, with a scale ranging from 0 (no signal) to 40 dB (excellent quality). Images exhibiting motion artifacts, fixation loss, segmentation errors, and OCTA quality less than 30 dB were excluded.

Statistical analyses

After confirming that the data followed a normal distribution using the Kolmogorov-Smirnov test, demographic information and ocular parameters were compared among groups using ANOVA and the Bonferroni post hoc test. The chi-square test was utilized for comparisons of categorical data. To control the effect of covariates, including T2DM duration, analysis of covariance was conducted. In the DR + HTN group, univariate and multivariate linear regression analyses were performed to identify factors associated with the VD of the SCP, DCP, and CC. All statistical analyses were performed using SPSS software (version 18.0; IBM Corp., Armonk, New York, USA).

Results

Demographics

A total of 227 eyes were included; 66 eyes in the T2DM group, 78 eyes in the DR-HTN group, and 67 eyes in the DM + HTN group (Table 1). The ages of each group were 62.7 ± 12.8, 63.5 ± 9.4, and 66.1 ± 9.7 years, respectively (P = 0.120). The T2DM durations for the T2DM group, DR-HTN group, and DR + HTN group were 6.0 ± 6.0, 14.8 ± 7.9, and 15.7 ± 8.9 years, respectively (P < 0.001). The HTN duration in the DR + HTN group was 14.5 ± 9.0 years. Other factors, including HbA1c, sex, laterality, DR severity, BCVA, spherical equivalent, IOP, and axial length were not significantly different among the groups.

Table 1 Demographics and clinical characteristics
Full size table

Inner retinal layer thicknesses in each group

The mean GCC thicknesses for the T2DM, DR-HTN, and DR + HTN groups were 112.2 ± 12.3, 109.2 ± 13.7, and 106.2 ± 11.2 μm, respectively, showing significant differences after adjusting for T2DM duration (P = 0.045) (Table 2). The DR + HTN group had a significantly thinner GCC than the T2DM group in post hoc analysis (P = 0.020). The mean GCL thicknesses were 49.3 ± 6.6, 47.6 ± 7.2, and 44.9 ± 6.2 μm (P = 0.003), and IPL thicknesses were 40.3 ± 4.1, 39.1 ± 4.9, and 39.9 ± 4.1 μm (P = 0.015) in the DM, DR, and DR + HTN groups, respectively.

Table 2 Inner retinal layer thicknesses in each group
Full size table

VD in each group

The mean VDs of the SCP for the T2DM, DR-HTN, and DR + HTN groups were 25.4 ± 5.0, 24.3 ± 8.9, and 21.2 ± 7.1%, respectively, showing significant differences after adjusting for DM duration (P = 0.014) (Table 3). Post hoc analyses revealed significant differences when comparing the T2DM group vs. the DR + HTN group (P = 0.003), and the DR-HTN group vs. the DR + HTN group (P = 0.032). The mean VDs of the DCP in each group were 25.9 ± 4.3, 22.9 ± 8.5, and 20.2 ± 6.6%, respectively (P < 0.001 after adjustment for DM duration). In post hoc analyses, the T2DM group vs. the DR-HTN group (P = 0.031), the T2DM group vs. the DR + HTN group (P < 0.001), and the DR-HTN group vs. the DR + HTN group (P = 0.047) comparisons remained significant results. The mean VDs of the CC were 32.3 ± 4.2, 30.4 ± 8.2, and 27.6 ± 5.8% in each group, respectively (P < 0.001 after adjustment for DM duration). In post hoc analyses, the T2DM group vs. the DR + HTN group (P < 0.001) and the DR-HTN group vs. the DR + HTN group (P = 0.034) comparisons remained significant results (Fig. 1).

Table 3 Macular vessel density in each group
Full size table
Fig. 1
figure 1

Bar graphs depict the thickness of the ganglion cell complex and the vessel density of the superficial capillary plexus, deep capillary plexus, and choriocapillaris in each group. T2DM group: subjects with type 2 diabetes mellitus without clinical diabetic retinopathy (DR), DR-HTN group: subjects with DR without hypertension (HTN), DR + HTN group: subjects with both DR and HTN. Error bars indicate mean ± standard deviation. *Statistically significant differences

Full size image

Linear regression in the DR + HTN group

In the SCP, univariate analysis identified HTN duration (B = -0.27, P = 0.005) and BCVA (B = -14.25, P = 0.004) as significant factors associated with VD. In multivariate analysis, both HTN duration (B = -0.24, P = 0.010) and BCVA (B = -12.71, P = 0.008) continued to show significance (Table 4). In the DCP, univariate analysis revealed HTN duration (B = -0.23, P = 0.010) and BCVA (B = -9.47, P = 0.046) as significant, and multivariate analysis retained significance for HTN duration (B = -0.21, P = 0.016) (Fig. 2). No significant factors were associated with VD in the CC.

Table 4 Univariate and multivariate linear regression in the DR + HTN group
Full size table
Fig. 2
figure 2

Scatterplots paired with linear regression analyses demonstrate the relationship between vessel density and the duration of hypertension (HTN) and type 2 diabetes mellitus (T2DM) in patients with diabetic retinopathy and HTN

Full size image

Discussion

In patients with T2DM, HTN induces upregulation of the renin-angiotensin-aldosterone system, oxidative stress, and vascular inflammation. This leads to arterial remodeling, and consequently, an increased prevalence and severity of macrovascular diseases such as cardiovascular disease and stroke [18]. Moreover, HTN exacerbates diabetic microangiopathy, including endothelial permeability and vascular smooth muscle cell dysfunction. As such, the risk for microvascular diseases like nephropathy and retinopathy significantly escalates when HTN and T2DM coexist [19]. Indeed, our previous study found that the vessel length density in the control, patients with T2DM, and those with both T2DM and HTN was 20.43 ± 1.16, 19.50 ± 1.45, and 18.19 ± 20.06 mm− 1, respectively (P < 0.001), and the GC-IPL thickness was 84.9 ± 8.8, 83.5 ± 5.8, and 78.7 ± 8.7 μm, respectively (P < 0.001), indicating greater impairment of the microvasculature and inner retinal layers in patients with T2DM with HTN than those with T2DM alone. [20] However, few studies have reported damage to the inner retinal layer and microvasculature when HTN and DR coexist. In this study, we discovered that GCC thickness tended thinner and the VDs of the SCP, DCP, and CC were significantly lower in the DR + HTN group than in the DR-HTN group. Additionally, HTN duration was significantly associated with the VDs of SCP and DCP.

Ito et al. [21]. found that inner retinal thickness tended to decrease as DR progressed. Kong et al. [22]. found that inner retinal thickness was lower in subjects with systemic HTN than those with normotensive subjects. As such, while previous studies have reported thinning of the inner retinal layers due to HTN or DR individually, there are few studies on the impact of their coexistence on inner retinal layer damage. In our study, the GCC thickness was significantly different among the groups, with the DR + HTN group presenting the thinnest GCC thickness. A previous study reported more aggressive retinal inflammatory alterations with elevated expression levels of ICAM-1, VEGF, and NF-κB in the retina of hypertensive diabetic rats compared to normotensive diabetic rats in their experimental models [23]. This condition may amplify neural apoptosis, glial cell reactivity, and various pathological biochemical pathways linked to the mechanism of diabetic retinal neurodegeneration (DRN), which would result in a more pronounced reduction of inner retinal layer thickness. Apart from intensified DRN, ischemic damage caused by HTN could also contribute to more severe inner retinal damage. Nevertheless, our post hoc analyses did not yield significant results, suggesting that further studies with larger sample sizes are needed.

The VDs of the SCP, DCP, and CC were significantly different among the groups. In particular, that of the DR + HTN group was significantly lower than that of the DR-HTN and T2DM groups. In DR, the infiltration of circulating immune cells and serum proteins into the retina and vitreous, as well as endothelial cell injuries caused by chronic hyperglycemia, can result in acellular retinal capillaries [24, 25]. Such changes are closely linked to a compromised antioxidant defense system, along with increased production of free radicals and VEGF. [26] These effects could also be influenced by altered expressions of enzymes responsible for removing free radicals, as well as epigenetic changes, including DNA methylation and post-translational modifications in the histones. [27] This, in turn, may impair macular microvasculature. In HTN, atherosclerosis, increased vascular resistance and rigidity, and vasoconstriction as an autoregulatory response may lead to ischemic retina damage, which is reflected in the impaired macular microvasculature [28]. The convergence of these two mechanisms might lead to more severe retinal microvasculature impairment. The microvascular damage due to coexisting HTN in DR patients appears to be more prominent than structural changes in the inner retinal layer, as the GCC thickness was not significantly different between the DR-HTN and DR + HTN groups in post-hoc analysis. This severe microvascular damage might be linked to the increased risk or the initial development and progression of DR due to HTN [29,30,31]. Therefore, physicians should evaluate the retinal microvasculature in patients with DR, considering that microvascular damage may be more severe when HTN is comorbid.

The VDs of SCP and DCP were significantly associated with HTN duration in the DR + HTN group. Lim et al. [32]. reported that VD and perfusion density were significantly reduced and the foveal avascular zone tended to be larger in patients with prolonged HTN. Additionally, in our previous study, we found a significant correlation between HTN duration and peripapillary RNFL thickness in patients with both T2DM and HTN [33]. This suggests that HTN duration could have a significant impact on macular microvasculature, much like it does on the thickness of peripapillary RNFL, in DR patients. On the other hand, the T2DM duration, a known factor related to retinal microvasculature in T2DM patients, was not significantly associated with macular VD in the DR + HTN group in this study [34, 35]. Indeed, Durbin et al. [11]. reported that no correlation was detected between VD or foveal avascular zone metrics and duration of diabetes in patients with DR. Thus, the duration of HTN may be a more crucial factor than T2DM duration for macular microvasculature in patients with both DR and HTN.

Previous studies have documented the relationship between visual acuity and macular microvasculature parameters in patients with T2DM [36, 37]. Meshi et al. [37]. found that VDs of the SCP and DCP were negatively correlated with logMAR visual acuity in patients with diabetes without clinical DR. This association was also observed in patients with DR. Samara et al. [38]. found a positive correlation between VD and visual acuity at both the SCP and DCP levels in patients with DR. In our study, VDs of SCP and DCP were significantly associated with BCVA in univariate analyses, suggesting a continued link between VD and BCVA even when HTN is present. However, only the VD of the SCP remained a significant association in multivariate analyses. Some studies have indeed found a significant correlation between BCVA and the VD of the SCP but not the DCP in patients with T2DM or DR [39,40,41]. This might be related to the fact that the DCP, known to be more affected in the early stages of these diseases, does not cause noticeable visual impairment at this stage [36, 42]. Therefore, SCP may have a more direct connection to visual acuity than DCP in DR patients with HTN.

The HbA1c levels among the three groups were not significantly different, and they were not significantly associated with retinal VD in the DR + HTN group in this study. Although blood glucose control is crucial in preventing or reducing retinal damage caused by DR, other factors appear to have a greater association with retinal VD when blood glucose is relatively well-controlled.

This study had certain limitations. First, its retrospective nature may have introduced some selection bias. Second, we were unable to determine the relationship between impaired retinal microvasculature and various aspects of visual function due to the lack of tests such as microperimetry, color-vision testing, or contrast sensitivity testing. Further research is needed to understand how the observed differences in this study results have affected various visual functions. Third, due to the limitations of a retrospective analysis, we could not consider various systemic factors in our analysis. Future studies that include renal function, body mass index, dyslipidemia, detailed family history, fasting blood glucose, and the use of statins will enable a more accurate analysis. Fourth, due to data scarcity, we were unable to include more diverse groups such as patients with only HTN or those with both T2DM and HTN. If these groups and normal controls had been included, a more comprehensive analysis could have been conducted. Additionally, despite adjusting for it in the analysis, the difference in the duration of T2DM among groups and the wide confidence interval range in the linear regression analysis, which may indicate relatively lower accuracy, can be the statistical limitations of this study. However, a strength of our study was that we identified more severe impairment of macular microvasculature in DR patients when HTN is present, a finding not previously reported, to the best of our knowledge.

In conclusion, HTN leads to more severe impairment of macular microvasculature in addition to structural changes in the inner retina in patients with DR. This may be linked with the accelerated deterioration of DR when HTN is concomitant. Similar to other organs, the retina also experiences exacerbated damage when T2DM is complicated by HTN. Additionally, the duration of HTN was significantly associated with VDs of SCP and DCP in DR patients with HTN, whereas the duration of T2DM was not. This suggests that HTN duration may have a greater impact on macular microvasculature than T2DM duration when DR and HTN coexist. Based on these results, further research is needed to investigate how HTN actually influences the progression of DR and various aspects of visual function.