Longitudinal Changes in the Peripapillary Retinal Nerve Fiber Layer Thickness in Hypertension: 4-Year Prospective Observational Study

Abstract

Purpose: To determine longitudinal changes in the peripapillary retinal nerve fiber layer (pRNFL) thickness in patients with hypertension (HTN).

Methods: Participants without any ophthalmic disease were divided into two groups: an HTN group (50 eyes) that included patients with HTN for ≥5 years and a control group. After the initial visit, pRNFL thicknesses were measured four more times at 1-year intervals using spectral-domain optical coherence tomography. The pRNFL thickness was fitted using linear mixed models. Univariate and multivariate generalized linear mixed models were used to determine factors associated with pRNFL reductions over time.

Results: The mean ages of the HTN and control groups were 62.9 ± 6.2 and 60.7 ± 6.1 years, respectively, and they were not significantly different (P = 0.089). The baseline mean pRNFL thicknesses were 90.50 ± 8.31 and 93.90 ± 8.77 μm; they were significantly different (P = 0.049). The mean pRNFL reduction rate was −0.99 and −0.40 μm/y in the HTN and control groups, respectively; the interaction between group and duration was significant (P < 0.001). In the linear mixed model determination of factors associated with pRNFL reduction, there was no significant factor in the control group. In the HTN group, age, and axial length showed significant results in both univariate and multivariate analyses (estimate: −0.362 and −2.618; P = 0.042 and 0.026, respectively).

Conclusions: Patients with HTN had a significantly greater decrease in pRNFL than normal individuals. Additionally, age and axial length significantly affected the reduction in pRNFL in HTN patients.

Hypertension (HTN), which causes cardiovascular, cerebrovascular, and renal disease, is a major global health problem.^1,2 Mills et al.³ reported that globally, 31.1% of the adult population (1.39 billion people) had HTN in 2010, which was increased from 26.4% in 2000. Sustained HTN causes physical changes in blood vessels such as thickening of the elastic lamina and hyaline changes.⁴ These changes can cause various target organ damage such as myocardial infarction, stroke, and nephropathy. 

HTN is also a risk factor for various ophthalmic diseases such as retinal vascular occlusion, retinal macroaneurysm, ischemic optic neuropathy, and glaucoma. Uncontrolled HTN can cause hypertensive retinopathy, characterized by retinal hemorrhage, hard exudates, cotton wool spots, optic disc edema, and macular edema.⁵ Although hypertensive retinopathy alone may not cause visual impairment except in a severe form, chronic HTN can affect the retinal conformation for long periods. Lee et al.⁶ reported that chronic HTN patients without hypertensive retinopathy and patients with relieved hypertensive retinopathy showed significantly thinner macula, peripapillary retinal nerve fiber layer (pRNFL), and ganglion cell-inner plexiform layer when compared with normal controls. Additionally, the Baltimore Eye Survey showed that systemic HTN had a significantly positive association with glaucoma in older patients because narrowed vessel lumens may have decreased ocular perfusion pressures with subsequent glaucomatous damage to ocular structures.⁷ These changes may result in more pRNFL damage in HTN patients than healthy individuals over time. Because this pRNFL damage could be associated with various ophthalmic diseases such as glaucoma, it is important to carefully monitor changes of pRNFL in HTN patients. However, studies for longitudinal change of pRNFL in HTN patients has not been reported as far as we know. 

In the present study, we analyzed longitudinal changes in pRNFL thicknesses in HTN patients without retinal or optic nerve diseases, including hypertensive retinopathy, to evaluate the long-term effects of HTN on the pRNFL structure. 

Methods

Patients

This prospective, longitudinal, observational study adhered to the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board of Chungnam National University Hospital, Daejeon, Republic of Korea. Patients who visited our retinal clinic for a regular checkup were enrolled and examined between May 2011 and September 2018, and informed consent was obtained from all patients. 

The HTN group consisted of patients diagnosed with HTN ≥ 5 years. All patients were initially diagnosed with HTN at the Department of Internal Medicine of Chungnam National University Hospital. The diagnosis of HTN was made according to the Korean HTN treatment guideline.⁸ The blood pressure (BP) of all patients with HTN was well controlled. The control group enrolled patients with the unilateral epiretinal membrane, macular hole, or intraocular lens dislocation as well as subjects without any ophthalmic diseases. We checked BP using an automated BP device in our retina clinic, and we did not enroll subjects with systolic BP ≥ 140 or diastolic BP ≥ 90. The control group consisted of eyes with a spherical equivalent between +3.0 and −6.0 diopters (D). All patients had a complete ophthalmic examination including measurement of the best-corrected visual acuity (BCVA), intraocular pressure (IOP), spherical equivalent, axial length (using an IOL Master, version 5.02; Carl Zeiss, Jena, Germany), detailed fundus examination, and spectral-domain optical coherence tomography (SD-OCT). After the initial visit, the examinations were performed four more times at 1-year intervals. Patients with a medical history of diabetes; any ophthalmic disease that could affect pRNFL thickness, including glaucoma; retinal diseases; neuro-ophthalmic diseases; intraocular surgery (except cataract extraction); BCVA < 20/25; IOP > 21 mm Hg; axial length ≥ 26 mm; spherical equivalent ≤ −6.0; or ≥ +3.0 D were excluded. Patients with any changes associated with hypertensive retinopathy such as arteriovenous nicking, retinal hemorrhage, cotton wool spots, or optic disc edema were also excluded during the follow-up period. 

OCT Measurements

An experienced examiner obtained measurements from a 200 × 200 optic cube (measured in pixels) scanning protocol using OCT (Cirrus HD OCT 5000, version 10.0; Carl Zeiss Meditec, Dublin, CA, USA). The pRNFL thicknesses were measured using an optic disc cube scan. The optic disc nerve head was brought to the center of the scanned image, and a 200 × 200-pixel resolution axial scan was made over an area of 6 × 6 mm that included the optic nerve head and its surroundings. Images with a signal strength <7, any motion artifact, involuntary saccade, obvious decentration misalignment, or algorithm segmentation failure identified by the autosegmentation error on B-scan images were excluded. 

Statistical Analyses

Ocular parameters in the HTN and control groups were compared using an independent t-test. Linear mixed models were performed to show the reduction rate of the pRNFL in each group and the difference in the reduction of the pRNFL over time between the HTN and control groups. The pRNFL thickness was fitted with linear mixed models with age, sex, BCVA, IOP, axial length, spherical equivalent, HTN duration, baseline pRNFL thickness, and interaction between group and duration as fixed effects. A random intercept was included at eye levels. Univariate and multivariate generalized linear mixed models were used to determine the factors associated with pRNFL reduction over time in the HTN group. To analyze the impact of each class of antihypertensive medication for pRNFL change, patients were categorized into four mutually exclusive subgroups, reflecting the most commonly used regimens of antihypertensive treatment^9,10: (A) angiotensin-converting enzyme inhibitors (ACE-i) and/or angiotensin-receptor blockers (ARB); (B) beta-blockers (BB) and/or calcium-channel blockers (CCB); (C) diuretics alone or combined with other medications; and (D) other combinations. Statistical analyses were performed using software (SPSS version 18.0; IBM Corp., Armonk, NY, USA). 

Results

Demographics

Of the 105 eyes, three eyes were excluded in the HTN group during the follow-up period; two eyes for the occurrence of retinal hemorrhage and cotton wool spots and one eye for follow-up loss. Two eyes were excluded in the control group due to follow-up loss. In total, 50 eyes of 50 participants in each of the HTN and control groups were included. Their BP was well controlled and did not show a significant change to alter the medication. The mean age, spherical equivalent, IOP, and axial length of the HTN and control groups were 62.9 ± 6.2 and 60.7 ± 6.1 years, +0.16 ± 0.16 and −0.29 ± 1.30 D, 14.8 ± 2.3 and 15.4 ± 2.4 mm Hg, and 23.6 ± 0.9 and 23.8 ± 1.0 mm, respectively; they were not significantly different (Table 1). The mean HTN duration in the HTN group was 15.0 ± 6.8 years. In antihypertensive medication, ACE-i or ARB (50%), CCB (50%), BB (18%), and diuretics (14%) were used in the HTN group. In subgroups classified according to the HTN medication, the number of subgroup A, B, C, and D was 12 (24%), 18 (36%), 7 (14%), and 13 (26%), respectively. The blood level of total cholesterol was 192.93 ± 39.38 mg/dL in the HTN group, and 12 patients (24%) were taking statin. The mean pRNFL thicknesses of the HTN and control groups at the initial visit were 90.50 ± 8.31 and 93.90 ± 8.77 μm, respectively; they were significantly different (P = 0.049). None of the eyes showed any signs of retinal diseases, and the rim area, disc area, and cup/disc ratio did not be changed significantly during the follow-up period (P = 0.107, P = 0.256, and P = 0.362, respectively). 

Table 1

 

Demographic and Clinical Characteristics

 

pRNFL Thicknesses of the HTN and Control Groups at Each Visit

In the HTN group, the mean pRNFL thickness showed a constant reduction at every visit, and the reduction was significant over time using a linear mixed model (P < 0.001) (Table 2). Additionally, the pRNFL thicknesses of all four sectors showed a significant reduction over time. In the control group, the mean pRNFL thicknesses also showed a significant reduction over time (P = 0.003). However, only the superior sector showed a significant reduction over time among the four sectors (P = 0.003) (Figure). 

Table 2

 

pRNFL Thicknesses at Each Visit

 

Figure

 

Line graphs of averages with 95% confidence intervals (CIs) showing mean and sectoral pRNFL thicknesses at each visit. The mean and all sectors except the temporal area showed a significant interaction between group and duration in linear mixed models. (Mean, P < 0.001; superior, P = 0.009; nasal, P < 0.001; inferior, P = 0.022; temporal, P = 0.064) Circles: hypertension; triangles: control.

Reduction Rate in the pRNFL Thicknesses and Factors Associated With pRNFL Reductions

The reduction rates in the mean pRNFL thicknesses of the HTN and control groups were −0.99 and −0.40 μm/y, respectively; they showed a significant interaction between group and duration in linear mixed models (P < 0.001), indicating that the degree of reduction over time differed between the HTN and control groups (Table 3). All four sectors in the HTN group showed greater reduction rate in pRNFL thickness than those in the control group, and most of them showed significant interactions between group and duration. The superior and inferior sectors showed greater reduction rates than the nasal and temporal sectors in both groups. 

Table 3

 

Rate of Change in pRNFL Thickness

 

Using a linear mixed model determination of factors associated with pRNFL reductions, there were no significant factors in the control group, whereas in the HTN group, age and axial length showed significant results using both univariate and multivariate analyses (estimate: −0.43 and −3.59; P = 0.016 and P = 0.005, respectively) (Table 4). The subgroup classified according to the antihypertensive medication, total cholesterol, and statin medication did not show a significant relationship with pRNFL reduction (P = 0.149, P = 0.987, and P = 0.262, respectively). 

Table 4

 

Univariate and Multivariate Linear Mixed-Effect Model Determination of Factors Associated With Changes in the pRNFL in the Hypertension Group

 

Discussion

The structural changes in the retinal vessels of HTN patients could be a significant clue to a patient's future risk of heart disease or stroke.^11,12 Thus, observation of retinal vascular structure and its change should not be overlooked in HTN patients. Additionally, those changes might be associated with damage of the inner retina, causing high pRNFL reduction over time in HTN. In the present study, the HTN group showed a significantly thinner RNFL than the control group, consistent with previous studies. Lee et al.⁶ reported that chronic HTN patients, even with well-controlled BP, had a significantly thinner pRNFL than normal controls. A recent large meta-analysis reported a reduced pRNFL thickness in HTN patients.¹³ A thin pRNFL in HTN patients may be associated with the fact that HTN is one of the risk factors for glaucoma. Cantor et al.¹⁴ reported that in HTN patients glaucomatous damage may occur because of the ischemia in the optic nerve or retinal ganglion cells due to the reduction in perfusion pressure. It is therefore possible that microvascular pathology, such as atherosclerosis, increased resistance, rigidity, or insufficient autoregulation, and ischemia due to HTN could be the main cause of the reduction in the pRNFL, similar to the suggested mechanism underlying the reduction in pRNFL in diabetic patients without retinopathy.¹⁵ 

In addition to the HTN group having a thinner pRNFL, they also showed a greater reduction in pRNFL thickness than those in the control group every year. The rate of change in normal individuals was between −0.16 and −0.44 μm/y in most previous studies, although these values were based on cross-sectional estimates.^16–19 Leung et al.²⁰ performed a longitudinal prospective study of normal individuals and reported that the mean rate of change of the average pRNFL thickness was −0.52 μm/y, which is consistent with our study, whereas in the HTN group, the reduction rate of pRNFL was −0.99 μm/y, which is two times greater than in the control group. This result seemed too high considering the difference in baseline pRNFL thickness between two groups. This could be explained in that the vascular alteration would have occurred in the HTN group over time for at least 15 years, so the value of our study was induced from the patients who had accumulated those changes for at least 15 years. The result could be changed if younger or older subjects with different duration of HTN were included. Therefore, it may not be appropriate to apply directly this value to patients currently diagnosed with HTN. Meanwhile, the superior and inferior sectors showed greater reduction rates than the temporal and nasal sectors. Previous studies reported that the pattern of RNFL thinning in open-angle glaucoma preferentially involved superior and inferior RNFL losses.^21–23 It seemed that the pattern of pRNFL thinning in the HTN group was similar to that in open-angle glaucoma patients, which may help explain why HTN is a risk factor for open-angle glaucoma. 

Age was a significant factor associated with the reduction in pRNFL thickness in the HTN group. The risk for occurrence of various diseases associated with uncontrolled HTN such as cardiovascular, cerebrovascular, or renal diseases may increase with increasing age. However, Huang et al.²⁴ reported that in HTN patients aged 60 years or older without specific comorbidities, a more aggressive BP-lowering therapy significantly reduced the risk of stroke and cardiovascular mortalities but increased the risk of renal failure. They reported that age caused minor effects on most outcomes relating to BP-lowering treatment, except for renal failure, so close monitoring of renal function may be required. In addition to the kidney, the retina is a microcirculatory organ that can be affected by lowering BP more than necessary. Bowe et al.²⁵ reported that a nocturnal BP fall was a risk factor for progressive visual field loss in glaucoma patients. They reported that one of the most important factors was vascular dysfunction with disturbed autoregulation, leading to reduced perfusion toward the capillary network of the optic nerve. Thus, with increasing age in HTN patients, nocturnal BP fall caused by antihypertensive medication could affect the inner retina by low perfusion, which may also be one of the important mechanisms to induce a high pRNFL reduction in HTN. 

Axial length was also a significant factor associated with pRNFL reduction in the HTN group. We previously reported that highly myopic eyes showed a greater decrease in pRNFL than normal eyes.²⁶ Additionally, another study reported a negative relationship between axial length and pRNFL thickness in a population-based cohort study.²⁷ It seemed that the elongation of the globe leads to mechanical stretching and thinning of the retina. In the present study, axial length did not show a significant result associated with pRNFL reduction in the control group. The reduced pRNFL thickness in HTN patients may have been affected by the stretching effect of elongated eyes more than in normal controls. However, additional studies including subjects with a wider range of axial length are needed to determine how and how much the axial length affects the pRNFL reduction in HTN in more detail. 

Previous studies have reported the relationship between the classes of antihypertensive medication and glaucoma. In the Rotterdam Eye Study, subjects using CCB showed a 1.8-fold higher risk of developing incident open-angle glaucoma (95% CI, 1.04–3.2; P = 0.037).²⁸ They explained that although CCB dilates ophthalmic and posterior ciliary vessels, BP is decreased without decreasing IOP, thus reducing ocular perfusion pressure. In the European Glaucoma Prevention Study, the use of systemic diuretics in combination with any other antihypertensive medications was associated with the development of open-angle glaucoma.²⁹ However, Harris et al.⁹ reported that there may be no specific medication effect on the optic disc structure and that the association found is likely to be mediated through the hypotensive effect of antihypertensive treatment. Our study also showed no significant relationship between the classes of antihypertensive medication and pRNFL reduction in patients with well-controlled HTN, although the number of subjects was relatively small for analyzing the relationship. Further prospective studies are needed with a large number of well-controlled HTN patients. 

We predicted that the HTN duration would affect the reduction in pRNFL thickness; however, it did not show a significant result (P = 0.837). It is possible that before the treatment or diagnosis of HTN, patients might have already had HTN for an extended period, so the exact duration of HTN could be different. Additionally, HTN duration may not be a critical factor for patients whose nocturnal BP was not low enough to damage the optic nerve, even after long periods of HTN. However, further study is needed to identify the exact mechanism of pRNFL thinning in HTN patients. 

Study Limitations and Strengths

This study has several limitations. Although we hypothesized that both HTN itself and treatment for HTN would affect the pRNFL reduction in HTN, we could not identify how much each factor actually affected the pRNFL loss in this study. Further studies with 24-hour BP monitoring would be helpful to identify the mechanism of pRNFL reduction more in detail and the suitable BP variations that may minimize organ damage because of low perfusion. Second, we could not totally eliminate the possibility that there had been occurrence and regression of hypertensive retinopathy in HTN patients between follow-up periods. Third, we also could not exclude the possibility that patients with preperimetric glaucoma were enrolled at baseline, although we enrolled patients without diabetes, pRNFL defects, or histories of IOP >21 mm Hg during the follow-up periods. The visual field test would be useful in a future study. Fourth, our result may not be appropriate for application to much younger or older patients than our subjects because pRNFL reduction may not be linear with increasing age in subjects with a wider range of age.²⁵ Further prospective studies including subjects with a wider range of age are needed. 

In addition to the prospective design, a strength of our study is that we observed the change in pRNFL thickness in HTN patients for 4 years in 1-year intervals; this is a relatively long period and, to the best of our knowledge, has not been previously reported. Additionally, we enrolled patients without any systemic and ophthalmic diseases except for HTN to more precisely evaluate the effect of HTN on the pRNFL. With respect to understanding the effect of HTN on the pRNFL, the comparison of changes in the pRNFL thicknesses in normal individuals would also be a strength of the present study. 

In conclusion, patients with well-controlled HTN had a significantly greater decrease in pRNFL than normal individuals, and the pattern of pRNFL thinning in HTN was similar to that in open-angle glaucoma patients. Additionally, age and axial length significantly affected the reduction in pRNFL in HTN patients. These findings would be helpful for physicians to observe the pRNFL change in patients with HTN. 

Acknowledgments

Disclosure: M.-W. Lee, None; W.-H. Lee, None; G.-S. Park, None; H.-B. Lim, None; J.-Y. Kim, None