Association Between Systemic Hypertension and Macular Thickness Measured by Optical Coherence Tomography

Abstract

Purpose.: This study aimed to evaluate an association between hypertension and macular thickness.

Methods.: A total of 827 Korean adults composed of 163 pairs of twins and their family members were included in this population-based cross-sectional study. Macular thickness was measured with optical coherence tomography at nine macular subfields defined by the Early Treatment of Diabetic Retinopathy Study. Cardiometabolic risk factors, including body mass index (BMI), hypertension, diabetes, lipid profiles, and smoking status, were assessed. Linear mixed regression analysis was conducted with consideration of familial correlations and adjustment for covariates.

Results.: Age-, sex-, and axial length–adjusted analysis showed that systemic hypertension was associated with a significant change in macular thickness in most subfields except for the fovea. Compared with normotensive subjects, macular thickness was lower in subjects with systemic hypertension (P ≤ 0.05), with the highest difference (2.52%) in the outer temporal region and the lowest difference (1.44%) in the inner temporal region. This association persisted even after adjusting for other cardiometabolic risk factors. Other cardiometabolic risk factors were not independently associated with macular thickness in any subfields. Stratified analysis showed that the inverse association between macular thickness and hypertension was stronger in the group with elevated fasting glucose compared with the group with normal fasting glucose (P for interaction ≤ 0.05).

Conclusions.: Systemic hypertension was inversely associated with macular thickness in most macular subfields, particularly in subjects with an elevated fasting glucose level. This finding suggests that it may be necessary to consider the presence of hypertension when macular thickness and pericentral macular area volume are evaluated.

Macular thickness is commonly measured as a valuable index for diagnosis and treatment of macular pathology. Optical coherence tomography (OCT) provides noninvasive, reproducible, and reliable measurement of macular thickness^1,2 and, therefore, has been commonly used in clinical practice for the diagnosis and the evaluation of treatment efficacy for macular disorders. 

In addition to genetic factors, many other factors, including age, sex, axial length, fasting glucose, and ethnicity, have been suggested to affect macular thickness.^3–12 It is necessary to consider these characteristics in studies evaluating macular thickness. 

It is still unclear, however, whether high systemic blood pressure (BP) influences macular thickness because very few studies have been conducted with controversial findings. Although a prospective study reported that high BP was not associated with macular thickness,¹³ the study did not provide detailed findings regarding the influence of systemic hypertension on macular thickness. Another retrospective study in diabetic patients reported that macular central subfield thickness and total macular volume did not vary with the use of antihypertensive medications. However, the study did not measure blood pressure and could not evaluate the association between systemic blood pressure and macular thickness.¹⁴ 

This study aimed to evaluate an association between systemic hypertension and macular thickness with consideration of a wide range of other cardiometabolic risk factors, using data from a population-based Korean twin and family study. 

Methods

Study Subjects

This study used data from 827 participants (337 males and 490 females) in the Healthy Twin study. Details of the study design and protocol of the Healthy Twin study were published previously.¹² In brief, the Healthy Twin study has been conducted as a nationwide community-based cohort study since 2005 and has recruited Korean adult twins and their family members through nationwide advertisement and mailings, using the nationwide twin registry to investigate genetic and environmental determinants of a wide range of traits. None of the participants of this study was recruited from ophthalmology clinics. 

Optical coherence tomography images of both eyes were obtained after pupil dilation in 926 persons between May 2007 and January 2009 in the department of ophthalmology at Samsung Medical Center. Among these 926 persons, a total of 99 persons were excluded for the following reasons: 15 had low (<5) OCT signal strength in scans of both eyes, six exhibited highly skewed macular thickness caused by an unrecognized scanning error with motion artifact, 12 had either a history or evidence of pathologic features in the retina or treated retinal disease, four had glaucoma, and 62 had a high axial length (≥26 mm). We applied the axial length criteria to exclude high myopic eyes because thickness and the volume distribution of the macula in high myopic eyes were found to differ from those in eyes with normal axial length.¹⁵ 

Finally, OCT data of 827 persons composed of 129 pairs of monozygotic twins, 34 pairs of dizygotic twins, 125 mothers, 89 fathers, and 287 siblings from 232 families (average 3.6 [SD: 2.3] persons in each family) were included in the analysis. 

This study was approved by the institutional review board of Samsung Medical Center, and informed consent was obtained from all participants, adhering to the tenets of the Declaration of Helsinki. 

Ocular Measurement

Both eyes of participants were scanned using a Stratus OCT instrument (Carl Zeiss Meditec, Dublin, CA, USA) with the fast macular thickness map protocol. A single well-trained technician operated all OCT scans. The macula was scanned with six radial line scans, each 6 mm in length equally spaced, centered on the fovea. A retinal border detection algorithm contained in the OCT software allowed automatic calculation of macular thickness in the cross-sectional scan as the distance between the vitreoretinal interface and the junction of inner and outer photoreceptor segments.¹⁶ Then, a macular thickness map was made from the six radial line scans and interpolation using the OCT retinal mapping program (version 5.0). The program calculated a numeric average of the retinal thickness in nine subfields as defined by the Early Treatment of Diabetic Retinopathy Study: fovea, four inner quadrant subfields (within 1–3 mm of the center), and four outer quadrants subfields (within 3–6 mm of the center).¹⁷ 

Digital color fundus photography (TRC-50IX; Topcon Corp., Tokyo, Japan) was taken to identify any retinal pathology in all participants, and axial length was measured by corneal touch A-scan ultrasonography (model 820; Allergan-Humphrey, San Leandro, CA, USA). 

For analysis, we used data from the eye with the higher signal-to-noise ratio when compared with the contralateral eye. If both eyes had the same signal-to-noise ratio, the data of the eye with the thinner macula thickness were used. 

Measurement of Cardiometabolic Risk Factors

The concentrations of glucose, hemoglobin A1c, high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) were measured by enzymatic or homogeneous assay with commercial kits (ADVIA 1650; Siemens, Munich, Germany) in fresh serum collected after a minimum 12-hour overnight fast. Weight (kg) and height (cm) were measured in light clothing using standardized scales and stadiometers, and BMI was calculated by dividing the weight with the height squared (kg/m²). A trained research nurse measured BP manually using a standard mercury sphygmomanometer while participants were in a sitting position. All physical measurements were conducted twice for each participant, and the average value of the two measurements was used for statistical analyses. Data on past medical history and smoking status were collected using a self-administered questionnaire. The amount of lifetime cigarette smoking (pack-years) was calculated using the data of the duration of smoking and the mean amount of smoking per day. 

We defined hypertension on the basis of high systolic BP (≥140 mm Hg), high diastolic BP (≥90 mm Hg), or current use of a BP-lowering agent. We defined diabetes on the basis of fasting glucose level (≥6.99 mM), hemoglobin A1c level (≥6.5), or use of a glucose-lowering agent. 

Statistical Analysis

The difference of ocular measures and cardiometabolic risk factors was assessed by the t-test and χ² test between hypertensives and normotensives. 

The association between hypertension and macular thickness at each subfield was evaluated using a linear mixed regression analysis, in which the correlation structure from the intrafamilial relationship was accounted for by adjusting for household effect and sibling effects as random effects.¹² Prior to linear mixed regression analysis, macular thickness data at each subfield were log transformed because they did not show normal distribution. Initially, we did analysis with consideration of age, sex, and axial length as fixed effects. In the next step, we made further adjustment for known cardiometabolic risk factors, such as BMI, LDL-C, HDL-C, diabetes, and smoking, as fixed effects. 

We also evaluated the independent association of macular thickness with other cardiometabolic risk factors, such as BMI, LDL-C, HDL-C, diabetes mellitus, and smoking, also by linear mixed regression analysis. 

In addition, to evaluate whether the association between hypertension and macular thickness is influenced by the presence of other cardiometabolic risk factors or not, we conducted stratified analysis. For this analysis, subjects were categorized into two strata for each cardiometabolic risk factor: BMI (≥25, <25 kg/m²), LDL-C (≥3.36, <3.36 mM), HDL-C (≥1.03, <1.03 mM for men, and ≥1.29, <1.29 mM for women), glucose (≥5.55, <5.55 mM); and smoking (ever, never). For glucose, we applied the cutoff level of 5.55 mM used for diagnosing metabolic syndrome instead of the cutoff level for diagnosing diabetes because the number of subjects with diabetes was too small (67 persons) to conduct stratified analysis. 

All the hypotheses were tested bidirectionally with an alpha level set at 0.05. All the analyses were conducted with statistical software (SAS version 9.3; SAS Institute, Inc., Cary, NC, USA). 

Results

Table 1 shows the distribution of clinical characteristics according to the presence of systemic hypertension. A total of 140 of 827 subjects (16.9%) were identified as hypertensive. Compared with normotensives, hypertensives were older (P < 0.01) and had worse cardiometabolic profiles (P < 0.01). Axial length of hypertensives was borderline significantly shorter than normotensives (P = 0.06). 

Table 1

 

Clinical Characteristics* of Study Subjects According to the Presence of Hypertension

 

Table 2 shows the association between hypertension and macular thickness at each subfield. Generally, hypertensives had a significantly thinner macular thickness than normotensives at most macular subfields (P < 0.05), except for the fovea region. After adjusting for age, sex, and axial length, hypertension was associated with significantly reduced macular thickness at most subfields except for the fovea. When other measured cardiometabolic factors were additionally adjusted, the inverse association between hypertension and macular thickness did not materially change at all subfields. The percent decrease in macular thickness associated with hypertension was 1.56% at the inner superior, 1.86% at the inner inferior, 1.48% at the inner nasal, 1.50% at the inner temporal, 2.61% at the outer superior, 1.86% at the outer inferior, 2.08% at the outer nasal, and 2.84% at the outer temporal subfield. When Bonferroni's correction was applied by setting alpha level at 0.0045, the inverse association between hypertension and macular thickness remained significantly at inner superior, inner inferior, outer superior, and outer inferior fields. 

Table 2

 

Associations Between Hypertension and Macular Thickness

 

When we evaluated an independent association between each of the other cardiometabolic risk factors and macular thickness, after adjusting for all other measured cardiometabolic risk factors as covariates (Table 3), diabetes, LDL-C, HDL-C, BMI, and smoking were not associated with macular thickness. 

Table 3

 

Association* Between Cardiometabolic Factors and Macular Thickness

 

Table 4 shows the findings from the stratified analyses. There was no association between hypertension and macular thickness at the fovea region regardless of the presence of cardiometabolic risk factors, whereas macular thickness at the inner or outer region showed slightly different association with hypertension by the strata of each cardiometabolic risk factor. 

Table 4

 

Stratified Analysis for the Association* Between Hypertension and Macular Thickness According to the Level of Cardiometabolic Risk Factors

 

Macular thickness at a composite outer region was significantly associated with hypertension in both BMI strata, whereas significant association of macular thickness at a composite inner region with hypertension was found limited to subjects with higher BMI (≥25 kg/m²). For levels of LDL-C, significant association was found between hypertension and macular thickness at a composite outer field in both strata of LDL-C. The association between hypertension and macular thickness at a composite outer field was found regardless of HDL-C level, whereas the association between hypertension and macular thickness at a composite inner field was found only in subjects with higher HDL-C. The association between hypertension and macular thickness at composite outer and inner fields was found limited to never-smokers. However, there was no statistically significant interaction by BMI, LDL-C, HDL-C, and smoking on the association between hypertension and macular thickness. Although the association between hypertension and macular thickness at a composite outer region was significant at both strata of glucose level, the strength of association was higher in subjects with elevated glucose level (P for interaction = 0.05). The association between hypertension and macular thickness at a composite inner region was found only in subjects with a higher glucose level (P for interaction = 0.03). When Bonferroni's correction was applied by setting alpha level at 0.0033, the associations between hypertension and macular thickness remained significant only at composite outer region, in never-smokers or in subjects with higher BMI, higher glucose, or lower HDL-C. 

Discussion

Hypertensive retinopathy with acute exudative changes could induce abnormal thickening of macula, as seen in diabetic retinopathy and other occlusive retinal vascular diseases.¹⁸ However, with very few studies on this issue, it is still unclear whether systemic hypertension without exudative retinopathy also induces a change in macular thickness. A cross-sectional study including 92 diabetic patients with mild to no diabetic retinopathy and 92 healthy controls in the United States found no significant relationship between systolic or diastolic blood pressure and retinal thickness at all macular subfields.¹³ However, the previous study mainly focused on the influence of diabetes mellitus and included small numbers of subjects of relatively older age.⁴ Contrary to that study, our study found that systemic hypertension is significantly associated with decreased macular thickness except in the central foveal subfield, even after adjusting for age, sex, axial length, and other cardiometabolic risk factors. 

In the present study, none of the individual cardiometabolic risk factors except for hypertension showed significant association with macular thickness. 

The presence of diabetes mellitus without diabetic retinopathy showed no association with macular thickness at all subfields. This finding corresponds to the findings of previous studies reporting that there was no significant difference in central macular thickness between diabetics without diabetic retinopathy and normal healthy controls.^13,19–21 

There was a study reporting a significant association between BMI and macular thickness at the outer subfield,²² which differs from the finding of our study. However, the previous study was conducted in school-age children and cannot be directly compared with our study. 

There have been several studies that investigated the relationship of LDL-C and HDL-C with central macular thickness, although the studies were confined to diabetic patients.^23,24 One of the studies in Australian adults with diabetes found that serum LDL-C and HDL-C were associated with clinically significant macular edema but not with macular thickness,²³ which is compatible with the findings of our study. However, the other study in Japanese diabetic patients found that LDL-C had a positive association with central subfield macular thickness as well as central subfield macular volume.²⁴ 

Regarding the association between smoking and macular thickness, little is known, and, to the best of our knowledge, the current study is the only study examining the issue. Given the high prevalence of smoking throughout the world, studies are needed to evaluate the effect of smoking on eye health. 

The retina undergoes various pathophysiologic changes with elevated systemic blood pressure, such as vasoconstriction, vascular sclerosis, and exudative changes.^25–27 Although there is no generally accepted grading system for hypertensive retinopathy,²⁸ hypertensive retinopathy produces a wide spectrum of changes in the fundus, from signs of mild vascular narrowing to exudative retinopathy or vascular sclerosis.²⁹ All subjects in this study had no to mild signs of hypertensive retinopathy without any exudative changes or severe vascular sclerosis signs. Thus, it is unlikely that retinal exudation or severe perfusion defects complicated by systemic hypertension have induced change in macular thickness. 

The probable mechanism of macular thinning associated with systemic hypertension without exudative retinopathy is autoregulation of retinal blood vessels induced by increased blood pressure, followed by vascular constriction and mild reduction in macular thickness and volume. Systemic blood pressure was associated with a smaller diameter of the central retinal arteriole,^30,31 and probably such arteriolar change influenced macular thickness, as found in the current study. Interestingly, macular thickness at the central subfield did not show such an association with systemic hypertension. We think this finding is probably because the central subfield is an avascular area (foveal avascular zone), which might not be affected by hypertensive vascular changes. 

The present study has several clinically significant findings and strengths. First, the current study provides objective evidence supporting the concept that central macular thickness is less likely to be influenced by systemic hypertension. Thus, the study design of most clinical trials of treatment for macular disease that frequently use central macular thickness as one of the primary or secondary outcome measures without any strict exclusion criteria for the presence of systemic hypertension seems to be appropriate.^32,33 Second, the findings of the current study also indicate that additional caution is required for interpretation of macular thickness measurements of inner or outer macular subfields, especially when many study subjects have systemic hypertension and the detection of a small difference of macular thickness is needed. Third, the present study is a community-based study, which provides higher generalizability of study findings. Fourth, a wide range of cardiometabolic risk factors could be accounted for in evaluating the hypertension–macular thickness association. In the present study, subjects with systemic hypertension had significant differences in the distribution of age, axial length, and most cardiometabolic risk factors compared with normotensives. Although those variables may not fulfill the criteria for confounder in the association between hypertension and macular thickness, it seems necessary to consider them as covariates to more accurately evaluate an independent association of hypertension with macular thickness, given their relationship with hypertension. 

The current study had some limitations. First, the time-domain OCT used to measure macular thickness in the present study had an axial resolution of 10 μm, which may be inadequate to detect smaller (<10 μm) macular thickness. Thus, under- or overestimation of the association between macular thickness and cardiometabolic factors could have been possible. However, these biases seem less likely to be significant, considering that 1 SD of macular thickness distribution and 1 SD of difference in macular thickness between hypertensives and normotensives were much greater than the axial resolution (10 μm) at most subfields and most macular subfields consistently showed lower mean macular thickness in hypertensives than in normotensives. As a second limitation, the cross-sectional design of our study may have provided limited evidence for causal association between hypertension and macular thickness. Third, although the number of study subjects in our study was relatively larger than the size in other previous studies, we cannot rule out that the null association between macular thickness at those regions and hypertension could have been resulted from the lack of enough power. When we conducted a power analysis on the basis of the sample size of our study with P value setting at 0.05, the power for studying the association between hypertension and macular thickness at the fovea and outer nasal region was revealed to be less than 50%. Fourth, we defined the presence of systemic hypertension and diabetes on the basis of only two repeated measurements at one visit or use of medication. 

In conclusion, this community-based Korean study revealed that macular thickness was inversely associated with systemic hypertension at most macular subfields except for the fovea, and the association was significantly stronger in subjects with an elevated fasting glucose level. Future studies using measurements of extrafoveal macular thickness or macular volume for outcome measures should consider the influence of systemic hypertension on measurements in the analysis. 

Acknowledgments

Supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI #HI14C0064); a basic science research program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (#2012-0004255); and a global research network program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (#2011-220-E00006). The authors alone are responsible for the content and writing of the paper. 

Disclosure: M. Kong, None; Y. Kwun, None; J. Sung, None; D.-I. Ham, None; Y.-M. Song, None