Efficacy of ophthalmic examinations for predicting vascular calcification in patients undergoing maintenance hemodialysis
Article information
Abstract
Background
In maintenance hemodialysis (MHD) patients, vascular calcification can be detected not only in coronary vessels but also in ocular areas. However, ophthalmic examinations are not sufficiently validated to measure the degree of vascular calcification.
Methods
This study was performed prospectively, involving 32 MHD patients. Calcium deposition in the cornea and conjunctiva was checked using a slit lamp and anterior photography. Conjunctival and corneal calcification (CCC) score was calculated and the severity of CCC was graded. Extent of invasion in the corneal limbus and center was identified. Coronary artery calcium (CAC) deposit was scored using computer tomography, and cardiac function was investigated by echocardiogram. We divided patients into two groups: mild and moderate/severe groups according to the CCC scores.
Results
Mean CAC scores were 354.6 and 1,494.2 in the mild and moderate/severe groups. Mean extent of invasion in the corneal limbus and center was significantly higher in the moderate/severe groups than in the mild group. Parathyroid hormone was significantly higher in the moderate/severe groups than in the mild group and ejection fraction was significantly lower in the moderate/severe groups than in the mild group. The CCC score was positively associated with the CAC score, the extent of invasion in the corneal limbus and center, and the parathyroid hormone level. The extent of invasion in the corneal limbus and center was positively associated with the CAC score. The CCC score was negatively associated with ejection fraction.
Conclusion
The CCC score and the extent of invasion in the corneal limbus and center can predict vascular calcification in MHD patients.
Introduction
As kidney function decreases, changes in calcium and phosphorus concentrations in the blood and tissues occur, as well as changes in hormones such as parathyroid hormone (PTH) and vitamin D, which control these minerals. Abnormal mineral regulation associated with chronic kidney disease (CKD) increases in severity as CKD progresses to end-stage renal disease (ESRD). In particular, patients on hemodialysis have a very high mortality rate due to cardiovascular disease because vascular calcification is a risk factor for cardiovascular disease in patients with ESRD [1]. Therefore, it is important to detect vascular calcification early and stably control abnormalities in mineral regulation. Because blood vessel/valve calcification is highly correlated with the risk of cardiovascular complications, blood flow in the carotid artery is measured using carotid ultrasound as a screening test, or coronary artery calcification (CAC) is scored using coronary artery computed tomography (CT). The 2019 American College of Cardiology/American Heart Association guidelines state that the CAC score using coronary artery CT should be considered in patients with ESRD [2–5]. A simpler method is recommended to evaluate the degree of aortic calcification on chest or abdominal radiographs [6,7], or to determine whether valve calcification is performed using echocardiography [8]. However, when measuring the intima-media thickness of carotid ultrasound or performing echocardiography, the range of results varies greatly depending on the measurement by the ultrasound operator; moreover, coronary artery CT is very expensive and involves high radiation exposure. There is a risk, and in the case of chest and abdominal radiography, the specificity is low. Therefore, it is difficult to accurately reflect the blood vessel status of patients undergoing hemodialysis. It is reported that calcification can be detected not only in blood vessels but also in ophthalmic examinations in patients undergoing hemodialysis [9]. Especially, conjunctival and corneal calcification (CCC) is a well-known and easily detectable extraskeletal calcification that is the most common form of metastatic calcification in patients with ESRD [10]. Therefore, ophthalmic examinations might be used to measure the degree of vascular calcification more easily; however, they have not been sufficiently validated.
Therefore, we aimed to investigate the association between ophthalmic findings and vascular calcification, and to investigate the effectiveness of ophthalmic exam for the prediction of the risk of vascular calcification in patients undergoing maintenance hemodialysis (MHD).
Methods
Study population
The authors have 220 patients with maintenance dialysis in the dialysis center and enrolled 32 patients who have agreed to this study and without ophthalmologic treatment (n = 73). Thirty-two patients who had undergone MHD for >3 months at Keimyung University Dongsan Hospital were enrolled prospectively. After hemodialysis was started, all patients from other hospitals, those who had undergone kidney transplantation, those who were diagnosed and treated for cardiovascular diseases (especially those with a history of treatment for angina pectoris or myocardial infarction) before dialysis, those who underwent refraction correction surgery (laser-assisted in situ keratomileusis or laser-assisted subepithelial keratectomy), those who had uveitis or were currently being treated, those who had received retinal laser treatment, those who had undergone retinal surgery (excluding cataract surgery), and those accompanying other ophthalmic conditions (e.g., ocular inflammation, previous ocular trauma, intravitreal silicone use, corneal transplantation, chemical ocular burn) were excluded.
This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of Keimyung University Dongsan Hospital (No. 2018-07-024). Informed consent was obtained from all the participants.
Classification of conjunctival and corneal calcification
Ophthalmic examinations and grading were performed by two ophthalmologists. The anterior segment was checked for calcium deposition in the cornea and conjunctiva using a slit lamp and anterior photography. They used a slit-lamp biomicroscope to visually detect white, chalky calcifications on the conjunctiva and cornea. Different illumination techniques help distinguish the depth and location of the deposits. They captured the images for clinical documentation and progression tracking using high-resolution slit-lamp cameras. The severity of CCC was graded between 0 and 5 using the Porter and Crombie classification [11]: grade 0 means normal with no deposits in the conjunctiva or cornea (A, I); grade 1 means conjunctival calcium deposits only (B, I); grade 2 means irregular corneal deposits and conjunctival deposits (C, II); grade 3 means single line of corneal deposits and conjunctival deposits (D, II); grade 4 means increased corneal deposits, often as two lines (E, III), and conjunctival deposits; and grade 5 indicates extensive corneal deposits, often as three lines, and conjunctival deposits (F, III). The CCC score was calculated based on the Tokuyama method [12]. We divided the study population into three groups according to the CCC score: mild (0–2), moderate (3–5), and severe (6–10) [13]. The extent of invasion into the corneal limbus and center was identified using anterior segment optical coherence tomography (AS-OCT). AS-OCT imaging is a noncontact imaging technique that produces high-resolution images and quantitative measurements of the anterior segment and its anatomical structures. AS-OCT shows excellent repeatability and reproducibility for the measurement of corneal thickness, anterior chamber angle, and anterior chamber depth. With real-time video recording, AS-OCT can also effectively capture the dynamic changes of the iris and lens in response to light and eye accommodation. AS-OCT, specially designed for anterior segment imaging was developed. This new model employs a superluminescent diode of a longer wavelength (1,310 nm). With the use of this model, about 90% of the 1,310 nm light is absorbed by the ocular media before hitting the retina, allowing us to use light that is 20 times stronger before reaching the retinal exposure limit. With the improvements in speed and light penetration, the entire anterior segment can now be captured in one frame, in 0.125 seconds, and in 18-μm resolution (Visante OCT User Manual, 2006) [14,15].
Grading of coronary artery calcium score
CT images of four individuals with varying degrees of CAC were studied. The CAC score was calculated using the Agatston method using CT. On CT images, calcification was expressed as a white lesion (>130 Hounsfield units). Calcification in the left anterior descending artery is indicated by the arrows [16]. Qualitative CAC assessment (none, mild, moderate, or severe) is closely correlated with traditional CAC scores (0, 1–100, 101–400, and >400) [17].
Transthoracic echocardiography
To determine the extent to which ophthalmic problems and vascular calcification affect cardiac function, cardiovascular disease, and heart failure, transthoracic echocardiography (TTE) was performed during ophthalmic examinations and coronary CT. We performed TTE the day after hemodialysis to minimize the influence of volume status in patients undergoing dialysis. All enrolled patients maintained their usual dry weight through dialysis. We evaluated the diastolic left ventricular internal dimension (LVIDd), systolic left ventricular internal dimension (LVIDs), left ventricular ejection fraction (LVEF), left atrial (LA) diameter, early diastolic mitral inflow velocity/late diastolic mitral inflow velocity (E/A) ratio, and early diastolic mitral inflow velocity/early diastolic mitral annular tissue velocity (E/E’) ratio. We also checked the calcification status of cardiac valves.
Laboratory parameters
Patients with abnormal findings on investigation for secondary hyperparathyroidism, abnormal serum calcium and phosphorus levels, and lack or deficiency of vitamin D among those who had been receiving dialysis after 3 months and had started hemodialysis at Keimyung University Dongsan Hospital were enrolled. Dialysis adequacy (Kt/V) was done. Systolic blood pressure, diastolic blood pressure, hemoglobin, urea nitrogen, creatinine, estimated glomerular filtration rate, calcium, phosphorus, calcium × phosphorus product, PTH, 25-hydroxyvitamin D (25-OHD), ferritin, β2 microglobulin, albumin, uric acid, and total cholesterol were measured at the start of the study. During ophthalmic examination, the degree of calcium deposition in the cornea and conjunctiva was graded, and AS-OCT was performed to identify the area of the calcium deposits.
Statistical analysis
Continuous variables are expressed as means with standard deviations for the number of observations using the Kruskal-Wallis H test or Mann-Whitney U test, and categorical variables are expressed as numbers with percentages in brackets using the chi-square test or Fisher exact test. All data were routinely tested for normality of distribution and equality of standard deviations before analysis. The correlation between CCC score, CAC score, and the extent of invasion in the corneal limbus and center were evaluated by means of the Spearman rank correlation. The criterion for significance was a p-value <0.05. All statistical analyses were performed using the IBM SPSS version 19.0 (IBM Corp.).
Results
Baseline characteristics of the study population
The baseline characteristics of the study population are shown in Table 1. The mean age of the study population was 59.8 ± 11.8 years and the proportion of sex was the same. The mean dialysis vintage was 67.2 ± 61.3 months. The calcium levels were 9.2 ± 0.9 mg/dL; phosphate, 5.2 ± 1.7 mg/dL; PTH, 264.6 ± 169.5 pg/mL; 25-OHD, 21.6 ± 13.4 ng/mL; and total cholesterol, 136.9 ± 27.8 mg/dL. The incidence of CCC in this study was 43.8% (14 of 32), while the incidence reported in the reference literature was higher, at 87.3% (55 of 63). The mean CAC score was 696.9 ± 1,270.6. The CCC score was 1.7 ± 2.6, and the extent of invasion in the corneal limbus and center was 175.5 ± 333.9 mm (Table 2). Cardiac function in patients undergoing MHD is presented in Table 2. LVIDd was 5.4 ± 0.8 cm; LVIDs, 3.7 ± 1.1 cm; LVEF, 55.8% ± 15.3%; LA diameter, 4.5 cm ± 0.8 cm; E/A ratio, 0.9 ± 0.3; and E/E’ ratio, 12.1 ± 5.2. The baseline characteristics of the study population based on the CCC scores are shown in Table 1. As mentioned above, the study population was divided into two groups based on the CCC scores: mild, and moderate/severe. There were no significant differences in age, sex, body mass index, cause of ESRD, comorbidities, dialysis vintage, dialysis adequacy, systolic blood pressure, diastolic blood pressure, or medications for CKD-MBD based on the CCC scores. Laboratory findings showed that PTH levels were significantly higher in the moderate/severe groups than in the mild and moderate groups (p = 0.03). Other laboratory findings were not significantly different among the two groups.
Baseline characteristics of the study population based on the score of conjunctival and corneal calcification
Calcification status and echocardiographic findings based on the score of conjunctival and corneal calcification
Calcification status and echocardiographic findings based on the score of conjunctival and corneal calcification
The calcification status and echocardiographic findings based on the CCC scores are described in Table 2. We classified and compared the 32 patients according to the severity of their CCC grade and CAC score. The CCC mild group had 23 patients, and the moderate/severe groups had nine. The mean CAC score was 354.6 ± 765.8 in the mild group, and 1,494.2 ± 1,857.5 in the moderate/severe groups (p = 0.03). We made modifications in Table 2 according to the CAC scores (0, 1–100, 101–400, and >400) and qualitative CAC assessment (none, mild, moderate, or severe). The mean extent of invasion in the corneal limbus and center was significantly higher in the moderate/severe groups than in the mild group (p = 0.03). The mean CAC score was also significantly higher in the moderate/severe groups than in the mild group (p = 0.03). In the echocardiographic data, ejection fraction was significantly lower in the moderate/severe groups than in the mild group (p = 0.02). There was no significant difference in the proportion of valvular calcification among the groups.
Association among the conjunctival and corneal calcification scores and extent of invasion in the corneal limbus and center and coronary artery calcium score
The association between the CCC score and the extent of invasion in the corneal limbus and center, and the CAC score is shown in Fig. 1. The CCC score was positively associated with the CAC score (R2 = 0.297) and extent of invasion in the corneal limbus and center (R2 = 0.570). The extent of invasion in the corneal limbus and center was also positively associated with the CAC score (R2 = 0.365). The CCC score was only negatively associated with ejection fraction.
Association among conjunctival and corneal calcification score and extent of invasion in the corneal limbus and center and coronary calcium score.
Variables associated with scores of conjunctival and corneal calcifications in patients undergoing maintenance hemodialysis
The CCC score was positively associated with PTH level (R2 = 0.434, p = 0.01). The CCC, P or CCC and Ca × P values tended to be related (R2 = 0.333, p = 0.06; R2 = 0.316, p = 0.08). There were no significant differences between the CCC score and dialysis duration, dialysis adequacy, systolic blood pressure, diastolic blood pressure, medications for CKD-MBD, calcium levels, or 25-OHD levels. There was no significant difference between the CCC score and age.
Discussion
Vascular calcification is associated with high cardiovascular morbidity and mortality in patients undergoing MHD [18]. In this study, we evaluated the status of vascular calcification by using a coronary CT scan. The results showed that the mean CCC score was higher in patients with higher CAC scores than in those with lower scores. The CCC score was positively associated with the CAC score, extent of invasion in the corneal limbus and center, and PTH levels. The extent of invasion in the corneal limbus and center was also positively associated with the CAC score. The CCC score was only negatively associated with ejection fraction. When we enrolled the patients, we excluded the patients with a history of cardiovascular disease (especially those with a history of treatment [medical or interventional treatment] for angina pectoris or myocardial infarction) before dialysis. After enrollment, we performed echocardiography and there were four patients with less than 40% ejection fraction, and all were patients without the histories of coronary heart disease and related symptoms before enrollment. We think that their ejection fraction can be influenced by the problem of coronary vessels close to the enrollment time. Therefore, it is possible that many patients selected for this study had ischemic heart disease, which may explain the observed relationship with ejection fraction, as it reflects both coronary calcification and heart function. According to these results, CCC reflects vascular calcification. When investigating the risk of cardiovascular disease, coronary CT is the current investigation of choice. However, CT involves high costs and significant radiation exposure. Therefore, the clinical evaluation of ocular calcification is relatively inexpensive and noninvasive compared to coronary CT or TTE. The significant correlation between CCC and the extent of invasion in the corneal limbus and center and the CAC score might suggest some possible clinical applications of ocular examination. Importantly, for the CCC score to be used to predict vascular calcification in MHD patients, it would be meaningful if the incidence of CCC were high among the entire MHD patient population. The incidence of CCC in this study was 43.8% (14 of 32), while the incidence reported in the reference literature was higher, at 87.3% (55 of 63). Compared with the incidence in the reference literature, the incidence in this study was low, but it is not insufficient to prove that the CCC score is a helpful test for predicting vascular calcification.
On the contrary, a recent study reported no association between CCC and coronary calcium score [19]. We have conducted research to provide additional scientific evidence based on these studies, making it applicable in clinical practice. This study was a cross-sectional analysis with 34 patients, and while the primary outcomes are quite similar, the reviewer highlighted that a slit lamp, a method with lower sensitivity, was used. They noted that their study utilized AS-OCT, which has higher sensitivity, dividing the participants into an absent/mild group and a moderate/severe group for comparison. In our study, the anterior segment was examined for calcium deposits in the cornea and conjunctiva using a slit lamp and anterior photography, with the severity of CCC graded from 0 to 5 using the Porter and Crombie classification: grade 0 indicates normal with no deposits in the conjunctiva or cornea; grade 1 represents conjunctival calcium deposits only; grade 2 signifies irregular corneal deposits along with conjunctival deposits; grade 3 includes a single line of corneal deposits with conjunctival deposits; grade 4 involves increased corneal deposits, often as two lines, with conjunctival deposits; and grade 5 denotes extensive corneal deposits, often as three lines, with conjunctival deposits. The CCC score was calculated using the Tokuyama method, dividing the study population into mild, moderate, and severe groups for comparison: mild (0–2), moderate (3–5), and severe (6–10). While these classification methods differ, our study employed evidence-based classification and additionally utilized AS-OCT to offer more objective support for the existing method. Correlation analysis between the CCC score and the extent of invasion into the corneal limbus and center (measured using AS-OCT) demonstrated a significant relationship, further validating the existing slit lamp-based method. Ophthalmology experts noted that, due to the risks associated with contrast media and the high cost of AS-OCT, using the current methods and classifications, which have sufficient correlation and are easier to measure, would be advantageous in screening a larger number of patients. Furthermore, analysis according to this classification confirmed a correlation between CCC and CAC scores, and our paper also explored the correlation with cardiac function using TTE, which we consider a strength of this study.
Ophthalmic examination for assessing CCC has the advantage of being noninvasive, showing high sensitivity. Regular ophthalmological examinations are required to evaluate CCC, necessitating close cooperation with an ophthalmologist. It may be challenging to perform routinely in clinical practice because of the examination consultation with an ophthalmologist, but according to the ophthalmologist’s opinion, this test is straightforward and can be easily performed by an ophthalmologist in actual clinical setting. Therefore, this simple ophthalmological examination may be more effective in predicting vascular calcification in hemodialysis patients than invasive and expensive tests.
In patients undergoing MHD, most vascular calcifications are associated with CKD-MBD [20]. In other words, vascular calcification is associated with calcium, phosphorus, PTH, and 25-OHD levels. However, in our study, the PTH level was only associated with the CCC score (R2 = 0.434, p = 0.01), but calcium, phosphate, and Ca × P levels were not associated with the CCC score. None of the CKD-MBD laboratory findings were associated with the CAC score. Additionally, we examined the patients’ treatment histories for CKD-MBD in Table 1. None of the patients had a history of parathyroidectomy, and all patients, except one, were treated with phosphate binders, vitamin D, and calcimimetics. It is known that controlling CKD-MBD through these medications can help prevent the progression of vascular calcification in dialysis patients and improve vascular health. This study aimed to demonstrate the relationship between CCC and CAC through a cross-sectional analysis. However, it has limitations in assessing the impact of CKD-MBD treatment on vascular improvement. Although comparing the CCC of patients who underwent parathyroidectomy with those who did not could provide insights, none of the patients in this study had undergone parathyroidectomy. Furthermore, since each patient has a unique CKD-MBD status and uses different types and doses of treatment medications, there are limitations in comparing CCC before and after CKD-MBD treatment, as well as changes in CAC. While CCC scores are significantly associated with CKD-MBD parameters such as serum calcium, phosphate, and PTH levels, CAC scores are not, likely due to differences in pathophysiology. CCC reflects local calcium-phosphate metabolism in avascular tissues sensitive to biochemical changes, whereas CAC involves more complex vascular processes influenced by inflammation, lipid metabolism, and endothelial dysfunction, making it less directly related to CKD-MBD markers. I believe that conducting a prospective study to further examine this correlation could yield more meaningful results.
Dialysis vintage is a well-known risk factor for vascular calcification. However, our study did not demonstrate statistically significant results between CAC score and dialysis vintage (correlation coefficient 0.333, p = 0.06). Similarly, we did not find statistically significant results between dialysis vintage and CCC. While dialysis vintage is generally associated with vascular calcification and is believed to impact both CCC and CAC, statistical significance was not achieved between dialysis vintage, CCC, and CAC in this study. However, a positive correlation was observed when CCC and CAC scores were directly compared, which can be considered meaningful.
This study utilized existing evidence for classification and additionally used AS-OCT to provide more objective support for the current classification method. In this study, CCC scores and AS-OCT were employed to measure the extent of invasion into the corneal limbus and center, and correlation analysis demonstrated a significant relationship, sufficiently validating the existing slit lamp-based research method. According to ophthalmology experts, the high cost of AS-OCT is considered. The existing method and classification, which show a strong correlation and are easier to measure, offer an advantage for screening a larger number of patients without the need to perform AS-OCT on all patients. The strength of our paper lies in demonstrating a correlation between CCC and CAC scores.
Therefore, the CCC score and extent of invasion into the corneal limbus and center are convenient, noninvasive, and inexpensive methods to evaluate the extent of vascular calcification in patients undergoing MHD and might be used as indirect indicators to predict cardiovascular risks in these patients.
This study had some limitations. There were small sample sizes, different dialysis times, and prevalence of CKD-MBD among the study population, cross-sectional studies, and short follow-up durations. Therefore, the extent of vascular calcification may differ. We could not definitely explain the underlying mechanisms of increased the CAC score and CCC score because histological examination should be performed for the exact identification of calcifications. Further studies are required to better understand the association between CAC and CCC. However, this study had several strengths. We used more accurate and sensitive examination tools, such as coronary CT, TTE, slit lamp, and AS-OCT, to evaluate vascular calcification in patients undergoing MHD, compared to other studies using plain radiography. We can predict the relationship between the ocular state and vascular calcification. Therefore, the CCC score and extent of invasion in the corneal limbus and center might be used as markers of the presence of vascular calcifications.
In conclusion, we have to look into vascular calcification in patients undergoing MHD as they are a very high-risk group for cardiovascular disease. Although the treatment of CKD-MBD for the prevention of vascular calcification is important, early detection of vascular calcification is more important. There are many tools for screening, but they are expensive and time-consuming. Nowadays, the need for convenient and inexpensive methods for the detection of vascular calcification is emphasized. Our study showed that the CCC score and extent of invasion in the corneal limbus and center might reflect vascular calcification and predict the risk of cardiovascular disease in patients undergoing MHD. Larger populations and longitudinal studies are needed to evaluate the value of the CCC score as a representative marker of cardiovascular morbidity and mortality in patients undergoing MHD.
Notes
Conflicts of interest
All authors have no conflicts of interest to declare.
Funding
This study was supported by the Daewon Pham. A National Research Foundation of Korea (NRF) grant was funded by the Korean government (Ministry of Science and ICT, MSIT) (2021R1F1A1061572).
Data sharing statement
The data presented in this study are available from the corresponding author upon reasonable request.
Authors’ contributions
Conceptualization, Validation: WYP, KTK, JHJ, YCK, KJ
Data curation: YK, JHP, SH
Formal analysis, Project administration: WYP
Funding acquisition: WYP, KJ
Investigation, Software: WYP, YK, JHP, SH
Methodology, Resources: All authors
Supervision, Visualization: YCK, KJ
Writing–original draft: All authors
Writing–review & editing: All authors
All authors read and approved the final manuscript.