- Original investigation
- Open Access
Diabetic retinopathy is associated with diastolic dysfunction in type 2 diabetic patients with non-ischemic dilated cardiomyopathy
Cardiovascular Diabetologyvolume 16, Article number: 82 (2017)
To investigate the association between diabetic retinopathy (DR) and myocardial dysfunction in patients with type 2 diabetes and dilated cardiomyopathy (dCMP).
Data were collected retrospectively from 89 patients with dCMP (46 with type 2 diabetes and 43 without diabetes) and no evidence of coronary artery disease. Echocardiographic parameters and laboratory data, including lipid profiles and fundus findings, were obtained from medical records. A left ventricular ejection fraction (LVEF) less than 40% was considered impaired systolic function, while an E/E′ ratio greater than 15 was considered elevated left ventricular (LV) filling pressure.
Baseline characteristics show that LVEF was not significantly different between patients with and without diabetes or between diabetic patients with and without DR. Among the diastolic function parameters, patients with DR exhibited higher E/E′ ratios (left ventricular filling pressures) than patients without DR (23.75 ± 13.37 vs 11.71 ± 3.50, P = 0.022). Logistic regression analysis revealed that statin use lowered the risk of impaired systolic dysfunction in all patients (odds ratio (OR) 0.33, 95% confidence interval (CI) 0.12–0.92, P = 0.034) and in patients with diabetes (OR 0.273, 95% CI 0.08–0.99, P = 0.049), while the presence of DR was associated with a higher risk of elevated LV filling filling pressure in patients with diabetes (OR 18.00, 95% CI 1.50–216.62, P = 0.023).
In conclusion, DR was associated with elevated LV filling pressure in patients with dCMP. DR may not only represent microvascular long-term complications in patients with diabetes but may also be associated with more advanced form of diastolic dysfunction among diabetic patients with cardiomyopathy.
Cardiovascular disease (CVD), including coronary artery disease and stroke, is the leading cause of death in patients with diabetes and directly related to atherosclerosis [1, 2]. Although the excessive risk of CVD in patients with diabetes may be due to common comorbidities, such as dyslipidemia, hypertension, and smoking, it is well known that diabetes alone can induce molecular changes in the heart . The “common soil” hypothesis of diabetic complications has been introduced through several studies on the molecular mechanisms of diabetes [1,2,3]. Chronic hyperglycemia results in many microvascular complications in the eyes, nerves and kidneys, as well as higher risk of all macrovascular complications, including coronary and cerebrovascular disease [1, 4]. Diabetic retinopathy (DR), one of the major microvascular complications of diabetes, is also known to predict cardiovascular diseases and CVD-related death in individuals with type 2 diabetes .
Dilated cardiomyopathy (dCMP) is defined by the presence of left ventricular systolic dysfunction in the absence of an abnormal loading condition or significant coronary artery disease . Endocrine disorders, including diabetes, are known to be associated with dCMP. Recent studies have revealed the presence of diabetic cardiomyopathy, a rare condition of myocardial dysfunction without coronary artery disease [1, 2, 7]. This term was first introduced by Rubler et al.  in 1972 and described patients with diabetes and congestive heart failure with normal coronary arteries. The exact pathophysiological mechanisms are still under investigation, while oxidative stress, impaired mitochondrial function, activation of the renin-angiotensin system, and altered substrate metabolism have been suggested as possible contributors to the pathogenesis [2, 7]. Since these mechanisms share common pathways with diabetic microvascular complications, diabetic cardiomyopathy is considered to indicate a microvascular component .
The presence of DR in diabetic patients suggests that microvascular complications are manifested clinically. Accordingly, we performed a retrospective clinical study to investigate the association of diabetes or DR with myocardial function in patients with type 2 diabetes with non-ischemic dCMP compared with that in patients without diabetes.
The medical records of patients diagnosed with cardiomyopathy between 1994 and 2015 and followed by the Ophthalmology and Cardiology departments of Ajou University Hospital (Suwon, Korea) were retrospectively reviewed. This study complied with the Declaration of Helsinki and was approved by the Institutional Review Board of Ajou University Hospital (#AJIRB-MED-MDB-16-542). Patients were excluded if fundus examinations revealed any retinal vascular diseases other than diabetic retinopathy; without records of fundus findings and echocardiographic results; or with myocardial dysfunction related to secondary causes. Detailed exclusion criteria were as follows: the presence of significant coronary artery disease (>50% stenosis of at least one major coronary artery) confirmed by coronary angiography or non-performed angiography, or myocardial dysfunction resulted from significant valvular disease (symptomatic patients or asymptomatic patients with criteria of valvular heart disease by 2014 AHA/ACC guidelines ) or abnormal hemodynamic loading conditions. We also excluded patients with hypertrophic and restrictive cardiomyopathy to avoid heterogeneity in the echocardiographic data.
Demographic and clinical factors were obtained from medical records: age, gender, body mass index, general medical illness, presence of diabetes and DR, serum lipid profile, estimated glomerular filtration rate (eGFR), and medications including anticoagulants, β-blockers, angiotensin-converting enzyme inhibitor (ACEI), angiotensin receptor blocker (ARB), calcium channel blocker (CCB), diuretics and statins. Those without diabetes were defined by HbA1c <6.5%, fasting blood glucose <126 mg/dL, and lack of antidiabetic agents, verified from medical records. Fundus findings and echocardiographic parameters were also obtained from medical records. The collected echocardiographic data included left ventricular ejection fraction (LVEF), left ventricular fractional shortening (LVFS), left ventricular end-diastolic dimension (LVEDD), left ventricular end-systolic dimension (LVESD), left ventricular mass index (LVMI), posterior wall thickness at end-diastole (PWd) and at end-systole (PWs), relative wall thickness (RWT), peak early diastolic mitral inflow velocity (E), and peak late diastolic mitral inflow velocity (A). The early diastolic mitral annular velocity (E′) was averaged from the two measurements at septal and lateral sides of mitral annulus. The presence and severity of DR were obtained from medical records of ophthalmology department, based on the traditional Early Treatment Diabetic Retinopathy Study (ETDRS) grading system using fundus photos and fluorescence angiography of patients by an experienced retina specialist.
Statistical analysis was performed using SPSS software (version 23.0, SPSS, Chicago, IL). Shapiro–Wilk test was used for the assessment of the Gaussian distribution, and nonparametric tests were applied if the assumption of normality was violated. Categorical variables were compared using the Chi square test, and continuous variables were compared using independent t test, Mann–Whitney test, or Kruskal–Wallis test. Logistic regression analysis was performed to evaluate factors associated with impaired systolic function or elevated left ventricular (LV) filling pressure. Systolic dysfunction was defined as LVEF <40%, while elevated LV filling pressure was defined as E/E′ ratio >15 [10,11,12,13,14]. P < 0.05 were considered statistically significant.
89 patients with dCMP were included in this study, and their demographic characteristics are summarized in Table 1. In the laboratory profile, creatinine was increased in patients with diabetes, and eGFR was decreased in these patients. Hypertension was more common in patients with diabetes, while the use of statins was also more frequent in patients with diabetes. A subgroup analysis of the patients with diabetes was also performed, and 29 patients had DR, and 17 did not have DR (Table 2). Among the patients with DR, 18 presented with non-proliferative DR, and 11 presented with proliferative DR. The creatinine levels were significantly higher in the patients with DR. Total cholesterol and the levels of high-density lipoprotein cholesterol were higher in patients without DR.
The echocardiographic parameters are described in Table 3 (patients with or without diabetes) and Table 4 (patients with diabetes and with or without DR). There were no significant differences in the parameters between the patients with and without diabetes. However, the values of the E′ and E/E′ ratio among the diastolic parameters showed significant differences according to the presence of DR (P = 0.021 and P = 0.022, respectively). These parameters were also significantly different when compared among those without diabetes, those with diabetes but no DR, and those with DR (P = 0.019 for E′ and P = 0.022 for E/E′ ratio, respectively) (Additional file 1: Tables S1, S2).
Logistic regression analysis was performed to investigate the risk factors associated with impaired systolic or diastolic dysfunction in patients with dCMP for the following factors: age, gender, hypertension, diabetes, DR, eGFR, lipid profiles, and systemic medications. Among these factors, the regression analysis investigating the risk factors associated with systolic dysfunction showed that statin use significantly lowered the degree of impairment in LVEF in the whole study population (OR 0.33, 95% CI 0.12–0.92, P = 0.034, Table 5) and patients with diabetes (OR 0.273, 95% CI 0.08–0.99, P = 0.049, Table 6). For the factors associated with diastolic dysfunction, the presence of DR was significantly associated with the risk of an E/E′ ratio >15 in the patients with diabetes (OR 18.00, 95% CI 1.50–216.62, P = 0.023, Table 6), while the other factors were not significant.
Microvascular complication (DR) and macrovascular complication (CVD) of diabetes
Diabetes is responsible for various cardiovascular complications, such as myocardial infarction, stroke, and peripheral vascular disease . These diseases are at least twofold more common in patients with type 2 diabetes than in individuals without diabetes . CVD is the most important complication in diabetic people, and coronary artery disease is the main cause of death in over 50% of patients with type 2 diabetes . Traditionally, microvascular and macrovascular complications were studied and treated as distinct aspects of diabetes, while many evidences suggested common pathophysiological features between these diabetic complications . However, a close relationship between microvascular complication (DR) and macrovascular complication (CVD) recently. DR in patients with normal renal function and without cardiovascular disease was associated with a higher atherosclerotic burden in the carotid arteries . The presence of DR was independently associated with diastolic and systolic impaired function, both at rest and stress, evaluated by global longitudinal strain and diastolic function reserve index, and might be a useful predictor of major adverse cardiac events such as cardiac death, myocardial infarction, and acute heart failure following percutaneous coronary intervention [21, 22]. The progression of DR, the presence of proliferative DR (PDR), was associated with higher risk of having coronary heart disease (CHD), and was correlated with the severity of CHD . In similar contexts, the impairment of the heart muscle perfusion at stress and rest in PDR patients was more frequent than in the non-proliferative DR (NPDR) patients and diabetic patients without DR suggesting PDR as a useful indicator of heart muscle perfusion disturbance in SPECT studies .
Microvascular complication (DR) and cardiac microangiopathy (diabetic cardiomyopathy)
In addition to these macrovascular complications, there is evidence of microvascular damage in the hearts of patients with diabetes [1, 7]. The disease entity of “diabetic cardiomyopathy” has been introduced and is characterized by the presence of myocardial dysfunction with non-significant coronary arteries [2, 7, 8]. The exact pathophysiology of diabetic cardiomyopathy is unknown, but metabolic changes such as hyperglycemia, dyslipidemia, insulin resistance, and activation of the renin-angiotensin system are thought to result in myocardial fibrosis [1, 2]. Structural changes such as endothelial swelling and/or degeneration and thickening of the capillary basement membrane are also present in diabetic cardiomyopathy, which suggests a similar pathogenesis with microangiopathy [1, 23]. Based on these findings, the clinical phenotype of diabetic cardiomyopathy mostly corresponded to dCMP characterized by left ventricular dilatation and left ventricular systolic dysfunction [2, 6, 24].
The presence of microangiopathy in the heart was reported to show thickening of the capillary basement membrane, microvascular spasm, and capillary microaneurysms . These are also representative features of DR, which suggests a common pathophysiology in the heart. The metabolic complications of diabetes originate from hyperglycemia, which results in the formation of advanced glycation end products and production of reactive oxygen species, followed by vascular endothelial dysfunction [25,26,27]. Moreover, recent study proposed that hyperglycemia-related hyperosmolarity promoted inflammation and angiogenesis by COX-2 expression, and may have a role not only in microvascular disease but also in macrovascular disease . These factors act alone or in combination to promote myocardial fibrosis due to diabetes-induced microangiopathy [25,26,27].
DR and diastolic dysfunction in dCMP
Our study focused on the association of diabetes or DR with dCMP, and we especially investigated the detailed changes in echocardiographic findings. The systolic and diastolic parameters revealed no significant difference according to the presence of diabetes because the study population was confined to those diagnosed with dCMP. However, the parameters, such as E/A ratio, E′ or E/E′ ratio, that represent diastolic dysfunction, were impaired in patients with DR compared to those in patients with diabetes and without DR, which suggests that DR rather than diabetes may be associated with worsened form of diastolic dysfunction in dCMP [7, 28, 29]. It is believed that the use of ACEI/ARB for patients with diabetes as the first choice of antihypertensive drugs in general may protect against the fibrotic changes in the left ventricle, since increased activation of the renin-angiotensin-aldosterone pathway leads to fibrosis formation . It is only after the small vessel diseases, such as retinopathy or nephropathy, are present, which indicate widespread systemic microcirculation diseases, that the characteristics of diabetic cardiomyopathy are prominently evident.
The presence of DR suggests that the patient is experiencing microvascular complications. DR, one of the major microvascular complications in patients with diabetes, has been investigated as a potential predictor for cardiovascular diseases [5, 19, 30, 31]. Studies performed with patients with type 2 diabetes have revealed that the presence of DR is associated with an excess risk of heart failure or cardiovascular mortality [5, 30]. Furthermore, decreased eGFR was noted in patients with diabetes compared to that in non-diabetic patients, as well as in patients with DR compared to patients with diabetes and without DR. The decreased eGFR shown in this study may suggest systemic microvascular damage along with DR, since the associations between diabetic nephropathy and retinopathy have been well documented . These studies suggest a possible contribution of microvascular damage to macrovascular diseases in patients with diabetes [5, 30]. However, lower eGFR did not significantly increase the risk of diastolic dysfunction (OR 0.969, 95% CI 0.941–0.998, P = 0.039). Altogether, the presence of retinopathy rather than nephropathy could be used as a clinical feature to determine whether patients with diabetes are likely to have elevated LV filling pressure in dCMP (Fig. 1).
Statin use and dCMP
Among the systemic medications, statin use was associated with a lower degree of systolic dysfunction in patients with dCMP. There are several experimental studies that have reported the protective effect of statin use on diabetic cardiomyopathy [33, 34], while controversy exists in human studies [35, 36]. We previously demonstrated the protective effect of statin use in DR , which may also be protective in diabetes-induced cardiac microangiopathy as shown in this study. However, further larger scale studies are needed to verify the protective effect of statin on dCMP, as the small number of patients included in this study is a major limitation.
The present study has also other limitations with regard to the retrospective design; more detailed assessment of diastolic profiles such as pulmonary venous flow, left atrial volume, or tricuspid regurgitation pressure gradient was not available. Furthermore , the measurement of LV end-diastolic pressure by catheter examination would be informative, which was not available in this retrospective study. Cohort studies with larger numbers of subjects may be needed for further investigation of diabetes, DR, and cardiomyopathy. Some large ranges of 95% CI in the logistic regression analyses, which were especially common in the subgroup of patients with diabetes, may be narrowed with statistical significance if analyzed with a larger number of patients.
In conclusion, DR was associated with diastolic dysfunction in patients with dCMP. DR may not only represent microvascular long-term complications in patients with diabetes but may also be associated with more advanced form of diastolic dysfunction among diabetic patients with cardiomyopathy. Based on our findings, dCMP patients with DR should be encouraged to increase the frequency of cardiovascular follow-up, and intensive health education on life style modification and special attention to medication adherence are needed to avoid worsening the condition.
peak late diastolic mitral inflow velocity
angiotensin-converting enzyme inhibitor
angiotensin receptor blocker
calcium channel blocker
peak early diastolic mitral inflow velocity
early diastolic mitral annular velocity
estimated glomerular filtration rate
left ventricular end-diastolic dimension
left ventricular end-systolic dimension
left ventricular ejection fraction
left ventricular fractional shortening
left ventricular mass index
posterior wall thickness at end-diastole
posterior wall thickness at end-systole
relative wall thickness
Laakso M. Heart in diabetes: a microvascular disease. Diabetes Care. 2011;34(Suppl 2):S145–9.
Mizamtsidi M, Paschou SA, Grapsa J, Vryonidou A. Diabetic cardiomyopathy: a clinical entity or a cluster of molecular heart changes? Eur J Clin Invest. 2016;46:947–53.
Brownlee M. The pathobiology of diabetic complications: a unifying mechanism. Diabetes. 2005;54:1615–25.
Laakso M. Cardiovascular disease in type 2 diabetes from population to man to mechanisms: the Kelly West Award Lecture 2008. Diabetes Care. 2010;33:442–9.
Juutilainen A, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Retinopathy predicts cardiovascular mortality in type 2 diabetic men and women. Diabetes Care. 2007;30:292–9.
Elliott P, Andersson B, Arbustini E, Bilinska Z, Cecchi F, Charron P, Dubourg O, Kuhl U, Maisch B, McKenna WJ, et al. Classification of the cardiomyopathies: a position statement from the European Society Of Cardiology Working Group on Myocardial and Pericardial Diseases. Eur Heart J. 2008;29:270–6.
Boudina S, Abel ED. Diabetic cardiomyopathy revisited. Circulation. 2007;115:3213–23.
Rubler S, Dlugash J, Yuceoglu YZ, Kumral T, Branwood AW, Grishman A. New type of cardiomyopathy associated with diabetic glomerulosclerosis. Am J Cardiol. 1972;30:595–602.
Nishimura RA, Otto CM, Bonow RO, Carabello BA, Erwin JP 3rd, Guyton RA, O’Gara PT, Ruiz CE, Skubas NJ, Sorajja P, et al. 2014 AHA/ACC guideline for the management of patients with valvular heart disease: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2014;63:2438–88.
Mahadevan G, Davis RC, Frenneaux MP, Hobbs FD, Lip GY, Sanderson JE, Davies MK. Left ventricular ejection fraction: are the revised cut-off points for defining systolic dysfunction sufficiently evidence based? Heart. 2008;94:426–8.
Shin JH, Kang KW, Kim JH, Chin JY, Kim NY, Park SH, Kim WH, Choi YJ, Jung KT. Treadmill exercise-induced E/E′ elevation as a predictor of cardiovascular event in end-stage renal disease on peritoneal dialysis. Korean J Intern Med. 2016. doi:10.3904/kjim.2016.254.
Hillis GS, Moller JE, Pellikka PA, Gersh BJ, Wright RS, Ommen SR, Reeder GS, Oh JK. Noninvasive estimation of left ventricular filling pressure by E/E′ is a powerful predictor of survival after acute myocardial infarction. J Am Coll Cardiol. 2004;43:360–7.
Luers C, Edelmann F, Wachter R, Pieske B, Mende M, Angermann C, Ertl G, Dungen HD, Stork S. Prognostic impact of diastolic dysfunction in systolic heart failure-a cross-project analysis from the German Competence Network Heart Failure. Clin Cardiol. 2017. doi:10.1002/clc.22710.
Yamada H, Tanaka A, Kusunose K, Amano R, Matsuhisa M, Daida H, Ito M, Tsutsui H, Nanasato M, Kamiya H, et al. Effect of sitagliptin on the echocardiographic parameters of left ventricular diastolic function in patients with type 2 diabetes: a subgroup analysis of the PROLOGUE study. Cardiovasc Diabetol. 2017;16:63.
Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med. 1998;339:229–34.
Alonso N, Traveset A, Rubinat E, Ortega E, Alcubierre N, Sanahuja J, Hernandez M, Betriu A, Jurjo C, Fernandez E, et al. Type 2 diabetes-associated carotid plaque burden is increased in patients with retinopathy compared to those without retinopathy. Cardiovasc Diabetol. 2015;14:33.
Madonna R, Giovannelli G, Confalone P, Renna FV, Geng YJ, De Caterina R. High glucose-induced hyperosmolarity contributes to COX-2 expression and angiogenesis: implications for diabetic retinopathy. Cardiovasc Diabetol. 2016;15:18.
Tryniszewski W, Kusmierczyk J, Maziarz Z, Gos R, Mikhailidis DP, Banach M, Rysz J, Pesudovs K. Correlation of the severity of diabetic retinopathy and the heart muscle perfusion in patients with type 2 diabetes. J Diabetes Complicat. 2011;25:253–7.
Kramer CK, Rodrigues TC, Canani LH, Gross JL, Azevedo MJ. Diabetic retinopathy predicts all-cause mortality and cardiovascular events in both type 1 and 2 diabetes: meta-analysis of observational studies. Diabetes Care. 2011;34:1238–44.
Um T, Lee DH, Kang JW, Kim EY, Yoon YH. The degree of diabetic retinopathy in patients with type 2 diabetes correlates with the presence and severity of coronary heart disease. J Korean Med Sci. 2016;31:1292–9.
Tanaka A, Ishii H, Tatami Y, Shibata Y, Osugi N, Ota T, Okumura S, Suzuki S, Inoue Y, Murohara T. Impact of diabetic retinopathy on late cardiac events after percutaneous coronary intervention. J Cardiol. 2014;64:175–8.
Zhen Z, Chen Y, Shih K, Liu JH, Yuen M, Wong DS, Lam KS, Tse HF, Yiu KH. Altered myocardial response in patients with diabetic retinopathy: an exercise echocardiography study. Cardiovasc Diabetol. 2015;14:123.
Asghar O, Al-Sunni A, Khavandi K, Khavandi A, Withers S, Greenstein A, Heagerty AM, Malik RA. Diabetic cardiomyopathy. Clin Sci (Lond). 2009;116:741–60.
Seferovic PM, Paulus WJ. Clinical diabetic cardiomyopathy: a two-faced disease with restrictive and dilated phenotypes. Eur Heart J. 2015;36(1718–27):27a–27c.
Adameova A, Dhalla NS. Role of microangiopathy in diabetic cardiomyopathy. Heart Fail Rev. 2014;19:25–33.
Jia G, DeMarco VG, Sowers JR. Insulin resistance and hyperinsulinaemia in diabetic cardiomyopathy. Nat Rev Endocrinol. 2016;12:144–53.
Russo I, Frangogiannis NG. Diabetes-associated cardiac fibrosis: cellular effectors, molecular mechanisms and therapeutic opportunities. J Mol Cell Cardiol. 2016;90:84–93.
Boudina S, Abel ED. Diabetic cardiomyopathy, causes and effects. Rev Endocr Metab Disord. 2010;11:31–9.
Bell DS. Diabetic cardiomyopathy. Diabetes Care. 2003;26:2949–51.
Cheung N, Wang JJ, Rogers SL, Brancati F, Klein R, Sharrett AR, Wong TY. Diabetic retinopathy and risk of heart failure. J Am Coll Cardiol. 2008;51:1573–8.
Miettinen H, Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Retinopathy predicts coronary heart disease events in NIDDM patients. Diabetes Care. 1996;19:1445–8.
Wong CW, Wong TY, Cheng CY, Sabanayagam C. Kidney and eye diseases: common risk factors, etiological mechanisms, and pathways. Kidney Int. 2014;85:1290–302.
Abdel-Hamid AA, Firgany Ael D. Atorvastatin alleviates experimental diabetic cardiomyopathy by suppressing apoptosis and oxidative stress. J Mol Histol. 2015;46:337–45.
Hu W, Jiang WB. Pitavastatin-attenuated cardiac dysfunction in mice with dilated cardiomyopathy via regulation of myocardial calcium handling proteins. Acta Pharm. 2014;64:105–15.
Deo SV, Rababa’h A, Altarabsheh SE, Lim JY, Cho YH, Park SJ. Statin therapy improves long-term survival in non-ischaemic cardiomyopathy: a pooled analysis of 4500 patients. Heart Lung Circ. 2014;23:985–7.
Broch K, Askevold ET, Gjertsen E, Ueland T, Yndestad A, Godang K, Stueflotten W, Andreassen J, Svendsmark R, Smith HJ, et al. The effect of rosuvastatin on inflammation, matrix turnover and left ventricular remodeling in dilated cardiomyopathy: a randomized, controlled trial. PLoS ONE. 2014;9:e89732.
Chung YR, Park SW, Choi SY, Kim SW, Moon KY, Kim JH, Lee K. Association of statin use and hypertriglyceridemia with diabetic macular edema in patients with type 2 diabetes and diabetic retinopathy. Cardiovasc Diabetol. 2017;16:4.
YRC, SJP, and SWP wrote the manuscript and performed the research. KYM, SAC, and HSL performed the research. HSL, JHK, and KL contributed to discussion of the results and reviewed/edited the manuscript. All authors read and approved the final manuscript.
The English in this document has been checked by one or more native English-speaking editors at Nature Research Editing Service. For a certificate, please see: http://secure.authorservices.springernature.com/certificate/verify.
The authors declare that they have no competing interests.
Availability of data and materials
Not applicable. The conclusions of the manuscript are based on relevant data available in the manuscript.
Ethics approval and consent to participate
This retrospective study was approved by the Institutional Review Board of Ajou University Hospital (#AJIRB-MED-MDB-16-542) and complied with the Declaration of Helsinki.
This study was supported by the Pioneer Research Program of the National Research Foundation of Korea/Ministry of Education, Science and Technology (2012-0009544), and the Bio & Medical Technology Development Program of the National Research Foundation funded by the Korean government, MSIP (NRF-2015M3A9E6028949).
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.