The incidence of end-stage renal disease (ESRD) has increased worldwide in recent years, mainly because of the ageing population and the rising rates of diabetes and hypertension, increasing medical and social expenditure. According to the annual report of the US Renal Disease System in 2020, Taiwan had the second-highest incidence and the highest prevalence rate of ESRD globally. Reference Johansen, Chertow, Foley, Gilbertson, Herzog and Ishani1 Patients with ESRD have multiple comorbidities and worse overall survival than the general population. Reference Goodkin, Bragg-Gresham, Koenig, Wolfe, Akiba and Andreucci2 Cardiovascular disease (CVD) is the leading cause of death, followed by infections. Reference Al Wakeel, Mitwalli, Al Mohaya, Abu-Aisha, Tarif and Malik3,Reference Sarnak and Jaber4
Depression is the most common mental disorder in patients with chronic kidney disease (CKD) and ESRD. Reference Shirazian, Grant, Aina, Mattana, Khorassani and Ricardo5 Patients with CKD have an approximately three times higher risk for depression than those in the general population (5–9% in women; 2–3% in men). Reference Hedayati, Bosworth, Briley, Sloane, Pieper and Kimmel6,Reference Snow, Lascher and Mottur-Pilson7 Moreover, depression can result in a poor quality of life and many adverse medical outcomes in CKD patients, such as a faster decline in kidney function, early initiation of dialysis, hospital admissions and dialysis withdrawal. Reference Gerogianni, Kouzoupis and Grapsa8–Reference Kurella, Kimmel, Young and Chertow10 There are theories explaining the mechanism and pathways. For instance, through behavioural pathways, depression is associated with decreased motivation and increased fatigue, which can lead to poor adherence to medication regimens and dietary restriction. Non-adherence can exacerbate CKD progression and contribute to more frequent need for emergency care or hospital admission.
ESRD and depression significantly affect human health. It has been demonstrated that there is an independent relationship between depression and increased mortality in dialysis patients. Reference Farrokhi, Abedi, Beyene, Kurdyak and Jassal11–Reference Halen, Cukor, Constantiner and Kimmel12 A meta-analysis of individuals on dialysis, including 25 studies with 41 941 participants, reported a 1.66 times higher mortality risk with depressive symptoms. Reference Farrokhi, Abedi, Beyene, Kurdyak and Jassal11 Despite limitations in the quality and heterogeneity of the studies included in the analysis, a significant finding persisted even after multiple sensitivity analyses. In addition to the well-established causes of death in ESRD, several studies have discussed the probable interactions through which depression may affect survival in patients with ESRD, including chronic inflammation, the burden of physical illness, malnutrition, non-adherence to treatment and a higher risk of suicide. Reference Gerogianni, Kouzoupis and Grapsa8,Reference Farrokhi, Abedi, Beyene, Kurdyak and Jassal11–Reference Halen, Cukor, Constantiner and Kimmel12 However, few studies have comprehensively examined the factors contributing mortality in depressed patients with ESRD.
The strong association between depressed patients with ESRD and suicidal risks has been repeatedly emphasised. Reference Keskin and Engin13,Reference Chen, Tsai, Hsu, Wu, Sun and Chou14 However, most of the studies were cross-sectional in nature, making it difficult to demonstrate a definite causality. Suicide attempt remains a non-negligible public health threat among people with depression. Whether suicide attempts lead to subsequent suicide deaths, or even the relevant predictors, Reference Chen, Chou, Hsieh, Chang, Lee and Dewey15 has been the subject of various discussions in previous studies. Repetition of attempts, Reference Zahl and Hawton16 lethality of the suicide method Reference Bergen, Hawton, Waters, Ness, Cooper and Steeg17,Reference Finkelstein, Macdonald, Hollands, Sivilotti, Hutson and Mamdani18 and receipt of active aftercare were thought to play roles in completed suicide. Reference Hjorthøj, Madsen, Agerbo and Nordentoft19,Reference Pan, Chang, Lee, Chen, Liao and Caine20 However, to the best of our knowledge, none of the available data have been used to evaluate the impact of suicide attempts in the dialysis population with depression.
Therefore, we hypothesised that (a) the risk factors for mortality may be distinctive in patients with ESRD and depression; (b) these patients may have a heightened risk of suicide attempts; and (c) suicide attempts may be associated with mortality in dialysis patients. We used the Taiwan National Health Insurance Research Database (NHIRD) to conduct this longitudinal cohort study and thoroughly investigate the predictors of mortality and psychological long-term outcomes in incident dialysis patients with depression.
Method
Database
This study used Taiwan’s NHIRD. The NHIRD contains detailed healthcare information based on Taiwan’s National Health Insurance programme, which covers more than 99% of the Taiwanese population. All medical institutions that have contracts with the programme must submit standard computerised claims documents for medical services and medications used to treat their patients. Patients with ESRD are eligible for any type of renal replacement therapy free of charge, and all costs of treating patients undergoing chronic dialysis are covered by National Health Insurance. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2013. All procedures involving human participants/patients were approved by the Research Ethics Committee of Chi Mei Hospital (IRB No. 10409–E01).
Patient selection and definition
For this longitudinal cohort study, we selected incident adult (≥18 years old) ESRD patients on maintenance dialysis who began renal replacement therapy between 1 January 1997 and 31 December 2012. ESRD patients on maintenance dialysis were defined as having undergone dialysis for more than 90 days. We first examined the patients with baseline depression diagnosed before the initiation of dialysis. We established a study cohort that comprised two groups: a case group consisting of 3289 ESRD patients with baseline depression (depression group) and a reference group consisting of ESRD patients without depression (non-depression group). A propensity score was derived from age and gender for each ESRD patient. We employed the greedy matching algorithm to select matched reference members from ESRD patients without previous depression diagnoses. Reference Rassen, Shelat, Myers, Glynn, Rothman and Schneeweiss21 This process ensured that individuals with the closest score to the case member were chosen. The sample size of reference to the case group was 8:1, to maximise the statistical power.
To enhance the clinical relevance of both baseline depression severity and depression-related events during follow up, we refined our cohort by including only patients diagnosed with major depressive disorder (ICD-9 22 codes: 296.2x and 296.3x) in the subgroup analysis. The severity of depression was identified based on the final diagnostic code recorded before cohort entry. Patients were classified as having mild (ICD-9-CM: 296.21 or 296.31), moderate (ICD-9-CM: 296.22 or 296.32) or severe (ICD-9-CM: 296.23, 296.24, 296.33 or 296.34) depression. Reference Pilon, Sheehan, Szukis, Morrison, Zhdanava and Lefebvre23 Among patients with multiple severity codes, the last severity code was selected as the most relevant from a payer’s perspective. Patients with an unspecified severity status or without a severity status in the study period were excluded from the analysis. Depression-related events included emergency room visits and in-patient hospital admissions for depression.
Ascertaining the demographic and comorbid variables
We collected patients’ demographic data, survival status, date of death and baseline comorbidities relevant to the outcome in ESRD dialysis patients. Comorbid conditions are highly prevalent among ESRD patients, and comorbidity indexes have been widely used for describing comorbidity burden, predicting outcomes and adjusting as a confounder in ESRD patients’ analyses. However, most comorbidity indexes were developed for general populations or small patient cohorts. Jiannong Liu et al evaluated mortality analyses of ESRD dialysis patients based on the comorbid conditions used. Eleven comorbidities, namely diabetes mellitus, atherosclerotic heart disease (ASHD), congestive heart failure (CHF), cerebrovascular accident/transient ischaemic attack (CVA/TIA), peripheral vascular disease (PVD), chronic obstructive pulmonary disease (COPD), dysrhythmia, other cardiac diseases, gastrointestinal bleeding, liver disease and cancer, were defined as used by the US Renal Data System. The 11 comorbid conditions we adopted in this study were derived from an improved comorbidity index developed by Jiannong Liu et al in 2010, Reference Liu, Huang, Gilbertson, Foley and Collins24 having an excellent predictive capacity for mortality among dialysis patients. The method to identify comorbidities has been used extensively in various studies, and many articles have been published. Reference Tien, Lin, Chio, Wang, Chu and Sun25,Reference Kan, Wang, Wang, Sun, Hung and Chu26
For more accurate diagnoses of comorbidities, we linked to the diagnostic codes through the in-patient and out-patient claims databases of the National Health Insurance (NHI). Cases of baseline depression and those with comorbidities were identified according to one of the following definitions: (a) diagnostic codes in out-patient visits if the patient had an initial diagnosis at any time in the year leading up to the beginning of dialysis and then experienced one or more additional diagnoses within the subsequent 12 months, and the first and last out-patient visit within 1 year must be >30 days apart to avoid accidental inclusion of miscoded patients, or (b) diagnostic codes in hospital admission databases at least one time within the year leading up to start of dialysis. This method of identifying these comorbidities has been used extensively in various studies of the Taiwan NHIRD, and many articles have been published. Reference Tien, Lin, Chio, Wang, Chu and Sun25,Reference Kan, Wang, Wang, Sun, Hung and Chu26 ICD codes are provided in Supplementary Table 1 available at https://doi.org/10.1192/bjo.2025.10872.
Measurement
The outcome of this study included all-cause mortality and the cumulative incidence rate of suicide attempts after initiation of dialysis. Patients were followed up from the first reported date of dialysis initiation to the date of death, end of dialysis or 31 December 2013, the end of the follow-up period.
Although psychotic episodes were initially designated as a study outcome, this end-point was removed after expert review because of concerns about diagnostic ambiguity and potential misclassification with delirium in the context of ESRD.
Statistical analyses
Baseline characteristics between the group that had depression and that without depression were compared using Pearson’s χ 2 test. Age was entered as a categorical variable (18–44, 45–64 and ≥65 years). The significance level was set at P < 0.05. The cumulative proportion of survivors after the initiation of dialysis was calculated using the Kaplan–Meier method. The log-rank test was used to compare the difference of survival curves. Based on the matched cohort design of this study, we employed stratified Cox proportional hazard models to identify the risk factors of mortality and compare risks of mortality and suicide attempts after initiation of dialysis between the depression and non-depression groups, reporting hazard ratios and 95% confidence intervals. Because the age and gender were tightly matched between the depression and non-depression groups, we did not additionally adjust for age and gender in multivariate analyses. Unstratified Cox model was used in the subgroup analyses for depression and non-depression group, respectively. The data have been examined to meet the proportional hazard assumption for Cox model. Additionally, in subgroup analyses by depression severity, we reported the incidence rates of emergency room visits and the number of hospital admissions based on the Poisson distribution, and compared incidence rates using the Poisson regression model. The independent associations were examined using multivariate analysis. All data were analysed using SAS statistical software (SAS System for Windows, Version 9.4, SAS Institute Inc., Cary, NC, USA; see https://www.sas.com).
Results
Demographics and baseline characteristics by depression status
A total of 3289 ESRD patients with depression were enrolled in this study (Table 1). They were predominantly female (61.6%) and older (6.96% <45 years old; 38.01% 45–64 years old; 55.03% ≥65 years old). ESRD patients with depression were more likely to have comorbidities, including diabetes mellitus, coronary artery disease, CVD, dysrhythmia, PVD, chronic obstructive lung disease, gastrointestinal bleeding and cancer.
Table 1 Comparisons of baseline characteristics between groups with and without depression in patients receiving haemodialysis
a. Includes pericarditis, endocarditis, myocarditis, other complications of heart disease, heart transplant, heart valve replacement and cardiac devices.
Cumulative survival rate and risk factors for all-cause mortality after initiation of dialysis
The 1-, 3-, 5- and 7-year cumulative survival rates were as follows: 75.1%, 54.0%, 38.7% and 30.2% in ESRD patients with depression, and 79.2%, 61.4%, 48.6% and 38.3% in matched ESRD patients without depression, respectively (log-rank test: P < 0.001) (Fig. 1(a)). Patients with depression had a 15% higher mortality risk than those without depression (hazard ratio 1.15; 95% CI: 1.10–1.21) (Table 2). In addition, patients with cancer, CVD, diabetes mellitus, chronic liver disease, chronic obstructive lung disease, dysrhythmia, CHF and gastrointestinal bleeding had a significantly higher mortality rate.
Fig. 1 (a) Overall survival curves after initiation of dialysis stratified by patients with and without depression. (b) Cumulative incidence rate of suicide attempts.
Table 2 Risk factors for all-cause mortality after initiation of dialysis in end-stage renal disease dialysis patients by stratified Cox proportional hazard model
a. Including pericarditis, endocarditis, myocarditis, other complications of heart disease, heart transplant, heart valve replacement and cardiac devices.
*P < 0.05.
Risk factors for all-cause mortality after initiation of dialysis stratified by depression status
We further stratified the ESRD dialysis patients by depression (Table 3). Multivariate analysis revealed higher mortality for male and older individuals in both ESRD cohorts, regardless of the presence of depression. All 11 comorbid conditions were significantly associated with mortality in ESRD patients without depression. The depression cohort had a similar distribution of risk predictors for mortality, except for other cardiac diseases (hazard ratio 1.12; 95% CI: 0.98–1.29).
Table 3 Risk factors for all-cause mortality stratified by depression status in end-stage renal disease dialysis patients
a. Including pericarditis, endocarditis, myocarditis, other complications of heart disease, heart transplant, heart valve replacement and cardiac devices.
*P < 0.05.
Suicide attempts
The 1-, 3-, 5- and 7-year cumulative proportions of suicide attempts were 0.35%, 0.7%, 1.02% and 1.39% in the case cohort, and 0.08%, 0.24%, 0.37% and 0.45% in the matched controls, respectively (log-rank test: P < 0.001) (Fig. 1(b)). The depression cohort had a suicide attempt incidence of 2.19/1000 patient-years, and that of the non-depression group was 0.70/1000 patient-years (Table 4). The crude hazard ratio for suicide attempts in the depression cohort was 3.20 (hazard ratio 3.20, 95% CI: 1.88–5.43). Multivariate analysis revealed that the depression cohort had a 3.02-fold higher risk for suicide attempts than the non-depression cohort (hazard ratio 3.02, 95% CI: 1.68–5.42). We further examined the effects of suicide attempts on all-cause mortality in the entire dialysis cohort; suicide attempts were not associated with mortality risk after adjusting for depression and baseline covariates simultaneously (hazard ratio 1.10; 95% CI: 0.86–1.39).
Table 4 Associations of depression with incidence of all-cause death and suicide attempts
a. Per 1000 person-years.
Subgroup analyses for emergency room visits and in-patient visits for depression by depression severity status
Among the 413 dialysis patients with major depressive disorder, 244 (59.1%) had non-severe depression (mild or moderate) and 169 (40.9%) had severe depression. There were no significant differences between the two groups in gender, age or comorbidities (Supplementary Table 2). Severe depression was associated with a non-significant 39% higher rate of emergency room visits for depression compared to the non-severe depression group (incidence rate ratio (IRR) 1.39; 95% CI: 0.99–1.95). In addition, hospital admission rates for depression were significantly higher in the severe depression group (IRR 1.82; 95% CI: 1.14–2.93) (Supplementary Table 3). Regarding all-cause mortality, the severe depression group experienced a significantly higher mortality rate relative to those without depression (IRR 1.42; 95% CI: 1.15–1.76) (Supplementary Table 4).
Discussion
The current study demonstrated that depression was associated with increased all-cause mortality (hazard ratio 1.15; 95% CI: 1.10–1.21) in ESRD dialysis patients. Several characteristics, including older age, female gender and more chronic conditions, were identified in ESRD dialysis patients with depression. We further stratified the patients by depression status to explore the independent predictors of survival, which showed a similar risk distribution between the depressed and non-depressed groups. Moreover, this is the first study to evaluate the long-term psychological outcomes in this unique population. Our study revealed that depressed dialysis patients had a higher rate of suicide attempts (hazard ratio 3.02, 95% CI: 1.68–5.42) as compared to those without depression. However, suicide attempts were not associated with mortality risk in the whole dialysis cohort (hazard ratio 1.10; 95% CI: 0.86–1.39). Consequently, baseline comorbidities still played the most important role in the survival outcomes of ESRD patients with depression. Moreover, we found that severe depression was significantly associated with higher rates of hospital admission and mortality, suggesting the need for tailored interventions and careful monitoring based on depression severity in dialysis patients.
There are various risk factors for depression in the general population, ranging from sociodemographic to hereditary factors. Female gender and older age are the two sociodemographic factors most consistently supported by the evidence, Reference Cole and Dendukuri27,Reference Piccinelli and Wilkinson28 and underlying physical burdens affect the chronicity and severity of depression. Reference Egede29 Our study revealed that the dialysis patients with depression tended to be older, female and have more comorbidities, which portrayed a parallel phenomenon among community participants with depression.
Previous studies have revealed an independent relationship between depression and all-cause mortality in the dialysis population. Reference Farrokhi, Abedi, Beyene, Kurdyak and Jassal11,Reference Halen, Cukor, Constantiner and Kimmel12,Reference Chilcot, Davenport, Wellsted, Firth and Farrington30 Our study similarly demonstrated that dialysis patients with depression had increased death risk (hazard ratio 1.15; 95% CI: 1.10–1.21). Prior research has proposed that depression may worsen ESRD outcomes through a range of mechanisms, including biological pathways such as immune dysregulation, Reference Kiecolt-Glaser and Glaser31 malnutrition and exacerbation of somatic symptoms, Reference Tsai, Chiu, Hung, Hwang, Tsai and Wang32 as well as behavioural and social mechanisms such as lack of adherence to dialysis and medications, Reference Cukor, Rosenthal, Jindal, Brown and Kimmel33 social withdrawal and elevated suicide risk. Reference Martiny, de, Silva, Neto and Nardi34 All of these pathways could aggravate the general health decline of ESRD patients. Furthermore, depression is associated with alterations in the hypothalamic–pituitary–adrenal axis and increased inflammatory markers, which could contribute to cardiovascular risk and overall mortality. Reference Saglimbene, Palmer, Scardapane, Craig, Ruospo and Natale9
In several studies, the association between depression and mortality in ESRD patients on dialysis was attenuated after multivariate adjustment, Reference Fan, Sarnak, Tighiouart, Drew, Kantor and Lou35 reflecting the complexity of the contributing mediators. Our study showed similar trends. The hazard ratio of mortality risk reduced from 1.27-fold to 1.15-fold in patients with depression after the multivariate analysis. We further stratified these patients by depression to evaluation. We adopted 11 significant mortality predictors of ESRD patients on dialysis, including diabetes mellitus, CVD and cancer, as covariates. Reference Liu, Huang, Gilbertson, Foley and Collins24 The results showed no notable difference between the two groups, implying that the underlying medical condition was the most important mortality factor for ESRD patients on dialysis, regardless of depression status.
Depression is a robust indicator of suicide risk in the dialysis population. Reference Keskin and Engin13,Reference Chen, Tsai, Hsu, Wu, Sun and Chou14,Reference Fan, Sarnak, Tighiouart, Drew, Kantor and Lou35,Reference Patel, Sachan and Nischal36 Suicidal ideation increased with depression severity in ESRD patients on haemodialysis. Reference Keskin and Engin13 Nevertheless, most of these studies were cross-sectional studies using diverse questionnaires and scales for depression surveys. To address the association between depression and suicide risk in ESRD patients, we analysed suicide attempts in our longitudinal ESRD cohort study. We found that the cumulative rate of suicide attempts increased remarkably and was nearly three times higher in the depression group. However, increased suicide attempts were not associated with mortality in ESRD dialysis patients (hazard ratio 1.10; 95% CI: 0.86–1.39). A national registry study on the predictors of suicidal death among multiple attempters in Taiwan reported that only 1.5% (n = 861/55 560) of attempters eventually died by suicide. Reference Chen, Liao, Lee, Wu, Lin and Chen37 Similarly, in our dialysis cohort, the lack of a statistically significant association between suicide attempts and mortality suggested that the majority of attempts may not have been medically lethal. This observation could be because of several factors. Patients who attempt suicide receive immediate psychological and intensive medical attention, which reduces the short-term risk of mortality. In Taiwan, the aftercare for suicide attempts is active and timely. In addition, dialysis patients need to visit healthcare facilities thrice a week, which ensures they receive a great deal of medical attention and may prevent them from completing suicide. Lastly, prevalent multiple comorbidities of ESRD patients, rather than suicide attempts, were probably the most significant risk predictors for mortality in this vulnerable population.
The significantly higher incidence of suicide attempts among ESRD patients with depression underscored the need for integrated psychiatric care in this population. Routine psychological evaluations and timely interventions could be crucial in managing and improving these patients’ quality of life and treatment adherence. Our findings could inform health policies to implement standardised screening procedures for depression in dialysis centres.
Our study’s findings aligned with and expanded upon a 2023 Korean study, which identified a link between depression severity and mortality in ESRD patients. Reference Jeon, Lim, Jeon, Chung, Kim and Kang38 Both studies found that severe depression was a significant predictor of mortality. However, key differences in methodology and demographics distinguished our research. While we used ICD-9-CM codes for diagnosis and classification, the Korean study employed Beck Depression Inventory-II scores. Our study included a broader adult population, leading to a younger mean age, whereas the Korean study focused on older patients. This indicates that severe depression negatively affects survival across the entire adult dialysis population, not just geriatric patients. In addition, our study revealed that severe depression was associated with a higher risk of hospital admission compared to non-severe depression, a finding not assessed in the Korean study. Generally, while methodologies differ, our results corroborated the Korean study and highlighted depression severity as a key risk factor for mortality and hospital admission in adult ESRD patients.
Our study has several strengths. First, it is important to note that our study had both national representativeness and a large sample size, as we used the NHIRD system to ascertain complete comorbidities and outcomes. Second, we applied rigorous algorithms to identify cases, comorbidities and outcomes. Third, this study was the first to delineate mortality predictors and evaluate psychological outcomes in ESRD dialysis patients with depression.
This study had several limitations. Its retrospective and observational nature made it difficult to draw solid conclusions regarding causality. The lack of a gold standard for diagnosing depression and reliance on claims data (ICD-9-CM) with physician-diagnosed depression may result in misclassification bias and underestimation of depression prevalence and observed associations. In addition, potential confounding factors, such as episodic depression preceding dialysis, could influence mental health outcomes. Residual confounding variables, including marital status, educational level, underlying medications, socioeconomic status, lifestyle factors, working conditions, substance misuse and family history of depression, were not controlled. Similarly, like other studies using administrative data, we did not account for some unmeasured confounders, whether based on latent confounding or biological factors. Comorbidities were derived from claim data and ICD-9-CM diagnosis codes, potentially leading to disease misclassification and an inability to account for disease severity, which reduced our ability to show severity-related effects of comorbidities. Psychotic episodes, originally prespecified as an outcome, were excluded after expert discussion because of diagnostic ambiguity and possible misclassification with delirium in ESRD. Furthermore, we lacked specific data on dialysis adequacy, patient compliance, nutritional status, biochemical data and socioeconomic characteristics. Therefore, our results should be interpreted with caution, as residual confounding is present in all observational studies, limiting us to showing associations rather than causality between risk factors and mortality.
In conclusion, depression is associated with increased mortality risk in ESRD patients on dialysis. The relationship between depression and ESRD appears to be complex and multifactorial. The strongest predictors of death in dialysis patients with depression were underlying comorbidities, such as diabetes mellitus, CVD and cancer, similar to those without depression. Our results reemphasised the importance of baseline chronic illness in this unique population, for whom reasonable medical control may alleviate suffering and support overall well-being. In addition, ESRD patients with depression are more likely to experience suicide attempts, emergency room visits and hospital admissions, particularly in patients with severe depression. Therefore, more attention should be paid to awareness when treating these high-risk patients. Given the inherent limitations of retrospective designs, further prospective studies or randomised controlled trials are warranted to explore the intricate relationship among depression, psychological outcomes and mortality in dialysis patients.
Supplementary material
The supplementary material is available online at https://doi.org/10.1192/bjo.2025.10872
Data availability
The data that support the findings of this study are available on request from the corresponding author, C.-C.C. The data are not publicly available because of the policies of the National Health Insurance Research Database, which restrict the distribution of raw data to protect participants’ privacy.
Author contributions
All authors contributed to this study. I.-N.Y., C.-L.L. and C.-C.C. contributed to the conception and design of the study. I.-N.Y. and C.-C.C. drafted the manuscript. I.-N.Y., C.-L.L. and C.-C.C. contributed to the acquisition, analysis and interpretation of data. C.-L.L. modified the model structure. I.-N.Y., C.-L.L., J.-J.W., M.-C.H. and C.-C.C. revised the manuscript. All authors gave final approval for this manuscript and are accountable for the study’s integrity and accuracy.
Funding
This study was supported by grant 111CM-TMU-06 from Chi-Mei Hospital.
Declaration of interest
None.
Ethical standards
This study was conducted in compliance with the Declaration of Helsinki and was approved by the Research Ethics Committee of Chi Mei Hospital (IRB No. 10409–E01).
Open access
eLetters
No eLetters have been published for this article.