SEARCH

SEARCH BY CITATION

Keywords:

  • dementia;
  • National Health Insurance Research Database;
  • statin

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of conflicts of interest
  10. References

Background and purpose

Dementia is a neurodegenerative disorder that presents a progressive decline in cognitive function and loss of short-term memory with age. Several studies have shown that statin, an oral lipid-lowering drug, may reduce the risk of developing dementia. The objective of this study is to explore the association between statin and the development of dementia.

Methods

The data analyzed in this study were retrieved from the National Health Insurance Research Database in Taiwan. The sample consisted of 123 300 patients ≥ 20 years of age, including 61 650 dementia patients with statin use and 61 650 patients without statin use who were eligible for inclusion in this study. Univariate and multivariate Cox proportional hazard regression analyses were performed to measure the effects of statin use on the risk of dementia.

Results

The beneficial effect of statin on dementia was significant after adjusting for sociodemographic factors and comorbidities (adjusted hazard ratio of 0.92, 95% confidence interval 0.86–0.98). The sex- and age-specific analysis of adjusted hazard ratios showed a higher beneficial effect from statin treatment in women than in men, and the effect became more significant with age.

Conclusion

Statin therapy may help prevent the development of dementia, and both hydrophilic and lipophilic statins produce similar effects. However, the preventive characters and associated mechanisms must be further explored and identified.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of conflicts of interest
  10. References

Dementia, which is commonly caused by Alzheimer's disease (AD) and vascular disorders, is a neurodegenerative disorder characterized by progressive cognitive and functional decline. Initial symptoms include memory loss and progressive inability to perform the basic activities of daily life [1]. Dementia is one of the major causes of disability and dependence amongst older people worldwide. As the worldwide life expectancy increases, dementia is becoming a considerable burden on families and society in general. The number of people with dementia in 2010 was nearly 36 million worldwide, and this number is expected to increase to 66 million by 2030 and 115 million by 2050 [2].

Epidemiological studies have shown that the risk factors for atherosclerosis and stroke, including diabetes mellitus, hypertension and hyperlipidemia, are also risk factors for dementia. Epidemiological research has shown a correlation between increased cholesterol levels and the incidence of AD [3, 4]. The histopathological examination of AD brain slices shows β-amyloid peptide deposits and neurofibrillary tangles, both of which are hallmarks of AD. High cholesterol diets increase β-amyloid deposition in mice [5]. In the brain, cholesterol is synthesized through a multi-enzyme cascade that includes 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) synthase and HMG CoA reductase. Statins are HMG-CoA reductase inhibitors that reduce low density lipoprotein (LDL) cholesterol levels and also reduce cardiovascular ischaemic events. Lowering cholesterol levels in patients using statins has been associated with decreased β-amyloid levels in the cerebral spinal fluid of patients [6]. Because of the potential multifactorial actions of statins, statin therapy may reduce the risk of dementia and delay the onset of cognitive decline.

The potential association between statin use and dementia was first reported in a prevalence study by Wolozin et al. [7] and a nested case–control study by Jick et al. [8]. However, relatively little information is available on these associations for non-Caucasian ethnicities, and in particular in the Asia-Pacific region where dementia prevalence is increasing [9, 10]. Representative National Health Insurance data sets from Taiwan were used to form a cohort to assess the potential for statins to prevent dementia.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of conflicts of interest
  10. References

Data sources

The National Health Insurance program in Taiwan integrated all public insurance systems into a single-pay program in 1995, and covered 99% of Taiwan's 23.74 million residents by 2009. The National Health Research Institute cooperated with the Bureau of National Health Insurance to establish several data sets for public research. The claims data set of the Longitudinal Health Insurance Database was used, which consists of 1 million people randomly selected from all those insured, with claims data abstracted from 1996 to 2010. The National Health Research Institute reported no substantial differences in age and gender between the Longitudinal Health Insurance Database and all insurants. Information on these databases has been presented in previous papers [11, 12]. Other studies have also demonstrated the accuracy and high validity of diagnoses in the National Health Insurance Research Database (NHIRD) [13, 14]. The International Classification of Disease, Ninth Revision (ICD-9), was used for diagnosis.

Study patients

Based on claims data for ambulatory case visits or admission records from 1998 to 2010, patients diagnosed with hyperlipidemia (ICD-9-CM code 272) who were 20 years of age or older were identified. Patients with statin treatment were selected for the study cohort. The date of diagnosis was defined as the index date for initiating follow-up person-years measurement. The comparison cohort consisted of patients randomly selected from hyperlipidemia patients without a statin treatment history. For each statin treatment case, one comparison person frequency matching with sex, age (every 5-year span) and index year was randomly selected. Patients with any history of dementia before the index date were excluded.

Outcome measures

The incidence rates of dementia (ICD-9-CM codes 290, 294.1 and 331.0) or myalgia and myositis (ICD-9-CM code 729.1) in these two cohorts were calculated from the follow-up period until the end of 2010. The person-years of follow-up were calculated for each person until outcome was diagnosed or the person was censored for death, migration or discontinued enrollment in the insurance database. The baseline comorbidity history was determined for each subject. Comorbidities included diabetes (ICD-9-CM code 250), hypertension (ICD-9-CM codes 401−405), stroke (ICD-9-CM codes 430−438), coronary artery disease (CAD) (ICD-9-CM codes 410−414), head injury (ICD-9-CM codes 850−854, 959.01), depression (ICD-9-CM codes 296.2, 296.3, 300.4, 311) and schizophrenia (ICD-9-CM codes 295, 296.4, 298.9 and 301.2).

Statistical analysis

The distributions of age (<45, 45–60, 60–75 and ≥75 years), sex, income (less than New Taiwan Dollar 15 000, 15 000–19 999 and more than 20 000 per month) and comorbidities including diabetes, hypertension, CAD, head injury and depression between hyperlipidemia patients with statin treatment and those without statin treatment were first compared. Differences using the chi-squared test for categorical variables and the t test for continuous variables were examined. The incidence rates of dementia for the two cohorts were estimated based on demographic status. Univariate and multivariate Cox proportional hazard regression analyses were performed to measure the effects of statin use on the risk of developing dementia. The multivariate models were simultaneously adjusted for age, sex, monthly income, and comorbidities of diabetes, hypertension, stroke, CAD, head injury, depression and schizophrenia. The crude and adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using the Cox model. The dose−response of statin use on the risk of developing dementia based on cumulative days of statin treatment was also estimated. In addition, the Cox model was used to estimate the HRs of dementia associated with cumulative days of statin treatment compared with those without statin treatment. All analyses were performed using sas statistical package (version 9.1; SAS Institute Inc., Cary, NC, USA). A two-tailed P value of <0.05 was considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of conflicts of interest
  10. References

From patients with hyperlipidemia, 61 650 patients with statin treatment were selected as the study cohort and 61 650 patients without statin treatment as the comparison cohort. The two cohorts had mean ages of 54.6 (SD = 12.4) years and 54.1 (SD = 12.8) years, respectively (Table 1). The distributions of age and sex for study patients were similar between the two cohorts. Compared with subjects without statin treatment, the statin cohort was more likely to have diabetes (23.8% vs. 43.1%, < 0.0001), hypertension (49.9% vs. 67.5%, P < 0.0001), stroke (14.8% vs. 20.7%, < 0.0001), CAD (23.8% vs. 35.0%, < 0.0001) and depression (10.2% vs. 10.7%, = 0.005).

Table 1. Demographic characteristics of study subjects among medicine status
VariableStatinP value
No N = 61 650Yes N = 61 650
n % n %
  1. NTD, New Taiwan dollar. Chi-squared test. at test.

Sex
Women31 35450.931 35450.90.99
Men30 29649.130 29649.1
Age, years
<4513 67922.213 67922.20.99
45–6027 75245.027 75245.0
60–7516 96727.516 96727.5
≥7532525.2732525.27
Mean (SD)a54.112.854.612.4<0.0001
Monthly income (NTD)
<15 00011 73419.012 23919.90.0009
15 000–19 99930 03248.729 58447.8
≥20 00019 88432.319 82732.2
Medical history
Diabetes14 67923.835 09943.1<0.0001
Hypertension30 73749.941 58767.5<0.0001
Stroke913114.812 73220.7<0.0001
CAD14 69523.821 60335.0<0.0001
Head injury41346.7141866.790.56
Depression630310.2660210.70.005
Schizophrenia6171.006451.050.43

Comparison of the incidence and hazard ratio of dementia development between the statin and non-statin groups

Overall, the statin cohort had a higher incidence rate of dementia than the non-statin cohort, but the difference was not significant (4.68 vs. 4.40 per 1000 person-years, crude HR 1.06, 95% CI 1.00–1.13; Table 2). However, the beneficial effect of statin use was significant after adjusting for sociodemographic factors and comorbidities (adjusted HR 0.92, 95% CI 0.8–0.98). Table 3 shows the incidence rate and risk of dementia of the statin cohort and non-statin cohort. Women had a higher incidence rate of dementia than men did, and this difference increased with age in both cohorts. Sex- and age-specific analysis of adjusted HRs showed a higher beneficial effect from statin treatment in women than in men (adjusted HR 0.91, 95% CI 0.83–0.99 vs. adjusted HR 0.94, 95% CI 0.85–1.04), and the effect was more significant with increased age (<45 years of age, adjusted HR 0.94, 95% CI 0.49–1.80; 45–60 years of age, adjusted HR 1.19, 95% CI 0.99–1.43; 60–75 years of age, adjusted HR 0.92, 95% CI 0.84–0.99; ≥75 years of age, adjusted HR 0.79, 95% CI 0.69–0.90). In the interaction analysis, age significantly modified the association between statin treatment and dementia (P values for interaction < 0.0001). Stratified by monthly income, the incidence rate of dementia decreased with monthly income and was highest for a monthly income <15 000 (8.81 vs. 9.45 per 1000 person-years). Monthly income-specific analysis shows that patients with statin treatment, compared with non-statin treatment, exhibited the highest beneficial effect in the monthly income category <15 000 (adjusted HR 0.81, 95% CI 0.73–0.90). The effects of cumulative months of individual statin treatment on dementia showed a dose−response relationship, and the adjusted HR of dementia showed the beneficial effects of simvastatin (adjusted HR 0.97, 95% CI 0.95–0.99), fluvastatin (adjusted HR 0.97, 95% CI 0.94–1.00), atorvastatin (adjusted HR 0.98, 95% CI 0.96–0.99) and rosuvastatin (adjusted HR 0.95, 95% CI 0.92–0.98) (Table 4).

Table 2. Incidence rate and hazard ratios of dementia according to medication status for patients with hyperlipidemia
 Statin
NoYes
  1. Models included the following variables: Model 1, medication status; Model 2, model 1 variable plus age, sex, monthly income, comorbidities of diabetes, hypertension, stroke, CAD, head injury, depression and schizophrenia. *P < 0.05.

No. of events18211966
Person-years413 938420 476
Incidence rate/1000 person-years4.404.68
Hazard ratio (95% confidence interval)
Model 11.00 (reference)1.06 (1.00, 1.13)
Model 21.00 (reference)0.92 (0.86, 0.98)*
Table 3. Incidence rate and hazard ratios of dementia according to medication status for patients with hyperlipidemia by sex, age and monthly income
VariableNon-statinStatinModel 1Model 2
CasePerson-yearsIRCasePerson-yearsIRCrude HR (95% CI)Adjusted HR (95% CI)
  1. IR, incidence density rate per 1000 person-years; NTD, New Taiwan dollar. Models included the following variables: Model 1, medication status; Model 2, model 1 variable plus age, sex, monthly income, comorbidities of diabetes, hypertension, stroke, CAD, head injury, depression and schizophrenia. *P < 0.05, ***P < 0.001.

Sex
Female1064214 7144.961130218 1165.181.05 (0.96, 1.14)0.91 (0.83, 0.99)*
Male757199 2233.80836202 3604.131.08 (0.99, 1.20)0.94 (0.85, 1.04)
Age, years
<452098 5900.202398 3750.231.15 (0.63, 2.10)0.94 (0.49, 1.80)
45–60221188 8231.17329189 2351.741.49 (1.25, 1.76)1.19 (0.99, 1.43)
60–751085111 8239.701172116 26710.081.04 (0.96, 1.13)0.92 (0.84, 0.99)*
≥7549514 70133.6744216 59926.630.79 (0.70, 0.90)***0.79 (0.69, 0.90)***
P for interaction       < 00001
Monthly income (NTD)
<15 00072176 3099.4572782 5528.810.93 (0.84, 1.03)0.81 (0.73, 0.90)***
15 000–19 999887203 1714.37928200 1624.641.06 (0.97, 1.16)0.94 (0.85, 1.03)
≥20 000213134 4581.58311137 7622.261.43 (1.20, 1.70)***1.15 (0.96, 1.38)
Table 4. Risk of dementia and associated with cumulative exposure of using individual statin
 Model 1Model 2
Crude HR (95% CI)Adjusted HR (95% CI)
  1. Models included the following variables: Model 1, medication status; Model 2, model 1 variable plus age, sex, income, comorbidities of diabetes, hypertension, stroke, CAD, head injury, depression and schizophrenia. *P < 0.05, **P < 0.01, ***P < 0.001.

Individual statin cumulative exposure (months)
Non1 (Reference)1 (Reference)
Simvastatin0.97 (0.96, 0.99)*0.97 (0.95, 0.99)**
Fluvastatin0.99 (0.96, 1.02)0.97 (0.94, 1.00)*
Lovastatin1.04 (1.02, 1.05)***1.02 (1.00, 1.03)
Atorvastatin0.99 (0.97, 0.99)*0.98 (0.96, 0.99)**
Pravastatin1.03 (1.00, 1.05)*1.02 (0.99, 1.04)
Rosuvastatin0.94 (0.91, 0.97)***0.95 (0.92, 0.98)***

Comparison of the incidence and hazard ratio of myalgia and myositis development between the statin and non-statin groups

Table 5 shows the incidence rate and hazard ratios of myalgia and myositis for statins use. The use of statins is associated with increased risk of myalgia and myositis (adjusted HR 1.05, 95% CI 1.02–1.08). Compared with men in the non-statin cohort, men with statin treatment had a higher relative risk of myalgia and myositis (adjusted HR 1.07, 95% CI 1.02–1.13). The highest age-specific myalgia and myositis incidence rate for patients with statin treatment was in the 60–75 years age group (26.6 per 1000 person-years). The monthly income-specific analyses indicated that patients with statin treatment exhibited the highest incidence rate in the monthly income category of 15 000–19 999 (26.0 per 1000 person-years).

Table 5. Incidence rate and hazard ratios of myalgia and myositis according to medication status for patients with hyperlipidemia by sex, age and monthly income
VariablesNon-statinStatinModel 1Model 2
CasePerson-yearsIRCasePerson-yearsIRCrude HR (95% CI)Adjusted HR (95% CI)
  1. IR, incidence density rate per 1000 person-years; NTD, New Taiwan dollar. Models included the following variables: Model 1, medication status; Model 2, model 1 variable plus age, sex, monthly income, comorbidities of diabetes, hypertension, stroke, CAD, head injury, depression and schizophrenia. *< 0.05, **< 0.01, ***P < 0.001.

All7941371 77321.48968371 33924.21.13 (1.10, 1.17)***1.05 (1.02, 1.08)**
Sex
Female4697188 88524.95217188 35027.71.11 (1.07, 1.16)***1.03 (0.99, 1.08)
Male3244182 88817.73751182 98920.51.16 (1.10, 1.21)***1.07 (1.02, 1.13)**
Age, years
<45160888 72618.1191486 34922.21.22 (1.14, 1.31)***1.09 (1.02, 1.17)*
45–603652167 12221.93996165 22724.21.11 (1.06, 1.16)***1.01 (0.97, 1.06)
60–752399101 29023.72747103 48326.61.12 (1.06, 1.18)***1.07 (1.01, 1.13)*
≥7528214 63519.331116 28019.10.99 (0.84, 1.16)1.01 (0.86, 1.19)
Monthly income (NTD)
<15 000148770 09921.2181673 87524.61.16 (1.08, 1.24)***1.09 (1.02, 1.17)*
15 000–19 9994216180 19423.44545174 88326.01.11 (1.07, 1.16)***1.02 (0.98, 1.07)
≥20 0002238121 47918.42607122 58121.31.15 (1.09, 1.22)***1.07 (1.01, 1.14)*

The interaction of comorbidity and statin treatment on the risks of dementia development

Table 6 illustrates the interactions of comorbidities and statin treatment on outcome. A significant interaction was observed for patients with both statin treatment and hypertension (interaction P = 0.0002) and diabetes (interaction < 0.0001). It suggested that statins reduced the risk of dementia in hyperlipidemia patients, coexisting with hypertension or diabetes.

Table 6. Interactions between statin treatment and comorbidity (Hypertension, Diabetes, Schizophrenia) on the risk of dementia
Variables N Event, nAdjusted HR (95% CI)P valuea
  1. Adjusted HR: adjusted for age and sex. aP value for interaction. *< 0.05; ***P < 0.001.

Statin treatmentHypertension    
NoNo39 0133671 (Reference)0.0002
NoYes30 73714541.35 (1.20, 1.51)***
YesNo20 0632541.19 (1.01, 1.39)*
YesYes41 58717121.30 (1.16, 1.45)***
Statin treatmentDiabetes    
NoNo46 97111061 (Reference)<0.0001
NoYes14 6797151.32 (1.20, 1.45)***
YesNo35 0999401.03 (0.94, 1.12)
YesYes26 55110261.23 (1.13, 1.34)***
Statin treatmentSchizophrenia    
NoNo61 03317901 (Reference)0.08
NoYes617313.23 (2.26, 4.60)***
YesNo61 00519451.02 (0.96, 1.09)
YesYes645213.09 (2.01, 4.75)***

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of conflicts of interest
  10. References

The results of this cohort study show that the use of statins is associated with a decreased risk of dementia after adjusting for age, sex and comorbidities. These results show that there is a lower prevalence of dementia in patients prescribed with various HMG-CoA reductase inhibitors, including hydrophilic and lipophilic statins. The association between statin usage and dementia is stronger in women, older adults and low-income patients.

A study of 110 patients with brain autopsies showed an association between statin therapy and a reduced burden of neurofibrillary tangles [15]. The relationship between cholesterol homeostasis within and outside the blood–brain barrier is not clearly understood. Whether the effect of statins on brain cholesterol turnover occurs through local inhibition in the brain or is mediated by lowering cholesterol in the circulation remains unclear. An animal study involving guinea pigs treated with high doses of hydrophilic and lipophilic statins, pravastatin and simvastatin, showed that brain cholesterol synthesis can be influenced by high doses of lipophilic and hydrophilic statins [16]. The different statins used in this study had similar effects in preventing dementia whether they were hydrophilic or lipophilic statins.

There are few population-based studies on the epidemiology of young-onset dementia. A meta-analysis study [17] estimated that, in London boroughs in the UK with a total population of 567 500 people, the prevalence of dementia with onset between the ages of 30 and 65 years was 54 per 100 000 (95% CI 45–64), and the prevalence was 98 per 100 000 (95% CI 81–118) between the ages of 45 and 65 years. The most common single diagnosis was AD, followed by vascular disease and frontotemporal lobar degeneration [17]. Using a two-step postal survey with a population of 2 966 000 people, Ikejima et al. found a similar prevalence of 42.3 per 100 000 (95% CI 39–45) between the ages of 18–65 years in the Ibaraki prefecture in Japan. Vascular disease was the most common causes followed by AD in this Japanese study [18]. Kelley et al. analyzed a total of 235 patients with progressive cognitive decline who were younger than 45 years and found that patients with an onset age of less than 35 years mainly had metabolic causes. By contrast, in patents having an onset age of between 35 and 45 years, most causes of dementias were degenerative diseases [19].

In this study, the age-specific analysis shows that the older patients (>60 years of age) with statin treatment showed a much more beneficial effect on dementia than younger groups. In the interaction analysis, age significantly modified the association between statin treatment and dementia (P values for interaction <0.0001; Table 3). Patients with young-onset dementia differ from older patients by a greater number of rare causes (29%) [20], unusual features of dementia, such as non-amnestic phenotypes of AD, a high frequency of frontal symptoms, and possible genetic origin [21, 22]. This might partially explain the differing effects of statin on dementia in young and old patients.

A meta-analyses study reported that there are few associations between serum lipid levels and the risk of developing dementia [23]. High total cholesterol in midlife could predict AD. Nevertheless, late-life total cholesterol in relation to AD, vascular dementia and any dementia did not reveal any significant associations. This suggested the possibility that statins have effects on dementia beyond cholesterol reduction.

Two randomized controlled clinical trials of statins for the treatment of AD patients have failed to provide consistent results in cognitive measurement. Simons et al. treated 44 AD patients with 40 mg of simvastatin or a placebo for 26 weeks. Serum LDL cholesterol decreased by 52% in the simvastatin group, but simvastatin did not significantly alter the cerebrospinal fluid levels of β-amyloid. Post hoc analysis revealed that simvastatin reduced β-amyloid levels in the cerebrospinal fluid of patients with mild AD, but did not change β-amyloid levels in more severely affected patients [6]. Sparks et al. treated AD patients with 80 mg of atorvastatin or a placebo. They found evidence that patients taking statins achieved higher cognitive function if their serum cholesterol levels were high at the baseline and their Mini-Mental State Examination results were higher at the baseline or if they had an apolipoprotein E4 allele present. The benefit of statin usage appeared in mild to moderately affected AD patients in early treatment [24]. These findings suggest that statin use might delay the pathogenic neurodegenerative process early in the course of dementia but may not reverse neuronal damage at later stages.

The prevalence of stroke and dementia increase with increasing age. Epidemiological studies have shown that risk factors for atherosclerosis are associated with increased risk of dementia [25]. Hyperlipidemia is a major risk factor for cerebral atherosclerosis. There have been substantial data supporting the use of statins in the prevention of atherosclerosis and stroke [17]. These suggested that the role of statins for vascular protection may be associated with vascular dementia.

Several case reports and case series have described a potential association between statins and cognitive impairment [26-30]. Symptoms were noted as early as a few days after initiation of statin therapy. These effects could be reversible after medication discontinuation. In all cases the statins were discontinued and patients experienced full recovery of cognition, with recovery times varying from a few days to 1 month after drug discontinuation. Statins are commonly used, especially in older patients who are at higher risk for cognitive disorder. Physicians should therefore be able to properly assess potential adverse cognitive effects of statins.

The study was subject to some limitations, which must be mentioned. First, the NHIRD does not provide detailed information on patients such as their smoking habits, alcohol consumption, body mass index, physical activity and family history of dementia, all of which may be related with the risk of dementia. Secondly, all data in the NHIRD are anonymous. Thus, relevant clinical variables such as blood pressure, blood sugar, imaging results, lipid level and information about hospitalization and mortality were unavailable for analysis. However, the data regarding statin prescriptions and dementia diagnosis were highly reliable.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of conflicts of interest
  10. References

In conclusion, statin therapy may help to avoid the development of dementia, and both hydrophilic and lipophilic statins provide a similar effect. However, the associated mechanisms must be explored and identified. Further randomized controlled trials can be used to eliminate indication bias and provide a more definitive finding regarding the effect of statins in protecting against dementia.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of conflicts of interest
  10. References

This work was supported by study projects from China Medical University Hospital (DMR-102-014 and DMR-102-023), Taiwan Ministry of Health and Welfare Clinical Trial and Research Center for Excellence (DOH102-TD-B-111-004), Taiwan Ministry of Health and Welfare Cancer Research Center for Excellence (MOHW103-TD-B-111-03) and International Research-Intensive Centers of Excellence in Taiwan (I-RiCE) (NSC101-2911-I-002-303).

Disclosure of conflicts of interest

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of conflicts of interest
  10. References

The authors declare no financial or other conflicts of interest

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Disclosure of conflicts of interest
  10. References