Article Text

Download PDFPDF
Original research article
Plasma 7-ketocholesterol levels and the risk of incident cardiovascular events
  1. Min Wang1,
  2. Weiqing Long2,
  3. Di Li3,
  4. Duan Wang4,
  5. Yuan Zhong3,
  6. Di Mu3,
  7. Jiayi Song3,
  8. Min Xia3
  1. 1 Department of Clinical Nutrition, The Affiliated Hospital of Southwest Medical University, Luzhou, China
  2. 2 Department of Clinical Laboratory, The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China
  3. 3 Department of Nutrition, School of Public Health, Sun Yat-sen University, Guangzhou, China
  4. 4 Department of Clinical Nutrition, Guangdong General Hospital, Guangzhou, China
  1. Correspondence to Dr Min Xia, Department of Nutrition, School of Public Health, Sun Yat-sen University, 74 Zhongshan Road 2, Guangzhou, Guangdong Province 510080, China; xiamin{at}mail.sysu.edu.cn

Abstract

Objective 7-Ketocholesterol (7-KC), a major oxidation product of cholesterol, is found in human atherosclerotic plaque and more atherogenic than cholesterol in animal models. This study was designed to investigate the association of plasma 7-KC level with the incident cardiovascular disease (CVD) events in general population.

Methods We measured plasma 7-KC concentrations at baseline in 1944 participants free from CVD in a community-based cohort study. The primary endpoint was incident of a major adverse cardiovascular event. A Cox proportional hazards model was used to calculate the HRs with 95% CI.

Results A total of 101 incident CVD events were recorded during the 5.2 year median follow-up. The baseline plasma 7-KC levels were associated with a higher risk of incident CVD events; compared with quartile 1, participants in quartile 4 had an unadjusted HR of 2.38 (2.03–2.85, p<0.001) and an adjusted HR of 1.70 (1.45–1.91, p=0.004) after adjusting for traditional risk factors. Plasma 7-KC levels improved all of the metrics of discrimination and reclassification when added to the intima–media thickness (C-statistic: p=0.002; net reclassification improvement (NRI): p<0.001; integrated discrimination improvement (IDI): p<0.001), family history of myocardial infarction (C-statistic: p=0.011; NRI: p=0.004; IDI: p=0.003) and elevated high-sensitivity C reactive protein (C-statistic: p=0.008; NRI: p=0.015; IDI: p=0.009).

Conclusions Elevated plasma 7-KC levels are associated with the incident CVD events in a population-based cohort. Further studies are needed to confirm this observation.

  • 7-Ketocholesterol
  • Cardiovascular Disease
  • Risk
  • Primary end point

Statistics from Altmetric.com

Introduction

Atherosclerotic cardiovascular disease (CVD) is the leading cause of death and disability worldwide, and its prevalence is expected to rise sharply over the next several decades. The basic pathophysiological features of CVD are characterised by the development of intimal fibrolipid plaques,1 2 which are mainly attributed to proinflammatory and pro-oxidative mechanisms.2–4 Oxysterols, which are generated via enzymatic or non-enzymatic oxidation of cholesterol by oxygen free radicals,5 possess much higher biochemical reactivity than that of the parent compound itself. Oxysterols are present at low levels in the circulation and accumulate is plasma and tissues in some pathologies.6 7-Ketocholesterol (7-KC) is a well-studied oxysterol formed by the auto-oxidation of cholesterol and cholesterol esters present in lipoprotein deposits.7 The oxidation of cholesterol to 7-KC could be caused by catalytic levels of Fe2+ and/or Cu2+ (known as the Fenton’s reaction). 7-KC has been more abundant in the circulation of patients with hypercholesterolaemia and accumulates in atherosclerotic lesions than any other oxysterol.7 8 In vitro studies show that 7-KC decreases the activity of endothelial nitric oxide synthase and impairs nitric oxide bioavailability in endothelium, thus initiating the formation of atherosclerotic lesions. Furthermore, exposure of macrophages to 7-KC cause oxidative damage, apoptosis and necrotic core formation, resulting in plaque instability and rupture.9 10 Inhibition of 7-KC accumulation in endothelial cells and macrophages has been proven to be an effective approach to prevent and reverse the progression of atherosclerosis in animal studies.11 12

However, the clinical relevance of circulating 7-KC with the incidence of cardiovascular events has not been evaluated. We hypothesised that the elevated plasma 7-KC levels are a reflection of endogenous oxidative stress and could be reliable markers of CVD. Therefore, in the present study, we prospectively evaluated the relationship of the plasma 7-KC levels with the incidence of CVD events.

Methods

Study participants

The study participants in this cohort study were recruited as a part of an ongoing nutrition and health study of ageing populations in South China, which were designed to evaluate the relationship of dietary and genetic factors with ageing-related chronic diseases. From March 2008 to December 2009, 2629 participants aged 50–80 years were enrolled in this cohort if they were Guangzhou residents for at least 10 years. The study protocol was approved by the Institutional Review Board of the Sun Yat-sen University, and written informed consent was obtained from all participants at the time of enrolment. Participants with a history of CVD, including a self- reported history of myocardial infarction (MI), stroke, arterial revascularisation, heart failure or arrhythmia and niacin use were excluded. Of all the participants, 181 were excluded because of a history of CVD, 241 were lost to follow-up, 160 lacked valid serum 7-KC measurements and 103 had missing covariates, leaving 1944 subjects for the final analysis of cardiovascular outcomes. Participants were followed up prospectively up to 30 June 2014 with a median follow-up of 5.2 years.

Data collection

Participants were interviewed to collect information regarding demographic characteristics, medical history, medication use, behavioural habits and risk factor prevalence at the baseline visit. Current smoker was defined as at least one cigarette per day for at least 6 months. Alcohol use was defined as consumption of any type of alcoholic beverage at least once a week for at least 6 months. Medical records were reviewed to confirm a history of MI and other conditions as well as to document previous angiographic findings and prior coronary revascularisation. The body mass index (BMI) is calculated by an individual’s weight in kilograms divided by the square of height in metres. The systolic blood pressure and diastolic blood pressure were measured three times at 5 min intervals in the morning.

Covariate definition

Hypertension was defined as a mean blood pressure of ≥140/90/90 mm Hg or self-reported use of antihypertensive medication, and diabetes was diagnosed as a fasting plasma glucose level ≥7.0 mmol/L, postprandial plasma glucose level ≥11.1 mmol/L or self-reported use of antidiabetic medication. Hypercholesterolaemia was considered present if the following criteria were met: (1) total cholesterol ≥6.45 mmol/L, (2) self-reported hyperlipidaemia or (3) self-reported use of lipid lowering drugs. Family history (FH) was defined as any first-degree relative with a history of MI. The history was obtained by asking participants whether any member in their immediate family (first-degree relatives (parents, siblings or children)) had experienced either a fatal or non-fatal MI or a stroke.13

Intima–media thickness measurement

The intima–media thickness (IMT) of the carotid artery was measured by using a high-resolution B-mode ultrasound (ARIETTA V70; Hitachi Medico, Tokyo, Japan) with a linear array 5.5/7.5 MHz probe. An anterolateral approach was used to longitudinally capture the right common carotid, femoral and common iliac arteries. The best image was selected to show the far wall intimal–lumen interface as a continuous straight line. Three determinations of the IMT were made at 1 cm proximal to the bulb and at the site of greatest thickness. The values at each site were averaged, and the average IMT was used as the representative value for each subject. Subclinical atherosclerosis was defined as an IMT >1.0 mm and/or a plaque on one or more of the three arteries (common carotid artery, femoral artery and common iliac artery) but without clinical manifestations.14

Laboratory measurement at enrolment

Fasting blood samples were collected at the time of enrolment, by venipuncture into tubes containing EDTA and butylated hydroxytoluene (BHT) immediately centrifuged at 4°C and stored at −80°C before analysis. BHT was used to prevent the occasional formation of cholesterol oxidation products. Total blood cholesterol, high-density lipoprotein (HDL) cholesterol and triglycerides were measured enzymatically using an automatic analyser (Hitachi 747 autoanalyser, Hitachi) using a commercial assay kit (Wako Pure Chemical Industries, Osaka, Japan). Low-density lipoprotein (LDL) cholesterol was subsequently calculated using the Friedewald formula. Plasma levels of high-sensitivity C reactive protein (hsCRP) were determined by a high-sensitivity turbidimetric assay (Roche, Basel, Switzerland). The intra-assay and interassay coefficient of variation for all of the measured biochemical parameters was <5%.

Estimated glomerular filtration rate (eGFR) was calculated according to the levels of serum creatinine and cystatin C.15

Plasma 7-KC assessment

Plasma 7-KC levels were measured according as previously described.16 The 7-KC standard was purchased from Sigma Chemicals (St Louis, Missouri, USA). 24,25,25,25,26,26,26-d7-7-Ketocholesterol (d7-7-KC) was prepared by CDN Isotopes (Quebec, Canada) and used as an internal standard. The internal standard solution (d7-7-KC) was prepared in methanol at a concentration of 50 ng/mL and stored at −80°C until use. The endogenous 7-KC was removed by adding activated charcoal (100 mg/mL) to the isolated plasma and the mixture was stirred at room temperature overnight. Then, the charcoal was removed by centrifuging with a Beckman Optima MAX-E ultracentrifuge (Beckman Coulter, Brea, California, USA) at a speed of 80 000 rcf. The charcoal-stripped plasma was then used as blank plasma which was treated with a stock solution containing a 10 µg/mL concentration of 7-KC in methanol. The final concentrations of the calibration series were 2.5, 25 and 250 ng/mL of 7-KC. The calibrators were stored in aliquots at −80°C before analysis. Stored plasma samples were thawed and vortex mixed, and a 25 µL aliquot of plasma was mixed with 100 µL of internal standard solution. The mixture was vigorously mixed and centrifuged at a speed of 13 500 rcf for 5 min at room temperature, and then the supernatant was transferred to a 96-well plate. The measurement was performed on an Ultra Performance Liquid Chromatography system with a Xevo TQ MS detector (Waters, New York, USA). 7-KC and d7-7-KC were detected in positive electrospray ionisation through the multiple reaction monitoring modes. Mobile phase A consisted of water with 0.5% formic acid and mobile phase B contained high-performance liquid chromatography-grade methanol with 0.5% formic acid. The gradient was elevated from 80% mobile phase B to 95% of B over 3 min, then stepped to 100% of B and maintained for 1 min. The column was then reconditioned to the initial 80% of B for 1 min. The injection volume of each sample was 10 µL. All of the determinations were conducted within 2 months after the participant was enrolled in the cohort. The recovery of plasma 7-KC was in the approximately range of 85% to 95%. All plasma samples were run in duplicate and the values were averaged. The mean intra-assay and interassay coefficient of variation was 4.1% and 7.8%, respectively. The detection limit of this assay was 13.4 nmol/L. To evaluate the stability of the plasma 7-KC measurements, we measured 7-KC in the plasma samples that were either freshly isolated or stored for 6 months under different operating conditions: (1) in the presence or absence of BHT and (2) after overnight storage either at 4°C or at room temperature (at 25°C). We found that both BHT usage and long-term storage of plasma did not affect 7-KC stability.

Perspective follow-up

The primary endpoint was a composite CVD adverse outcome, which was defined as a composite of a first non-fatal MI, non-fatal stroke, coronary revascularisation (including percutaneous coronary intervention or coronary artery bypass grafting) or death from cardiovascular causes. MI was confirmed if symptoms of ischaemia were present and if either the event was associated with diagnostic changes in cardiac enzyme levels or there were diagnostic electrocardiographic changes. CVD death, which was confirmed by record review, death certificates, autopsy reports and information provided by family members, was defined as either death attributable to an ischaemic cardiovascular event (including fatal MI, stroke, peripheral arterial disease) or sudden death due to an unknown but presumed cardiovascular cause in high-risk participants. Patients experiencing either MI or revascularisation within a month of enrolment were not included. All of the endpoints were adjudicated by two independent cardiologists who were unaware of the measurements of plasma 7-KC. The National Death Index was used to determine the vital status for all of the participants through 31 December 2010. Death from cardiovascular causes was defined according to the International Classification of Diseases, 10th Revision, codes I00 to I99.

Statistical analysis

Baseline continuous variables are reported as the mean±SD, and categorical variables are presented as percentages. Using the Cox proportional hazards model, HRs of 7-KC with the primary endpoint (ie, the time to a first event for CVD), with 95% CIs, were assessed by unadjusted univariable models as well as multivariable adjusted models for potential confounding factors. The factors selected were established cardiovascular clinical and metabolic risk factors (such as age, sex, BMI, presence of diabetes, hypertension, current smoking, LDL and HDL cholesterol) that in previous studies were found to associate with plasma 7-KC levels. Statin use was additionally adjusted for the analysis. The proportional hazard assumption was tested by Schoenfeld residuals. All results reported from modelling 7-KC in relation to CVD outcome are based on participants with complete data on all variables included in each specific model. Effect modification of 7-KC by sex, IMT, FH or hsCRP was tested, respectively, by inserting interaction terms in the multiple adjusted models. Kaplan-Meier curves for plasma 7-KC levels below the median (vs above the median) for the overall cohort were compared by the log-rank test and stratified by the presence of a prevalent IMT (defined as an IMT >1.0 mm), FH and elevated hsCRP levels (defined as hsCRP >2 mg/L). The improvement of 7-KC levels over and above that of traditional risk factors in the prediction of the incidence of CVD events was analysed with the use of multiple metrics of biomarker performance,17 including discrimination (Harrell’s C-statistic), reclassification (integrated discrimination improvement index)18 and the net reclassification index.19 A two-sided p value of 0.05 or less was considered to indicate statistical significance. All statistical analyses were performed with the use of SPSS V.19.0 statistical packages.

Results

The baseline characteristics of the 1944 study participants are categorised into four groups according to the quartiles of baseline plasma 7-KC concentrations (table 1). The mean age of all the participants was 59.9 years, approximately 61% of the subjects were male and all were Chinese citizens. Subjects in the highest quartile of plasma 7-KC levels were often older in age, had a higher BMI, were more likely to smoke, had a higher prevalence of hypertension and diabetes mellitus, systolic blood pressure and hsCRP and had lower HDL cholesterol levels. Overall statin use was more frequent in the subjects in 7-KC quartiles 3 and 4 compared with those in quartiles 1 and 2. The mean plasma 7-KC concentration was 29.8 nmol/L with a median concentration of 28.4 nmol/L. The median value was higher in men than in women: 30.5 vs 27.3 nmol/L, but these values did not differ significantly between genders. Plasma 7-KC concentrations positively correlated with LDL cholesterol (r=0.116, p<0.001) but negatively correlated with HDL cholesterol (r=–0.107, p=0.006) (table 2).

Table 1

Baseline characteristics of the study subjects according to the quartile of plasma 7-KC

Table 2

Correlation of 7-KC with cholesterol level

During a mean follow-up of 4.8 years (median: 5.2; interquartile range: 4.3–6.2 years, representing 9342 person-years of follow-up), 101 of the 1944 participants had a primary CVD event with complete follow-up data. Univariate regression analysis revealed a significant association between plasma 7-KC levels at baseline and the incidence of CVD events (unadjusted HR 2.06, 95% CI 1.70 to 2.41, p<0.001) per SD increase (6.9 nmol/L) of plasma 7-KC at baseline. This association was still applicable even after adjusting for multiple established CVD risk factors and statin use (adjusted HR 1.38, 95% CI 1.08 to 1.71, p=0.014; table 3). Furthermore, this association was still significantly evident when the baseline plasma 7-KC levels were categorised into quartiles. There is a gradually inverse association between the increasing quartiles of 7-KC and the CVD endpoint. The fully adjusted HR (95% CI) from the lowest to highest quartiles of plasma 7-KC levels were 0.97 (95% CI 0.74 to 1.22), 1.27 (95% CI 1.01 to 1.53) and 1.70 (95% CI 1.45 to 1.91), respectively, using quartile 1 as a reference (p value for trend=0.004; table 3).

Table 3

Association of plasma 7-KC concentrations with incident CVD event

Kaplan-Meier curves showed participants with plasma 7-KC above median versus below median had a higher risk of the incident CVD event during the follow-up period (figure 1A). Among participants with prevalent IMT (n=673, figure 1B), FH (n=422, figure 1C), elevated hsCRP (n=387, figure 1D), those with plasma 7-KC above median versus below median had a higher risk of the incident CVD event. The relationship between plasma 7-KC and the incidence of CVD was also analysed in men and women independently. The positive associations were similar across genders (data not shown). The p value for the 7-KC interaction with regard to gender was 0.87.

Figure 1

Kaplan-Meier analysis of plasma 7-KC level with major adverse cardiovascular events. (A) Kaplan-Meier curves for CVD events in participants with 7-KC below median versus above median among the overall study participants. (B) Kaplan-Meier curves for CVD events in participants with 7-KC below median versus above median among those with prevalent IMT (>1.0 mm). (C) Kaplan-Meier curves for CVD events in participants with 7-KC below median versus above median among those with prevalent FH of myocardial infarction. (D) Kaplan-Meier curves for CVD events in participants with 7-KC below median versus above median among those with elevated hsCRP (>2 mg/L). 7-KC, 7-ketocholesterol; CVD, cardiovascular disease; FH, family history; IMT, intima–media thickness; high-sensitivity C reactive protein.

As both 7-KC and total cholesterol were independently associated with the incidence of CVD events in participants, we used additional high/low categories to identify the potential value of combining 7-KC and total cholesterol measurements, which were defined as values above or below the median value of each parameter, for predicting future CVD events (table 4). In comparison with patients with low 7-KC/low total cholesterol, those with high 7-KC/low total cholesterol had a multivariate adjusted HR of 1.47 (95% CI, 1.20 to 1.84, p<0.001) for risk of incident events. Patients with high 7-KC/high total cholesterol had a fully adjusted HR of 1.51 (95% CI 1.24 to 1.88, p<0.001) compared with those with low 7-KC/low total cholesterol for the risk of future events.

Table 4

HR for incident CVD event associated with 7-KC by level of total cholesterol

The inclusion of 7-KC to improve CVD risk prediction beyond using IMT, FH and hsCRP was assessed using the metrics of calibration, discrimination and reclassification. All of the models including 7-KC were well calibrated. 7-KC levels above the median versus below the median improved the discrimination indexes as determined by the C-statistic and IDI when added to risk factor-adjusted models, including prevalent IMT, FH or elevated hsCRP (table 5). With respect to reclassification according to the NRI, the inclusion of 7-KC resulted in the significant reclassification of all models (table 5).

Table 5

Risk prediction metrics with for 7-KC in the cohort for incident CVD event

Discussion

In the present study, we show that baseline plasma 7-KC levels were significantly associated with adverse CVD outcomes in a fully adjusted multivariate model. Despite the well-known biological proatherogenic properties of 7-KC, to our knowledge, this is the first study to evaluate the relationship of 7-KC formation with CVD risk in the general population. The positive association of 7-KC with the incidence of CVD was still significant when accounting for three established risk markers: IMT, FH and elevated hsCRP. Furthermore, among participants with prevalent IMT, FH or hsCRP, 7-KC was able to significantly stratify CVD risk. Thus, this cholesterol oxidation product could serve as a biochemical marker and may prove useful for the early prediction of CVD.

It has been proposed that higher 7-KC levels have biological roles, including impairing endothelial-dependent vasorelaxation, promoting macrophage-derived foam cells and enhancing oxidative damage.9 10 Elevated levels of circulating 7-KC might mirror metabolic and cardiovascular effects via such mechanisms. However, the clinical significance of blood 7-KC levels is not fully understood. Although the relationship of 7-KC with disease progression is limited by the pleiotropic nature of the disease, we were able to demonstrate the pronounced positive association of plasma 7-KC levels with future CVD risk, which indicates the pathophysiological role of 7-KC in CVD in the general population.

We further examined the incremental ability of 7-KC to improve CVD risk prediction beyond using the IMT, FH or hsCRP in this population-based cohort. IMT, FH and hsCRP, which are three validated biomarkers of CVD risk, may reflect different aspects of the development of CVD (subclinical atherosclerosis, inherited risk and inflammation, respectively). The IMT, a surrogate marker of atherosclerosis in adults, is a non-invasive measurement and an inexpensive tool for risk prediction of subclinical atherosclerosis.20 21 FH represents a heritable marker of genetic risk for CVD.22 23 hsCRP is a circulating marker of inflammation that correlates with the future risk of CVD.24 25 Multiple large population-based prospective studies have demonstrated the association of elevated hsCRP levels with cardiovascular events and ability of these levels to improve CVD risk prediction after accounting for traditional cardiovascular risk factors.26 27 A Cox proportional hazards analysis demonstrated that 7-KC improved risk stratification among participants with a prevalent IMT, FH or elevated hsCRP, providing strong evidence that 7-KC may improve the ability to predict the incidence of CVD events beyond IMT, FH and hsCRP. Thus, these findings demonstrate that the pathways that mediate the association of 7-KC and CVD in this study population are not fully reflected by the atherosclerotic pathways reflected by the IMT, the inherited risk reflected by family history or the inflammatory pathways reflected by elevated CRP. Taken together, we conclude that the identification of increased 7-KC levels among individuals may contribute to a more correct classification of risk of CVD.

Clinical relevance

Plasma 7-KC is a stable cholesterol metabolite, can be reproducibly measured and exhibits only modest variations between men and women, which augurs well for its use as a clinical biomarker for the prediction of clinical adverse CVD endpoints. Cholesterol serves as a precursor in the synthesis of 7-KC which is produced by the oxidation of cholesterol.28 29 To further calibrate these findings and discriminate the influence of circulating total cholesterol on association of 7-KC with CVD risk, we introduced a ratio between the oxidation product and precursor (7-KC/total cholesterol) as a metric for further analyses with the incidence of a CVD event. Consistently, a similar association was found between the 7-KC/total cholesterol ratio and the primary CVD endpoints in a multivariable adjusted model.

The strengths of this study include (1) a population-based prospective design with a 60-year-old participant cohort, which was considered to be a relatively homogenous study population; (2) a relatively strict protocol for plasma sample collection with high-quality annual follow-up of study subjects; (3) a high participation rate, which may increase generalisability and reduce the risks of selection bias and (4) adjusted multiple variables, including traditional risk factors to explore the relationship of 7-KC with the incidence of CVD risk.

Limitations

Several limitations of the present study merit acknowledgement. Due to the limited number of CVD events, we could not perform subgroup analyses for MI and stroke. The race/ethnicity of our study sample is exclusively Chinese, and the findings need to be replicated and extended to other ethnicities and different age groups. We did not have dietary information from our subjects, and consumption of a high cholesterol diet may influence plasma 7-KC levels. However, we examined an older population who tend to maintain healthy dietary patterns, which could minimise the possible moderate fluctuations of 7-KC levels induced by dietary changes. Ankle–brachial index and coronary artery calcium (are other two non-conventional risk factors for risk factor assessment in addition to IMT, FH and hsCRP but were not available in this cohort. Finally, the reported NRI measures should be interpreted with caution considering that they are based on several arbitrary choices and risk category cut points and typically cannot be generalised to other studies.30

Conclusion

Plasma 7-KC levels are positively and independently associated with CVD risk in a population-based cohort. Our findings support the importance of 7-KC in cardiovascular risk prediction beyond the established cardiovascular risk factors.

Key messages

What is already known on this subject?

7-Ketocholesterol (7-KC) is a type of oxysterol formed by the auto-oxidation of cholesterol and cholesterol esters present in lipoprotein deposits. 7-KC was found to be abundantly accumulated in the circulation of patients with hypercholesterolaemia and in atherosclerotic lesions. However, prospective epidemiological studies relating circulating 7-KC levels to risk of cardiovascular disease (CVD) are limited.

What might this study add?

The present study show that baseline plasma 7-KC levels were significantly associated with adverse CVD outcomes in a fully adjusted multivariate model. The positive association of 7-KC with the incidence of CVD was still significant when accounting for three known risk markers: carotid intima media (IMT), family history (FH) and elevated high-sensitivity C reactive protein (hsCRP). Furthermore, among participants with prevalent IMT, FH or hsCRP, 7-KC was able to significantly stratify CVD risk. Thus, 7-KC may serve as a biochemical marker in the prediction of incident CVD.

How might this impact on clinical practice?

Our findings support the importance of 7-KC in cardiovascular risk prediction beyond the established cardiovascular risk factors in a population-based cohort.

References

Footnotes

  • Contributors MW, WQL, DL and DW contributed to data acquisition. MW, WQL, DL, DW, YZ, DM and JYS contributed to data analysis and statistical analyses. MX drafted the manuscript. MW and DL contributed to critical revision of the manuscript.

  • Funding This study was supported by the Funding from the National Natural Science Foundation from Guangdong Province (No. S2012020011104).

  • Competing interests None declared.

  • Ethics approval Institutional Review Board of the Sun Yat-sen University.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.