Author + information
- Received September 4, 2015
- Revision received September 24, 2015
- Accepted September 28, 2015
- Published online December 22, 2015.
- Gert-Jan R. ten Kate, MD∗,†,‡,
- Sven Bos, MD§,
- Admir Dedic, MD∗,†,‡,
- Lisan A. Neefjes, MD, PhD∗,†,‡,
- Akira Kurata, MD, PhD∗,
- Janneke G. Langendonk, MD, PhD§,
- Anho Liem, MD, PhD¶,
- Adriaan Moelker, MD, PhD∗,
- Gabriel P. Krestin, MD, PhD∗,
- Pim J. de Feyter, MD, PhD∗,‡,
- Jeanine E. Roeters van Lennep, MD, PhD§∗ (, )
- Koen Nieman, MD, PhD∗,‡ and
- Eric J. Sijbrands, MD, PhD§
- ∗Department of Radiology, Erasmus Medical Centre, Rotterdam, the Netherlands
- †Interuniversitair Cardiologisch Instituut Nederland, Utrecht, the Netherlands
- ‡Department of Cardiology, Thorax Centre Rotterdam, Rotterdam, the Netherlands
- §Department of Internal Medicine, Erasmus Medical Centre Rotterdam, Rotterdam, the Netherlands
- ¶Department of Cardiology, Sint Franciscus Gasthuis, Rotterdam, the Netherlands
- ↵∗Reprint requests and correspondence:
Dr. Jeanine E. Roeters van Lennep, Erasmus Medical Centre, Department of Internal Medicine, Room Bd-424, ‘s-Gravendijkwal 230, 3015 CE Rotterdam, P.O. Box 2040, 3000 CA, Rotterdam, the Netherlands.
Background Familial hypercholesterolemia is typically caused by LDL receptor (LDLR) mutations that result in elevated levels of LDL cholesterol (LDL-C). In homozygous FH, the prevalence of aortic valve calcification (AoVC) reaches 100% and is often symptomatic.
Objectives The objective of this study was to investigate the prevalence, extent, and risk-modifiers of AoVC in heterozygous FH (he-FH) that are presently unknown.
Methods Asymptomatic patients with he-FH and 131 non-familial hypercholesterolemia controls underwent CT computed tomography calcium scoring. AoVC was defined as the presence of calcium at the aortic valve leaflets. The extent of AoVC was expressed in Agatston units, as the AoVC-score. We compared the prevalence and extent of AoVC between cases and controls. In addition, we investigated risk modifiers of AoVC, including the presence of LDLR mutations without residual function (LDLR-negative mutations), maximum untreated LDL-cholesterol (maxLDL), LDL-C, blood pressure, and coronary artery calcification (CAC).
Results We included 145 asymptomatic patients with he-FH (93 men; mean age 52 ± 8 years) and 131 non-familial hypercholesterolemia controls. The prevalence (%) and AoVC-score (median, IQR) were higher in he-FH patients than in controls: 41%, 51 (9–117); and 21%, 21 (3–49) (p < 0.001 and p = 0.007). Age, untreated maxLDL, CAC, and diastolic blood pressure were independently associated with AoVC. LDLR-negative mutational he-FH was the strongest predictor of the AoVC-score (OR: 4.81; 95% CI: 2.22 to 10.40; p = <0.001).
Conclusions Compared to controls, he-FH is associated with a high prevalence and a large extent of subclinical AoVC, especially in patients with LDLR-negative mutations, highlighting the critical role of LDL-C metabolism in AoVC etiology.
- aortic valve calcification
- calcific aortic stenosis
- coronary artery calcification
- familial hypercholesterolemia
- LDLR-negative mutation
- low-density lipoprotein receptor
Aortic valve calcification (AoVC) has an estimated prevalence of >50% in the elderly (i.e., those aged >75 years) and is associated with an elevated risk of coronary (72%) and cardiovascular (50%) events (1,2). In addition, the degree of AoVC correlates with severity of stenosis, disease progression, and the development of coronary and cardiovascular events (3–5).
In the general population, AoVC is associated with age, male sex, smoking, hypertension, diabetes, obesity, and hypercholesterolemia (6,7). Patients with familial hypercholesterolemia (FH) have extremely high levels of low-density lipoprotein cholesterol (LDL-C) and may be at high risk for developing AoVC. FH is an autosomal inherited disorder caused by mutations in the LDL receptor (LDLR) gene, the apolipoprotein B (APOB) gene, or the proprotein convertase subtilisin/kexin type 9 (PCSK9) gene (8). LDLR mutations can be classified as mutations with residual LDLR function (LDLR-defective mutations) or without LDLR function (LDLR-negative mutations) (9).
In patients who are homozygous for FH, the prevalence of AoVC reaches 100%, and surgical intervention of functional valvular disease is often needed (10,11). Compared with homozygous FH, heterozygous familial hypercholesterolemia (he-FH) is associated with less aortic valve dysfunction on echocardiography (12–15). However, the prevalence of AoVC in he-FH is unknown.
The purpose of the present single-center study was to determine the prevalence and extent of AoVC in asymptomatic statin-treated patients, heterozygous for FH. In addition, we evaluated which variables were associated with the presence and extent of AoVC. In the molecular context of the patients, AoVC was compared between he-FH patients with and without LDLR-negative mutations.
Between February 2008 and June 2011, a total of 145 consecutive patients with he-FH were included in the study. Between November 2006 and January 2011, we also included 131 consecutive patients with nonanginal chest pain (NACP) as a control group. Patients with NACP were used as a substitute for asymptomatic patients without he-FH because the radiation exposure limits the choice of control subjects to patients with an indication for cardiac computed tomography (CT) scanning.
Patients with NACP were referred by their general practitioner for the evaluation of chest pain and underwent stress testing and cardiac CT scanning. They had no history of coronary artery disease (CAD). NACP was defined as chest pain or discomfort that was not provoked by exertion or emotional stress or relieved by rest or nitroglycerin (16).
Patients with he-FH were recruited from our tertiary outpatient lipid clinic. he-FH was determined either by the presence of a confirmed LDLR or APOB gene mutation (the patients did not have PCSK9 mutations) or clinically as having untreated LDL-C levels above the 95th percentile for sex and age in combination with at least 1 of the following: the presence of typical tendon xanthomas in the patient or a first-degree relative; an LDL-C level above the 95th percentile for sex and age in a first-degree relative; or proven CAD in a first-degree relative aged <60 years (17).
DNA samples were taken of all patients with a clinical suspicion of he-FH and were sent to a central laboratory for mutational screening (18). A complete overview of the mutations found and clinical characteristics of both LDLR-negative and LDLR-defective he-FH has been published previously (19). Plasma lipid levels were measured by using the fasting blood samples at the time of inclusion. Cholesterol levels before statin treatment were obtained from patient medical records, and they were used as the variable untreated maximum total cholesterol and untreated maximum low-density lipoprotein cholesterol (maxLDL).
Exclusion criteria were: symptoms of CAD, history of CAD, rheumatic fever, or known aortic valve pathology (although cardiac ultrasounds were not routinely performed before study inclusion). Patients with a secondary cause of hypercholesterolemia (e.g., renal, liver, or thyroid disease) were also excluded from the study. Further exclusion criteria were renal insufficiency (serum creatinine >120 μmol/l), known contrast allergy, and irregular heart rhythm (atrial fibrillation). In patients with asymptomatic he-FH, the inclusion age was 40 to 70 years for men; women were included if they were older than childbearing age because of potential radiation-induced harm to the fetus or ovaries. Women’s inclusion age was thus 45 to 70 years.
This study complied with the Declaration of Helsinki, and the institution’s human research committee approved the study protocol. All patients provided written informed consent.
CT calcium score
To quantify the AoVC, as well as the coronary calcium score, a cardiac CT scan without contrast medium was performed, which enabled calcium scoring with high accuracy and reproducibility (20,21). All CT scans were performed on a dual-source CT scanner (first 232 scans: Somatom Definition; last 44 scans: Somatom Definition FLASH [Siemens Medical Solutions, Forchheim, Germany]), with application of a prospectively electrocardiogram-triggered scan protocol with a tube current of 76 mAs at 70% of the RR-interval. Images were reconstructed with a slice thickness of 3 mm and an increment of 1.5 mm by using a medium convolution kernel (B35f).
Lesions were classified as AoVC if located within the aortic valve leaflets, exclusive of the aortic annulus or coronary arteries, and contained ≥3 contiguous pixels with an attenuation value >130 HU (2,21). The AoVC score was defined as the quantity of AoVC expressed in Agatston units, according to the same lesion definition as for coronary artery calcification (CAC), using dedicated software (MMWP, Siemens Medical Solutions) (22). The CT reading was performed blinded with regard to patient characteristics. The absence of AoVC was assigned a score of 0. Additional information regarding the scan protocol, including the quantification of CAC, has been previously published (23).
Contrast-enhanced scans were consulted if the exact location of calcified lesions (in the valve or aortic root) were unclear.
Categorical variables are expressed as n (%). Normally distributed continuous variables are shown as mean ± SD and skewed variables as median (interquartile range). To determine the differences between he-FH patients and NACP patients, a Pearson chi-square test was used to compare binary variables. Continuous variables with a normal distribution were analyzed with a Student t test, and skewed variables were assessed by using a Mann-Whitney U test. Statistical significance was considered at a 2-sided p value <0.05.
We compared the prevalence of AoVC and the AoVC scores between he-FH and NACP patients in relationship to age. Age categories were chosen on the basis of equal patient numbers in all groups (n = 92 [age range: 40 to 50 years], n = 92 [age range: 51 to 58 years], and n = 92 [age range: 59 to 70 years]).
To evaluate which variables were associated with AoVC, a univariable ordinal logistic regression model was used. Subsequently, the AoVC score (Agatston units) was divided into 3 groups based on equal distribution of patients in whom AoVC was present: 1) AoVC score = 0 (n = 190); 2) AoVC score >0 to 37 (n = 43); and 3) AoVC score >37 (n = 43). Variables associated with AoVC where analyzed in the entire cohort with a multivariable ordinal logistic regression model to identify a set of predictors of AoVC. An ordinal regression model was chosen instead of a linear regression model to investigate the dose–response relationship between the highly skewed AoVC variable and other variables. Although correcting the skewness by logistically transformation would have also been possible, we did not want to change the data into an artificial score in the 190 subjects whose Agatston score was 0.
In addition, the association between the presence of CAC and AoVC was examined for patients with he-FH and the control subjects. The differences in distribution of AoVC in the presence and absence of CAC were analyzed with a Pearson chi-square test, and differences between the AoVC score in the presence and absence of CAC were assessed by using the Mann-Whitney U test. Finally, the influence of LDLR-negative and LDLR-defective mutational he-FH on AoVC was compared. Data were analyzed by using SPSS version 22 (IBM SPSS Statistics, IBM Corporation, Armonk, New York).
Baseline characteristics of patients with he-FH and control subjects
Age, systolic blood pressure, and the prevalence of hypertension and diabetes were higher in the control group than in the he-FH group (Table 1). Patients with he-FH had higher untreated maxLDL levels, more frequently used statins, and more often reported a positive family history for CAD compared with the control subjects. Sex, body mass index, and treated cholesterol levels were similar in both groups.
AoVC in patients with he-FH and control subjects
AoVC was compared between patients with he-FH and control subjects (Table 2). AoVC was more prevalent in he-FH patients (41%, n = 59) than in control subjects (21%, n = 27; p < 0.001), irrespective of the age category. Limiting the analysis to patients with AoVC present, the AoVC score (median [interquartile range]) was higher in he-FH patients than in control subjects: 51 (9 to 117) and 21 (3 to 49), respectively (p = 0.007).
Risk factors for AoVC
Risk factors for AoVC are shown in Table 3. The AoVC burden according to the Agatston score was associated with age, untreated maxLDL, LDLR-negative mutational he-FH, CAC, and diastolic blood pressure. Sex, smoking, hypertension, diabetes mellitus, and obesity were not associated with the extent of AoVC.
In the multivariable ordinal regression model, all variables explained 27% of the variance of AoVC, and all remained significantly associated with AoVC. Among the variables, LDLR-negative mutation carrier status was a strong predictor of the extent of AoVC (odds ratio: 4.81; 95% confidence interval: 2.22 to 10.40; p < 0.001). Analyses restricted to the he-FH, LDLR-defective, and LDLR-negative groups had similar results (data not shown).
Association between CAC and AoVC
The presence of CAC was associated with a higher prevalence of AoVC, both in he-FH patients and in the control subjects (Table 4). Of the patients without CAC, ≤4% exhibited AoVC. However, in the absence of AoVC, >39% of patients exhibited CAC.
AoVC and LDLR mutational status
Of 145 patients with he-FH, 59 (41%) had an LDLR-negative mutation. Compared with the he-FH patients with LDLR-defective mutations, LDLR-negative mutational he-FH was associated with: higher total cholesterol (5.8 ± 1.6 mmol/l and 5.3 ± 1.3 mmol/l, respectively; p = 0.026), LDL-C (3.9 ± 1.4 mmol/l and 3.2 ±1.1 mmol/l; p = 0.002), and untreated maxLDL (8.0 ± 2.5 mmol/l and 6.6 ± 1.7 mmol/l; p < 0.001). In addition, he-FH patients with LDLR-negative mutations were younger (51 ± 7 years and 53 ± 8 years; p = 0.040), started statin treatment at a younger age (40 ± 9.8 years and 46 ± 9.4 years; p < 0.001), and used statins for a longer period of time (10 ± 7 years and 7 ± 7 years; p = 0.010). None of the other variables from Table 1 were statistically different between groups.
he-FH patients with LDLR-negative mutations had a higher prevalence of AoVC (n = 31 [53%]) compared with those with LDLR-defective mutations (n = 28 [33%]; p < 0.001) and control subjects (n = 27 [21%]; p = 0.016). The difference in AoVC prevalence between patients with LDLR-defective mutational he-FH and the control subjects was also significant (p = 0.048). In addition, AoVC scores increased faster with age in LDLR-negative he-FH patients than in those with LDLR-defective he-FH (data not shown).
The main findings of the present study are: 1) the prevalence and extent of AoVC were higher in patients with he-FH than in those with nonfamilial hypercholesterolemia; 2) age, untreated maxLDL, LDLR-negative mutational he-FH, and diastolic blood pressure were positively associated with AoVC; 3) the level of treated LDL-C was not predictive of the prevalence and extent of AoVC; and 4) the absence of CAC was associated with a low prevalence of AoVC (Central Illustration).
A recent study by Smith et al. (24) using Mendelian randomization found that a genetic predisposition to elevated LDL-C was associated with the presence of AoVC and the incidence of functional aortic stenosis in large community-based cohorts. According to the authors, their results provide evidence supportive of a causal association between LDL-C and aortic valve disease. We also performed a Mendelian randomization assessment, which is a combination of: 1) an association of the genetic background (the LDLR mutation), with the intermediate trait (untreated maxLDL) and the outcome parameter (AoVC); and 2) an association between the intermediate trait (untreated maxLDL) and the outcome parameter (AoVC), corrected for confounding by multiple regression analysis. This Mendelian randomization approach mimics a randomized controlled trial on a genetic level and suggests a causal role of LDL-C in beginning aortic valve pathology.
In the present study, patients with he-FH were exposed to extremely high levels of LDL before their statin treatment, especially those with LDLR-negative mutational he-FH. This approach could have caused the higher prevalence and quantity of AoVC that we found in patients with he-FH, and particularly those with LDLR-negative mutations. The patients were treated with statins since they were diagnosed with hypercholesterolemia, which dramatically lowered LDL-C levels; this approach thereby also reduced the predictive value of LDL-C toward AoVC. This could explain why untreated maxLDL and LDLR-negative mutational he-FH were predictive of AoVC and why the level of statin-treated LDL-C was not predictive of AoVC in our study.
Diastolic blood pressure was mildly but significantly associated with AoVC in our overall normotensive subjects. One might speculate that the increased diastolic blood pressure promotes stress on the aortic side of the valve leaflets, which is where aortic valve lesions are most commonly found (25). This increased stress on the aortic valve can lead to tissue remodeling and promote inflammation, resulting in calcification, stenosis, and ultimately valve failure (26).
Using electron-beam CT scans, Messika-Zeitoun et al. (27) investigated determinants and progression of AoVC in a population-based follow-up study. New AoVC was found to be associated with elevated LDL-C levels, whereas established AoVC progressed independently of atherosclerotic risk factors and faster with increasing initial extent of AoVC. This led to the hypothesis that elevated levels of LDL-C produce their atherogenic effect during the early phase of AoVC, before the start of statin treatment.
As shown in the Results section, patients with LDLR-negative mutational he-FH started statin treatment at a younger age and used statins for a longer period of time. However, despite their more intense statin treatment, these patients exhibited a higher prevalence of AoVC, which increased more rapidly with age. To investigate the effect of statin treatment on AoVC in our cohort of he-FH patients, whom we knew were exposed to high levels of LDL-C early in life, we included “duration of statin use” in a multivariable ordinal regression model. Duration of statin use, however, was not associated with AoVC after correction for age, untreated maxLDL, LDLR-negative mutational he-FH, and diastolic blood pressure. All other variables remained statistically significant (data not shown).
In addition, 3 major prospective randomized trials found no impact of lipid-lowering therapy on the rate of progression of AoVC (28–30). However, macrophage and osteoclast infiltration of the AoVC were reduced by atorvastatin in cholesterol-fed mice (31). Apparently, other pathogenic risk mechanisms prevail once AoVC has been established. It is known that during the later stages of calcific aortic stenosis, a process of osteoblastic activity prevails over the initial atherosclerotic process, resulting in progressive calcification of the valve that seems unrelated to LDL-C levels or statin treatment and fits the observed independence of AoVC from lipid profile or statin treatment. The extensive AoVC in our young study population suggests that statins create their main effect in preventing aortic valve pathology before the development of aortic valve stenosis. The 3 prospective randomized trials were restricted to patients with beginning aortic valve pathology in whom statins could no longer exert a preventive effect. Although the exact role of serum lipids in the pathogenesis of aortic valve disease is unknown, it is evident that lipid depositions are found within and in proximity to aortic valve lesion, which is not the case in healthy valve leaflets (25). This finding suggests a critical role for lipids in the early onset of aortic valve pathology.
The concept of 2 different phases in the development of AoVC progression is not only essential in comprehending the effect of statin treatment but could also explain the discordant association between AoVC and CAC. As shown in Table 4, the absence of CAC was associated with a very low prevalence of AoVC. The absence of AoVC was not predictive of the absence of CAC, however. It is possible, during the early phase of AoVC, that risk factors for CAC are preconditions for the development of AoVC. However, if AoVC develops after the initial atherosclerotic phase, its progression seems to be regulated by risk factors that differ from those causing CAC (27,32–34).
We found that the prevalence and extent of subclinical AoVC are clearly increased in patients with he-FH, especially in those carrying LDLR-negative mutations. It should be emphasized that AoVC is generally without symptoms, and only a fraction of patients with AoVC ultimately develop clinical aortic stenosis. The reported prevalence of hemodynamically significant aortic valve stenosis on echocardiography is low in he-FH (15). Since statin therapy became available, the risk of cardiovascular disease mortality has been substantially reduced in patients with he-FH (35).
Detection and treatment of patients with he-FH at a young age may not only slow progression of CAD but could also be effective in preventing or slowing the development of AoVC during the early phase of disease. This outcome underlines the clinical importance of studies on the effectiveness of statin use for the primary prevention of AoVC, especially in patients with LDLR-negative mutational he-FH.
This study was a cross-sectional observation of AoVC in patients aged between 40 and 70 years, without clinical outcome data and without functional assessment of stenosis with echocardiography. The cross-sectional design of our study did not allow for proper evaluation of the effects of statins on AoVC, which would require a prospective study and sequential imaging. In addition, we do not have sufficient follow-up in our cohort to assess the clinical consequences of the observed AoVC, which is a limitation of the study’s observational design.
Only asymptomatic patients were selected for study. It remains to be seen if patients with AoVC will eventually develop clinical aortic valve disease, given that the majority of AoVC will not lead to aortic stenosis (36).
The current selection of patients with he-FH who were referred to our university lipid clinic may have more severe AoVC compared with he-FH patients in the general population. This single-center study from a tertiary hospital could have resulted in overestimation of the total prevalence and extent of AoVC. However, this potential selection bias should equally hold for LDLR-negative and LDLR-defective mutations.
We have analyzed patients with a major locus effect in the cholesterol metabolism, but we cannot exclude contributions from variants of other genes, such as NOTCH1 (37,38); calcified bicuspid aortic valves have an especially high heritability. Bicuspid aortic valves have an estimated prevalence of 1% to 2%. The linkage peaks of aortic stenosis and the NOTCH1 gene have not been found on chromosome 19, which is where the LDLR is located (19p13.2). Without co-segregation with the mutations in the LDLR, it is unlikely that our cohort was enriched with variants of these other genes.
Recent studies on coronary atherosclerosis showed that the calcified plaque component increased after long-term statin therapy (39,40). In our study, patients with he-FH and particularly patients with LDLR-negative mutations received higher dosages of statins for longer periods of time, and statin use could therefore be a more complex confounder in our analyses. However, in our multivariable ordinal regression model, the time of statin treatment was not significantly associated with AoVC in patients with he-FH.
Male sex was not found to be a risk modifier of AoVC in our study patients compared with the general population. This is likely caused by the relatively small size of our study compared with the large population-based Heinz Nixdorf RECALL (Risk Factors, Evaluation of Coronary Calcium and Lifestyle Factors) study or the Multi-Ethnic Study of Atherosclerosis (2,5,27).
According to the results of the present study, we concluded that: he-FH was associated with a high prevalence and a large extent of subclinical AoVC; age, diastolic blood pressure, untreated maxLDL, and LDLR-negative mutations were associated with the extent of AoVC; the difference between LDLR-negative and LDLR-defective mutations provides important evidence for the critical role of LDL-C metabolism for the pathogenesis of AoVC; and the absence of CAC was associated with a low prevalence of AoVC, suggesting shared pathophysiological determinants. Worldwide, the majority of patients with he-FH who have been treated with statins for a substantial period of their life are still too young to have developed valve disease. However, due to the prolonged survival in these patients because of statin treatment, our results suggest that aortic valve pathology will be a common problem in aging patients with he-FH, especially in those with LDLR-negative mutations.
COMPETENCY IN MEDICAL KNOWLEDGE: The high prevalence and large extent of aortic valve calcification in treated heterozygous familial hypercholesterolemia (FH) patients suggests that aortic valve pathology will be a common problem among the aging statin-treated heterozygous FH patients.
TRANSLATIONAL OUTLOOK: Future research should be directed toward developing more accurate cardiovascular risk prediction tools.
This work was supported by grant 2006T102 from the Dutch Heart Foundation and the Interuniversitair Cardiologisch Instituut Nederland. Dr. Krestin has served as a consultant for Bracco Diagnostics. Dr. Nieman has received institutional research support from Siemens Medical Solutions, GE Healthcare, and Bayer Healthcare. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. ten Kate and Bos contributed equally to this work.
- Abbreviations and Acronyms
- aortic valve calcification
- coronary artery calcification
- coronary artery disease
- computed tomography
- familial hypercholesterolemia
- heterozygous familial hypercholesterolemia
- low-density lipoprotein cholesterol
- low-density lipoprotein receptor
- maximum low-density lipoprotein cholesterol
- nonanginal chest pain
- Received September 4, 2015.
- Revision received September 24, 2015.
- Accepted September 28, 2015.
- 2015 American College of Cardiology Foundation
- Lindroos M.,
- Kupari M.,
- Heikkila J.,
- Tilvis R.
- Kalsch H.,
- Lehmann N.,
- Mahabadi A.A.,
- et al.
- Stewart B.F.,
- Siscovick D.,
- Lind B.K.,
- et al.
- Brown M.S.,
- Goldstein J.L.
- Hoeg J.M.,
- Feuerstein I.M.,
- Tucker E.E.
- Awan Z.,
- Alrasadi K.,
- Francis G.A.,
- et al.
- Rallidis L.,
- Naoumova R.P.,
- Thompson G.R.,
- Nihoyannopoulos P.
- Diamond G.A.
- Agatston A.S.,
- Janowitz W.R.,
- Hildner F.J.,
- Zusmer N.R.,
- Viamotne M. Jr..,
- Detrano R.
- Neefjes L.A.,
- Ten Kate G.J.,
- Rossi A.,
- et al.
- Otto C.M.,
- Kuusisto J.,
- Reichenbach D.D.,
- Gown A.M.,
- O'Brien K.D.
- Rajamannan N.M.
- Messika-Zeitoun D.,
- Bielak L.F.,
- Peyser P.A.,
- et al.
- Chan K.L.,
- Teo K.,
- Dumesnil J.G.,
- Ni A.,
- Tam J.,
- ASTRONOMER Investigators
- Mohler E.R. 3rd.,
- Gannon F.,
- Reynolds C.,
- Zimmerman R.,
- Keane M.G.,
- Kaplan F.S.
- Versmissen J.,
- Oosterveer D.M.,
- Yazdanpanah M.,
- et al.