Author + information
- Received March 9, 2020
- Revision received June 30, 2020
- Accepted July 7, 2020
- Published online August 31, 2020.
- Shao-Wei Chen, MD, PhDa,b,∗,
- Chang-Fu Kuo, MD, PhDc,d,∗,
- Yu-Tung Huang, PhDb,
- Wan-Ting Lin, MSb,
- Victor Chien-Chia Wu, MDe,
- An-Hsun Chou, MD, PhDf,
- Pyng-Jing Lin, MDa,
- Shang-Hung Chang, MD, PhDb,e and
- Pao-Hsien Chu, MDe,∗ (, )@PoChu3
- aDivision of Thoracic and Cardiovascular Surgery, Department of Surgery, Chang Gung Memorial Hospital, Linkou Medical Center, Chang Gung University, Taoyuan City, Taiwan
- bCenter for Big Data Analytics and Statistics, Chang Gung Memorial Hospital, Linkou Medical Center, Taoyuan City, Taiwan
- cDivision of Rheumatology, Allergy and Immunology, and Center for Artificial Intelligence in Medicine, Chang Gung Memorial Hospital, Taoyuan City, Taiwan
- dDivision of Rheumatology, Orthopaedics, and Dermatology, School of Medicine, University of Nottingham, Nottingham, United Kingdom
- eDepartment of Cardiology, Chang Gung Memorial Hospital, Linkou Medical Center, Taoyuan City, Taiwan
- fDepartment of Anesthesiology, Chang Gung Memorial Hospital, Linkou Medical Center, Chang Gung University, Taoyuan City, Taiwan
- ↵∗Address for correspondence:
Dr. Pao-Hsien Chu, Department of Cardiology, Chang Gung Memorial Hospital, Linkou, No. 5 Fuxing Street, Guishan District, Taoyuan City 33305, Taiwan.
Background Aortic dissection (AD) is a life-threatening emergency. However, the heritability and association of family history with late outcomes are unclear.
Objectives The purpose of this study was to evaluate the effect of family history of AD on the incidence and prognosis of AD and estimate the heritability and environmental contribution in AD in Taiwan.
Methods Both cross-sectional and cohort studies were conducted using Taiwan National Health Insurance database. A registry parent–offspring relationship algorithm was used to reconstruct the genealogy of this population for heritability estimation. The cross-sectional study included 23,868 patients with a diagnosis of AD in 2015. The prevalence and adjusted relative risks (RRs) were evaluated, and the liability threshold model was used to examine the effects of heritability and environmental factors. Furthermore, a 1:10 propensity score–matched cohort comprising AD patients with or without a family history of AD was included to compare late outcomes in the cohort study.
Results A family history of AD in first-degree relatives was associated with an RR of 6.82 (95% confidence interval [CI]: 5.12 to 9.07). The heritability of AD was estimated to be 57.0% for genetic factors, and 3.1% and 40.0% for shared and nonshared environmental factors, respectively. After excluding individuals with Marfan syndrome or bicuspid aortic valve, a family history of AD was associated with an RR of 6.56 (95% CI: 4.92 to 8.77) for AD. Furthermore, patients with AD and a family history of AD had a higher risk of later aortic surgery than those with AD without a family history (subdistribution hazard ratio: 1.40; 95% CI: 1.12 to 1.76).
Conclusions A family history of AD was a strong risk factor for AD. Furthermore, patients with AD with a family history of AD had a higher risk of later aortic surgery than those with no family history of AD.
Aortic dissection (AD) is a life-threatening disease, with a sudden death rate of up to 50% in patients with the acute type of AD (type A) (1). A recent analysis from the International Registry of Acute Aortic Dissection, which included 4,428 patients between 1995 and 2013, reported that the mortality rate was still as high as 22% for patients with acute type A AD and 14% for patients with acute type B AD despite recent advances in diagnostic strategies, medical therapy, and surgical treatment (2). Epidemiology studies from Western countries have reported that AD is an uncommon disease with an incidence of 3.0 to 16 per 100,000 patient-years (3–5). In addition, a previous population-based study from Taiwan reported an average annual incidence of AD of 5.6 per 100,000 persons, and that the incidence increased over time (6). AD is usually asymptomatic until it occurs, and therefore identifying risk factors followed by close surveillance and aggressive control of the risk factors or even preventive ascending aortic replacement in high-risk patients may decrease the frequency of dissection and death.
Several genetic disorders including Marfan syndrome, Ehlers-Danlos syndrome, Turner’s Loeys-Dietz’s syndrome, and bicuspid aortic valve have been shown to be risk factors for aortic aneurysm and AD (1). Other associated disorders include arterial hypertension, smoking, and vascular inflammation (3). Recently, Ma et al. (7) reported that a positive family history of AD was associated with a higher risk of dissection in family members, with an annual probability of AD per first-degree relative 2.77 times higher in patients with a positive rather than negative family history (7). Accordingly, we hypothesized that there may also be an association between a family history and the prognosis of AD.
Because no previous study has thoroughly investigated this issue, we conducted this nationwide family study with 2 study designs on nearly the whole population of Taiwan by using genealogy and linked health information. This study had 3 main aims: the first was to evaluate the extent of family aggregation of AD and estimate the relative risk (RR) of AD in individuals with affected first-degree relatives either with or without diagnosed genetic disorders related to aortic disease using a cross-sectional study design with nearly the whole population of Taiwan in 2015; the second was to assess the relative contributions of genetic and environmental factors of AD using a liability threshold model in the cross-sectional population study; and the third was to conduct a longitudinal study to compare the late outcomes between patients with AD and a family history of AD and those without a family history of AD using a cohort of patients who received a diagnosis of AD in Taiwan from 2000 to 2015.
This study was approved by the Institutional Review Board of Chang Gung Memorial Hospital (approval number 201801531B0) and also by the data holder, the National Health Insurance (NHI) Administration. We identified all people who were actively registered under the NHI program between 2000 and 2015 from the Applied Health Research Data Integration Service provided by the NHI Administration. The NHI program is a government-run universal single-payer health insurance program launched in 1995 and covers over 99% of the population in Taiwan (>24 million residents in 2015). The National Health Insurance Research Database (NHIRD) contains data on all registered beneficiaries, including all outpatient, inpatient, and pharmacy dispensing claims from medical providers contracted with the NHI Administration. All personal registration information in the NHIRD is encrypted and deidentified prior to release to researchers but remains linkable.
Study design, population, and identification of family ascertainment
This study was conducted in 2 parts: study 1 had a cross-sectional study design; and study 2 had a matched cohort study design (Central Illustration). A claims-based registry parent–offspring relationship algorithm was used to reconstruct the genealogy of the study populations for heritability estimation in both study designs. The methods used have been described previously (8,9). Among 38,245,965 beneficiary records included in the registry, 9,296,840 were registered by individuals with no relationships with other people over a span of 20 years (from 1996 to 2015). The remaining beneficiaries were clustered into 4,580,177 families, with a mean family size of 4.74 persons.
In the cross-sectional study (study 1), we identified 23,868 patients with a diagnosis of AD among all beneficiaries under the NHI program in 2015. Furthermore, we identified these patients’ family history using the aforementioned rule and compared the prevalence and adjusted RRs of AD and other syndromic diseases associated with aortic disease in individuals with affected first-degree relatives. In addition, we used the standard ACE (additive genetic [A], common environmental factors shared by family members [C], and nonshared environmental factors [E]) liability threshold model to examine the effects of heritability and/or environmental factors (8–10).
The purpose of study 2 was to compare the late outcomes of patients with AD and a family history of AD with the late outcomes of those without a family history of AD. For this purpose, we recruited a propensity score–matched cohort of AD patients with and without a family history of AD in 2000 at a 1:10 ratio and followed them up to 2015 (Supplemental Figure 1). The variables included in the propensity score matching were age at diagnosis of AD, sex, comorbidities, and medications (Supplemental Table 1). Because of the limited number of AD cases, some cases could not be matched with 10 control subjects. Propensity score matching (PSM) by using greedy nearest neighbor matching (random order) without replacement with a caliper of 0.2 times the SD of the logit of propensity score was used. Finally, 93 cases and 894 matched control subjects were included in the analysis.
Ascertainment of AD
International Classification of Diseases-9th Revision-Clinical Modification (ICD-9-CM) codes were used to identify the patients diagnosed with AD (441.0x). This type of case definition has been applied in previous studies based on administration databases (6,11,12). As AD is a life-threatening acute disease that requires prompt management, most of the patients diagnosed with AD were transferred to a medical center with cardiovascular surgeons. Theoretically, the accuracy of ICD-9-CM coding for AD is higher than for other diseases. Furthermore, the NHI program has strict regulations regarding reimbursements, which include a comprehensive review of all medical records including imaging studies, especially for high-cost diseases such as AD. Reimbursements are only given for cases with an accurate diagnosis as supported by such evidence.
To verify the accuracy of the definition of the included AD population (ICD-9 CM diagnostic code: 441.0) and ensure the internal validity of this study, we performed an internal validation study. This was conducted using a chart review of consecutive patients who were diagnosed with AD at Linkou Chang Gung Memorial Hospital from 2011 to 2013. The data in the NHIRD are anonymized and were therefore linked to our data by sex, birth date, and admission and discharge dates. A total of 306 patients in the NHIRD and our validation cohort were linked. The positive predictive value was 97.06% (297 of 306) (data not shown).
Definition of outcomes
For the secondary aim of this study, there were 4 outcomes: all-cause mortality, major adverse cardiac and cerebral events (MACCE), later aortic surgery, and readmission due to any cause. Cases of all-cause mortality were identified from the Taiwan Death Registry, which can be linked to NHI data using encrypted civil identification numbers. MACCE included acute myocardial infarction, heart failure, stroke, and cardiovascular death. AMI, heart failure, and stroke were defined as a principal diagnosis of admission. Cardiovascular death was defined according to the criteria of the Standardized Definitions for Cardiovascular and Stroke Endpoint Events in Clinical Trials by the U.S. Food and Drug Administration. Aortic surgery was identified according to NHI reimbursement codes from the inpatient claims data. All-cause readmissions were also identified in the inpatient claims data. Each patient was followed from the discharge date of the index AD admission to the date of an event, death, or December 31, 2015, whichever occurred first. In addition, the ICD-9 diagnostic codes for detecting comorbidities and outcomes are provided in Supplemental Table 2.
In the cross-sectional study, the prevalence of AD was calculated for the general population with affected first-degree relatives. We calculated the adjusted RRs of AD by sex and relationship (parent, offspring, sibling, twin, and spouse) of the affected relative. In addition, we calculated the adjusted RRs of other diseases (Marfan, Bicuspid, Turner, and Ehlers-Danlos syndromes; inflammatory diseases; type 1 and 2 diabetes mellitus; and hypertension) in first-degree relatives of the patients with AD. Moreover, we applied the liability threshold model to examine the influences of heritability and/or environmental factors.
In the cohort study, the risk of time to an event outcome was compared between the patients with AD with and without a family history of AD using both Cox proportional hazard regression and a Fine and Gray model, which considered all-cause mortality as a competing risk. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, North Carolina). A 2-sided p value <0.05 was considered statistically significant, and no adjustments for multiple testing (multiplicity) were made in this study.
AD prevalence in individuals with affected first-degree family members and in the general population
The incidence rate of AD significantly increased from 5.2 to 8.5 events per 100,000 person-years from 2001 to 2015 (p trend <0.001) (Supplemental Table 3). Individuals with an affected relative had a higher prevalence of AD than the general population in both women and men (0.21% vs. 0.03% in women; 0.41% vs. 0.08% in men) (Table 1). As shown in Figure 1 and Supplemental Table 4, the age-specific prevalence of AD was higher in the individuals with an affected relative with AD than in the general population. The difference in the prevalence of AD between the 2 groups increased with increasing age (Supplemental Table 4).
The RR for AD in individuals with affected relatives
Table 2 lists the prevalence of AD in individuals with affected first-degree relatives. The RRs (95% confidence intervals [CIs]) for AD were associated with the degree of genetic distance between family relatives. The results showed that first-degree relatives (on average 50% genetic similarity) were associated with an adjusted RR of 6.82 (95% CI: 5.12 to 9.07). After splitting relatives into specific types, the adjusted RRs were significant for parents (RR: 5.32; 95% CI: 3.85 to 7.36), offspring (RR: 8.48; 95% CI: 5.85 to 12.29), and siblings (RR: 18.83; 95% CI: 8.64 to 41.07). There were no AD events in individuals with an affected twin. In contrast, familial aggregation of AD was not seen in individuals with affected spouses without genetic similarities (RR: 1.23; 95% CI: 0.59 to 2.55).
The RR for other diseases associated with AD in individuals with affected relatives
Table 3 lists the RRs of other diseases in first-degree relatives of individuals with AD compared to the general population. The results showed that first-degree relatives were associated with an adjusted RR of 31.92 (95% CI: 26.19 to 38.90) for Marfan disease 5.88 (95% CI: 3.62 to 9.53) for bicuspid aortic valve, and 1.19 (95% CI: 1.14 to 1.24) for hypertension. There were no increased risks of Turner or Ehlers-Danlos syndrome and inflammatory diseases in the individuals with affected relatives.
Nonsyndromic familial AD (sensitivity analysis)
Table 4 lists the results of sensitivity analysis. After excluding individuals with Marfan syndrome, bicuspid aortic valve, and Marfan syndrome or bicuspid aortic valve, a family history of AD was associated with RRs of 5.29 (95% CI: 3.82 to 7.31), 6.32 (95% CI: 4.71 to 8.47), and 6.56 (4.92 to 8.77) for AD, respectively. The results remained the same in subgroups of parents, offspring, and siblings.
Familial resemblance and heritability of AD
Using a liability threshold mode, we estimated that the accountability for the phenotypic variance of AD was 57.0% for genetic factors (heritability), 3.1% for shared environmental factors, and 40.0% for nonshared environmental factors. Given the parameters previously estimated, the probability that a patient had sporadic AD was 84.0%.
Age at a diagnosis of AD in patients with a family history of AD
Figure 2 demonstrates the distribution of the age at a diagnosis of AD in those with and without a family history of AD. The age of AD in the patients with a family history was younger than in those without a family history (51.52 years vs. 59.72 years).
Late outcomes in the patients with AD with a family history of AD
A total of 93 and 12,633 patients with a new diagnosis of AD with and without a first-degree relative affected by AD between 2000 and 2015, respectively, were identified. Among these patients, the proportion of positive recording of first-degree relatives with dissection showed a borderline increase across the study period. By contrast, the prevalence of known risk factors was significantly decreased across the study period (Supplemental Table 5).
After matching, 894 matched control subjects were included in the analysis. No substantial difference in baseline characteristics was noted between the 2 groups (Supplemental Table 1). Among the 987 matched patients, 143 (14.5%) died during the index AD admission. No significant difference in in-hospital mortality rate was observed between patients with and without first-degree relatives with dissection (12.9% vs. 14.7%; data not shown). The mean follow-up times were 3.00 and 3.71 years in the AD patients with or without a family history, respectively. In terms of surgical details, patients with a family history of AD may have significantly higher risks of receiving aortic root replacement in type A aortic dissection (Supplemental Table 6).
The hazard ratio of all-cause mortality in the patients who had a first-degree relative with AD was 0.98 (95% CI: 0.70 to 1.37). The subdistribution hazard ratios for the patients with a family history were 0.72 (95% CI: 0.40 to 1.32) for MACCE, 1.40 (95% CI: 1.12 to 1.76) for later aortic surgery, and 0.96 (95% CI: 0.77 to 1.21) for readmission due to any cause (Figure 3, Table 5).
AD is an acute and life-threatening disease associated with a high mortality rate. However, most patients with AD are asymptomatic before onset, and therefore, identifying the risk factors for AD followed by close surveillance and aggressive control of the risk factors or even aggressive treatment in high-risk populations may decrease the risk of AD. To the best of our knowledge, this is the first population-based family study to report that a family history of AD is a strong risk factor for this disease, regardless of the coexistence of genetic disorders associated with aortic disease. In addition, we demonstrated a 6-fold higher risk in the individuals who had first-degree relatives with AD compared with the general population and that genetic relatedness was associated with the magnitude of the risk of AD. Furthermore, we estimated the disease heritability of AD to be 57% and that nonshared environmental factors also contributed to 40% of the cases of AD. Finally, a family history in the patients with AD was associated with a higher risk of aortic surgery and a younger age at onset of AD. The findings of this study may be valuable when counseling families with affected members, and may be applicable for clinical screening including image studies and genetic tests in a subject with an affected family member with AD. This may allow for more aggressive treatment such as preventive aortic replacement to avoid the adverse outcomes of AD.
The risk factors for AD can basically be classified into 2 categories (13): 1) increased aortic wall stress in which the most common condition is hypertension; and 2) genetic risks associated with medial degeneration including Marfan syndrome, congenital bicuspid aortic valve, Loeys-Dietz syndrome, Ehlers-Danlos syndrome, and Turner syndrome (3,13), with the associated genes including FBN1, EFEMP2, TGFBR1, TGFBR2, TGFB3, ACTA2, COL3A1, and SMAD3 (14–18). However, the altered genes can lead to aortic disease without syndromic features, termed nonsyndromic familial thoracic aortic aneurysm or dissections. Our results also demonstrated that family history was a strong predictor of AD after excluding all known associated syndromes related to aortic disease. There are many previous studies regarding the management of syndromic aortic diseases such as Marfan syndrome to prevent AD (19,20). However, the absence of syndromic features can cause clinicians to overlook the potential risk of aortic emergency in individuals with a family history of AD with genetic alterations associated with aortic disease. Moreover, few previous studies have addressed screening, prevention, and management in these patients. Therefore, further genetic studies and specific management protocols for individuals with a family history of AD are warranted.
Previous familial studies of patients referred for repair of thoracic aortic dissection without a known genetic syndrome have indicated that nearly 20% of these patients have a first-degree relative with dissection, suggesting familial clustering without specific identifiable genetic defect syndromes (21,22). Albornoz et al. (23) reported that thoracic aneurysm and dissection are common familial diseases with a younger age at onset. In addition, a recent study by Ma et al. (7) demonstrated that a positive family history of AD dramatically increased the risk of dissection in family members, with an annual probability of AD per first-degree relative of 2.77 times higher in patients with a positive family history compared with those with a negative family history (7). However, their study was limited by the small sample size and single-center cohort study. In addition, one-third of the patients with a positive family history were syndromic patients, and their results may not represent the risk in nonsyndromic patients. In this study, we demonstrated a 6-fold increased risk of AD in the first-degree relatives of the AD patients after taking associated syndromes related to AD into consideration. To the best of our knowledge, this is the first study to show that a family history of AD is associated with adverse outcomes with a higher risk of later aortic surgery.
There are several strengths to this study. First, we used systematic methods to ascertain first-degree relatives and identify kinship, which allowed for precise estimations of the prevalence and RRs of AD with minimal selection bias. Second, consistent case definitions for individuals at risk and their relatives and complete identification of AD in a large population of Taiwanese patients ensured the absence of information bias. Finally, the nature of prospectively recorded data in the NHIRD for the construction of family relationships and ascertainment of socioeconomic information minimized the errors associated with self-reporting and recall bias.
With regard to the clinical perspectives of this study, given the high mortality rate associated with AD and that it can remain asymptomatic until it occurs, early consultations for aortic disease or genetic disorders, aggressive management of hypertension and smoking cessation, and routine screening with computed tomography scans for individuals with a family history of AD are recommended (24). Moreover, screening should be conducted from a young age because the onset of AD occurs in affected family members at a younger age than in patients with no family history based on our results. Prophylactic resection of ascending aortic aneurysms with a lower trigger of the aortic size in patients with a positive family history may be reasonable for this life-threatening disease (25). Furthermore, extensive surgical treatment during the first onset in AD patients with a family history may be needed due to the higher risk of later aortic surgery in this population. To further apply our findings to clinical perspectives, studies comparing the effect of different treatment strategies including medications and interventions for patients with a family history of AD are warranted.
First, the diagnoses of AD and other covariates were based on ICD-9-CM codes, and misclassification of the diagnosis and coding errors could have occurred. However, we performed internal validation to verify the accuracy of the diagnosis of AD. In addition, the accuracy of the diagnosis codes of other major cardiovascular diseases has been validated in previous studies, and the NHIRD appears to be a valid resource for population research in cardiovascular diseases (26). Second, the NHIRD does not include data regarding anatomic imaging such as CT scans. CT of the aorta is the gold standard to diagnose AD, and it provides detailed data for the classification of all aortic diseases and extension of dissected segments of the aorta. This lack of imaging data, which would have indicated the severity or subtype of AD, could have had an effect on the outcomes. Notably, we could not identify the type of dissection in patients who did not undergo surgery using the Taiwan NHIRD. However, the NHI program has strict regulations regarding examinations and treatment reimbursements through comprehensive reviews of all medical records including imaging studies and laboratory data, and reimbursements are only granted to patients who are indicated for specific examinations or treatments due to an accurate diagnosis. Thus, this may have limited the potential bias of unavailable data. Last, this was a population-based study conducted in Asia, and the incidence and characteristics of AD may differ between different populations. Therefore, our results may not be directly generalizable to other patient populations and environments. Despite these limitations, we believe that our results provide a valuable contribution to the knowledge gap in family studies on AD.
This large Taiwanese population-based family study demonstrated family aggregation in AD. A family history was a strong risk factor for AD with or without syndromes related to aortic disease, and both genetic and environmental factors contributed to the overall risk of AD. Furthermore, a family history of AD in the patients with AD was associated with a higher risk of later aortic surgery. We recommend routine imaging and genetic screening of individuals with a family history of AD. Close surveillance and aggressive control of risk factors or even preventive aortic replacement in high-risk populations may decrease the frequency of dissection and possibly reduce mortality through an earlier preclinical diagnosis in affected families.
COMPETENCY IN MEDICAL KNOWLEDGE: A family history of AD is a powerful risk factor for AD in patients with or without syndromic aortic disease. Both genetic and environmental factors contribute to overall risk.
TRANSLATIONAL OUTLOOK: Future research should focus on defining the optimum screening and imaging surveillance strategies and the aortic diameter at which prophylactic surgical intervention is advantageous for patients with a family history of AD.
The authors thank and acknowledge the support of the Maintenance Project of the Center for Big Data Analytics and Statistics (Grant CLRPG3D0045) at Chang Gung Memorial Hospital for statistical consultation and data analysis. The authors also thank Alfred Hsing-Fen Lin and Hui-Tzu Tu for their assistance with the statistical analysis.
↵∗ Drs. Chen and Kuo contributed equally to this work.
This work was supported by grants from Chang Gung Memorial Hospital, Taiwan (CMRPG3H1511 and CMRPG3J0661 to Dr. Chen) (CLRPG3H0012 to Dr. Kuo) and the Ministry of Science and Technology (Most 108-2314-B-182A-141 to Dr. Chen) (MOST106-2314-182A-115-MY3 to Dr. Chu). This study was based on data from the National Health Insurance Research Database provided by the National Health Insurance Administration, Ministry of Health and Welfare of Taiwan, and managed by the National Health Research Institutes of Taiwan. However, the interpretation and conclusions contained in this paper only represent the authors. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC author instructions page.
- Abbreviations and Acronyms
- aortic dissection
- confidence interval
- International Classification of Diseases-9th Revision-Clinical Modification
- major adverse cardiac and cerebral events
- National Health Insurance
- National Health Insurance Research Database
- relative risk
- Received March 9, 2020.
- Revision received June 30, 2020.
- Accepted July 7, 2020.
- 2020 The Authors
- Pape L.A.,
- Awais M.,
- Woznicki E.M.,
- et al.
- Howard D.P.,
- Banerjee A.,
- Fairhead J.F.,
- Perkins J.,
- Silver L.E.,
- Rothwell P.M.
- Olsson C.,
- Thelin S.,
- Stahle E.,
- Ekbom A.,
- Granath F.
- Yeh T.Y.,
- Chen C.Y.,
- Huang J.W.,
- Chiu C.C.,
- Lai W.T.,
- Huang Y.B.
- Ma W.G.,
- Chou A.S.,
- Mok S.C.M.,
- et al.
- Zimmerman K.P.,
- Oderich G.,
- Pochettino A.,
- et al.
- Sidloff D.,
- Choke E.,
- Stather P.,
- Bown M.,
- Thompson J.,
- Sayers R.
- Goldfinger J.Z.,
- Halperin J.L.,
- Marin M.L.,
- Stewart A.S.,
- Eagle K.A.,
- Fuster V.
- Bertoli-Avella A.M.,
- Gillis E.,
- Morisaki H.,
- et al.
- Martin M.,
- Rodriguez I.,
- Palacin M.,
- Rios-Gomez E.,
- Coto E.
- Isselbacher E.M.,
- Lino Cardenas C.L.,
- Lindsay M.E.
- Attias D.,
- Stheneur C.,
- Roy C.,
- et al.
- Hasham S.N.,
- Willing M.C.,
- Guo D.C.,
- et al.
- Teixido-Tura G.,
- Forteza A.,
- Rodriguez-Palomares J.,
- et al.
- Robertson E.N.,
- van der Linde D.,
- Sherrah A.G.,
- et al.
- Verhagen J.M.A.,
- Kempers M.,
- Cozijnsen L.,
- et al.
- Chou A.S.,
- Ma W.G.,
- Mok S.C.,
- et al.