Ethical issues in susceptibility genetic testing for late‐onset neurodegenerative diseases

Genome‐wide association studies have revolutionized our understanding of the genetic architecture of complex traits and diseases over the last decade. This knowledge is enabling clinicians, researchers, and direct‐to‐consumer genetics companies to conduct disease susceptibility testing based on powerful methods such as polygenic risk scoring. However, these technologies raise a set of complex ethical, legal, social, and policy considerations. Here we review and discuss a series of ethical dilemmas associated with susceptibility genetic testing for the two most common late‐onset neurodegenerative diseases, Alzheimer's and Parkinson's disease, including testing in asymptomatic individuals. Among others, these include informed consent, disclosure of results and unexpected findings, mandatory screening, privacy and confidentiality, and stigma and genetic discrimination. Importantly, appropriate counseling is a deciding factor for the ethical soundness of genetic testing, which poses a challenge for the regulation of these tests and the training of healthcare professionals. As genetic knowledge about these diseases continues growing and genetic testing becomes more widespread, it is increasingly important to raise awareness among researchers, medical practitioners, genetic counselors, and decision makers about the ethical, legal, and social issues associated with genetic testing for polygenic diseases.

increased susceptibility, and the ability to convey information about their relatives, creates a set of complex ethical, legal and social issues.
Here, we aim to contribute to the debate, discussing issues related to the use of susceptibility genetic testing to assess genetic predisposition to late-onset neurodegenerative diseases, including its use in asymptomatic individuals, and its impact on different social spheres.

| LATE-ONSET NEURODEGENERATIVE DISEASE IN THE GENOMICS ERA
This first section briefly details the underlying mechanisms, pathology and genetic components of each of the two most common neurodegenerative disorders: AD and PD. In addition, we detail their incidence and prevalence, and the resulting global economic impact. Finally, we outline how genomic technologies have and will continue impacting the clinical handling of AD and PD.

| Alzheimer's disease
AD is a chronic and progressive disorder which leads to cognitive impairment and neuropsychiatric abnormalities. It is characterized by the accumulation of insoluble forms of the amyloid beta peptide (Aβ) and aggregation of tau proteins in wide areas of the cerebral cortex and the hippocampus.
AD is considered the most common neurodegenerative disorder, with an estimated prevalence of 10-30% in the global population over 65 years of age (Masters et al., 2015). While evidence suggests that sporadic AD affects all of the world's geographical populations in similar rates, high-income countries currently account for about half of the global prevalence (Alzheimer's Disease International [ADI], 2010).
Importantly, dementia seems to be more prevalent in women than men (Masters et al., 2015). Alzheimer's can dramatically shorten life expectancy: median survival with AD is approximately 7.1 years after onset (Duthey, 2013). It is difficult to determine the independent contribution of Alzheimer's to mortality, as people suffering from AD often develop comorbid health conditions, related or not to the dementia process itself. Based on predicted increases in the global prevalence of dementia, partly due to the aging population, costs of dementia in 2030 are expected to be 85% higher (Alzheimer's Disease International [ADI], 2010). Just in Europe, estimates of the future cost of AD predict a rise of 43% (Duthey, 2013). However, rising costs in developing countries are expected to be much sharper, because increased life-expectancy related to development will result in higher levels of AD. For example, the prevalence of dementia in Latin America is expected to rise by 435% by 2050 in relation to 2010 (Alzheimer's Disease International [ADI], 2010).
The vast majority of AD patients have the sporadic form of the disease, characterized by late onset; a small proportion of patients present early-onset inherited forms. Numerous genetic factors have been identified which determine the development of both sporadic and inherited Alzheimer's. Familial AD follows an autosomal dominant pattern of inheritance, involving three main genes implicated in the genesis of Aβ: presenilin 1 (PSEN1), presenilin 2 (PSEN2), and amyloid precursor protein (Verlinsky et al., 2002). Late-onset AD is much more genetically complex. The most common risk allele is found within the gene coding for apolipoprotein E (APOE), which is a major determinant of age of onset (Escott-Price, Shoai, Pither, Williams, & Hardy, 2017). Depending on its isoform (APOE2, APOE3, or APO4), the risk of developing Alzheimer's increases or decreases. For example, one allele with the APOE4 polymorphism imparts a threefold increase in risk, whereas APOE2 is considered a protective factor (Masters et al., 2015). Other genetic determinants are involved in Aβ clearance pathways. Overall, approximately 24 common loci have been confirmed as susceptibility factors for sporadic AD, and several more show some evidence of association (Escott-Price et al., 2017;Kunkle et al., 2018). Importantly, the majority of known genetic contributing factors were discovered amongst populations of European ancestry, as is the case for many diseases. However, the rate of decline has not consistently been shown to be linked to APOE4 in nonwhite populations (Marden et al., 2016) and also differs across gender (Goldman et al., 2011).
Nongenetic risk factors for AD have also been identified. For example, diabetes, depression, smoking, and low levels of education are associated with increased AD risk (Masters et al., 2015). This highlights the relevance of environmental factors on the development and progress of complex disease.

| Parkinson's disease
PD is a slowly progressing neurodegenerative disorder known for affecting movement. The most common motor symptoms include tremors, rigidity and bradykinesia (Farlow, Pankratz, Wojcieszek, & Fouroud, 2004). Nonmotor symptoms and comorbid conditions are also prevalent, such as cognitive decline, depression, dysphagia, and symptomatic postural hypotension (von Campenhausen et al., 2011). PD is characterized by the loss of dopaminergic neurons in the substantia nigra, partly due to the accumulation of the α-synuclein protein (Poewe et al., 2017).
Parkinson's is the second most prevalent late-onset neurodegenerative disorder worldwide, with an overall prevalence of 1-3% of the over-60 population (de Lau & Breteler, 2006). Importantly, the number of PD patients is expected to double between 2005 and 2030 (Poewe et al., 2017). In contrast to AD, PD is more prevalent in men than women (Poewe et al., 2017). Evidence suggests that race and ethnicity might have an important effect on disease incidence (Poewe et al., 2017). However, it is currently difficult to discern whether these disparities are due to purely biological causes or sociodemographic factors related to health (Roberts & Uhlmann, 2013).
PD is considered one of the costliest diseases in some countries (von Campenhausen et al., 2011), as it represents a high economic burden on the patients, their close family, and society as a whole. In the United Kingdom, the expenditure on PD is estimated to be between 449 million and 3.3 billion pounds every year (Findley, 2007). One important characteristic of a neurodegenerative disease such as PD is that the cost escalates in proportion to the progression of the disease. Direct costs, such as inpatient care and medication, increase with the development of comorbid conditions and the need for more healthcare resources. Similarly, the loss of independence of the patients results in increased indirect costs, shown by decreased productivity.
PD exists in several forms. Familial forms of PD are monogenic, inherited in autosomal dominant, autosomal recessive, or even Xlinked patterns (Farlow et al., 2004). Sporadic or idiopathic forms of PD are more common and caused by the cumulative action of multiple genes. Studies confirm that some genetic variants responsible for familial PD also contribute to risk for sporadic PD (Poewe et al., 2017). For example, mutations in the LRKK2 gene represent a frequent cause of all forms of Parkinson's, accounting for 1% of sporadic cases and 4% of familial cases (Kalia & Lang, 2015). Risk-altering polymorphisms for sporadic disease can be found in PARK loci (initially discovered in familial PD), along with several other genes (Ibanez et al., 2017). Mutations in GBA, for example, are common genetic risk factors (Poewe et al., 2017). Overall, as many as 92 independent genome-wide significant signals have been identified (Nalls et al., 2018). Genetic factors contributing to disease are mainly associated with pathways related to α-synuclein proteostasis and degradation, mitochondrial function and immune regulation (Poewe et al., 2017).
In addition to genetic risk factors, environmental and lifestyle influences over disease development and progression are also relevant. For example, exposure to pollutants, physical activity, vitamin D levels, years of education, smoking status, and urate levels might influence overall risk for PD (Kalia & Lang, 2015). Arguably, this might be mediated by environmentally-triggered epigenetic changes in the nervous system (Poewe et al., 2017).

| Impact of new technologies on the clinical handling of AD and PD
Accurate clinical diagnosis and treatment of neurodegenerative diseases such as AD and PD is a big challenge, given the complexity of said conditions. This is due in part to the difficulty of measuring and integrating all the factors that influence the expression of a disease across a patient's lifespan, as well as differentiating these diseases from other disorders and common comorbidities. For example, over one third of clinically diagnosed patients with AD in a large clinical study were later found to be misdiagnosed (Masters et al., 2015). Similarly, a systematic review of accuracy of clinical diagnosis of Parkinson's estimated a pooled diagnostic accuracy of 80.6% in 11 studies over the past 25 years (Rizzo et al., 2016). Importantly, accuracy has not significantly improved in recent times, particularly diagnosis during the early stages of disease.
New technologies, such as neuroimaging and biomarkers, have emerged as essential tools to aid in the diagnosis of neurodegenerative disorders. Several biomarkers in cerebrospinal fluid (CSF), for instance, are promising diagnostic tools for AD, and could potentially enable diagnosis even before the development of clinical symptoms (Masters et al., 2015). Further, CSF-based diagnostic tests have also been suggested for PD (Poewe et al., 2017).
Genetic testing is another tool used to estimate susceptibility and aid diagnosis for several diseases, as well as an important means for research. Genetic tests are sometimes used to assess multiple diseases at once, instead of being targeted for a specific condition. Depending on the disease of interest and the purpose of the test, a couple of main questions must be addressed to establish its usefulness: which gene (or genes) should be analyzed, and whether the test has sufficient sensitivity and specificity (Cozaru, Aşchie, Mitroi, Poin areanu, & Gorduza, 2016). In the case of complex diseases such as AD and PD, genetic testing often yields ambiguous results. This is partly because the effect of multiple genetic mutations can be altered by factors such as penetrance, expressivity, pleiotropy, epistasis or other known or unknown regulatory interactions (Roberts & Uhlmann, 2013). In addition, even though current technologies are highly improved, there are still technical limitations which make results uncertain. For example, single-nucleotide polymorphism microarrays usually encompass common variants, but a considerable component of disease risk may come from rare or new variants. In fact, genetic testing for complex diseases assesses relative susceptibility rather than being a predictive method, meaning that it is mostly probabilistic and highly contingent (Arribas-Ayllon, 2011).
Inheritance and susceptibility for AD and PD is difficult to predict, especially when only a single variant is assessed. Polygenic risk scores summarize genotype data to inform about the genetic architecture of a complex trait or a disease (Lewis & Vassos, 2017). Each score is based on known risk and protective alleles in genome-wide data, and it is calculated from GWAS summary statistics for the trait of interest (Lewis & Vassos, 2017). Discovery GWAS have estimated effects and effect sizes for hundreds of thousands of common variants, in addition to genome-wide significant loci. These additional polymorphisms also account for a fraction of the heritability of the disease and act in an additive manner (Escott-Price et al., 2017;Poewe et al., 2017). PRS is a powerful tool, precisely because it is able to account for the cumulative effects of genome-wide common variants (Ibanez et al., 2017).
Indeed, it has been used to successfully identify at-risk individuals and estimate their overall risk.
The actual maximum prediction accuracy of PRS for AD has been estimated to be of 74%, compared to the 82% theoretical maximum that can be achieved by predictors of risk based on genotype data (Escott-Price et al., 2017). Mild cognitive impairment (MCI) is an intermediate stage in the progression from expected cognitive decline to dementia. Recently, an AD PRS was used to identify MCI in subjects in their 50s (Logue et al., 2018). Other AD PRSs, both including and excluding APOE, were able to predict the level and rate of memory decline (Marden et al., 2016). The average PRS in Parkinson's patients is indirectly proportional to age at onset (Escott-Price et al., 2015). A PD PRS has also been associated with motor and cognitive decline in patients (Paul, Schulz, Bronstein, Lill, & Ritz, 2018), but not with levels of α-synuclein in CSF (Ibanez et al., 2017).
Ideally, PRS could capture the combined effects of genome-wide genetic variation in a single value, applicable in a clinical setting as an indicator of genetic susceptibility. Importantly, PRSs remain constant throughout life, which enables prediction from any age. Early identification of individuals at increased risk is imperative, because the pathological processes of AD and PD begin long before symptom onset. For example, abnormalities in biomarker measurements are seen as early as 20 years before AD onset (Masters et al., 2015). Additionally, PRSs could also be useful to assess several aspects of disease progression, as shown by some of the studies described above, which could help define disease prognosis and treatment options. Furthermore, a better understanding of the relationship between genotype and clinical phenotypes could also inform patient selection for clinical trials and other research endeavors to accelerate the development of precision therapies (Tan & Jankovic, 2006).
It must be noted that, even though PRSs are clearly attractive for clinical implementation, their accuracy may not yet be significant enough. PRSs will potentially become more informative with larger GWAS and knowledge of more variables (Ibanez et al., 2017). Conversely, other studies in AD indicate the contributions of any new findings are likely to be small and thus unnecessary for the overall genetic prediction of disease (Escott-Price et al., 2017). Either way, the accuracy of PRS may be improved by their combination with clinical and environmental variables (Lewis & Vassos, 2017), as well as more varied data sets to account for different ethnicities.
Neither PRSs nor any other forms of genetic testing are currently part of the routine diagnostic process for AD or PD. However, this might very well change in a not so distant future as the field continues evolving. The correct implementation of these tools will largely depend on the available legal safeguards worldwide. For the purpose of the ethical discussion in this manuscript, "positive results" and "risk-positive" will refer to high predisposition for a disease according to a genetic test.

| ETHICAL CONSIDERATIONS ASSOCIATED WITH SUSCEPTIBILITY GENETIC TESTING FOR NEURODEGENERATIVE DISEASES
The ethical debate about issues associated with genetic testing covers everything from the research which enables the creation of the technology to what uses are made of its results. Some relevant questions in the context of neurodegenerative disorders include: Under which conditions should patients, including asymptomatic individuals, have access to genetic testing? Could genetic testing for late onset complex conditions be mandatory under any circumstance? Should genetic test results ever be released to third parties, including the user's family?
How can individuals be protected from unfair treatment because of their genetic status? These and other questions can be approached from four basic ethical principles: autonomy, nonmaleficence, beneficence, and justice (Beauchamp & Childress, 2001). We shall discuss the ethical issues related to genetic susceptibility testing for AD and PD in respect to their corresponding principles (Table 1).

| Autonomy
Individuals are autonomous to the extent that they are able to make willing and reasoned choices and take action without outside control.
In the context of genetic testing, respect for autonomy can be summed up in the right of individuals to make informed decisions on whether they wish to be tested, and what they wish to do with the results of said tests. and clinical trials prior to or following a genetic test, with the addition of specific requirements to each case.

| Informed consent
Meeting even just the minimum requirements for an informed consent form proves difficult with many recent technologies, including the use of susceptibility testing in which results may be ambiguous. It is important to note that both providers and clinicians who refer patients to testing, have the duty to find alternatives themselves when faced with obstacles applying the traditional mechanisms of consent; the right to be informed must be respected first and foremost (CAGR, 1994). Collective experiences on what users of these new technologies want and need to know will be fundamental to legally safeguard their autonomy (Roche & Berg, 2015).
It is important to emphasize that there exists an intrinsic power imbalance in situations where genetic testing is involved. Participants in a clinical trial, for example, have less technical knowledge than the researchers, placing them in a position of vulnerability. This imbalance may be aggravated when taking into account the gender, ethnicity, socioeconomic status, or other social conditions of a subject. AD and PD patients in particular are vulnerable given their dependency on others. Similarly, asymptomatic individuals wishing to test their susceptibility to AD or PD might do so because of known family history, placing them in a position of emotional and mental distress even before the test. Furthermore, some patients likely to be tested for susceptibility to a neurodegenerative disease may already be suffering from cognitive difficulties, compromising their full understanding of the test (Roberts & Uhlmann, 2013). This kind of compromised auton- Underage individuals are considered agents with reduced autonomy, and there are further ethical dilemmas specific to genetic testing of children. However, as susceptibility testing for AD and PD renders only probabilistic results and there are currently no cures of preventive measures available, we argue that there are currently no benefits to a child knowing their genetic susceptibility for AD or PD (Goldman et al., 2011). Furthermore, results might trigger anxiety or depression in the child, or even cause a sense of guilt in the parents due to the hereditary component (Cozaru et al., 2016). Therefore, genetic testing for AD and PD should be avoided for underage individuals. This point should be reconsidered in the future, when and if more treatment options are available, or when predictive technologies are more accurate.

| Disclosure of results and unexpected findings
Individuals have the right to decide what they want or do not want to know about the results of a genetic test. This right is protected by national and international legislation around the world (Pont-Sunyer et al., 2015). In the case of genetic tests that provide information about susceptibility for multiple disorders at once, the assessment of the risks and benefits of learning a result could differ between those relevant to the original purpose and secondary findings. Furthermore, the decision to know would probably depend on the nature of the secondary finding (Roche & Berg, 2015).
Due to the complex nature of AD and PD, we would propose the following recommendations regarding disclosure: First, individuals should be made aware of their right not to know; even when they have specifically decided to take a susceptibility test for AD or PD, they retain the ability to change their decision of knowing. Second, given the lack of actionability for such diseases, we argue there is currently no justification to override a preference for nondisclosure.
When and if this situation changes in the future, we would argue that, since health is defined as more than just physical well-being, an individual's informed analysis of what constitutes their own well-being should prevail either way. Third, in order to avoid issues with unexpected findings, we propose that these diseases be considered exclusively in specifically targeted tests. Alternatively, we argue that these diseases should, for now, be excluded "by default" from multiple disorder susceptibility tests. Individuals wishing to know their risk status should explicitly request to "opt-in", with the condition that they are given full information of what these particular diseases entail, ideally accompanied by genetic counseling.
Nondisclosure can work differently in a research setting (i.e., clinical trials), in contrast to patients being referred to a genetic Autonomy Informed consent • Users should be protected against abuse, and their personal integrity respected.
• Genetic testing for AD and PD should be avoided for underage individuals.

Disclosure of results and unexpected findings
• Individuals should be made aware of their right not to know.
• The preference for nondisclosure should prevail, given the lack of actionability for AD and PD. This includes research settings as well. • AD and PD should be excluded by default from multiple disorder susceptibility tests.

Mandatory population screening
• Screening programs targeting AD and PD are only justified to direct individuals to early action; however, such measures are not currently readily available. • Prenatal screening programs can be viewed as a form of discrimination, given the lack of actionability.
• Collective benefits should be weighed against individual autonomy. However, population screenings are not necessary to inform public health decisions in a present time.
Privacy and confidentiality • There is no clear benefit of disclosure of probabilistic results to family members.
• Data protection measures will depend on the degree of sensitivity determined for probabilistic disorders.

Nonmaleficence and beneficence
Stigma and genetic discrimination • Legal safeguards against discrimination must be in place, regardless of whether or not individuals susceptible to these diseases are currently considered vulnerable to suffer it.
Genetic counseling • Genetic counseling is essential to preserve autonomy in the informed consent process.
• Appropriate counseling must be provided to aid the interpretation of results.
• Counselors should be up-to-date on continuously evolving knowledge regarding prognosis and treatment options Justice Diversity in clinical and biomedical research • Inclusion of populations of non-European ancestry in GWAS is necessary to improve predictive capability and to ensure widespread applicability of resulting technologies.
Availability and accessibility • Stricter controls need to be put in place to regulate companies offering direct-to-consumer genetic tests.
• Information regarding genetics, statistics and medicine should be made available to the general public.
• Training of healthcare professionals should also include education in effective communication.
Allocation of resources • Due to limited resources, and given the lack of treatment options or immediate burdens, public expenditure on susceptibility genetic testing should not be considered a priority at this time.
AD = Alzheimer's disease; PD = Parkinson's disease; GWAS = genome-wide association study. test by their clinician or choosing to take one privately. Importantly, new forms of sequencing make it difficult to clearly define "unexpected" or "incidental" findings in these contexts (Christenhusz, Devriendt, & Dierickx, 2013), in turn obstructing disclosure guidelines.
A relevant situation regarding disclosure in research is the case of prevention trials. This kind of trial consists of studying subjects who present risk markers but who show no symptoms of disease. For example, patients with REM sleep behavior disorder (RBD) are extremely interesting subjects for Parkinson's prevention research (Postuma, Gagnon, Bertrand, Génier Marchand, & Montplaisir, 2015).
A pool of RBD patients might be tested for genetic susceptibility and included in a trial. The dilemma arises from the fact that people who wish to participate in a trial do not necessarily want information about their genetic status.
Enrollment protocols in prevention trials are said to be blinded when they accommodate the choice of not knowing (Kim, Karlawish, & Berkman, 2015). For example, they include risk-negative subjects in the study so that participation does not automatically equate to being a carrier. In contrast, transparent enrollment protocols do not consider the option of nondisclosure, and most trials of these type only involve risk-positive subjects (Kim et al., 2015). Some critics of transparent enrollment pose that such protocols are coercive, arguing that blinded enrollment is a requisite for autonomy. Nevertheless, others argue that blinded enrollment is not ethically required, and point out that transparent protocols could in fact be of greater benefit, as it spares risk-negative subjects from the burden of participation without any personal gain (Kim et al., 2015). This last argument might hold more weight when at-risk individuals are defined solely by genetic susceptibility, instead of presenting other markers. That is, looking for PD riskpositive subjects in a general pool of individuals instead of RBD patients, for example.
Regarding disclosure in research settings, we argue that participation in research trials can be an altruistic act and individuals should be allowed to participate, whether or not there is a personal gain. It follows that their decision not to know should be respected, especially for unactionable disorders such as AD and PD, considering the trial might not work or they might be placed as a negative control. In the particular case of prevention trials, we find no justification for transparent enrollment protocols, as they disregard these principles.  While it is recognized that the decision to participate in any genetic test must be voluntary, some decision-makers might argue in favor of a mandatory testing in some cases. This is due to the fact that it is a minimally invasive procedure whose results could contribute to better resource allocation and more efficient public health planning.  (Lewis & Vassos, 2017). Importantly, all these measures should be readily available to all the population before a mandatory screening takes place and individuals receive feedback regarding their genetic status. Given that this is not presently true for AD and PD-current data regarding risk-reducing behaviors is uncertain and evolving, and treatment is unavailable-we would find it hard to argue in favor of screenings.
A particular type of mandatory screening could be prenatal susceptibility testing. Testing for AD and PD is more technically challenging than for other conditions. Therefore, prenatal testing is usually strongly discouraged (Tan & Jankovic, 2006), especially if the user intends to carry a pregnancy to term regardless of the results (Goldman et al., 2011). In fact, requests for prenatal diagnosis of lateonset diseases are uncommon (Farlow et al., 2004), which indicates that a mandatory program would not be well-received by the public.
However, these views might change as technologies become more reliable. For example, preimplantation genetic diagnosis has already been used to ensure the implantation of familial AD risk-free embryos (Verlinsky et al., 2002). Importantly, the perception and ethical soundness would depend on whether a screening program was widespread or specifically targeted to couples with known genetic risk or family history of neurodegenerative disorders. That is, if there was additional evidence to support a prenatal diagnosis or not.
The issue of prenatal screening is further complicated by the possibility of abortion. As there are currently no preventive measures available for AD and PD, the question arises if prenatal testing is a way to avoid the birth of particular individuals with specific risk alleles, rather than to prevent the conditions themselves (CAGR, 1994). Medical professionals have many different perspectives regarding genetic testing which could influence the termination of a pregnancy (Farlow et al., 2004). Regardless of personal beliefs, in countries where the legal right to elective abortion has been granted, selective abortion is a viable option.
It is important to note that legal rights do not resolve the related ethical dilemmas. Selective abortion in these cases is based on maternal considerations of the future child's quality of life (Post, 1994).
While it is exclusively a woman's choice to terminate a pregnancy, Even with the safeguard of information, it is still difficult to justify a mandatory screening program, regardless of whether it is planned for adults or as prenatal testing. This is partly due to the fact that even the most advanced genetic diagnostic techniques like PRS are insufficient to determine whether or not an individual will definitely develop AD or PD. Even though it is undeniable that the onus of caring for patients with AD and PD will greatly affect future generations, other currently available information should, in our opinion, be enough to push decision-makers to preventive action. When the time comes, it will be important to note that some societies value collective over individual benefits because of their cultural context, and this should be taken into account when weighing the ethical merit of such measures.

| Privacy and confidentiality
According to international declarations, a person's biological information, including human genetic data resulting from a test, must be held confidential in the conditions set by law and should not be disclosed or made accessible to third parties (i.e., employers, insurance companies, educational institutions, or the family) (UNESCO, 2005). Additionally, individuals are granted the right of respect for their private life, in particular to protection of their personal data derived from a genetic test (AP-GTHP, 2008). The only exception to these rights relates to cases when the person concerned has provided prior, free, informed and express consent for disclosure.
Challenging these ideas, some argue that genetic testing not only reveals the risk of an individual, but also indirectly that of blood relatives, raising the question of how personal genetic information actually is. A positive result can have considerable consequences for spouses (i.e., reproductive decisions) and other members of the patient's close social context (e.g., potential caregivers). As such, disclosing genetic susceptibility results might be considered a moral obligation. However, ethical reasoning should be made in the context of the genetic complexity of a specific condition (Arribas-Ayllon, 2011).
In the case of AD and PD, for which genetic tests are largely probabilistic, there is currently no clear benefit of disclosure to family members.
Although all genetic information falls under the umbrella of protection mentioned above, it is worth discussing that, as discussed, AD and PD represent specific ethical issues linked to the probabilistic nature of their genomics. Derived health data from a genetic test in these cases will not necessarily determine the future onset of a neurodegenerative disease, which could make this information less sensitive than, for example, genetic data related to monogenic disorders. Being probabilistic, genetic data related to neurodegenerative diseases remains difficult to regulate, and the doubt also arises of who will decide when the probability has turned into a declared neurodegenerative disease, even before onset, and how that relates to data protection. Importantly, data-sharing approaches should be established in order to encourage joining efforts favoring research and protecting the interests of those affected by the disease, also taking into account the degree of such affection at the same time.
Finally, determining the scope of a right to health privacy within neurodegenerative diseases will prove to be challenging. We have already discussed the potential benefits of inclusion of these disorders in population screenings, thus challenging a right to health privacy in this regard. However, the fact remains that the promise of social longterm benefits might not be sufficient justification for disregarding the individual right to choose, especially given the nature of genetic information. Experience and the increasing debates will most likely direct future guidelines in this matter. Beyond stigmatization, the disclosure of a person's genetic information can lead to discriminatory behavior. Genetic discrimination involves the differential treatment of a person on the grounds of their actual or presumed genetic differences, known either through genetic testing, family history, or even information about the communities they belong to (Otlowski, 2005). Several international legal instruments prohibit any form of discrimination based on genetic heritage in a way that infringes a person's basic human rights and fundamental freedoms, or for purposes that lead to the stigmatization of an individual, a family, a group, or a whole community (AP-GTHP, 2008; Convention on Human Rights and Biomedicine, 1997;UNESCO, 1997;UNESCO, 2003;UNESCO, 2005).

| Nonmaleficence and beneficence
Genetic discrimination causes disadvantages with respect to access to health services, education and employment. In the United States, the Genetic Information Nondiscrimination Act (GINA, 2008) prohibits health insurers from denying coverage or charging higher premiums based solely on a genetic predisposition. It also prohibits employers from using a person's genetic information as a factor in decisions such as hiring, firing, assigning jobs, or any other terms of employment. Finally, it prohibits employers from requesting, requiring, or purchasing genetic information about their employees and their family members (Hudson, 2011). A number of developing countries have also introduced legislative measures to protect citizens against discrimination and stigmatization related to medical conditions and genetic status (WHO, 2006).
Importantly, only a small percentage of people in surveys express concerns over third-party access to their genetic status regarding lateonset neurodegenerative disorders (Neumann et al., 2001). Indeed, it is essential to determine the likelihood for abuse to occur. For example, several events would have to take place for genetic discrimination to occur in the case of a patient who undergoes testing to determine susceptibility to AD or PD. First, the test would have to prove a high degree of genetic susceptibility to the disease for it to be relevant.
Then, a potential employer or insurer would have to ascertain that a test has taken place and to somehow access the information (Brazier & Cave, 2011). For the latter to happen, considering the data protection safeguards internationally regulated, it would most surely have to be the patient the one to disclose such result. Finally, the consequence would have to end up in the patient being turned down for a job, or having insurance refused.
Even though it appears that information about conditions such as AD and PD is perceived as less sensitive, we argue that this might be mostly due to the fact that usual onset does not occur in the economically productive population. Considering that the aging population will lead to overall later retirement, it is possible that AD and PD will increasingly overlap with productive ages. Therefore, these diseases will arguably become more relevant for employability and, consequently, genetically susceptible individuals more vulnerable to discrimination. In the case of discrimination in insurance, current legislation primarily covers health insurance, and not long-term care insurance or other domains which are more relevant to AD and PD (Roberts, Christensen, & Green, 2011). Therefore, even though genetic discrimination seems currently unlikely, we argue that measures must still be taken to protect genetic information related to these diseases because of their overall sensitive nature and probable future consequences.

| The importance of genetic counseling
A great concern of genetic testing is the impact on the individual's self-image. Unlike infectious disease, genetic disease can be considered a part of the patient's intrinsic identity, leading to selfclassification as "defective" (CAGR, 1994 Conversely, not having clear answers is also a source of anxiety (Cozaru et al., 2016). Disclosure of genotyping information can actually represent a benefit for those negative, and even for some riskpositive individuals, as long as some constraints are established (Green et al., 2009). The main and most relevant requirement is the provision of appropriate counseling by trained professionals.
Genetic counseling is essential to preserve autonomy in the informed consent process prior to testing, and it represents the difference between favorable and adverse reactions to the results following the test. In the case of complex diseases like AD and PD, genetic counseling should be done on a case-by-case basis, in order to account for the context of the individual (Farlow et al., 2004). When proper post-test counseling is guaranteed, the risk of further psychological risk is minimized. Research demonstrates that using standardized counseling protocols after an AD susceptibility test ensures effective coping skills (Goldman et al., 2011). Similarly, asymptomatic testing for PD, when done with the appropriate genetic counseling, may be useful to ease the anxiety of an at-risk individual when negative, or help them cope with transient distress when positive (Tan & Jankovic, 2006).
Genetic counseling is also important to avoid unreasoned decisions stemming from misunderstood technical terms, including the meaning of susceptibility itself. For example, the results of genetic testing have been shown to influence a person's finances, reproductive planning, and interpersonal relationships (Cozaru et al., 2016).
Therefore, counselors should convey the limitations of genetic testing for AD and PD with the technologies currently available, as well as the incomplete knowledge regarding protective measures patients might feel encouraged to take (i.e., behavioral changes thought to decrease risk for developing disease). Importantly, genetic risk information with proper counseling might encourage the purchase of longterm insurance. This is particularly relevant for neurodegenerative disorders, because patients are often placed in nursing homes and go through lengthy inpatient stays (Paulsen et al., 2013;Roberts et al., 2011).
It is important to note that genetic counseling becomes challenging in diseases such as AD and PD, in which knowledge is continuously evolving: there are constantly new risk loci being identified, as well as related environmental risk factors. It will become essential to develop complete and up-to-date databases which can serve as reliable references for genetic counselors.

| Justice
The concept of justice encompasses both the access and contribution to genetic technologies. Distributive and contributive justice are relevant to public planning, including questions of fair allocation of limited resources, how to deal with competing needs, and how to ensure the accessibility to technologies. groups, such as AD and PD (Goldman et al., 2011;Marden et al., 2016;Poewe et al., 2017;Roberts & Uhlmann, 2013). Particularly, it hinders knowledge of the interactions between genetic status and certain environmental risks to confer a prognosis (Oh et al., 2015). To date, GWAS of AD and PD have been largely focused on populations of European ancestry, even though it is possible that the same genetic markers may not necessarily show the same effect in other populations. Racial diversity in GWAS could help improve the predictive capability of PRS (Marden et al., 2016), as well as ensure that these technologies can be accurately applied to people of diverse ethnicities. Efforts must be made so that all members of society should equally share in the burden of research (i.e., participating in a clinical study) and equally reap its benefits (i.e., having access to genetic testing).

| Availability and accessibility of genetic tests
Two essential requirements for a just distribution of healthcare are its availability and accessibility. The former refers to the sufficient quantity of goods and services, whereas the latter implies financial and nondiscriminatory access to healthcare. Accessibility also refers to the right to seek, receive and impart health-related information, without impairing the right to confidentiality of others (OHCHR, 2008).
Those with high levels of perceived risk (i.e., because of family history or caring for a patient) are most likely to pursue genetic testing (Wikler, Blendon, & Benson, 2013). Considering that the prevalence of neurodegenerative diseases will continue to rise, and more people will have close proximity to patients within their social environment, it stands to reason that the demand for susceptibility testing for AD and PD will increase. Taking that into account, some might argue that clinical and research settings obstruct, rather than facilitate, both the availability to and accessibility of genetic testing.
The marketing of direct-to-consumer (DTC) tests has emerged as an answer to the general public's desire to knowing as much about their genes and their bodies as they choose, through simple and affordable services (Ramani & Saviane, 2010). DTC tests aid availability by increasing the supply of genetic tests in the market, which in turn helps lower prices to make these services more accessible. However, another key aspect to consider for a fair allocation of healthcare is the quality: What is the scientific and medical suitability of the tests, and whose responsibility it is to ensure it? Companies currently providing results related to AD and PD do so largely without the provision of direct genetic counseling by healthcare professionals, which puts users at risk of psychological distress if they fail to understand the significance-or lack thereof-of the results provided (Roberts et al., 2011). Furthermore, the results do not consider information relative to family history, rare genetic variation and environmental risks (Kaye, 2008). Finally, it is worth asking how up-to-date the information DTC tests use to determine relative risk actually is, considering that more recent independent genome-wide significant signals related to AD and PD are being reported constantly. Should these diseases be excluded from DTCs completely? Perhaps, but that would completely disregard consumer choice. Rather, stricter controls should be put in place for companies offering these kinds of tests. For example, requiring them to offer genetic counseling as part of their services, as well as have controls in place to assess the validity of tests before they are marketed.
Regarding the accessibility of information, we argue that it is imperative to design policies targeted toward increased scientific literacy worldwide. In the case of susceptibility testing for AD and PD, this entails more accessible information regarding genetics (i.e., heritability, genetic determinism), statistics (i.e., what does a probabilistic result actually mean), and medicine (i.e., what these diseases entail, possible and probable prognosis, currently available treatment options, etc.).
We additionally argue that, even though there is some correlation between knowledge and the general attitude toward genetic testing, public perception is also defined by other factors, including social and cultural ones. Therefore, information aimed at changing stigma related to aging and mental health will also be necessary in many contexts.
In addition to widespread information for the general public, it is imperative that potential users of genetic testing are further guided by the medical community (Tan & Jankovic, 2006 (Brazier & Cave, 2011).
In order to develop appropriate educational materials for both healthcare providers and the general public, the following must be taken into account: baseline knowledge of basic principles, interest in topics of genetics and statistics, and opinions regarding genetic testing (Falcone, Wood, Xie, Siderowf, & Van Deerlin, 2011). Additionally, existing preconceptions surrounding mental health and aging should also be considered. It is imperative to mention that the onus of providing said education falls mainly on public institutions. Additionally, resources are limited, and decision-makers need to take as many factors as possible into consideration when budgeting.

| Allocation of resources
For example, would widespread investment on genetic disease raise health costs? Or would it mean a decrease in budget for another area, such as infectious disease? Importantly, conditions with a hereditary component will have increased incidence in particular groups, in contrast to infectious disease which generally affects the entire population. If accessibility to genetic testing for AD and PD is ensured, some might argue that this benefits only certain populations with a predisposition, aggravating inequity amongst these groups and "healthy" individuals. However, we argue that such inequities could be morally permissible, as they are in the benefit of the least well-off group.
In general, access to genetic testing could potentially improve the effectiveness and efficiency of public health policies and expand the delineation of the principle of justice. However, even assuming that public expenditure on genetic testing should be approved, we argue that testing for late-onset neurodegenerative diseases should presently be of low priority, at least until treatment is available or burden becomes more evident.

| CONCLUSIONS
Beyond understanding the biology of disease, some of the ultimate goals of research in genomics include advancing the science of medicine and improving the effectiveness of healthcare. Undoubtedly, integration poses great challenges. However, it is an ethical obligation to make the best possible use of scientific progress to advance society (Brand, Brand, & Schulte in den Bäumen, 2008).
In the case of genetic testing for late-onset neurodegenerative disorders, a plethora of issues arises, many of which are common to other complex diseases. In this manuscript we have established the importance of governments strengthening legislation to support the principles of autonomy, beneficence and nonmaleficence, and justice.
We argue that users of susceptibility genetic testing for AD and PD should be considered to be in a condition of vulnerability, particularly those who approach genetic testing because of a clinician's referral and those who participate in research. As such, strong measures must be taken to ensure that their decisions regarding what to do with their results are respected, and to protect them from any discrimination stemming from their genetic status. This includes the development of standardized regulations specific to clinical and research settings, as well as the regulation of DTC tests. Importantly, we argue that susceptibility testing for underage individuals should be discouraged in all three contexts, given the probabilistic nature of tests for these diseases and the current lack of actionability.
Due to the complex nature of AD and PD, and in order to avoid unnecessary issues with unexpected findings, we also argue that, ideally, they should only be included in multiple disorder screenings when individuals explicitly request them and with the support of a genetic counselor. This is particularly relevant in the context of DTC, as the provision of counseling is not currently taken into consideration overall. Appropriate genetic counseling is imperative to maximize the benefits of susceptibility genetic testing for neurodegenerative disorders, which is why we pose the importance of complete, up-to-date and accessible reference databases, in order to deal with the continuously evolving knowledge of these diseases. In addition, we also argue the importance of accessible information for the general public and improved training of healthcare professionals regarding topics of genetics and statistics. We assert that education is mainly the responsibility of public institutions.
Importantly, said institutions are also responsible for protecting genetic test users from discrimination. Even though data related to AD and PD might currently be considered less sensitive, we argue that measures must be taken now to avoid any future negative consequences for risk-positive individuals.
Regarding the potential benefits of including these disorders in population screenings, we argue that the promise of future benefits is not sufficient justification in this case, and the right to health privacy should prevail. This is mostly due to the fact that epidemiological data and other currently available information should provide enough information for decision-makers to design appropriate health programs that will reduce the increasing burden of AD and PD in the future. Similarly, we argue that public expenditure should not presently consider susceptibility testing for these diseases. Rather, public health actions should prioritize AD and PD in other ways, such as increasing investment in community care facilities, launching campaigns to educate the public, and taking the time to carefully analyze the implications of new technologies related to the diagnosis and treatment of neurodegenerative disorders.
It must be noted that what we have discussed is just the tip of the iceberg: we have not touched upon matters related to, for example, advanced requests for euthanasia and physician-assisted suicide after early diagnosis of neurodegenerative disease. Importantly, all of these considerations must also be taken into account in the ethical debate of these disorders.
Finally, even though evidence should be the base for science and health policy, trust will only be strengthened if public participation is an integral part of the discussion. We assert that it is imperative to bridge the communication gaps amongst researchers, clinicians, decision-makers, and the general public. We must maximize the bene-