Within days of birth, a few drops of blood are collected from every newborn in California – and across the United States – which are then stored on filter paper and screened for dozens of genetic and congenital disorders, such as phenylketonuria (PKU), an inherited metabolic disorder that can result in intellectual disability, seizures, heart and behavioral problems.
Researchers at University of California San Diego School of Medicine have launched a Phase II research study to look for signs of another similarly devastating disorder, one that typically does not appear in seemingly healthy children until years later: autism spectrum disorder or ASD.
The UC San Diego Newborn Screening-Autism Risk Study is designed to determine whether the dried and stored blood drops of children later diagnosed with ASD contain within them the tell-tale presence and combinations of biological molecules and environmental chemicals that might predict the risk of a future ASD diagnosis.
“We know from the history of certain genetic diseases, such as PKU, that if children can be identified before the first symptoms have appeared, then the disease can be prevented, even though the children have the DNA mutations,” said Principal Investigator Robert Naviaux, MD, PhD, professor of medicine, pediatrics and pathology at UC San Diego School of Medicine.
“I believe that over half of autism cases may be preventable if only we had a way to identify the children at risk before the first symptoms appear.”
Naviaux said the new study is important for two reasons: the dramatic rise in diagnosed cases of ASD and increasing evidence that early intervention in children at risk of ASD can significantly improve outcomes.
The prevalence of ASD has risen from 20 in 100,000 births in the 1970s to 1,700 in 100,000 in 2014, according to the U.S. Centers for Disease Control and Prevention — an 84-times increase.
Approximately one in 59 children is diagnosed with ASD. Statistics from the U.S. Department of Education and other government agencies indicate autism diagnoses are increasing at the rate of 10 to 17 percent per year.
Changes in diagnostic criteria and reporting practices account for 60 percent of the rise, at most, according to previously published research.
“This means that even by the most conservative estimates, the prevalence of ASD has increased at least 34 times,” said Naviaux.
The overarching question for Naviaux and others is why? Is it genetics? The environment?
“Our genes have not changed significantly in the past 50 years,” said Naviaux. Single gene mutations play a causal role in approximately 10 percent of ASD cases.
The vast majority of ASD cases are idiopathic or of unknown cause, most likely the result of a combination of genes, environmental factors or something yet to be identified.
“More than 1,000 genes can contribute to the risk and resistance a child has to ASD, but more than 95 percent of these genes are common variations also present in asymptomatic parents and children who don’t have ASD,” Naviaux said.
“A clue to how the genetics of ASD is misinterpreted is the fact that many of the genes that contribute to ASD are the same genes that contribute to other disorders like schizophrenia and bipolar depression.
In most cases, DNA only sets what is possible, not what is destined.”
The new Phase II study will focus on exposure and possible roles of chemicals and compounds (detected in blood) and how they might interact with genes.
Researchers will use a blood test developed in Naviaux’s lab to analyze the presence of more than 600 metabolites –typically small molecules produced by metabolism, the life-sustaining chemical reactions in all organisms.
Metabolites from amino acids and antioxidants to vitamins and lipids serve diverse, crucial functions, including as fuel, signal carriers, structure providers, defenders and regulators among them.
Earlier research by Naviaux and others has found that persons with ASD appear to have a shared “metabolic signature.” That is, their biological chemistry is comparable, though their genetics are unique.
Testing will also look at more than 400 environmental chemicals in each dried blood drop.
Exposure to these chemicals, such as commonly used pesticides, flame retardants, air pollutants, lead, mercury and polychlorinated biphenyls or PCBs, has been linked to several neurodevelopmental disorders, including ASD.
At birth, blood samples are taken from newborns and used to screen for genetic diseases. A new research study will help determine whether these drops can also help predict autism risk.
Naviaux and colleagues believe that the majority of ASD symptoms are the result of a treatable metabolic syndrome triggered by persistence activation of the cell danger response (CDR), a natural and universal cellular reaction to injury or stress.
Chronic CDR, they suggest, results in disrupted and incomplete healing at the metabolic and cellular levels. In ASD, the consequence may be dysfunctional neural circuits and internal systems, producing autism’s well-documented symptoms and behaviors.
“Metabolism is the real-time result of our genes interacting with the environment,” said Naviaux.
“Environmental chemical or biotoxin exposures –the ‘exposome’ — at critical developmental windows can produce delayed effects that become apparent only after months or years.
By measuring metabolism and the exposome, it may be possible to identify children at risk for developing autism before the first behavioral symptoms appear.”
The study seeks 400 participants between the ages of three and 10 years old, meeting these requirements:
- Born in California
- Have a confirmed diagnosis of ASD from a licensed clinician or be a healthy child not taking any prescription medications (200 participants from each group)
- Born after a normal term pregnancy of 37 to 42 weeks
- Have not had a medical issue that required readmission to the hospital in the first month of life
The study requires parents of participating children to answer questionnaires covering pregnancy, labor and delivery, the child’s health history and that of the family. Consented analyses will be conducted of dried blood drops recorded as part of California’s Newborn Screening program, which began in 1966 and now screens for 80 different genetic and congenital disorders. Blood spots have been saved and stored by the California Department of Public Health since 1982. No new blood tests or behavioral testing will be required for the Phase II study.
Naviaux said he hopes to screen and enroll the full complement of participants by June 2020. Analyses of the identified and retrieved blood spots is expected to be complete by June 2021.
“We then hope to expand the testing program to states like New Jersey, New York, Pennsylvania and Washington by enlisting collaborators in each of those states who will be able to apply the new methods we have developed.
“Each new state has slightly different policies and regulations regarding the collection and storage of dried blood spots from universal newborn screening programs, so this medium-scale expansion study will teach us what will be needed to launch a national study.”
Note: For more information on the study or to apply for enrollment, see the study web site.
Early intervention and disease modification are the future of health care worldwide. Rather than the technical and regulatory concerns, the greatest threat to this effort of detecting prodromal and preclinical states may in fact be ethical issues.
Detecting diseases and disorders before clinical symptoms manifest enables earlier intervention and offers the hope of improved health outcomes. For example, screening for markers of breast cancer before symptoms arise is both widespread and recommended by many physician groups (Monticciolo et al., 2017; Sardanelli et al., 2017).
Early-stage interventions reduce average patient cost by more than $100,000 over two years (Blumen et al., 2016) and decrease mortality (Howlader et al., 2017). Such a large positive effect of early detection and treatment provide an almost incontrovertible argument for regular early screenings.
Even so, the method of arriving at early intervention is controversial. There is conflicting evidence on the efficacy of routine mammograms in decreasing breast cancer mortality (Berry et al., 2005; Domchek et al., 2010; Narod et al., 2014; Harding et al., 2015; Monticciolo et al., 2017).
Whether regular screenings for breast cancer are necessary is an ongoing debate, demonstrating the complexities that arise from early detection efforts, even when treatments are widely available and effective.
The debate becomes more complicated with disorders where effective treatments are not yet developed, as with brain disorders.
With the considerable global burden of brain disease, the promise of early detection and early intervention cannot be overstated.
That being said, preclinical detection of brain disorders encompasses a unique suite of ethical concerns, as dysfunctions in the brain directly impact behavior and are intrinsically linked to identity and autonomy.
In other words, when we predict a future brain disorder, we not only predict a health diagnosis but also predict who a person may become.
This review will discuss the considerations surrounding the ethics of preclinical detection through the lens of three brain disorders that typically present at distinct time points across the life span: autism spectrum disorder (ASD) in early childhood, schizophrenia in adolescence, and Alzheimer’s disease with aging populations.
A patient is similarly impacted whether the etiology of a disorder is an acute biological or a multifactorial biopsychosocial one, so disorders from both categories will be discussed together. Related discussions of the ethics of preclinical detection have been started in other venues, such as Baum (2016) and Chneiweiss (2017).
We will expand the discussion and place a greater emphasis on the implications for patients in a medicalized preclinical state. The disorders we focus on demonstrate the unique ethical quandaries in:
(1) risk/benefit analysis,
(2) the possibility of stigma and discrimination, and
(3) responsibility and communication of risk.
The review will conclude with recommendations for addressing these ethical challenges, which we do not intend to hinder research but to anticipate and mitigate potential roadblocks.
As medical screenings and diagnostic tools continue to expand in scope and accuracy, an ethical framework will be necessary, even in research and clinical settings where preclinical detection of brain disorders is not the primary goal.
The nature of preclinical detection is inherently probabilistic, so certainty can never be fully achieved with these strategies, but citizens worldwide stand to greatly benefit from the scientific advancements offered by preclinical detection if interventions and regulation are developed with careful ethical reflection.
We believe addressing these ethical concerns in anticipation and as part of the improvements to preclinical detection technology will help ensure the promise of improved health that predictive technologies aspire to offer.
Terminology: Preclinical or Prodromal Brain Disorders
Brain disorders are contextualized states, regardless of their etiology. Disordered states that lead to disordered behavior are diverse in their development and manifestation, and some of these states are not universally seen as truly disordered (e.g., the prominent neurodiversity movement in the ASD community; Armstrong, 2015).
That said, all cases discussed here, and all cases in which preclinical detection could be used to identify patients before symptom onset, are medicalized, and are therefore subject to the same protection, concerns, and risks.
The preclinical label is defined by the presence of predictive markers in the absence of symptoms that currently define the disease. Preclinical states are distinct from prodromal or subclinical states, in which some clinical presentations (such as a mood disorder) are present but do not satisfy criteria for diagnosing a disorder (like schizophrenia; Gourzis et al., 2002; Meyer et al., 2005; for examples of preclinical and prodromal markers, see Table 1, adapted from Arias et al., 2018).
Early interventions of schizophrenia currently target the prodromal stage. In ASD, the hope is that early interventions begin at the age when the child’s behavioral symptoms do not yet reach diagnostic criteria. Efforts in Alzheimer’s disease are unique, in that the preclinical stage has been defined by an absence of behavioral or cognitive symptoms, well before the onset of mild cognitive impairment (MCI).
The definition and detection of preclinical stages are more accessible in disorders like Alzheimer’s disease, which have established molecular biomarkers (e.g., measuring amyloid levels with positron emission tomography and measuring τ levels in cerebral spinal fluid (Dubois et al., 2016; see Table 1) arising well before behavioral symptoms. Preclinical Alzheimer’s disease is defined as the presence of one or more of these molecular biomarkers, in the absence of cognitive impairment.
The diagnosis is often subdivided into two differential diagnoses: presymptomatic, for those who will develop clinical Alzheimer’s disease with pathogenic autosomal mutations, and asymptomatic, for those at risk of developing clinical Alzheimer’s disease with predictive biomarkers (Dubois et al., 2010). The reliability and validity of such tests will be further explored in the following section.
In contrast to Alzheimer’s disease, no preclinical biomarkers for ASD or schizophrenia have been validated to date, although many genetic and environmental factors have been identified. Current efforts for early detection in these diseases focus on identifying subclinical symptoms in the prodrome (Gourzis et al., 2002; Christensen et al., 2016)
Table 1.
Recognized biomarkers, symptoms, and methods for detection
| Preclinical biomarkers | Prodromal symptoms | Techniques for measuring markers or symptoms | |
|---|---|---|---|
| Autism | None identified | Decreased social engagement and eye focus (Jones and Klin, 2013) | Eye tracking (Klin et al., 2002), naturalistic observation (Baranek, 1999), structural brain scan (Hazlett et al., 2017) |
| Schizophrenia | None identified | Subclinical positive, negative, and cognitive symptoms (Goulding et al., 2013) | Clinical interview (Goulding et al., 2013), genomic analysis (Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014) |
| Alzheimer’s | Low CSF Aβ1-42 with high CSF P-τ or T-τ, increased amyloid PET retention, autosomal dominant mutation (e.g., APP, PSEN1/2; Jack et al., 2011; Dubois et al., 2014, 2016) | Mild cognitive impairment | PET scan with injectable tracer, lumbar puncture, memory assessment (e.g., FCSRT; Dubois et al., 2016) |
Alzheimer’s is the only disease of those discussed with recognized preclinical markers. Adapted from Arias et al. (2018). CSF: cerebrospinal fluid, PET: positron emission tomography, FCSRT: Free and Cued Selective Reminding Test, APP: amyloid protein precursor, PSEN: presenilin.
ASD
ASD encompasses a range of phenotypes, from mild social impairment to an inability for self-sufficiency (American Psychiatric Association, 2013). ASD is now estimated to affect one in 160 children globally (World Health Organization, 2017) and is the leading cause of disability in children under the age of five (Baxter et al., 2015).
The average age of diagnosis is approximately four years old (Christensen et al., 2016), which makes the needs of patients and their caregiver(s) a public health concern (Khanna et al., 2011; Cadman et al., 2012).
Studies have shown that infants who will develop autism prefer looking at mouths versus eyes during social engagement (Jones and Klin, 2013). Early screening attempts for ASD rely on eye-tracking in infants to detect atypical patterns of social gaze. Retrospective analyses of eye tracking behavior have identified infants as young as six months of age who would later develop ASD (Chawarska et al., 2013; Jones and Klin, 2013; Shic et al., 2014).
To date, these studies test the value of eye-tracking as a relatively non-invasive, easy, and inexpensive screening tool. These studies target high-risk populations (siblings of children with autism) of infants and children whose parents express concern over their child’s social development (Sandin et al., 2014; Rowberry et al., 2015).
Eventually, the hope is that such a tool could be implemented in routine wellness visits in all infants (high risk or not). Preliminary studies have also found differences in cortical development between infants who do and do not develop ASD (Hazlett et al., 2017).
While brain scans may provide an opportunity for another preclinical biomarker of the disorder, neuroimaging is likely less accessible and too expensive to be considered for widespread screening.
Early interventions to address early diagnoses are currently being designed. Perhaps unique to ASD treatment, the proposed behavioral interventions are beneficial for both autistic children and typically developing children (Institutes of Medicine and National Research Council, 2013), which minimizes the risk of false positives in this specific context.
Schizophrenia
Schizophrenia develops later in life than ASD, with the first symptoms usually appearing in late adolescence/early adulthood or during the peri-menopausal phase (Castle and Murray, 1993; World Health Organization, 2001).
Positive symptoms (such as psychosis), negative symptoms (such as anhedonia), and cognitive deficits contribute to the severe disability and loss of productivity associated with the disorder (World Health Organization, 2001).
Although the lifetime prevalence of schizophrenia is ∼1% of the world population, the World Health Organization (WHO) estimates that schizophrenia is the eighth leading cause of disability-adjusted life years (DALYs) in 15–44 year olds (World Health Organization, 2001). Many risk factors of schizophrenia have been identified, including environmental (Cornblatt et al., 2003) and genetic (Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014) contributors.
Despite the genetic factors, genome-wide association studies (GWAS) show low sensitivity and specificity in identifying those who will develop schizophrenia, which has led some teams to warn against using genetic analyses as predictive tests (Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014).
No preclinical markers of schizophrenia have been identified; as such, clinicians rely on prodromal symptoms like anxiety, sleep disturbances, and depressive mood, to identify at-risk patients (Goulding et al., 2013).
At-risk patients are often identified because of treatment sought by the patient or caregiver, not by routine appointments.
People often seek treatment for prodromal symptoms for schizophrenia, which are themselves clinical symptoms for other disorders (Gourzis et al., 2002; Meyer et al., 2005; Rosen et al., 2006). At this early stage, symptoms, family history, and genetic risk factors can put the patient at a high-risk for developing schizophrenia (Larson et al., 2010; Seidman et al., 2010; Goulding et al., 2013).
This categorization presents the opportunity to intervene before clinical schizophrenia develops, in the interest of instigating preventative interventions. Prodromal symptoms do not always transition into clinical schizophrenia.
Symptoms are often non-specific to psychosis (Gourzis et al., 2002; Rosen et al., 2006), and this has hindered success in designing early interventions. Prodromal interventions, such as the use of atypical antipsychotics (McGorry et al., 2009), antidepressants (Cornblatt et al., 2007), and alternative treatments like omega-3 fatty acids (Amminger et al., 2010), have produced mixed success in reducing transition rates (Larson et al., 2010). The uncertainty of a prodromal diagnosis further limits the confidence of successfully intervening before clinical symptoms develop, especially given the severity of side effects of anti-psychotic medications (Patel et al., 2014).
Alzheimer’s disease
Alzheimer’s disease is unique among the three disorders discussed here, in that there is a generally accepted symptomatic subclinical stage for this disorder (MCI), which is often preceded by the presence of amyloid-β (Aβ) plaques, tau, and neurodegenerative biomarkers (Dubois et al., 2014; Jack et al., 2016; Racine et al., 2017). The research has progressed to the point that many organizations are advocating for the inclusion of a preclinical (fully asymptomatic) diagnosis being integrated into regular clinical practice (Dubois et al., 2014; Alzheimer’s Association, 2019). Alzheimer’s disease is the leading cause of dementia, and risk for this disorder increases dramatically with age (Hebert et al., 2013). Occurrence of the disorder is expected to double in the next 20 years, driven largely by the impending boom in population of those aged 65 or older (He et al., 2016). Ranked as the 25th most burdensome disorder in 1990, the increasing prevalence has driven Alzheimer’s disease to become the 12th most burdensome disorder in the United States over the past 20 years (Alzheimer’s Association, 2019). Similar increases in prevalence and burden are recorded throughout Europe (Wittchen et al., 2011). The protracted development of the disorder creates an enormous burden on the primary caregiver(s), as many as 40% of whom suffer from depression (Alzheimer’s Association, 2019).
In recent years, preclinical trials have commanded more of the industry’s effort, given the poor success rate of pharmaceutical trials in clinical interventions (Cummings et al., 2014; Hung and Fu, 2017). Dementia is thought to develop 20–30 years after the onset of Aβ deposits in the brain (Hubbard et al., 1990; Jansen et al., 2015), strongly supporting the idea that effective treatments may require intervening at the preclinical stage. Multiple ongoing clinical trials for pharmaceutical interventions now target high-risk populations not yet diagnosed with any cognitive impairment. For example, many drugs that previously failed efficacy trials in patients with mild to moderate Alzheimer’s disease are now being retested in preclinical populations (Hung and Fu, 2017). High-risk populations are defined as individuals with a family history of Alzheimer’s disease (Honea et al., 2012), the ε4 allele of the APOE gene (Bonham et al., 2016), or the presence of biomarkers, like elevated τ and a high Aβ1-42/Aβ1-40 ratio (Holland et al., 2012).
Conclusion
Brain disorders are becoming statistically more prevalent in a population that is living longer and that is less affected by communicable diseases (Borlongan et al., 2013; Effertz and Mann, 2013). We must recognize that everyone is a patient in waiting. All disorders are developmental in nature, and therefore many more disorders than those discussed above have discrete, if currently undiscovered, preclinical stages. Risk modification will be the future of health care as the science of preclinical detection progresses. A thorough investigation of best ethical practices is needed to manage the use of new tools in the clinic and beyond. Regulatory hurdles and public distrust can easily stymie or corrupt these advancements if scientists and clinicians fail to engage in conversations with policymakers and the wider public. Most importantly, we must recognize that the best practices will not be consistent across conditions or cultures. True appreciation for the risks of preclinical research requires the acknowledgment that the risks (be they stigma, impact on interpersonal relationships, or individual anxiety) are influenced by cultural norms. The need for empirical research to measure public attitudes is never more important than when identity and autonomy are directly impacted. We can maximize scientific advances and public acceptance by responding to, and not dictating, public views on the matter. Such a dialogue will help the scientific community protect patients before the harms of uninformed preclinical detection are inflicted on them.
Source:
UCSD
