NEW YORK (GenomeWeb) – Christian Millare had a severe seizure on Jan. 5, 2008, and died. He was two years old.
His mother Amy Williams is convinced, based on his medical records, the opinions of experts, and the published literature, that her son's life didn't have to come to such a premature end. Eight years later, Williams is suing Quest Diagnostics, one of the largest reference labs in the US, and its subsidiary Athena Diagnostics, which in 2007 tested Christian for mutations in the SCN1A gene.
In a lawsuit filed last month in the fifth judicial circuit court in Richland County, South Carolina, Williams alleges that because Athena failed to follow federal lab regulations and accurately classify the genetic mutation causing her son's epileptic seizures, he continued to receive treatment that worsened his condition and caused his death.
In 2007, Christian's doctors sent his blood sample to Athena to gauge if he had mutations in the SCN1A gene, which is involved in the mechanism that controls the flow of sodium ions from neuron to neuron. Defects in SCN1A can throw off this process, creating an imbalance of excitatory and inhibitory electrical impulses in the brain and causing seizures. Mutations in SCN1A are well known in the literature to cause Dravet syndrome, a severe form of epilepsy that impacts one in 21,000 infants. Dravet syndrome babies start having seizures a few months after birth and have developmental delays. The Dravet Syndrome Foundation estimates that 80 percent of patients will have an SCN1A mutation.
In the complaint, Williams accuses Athena of misclassifying her son's SCN1A mutation as a variant of unknown significance (VUS), meaning that the lab determined there wasn't sufficient evidence in 2007 to link the mutation to epilepsy or determine it was benign. Williams asserts there was enough evidence at the time that her son's mutation was disease-causing. The complaint cites two papers, one published in June 2006 and one published in March 2007, which mention Christian's specific mutation had been studied and seen in another patient who had epileptic encephalopathy.
One of the authors of the 2007 paper, published in Brain, is Sat Dev Batish, who then was and still is Athena's chief director of genetics. Moreover, a patent that Australian firm Bionomics licensed to Athena for the development of its SCN1A test lists the mutation (1237T>A) that Christian had as one that "disrupts the functioning of an assembled ion channel so as to produce an epilepsy phenotype."
Further complicating matters, amid the litany of workups and tests performed on Christian, Williams claims the report that Athena issued in June 2007 with the VUS finding never made its way to her. She learned about it in 2014, while searching for more definitive answers to why Christian passed away. Newly married, Williams was fearfully considering whether she could have a child again. Uncertain if she had passed on a genetic abnormality to Christian, she didn't want to risk it. "My body can't go through it again," Williams told GenomeWeb. "My heart can't go through it."
Williams was able to obtain the original VUS report from Athena in the fall of 2014, and learned that the company now had classified the SCN1A mutation as disease-causing. Athena issued an amended report in 2015, but the published studies listed in the document as evidence the lab used to make the variant classification are the same as the 2007 report.
According to the complaint, Athena's failure to provide additional information to support this variant reclassification is one of many violations of the Clinical Laboratory Improvement Amendments — the federal regulatory standards that all labs must follow when testing humans. CLIA requires that labs have adequate systems for delivering test results to physicians and tracking the entire analytical process. Moreover, the complaint also highlights sections of CLIA stating that labs must furnish additional information in a timely manner to ensure the test results are accurate.
"Athena seemed to make a series of mistakes, according to the complaint. Although the initial mistake, while extremely unfortunate and ultimately tragic, may be chalked up to human error, it does appear that they did not follow their own variant classification scheme," said Girish Putcha, who has directed a number of clinical labs. He reviewed the plaintiff's complaint for this article but is not involved in the case.
The complaint notes that, based on Athena's own variant classification criteria described in the 2007 report, Christian's mutation should have been deemed an "amino acid change of unknown significance," instead of a VUS. In 2010, Athena launched a program called Athena Insight, which investigates the pathogenicity of variants and categorizes them on a seven-point scale.
"The subsequent failure to notify the patient and his caregivers in a timely manner when the error was noticed and a change [was] made to the final test report, followed by what appears to be a deliberate attempt to delay such notification and/or conceal the mistake altogether, not just for this patient but perhaps for others as well, is especially disturbing," said Putcha.
Christian's 2007 and 2015 reports indicate they were reviewed by Batish, and two other directors at Athena, Narasimhan Nagan and Hui Zhu. According to publicly available information, Nagan and Zhu left Athena to work at other labs in 2007 and 2008, respectively. As such, Williams' lawyer Bradford Cranshaw questioned how Zhu and Nagan could have reviewed the 2015 amended report, and pointed out that CLIA requires labs to record the identity of personnel overseeing testing.
A kid dies from a misdiagnosis, and the system doesn't even notice that because it doesn't get incorporated into the data structures any place.
Further, "upon information and belief, prior to 2015, [Quest and Athena] recognized that a substantial number of patients were affected by similar erroneous SCN1A … sequencing reports," the complaint states.
If the case goes forward, a jury will have to decide whether there was sufficient evidence in the literature in 2007 for Athena to determine a pathogenic link between Christian's mutation and Dravet. "We can demonstrate that the change was made between the 2007 report and what we received in 2015. We do not know when that change was made by Quest and Athena." Cranshaw told GenomeWeb. "That's why we've brought it to the court and why we think a jury should consider the facts in the case."
John Conley, a law professor at the University of North Carolina, told GenomeWeb that if this case goes forward it could impact the standard of care provided by a genetic testing lab, but first, plaintiffs will have to prove that if Athena hadn't been negligent, Christian wouldn't have died. "That to me is their biggest challenge," said Conley, an expert in biotech and IP law. "The link between a genetic test and an outcome is so complex and attenuated, that's going to be hard to explain factually and hard to prove."
When asked to provide comment for this article, Athena's parent company Quest declined to discuss pending litigation. Williams is suing the company for actual and punitive damages. She would also like to see some change in the law that holds labs more accountable.
Right now, there are no mandated standards for how labs classify and interpret detected variants, and report genetic test results. The American College of Medical Genetics and Genomics published updated guidelines in this regard in 2008 and again in 2015, and while accrediting bodies say labs should adopt these standards, it is voluntary. There's no requirement that labs be transparent about their variant classification process and report data from genetic test results into publicly accessible databases. And although recent laws now allow patients access to their lab results, in the field of genetics, where variant interpretation and reporting standards are still evolving, there is little agreement among industry players, regulators, and medical professionals about what information to return to patients, how to present this information in reports, and whether it should go directly to them or through a doctor, particularly when VUS are involved.
Robert Cook-Deegan, an expert in genomic testing policy and ethics at Duke University, said this case "exposes an incredible flaw" in the genetic testing field right now, in the way labs interpret whether a particular genetic variant is associated with disease and how that information gets published, finds its way into public databases for others in the field to evaluate, and makes its way into providers' and patients' hands. "As far as I know, [Christian's] case has never been reported in the literature," said Deegan, who filed an affidavit on behalf of the plaintiff stating his opinion that Athena breached the standards of care for a high-complexity genetic testing laboratory. "A kid dies from a misdiagnosis, and the system doesn't even notice that because it doesn't get incorporated into the data structures any place."
Williams was 20 years old and in her junior year studying early childhood education at the University of South Carolina, Aiken, when she gave birth to Christian. It was an uncomplicated pregnancy and Christian, in the first months of his life, was a healthy baby, with blue eyes and a button nose.
He had his first seizure at four months old, in December 2005. He had a slight fever and started shaking. Williams had no idea what was wrong. "I thought he was cold. I'd never seen a seizure before," she said. In subsequent months, as Christian's seizures worsened, Williams kept a journal.
"I often felt like an investigator tracking a serial killer trying to figure out what the triggers were, but eventually gave up because nothing made sense," she said. Christian's immune system was weak, so any illness could set off a seizure. As a precaution, he had his tonsils and adenoids removed. Still, nothing seemed to help. Sometimes he'd shake. Other times one side of his body would stiffen like during a stroke. He'd seize at the grocery store, when he was overexcited, or too tired. The visits to the hospital grew more frequent, and there were endless tests and workups.
One of these evaluations was to see if a mitochondrial disorder could be causing his seizures. The results suggested this might be the case, so Christian was treated with increasing doses of the sodium channel-blocking drugs carbamazepine and lamotrigine, which are standard treatments for epileptic seizures when they are not due to Dravet syndrome.
Naturally, I typed the mutation into Google and obviously the top hits were those papers.
Studies show that these kinds of drugs worsen seizures in Dravet patients. According to Christian's autopsy report, he died from a seizure due to complications from an unspecified mitochondrial disorder. There was a bite mark on his tongue and carbamazepine in his blood. According to the Epilepsy Foundation's estimates sudden death occurs in more than one out of 1,000 people with epilepsy, and when the epilepsy is uncontrolled sudden death occurs in one out of 150. Sudden death due to epilepsy is rare in children.
Cook-Deegan writes in his affidavit that Christian's final and fatal seizure in January 2008 was due to Athena's "negligent diagnosis and failure to accurately advise selection of appropriate therapy."
It took Williams a long time to come to terms with her son's death and muster the courage to try to piece together what went wrong. "She didn't mention Christian a whole lot," said Laurel Coons, a high school friend who reconnected with Williams several months after her son died in 2008. "I'm quite certain that even today most people who know Amy do not know she had a child."
By 2009, the two women had become close and Williams opened up about Christian. Coons admits jokingly that she probably asked her friend too many questions. "[Amy] always thought that Christian died of a mitochondrial disorder that she gave him," recalled Coons, who never got a chance to meet Christian. "I used to give her a hard time because the stuff she was saying didn't make sense to me."
When mitochondria, the energy-producing factories in our cells, don't work properly it can cause a range of disorders. Mitochondrial diseases may be due to defects in the mitochondrial DNA, inherited entirely from the mother, or defects in nuclear DNA, which can come from the mother or father. After Christian's death, Williams didn't want to have genetic testing. "I didn't want someone to confirm that it was me that killed my child," she said. She resigned herself to never having children again.
Eventually, Williams shared her son's medical reports with Coons. Within these documents, Coons noticed that Christian had been tested for SCN1A mutations, but there were no results.
"Nothing much ever came of that," said Coons, who has a biology background and is a PhD candidate at Duke University's Integrated Toxicology and Environmental Health. By this point, however, Williams was trying to actively piece together precisely why her son died, and had been contacting mitochondrial disease researchers around the world to see if they would test his tissue samples.
Two years ago, one of these researchers brought up the missing SCN1A test report again. "I had no idea what he was talking about," Williams said. Through one of Christian's former genetic counselors, she learned Athena had performed the test. But when Williams contacted the firm, Athena would only release a copy of the results to a medical professional. By then the physician who had ordered the test was no longer seeing patients and wouldn't request the results for her, Williams said.
With the help of a genetic counselor, she was able to obtain a copy of the 2007 results. The day in September 2014 Williams received the report from Athena with the VUS finding, she sent it to Coons. "Naturally, I typed the mutation into Google and obviously the top hits were those papers," she said, referencing the two published papers from 2006 and 2007 that link Christian's SCN1A mutation to epilepsy in at least one other patient.
Athena indicated when releasing the 2007 VUS report that it would provide an amended report in six weeks because the variant had been reclassified as disease-causing. Months later, in January 2015, Williams received the new report with the SCN1A variant classified as pathogenic, but she noted that Athena provided no new evidence for the change.
"This is a heartbreaking case," Cook-Deegan said. "This case really brought home for me how important it is that we solve these problems of making information flow, not just from research laboratories and academic health centers, but from the whole system, and making sure we're capturing the information about what these genomic variants mean and getting them into databases where everyone can see them."
Remember the TJ Hooper
In March 1928, a tug boat called the TJ Hooper was towing a barge full of coal from Norfolk, Virginia, to New York City and got caught in an "easterly gale" off the Jersey Coast. The barge sank and took down the coal with it. The barge owners sued the tug boat owners, saying the tugboat should have had radios on board so the crew could have been warned about the storm.
A statute at the time required passenger steamers, but not tug boats, to be equipped with such radios. The tug boat owners defended that no other tug boats were using radios, so that constituted the standard. An appellate court judge ruled against the tugboat owners, calling the TJ Hooper unseaworthy. "There are precautions so imperative that even their universal disregard will not excuse their omission," the judge wrote.
Like the TJ Hooper case, Williams v Quest/Athena could have an impact on what constitutes the standard of care for a genetic testing lab, according to Conley, who is editor of Genomics Law Report, a blog that explores legal issues impacting the genomics field. "The TJ Hooper is a good model for a novel case like this, because a court can say in hindsight that although everyone was doing things a certain way, they shouldn't have been," said Conley. "It could turn up the pressure [on labs'] handling of VUS. It could force labs to be more collaborative and share data out of fear."
After the US Supreme Court in 2013 struck down several of Myriad Genetics' patent claims on BRCA1/2 genes, competing labs began to highlight that unlike Myriad — which has a large proprietary database — they would share their variant data in public variant repositories. In response, Myriad has highlighted the inaccuracies and lack of agreement on variant classifications within public databases, and has warned that labs would be taking on liability by using them to inform patient care.
Indeed, many variant databases contain disclaimer language informing users to not make diagnostic decisions based solely on the information in the repository. Even so, the life sciences community as a whole seems to be increasingly in favor of public variant resources, such as ClinVar — an NIH-funded database launched in 2013 where labs can submit variant interpretations and look for areas of agreement and disagreement. "Steps need to be taken to help ensure that other patients, families, providers, and laboratories do not repeat this sad tale," genetic counselor Robert Resta wrote in The DNA Exchange blog last week, discussing the circumstances around Christian's death and urging for the need to fund well-curated, freely-accessible databases.
"We, as a community, need to come together and create a common reliable source of all this information, where we can curate papers and share observations in a publicly accessible place, and track knowledge as it evolves," said Heidi Rehm, director of the Laboratory for Molecular Medicine at Partners Healthcare Personalized Medicine. "So, every time you look at a variant, you can go to a reliable source to understand … the current thinking on that variant at least from the community and the literature."
Rehm is also one of the principal investigators of the Clinical Genome Resource (ClinGen), a broad effort to create a shared knowledgebase of genes and genetic variants important to human health. Within this effort, experts have been able to collaborate on variant classifications and, in most cases, come to a consensus about their clinical significance. If ClinVar and ClinGen were available back in 2007, would the outcome have been different in Christian's case? That's hard to determine, because the entire field was so different. Gene patents were still in effect. Data sharing wasn't the badge of good industry practice that it is becoming today. Most doctors wouldn't have thought to share genetic test results with patients.
I wish I never knew about genetics. It overwhelming, but it's the uncertainty that is the most overwhelming.
There are now databases dedicated to SCN1A, which list Christian's mutation as causing Dravet syndrome. There is one submission on Christian's mutation in ClinVar from the UniProt Knowledgebase, a database run by several European institutions (see PDF below). Athena has made 25 submissions in ClinVar on SCN1A variants, but not on the one Christian had.
Variant interpretation and classification requires a significant investment in resources, time, and personnel commitment from labs, which are under increasing reimbursement and competitive pressure to lower costs of complex genetic testing like whole-genome and exome sequencing. Still, the fact that consumers can now have their whole genome sequenced for less than $1,000, through which they will undoubtedly learn about the many VUS picked up by such analyses, necessitates more rigorous standards and regulatory oversight, some experts would argue.
Rehm is of the view that lab accrediting organizations like the College of American Pathologists (CAP) should require labs to submit their variant interpretations as a way to check quality control; journals should require data submission for publications; insurers should mandate variant submissions for reimbursement; and the US Food and Drug Administration should deem labs that don't share variant interpretations for peer review to be high risk and regulate them. At the same time, she believes that patients should have unfettered access to medical reports and test results, even VUS if done so appropriately, to allow patients to be advocates for understanding their genomes.
But at what point should doctors be consulted and are they responsible for failing to further investigate a VUS result that might be potentially important to a patient's health? "We're going to have to figure that out. I don't think anyone has a crisp, clean answer," Cook-Deegan said. While he also believes that people should have access to their raw genomic data and variant call files, Cook-Deegan added, "It's reasonable for healthcare professionals to say we don't know what to say about VUS, other than that's what they are."
Furthermore, malpractice cases against doctors are expensive and often fail. The plaintiff in this case may not be suing Christian's doctors, Conley reasoned, because there may not be any evidence they did anything wrong. "Bottom line being that the doctor was entitled to rely on the lab," he said.
The case against Athena also comes as the US Food and Drug Administration is readying to finalize guidelines for regulating lab-developed tests (LDTs), like the SCN1A test, which have been traditionally regulated under CLIA. The FDA has said that CLIA isn't sufficient to protect the public's health and last year issued a report highlighting 20 cases where tests not overseen by the agency caused harm. Christian's case fits within this narrative of patient harm that the FDA is building, but would the agency's regulation have helped avoid the tragedy?
Based on the information in the complaint, Athena was able to detect the SCN1A mutation, but made an error in the interpretation, which ventures in to "the practice of medicine," an area not under FDA's oversight authority. The FDA can still help, Putcha noted, as the agency is already doing, in developing and defining best practices for databases used for clinical interpretation of variants, and may even be able to create a path for "certifying" databases that meet certain criteria. The agency could also provide guidance in collaboration with groups like ACMG to establish standards or at least best practices for genetic test reports and variant classification, he added.
Lab certification and accrediting bodies, such as CAP and the New York State Department of Health, can check labs' variant classification processes, but the lawsuit against Athena highlights, according to Putcha, that their oversight should be strengthened. In addition to CLIA certification, Athena's lab has CAP accreditation and is licensed in NY. "It's disappointing that it appears neither CAP nor NY, which ostensibly provide the highest level of oversight for LDTs in this country, found this and other problems highlighted in the complaint, even eight years after it happened," said Putcha. "To be fair, finding a single violation or even several of them can be difficult when the test volume is in the thousands, if not more, and given the time, resources and expertise available to the inspection teams."
Also concerning, he continued, is the language in the complaint where the plaintiff asserts that at least one of Athena's lab directors should have known about the clinical significance of the variant, because he was co-author on a paper submitted and published before Christian's test report was issued. "While it's always possible, it's hard to imagine a case where the plaintiff could connect the dots, from what should have been known to be an erroneous molecular diagnostic test result to patient harm, much more directly than this," he said. "But even here, it's unclear whether the right test result would have changed the outcome for this little boy, but it certainly seems more than plausible that it might have."
Williams, now 31 years old, said her experience over the last eight years has given her post-traumatic stress disorder when it comes to genetic testing and about having her own family. She's not sure yet if she will have another child, but she's thinking about it. "I'd be a pediatrician's worst nightmare," she said, because she has lost trust in the medical establishment.
Even though Christian's illness must have taken a devastating toll on Williams, she was still planning for his future and trying to lead a normal life. She kept teaching. She was planning to send Christian to a school for kids with special needs. Because Christian couldn't talk, she was teaching him to use sign language. In the weeks before he died, he called her "Mom" and said, "I love you."
Williams and Christian's father had planned to get married, but after the baby died, they only stayed together for four months. Williams' own family had been very supportive while she was pregnant and after Christian's birth, but his death initially drove a wedge between them, too. There were uncomfortable discussions about genetics and mutations and who had inherited what from whom.
Williams ultimately got genetic testing from Athena after she received the amended report, and also had her exome sequenced as part of a research study, which revealed that she doesn't have an SCN1A mutation or mitochondrial defects. Williams is actively involved with families with Dravet syndrome kids, but said, "I wish I never knew about genetics. It's overwhelming, but it's the uncertainty that is the most overwhelming."