NEW YORK – Adding a polygenic risk score (PRS) to routine clinical care in England was well accepted by patients and physicians, and in several cases triggered changes in care plans, suggesting clinical utility.
Precision health company Genomics plc published the results of its Healthcare Evaluation of Absolute Risk Testing (HEART) Study last month in the European Journal of Preventive Cardiology as a step in a longer-term effort to make PRS an accepted clinical tool within the UK National Health Service (NHS).
"The heart study is the first trial that puts genetic risk tools into the hands of primary care physicians in the NHS," said Seamus Harrison, VP of global medical affairs and clinical development at Genomics plc.
"We have enormous advances in our understanding of the genetics of cardiovascular risk," he added, "but being able to operationalize this in primary care is the key challenge."
To assess the acceptability and potential utility of adding cardiovascular PRS to clinical care routines and whether its use changed treatment decisions, Genomics plc asked physicians in primary healthcare settings to administer both the commonly used QRISK2 tool and QRISK2 combined with a custom cardiovascular integrated risk tool (CVD IRT) that includes PRS information.
Physicians reported how easily they could incorporate the CVD IRT into routine primary care. As a secondary measure in the study, Genomics plc also assessed how the CVD IRT compared to QRISK2 in impacting physicians' patient care decisions.
QRISK2 is a multivariate risk score that takes into consideration traditional risk factors such as age, sex, systolic blood pressure, smoking status, and total serum cholesterol-to-high density lipoprotein ratio, as well as self-assigned ethnicity and conditions associated with cardiovascular risk, such as diagnosed type 2 diabetes, treated hypertension, rheumatoid arthritis, renal disease, and atrial fibrillation.
Genomics plc has developed PRSs for dozens of conditions but a key criterion for testing one in a clinical setting is the need to be able to take action that mitigates genetic risk.
"That's why cardiovascular disease was our first clinical study," Harrison said.
As an example, Harrison pointed to evidence that people with high genetic risk who take statins receive just as much benefit as people with environmental risk factors who take statins.
A review published in 2021, for instance, recommended stratifying people by genetic risk for CVD due to the large clinical benefit of statin therapy observed among them.
The HEART Study enrolled 836 participants at 12 sites in northeast England. Of those participants, who had a median age of approximately 55 years and were predominantly white (approximately 98 percent), 832 were eligible to receive a CVD IRT score, and 824 received a post-result questionnaire.
CVD IRT testing consisted of acquiring a patient blood sample, which was genotyped by Thermo Fisher Scientific using a custom Thermo Fisher Axiom array, followed by PRS and integrated cardiovascular risk scoring calculated by Genomics plc.
Use of the CVD IRT was deemed largely acceptable to patients, as approximately 87 percent of them replied that they were either likely or highly likely to recommend the test to friends or family in similar situations. Similarly, 17 of the 23 participating healthcare providers responded the same, overwhelmingly reporting that they could incorporate the test into their routine primary care in a straightforward manner.
Although overall mean QRISK2 and CVD IRD scores were highly correlated, researchers noted some marked differences at the individual patient level. Approximately 5 percent of participants were up-classified to high risk by CVD IRT and another 5 percent were down-classified to low risk.
These shifts appeared to impact clinical decision-making. In approximately 28 percent of cases where the participant’s CVD IRT score was greater than their QRISK2 score, physicians reported that the change influenced their management decision.
"Physicians told us [that] this made the conversation about starting treatment in otherwise healthy people much easier because it was more personalized," Harrison said.
"CVD is arguably the most promising indication for incorporation of PRS into clinical practice because the PRS is reasonably informative compared to other indications, and there is an existing absolute risk model with associated clinical management guidelines in widespread clinical use," said Anna Lewis, a bioethicist at Harvard Medical School who was not involved in the study.
Lewis said that the HEART implementation study showed that logistical barriers to incorporating PRS into routine care could be overcome in the primary care setting in the North of England.
She noted, however, that this acceptance came from an overwhelmingly white population and that the acceptability shown by this group may not generalize to others.
The study itself does note that attitudes to genetic testing vary, with minorities typically being more skeptical.
"Equity considerations have been central in the literature of the possible implementation of PRS in clinical practice," she said, "because of the differential predictive performance of PRS in different groups."
Lewis also drew attention to the study design, noting that it is not a randomized, controlled trial and as such, does not demonstrate that sharing PRS information with patients improves morbidity or mortality.
"An open question is whether such studies will or should be deemed necessary for widespread clinical adoption," she said.
Harrison agreed that this is a critical point for future studies and adoption, noting that typically, tools used to predict cardiovascular diseases, such as QRISK2, aren't subject to long-term randomized trials prior to implementation in the same way novel drugs are. Rather, he wrote, the clinical utility of novel prognostic tools is measured by the impact they have on important clinical decisions.
"In the HEART study," Harrison said, "when the combined score –– with genetics and clinical factors ––– was greater than the clinical score alone, healthcare providers planned changes of management in 27.8 percent of cases. This is strong evidence that the tool can influence clinical outcomes through supporting clinical decisions. As PRS can be applied across all the major common chronic diseases, we need to determine the optimal implementation scenarios to deliver the best outcomes in the most cost-effective way."
Lewis added that the existence of a widely used absolute risk model with associated clinical management guidelines was key to the study, and there are few indications which enjoy that benefit, making the results not generalizable.
Harrison agreed, noting that the HEART study demonstrates feasibility only in this particular use case.
"I think there will be learnings which are applicable across diseases," he said, "but we'd have to take it on a use-case or disease-by-disease basis to understand the best implementation scenarios within the different healthcare settings that we're working in."
Finally, Lewis noted that the published study gave very few details on the informed consent process.
"What fraction of approached individuals opted into this study?" she asked. "What level of understanding of PRS did the participants have at the time of consent? Prior work shows that the details of the consent process matter hugely to enrollment, so this is a really key part of the workflow to better understand."
Harrison replied that as the trial was observational, the company didn't maintain a screening log to know what fraction of people opted in out of those who were presented with the option.
"Patients were invited to participate in the study when they were invited to attend routine health checks in the NHS," he said via email.
Harrison commented, however, that participating clinical investigators reported high levels of enthusiasm for inclusion, which he found "particularly noteworthy," as the trial was initiated toward the end of the COVID-19 pandemic.
Although all patients were required to be able to consent for the study, Harrison commented that they were not required to have any specific knowledge of PRS. After returning the results, however, nearly 95 percent of participants reported feeling that the information presented was easy to understand, and approximately 99 percent found the information useful. Furthermore, Harrison said that Genomics plc provided physicians with educational content throughout the study.
Although Genomics plc plans to commercialize the CVD IRT, it is doing so in measured steps.
"At the moment, we're not commercializing in the United Kingdom," Harrison said. "We are commercializing in the United States as part of a wellness program with a commercial partner. And we're also working with healthcare systems in both the US and the UK to understand what the best scenario for commercialization would be."
Harrison declined to name the company's US commercialization partner but said that it would be disclosed in an upcoming announcement.
Meanwhile, the company is working with the NHS on the Our Future Health initiative, a public health research project seeking to recruit 5 million people to develop new ways to prevent, detect, and treat diseases.
"Genomics will be working with Our Future Health to calculate polygenic risk scores for a range of diseases," Harrison said. "Most of these will be used in a research context, but in a smaller number of diseases, we'll return this risk [information] back to individuals in a way similar to what we've done in HEART."
Harrison said that it was "exciting" to be able to integrate an understanding of genetic risk with routine clinical pathways in the NHS and to translate that understanding of genetic risk to a real-world setting.
"Now that we have feasibility that this can be operationalized," Harrison said, "we're really excited about developing it in a range of use cases and over and above cardiovascular disease alone."