NEW YORK — The Center for Personalized Therapeutics at the University of Chicago is assessing whether pharmacogenomics can tailor drugs for patients before, during, and after surgery.
The clinical trial, dubbed Implementation of Pharmacogenomic Decision Support in Surgery, or ImPreSS, grew out of the center's recognition that while pharmacogenomics has been gaining traction in certain areas of medicine, such as psychiatry, it hadn't been considered much in the perioperative setting, said Peter H. O'Donnell, deputy director of UChicago's Center for Personalized Therapeutics (CPT).
The center, which researches and develops systems for integrating pharmacogenomics into clinical care, kicked off its first clinical trial in 2011 with the 1200 Patients Project, in which outpatients across multiple medical specialties preemptively underwent pharmacogenomic testing to inform future prescriptions. The center has since launched pharmacogenomic implementation projects focused on certain specialties like cardiology and oncology.
While the 1200 Patients Project provided insight into how physicians use pharmacogenomic test results and the feasibility of incorporating such testing into routine clinical care, the impact on patient care in the outpatient setting was modest, since most patients aren't starting new medications during a visit, according to O'Donnell. That led researchers to question whether there was greater potential to show pharmacogenomics' usefulness in settings — during a surgical encounter, for example — where patients are exposed to multiple medications they've never received previously.
Although it's uncommon for surgical patients to get pharmacogenomic test results indicating they're at high-risk for a drug-related adverse event, if something does go wrong, the results can have significant patient safety implications. For example, patients with pharmacogenomic variants in the BCHE gene can have an increased risk of prolonged apnea or of experiencing paralysis after being administered certain muscle relaxants used in the operating room.
Perioperative medicine is a "high-stakes, high-prescribing environment," O'Donnell said.
UChicago has wrapped up the first phase of a pilot of the ImPreSS trial and published on the feasibility of pharmacogenomic testing in this setting in Anesthesia & Analgesia in November. The second phase of the project currently underway involves randomizing 1,800 patients to receive either pharmacogenomic testing to inform perioperative prescribing or to standard care, without pharmacogenomics. Researchers will compare the rate of use of genomically discordant drugs during perioperative care and how frequently providers access pharmacogenomic results, as well as assess pain management and the rate of adverse drug events between the two study arms.
As long as patients are receiving pharmacogenomic testing within the context of this trial, the associated costs will be covered by a grant from the National Institutes of Health. This means that the center won't bill patients' insurance for testing, which is good news for patients since not all payors would cover the cost.
Limited insurance coverage of pharmacogenomic testing has historically been a barrier to integration. O’Donnell acknowledged this remains a challenge in this space, though he is optimistic that studies like ImPreSS will eventually demonstrate the type of evidence insurers want to see for coverage.
In the meantime, since its founding in 2010, CPT has garnered several millions of dollars in funding from UChicago, the NIH, and philanthropy. It's put the resources toward building the infrastructure necessary to integrate pharmacogenomic testing into research projects like ImPreSS, which has received grants from the NIH and the Benjamin McAllister Research Fellowship. With institutional funds, CPT established an in-house clinical laboratory within the pathology department that performs pharmacogenomic testing and developed the Genomic Prescribing System (GPS), a web portal and pharmacogenomic clinical decision support system that is integrated into the electronic health record (EHR) system.
EHR-based clinical decision support is critical for executing the type of implementation studies CPT is undertaking and for measuring the impact of perioperative pharmacogenomics. Within the ImPreSS pilot, 71 patients undergoing elective surgeries provided a blood draw or cheek swab sample for genotyping. Results were brought into the GPS portal before surgery and anesthesia providers were alerted to their availability within the EHR.
In the portal, providers received recommendations about how pharmacogenomic variants detected in patients would impact their ability to metabolize drugs that may be prescribed in the preoperative, intraoperative, recovery, and hospitalization phases, as well as at discharge. These recommendations are based on research reviewed by the CPT team and align with guidelines from the Clinical Pharmacogenetics Implementation Consortium.
Overall, providers prescribed 2,371 drugs to patients during the pilot. Each patient received a median of 35 medications of which a median of seven drugs had associated pharmacogenomic results — most of which were already congruent with providers' prescribing decisions. There were relatively few instances of pharmacogenomic results that would alter prescribing, the study authors reported.
Providers accessed pharmacogenomic results for 41 — or 58 percent of — procedures. In follow-up surveys, 70 percent of providers who accessed GPS said they had enough time to evaluate results and 76 percent said the portal was simple to use. The most common reason for not accessing results in GPS was not remembering to do so, since pharmacogenomics is not part of providers' typical clinical workflow, according to the study authors.
"Our pilot data for result access rates suggest interest in pharmacogenomics by anesthesia providers, even if opportunities to alter prescribing in response to high-risk genotypes were infrequent," the authors wrote in their paper.
In the randomization phase of the ImPreSS trial, researchers are also making pharmacogenomic results available to surgeons to inform prescriptions for postoperative care and are hoping to encourage greater use of this information by embedding pharmacists within some clinical care settings to support prescribing decisions and by targeting patients who might need intensive care or dedicated pain care after surgery.
But even if this part of the study shows the utility of this testing, healthcare systems will weigh a number of factors when assessing whether to invest in pharmacogenomic testing infrastructure, such as how common high-risk pharmacogenomic variants are in the population and current methods for avoiding drug-related adverse events.
In an editorial accompanying the Anesthesia & Analgesia paper on the pilot results, two anesthesiology professors questioned the utility of widespread pharmacogenomic testing in the perioperative setting, given that adverse reactions to anesthetic drugs based on genetics may prove rare and that developing clinical decision support systems like GPS is likely expensive.
In CPT's pilot, researchers tested patients for variants in more than a dozen genes that affect drug response or toxicity. For example, researchers tested for variants in the CYP2D6 gene that have been associated with nonresponse to opioids like tramadol, a pain medication often used in postoperative hospitalization and at discharge. Normal metabolizers have a response rate of 78 percent for tramadol, compared to 53 percent for poor metabolizers, the study authors wrote in the paper. Among the patients enrolled in the pilot, between 69 percent and 73 percent were normal metabolizers and 7 percent were poor metabolizers.
The most prescribed drug with pharmacogenomic results in the pilot was sevoflurane, a general anesthetic used in the operating room. Sevoflurane has been linked with susceptibility to malignant hyperthermia in less than 1 percent of individuals who have mutations in the RYR1 gene. Researchers did not observe any patients in the pilot with such mutations.
There weren't many instances in the pilot of patients having pharmacogenomic risk associated with drugs they were prescribed. Moreover, in the intraoperative setting, medications that were administered to patients aligned with pharmacogenomic recommendations, aside from one patient who received succinylcholine despite having an increased risk for prolonged apnea; the patient did not have an adverse outcome.
"Because of the rarity of those outcomes, which are, of course, potentially of great health significance to patients, it's not clear that genetic testing is the most effective approach," said Stuart Forman, a professor of anesthesiology at Harvard Medical School who coauthored the editorial. He noted that patients are already asked questions about family history of adverse reactions to general anesthesia during preoperative evaluation.
Perhaps pharmacogenomic recommendations could be more useful for surgeons managing postoperative care, he suggested, as opposed to anesthesiologists prescribing during surgery, where patients are already only exposed to drugs for a limited time and are closely monitored.
There are also workflow considerations, said Ken Johnson, a professor in the anesthesiology department at the University of Utah, who also coauthored the editorial. After pharmacogenomic testing, it can take between a few days to up to two weeks to get results, which could be a barrier for patients whose surgeries are scheduled quickly. In the pilot, patients were only eligible if their procedure was at least two weeks away to ensure their providers would have test results available in the EHR. CPT has since decreased its genotyping turnaround time to one week, the study authors noted in their paper.
Anesthesia providers also tend to administer drugs and then document them afterward, so they might not see EHR-based clinical decision support alerts before prescribing, Johnson said. Still, pharmacogenomics "has a lot of potential down the road to improve safety and personalize how we take care of patients," he said. "We're still trying to figure out what it is and how it might be useful."
While Johnson and Forman expressed doubt in their editorial that genomic testing will be widely adopted to inform perioperative drug prescribing in the near term, pharmacogenomics could become more broadly applied — particularly in large healthcare systems — as testing becomes more affordable and accessible. And anesthesiologists may benefit from greater access to pharmacogenomic results as more genomic testing is performed and results are added to patients' medical records for other reasons.
"When that situation emerges, systems like GPS that inform anesthesiologists about potentially impactful gene-drug interactions will likely be implemented, regardless of the frequency of life-threatening interactions," they wrote in the editorial.
O'Donnell said he does see a place for perioperative pharmacogenomics and hopes that — if research shows improvements to patient outcomes and utility for providers — other healthcare systems and providers will adopt it.
The CPT team has approached each of its pharmacogenomic implementation projects with the idea that if a program is determined to be useful, the next step would be figuring out how to make it a sustainable part of clinical care, O'Donnell said. "Will it be for certain patient populations? Certain care settings? Certain drugs?” he added. "How will it become standard of care? That remains to be seen."