Screening Tool Helps Determine Opioid Overdose Risk

NATIONAL HARBOR —  A questionnaire designed to detect serious opioid-induced respiratory depression or overdose performed well in identifying patients at high risk for these life-threatening outcomes, study findings indicate. 

“There are a number of tools to assess the risk of developing opioid-related addiction and abuse, but nothing is available to provide real-time, evidence-based, practical assessment for patients at risk for the most serious side effects associated with opioids,” said Barbara K. Zedler, MD, an internist and chief medical officer at Venebio Group in Richmond, Virginia, during a late-breaking poster session at the American Academy of Pain Medicine 2015 meeting.

She and colleagues tested the validity of the  Risk Index for Overdose or Serious Opioid-Induced Respiratory Depression (RIOSORD) previously developed using  U.S. Veterans Health Administration data.

They conducted a retrospective, nested case-control study involving data in the largest U.S. commercial health claims database (IMS PharMetrics Plus) from more than 18 million U.S. patients who were prescribed an opioid from 2009 to 2013. “We know that the VA population may not be representative of the broader U.S. population, so we wanted to see how well the tool would perform in more typical medical users of prescription opioids.”

“Overall, 7,234 case patients experienced an overdose or serious opioid-induced respiratory event (OSORD) during the five-year period. Four control patients were assigned to each case patient. We then tested how well the VA RIOSORD model fit the IMS data. The model had a C-statistic of 0.90, which indicates excellent discrimination between cases and controls,” Zedler said.

“The variables most strongly associated with experiencing OSORD were very consistent with what we saw in the VA,” Zedler explained. “To make this a practical screening risk instrument that busy clinicians will actually complete, we narrowed 32 significant predictors to the 16 most associated with OSORD and scientifically supported in the literature.”

The predictors most strongly associated with OSORD in the general population included eight health conditions: substance use disorder, bipolar disorder or schizophrenia, stroke, and chronic headache as well as chronic kidney, lung or heart disease with impaired function, and nonmalignant pancreatic disease.

The remaining eight predictors were prescription drug-related and included opioid characteristics (fentanyl, morphine, methadone, hydromorphone, extended-release or long-acting opioids, and a total maximum prescribed daily dose of opioids of 100 mg or more morphine equivalents) and prescription benzodiazepine or antidepressant use.

The researchers assigned a point values ranging from 5 to 25 to each risk factor based on its weight in the model. Each patient’s RIOSORD score was the sum of points for their risk index items, Zedler explained. They modeled the total scores to produce predicted probabilities for experiencing OSORD.

Among the seven risk classes, the average predicted probability of an event ranged from 2% in the lowest risk class (RIOSORD score of 0-4 points) to 83% in the highest (42 points or more). “The observed occurrence of OSORD matched almost perfectly with what was predicted,” noted Zedler.

“The most important thing is to identify a patient’s risk for overdose, but just as valuable is using the individual patient’s risk profile to support clinician decision-making on a personalized basis,” she added.

Patients identified with elevated risk are most likely to benefit from risk-mitigation interventions, such as increased caution in opioid selection and dose escalation, close monitoring for OSORD and changes in risk factors, and a prescription for rescue naloxone to be administered by family members or caregivers in the event of an emergency.

Prospective evaluation of RIOSORD in real-world clinical settings and further refinement in patient subgroups are needed, according to Zedler.