Brief trials, safety omitted

The FDA approvals for opioids over the past two decades appear rushed and lacked information on safety considerations, according to a new cross-sectional analysis of new drug approvals (NDAs) submitted to the agency.

“We found that approvals for chronic pain indications were generally based on a few trials of no more than 12 weeks, and few approvals for chronic pain included or referenced pooled safety analyses that incorporated systematic assessments of opioid-associated risks, such as tolerance, drug diversion, and nonmedical use,” James Heyward, MPH, from the Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, and colleagues wrote in the Annals of Internal Medicine. “Many trials for products indicated for chronic pain used an [enriched enrollment randomized withdrawal (EERW)] design with a run-in period to exclude persons who did not respond to or otherwise tolerate a product. These findings are important because of the central role that prescription opioids have played in driving the opioid epidemic, as well as the FDA’s vital role as the regulator of the pharmaceutical marketplace in the United States.”

To that point, the study authors noted that opioid overdose was responsible for the deaths of more than 46,000 Americans in 2018, with prescription opioids responsible for 15,000 of those deaths. Despite the decline in opioid sales since 2010, “per capita consumption in the United States remains far above pre-epidemic levels, as well as levels in other countries.” And, while heroin and fentanyl now account for most overdose deaths, prescription opioids appear to be the entry drug, Heyward and colleagues noted.

In their analysis, Heyward and colleagues culled data from ClinicalTrials.gov, FDA reviews, and peer-reviewed publications to assess the evidentiary thresholds used by the FDA in the approval of prescription opioids in phase III pivotal trials. The NDAs selected were of opioid products approved for the treatment of acute and chronic pain from 1997 to 2018.

They also looked at the “key characteristics of each NDA, including the number, size, and duration of pivotal trials; trial control groups; the use of enrished trial populations; and systematically measured safety outcomes.”

Among their results:

  • 47 NDAs were for new dosages (n=25; 52.1%) or new formuations (n=9; 18.8%).
  • 1 NDA was for a new molecular entity (NME) for tapentadol.
  • 48 NDAs were supported by 47 pivotal trials (median 1 pivotal trials [interquartile range, 0-1 trials]).
  • The number of pivotal trials differed depending on NDA type, with tapentadol supported by 2 pivotal trials; 21 of 25 trials for new dosage forms had at least 1 pivotal trial, which was also the case for 5 of 6 new drug combinations.
  • It was rare for a new formulation of previously approved drugs to have a new pivotal trial (1 of 9), as well as for products previously marked without an NDA (0 of 6).
  • Of the 21 products approved for the treatment of chronic pain, the trials had a median duration of 84 days (IQR 25-84 days), with a median enrollment of 299 patients (IQR, 174 to 525 patients).
  • 17 of the trials for the treatment of chronic pain used the EERW design.
  • In 81% of the NDAs for the treatment of chronic pain, the pivotal trials excluded patients who couldn’t tolerate the drugs, had early adverse effects, or did not report immediate benefit.
  • Of the 39 NDAs approved for the treatment of chronic pain, 10 did not include a safety review.
  • 38.5% monitored drug tolerance, 20.5% monitored diversion, 17.9% measured aberrant use, and 12.8% monitored overdose symptoms.
  • For acute pain, there were 19 trials with a median duration of 1 day (IQR, 1-2 days) and there were a median of 329 patients enrolled (IQR, 199-456 patients).

Heyward and colleagues noted that, “among NDAs for chronic pain, 8 (20.5%) included pooled safety reviews that reported systematic assessment of diversion, 7 (17.9%) reported systematic measurement of nonmedical use, and 15 (38.5%) assessed development of tolerance.

“Although all but 1 of the 48 approved NDAs were for previously approved moieties, analysis of available NDAs for referent products yielded similar findings,” the study authors noted.

Heyward and colleagues offered some recommendations the FDA could undertake based on their findings:

  • “[B]olster existing regulatory guidance for opioid applications to include explicit requirements regarding populations, duration of therapy, and efficacy and safety outcomes measured.”
  • “Information regarding frequently occurring adverse events, such as tolerance, mausea, and vomiting; symptoms of systemic overdose …should be systematically defined and assessed in new pivotal trials rather than gathered on the basis of open-ended adverse event reports alone.”
  • “The FDA should stop relying on EERQ designs to assess opioid efficacy…”
  • “Given the paucity of evidence regarding the long-term safety of opioids included in the NDAs we examined, the FDA should revise the labeling of products used for chronic pain by removing claims about ’daily, around-the-clock, long-term’ efficacy ….to more suitably align it with the existing clinical trial evidence.”
  • “Important safety information, especially for outcomes too uncommon to be measured reliable in a randomized controlled trials, such as aberrant use, addition, and diversion, should be captured robustly during the postmarketing period …”

The study authors noted that their study has several limitations — animal studies and nonpivotal trials were excluded, and the safety evidence for NCAs was only presented for chronic pain trials.

  1. The FDA’s NDAs for opioids for treating pain were generally based on a few trials of no more than 12 weeks.

  2. Few approvals for chronic pain included or referenced pooled safety analyses that incorporated systematic assessments of opioid-associated risks, such as tolerance, drug diversion, and nonmedical use.

Candace Hoffmann, Managing Editor, BreakingMED™

Heyward disclosed no relevant relationships.

Cat ID: 925

Topic ID: 915,925,730,192,144,725,925