In reply
We thank McCarberg and colleagues for their thoughtful comments on the difficult issue of optimal use of opioids for patients with chronic pain. They make 2 general points, first that opioids in general and high-dose opioids specifically are not as dangerous as we claimed, and second, opioids are an effective treatment for chronic pain. McCarberg and colleagues state that overdose death is “an important but rare outcome” of opioid use and “probably only a secondary factor at most” in the cause of opioid overdose death. We respectfully disagree. The reference that MCarberg and colleagues provide to document that opioid use is a rare outcome is based on a large case-cohort study from the Veterans Health Association. In that study, the rate of opioid overdose death among patients taking any dose of opioids was estimated to be 0.04% over an approximately 3-year period.1 But the hazard ratio for patients taking dose equivalents of 100 mg or more of morphine compared with those taking low doses was 7.2, indicating that the rate of opioid overdose death among these patients was approximately 0.3%, or 3 opioid-related overdose deaths per 1000 patients per 3 years—1 per 1000 per year.