The ghost of Heisenberg, or: methamphetamine, again

Abstract

M any of you will recall the television series Breaking Bad, which ran from 2008 to 2013. Whether you were a fan or not, it was hard to escape it as a cultural phenomenon, gathering accolades and provoking next-day recapitulations around countless office water coolers (remember office water coolers?). For those who somehow missed it, the story revolved around an erstwhile hapless chemistry teacher, Walter White, who finds his inner sociopath by becoming a major methamphetamine producer. The series captured a certain zeitgeist of illicit drug use in the early 21st century, a time when methamphetamine abuse was the focus of considerable attention. But while Walter White has long since departed from the scene, the methamphetamine epidemic has not. Especially in the American Midwest and West, methamphetamine remains a major, and now skyrocketing, cause of overdose deaths, relatively obscured to those of us in the East by the opioid epidemic. And unlike opioid use disorder, which can be addressed by several evidence-based treatments (both agonists, such as methadone and buprenorphine, and antagonists, such as naltrexone), there are no Food and Drug Administration-approved treatments for methamphetamine dependence. That, in turn, reflects the absence of established treatments for any form of stimulant dependence. Having worked on pharmacological approaches to cocaine abuse earlier in my career, I can state with great humility that the challenges in this area are daunting. It’s against this backdrop that we have to consider the controlled clinical trial by Trivedi and colleagues1 appearing in the Jan. 14 issue of the New England Journal of Medicine and covered in our lead article in this month’s Update. They studied the efficacy of a combination treatment consisting of extended-release injectable naltrexone plus oral bupropion in 403 participants with moderate or severe methamphetamine use disorder. The sequential parallel comparison design they used is a bit complicated, so I’ll cut to the chase: the active treatment was more effective than placebo at the p<0.001 level, more than five times more effective in terms of relative response rates. Moreover, this study had a large sample size, low attrition rate, stateof-the art assessment methods, and high protocol adherence by participants — features that are frequently missing in studies of this disorder. So: problem solved, right? If only. That five-times-higher response rate I mentioned? It was 13.6% (the response to placebo was 2.5%). Imagine this conversation with a prospective patient in your office: “Is it effective? Yes, well, this treatment gives you almost a 1 in 7 chance of responding. Oh, and did I mention that you’ll have to come in every three weeks to get your naltrexone injection?”