This report re-examined and extended upon the seminal report of Bachhuber et al. 2014. The findings may be considered a partial replication. Their result that states with MC from 1999-2010 had elevated opioid mortality, relative to those without MC, was confirmed with the inclusion of seven additional years (Figure 1A). Their finding of opioid overdoses decreasing in the years subsequent to adopting MC as compared to states that did not, was not replicated. In fact, states that previously adopted MC had a significantly greater overdose slopes than those that did not (Figure 1B). The preclinical (Smith et al. 1998) and epidemiological evidence base (Bradford & Bradford, 2017, 2018; National Academy of Sciences, 2017; Stith et al. 2018) indicates that it is plausible that MC could decrease opioid use with its synergistic anti-nociceptive properties (Abrams et al. 2011; Cooper et al. 2018) and therefore limit misuse and overdose potential. However, investigations like Bachhuber et al. 2014 assume that overdose determinations are consistently made by medical examiners and coroners across the US (Supplemental Figure 1). At the very least, any differences should be random and not systematically different based on MC laws. An additional key finding was that this assumption was not warranted (Figure 1C) in that states without MC had lower overdose death determination reporting quality using the CDC criteria (Rudd et al. 2016).

The continued pronounced elevations in opioid overdoses, typically involving poly-substances (Simpson et al. 2019) is concerning. These findings, although largely contradictory to Bachhuber et al. 2014), should not be used to discount that the risk to benefit ratio may still favor MC as another tool that may be endorsed by some chronic pain patients, especially relative to the overdose potential of prescription opioids. As others have recently noted (Shover et al. 2019), the ecological research design where the unit of analysis is the state might not be best-suited to address this question. State-level MC systems, either established through voter initiative or through their state legislation, may differ on a wide-variety of socio-political characteristics. For example, Medicaid expansion was much more common among MC+ states. The additional resources provided by Medicaid could have provided for additional buprenorphine (Wen et al. 2017) which initially decreased opioid overdoses (Schwartz et al. 2013). Later, the potency of fentanyl and many fentanyl analogues (Simpson et al. 2019) could overcome any cannabis substitution effect (Piper et al. 2017). The increase in China white variety of heroin which is more common in the eastern US (Mars et al. 2016) and is more susceptible to contamination with white fentanyl than black tar heroin likely also contributes to regional differences in the recent overdose surge. This may be crucial because the majority (83.3%) of the MC+ states were east of the Mississippi.

Some caveats and future areas of study should be noted. This report was based on US overdoses as reported by the CDC. Five states had more than one-third of their death certificates where the substance involved was unspecified (Buchanich et al. 2018). The non-uniformity of autopsy procedures (e.g. When did each state, county, or city start testing for fentanyl? Were screens or the more expensive confirmatory testing employed? How many of the two-hundred fentanyl analogues were tested in each municipality and when?) is a substantial challenge to research, and barrier for empirically informed public policy which utilizes this information. Although this would limit the power to detect differences, future research might consider focusing only on the subset of areas where the death determination procedures have been consistently high (e.g. Simpson et al. 2019). This study investigated MC and may not generalize to recreational marijuana laws.