Commentary on Cousins et al. (2016): Accumulating evidence on risk of mortality on and off opioid substitution treatment.

Abstract

The National Irish cohort study strengthens evidence on the hazards of leaving opioid substitution treatment (OST) and dangers of short duration treatment and provides the first piece of evidence for future syntheses of the effect of supervised consumption on mortality risk in people on OST We know that methadone and buprenorphine are ‘essential medicines’ for the treatment of heroin and other opioid dependence and the prevention of drug-related harm 1-3. There is good evidence from multiple studies that patients have better outcomes, including a reduced risk of mortality during opioid substitution treatment (OST) 4-8. Heroin and other opioid use, especially if injected, are inherently risky, and the risk of mortality is heightened at specific times—such as in periods of low tolerance to the effect of opioids immediately after prison release or treatment discharge, or when taken in combination with alcohol, benzodiazepines and other drugs 6, 9-12. Clinicians and policymakers need evidence both on how best to deliver treatment and manage opioid dependence in order to reduce the mortality risk and number of drug-related deaths in the population and, critically, on whether changes to drug treatment have had an impact upon mortality risk and other health outcomes. As in many situations, randomized controlled trials (RCT) may not be best placed to provide this evidence. Typically they are underpowered to detect small, but important, effects on relatively rare outcomes such as mortality. Therefore, policy makers are increasingly reliant upon evidence from large clinical administrative databases of patients in receipt of interventions in the real world. For instance, recent studies have shown how the combination of OST and HIV can reduce HIV and all-cause mortality among HIV-positive opioid-dependent people; that initiation onto buprenorphine may reduce elevations in mortality seen during induction onto methadone; and direct criminal justice referral into treatment may lead to reduced mortality risk during treatment compared to other routes into treatment 13-16. The study by Cousins and colleagues uses administrative data on the delivery of OST in primary care in Ireland to investigate two issues 17. The first is whether treatment transitions, the period at the start and immediately after the end of episodes of opiate substitution therapy, are associated with increased mortality risk. The second issue tests specifically whether ‘supervised consumption’ (i.e. direct observation of methadone or buprenorphine consumption by the pharmacist) is associated with a reduction in mortality risk. The study corroborates the extreme transitory risk in mortality in the period immediately following treatment cessation and highlights a problem with treatment duration and retention 5-7. If the mortality risk is as high as 4% (four per 100 person-years) in the first month after treatment—eight times higher than risk during treatment and four times higher than other periods out of treatment—then it is highly likely that treatment duration needs to be longer than the mean or median 168 and 83 days (5.5 and 2.7 months), respectively, that were provided to patients in National Irish cohort 17. Cousins and colleagues do not find evidence for the increased mortality risk at the beginning of treatment observed in other settings 5, 6, 15, 18. However, as the authors point out, this may be because a proportion of patients in their database initiated their treatment in specialist services before being transferred to primary care. Ecological evidence from Great Britain suggests that after supervised consumption was introduced patient safety was improved, as measured by the OD4 index that calculates the number of methadone related deaths per amount of methadone prescribed 19. However, there was no direct evidence that people exposed to supervised consumption have a reduced mortality risk. This study provides weak evidence in favour of supervised consumption, and forms the first piece of evidence which can contribute to a future synthesis of the effect of supervised consumption including, as the authors highlight, the impact upon the diversion of methadone. We need and look forward to more studies such as that by Cousins et al., so we can develop ‘stratified medicine’ for treatment of opioid dependence and can pool experience and evidence from different health-care systems in order to tailor specific treatments and ensure that ‘the right patient gets the right treatment at the right time’. M.H. and P.V. have received unrestricted research grants as co-applicants from Gillead, and received honoraria to speak at conferences from Abbvie, Gillead, Merck and Jansenn. L.D. has received untied educational grants from Reckitt Benckiser and Mundipharma.