Dr. Heilig Replies

To The Editor: Buprenorphine and methadone are both effective treatments for heroin dependence (1) . Counting studies on these medications is obviously not a valid method for comparing their efficacy. For retention in treatment, a meta-analysis yielded only a tendency-level advantage for methadone in high-dose studies (relative risk=0.79; 95% confidence interval [CI]=0.62–1.01) (1) . In flexible-dose studies, the relative risk was similar, but reached significance (relative risk=0.82; 95% CI=0.69–0.96). For drug use and criminality, the two treatments were reported to be equivalent. Thus, methadone provides a slight advantage over buprenorphine for retention in treatment, and the two medications are equivalent on other relevant outcomes.

Methadone treatment is essential, but also has distinct limitations. As pointed out by Drs. Byrne and Wodak, methadone is “more toxic,” i.e., methadone has sufficient mu-opioid receptor activity to induce lethal respiratory suppression, whereas buprenorphine, a partial agonist, does not. Safety itself aside, the monitoring necessitated by methadone use somewhat detracts focus away from building a therapeutic alliance, which offsets the cost advantage of the medication.

Given the complementary profiles of the two medications, pitching one against the other is not meaningful. The field needs rational ways of using both. In this context, we are unaware of any other therapeutic area in which a safer, albeit somewhat less effective medication, would be reserved for second-line treatment. Optimal balance between efficacy and safety is typically achieved by doing exactly the opposite. For example, few would consider using chloramphenicol for an infection before trying penicillin.

But what if stepping up treatment as needed rather than giving everyone methadone right away led to losing more patients overall? That would indeed mean that the safety gains must be carefully weighed against efficacy losses, an exceedingly difficult tradeoff. Our study was designed to assess whether this is a concern and clearly showed that it is not. Nothing is lost by first trying the safer medication.

In that perspective, the exact proportion of patients who ultimately transfer to methadone is irrelevant. But let us be correct. In our study, among 48 subjects randomly assigned to stepped treatment, 37 remained. Of those, 20 transferred to methadone. That is 54%, which is what we reported. The 65% given by Drs. Byrne and Wodak is a misrepresentation of our data.

In summary, excellent outcomes can be achieved by starting every heroin-dependent patient with buprenorphine and progressing to methadone only if needed. These outcomes are as good as those achieved with the best possible methadone treatment. Among unselected individuals addicted to heroin who are retained in treatment, close to one-half do well without progressing to methadone. Each of these individuals represents a safety gain worth capturing.

Finally, our study disclosed an unrestricted research grant from industry that accounted for approximately 25% of the budget. The remaining funding came from the Swedish Government and Stockholm County. It is unclear how this could invalidate our results. The reference cited (2) by Drs. Byrne and Wodak in support of this notion deals with meta-analyses, which our study clearly is not.