Chronic use of tramadol after acute pain episode: cohort study

Dear Editor
We thank the respondents for their responses to our work and would like to address some of their concerns and further engage in this important conversation.

Dr. Newman was concerned that our population base is Medicare patients, who are more medically complex and more likely to experience chronic pain than the general population. We would like to clarify that our population included both Medicare and commercially insured enrollees and that we saw no indication that our population had unusually high rates of chronic pain compared to the general population receiving medical care. In fact, our population likely included fewer people with chronic pain because we excluded anyone with any opioid fills in the prior 6 months.

Dr. Newman notes that without data on overdose or abuse, we are unable to conclude that tramadol is a drug of abuse. We agree. We do not make any claims in this study that tramadol is a drug of abuse.

Regarding the comment on whether 7% continued use is important: we would note that our study included only people who were not previously taking opioids, who had received a curative/therapeutic procedure performed on an elective basis, with no other procedures on the same day, short pre- and post-operative stays, discharged to home, and with no other surgical procedures in the follow-up period. Readers must make up their own minds on whether this is a high rate of continued use. We consider the absolute rate of continued use to be higher than expected for this cohort.

Dr. Newman also writes of her concern for people with chronic pain who have difficulty receiving treatment. We share her concerns and agree that all patients who are suffering deserve patient-centered, responsive, and humane treatment, including access to scheduled drugs where that is appropriate and effective. We neither believe nor argue that all people on long-term opioid therapy are “drug abusers” or that they are addicted to their medication. Opioid use disorder is a disease that is characterized by maladaptive behaviors that negatively affect a person’s social, emotional, and physical functioning. People taking opioids as directed by their health care provider and whose functioning is improved by their medication should certainly be able to continue this treatment.

Finally, Dr. Newman notes that scheduling tramadol will not stop people from dying. We would like to clarify that tramadol is already a scheduled substance in the United States—it falls in schedule IV. Our conclusion encourages policymakers to consider reclassifying tramadol from schedule IV to a higher schedule, as was done for hydrocodone combination products in 2014, for example.

Dr. Mastroianni notes that tramadol is intended to treat chronic pain and is surprised that it is being used in a post-op setting. She is concerned that rescheduling tramadol will force people to use street drugs. As noted above in response to Dr. Newman, our population was limited to patients without any opioid fills in the 6 months prior to surgery, which excluded all patients taking tramadol in the pre-operative period for chronic pain. Further, in our manuscript we note that tramadol is included in the post-surgery prescribing guidelines at many institutions, including our own. As was noted in the accompanying editorial, there is a sense among many providers that tramadol is an “opioid-lite”—and therefore better for patients than hydrocodone or oxycodone. This has been a perception from the first introduction of tramadol in the US in 1995. This may be because before being metabolized, tramadol is indeed only a very weak opioid. However, as Dr. Mastroianni no doubt knows, once it passes through the liver, tramadol’s main metabolite has 700 times the affinity for the mu-receptor as the parent drug. Regardless of the reasoning, there is a persistent sense among many prescribers that tramadol is not a real opioid, and that is reinforced by the fact that until 2014, it was not a scheduled substance in the US. It remains unscheduled in many countries.

Dr. Hochman has two concerns about the study design and a conclusion that tramadol presents a similar or greater risk of opioid dependence.
1. Groups were not matched and there may be preferential use of tramadol for patients whom physicians fear are at greater risk for OUD.
We did not see evidence that people with substance use disorders were more likely to receive tramadol vs. other short-acting opioids. In fact, in each of the three insurance coverage groups (commercial, Medicare Advantage aged 65+, and Medicare Advantage disabled/<65 years), the Elixhauser drug abuse diagnosis codes were equally prevalent among those receiving tramadol alone and those receiving tramadol plus another short-acting opioid. (Additional analysis after publication) In the Medicare populations, the prevalence of the drug abuse diagnosis codes were also similar (not statistically distinguishable) between the tramadol groups and the group receiving no opioid fill. Commercial beneficiaries with drug abuse diagnosis codes were more likely to receive tramadol than no opioid fill. These data are certainly not conclusive evidence of prescribers’ medical reasoning, but they suggest that an individual patient’s risk of future abuse is not the major driver of tramadol prescribing.

2. Persistent treatment for pain is not the same as abuse or addiction. Without a non-addictive control group, the study cannot conclude addiction or abuse is responsible for continued use.
We do not at any point in this study conclude addiction or abuse is responsible for continued use. We completely agree that our study cannot and does not address this question.
We would, however, encourage Dr. Hochman to review Tables 2 and 3, which report the later opioid use of people who did not receive an opioid at discharge (19.5% of the total study population) for some limited evidence on the experience of people who received over-the-counter or non-opioid pain medications after surgery. This group experienced a lower, but non-negligible risk of all three definitions of prolonged opioid use compared to those receiving an opioid.

Finally, a drug needn’t be the preferred street drug in order to be problematic, to have the risk for dependence, or to cause harm.