Overdose from prescription opioids (e.g., Oxycodone, Hydrocodone) has become one of the most common causes of accidental death in the United States. Two new articles in BMJ suggest that overdose is not the only risk about which patients, prescribers and policy makers should be concerned.
Khademi and colleagues conducted a prospective study of a cohort of 50,045 Iranians. They followed up over 99% of the sample and then assessed the impact of opium use on mortality. After statistically adjusting for cigarette smoking, education, age and other factors, the research team reported that opium use nearly doubled the risk of death. The number of diseases with increased incidence among opium users was large, and included tuberculosis, cancer, and chronic obstructive pulmonary disease. The results held even when the researchers excluded from analysis individuals who started using opium in response to the onset of a chronic illness.
These results do not necessarily generalize to prescription painkillers such as Oxycontin. Unlike opium, which comes directly from the poppy flower, modern, synthetic opioids are free of impurities are not prescribed in a smoked form. Further, opium use in Iran may be a marker for other risk factors (e.g., poor self-care habits, social isolation) for which the epidemiological study could not fully adjust.
That said, in an accompanying commentary in BMJ, Dhalla notes that preliminary studies have found indications of higher death rates in patients with the same medical condition who take opioid medications versus non-steroidal anti-inflammatory drugs. The higher mortality risk is not simply attributable to accidental overdoses. None of these studies of prescription opioids is definitive, but they certainly justify a larger replication research effort along the lines of the Iranian study of opium users.
The worrisome fact about prescription opioids is that their use has grown extraordinarily rapidly in a short period in the United States, to over 200 million prescriptions in 2010 alone. As a result, any adverse impacts of opioids that take a few years to accrue may hit the population in a tidal wave before there is time to understand and prevent the damage.
Cross-posted at Stanford Medical School’s SCOPE Blog.
” . . . included tuberculosis, cancer, and chronic obstructive pulmonary disease.” Doesn’t this suggest that the danger comes from smoking opium? Or were the Iranians taking the drug in some other way?
The data I am aware of indicate that oral consumption of opium is common in Iran. Nonetheless, the range of diseases with heightened risk makes me think that some factors that contribute to globally bad health (e.g., poverty, poor health habits) may have been not fully controlled by the study (and perhaps as you note smoking among those who consumed that way was not fully adjusted for either. As you know all covariate measures have error in them and in that sense tend to under control). That’s another reason I consider the findings intriguing but not definitive.
Is there a suspected causal relation? I know opioids slow down the digestive tract, dull the senses more, etc. I’m specifically wondering about the higher death rates for opiate pain killer use. A study would have to be carefully designed to ensure fair sorting based on severity: I imagine that when opiates are prescribed, it’s likely a more severe instanced of said condition, with higher mortality inherent.
Yes — one would have to start within one illness rather than go across them (e.g., opiates for cancer, NSAIDs for lower back pain). That is what the osteoarthritis study did, but that of course is just one study with one condition. A better study would have, say, five common diseases for which opiates are one possible therapeutic choice in populations where we are at equipoise as to whether that is the best therapeutic choice or not, and then randomize within disease to opiates or an alternative regimen before following people over time.
This is a question, not a political statement. I had heard that previous reports on accidental death counted all deaths not in a hospital where the person had traces of an opiate in their system whether or not a high dose led to death. Is that true in these reports?
Doing a full toxicological exam on deceased persons costs money and many states/counties do not invest money in it. The best example is someone who is drunk and dies of an overdose or suicide or car accident also has, say, oxy and benzos on board as well — detecting the alcohol is easy and that is often all that is done in death reporting. I would therefore take overdose data on opioids as conservative.
Keith