Do Higher Rx-Opioid Doses Raise Death Risks?

An analysis of opioid analgesic prescription patterns and deaths in a national sample of Veterans Health Administration (VHA) patients found an association of higher prescribed opioid doses with increased risks of overdose death. However, there were many limitations of this research and the risks were so small that one must question what, if anything, this study actually demonstrates.

Reporting in the April 6, 2011 edition of the Journal of the American Medical Association (JAMA), researchers culled a large national database of patients attending VHA facilities from 2004 through 2008 to construct a retrospective case-cohort study [Bohnert et al. 2011]. Cases included all unintentional prescription opioid overdose deaths (n = 750) during that 5-year period compared with a cohort of patients randomly drawn from individuals who used VHA medical services in 2004 or 2005: n = 154,684, 32% of whom had been treated with opioids for pain. Opioids had been prescribed for either cancer pain, chronic bodily pain, headache, neuropathy, or injuries/acute pain.

Subjects were almost all male (93%), white (83%), ranging in age from 18 to ≥70 years (with 81% being 40-59 years of age), the majority were treated for chronic bodily pains (78%), and approximately 40% had co-occurring substance-use disorders while 66% had other psychiatric disorders (possibly in addition to substance-use problems). Prescribed opioids included codeine, morphine, oxycodone, hydrocodone, and hydromorphone; methadone, buprenorphine, and fentanyl were not included in the analyses.

Based on outcome analyses, the researchers approximated the total rate of fatal overdose among individuals treated with opioids as 0.04% (1/2,500). Opioid overdose decedents were statistically significantly more likely to be (a) middle-aged and white, (b) have chronic or acute pain, (c) have a substance use and/or other psychiatric diagnosis, and (d) less likely to have cancer. [This corresponds directly with the largest subgroups represented in the population.]

The authors found that the overdose rate was greater at higher maximum daily opioid doses compared with lower maximum daily doses — that is, ≥100 mg/day versus 1 to <20 mg/day — across all subgroups examined. However, in 43.5% of overdose deaths (326 of 750 cases) the maximum prescribed dose was listed as being zero (0 mg/day), according to data in the report, and overall rates of overdose were greatest in patients with substance-use disorders.

Absolute overdose risk differences between opioids prescribed for as-needed (PRN) dosing, regularly scheduled dosing, or both were small and of little or no statistical significance. Data were unavailable to account for whether subjects actually were consuming their opioids as prescribed and/or the use of other drugs or alcohol that might have impacted overdose incidences. Also, the authors concede that some of the deaths may have been suicides misclassified as unintentional overdose.

The authors conclude, “The present findings highlight the importance of implementing strategies for reducing opioid overdose among patients being treated for pain. …. This study documents a relationship between opioid prescribing and opioid overdose in a large, national, prospective[sic] cohort of individuals receiving opioid therapy for a variety of medical conditions. The risk of opioid overdose should continue to be evaluated relative to the need to reduce pain and suffering and be considered along with other risk factors.”

COMMENTARY: While the relative risks are low, reported rates of overdose deaths associated with prescribed opioid analgesics have increased sharply in the United States during the past decade and this has become a pressing public health concern. Therefore, achieving a better understanding of factors contributing to such deaths is an important step toward addressing this problem. However, studies like this are unhelpful and possibly misleading, in our opinion.

We have previously examined studies addressing the impact of opioid-analgesic dosing on overdoses and/or deaths. In one, by Dunn et al. 2010 [discussed here] and also noted by the authors of the present JAMA report, overall incidence rates of overdose or death were small and there were many potentially confounding factors, such as mis-prescribing, co-occurring conditions, and polysubstance abuse. A more recent study, by Naliboff et al. 2011 [here], found that a liberal opioid-dose titration approach in a VHA population fostered better pain control while not incurring significantly greater substance misuse.

However, the quality of evidence from those two earlier studies and this present one should be rated as low and their external validity is doubtful. Perhaps foremost, this present study by Bohnert et al. and the Naliboff et al. trial focused on select VHA populations that were almost exclusively white middle-aged males with high prevalence rates of comorbid substance use and psychiatric disorders, placing the subjects at significant risk for potential problems. Along with that, Bohnert and colleagues do concede, “… the estimated overall risk of opioid overdose among individuals treated with opioids (0.04%) and the approximated absolute risk increases for significant associations, which ranged between 0.072% and 0.45%, were small. Opioid overdose death represents a particularly important outcome, but a rare one, and the findings should be interpreted accordingly” [emphasis added].

It also is curious in the present study that the greatest absolute number of overdose deaths (43.5%) occurred when the maximum prescribed daily opioid dose was listed as 0 mg/day. The authors had little explanation for this, other than many patients might have obtained opioids from non-VHA healthcare providers, and some might have saved opioids from a prior prescription or obtained them from illicit sources. These are important unknown factors that highlight an important failing of retrospective “data mining” approaches that rely upon vast storehouses of information that were not designed for the research at hand [previously discussed here].

In sum, while news media headlines touted this study as suggesting a dire threat of overdose and death from so-called “higher doses” of prescribed opioid analgesics, the actual data are unimpressive. And, while deaths associated with any medications must be taken seriously as public health concerns, it is hoped that better and less biased research will be forthcoming, comparing the consequences of undertreated pain with potential but rare risks of adequate opioid therapy in a broader population of properly selected patients.

REFERENCE: Bohnert ASB, Valenstrein M, Bair MJ, et al. Association Between Opioid Prescribing Patterns and Opioid Overdose-Related Deaths. JAMA. 2011;305(13):1315-1321 [abstract here].

ADDENDUM 4/16/11: During this past week another article similar to the above (except among public aid recipients in Canada) was published, concluding that higher opioid dosing for chronic noncancer pain is associated with greater opioid-related mortality. In this case, the researchers found that >200 mg/day of morphine (or equivalent) were associated with nearly 3 times the risk of opioid-related mortality compared with doses lower than 20 mg. This latest data-mining research has the same flaws and biases as the above study, and more. During the 10 years of the Canadian study, which identified more than 600,000 persons prescribed opioid analgesics and 593 associated fatalities in noncancer/non-palliative care patients, the annualized death rate was only 0.010%. Among the many confounding factors, benzodiazepines and/or alcohol were detected in 80% of decedents taking opioids. The authors concede that their results may have little external validity in broader populations. REFERENCE: Gomes T, Mamdani MM, Dhalla IA, et al. Opioid Dose and Drug-Related Mortality in Patients With Nonmalignant Pain. Arch Intern Med. 2011;171(7):686-691 [abstract here].