Scant Evidence for Long-Term Opioid Therapy in Chronic Pain

Pauline Anderson

January 13, 2015

There is insufficient evidence to determine the effectiveness of long-term opioid therapy for improving chronic pain, but emerging data support a dose-dependent risk for serious harms, such as overdose, mortality, and possibly fractures and cardiovascular events, according to a new review.

Most opioid-related studies lasted less than 12 weeks (and many less than 6 weeks), and no placebo-controlled trial continued for even 6 months.

"Unfortunately, a lot of these drugs are approved on the basis of short-term trials," Roger Chou, MD, associate professor, medicine, Division of General Internal Medicine and Geriatrics, Oregon Health & Science University, Portland, said in an interview with Medscape Medical News.

Dr Chou led the research group that prepared this review for a National Institutes of Health (NIH) workshop on the role of opioids in treating chronic pain. It is published online January 13 in Annals of Internal Medicine, along with the final report from the NIH workshop, also reported by Medscape Medical News.

Difficult Trials

Long-term controlled studies are difficult to carry out — they're relatively expensive, patients may not want to be put on a placebo, and many patients drop out of such studies. There's also not a strong incentive for the pharmaceutical industry to carry out such studies, said Dr Chou. "When you have a big market share, there's not a big incentive to conduct more studies," he said.

Chronic pain is often defined as pain lasting longer than 3 months or past the normal time for tissue healing. Chronic pain is common and a major cause of decreased quality of life and disability.

For the review, Dr Chou and his colleagues searched MEDLINE, the Cochrane Central Register of Controlled Trials, the Cochrane Database of Systemic Reviews, PsycINFO, and CINAHI for English-language articles published from January 2008 to August 2014. They also included relevant studies from a prior review, investigated reference lists, and searched

They looked for studies of adults with chronic pain who were prescribed long-term (lasting longer than 3 months) opioid therapy vs placebo, no therapy, another drug, or nondrug therapy. They also searched for studies of different opioid dosing strategies and of risk-mitigation strategies.

Although the investigators focused on long-term outcomes of opioid therapy, for overdoses and injuries they included studies with any duration of opioid prescription for chronic pain.

The review included 39 studies (in 40 publications). Five studies that assessed immediate effects of opioids used to treat acute pain exacerbations were omitted from the paper but were discussed in the full report (posted on the Agency for Healthcare Research and Quality site,

The authors found no study of opioid therapy versus placebo (no opioid therapy or nonopioid therapy) that evaluated long-term outcomes related to pain, function, or quality of life. Most placebo-controlled randomized trials were shorter than 6 weeks, and almost all lasted less than 16 weeks.

"We did not include uncontrolled studies for these outcomes; reliable conclusions cannot be drawn from such studies because of the lack of a nonopioid comparison group and heterogeneity of the results," commented the authors.

As well, no randomized trials evaluated opioid abuse, addiction, or related outcomes with long-term opioid therapy vs placebo or no opioid therapy.

Varying Definitions

Definitions of opioid abuse, addiction, and related outcomes and methods used to identify these events varied between studies, which was "a big problem," said Dr Chou. "Studies have not used standardized definitions."

Methods used to gather related information also varied, he said. "Some studies would do a really detailed interview of patients and spend an hour or 2 trying to get at how they were using their opioid. Others were kind of based on what was reported in the charts, so it was pretty ad hoc."

Another complicating factor was that randomized trials typically excluded patients at high risk for abuse, said Dr Chou. "A lot of studies will exclude patients who have any history of substance abuse or have significant psychological comorbidities. But we know that these patients are actually more likely to get opioids in the real world than people who don't have those kinds of risk factors."

The review collected better data on potential harm related to long-term opioid use. One explanation for the better data was that some of it came from administrative or "big payer" databases that could be linked to coroner or medical examiner reports, said Dr Chou.

One of the biggest areas of advancement since the last review of the topic about 5 years ago was collection of data on an association between opioid dose and overdose, said Dr Chou. "There are several studies that have consistently shown that patients who are prescribed higher doses of opioids are at higher risk for death," even after controlling for factors such as age, sex, use of other medications, and comorbidities.

Included in this review was a large, fair-quality retrospective cohort study of 9940 patients receiving opioid prescriptions within 90 days for chronic noncancer pain. It found that compared with nonuse, recent opioid use was associated with an increased risk for any overdose events (adjusted hazard ratio [HR], 5.2; 95% confidence interval [CI], 2.1 - 12.5) and serious overdose events (adjusted HR, 8.4; 95% CI, 2.5 - 28).

In this study, higher doses were associated with increased risk. Compared with the group receiving 1 to 19 mg/d, the adjusted HRs for overdose were 1.44 (95% CI, 0.57 - 3.62) for those getting 20 to 49 mg/d and 8.87 (95% CI, 3.99 - 19.72) for those taking at least 100 mg/d.

The problem with such studies, however, is that they're not randomized controlled trials, said Dr Chou. "The concern always is that maybe people prescribed higher doses are those who are sicker to begin with."

Dr Chou noted that for overdose, the risk is probably higher early on because patients haven't yet developed tolerance.

Data on fractures were weaker and less clear than those on overdose. A cohort study of adults aged 60 years and older found a higher fracture rate among opioid users (6%) than among nonusers (4%) after a mean follow-up of 33 months, although the difference wasn't statistically significant. A case-control study found that current opioid use was associated with increased risk for hip, humerus, or wrist facture compared with nonuse (adjusted odds ratio [OR], 1.27; 95% CI, 1.21 - 1.33).

The related data also suggested a timing issue, said Dr Chou. "If someone is given an opioid and they're not used to it, they may be more likely to fall. We don't have a whole lot of data to know for sure what that relationship is, but it's a reasonable thing to consider."

As for cardiovascular events, a cohort study that included 148,657 adult users of long-term opioid therapy found that a cumulative opioid supply of at least 180 days over a 3.5-year period was associated with an increased risk for myocardial infarction vs no long-term opioid therapy (adjusted incidence rate ratio, 2.66; 95% CI, 2.30 - 3.08).

The review also uncovered endocrine-related harms. One cross-sectional study of men with back pain found that compared with nonuse, long-term opioid use was associated with increased use of medications for erectile dysfunction or testosterone replacement (adjusted OR, 1.45; 95% CI, 1.12 - 1.87).

And there was some evidence pertaining to motor vehicle accidents. A study of 5300 cases found that doses of at least 20 mg/d were associated with increased odds of road trauma among drivers.

Dosing Strategies

There was little evidence on optimal dosing strategies. "There were a lot of differences in how opioids were dosed," so the information wasn't very useful in determining whether short-acting or long-acting drugs were more effective, said Dr Chou.

"We were taught that everyone with chronic pain should be transitioned to sustained-release, round-the-clock opioids, and it turns out that that's not based on any strong data."

The reviewers also looked at different opioid agents, including methadone, which Dr Chou called "unique" because, among other things, it has a relatively long half-life. Here, the best evidence came from a Veterans Affairs (VA) study looking at all-cause mortality in patients with chronic pain prescribed long-acting morphine vs methadone. It found that the overall risk for death was lower with methadone (adjusted HR, 0.56; 95% CI, 0.51 - 0.62).

"I don't think it necessarily means that methadone is safer than morphine or as safe as morphine; it might mean that in some settings they've implemented things to make methadone prescribing safer," said Dr Chou. "We need to understand what those things are."

The data on risk assessment tools were "pretty poor," commented Dr Chou. Two studies of the 10-item Opioid Risk Tool that reported sensitivity and specificity had estimates that were "all over the place," he said. He added that it's disappointing that "we can't estimate the accuracy of that tool with any kind of certainty."

As for the 14-item Screener and Opioid Assessment for Patients with Pain instrument, again, there were few data. Although two studies reported similar sensitivity, which was relatively reassuring to Dr Chou, the study quality was not high.

"We really do need a lot more research to understand how to use these tools, what their accuracy is, and then most importantly, how using them impacts clinical outcomes," he said.

Ideally, he added, clinicians should be able to use these tools to determine which patients are at high risk and which at low risk and then use that information to decide who to prescribe to, and how, to reduce harms. "But nobody has shown that; it's all hypothetical."

How we use these drugs has kind of blown up despite the fact that there is very little data. Dr Roger Chou

In Dr Chou's view, the dramatic increase in opioid prescribing has outpaced to some degree the evidence that's available. "How we use these drugs has kind of blown up despite the fact that there is very little data."

Recent media coverage of increased opioid use and the accompanying problems of overdose, abuse, addiction, and diversion may be having an effect, though. Some emerging evidence suggests that deaths related to opioid use may have plateaued, although it's still too early to determine trends, said Dr Chou.

Meanwhile, chronic pain appears to be on the rise. A North Carolina survey on the prevalence of low back pain showed that it more than doubled from 1992 and 2006. But according to Dr Chou, it's impossible to know whether factors such as an increasingly sedentary lifestyle are to blame.

Possible limitations of the review were that it excluded non–English-language articles, a meta-analysis couldn't be done, and publication bias couldn't be assessed. In addition to there being no placebo-controlled trials that met inclusion criteria, evidence was lacking for many comparisons and outcomes.

Reframe the Discussion

For comment, Medscape Medical News contacted Lynn Webster, MD, vice president, scientific affairs, PRA Healthsciences, Salt Lake City, Utah, and immediate past president, American Academy of Pain Medicine, whose research was included in the review.

Dr. Webster pointed out that although there's insufficient evidence to determine the effectiveness of long-term opioids, some data suggest that a subset of patients do improve with such therapy.  

"Most physicians who have treated patients with chronic opioid therapy know there are some people whose lives are better after treatment with opioids," said Dr Webster. "But since the risks are significant, we should reframe the discussion and ask why there isn't more research and investment in nonopioid analgesics. Let's not ignore the real crisis of the unmet need for better and more effective analgesics."

A research priority should be to determine how the human genome contributes to opioid response, he added.

"Not all opioids are alike and people may respond to any given opioid molecule differently," he said. "We need more research to determine the risk of harm and benefit before prescribing opioids, not after they have been prescribed."

He also urged caution in interpreting data that suggest that higher opioid doses are associated with more complications. "Higher doses can contribute to more risk, but higher doses usually are prescribed for people who have more pain. The higher levels of pain and other comorbid conditions may be the underlying reason for the increased complications seen at higher opioid doses."

Dr Chou reports grants from the Agency for Healthcare Research and Quality during the conduct of the study and consultancies for the US Department of Health and Human Services, the Physicians' Clinical Support System for Opioids (funded by the Substance Abuse and Mental Health Services Administration), the Mayday Foundation, the Collaborative Opioid Prescribing Education for REMS (funded by the University of Washington), and the University of Wisconsin outside the submitted work. Dr Webster has received consultation fees, advisory board honoraria, and travel expenses from various opioid manufacturers, including AstraZeneca, Mallinckrodt Pharmaceuticals, and others; he is a fulltime employee of PRA Healthsciences and does not receive funds from industry for research.

Ann Intern Med. Published online January 13, 2015. Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.