Share this article on:

Landmark Trial Punctures the Myth That Opioids Provide Powerful Relief of Chronic Pain

doi: 10.1097/01.BACK.0000520970.46118.bc

The deadly opioid overtreatment epidemic picked up steam in the late 1980s and early 1990s with the misguided notion that opioids are painkillers that can be used safely and effectively in the long-term treatment of chronic back pain—or other forms of noncancer chronic pain.

The intervening years—and as many as 300,000 deaths in the related opioid overdose epidemic—have rebutted the idea that opioids can be used safely on a mass basis. More than 30,000 Americans will die in 2017 as the result of opioid overdoses. And the opioid overdose epidemic is still increasing in ferocity in many quarters.

Now a groundbreaking new study suggests that opioids may not be powerful pain relievers at all, at least in the long-term management of chronic pain.

In the first randomized controlled trial (RCT) with long-term follow-up comparing opioids with non-opioid medications, Erin E. Krebs, MD, and colleagues from the Minneapolis Veterans Health Care System found that opioids provided no better pain relief for patients with low back pain or painful osteoarthritis than safer analgesics such as nonsteroidal anti-inflammatory drugs and acetaminophen—and other nonopioid pain medications.

“Opioids are perceived as strong pain relievers, but our data showed no benefits of opioid therapy over non-opioid medication therapy for pain,” said Krebs in presenting the unpublished study at the 2017 meeting of the Society for General Internal Medicine (SGIM) in Washington, DC. (See Krebs et al., 2017.)

Opioids provided no advantage in terms of function at the 12-month follow-up mark, and patients in the opioid wing of the study actually reported marginally more pain at 12 months than those in the nonopioid group.

“The data do not support opioids' reputation as ‘powerful painkillers,’” said Krebs. “The results support CDC [Centers for Disease Control and Prevention] guideline recommendation: that non-opioid medications are preferred for chronic pain.”

They also support the recent recommendation in the American College of Physicians guideline that opioids should be an uncommon treatment—a treatment of last resort—for patients with low back pain.

After Krebs presented the study at the annual meeting of the SGIM in Washington, DC, it sparked immediate comments across the Twittersphere. “Long-term opioids no better than Motrin for low back pain at 12 months. Just don't start them,” tweeted Steven Asch, MD, of Stanford University.

Asch is Professor of Medicine at Stanford and Chief of Health Services Research at the VA Palo Alto Health Care System. He offered additional comments via email.

“This study fills a gaping hole in the literature. Many have asked, ‘Do opioids work for chronic musculoskeletal pain?’ but until now, there have been few if any high-quality trials. Now we know better. The best evidence argues against the use of chronic opiates for patients with long term back or other musculoskeletal pain.”

Asked if he had any advice for patients and physicians in the wake of this study, Asch suggested caution. “Think twice before starting opiates in chronic musculoskeletal pain, then think again. They will work for a while, but this study says they don't do any better than Motrin or other [non-opioid drug] therapies in the long run.”

Many had argued that long-term RCTs comparing opioids to non-opioids would be impossible to perform, because patients with chronic pain wouldn't want to be assigned to the wing receiving non-opioid pain medications. Some even suggested it was unethical to enroll patients in a RCT where they wouldn't receive opioids. The new study suggests that both these concerns were misplaced.

Asch suggested that the new study provides a lesson for researchers—to persevere even in difficult research areas. “The lesson is that difficult studies like this are worth doing if they can change practice.”

Back to Top | Article Outline

Important Milestone in Opioid Research

Roger Chou, MD, of Oregon Health and Science University, coauthored the CDC Guideline on Opioids for Chronic Pain and was lead author of the evidence reviews supporting the recent American College of Physicians Guideline on Non-invasive Treatments for Low Back Pain. (See Dowell et al., 2016; Chou et al., 2017.) He also believes this is an important milestone in opioid research.

“This is an important study because it is the first trial to look at long-term outcomes of opioids vs. non-opioid therapy for chronic pain (in this case, LBP and painful osteoarthritis of the hip or knee). It was done in the context of a collaborative care model, which I think is also a plus,” according to Chou via e-mail. “It goes beyond previous trials by showing the lack of even short-term benefits in pain or function, with slightly worse pain intensity at 9–12 months. This is striking because patients were not blinded in this trial—i.e., the patients on non-opioid therapy knew they were not on opioids, yet they still reported better long-term pain intensity.

“The short-term placebo-controlled trials of opioids vs. placebo suggested that we should see modest short-term effects on pain intensity. I think this trial suggests that even these relatively small benefits are not observed when patients are managed in a collaborative care model. Patients improved with respect to pain and function in both arms of the trial. The trial also confirmed that, despite the lack of benefits, that opioids are associated with more side-effects,” Chou wrote.

He also believes it is an important research achievement and may open the door to further research. “I think the trial shows that a long-term trial can be done and that there was equipoise among participants,” according to Chou. “It should put to rest concerns that it is unethical to enroll patients in similar trials going forward,” he suggested.

“The trial should open the door to more research,” said Chou. He pointed out there is still a need for research in many areas. “It would be useful to know if there is a way to identify opioid “responders” and to investigate how factors such as degree of central sensitization may impact the effectiveness of opioids. We also need more data on the other side of things—i.e., randomized trials of tapering vs. continued opioids in people who are on high doses, but don't meet clear criteria for tapering—this is a big challenge for many clinicians.”

Back to Top | Article Outline

Study Confirmed Growing Skepticism About Opioids

“This is an impressive study,” said internist Anil Makam, MD, MAS. Makam is a primary care researcher and clinician at the University of Texas Southwestern Medical Center.

“It is the first clinical trial comparing opioid and non-opioid medications with long-term follow-up. It provides strong evidence that opioids should not be the first line of treatment for chronic musculoskeletal pain, given that there were similar changes in pain and function with non-opioids,” said Makam in a telephone interview.

“I think a lot of people at the conference were impressed with this study, as it confirmed the clinical impressions of many physicians that opioids are not as effective as advertised.

“There is a growing evidence base from randomized trials, other prospective trials, and observational studies on the risks and benefits of opioids in the long-term treatment of chronic pain,” Makam noted. “This evidence base suggests that opioids are not achieving the benefits for which they are marketed. And everyone is now well aware of the adverse effects of opioids.

“There are always trade-offs between the benefits and risks of treatments applied in clinical practice. But trials like this are making us question the overall impact of opioids. We don't appear to be seeing the intended benefits of opioids. So why incur the downstream costs and adverse events?” he asked.

Makam suggested some of the strong points of the study are the randomized trial design, the long-term followup, and the pragmatic nature of the study. “That is really compelling,” said Makam. “The researchers didn't compare fixed doses of opioids to fixed doses of non-opioids. Choices of medications and dosage were left up to patients and physicians. And there was flexible dosing of the medications to achieve adequate pain control.” This is an important issue, he suggested, and makes the study relevant to real-world clinical practice.

But Makam also introduced a note of caution regarding the new RCT. “This was a meeting proceeding,” he noted. The study has not yet entered the peer review and publication process. “The devil is often in the details” when it comes to unpublished studies presented at meetings, he explained.

“I think the main findings are going to hold.” he commented. But he has some reservations about the generalizability of this trial.

The study diagram presented at the meeting drew his attention. “I think any time you see a study that goes from 4500 patients being contacted to 250 patients being enrolled and randomized, it creates some concerns.

“About half the subjects who were contacted weren't eligible. This speaks to the generalizability of the inclusion/exclusion criteria. They restricted the study to a fairly narrow group of patients with chronic pain. And then, a sizeable number of patients who were eligible didn't want to participate.

“So only 5% of the people who were initially contacted were ultimately enrolled and randomized. This raises a fundamental question: ‘What about the other 95% of patients? Would we see a similar effect of opioids in people with chronic pain who were not randomized?’ ” asked Makam. He suggests waiting for the final published paper to see how these issues are addressed.

Back to Top | Article Outline

Study Aimed to Fill a Major Gap in the Evidence on Opioids For Chronic Pain

Krebs et al. performed their study because of looming questions about the risks and benefits of long-term opioid therapy for chronic pain. “Over the past two decades, improved clinical attention to pain has been associated with exponentially greater use of long-term opioid therapy for chronic non-cancer pain, both within and outside the VA system,” noted the researchers in a posting of the study design at

“Despite this change in practice, the proper place of opioids in chronic pain management continues to be controversial because research has not demonstrated the long-term safety and effectiveness of opioids for chronic musculoskeletal pain,” they added. (See VA Office of Research and Development, 2017.)

Despite a modest recent decline in numbers of prescriptions, opioids are still among the most heavily prescribed medications in the United States. In 2013, about 30% of patients with back pain received a prescription for an opioid. (See Mafi et al., 2013.) The recent CDC guideline on opioids for chronic pain estimated that 20% of patients with chronic pain were prescribed an opioid. (See Dowell et al., 2016.) A recent study in the BMJ listed the 25 most prescribed drugs in the US in 2013. The opioid combination hydrocodone/acetaminophen was number one, and two other opioids made the top 25. (See Greenway and Ross, 2017.)

Back to Top | Article Outline

What Do Systematic Reviews Say About the Long-Term Effectiveness of Opioids?

Several recent systematic reviews have highlighted a major gap in the evidence on opioids for chronic back and other forms of chronic pain. Their effectiveness over the long-term is completely unknown. In preparation for its recent guideline, the CDC recently commissioned a systematic review to evaluate the long-term impact (≥ 1 year) of opioids for chronic pain. Deborah Dowell, MD, and colleagues performed an extensive literature search and could not find a single RCT with long-term follow-up.

“No study of opioid therapy versus placebo, no opioid therapy, or nonopioid therapy for chronic pain evaluated long-term (≥1 year) outcomes related to pain, function, or quality of life,” according to Dowell et al. “Most placebo-controlled randomized clinical trials were ≤6 weeks in duration.” (See Dowell et al., 2016.)

Chou and colleagues performed an evidence review in preparation for the recent American College of Physicians guideline on noninvasive management of low back pain. They also could find only short-term trials addressing the effectiveness of opioids in relieving pain and enhancing function. “For opioids, evidence remains limited to short-term trials showing modest effects versus placebo for chronic low back pain.” (See Chou et al., 2017.) The most recent Cochrane Collaboration systematic review on opioids for low back pain came to a similar result. (See Chaparro et al., 2014.)

Back to Top | Article Outline

Multicenter Pragmatic RCT

To fill this gap, Krebs and colleagues performed a pragmatic RCT involving patients from multiple clinics in the Veteran's Administration Health Care system. To qualify for inclusion, subjects had to have moderate-to-severe pain despite analgesic use. The study excluded individuals who were already on long-term daily opioid therapy.

The researchers' working hypothesis was that patients in the opioid group would experience greater improvement in pain intensity and pain-related function than those in the nonopioid group. But they expected that opioids would cause more medication-related symptoms and adverse events than nonopioid medications.

The research team contacted 4485 patients. Roughly 41% of the potential study subjects declined to join the study. Approximately 53% did not meet the study inclusion criteria. This left 240 patients who were randomly allocated to opioid therapy or nonopioid drug therapy.

The study subjects were predominantly male—only 13% were women. The mean age of the opioid group was 56.8 years and the nonopioid group 59.7 years. A little over half the subjects in both groups had a primary complaint of low back pain.

Patients in each wing of the study had access to a flexible menu of drug options. Step I in the opioid group was morphine IR (the preferred option) or hydrocodone, Step II involved morphine slow release or oxycodone sustained action (i.e., time-released opioids), and Step III was transdermal fentanyl patches. The study protocol limited the daily dose of opioids to 100 morphine equivalents per day.

In the nonopioid arm, Step I offered acetaminophen (the preferred option) or naproxen, Step 2 the tricyclic antidepressant nortriptyline, topical capsaicin (an ointment derived from hot peppers), or the antispasmodic medication gabapentin (Neurontin), and Step 3 involved the atypical opioid analgesic Tramadol.

The VA research team explained the drug strategy and collaborative care model at “Medications in each arm will be adjusted to target improvement in pain, while considering individual patient preferences and responses. Interventions will be delivered in a [telecare] management model using the randomly assigned prescribing strategies, automated symptom monitoring, and a structured decision-making approach to guide medication adjustment. Outcome assessors masked to treatment assignment will conduct interviews to assess patient-reported outcomes at 0, 3, 6, 9, and 12 months and will assess physical performance and cognitive function at 0, 6, and 12 months,” they wrote. There was an exceptional level of follow-up in the study. Only a handful of patients were not available for long-term follow-up.

Back to Top | Article Outline

Study Results Confounded Expectations

The study results, needless to say, confounded the researchers' expectations. As mentioned above, the patients in the opioid group had no advantage in the primary outcome measures. In terms of pain intensity, members of both groups had a mean baseline score of 5.4 (out of 10, 10 indicating maximum pain severity). At 12-month follow-up, the mean score fell to 4.0/10 in the opioid group. However, the nonopioid group exhibited greater progress at 12 months, with a mean score of 3.5/10. The differences were statistically significant.

The nonopioid group had a larger proportion of patients with clinically significant improvement in Brief Pain Inventory (BPI) pain severity scores: 53.9% compared with 41% in the opioid group.

Both groups made similar improvements in terms of pain interference with function. The opioid group had a mean baseline BPI interference with function score of 5.4 (out of 10, 10 signifying maximal interference with function). This fell to 3.4/10 at 12-month follow-up. The non-opioid arm had a mean baseline score of 5.6/10, which fell to 3.3/10 at 12-month follow-up. Roughly 60% of both groups had clinically significant improvements in function, as measured by the BPI interference with function measure.

The opioid group had a higher level of total adverse symptoms and total medication-related symptoms. Fifty percent of members of the opioid group made an emergency department visit over the 12-month follow-up period vs. 39% in the non-opioid arm. Equal proportions of both groups—18%—had one or more hospitalizations.

Disclosures: None declared

Back to Top | Article Outline


Figure. No caption a...
Chou R, et al, Systemic pharmacologic therapies for low back pain: A systematic review for an American College of Physicians clinical practice guideline [published online ahead of print February 14, 2017], Annals of Internal Medicine. doi:10.7326/M16-2458.
Dowell D, et al, CDC Guideline for Prescribing Opioids for Chronic Pain—United States, 2016, JAMA, 2016;315(15):1624–45. doi:10.1001/jama.2016.1464.
Greenway T, Ross JS, US drug marketing: How does promotion correspond with health value? BMJ, 2017; 357:j1855. doi:10.1136/bmj.j1855.
Johnson C, The other tragedy behind Prince's death,, July 11, 2016; http://
Krebs EE, et al, Effectiveness of opioid therapy vs. non-opioid medication therapy for chronic back & osteoarthritis pain over 12 months. Presented at: the annual meeting, Society for General Internal Medicine, Washington DC, 2017, as yet unpublished.
Olson J, Minnesota VA study challenges wide use of opioids, Star Tribune, May 4, 2017; http://
Mafi JN, et al, Worsening trends in the management and treatment of back pain, JAMA Internal Medicine, 2013; 173:1573–81. doi:10.1001/jamainternmed.2013.8992.
Portenoy RK, Foley KM, Chronic use of opioid analgesics in non-malignant pain: Report of 38 cases, Pain, 1986; 25:171–86.
Porter J, Jick H, New England Journal of Medicine, 1980, 302(2): 123.
Qaseem A, et al, Noninvasive treatments for acute, subacute, and chronic low back pain: A clinical practice guideline from the American College of Physicians [published online ahead of print February 14, 2017], Annals of Internal Medicine. doi:10.7326/M16-2367.
VA Office of Research and Development, Strategies for Prescribing Analgesics Comparative Effectiveness Trial (SPACE),, 2017;
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.