Major Finding: Opioid users had higher rates of overall adverse events, severe adverse events, cardiovascular events, fractures, and all-cause mortality than did NSAID users or coxib users.

Data Source: A propensity score–matched cohort analysis involving 12,840 elderly patients taking opioids, NSAIDs, or coxibs for rheumatoid arthritis or osteoarthritis pain.

Disclosures: The study was supported by the Agency for Healthcare Research and Quality. Dr. Solomon reported being an unpaid member of a celecoxib trial executive committee sponsored by Pfizer and an unpaid member of the data safety monitoring board for an analgesic trial sponsored by Pfizer.

Opioids were associated with more risks than were other analgesics in elderly patients taking the drugs for arthritis pain.

Although NSAIDs are known to pose certain risks, the results of the study “support the safety of [NSAIDs] compared with other analgesics,” said Dr. Daniel H. Solomon and his associates at Brigham and Women's Hospital, Boston.

Few studies have examined the relative risks of the three major analgesic groups: NSAIDs, opioids, and coxibs (selective cyclooxygenase-2 inhibitors). “Postmarketing surveillance data from usual care cohorts provide an opportunity to examine comparative safety across a wid-range of events and can complement safety data from randomized controlled trials. However, imbalance in baseline population characteristics confounds many postmarketing surveillance studies,” Dr. Solomon and his colleagues noted (Arch. Intern. Med. 2010;170:1968-78).

“Propensity score–matched analyses may provide better balance of confounders and facilitate relatively straightforward comparative safety analyses,” they added.

To compare safety, the researchers performed a propensity-matched cohort analysis using information from a Medicare database of pharmaceutical coverage for low-income elderly residents of Pennsylvania and New Jersey in 1999-2005. The study population included 12,840 adults with rheumatoid arthritis or osteoarthritis who began using one of the three types of analgesics during the study period and were followed for at least 1 year.

Overall, adverse safety event rates were high for all three groups, with the rate of adverse events leading to hospitalization being greater than 100 per 1,000 person-years for all three types of analgesics. Opioid users had the highest rates of serious adverse events. In particular, the rate of hip, pelvis, wrist, or humerus fractures was 101 per 1,000 person-years in the opioid group, compared with 19 per 1,000 person-years in the coxib group and 26 per 1,000 person-years in the NSAID group.

Even though a link between opioids and fractures has been reported previously, “the strength of the association we observed is larger than in previous reports,” Dr. Solomon and his associates said.

Compared with NSAIDs, opioids (hazard ratio, 1.77) and coxibs (HR, 1.28) were associated with elevated risk for cardiovascular events such as MI, stroke, heart failure, revascularization, and cardiac death. That “unexpected” finding regarding such severe events warrants further study, the investigators noted.

Compared with NSAIDs, risk for upper GI bleeding, lower GI bleeding, or bowel obstruction was similarly high for opioid users (HR, 1.07) but was lower for coxib users (HR, 0.60).

In general, opioid users experienced the most adverse events over time, while NSAID users experienced the fewest.

In addition, “opioid users experienced moderate risk early in treatment,” the investigators noted, while the other groups did not.

Opioid users had significantly higher all-cause mortality (75 deaths per 1,000 person-years) than did either NSAID users (48 deaths per 1,000 person-years) or coxib users (47 deaths per 1,000 person-years).

View on the news

Oversight in the Statistical Methods

The propensity matching was meticulous in this study, leading to well-balanced baseline characteristics among the three treatment groups.

This reassures readers that the observational study design is as robust as possible, and that treatment effects alone account for the observed differences among opioid, NSAID, and coxib users.

However, there is a single unmeasured confounder that calls into question the validity of some of the study findings: the use of over-the-counter NSAIDs, said Dr. William C. Becker and Dr. Patrick G. O'Connor.

It is likely that “a significant proportion” of patients in the opioid group were also taking NSAIDs, because “physicians routinely recommend antiinflammatory medication in addition to opioids to achieve therapeutic synergy in the treatment of arthritis,” they noted.

“It seems implausible that a group of 'opioid-only' elderly patients” who were actually taking supplemental NSAIDs would have a higher risk of adverse cardiovascular events, GI bleeding, and acute kidney injury than would a group taking NSAIDs alone, because these are all known effects of NSAID therapy.

Despite this limitation, “the data on falls and fracture from this … study are nonetheless compelling and carry important clinical implications,” Dr. Becker and Dr. O'Connor said.

DR. BECKER and DR. O'CONNOR are in general internal medicine at Yale University, New Haven, Conn. They reported no financial disclosures. These comments were taken from their invited commentary accompanying Dr. Solomon's report (Arch. Intern. Med. 2010;170:1986-8).

Elderly With Arthritis: Opioids Riskier Than Other Analgesics