The risks of opioid use for elderly patients with nonmalignant pain vary considerably by different agents and by different durations of use, according to an analysis of Medicare databases.
Patients taking codeine for more than 180 days are at increased risk for cardiovascular events, and those taking oxycodone or codeine for only 30 days are at increased risk for any-cause mortality, said Dr. Daniel H. Solomon and his associates in the rheumatology department and the pharmacoepidemiology division at Brigham and Women's Hospital, Boston.
“This study's findings do not agree with a commonly held belief that all opioids are associated with similar risk,” the investigators noted.
They compared the safety profiles of different opioids for the treatment of nonmalignant pain in elderly patients because “relatively little attention has been paid” to this issue even though the use of these drugs has risen by 50%-100% in recent years. In contrast, patients and physicians are relatively well informed about the toxicities of NSAIDs used for the same indications.
Dr. Solomon and his colleagues analyzed information in two states' Medicare databases of pharmaceutical coverage for low-income patients (mean age, 79 years) between 1995 and 2005. They matched 6,275 subjects who took five of the most commonly prescribed opioids for a variety of baseline factors using propensity scores.
Hydrocodone was used as the reference exposure, to which codeine, oxycodone, propoxyphene, and tramadol were compared.
The risk of cardiovascular events including MI, stroke, heart failure, revascularization, and cardiac death was similar across the five opioid groups after 30 days of use. However, by 180 days the cardiovascular risk with codeine was significantly higher (risk ratio 1.62) than with the other four opioids.
This finding is surprising and requires validation in other study populations, the investigators said (Arch. Intern. Med. 2010;170:1979-86).
All-cause mortality was elevated after only 30 days of use for patients taking oxycodone (RR 2.43) or codeine (RR 2.05), but not for those taking other opioids.
In contrast, the risk of fracture of the hip, pelvis, wrist, or humerus was significantly reduced after 30 days of treatment for patients taking tramadol (RR 0.21) or propoxyphene (RR 0.54)
The risk of gastrointestinal adverse events – including upper GI bleeding, lower GI bleeding, and bowel obstruction – did not differ across opioid groups.
These risks remained consistent through a range of sensitivity analyses either 30 days or 180 days and regardless of whether patients were taking low, medium, high, or very high doses of the drugs.
View on the News
Opioids Are Not 'Interchangeable'
The findings by Dr. Solomon and his associates “challenge the conventional notion that the safety profiles of opioids are generally interchangeable” and carry two important clinical implications, said Dr. William C. Becker and Dr. Patrick G. O'Connor.
The first and most crucial implication is that the frequent use of codeine must be reexamined. “The untested but widespread assumption that codeine is safer from an addiction standpoint because of its lower potency may need to give way to these data demonstrating increased risk of cardiovascular events and all-cause mortality. If codeine is of middling efficacy for pain and is more risky than other opioids, there would be little reason to use it,” they noted.
Secondly, the elevated risk of fractures with opioid use “has solid biological plausibility” by two mechanisms of action: Opioids raise the rate of falls, and they suppress the production of androgen and estradiol, imperiling bone health. The study findings suggest that basic safety measures to counterbalance these effects are not being implemented.
“Efforts to improve patients' understanding of safe medication-taking practices, providers' understanding of safe prescribing practices, and standardization of safety-oriented follow-up are sorely needed,” said Dr. Becker and Dr. O'Connor.
DR. BECKER and DR. O'CONNOR are internists at Yale University, New Haven, Conn. They reported no relevant financial disclosures. These comments were taken from their invited commentary accompanying Dr. Solomon's report (Arch. Intern. Med. 2010; 170:1986-8).