番茄社区

One of the enduring enigmas in medicine is why patients with osteoarthritis (OA), whose joints are clearly diseased enough to cause substantial pain, nevertheless find relief with sugar pills. Now a meta-analysis from China has unraveled some of the factors associated with placebo responses in clinical studies of oral drugs.

Analyzing data from 130 trials involving 12,673 OA patients, not only were placebo responses strongest in those with the most severe symptoms at baseline (as expected from previous studies), but the magnitude of response also increased with a study's sample size and the potency of the active agent being evaluated.

This was true not only for pain but also for ratings of stiffness and functional ability, according to Lujin Li, PhD, of Shanghai University of Traditional Chinese Medicine, and colleagues, .

On the other hand, the analysis yielded no significant association between the strength of placebo responses and any of the following factors that had previously been suggested as influencing placebo responses:

• Funding source
• Radiographic disease severity ()
• Dosage form (tablet, caplet, capsule)
• Participants' race or ethnicity
• Year of publication

Moreover, Li and colleagues found that placebo responses in studies of oral medications typically approached their maximum after 5-7 weeks of treatment, meaning that trials shorter than 8 weeks probably underestimated these responses.

The findings should help inform designs of future trials, the researchers said, as well as clinical decision-making.

Trials included in the meta-analysis were published from 1999 to May 2022. The vast majority had corporate sponsorship, but Li and colleagues determined that most had low risk of bias in terms of blinding and outcome reporting. Many, however, did probably come with bias related to baseline differences between the placebo and active-drug groups.

Key to the analysis was the use of "model-based" meta-analytic techniques. This, the researchers said, "can quantitatively and accurately describe the time-course association of the placebo response and identify various influencing factors in clinical trials."

Li and colleagues were particularly interested in establishing the trajectory of placebo responses over time, which they said had not been a focus of previous research but is important in understanding why promising drugs so often fail to prove superior to placebo in randomized trials.

Li and colleagues noted that had found astonishingly strong placebo responses in OA trials, in the realm of 70-80% reductions from baseline in pain, stiffness, and disability scores. The current analysis showed more modest improvements: 11-22% in pain as assessed in the Western Ontario and McMaster Universities OA Index (WOMAC) system, 15-18% in stiffness, and 12-15% in disability.

Absolute magnitude of symptom relief was strongly related to baseline scores, such that patients with more pain, stiffness, and disability showed greater improvements. This was also true for the percentage improvement in pain -- the 11% reduction was seen in patients rated at WOMAC level 15, while the 22% decrease was in patients with baseline WOMAC scores of 35 -- but less so for stiffness and disability, where percentage improvements did not differ so much.

Another important finding was that placebo responses were greater when synthetic drugs were the focus -- such as nonsteroidal anti-inflammatory drugs (NSAIDs), acetaminophen, or diacerein -- compared with "natural" herbal and plant extracts. Li and colleagues attributed that finding to "participants' psychological expectation of the tested efficacy of the drugs." Previous NSAID use was also a factor, with smaller placebo responses when more than half versus less than half of participants had been taking these agents (e.g., -4.5 vs -7 WOMAC pain score points).

Limitations to the analysis included the lack of no-treatment control groups in most of the included studies. Most of the reports also had no information on many details of patients' disease or other potentially relevant factors, such as "the time and attention that the study staff devoted to the participants," Li and colleagues noted.

Also, because the analysis focused only on oral medications, the results cannot be generalized to biologic drugs or other injectables.

Comment