In Korea's pharmaceutical and biotech sector, the "placebo effect" is again drawing attention as the biggest variable in late-stage clinical trials. Cases are mounting in which the efficacy of a drug candidate is confirmed but the placebo group shows greater-than-expected symptom improvement, preventing statistical significance.
On the 28th, according to the pharmaceutical and biotech sector, tension is rising after Kolon TissueGene(950160)'s knee osteoarthritis treatment TG-C, which underwent a U.S. phase 3 trial, failed to demonstrate a meaningful difference versus placebo. With many domestic corporations also set to announce large-scale late-stage results in the second half, including for osteoarthritis and Alzheimer's disease, observers say the placebo effect will again emerge as a swing factor.
◇ TG-C again exposes the "placebo wall"
Kolon TissueGene said on the 20th that in the U.S. phase 3, its knee osteoarthritis treatment TG-C did not secure statistical significance versus the placebo group. Although both pain and joint function improved in the TG-C arm, a similar level of improvement appeared in the placebo group, effectively erasing the difference between the two arms.
The company cited a stronger-than-expected placebo effect as the main reason for the failure. Chief Executive No Moon-jong of Kolon TissueGene said at a media briefing the next day, "We confirmed TG-C's osteoarthritis improvement effect itself, but we could not demonstrate differentiation because the placebo group also showed significant symptom improvement," and added, "The placebo effect, generally known to diminish after six months, unusually persisted up to 24 months in this trial." The company is conducting additional analyses, including by clinical site and subgroup, to identify why the placebo response increased.
A placebo is a so-called "dummy pill" with no therapeutic effect, used as a control to evaluate a new drug's efficacy. Especially in diseases like osteoarthritis, where patients self-assess pain or symptoms, not a few feel better due to the expectation of receiving treatment or the psychological impact of an intra-articular injection itself. As a result, even if actual efficacy is confirmed, the gap with the placebo group can narrow, making it difficult to secure statistical significance.
This is not the first such case in Korea. Helixmith(084990) failed to secure significance versus placebo in the U.S. phase 3 of its diabetic neuropathy treatment candidate Engensis in 2019, causing major delays to development. It then failed to meet the primary endpoints in phases 3-2 and 3-2b, and phase 3-3 has yet to begin.
Shin Poong Pharm also failed to demonstrate superiority over a comparator in last year's phase 3 of its osteoarthritis treatment SP5M002. CHA Vaccine Research Institute (now AriBio LAB(261780)) likewise is redesigning its follow-up clinical strategy after its chronic hepatitis B therapeutic vaccine candidate CVI-HBV-002 did not show a meaningful difference from placebo in phase 2b.
The placebo effect has not spared global big pharma either. Merck of Germany confirmed cartilage-regeneration effects for its osteoarthritis treatment sprifermin in a 2019 phase 2, but failed to prove a difference between groups as the placebo arm saw a large reduction in pain with saline injections alone. Japan's Daiichi Sankyo also failed all three global phase 3 trials in 2017 for its fibromyalgia treatment mirogabalin due to a larger-than-expected placebo response.
◇ "Late-stage trials are a 'design fight'… the 'placebo effect' decides wins and losses"
As a result, some in the industry say the success or failure of late-stage trials depends less on the candidate itself than on how well the placebo effect is controlled. To prove a new drug's efficacy, it is not enough that the drug works; the degree to which symptom improvement in the placebo group is reduced determines statistical significance.
In fact, the placebo effect is hard to dismiss as mere "psychological illusion." Studies keep showing that real physiological changes occur simply from patients expecting treatment or trusting clinicians.
A representative example is a 2005 University of Michigan study. The team first demonstrated that placebos activate the brain's natural pain-relief system to induce a real analgesic response. In other words, even a drug with no therapeutic effect can lead to actual symptom improvement through patients' expectations and beliefs.
Industry voices agree that designing from the outset to minimize the placebo effect is paramount.
An executive at a biotech developing immuno-oncology drugs in Korea said, "We used to think making a good drug was enough, but now designing a good trial has become as important as making a good drug," adding, "Small differences like patient expectations, how investigators explain the study, and assessors' experience can amplify the placebo effect and flip the trial outcome."
The person added, "Especially for patient-reported measures like pain or quality of life, the placebo response is likely to be large," and said, "If you don't craft precise enrollment criteria from phase 2 and standardize assessor training, you may fail to prove efficacy in phase 3 even after investing hundreds of billions of won."
An executive in research and development (R&D) at a biotech developing an Alzheimer's treatment also said, "Late-stage trials are no longer a fight to develop a drug but a fight to design the trial," adding, "If many patients sensitive to placebo enroll, the gap with the placebo group can narrow even when there is real efficacy, so deciding which patients to enroll becomes a core strategy."
Academia is also discussing excluding high placebo responders before trials begin. A representative method is the sequential parallel comparison design (SPCD), in which placebo is given first and only nonresponders are included in the main trial. However, because the placebo response is also a reaction seen in real-world care, there is considerable pushback that artificially excluding it may not adequately reflect real patient populations.
There are also studies showing the placebo effect itself is growing. A McGill University team in Canada released a 2015 paper in the international journal Pain analyzing that while actual drug responses have not changed much, placebo responses have steadily increased, gradually narrowing the gap in effect between drugs and placebos.
◇ Late-stage readouts line up in the second half… will they clear the "placebo effect"?
That is why the industry's attention is on late-stage results from domestic biotechs set for release in the second half of this year. Because many indications involve patient-reported symptoms—such as osteoarthritis, Alzheimer's disease and dry eye—the outcome will hinge not only on the candidates' efficacy but also on how effectively the placebo effect was managed.
The most attention is on Kangstem Biotech(217730), which is slated to release results this week. The company's knee osteoarthritis candidate Oscar targets the same condition as Kolon TissueGene's TG-C. Because this phase 2a was designed as a placebo-controlled study, the key questions are how much placebo response will appear and whether statistical significance can be achieved.
In September, AriBio will release key results from the global phase 3 of AR1001, an oral Alzheimer's disease candidate conducted in 13 countries including Korea, the United States and Europe.
Central nervous system (CNS) diseases, including Alzheimer's disease, are also representative areas with large placebo responses.
In fact, Eisai's Leqembi and Eli Lilly and Company's Kisunla succeeded in proving differences versus placebo by selecting patients with relatively clear early Alzheimer's symptoms.
By contrast, Biogen's Aduhelm struggled to prove efficacy because the placebo group's rate of symptom worsening was slower than expected. Industry watchers say AriBio also applied patient selection criteria similar to successful global cases, so how well it controlled the placebo effect will likely determine the outcome.
HanAll Biopharma(009420) also plans to secure key U.S. phase 3 results within the year for HL036, a dry eye treatment co-developed with Daewoong Pharmaceutical(069620).
In the prior U.S. phase 3, the company failed to secure statistical significance versus placebo on the primary endpoints—the corneal central staining score (CCSS) and eye dryness score (EDS). However, because it confirmed meaningful improvement over placebo on the secondary endpoint, the Schirmer test (tear secretion measurement), whether it achieves the primary endpoints in the follow-up trial will be pivotal.