Placebo Effects Are Small on Average in the 7.5% CO2 Inhalational Model of Generalised Anxiety

Background Anxiety disorders are highly prevalent and socio-economically costly. Novel pharmacological treatments for these disorders are needed as many patients do not respond to current agents or experience unwanted side-effects. However, a barrier to treatment development is the variable and large placebo response rate seen in trials of novel anxiolytics. Despite this, the mechanisms that drive placebo responses in anxiety disorders have been little investigated, possibly due to low availability of convenient experimental paradigms. We aimed to develop and test a novel protocol for inducing placebo anxiolysis in the 7.5% CO2 inhalational model of generalised anxiety in healthy volunteers. Methods Following a baseline 20-minute CO2 challenge, 32 healthy volunteers were administered a placebo intranasal spray labelled as either the anxiolytic ‘lorazepam’ or ‘saline’. Following this, participants surreptitiously underwent a 20-minute inhalation of normal air. Post-conditioning, a second dose of the placebo was administered, after which participants completed another CO2 challenge. Results Participants administered sham ‘lorazepam’ reported significant positive expectations of reduced anxiety (p = 0.001) but there was no group-level placebo effect on anxiety following CO2 challenge post-conditioning (p’s > 0.350). Surprisingly, we found many participants exhibited unexpected worsening of anxiety, despite positive expectations. Conclusions Contrary to our hypothesis, our novel paradigm did not induce a placebo response, on average. It is possible that effects of 7.5% CO2 inhalation on prefrontal cortex function, or behaviour in line with a Bayesian predictive coding framework, attenuated the effect of expectations on subsequent placebo response. Future studies are needed to explore these possibilities.