Despite the findings regarding conditioned responses in different components of the endocrine system, the majority of studies have certain limitations. The most remarkable one is the fact that the majority of assays were performed in men, without considering the possible moderation of the conditioned response according to gender. Therefore, future research should consider gender specificity in endocrine responses as well as evaluate other endocrine parameters that have not been entirely explored.
6. Preclinical and Clinical Implications of the Placebo Effect
During the last decade, there have been significant advances in the development of preclinical evidence regarding the placebo response, with the main goal of creating reproducible animal models that would bring great advantages to this field. Among these, there are advances in molecular mechanisms involved in the placebo response as well as experimental manipulations that cannot be performed in humans for technical or ethical reasons
[126][115].
The current discussion about ethical considerations is based on aspects related to deceptive placebos and placebos without deception
[127][116]. The first is the one that has been more extensively prohibited according to international ethical guidelines, mainly under the policy emitted by the American Medical Association in which it is declared that “Doctors can use placebos for diagnosis or treatment only if the patient is informed and accepts its use”
[128][117]. On the other hand, a considerable number of research studies about placebos without deception have questioned the widely shared assumption that placebos require deception to be effective. Usually, in this type of study, denominated “open-label placebo”, individuals are assigned to either a group that does not receive treatment or another group that will receive a placebo pill
[129][118]. Furthermore, patients are informed of the fact that the pill does not have any active medication and the researchers read a script to the patient informing them about placebo response and explaining the justification for the study. Recently, this methodology has provided information about statistically significant improvements in patients with chronic lumbar pain
[130][119], IBS
[131][120], depression
[132][121], and recurrent migraines
[133][122]. It has been suggested that administering placebo medication can have beneficial effects even if it is not deceptively presented as an efficient treatment.
A new type of treatment protocol has been applied in the open-label placebo model called pharmaco-conditioning to resolve possible ethical implications. In this therapeutic regime, an open-label placebo is coupled with an active drug until the administration of the open-label placebo alone induces a conditioned placebo response. The effectiveness found in different studies
[134,135][123][124] suggests that this method can be effective to maintain the therapeutic response while the secondary effects of active drugs decrease. This could be a less controversial way to incorporate placebos in the clinical setting.
Currently, the available clinical evidence regarding the placebo effect is extensive and variable, especially in the case of neurologic, psychiatric, and immune disorders. The next section summarizes the key clinical evidence regarding the impact of the placebo effect as treatment for these disorders.
6.1. Neurological Disorders
The placebo effect has been reported to possibly improve various neurological disorders. In this sense, the use of classical conditioning for the induction of analgesia has been extensively studied in the last 20 years. The first attempts to materialize placebo analgesia in mice led to authors reporting that taste and olfactory stimuli coupled with morphine as a conditioned stimulus caused analgesia in mice
[139,140,141][125][126][127]. Afterward, it was demonstrated that tactile and visual stimuli coupled with the administration of morphine as a conditioned stimulus can generate placebo analgesia in female mice
[58][46]. Zhang et al.
[49][37] were able to replicate the results in Sprague–Dawley male rats. On the other hand, Lee et al.
[59][47] recently proposed a new animal model of placebo analgesia. In their study, they used a conditioning paradigm in which a neutral signal was conditioned to different pain intensities in an attempt to avoid the possible stress associated with analgesia injections during conditioning phases. The authors found that in this drug-free conditioning process, there was a decrease in the nociceptive response to heat in which animals learned to associate their conditioned space with lower exposure to heat. However, studies with a larger sample and more rigorous analyses that replicate these findings are needed to determine if it is a reliable method to cause placebo analgesia in mice.
The majority of preclinical studies have been mainly based on acute pain models, with only three published animal assays attempting to examine placebo analgesia in chronic pain
[142,143,144][128][129][130]. First, McNabb et al.
[143][129] evaluated placebo analgesia in female mice who received a spinal clamping of the L5 nerve to induce a condition of neuropathic pain. Contextual stimuli such as the environment, time, smell, touch, and sight were used as conditioning stimuli; however, no significant differences were found. Alternatively, Zeng et al.
[144][130] reported the induction of pharmacologically conditioned placebo analgesia using a model of spinal nerve clamping. However, this study did not include proper control groups to separate the effects of placebo from the non-specific responses that can be caused by other factors. More recently, Akintola et al.
[142][128] approached these limitations in a rodent model of neuropathic chronic pain, finding that chronic pain in mice could be non-responsive to placebo analgesia.
Beyond the numerous studies in animals, classic conditioning has also been proven as an analgesic in humans in different pathologies associated with pain
[145,146,147,148,149][131][132][133][134][135]. Multiple meta-analyses from clinical studies report a weak therapeutic effect on central neuropathic pain
[148][134] and the complex regional pain syndrome, and a moderate effect in postherpetic neuralgia
[150][136], peripheral diabetic neuropathy
[150][136], VIH associated pain
[150][136], fibromyalgia
[151][137], and migraines
[146,147,152][132][133][138]. On the other hand, only three open-label place studies related to pain have been performed to this day
[130,133,153][119][122][139]. Carvalho et al.
[130][119] performed an open-label, controlled, randomized study finding that placebos presented in a positive context can be used in chronic lumbar pain. In the study, patients with at least three months of chronic lumbar pain were randomly assigned to receive two tablets of placebo taken twice a day or their usual treatment for three weeks, reporting a significant decrease in their severity of pain score (95% interval confidence: 1.0–2.0). Likewise, Kam-Hansen et al.
[133][122] performed an open-label placebo study evaluating episodic migraines, reporting superior efficiency in individuals treated with placebos compared to those that did not receive any treatment.
The placebo effects of each treatment can be used to design therapeutic strategies that improve the clinical results of the analgesic and limit its adverse effects
[154][140]. In this context, the placebo effect induces the release of endogenous opioids that facilitate the analgesic action of exogenous opioids; therefore, it is possible to improve the response to analgesic treatments by increasing the additional placebo effect
[155][141]. Thus, through the development of interventions that optimize the placebo effect towards the adaptation of the CNS for pain relief, a potential progressive reduction in the administration of exogenous opioids is possible
[156][142]. There are various possibilities for taking advantage of placebo effects in the context of pain, adapting the information on analgesic treatment, and associating its intake with a positive context
[155][141]. Based on the above, it is to be expected that the combination of analgesic drugs and placebos would have better results in reducing pain than using each strategy separately.
Recently, it has been described that the placebo effect in humans can generate an increase in dopamine release in the dorsal and ventral striatum
[74,75,76][63][64][65], reporting that even 50% of patients with PD have shown response to placebo characterized by significant motor manifestations
[77,157,158,159,160,161][66][143][144][145][146][147]. Shetty et al.
[162][148] reported that from 36 clinical assays included in their study, 12 reported improvement after placebo treatment with PD, with a variation in the improvement from 9% to 59%. Likewise, a double-blind study found significant improvement in the group treated with pergolide and in the placebo group
[163][149]. Alternatively, Goetz et al.
[164][150] performed a randomized, multicenter, placebo-controlled study in which they found that 14% of patients achieved motor function improvement while they were on placebo. Another study performed by Goetz et al.
[165][151] involved data from 11 medical and surgical assays in patients with PD. They showed that the placebo effect can be significant, especially with surgical intervention. Regarding motor symptoms, bradykinesia is the one that has the greatest response to placebo
[164[150][152],
166], followed by rigidity
[164][150], gait, and tremor
[164][150], respectively. There was a 94% improvement in bradykinesia and 59% in gait
[164][150]. Likewise, Bennedetti et al.
[167][153] demonstrated that the administration of placebo-induced clinical responses as large as the one from apomorphine in rigidity.
The placebo effect has also been explored in other neurological disorders. Multiple sclerosis has an unpredictable remission–relapse pattern, making it a challenge to separate the placebo effect from the natural history of the disease in clinical assays. Despite this, different neuroimaging studies have shown a decrease in the number of injuries observed in magnetic resonance in the placebo groups
[136,168,169][154][155][156]. In an assay performed by Jacobs et al.
[136][154] there was an improvement in the magnetic resonance of the placebo group according to what was evaluated by the number of lesions potentiated by gadolinium. In addition, a meta-analysis was performed by Beyenburg et al.
[170][157], which included 54 studies examining anticonvulsant drugs versus placebos in over 11,106 adults and children with refractory epilepsy. They reported that there was a small difference between anticonvulsant drugs and the placebo effect
[170][157]. Similarly, a systematic review that included 28 clinical assays evaluating multiple anticonvulsant drugs versus placebos as refractory epilepsy treatment found a response in 18% of the patients receiving placebos
[171][158]. These results are similar to what was reported by Guekht et al., who conducted a meta-analysis that included 27 assays evaluating anticonvulsant drugs versus placebos in adults with focal epilepsy, reporting response to placebos in 12.5% of the patients
[172][159].
6.2. Psychiatric Disorders
Numerous studies have researched the placebo effect in the context of current psychiatry, especially in depression. Although antidepressants offer a clear advantage over placebos in patients with severe depression, the same is not true for those with mild depression. These patients have shown a response rate to placebos close to 50%. Often, the response rate between placebos and antidepressants cannot be differentiated
[173][160]. Furthermore, no type of psychotherapy has consistently proven to be better than placebo
[174][161]. It has been hypothesized that common and possibly therapeutic characteristics of psychotherapy, which include improvement expectation, support, and hope mobilization, are often provided together with placebo. Different studies have estimated that the double-blind response to placebo has 80% of the strength of double-blind antidepressant response in patients with major depressive disorder in randomized controlled assays
[132][121]. In this sense, in an open-label, randomized placebo study, a positive difference was seen among patients with major depressive disorder treated with open-label placebo and the control group. However, the difference was not statistically significant
[175][162].
The impact of the placebo effect on anxiety disorders has been explored. It has been reported that the placebo effect in clinical assays involving this disorder ranges from 10% to 60%
[176,177,178,179,180,181][163][164][165][166][167][168]. Different randomized assays have shown that the placebo response in anxiety disorders can be relevant and long-lasting. In this sense, improvement in the placebo group in clinical trials has been stable and maintained after the use of the placebo was suspended, while the patients using anxiolytic drugs suffered relapses
[182,183][169][170]. On the other hand, Faria et al.
[137][171] performed a study in which it was shown that telling patients who had been diagnosed with social anxiety disorder (SAD) that they were being treated with an active drug doubled the efficacy and tripled the response rate.
Likewise, Sandler et al.
[135][124] demonstrated that treatment with an open-label placebo was acceptable and efficient in the short term in the case of some children with attention-deficit and hyperactivity disorder (ADHD). In the study, the behavior of kids with ADHD remained the same when the dose of the stimulus drug with the placebo was reduced, but it deteriorated when the dose without the placebo was reduced. Alternatively, Weiss et al.
[184][172] examined the nature of the effect of placebo medication with medical treatment in alcohol dependency. It was found that the groups receiving a placebo along with medical treatment were more likely to go to Alcoholic Anonymous meetings during the treatment (32.7% and 32% vs. 20.4%) and were less likely to withdraw from treatment (14.1% and 22.9% 553 vs. 29.3%). However, more studies are needed in psychiatric settings to confirm these findings.
6.3. Immunological Disorders
A great number of assays evaluating classical conditioning in different immune diseases have been performed, showing its efficacy in animal models of Systemic Lupus Erythematous (SLE)
[185][173], rheumatoid arthritis
[186][174], and asthma
[187][175]. In a study involving rats with experimentally induced rheumatoid arthritis, re-exposure to a solution with saccharine and vanilla flavor that had been previously combined with cyclophosphamide resulted in a decrease in inflammatory processes
[188][176]. Likewise, in a model in rodents with SLE, mice with conditioned behavior showed relatively prolonged latency and survival time when compared to the control group animals
[189][177].
Numerous studies have highlighted the effects of placebo response in allergies, which seems to be mediated by cognitive factors such as expectations. A decrease in symptoms of type 1 allergic reactions in people treated with placebo with previous conditioning has been reported
[190][178]. Similarly, Goeber et al.
[138][179] reported a placebo response in patients with allergic rhinitis. These individuals were exposed to a conditioning protocol, receiving desloratadine and a beverage for 5 days. Afterward, the patients were exposed to the beverage and a placebo, showing improvement in the symptoms after this last exposure. Different assays have also been able to show that placebo responses imitate the effects of a drug to which the subjects have been previously exposed
[190,191,192][178][180][181]. A randomized, open-label placebo study evaluated two groups of 25 patients with allergic rhinitis comparing the use of the open-label placebo with their usual treatment for two weeks. It was observed that, at two weeks, there was a significant effect on the subjective experiences of 11 physical symptoms with significant improvement in subjective well-being (
p = 0.009). In addition, a statistically significant reduction in symptoms was observed in the open-label placebo group when compared with the group receiving their usual treatment
[193][182].
Likewise, different studies have reported that placebo administration leads to an improvement in objective parameters of lung function in asthma patients. These include the forced expiratory volume in 1 s (FEV1), bronchial hyperactivity, and peak expiratory flow (PEF)
[187,188,194,195][175][176][183][184]. A second study found that the administration of placebowith an inhaler was beneficial according to self-reported results, with an effect similar to that of albuterol without the need for conditioning. However, no increase in FEV1 was observed in asthmatic patients treated with placebo, unlike patients treated with albuterol
[196][185].