Expectation‐induced placebo responses fail to accelerate wound healing in healthy volunteers: results from a prospective controlled experimental trial

Abstract

Placebo responses have been shown to affect the symptomatology of skin diseases. However, expectation‐induced placebo effects on wound healing processes have not been investigated yet. We analysed whether subjects' expectation of receiving an active drug accelerates the healing process of experimentally induced wounds. In 22 healthy men (experimental group, n = 11; control group, n = 11) wounds were induced by ablative laser on both thighs. Using a deceptive paradigm, participants in the experimental group were informed that an innovative ‘wound gel’ was applied on one of the two wounds, whereas a ‘non‐active gel’ was applied on the wound of the other thigh. In fact, both gels were identical hydrogels without any active components. A control group was informed to receive a non‐active gel on both wounds. Progress in wound healing was documented via planimetry on days 1, 4 and 7 after wound induction. From day 9 onwards wound inspections were performed daily accompanied by a change of the dressing and a new application of the gel. No significant differences could be observed with regard to duration or process of wound healing, either by intraindividual or by interindividual comparisons. These data document no expectation‐induced placebo effect on the healing process of experimentally induced wounds in healthy volunteers.

Introduction

Wound healing is a complex biological process that can be affected by various factors, which can be classified into local factors such as wound exudate or wound contamination, and systemic factors such as hormones, drugs, nutrition or smoking 1. In addition, there is a growing evidence that psychological factors may also affect the wound healing processes 2. Different studies in humans have shown that psychological stress such as caregiving or examination stress, low levels of anger control 3 and marital conflicts 4 slowed wound healing time 5. On the contrary, psychological interventions such as relaxation techniques or emotional self‐disclosure can accelerate wound healing 6, 7, demonstrating that emotional and cognitive factors affect the wound healing processes.

Placebo responses have been shown to be steered by two major neuropsychological mechanisms: associative learning (behavioural conditioning) and the subject's or patient's positive expectation towards a treatment 8, 9, 10, 11. Placebo effects in autonomous physiological responses such as hormone secretion or immune functions appeared to be predominantly mediated by learning processes 12, 13. Cognitive factors such as expectation have been shown to induce placebo effects not only in subjective outcome measures such as placebo analgesia 14, 15 but also mediate placebo effects in objective outcomes such as blood pressure and type 1 hypersensitivity reactions 16, 17. However, whether and to what extent cognitive factors such as expectation are inducing placebo responses in wound healing has never been experimentally tested in vivo before. Thus, in this study, we analysed whether the mere expectation of being treated with a pretended potent wound gel influences the duration and process of cutaneous wound healing in healthy volunteers.

Materials and methods

Subjects and inclusion criteria

A total of 22 healthy male subjects were recruited through advertisements at the University Hospital in Essen. The mean age was 29·6 years (SD 5·9). Exclusion criteria were age below 18 years, smoking, regular medication, obesity and diseases of the immune system. The study was approved by the Ethics Committee at the University Hospital in Essen and follows the rules stated in the Declaration of Helsinki. All subjects provided written informed consent.

Study design

A deceptive paradigm was used to investigate expectation‐induced placebo effects on wound healing. Participants were informed that they take part in a clinical open‐label trial that tested a new and innovative wound gel including epidermal growth factors. They were also informed that the gel strongly accelerated wound healing and had an analgesic effect in former pilot studies and that the purpose of this study was to include a larger population to confirm these positive results.

Subjects were randomly allocated to the experimental group or to the control group. In both groups, wounds were induced on both thighs. The experimental group was informed that one of the two wounds would be continuously treated with the assumed wound gel and the other wound with a non‐active gel. However, both gels were identical hydrogels. The pretended difference between the gels was made salient for the participants by using a green cover for the syringe in which the ‘wound gel’ was delivered and a white cover for the neutral syringe. To control for possible lateralisation effects on wound healing, the side of ‘wound‐gel’ application was counterbalanced between the participants; thus, approximately half of the subjects in the experimental group were treated with the sham wound gel on the left side (n = 5) and the other half on the right side (n = 6).

In addition to the intraindividual control condition, a control group for between‐subject comparison was included. This group was informed that a non‐active gel would be applied on both wounds.

Participants were asked to come to the laboratory on days 1, 4, 7 and 9 after wound induction. On each visit, the dressings were removed, the wound healing process was documented and subsequently new gel and new dressings were applied. From day 9 onwards, they came to the laboratory every day and underwent the same procedure until full reepithelialisation of the wound was reached.

Wound induction and treatment

Wounds were induced with an ablative laser (Lumenis UltraPulse^®, 225 mJ, 60 W, Lumenis Ltd, Yokneam, Israel) on both thighs with a size of 1 cm² and a depth of 2 mm (Figure 1). They were treated with a sterile hydrogel (Suprasorb G^®; Lohman & Rauscher, Neuwied, Germany) and covered with a sterile dressing (Curapor Transparent^®; Lohmann & Rauscher).

Figure 1.

Wound with a size of 1 cm² and a depth of 2 mm.

Measures

Progress in wound healing was documented via planimetry by Fotofinder Dermoscope II (Fotofinder Systems GmbH, Bad Birnbach, Germany) on every visit. The area of the wound that was not epithelialised was calculated. Wound healing was considered to be completed with full reepithelialisation, which was determined by an experienced dermatologist unaware of the participant's group allocation.

Pain perceived with the change of the dressings at each visit was documented with a visual analogue scale (VAS) ranging from 0 (no pain at all) to 100 (very intense pain).

Statistical analyses

Within the experimental and the control groups, a paired t‐test was used to detect differences in the duration of wound healing. Mean values of right and left wound side in the control group were subsequently used for between‐subject comparison of ‘wound‐gel’ treatment and control treatment (t‐test).

Process of reepithelialisation was documented by calculating the percentage of remaining wound area at each visit. Repeated‐measure analysis of variance (ANOVA) was used to detect any differences of wound healing progress within the experimental group and within the control group and between the experimental and control groups. The same procedure was used for analysing subjective pain measures.

All data are expressed as mean ± standard error of the mean (SEM). SPSS 20.0 was used for all analyses. Statistical significance was defined as P‐value <0·05.

Results

Time until full reepithelialisation

Reepithelialisation of wounds was analysed during each visit day. In the experimental group, there was no significant difference in the duration of wound healing between the wounds treated with ‘active gel’ and those treated with the ‘non‐active gel’ (t = 0·36, n.s.). In the control group, mean duration for wound healing showed no significant differences between left and right leg (t = 0·56, n.s.; Figure 2). Mean values of both legs in the control group used for between‐group comparison did not show a statistically significant differences (t = −0·44, n.s.).

Figure 2.

Days until full reepithelialisation in the experimental group (•) and in the control group (Δ) for each thigh. Statistical analysis did not show any significant differences, neither between the wounds treated with ‘wound gel’ and ‘non‐active’ gel in the experimental group nor between right and left leg both treated with ‘non‐active’ gel in the control group.

Process of reepithelialisation

In addition to monitoring the time until full reepithelialisation of experimentally induced wounds, the reepithelialisation process was analysed by planimetry. However, there was neither a significant difference between the wounds treated with the ‘active’ and ‘non‐active’ gel within the experimental group (F = 1·78, n.s.) nor in the between‐group (control versus experimental) comparison (F = 0·17, n.s.; Figure 3).

Figure 3.

(A) Mean percentage of wound area for each visit for the wound gel‐treated leg and the non‐active gel‐treated leg in the experimental group. Repeated‐measure ANOVA did not show any significant differences between the two conditions during the wound healing process. Data are expressed as mean ± SEM. (B) Mean percentage of wound area for each visit for the wound gel‐treated leg in the experimental group and for mean values in the control group. Repeated‐measure ANOVA did not show any significant differences between the groups during the wound healing process. Data are expressed as mean ± SEM.

In addition, pain ratings did not show any significant differences between the wounds treated with ‘active gel’ and ‘non‐active gel’ in the experimental group and between experimental and control groups (data not shown).

Discussion

Current scientific data in humans provide evidence that wound healing can be influenced by psychological factors 2. In addition, disease progress in psoriasis and allergic skin reactions is affected by placebo responses 17, 18, 19. However, no study to date investigated the placebo effects on the healing time in experimentally induced wounds. Thus, in this study, we analysed whether and to what extent the wound healing process in experimentally induced wounds can be affected by expectation‐induced placebo responses. Intraindividual and interindividual comparisons did not show a significant placebo response, either on the duration until wound healing or on the reepithelialisation process.

Two major neuropsychological factors such as the expectation of subjects or patients towards the benefit of a treatment as well as behavioural conditioning processes are mediating the placebo response across various experimental conditions and diseases 8, 10 However, the knowledge of how these neuropsychological factors are steering the placebo response in the different physiological systems and end organ functioning is still incomplete 9.

In patients allergic to house dust mite, placebo responses on immunological functions such as basophil activation can be influenced through classical conditioning, whereas the allergic skin reaction documented by prick testing appeared to be also affected by patients' expectation 17. In addition, conditioned placebo responses on the disease progress were reported in psoriasis patients under glucocorticoid treatment 18. However, a recent study failed to show expectation‐induced placebo effects on inflammatory skin reactions 20.

In this study, the expectation of receiving an active drug (wound gel) did not affect the duration or the process of wound healing. Several reasons may account for these negative results. Firstly, it has been proposed that unconscious physiological processes such as endocrine or immunological alterations can be induced by behavioural conditioning, while placebo responses regarding conscious processes such as pain perception or motor performances are influenced by cognitive factors such as expectation 12, 13. The results of this experiment appear to support this view, suggesting that wound healing, which involves immunological processes such as cytokine release and growth factor secretion 21, is indeed not being affected by expectation‐induced placebo responses. It remains to be analysed if wound healing can be accelerated by behavioural conditioning. However, a problem in testing this hypothesis will be to identify a wound gel or a dressing that significantly accelerates the wound healing process and that can be used as an unconditioned stimulus within a conditioning protocol.

Secondly, the inclusion of an intraindividual as well as interindividual control condition might be a methodological strength of our study; however, it might also be at least partially responsible for the negative outcome. Participants in the experimental group were informed that on one leg the wound gel would be applied and on the other leg a non‐active gel as control. These two expectations at the same time (100% versus 0%) might have induced cognitive interferences that might have prevented a convincing expectation with subsequent neuropsychological and significant effects on the wound healing. Further studies with a mere between‐subject design are required to clarify this issue.

Thirdly, we observed neither a placebo response regarding the wound healing nor an expectation‐induced placebo response in pain ratings, indicating that wound induction did not cause sufficient enough discomfort or pain perception in the experimental subjects. Placebo effects on subjective symptoms such as pain have been shown to be influenced through somatic focus and the degree of attention that patients focus on their symptoms 22, 23. For ethical reasons, it is not possible to experimentally induce wounds that are causing substantial pain perception and discomfort. As the grade of pain perception that subjects reporting at each visit was rather low and as the wound was covered by a dressing, the degree of attention that subjects focused on the wounds has been most likely very low.

In addition, our results are limited to our population of young and healthy participants. Future studies will have to consider placebo effects in wound healing in older participants and patients with chronic wounds.

In summary, the setting expectation‐induced placebo responses did not affect the process of cutaneous wound healing in healthy subjects in our study. Further studies are required to investigate whether and to what extent wound healing might be affected by learned placebo responses.