ERCP is a technically demanding endoscopic procedure requiring a high level of expertise to provide effective and safe patient care. Somehow, trainees must be able to practice effectively without putting the safety of the patient in jeopardy. Simulator-based training is thus highly recommended (1).

Virtual reality simulation is an established method for acquiring and improving technical and non-technical skills in a controlled, reproducible, and quantitative environment that replicates real psychological challenges and mental stress. Simulation training has been used to assess stress and to develop intraprocedural stress management (2). The complexity of stress mechanisms makes acute stress measurement difficult to quantify and interpret. There is no universally recognised non-invasive gold standard technique for the assessment of stress. Instead, subjective, and objective surrogate methods have been proposed, such as measuring acute changes in the autonomic nervous system (ANS). Furthermore, non-invasive measurement of actual stress during a clinical procedure is not practically feasible (3).

There are objective biomarkers of stress in the blood and saliva. The use of blood biomarkers involves the invasive procedure of taking blood, whereas saliva sampling is relatively non-invasive and thus a better way to analyse stress. According to the literature, the most effective salivary biomarkers of stress are cortisol (sC), alpha-amylase (sAA), and chromogranin A (sCgA) (4). Since stress is not dichotomous there are no specific thresholds for these biomarkers that indicate a high level of stress (5). Furthermore, stress is usually associated with impaired performance which, in the clinical situation, could lead to complications. Moreover, it has been shown that trainees suffer greater stress than more experienced practitioners (6).

When training individuals in colonoscopy, steeper learning curves and fewer complications have been observed when employing a simulator-based program in training (7). These benefits have yet to be shown for ERCP (8). Furthermore, in a recent systematic review on the use of simulators to acquire ERCP skills, only one study conformed with validation criteria (9).

Our primary objective was thus to non-invasively measure stress levels of ERCP trainees by means of saliva stress biomarkers before and after familiarisation with virtual reality (VR) ERCP simulation. A secondary outcome was to assess any differences in stress between men and women.

Twenty-five of the 51 participants were unable to complete Phase 1 and were therefore excluded from the study. Thus, our final group consisted of 26 participants (12 men and 14 women), 7 were 20–30 years of age and the remaining (n = 19) 30–40 years of age.

All participants experienced saliva collection to be problem-free and did not cause distraction. The distribution of the collected data was tested using the Kolmogorov–Smirnov test and found to be normal.

Diff_ex and Diff_comp values were calculated from the Phase 1 and Phase 3 saliva biomarker data. Thus, six percentage differences Diff_ex and Diff_comp for the three saliva biomarkers were estimated in Phase 1 and six in Phase 3 (see Tables 1–3).

The saliva stress biomarkers sAA, sCgA, and sC reliably correlated with mental stress while training on an ERCP simulator. The most accurate prediction of degree of change is obtained by sAA. The response to acute stress involves a complex process which is mediated by the hypothalamic-pituitary-adrenal (HPA) axis, as well as psychological and social reactions.

In the clinical setting, acute stress has a direct impact on performance and patient safety. Endoscopists performing ERCP, frequently encounter highly complex and thus stressful situations. Stress assessment and coping is thus relevant and necessary in this field (2). Thus, the ability to implement a coping strategy to deal with stress is thus important for enhancing performance (11).

Stress is a psychological construct and thus inherently difficult to measure objectively in terms of physiological parameters. Methods assessing autonomic nervous system (ANS) responses in various organ systems have been suggested as surrogate stress markers. These markers include: (a) changes in heart rhythm, measured by heart rate (HR), heart rate variability (HRV), or inter-beat interval (IBI); (b) electrodermal activity (EDA) levels; (c) thermal activity; and (d) saliva stress biomarkers (i.e., sAA, sCgA, sC, and secretory immunoglobulin A). There is a lack of consistent methodology that has led to rather inconclusive and, in certain cases, conflicting results (12).

To our knowledge, this study is the first to use three commercially available saliva stress biomarkers in an ERCP simulation setting. No previous study has reported acute mental strain measurements during endoscopy in clinical and simulation settings (13). In this study, we concomitantly measured saliva a-amylase, saliva chromogranin A, and saliva cortisol, all of them potential stress biomarkers, in a reproducible virtual simulation setting (14–16).

A previous study showed that the biomarkers we used have strong correlations with stress (3). sAA is secreted from the salivary glands but great intra-individual variations are observed (17). It has been suggested that sAA could be used as a surrogate marker of norepinephrine in a variety of stressful conditions (18). In contrast to sC, sAA activity is affected by salivary flow rate and pH (19). Furthermore, in a study stressful situations were associated with higher sAA levels (20).

Chromogranin A (CgA) is a glycoprotein that mediates intracellular storage of catecholamines and is released together with these by the sympathetic nervous system into the blood circulation (21). Previous studies have shown that sCgA responds rapidly to mental stress such as psychosomatic stress [25], academic assessment stress, and computer operation psychological stress (22, 23). Furthermore, it has been observed that sCgA levels increase during mental stress but decrease during intermissions, suggesting that sCgA can be used for short-term assessment of mental workload (24). sCgA is a more accurate indicator than sC since it responds more rapidly and more sensitively to psychological stressors (12). Others, however, have questioned the ability of sCgA to measure stress and/or ANS activity (25).

Plasma cortisol enters the saliva through passive diffusion leading to a stable plasma/saliva ratio. Saliva cortisol levels can thus be used to assess stress related HPA activity, and this has become the most widely used biomarker for studies on mental stress. Saliva cortisol levels begin to rise within 5 min of an increase in plasma cortisol reaching a peak 31–40 min after the onset of the stressor, and saliva levels correlate strongly with plasma cortisol concentrations. Studies have shown that acute stress increases sC levels (3). Some studies have failed to observe such an increase (26) and even a reduction in cortisol level after stress has been reported (11).

The use of saliva stress biomarkers in this study, proved to be simple and without distraction and preferable to invasive methods such as blood sampling (3). Biochemical stress markers were elevated during Phase 1 i.e., during the first attempts to perform ERCP simulator exercises, and decreased as the trainees became acquainted with the simulation environment. This should be considered when planning simulator and clinical training. There were also some differences between the male and female trainees.

As Diff_ex and Diff_comp in Phase 1 were all significantly positive, we conclude that sAA, sCgA, and sC may be used as stress biomarkers. Of these, the best stress marker appeared to be sAA as it showed the greater percentage increase compared to increases in sCgA and sC (Table 1). It is also evident that familiarisation with the ERCP simulator during Phase 2 led to a reduction in stress while performing exercise in Phase 3, as indicated by the biomarkers used.

Throughout the world, women are highly underrepresented in the field of advanced endoscopy (8). In this study, however, more women participated than men. The stress response appears to be lower in women than in men (27). It has been suggested that this could be caused by a stronger cortisol response in women in the luteal phase than in those in the follicular phase. The menstrual cycle should thus be taken into consideration when assessing the reaction to stress. The cortisol levels may also be affected by oral contraceptives (28). Thus, saliva secretion differs between the sexes, and may be due to variations in the secretion of gonadal steroids and ANS regulation (29).

Our results partially support this observation since sAA Diff_ex and Diff_comp in the women during Phase 1 were greater than the corresponding values in men. However, sCgA Diff_ex and Diff_comp were higher in men, and no difference was seen between men and women for sC (Tables 2, 3). When comparing saliva results for sAA, sCgA, sC between Phase 1 and Phase 3, there were no differences in the male group, but sAA showed a statistically significant increase in the female group. Moreover, as can be seen from Tables 2, 3, during Phase 3 both sexes had similar Diff_ex and Diff_comp for all saliva biomarkers measured (sAA, sCgA, sC).

It should be noted that performance in basic endoscopy does not necessarily correlate with ERCP performance, since no relationship between basic handling skills and therapeutic skills has been demonstrated (30). Furthermore, extrapolation of results from a simulator to the clinical setting should be made with extreme caution. For practical reasons, in vivo measurements of physical examination cannot routinely be measured. Endoscopy simulators could, however, play a role in the trainee screening process (31).

In this study, mean sCgA Diff_ex and Diff_comp during Phase 1 were significantly positive (p < 0.05), suggesting that Chromogranin A may be used as a stress biomarker. We also saw that the percentage difference in sCgA between exercise and baseline in Phase 1 was significantly greater than the corresponding difference in Phase 3, indicating that familiarisation in Phase 2 reduced stress in Phase 3.

During Phase 1, mean sC Diff_ex and Diff_comp were significantly positive (p < 0.05), suggesting that saliva cortisol may be used as a stress indicator. This finding concurs with the findings of a study assessing sAA and sC during acute mental stress in 51 surgeons in the OR (25).

This study has some limitations. The sample size was small, and our findings must be interpreted with caution. Secondly, any delayed salivary cortisol response would have been missed and delay in saliva sampling may be necessary to fully detect stress-induced cortisol response (32). No comparison with the FES score was made, and finally, the study focused on simulated ERCP training only, so extrapolation to clinical practice is not feasible. Furthermore, another limitation of the study is that we did not check the oral hygiene of the participants. As it is suggested the latter could influence the accuracy of the saliva biomarkers.

Nevertheless, this study sheds some light on the feasibility of using non-invasive assessment of stress experienced by trainee endoscopists using easy to collect saliva biomarkers. Larger, controlled studies with participants that have different clinical experience is needed to further evaluate and monitor stress during simulation training using saliva biomarkers.

Finally, this was a laboratory study conducted in a controlled environment, whereas stress monitoring in the clinical setting is more complex due to the influence of social, cultural, and psychological factors (33).

In conclusion, the use of saliva biomarkers for assessing mental stress was easy to implement and well-accepted by all participants in this virtual simulation ERCP training setting. Familiarisation with the ERCP simulator greatly reduced stress when performing ERCP exercises afterwards. The saliva stress biomarkers sAA, sCgA, and sC may all be used to assess mental stress while training on an ERCP simulator, but the best of the three, judging by the degree of change, appears to be sAA. No conclusive difference in stress response between men and women was observed. Further studies including a larger number of trainees with different levels of ERCP experience are needed to exploit performance enhancement and error reduction techniques in ERCP training.