Daily physical activity habits influence surgeons’ stress in the operating room: a prospective cohort study

Abstract

Background:

Exercise training improves heart rate variability, and evidence suggests it can mitigate the detrimental effects of stress. This study aims to evaluate the relationship between surgeons’ physical activity habits and their stress, assessed as heart rate variability, at the start of surgery.

Materials and methods:

This multispecialty prospective cohort study included surgeons from fourteen cardiac, endocrine, digestive, gynecologic, orthopedic, thoracic, and urologic surgical departments of four university hospitals. Surgeons wore accelerometers 24/7 from 1 November 2020 to 31 December 2021 to quantify the mean daily step counts and daily sedentary time for 7 days prior to each operation. RMSSD, the root mean square of successive differences between normal heart beats, is a heart rate variability (HRV) metric that reflects cardiac vagal tone. We evaluated RMSSD during the first 5 minutes of surgeries performed over five 15-day periods. Data were analyzed using a multivariable linear mixed model with a random effect for surgeons.

Results:

We analyzed 722 surgeries performed by 37 surgeons (median age = 47 (IQR 42–55); 29 (78.4%) male). On average (SD), surgeons walked 9762 (2447) steps and were sedentary 391 (102) minutes daily. The model showed a positive relationship between steps and cardiac vagal tone, with an increase in lnRMSSD (0.028, 95% CI 0.003 to 0.053, P = 0.026) for every 1000 more steps per day, but not for sedentary behavior. Surgeon professors presented lower lnRMSSD (−0.437, −0.749 to −0.126, P = 0.006), as did surgeons who spent less total time operating over the study period (−0.337, −0.646 to −0.027, P = 0.033), compared to their counterparts.

Conclusions:

Higher daily step counts the week before surgery were associated with increased cardiac vagal tone, indicating lower stress levels at the beginning of surgery. This relationship highlights the influence of physical activity on surgeons’ stress in the operating room.

Introduction

Surgery is an undeniably demanding medical profession that requires surgeons to manage high-risk situations with a multi-professional team while working long hours and handling intraoperative stressors. These stressors include complex procedures, time constraints, interruptions, environmental distractions, and equipment problems^[1–3]. Moreover, current changes in healthcare systems have created additional stressors outside of the operating room. These are related to increased demands on clinical productivity, inefficient electronic filing systems, and administrative requirements, which further impair work-life balance^[4,5]. Heightened levels of stress can negatively influence both technical^[6] and non-technical skills, such as decision-making, judgment, and teamwork, on both the individual and team levels^[7,8]. Prolonged exposure to stressors can predispose surgeons to burnout and depression^[1,9]. Both are independently associated with major medical errors in surgery^[2,10] and can negatively impact physicians’ physical health^[11,12].

The term “stress” can be defined as a state of worry or mental tension that occurs when the demands of a situation exceed an individual’s available resources^[13,14]. Evaluating intraoperative stress is complex because each surgeon perceives and copes with it differently. Furthermore, measurement techniques include both subjective self-report questionnaires and objective assessments of physiological states, such as heart rate variability (HRV), stress-related hormones, ocular activity, electrodermal activity, and blood-related metrics^[15]. Although self-report questionnaires are cost-effective and are widely used to gather information about surgeons’ perceived stress levels, they are time-consuming for surgeons to complete and can be subject to bias^[15–17]. Past studies have demonstrated that HRV can identify stressors during an operation, determine which operating techniques generate the most stress for surgeons, and show differing stress levels between performing and assisting surgeons^[18]. HRV metrics correlate with surgeons’ subjective stress reports and procedural difficulty^[19,20]. Furthermore, surgeons with lower HRV during technically challenging parts of surgery tend to perform worse^[19,20].

HRV assesses autonomic activation by measuring the fluctuations in the time intervals between two consecutive heartbeats (IBIs)^[15,21]. The IBIs of a healthy heart constantly and rapidly change to adapt to physiological and psychological changes to homeostasis. These modulations are driven by the autonomic nervous system, which controls all “automatic” processes of the body and regulates the physical action of the heart via the sinoatrial node. The autonomic nervous system is composed of two branches, the parasympathetic (PNS) and the sympathetic (SNS) nervous systems. The PNS is generally downregulated during times of stress, while the SNS is upregulated. As a result, the heart rate increases, and HRV decreases. A variety of HRV metrics are used to assess fluctuations in IBIs. RMSSD (root mean square of successive differences between normal heartbeats) is a linear domain HRV metric calculated by taking the square root of the mean of the squared differences between successive IBIs^[21,22]. RMSSD focuses on short-term variations in the heart rate and is considered a reliable indicator of parasympathetic (vagal) activity. Higher RMSSD values indicate greater parasympathetic activity and better autonomic function^[21,23].

Cardiac vagal tone refers to the activity of the vagus nerve, which is the main nerve of the parasympathetic nervous system, in regulating heart function^[24,25]. Heart rate variability is mainly under the influence of cardiac vagal control as the SNS influence on beat-to-beat changes is much slower than the PNS influence^[21,26,27]. As summarized by Larkin et al, findings from a variety of studies using animal models, case-control studies of humans exposed to stressful situations, and samples of humans diagnosed with a range of psychiatric disorders have supported the link between exposure to stress and reduced cardiac vagal tone^[28]. Although stress is often considered a psychological phenomenon, the multiple physiological manifestations of stress, including changes in the heart rate and heart rate variability, are well-recognized^[29]. During times of stress, sympathetic activity generally increases, whereas vagal activity is inhibited to allow the heart rate to increase and HRV to decrease^[28]. The “vagal tank theory,” proposed by Laborde and colleagues, suggests that higher resting cardiac control is linked to better executive functioning, stress management, emotional regulation, and overall health^[23,24,30]. Additional studies demonstrate that individuals with higher vagal tone at rest tend to be more resilient to stress and adapt better to changing situations^[31–33]. On the other hand, a reduction in cardiac vagal tone could indicate an impaired ability to respond flexibly to changing demands and psychological challenges^[21].

There is growing evidence that regular physical activity can help protect against the harmful effects of stress^[34]. On the other hand, sedentary behavior, defined as sitting or reclining waking behaviors requiring minimal energy expenditure (1.5 metabolic equivalents or less)^[35], has been linked to anxiety and depression^[36] and may increase perceived stress^[37]. The positive effects of long-term physical activity also impact autonomic balance^[38–40], with higher levels of chronic exercise leading to increased cardiac vagal tone and decreased sympathetic activity^[41]. In surgeons, higher levels of aerobic exercise have been identified as an independent factor in preventing burnout and are associated with improved quality of life scores^[42]. Furthermore, regular exercise has been identified as an important factor that helps reduce occupational stress^[43,44]. Effective stress management is a non-technical skill that can support decision-making in the operating room and possibly improve patient care^[45].

However, no previous studies have objectively assessed the association between physical activity or sedentary time and stress in surgeons when operating on patients. Therefore, this multicenter, observational study aimed to investigate the relationship between surgeons’ physical activity habits, considering both daily step count and sedentary time, and their intraoperative stress, assessed using RMSSD, which indexes cardiac vagal tone, during the first 5 minutes of surgery. We focused on this period because higher RMSSD at the beginning of surgery would be linked to enhanced executive performance and emotional regulation, which could help surgeons respond better to any unexpected stressors during surgery^[24,30].

Methods

Study setting and population

The cohort of this multicenter, prospective cohort study was composed of surgeons operating in fourteen cardiac, endocrine, digestive, gynecologic, orthopedic, thoracic, and urologic surgical departments of four university hospitals in France. Surgeons were excluded from the cohort if they were still in training (i.e. assistant specialists, residents, medical students, visiting interns, etc.), performed fewer than 50 surgeries per year, or refused to share personal data for the study purpose. Participating surgeons consented to heart rate variability monitoring during five 15-day periods regularly spaced throughout the study period between 1 November 2020 and 31 December 2021, and the surgeries they conducted during these sessions were eligible for inclusion. Surgeries conducted on patients who were under the age of 18, who refused to share their data for research purposes, who received surgeries as part of palliative care or organ donation procedures, and who were missing operative timestamping were excluded.

The study used pseudonymized data in accordance with European General Data Protection Regulation No. 2016/679. The French National Data Protection Authority (DR-2020-055 CNIL) and the European Research Council Executive Agency (801660 ERCEA) approved the study protocol. It was deemed exempt from formal oversight by the Mass General Brigham Institutional Review Board (Protocol 2023P002266). Surgeons gave written informed consent to participate in the study and for the use of their data. Patients were informed that their data could be used in research studies and were given the option to refuse to participate. The study protocol was registered on clinicaltrials.gov, and the work was reported in line with the STROCSS criteria^[46].

Physical activity assessment

Accelerometers are widely regarded as one of the most reliable and objective tools for measuring physical activity, especially in free-living conditions^[47]. To quantify the daily step count and sedentary time, surgeons continuously wore Actigraph wGT3X-BT accelerometers (ActiGraph, Pensacola, FL, USA) on the ankle 24 hours a day throughout the entire study period. Accelerometer data were analyzed using ActiLife software (version 6.13.3). A sampling frequency of 30 Hz and an epoch length of 60 seconds were used. Non-wear periods were defined as zero counts per minute for ≥180 minutes. Sedentary time was defined as 0–100 activity counts per minute, and sedentary bouts were defined as periods of uninterrupted sedentary time ≥10 minutes. Daily sedentary time was calculated as the sum of all sedentary bouts each day. The mean of daily step counts and sedentary time were calculated for the 7-day period prior to the surgery. Only 7-day periods prior to surgery with ≥4 days of valid accelerometer data were included in the analyses.

Assessment of stress using heart rate variability

We evaluated surgeons’ stress levels, assessed as cardiac vagal tone, at the beginning of surgery using short-term HRV measured using Polar H10 heart monitor chest straps (Polar Electro, Kempele, Finland). When an individual moves, their HRV is affected because both the parasympathetic and sympathetic nervous systems are involved in movement^[23]. To confirm that surgeons were static during HRV data collection, the Polar H10 heart monitors were paired with an Actigraph GT3x-BT accelerometer, as suggested in Laborde et al^[21,23]. The surgeon was considered static when the number of steps per minute measured by the accelerometer was ≤10. Measuring HRV during the initial 5 minutes of surgery allowed us to evaluate the heart rate before the potential onset of stressors related to the operation that we would not be able to integrate into the model. The recording length of 5 minutes was chosen in accordance with the 1996 HRV Task Force Guidelines^[23].

Inter-beat intervals recorded during the first 5 minutes of each surgery following incision were validated, corrected for artifacts and noise segments, and analyzed using Kubios HRV Scientific 4.0 software (Mika 2021, User’s Guide)^[48]. Surgeries that were less than 20 minutes in length, had no period of immobility ≥20 minutes, had no Polar data recorded for ≥5 minutes, or had recordings with >5% beat correction were excluded from analyses.

RMSSD (ms) was chosen as the primary outcome to assess cardiac vagal tone as it is relatively free of respiratory influences, unlike high frequency parameters^[21,23]. A reduction in RMSSD is related to reduced parasympathetic activity and thus greater reactions of anxiety or stress. Higher levels of resting vagally-mediated HRV are linked to performance of executive functions like attention and emotional processing by the prefrontal cortex^[21].

Data sources

Data related to the patients’ socio-demographic characteristics, the primary diagnosis related to the operation, and the type of surgery conducted was collected from the electronic record-keeping system used by all hospitals. Additional data related to patient comorbidities, operating room logistics, including scheduling and times of operative events, intraoperative and postoperative patient outcomes, and anesthetic details were manually collected from the patients’ electronic health records. The age and professional status of participating surgeons were determined using human resource records. The data for each surgery were integrated into the study database and then linked with the surgeons’ physical activity data and HRV data using anonymous identifiers.

Statistical analyses

Characteristics of patients, surgeries, and surgeons were summarized as numbers (percentages) for qualitative variables and as means with their standard deviations (SD) or medians with their interquartile ranges (IQR), depending on the normality of the distribution, tested using the Kolmogorov–Smirnov test, for quantitative variables.

After testing for normality (using the Kolmogorov–Smirnov test and visual inspection), the primary outcome (RMSSD) was natural log-transformed (ln) to meet normality assumptions for linear modeling. The relationships between the two physical activity variables and lnRMSSD were first examined using Spearman’s rank correlations. Correlation coefficients of ≥ ± 0.40 were considered to be of moderate strength^[49]. The shapes of these relationships were studied to determine the appropriate model to use. Next, a linear mixed model was used to evaluate the influence of the physical activity metrics on surgeons’ cardiac vagal tone during the first 5 minutes of surgery. Random effects for surgeons were included in the model to account for the clustering of operations within surgeons. The final model controlled for potential confounding variables that were chosen a priori based on a review of the literature. Surgeon characteristics, such as surgeon age, sex, professional status (surgeon professor or surgeon non-professor), and total cumulative operative time during the entire study period (categorized as low or high based on the median of the study population), were included as we considered these factors could influence stress and/or heart rate variability. Additionally, we considered that patient characteristics and the complexity of the surgery could influence surgeons’ stress levels. These factors were considered together in the model using a composite risk score developed from an independent dataset of patients operated on by the same surgeons between 1 January 2022 and 31 October 2022 (details in Supplementary Digital Content, Appendix 1, http://links.lww.com/JS9/D769). Finally, circadian rhythms affect heart rate variability. To control for time of measurement, we included time of incision in the model. A correlation matrix was used to identify possible collinearity between exposure variables. We tested the interactions between steps and sedentary time and steps and age. In addition, we tested the interactions between steps and the categorical variables, surgeon professor vs. non-professor surgeon, and surgeon sex. None of these interactions were significant, so no interactions were added to the final model. We conducted sensitivity analyses to evaluate the influence of the physical activity metrics on surgeons’ lnRMSSD (1) during the middle 5 minutes around the midpoint of the surgery and (2) during the final 5 minutes of surgery prior to wound closure. We conducted additional sensitivity analyses to determine whether surgical specialty influenced the relationship between daily step counts and lnRMSSD. We grouped the specialties into three groups ((1) cardiothoracic surgery, (2) general surgery (digestive and endocrine), and (3) orthopedic, urologic, and gynecologic surgery) according to the risk of surgical complications associated with the type of surgery. We then tested our final model for each specialty group, including the same adjustments for all potential confounding variables. For all analyses, a two-sided P-value of <0.05 defined statistical significance. All data analyses were conducted using SAS V.9.4 (SAS Institute, Cary, North Carolina, USA). Results are presented as frequency (%) or mean (standard deviation), unless specified otherwise.

Results

A cohort of 37 surgeons participated in the study, of whom 29 (78.4%) were men and 22 (59.5%) were surgeon professors. The median cumulative operating time per surgeon was 22 039 minutes (IQR 13 647–28 073), and the median number of surgeries per surgeon was 231 (IQR 159–281) during the study period. The surgeons were between the ages of 33 and 66, with a median age of 47 (IQR 42–55).

A total of 722 surgeries were included in analyses (Fig. 1). The characteristics of the patients operated on by the surgeons during these surgeries are shown in Table 1. The mean patient age was 58.1 (SD 16.7), 379 (52.5%) were women, and patients had a mean of 2.5 (2.1) comorbidities, with a mean risk score of 0.2 (0.2). The characteristics of the surgeries are shown in Table 2. Table 3 shows heart rate variability and accelerometer data. During the 7-day period prior to surgery, surgeons were sedentary for a mean of 391.0 (102.3) minutes a day and walked 9762.1 (2446.6) steps per day. The mean RMSSD (milliseconds) for the first 5 minutes after incision was 19.4 ms (21.6). The distribution of the mean daily step count during the 7-day period before surgery, which ranged from 5433.6 to 19134.6 steps per day, is shown in Figure 2.

Figure 1.

Flow chart.

Table 1.

Characteristics of patients who the surgeons operated on during the study period

	Total = 722
Age of patient in years, mean (SD)	58.1 (16.7)
Female patient, N (%)	379 (52.5)
ASA physical status classification system, mean (SD)	2.1 (0.8)
Number of comorbiditiesa, mean (SD)	2.5 (2.1)
Risk score (missing = 6)b, mean (SD)	0.2 (0.2)

^aThe following comorbidities were considered: critical condition, current pregnancy, obesity, malnourishment, tobacco addiction, alcohol addiction, other addiction, open wound, surgical site infection, sepsis, endocarditis, cancer, neoadjuvant treatment, immune deficiency, coagulopathy, anticoagulant treatment, anti-aggregation treatment, blood transfusion, coma, limb paralysis, other neurological disorder, confusion, dementia, depression, cardiovascular disease, neurovascular disease, peripheral arterial disease, cardiac arrhythmia, chronic heart failure, hypertension, diabetes, dyslipidemia, high pulmonary artery systolic pressure, chronic renal failure, acute renal failure, chronic respiratory failure, chronic obstructive pulmonary disease, liver disease, rheumatic pathology, and hypoparathyroidism.

^bIn order to take into account the surgical risk of each procedure, a risk score based on all the variables presented in this table as well as the comorbidities listed above was constructed.

Table 2.

Characteristics of the surgeries performed by the surgeons during the study period

	Total = 722
Surgical specialty of the operation, N (%)
Orthopedic	164 (22.7)
Digestive	146 (20.2)
Gynecologic	124 (17.2)
Urologic	90 (12.5)
Endocrine	89 (12.3)
Cardiac	83 (11.5)
Thoracic	26 (3.6)
Principal anesthesia technique, N (%)
General	597 (82.7)
Locoregional	125 (17.3)
Surgical operative time in minutes between patient’s incision and closure of the wound, mean (SD)	108.2 (88.5)
Mean hour of incision time, mean (SD)	11:53 (2:54)
Scheduling of the operation (missing = 2), N (%)
Elective	653 (90.7)
Semi-urgent	7 (1.0)
Urgent	60 (8.3)
Initial surgical approach (missing = 3), N (%)
Open	459 (63.8)
Videoscopic	172 (23.9)
Endoscopic	56 (7.8)
Robot	32 (4.5)

Table 3.

Surgeons’ heart rate variability and accelerometer data for all of the surgeries they performed during the study period

	Total = 722
RMSSD, root mean square of successive RR interval differences the first 5 minutes after incision, mean (SD)	19.4 (21.6)
Mean daily step count 7 days before operation (daily average)a, mean (SD)	9762.1 (2446.6)
Mean daily sedentary time in minutes 7 days before operation (daily average)b, mean (SD)	391.0 (102.3)

^aMean step count per day calculated for the 7-day period prior to surgery.

^bTotal number of minutes spent in sedentary bouts (1 bout = a period of uninterrupted sedentary time of ≥10 minutes). This measure is equal to total cumulated sedentary time in a day. Sedentary time is defined as 0–100 counts per minute (CPM) using the vertical axis.

Figure 2.

(A) Distribution of mean daily step count 7 days before the operation (by thousands of steps) and (B) distribution of daily sedentary time in minutes 7 days before the operation.

There was a positive unadjusted correlation between the mean daily step count and lnRMSSD (ρ = 0.413, P < 0.0001) and a negative unadjusted correlation between the mean daily sedentary time and lnRMSSD (r = − 0.211, P < 0.0001). After adjustment, the results of the multivariable model in Figure 3 showed a positive linear relationship between the mean daily step count 7 days before the operation and cardiac vagal tone, with an increase of 0.028 lnRMSSD (95% CI: 0.003 to 0.053, P = 0.026) for every 1000 more steps taken per day. The relationship between the mean daily sedentary time 7 days before the operation and cardiac vagal tone was not significant. Surgeons who spent less total time operating over the course of the study period presented lower cardiac vagal tone (−0.337 lnRMSSD decrease, 95% CI: −0.646 to −0.027, P = 0.033), as did surgeon professors (−0.437 lnRMSSD decrease, 95% CI: −0.749 to −0.126, P = 0.006) compared to their counterparts. Figure 4 illustrates the relationship between the mean daily step count 7 days before the operation and the surgeons’ adjusted cardiac vagal tone at the beginning of the operation.

Figure 3.

Linear mixed model for surgeon ln(RMSSD).

Linear mixed model relating the exposures of interest to the natural logarithm of surgeon RMSSD with adjustment for potentially confounding surgeon, patient, and surgery characteristics. The forest plot on the right visually displays each predictor variable’s coefficient and 95% confidence interval. Blue: significant increase in ln(RMSSD); black: marginally significant (|P < 0.10|); red: significant decrease in ln(RMSSD); gray: non-significant association with ln(RMSSD).

Figure 4.

Relationship between the mean number of steps per day during the 7-day period and surgeons’ adjusted cardiac vagal tone during the first 5 minutes of surgery.

A linear mixed effects model was used to model the influence of physical activity habits on surgeons’ vagal tone, assessed as RMSSD, during the first 5 minutes of surgery. Adjusted ln(RMSSD) and 95% confidence intervals (CI) were calculated according to daily step count values 7 days before the operation, using beta-estimates derived from the model (see Figure 3) and setting each confounder to its mean value.

The results of the sensitivity analyses exploring the influence of the physical activity metrics on surgeons’ lnRMSSD during the middle 5 minutes around the midpoint of the surgery and during the final 5 minutes of surgery prior to wound closure are shown in Table 4. The results of the sensitivity analyses exploring the impact of surgical specialty on lnRMSSD are shown in Figure 5.

Table 4.

Results of sensitivity analyses comparing the results of the linear mixed effects models showing the influence of physical activity habits on surgeons’ vagal tone (lnRMSSD) during (A) the middle 5 minutes of surgery and the (B) final 5 minutes of surgery

	(A)Middle 5 minutes			(A)Final 5 minutes
	Estimate	95% confidence intervals	P-values	Estimate	95% confidence intervals	P-values
Surgeon daily step count 7 days before operation (+1000 steps)	0.029	(0.002 to 0.057)	0.034	0.034	(0.009 to 0.058)	0.006
Surgeon daily sedentary time 7 days before operation (+60 minutes)	−0.016	(−0.061 to 0.027)	0.451	−0.012	(−0.051 to 0.026)	0.539
Professor vs. non-professor	−0.338	(−0.654 to −0.022)	0.035	−0.353	(−0.687 to −0.019)	0.038
Surgeon cumulative operating time (low vs high) a	−0.342	(−0.655 to −0.029)	0.032	−0.273	(−0.605 to 0.058)	0.105
Operation hour of incision time (+1 hour)	−0.009	(−0.020 to 0.000)	0.058	−0.010	(−0.019 to −0.001)	0.023
Surgeon sex (man vs. woman)	−0.193	(−0.588 to 0.201)	0.335	−0.211	(−0.629 to 0.206)	0.321
Patient risk score (+1 score unit)	−0.276	(−0.469 to −0.082)	0.005	−0.249	(−0.419 to −0.080)	0.003
Surgeon age at the time of the operation (+1 year)	−0.006	(−0.024 to 0.011)	0.476	0.000	(−0.019 to 0.019)	0.999

Linear mixed model relating the exposures of interest to the natural logarithm of surgeon RMSSD either (A) during the middle 5 minutes around the midpoint of surgery or (B) during the final 5 minutes of surgery before wound closure, with adjustment for potentially confounding surgeon, patient, and surgery characteristics. Significant relationships are highlighted in green.

Figure 5.

Results of sensitivity analyses examining the relationship between daily step counts and surgeon lnRMSSD for three groups of surgical specialties (cardiothoracic surgery, general surgery (digestive and endocrine), and orthopedic, urologic, and gynecologic).

Linear mixed models relating the exposures of interest to the natural logarithm of surgeon RMSSD with adjustment for potentially confounding surgeon, patient, and surgery characteristics were conducted for each specialty group. The beta estimates, 95% CI, and P-values for daily step count (+1000 steps) for each specialty group are shown in the table.

Discussion

This prospective multispecialty cohort study indicated that surgeons who had higher daily step counts the week leading up to a surgery had lower stress, as indicated by cardiac vagal tone, during the first 5 minutes of the surgery. However, no significant relationship was found between sedentary time and stress. Additionally, our results demonstrated that a greater total time spent practicing surgery in the operating room was associated with higher cardiac vagal tone at the beginning of surgery, while surgeon professors had lower cardiac vagal tone compared to their counterparts. Overall, our results underscore the complex relationships between physical activity and physiological stress, while highlighting the potential role of physical activity in buffering against the stressors encountered by surgeons.

This study is the first to explore the relationship between objectively measured daily step counts or sedentary time and cardiac vagal tone in surgeons. However, our results are consistent with prior research in the general population. Indeed, it is well-established that exercise training alters the autonomic nervous system by increasing parasympathetic tone, resulting in physiological adaptations like resting bradycardia and reduced heart rate at submaximal intensities^[50]. Numerous studies have demonstrated that physical training using a variety of modalities, including walking, improves autonomic function^[51]. Our findings are also in alignment with previous research in surgeons that demonstrates that regular exercise is associated with improved quality of life^[42] and may help combat occupational stress^[44].

Our finding that daily sedentary time was not associated with cardiac vagal tone is not surprising in light of past research highlighting the association between sedentary time and increased risk of cardiovascular disease, diabetes, and all-cause mortality^[52]. Research also suggests that sedentary behavior may be associated with higher levels of depression and anxiety^[36]. The World Health Organization recognizes the negative health effects associated with sedentary behavior and recommends limiting sedentary time and participating in regular physical activity to mitigate these health risks^[53]. However, the relationship between sedentary time and stress is less straightforward. A systematic review identified a lack of sufficient evidence attesting to an association between sedentary behavior and stress measured using subjective methods, and no relationship between sedentary behavior and stress when stress was evaluated using objective measures^[36].

The present study identified a significant association between professional status and heart rate variability, with surgeon professors exhibiting lower cardiac vagal tone compared to their counterparts. In France, surgeons with this status are primarily employed by the university and are required to perform teaching and research in addition to surgery and clinical consultations, whereas non-professor surgeons’ duties are limited to surgery and clinical consultations. Surgeon professors are not necessarily older than non-professor surgeons. While the underlying mechanisms driving this disparity warrant further investigation, it is plausible that academic and research responsibilities, in addition to leadership and clinical duties, contribute to heightened stress levels among surgeon professors. The multifaceted nature of academic roles may impose additional workload with cognitive and emotional demands, thereby impacting autonomic nervous system function and cardiac vagal tone. Indeed, increased workload, longer working hours, and intraoperative stressors have previously been shown to contribute to surgeons’ stress^[1-3]. Furthermore, a study on surgeons identified teaching responsibilities as a risk factor for burnout among surgeons^[54]. Interestingly, our analyses also revealed that surgeons with higher cumulative operative time exhibited higher cardiac vagal tone. When a surgeon spends more time operating, it is possible that he or she will become more familiar with the operating room environment, which could in turn help reduce stressors related to the operating room and equipment, or enable surgeons to better anticipate and cope with them.

Several elements of our methodology strengthen the validity of our results. Unlike many studies that rely on simulated surgeries conducted by residents, our data were collected in real-life, on primary surgeons doing routine surgery on real patients. Our choice to use objective methods to evaluate stress and quantify physical activity and sedentary behavior allowed us to avoid potential bias related to self-report questionnaires^[15,16]. Step counting is a straightforward physical activity metric that is easy to understand and that facilitates physical activity promotion^[55]. We chose to use RMSSD as our heart rate variability parameter because it clearly reflects cardiac vagal tone and is free of respiratory influences^[23]. Focusing on the first 5 minutes of surgery enabled us to evaluate surgeons’ cardiac vagal tone before the occurrence of unpredictable intra-operative stressors that would be difficult to include in our model. Our sensitivity analyses demonstrated that our results are robust, as higher daily step counts the week before surgery were associated with lower levels of stress, as indicated by higher cardiac vagal tone, during both the middle 5 minutes and the last 5 minutes of surgery. We hypothesize that, compared to the middle and end of surgery, the first 5 minutes of surgery are more likely to reflect surgeons’ physiological and mental states related to the surgeon’s operating environment, lifestyle, and mental preparation. Regardless, higher cardiac vagal tone has previously been linked to improved executive function, enhanced stress management, and increased emotional regulation^[24,30]. These benefits could translate to improved responses to unexpected intraoperative stressors (related to the patient and the operating environment) surgeons must manage during surgery. Future studies are warranted to explore these possibilities.

Another strength of our study was that participating surgeons from seven different specialties operating in 14 surgical departments of four university hospitals wore accelerometers for over a year, and the heart rate monitors for a large sample of surgeries covering the study period. As a result of all of these factors, our data are both rich and unique. Additionally, our statistical approach enabled us to adjust for potential confounders related to the surgeon, the surgery, and the patient, as well as the clustered nature of our data. Our sensitivity analyses revealed a trend for a relationship between daily step counts and lnRMSSD for the two specialty groups with the highest risk of surgical complications (cardiothoracic and general surgery), but not for orthopedic, urologic, and gynecological surgery. The small sample sizes of these subgroups likely limited the power of these sensitivity analyses. Future studies could explore whether physical activity could be particularly useful for lowering stress among surgeons who practice the most difficult surgeries.

Our study also contains several limitations. First, we were unable to control for certain lifestyle factors, including alcohol consumption, tobacco use, caffeine consumption, or medication use, which can influence heart rate variability^[21]. However, we did reduce the potential impact of these factors indirectly by using a within-subject design, and by controlling for certain known confounders, identified a priori. For example, time of day or circadian rhythms can influence heart rate variability^[23]. Therefore, the time when each surgeon made the first incision on the patient was included in the model to account for any variations related to the time of day when the measurement was taken. Additionally, individual characteristics, including age and sex, can impact heart rate variability parameters and were thus included in our model^[21,23]. Second, we did not observe a normal distribution of RMSSD, as is common in heart rate variability studies. As a result, and in line with recommendations for conducting heart rate variability studies, we performed a natural log transformation of RMSSD data before modeling^[21,23]. Although necessary, this data transformation does mean that caution should be exercised when interpreting our findings since we modeled lnRMSSD, not RMSSD. Third, we cannot exclude the possibility that, due to the Hawthorne effect, the study participants changed their physical activity patterns. However, the long duration of the observation period and within-subject design of the study would limit possible impacts of the Hawthorne effect on the interpretation of our findings. Another limitation related to the duration and in situ design of the study was the placement of the accelerometers on the ankle. The hip and wrist are the most common accelerometer wear sites. However, these placements were not considered feasible for 14 months of wear, as they are not compatible with the strict sterile techniques used to maintain antisepsis in the operating room. A prior study concluded that vector magnitude counts recorded using accelerometers worn on the ankle, wrist, and hip during activities of daily living were highly correlated and reliable^[56]. Therefore, in order to improve study compliance, surgeons wore accelerometers on the ankle. Finally, the present study was observational so we cannot draw any conclusions regarding causality.

When stressors accumulate, and a surgeon’s capacity to manage them is exceeded, their surgical technical and non-technical performance can be altered, resulting in medical errors and higher risk of adverse patient outcomes^[2,57]. Therefore, any approach that surgeons can use to help deal with stressors could potentially improve perioperative safety. A large body of evidence indicates that exercise can improve wellbeing and one’s ability to cope with stress^[58]. Exercise has also been proposed as an element of a holistic performance enhancement approach that could enable surgeons to continuously perfect their skills while promoting their own wellbeing and career longevity^[59]. Our results support the idea that regular exercise could be an effective, simple, and inexpensive coping strategy for stress management in practicing surgeons. Prior studies have shown that exercise interventions can effectively increase physical activity and improve quality of life in surgical fellows and residents^[60,61]. Future studies could investigate whether similar programs could be equally beneficial for surgeons, or even improve patient care. An exercise program designed for surgeons would need to help participants overcome the various barriers that prevent surgeons from participating in physical activity, including both environmental factors (the country where they live, the type of hospital, and the health system where they practice) and individual factors (family and other social obligations, etc.). Surgical coaching is a promising approach that can help surgeons achieve goals related to both performance and overall wellbeing^[62]. The potential utility of surgical coaching as a complementary tool for surgeons who would like to incorporate more physical activity into their daily routines could also be explored. When considering how surgeons’ lifestyles and habits influence their physiology and performance inside the operating room, the context in which they operate should also be taken into account. Although surgeons can employ positive coping strategies to reduce their perceived stress, institutions have the potential to reduce stressors, including inefficient operating room layouts that impede workflow, distractions in the operating room, administrative concerns, and time and productivity pressures, by making structural, organizational changes that promote positive work environments and cultures^[43,63–65].

Conclusions

Overall, this study demonstrates that higher levels of physical activity are associated with greater cardiac vagal tone and improved autonomic nervous system regulation. Importantly, our study highlights the potential role of physical activity in buffering against the occupational stressors encountered by surgeons. As past studies have shown that higher cardiac vagal tone is associated with improved stress response and greater cognitive function, future studies should confirm the relationship between surgeon stress and patient outcomes and explore whether prioritizing physical activity could favorably influence patient care.

Provenance and peer review not commissioned, externally peer-reviewed.

Ethical approval

The French National Data Protection Authority (DR-2020-055 CNIL) and the European Research Council Executive Agency (801660 ERCEA) approved the study protocol. It was deemed exempt from formal oversight by the Mass General Brigham Institutional Review Board (Protocol 2023P002266).

Consent

The study used pseudonymized data in accordance with European General Data Protection Regulation No. 2016/679. Surgeons gave written informed consent to participate in the study, and for the use of their data. Patients were informed that their data could be used in research studies, and were given the option to refuse to participate.

Source of funding

This study was supported by European Research Council (ERC) Starting Grant under the European Union’s Horizon 2020 research and innovation program (Grant Agreement No. 801660 – TopSurgeons – ERC-2018-STG). This study was also funded by a public grant from the French Ministry of Health (Programme de Recherche sur la Performance du Système des Soins [PREPS-17-0008]).

Author’s contribution

Concept and design: S.S., L.P., A.D.; acquisition, analysis, or interpretation of data: S.S., J.A., L.P., S.P., J.C.L., A.D.; drafting of the manuscript: S.S., J.A., A.D.; critical revision of the manuscript for important intellectual content: S.S., J.A., L.P., S.P., J.C.L., A.D.; administrative, technical, or material support: A.D., S.P.; supervision: A.D.

Conflicts of interest disclosure

None.

Research registration unique identifying number (UIN)

This study was conducted as the observational part of a randomized control trial, registered on clinicaltrials.gov. The clinicalTrials.gov ID is NCT04532658.

Guarantor

Antoine Duclos.

Provenance and peer review

This paper was not invited.

Presentation

None.

Assistance with study

The authors would like to thank the TopSurgeons Study Group for their assistance.