Thirty-eight individuals with burnout (13 men and 25 women) in the age 24–55 were included in the study. The participants were recruited via advertisements in waiting rooms at occupational health care and primary care units in Gothenburg, Sweden. Inclusion criteria for burnout cases were a mean score ≥4.4 on the Shirom-Melamed Burnout Questionnaire (SMBQ) and self-rated symptoms for severe exhaustion disorder assessed with the self-rated exhaustion disorder (s-ED) questionnaire. A mean SMBQ score above 4.4 has previously been shown to discriminate patients with clinical burnout from healthy controls (17). The s-ED questionnaire is based on the Swedish diagnostic criteria for Exhaustion Disorder, which is the clinical diagnosis used for burnout in Sweden (18). Before inclusion, the subjects underwent a screening test, including anthropometric measurements. Blood samples were obtained. The screening was performed to assess the following exclusion criteria: having a body mass index less than 18.5kg/m² or higher than 30kg/m², hypertension (blood pressure>140/90 mmHg), infection, anemia, vitamin B12 deficiency (high homocysteine), known systemic disease such as diabetes or thyroid disease, known psychiatric disease (except self-rated stress-related exhaustion, depression and anxiety), alcohol abuse, pregnancy, breast feeding, menopause or use of drugs with systemic effects, including oral contraceptives but excluding antidepressants for the burnout cases. All participants gave written informed consent before entering the study and were informed that they could withdraw their participation at any time. The study was conducted according to the Declaration of Helsinki and approved by the Regional Ethical Board, Gothenburg, Sweden, Dnr 755-15.

The participants underwent an OGTT performed according to the World Health Organization (WHO) criteria of 1985. While a single glucose measurement can only define impaired fasting glucose (iFG), OGTT mirrors the patient’s ability to normalize the glucose levels after glucose intake, thus it detects impaired glucose tolerance (iGt). Furthermore, OGTT offers the opportunity to study insulin resistance and beta-cell function, by HOMA indexes, using fasting levels of glucose and insulin. A venous catheter was inserted to enable blood sampling at regular intervals. During the test, study participants rested sitting in an armchair reading or listening to the radio. Directly before the glucose load, the first blood samples were taken, constituting the fasting samples. The subjects then received 200 ml of glucose dissolved in water (75 g of glucose) and blood samples were drawn 30, 60, 90 and 120 minutes after the glucose load. Total blood loss was about 110 ml. Blood samples for analysis of plasma glucose and serum insulin were sent directly to the Laboratory for Clinical Chemistry at Sahlgrenska University Hospital for analysis.

The standard American Diabetes Association (ADA) criteria were used for the diagnosis of diabetes and pre-diabetes. 2-hour venous plasma glucose value ≥ 11.1 mmol/L was classified as diabetes and 7.8-11.0 mmol/L was classified as impaired glucose tolerance (IGT). HOMA-IR (Homeostatic Model Assessment for Insulin Resistance) (19) and the Matsuda Index (20), indicators of insulin resistance and insulin sensitivity, were calculated (fasting glucose x fasting insulin)/22.5 and 10000/√(fasting glucose x fasting insulin x mean glucose during OGTT x mean insulin during OGTT), respectively. Area under the curve (AUC) for glucose levels during the OGTT was calculated, with the formula; AUC(mmol/L*min) = 1/2 ×(PG 0 min+PG 30 min)×30 min +1/2×(PG 30 min+PG 60 min)×30 min +1/2×(PG 60 min+PG 90 min)×30 min +1/2×(PG 90 min+PG 120 min)×30 min. AUC for insulin was calculated in correspondingly.

The participants answered several questionnaires. Besides assessing symptoms of burnout, anxiety, and depression, they also answered questionnaires regarding their physical activity level. The Shirom-Melamed Burnout Questionnaire (SMBQ) (1) was used to measure severity of burnout symptoms. SMBQ contains 22 items (graded 1-7) measuring the different aspects of burnout: emotional and physical exhaustion, tension, listlessness, and cognitive weariness. A mean burnout score was calculated for each participant. The Hospital Anxiety and Depression (HAD) scale was used to assess self-reported symptoms of depression and anxiety (21, 22). HAD contains 14 items (7 items for each subscale). The scores for each subscale were used to classify “non-cases” (0-7), “possible cases” (8-10), and “cases” (above 10) of anxiety and depression, respectively. Participants reported their physical activity level (the past year) with the Saltin-Grimby Physical Activity Level Scale (23), which has been shown to be a useful screening measure of physical activity (24). This scale is a single-item question with four response options. The first level corresponds to a sedentary lifestyle, while levels 2–4 represent activity levels from light to strenuous exercise training.

Kolmogorov-Smirnov test was used for each study variable to test whether data were normally distributed. The variables which showed a non-normal distribution underwent logarithmic transformation. After logarithmic transformation (ln), the Kolmogorov-Smirnov test was used again to control whether the new variable showed a normal distribution.

Of the 38 individuals with burnout, 11 reported that they were sedentary, 13 reported some light physical activity, 13 reported regular physical activity/training and one reported regular hard physical training for competitive sports. The single participant who reported hard physical training was merged with the group who reported regular physical activity/training. Thus, three physical activity groups were formed; Sedentary (n=11), Light physical activity (n=13) and Moderate to hard physical activity (n=14). Background variables were compared between these physical activity groups. Two-way ANOVA was used analyzing possible differences in age, BMI, and WHR; Kruskal-Wallis H test was used for burnout score, depression score, and anxiety score, and Chi-Square test for proportion of males/females, tobacco use and use of antidepressants. Figures were created and these showed that the two groups who were physically active exhibited very similar values all throughout. Further statistical comparisons were therefore performed between two groups; those who were sedentary and those who were active (response option 1 vs. 2–4 on the Saltin-Grimby Physical Activity Level Scale). The distribution of men and women in the two groups was compared using the Chi-square test. T-tests were used to compare age and BMI between the groups. The Mann-Whitney U test was used to compare scores of burnout and depression between sedentary and physically active burnout cases. Levels of HbA1c and fasting glucose (measured at screening) were compared between sedentary and active participants using t-tests. To investigate possible differences in response to OGTT between sedentary and physically active burnout cases, mixed between and within analysis of variance (ANOVA) including interaction was computed with hormonal level (insulin and glucose, respectively) at the five different time points as the within variable and Group (Sedentary vs Active) as the between variable. Log values of glucose and insulin were used. AUC glucose (log), AUC insulin (log), the Matsuda Index and HOMA-IR (log) were compared between sedentary and active subjects using t-tests. Regression analyses were performed to investigate whether the observed effect of physical activity (Sedentary vs. Active) on insulin sensitivity (AUC insulin, HOMA-IR, and Matsuda Index) persisted after adjusting for age, BMI, gender.

In a previous study (16) we found that the burnout cases with more severe burnout symptoms exhibited significantly higher levels of both glucose and insulin levels during the OGTT compared to burnout cases reporting lower severity of symptoms. In addition, the group of burnout cases who reported symptoms of depression exhibited higher insulin levels during OGTT compared to the burnout cases without depressive symptoms. It could be speculated that the relation between physical activity and glucose control may be disturbed in this patient group due to previous extensive stress exposure and that the symptoms of burnout and depression might have a greater impact on insulin sensitivity than the level of physical activity. However, the burnout cases who reported severe burnout and depression reported equal level of self-reported physical activity. Accordingly, there was no statistical difference between the sedentary and physically active in scores of burnout or depression. Thus, in this study, reported physical activity seems to be independent of the severity of symptoms in burnout subjects. However, previously we have shown that better compliance to physical activity recommendation among patients with burnout was shown to be related to symptom reduction of both burnout and depression (29).

Prospective studies have shown that burnout increase risk of incident of type 2 diabetes by twofold or threefold (5, 9). The sedentary burnout cases in the present study show signs towards pre-diabetes, as indicated by elevated insulin levels, higher HOMA-IR, and lower Matsuda index. This adds to the rationale for promoting physical activity in burnout patients. There is evidence for good effect of physical activity on reducing exhaustion in intervention studies (30) This study points towards additional reasons for the importance of physical activity in burnout patients, thus, in preventing type 2 diabetes.

Firstly, it should be mentioned that the small study sample is to be considered as a limitation, but even with the limited number of participants the results were significant. The original plan was to recruit additional patients, but many individuals with burnout are simply too exhausted to participate in research studies that require efforts such as the OGTT challenge. One methodological consideration to be raised is the measurement of physical activity since a simple self-rated measure has been used. The Saltin-Grimby Physical Activity Level Scale (31) to assess physical activity level among the subjects during the past year and related their responses to glucose and insulin metabolism. This is a single item question with four response options. The first level corresponds to a sedentary lifestyle, while levels 2–4 represent graded increases in activity level from light to strenuous exercise training. This measure has been shown to be useful as screening measure and the different level of physical activity was shown to be clearly related to different cardiometabolic measures in both women and men (24). An optional methodology to investigate the research question would be to use for example pedometers to objectively measure physical activity level (32). Since 3 participants were tobacco users it should be mentioned that smoking and snuff (33, 34) may increase the risk of type 2 diabetes and thus may influence the results. Anti-depressant medication may also potentially affect insulin sensitivity. However, there were equal proportions of tobacco users and antidepressant users, respectively, between sedentary and active (shown in Table 1) and post hoc analysis performed excluding the three tobacco users or antidepressant users did not change the results (data not shown). The study individuals were selected to have normal BMI and to be otherwise healthy, thus we do not know if the results would be valid for individuals with burnout who were overweight, underweight, or having other health issues.