Dyspepsia is prevalent and affects 27–40% of the general population 1–3, with symptoms of bothersome postprandial fullness, early satiation, epigastric pain, and epigastric burning 4,5. Dyspepsia affects the quality of life 6,7 and increases the cost of healthcare 5. The initial pharmaceutical treatment recommended for these patients is treatment with a proton-pump inhibitor (PPI) or an H2-receptor antagonist for up to four weeks. A Helicobacter pylori test is used to determine whether treatment should further include a course of two antibiotics (triple therapy) 5,8. In scientific research, NSAIDs 3,9, exercise 1, peptic ulcer 3, age 3,9–13, sex 1,9–11,14,15, smoking 9,16, marital status 10, depression 9,11,13,17,18, sleep 1, BMI, and hypertension 9 have all been associated with dyspepsia. Psychological stress has been investigated as a risk factor for dyspepsia, with contradictory results. Different measures of stress have been associated with dyspepsia, most often in nonwestern populations. The measures of stress in research are frequently based not on everyday life stress but on life events 10–12,15,16,19 or induced stress 20–22. It has been suggested that psychological stress affects the symptom severity rather than the development of dyspepsia 2,23, but this suggestion has been disputed for perceived stress 13. Stress management programs have also been investigated as a way to reduce the symptoms of dyspepsia and have led to encouraging results 2.
These results suggest an association between stress and dyspepsia, although the measures of stress are most commonly based on life events or induced stress. The effect of everyday life stress on the development of dyspepsia over time has not been examined in detail.
Therefore, the aim of this study was to examine whether a high self-reported perceived stress level at baseline, as measured by Cohen’s Perceived Stress Scale (PSS-10), is associated with an increased risk of treatment with proton-pump inhibitors or H2-receptor antagonists in subsequent years.
Participants and methods
We linked data from the North Denmark Health Profile 2010 survey 24 with existing Danish registers in a cohort study. The primary aim of the survey was to describe citizens’ health states. There were 570 000 inhabitants in the Region of North Denmark and there were 35 700 citizens older than 16 years of age across 11 municipalities covering the entire northern Jutland who received the questionnaire. The selection of participants was stratified on the basis of municipality. Data collection was performed from 5 February to 22 March 2010. Nonresponders received two reminders by mail 24. Included in the North Denmark Health Profile 2010 was Cohen’s PSS-10 25.
Data from the survey were linked to several Danish registers. The Danish Civil Registration System included information on unique personal identification numbers [Civil Personal Registration (CPR)]. A CPR number is assigned to all individuals in Denmark at birth or immigration 26. These CPR numbers made it possible to link data from all other included registers. The Danish National Prescription Registry had information on all prescriptions redeemed in Denmark and included the date and specific anatomical therapeutic chemicals (ATC)-codes of the drugs that were redeemed 27. The Danish National Patient Register provided ICD-10 diagnoses for all hospital contacts 28. Using data gathered from the Central Taxpayers’ Register and the Salary Information Register, the Income Statistics Register recorded individual income 29. The Population’s Education Register 30 had information on each citizen’s educational history.
The PSS-10 25 score was generated from the respondents’ answers to the North Denmark Health Profile 2010. The PSS-10 consisted of 10 items on predictability, controllability, and life overload, as perceived by the individual during the last month 31,32. Each question had five possible answers scaled from never to very often. Each answer was coded from 0 to 4. The cumulative score for all ten items was the total PSS-10 score and ranged from 0 to 40. Respondents were then divided into quintiles on the basis of their PSS-10 score. In accordance with the description of the PSS-10, we compared quintiles of stress level 33. Higher PSS-10 scores indicated a greater feeling of psychological stress 32.
Dyspepsia was defined as the first redeemed prescription after baseline for a PPI or an H2-receptor antagonist. The ATC-codes A02BC01 through A02BC05 identified PPIs. The ATC-codes A02BA01 through A02BA08 identified H2-receptor antagonists. Individuals who were diagnosed in the hospital with a peptic ulcer during follow-up were censored. If an individual had redeemed a prescription and was diagnosed with peptic ulcer within the 60 days after redeeming the prescription, they were not categorized as a dyspepsia patient and were censored. Peptic ulcer diagnoses were identified using ICD-10 codes 28 K25 through K279 in the National Patient Register.
Age was derived from the Danish Civil Registration System 26 and was included in the model as a continuous variable.
Sex was derived from The Danish Civil Registration System 26.
Educational level indicates the highest completed educational level at baseline and was grouped as primary (basic school of <10 years), secondary (high school education of +3 years or vocational education of +4 years), or higher (short/medium length higher education of +2 to 4 years or long length higher education of +≥5 years). Educational data were identified through the Population’s Education Register 30.
NSAID usage was included in the model as a dichotomous variable. Respondents with NSAID usage during the 3 months preceding baseline measurements were identified through the prescription database and were grouped with respondents who reported using nonprescription painkillers in the North Denmark Health Profile 2010.
Smoking data were generated from respondents’ answers on the North Denmark Health Profile 2010. The respondents answered whether they smoked or used to smoke on a daily basis. If confirmed, they were asked how many cigarettes, cigars, pipe bowls, and/or cheroots they smoked per day on average 24. The smoking variable was included as a categorical variable and was grouped as never smoked, former smoker, smoking 1–14 cigarettes/day and/or cheroots, cigars, or pipe bowls daily, and smoking more than 14 cigarettes/day.
Alcohol consumption was included as a dichotomous variable at baseline on the basis of the recommendations for moderate alcohol intake from the Danish Health Authorities 34. Respondents were identified as within recommendations (≤168 g/week for women and ≤252 g/week for men) or over recommendations (>168 g/week for women and >252 g/week for men). Respondents reported units of alcohol per week in the North Denmark Health Profile 2010. One unit was calculated as equivalent to 12 g of alcohol.
BMI was grouped as underweight (BMI<18.5), normal weight (BMI=18.5–25.0), or overweight (BMI>25.0). BMI calculations were based on self-reported height and weight in the North Denmark Health profile 2010.
Sleep was grouped as less than 7, 7, and more than 7 h/day. Respondents reported in the North Denmark Health Profile 2010 on how many hours of sleep they had in a typical weekday.
Income was examined as quartiles. Income was a measure of the total household income in the year 2009 and was used to estimate the economic status for the respondents at baseline. Household income was identified through the income statistics register 29 and was an equalized income estimate on the basis of the number of adults and children in the household 35.
Marital status was included as a categorical variable on the basis of two items in the North Denmark Health Profile 2010, which asked for marital status and whether they were living with a partner. Marital status was grouped as 1 – married, in a registered partnership, or living with a partner, 2 – separated, divorced, or widowed, or 3 – unmarried and single.
Anxiety or depressive symptoms were based on an item from the North Denmark Health Profile 2010. Respondents were grouped as not having symptoms if they reported not being anxious or depressed and as having symptoms if they reported being moderately or extremely anxious or depressed.
Hypertension was included as a dichotomous variable grouped as hypertension/side effects from former hypertension and no current hypertension/no side effects from former hypertension. Data were based on items in the North Denmark Health Profile 2010.
Exercise was included as a dichotomous variable. The respondents in the North Denmark Health Profile 2010 were asked which description most accurately described their physical activity level in the last year. The possible answers were 1 – exercising hard and participating in competitive sports regularly and several times a week; 2 – participating in sports or doing heavy gardening or the like for at least 4 h a week; 3 – walking, riding bicycles, or doing other light exercise for at least 4 h a week; or 4 – reading, watching television, or doing another type of sedentary occupation. Respondents who answered participating in sports or more were grouped as exercising and the rest as not exercising because the Danish Health Authorities recommend that adults have more than 30 min a day of moderate-intensity to high-intensity exercise 36.
Previous ulcer was based on ICD-10 codes 28 K25–K27.9 for diagnoses of any type of peptic ulcer before baseline and as far back as January 1989.
Previous PPI use was included as a dichotomous variable on the basis of whether respondents had redeemed a prescription for a PPI or an H2-receptor antagonist before baseline. Prescriptions were identified by the same procedure as the outcome variable. PPI use was identified as far back as permitted by the register, which was from 1 January 1995.
χ2-tests were used to compare the baseline characteristics for categorical variables and a t-test for continuous age variable. Cumulative incidence proportion curves of the first redeemed prescription after baseline were created, in which individuals who died or were diagnosed in hospital with peptic ulcer during follow-up were censored. Descriptive statistics were calculated, describing the distribution of first redeemed prescriptions during follow-up. Cox proportional hazards regression was used to estimate the association between stress quintiles at baseline and redeeming a prescription of a PPI or a H2 antagonist during 33 months of follow-up. When calculating the estimates, the stratified sampling design was taken into account using the R-package Svycoxph 37. Schöenfeld residuals were examined to verify the proportional hazard assumption. Age, sex, NSAID use, smoking, alcohol consumption, BMI, sleep, income, marital status, depression, hypertension, exercise, previous ulcer, and previous PPI use were included in the analysis as covariates and the analysis was divided into four models. Age was included as a continuous variable after checking the linearity assumption. Tests for interactions between PSS-10 quintiles and covariates for previous PPI use and sex on the risk of nonulcer dyspepsia showed no statistically significant interactions. As psychological disorders may be associated with level of stress, a sensitivity analysis was carried out, testing a possible interaction between PSS-10 quintiles and anxiety and depressive symptoms. To examine the severity of dyspepsia, total cumulated defined daily doses (DDD) of PPIs and H2-receptor antagonists were calculated and tested for any differences between PSS-10 quintiles by carrying out an analysis of covariance, still taking the stratified sampling design into account and including full model covariates. Because of the large exclusion of respondents because of missing data on smoking, alcohol consumption, BMI, sleep, exercise, marital status, depression/anxiety, and hypertension, imputation was performed as a sensitivity analysis. Sensitivity analyses were also carried out either with only PPI use as the outcome or with a sample that censored individuals receiving triple therapy. Further, an analysis including all individuals redeeming prescriptions, without censoring those diagnosed with peptic ulcer during follow-up, was carried out. Data management was performed using SAS software, version 9.4 (SAS Institute Inc., Cary, North Carolina, USA). Statistical analysis was carried out using R statistical software package, version 3.2.2 (R Development Core Team, Vienna, Austria).
The Danish Data Protection Agency approved this study (Ref. GEH-2014-014). All data used were linked and stored in computers held by Statistics Denmark. Anonymity was maintained as data were made available with deidentified personal information. Only aggregated statistical analyses and results were published in accordance with the Act on Processing of Personal Data 38,39. Retrospective anonymized register-based studies do not require written informed consent and ethical approval 38,39.
In total, 23 392 (65.52%) individuals responded to the questionnaire, 21 842 of whom answered all 10 PSS-10 items. Thirteen respondents were excluded because they were registered as deceased before the start of follow-up and 5705 were excluded because they had not completed all questions on the covariates for full model adjustment. The numbers of respondents missing data on covariates were 444 for educational level, 346 for smoking, 2993 for alcohol consumption, 484 for BMI, 523 for sleep, 538 for exercise, 526 for marital status, 199 for anxiety or depressive symptoms, and 1085 for hypertension. This resulted in a sample of 16 124 individuals for statistical analysis, with 2703 individuals who redeemed a PPI or an H2-receptor antagonist prescription during follow-up (Fig. 1).
Individuals redeeming a drug during follow-up were, on average, 8.6 years older, were more often female, had a lower educational level, smoked more, used NSAIDs more often, were more often overweight, slept less, had a lower income, had been diagnosed with peptic ulcer before baseline more frequently, exercised less, reported anxiety or depressive symptoms more often, were more often married or separated, had a higher frequency of hypertension, and had redeemed prescriptions for the same group of drugs preceding baseline more frequently compared with the individuals who did not redeem a drug during follow-up. There were no significant differences in alcohol consumption between the groups (Table 1).
The cumulative incidence proportion was higher with an increasing level of self-perceived stress. The risk of redeeming a drug over 33 months was ~11.6, 13.5, 15.7, 17.9, and 24.9% from the lowest to the highest stress quintile. The cumulative incidence proportions increased throughout follow-up for all PSS-10 groups (Fig. 2). The mean interval between baseline and redemption of first prescription during follow-up was 341 days, with a lower quartile of 67 days and an upper quartile of 578 days. The distribution of redeemed prescriptions during follow-up has been included in Appendix A (Supplemental digital content 1, https://links.lww.com/EJGH/A171).
The results of Cox proportional hazard regression analyses are shown in Fig. 3. At the univariate level (model 1), the four highest stress quintiles all had a statistically significantly higher risk of redeeming medications than did the lowest stress quintile, with hazard ratios (HRs) of 1.18, 1.40, 1.61, and 2.37, respectively. After adjusting for age and sex (model 2), all HRs increased slightly. On further adjusting for previous PPI use and previous ulcers (model 3), the HRs decreased considerably. After full model adjustment for all covariates (model 4), the HRs decreased even further, but the risk of redeeming medicine for dyspepsia remained significantly higher in the four highest stress quintiles compared with the lowest quintile. The estimate for the second quintile had an HR of 1.16 [95% confidence interval (CI): 1.00–1.34], the third quintile had an HR of 1.21 (95% CI: 1.06–1.39), the fourth quintile had an HR of 1.20 (95% CI: 1.05–1.38), and the fifth quintile had an HR of 1.30 (95% CI: 1.12–1.50).
When using only PPIs as a measure for dyspepsia, the total number of individuals redeeming prescriptions was 2688. The HR for the highest stress quintile was 1.31 (95% CI: 1.14–1.51) after full model adjustment. When censoring individuals who received triple therapy for Helicobacter treatment (a PPI or an H2-receptor antagonist combined with two antibiotics), the total number of individuals redeeming prescriptions was 2675. The HR for the highest stress quintile was 1.31 (95% CI: 1.14–1.51) after full model adjustment. Tests for interaction between PSS-quintiles and anxiety and depressive symptoms were not statistically significant and the inclusion of the interaction term in the full model analysis did not alter the conclusion as HRs ranged from 1.17 to 1.34. Including all individuals who had redeemed prescriptions for PPI or H2-receptor antagonists during follow-up, irrespective of peptic ulcer diagnosise, did not alter the conclusion as HRs ranged from 1.18 to 1.31. The severity of dyspepsia was examined by calculating the total DDD redeemed during follow-up for each individual, with no statistical differences between PSS-10 quintiles. Among those 2703 individuals redeeming at least one prescription during follow-up, the mean cumulated number of PPIs and H2-receptor antagonists was 356.5 DDD. By imputing missing values for the covariates of smoking, alcohol consumption, BMI, sleep, exercise, marital status, anxiety or depressive symptoms, and hypertension, the sample remained at 21 829 individuals, with 3917 redeeming prescriptions. The HR for the highest stress quintile was 1.34 (95% CI: 1.19–1.51) after full model adjustment and the second stress quintile was no longer statistically significantly different from the lowest stress quintile (HR: 1.11; 95% CI: 0.98–1.26).
The main result of the study was that after full model adjustment (model 4, Fig. 3), the four highest stress quintiles had a statistically significantly increased risk of dyspepsia treatment during 33 months of follow-up. In particular, adjusting for previous PPI use and previous peptic ulcer affected the HRs, and it is recommended that these factors be included in future research.
In a cross-sectional study, 15 dyspepsia patients had significantly higher median stress scores than did controls on the basis of a stress profile test 18. The 2547 individuals who fulfilled the ROME III criteria for functional dyspepsia or irritable bowel syndrome felt both more stressed and more susceptible to stress than did controls 1. Stress, as measured by the BEPSI-K 5 item questionnaire, was calculated as an independent risk of functional dyspepsia (odds ratio: 1.713) for the 9.4% of the patients with the highest stress level compared with the rest in a sample of 23 698 respondents 9. These findings support the results from our study that stress increased the risk of dyspepsia; however, all of these previous studies used a cross-sectional design, whereas the individuals in our sample felt stressed before dyspepsia treatment. Only one of those studies could contribute with an estimate of association. The discrepancy between our increased risk of 1.3 for the highest quintile of stress and that of 1.713 may be because of our different choice of questionnaires used for stress measurement.
The perceived stress level in this study was measured using Cohen’s PSS. This measurement tool has been validated and is often used as an instrument to estimate everyday life stress 31. Because the outcome measure for dyspepsia is based on collected prescriptions, it cannot be confirmed that they actually had dyspepsia, but it is likely that they had symptoms because a physician prescribed the drug. Because dyspepsia patients are not hospitalized with dyspepsia, using prescriptions for drugs used in treatment is considered more valid in this sample than using a hospital diagnosis for functional dyspepsia to identify patients. Only two individuals were diagnosed with dyspepsia in the hospital during follow-up. As a proxy for severity of dyspepsia, we examined individual cumulated daily doses redeemed among those redeeming prescriptions. This showed no difference in severity between PSS-10 quintiles, indicating no difference in the severity of dyspepsia symptom severity on the basis of stress level. Furthermore, a recent study did not find differences in distress in patients who were referred for the gastric emptying test and did or did not fulfill the ROME III criteria 40, which supports our method of investigating this group of individuals collectively.
The implications of this study should be an increased focus on dyspepsia symptoms in individuals with stress and an increased focus on stress signals in dyspepsia patients as a possible factor in the development of their symptoms. Even modest increases in stress levels may be associated with an increased risk of redeeming prescriptions for PPIs. The results support future directions in risk stratification and strengthen the notion of an association between perceived stress and dyspepsia.
Drugs used for treating dyspepsia are also prescribed for patients with gastroesophageal reflux disease. As these two diseases are often treated without a confirmed diagnosis and share some symptoms, they can be difficult to distinguish. Some of the patients identified in our study could have suffered from gastroesophageal reflux disease instead or from both. Therefore, there is a possibility that some of the effects from stress on gastrointestinal symptoms will result in gastroesophageal reflux disease rather than dyspepsia. This is also a possibility for individuals with an undiagnosed peptic ulcer. Individuals with peptic ulcer disease whose symptoms are eliminated with treatment will not be diagnosed in hospital and could therefore be included as dyspepsia patients in this study. Dyspepsia patients are liable to self-medication 14, and Danish patients can buy smaller quantities of H2-receptor antagonists without a prescription. Therefore, it is possible that some respondents in the sample consumed H2-receptor antagonists or bismuth-containing compounds. These individuals will not be identified in this study as dyspepsia patients. Therefore, the total incidence of dyspepsia, in this study, is most likely an underestimation of the incidence in the Danish population. To affect the conclusion, there would have had to be an uneven distribution of individuals using nonprescription H2-receptor antagonist among the PSS-10 quintiles, which seems unlikely. Testing the hypothesis with only PPIs as the outcome measure did not alter the conclusion. It has been suggested that stressed individuals visit their general practitioner more often than others 17,41, but this finding has been disputed for dyspepsia patients elsewhere 12,19.
By using Danish national registries, information bias is probably limited because registries limit the self-reported data to the North Denmark Health Profile 2010 questionnaire. The stratified administration of the North Denmark Health Profile 2010 by municipality helped maintain the large sample and increased the generalizability of the results. The comprehensive adjustment in the four models suggests that confounding was not responsible for our results. The response rate of the North Denmark Health Profile 2010 was 65.5% 24. There was a possibility that the nonresponders did not respond because of a high stress level or that the nonresponders had different stress levels than the responders. This issue would not lead to selection bias unless the nonresponders also differed from responders in terms of dyspepsia risk. This eventuality was considered unlikely. The Danish National Prescription Registry, the Danish registers on personal income, and the Danish Civil Registration System are of high quality 26,27,29. The Danish National Patient register may be affected by changes over time in the organization and provision of health services 28, but no relevant changes were made from baseline until the end of follow-up in this study, and the register is internationally recognized as one of the most comprehensive of its kind 28. The registries eliminate the risk of selection bias because there is no loss to follow-up, which also adds power to the analysis as the large sample size is maintained.
Higher levels of self-perceived everyday life stress increased the risk of redeeming a drug for dyspepsia significantly during 33 months of follow-up. The group with the second lowest stress level had an increased risk of dyspepsia of 1.16. The group with the third and second highest stress levels had an increased risk of dyspepsia of 1.2 and the group with the highest stress level had an increased risk of dyspepsia of 1.3. The proportions of dyspepsia patients in the stress level groups ranged from 11.6 to 24.9%. Therefore, even modest increases in risk resulted in large numbers of extra patients, and practitioners should be aware of stress signals in patients with dyspepsia symptoms.
The North Denmark Health Profile 2010 was funded by the North Denmark Region.
The authors are grateful for the support received from the Public Health and Epidemiology Group at the Department of Health Science and Technology, Aalborg University, Denmark.
H.B. and U.D. developed the idea for this study. U.D. was the main author of the manuscript and carried out data management and statistical analysis with guidance and advice from H.B. and C.T.P. All authors contributed toward the interpretation of results and revised and approved the final manuscript.
Conflicts of interest
There are no conflicts of interest.
1. Miwa H. Life style in persons with functional gastrointestinal disorders – large-scale internet survey of lifestyle in Japan. Neurogastroenterol Motil 2012; 24:464–471.
2. Deghanizade Z, Zargar Y, Honarmand MM, Kadkhodaie A, Baygi ME. The effectiveness of cognitive behavior stress management on functional dyspepsia symptoms. J Adv Med Educ Prof 2015; 3:45–49.
3. Bernersen B, Johnsen R, Straume B. Non-ulcer dyspepsia and peptic ulcer: the distribution in a population and their relation to risk factors. Gut 1996; 38:822–825.
4. Talley NJ, Stanghellini V, Heading RC, Koch KL, Malagelada JR, Tytgat GNJ. Functional gastroduodenal disorders. Gut 1999; 45:II37–II42.
5. Klinisk vejledning for almen praksis Dyspepsi Udredning og behandling af voksne (Clinical Guidelines for general practice in determining and treating dyspepsia in adults). Available at: http://www.dsam.dk/files/9/dyspepsi_med_links_og_sidetilpasset.pdf
; 2009. pp. 1–46. [Accessed 27 January 2016].
6. De Vries DR, Van Herwaarden MA, Baron A, Smout AJ, Samsom M. Concomitant functional dyspepsia and irritable bowel syndrome decrease health-related quality of life in gastroesophageal reflux disease. Scand J Gastroenterol 2007; 42:951–956.
7. Talley NJ, Locke GR 3rd, Lahr BD, Zinsmeister AR, Tougas G, Ligozio G, et al. Functional dyspepsia, delayed gastric emptying, and impaired quality of life. Gut 2006; 55:933–939.
8. Ables AZ, Simon I, Melton ER. Update on Helicobacter pylori
treatment. Am Fam Physician 2007; 75:351–358.
9. Lee SP, Sung I-K, Kim JH, Lee S-Y, Park HS, Shim CS. The effect of emotional stress and depression on the prevalence of digestive diseases. J Neurogastroenterol Motil 2015; 21:273–282.
10. Hafeiz HB, Al-Quorain A, Mangoor S, Karim AA. Life events stress in functional dyspepsia: a case control study. Eur J Gastroenterol Hepatol 1997; 9:21–26.
11. Lee S, Park M, Choi S, Nah Y, Abbey SE, Rodin G. Stress coping, and depression in non-ulcer dyspepsia patients. J Psychosom Res 2000; 49:93–99.
12. Locke GR, Weaver AL, Melton LJ, Talley NJ. Psychosocial factors are linked to functional gastrointestinal disorders: a population based nested case–control study. Am J Gastroenterol 2004; 99:350–357.
13. Van Oudenhove L, Vandenberghe J, Geeraerts B, Vos R, Persoons P, Fischler B, et al. Determinants of symptoms in functional dyspepsia: gastric sensorimotor function, psychosocial factors or somatisation? Gut 2008; 57:1666–1673.
14. Rosenstock S, Kay L, Rosenstock C, Andersen LP, Bonnevie O, Jørgensen T. Relation between Helicobacter pylori
infection and gastrointestinal symptoms and syndromes. Gut 1997; 41:169–176.
15. Devanarayana NM, Mettananda S, Liyanarachchi C, Nanayakkara N, Mendis N, Perera N, et al. Abdominal pain-predominant functional gastrointestinal diseases in children and adolescents: prevalence, symptomatology, and association with emotional stress. J Pediatr Gastroenterol Nutr 2011; 53:659–665.
16. Stanghellini V. Relationship between upper gastrointestinal symptoms and lifestyle, psychosocial factors, comorbidity in the general population: results from the Domestic/International Gastroenterology Surveillance Study (DIGEST). Scand J Gastroenterol 1999; 34:29–37.
17. Herschbach P, Henrich G, Von Rad M. Psychological factors in functional gastrointestinal disorders: characteristics of the disorder or of the illness behavior? Psychosom Med 1999; 61:148–153.
18. De la Roca-Chiapas JM, Solís-Ortiz S, Fajardo-Araujo M, Sosa M, Córdova-Fraga T, Rosa-Zarate A. Stress profile, coping style, anxiety, depression, and gastric emptying as predictors of functional dyspepsia: a case–control study. J Psychosom Res 2010; 68:73–81.
19. Koloski NA, Boyce PM, Talley NJ. Somatization an independent psychosocial risk factor for irritable bowel syndrome but not dyspepsia: a population-based study. Eur J Gastroenterol Hepatol 2006; 18:1101–1109.
20. Nesland A, Øktedalen O, Opstad PK, Serck-Hanssen A, Aase S, Berstad A. Erosive prepyloric changes – a manifestation of stress? Scand J Gastroenterol 1989; 24:522–528.
21. Hveem K, Svebak S, Hausken T, Berstad A. Effect of mental stress and cisapride on autonomic nerve functions in functional dyspepsia. Scand J Gastroenterol 1998; 33:123–127.
22. Hveem K, Hausken T, Svebak S, Berstad A. Gastric antral motility in functional dyspepsia. Effect of mental stress and cisapride. Scand J Gastroenterol 1996; 31:452–457.
23. De Giorgi F, Sarnelli G, Cirillo C, Savino IG, Turco F, Nardone G, et al. Increased severity of dyspeptic symptoms related to mental stress is associated with sympathetic hyperactivity and enhanced endocrine response in patients with postprandial distress syndrome. Neurogastroenterol Motil 2013; 25:2–4.
25. Cohen S, Williamson G. Spacapan S, Oskamp S. Perceived stress in a probability sample of the United States. The social psychology of health. Newbury Park, CA: Sage; 1988. 31–67.
26. Pedersen CB. The Danish Civil Registration System. Scand J Public Health 2011; 39:22–25.
27. Kildemoes HW, Sørensen HT, Hallas J. The Danish National Prescription Registry. Scand J Public Health 2011; 39:38–41.
28. Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J Public Health 2011; 39:30–33.
29. Baadsgaard M, Quitzau J. Danish registers on personal income and transfer payments. Scand J Public Health 2011; 39:103–105.
30. Jensen VM, Rasmussen AW. The Danish Education Registers. Scand J Public Health 2011; 39:91–94.
31. Taylor JM. Psychometric analysis of the ten-item perceived stress scale. Psychol Assess 2015; 27:90–101.
32. Cohen S, Janicki-Deverts D. Who’s stressed? Distributions of psychological stress
in the United States in probability samples from 1983, 2006, and 2009. J Appl Soc Psychol 2012; 42:1320–1334.
33. Cohen S. Frequently asked questions. Available at: http://www.psy.cmu.edu/~scohen/scales.html
; 2013. [Accessed 21 March 2016].
34. Sundhedsstyrelsen (The Danish Health Authorities). Alkohol og Helbred (Alcohol and Health). Available at: http://www.sst.dk/~/media/149578E6F82D47C28077B692E4642905.ashx
; 2008. p. 1–65. [Accessed 30 January 2016].
35. OECD. OECD Framework for statistics on the distribution of household income, consumption and wealth. Available at: http://dx.doi.org/10.1787/9789264194830-en
; 2013. [Accessed 15 June 2016].
36. Sundhedsstyrelsen (The Danish Health Authorities). Recommendations for exercise, adults. Available at: https://sundhedsstyrelsen.dk/da/sundhed-og-livsstil/fysisk-aktivitet/anbefalinger/18-64-aar
. [Accessed 18 May 2016].
37. Lumley T. Analysis of complex survey samples. J Stat Softw 2004; 9:1–19.
38. Thygesen LC, Daasnes C, Thaulow I, Bronnum-Hansen H. Introduction to Danish (nationwide) registers on health and social issues: structure, access, legislation, and archiving. Scand J Public Health 2011; 39:12–16.
40. Dibaise JK, Islam RS, Dueck AC, Roarke MC, Crowell MD. Psychological distress in Rome III functional dyspepsia patients presenting for testing of gastric emptying. Neurogastroenterol Motil 2016; 28:196–205.
41. Jebbink HJ, Smout AJ, van Berge-Henegouwen GP. Pathophysiology and treatment of functional dyspepsia. Scand J Gastroenterol 1993; 28:8–14.