Severe Gastro-oesophageal Reflux Symptoms in Relation to Anxiety, Depression and Coping in a Population-Based Study

C. Jansson; H. Nordenstedt; M.-A. Wallander; S. Johansson; R. Johnsen; K. Hveem; J. Lagergren


Aliment Pharmacol Ther. 2007;26(5):683-691. 

In This Article

Material and Methods

Study Design

The design of our cross-sectional case-control study has been described in detail elsewhere.[12] Briefly, the Nord-Trondelag Health Study consists of two large population-based health surveys performed in the Norwegian county of Nord-Trondelag in 1984-1986 (Helseundersokelse i Nord-Trondelag, HUNT 1) and in 1995-1997 (HUNT 2).[26] The residents of Nord-Trondelag are essentially representative of the Norwegian population as a whole.[27] HUNT 2 included 65 333 county residents of ages 20 years or older (70% of those 92 808 eligible to participate in the study), and 47 556 subjects participated in both HUNT 1 and HUNT 2. Data were collected in questionnaires and through physical examinations. Informed consent was obtained from each participant. The research was approved by the Regional Committee for Medical Research Ethics, Region IV, Norway. The Norwegian Data Inspectorate approved the establishment of a research register.

Definition of Reflux Symptoms, Cases and Controls

In this study, reflux symptoms were defined as severe symptoms of recurrent heartburn or acid regurgitation, and this information was assessed in HUNT 2 only. The participants were asked to report their experience of such symptoms during the past 12 months, and the response alternatives were: (i) no symptoms, (ii) minor symptoms and (iii) severe symptoms. Of the 58 596 subjects (90%) answering this question, the 3153 subjects (5%) reporting severe reflux symptoms were defined as cases and the 40 210 subjects (69%) reporting no such symptoms were defined as controls.[12] The 15 233 subjects (26%) reporting minor symptoms were excluded from the study, an exclusion based on our validation study described in detail previously,[12] where we found that misclassification of the outcome was high in this group. In the validation study we further found that 95% of the patients with severe reflux symptoms (corresponding to our case group) had experienced reflux symptoms at least once weekly, resulting in a high specificity (99%) for reflux symptoms occurring at least once a week in our case group.[12]

Data on Psychiatric Disorders and Psychological Factors

Data on anxiety and depression were included in both HUNT 2 and HUNT 1, while data on coping were included in HUNT 2 only.

Anxiety and Depression. Anxiety involves feelings of fear, worry and apprehension, while depression is associated with feelings of sadness, sorrow, hopelessness and gloom.[28] The Hospital Anxiety and Depression Scale (HADS) is a widely used self-assessment scale originally designed for detecting depression and anxiety among hospital medical out-patients.[18,28,29] HADS is a validated measure that performs well in screening for the separate dimensions of anxiety and depression, but also as a measure of co-morbid anxiety and depression.[18,28,30] HADS consists of two subscales, the anxiety subscale (HADS-A) and the depression subscale (HADS-D). Each subscale consists of seven items, and each item has a 4-point ordinal scale describing symptom severity, ranging from 0 (not present) to 3 (maximum). These scales are summarized into total scores, and sum score ranges from 0 to 21 on HADS-A and HADS-D, respectively. Possible diagnosis of anxiety and depression is defined by a score of 8 or above on each subscale.[23,28,31]

We examined three exposure groups based on HUNT 2 data; (i) anxiety (HADS-A ≥ 8), (ii) depression (HADS-D ≥ 8) and (iii) combined anxiety and depression (HADS-A ≥ 8 and HADS-D ≥ 8).[32] The HUNT 2 questionnaire included 14 HADS items regarding the participant's feelings during the last week.

First, the total score for each of the two different subscales was dichotomized and analysed separately as:

  1. (i) no anxiety (score 0-7, reference group) and (ii) anxiety (score ≥8) and

  2. (i) no depression (score 0-7, reference group) and (ii) depression (score ≥8).

Secondly, the total score for each subscale was categorized into three categories, with the third category set to the 99th percentile of the study population, i.e. anxiety and depression 'high-risk groups'.[33] Thus, these variables were categorized as: (i) no anxiety/no depression (score 0-7, reference group), (ii) anxiety/depression (score 8-12) and (iii) high-risk anxiety/high-risk depression (score 13-21).

Finally, combined anxiety and depression was analysed in one variable and categorized as: (i) no anxiety or depression (reference group), (ii) anxiety without depression, (iii) depression without anxiety and (iv) both anxiety and depression.

The Hospital Anxiety and Depression Scale was not included in HUNT 1, but four questions concerning symptoms of anxiety and depression, addressing issues of nervousness, calmness, mood and vitality during the last month, were included in both HUNT 1 and HUNT 2. These items could not differentiate sufficiently between anxiety and depression, however. Therefore, in a previous HUNT study, Bjerkeset and colleagues validated a compound index (i.e. anxiety and depression index) based on these four items against HADS in the total HUNT 2 population.[31] Cut-offs of the total index score were set to the 88th percentile of the HUNT 2 study population to match HADS cut-offs. This provided an acceptable indicator whether psychiatric caseness was present in HUNT 1.[31] Thus, we added four items included in HUNT 1 into a total score (sum score range: 0-18) that was dichotomized into: (i) no anxiety or depression (score 0-7, reference group) and (ii) anxiety and depression (score ≥8).

The total anxiety and depression score was further categorized into three categories, with the third category set to the 99th percentile of the study population.[33] Thus, this variable was categorized as: (i) no anxiety or depression (score 0-7, reference group), (ii) anxiety and depression (score 8-10) and (iii) high-risk anxiety and depression (score 11-18).

Subjects not answering all questions that were part of the anxiety and depression indices were excluded from the analyses.

Coping. Two different coping styles in relation to anger control have been defined; open coping, primarily directed towards the aggressor and covert (or passive) coping, directed inwards or towards others.[25]

The HUNT 2 questionnaire contained two items, previously employed and derived from epidemiological studies of hypertension,[25] where the study participant was asked to describe his or her response in relation to two statements. The first statement concerned how the participant expressed his or her anger, and if other people knew whether they were angry. Covert coping was indicated by the response alternatives 'almost never' or 'sometimes', while overt coping was indicated by 'rather often' or 'almost always'. The second statement dealt with whether the participant when boiling with rage, did not show this rage to other people. The response alternatives 'rather often' or 'almost always' indicated covert coping, and 'almost never' or 'sometimes' indicated overt coping. We dichotomized these two variables into: (i) overt coping (reference group) and (ii) covert coping.

Statistical Analyses

We used odds ratios (OR) and 95% confidence intervals (CI), estimated by unconditional logistic regression,[34] to assess the association between anxiety, depression and coping and risk of reflux symptoms. Data were analysed using the PROC GENMOD procedure in SAS.[35] Each psychiatric or psychological factor was evaluated separately in crude models and in models adjusted for variables known a priori to be independently associated with risk for reflux symptoms, i.e. age (in 10-year age groups), sex, tobacco smoking [duration in years of daily smoking in three categories: (i) never or <1 year (reference group), (ii) 1-9 years and (iii) ≥10 years], BMI [kg/m2, in four categories based on the World Health Organization (WHO) classification of overweight and obesity: (i) <25 (reference group), (ii) 25-30 (overweight), (iii) 31-35 (obesity) and (iv) >35 (severe obesity)], and socioeconomic status based on education [in three categories: (i) high education (reference group), (ii) medium education and (iii) low education]. These variables were cross-sectional except for tobacco smoking, where lifetime exposure was assessed. Observations with missing data on any of the covariates included in the models were excluded from the analyses. Exposures including three categories were evaluated using the Wald test, which considers all categories of the variable and not just pairwise comparisons with the reference category.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.