The use of questionnaires concerning heartburn and acid regurgitation is a well-validated method of assessing the true occurrence of reflux. Nonetheless, we conducted a separate validation study to compare the question regarding reflux symptoms in HUNT-2 with that in another, more comprehensive questionnaire. In this validation study, 1,102 outpatients from the county of Nord-Tr0ndelag in Norway and Karolinska Hospital in Sweden were included as previously described. In summary, 95% of the participants with severe reflux symptoms (corresponding to our case group) had reflux symptoms at least once a week, which resulted in a specificity of 99.5% for reflux symptoms occurring at least once per week in our case group. comments
The participants were asked about various respiratory disorders experienced during the past 12 months: (1) daily cough and, in the case of a positive answer, for how many months and whether the cough was productive; (2) breathlessness, with three alternative answers: no symptoms, minor symptoms, or major symptoms; (3) attacks of heavy breathing or wheezing during the past 12 months; (4) occurrence of asthma; and (5) use of asthma medication.
All respiratory variables under study except breathlessness were encoded as dichotomous, indicating presence or absence of the symptom or condition. In the logistic regression analysis, severe reflux symptoms, as defined above, served as the outcome variable. Chosen as potentially confounding factors were the following: age (categorized into 10-year intervals); sex; body mass index (BMI) [body weight in kilograms divided by the square of body height in meters; grouped as < 25 kg/m2, 25 to 30 kg/m2, 30 to 35 kg/m2, and > 35 kg/m2]; smoking (years of daily smoking classified into < 1 year, 1 to 10 years, and > 10 years); use of alcohol (number of occasions of drinking alcoholic beverages during the last 2 weeks classified as none, 1 to 4 drinks, 5 to 10 drinks, and > 10 drinks); occurrence of asthma (yes or no); and use of asthma medication (yes or no). For all these variables, the data were cross-sectional, except for smoking, for which we used data representing lifetime exposure. A multivariable, unconditional logistic regression analysis (GENMOD, SAS 8e; SAS Institute; Cary, NC) was performed to calculate odds ratios (ORs) and their 95% confidence intervals (CIs) that were used to estimate the influence of each predictor variable. The potential confounders were tested by introducing them one by one into the model. Since occurrence of asthma and alcohol consumption did not contribute to explaining the variance, these variables were left out of the final multivariable model.