countries18-20. The Joint National Commission Seventh Report on Prevention, Detection, Evaluation, and Treatment of Hypertension recommends thiazide-type diuretics as initial therapy in most patients either alone or in combination with another agent21. Because of the favorable cost profile and efficacy, thiazide diuretics are commonly prescribed, ranking seventh among the most prescribed medications in the US in 200322, and this number does not include the use of thiazides in combination therapy. Thus, both increasing prevalence of hypertension20 and use of thiazide diuretic prescriptions are likely to lead to further increases in recurrent gout attacks.
Considering the relatively high prevalence of hypertension among gout patients and the widespread use of diuretics for the treatment of hypertension, we conducted an internet-based case-crossover study to assess the relation of diuretic use to the risk of recurrent gout attacks and evaluated whether the effect varies according to types of diuretic use. The case- crossover study uses each subject as his/her own control and compares the frequency of exposure to a suspected precipitat- ing factor immediately prior to disease onset (hazard period) to that during the control periods. Self-matching of each sub- ject eliminates bias in control selection and removes con- founding effects of factors that are constant over time. Thus, it is an optimal study design for examining the effect of risk factors triggering recurrent attacks, such as repeated gout attacks.
MATERIALS AND METHODS Study design. We constructed a website for this study (https://dcc2. bumc.bu.edu/GOUT) on an independent secure server within the Boston University School of Medicine domain. The study website provided informa- tion about the study, invited applicants to participate, administered a screen- ing questionnaire, linked eligible respondents to an online consent form, and administered additional questionnaires to assess risk factors and features of respondents’ recurrent gout attacks.
The study was advertised on the Google search engine (www.Google.com) by linking an advertisement to the search term “gout.” When a search was conducted containing the keyword “gout,” a study advertisement appeared on the screen. Interested individuals who clicked on the box containing our advertisement were immediately directed to the study website. To be eligible for the study, a subject had to report a diagnosis of gout by a physician, have had a gout attack within the past 12 months, be at least 18 years of age, reside in the US, and be willing to release medical records pertaining to gout diag- nosis and treatment.
Questionnaires. Eligible subjects were asked to complete the following ques- tionnaires: (1) sociodemographic questionnaire: name, age, sex, home address, home and work phone number, e-mail address, date of birth, years of education, and household income; (2) medical history questionnaire: medica- tion use, self-reported comorbidities, and history of gout attacks; (3) control- period questionnaire: frequency and quantity of potential risk factors, includ- ing all medications taken, for each day over the preceding 2-day period dur- ing the intercritical gout period. The questionnaire was administered at study entry, and at 3, 6, and 9 months of followup. Subjects were asked to report the occurrence of a gout attack within 48 hours after its onset and complete (4) a hazard-period questionnaire: date of recurrent gout attack, clinical symptoms, medication used to treat the recurrent gout attack, and frequency and quanti- ty of potential risk factors, including all medications taken, for each 24-hour period over the 48 hours prior to the attack. As part of the medical history par- ticipants were asked, ‘Did you take any of the following diuretics (“water
pills’) yesterday or the day before yesterday?” They were then provided with different options to check on the website, including: (1) hydrochlorothiazide or thiazide for hypertension (e.g., Dyazide, Esidrix, Hydrodiuril); (2) furosemide (“loop” diuretics) for congestive heart failure (e.g., Lasix); (3) amiloride, spironolactone, or triamterene (e.g., Midamor, Aldactone, Dyrenium); or (4) other diuretics (please specify).
Diagnosis of gout, and assessment of risk factors. We obtained medical records pertaining to the participant’s gout history and/or a checklist of gout symptoms completed by the subject’s physician. To confirm a subject’s diag- nosis of gout, we used information abstracted from medical records or the checklist, and data collected from the questionnaires submitted by study par- ticipant. A rheumatologist reviewed all medical records and the checklist information and assessed whether the subject had a diagnosis of gout by a physician, and a history of gout according to American College of Rheumatology (ACR) criteria for gout23.
Self-reported acute gout attacks were defined on the basis of: (1) moder- ate to severe joint pain developed within a 24-hour period; (2) 3 or fewer joints involved at time of the acute attack; (3) redness observed over joints;
joint(s) feels warm or swollen; and (5) attack requires specific therapy,
g., a nonsteroidal antiinflammatory drug, or colchicine. These modified cri-
teria are based on criteria published by the ACR23.
We also assessed a set of other putative risk factors, including alcohol con- sumption and total amount of purine intake. Questions on alcohol consumption included the number of servings of beer, wine, or spirits (either straight or in a mixed drink) consumed on each day over the control or hazard periods. The serving sizes were provided along with color images of standard drink sizes and containers, i.e., 12-ounce bottle or can of beer; 5-ounce glass of wine; and 1–1.5 ounces of spirits. Total purine intake from food for each day over 2-day period was estimated using established food purine contents24.
Statistical analysis. Using a conditional logistic regression model, we exam- ined the relation of all diuretic use over the 2-day period to the risk of recur- rent gout attacks while adjusting for alcohol consumption and purine intake using well described methods25. Data on diuretic use, purine consumption from food, and alcohol intake were collected during the same time period at each hazard and control period. In addition, we also evaluated the effect of specific type of diuretic use, i.e., thiazide or loop class, on the risk of recur- rent gout attacks.
RESULTS Of 197 subjects who completed both hazard-period and con- trol-period questionnaires, 179 (91%) subjects fulfilled the ACR criteria for gout. One hundred eighty-six subjects (94.4%) returned a signed medical record release form. Of those, we obtained 172 subjects’ medical records or physi- cian’s checklists from their physicians, and 163 (94.8%) met ACR criteria for gout.
Characteristics of the 197 participants who completed both hazard-period and control-period questionnaires are presented in Table 1. The average age of participants was 52 years. Participants were predominantly male (80%) and Caucasian (88%), and over half had a college education. Subjects were recruited from 40 states and the District of Columbia. During the one-year followup period, 321 recurrent gout attacks occurred among participants. The median time between the date of the gout attack and date of completion of hazard-peri- od questionnaire was 2 days. All participants included in this analysis completed at least one control-period questionnaire. The mean number of control-period questionnaires completed was 2.4 (SD 1.4, range 1–5) and hazard period questionnaires was 1.7 (SD 1.1). Among the subset who took diuretics, the
The Journal of Rheumatology 2006; 33:7