Table 2. Association between any diuretic use as well as specific type of diuretic use over the last 48 hours and risk of recurrent gout attacks.
Thiazide Only (n = 39)
Loop Only (n = 14)
Loop or Thiazide or Both (n
Adjusted for alcohol consumption and purine intake.
Adjusted OR* (95% CI)
for such an increase is the increases in the prevalence of hypertension and concomitant diuretic therapy18-20. For instance, the prevalence of hypertension in the US has increased from 25.0% during 1991-98 to 28.7% during 1999- 200020. Recently, the Joint National Commission 7th Report on Prevention, Detection, Evaluation and Treatment of Hypertension recommended thiazide-type diuretics as initial therapy in most patients either alone or in combination with another agent21. Since hypertension impairs urate secretion and promotes hyperuricemia26, the association between hypertension and gout is further compounded by the adminis- tration of diuretic therapy. Clarification of the effect of diuret- ic use on the risk of recurrent gout attacks has important clin- ical implications. We hypothesize that abrupt changes in diuretic use may lead to acute increases in uric acid concen- tration, thus increasing the appropriate conditions for crystal formation or precipitation and modulating their phlogistic potential through perturbation of the local synovial microen- vironment27.
There are a number of alternative effective agents for the treatment of hypertension and congestive heart failure in per- sons with gout that are unlikely to predispose to recurrent gout attacks. Clinicians have ample ability to individualize man- agement for special populations28 and could reduce the risk of gout attacks by avoiding the use of diuretics in persons with preexisting gout. In our study, about 28% of participants with a history of gout were taking diuretics, and the effect of these diuretics on recurrent gout attacks was not trivial.
Several characteristics of our study are noteworthy. First, studying the triggering effect of diuretics on the risk of recur- rent gout attacks is challenging. Neither case-control nor cohort studies are ideal study designs for such research ques- tions. In our study, we applied 2 innovative approaches, a case-crossover study design and use of the Internet, to exam- ine whether diuretic use could trigger recurrent gout attacks. The case-crossover study design uses each subject as his/her own control and compares the frequency of exposure to a sus- pected precipitating factor immediately prior to disease onset 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. Second, using the Internet as a medium, we were able to reach patients with gout in the entire US. Indeed, participants in our study were recruited from 40 states and the District of Columbia. Further, we showed that the majority of subjects’ history of gout could be verified by their medical records and these subjects can be efficiently followed over time. Third, we showed that both risk factors and disease occurrence can be assessed in real time, which should minimize the potential recall bias.
Our study has some limitations as well. We did not collect information on dosages of medications used; we were unable to assess a dose-response relationship between diuretic use and risk of recurrent gout attacks. Also, the number of subjects using diuretics in our study was relatively small; this was especially the case for loop users.
While we used validated questionnaires to assess risk fac- tors for gout, including diuretic use, and subjects were asked to recall these putative risk factors occurring within the last 48 hours, it is still possible that misclassification of risk factors may have occurred. Such misclassification, if it occurred, is likely to be nondifferential, and would bias the results toward the null. Thus, we postulate that the true effect of diuretics may actually be larger than we observed.
There are other potential important confounders that could have varied between control and hazard periods including adi- posity (e.g., weight and body mass index) and use of other medications known to affect serum urate levels (e.g., beta- blockers, low-dose aspirin use, and allopurinol). Further, our study design was not immune to “confounding by indication” for diuretic use, such as developing or worsening of hyperten- sion, congestive heart failure, and renal insufficiency, that may also trigger recurrent gout attacks. While it is possible these risk factors may occur differently during the hazard peri- od compared to the control period, we believed the frequency of these risk factors, such as congestive heart failure or renal insufficiency, would be very low in our study participants. Developing or worsening of hypertension is likely to be a
The Journal of Rheumatology 2006; 33:7