Sarawate, Chaitanya A. MS*; Patel, Pankaj A. MS, PharmD†; Schumacher, H Ralph MD‡; Yang, Wenya MPA*; Brewer, Kathleen K. MS, PhD*; Bakst, Alan W. MBA, PharmD†
Gout is characterized by recurrent episodes of acute inflammatory arthritis, tophi, and/or urinary tract stones. The prevalence of gout appears to be increasing, eg, from 0.5% of patients with self-reported gout cases in 1969 to 1.0% in 1996.1 In a managed care population, annual prevalence rates increased from 0.29% from 1990 to 0.52% in 1999.2 This increase has been attributed to increasing longevity, an increase in obesity, and use of drugs that may increase urate levels.3
The biochemical basis of gout is supersaturation of serum urate (SUA) in extracellular fluids resulting in precipitation of urate crystals. Although hyperuricemia does not always result in gout, higher levels have been associated with gout flares.4 The target goal recommended in treatment of gout as well as in clinical trials of prospective gout medications is to lower the SUA to the subsaturating range, <6.0 mg/dL.5,6
The relationship between SUA and gout flares was studied in a retrospective study of Japanese patients with gout who were followed for up to 3 years.4 Mean SUA was 7.2 mg/dL in the flare subgroup and 6.5 mg/dL in the no-flare subgroup. In adjusted analyses, patients with lower SUA were 58% less likely to have recurrent gout flares (odds ratio [OR], 0.42; 95% confidence interval [CI], 0.31–0.57). Patients on antihyperuricemic medication therapy were 78% less likely to have gout flare (OR, 0.22; CI, 0.10–0.47).
The present study was a 2-year, nonrandomized, retrospective, database analysis of patients identified as having gout from a managed care perspective. The objectives of the study were to determine the relationship among gout-specific prescription drug therapy, SUA level, and gout flares among adult gout patients. This study addressed a gap in the literature of the relationship between SUA level and gout attacks in a naturalistic setting.
MATERIALS AND METHODS
The managed care database was a large southeastern U.S. health plan with 2.2 million lives at study initiation. The lines of business within the plan included a health maintenance organization and preferred provider organizations. The database included medical claims, pharmacy claims, electronic laboratory results, and health plan eligibility. The data elements were similar among the lines of business. Data from January 1, 1999, to March 31, 2004, were used in this study.
Patients with either of the following 2 criteria were selected: 1) ≥2 visits with an International Classification of Disease, 9th Revision, Clinical Modification (ICD-9) code for gout (274.xx) or 2) ≥1 pharmacy claim for medications with gout-specific indication (colchicine, allopurinol, probenecid, or sulfinpyrazone). The date of the first such claim between January 1, 2000, and December 31, 2002, was the index date. Patients must have been ≥18 years of age as of the index date. A study schema is shown in Figure 1.
Insurance eligibility criteria must have been met continuously for 1 year pre- and 1 year postindex date. Patients taking allopurinol who had an ICD-9 code for neoplasm or nephrolithiasis were excluded to increase the specificity of the gout diagnosis.
Patients in the gout cohort were differentiated as newly or previously diagnosed with gout. Differences in severity and treatment patterns were expected between patients with chronic gout versus a new diagnosis of gout. Newly diagnosed patients were defined as those who did not have a gout medical or pharmacy claim within the 12-month period before the index date. Patients with previously diagnosed gout had a gout medical or pharmacy claim within the 12-month period before the index date.
Preindex comorbidities identified from ICD-9 codes and selected gout-relevant specific diagnoses were present for a minimum of 1 year preindex date. The Deyo-Charlson Comorbidity Index (DCI), which is based on predetermined ICD-9 codes,7,8 was used to estimate the baseline chronic disease burden in the patient cohort.
Allopurinol dose was calculated as quantity dispensed multiplied by strength divided by days’ supply for each pharmacy fill from initiation to discontinuation or end of study during continuous therapy. Mean dose was the average of all fills, modal dose was the most frequent dose, median was the midpoint dose, and last dose was the final allopurinol fill during continuous therapy. Because the allopurinol dosage was almost identical regardless of the method of calculation, we used modal dose in all analyses.
Serum Urate Level
Actual SUA results were obtained from a single national laboratory provider. Mean baseline, postindex SUA, and mean change were calculated for patients with SUA results available before and postindex date.
Gout Flare Algorithm
Gout flare was defined as an office or emergency department (ED) visit with medical claim for gout or joint pain (ICD-9 code 719.4x) that may have occurred on the index date (date of first claim). Patients must also have had ≥1 of the following claims with these pharmacy or procedure codes within 7 days postoffice or ED visit: nonsteroidal antiinflammatory drug, colchicine, oral corticosteroid, adrenocorticotrophic hormone, intraarticular aspiration/injection (CPT codes 20600, 20605, 20610), or microscopic examination of specimen from musculoskeletal system/joint fluid (ICD-9 procedure code 91.5).
Descriptive statistics included means (± standard deviation [SD]) and relative frequencies for continuous and categorical data, respectively. For any inferential analyses, an a priori 2-tailed level of significance was set at the 0.05 level.
Multivariable logistic regression and negative binomial regression analyses were conducted to evaluate the association between SUA levels ≥6.0 mg/dL and risk and rate of gout flares, respectively. Risk of gout flare was expressed as an odds ratio; rate of gout flare was expressed as an incidence rate ratio (IRR).
All statistical analyses were performed using Stata version 8.0 (Stata Corp., College Station, TX).
A study cohort of 5942 gout patients ≥18 years of age was identified from the database. The prevalence of gout as defined above among all eligible patients in the population was 1.6% (5942 of 369,356 eligible patients ≥18 years of age). The majority of patients (3651; 61.1%) were considered newly diagnosed (no medical claim for gout or pharmacy claim for a gout-specific medication within 12 months before index). Mean postindex follow-up time was 29 months; maximum was 51 months.
Within the gout cohort, 2075 patients (35%) had a medical encounter(s) for gout flare. Among these patients, 3896 flare events were captured in claims data; mean of 0.66 ± 1.30 flare per gout patient; mean of 1.86 ± 1.60flares among patients with flare. Gout flare was captured for 1437 (40%) of newly diagnosed and 638 (28%) of patients with previously diagnosed gout.
Among patients with flare, there were significantly more males (82%) compared with the nonflare group 74% (P < 0.001) (Table 1). Mean age of patients with flares was 56 ± 14 years and was greater than the mean age of patients in the entire health plan database (41 years). The greatest frequency of patients with flares was among middle-aged patients 50 to 69 years (43%). Within the gout cohort, there were statistically significant differences in mean age between patients with and without flare (P < 0.001). As indicated by the mean DCI score, an indicator of chronic disease burden, median DCI was zero for both groups, indicating that the flare and nonflare cohorts did not differ in terms of their comorbidities.
Significantly more patients with flares had tophi captured in medical claims data postindex (47%) compared with patients with gout without flare (18%) (P < 0.001). An ICD-9 code for hyperuricemia (790.6) was only reported for 7.8% of patients with gout with flare.
Allopurinol Dose and Other Gout Therapies
Allopurinol was the most prevalent gout therapy (n = 2405); modal daily dose ranged from ≤100 mg/day to ≥600 mg/day. Among patients on allopurinol, the modal daily dose taken by the greatest percentage of patients (1560; 64.9%) was 300 mg/d. Modal allopurinol doses of <300 mg/day were taken by 776 patients (32.3%). A very small minority of patients (69; 2.9%) received modal allopurinol doses of >300 mg/d. Results were similar by last allopurinol dose taken over follow up, eg, only 84 patients (3.5%) patients received allopurinol doses of >300 mg/d as the last dose during the study. Other urate-lowering medications (probenecid and sulfinpyrazone) were taken by <10% of patients.
Baseline and Postindex Serum Urate Levels
SUA levels at baseline and after the start of gout drug(s) were examined. Among patients with gout who received gout-specific drug therapy, 266 had pre- and postindex SUA result data. Among patients on allopurinol with pre- and postindex SUA data (n = 162), mean preindex SUA was 8.7 (±1.9) mg/dL (median, 8.9 mg/dL) and mean postindex SUA was 7.1 (± 1.8) mg/dL (median, 7.1 mg/dL). The reduction of SUA from the preindex to postindex period was significant (P < 0.001) (data not shown). However, only 25% (36 of 147) of allopurinol users with SUA ≥6.0 mg/dL preindex achieved SUA <6.0 mg/dL postindex (Table 2). Similarly, only 21% (49 of 239) of total patients on any gout medication(s) at index who had SUA ≥6.0 mg/dL preindex achieved SUA <6.0 mg/dL postindex.
Serum Urate Level and Gout Flare
For total patients with gout with flares and SUA data, median time between last SUA test and gout flare was 5 months. Mean preflare SUA level was 8.0 ± 2.0 mg/dL. Median SUA testing followed gout flare by approximately 1 month. Mean postflare SUA was 7.7 ± 2.1 mg/dL, which was higher than the target of <6.0 mg/dL (data not shown).
Among patients on allopurinol, significantly fewer patients with postindex SUA <6.0 mg/dL (23%) had flares relative to patients with postindex SUA ≥6.0 to 8.0 mg/dL (33%) or ≥8.0 mg/dL (45%) (P < 0.05). Results were very similar in all other gout-specific medication therapies (data not shown).
Among patients with postindex SUA results on gout medication therapy (n = 1315), those with SUA ≥6.0 mg/dL were 59% more likely to have gout flare (OR, 1.59; 95% CI, 1.21–2.09) than patients with SUA <6.0 mg/dL (Fig. 2). In addition, patients with postindex SUA ≥6.0 mg/dL were 49% more likely to experience additional flares (IRR, 1.49; 95% CI, 1.20–1.84) as compared with patients with postindex SUA <6.0 mg/dL regardless of gout-specific therapy.
Among allopurinol users with postindex SUA test (n = 735), patients with SUA ≥6.0 mg/dL were 75% more likely to have a gout flare than patients who achieved target SUA <6.0 mg/dL (OR, 1.75; 95% CI, 1.21–2.52) (Fig. 2). In addition, patients on allopurinol with postindex SUA ≥6.0 mg/dL were 72% more likely to experience additional flares (IRR, 1.72; 95% CI, 1.26–2.35) as compared with patients with postindex SUA <6.0 mg/dL.
In this retrospective, observational study, a minority of gout patients underwent SUA testing or attained target SUA level <6.0 mg/dL. In addition, 80% of patients with symptomatic flare on all urate-lowering therapies in the gout cohort were biochemically hyperuricemic. Furthermore, SUA levels ≥6.0mg/dL were associated with 59% greater likelihood of gout flare, irrespective of gout therapy. In the present study, the median allopurinol dose was 300 mg daily. Too few patients received allopurinol dose >300 mg/day for analysis by dose. In a naturalistic setting, suboptimal dosing of allopurinol appears to be occurring.
Failure to attain target SUA levels and the occurrence of gout flares in the present study may have multifactorial origins. Physicians and patients may be unaware of the need for testing and the current recommended SUA target. Current practice patterns for allopurinol such as suboptimal dosing patterns and patient compliance may also be contributing. Excessive alcohol use may also have been a factor for some patients. The possibility that allopurinol efficacy is not optimal is also a possibility.
In a prospective study of allopurinol and benzbromarone in male patients with gout, Perez-Ruiz et al9 found that 53% of patients achieved SUA <6.0 mg/dL on 300 mg allopurinol daily. Patients with poor results on 300 mg allopurinol daily were titrated to 450 to 600 mg daily. Although some similarities were evident, the Perez-Ruiz study differed from the present study in that it was prospective, patients had regular clinic visits and could be determined as urate under- or normal excretors, obese patients were on a reducing diet, and noncompliant patients were excluded.
The methodology used in the present study differs from a Japanese study in which the study sample was culturally different, different urate-lowering medications were available, gout flares were self-reported, and routine monthly SUA levels were obtained.4 The Japanese study was also biased to patients with ≥1 prior gout flare. However, similar to this study's results, the Japanese study also demonstrated that currently available urate-lowering medications as used were often not effective. Among patients with gout on urate-lowering therapy who reported flares (n = 69), mean SUA over the study period was 7.01 ± 0.10 mg/dL. A significant association between SUA level and gout flare was also reported in the Japanese study.
This was a nonrandomized, retrospective study using medical, pharmacy, and laboratory claims data from a single health plan. This study was biased to patients with health plan coverage; no Medicaid patients were captured. Patients in other regions of the United States with different ethnic and occupational mixes may respond differently to hyperuricemic medication therapy. However, gout treatment patterns are unlikely to be regionalized because there are presently few medication treatment options for hyperuricemia.
Gout diagnosis and gout flares were detected in claims data. Clinical validation, including crystal analysis, of gout and flares was not performed. An unvalidated proxy for gout flare was used in this study, which may have biased the study to patients with more severe gout or patients with other comorbidities that required direct medical intervention as opposed to patient self-management. Acute gout flares were most likely underrepresented, because many patients with gout self-manage acute episodes.
Unlike a controlled clinical trial, SUA results were not available for all patients pre- and postindex in this observational study. Testing for SUA may have been performed more aggressively for patients with known hyperuricemia or more severe gout. All patients did not have an administrative claim for SUA testing. Some patients may have had testing performed at an out-of-network laboratory; which could not be detected in the managed care organization's database.
This study analyzed only interventions for hyperuricemia and gout flares. First-line interventions for gout management also include dietary modifications, reduction in alcohol use, weight loss, and use of nondiuretic therapies for hypertension,3 which were not evaluated in this study.
There is significant medical need for more effective education and compliance, and possibly other pharmacotherapy for the management of gout and hyperuricemia. Significant percentages of patients are not attaining target SUA control with current urate-lowering medication therapy. Patients with nontarget SUA levels (≤6.0 mg/dL) on urate-lowering medication have greater likelihood of gout flares. Failure to attain target SUA levels may have multifactorial origins. This study provides benchmark data for assessing the impact of future uric acid-lowering strategies and medications from the clinical and managed care organization perspective.
Christy Fang, MS, contributed to the data extraction phase of this study. Lauren Burawski, MS, contributed to manuscript preparation.
© 2006 Lippincott Williams & Wilkins, Inc.