Frequency, Risk, and Cost of Gout-related Episodes Among the Elderly: Does Serum Uric Acid Level Matter?

DISCUSSION

The existing literature has not examined the relationship between hyperuricemia and gouty episodes, or flares, from an epidemiological perspective, especially among the high-risk elderly, nor have studies investigated the economic burden of gout flares. Our study addressed this void by examining the association between sUA and the frequency, risk, and costs of gout flares among 2,237 elderly patients from a large nationally representative claims database spanning the years 1999 through 2005. The bulk of the reduction in sample size was due to the 2-year continuous eligibility criteria and the age cutoff, 65 years. A sizable reduction was also due to the requirement that patients have at least one SUA laboratory test value. Among the 2,237 patients who met the inclusion criteria, we found that higher SUA levels were associated with increased risk and frequency of gout flares among elderly patients, as well as higher total and gout-related healthcare costs per flare.

More specifically, we found that about 40.9% of patients experienced at least one gouty episode in the 12-month period following the date of the earliest observed SUA laboratory test since January 1, 2000 (the index date). The average number of gout flares was 1.5 (SD 0.8) for patients with an SUA level < 6 mg/dl, 1.6 (SD 0.9) for those with SUA 6–8.99 mg/dl, and 1.7 (SD 1.2) for those with SUA ≥ 9 mg/dl. Using negative binomial regression, we found that among patients with SUA values > 6 mg/dl, the average number of gout flares increased by 11.9% (95% CI 7.1% to 17.0%) with each unit-increase in SUA. The study also addressed the economic implications of having a high SUA level. We found that a high SUA level was associated with increases in both total and gout-related direct healthcare costs. Compared to gout flares among patients with normal SUA levels (< 6 mg/dl), gout flares among patients with very high SUA levels (≥ 9 mg/dl) were associated with approximately $2,500 in additional total healthcare costs during the 30-day period following the index date.

Our findings are consistent with those from Campion, et al, whose analysis of 2,046 healthy men from the Normative Aging Study showed that SUA level was a clear predictor of acute gouty arthritis¹⁴. Our findings are also generally consistent with 2 retrospective studies^16,17. One, by Shoji and colleagues, was based on an outpatient clinic in Japan¹⁶. Their analysis found that each unit-reduction in SUA was associated with a reduced risk of recurrent gout flares during a followup period of up to 2 years (OR 0.42, 95% CI 0.31–0.57). The second study, Sarawate, et al, investigated the association between SUA and risk and rate of gout flares using managed-care data in the US¹⁷. Their results demonstrated that patients with SUA levels ≥ 6 mg/dl were more likely to experience a gout flare within one year compared with patients who had SUA levels < 6 mg/dl (OR 1.6, 95% CI 1.2–2.1).

Given the clinical and economic benefits of maintaining an optimal SUA level, appropriate therapies for lowering SUA levels should be considered. Behavioral therapy typically attempts to reduce SUA through changing modifiable risk factors, such as diet, alcohol consumption, and incorrect medications, but these approaches frequently fail due to poor compliance³⁵. Pharmacological treatment is then often required as part of the therapy. Researchers have found that in order to prevent recurrent attacks of gout, SUA should be maintained at 4–6 mg/dl³². Use of urate-lowering agents can effectively reduce SUA level, and can also reverse the deposition of monosodium urate, which causes the inflammatory response in gout flares³⁶. Previous studies demonstrated that urate-lowering drug treatment, usually a lifetime commitment, can be cost-effective or even cost-saving in cases of nontophaceous recurrent gouty arthritis²⁰. The development of new pharmacotherapies in treatment of hyperuricemia may present a good opportunity for prevention of gout flares and reduction in healthcare costs related to gout.

Our study has limitations. First, our findings may be specific to elderly patients with gout, as our study population consisted only of individuals aged 65 years and older. Also, it is important to note that the risk of flares is not zero among those patients with SUA levels below the normal range, even though this range is considered to be below the saturation point for monosodium urate. One reason is because SUA is an indirect indication of joint tissue urate levels. SUA levels can vary over time, and normal SUA ranges are influenced by other factors, such as sex and obesity³³.

A second limitation pertains to the definition of gout flares. To our knowledge, no standard epidemiological definition for gout flares exists. Therefore, we used a definition based on ICD-9 codes and medications, and included 2 definitions to address the fact that some information may have been missing in the claims and laboratory data. Neither definition is ideal, however. For example, the first definition could result in false-positives if NSAID are used as prophylaxis, and the second could result in false-negatives because the criteria may be too stringent. Thus, the goal of our study was to assess the risk of gout flares by SUA level. As long as the frequency of any false-positives or false-negatives is consistent across each of the SUA levels, the main findings will remain unbiased. The Outcome Measures in Rheumatology (OMERACT) group, an informal international network studying outcome measurement in rheumatology, is currently working on an appropriate epidemiological definition for gout flare.

Further, gout flares were measured using claims-based diagnoses only, and could not be verified using a second source, such as chart review or direct contact with patients. It is possible, therefore, that some gout patients with flares were not identified in our database because a gout claim did not exist, possibly because the patient did not contact a physician and instead used an over-the-counter NSAID. On the other hand, some patients may have been characterized as having had a gout flare if a physician wrote a prescription for an acute gout medication for the patient to use in case of emergency. This latter issue may be particularly relevant for our study if physicians are more likely to prescribe an acute gout medication when a gout patient has a high SUA level compared to when a patient has low SUA. In such instances, our analysis may expose the effects of gout care, as opposed to gout flares per se. Finally, it is possible that patients were prescribed an NSAID or steroids for the treatment of another condition other than gout, such as low back pain. To the extent that such instances exist among our sample of gout patients, our estimated incremental costs of gout would likely be underestimated due to the presence of patients without a gout flare in the gout cohort.

A recent report evaluated the validity of gout diagnoses based on administrative claims databases by cross-checking administrative claims information with a medical chart review conducted by 2 rheumatologists³⁷. Their findings showed that 39% of patients with at least 2 claims-based gout diagnoses over a 5-year period did not have gout or did not have sufficient medical records information to confirm the gout diagnosis (i.e., false-positive). Again, if this was the case in our analyses, then our estimated incremental costs of gout would likely be underestimated due to the presence of gout-free patients in the gout cohort. Separately, use of claims data may also underestimate the true number of gout flares because many flares may be cared for by patients without seeing a physician and without prescription medication (e.g., patients who use over-the-counter medications).

Another potential limitation is that the sample may not be representative of all Medicare patients because Medicare fee-for-service patients, most of whom would not have prescription drug coverage, were not included. Further, most gout patients with a valid SUA value from electronic laboratory data were concentrated in the mid and south Atlantic region, which resulted in a study sample that was not representative of all geographic locations. The sample with an SUA also had fewer comorbidities on average than all gout patients, as indicated by the Deyo-Charlson comorbidity index (1.061 vs 1.146, respectively).

Each of the regression models controlled for age, sex, index year, region, prior comorbidities (hypertension, diabetes, hyperlipidemia, renal impairment, rheumatoid arthritis, lupus), and the Deyo-Charlson comorbidity index. Other confounding factors, such as body mass index, may also influence the relationship between SUA levels and the risk of gout flares, the number of gout flares, or the direct healthcare costs of gout flares. Further, costs were measured as those incurred by patients during the 30-day period following the onset of a flare. These costs could reflect factors beyond gout severity, such as increased healthcare costs due to nonadherence. The relationship between SUA level and cost could also be partly explained by a pattern of behavior in which some physicians requested to see high SUA patients more often than others as a precaution, rather than the decision being a patient-driven one in which healthcare was sought due to illness per se. Regardless of the underlying decision, increased SUA levels were associated with higher costs per flare. Finally, the costs per gout flare reported here could be higher than for a typical flare observed in a “real-world” setting, because mild flares that do not require the attention of a physician may have been excluded. Nonetheless, because the cost of a flare was defined consistently across all SUA levels, our findings in terms of the relationship between SUA and cost per flare remain valid.

Gout is a common disease in the US, and is particularly prevalent among the elderly population. Using claims data for a group of elderly (≥ 65 yrs) patients with gout, we found that patients with very high SUA levels (≥ 9 mg/dl) had more frequent gout flares compared to those with normal SUA levels (< 6 mg/dl) and had a significantly higher risk of experiencing a gout flare. A very high SUA level was also associated with increased direct healthcare costs of gout flares. Reductions in SUA are therefore likely to lower the economic burden of gout through the reduction of both the number of flares and the average cost per flare. Given the substantial effects of higher SUA levels observed in this study, it is desirable from both a clinical and an economic perspective to focus on controlling SUA levels as a method of preventing gout flares.