Patient-reported Outcomes in Chronic Gout: A Report from OMERACT 10

Pain

Pain is a cardinal symptom of acute gout flare. Pain is an important feature of chronic gout as it is for other inflammatory arthritides including RA and spondyloarthropathies. The importance of pain as a feature of chronic gout is exemplified by its inclusion in the inner circle for the domains (Figure 1).

Pain measurement by VAS was endorsed by 85% of voters as meeting the OMERACT filters of truth, discrimination, and feasibility based on data presented (Table 1). Data presented in preparation for discussion and voting came from 2 replicate randomized controlled trials (RCT) of pegloticase versus placebo¹² that enrolled 225 patients with tophaceous gout, across 56 centers. PRO data from these 2 RCT were combined (as per FDA and company’s agreement; note that the primary endpoint, as opposed to clinical endpoints, was separately analyzed) since the trials had the same methodology, design, duration (6 months), treatment interventions, and outcomes. Data presented were from baseline and end of the study visits. VAS pain has been used extensively, and therefore, has face and content validity; the specific question used in the studies is shown in Figure 2. Divergent validity was examined by comparing VAS pain scores across varying baseline tender and swollen joint counts, and SF-36 Physical Component Summary (PCS) and Mental Component Summary (MCS) score quartiles. For discrimination, VAS pain scores were compared between the treatment arms and placebo, and estimates of clinically important differences were presented. As a single question, VAS pain was regarded as feasible for the clinical trial setting.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Measures for pain and patient global assessment used in 2 RCT of pegloticase¹².

Although not formally voted upon, a separate analysis was also presented for SF-36 bodily pain domain as a measure of pain. The SF-36 instrument as a measure of HRQOL has been validated in gout and endorsed at OMERACT 9. Analyses were derived from the above RCT of pegloticase¹² and 2 RCT with febuxostat^13,14. Assessment of the “truth” component of the OMERACT filter for this measure included an assessment of divergent validity by baseline number of tender and swollen joint counts, baseline disease duration, and presence of tophi. Discrimination included comparison of treatment arms.

During the breakouts, several important aspects were discussed. Discussants agreed with using the SF-36 bodily pain domain as a measure of pain severity in chronic gout. Discussants agreed that a numeric rating scale (NRS) could be substituted for the VAS. Discussants also commented that pain could be influenced by acute flares, as well as by other chronic pain conditions such as concomitant osteoarthritis and/or back pain or fibromyalgia.

Patient Global Assessment Scale

Patient global assessment of disease scales are commonly used in RCT and longitudinal observational studies (LOS) and often used as a gold standard to determine clinically meaningful changes. Validation data were derived from the 2 above pegloticase RCT¹² and were similar to those presented for pain VAS (Table 2). Patient global met OMERACT filters of truth, discrimination, and feasibility and was endorsed by 73% of voters (Table 1).

Activity Limitations

Chronic gout is associated with chronic pain and inflammation, joint deformities and/or joint destruction, and deposition of urate as tophi in joints and subcutaneous tissues. Each of these can result in activity limitations in patients with chronic gout. These can be measured by general instruments such as Katz activities of daily living (ADL)¹⁵. Another commonly used instrument in rheumatic conditions is the Health Assessment Questionnaire (HAQ)¹⁶. The disability index of HAQ, HAQ-DI, has been used commonly in patients with RA and other rheumatic conditions¹⁷.

Data regarding HAQ-DI were derived from the pegloticase RCT¹² and a longitudinal observational study¹⁸. In the observational study, HAQ-DI scores were significantly higher in subjects with any tender and/or swollen joints (vs none) or presence versus absence of tophi, and were significantly different between patients who improved versus those with no change or deterioration¹⁸. HAQ-DI was also shown to fit a Rasch measurement model in gout patients and to correlate in expected ways with other measures of physical performance¹⁹. HAQ-DI was felt to meet the OMERACT filters of truth, discrimination, and feasibility with 71% of voters endorsing the measure.

Disease-specific Measure: GAQ v2.0

Recently, a disease-specific instrument to measure HRQOL, impact of gout during acute attacks, unmet patient needs, and medication side effects and concerns has been developed²⁰. A revised GAQ v2.0 was redeveloped following patient focus groups to assess various aspects of gout considered important to them.

The initial validation study of GAQ v2.0 was a 3-center cross-sectional study of 300 gout patients performed in the US from 2006 to 2008, focusing on the Gout Impact Section (GIS; Table 3)²¹. GAQ v2.0 was shown to have good face and content validity. Internal consistency for the Gout Concern Overall scale was good, with unadjusted Cronbach’s alpha of 0.88 and adjusted Cronbach’s alpha 0.94. Lower internal consistency was noted for 2 subscales: Medication Side Effects (0.60 and 0.86, respectively) and Unmet Gout Need (0.65 and 0.86, respectively)¹⁷, and 0.35 in the RCT²². Test-retest reliability for the Gout Overall Concern subscale was 0.77, and for Unmet Treatment Need 0.76²¹. Construct validity showed correlation coefficients of −0.16 for Gout Concern Overall with PCS and −0.28 with MCS; for the Unmet Gout Treatment Need, respective correlation coefficients were –0.15 and −0.24²¹. The poor correlation with generic HRQOL instruments raised concerns regarding construct validity (truth). In an unpublished study, Gout Concern Overall showed sensitivity to change, with change scores for “markedly improved” much higher than for those who reported lesser or no change on a global scale²². Validation data for the Gout Impact Section (Gout Concern Overall, Unmet Gout Treatment Need) of GAQ v2.0 was thought to be acceptable by 30% of voters, thus indicating that GAQ v2.0 was not yet endorsed as a validated measure in gout.

Breakout session discussions revealed several questions and problems regarding GAQ v2.0:

1. Lack of familiarity with the questionnaire — most people had not heard of it and the majority had not seen the questionnaire.

2. What does it measure? Discussants were not sure which domains it measured — activity limitations, disability, HRQOL, acute versus chronic gout impact, need, adverse events? They found the structure of the GAQ confusing — and considered only the Gout Impact Section (GIS) relevant.

3. Issues with internal consistency — Discussants expressed concerns regarding the internal consistency and construct validity of the Unmet Gout Treatment Need subscale, but felt comfortable with the Gout Concern Overall subscale.

Research Agenda

Several items are proposed for the research agenda related to PRO in gout. Pain, patient global, HRQOL, and activity limitations assessed by HAQ-DI or similar instruments are considered validated in chronic gout, and should be validated in acute gout — a focus for the next OMERACT. Several questions of the GAQ v2.0 make it a good subject for future research and validation. Other important suggestions from breakout groups include the following:

1. Items should be framed in relation to acute versus chronic gout, and functional limitation and other items should be framed specifically for gout.

2. The PROMIS network should be explored for additional measures of PRO.

3. Minimum clinically important differences should be anchored by perceived change scores rather than the change in disease status scores.

4. Further measurement of intermittent versus persistent pain, as well as impact of pain, should be considered.

5. Other PRO, such as fatigue, sleep, and social/work participation, should be explored to assess the impact of chronic as well as acute gout on the entire range of PRO.

6. In patients with chronic gout, distinction of PRO obtained during an acute flare versus between flares should be noted.

Summary and Conclusions

Of the PRO, HRQOL measurement with SF-36 previously met the OMERACT filters and was endorsed at OMERACT 9. Based on data from 5 RCT and 2 longterm observational studies, PRO measures that met the OMERACT filters of truth, discrimination, and feasibility included the following: pain and patient global assessments by VAS and activity limitations by HAQ-DI. Additionally, most voters agreed that NRS could substitute for the VAS pain scale to allow data collection over the telephone, using personal digital assistants, and mailed surveys. The disease-specific composite measure, GAQ v2.0, did not meet the OMERACT filters; more research is required for its validation. With the endorsements obtained at OMERACT 9 and OMERACT 10, the gout working group has made significant advances. All 4 PRO domains included in the inner circle of chronic gout outcomes now have at least one validated outcome measure that both meets the OMERACT filters and has been endorsed by OMERACT. These PRO measures should be used in both RCT and LOS to assess the impact of disease and treatment interventions for gout. Nevertheless, this endorsement should not be interpreted as implying that additional measures of these domains should not be considered. However, any additional measures need validation data and endorsement. An important item on the research agenda for the PRO subgroup of the gout group includes the examination of PRO measures in acute gout.