Development of a Multivariable Improvement Measure for Gout

Naomi Schlesinger; N. Lawrence Edwards; Anthony E. Yeo; Peter E. Lipsky

Disclosures

Arthritis Res Ther. 2020;22(164) 

In This Article

Results

GMIM components (i.e., sU < 6 mg/dL, flares, tophi, PGA, pain, and swollen and tender joints) were determined from previous analyses of the patient data in the RCTs of pegloticase. A combination of criteria from the proposed remission criteria[33] and also criteria for a complete response[34] were considered. Although some criteria contributed little to the complete response model, such as pain and flares,[34] they were included in the final GMIM because they were thought to be important in evaluating patients with gout. Moreover, it was felt that the way the data had been collected in the RCTs may have decreased its discriminative value.[34] For example, pain was collected as a general condition and not as related to gout. The previous exercise had evaluated the criteria strictly as a means to evaluate the induction of remission or complete response.[34] Here, we employed these criteria to determine the degree of response. Initially, the entire group of 85 subjects receiving biweekly pegloticase was assessed for the frequency of subjects achieving GMIM20, 50, and 70 status and compared to the outcome of subjects receiving placebo (Figure 1a, Supplementary Table 1). No placebo-treated subjects achieved GMIM status, whereas approximately 40% of pegloticase-treated subjects became GMIM20 responders after the 6-month RCT. Since some pegloticase-treated subjects only had a transient urate-lowering effect, we were interested in determining whether pegloticase might have a clinical benefit even if the contemporaneous sU was not < 6 mg/dL. To accomplish this, we relaxed the GMIM status by omitting the requirement for a sU < 6 mg/dL. As can be seen in Figure 1b and Supplementary Table 1, the frequency of GMIM responders was considerably higher in the pegloticase-treated subjects when the urate lowering requirement was removed, with approximately 60% of pegloticase-treated subjects achieving a GMIM20 and 30% a GMIM70 after 6 months of treatment. However, in this circumstance, nearly 30% of placebo-treated subjects also achieved a GMIM20 by 6 months, but only infrequently achieved a higher level of response.

To explore the relationship between persistent urate lowering caused by pegloticase and the percentage of subjects achieving GMIM responses in greater detail, we divided the pegloticase-treated subjects into those with persistent urate lowering (responders) and those with only transient urate lowering (nonresponders) and analyzed the percentage achieving GMIM responses. Thirty-four of 85 subjects treated with biweekly pegloticase had persistent urate lowering (responders) and also entered the OLE. As can be seen in Figure 2 and Supplementary Table 2, GMIM responder status was mainly limited to the subjects with persistent urate lowering, with 36.1% and 58.3% achieving GMIM20 responses by 3 and 6 months, respectively. Less than 10% of subjects who failed to maintain persistent urate lowering (nonresponders) attained GMIM20 responses. It should be noted that many of these nonresponders still had urate lowering through 3 months of therapy and a small number had urate lowering at 6 months even though they did not meet the strict criteria of being a responder.[34]

Figure 1.

Percentage of subjects treated with pegloticase q2w or placebo reaching GMIM criteria at 3 and 6 months in the RCTs. Number of evaluable subjects is shown for each group. The percentage of responders is shown above the bars. a Results for all GMIM criteria and b Results for all criteria except urate < 6.0 mg/dL. The statistical analysis of the data is shown in the supplementary material

Figure 2.

Percentage of responders and nonresponders to pegloticase or placebo-treated subjects meeting GMIM criteria at 3 and 6 months. Data show percentage of subjects meeting GMIM criteria including serum urate < 6 mg/dL. Number of evaluable subjects is shown for each group. The percentage of responders is shown above the bars. The statistical analysis of the data is shown in the supplementary material

To explore the clinical response in the nonresponders in greater detail, the criterion for achievement of sU < 6 mg/dL was omitted (Figure 3, Supplementary Table 3). As with the analysis shown in Figure 1, this resulted in some subjects in the placebo group achieving GMIM20, 50, and 70 responses at 3 months and GMIM20 and 50 responses at 6 months. There were also increases in GMIM20, 50, and 70 responses at 3 and 6 months for the q2w nonresponders. At 6 months, q2w responders achieved GMIM20 significantly more often vs placebo or nonresponders, whereas the nonresponders exhibited more frequent responses compared to placebo-treated subjects.

Figure 3.

Percentage of responders and nonresponders to pegloticase or placebo-treated subjects meeting GMIM criteria at 3 and 6 months. Data show percentage of subjects meeting GMIM criteria excluding serum urate < 6 mg/dL. Number of evaluable subjects is shown for each group. The percentage of responders is shown above the bars. The statistical analysis of the data is shown in the supplementary material

Sensitivity Analysis

A sensitivity analysis was carried out by recalculating GMIM20, 50, and 70 for each group with elimination (one at a time) of each of the clinical parameters. The requirement to achieve sU < 6 mg/dL was also not required for these analyses (Figure 4, Supplementary Table 4). As can be seen, most variables contributed to the model with the exception of flares, whose elimination had little impact on the percentage of subjects meeting GMIM criteria. To explore this further, we carried out regression analysis of all of the GMIM components. Multiple linear regression analysis indicated that the components that consistently contributed information to GMIM were sU, PGA, tender joints, pain, and swollen joints. Tophus area and flare contributed only inconsistently. Because only a subset of patients had measurable tophi, we evaluated the performance of GMIM in those subjects independently (Supplementary Figure 1). GMIM effectively separated responses in the tophaceous patients. Moreover, eliminating resolution of tophus area as a component of GMIM significantly decreased the measured responses, implying that tophus resolution significantly contributed to the GMIM model when only those subjects with tophi were considered.

Figure 4.

Sensitivity analysis of the GMIM model. Data show the percentage of responders and nonresponders to pegloticase or placebo-treated subjects meeting GMIM criteria at 3 and 6 months. Data show percentage of subjects meeting the GMIM criteria excluding serum urate < 6 mg/dL. In each figure, data are calculated omitting one criterion. Number of evaluable subjects is shown for each group. The percentage of responders is shown above the bars. Data are shown in supplementary table 4

Achievement and Maintenance of GMIM Responses

Subjects who had persistent lowering of urate during the RCT were followed into the OLE. As can be seen, GMIM responses were maintained for the 2 years of the OLE (Figure 5a, b). Moreover, patients who received placebo in the RCTs and were then switched to pegloticase q2w in the OLE achieved GMIM responses comparable to those treated with biweekly pegloticase in the RCTs (Figure 6a, b).

Figure 5.

Percentage of subjects with GMIM20, 50, and 70 responses during the OLE for patients treated with biweekly pegloticase in the RCT and OLE. a Results for all criteria including urate < 6.0 mg/dL and b Results for all GMIM criteria excluding urate < 6 mg/dL

Figure 6.

Percentage of subjects with GMIM20, 50, and 70 responses during the OLE for patients who were treated with placebo in the RCT and biweekly pegloticase in the OLE. a Results for all GMIM criteria including urate < 6 mg/dL and b Results for all criteria excluding urate < 6.0 mg/dL

processing....