Determinants of Achieving Serum Urate Goal with Treat-to-Target Urate-Lowering Therapy in Gout

Abstract

Objective:

Using trial data comparing treat-to-target allopurinol and febuxostat in gout, we examined participant characteristics associated with serum urate (SU) goal achievement.

Methods:

Participants with gout and SU ≥6.8 mg/dl were randomized to allopurinol or febuxostat, titrated during weeks 0–24 and maintained weeks 25–48. Participants were considered to achieve SU goal if the mean SU from weeks 36, 42 and 48 was <6.0mg/dl or <5 mg/dl if tophi present. Possible determinants of treatment response were preselected and included sociodemographics, comorbidities, diuretic use, health-related quality of life (HRQoL), body mass index, and gout measures. Determinants of SU response were assessed using multivariable logistic regression with additional analyses to account for treatment adherence.

Results:

Of 764 study participants completing week 48, 618 (81%) achieved SU goal. After multivariable adjustment, factors associated with a greater likelihood of SU goal achievement included older age (aOR 1.40/10y), higher education (aOR 2.02) and better HRQoL (aOR 1.17/0.1 u). Factors associated with a lower odds of SU goal achievement included non-White race (aORs 0.32–0.47), higher baseline SU (aOR 0.83/1 mg/dl), presence of tophi (aOR 0.29), and the use of diuretics (aOR 0.52). Comorbidities including chronic kidney disease, hypertension, diabetes and cardiovascular disease were not associated with SU goal achievement. Results were not meaningfully changed in analyses accounting for adherence.

Conclusions:

Several patient-level factors were predictive of SU goal achievement among gout patients administered treat-to-target ULT. Approaches that accurately predict individual responses to treat-to-target ULT hold promise in facilitating personalized management and improving outcomes in patients with gout.

Recognizing hyperuricemia as a causal risk factor in the development and progression of gout, urate-lowering therapy (ULT) represents a cornerstone of disease management. Optimal ULT administration has been proposed by some to represent a potential ‘cure’ in gout [1]. Subspecialty rheumatology societies, including both the American College of Rheumatology (ACR) and European Alliance of Associations for Rheumatology (EULAR), have endorsed a treat-to-target approach as representing a best practice in ULT administration as a means of optimizing patient outcomes [2, 3].

In 2018, Doherty and colleagues reported results from a study comparing nurse-led management incorporating treat-to-target ULT to usual care in a sample of more than 500 patients with gout in the United Kingdom (U.K.) [4]. After two years, nurse-led treat-to-target ULT demonstrated clear superiority to usual care. Those receiving guideline concurrent treat-to-target care were more than three-times as likely to achieve serum urate (SU) goals of <360 umol/L (<6.0 mg/dl) and importantly were 67% less likely to experience two or more gout flares during the second year of follow-up. These results parallel those from earlier reports examining pegloticase, demonstrating that achievement of SU goals using this potent form of ULT administered in a treat-to-target-like framework led to significant improvements in physical function, health-related quality of life (HRQoL), and patient-reported pain at time points as early as 6 months [5]. Collectively, these data led to the strong endorsement of treat-to-target urate-lowering, one that encompasses the initiation of low-dose ULT followed by gradual dose titration guided by serial laboratory assessments to achieve and maintain SU concentrations below 6 mg/dl [2].

To date, several observational studies have examined potential determinants of SU goal achievement using real-world data. In patients with gout initiating ULT as part of usual care, factors associated with SU goal achievement have included age, sex, race, comorbidity, and concomitant medication use, among others [6, 7]. Importantly, these studies were completed in the context of usual care predominantly in primary care settings, where ULT administration is typically characterized by a ‘fixed dose’ approach and suboptimal outcomes [6, 8, 9]. Preliminary efforts to identify predictors of efficacy with treat-to-target ULT have been limited in scope and sample size [10, 11]. As treat-to-target ULT represents an optimal management strategy in gout, further efforts to identify factors predictive of efficacy could help to identify patient subgroups less likely to experience optimal benefit from this approach or for whom alternative or adjunctive treatment strategies (e.g., combination ULT or more intensive lifestyle modification) might be necessary early in the course of treatment.

In this post-hoc analysis of a large randomized, placebo-controlled study that compared allopurinol to febuxostat in gout management with both agents administered following a standardized treat-to-target protocol [12], we examined patient characteristics at enrollment as predictors of achieving recommended SU thresholds at 48 weeks. Based on results of previous studies detailed above, we tested the hypotheses that select demographics (i.e., younger age, non-White race) and measures of gout severity (i.e., higher SU, longer disease duration and tophi) at enrollment would predict reduced ULT efficacy in the context of treat-to-target ULT.

Methods

Parent Trial Design and Protocol.

This study is a post-hoc analysis of the STOP Gout trial (NCT02579096), a multicenter, randomized, double-blind, non-inferiority comparative effectiveness trial of allopurinol and febuxostat administered following a treat-to-target strategy in 940 patients with gout [12, 13].

Eligible participants in the trial fulfilled the 2015 ACR gout classification criteria [14] and were required to have a SU at the time of screening ≥6.8 mg/dL. Participants with a previous history of allopurinol in daily doses ≤300 mg were allowed to participate. The study protocol specified that at least one-third of participants would have stage 3 chronic kidney disease (CKD; defined by estimated glomerular filtration rate [eGFR] of <60 and ≥30 mL/min/1.73 m²). Patients with stage 4 or 5 CKD were excluded. At enrollment, occurring between 2017 and 2019, participants were randomized 1:1 to receive allopurinol or febuxostat. The last study visit occurred in February of 2021.

The study was divided into three phases with a total follow-up period of 72 weeks. During Phase 1 (weeks 0–24) ULT was titrated to reach goal SU, defined as <6 mg/dl or <5 mg/dl for those with tophi. This threshold was based on 2012 ACR recommendations available at the time of trial design [15]. During Phase 2 (weeks 25–48), ULT dosing was maintained with an allowance for continued titration early in this phase if participants had not yet achieved SU goal. During Phase 3 (weeks 49–72), participants were observed on stable doses of ULT as no further dose escalations were allowed during this phase. The maximum allowable daily dose was 800 mg for allopurinol and 120 mg for febuxostat; the latter decreased to 80 mg at the request of the Food and Drug Administration following issuance of a black-box warning related to cardiovascular risks in 2019 [16]. All participants received anti-inflammatory prophylaxis with colchicine, nonsteroidal anti-inflammatory drugs (NSAIDs), or glucocorticoids at the discretion of site investigators per 2012 ACR guidelines [15]. The occurrence of ≥1 gout flare between weeks 49 and 72 served as the primary study outcome of the parent trial. Flare data was collected using regular structured interviews and defined to have occurred if at least 3 of 4 participant-reported criteria were satisfied including warm joint(s), swollen joint(s), pain rating of >3 (scale of 0 to10 with higher numbers equating to greater pain), or self-identified flare [17]. The achievement of SU goal during Phase 2 of the study was a secondary outcome of the parent trial and defined by a mean SU value of <6.0mg/dL (or < 5 mg/dl if tophi) at weeks 36, 42, and 48 [12, 13].

ULT adherence was quantified using a participant-maintained diary wherein participants were instructed to record days that study medications were taken or missed. To reduce potential misclassification of adherence during periods of ULT dose escalation, the adherence assessment was limited to Phase 2, during which only limited escalation occurred. Patients were considered adherent if they reported taking assigned ULT doses on ≥80% of days.

Funding and Ethical Considerations.

The STOP Gout study protocol was funded by the Cooperative Studies Program of the Department of Veterans Affairs Office of Research and Development and approved by the Veterans Affairs Central Institutional Review Board (IRB). All participants provided written informed consent prior to enrollment. This post-hoc analysis was approved by the STOP Gout Steering Committee.

Post-hoc analyses.

The primary outcome in this post-hoc analysis was SU goal achievement during Phase 2 as described above. Analyses were restricted to 764 participants with available SU outcome data (n=176 excluded). Standardized differences were calculated comparing those included and excluded for each characteristic examined. Enrollment characteristics and associations of SU goal achievement with flare occurring during Phase 3 were summarized and compared using descriptive statistics. Potential determinants of achieving SU goal were selected a priori and included age, sex, race, education level, self-reported comorbidities (CKD, hypertension, diabetes, cardiovascular disease), body mass index (BMI; kg/m²) categories, HRQoL (using EuroQol 5 Dimension-3 Level score; EQ-5D-3L), SU concentration, gout duration, presence of tophi, prior allopurinol use, and current diuretic use. ULT type was not included as a covariate as we have previously demonstrated near identical SU goal achievement for those receiving allopurinol and febuxostat [12]. Though race is a social construct, it was examined in this analysis based on prior reports showing that it may influence gout outcomes including responses to treatment [18, 19], likely reflecting social determinants of health and other healthcare related factors. Race was collected by self-report according to the following categories: American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or other Pacific Islander, White/Caucasian or none of the above. The ‘other’ category used in this analysis combined Asian, American Indian or Alaska Native, Native Hawaiian or other Pacific Islander and none of the above.

Associations of patient factors with the achievement of SU goal were examined in unadjusted and adjusted logistic regression models. Recognizing that explanatory variables can simultaneously exert a causal influence and not achieve a statistical threshold of significance [20], all covariates identified to be potentially informative a priori were retained in multivariable models without the use of stepwise selection. C-statistics were calculated for multivariable models to determine discriminative ability. A secondary analysis was completed examining models that also included ULT adherence during the trial as a covariate. Given missing pill count data for a proportion of participants (n=132; 17%), we performed an additional sensitivity analysis that imputed non-adherence for those with missing pill count data.

Results

Participant Selection and Characteristics.

A total of 176 study participants were excluded from analyses because they terminated trial participation during Phases 1 or 2. The most common reasons for early study termination during Phases 1 and 2 were participant or provider decision (n=96), lost to follow-up (n=26), and the occurrence of adverse effects (n=20). Six participants died during Phase 2 with no deaths occurring in Phase 1. Of the 764 study participants included in the analysis, most were male (98%; reflecting the national VA gout population [21]), self-reported White race (70%), and the mean age was 62 years (Table 1). Consistent with the study design, slightly more than one-third (38%) had stage 3 CKD. Additional comorbidities were highly prevalent, including hypertension (77%), diabetes (34%), CVD (27%), and obesity defined by a BMI ≥ 30 kg/m² (68%). At enrollment, participants had mean SU of 8.6 mg/dl, a mean duration of gout approaching 10 years, and 16% had tophi. Characteristics differing between those included (n=764) and excluded (n=176), defined by a standardized difference >0.1, included race, education, BMI, EQ-5D-3L score, and diuretic use.

Table 1.

STOP Gout Trial participant characteristics among those included and not included in post-hoc analysis

Characteristic	Participants Included in Post-hoc Analysis	Participants Not Included in Post-hoc Analysis
	N=764	N=176
DEMOGRAPHICS
Age, years, mean (SD)	62.2 (12.3)	61.6 (13.1)
Male, %	98.3	98.9
Race, %
Black/African American	19.8	31.3
Other	10.6	9.1
White	69.6	59.7
Education, %
High school or less	25.8	30.7
Some college or associate degree	45.7	48.9
Bachelor’s degree or higher	26.4	19.3
Other or not stated	2.1	1.1
COMORBIDITY & HEALTH FACTORS
Chronic kidney disease, Stage 3, %	37.6	36.4
Hypertension, %	76.6	75.6
Diabetes, %	33.6	31.8
Cardiovascular disease, %	26.6	27.8
Body mass index, kg/m2, %
< 25 (healthy)	5.4	4.5
25 ≤ BMI < 30 (overweight)	27.5	27.8
30 ≤ BMI < 35 (obese)	30.1	38.6
≥ 35 (morbidly obese)	36.8	29.0
EQ-5D-3L index, mean (SD)	0.7 (0.2)	0.7 (0.2)
GOUT RELATED FACTORS
SU, mg/dl, mean (SD)	8.6 (1.4)	8.5 (1.2)
Duration of gout, years, mean (SD)	9.8 (11.0)	10.8 (11.1)
Presence of tophi, %	15.8	17.6
Prior allopurinol use, %	37.6	33.0
Diuretic use, %	38.5	33.0
ULT Treatment Assignment, %
Allopurinol	49.7	50.0
Febuxostat	50.3	50.0

Abbreviations: SD, standard deviation; EQ-5D-3L, EuroQol 5 Dimension – 3 Level; SU, serum urate; ULT, urate-lowering therapy; data missing for BMI in 2 participants and EQ-5D-3L for 1 participant; other race category combines Asian, American Indian or Alaska Native, Native Hawaiian or or other Pacific Islander, and ‘none of the above’; standardized difference (those included vs. those not included in post-hoc analysis) >0.1 for race, education, BMI, and diuretic use

Determinants of SU Goal Achievement.

Of the 764 participants included, 618 (81%) achieved the target SU goal during Phase 2. Participants achieving SU goal were more likely than participants not achieving this goal to remain flare free during Phase 3 (62% vs. 49%; p=0.004). The proportion achieving SU goal was similar between those assigned allopurinol (82.8%) vs. febuxostat (78.9%) (p=0.16; data not shown). Participant characteristics at enrollment among those achieving and not achieving SU goal during Phase 2 are shown in Table 2. Compared to those not achieving target goal, participants with gout achieving SU goal with treat-to-target ULT were older (mean age 63 vs. 59 years), more often self-reported white race (74% vs. 52%), had higher EQ-5D-3L scores suggesting better HRQoL (0.7 vs. 0.6 in non-responder group), and had disease characteristics suggesting lower gout severity. Specifically, compared to non-responders, those achieving threshold SU goals had lower enrollment SU concentrations (mean 8.4 mg/dl vs. 9.1 mg/dl), a lower frequency of tophi (13% vs. 28%) and a trend towards shorter gout disease duration (mean 9.4 vs. 11.3 years) at enrollment (Table 2).

Table 2:

Enrollment characteristics of STOP Gout trial participants by serum urate (SU) outcome achieved during Phase 2 (n=764)

Characteristic	SU Goal Achieved at 48 Weeks	SU Goal Not Achieved at 48 Weeks	P-value
	N=618	N=146
DEMOGRAPHICS
Age, year, mean (SD)	63.1 (12.1)	58.5 (12.5)	<0.001
Male, %	97.9	100.0	0.14
Race, %			<0.001
Black/African American	17.2	30.8
Other	9.1	17.1
White/Caucasian	73.8	52.1
Education, %			0.17
High School or less	25.4	27.4
Some College or Associates Degree	44.5	50.7
Bachelor’s Degree or higher	28.2	19.2
Other/Not Stated	1.9	2.7
CORMOBIDITY & HEALTH FACTORS
Chronic kidney disease – Stage III, %	37.9	36.3	0.73
Hypertension, %	77.3	73.3	0.30
Diabetes, %	33.5	34.2	0.86
Cardiovascular disease, %	25.4	31.5	0.13
Body mass index, kg/m2, %			0.48
< 25 (healthy)	5.5	4.8
25 ≤ BMI < 30 (overweight)	28.5	23.3
30 ≤ BMI < 35 (obese)	30.1	30.1
≥ 35 (morbidly obese)	35.6	41.8
EQ-5D-3L index, mean (SD)	0.7 (0.2)	0.6 (0.2)	<0.001
GOUT RELATED FACTORS
SU mg/dL, mean (SD)	8.4 (1.3)	9.1 (1.7)	<0.001
Duration of gout, years, mean (SD)	9.4 (10.7)	11.3 (12.2)	0.07
Presence of tophi, %	12.9	28.1	<0.001
Prior Allopurinol use, %	37.2	39.0	0.68
Diuretic use, %	37.2	43.8	0.14

Abbreviations: SD, standard deviation; BMI, body mass index; EQ-5D-3L, EuroQol 5 Dimension-3 Level; SU, serum urate; SU goal < 6 mg/dl or < 5 mg/dl if tophi

In multivariable models, patient characteristics at enrollment associated with higher odds of SU goal achievement included older age (aOR 1.04; 95% CI 1.02, 1.07 per year), higher levels of education (aOR 2.02; 95% CI 1.10, 3.69; vs. high school education or less), and better HRQoL (aOR 1.17; 95% CI 1.07, 1.29 per 0.1 unit EQ-5D-3L). Characteristics associated with a lower odds of SU achievement included higher SU concentration at enrollment (aOR 0.83; 95% CI 0.72, 0.96 per 1 mg/dl), the presence of tophi (aOR 0.29; 95% CI 0.17, 0.49), and concomitant diuretic use (aOR 0.52; 95% CI 0.32, 0.87) (Figure 1). The multivariable model demonstrated good discriminative ability (C-statistic 0.76).

Figure 1: Associations of participant characteristics at enrollment and serum urate response vs. non-response at 48-weeks in patients with gout receiving treat-to-target urate-lowering therapy.

Treatment response defined as achievement of SU < 6 mg/dl (or < 5 mg/dl if tophi) based on mean concentrations at weeks 36, 42, and 48. Abbreviations: BMI, body mass index; EQ-5D-3L, EuroQol 5 Dimension-3 Level; SU, serum urate; C-statistic of multivariable model = 0.76; excludes 2 with missing BMI and 1 with missing EQ-5D-3L score.

On-trial Adherence to ULT.

Phase 2 ULT adherence data was available for 632 (83%) of the 764 participants. Of the participants with adherence data, the mean proportion of days participants reported taking the allotted study medication was 96.1% (SD 12.1%); 532 (84.2%) reported taking assigned study doses ≥80% of days.

To examine the possible impact of ULT adherence on the aforementioned findings, we evaluated predictors of SU goal achievement with adjustment for adherence in a multivariable logistic regression model limited to the 632 participants with Phase 2 pill count data (Table 3, Model A). In this model, participants adherent to ULT were over 2-fold more likely to reach target SU (OR 2.29; 95% CI 1.05, 4.98). In a sensitivity analysis imputing non-adherence for those with missing pill count data, ULT adherence was associated with an approximate 3-fold higher odds of achieving SU goal (OR of 2.81; 95% CI 1.82, 4.35) (Table 3, Model B). Inclusion of adherence did not meaningfully change the associations between patient characteristics and achieving goal SU noted in the primary analyses. The C-statistic for these models was 0.79.

Table 3:

Multivariable models examining participant characteristics associated with serum urate goal achievement including treatment adherence as a covariate

	A. Participants with Missing Adherence Records Excluded (N=632)		B. Non-adherence Imputed for Missing Adherence (N=764)
Characteristic	Odds Ratio (95% CI)	P-value	Odds Ratio (95% CI)	P-value
DEMOGRAPHICS
Age, per year	1.05 (1.03, 1.08)	<0.001	1.04 (1.02, 1.07)	<0.001
Male (vs. Female)	0.00 (0.00, I)	0.98	0.00 (0.00, I)	0.98
Race		0.002		<0.001
White/Caucasian	Referent		Referent
Black/African American	0.34 (0.19, 0.62)		0.34 (0.20, 0.56)
Other	0.74 (0.32, 1.70)		0.50 (0.27, 0.95)
Education		0.15		0.11
High School or less	Referent		Referent
Some College or Associates Degree	1.20 (0.66, 2.17)		1.13 (0.68, 1.89)
Bachelor’s Degree or higher	2.23 (1.08, 4.61)		2.00 (1.09, 3.68)
Other/Not Stated	1.32 (0.23, 7.45)		1.31 (0.30, 5.66)
CORMOBIDITY & HEALTH FACTORS
Chronic kidney disease – Stage III	1.13 (0.65, 1.99)	0.66	1.14 (0.70, 1.86)	0.59
Hypertension	1.59 (0.83, 3.06)	0.16	1.33 (0.78, 2.25)	0.30
Diabetes	1.24 (0.69, 2.23)	0.46	1.29 (0.78, 2.13)	0.33
Cardiovascular disease	0.52 (0.28, 0.95)	0.03	0.61 (0.36, 1.02)	0.06
Body mass index, kg/m2		0.33		0.77
< 25 (healthy)	Referent		Referent
25 ≤ BMI < 30 (overweight)	0.63 (0.17, 2.39)		0.83 (0.30, 2.32)
30 ≤ BMI < 35 (obese)	0.39 (0.11, 1.39)		0.66 (0.24, 1.80)
≥ 35 (morbidly obese)	0.45 (0.13, 1.58)		0.78 (0.29, 2.13)
EQ-5D-3L index, per 0.1 unit	1.21 (1.08, 1.35)	0.001	1.19 (1.08, 1.31)	<0.001
GOUT RELATED FACTORS
SU, per 1 mg/dl	0.84 (0.70, 1.00)	0.05	0.84 (0.73, 0.98)	0.03
Duration of gout, per 1 year	0.99 (0.96, 1.01)	0.22	0.98 (0.96, 1.00)	0.06
Presence of tophi	0.21 (0.11, 0.39)	<0.001	0.29 (0.17, 0.49)	<0.001
Prior allopurinol use	0.99 (0.59, 1.68)	0.98	1.06 (0.68, 1.66)	0.80
Diuretic use	0.56 (0.31, 1.04)	0.07	0.50 (0.30, 0.84)	0.009
Adherence (≥80% days taken)	2.29 (1.05, 4.98)	0.04	2.81 (1.82, 4.35)	<0.001

Abbreviations: I, infinity; BMI, body mass index; EQ-5D-3L, EuroQol 5 Dimension-3 Level; SU, serum urate; Model A excludes participants with missing pill count data in Phase 2; Model B includes all 764 participants from post-hoc analysis, imputing non-adherent status for all patients with Phase 2 missing pill count data; 2 records missing BMI covariate; SU goal < 6 mg/dl or < 5 mg/dl if tophi

C-statistics 0.79 (both models)

Discussion

In the STOP Gout Study, 81% of participants with gout completing 48 weeks of follow-up achieved SU goal when treated via a treat-to-target ULT strategy, demonstrating non-inferiority of allopurinol to febuxostat [12]. In addition to reconfirming the importance of SU goal achievement in the reduction of flare risk, we have further extended findings from this randomized controlled trial by identifying participant characteristics associated with the achievement of SU goal at 48 weeks in the context of highly protocolized treat-to-target ULT, a strategy that eliminates the effects of provider prescribing patterns that are likely to confound findings from observational, real-world experience [6, 7]. Factors associated with a lower likelihood of achieving this recommended SU threshold using treat-to-target care included younger age, non-White race, worse HRQoL, higher enrollment SU values, the presence of tophi, concomitant use of diuretics, and reduced ULT adherence. Of added importance, we found that comorbidities common in gout such as CKD, diabetes, hypertension, obesity, and cardiovascular disease were not predictors of SU response. Results from our analysis contrast with findings from observational studies suggesting that these comorbid conditions, including CKD, may be associated with lower rates of SU goal achievement [6, 7]. Taken together, these data and those from the STOP Gout trial suggest that findings from previous real-world experience may relate more to provider practice patterns characterized by less intensive urate lowering and that the adoption of a protocolized treat-to-target approach may be an efficacious and well tolerated strategy to optimize outcomes in patients with gout and comorbidity.

At least two previous studies have been undertaken to identify factors predictive of treatment response in the context of treat-to-target ULT [10, 11]. The first by Stamp and colleagues examined data from 150 participants with gout undergoing protocolized allopurinol dose escalation with outcomes assessed after one year [11]. Participants were categorized as achieving inadequate response (no observations with goal SU achieved) versus partial or complete response, with the latter group including patients at SU goal after both 9 and 12 months of observation. Similar to findings of our study, factors associated with inadequate response (vs. partial or complete response) included younger age and higher SU concentration at enrollment. In additional analyses, these investigators found that compared to individuals reporting other ethnicities (primarily Māori or Pacific Island status), participants of European ancestry were significantly more likely to achieve a complete response. This latter result parallels findings of our study showing that patients with gout reporting non-White race (primarily Black or African American race) were 53% to 68% less likely to achieve SU goal with treat-to-target ULT after accounting for confounding factors. Whether there are other healthcare system barriers, systemic bias, or other social determinants of health mediating the relationship between race and treatment outcome are unknown, important questions that warrant further study.

In a separate single-center cohort study, Uhlig and colleagues reported the results of treat-to-target ULT administration (primarily allopurinol) in 186 gout patients with 85.5% achieving SU goal at 12 months [10]. Although not included in their predictive model due to limited sample sizes, only 55% of patients with tophaceous disease (comprising 17% of study population) achieved target SU concentrations at 12 months, consistent with findings from our study showing 71% lower odds of achieving this goal in this patient subgroup. Other significant predictors of inadequate treatment response in multivariable models from the study by Uhlig et al. included regular alcohol use, reduced self-efficacy, and patient beliefs that medications are often “overused”, factors that were not available in our analyses.

Uhlig et al. [10] speculated that their identified associations were likely attributable to lower ULT adherence. Treatment adherence was not included in their study but shown in our analysis to strongly predict treatment response, with adherent patients more than twice as likely to achieve target SU goal at 48 weeks. We measured adherence based on participant diaries, an approach that could be prone to misclassification. Direct measurement of medication or metabolites (i.e., febuxostat or oxypurinol in allopurinol users) was not done for this study, but would be relevant as such measures could be used to more precisely define adherence [22]. As anticipated in a clinical trial, reported adherence was high in the STOP Gout Study with rates approaching those defined using oxypurinol measurement among participants undergoing allopurinol escalation in a previous clinical trial [22].

Collectively, factors comprising our predictive models demonstrated good discriminative capacity with C-statistics of 0.76 (models excluding adherence) to 0.79 (models including adherence). Accurate prediction could facilitate a more personalized approach in gout management including the addition of adjunctive therapies or tailoring medications used to treat comorbid disease that might impact SU concentration. This could include steps such as the addition of uricosuric therapy [23] or, when feasible, stopping diuretic therapy in favor of alternate treatment options as diuretic treatment was shown in our study to be associated with an approximate 50% lower likelihood of achieving SU goal, likely related to the tendency of many diuretics to decrease renal urate excretion [24]. Our results suggest that even modest gains in urate-lowering effect (additional lowering by 1 mg/dl) would capture a majority of patients categorized as non-responders, as the final median SU level in this group was 6.3 mg/dl with a mean of 6.8 mg/dl.

Reasons for the association of younger age with reduced treatment response observed in our study and in the study from Stamp et al [11] are unknown. The consistency of findings with respect to age and SU goal achievement across models with and without treatment adherence as a covariate would suggest that age-related differences in ULT compliance are unlikely to explain this association. Similar associations of older age with improved SU goal achievement with ULT was recently reported using registry data from a large cohort of patients with gout derived from rheumatology practices across the U.S. [25]. It is possible that increased drug metabolism and/or elimination in younger individuals or age-related changes in xanthine oxidase expression and susceptibility to drug effects could drive this association. A novel finding from this analysis is the association of reduced HRQoL in this study with a lower treatment response. While the reasons for this association are not known, it suggests that such measures should be more widely integrated as part of holistic assessment of patients with gout initiating ULT.

Beyond SU measurement, it is possible that other biomarkers could yield needed improvements in efforts to identify patients at risk for suboptimal treatment response. The ABCG2 (Q141K) loss-of-function polymorphism, for example, is associated with reduced allopurinol response with patients carrying this mutation being more than two-fold less likely to achieve target SU thresholds [26, 27]. Although this variant is more common in select at-risk populations, it is found in less than 10% of Black or African American individuals [28] suggesting that this alone is unlikely to explain the lower rate of treatment success observed in this population. Whether extended genotyping of ABCG2 gene and whole genome, or the addition of other circulating biomarkers, might prove informative in this population will require further study.

There are additional limitations to this investigation. As the STOP Gout Study was completed predominantly in the U.S. Veterans Health Administration (VHA), findings from our analysis may not be generalizable to other populations. Given the demographics of the U.S. Veteran population and in those with gout, study participants were overwhelmingly male (>98%), prohibiting any meaningful examination of sex-based differences in response. This is highly relevant given the increasing incidence of gout reported in women both in the VHA and elsewhere [29, 30] in addition to reports of real-world experience suggesting that men may be substantially less likely than women to achieve target SU levels with ULT [6]. Beyond education status, other socioeconomic measures such as social deprivation and household income were not available for this analysis. As with other recent efforts to identify patient factors associated with response to treat-to-target ULT [31], our study focused on study completers. Thus, these analyses should be interpreted in light of this caveat.

In summary, we identified several patient-level factors associated with SU response among gout patients administered ULT using a treat-to-target strategy. Predictive factors were derived from different domains and included sociodemographic measures, HRQoL, concomitant diuretic use, and measures of gout severity. With continued refinement and external validation, approaches that accurately predict individuals at risk for inadequate response to treat-to-target ULT holds the promise of facilitating personalized management and improving outcomes in patients with gout.

Key Messages:

• Treat-to-target urate-lowering therapy (ULT) is supported as a best practice in gout management.

• In this post-hoc analysis of data from a randomized trial of treat-to-target ULT, we identified several factors predictive of SU goal achievement among gout patients administered treat-to-target ULT.

• Baseline factors predictive of SU goal achievement included sociodemographic factors, health-related quality of life, diuretic use, and measures of gout severity.