Tophi and Frequent Gout Flares Are Associated With Impairments to Quality of Life, Productivity, and Increased Healthcare Resource Use

Results From a Cross-Sectional Survey

Puja P Khanna; George Nuki; Thomas Bardin; Anne-Kathrin Tausche; Anna Forsythe; Amir Goren; Jeffrey Vietri; Dinesh Khanna


Health and Quality of Life Outcomes. 2012;10(117) 

In This Article


Demographics & Patient Characteristics

The characteristics of the sample are presented in Table 1 and Table 2. The average age was 60.9 years, and patients had been diagnosed an average of 12 years. Most were male (81%), and 38% were employed. Approximately half were on allopurinol (51%). Thirteen percent had at least one tophus, and an additional 31% were unsure of whether they had any tophi. One quarter (25%) of the sample had been flare-free for the past 12 months, while 38% experienced at least 3 flares in the same time period.

The subsample of 563 respondents sourced from the NHWS was similar to the overall sample. Mean age was 61.2 years, 82% were male, and mean duration of gout was 12 years. The distribution of flares in the past 12 months was the same as in the full sample, and the proportion of patients reporting tophi was similar (12% reporting at least 1 tophus, 33% unsure, and 55% reporting none). Information about BMI and comorbidities were also available for this subgroup. Most were obese (53.8%) or overweight (33.9%), with 9.9% underweight or of normal weight, and 2.3% declining to answer questions about weight or height. In terms of comorbidities, 5% reported chronic kidney disease, 26% reported type 2 diabetes, and 63% reported hypertension.

sUA Levels

Nearly three quarters of the patients (72%) did not know their most recent sUA level; of those who knew their levels (n = 172), 62 (36%) reported sUA below 6 mg/dL/360/μmol/L. sUA level was positively correlated with flares experienced in the last 12 months, r s = 0.36, as well as the number of tophi currently present, r s = 0.37, both p < 0.001. Tophi were also associated with a higher number of flares in the last 12 months, r s = 0.29, p < 0.001. Being unsure of tophi was also associated with higher sUA, r s = 0.19, p < 0.05, as well as experiencing more frequent flares, r s = 0.20, p < 0.001.

The Burden of Tophi

Tophi were associated with considerable impairment of HRQOL. Patients with tophi had lower MCS, PCS, and SF-6D scores than patients without tophi, while the scores of those patients unsure of tophi fell in-between the two other groups. Post hoc tests revealed that both the tophi group (≥ 1 tophus) and the unsure group had significantly impaired HRQOL relative to the no tophi group for all three measures (all p < 0.01). The magnitude of these decrements for all three measures exceeded the commonly accepted MIDs.[29,30,32] Mean values and omnibus tests of statistical significance are presented in Table 3 .

Patients with tophi also had significantly greater impairment of work productivity and activity. Tophaceous gout was associated with greater absenteeism and overall work impairment than both the group without tophi and the unsure group (both p < 0.05). The presence of tophi was also associated with significantly greater impairment in daily activities as compared with those free of tophi and those unsure of their presence (both p < 0.01).

The Burden of Flares

Unadjusted comparisons revealed that HRQOL varied across the different flare groups, with more flares associated with lower HRQOL in a dose-dependent manner. SF-12 MCS, PCS, and SF-6D health utilities were all statistically significantly lower in patients with more frequent flares (all p < 0.001). Although flares were associated with an apparent numerical increase in work impairment, this difference did not reach statistical significance within the employed subsample. However, more frequent flares were associated with statistically significantly greater impairment of activity in the total sample, with those having the most frequent flares suffering nearly double the activity impairment of those without. Means and tests of significance are presented in Table 4.

Multivariable Models

Multivariable models, adjusting for covariates and simultaneous presence of flares and tophi, showed that tophi (vs. no tophi) were associated with statistically significantly lower MCS (p < 0.01) and PCS (p < 0.01) scores, as was being unsure of the presence of tophi (both p < 0.01). More frequent flares (≥ 4) were also associated with lower MCS (p < 0.05); those with 3 or more flares also had lower PCS scores (p < 0.05). The adjusted means are presented in Table 5. A similar pattern emerged in the analyses of health utilities, presented in Figure 1. As in the analysis of MCS and PCS scores, both confirmed tophi and being unsure of tophi were associated with significant impairment of HRQOL (both p < 0.01). More frequent flares (≥ 3) were associated with additional decrements in SF-6D utilities, after adjustment for age, gender, and length of diagnosis (p = 0.001).

Figure 1.

Mean SF-6D health utilities by flare frequency and the presence of tophi, adjusted for age, gender, and length of illness. Lower values indicate worse health; error bars represent standard error of the mean. Tophi (1+ or not sure) and flares (4+) are both associated with lower health utilities relative to no tophi and no flares/don't recall, respectively (all p < 0.05).

Patients with 1–2 flares in the past year showed significant activity impairment compared with those without, even after adjustment for covariates (p < 0.05), and impairment increased with frequency of flares. Those unsure of tophi were significantly more impaired in non-work activities than those without (p < 0.05), though lack of power prevented this effect from reaching conventional statistical significance among those with tophi (p = 0.07). The adjusted means are displayed in Figure 2.

Figure 2.

Mean overall activity impairment by flare frequency and presence of tophi, after adjusting for age, gender, and length of illness. Higher values indicate greater impairment; error bars represent standard error of the mean. Those with any flares are significantly more impaired than those without flares after adjusting for age, gender, and length of illness. Compared with those without tophi, those unsure of tophi were also more impaired (p < 0.05) and there was a trend for greater impairment among those with tophi (p = 0.07).

No associations between flares or tophi and measures of work productivity were observed after adjusting for covariates, but numbers were small as these analyses were of necessity limited to the employed 38% of the sample.

We also examined the data from the smaller sample of respondents re-contacted from NHWS, and adjusting for additional covariates (i.e., BMI, diagnosis with chronic kidney disease, type 2 diabetes, and hypertension), the presence of tophi, uncertainty about tophi, and frequent flares (≥ 4) were all significantly associated with decreased MCS, PCS, and health utilities, as well as increased activity impairment. Specifically, having more than 4 flares in the past year (vs. 0/unknown) was associated with lower MCS (p < 0.05), PCS (p < 0.05), SF-6D (p < 0.05), and activity impairment (p < 0.05) in the supplementary regressions. Likewise, those who were not sure of whether they had tophi exhibited lower MCS (p < 0.01), PCS (p < 0.001), SF-6D (p < 0.001), and activity impairment (p < 0.01) than those with no tophi. Those reporting at least one tophus (vs. no tophi) also demonstrated lower MCS (p < 0.05), PCS (p < 0.01), SF-6D (p < 0.01), and activity impairment (p < 0.05). Despite accounting for potential confounders and with reduced power, almost all results were replicated in terms of magnitude, direction, and significance. The only discrepancies in results were that 3 flares became non-significantly associated with a health utilities decrement (p = 0.31), and the associations between 1–3 flares and greater activity impairment (relative to no flares) fell below conventional levels of significance (1–2 flares: p = 0.051; 3 flares: p = 0.075).

Comparisons Across the Spectrum of Disease Phenotypes

Patients were categorized into three subgroups based on frequency of flares and the presence of tophi, without regard to treatment; 1) asymptomatic gout (no flares in the past year, no tophi), 2) severe tophaceous gout (defined as confirmed tophi and ≥ 3 flares in the past year) and 3) very severe tophaceous gout (defined as confirmed tophi and ≥ 6 flares in the past year). The groupings were similar to those employed in a previous study.[33] Both the severe group (n = 44) and very severe subgroup (n = 11) had significantly lower SF-6D health utilities than did the asymptomatic group (both p < 0.01).

The average SF-6D scores of the severe and very severe gout patients were also compared with the average SF-6D scores of US patients with other chronic rheumatic diseases. Average SF-6D scores for gout patients who did not report rheumatoid arthritis (RA), osteoarthritis (OA), or systemic lupus erythematosus (SLE), and average utilities of patients with RA, OA, and SLE patients without comorbid gout were calculated from the 2010 US NHWS. The means are presented in Figure 3. Independent-sample t-tests revealed that asymptomatic gout patients had higher health utilities than all other patient groups (all p < 0.05). In contrast to the asymptomatic patients, those in our sample with severe gout had health utilities similar to the average SLE or RA patient (both p > 0.05), and significantly worse than the average gout or osteoarthritis patient in the NHWS (p < 0.05). Patients with very severe gout had significantly lower health utilities than any comparison group (p < 0.01).

Figure 3.

Average health utilities of patients with gout and patients with other rheumatic diseases. Average health utilities are unadjusted values taken from 2010 US NHWS (red) and the current study (orange). Error bars represent standard error of the mean. Mean utility for gout with 6+ flares and tophi is significantly worse than all comparison conditions, p < 0.05. RA: rheumatoid arthritis; OA: osteoarthritis; SLE: systemic lupus erythematosus.