|Home | About | Journals | Submit | Contact Us | Français|
Objective. Short Form-36 (SF-36) is a validated outcome measure to assess health-related quality of life (HRQOL) in patients with gout. We assessed responsiveness to change of SF-36 in patients with gout.
Methods. SF-36 was administered at baseline and at yearly intervals. We assessed the minimal clinically important differences (MCIDs) at the first and second year. We also assessed the responsiveness to change (effect size) and interpreted it based on Cohen’s criteria. We modelled the improvement (defined as ≥MCID) in SF-36 scales and summary scores. Covariates included age, presence of tophi, comorbidities, baseline joint involvement, baseline serum urate, change in serum urate and the number of flares from baseline to 12 months.
Results. Of 99 subjects, 96 were male, mean age was 57.1 years, disease duration was 8.2 years and 40.4% had tophi. Ninety-two patients were treated with urate-lowering therapy (ULT) and daily colchicine, and seven were only on colchicine. Baseline mean serum urate level was 8.9mg/dl and mean number of flares was 4.7 over last year. ULTs were associated with reduction in serum uric acid and number of flares (P<0.001 for both) over 12 months. Therapy was associated with 22–70% of the patients achieving MCID in SF-36 scores at 12 months. Effect size estimates ranged from negligible to large (SF-36 mental component summary 0.08–bodily pain 1.09). Reduction in flares independently predicted improvements in three SF-36 physical scales (P=0.001–0.06). Improvement in SF-36 scores was maintained at 2 years.
Conclusion. In our real-life observational cohort, chronic urate lowering therapy and colchicine was associated with clinically meaningful improvements in HRQOL at 1 year and then maintained at 2 years. SF-36, especially physical domains and physical component summary, are responsive to change in gout.
Gout is a debilitating disease characterized by recurrent attacks of acute inflammatory arthritis and, if untreated, can result in chronic progressive disease including tophaceous gout . The biochemical basis of gout is oversaturation of urate levels in the extracellular fluids that result in precipitation of urate crystals in extra-articular tissues. Although hyperuricaemia has not been shown to always result in gouty attacks, elevation of serum urate levels >8.0mg/dl have been associated with more frequent attacks of gout . The prevalence of gout is increasing (e.g. from 0.5% of patients with self-reported gout cases in 1969 to 1.0% in 1996) .
Recent data suggest that gout is associated with detrimental effects on health-related quality of life (HRQOL) [4–6]. Patients with gout have lower scores on the medical outcomes Short Form-36 (SF-36) in physical and social scales when compared with the US general population [4–7]. However, since patients with gout often have other associated comorbidities (such as hypertension, diabetes mellitus and coronary artery disease), it is conceivable that the detriment in HRQOL is not a result of gout per se, but may be due to associated comorbidities. Different cross-sectional observational studies have demonstrated detrimental impact of gout on HRQOL after adjustment for multiple comorbidities [6–8]. However, no study has assessed the impact of gout on HRQOL in a prospective observational cohort.
A recent advancement in HRQOL is the estimation of minimal clinically important differences (MCIDs)—the smallest improvement in score of an HRQOL instrument that patients perceive as beneficial and that may lead to a change in the patient’s management . MCID can provide a benchmark for future design of gout clinical trials by helping researchers and clinicians to understand whether HRQOL score differences between two treatment groups are meaningful, or if changes within one group over time are clinically meaningful.
Therefore, the objectives of our study were: (i) to assess HRQOL using SF-36 in an observational cohort; and (ii) to assess if long-term chronic therapy with urate-lowering therapy (ULT) and colchicine leads to an improvement in HRQOL. Our hypothesis was that SF-36 is responsive to change in chronic gout assessed in a real-life observational cohort.
Participants were patients with crystal-proven gout who were seen at their initial visit in a gout clinic. Ethical approval was received from the Institutional Review Board (IRB; Comité de Ética e investigación Clínica del Hospital de Cruces) to get general data and HRQOL. Written consent was not needed for such purpose, as stated by the IRB following local regulations. The study was done according to the Declaration of Helsinki.
We collected the following data at each visit: self-reported number of flares in the year previous to the first visit; number of self-reported joints previously involved by gout flares (1, 1–4, >4); previous ULTs and doses; and radiographs of previously involved joints (normal, erosions only or erosions and joint space involvement characteristic of gout). We assessed comorbidities by conducting a detailed patient chart review and these are as listed in Table 1. Patients who had suffered a gout flare in the past 4 weeks before the evaluation or those returning for follow-up visit were not included in the study. All patients whose serum urate was >6mg/dl were offered ULT in increasing doses, to achieve the target serum urate <6mg/dl along with colchicine prophylaxis (0.5–1.0mg/day depending on renal function) for 1 year [10, 11]. Patients with serum urate of <6mg/dl were not included in this cohort. If a patient declined ULT, colchicine at prophylactic doses was maintained to obtain a non-ULT group (n = 7).
The SF-36 is a generic health status measure consisting of 36 items assessing eight scales [12, 13]. Each of the SF-36 subscales is scored from 0 to 100, with a higher score representing better health. The eight SF-36 scales can be summarized into a physical component summary (PCS) and a mental component summary (MCS) scores. The scales were normed to Spanish population and summary scores are normed to the US general population [since Spanish algorithms are not available (per e-mail communication with Dr Jordi Alonso)], where the mean (s.d.) score is 50 (10). Spanish norms are very similar to those obtained in the US population . We used version 2 of the SF-36 and a standard (4-week) recall period . The MCIDs for the SF-36 summary scores are between 2.5 and 5.0, and for individual scales are between 5.0 and 10.0 in different arthritides [16, 17]. Since SF-36 asks about health over the past 4 weeks, patients who had a gout flare within the past 1 month were excluded to avoid capturing impact of acute gout on HRQOL.
Baseline descriptive statistics for the SF-36 scale scores and percentage with floor and ceiling effects were calculated. Floor and ceiling effects are the percentages of respondents scoring at the lowest and highest possible scale level. These effects can influence responsiveness, as they may limit a change of score over time. Statistical significance was evaluated using paired t-test for changed scores of the SF-36 scales and summary scores. MCID for the SF-36 scales and summary scores was defined as improvement of ≥5 points and ≥2.5 points at 12 months, respectively [17, 18]. Responsiveness to change (effect size) was also assessed and interpreted based on Cohen’s criteria: 0.00–0.19 as negligible; 0.20–0.49 as small; 0.50–0.79 as moderate; and ≥0.80 as large .
We also generated logistic regression models to assess improvement (defined as ≥MCID) in eight SF-36 scales and two summary scales. Independent variables included age, tophi (presence/absence), comorbidities (presence/absence), baseline joint involvement, baseline serum uric acid levels, change in serum uric acid level and the number of flares from baseline to 12 months. All analyses were done using Stata 10.1 (College Station, TX, USA); P=0.05 was indicative of statistical significance.
Of 99 subjects with gout, 96 were male, mean (s.d.) age was 57.1 (11.0) years and mean (s.d.) disease duration was 8.2 (7.7) years. The patients had moderate-to-severe gout with 46 (46.9%) who were associated were medical comorbidities, 40 (40.4%) had tophi, 63 (63.6%) had mono- or oligo-articular disease and 63 (63.6%) had evidence of radiographic damage due to gout (Table 1). Fifty-three (54%) patients had at least one prescription of allopurinol before their baseline visit. Baseline mean (s.d.) serum uric acid level was 8.9 (1.36) mg/dl and number of mean (s.d.) flares were 4.7 (4.5) over the last year.
Patients had marked detriments in their SF-36 PCS scores that were 1.0 s.d. below US population (Table 2). Surprisingly, the MCS scores were 0.2 s.d. above US population. Follow-up data were available for 77 patients at 1 year and 36 patients at 2 years. Patients were treated with ULT including allopurinol (n=64), benzbromarone (n=4) or combination therapy (n=2); only seven patients were on oral colchicine (mild, mono-articular gout not willing to start ULT). As stated above, the goal of ULT was to reach a serum uric acid goal of <6.0mg/dl. Of the 77 patients at the end of 12 months, 54 (70%) showed a decrease in their serum uric acid levels to <6.0mg/dl; these included seven patients on oral colchicine. ULTs were associated with reduction in mean (s.d.) serum uric acid to 5.46 (1.1), [−3.44 (1.73); P<0.001] and the number of mean (s.d.) flares [−4.00 (4.32); P<0.001] over 12 months. Chronic therapy with ULT and colchicine or colchicine alone were associated with a statistically significant physical functioning, role physical, bodily pain, general health, vitality and social functioning scales and PCS summary scores at 12 months (Table 2). Large effect sizes were seen in SF-36 PCS (0.91) and bodily pain scale (1.09); moderate effect sizes were seen in role physical (0.50); and small effect sizes were seen in the remaining scales including physical function, general health, mental health, vitality, social functioning and role emotional scales (0.20–0.49); and negligible effect sizes were seen in SF-36 MCS (0.08). The proportion of patients achieving MCID also ranged from 38% for MCS to 79% for bodily pain scale. Improvement in SF-36 scores was maintained at 2 years (Table 2). Floor effect ranged from 0.0 to 6.1% for bodily pain and ceiling effect ranged from 1% (general health) to 58.6% (role emotional).
As stated before, seven patients declined to take ULT and were continued on daily colchicine (Table 3). We compared the colchicine-only group with the ULT group. In comparison with the ULT group, the use of colchicine was associated with similar effect sizes in the physical function and mental health scales and SF-36 MCS; smaller effect sizes were seen for role physical, bodily pain, vitality and social functioning scales and SF-36 PCS; and larger effect size was seen for role emotional scale compared with the ULT group.
Patients with the presence of tophi, comorbidities, polyarticular disease and radiographic damage had lower baseline SF-36 scores compared with patients without (Table 4). Physical functioning and role physical scales and SF-36 PCS were significantly lower in patients with the presence of tophi, comorbidities, polyarticular disease and radiographic damage.
In the multivariable models predicting improvement in the SF-36 scales and summary scores (defined as ≥MCID), reduction in flares independently predicted improvements in the three SF-36 scales (bodily pain, physical functioning and role physical) at 1 year after adjusting for the covariates (P=0.001–0.06). Other significant predictors included baseline serum uric acid for bodily pain and social functioning scales and SF-36 PCS scores (P=0.001–0.04), and age for general health scale and SF-36 PCS score (P=0.04 for both). Presence of tophi, comorbidities and joint involvement were not independent predictors of improvement in HRQOL.
Chronic gout is associated with a detrimental effect on HRQOL [4–7]. Our study suggests that chronic long-term therapy (majority on ULT) and daily colchicine started for chronic gout (to a predefined goal of serum uric acid level of <6mg/dl) is associated with reduction in gout flares and improvement in HRQOL. This effect was seen at 1 year and maintained at 2 years. We also show in a small number of patients (n=7) that daily colchicine therapy is associated with improvement in HRQOL.
SF-36 is a generic HRQOL (i.e. the concepts are not specific for any age, disease or treatment group) and is proposed as an outcome measure for clinical trials in gout . Since generic HRQOL measures are generally less sensitive to change in a specific disease, disease-specific HRQOL instruments have been proposed to capture smaller differences and smaller changes over time and have more face validity. Gout assessment questionnaire is an example of disease-specific measure for gout . Being a generic HRQOL, one of the concerns is that the SF-36 scales’ scores may be reflecting associated comorbidities (such as hypertension, diabetes and coronary heart disease) seen in chronic gout and not due to gout per se. This was evident in a single-centre study from the USA, where 80 patients with gout were interviewed and health utilities were assessed . Health utilities assess the value of desirability of the health state. Although patients had marked physical limitations and co-existing comorbidities, gout per se was associated with small disutility. Also, it is conceivable that the co-existing comorbidities may limit responsiveness to change of SF-36 as improvement in gout may not have a clinically meaningful improvement in generic HRQOL. However, our study convincingly shows that physical scales of SF-36 are sensitive to change at the end of the 12-month period in our observational study; scales of bodily pain, role physical and SF-36 PCS showed moderate-to-large effect sizes. In addition, small-to-moderate effect size improvements were seen in the remaining five scales. The scales relating to mental health and MCS had negligible-to-small effect sizes that are likely due to the ceiling effect. In other words, adjusted scores for mental scales and MCS were normal or near normal to the general population as compared with the physical health scales, which were significantly lower than those of the general population (Table 2). Our data are substantiated by two large randomized controlled trials using pegloticase (presented as abstracts), which have shown clinically meaningful and statistically significant improvement in SF-36 scales and summary scores. Based on these and other published data [4–7], SF-36 meets OMERACT filters of feasibility, reliability and validity and endorsed at an outcome measure .
As hypothesized, patients with the presence of tophi, comorbidities, polyarticular disease and radiographic damage had lower SF-36 scores and this provides construct validity of the SF-36 (Table 4). In a large study of US veterans, patients with comorbidities reported greater decrements in their SF-36 scores . Another large observational study from the USA assessed treatment-failure gout population and showed decrements in the SF-36 scales and SF-36 PCS scores, and comorbidities were associated with worse physical functioning .
In our multivariable models predicting improvement in SF-36 scales and summary scores (defined as ≥MCID scores), reduction in flares over 12 months was independently associated with improvements in the three SF-36 scales (bodily pain, physical functioning and role physical) at 1 year. Change in serum uric acid was not associated with an improvement in HRQOL. Since serum uric acid levels of <6mg/dl are associated with the decline in number of flares [10, 11], and greater frequency and intensity of flares are associated with poor HRQOL , we believe the goal of therapy should be to reduce uric acid levels <6mg/dl. Although presence of tophi, comorbidities and polyarticular joint involvement were associated with poor HRQOL, they did not predict an improvement in HRQOL in the multivariable models.
We compared our SF-36 data with other publications assessing SF-36 in chronic gout observational studies (Table 5). In general, patients with chronic gout had marked decrements in their SF-36 PCS scores (ranging from 32.2 to 40.3), but surprisingly near-normal SF-36 MCS scores (ranging from 46.6 to 52.2). This has also been seen in other chronic arthritides such as RA and scleroderma  and is likely related to psychological adjustment to their chronic disease. It is likely that patients with recurrent or recent attacks will lead to lower SF-36 MCS scores. This was shown in a US cohort study where recurrent gout attacks over the past year or having an attack over the past 3 months was associated with lower SF-36 MCS scores .
We also assessed ceiling and floor effects of the SF-36 scales. Ceiling effects are important as they can limit the responsiveness to change because subjects, who have achieved the maximum scores, cannot improve any further. Ours and another US study  showed low floor effects in the SF-36 scales but high ceiling effects, especially in the social functioning, role physical and role emotional scales. However, in the current study, effect sizes were similar between scales that had lower ceiling effects vs these three scales, suggesting that high ceiling effects did not have an impact on responsiveness to change.
Our study has several strengths. First, this is the first study that evaluates responsiveness to change of SF-36 in a real-life cohort. Secondly, this study comprehensively assesses SF-36 in a cohort outside USA. Although another study from Mexico assessed SF-36, they did not provide detailed information on SF-36 scores . Thirdly, our cohort comprehensively collected clinical, radiological and laboratory data and treated patients with ULT to a goal of <6.0mg/dl. Finally, the seven patients who were on prophylaxis with colchicine alone showed improvement in their HRQOL. These trends are interesting and thought provoking, but should be considered preliminary.
Our study is not without limitations. First, this is a single academic centre-based study, which limits the generalizability as most gout is treated in primary care. Also, our data need to be replicated in larger cohorts. Secondly, we did not evaluate other HRQOL measures such as HAQ-disability index that are responsive to change in gout [4, 24]. Thirdly, we did not include a placebo group; the improvement in HRQOL measures may need to be validated in a placebo-controlled study. Finally, our MCID cut-off for the SF-36 is based on the data from studies in RA and scleroderma [17, 25]. There are no MCID values currently available in gout.
To conclude, in this real-life observational cohort, long-term treatment with ULT or colchicine was associated with statistical and clinically meaningful improvements in the number of flares and HRQOL at 1 year, and was maintained at 2 years. SF-36, especially physical domains and PCS, are responsive to change with treatment in gout over time.
P.P.K. was supported by Ruth L. Kirschstein National Research Service Award (NRSA) Institutional Research Training Grant National Institute of Arthritis, Musculoskeletal, and Skin Diseases (NIAMS) 1 T32 AR053463 and ACR Research and Education Foundation Clinical Investigator Fellowship Award 2009–11. F.P.-R. was supported by a Asociación de Reumatólogos del Hospital de Cruces Grant for Clinical Investigation.
Disclosure statement: D.K. was supported by a National Institutes of Health Award (NIAMS K23 AR053858-04). D.K. is Principal Investigator of the CRYSTAL Registry, a Veteran Affairs registry supported by different pharmaceutical companies. All other authors have declared no conflicts of interest.