Abstract
Objective. To analyze prophylaxis using the CORRONA (COnsortium of Rheumatology Researchers Of North America) Gout Registry according to the American College of Rheumatology (ACR) guidelines, and to evaluate whether differences in disease characteristics influenced prophylaxis.
Methods. All patients with gout in the CORRONA Gout Registry between November 1, 2012, and November 26, 2013, were included. They were divided into 2 groups: “receiving prophylaxis” versus “not receiving prophylaxis” at the time of enrollment. Patients having a flare at time of visit were excluded. Descriptive statistics and multivariable logistic regression models were performed to evaluate the factors associated with prophylaxis.
Results. There were 1049 patients with gout available for analysis. There were 441 patients (42%) receiving prophylaxis and 608 (58%) not receiving prophylaxis. The most common drugs used for prophylaxis were colchicine (78%) and nonsteroidal antiinflammatory drugs (32%). Prophylaxis drug combination was used by 45 patients (10.2%). Patients in the “receiving prophylaxis” group were more likely to have a gout duration of ≤ 1 year (n = 68, p < 0.001), ≥ 1 flare in the year previous to enrollment (p < 0.001), ≥ 1 healthcare uses in the last year [Emergency Department (p = 0.029); outpatient visit to primary care, rheumatologist, or urgent care clinic (p < 0.001)], have tophi (p < 0.001), report pain > 3 (p = 0.001), and have disease activity > 10 (p < 0.001) compared with patients in the “not receiving prophylaxis” group.
Conclusion. Forty-two percent of patients with gout in the CORRONA Gout Registry were receiving prophylaxis. Prophylaxis was significantly more common in patients with a higher disease burden and activity, which is in agreement with the ACR guidelines. Our study highlights disease characteristics influencing prophylaxis and furthers our knowledge on current use of flare prophylaxis.
Gout is the most common inflammatory arthritis in men and older women, affecting an estimated 8.3 million adults in the United States1. The management of gout is focused on treating pain and inflammation associated with acute flares and preventing further acute flares, as well as monosodium urate (MSU) crystal deposition2,3,4.
A challenge associated with the successful management of gout is the increase in acute gout flares during the first months after initiation of urate-lowering therapies (ULT) as a result of rapid changes in serum urate (SU) levels5. This increase in flare frequency has been observed regardless of the choice of ULT (e.g., allopurinol, febuxostat, probenecid, pegloticase)6,7 and has been linked to suboptimal patient adherence to ULT4,8,9. Failure to adequately control gout flares during the initiation of ULT can result in suboptimal patient adherence to treatment5,10,11.
The exact mechanism by which ULT trigger acute flares is not well understood. It has been suggested that the increased flare rate may result from alterations in the chemical and/or physical state of preexisting MSU crystals when ULT induce rapid changes in SU levels12. Thus, superficial MSU crystals become solubilized, exposing uncoated (i.e., lacking protein coat) MSU crystals to monocytes and synoviocytes, stimulating activation of the NALP-3 inflammasome and increasing the expression of proinflammatory cytokines such as interleukin (IL)-113.
According to the 2012 American College of Rheumatology (ACR) guidelines14, antiinflammatory drugs for flare prophylaxis are recommended when ULT are initiated and should be continued if there is continuing gout disease activity and/or the SU target has not been achieved. The guidelines14 recommend low-dose colchicine (0.5 mg or 0.6 mg orally) once or twice daily. Colchicine is considered the standard of care for flare prophylaxis during the initiation of ULT, and is currently the only US Food and Drug Administration–approved therapy for gout flare prophylaxis. Other options were giving nonsteroidal antiinflammatory drugs (NSAID) with a proton pump inhibitor, low-dose corticosteroids [prednisone (PRED) ≤ 10 mg/d], or IL-1 inhibitors (anakinra in United States and Europe, and canakinumab in Europe). Recommended was the initiation of these drugs when starting ULT. Duration of prophylaxis was recommended for ≥ 6 months: 3 months posttarget SU reached (when no tophi) and 6 months posttarget SU reached when there were resolution tophi.
Management of gout includes treating chronic hyperuricemia and reducing the MSU crystal burden, as well as treating and preventing flares. Despite the current treatment recommendations, we suspect that in practice, antiinflammatory prophylaxis is not being prescribed regularly. Flares occur mainly during the first months, because of tophus mobilization and the rapid changes in SU. Our aim was to assess the frequency of gout flare prophylaxis, and record what drugs are commonly used for prophylaxis and if they are used according to the 2012 ACR recommendations14. In addition, we wanted to assess whether there were differences in disease characteristics such as the number of flares, pain in the past week, disease activity, medications used, and doses, as well as adverse events and hospitalizations in patients receiving prophylaxis compared with those who were not receiving prophylaxis.
MATERIALS AND METHODS
Data source and population
The COnsortium of Rheumatology Researchers Of North America (CORRONA) is a prospective observational cohort of patients with arthritis who are enrolled by participating rheumatologists in both academic and private practice sites; the details have been previously published15,16.
For the Gout Registry, rheumatologists were asked to enroll patients aged 21 years and older who met the criteria for the diagnosis of gout based on the ACR criteria. Data were collected from patients and their treating rheumatologists using standard clinical research forms. Information collected included demographics, comorbid conditions, gout presentation, disease severity and activity, family history of gout, body mass index (BMI), dietary intake over the past week, use of medications that can raise SU level (e.g., diuretics), use of medications for acute gouty inflammation (NSAID, colchicine, corticosteroids, and anakinra), and ULT including uricosurics, xanthine oxidase inhibitors, and recombinant uricase (pegloticase). Documentation included physician’s examination findings of tophi and inflamed joints, physician’s and patient’s global assessments of disease activity, patient’s assessment of pain, the Health Assessment Questionnaire (HAQ) assessing physical function, and serum uric acid levels from laboratory tests obtained within 10 days of the clinical encounter (these data were not mandated by the study protocol). Patients reported the number of days in the past 3 months they were unable to perform their usual activity. Healthcare use data were gathered, including gout hospitalizations within the last 3 years as well as Emergency Room (ER) and/or outpatient visits in the past 12 months for gout flares. Additionally, patients reported how many flares they managed themselves without seeing a healthcare professional.
There were 1049 patient visits, aged 18 and older, entered in the CORRONA Gout Registry database for this population between November 1, 2012, and November 26, 2013, and available for analysis. There are 34 rheumatology practices participating in the registry with > 80 rheumatologists involved, of whom 88% are in private practice. There are no disease activity requirements or comorbidity exclusion criteria. Approvals for data collection and analyses were obtained for academic and private practice sites from local and central institutional review boards, respectively.
Measures and data collection
We identified patients receiving prophylaxis by using the physician’s report of medication prescription. Among the data elements collected in the registry relevant to our study, there were physician’s examination findings of tophi and inflamed joints, physician’s and patient’s global assessments of disease activity, patient’s assessment of pain, the HAQ assessing physical function, and SU levels. SU data were recorded from laboratory tests obtained within 10 days of the clinical encounter. Chronic kidney disease (CKD) is defined as the presence of kidney damage, or a decreased level of kidney function, for a period of 3 months or more. Glomerular filtration rate (GFR) was calculated using the Modification of Diet in Renal Disease. CKD can be divided into 5 stages, based on GFR: normal (stage 1), mild CKD (stage 2) 60–89 ml/min, moderate CKD (stage 3) 30–59 ml/min, severe CKD (stage 4) 15–29 ml/min, and endstage renal disease (stage 5) < 15ml/min. Collection of laboratory data, however, was not mandated by the study protocol.
Prophylaxis medication exposure cohorts
Patients were categorized based on the use of chronic suppressive therapy and dichotomized into 2 groups: “receiving prophylaxis” versus “not receiving prophylaxis.” Patients with gout were considered in the “receiving prophylaxis” group if at the time of enrollment they were receiving colchicine, an NSAID, PRED (≤ 10 mg), or IL-1 inhibitors (anakinra or canakinumab), and were reported by the treating rheumatologist to not have a flare. Patients in the “not receiving prophylaxis” group included patients receiving PRED at a dose of > 10 mg or receiving no acute gout medications. Patients having a flare at time of visit were excluded from our study.
Covariates for comparison
We compared patients “receiving prophylaxis” to those “not receiving prophylaxis” based on baseline characteristics at the time of enrollment. These included sociodemographic characteristics such as age, sex, race, ethnicity, education, work status, insurance status, and marital status. BMI was evaluated and compared based on the World Health Organization categories of underweight (< 18.5), normal (18.5–24.9), overweight (25–29.9), obesity class 1 (30–34.9), and obesity class II/III (35+). Gout characteristics based on physician documentation included age at gout onset, duration of gout, presence of tophi, comorbid conditions, hospitalizations for gout, and medications (both acute and chronic medications). Patient-reported variables included disease activity [0–100 on a visual analog scale (VAS)], pain (0–100 on a VAS), number of gouty flares in the prior 12 months, and healthcare use for gout. Laboratory data included serum uric acid level, C-reactive protein (CRP), and erythrocyte sedimentation rate (ESR).
Statistical analysis
The Student 2-sample t test (continuous variables) and chi-square test (categorical variables) were used to compare factors between the “receiving prophylaxis” and “not receiving prophylaxis” groups. If cell sizes were small (< 5 found in a cell), the Fisher’s exact test was used in place of the chi-square test, and this is indicated in the tables.
Logistic regression was used to estimate the association of factors with prophylaxis treatment. First, each factor of interest was evaluated in a univariable model. Those factors with a resulting p value < 0.20 were further evaluated simultaneously in a multivariable logistic model. Prior to the multivariable model, the possible existence of multicollinearity among the predictor variables in the multivariable model was evaluated. Variables found to be pairwise correlated at a level > 0.20 or −0.20 were flagged. These flagged pairs included the number of flare attacks and pain, number of flare attacks and disease activity, and pain and disease activity. Because a patient would not be receiving joint therapy of allopurinol and febuxostat, the following categorical variables were created: not receiving either drug (reference), allopurinol 50–200 mg, allopurinol > 200–300 mg, allopurinol > 300 mg, febuxostat 40 mg, and febuxostat 80 mg/other.
Two final multivariable logistic models, 1 each for the number of flare attacks and disease activity, were fit. Each of the 2 models included marital status, years of gout, hospitalization in the last 3 years, serum acid level, combined allopurinol and febuxostat use, and presence of tophi because these were significant at the 0.2 level in univariable analyses. Sex and age at enrollment were also included in the final models. To account for the clustering of patients by physician, we evaluated the practice site as a random effect in the final multivariable model.
Because of the potentially small group sizes, we collapsed the categories for the following variables prior to inclusion in the regression models: race (white vs non-white), education (> 12 yrs vs ≤ 12 yrs), BMI categories (underweight/normal, overweight, and obese), work status (full/part-time vs others), marital status (single, married/partners, and others), adverse effect to any medication (allergy/side effect vs other), allopurinol (not receiving drug, 50–200 mg, > 200–300 mg, > 300 mg), febuxostat (not receiving drug, 40 mg, 80 mg/other), and probenecid (not receiving drug, 250–750 mg, 1000–2500 mg).
In addition, to avoid any modeling issues because of sparseness in the distributions of these continuous variables, the following continuous variables were dichotomized at the median of the “not receiving prophylaxis” group or as otherwise noted to have meaningful group comparisons in the final model: duration of gout (dichotomized as newly diagnosed ≤ 1 reference vs other, and median > 7 vs ≤ 7); number of flares in past year (> 1 vs ≤ 1); patient’s reported pain (> 3 vs ≤ 2); disease activity (> 10 vs ≤ 9); healthcare use (no. visits) in the past 12 months to the ER (≥ 1 vs 0); outpatient visit to rheumatologist, primary care provider, or urgent care clinic (> 1 vs ≤ 1); no care (≥ 1 vs 0); CRP at first visit (≤ 0.42 mg/dl vs > 0.42); ESR at first visit (≤ 12 mm/h vs > 12); and SU at first visit (≥ 5 vs 5, and also ≥ 6 vs < 6).
RESULTS
One thousand forty-nine patients with gout were available for analysis in the CORRONA Gout Registry. There were 441 patients with gout (42%) who were receiving prophylaxis and 608 (58%) who were not receiving prophylaxis. Characteristics of patients with gout in the CORRONA Registry at the time of enrollment by prophylaxis group compared among the 2 groups are shown in Table 1. Disease characteristics of patients with gout in the CORRONA Registry at the time of enrollment by prophylaxis group compared among the 2 groups are shown in Table 2. The history of comorbid conditions, alcohol consumption, and other dietary factors at the time of enrollment by prophylaxis group are shown in Table 3.
Characteristics of gout patients in CORRONA registry at time of enrollment by prophylaxis group. P values are from the 2-sample Student t test for continuous variables and the chi-square test of association for categorical variables, except where indicated with an asterisk. Values are n (%) unless otherwise specified.
Disease characteristics of patients with gout in CORRONA Registry at time of enrollment by the prophylaxis group. P values are from the 2-sample Student t test for continuous variables and the chi-square test of association for categorical variables, except where indicated with an asterisk. Values are n (%) unless otherwise specified.
History of comorbid conditions, alcohol consumption, and other dietary factors at time of enrollment by the prophylaxis group. P values are from the chi-square test of association, except where indicated with an asterisk. Values are n (%) unless otherwise specified.
Prophylaxis included colchicine (n = 345, 78%); NSAID (n = 140, 32%; ibuprofen, n = 56; indomethacin, n = 51; celecoxib, n = 44); PRED, n = 30 (7%); methylprednisolone, n = 4 (< 1%); and anakinra, n = 3 (< 1%). Prophylaxis drug combination was used by 45 patients (10.2%). Ninety-six percent of colchicine users were receiving 0.5 mg or 0.6 mg, and 53% were receiving a once-daily dosing. There were 918 patients (88%) receiving a ULT (allopurinol 79%, n = 735; febuxostat 16%, n = 162; probenecid < 1%, n = 17; and pegloticase < 1%, n = 7). Of the patients, 292 (67%) receiving prophylaxis had a disease duration of ≥ 1 year.
Within both subgroups not receiving/receiving ULT, stratified by patients receiving prophylaxis versus those not receiving prophylaxis, SU level was associated with ULT use: p < 0.001 within both groups (Table 4).
SU Levels by ULT use stratified by patients receiving prophylaxis versus not receiving prophylaxis. P values are from the chi-square test of association. Values are n (%) unless otherwise specified.
Of the 441 patients receiving prophylaxis, 200 (46%) had CKD reported (Table 2). Four patients with severe CKD [creatinine clearance (CrCl) < 30 ml/min] were receiving colchicine prophylaxis; 2 (50%) of them had their colchicine dose adjusted accordingly to 0.3 mg/day. NSAID were used for prophylaxis in patients with CKD: 17% of patients (56/329) with CKD stage 2 and 14% of patients (19/132) with CKD stage 3, and 4% of patients (1/23) with CKD stage 4 were receiving NSAID prophylaxis.
Patients in the “receiving prophylaxis” group were more likely to have a gout duration of ≤ 1 year (n = 68, p < 0.001), ≥ 1 flare in year previous to enrollment (p < 0.001), and ≥ 1 healthcare use in the last year (ER, p = 0.029; outpatient visit to primary care, rheumatologist, or urgent care clinic, p < 0.001; or ≥ 1 flare but did not seek care, p < 0.001) compared with patients in the “not receiving prophylaxis” group. In addition, patients in the “receiving prophylaxis” group were less likely to have SU < 6 mg/dl (p < 0.001) and more likely to have tophi (p < 0.001), report pain > 3 (p = 0.001), and have disease activity > 10 (p < 0.001) compared with patients in the “not receiving prophylaxis” group. Comorbidities did not have a significant effect on whether the patient was treated with prophylaxis.
Logistic regression was used to estimate the association of factors with prophylaxis treatment. First, each factor of interest was evaluated in a univariable model and those factors with a resulting p value < 0.20 were further evaluated in the multivariable logistic models (Table 5). Although sex and age at enrollment were not significant at the 20% α level in univariable models, they were included in the final multivariable models.
Multivariable OR and 95% CI from logistic regression. Sex and age at enrollment are included in model although not significant at the 0.020 level in the univariable models. Values are AOR (95% CI).
Those factors with a resulting p value < 0.20 were further evaluated in a multivariable logistic model. Prior to the multivariable model, we evaluated the possible existence of multicollinearity among the predictor variables in the multivariable model. Variables found to be pairwise correlated at a level > 0.20 or −0.20 were flagged. As expected, the number of flares (because it is the sum of flares resulting in an ER visit, an MD visit, or flares with no medical attention sought) was correlated with its 3 components, and the correlation among the 3 components ranged from 0.24 to 0.40. In addition, the number of flares was also correlated with pain (0.25) and disease activity (0.32), and pain and disease activity were correlated (0.56). Variables that indicate the use of allopurinol and febuxostat were also correlated (−0.65). We therefore created a combined variable for allopurinol and febuxostat use with the following categories: not receiving either drug, allopurinol 50–200 mg, allopurinol > 200–300 mg, allopurinol > 300 mg, febuxostat 40 mg, and febuxostat 80 mg/other. We further fit 2 final models, 1 each for the number of flares and disease activity. Each of the 2 models further included marital status, years of gout, hospitalization in the last 3 years, serum acid level, combined allopurinol and febuxostat use, and presence of tophi because these were significant at the 0.2 level in univariate analyses. Sex and age at enrollment were also included in the final models. To account for the clustering of patients by physician, we evaluated practice site as a random effect in the final multivariable model. Because this effect proved to be significant, the final models shown in Table 4 include adjustment for physician effect.
DISCUSSION
According to the 2012 ACR guidelines14, antiinflammatory drugs for flare prophylaxis are recommended when ULT are initiated and should be continued if there is continuing gout disease activity and/or the SU target has not been achieved. In our study, we found that only 42% of patients with gout in the CORRONA Gout Registry were receiving gout flare prophylaxis. Because ours was a retrospective study, we may not have identified patients previously receiving prophylaxis. The most common drugs used for prophylaxis among the CORRONA Gout Registry patients were colchicine (78%) and NSAID (32%). Prophylaxis drug combination was used in 45 patients (10.2%). Only 26% (10% colchicine and 16% NSAID) of patients with gout were receiving gout flare prophylaxis in a study of 643 patients receiving a new allopurinol prescription17. In a 2002 study of gout treatment patterns in the United States, whereas about 2.8 million prescriptions for allopurinol were issued, only 381,000 prescriptions for colchicine and 700,000 prescriptions for NSAID were issued, which may or may not have been for prophylaxis18.
To our knowledge, our study is the first to evaluate whether differences in disease characteristics influenced the use of prophylaxis. We found prophylaxis to be significantly more common in patients with higher disease activity. Thus patients with ≥ 1 flare in the year previous to enrollment, greater healthcare use, higher ULT doses, and presence of tophi contributed to increased use of prophylaxis. Patients who reported greater disease activity and pain were also more likely to receive prophylaxis. This was in agreement with the ACR guidelines, which support prophylaxis in patients with continuing disease activity. Interestingly, many patients receiving prophylaxis had a disease duration of ≥ 1 year, and although length of observation was not specified, secondary to our study’s retrospective design, we suggest that further studies are needed to assess appropriate length of prophylaxis.
Prophylaxis during ULT initiation can reduce the incidence and severity of gout flares. The efficacy of colchicine prophylaxis has been established, with lower rates of flare recurrence and less-severe flares in patients who received colchicine compared with placebo5,7,14,17,18. The guidelines12 recommend low-dose colchicine (0.5 mg or 0.6 mg orally) once or twice daily. Prophylactic doses of colchicine are generally well tolerated5,7; however, for patients with severe CKD (creatinine clearance < 30 ml/min), the recommended colchicine dose is 0.3 mg/day and was given to 2 of the 4 patients receiving colchicine prophylaxis in our cohort19. We found patients to be commonly treated with NSAID prophylaxis despite having CKD. Although current treatment recommendations suggest NSAID as an option for gout prophylaxis, the longterm safety of NSAID may be an issue20 and contributes to them not being an appropriate choice for gout flare prophylaxis, although no guidelines for safe dosing have been established in the CKD population.
IL-1 inhibitors may be a useful alternative for patients who are intolerant to or have contraindications for colchicine or NSAID. The proinflammatory cytokine IL-1 is involved in mediating the inflammation in gout13,21. In our study, anakinra, an IL-1 receptor antagonist, was used for prophylaxis in less than 1% of patients. In clinical trials, rilonacept and canakinumab, IL-1 inhibitors, demonstrated significant flare prevention during ULT initiation22,23.
There are a number of strengths and potential limitations in our study. Our analysis represents a large US-based observational study of gout treatment. Moreover, the data collection covered 34 practices and more than 80 rheumatologists, with the majority of the data collected from community-based rheumatology practices. This approach is consistent with the recommendations from the National Institutes of Health Roadmap Initiative to reengineer the clinical research enterprise, including community-based investigators who can expedite study recruitment. In addition, we were able to examine both physician-derived and patient-derived outcome measures, and account for a broad set of potential confounding variables. Nevertheless, the generalizability of our study results remains a potential limitation of our study10. It is possible that potential explanatory factors that were not part of the study data collection could influence the study results, such as medication adherence and patient literacy.
Effective ULT treatment and SU reduction can be achieved with old ULT as well as with newer ULT and those in development4,18. However, initiation of these therapies is also associated with a high incidence of flares. Therefore, it is increasingly important that prophylaxis be an integral part of chronic gout treatment. Our study highlights disease characteristics influencing the use of prophylaxis and advances our knowledge on the current use of gout flare prophylaxis. Despite the 2012 ACR guidelines for use of antiinflammatory drugs for flare prophylaxis14, antiinflammatory prophylaxis is not commonly prescribed.
Gout flare prophylaxis needs to be an integral part of chronic gout treatment. However, gout prophylaxis is used uncommonly. Our study highlights disease characteristics influencing the use of prophylaxis and furthers our knowledge on the current use of gout flare prophylaxis.
Footnotes
Carol J. Etzel is a full-time employee at Corrona LLC. Jeffrey D. Greenberg is an employee and shareholder in Corrona LLC. Joel M. Kremer is a shareholder in Corrona LLC and receives employment compensation. Leslie R. Harrold is an employee and has a research contract with Corrona LLC.
- Accepted for publication January 22, 2016.
REFERENCES
In this issue
