Arthritis & Rheumatism, Volume 63,
November 2011 Abstract Supplement

Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Chicago, Illinois November 4-9, 2011.


Clinical Characteristics of Difficult-to-Treat Gout Patients: a Principal Components Analysis.

Vaysbrot1,  Elizaveta, Lee1,  Yoojin, McLaughlin1,  Sarah, Agashivala2,  Neetu, Yadao3,  Anthony, McAlindon1,  Timothy E., Harvey1,  William F.

Tufts Medical Center, Boston, MA
Novartis Pharmaceutical Corporation, East Hanover, NJ
Novartis Pharmaceuticals Corporation, East Hanover, NJ

Background/Purpose:

Difficult-to-treat gout patients have unmet medical need in symptom control due to limited treatment options and may require additional healthcare services, including ER visits and hospitalization. Understanding which clinical characteristics of this population are significant and how they are intercorrelated is vital for future research but limited at this point. Our objective was to uncover the correlation between the clinical characteristics of difficult-to-treat gout patients to identify key variables for future research.

Methods:

Tufts Medical Center charts were reviewed for fiscal years 2008–2010. As no definition of difficult-to-treat gout exists, we chose inclusion criteria based on ease of electronic searching and on the principle that poor control of gout symptoms results in frequent visits/admissions. Inpatients with difficult-to-treat gout were defined as persons with >=2 inpatient consultations in one year by a rheumatologist for gout (ICD9 274.xx) regardless of admission diagnosis. Outpatients were selected on the basis of a new rheumatologic outpatient consult and >=5 outpatient rheumatology visits for gout in a year. From each of these two criteria sets, 75 inpatients and 75 outpatients with the highest billing codes were selected. Data were extracted on patient demographics, referral data, disease characteristics, comorbidities, and medication history over a 12 months period. Nominal dichotomous (Yes/No) variables such as "diuretics use" and comorbid conditions data were selected to perform a principal components analysis (PCA) using an orthogonal rotation (SAS version 9.2. Cary, NC. SAS Institute Inc.). Factors were selected based on the Kaiser criterion (eigenvalue >1), and components were included if their loading value was >0.4.

Results:

Mean age was 63±13.7 years, 71% were male, and 67% were white. The most common referral source was primary care (43.2%). 9.3% had >=3 attacks in the previous year. 72% were on >=5 concomitant medications, excluding those for gout. Approximately 20% had an emergency room visit and 17% were hospitalized specifically for a gout attack. Five factors were identified in PCA (see table).

Factors (Eigenvalues) and principal components (loading values)Patient Records (n=150)
Factor 1 (3.37): Chronic Heart Failure (0.81), Coronary Heart Disease (0.73), Cerebrovascular Disease (0.64), Diuretic Use (0.53), Dyslipidemia (0.52) 
Patients with 5 components (%(n))5.3% (8)
Patients with 4 components (% (n))10.7% (16)
Patients with 3 components (% (n))4% (6)
Patients with 2 components (% (n))19.3% (29)
Patients with 1 component (% (n))31.3% (47)
Patients with 0 components (% (n))29.3% (44)
Factor 2 (2.55): Cirrhosis (0.87), Chronic Liver Disease (0.86), Upper GI Ulcer (0.66), GI Bleeding history (0.48) 
Patients with 4 components (% (n))1.3% (2)
Patients with 3 components (% (n))0
Patients with 2 components (% (n))4% (6)
Patients with 1 component (% (n))8% (12)
Patients with 0 components (% (n))86.7% (130)
Factor 3 (1.80): Hypertension (0.66), Metabolic Syndrome (0.61), Diabetes Mellitus (0.54), Chronic Kidney Disease (0.53), Arthritis other than gout (0.50) 
Patients with 5 components (% (n))6% (9)
Patients with 4 components (% (n))8% (12)
Patients with 3 components (% (n))18.7% (28)
Patients with 2 components (% (n))24.7% (37)
Patients with 1 component (% (n))27.3% (41)
Patients with 0 components (% (n))15.3% (23)
Factor 4 (1.37): Obesity (0.56), Alcohol Use (-0.55), Peripheral Vascular Disease (-0.51) 
Patients with 3 components (% (n))0
Patients with 2 components (% (n))3.3% (5)
Patients with 1 component (% (n))52.% (79)
Patients with 0 components (% (n))44% (66)
Factor 5 (1.14): Renal Stones (0.56), Tophi (-0.42) 
Patients with 2 components (% (n))1.3% (2)
Patients with 1 components (% (n))26.7% (40)
Patients with 0 components (% (n))72% (108)
* negative loading values indicate that absence of that component is correlated with the other components.

Conclusion:

The four out of five factors had components with clinical relationships: cardiovascular diseases, gastrointestinal disorders, metabolic syndrome and features of advanced gout. The components in factor 4 did not have a clear clinical relationship. The study was limited by its retrospective nature, small sample size, and an inexact, exploratory nature of PCA method. Characteristics of this population may vary if using different definitions of difficult-to-treat. Despite the limitations, our results confirm the clinical intuition that the above conditions are linked to difficult-to-treat gout and may help design future research.

To cite this abstract, please use the following information:
Vaysbrot, Elizaveta, Lee, Yoojin, McLaughlin, Sarah, Agashivala, Neetu, Yadao, Anthony, McAlindon, Timothy E., et al; Clinical Characteristics of Difficult-to-Treat Gout Patients: a Principal Components Analysis. [abstract]. Arthritis Rheum 2011;63 Suppl 10 :1033
DOI:

Abstract Supplement

Meeting Menu

2011 ACR/ARHP