The Determinants and Disparities of Gout within the Multiethnic Cohort

Abstract

BACKGROUND: While prior studies have identified differences in gout prevalence estimates by ethnic groups, especially among Polynesian and Black populations relative to Whites, the incidence and effect of behavioral factors on risk of gout within these groups remains severely understudied. In addition, few have investigated the association between biomolecules and gout, which may serve as early risk indicators. The purpose of the following Master’s Thesis was to ascertain the incidence, the effects and clustering of behavioral risk factors, and associations between biomolecules, and incident gout within a large multiethnic populations of older adults from the Multiethnic Cohort (MEC). METHODS: Two studies were performed using secondary data from the MEC linked to Medicare claims data on gout (ICD-9: 274.9; ICD-10: M10.9) to assess incident gout cases. Study 1 assessed the incidence rates, effects of behavioral factors (alcohol use, smoking, vitamin C supplementation (VC), physical activity, and dietary quality score), and risk profiles of gout among ethnically disaggregated samples of Black (B, N = 15,660), Native Hawaiian (NH, N = 7,600), Japanese (JA, N = 32,923), Latino (L, N = 21,793), and White (W, N = 29,129) participants from the MEC utilizing ethnic-specific Cox regression modeling and conditional inference survival tree analyses. Study 2 examine the associations between serum concentrations of biomolecules (C-reactive protein [CRP], cholesterols, triglycerides, tocopherols, carotenes, adiponectin, and leptin) and incident gout using an ethnically aggregated Cox regression. All covariates in Study 1 were ascertain from self-reported responses to the MEC baseline questionnaire, while biomolecules in Study 2 were assessed through high-performance liquid chromatography and an autoanalyzer (for CRP). RESULTS: NH had the highest incidence of gout (IDR: 9.97 per 1,000 person-years), followed by B, JA, W, and L participants. Three or more alcoholic drinks per day was associated with an increased risk among all ethnic groups except Ls, with the largest effect observed among B participants (HR: 1.55, 95%CI: 1.23, 1.94). Similarly, current smoking was associated with a 28% (95%CI: 10-49%) increased risk among B, and a 17% (95%CI: 3-32%) increased risk among JA, participants. Higher DASH tertiles were associated with a decreased risk of gout among all ethnic groups, apart from L, with the largest effect observed among Ws in the highest versus lowest tertile comparison (HR1vs3: 0.70, 95%CI: 0.63, 0.78), while VC was weakly associated with a decreased risk of gout only among NH (HR: 0.85, 95%CI: 0.75, 0.98) and JA (HR: 0.91, 95%CI: 0.85, 0.98) participants. Survival tree analysis identified several risk clusters, with BMI identified as the most important factor in predicting gout risk (cut-point: 25.843, p < 0.001). The highest risk profile was identified as NHs with a BMI > 25.843 and a history of hypertension, that consumed greater than 285.526 grams of white rice per day. The highest tertile of CRP (HR1vs3: 1.84, 95%CI: 1.51, 2.25) and γ-tocopherol (HR1vs3: 1.69, 95%CI: 1.35, 2.11) were associated with an increased risk of gout compared to the lowest tertile serum concentrations. Compared to the lowest tertile, elevated concentrations of leptin (HR1vs2: 1.55, 95%CI: 1.26, 1.91; HR1vs3: 2.73, 95%CI: 2.14, 3.47) and triglycerides (HR1vs2: 1.34, 95%CI: 1.10, 1.65; HR1vs3: 1.83, 95%CI: 1.50, 2.23) were generally associated with an elevated risk of gout. Conversely, HDL-cholesterol (HR1vs3: 0.62, 95%CI: 0.51, 0.77), adiponectin (HR1vs2: 0.76, 95%CI: 0.63, 0.92; HR1vs3: 0.67, 95%CI: 0.54, 0.82), α-carotene (HR1vs2: 0.68, 95%CI: 0.55, 0.84; HR1vs3: 0.56, 95%CI: 0.44, 0.70), and β-carotene (HR1vs2: 0.62, 95%CI: 0.50, 0.76; HR1vs3: 0.52, 95%CI: 0.41, 0.66) were associated with a decreased risk of gout in an apparent ‘dose-response’ gradient. DISCUSSION: Overall, ethnic differences in both the incidence and effects of modifiable risk factors in gout were observed, with NH having the highest rates of gout and L having the lowest. Alcohol use and smoking were associated with an increased risk, while physical activity, VC, and DASH scores were associated with a decreased risk. Specific and novel dietary factors were identified through the exploratory risk profile analysis, and high-risk profiles were identified. In addition, several biomolecular associations were observed that were consistent with prior literature, possibly elucidating objectively measured early-warning indicators for gout. Limitations are discussed, including self-reporting of demographic characteristics, medical histories, and behaviors, along with disadvantages in using Medicare claims as a proxy for incident gout. Finally, future directions are recommended to address issues that fell outside of the scope of the current investigation.