514 FRAILTY IN OSTEOARTHRITIS AND THE INFLUENCE OF CO-MORBIDITY

Introduction Risk factors for frailty, including low physical activity and chronic pain, are common among people with osteoarthritis. The aim of this analysis was to determine the association between osteoarthritis and frailty and to determine whether comorbidities interact additively with OA to increase the likelihood of frailty. Methods Men and women aged 40-69 years who contributed to the UK Biobank were analysed. Data about self-reported physician-diagnosed diseases was collected, as well physical measurements, including hand-grip strength. Frailty (robust, pre-frail, frail) was assessed using a modified frailty phenotype, comprising five components: low grip strength, slow walking speed, weight loss, low physical activity, and exhaustion. The association between osteoarthritis and the frailty phenotype was determined using negative binomial regression, adjusting for age, sex, body mass index, smoking status, and Townsend deprivation score. We calculated the attributable proportion of risk of frailty due to additive interaction between osteoarthritis and common co-morbidities (cardiovascular disease, diabetes, COPD, and depression). Results 457,561 people were included, 35,884 (7.8%) had osteoarthritis. The adjusted relative risk ratio (95% CI) for pre-frailty and frailty (versus robust), respectively was higher among people with (versus without) osteoarthritis: 1.58 (1.54, 1.62) and 3.41 (3.26, 3.56). There was significant additive interaction between the presence of osteoarthritis and each of the co-morbidities considered in increasing risk of frailty, particularly diabetes (attributable proportion of risk due to additive interaction with osteoarthritis (95% CI)), 0.49 (0.42, 0.55), coronary heart disease 0.48 (0.41, 0.55), and depression 0.47 (0.41, 0.53). Conclusions Our results suggest that people with OA are at increased risk of pre-frailty and frailty. The mechanisms are not fully understood, though co-morbidity appears to contribute to the risk of frailty beyond the expected additivity of risk due to OA and co-morbidity. Early diagnosis and optimal management of co-morbidities in people with OA may be beneficial.

Diabetes, Hypertension, and Incidence of Chronic Kidney Disease: Is There any Multiplicative or Additive Interaction?

Background: The burden of chronic kidney disease (CKD) is on the rise worldwide; diabetes and hypertension are mentioned as the main contributors. Objectives: The current study aimed to investigate the multiplicative and additive interaction of diabetes and hypertension in the incidence of CKD. Methods: In this population-based cohort study, 7342 subjects aged 20 years or above (46.8% male) were divided into four groups: no diabetes and hypertension; diabetes and no hypertension; no diabetes but suffer from hypertension; and both diabetes and hypertension. The multivariable Cox regression was used to determine the effect of diabetes, hypertension, and their multiplicative interaction on CKD. The following indices were used to determine the additive interaction of diabetes and hypertension: the relative excess risk of interaction, the attributable proportion due to interaction, and the synergism index. Results: Diabetes and hypertension had no significant multiplicative interaction in men (hazard ratio of 0.93, P value: 0.764) and women (hazard ratio of 0.79, P value: 0.198); furthermore, no additive interaction was found in men (Relative Excess Risk due to Interaction of 0.79, P value: 0.199; Attributable Proportion due to Interaction of 0.22, P value: 0.130; Synergy index of 1.44, P value: 0.183) and women (Relative Excess Risk due to Interaction of -0.26, P value: 0.233, Attributable Proportion due to Interaction of -0.21, P value: 0.266; Synergy index of 0.48, P value: 0.254). Conclusions: This study demonstrated no synergic effect between diabetes and hypertension on the incidence of CKD.

Relationship between Annual Average Concentration of Ambient PM10 and Out-patients with Respiratory Disease: Thailand Case Study

The objectives of this research were to identify (1) correlation between annual average concentration of ambient PM10 and prevalence of out-patients with respiratory diseases; (2) the relative risk and attributable proportion of out-patients with respiratory disease due to long-term exposure to ambient PM10; and (3) the correlation between annual average concentration ofambient PM10 and the relative risk of out-patients with respiratory disease. Time-series data of annual average concentration of ambient PM10 and prevalence rate of out-patients with respiratory disease during an 11- year period (2004-2014) in the study area were obtained from the Pollution Control Department, Ministry of Natural Resources and Environment, and the Bureau of Policy and Strategy, Ministry of Public Health, respectively. To estimate the relationship, Pearson’s product moment correlations between variables were calculated and significance testing of correlation coefficients were carried out. A retrospective cohort method was used to study the annual average concentration of ambient PM10 ratio, the relative risk and the attributable proportion. No association was found between the annual average concentration of ambient PM10 and the prevalence rate of out-patients with respiratory disease; however, there was a highly significant positive relationship between the annual average concentration of ambient PM10 ratio and the relative risk of out-patients with respiratory disease (r = 0.852, df = 4, p < 0.05). The relative risk of out-patients with respiratory disease due to exposure to the annual average concentration of ambient PM10 equal to or higher than 30 μg/m3 varied within the range of 0.99 - 1.14, and the attributable proportion of out-patients with respiratory disease were about 10 % and 3 % attributed to exposure to the annual average concentration of ambient PM10 higher than 30 μg m-3 and 40 μg m-3, respectively.

A General Framework for and New Normalization of Attributable Proportion

A unified theory is developed for attributable proportion (AP) and population attributable fraction (PAF) of joint effects, marginal effects or interaction among factors. We use a novel normalization with a range between –1 and 1 that gives the traditional definitions of AP or PAF when positive, but is different when they are negative. We also allow for an arbitrary number of factors, both those of primary interest and confounders, and quantify interaction as departure from a given model, such as a multiplicative, additive odds or disjunctive one. In particular, this makes it possible to compare different types of threeway or higher order interactions. Effect parameters are estimated on a linear or logit scale in order to find point estimates and confidence intervals for the various versions of AP and PAF, for prospective or retrospective studies. We investigate the accuracy of three confidence intervals; two of which use the delta method and a third bootstrapped interval. It is found that the delta method with logit type transformations, and the bootstrap, perform well for a wide range of models. The methodology is also applied to a multiple sclerosis (MS) data set, with smoking and two genetic variables as risk factors.

The Attributable Proportion of Specific Leisure-Time Physical Activities to Total Leisure Activity Volume Among US Adults, National Health and Nutrition Examination Survey 1999–2006

Background:Previous studies have examined participation in specific leisure-time physical activities (PA) among US adults. The purpose of this study was to identify specific activities that contribute substantially to total volume of leisure-time PA in US adults.Methods:Proportion of total volume of leisure-time PA moderate-equivalent minutes attributable to 9 specific types of activities was estimated using self-reported data from 21,685 adult participants (≥ 18 years) in the National Health and Nutrition Examination Survey 1999–2006.Results:Overall, walking (28%), sports (22%), and dancing (9%) contributed most to PA volume. Attributable proportion was higher among men than women for sports (30% vs. 11%) and higher among women than men for walking (36% vs. 23%), dancing (16% vs. 4%), and conditioning exercises (10% vs. 5%). The proportion was lower for walking, but higher for sports, among active adults than those insufficiently active and increased with age for walking. Compared with other racial/ethnic groups, the proportion was lower for sports among non-Hispanic white men and for dancing among non-Hispanic white women.Conclusions:Walking, sports, and dance account for the most activity time among US adults overall, yet some demographic variations exist. Strategies for PA promotion should be tailored to differences across population subgroups.

Abstract 3441: Paraoxonase-192 Polymorphism Interaction with Low HDL Cholesterol in Coronary Artery Risk

Background: Increased lipid peroxidation is associated with accelerated progression of atherosclerosis. Paraoxonase (PON1) is an antioxidative enzyme in high density lipoproteins (HDL), which protects against lipid per oxidation and Coronary Artery Disease (CAD). PON 1 activity is under genetic control and its molecular basis is a polymorphism in the PON 1 gene that shows two common isoforms: Q (192 Gln) with high ability to protect LDL from lipid peroxidation in vitro, and R (Arg) with lower ability. Aim: To explore the interaction of the R allele paraoxonase gene and low HDL cholesterol concentrations, in CAD risk emergence. Methods: 818 individuals participated in the study, 298 coronary patients, 55.0±10.3 years, 78.9% male sex, and 520 healthy controls, 47.3±12.6 years, 57.5% male sex, age and sex matched. Low HDL-C was defined as <0.90 mmol/L in men and < 1.11 mmol/L in the women. Comparisons of genotypes between cases and control subjects were performed by a chi-square test. Statistical significance was accepted at p<0.05. Odds ratio as well as their 95% confidence intervals for the RR genotypes and HDL deficient subjects were computed using univariate analysis (2x2 tables). To determine the interaction between the RR paraoxonase genotype and HDL-deficient subjects, we used the epidemiologic tables 4x2 and the synergy measures: the additive model (Rothman’s synergy index) and the multiplicative model (Khoury’s synergy index). We calculated the relative excess risk (RERI) and the attributable proportion (AP) due to interaction (Rothman). Results: The PON 1 RR192 is associated with coronary heart disease [OR=1.61; p=0.043] in whole population. The HDL-deficient subjects 192 RR homozygotes showed an increased risk of CAD (OR=17.38; p< 0.0001) compared to normal HDL 192 RR (OR=1.39; p=0.348) and HDL-deficient subjects not carrying RR genotype (OR=7.79; p<0.0001). The genotype PON 192 RR increases the risk of CAD in the HDL-deficient populations (SI=2.3, SIM=1.6). The attributable proportion due to this interaction (AP) was 0.53, meaning that 53% of CAD was explained by this interaction. Conclusion: These data suggest the existence of a synergistic effect of the RR 192 PON 1 genotype (with lower ability) and HDL-deficient subjects in CAD emergence.