Combining zip code-based population data and pharmacy administrative claims data to create measures of social determinants of health.

BACKGROUND: Social determinants of health (SDoH) are key factors that impact health outcomes. However, there are many barriers to collecting SDoH data (eg, cost of data collection, technological barriers, and lack of standardized measures). Population data may provide an accessible alternative to collecting SDoH data for patients. OBJECTIVE: To explain how population data can be leveraged to create SDoH measures, assess the association of population SDoH measures with diabetic medication adherence, and discuss how understanding a patient's SDoH can inform care plans and patient engagement. METHODS: A nationally representative commercial sample of patients who were aged 18 years and older and met Pharmacy Quality Alliance inclusion criteria for diabetes mellitus were analyzed (N = 37,789). US Census and North American Industry Classification System data were combined with pharmacy administrative claims data to create SDoH measures. Derived measures represent 2 SDoH domains: (1) economic stability (housing density, housing relocation, jobs per resident, and average salary) and (2) health care access and quality (urban/rural classification, distance traveled to prescriber and pharmacy, use of a primary care provider [PCP], and residents per PCP). The association of population SDoH measures with diabetic medication adherence (proportion of days covered) was assessed via logistic regression, which included covariates (eg, sex, age, comorbidities, and prescription plan attributes). RESULTS: As housing density (houses per resident) increased, so did the likelihood of adherence (odds ratio = 1.54, 95% CI = 1.21-1.97, P = 0.001). Relative to patients who did not move, patients who moved once had 0.87 (95% CI = 0.81-0.93, P < 0.001) the odds of being adherent, and patients who moved 2 or more times had 0.82 (95% CI = 0.71-0.95, P = 0.008) the odds of being adherent. Compared with areas with fewer jobs per resident, patients living within a zip code with 0.16 to 0.26 jobs per resident were 1.12 (95% CI = 1.04-1.20, P = 0.002) times more likely to be adherent. Patients who lived in an urban cluster were 1.11 (95% CI = 1.01-1.22, P = 0.037) times more likely to be adherent than patients living in a rural area. Patients who travel at least 25 miles to their prescriber had 0.82 (95% CI = 0.77-0.86, P < 0.001) the odds of being adherent. Community pharmacy users had 0.65 (95% CI = 0.59-0.71, P < 0.001) the odds of being adherent compared with mail order pharmacy users. Patients who had a PCP were 1.26 (95% CI = 1.18-1.34, P < 0.001) times more likely to be adherent to their medication. CONCLUSIONS: Leveraging publicly available population data to create SDoH measures is an accessible option to overcome barriers to SDoH data collection. Derived measures can be used to increase equity in care received by identifying patients who could benefit from assistance with medication adherence.

Discovery of gene-gene interactions across multiple independent data sets of late onset Alzheimer disease from the Alzheimer Disease Genetics Consortium.

Late-onset Alzheimer disease (AD) has a complex genetic etiology, involving locus heterogeneity, polygenic inheritance, and gene-gene interactions; however, the investigation of interactions in recent genome-wide association studies has been limited. We used a biological knowledge-driven approach to evaluate gene-gene interactions for consistency across 13 data sets from the Alzheimer Disease Genetics Consortium. Fifteen single nucleotide polymorphism (SNP)-SNP pairs within 3 gene-gene combinations were identified: SIRT1 × ABCB1, PSAP × PEBP4, and GRIN2B × ADRA1A. In addition, we extend a previously identified interaction from an endophenotype analysis between RYR3 × CACNA1C. Finally, post hoc gene expression analyses of the implicated SNPs further implicate SIRT1 and ABCB1, and implicate CDH23 which was most recently identified as an AD risk locus in an epigenetic analysis of AD. The observed interactions in this article highlight ways in which genotypic variation related to disease may depend on the genetic context in which it occurs. Further, our results highlight the utility of evaluating genetic interactions to explain additional variance in AD risk and identify novel molecular mechanisms of AD pathogenesis.

Detection of pleiotropy through a Phenome-wide association study (PheWAS) of epidemiologic data as part of the Environmental Architecture for Genes Linked to Environment (EAGLE) study.

We performed a Phenome-wide association study (PheWAS) utilizing diverse genotypic and phenotypic data existing across multiple populations in the National Health and Nutrition Examination Surveys (NHANES), conducted by the Centers for Disease Control and Prevention (CDC), and accessed by the Epidemiological Architecture for Genes Linked to Environment (EAGLE) study. We calculated comprehensive tests of association in Genetic NHANES using 80 SNPs and 1,008 phenotypes (grouped into 184 phenotype classes), stratified by race-ethnicity. Genetic NHANES includes three surveys (NHANES III, 1999-2000, and 2001-2002) and three race-ethnicities: non-Hispanic whites (n = 6,634), non-Hispanic blacks (n = 3,458), and Mexican Americans (n = 3,950). We identified 69 PheWAS associations replicating across surveys for the same SNP, phenotype-class, direction of effect, and race-ethnicity at p<0.01, allele frequency >0.01, and sample size >200. Of these 69 PheWAS associations, 39 replicated previously reported SNP-phenotype associations, 9 were related to previously reported associations, and 21 were novel associations. Fourteen results had the same direction of effect across more than one race-ethnicity: one result was novel, 11 replicated previously reported associations, and two were related to previously reported results. Thirteen SNPs showed evidence of pleiotropy. We further explored results with gene-based biological networks, contrasting the direction of effect for pleiotropic associations across phenotypes. One PheWAS result was ABCG2 missense SNP rs2231142, associated with uric acid levels in both non-Hispanic whites and Mexican Americans, protoporphyrin levels in non-Hispanic whites and Mexican Americans, and blood pressure levels in Mexican Americans. Another example was SNP rs1800588 near LIPC, significantly associated with the novel phenotypes of folate levels (Mexican Americans), vitamin E levels (non-Hispanic whites) and triglyceride levels (non-Hispanic whites), and replication for cholesterol levels. The results of this PheWAS show the utility of this approach for exposing more of the complex genetic architecture underlying multiple traits, through generating novel hypotheses for future research.

Lack of associations of ten candidate coronary heart disease risk genetic variants and subclinical atherosclerosis in four US populations: the Population Architecture using Genomics and Epidemiology (PAGE) study.

BACKGROUND: A number of genetic variants have been discovered by recent genome-wide association studies for their associations with clinical coronary heart disease (CHD). However, it is unclear whether these variants are also associated with the development of CHD as measured by subclinical atherosclerosis phenotypes, ankle brachial index (ABI), carotid artery intima-media thickness (cIMT) and carotid plaque. METHODS: Ten CHD risk single nucleotide polymorphisms (SNPs) were genotyped in individuals of European American (EA), African American (AA), American Indian (AI), and Mexican American (MA) ancestry in the Population Architecture using Genomics and Epidemiology (PAGE) study. In each individual study, we performed linear or logistic regression to examine population-specific associations between SNPs and ABI, common and internal cIMT, and plaque. The results from individual studies were meta-analyzed using a fixed effect inverse variance weighted model. RESULTS: None of the ten SNPs was significantly associated with ABI and common or internal cIMT, after Bonferroni correction. In the sample of 13,337 EA, 3809 AA, and 5353 AI individuals with carotid plaque measurement, the GCKR SNP rs780094 was significantly associated with the presence of plaque in AI only (OR = 1.32, 95% confidence interval: 1.17, 1.49, P = 1.08 × 10(-5)), but not in the other populations (P = 0.90 in EA and P = 0.99 in AA). A 9p21 region SNP, rs1333049, was nominally associated with plaque in EA (OR = 1.07, P = 0.02) and in AI (OR = 1.10, P = 0.05). CONCLUSIONS: We identified a significant association between rs780094 and plaque in AI populations, which needs to be replicated in future studies. There was little evidence that the index CHD risk variants identified through genome-wide association studies in EA influence the development of CHD through subclinical atherosclerosis as assessed by cIMT and ABI across ancestries.

Association of functional polymorphism rs2231142 (Q141K) in the ABCG2 gene with serum uric acid and gout in 4 US populations: the PAGE Study.

A loss-of-function mutation (Q141K, rs2231142) in the ATP-binding cassette, subfamily G, member 2 gene (ABCG2) has been shown to be associated with serum uric acid levels and gout in Asians, Europeans, and European and African Americans; however, less is known about these associations in other populations. Rs2231142 was genotyped in 22,734 European Americans, 9,720 African Americans, 3,849 Mexican Americans, and 3,550 American Indians in the Population Architecture using Genomics and Epidemiology (PAGE) Study (2008-2012). Rs2231142 was significantly associated with serum uric acid levels (P = 2.37 × 10(-67), P = 3.98 × 10(-5), P = 6.97 × 10(-9), and P = 5.33 × 10(-4) in European Americans, African Americans, Mexican Americans, and American Indians, respectively) and gout (P = 2.83 × 10(-10), P = 0.01, and P = 0.01 in European Americans, African Americans, and Mexican Americans, respectively). Overall, the T allele was associated with a 0.24-mg/dL increase in serum uric acid level (P = 1.37 × 10(-80)) and a 1.75-fold increase in the odds of gout (P = 1.09 × 10(-12)). The association between rs2231142 and serum uric acid was significantly stronger in men, postmenopausal women, and hormone therapy users compared with their counterparts. The association with gout was also significantly stronger in men than in women. These results highlight a possible role of sex hormones in the regulation of ABCG2 urate transporter and its potential implications for the prevention, diagnosis, and treatment of hyperuricemia and gout.

Genetic variation in donor CTLA-4 regulatory region is a strong predictor of outcome after allogeneic hematopoietic cell transplantation for hematologic malignancies.

Relapse remains a major cause of death after allogeneic hematopoietic cell transplantation (allo-HCT). Graft-versus-tumor effect is primarily mediated by donor T cells. Cytotoxic T lymphocyte antigen-4 (CTLA-4) is a critical inhibitor of T cell proliferation. Single nucleotide polymorphisms (SNPs) in CTLA-4 may affect immune responses. We hypothesized that CTLA-4 SNPs will be associated with disease control after allo-HCT. One hundred sixty-four adult patients with the availability of pretransplantation recipient and donor DNA samples were included in this analysis. Ten tagSNPs of the CTLA-4 gene were identified. Donor CTLA-4 SNP rs4553808 was associated with decreased relapse-free survival (RFS) (P = .019) and overall survival (OS) (P = .033). In multivariable analysis of an additive genetic model, genotype of CTLA-4 SNP rs4553808 was an independent risk factor for inferior RFS (hazard ratio [HR] = 1.73, 95% confidence interval [CI] 1.10-2.71, P = .017) and OS (HR = 1.84, 95% CI 1.13-3.0, P = .015). CTLA-4 SNPs can be used to identify high-risk patient subsets that may benefit from preemptive immunomodulation to decrease relapse rates and improve survival.

Genetic variation in recipient B-cell activating factor modulates phenotype of GVHD.

B-cell activating factor (BAFF) single nucleotide polymorphisms (SNPs) are associated with autoimmune diseases. Because patients with classic and overlap chronic GVHD (cGVHD) have features of autoimmune diseases, we studied the association of recipient and/or donor BAFF SNPs with the phenotype of GVHD after allogeneic stem cell transplantation. Twenty tagSNPs of the BAFF gene were genotyped in 164 recipient/donor pairs. GVHD after day 100 occurred in 124 (76%) patients: acute GVHD (aGVHD) subtypes (n = 23), overlap GVHD (n = 29), and classic cGVHD (n = 72). In SNP analyses, 9 of the 20 tag SNPs were significant comparing classic/overlap cGVHD versus aGVHD subtypes/no GVHD. In multivariate analyses, 4 recipient BAFF SNPs (rs16972217 [odds ratio = 2.72, P = .004], rs7993590 [odds ratio = 2.35, P = .011], rs12428930 [odds ratio2.53, P = .008], and rs2893321 [odds ratio = 2.48, P = .009]) were independent predictors of GVHD subtypes, adjusted for conventional predictors of cGVHD. This study shows that genetic variation of BAFF modulates GVHD phenotype after allogeneic stem cell transplantation.