Further development of the case-only design for assessing gene-environment interaction: evaluation of and adjustment for bias

Assessing Vulnerability to Surges in Suicide-Related Tweets Using Japan Census Data: Case-Only Study

BACKGROUND

As the use of social media becomes more widespread, its impact on health cannot be ignored. However, limited research has been conducted on the relationship between social media and suicide. Little is known about individuals' vulnerable to suicide, especially when social media suicide information is extremely prevalent.

OBJECTIVE

This study aims to identify the characteristics underlying individuals' vulnerability to suicide brought about by an increase in suicide-related tweets, thereby contributing to public health.

METHODS

A case-only design was used to investigate vulnerability to suicide using individual data of people who died by suicide and tweet data from January 1, 2011, through December 31, 2014. Mortality data were obtained from Japanese government statistics, and tweet data were provided by a commercial service. Tweet data identified the days when suicide-related tweets surged, and the date-keyed merging was performed by considering 3 and 7 lag days. For the merged data set for analysis, the logistic regression model was fitted with one of the personal characteristics of interest as a dependent variable and the dichotomous exposure variable. This analysis was performed to estimate the interaction between the surges in suicide-related tweets and personal characteristics of the suicide victims as case-only odds ratios (ORs) with 95% CIs. For the sensitivity analysis, unexpected deaths other than suicide were considered.

RESULTS

During the study period, there were 159,490 suicides and 115,072 unexpected deaths, and the number of suicide-related tweets was 2,804,999. Following the 3-day lag of a highly tweeted day, there were significant interactions for those who were aged 40 years or younger (OR 1.09, 95% CI 1.03-1.15), male (OR 1.12, 95% CI 1.07-1.18), divorced (OR 1.11, 95% CI 1.03 1.19), unemployed (OR 1.12, 95% CI 1.02-1.22), and living in urban areas (OR 1.26, 95% CI 1.17 1.35). By contrast, widowed individuals had significantly lower interactions (OR 0.83, 95% CI 0.77-0.89). Except for unemployment, significant relationships were also observed for the 7-day lag. For the sensitivity analysis, no significant interactions were observed for other unexpected deaths in the 3-day lag, and only the widowed had a significantly larger interaction than those who were married (OR 1.08, 95% CI 1.02-1.15) in the 7-day lag.

CONCLUSIONS

This study revealed the interactions of personal characteristics associated with susceptibility to suicide-related tweets. In addition, a few significant relationships were observed in the sensitivity analysis, suggesting that such an interaction is specific to suicide deaths. In other words, individuals with these characteristics, such as being young, male, unemployed, and divorced, may be vulnerable to surges in suicide-related tweets. Thus, minimizing public health strain by identifying people who are vulnerable and susceptible to a surge in suicide-related information on the internet is necessary.

Registry Data in Injury Research: Study Designs and Interpretation

Purpose of Review

Injury data is frequently captured in registries that form a census of 100% of known cases that meet specified inclusion criteria. These data are routinely used in injury research with a variety of study designs. We reviewed study designs commonly used with data extracted from injury registries and evaluated the advantages and disadvantages of each design type.

Recent Findings

Registry data are suited to 5 major design types: (1) Description, (2) Ecologic (with Ecologic Cohort as a particularly informative sub-type), (3) Case-control (with location-based and culpability studies as salient subtypes), (4) Case-only (including case-case and case-crossover subtypes), and (5) Outcomes.

Summary

Registries are an important resource for injury research. Investigators considering use of a registry should be aware of the advantages and disadvantages of available study designs.

Causal Inference with Case-Only Studies in Injury Epidemiology Research

Purpose of Review

We review the application and limitations of two implementations of the "case-only design" in injury epidemiology with example analyses of Fatality Analysis Reporting System data.

Recent Findings

The term "case-only design" covers a variety of epidemiologic designs; here, two implementations of the design are reviewed: (1) studies to uncover etiological heterogeneity and (2) studies to measure exposure effect modification. These two designs produce results that require different interpretations and rely upon different assumptions. The key assumption of case-only designs for exposure effect modification, the more commonly used of the two designs, does not commonly hold for injuries and so results from studies using this design cannot be interpreted. Case-only designs to identify etiological heterogeneity in injury risk are interpretable but only when the case-series is conceptualized as arising from an underlying cohort.

Summary

The results of studies using case-only designs are commonly misinterpreted in the injury literature.

Case-only approach applied in environmental epidemiology: 2 examples of interaction effect using the US National Health and Nutrition Examination Survey (NHANES) datasets

INTRODUCTION

By substituting the general 'susceptibility factor' concept for the conventional 'gene' concept in the case-only approach for gene-environment interaction, the case-only approach can also be used in environmental epidemiology. Under the independence between the susceptibility factor and environmental exposure, the case-only approach can provide a more precise estimate of an interaction effect.

METHODS

Two analysis examples of the case-only approach in environmental epidemiology are provided using the 2015-2016 and 2017-2018 US National Health and Nutritional Examination Survey (NHANES): (i) the negative interaction effect between blood chromium level and glycohemoglobin level on albuminuria and (ii) the positive interaction effect between blood cobalt level and old age on albuminuria. The second part of the methods (theoretical backgrounds) summarized the logic and equations provided in previous studies about the case-only approach.

RESULTS

(i) When a 1 μg/L difference of both blood chromium level (mcg/L) and a 1% difference in blood glycohemoglobin level coincide, the multiplicative interaction contrast ratio (ICR_c/nc) was 0.72 (95% CI 0.35-1.60), with no statistical significance. However, when only the cases were analyzed, the case-only ICR (ICR_CO) was 0.59 (95% CI 0.28-0.95), with a statistical significance (a negative interaction effect). (ii) When a 1 μg/L difference of both blood cobalt levels and a 1-year difference in age coincide, the multiplicative interaction contrast ratio (ICR_c/nc) was 1.13 (95% CI 0.99-1.37), with no statistical significance. However, when only the cases were analyzed, the case-only ICR (ICR_CO) was 1.21 (95% CI 1.06-1.51), with a statistical significance (a positive interaction effect).

DISCUSSION

The discussion suggested the theoretical background and previous literature about the possible protective interaction effect between blood chromium levels and blood glycohemoglobin levels on the incidence of albuminuria and the possible aggravating interaction effect between blood cobalt levels and increasing ages on the incidence of albuminuria. If the independence assumption between a susceptibility factor and environmental exposure in a study with cases and non-cases is kept, the case-only approach can provide a more precise interaction effect estimate than conventional approaches with both cases and non-cases.

Black and unarmed: statistical interaction between age, perceived mental illness, and geographic region among males fatally shot by police using case-only design

PURPOSE

We examine whether the race and armed status interact to modify the risk of being fatally shot by police within categories of civilian age and mental illness status, and U.S. region.

METHODS

Data are from The Washington Post online public-use database of all U.S. police-involved shooting deaths. The sample includes black and white males with known armed status who were killed from 1/1/2015 through 12/31/2019 (n = 3090). A case-only design is used to assess multiplicative interaction using adjusted logistic regression.

RESULTS

The fully adjusted interaction estimate is null (SOR = 0.75; 95% confidence interval [CI] = 0.55-1.04). However, adjusted estimates within strata show that the risk of being armed versus unarmed when fatally shot is smaller for black than white males older than 54 years (SOR = 0.18; 95% CI = 0.06-0.65), those showing mental illness signs (SOR = 0.50; 95% CI = 0.26-0.98), and those killed in the South (SOR = 0.52; 95% CI = 0.33-0.83), and that the risk is greater in the Midwest (SOR = 2.42; 95% CI = 1.11-5.26). Notably, there is no black-white difference in armed status among younger age groups (SOR≈0.89).

CONCLUSION

The race and armed status may interact leaving black males at a higher risk of being unarmed than white males when fatally shot by police among those older than 54 years, mentally impaired, and residing in the South. Causal interaction suggests a lower risk for unarmed blacks in the Midwest. Researchers should further explore the utility of the case-only design to study social-environmental interaction.

Novel methods for epistasis detection in genome-wide association studies

More and more genome-wide association studies are being designed to uncover the full genetic basis of common diseases. Nonetheless, the resulting loci are often insufficient to fully recover the observed heritability. Epistasis, or gene-gene interaction, is one of many hypotheses put forward to explain this missing heritability. In the present work, we propose epiGWAS, a new approach for epistasis detection that identifies interactions between a target SNP and the rest of the genome. This contrasts with the classical strategy of epistasis detection through exhaustive pairwise SNP testing. We draw inspiration from causal inference in randomized clinical trials, which allows us to take into account linkage disequilibrium. EpiGWAS encompasses several methods, which we compare to state-of-the-art techniques for epistasis detection on simulated and real data. The promising results demonstrate empirically the benefits of EpiGWAS to identify pairwise interactions.

Use of Spontaneous Reporting Systems to Detect Host-Medication Interactions: Sex Differences in Oral Anti-Diabetic Drug-Associated Myocardial Infarction

Background Medical treatment should be tailored to an individual's characteristics to optimize treatment benefits. We examined whether case-only analyses from spontaneous reporting systems can detect host-medication interactions in oral antidiabetic drug-associated myocardial infarction. Methods and Results Interaction between sex and use of oral antidiabetic drugs was mined among patients with myocardial infarction in the US Food and Drug Administration Adverse Event Reporting System from 2004 to 2014, including 55 718 males and 42 428 females. The odds ratio ( OR ) of multiplicative interactions was used to estimate sex-drug interaction. Detected signs of these interactions were then validated by a nested case-control study utilizing a healthcare record database, Taiwan's National Health Insurance Research Database, from 2001 to 2014, including 31 585 cases and 126 340 controls. In the US Food and Drug Administration Adverse Event Reporting System, a higher proportion of male than female patients used metformin (10.32% in males versus 7.82% in females) and sulfonylureas (4.75% in males versus 3.43% in females); after adjusting for patients' pharmacy-based chronic disease score, males had a higher risk of metformin-associated ( OR =1.07; 99% confidence interval, 1.00-1.14) and sulfonylureas-associated ( OR =1.21; 99% confidence interval, 1.10-1.33) myocardial infarction than females. Detected signs of sex-drug interactions were validated in the National Health Insurance Research Database ( OR for metformin=1.14; 99% confidence interval, 1.03-1.26; OR for sulfonylureas=1.13; 99% confidence interval, 1.02-1.25). Conclusions Males have a higher risk of metformin- and sulfonylureas-associated myocardial infarction than females, which suggests that sex-drug interactions are a key issue in diabetes mellitus treatment plan development. This case-only approach using information from spontaneous reporting systems may be a potential tool for screening host-medication interactions that cause adverse events.

Enriched power of disease-concordant twin-case-only design in detecting interactions in genome-wide association studies

Genetic interaction is a crucial issue in the understanding of functional pathways underlying complex diseases. However, detecting such interaction effects is challenging in terms of both methodology and statistical power. We address this issue by introducing a disease-concordant twin-case-only design, which applies to both monozygotic and dizygotic twins. To investigate the power, we conducted a computer simulation study by setting a series of parameter schemes with different minor allele frequencies and relative risks. Results from the simulation study reveals that the disease-concordant twin-case-only design largely reduces sample size required for sufficient power compared to the ordinary case-only design for detecting gene-gene interaction using unrelated individuals. Sample sizes for dizygotic and monozygotic twins were roughly 1/2 and 1/4 of sample sizes in the ordinary case-only design. Since dizygotic twins are genetically similar as siblings, the enriched power for dizygotic twins also applies to affected siblings, which could help to largely extend the application of the powerful twin-case-only design. In summary, our simulation reveals high value of disease-concordant twins and siblings in efficiently detecting gene-by-gene interactions.

An Analytic Approach Using Candidate Gene Selection and Logic Forest to Identify Gene by Environment Interactions (G × E) for Systemic Lupus Erythematosus in African Americans

Development and progression of many human diseases, such as systemic lupus erythematosus (SLE), are hypothesized to result from interactions between genetic and environmental factors. Current approaches to identify and evaluate interactions are limited, most often focusing on main effects and two-way interactions. While higher order interactions associated with disease are documented, they are difficult to detect since expanding the search space to all possible interactions of p predictors means evaluating 2^p − 1 terms. For example, data with 150 candidate predictors requires considering over 10⁴⁵ main effects and interactions. In this study, we present an analytical approach involving selection of candidate single nucleotide polymorphisms (SNPs) and environmental and/or clinical factors and use of Logic Forest to identify predictors of disease, including higher order interactions, followed by confirmation of the association between those predictors and interactions identified with disease outcome using logistic regression. We applied this approach to a study investigating whether smoking and/or secondhand smoke exposure interacts with candidate SNPs resulting in elevated risk of SLE. The approach identified both genetic and environmental risk factors, with evidence suggesting potential interactions between exposure to secondhand smoke as a child and genetic variation in the ITGAM gene associated with increased risk of SLE.

In the Wrong Place with the Wrong SNP: The Association Between Stressful Neighborhoods and Cardiac Arrest Within Beta-2-adrenergic Receptor Variants

BACKGROUND

Sudden cardiac arrest has been linked independently both to stressful neighborhood conditions and to polymorphisms in the ADRB2 gene. The ADRB2 gene mediates sympathetic activation in response to stress. Therefore, if neighborhood conditions cause cardiac arrest through the stress pathway, the ADRB2 variant may modify the association between neighborhood conditions, such as socioeconomic deprivation and incidence of cardiac arrest.

METHODS

The Cardiac Arrest Blood Study Repository is a population-based repository of specimens and other data from adult cardiac arrest patients residing in King County, Washington. Cases (n = 1,539) were 25- to 100-year-old individuals of European descent who experienced out-of-hospital cardiac arrest from 1988 to 2004. Interactions between neighborhood conditions and the ADRB2 genotype on cardiac arrest risk were assessed in a case-only study design. Gene-environment independence was assessed in blood samples obtained from King County residents initially contacted by random-digit dialing.

RESULTS

Fewer than 4% of study subjects resided in socioeconomically deprived neighborhoods. Nonetheless, the case-only analysis indicated the presence of supramultiplicative interaction of socioeconomic deprivation and the homozygous Gln27Glu variant (case-only odds ratio: 1.8 [95% confidence interval: 1.0, 2.9]). Interactions between population density and the homozygous Gln27Glu variant were weaker (case-only odds ratio: 1.2 [95% confidence interval: 0.97, 1.5]).

CONCLUSIONS

Findings support a supramultiplicative interaction between the Gln27Glu ADRB2 variant and socioeconomic deprivation among individuals of European descent. This result is consistent with the hypothesis that the elevation in cardiac arrest risk associated with socioeconomic deprivation operates through the stress pathway.

The complement system: a gateway to gene-environment interactions in schizophrenia pathogenesis

The pathogenesis of schizophrenia is considered to be multi-factorial, with likely gene-environment interactions (GEI). Genetic and environmental risk factors are being identified with increasing frequency, yet their very number vastly increases the scope of possible GEI, making it difficult to identify them with certainty. Accumulating evidence suggests a dysregulated complement pathway among the pathogenic processes of schizophrenia. The complement pathway mediates innate and acquired immunity, and its activation drives the removal of damaged cells, autoantigens and environmentally derived antigens. Abnormalities in complement functions occur in many infectious and autoimmune disorders that have been linked to schizophrenia. Many older reports indicate altered serum complement activity in schizophrenia, though the data are inconclusive. Compellingly, recent genome-wide association studies suggest repeat polymorphisms incorporating the complement 4A (C4A) and 4B (C4B) genes as risk factors for schizophrenia. The C4A/C4B genetic associations have re-ignited interest not only in inflammation-related models for schizophrenia pathogenesis, but also in neurodevelopmental theories, because rodent models indicate a role for complement proteins in synaptic pruning and neurodevelopment. Thus, the complement system could be used as one of the 'staging posts' for a variety of focused studies of schizophrenia pathogenesis. They include GEI studies of the C4A/C4B repeat polymorphisms in relation to inflammation-related or infectious processes, animal model studies and tests of hypotheses linked to autoimmune diseases that can co-segregate with schizophrenia. If they can be replicated, such studies would vastly improve our understanding of pathogenic processes in schizophrenia through GEI analyses and open new avenues for therapy.

Excess psychosocial burden in women with diabetes and premature acute coronary syndrome

AIM

Diabetes is a stronger risk factor for acute coronary syndrome for women than men. We investigate whether behavioural and psychosocial factors contribute to the disparity in acute coronary syndrome risk and outcomes among women with diabetes relative to women without diabetes and men.

METHODS

Among 939 participants in the GENESIS-PRAXY cohort study of premature acute coronary syndrome (age ≤ 55 years), we compared the prevalence of traditional and non-traditional factors by sex and Type 2 diabetes status. In a case-only analysis, we used generalized logit models to investigate the influence of traditional and non-traditional factors on the interaction of sex and diabetes.

RESULTS

In 287 women (14.3% with diabetes) and 652 men (10.4% with diabetes), women and men with diabetes showed a heavier burden of traditional cardiac risk factors compared with individuals without diabetes. Women with diabetes were more likely to be the primary earner and have more anxiety relative to women without diabetes, and reported worse perceived health compared with women without diabetes and men with diabetes. The interaction term for sex and diabetes (odds ratio (OR) 1.40, 95% confidence intervals (95% CI) 0.83-2.36) was diminished after additional adjustment for non-traditional factors (OR 1.12, 95% CI 0.54-2.32), but not traditional factors alone (OR 1.41, 95% CI 0.84-2.36).

CONCLUSIONS

We observed trends toward a more adverse psychosocial profile among women with diabetes and incident acute coronary syndrome compared with women without diabetes and men with diabetes, which may explain the increased risk of acute coronary syndrome in women with diabetes and may also contribute to worse outcomes.

Genetic Factors Interact With Tobacco Smoke to Modify Risk for Inflammatory Bowel Disease in Humans and Mice

BACKGROUND & AIMS

The role of tobacco smoke in the etiology of inflammatory bowel disease (IBD) is unclear. We investigated interactions between genes and smoking (gene-smoking interactions) that affect risk for Crohn's disease (CD) and ulcerative colitis (UC) in a case-only study of patients and in mouse models of IBD.

METHODS

We used 55 Immunochip-wide datasets that included 19,735 IBD cases (10,856 CD cases and 8879 UC cases) of known smoking status. We performed 3 meta-analyses each for CD, UC, and IBD (CD and UC combined), comparing data for never vs ever smokers, never vs current smokers, and never vs former smokers. We studied the effects of exposure to cigarette smoke in Il10^-/- and Nod2^-/- mice, as well as in Balb/c mice without disruption of these genes (wild-type mice). Mice were exposed to the smoke of 5 cigarettes per day, 5 days a week, for 8 weeks, in a ventilated smoking chamber, or ambient air (controls). Intestines were collected and analyzed histologically and by reverse transcription polymerase chain reaction.

RESULTS

We identified 64 single nucleotide polymorphisms (SNPs) for which the association between the SNP and IBD were modified by smoking behavior (meta-analysis Wald test P < 5.0 × 10^-5; heterogeneity Cochrane Q test P > .05). Twenty of these variants were located within the HLA region at 6p21. Analysis of classical HLA alleles (imputed from SNP genotypes) revealed an interaction with smoking. We replicated the interaction of a variant in NOD2 with current smoking in relation to the risk for CD (frameshift variant fs1007insC; rs5743293). We identified 2 variants in the same genomic region (rs2270368 and rs17221417) that interact with smoking in relation to CD risk. Approximately 45% of the SNPs that interact with smoking were in close vicinity (≤1 Mb) to SNPs previously associated with IBD; many were located near or within genes that regulate mucosal barrier function and immune tolerance. Smoking modified the disease risk of some variants in opposite directions for CD vs UC. Exposure of Interleukin 10 (il10)-deficient mice to cigarette smoke accelerated development of colitis and increased expression of interferon gamma in the small intestine compared to wild-type mice exposed to smoke. NOD2-deficient mice exposed to cigarette smoke developed ileitis, characterized by increased expression of interferon gamma, compared to wild-type mice exposed to smoke.

CONCLUSIONS

In an analysis of 55 Immunochip-wide datasets, we identified 64 SNPs whose association with risk for IBD is modified by tobacco smoking. Gene-smoking interactions were confirmed in mice with disruption of Il10 and Nod2-variants of these genes have been associated with risk for IBD. Our findings from mice and humans revealed that the effects of smoking on risk for IBD depend on genetic variants.

Larger men have larger prostates: Detection bias in epidemiologic studies of obesity and prostate cancer risk

BACKGROUND

Obesity is associated with risk of aggressive prostate cancer (PCa), but not with over-all PCa risk. However, obese men have larger prostates which may lower biopsy accuracy and cause a systematic bias toward the null in epidemiologic studies of over-all risk.

METHODS

Within a cohort of 6692 men followed-up after a biopsy or transurethral resection of the prostate (TURP) with benign findings, a nested case-control study was conducted of 495 prostate cancer cases and controls matched on age, race, follow-up duration, biopsy versus TURP, and procedure date. Data on body mass index and prostate volume at the time of the initial procedure were abstracted from medical records.

RESULTS

Prior to consideration of differences in prostate volume, overweight (OR = 1.41; 95%CI 1.01, 1.97), and obese status (OR = 1.59; 95%CI 1.09, 2.33) at the time of the original benign biopsy or TURP were associated with PCa incidence during follow-up. Prostate volume did not significantly moderate the association between body-size and PCa, however it did act as an inverse confounder; adjustment for prostate volume increased the effect size for overweight by 22% (adjusted OR = 1.52; 95%CI 1.08, 2.14) and for obese status by 23% (adjusted OR = 1.77; 95%CI 1.20, 2.62). Larger prostate volume at the time of the original benign biopsy or TURP was inversely associated with PCa incidence during follow-up (OR = 0.92 per 10 cc difference in volume; 95%CI 0.88, 0.97). In analyses that stratified case-control pairs by tumor aggressiveness of the case, prostate volume acted as an inverse confounder in analyses of non-aggressive PCa but not in analyses of aggressive PCa.

CONCLUSIONS

In studies of obesity and PCa, differences in prostate volume cause a bias toward the null, particularly in analyses of non-aggressive PCa. A pervasive underestimation of the association between obesity and overall PCa risk may exist in the literature.

Genetic endowments, parental resources and adult health: Evidence from the Young Finns Study

This paper uses longitudinal survey data linked to administrative registers to examine socioeconomic gradients in health, particularly whether the effects of genetic endowments interact with the socioeconomic resources of the parental household. We find that genetic risk scores contribute to adult health measured by biomarkers. This result is consistent with the findings from genome-wide association studies. Socioeconomic gradients in health differ based on biomarker and resource measures. Family education is negatively related to obesity and the waist-hip ratio, and family income is negatively related to low-density lipoprotein cholesterol and triglyceride levels. Parental resources do not modify the effects of genetic endowment on adult health. However, there is evidence for gene-family income interactions for triglyceride levels, particularly among women.

Interactions Between Genome-Wide Significant Genetic Variants and Circulating Concentrations of 25-Hydroxyvitamin D in Relation to Prostate Cancer Risk in the National Cancer Institute BPC3

Genome-wide association studies (GWAS) have identified over 100 single nucleotide polymorphisms (SNPs) associated with prostate cancer. However, information on the mechanistic basis for some associations is limited. Recent research has been directed towards the potential association of vitamin D concentrations and prostate cancer, but little is known about whether the aforementioned genetic associations are modified by vitamin D. We investigated the associations of 46 GWAS-identified SNPs, circulating concentrations of 25-hydroxyvitamin D (25(OH)D), and prostate cancer (3,811 cases, 511 of whom died from the disease, compared with 2,980 controls-from 5 cohort studies that recruited participants over several periods beginning in the 1980s). We used logistic regression models with data from the National Cancer Institute Breast and Prostate Cancer Cohort Consortium (BPC3) to evaluate interactions on the multiplicative and additive scales. After allowing for multiple testing, none of the SNPs examined was significantly associated with 25(OH)D concentration, and the SNP-prostate cancer associations did not differ by these concentrations. A statistically significant interaction was observed for each of 2 SNPs in the 8q24 region (rs620861 and rs16902094), 25(OH)D concentration, and fatal prostate cancer on both multiplicative and additive scales (P ≤ 0.001). We did not find strong evidence that associations between GWAS-identified SNPs and prostate cancer are modified by circulating concentrations of 25(OH)D. The intriguing interactions between rs620861 and rs16902094, 25(OH)D concentration, and fatal prostate cancer warrant replication.

Corticosteroid receptor genes and childhood neglect influence susceptibility to crack/cocaine addiction and response to detoxification treatment

The aim of this study was to analyze hypotheses-driven gene-environment and gene-gene interactions in smoked (crack) cocaine addiction by evaluating childhood neglect and polymorphisms in mineralocorticoid and glucocorticoid receptor genes (NR3C2 and NR3C1, respectively). One hundred thirty-nine crack/cocaine-addicted women who completed 3 weeks of follow-up during early abstinence composed our sample. Childhood adversities were assessed using the Childhood Trauma Questionnaire (CTQ), and withdrawal symptoms were assessed using the Cocaine Selective Severity Assessment (CSSA) scale. Conditional logistic regression with counterfactuals and generalized estimating equation modeling were used to test gene-environment and gene-gene interactions. We found an interaction between the rs5522-Val allele and childhood physical neglect, which altered the risk of crack/cocaine addiction (Odds ratio = 4.0, P = 0.001). Moreover, a NR3C2-NR3C1 interaction (P = 0.002) was found modulating the severity of crack/cocaine withdrawal symptoms. In the post hoc analysis, concomitant carriers of the NR3C2 rs5522-Val and NR3C1 rs6198-G alleles showed lower overall severity scores when compared to other genotype groups (P-values ≤ 0.035). This gene-environment interaction is consistent with epidemiological and human experimental findings demonstrating a strong relationship between early life stress and the hypothalamic-pituitary-adrenal (HPA) axis dysregulation in cocaine addiction. Additionally, this study extended in crack/cocaine addiction the findings previously reported for tobacco smoking involving an interaction between NR3C2 and NR3C1 genes.

Allowing for population stratification in case-only studies of gene-environment interaction, using genomic control

Gene-environment interactions (G × E) have attracted considerable research interest in the past owing to their scientific and public health implications, but powerful statistical methods are required to successfully track down G × E, particularly at a genome-wide level. Previously, a case-only (CO) design has been proposed as a means to identify G × E with greater efficiency than traditional case-control or cohort studies. However, as with genotype-phenotype association studies themselves, hidden population stratification (PS) can impact the validity of G × E studies using a CO design. Since this problem has been subject to little research to date, we used comprehensive simulation to systematically assess the type I error rate, power and effect size bias of CO studies of G × E in the presence of PS. Three types of PS were considered, namely genetic-only (PSG), environment-only (PSE), and joint genetic and environmental stratification (PSGE). Our results reveal that the type I error rate of an unadjusted Wald test, appropriate for the CO design, would be close to its nominal level (0.05 in our study) as long as PS involves only one interaction partner (i.e., either PSG or PSE). In contrast, if the study population is stratified with respect to both G and E (i.e., if there is PSGE), then the type I error rate is seriously inflated and estimates of the underlying G × E interaction are biased. Comparison of CO to a family-based case-parents design confirmed that the latter is more robust against PSGE, as expected. However, case-parent trios may be particularly unsuitable for G × E studies in view of the fact that they require genotype data from parents and that many diseases with an environmental component are likely to be of late onset. An alternative approach to adjusting for PS is principal component analysis (PCA), which has been widely used for this very purpose in past genome-wide association studies (GWAS). However, resolving genetic PS properly by PCA requires genetic data at the population level, the availability of which would conflict with the basic idea of the CO design. Therefore, we explored three modified Wald test statistics, inspired by the genomic control (GC) approach to GWAS, as an alternative means to allow for PSGE. The modified statistics were benchmarked against a stratified Wald test assuming known population affiliation, which should provide maximum power under PS. Our results confirm that GC is capable of successfully and efficiently correcting the PS-induced inflation of the type I error rate in CO studies of G × E.

Detecting gene-environment interactions in human birth defects: Study designs and statistical methods

BACKGROUND

The National Birth Defects Prevention Study (NBDPS) contains a wealth of information on affected and unaffected family triads, and thus provides numerous opportunities to study gene-environment interactions (G×E) in the etiology of birth defect outcomes. Depending on the research objective, several analytic options exist to estimate G×E effects that use varying combinations of individuals drawn from available triads.

METHODS

In this study, we discuss important considerations in the collection of genetic data and environmental exposures.

RESULTS

We will also present several population- and family-based approaches that can be applied to data from the NBDPS including case-control, case-only, family-based trio, and maternal versus fetal effects. For each, we describe the data requirements, applicable statistical methods, advantages, and disadvantages.

CONCLUSION

A range of approaches can be used to evaluate potentially important G×E effects in the NBDPS. Investigators should be aware of the limitations inherent to each approach when choosing a study design and interpreting results.

Gene-environment interactions and obesity: recent developments and future directions

Obesity, a major public health concern, is a multifactorial disease caused by both environmental and genetic factors. Although recent genome-wide association studies have identified many loci related to obesity or body mass index, the identified variants explain only a small proportion of the heritability of obesity. Better understanding of the interplay between genetic and environmental factors is the basis for developing effective personalized obesity prevention and management strategies. This article reviews recent advances in identifying gene-environment interactions related to obesity and describes epidemiological designs and newly developed statistical approaches to characterizing and discovering gene-environment interactions on obesity risk.

Genetic epidemiology and insights into interactive genetic and environmental effects in autism spectrum disorders

Understanding the pathogenesis of neurodevelopmental disorders has proven to be challenging. Using autism spectrum disorder (ASD) as a paradigmatic neurodevelopmental disorder, this article reviews the existing literature on the etiological substrates of ASD and explores how genetic epidemiology approaches including gene-environment interactions (G×E) can play a role in identifying factors associated with ASD etiology. New genetic and bioinformatics strategies have yielded important clues to ASD genetic substrates. The next steps for understanding ASD pathogenesis require significant effort to focus on how genes and environment interact with one another in typical development and its perturbations. Along with larger sample sizes, future study designs should include sample ascertainment that is epidemiologic and population-based to capture the entire ASD spectrum with both categorical and dimensional phenotypic characterization; environmental measurements with accuracy, validity, and biomarkers; statistical methods to address population stratification, multiple comparisons, and G×E of rare variants; animal models to test hypotheses; and new methods to broaden the capacity to search for G×E, including genome-wide and environment-wide association studies, precise estimation of heritability using dense genetic markers, and consideration of G×E both as the disease cause and a disease course modifier. Although examination of G×E appears to be a daunting task, tremendous recent progress in gene discovery has opened new horizons for advancing our understanding of the role of G×E in the pathogenesis of ASD and ultimately identifying the causes, treatments, and even preventive measures for ASD and other neurodevelopmental disorders.

Omega-3 fatty acids and the genetic risk of early onset acute coronary syndrome

BACKGROUND AND AIMS

Recent gene-environment interaction studies suggest that diet may influence an individual's genetic predisposition to cardiovascular risk. We evaluated whether omega-3 fatty acid intake may influence the risk for acute coronary syndrome (ACS) conferred by genetic polymorphisms among patients with early onset ACS.

METHODS AND RESULTS

Our population consisted of 705 patients of white European descent enrolled in GENESIS-PRAXY, a multicenter cohort study of patients aged 18-55 years and hospitalized with ACS. We used a case-only design to investigate interactions between the omega-3 index (a validated biomarker of omega-3 fatty acid intake) and 30 single nucleotide polymorphisms (SNPs) robustly associated with ACS. We used logistic regression to assess the interaction between each SNP and the omega-3 index. Interaction was also assessed between the omega-3 index and a genetic risk score generated from the 30 SNPs. All models were adjusted for age and sex. An interaction for increased ACS risk was found between carriers of the chromosome 9p21 variant rs4977574 and low omega-3 index (OR 1.57, 95% CI 1.07-2.32, p = 0.02), but this was not significant after correction for multiple testing. Similar results were obtained in the adjusted model (OR 1.55, 95% CI 1.05-2.29, p = 0.03). We did not observe any interaction between the genetic risk score or any of the other SNPs and the omega-3 index.

CONCLUSION

Our results suggest that omega-3 fatty acid intake may modify the genetic risk conferred by chromosome 9p21 variation in the development of early onset ACS and requires independent replication.

The role of linkage disequilibrium in case-only studies of gene-environment interactions

Gene-environment (G × E) interactions have been invoked to account, at least in part, for the gap between the known heritability of common human diseases and the phenotypic variation hitherto explained by genetic variants. Noteworthy in this context, a case-only (CO) design has been proposed in the past as a means to detect G × E interactions possibly more efficiently than by using classical case-control and cohort designs. So far, however, most CO studies have followed a candidate (or single) gene approach, and the genome-wide utility of the CO design is still more or less unknown. In particular, the way in which linkage disequilibrium (LD) impacts upon the chance to detect G × E interaction through the analysis of proxy markers has not been studied in much detail before. Therefore, we systematically assessed the power to indirectly detect a given G × E interaction through exploiting LD in a CO design. Our simulations revealed a strong relationship between LD and detection power that was subsequently validated in a real colorectal cancer data set.

Application of routine electronic health record databases for pharmacogenetic research

Inter-individual variability in drug responses is a common problem in pharmacotherapy. Several factors (non-genetic and genetic) influence drug responses in patients. When aiming to obtain an optimal benefit-risk ratio of medicines and with the emergence of genotyping technology, pharmacogenetic studies are important for providing recommendations on drug treatments. Advances in electronic healthcare information systems can contribute to increasing the quality and efficiency of such studies. This review describes the definition of pharmacogenetics, gene selection and study design for pharmacogenetic research. It also summarizes the potential of linking pharmacoepidemiology and pharmacogenetics (along with its strengths and limitations) and provides examples of pharmacogenetic studies utilizing electronic health record databases.

The ontology of genetic susceptibility factors (OGSF) and its application in modeling genetic susceptibility to vaccine adverse events

Background

Due to human variations in genetic susceptibility, vaccination often triggers adverse events in a small population of vaccinees. Based on our previous work on ontological modeling of genetic susceptibility to disease, we developed an Ontology of Genetic Susceptibility Factors (OGSF), a biomedical ontology in the domain of genetic susceptibility and genetic susceptibility factors. The OGSF framework was then applied in the area of vaccine adverse events (VAEs).

Results

OGSF aligns with the Basic Formal Ontology (BFO). OGSF defines ‘genetic susceptibility’ as a subclass of BFO:disposition and has a material basis ‘genetic susceptibility factor’. The ‘genetic susceptibility to pathological bodily process’ is a subclasses of ‘genetic susceptibility’. A VAE is a type of pathological bodily process. OGSF represents different types of genetic susceptibility factors including various susceptibility alleles (e.g., SNP and gene). A general OGSF design pattern was developed to represent genetic susceptibility to VAE and associated genetic susceptibility factors using experimental results in genetic association studies. To test and validate the design pattern, two case studies were populated in OGSF. In the first case study, human gene allele DBR*15:01 is susceptible to influenza vaccine Pandemrix-induced Multiple Sclerosis. The second case study reports genetic susceptibility polymorphisms associated with systemic smallpox VAEs. After the data of the Case Study 2 were represented using OGSF-based axioms, SPARQL was successfully developed to retrieve the susceptibility factors stored in the populated OGSF. A network of data from the Case Study 2 was constructed by using ontology terms and individuals as nodes and ontology relations as edges. Different social network analys is (SNA) methods were then applied to verify core OGSF terms. Interestingly, a SNA hub analysis verified all susceptibility alleles of SNPs and a SNA closeness analysis verified the susceptibility genes in Case Study 2. These results validated the proper OGSF structure identified different ontology aspects with SNA methods.

Conclusions

OGSF provides a verified and robust framework for representing various genetic susceptibility types and genetic susceptibility factors annotated from experimental VAE genetic association studies. The RDF/OWL formulated ontology data can be queried using SPARQL and analyzed using centrality-based network analysis methods.

Gene-environment interactions and the risk of childhood acute lymphoblastic leukemia: exploring the role of maternal folate genes and folic Acid fortification

Few studies have evaluated the interaction of folic acid fortification and folate metabolic genes on the risk of childhood acute lymphoblastic leukemia (ALL). Because folate status is influenced by both intake and genetic variation, the objective of this study was to explore maternal folate metabolic gene-folic acid fortification interactions and the risk of childhood ALL. The study population consisted of 120 ALL case-parent triads recruited from Texas Children's Cancer Center between 2003 and 2010. For this analysis, we focused on 13 maternal single nucleotide polymorphisms (SNPs) in 5-methyltetrahydrofolate-homocysteine methyltransferase (MTR). Prefortification was defined as delivery before January 1997 and postfortification as delivery in or after January 1997. We used a two-step approach to evaluate gene-environment interactions. First, a case-only approach was used, as this design provides greater power in the assessment of gene-environment interactions compared to other approaches. Second, we confirmed all statistically significant interactions using a log-linear approach among case-parent triads. Only one of 13 interactions evaluated was confirmed in step 2. Specifically, mothers with the minor allele of MTR rs1804742 and who delivered during the prefortification period were at a greater risk of having a child with ALL (OR = 1.54, 95% CI: 0.82-2.88), compared to those mothers who delivered during the postfortification period (OR = 0.81, 95% CI: 0.22-2.99, P for interaction = .03). In one of the few studies to evaluate maternal folate metabolic genotype-folic acid interactions, we found limited evidence that the maternal MTR rs1804742 appeared to interact with higher folic acid levels to influence childhood ALL risk.

Maternal-fetal metabolic gene-gene interactions and risk of neural tube defects

Single-gene analyses indicate that maternal genes associated with metabolic conditions (e.g., obesity) may influence the risk of neural tube defects (NTDs). However, to our knowledge, there have been no assessments of maternal-fetal metabolic gene-gene interactions and NTDs. We investigated 23 single nucleotide polymorphisms among 7 maternal metabolic genes (ADRB3, ENPP1, FTO, LEP, PPARG, PPARGC1A, and TCF7L2) and 2 fetal metabolic genes (SLC2A2 and UCP2). Samples were obtained from 737 NTD case-parent triads included in the National Birth Defects Prevention Study for birth years 1999-2007. We used a 2-step approach to evaluate maternal-fetal gene-gene interactions. First, a case-only approach was applied to screen all potential maternal and fetal interactions (n = 76), as this design provides greater power in the assessment of gene-gene interactions compared to other approaches. Specifically, ordinal logistic regression was used to calculate the odds ratio (OR) and 95% confidence interval (CI) for each maternal-fetal gene-gene interaction, assuming a log-additive model of inheritance. Due to the number of comparisons, we calculated a corrected p-value (q-value) using the false discovery rate. Second, we confirmed all statistically significant interactions (q < 0.05) using a log-linear approach among case-parent triads. In step 1, there were 5 maternal-fetal gene-gene interactions with q < 0.05. The "top hit" was an interaction between maternal ENPP1 rs1044498 and fetal SLC2A2 rs6785233 (interaction OR = 3.65, 95% CI: 2.32-5.74, p = 2.09×10(-8), q=0.001), which was confirmed in step 2 (p = 0.00004). Our findings suggest that maternal metabolic genes associated with hyperglycemia and insulin resistance and fetal metabolic genes involved in glucose homeostasis may interact to increase the risk of NTDs.

Meta-analyses to investigate gene-environment interactions in neuroepidemiology

BACKGROUND

Most chronic neurological diseases are caused by a combination of multiple genetic and environmental factors. Increasingly, gene-environment interactions (GxE) are being examined, providing opportunities to combine studies systematically using meta-analysis.

METHODS

Systematic review of the literature on how to examine GxE using observational study designs, and how to conduct a meta-analysis of studies on GxE.

RESULTS

Most methods and challenges related to a standard meta-analysis apply to a GxE meta-analysis. There are, however, some substantive differences. With GxE, there is the capability of using a case-only design. Research on GxE interactions may be more prone to publication bias, since interactions are usually not the primary hypothesis and only 'exciting' significant GxE findings are reported out of a range of secondary analyses. In disease aetiology research, there has been debate whether to measure interaction on a multiplicative or additive scale. There are some significant challenges associated with measuring interaction on an additive scale, and thus the uptake of the measures of additive interaction has been limited. As a result, the methods of analysing interaction have been less consistent and reporting has been highly variable. We suggest using the STROBE/STREGA reporting guidelines to allow evaluation of interaction on both scales.

CONCLUSIONS

We identified a number of differences of a GxE meta-analysis over a standard meta-analysis. Awareness of these issues is important. Using established reporting guidelines for GxE studies is recommended. The development of consortia for neurological disorders that include both genetic and environmental data might offer benefits for GxE meta-analyses in the future.

Gene-environment correlation: difficulties and a natural experiment-based strategy

OBJECTIVES

We explored how gene-environment correlations can result in endogenous models, how natural experiments can protect against this threat, and if unbiased estimates from natural experiments are generalizable to other contexts.

METHODS

We compared a natural experiment, the College Roommate Study, which measured genes and behaviors of college students and their randomly assigned roommates in a southern public university, with observational data from the National Longitudinal Study of Adolescent Health in 2008. We predicted exposure to exercising peers using genetic markers and estimated environmental effects on alcohol consumption. A mixed-linear model estimated an alcohol consumption variance that was attributable to genetic markers and across peer environments.

RESULTS

Peer exercise environment was associated with respondent genotype in observational data, but not in the natural experiment. The effects of peer drinking and presence of a general gene-environment interaction were similar between data sets.

CONCLUSIONS

Natural experiments, like random roommate assignment, could protect against potential bias introduced by gene-environment correlations. When combined with representative observational data, unbiased and generalizable causal effects could be estimated.

An exploratory case-only analysis of gene-hazardous air pollutant interactions and the risk of childhood medulloblastoma

BACKGROUND

There is evidence that exposure to chlorinated solvents may be associated with childhood medulloblastoma and primitive neuroectodermal tumor (M/PNET) risk. Animal models suggest genes related to detoxification and DNA repair are important in the carcinogenicity of these pollutants; however, there have been no human studies assessing the modifying effects of these genotypes on the association between chlorinated solvents and childhood M/PNET risk.

PROCEDURE

We conducted a case-only study to evaluate census tract-level exposure to chlorinated solvents and the risk of childhood M/PNET in the context of detoxification and DNA repair genotypes. Cases (n = 98) were obtained from Texas Children's Hospital and MD Anderson Cancer Center. Key genotypes (n = 22) were selected from the Illumina Human 1M Quad SNP Chip. Exposure to chlorinated solvents (methylene chloride, perchloroethylene, trichloroethylene, and vinyl chloride) was estimated from the US EPA's 1999 Assessment System for Population Exposure Nationwide (ASPEN). Logistic regression was used to estimate the case-only odds ratios and 95% confidence intervals (CIs).

RESULTS

There were 11 significant gene-environment interactions associated with childhood M/PNET risk. However, after correcting for multiple comparisons, only the interaction between high trichloroethylene levels and OGG1 rs293795 significantly increased the risk of childhood M/PNET risk (OR = 9.24, 95% CI: 2.24, 38.24, Q = 0.04).

CONCLUSIONS

This study provides an initial assessment of the interaction between ambient levels of chlorinated solvents and potentially relevant genotypes on childhood M/PNET risk. Our results are exploratory and must be validated in animal models, as well as additional human studies.

Recommendations and proposed guidelines for assessing the cumulative evidence on joint effects of genes and environments on cancer occurrence in humans

We propose guidelines to evaluate the cumulative evidence of gene-environment (G × E) interactions in the causation of human cancer. Our approach has its roots in the HuGENet and IARC Monographs evaluation processes for genetic and environmental risk factors, respectively, and can be applied to common chronic diseases other than cancer. We first review issues of definitions of G × E interactions, discovery and modelling methods for G × E interactions, and issues in systematic reviews of evidence for G × E interactions, since these form the foundation for appraising the credibility of evidence in this contentious field. We then propose guidelines that include four steps: (i) score the strength of the evidence for main effects of the (a) environmental exposure and (b) genetic variant; (ii) establish a prior score category and decide on the pattern of interaction to be expected; (iii) score the strength of the evidence for interaction between the environmental exposure and the genetic variant; and (iv) examine the overall plausibility of interaction by combining the prior score and the strength of the evidence and interpret results. We finally apply the scheme to the interaction between NAT2 polymorphism and tobacco smoking in determining bladder cancer risk.

Breast cancer risk in relation to alcohol consumption and BRCA gene mutations--a case-only study of gene-environment interaction

The variable penetrance of the BRCA1 and BRCA2 genes suggests that other genetic or environmental factors may interact with these mutations to modify breast cancer risk. The objective of this study was to measure departures from multiplicative effects of alcohol consumption and BRCA gene mutations. A cohort of French-Canadian breast cancer patients was tested for BRCA gene mutations and completed a food frequency questionnaire. The case-only odds ratio (COR) was calculated. A total of 857 women, including 10 BRCA1 and 33 BRCA2 mutation carriers, participated in the study. No significant interaction between alcohol consumption and BRCA1 mutations was detected, although the interaction with wine consumption suggested a sub-multiplicative effect (COR = 0.38, 95% CI: 0.08-1.81). Consumption of alcohol other than wine interacted significantly with BRCA2 mutations (COR = 2.15, 95% CI: 1.03-4.49). Consumption of wine may protect against BRCA1-associated tumors, while women with BRCA2 mutations may be at greater risk of alcohol-induced breast cancer.

Bias in the case-only design applied to studies of gene-environment and gene-gene interaction: a systematic review and meta-analysis

BACKGROUND

The case-only study, proposed as a design specifically for assessing departure from multiplicative gene-environment and gene-gene interactions, is of considerable potential value but there are concerns about its validity. The objective of this study was to evaluate the extent and sources of bias in the case-only design by means of a systematic review and meta-regression analysis.

METHODS

The MEDLINE, CINAHL, EMBASE and PUBMED databases were searched through to 7 October 2009. Studies that assessed bias in the case-only design applied to the study of gene-environment and gene-gene interaction were identified. Qualitative comments on the sources and extent of bias were extracted. A meta-regression analysis of the ratio (IOR(CC)/IOR(CO)) of the case-control (IOR(CC)) and case-only (IOR(CO)) interaction odds ratios was conducted based on studies in which both methods were applied to the same data set.

RESULTS

The search yielded 365 unique articles of which 38 met the inclusion criteria. Potential sources of bias in the case-only design included non-independence of genotype and exposure in the source population. Meta-regression analysis, based on 24 evaluations, produced a mean IOR(CC)/IOR(CO) of 1.06 [95% confidence interval (95% CI) 0.93-1.22], suggesting that bias in case-only designs is not common in practice. The I(2) statistic indicated that 23.9% (95% uncertainty interval 0-53.9%) of the observed variation was due to heterogeneity between studies, which was not explained by any methodological characteristics of the included studies.

CONCLUSION

As understanding of the relationships between genes and environmental exposures in the population improves, the case-only design may prove to be of considerable value.

The semiparametric case-only estimator

We propose a semiparametric case-only estimator of multiplicative gene-environment or gene-gene interactions, under the assumption of conditional independence of the two factors given a vector of potential confounding variables. Our estimator yields valid inferences on the interaction function if either but not necessarily both of two unknown baseline functions of the confounders is correctly modeled. Furthermore, when both models are correct, our estimator has the smallest possible asymptotic variance for estimating the interaction parameter in a semiparametric model that assumes that at least one but not necessarily both baseline models are correct.

An easy-to-implement approach for analyzing case-control and case-only studies assuming gene-environment independence and Hardy-Weinberg equilibrium

The case-control study is a simple and an useful method to characterize the effect of a gene, the effect of an exposure, as well as the interaction between the two. The control-free case-only study is yet an even simpler design, if interest is centered on gene-environment interaction only. It requires the sometimes plausible assumption that the gene under study is independent of exposures among the non-diseased in the study populations. The Hardy-Weinberg equilibrium is also sometimes reasonable to assume. This paper presents an easy-to-implement approach for analyzing case-control and case-only studies under the above dual assumptions. The proposed approach, the 'conditional logistic regression with counterfactuals', offers the flexibility for complex modeling yet remains well within the reach to the practicing epidemiologists. When the dual assumptions are met, the conditional logistic regression with counterfactuals is unbiased and has the correct type I error rates. It also results in smaller variances and achieves higher powers as compared with using the conventional analysis (unconditional logistic regression).

Sample size requirements to detect gene-environment interactions in genome-wide association studies

Many complex diseases are likely to be a result of the interplay of genes and environmental exposures. The standard analysis in a genome-wide association study (GWAS) scans for main effects and ignores the potentially useful information in the available exposure data. Two recently proposed methods that exploit environmental exposure information involve a two-step analysis aimed at prioritizing the large number of SNPs tested to highlight those most likely to be involved in a GE interaction. For example, Murcray et al. ([2009] Am J Epidemiol 169:219–226) proposed screening on a test that models the G-E association induced by an interaction in the combined case-control sample. Alternatively, Kooperberg and LeBlanc ([2008] Genet Epidemiol 32:255–263) suggested screening on genetic marginal effects. In both methods, SNPs that pass the respective screening step at a pre-specified significance threshold are followed up with a formal test of interaction in the second step. We propose a hybrid method that combines these two screening approaches by allocating a proportion of the overall genomewide significance level to each test. We show that the Murcray et al. approach is often the most efficient method, but that the hybrid approach is a powerful and robust method for nearly any underlying model. As an example, for a GWAS of 1 million markers including a single true disease SNP with minor allele frequency of 0.15, and a binary exposure with prevalence 0.3, the Murcray, Kooperberg and hybrid methods are 1.90, 1.27, and 1.87 times as efficient, respectively, as the traditional case-control analysis to detect an interaction effect size of 2.0.

Epistasis between neurochemical gene polymorphisms and risk for ADHD

A number of genes with function related to synaptic neurochemistry have been genetically associated with attention deficit/hyperactivity disorder. However, susceptibility to the development of common psychiatric disorders by single variants acting alone, can so far only explain a small proportion of the heritability of the phenotype. It has been postulated that the unexplained 'dark heritability' may at least in part be due to epistatic effects, which may account for the small observed marginal associations, and the difficulties with replication of positive findings. We undertook a comprehensive exploration of pair-wise interactions between genetic variants in 24 candidate genic regions involved in monoaminergic catabolism, anabolism, release, re-uptake and signal transmission in a sample of 177 parent-affected child trios using a case-only design and a case-pseudocontrol design using conditional logistic regression. Marker-pairs thresholded on interaction odds ratio (OR) and P-value are presented. We detected a number of interaction ORs >4.0, including an interesting correlation between markers in the ADRA1B and DBH genes in affected individuals, and several further interesting but smaller effects. These effects are no larger than you would expect by chance under the assumption of independence of all pair-wise relations; however, independence is unlikely. Furthermore, the size of these effects is of interest and attempts to replicate these results in other samples are anticipated.

Genetic association and gene-environment interaction: a new method for overcoming the lack of exposure information in controls

The use of a reference control panel in genome-wide association studies is an interesting solution to the problem of how to reduce costs. In such designs, data on relevant environmental factors are usually collected only in cases, making it more difficult to deal with potential gene-environment interactions when testing for genetic association. However, under certain circumstances, neglecting an existing interaction with the environment may be detrimental in terms of statistical power to detect the genetic factor. In this paper, the authors propose a novel method based on a multinomial logistic regression model to overcome the lack of environmental exposure information in controls, by contrasting both exposed and unexposed cases with the control sample. For each case group, a genetic effect-size parameter is estimated, and the genetic association and the gene-environment interaction are tested jointly. The authors evaluate the performance of this method through asymptotic computations and simulations of cases and population controls under different models. In the presence of a gene-environment interaction, this approach outperforms other available methods that test for genetic association and gene-environment interaction either separately or jointly. Interestingly, it even has better power than the joint test requiring full knowledge of the environmental information in both cases and controls.

Case-only gene-environment interaction studies: when does association imply mechanistic interaction?

Case-only studies are often used to identify interactions between a genetic factor and an environmental factor under the assumption both factors are independent in the population. However, interpreting a statistical association between the genetic and the environmental factors among the cases, as evidence of a mechanistic gene-environment interaction, is not always warranted. Using a mechanistic approach based on the sufficient cause framework, we show association amongst cases can arise between the genetic and environmental factors when there is in fact no mechanistic gene-environment interaction. However, when it can be assumed the genetic and environmental factors themselves can never prevent the outcome, we show a positive association amongst cases implies a mechanistic gene-environment interaction. Without this assumption that the effects of the two factors are never preventive, a multiplicative interaction greater than two is needed to conclude the presence of a mechanistic interaction. We furthermore show these tests for mechanistic interaction can be extended to scenarios in which the genetic and environmental factors are negatively associated in the population rather than independent.

Genetic polymorphisms and the cardiovascular risk of non-steroidal anti-inflammatory drugs

The cardiovascular safety of cyclooxygenase-2-selective (coxibs) and nonselective nonsteroidal anti-inflammatory drugs (NSAIDs) is of concern, although most users remain free of adverse outcomes. A gene-drug interaction could modulate this cardiovascular risk through prostaglandin synthesis or inflammatory pathways. From an existing acute coronary syndrome cohort (Recurrence and Inflammation in the Acute Coronary Syndromes Study) (n = 1,210), a case-only study was performed by identifying 115 patients exposed to NSAIDs (rofecoxib [n = 43], celecoxib [n = 49], or nonselective NSAIDs [n = 23]) and 345 unexposed patients matched for age, gender, and hospital center. These patients were genotyped for 115 candidate single-nucleotide polymorphisms (SNPs). Statistically significant associations between NSAID exposure and 9 SNPs in 6 genes were observed. Analyzing patients exposed only to coxibs and their matched unexposed cases, significant associations remained for 5 SNPs at 4 loci (prostaglandin-endoperoxide synthase-1 [PTGS1], chromosome 9p21.3, C-reactive protein [CRP], and klotho [KL]). Two independent SNPs from the PTGS1 gene gave similar results under a recessive model, with odds ratios for the association with NSAID exposure of 6.94 (95% confidence interval 1.35 to 35.65, p = 0.016) and 7.11 (95% confidence interval 1.38 to 36.74, p = 0.033). A significant association was also observed for a SNP in the CRP gene (rs1205) (additive odds ratio 1.64, 95% confidence interval 1.18 to 2.27, p = 0.003). In conclusion, these findings suggest that genetic variability may contribute to the susceptibility for acute coronary syndromes observed in some NSAID users. In particular, genetic polymorphisms in the PTGS1 and CRP genes appear to be candidates for a possible gene-drug interaction influencing the acute coronary risk associated with NSAID use, but these findings will require confirmation in larger cohorts.

Case-only genome-wide interaction study of disease risk, prognosis and treatment

Case-control genome-wide association (GWA) studies have facilitated the identification of susceptibility loci for many complex diseases; however, these studies are often not adequately powered to detect gene-environment (G x E) and gene-gene (G x G) interactions. Case-only studies are more efficient than case-control studies for detecting interactions and require no data on control subjects. In this article, we discuss the concept and utility of the case-only genome-wide interaction (COGWI) study, in which common genetic variants, measured genome-wide, are screened for association with environmental exposures or genetic variants of interest. An observed G-E (or G-G) association, as measured by the case-only odds ratio (OR), suggests interaction, but only if the interacting factors are unassociated in the population from which the cases were drawn. The case-only OR is equivalent to the interaction risk ratio. In addition to risk-related interactions, we discuss how the COGWI design can be used to efficiently detect G x G, G x E and pharmacogenetic interactions related to disease outcomes in the context of observational clinical studies or randomized clinical trials. Such studies can be conducted using only data on individuals experiencing an outcome of interest or individuals not experiencing the outcome of interest. Sharing data among GWA and COGWI studies of disease risk and outcome can further enhance efficiency. Sample size requirements for COGWI studies, as compared to case-control GWA studies, are provided. In the current era of genome-wide analyses, the COGWI design is an efficient and straightforward method for detecting G x G, G x E and pharmacogenetic interactions related to disease risk, prognosis and treatment response.

Two-stage analysis for gene-environment interaction utilizing both case-only and family-based analysis

The case-only study and family-based study are two popular study designs for detecting gene-environment interactions. It is well known that the case-only analysis is efficient, but its validity relies crucially on the assumption of gene-environment independence in the study population. In contrast, the family-based analysis is robust to the violation of such an assumption, but is less efficient. We propose a two-stage study design for detecting gene-environment interactions, where a case-only study is performed at the first stage, and a case-parent/case-sibling study is performed at the second stage on a random subsample of the first-stage case sample as well as their parents/unaffected siblings. Statistical inference procedures are developed for the proposed two-stage study designs, which not only preserve the robustness property of the family-based analysis, but also utilize information from the case-only analysis to enhance estimation efficiency and testing power. Simulation results reveal both the robustness and efficiency of the proposed strategies.

Safety of short-term discontinuation of antiplatelet therapy in patients with drug-eluting stents

BACKGROUND

Antiplatelet therapy is often discontinued in patients with drug-eluting stents who are undergoing surgical procedures. However, discontinuation of antiplatelet therapy is an important risk factor for late stent thrombosis. Our objective was to examine the safety of short-term discontinuation of antiplatelet therapy.

METHODS AND RESULTS

We systematically searched Medline for reported cases of late stent thrombosis and very late stent thrombosis published between January 2001 and July 2008. We restricted our search to Academic Research Consortium-defined definite cases. We identified 161 cases of late stent thrombosis or very late stent thrombosis from 84 articles (79 from case reports, 61 from registries, and 21 from randomized clinical trials). Patients had a mean age of 58.4+/-13.4 years, and 88% were male. A total of 19 cases occurred in patients who were receiving dual antiplatelet therapy at the time of the event. If patients stopped both antiplatelet agents simultaneously, the median time to event was 7 days. If patients had previously stopped a thienopyridine with no ill effect and subsequently stopped acetylsalicylic acid, the median time to event was also 7 days from the time of acetylsalicylic acid cessation. If the thienopyridine was stopped but acetylsalicylic acid was maintained, the median time to event was 122 days. Among the 48 patients who stopped both agents, 36 cases (75%) occurred within 10 days. Among the 94 patients who discontinued a thienopyridine but continued acetylsalicylic acid, only 6 cases (6%) occurred within 10 days.

CONCLUSIONS

If acetylsalicylic acid therapy is maintained, short-term discontinuation of a thienopyridine may be relatively safe in patients with drug-eluting stents.

Detecting gene-environment interactions using a combined case-only and case-control approach

The conventional method of detecting gene-environment interactions, the case-control analysis, suffers from low statistical power. In contrast, the case-only analysis/design can be powerful in certain scenarios, although violation of the assumption of independence between the genetic and environmental factors can greatly bias the results. As an alternative, Bayes model averaging may be used to combine the case-control and case-only analyses. This approach first frames the case-control and case-only analyses as variations of a log-linear model. The weighting between these 2 models is then a function of the data and prior beliefs on the independence of the 2 potentially interacting factors. In this paper, the authors demonstrate via simulations that when there is no prior information on the independence of the genetic and environmental factors, this approach tends to be more powerful than the case-control analysis. Additionally, when the genetic and environmental factors are not independent in the population, bias is substantially reduced, with a corresponding reduction in type I error in comparison with the case-only analysis. Increased power or increased robustness to violations of the independence assumption may be obtained with more appropriate prior specification. The authors use an example data analysis to demonstrate the advantages of this approach.

Population stratification bias in the case-only study for gene-environment interactions

The case-only study is a convenient approach and provides increased statistical efficiency in detecting gene-environment interactions. The validity of a case-only study hinges on one well-recognized assumption: The susceptibility genotypes and the environmental exposures of interest are independent in the population. Otherwise, the study will be biased. The authors show that hidden stratification in the study population could also ruin a case-only study. They derive the formulas for population stratification bias. The bias involves three terms: 1) the coefficient of variation of the exposure prevalence odds, 2) the coefficient of variation of the genotype frequency odds, and 3) the correlation coefficient between the exposure prevalence odds and the genotype frequency odds. The authors perform simulation to investigate the magnitude of bias over a wide range of realistic scenarios. It is found that the estimated interaction effect is frequently biased by more than 5%. For a rarer gene and a rarer exposure, the bias becomes even larger (>30%). Because of the potentially large bias, researchers conducting case-only studies should use the boundary formula presented in this paper to make more prudent interpretations of their results, or they should use stratified analysis or a modeling approach to adjust for population stratification bias in their studies.

Gene-environment interactions for complex traits: definitions, methodological requirements and challenges

Genetic and environmental risk factors and their interactions contribute to the development of complex diseases. In this review, we discuss methodological issues involved in investigating gene-environment (G x E) interactions in genetic-epidemiological studies of complex diseases and their potential relevance for clinical application. Although there are some important examples of interactions and applications, the widespread use of the knowledge about G x E interaction for targeted intervention or personalized treatment (pharmacogenetics) is still beyond current means. This is due to the fact that convincing evidence and high predictive or discriminative power are necessary conditions for usefulness in clinical practice. We attempt to clarify conceptual differences of the term 'interaction' in the statistical and biological sciences, since precise definitions are important for the interpretation of results. We argue that the investigation of G x E interactions is more rewarding for the detailed characterization of identified disease genes (ie at advanced stages of genetic research) and the stratified analysis of environmental effects by genotype or vice versa. Advantages and disadvantages of different epidemiological study designs are given and sample size requirements are exemplified. These issues as well as a critical appraisal of common methodological concerns are finally discussed.

Genes and Environmental Exposures in Veterans with Amyotrophic Lateral Sclerosis: the GENEVA study. Rationale, study design and demographic characteristics

Recent reports of a potentially increased risk of amyotrophic lateral sclerosis (ALS) for veterans deployed to the 1990-1991 Persian Gulf War prompted the Department of Veterans Affairs to establish a National Registry of Veterans with ALS, charged with the goal of enrolling all US veterans with a neurologist-confirmed diagnosis of ALS. The Genes and Environmental Exposures in Veterans with ALS study (GENEVA) is a case-control study presently enrolling cases from the Department of Veterans Affairs registry and a representative sample of veteran controls to evaluate the joint contributions of genetic susceptibility and environmental exposures to the risk of sporadic ALS. The GENEVA study design, recruitment strategies, methods of collecting DNA samples and environmental risk factor information are described here, along with a summary of demographic characteristics of the participants (537 cases, 292 controls) enrolled to date.

Comprehensive evaluation of positional candidates in the IL-18 pathway reveals suggestive associations with schizophrenia and herpes virus seropositivity

Interactions between genetic variation and environmental factors have been invoked in schizophrenia genesis, but pathways linking them are uncertain. We used a pathway-oriented approach to evaluate six genes mediating IL18 function (IL-18, IL18BP, IL18R1, IL18RAP, IL12B, and IL12A). The first five are also localized to regions previously linked with schizophrenia. Fifty-four representative tag SNPs were selected from comprehensive sequence data and genotyped in 478 patients with schizophrenia/schizoaffective disorder (DSM IV criteria) and 501 unscreened control individuals. Exposure to three herpes viruses previously suggested as risk factors for schizophrenia was estimated simultaneously among the cases. Five SNPs in four genes were associated with schizophrenia, most prominently rs2272127 at IL18RAP (P = 0.0007, odds ratio for C allele 1.49, 95% CI: 1.18-1.87; P = 0.03 following correction for multiple comparisons). Exploratory analysis revealed that rs2272127 was also associated with herpes simplex virus 1 (HSV1) seropositivity in cases (P = 0.04, OR for G allele 1.58, 95% CI: 1.04-2.39). Similar patterns were observed at another correlated SNP (rs11465702, P = 0.005 and 0.006, respectively for associations with schizophrenia and HSV1 seropositivity). We suggest plausible, testable hypotheses linking IL-18 signaling and HSV1 in schizophrenia pathogenesis.

Polymorphisms in genes related to activation or detoxification of carcinogens might interact with smoking to increase renal cancer risk: results from The Netherlands Cohort Study on diet and cancer

Metabolic gene polymorphisms have previously been suggested as risk factors for renal cell carcinoma (RCC). These polymorphisms are involved in activation or detoxification of carcinogens in cigarette smoke which is another RCC risk factor. We evaluated gene-environment interactions between CYP1A1, GSTmicro1 and smoking in a large population-based RCC case group. The Netherlands Cohort Study on diet and cancer (NLCS) comprises 120,852 persons who completed a questionnaire on smoking and other risk factors at baseline. After 11.3 years of follow-up, 337 incident RCC cases were identified. DNA was collected for 245 cases. In a case-only analysis, interaction-odds ratios (OR) and 95% confidence intervals (95% CI) were calculated using logistic regression. We observed a moderate, not statistically significant, interaction between current smoking and CYP1A1*2C (OR 1.42; 95% CI 0.70-2.89) and GSTmicro1 null (OR 1.35; 95% CI 0.65-2.79). For current smokers with both a variant (heterozygous or homozygous) in CYP1A1 and GSTmicro1 null, risk was also increased (OR 1.63; 95% CI 0.63-4.24). No interaction was observed between ever smokers, smoking duration (increments of 10 smoking years) or amount (increments of 5 cigarettes/day) and CYP1A or GSTmicro1. Our results show a modest trend towards a statistically significant gene-environment interaction between CYP1A1, GSTmicro1 and smoking in RCC. This could indicate that RCC risk among smokers might be more increased with the CYP1A1*2C genotype, GSTmicro1 null, or both a CYP1A1 variant and GSTmicro1 null.

Breast cancer in first-degree relatives and risk of lung cancer: assessment of the existence of gene sex interactions

BACKGROUND

Previous studies have shown the sex differences in lung cancer and the associations between estrogen-related genes and non-small cell lung cancer. In the present study, we assumed the existence of shared candidate genes that are common in lung and breast cancers, and examined whether women with a family history of breast cancer are at increased risk of lung cancer compared with men, especially adenocarcinoma, in a case-only study.

METHODS

This case-only study was conducted based on the Lung Cancer Database Project at the National Cancer Center Hospital East. A total of 1566 patients with newly diagnosed primary lung cancer were consecutively recruited between 1999 and 2003. Information on their family history of cancer and smoking habit was obtained from a self-administered questionnaire. To assess an interactions between two factors, odds ratios for interaction (ORis) and 95% confidence intervals (CIs) were calculated by case-only contingency table.

RESULTS

A statistically significant ORi was observed between a family history of breast cancer in first-degree relatives (parent and siblings, not including children) and the sex of a patient (ORi: 2.22, 95% CI: 1.02-4.81). A stratified analysis by histologic subtypes showed a statistically significant ORi only for adenocarcinoma (ORi: 3.27, 95% CI: 1.19-8.98). No other family history of cancer, such as stomach, colon and lung cancer, showed a statistically significant ORi.

CONCLUSION

This study suggests the possibility of gene-sex interaction in lung cancer.