Causation and familial confounding as explanations for the associations of polygenic risk scores with breast cancer: Evidence from innovative ICE FALCON and ICE CRISTAL analyses.

A polygenic risk score (PRS) combines the associations of multiple genetic variants that could be due to direct causal effects, indirect genetic effects, or other sources of familial confounding. We have developed new approaches to assess evidence for and against causation by using family data for pairs of relatives (Inference about Causation from Examination of FAmiliaL CONfounding [ICE FALCON]) or measures of family history (Inference about Causation from Examining Changes in Regression coefficients and Innovative STatistical AnaLyses [ICE CRISTAL]). Inference is made from the changes in regression coefficients of relatives' PRSs or PRS and family history before and after adjusting for each other. We applied these approaches to two breast cancer PRSs and multiple studies and found that (a) for breast cancer diagnosed at a young age, for example, <50 years, there was no evidence that the PRSs were causal, while (b) for breast cancer diagnosed at later ages, there was consistent evidence for causation explaining increasing amounts of the PRS-disease association. The genetic variants in the PRS might be in linkage disequilibrium with truly causal variants and not causal themselves. These PRSs cause minimal heritability of breast cancer at younger ages. There is also evidence for nongenetic factors shared by first-degree relatives that explain breast cancer familial aggregation. Familial associations are not necessarily due to genes, and genetic associations are not necessarily causal.

Breast and bowel cancers diagnosed in people 'too young to have cancer': A blueprint for research using family and twin studies.

Young breast and bowel cancers (e.g., those diagnosed before age 40 or 50 years) have far greater morbidity and mortality in terms of years of life lost, and are increasing in incidence, but have been less studied. For breast and bowel cancers, the familial relative risks, and therefore the familial variances in age-specific log(incidence), are much greater at younger ages, but little of these familial variances has been explained. Studies of families and twins can address questions not easily answered by studies of unrelated individuals alone. We describe existing and emerging family and twin data that can provide special opportunities for discovery. We present designs and statistical analyses, including novel ideas such as the VALID (Variance in Age-specific Log Incidence Decomposition) model for causes of variation in risk, the DEPTH (DEPendency of association on the number of Top Hits) and other approaches to analyse genome-wide association study data, and the within-pair, ICE FALCON (Inference about Causation from Examining FAmiliaL CONfounding) and ICE CRISTAL (Inference about Causation from Examining Changes in Regression coefficients and Innovative STatistical AnaLysis) approaches to causation and familial confounding. Example applications to breast and colorectal cancer are presented. Motivated by the availability of the resources of the Breast and Colon Cancer Family Registries, we also present some ideas for future studies that could be applied to, and compared with, cancers diagnosed at older ages and address the challenges posed by young breast and bowel cancers.

Divorce and risk of suicide attempt: a Swedish national study.

BACKGROUND: Prior research has reported an association between divorce and suicide attempt. We aimed to clarify this complex relationship, considering sex differences, temporal factors, and underlying etiologic pathways. METHODS: We used Swedish longitudinal national registry data for a cohort born 1960-1990 that was registered as married between 1978 and 2018 (N = 1 601 075). We used Cox proportional hazards models to estimate the association between divorce and suicide attempt. To assess whether observed associations were attributable to familial confounders or potentially causal in nature, we conducted co-relative analyses. RESULTS: In the overall sample and in sex-stratified analyses, divorce was associated with increased risk of suicide attempt (adjusted hazard ratios [HRs] 1.66-1.77). Risk was highest in the year immediately following divorce (HRs 2.20-2.91) and declined thereafter, but remained elevated 5 or more years later (HRs 1.41-1.51). Divorcees from shorter marriages were at higher risk for suicide attempt than those from longer marriages (HRs 3.33-3.40 and 1.20-1.36, respectively). In general, HRs were higher for divorced females than for divorced males. Co-relative analyses suggested that familial confounders and a causal pathway contribute to the observed associations. CONCLUSIONS: The association between divorce and risk of suicide attempt is complex, varying as a function of sex and time-related variables. Given evidence that the observed association is due in part to a causal pathway from divorce to suicide attempt, intervention or prevention efforts, such as behavioral therapy, could be most effective early in the divorce process, and in particular among females and those whose marriages were of short duration.

Maternal polycystic ovary syndrome and risk of neuropsychiatric disorders in offspring: prenatal androgen exposure or genetic confounding?

BACKGROUND: Maternal polycystic ovary syndrome (PCOS) has been proposed as a model for investigating the role of prenatal androgen exposure in the development of neuropsychiatric disorders. However, women with PCOS are at higher risk of developing psychiatric conditions and previous studies are likely confounded by genetic influences. METHODS: A Swedish nationwide register-based cohort study was conducted to disentangle the influence of prenatal androgen exposure from familial confounding in the association between maternal PCOS and offspring attention-deficit/hyperactivity disorder (ADHD), autism spectrum disorders (ASD), and Tourette's disorder and chronic tic disorders (TD/CTD). PCOS-exposed offspring (n = 21 280) were compared with unrelated PCOS-unexposed offspring (n = 200 816) and PCOS-unexposed cousins (n = 17 295). Associations were estimated with stratified Cox regression models. RESULTS: PCOS-exposed offspring had increased risk of being diagnosed with ADHD, ASD, and TD/CTD compared with unrelated PCOS-unexposed offspring. Associations were stronger in girls for ADHD and ASD but not TD/CTD [ADHD: adjusted hazard ratio (aHR) = 1.61 (95% confidence interval (CI) 1.31-1.99), ASD: aHR = 2.02 (95% CI 1.45-2.82)] than boys [ADHD: aHR = 1.37 (95% CI 1.19-1.57), ASD: aHR = 1.46 (95% CI 1.21-1.76)]. For ADHD and ASD, aHRs for girls were stronger when compared with PCOS-unexposed cousins, but slightly attenuated for boys. CONCLUSIONS: Estimates were similar when accounting for familial confounding (i.e. genetics and environmental factors shared by cousins) and stronger in girls for ADHD and ASD, potentially indicating a differential influence of prenatal androgen exposure v. genetic factors. These results strengthen evidence for a potential causal influence of prenatal androgen exposure on the development of male-predominant neuropsychiatric disorders in female offspring of women with PCOS.

Education differences in sickness absence and the role of health behaviors: a prospective twin study.

BACKGROUND: Long-term sickness absences burden the economy in many industrialized countries. Both educational attainment and health behaviors are well-known predictors of sickness absence. It remains, however, unclear whether these associations are causal or due to confounding factors. The co-twin control method allows examining causal hypotheses by controlling for familial confounding (shared genes and environment). In this study, we applied this design to study the role of education and health behaviors in sickness absence, taking sex and cohort differences into account. METHODS: Participants were two cohorts of in total 8806 Norwegian twins born 1948 to 1960 (older cohort, mean age at questionnaire = 40.3, 55.8% women), and 1967 to 1979 (younger cohort, mean age at questionnaire = 25.6, 58.9% women). Both cohorts had reported their health behaviors (smoking, physical activity and body mass index (BMI)) through a questionnaire during the 1990s. Data on the twins' educational attainment and long-term sickness absences between 2000 and 2014 were retrieved from Norwegian national registries. Random (individual-level) and fixed (within-twin pair) effects regression models were used to measure the associations between educational attainment, health behaviours and sickness absence and to test the effects of possible familial confounding. RESULTS: Low education and poor health behaviors were associated with a higher proportion of sickness absence at the individual level. There were stronger effects of health behaviors on sickness absence in women, and in the older cohort, whereas the effect of educational attainment was similar across sex and cohorts. After adjustment for unobserved familial factors (genetic and environmental factors shared by twin pairs), the associations were strongly attenuated and non-significant, with the exception of health behaviors and sickness absence among men in the older cohort. CONCLUSIONS: The associations between educational attainment, health behaviors, and sickness absence seem to be confounded by unobserved familial factors shared by co-twins. However, the association between health behaviors and sickness absence was consistent with a causal effect among men in the older cohort. Future studies should consider familial confounding, as well as sex and age/cohort differences, when assessing associations between education, health behaviors and sickness absence.

Maternal infection requiring hospitalization during pregnancy and attention-deficit hyperactivity disorder in offspring: a quasi-experimental family-based study.

BACKGROUND: Maternal infection during pregnancy (IDP) has been associated with increased risk of attention-deficit/hyperactivity disorder (ADHD) in offspring. However, infection is associated with social adversity, poor living conditions and other background familial factors. As such, there is a need to rule out whether the observed association between maternal IDP and ADHD might be attributed to such confounding. METHODS: This nationwide population-based cohort study using a family-based, quasi-experimental design included 1,066,956 individuals born in Sweden between 1992 and 2002. Data on maternal IDP (bacterial or viral) requiring hospitalization and ADHD diagnosis in offspring were gathered from Swedish National Registers, with individuals followed up through the end of 2009. Ordinary and stratified Cox regression models were used for estimation of hazard ratios (HRs) and several measured covariates were considered. Cousin- and sibling-comparisons accounted for unmeasured genetic and environmental factors shared by cousins and siblings. RESULTS: In the entire population, maternal IDP was associated with ADHD in offspring (HR = 2.31, 95% CI = 2.04-2.61). This association was attenuated when accounting for measured covariates (HR = 1.86, 95% CI = 1.65-2.10). The association was further attenuated when adjusting for unmeasured factors shared between cousins (HR = 1.52, 95% CI = 1.12-2.07). Finally, the association was fully attenuated in sibling comparisons (HR = 1.03, 95% CI = 0.76-1.41). CONCLUSIONS: This study suggests that the association between maternal IDP and offspring ADHD is largely due to unmeasured familial confounding. Our results underscore the importance of adjusting for unobserved familial risk factors when exploring risk factors for ADHD.

Familial confounding of the association between maternal smoking during pregnancy and internalizing disorders in offspring.

BACKGROUND: Maternal smoking has consistently been associated with multiple adverse childhood outcomes including externalizing disorders. In contrast the association between maternal smoking during pregnancy (MSDP) and internalizing (anxiety and depressive) disorders in offspring has received less investigation. METHOD: We conducted a nationwide cohort study including 957635 individuals born in Denmark between 1991 and 2007. Data on MSDP and diagnoses of depression or anxiety disorders were derived from national registers and patients were followed up from the age of 5 years to the end of 2012. Hazard rate ratios (HRRs) were estimated using stratified Cox regression models. Sibling data were used to disentangle individual- and familial-level effects of MSDP and to control for unmeasured familial confounding. RESULTS: At the population level, offspring exposed to MSDP were at increased risk for both severe depression [HRR 1.29, 95% confidence interval (CI) 1.22-1.36] and severe anxiety disorders (HRR 1.26, 95% CI 1.20-1.32) even when controlling for maternal and paternal traits. However, there was no association between MSDP and internalizing disorders when controlling for the mother's propensity for MSDP (depression: HRR 1.11, 95% CI 0.94-1.30; anxiety disorders: HRR 0.94, 95% CI 0.80-1.11) or comparing differentially exposed siblings (depression: HRR 1.18, 95% CI 0.75-1.89; anxiety disorders: HRR 0.87, 95% CI 0.55-1.36). CONCLUSIONS: The results suggest that familial background factors account for the association between MSDP and severe internalizing disorders not the specific exposure to MSDP.

High Maternal Body Mass Index in Early Pregnancy and Risks of Stillbirth and Infant Mortality-A Population-Based Sibling Study in Sweden.

In a population-based case-control study, we investigated whether familial confounding influenced the associations between maternal overweight/obesity and risks of stillbirth and infant mortality by including both population and sister controls. Using nationwide data from the Swedish Medical Birth Register (1992-2011), we included all primiparous women with singleton births who also had a sister with a first birth during that time period. We used logistic regression analyses to calculate odds ratios (and 95% confidence intervals) adjusted for maternal age, height, smoking habits, education, and time period (5-year groups) of child's birth. Body mass index (BMI) was calculated as weight (kg)/height (m)(2) Compared with population controls with a normal BMI (18.5-24.9), stillbirth risk increased with increasing BMI (BMI 25-29.9: odds ratio (OR) = 1.51 (95% confidence interval (CI): 1.21, 1.89); BMI 30-34.9: OR = 1.77 (95% CI: 1.24, 2.50); BMI ≥35: OR = 3.16 (95% CI: 2.10, 4.76)). The sister case-control analyses revealed similar results. Offspring of obese women (BMI ≥30) had an increased risk of infant mortality when population controls were used (OR = 2.41, 95% CI: 1.83, 3.16), and an even higher risk was obtained when sister controls were used (OR = 4.04, 95% CI: 2.25, 7.25). We conclude that obesity in early pregnancy is associated with increased risks of stillbirth and infant mortality independently of genetic and early environmental risk factors shared within families.

Childhood maltreatment and attention deficit hyperactivity disorder symptoms in adults: a large twin study.

BACKGROUND: Childhood maltreatment (CM) has been associated with increased risk of attention deficit hyperactivity disorder (ADHD) in children and adults. It is, however, unclear whether this association is causal or due to familial confounding. METHOD: Data from 18 168 adult twins, aged 20-46 years, were drawn from the population-based Swedish twin registry. Retrospective self-ratings of CM (emotional and physical neglect, physical and sexual abuse and witnessing family violence), and self-ratings for DSM-IV ADHD symptoms in adulthood were analysed. Possible familial confounding was investigated using a within twin-pair design based on monozygotic (MZ) and dizygotic (DZ) twins. RESULTS: CM was significantly associated with increased levels of ADHD symptom scores in adults [regression coefficient: 0.40 standard deviations, 95% confidence interval (CI) 0.37-0.43]. Within twin-pair analyses showed attenuated but significant estimates within DZ (0.29, 95% CI 0.21-0.36) and MZ (0.18, 95% CI 0.10-0.25) twin pairs. Similar results emerged for hyperactive/impulsive and inattentive ADHD symptom scores separately in association with CM. We conducted sensitivity analyses for early maltreatment, before age 7, and for abuse and neglect separately, and found similarly reduced estimates in DZ and MZ pairs. Re-traumatization after age 7 did not significantly influence results. CONCLUSIONS: CM was significantly associated with increased ADHD symptoms in adults. Associations were partly due to familial confounding, but also consistent with a causal interpretation. Our findings support cognitive neuroscience studies investigating neural pathways through which exposure to CM may influence ADHD. Clinicians treating adults with ADHD should be aware of the association with maltreatment.

Age at fatherhood: heritability and associations with psychiatric disorders.

BACKGROUND: Advancing paternal age has been linked to psychiatric disorders. These associations might be caused by the increased number of de novo mutations transmitted to offspring of older men. It has also been suggested that the associations are confounded by a genetic liability for psychiatric disorders in parents. The aim of this study was to indirectly test the confounding hypotheses by examining if there is a genetic component to advancing paternal age and if men with a genetic liability for psychiatric disorders have children at older ages. METHOD: We examined the genetic component to advancing paternal age by utilizing the twin model in a cohort of male twins (N = 14 679). We also studied ages at childbirth in men with or without schizophrenia, bipolar disorder and/or autism spectrum disorder. Ages were examined in: (1) healthy men, (2) affected men, (3) healthy men with an affected sibling, (4) men with healthy spouses, (5) men with affected spouses, and (6) men with healthy spouses with an affected sibling. RESULTS: The twin analyses showed that late fatherhood is under genetic influence (heritability = 0.33). However, affected men or men with affected spouses did not have children at older ages. The same was found for healthy individuals with affected siblings. Instead, these men were generally having children at younger ages. CONCLUSION: Although there is a genetic component influencing late fatherhood, our data suggest that the associations are not explained by psychiatric disorders or a genetic liability for psychiatric disorders in the parent.

Commentary: How can family-based quasi-experimental designs and national registers be used to address confounding in risk factor studies of psychopathology? A reflection on Obel et al. (2016).

Standard observational studies have reported a robust correlation between maternal smoking during pregnancy and risk of ADHD in offspring. In the accompanying article, Obel et al. used sibling-comparisons to explore the extent to which unmeasured familial confounding explains this association. This commentary highlights three important implications of the study. At a general level, Obel et al. illustrates how (1) family-based quasi-experimental designs and (2) national registers can be used to address confounding in risk factor studies of psychopathology. At a more specific level, the study suggests that maternal smoking during pregnancy is probably not a causal risk factor for ADHD.

Birthweight, gestational age and familial confounding in sex differences in infant mortality: a matched co-twin control study of Brazilian male-female twin pairs identified by population data linkage.

BACKGROUND: In infancy, males are at higher risk of dying than females. Birthweight and gestational age are potential confounders or mediators but are also familial and correlated, posing epidemiological challenges that can be addressed by studying male-female twin pairs. METHODS: We studied 28 558 male-female twin pairs born in Brazil between 2012 and 2016, by linking their birth and death records. Using a co-twin control study matched for gestational age and familial factors, we applied logistic regression with random effects (to account for paired data) to study the association between male sex and infant death, adjusting for: birthweight, within- and between-pair effects of birthweight, birth order and gestational age, including interactions. The main outcome was infant mortality (0-365 days) stratified by neonatal (early and late) and postneonatal deaths. RESULTS: Males were 100 g heavier and more at risk of infant death than their female co-twins before [odds ratio (OR) = 1.28, 95% confidence interval (CI): 1.11-1.49, P = 0.001] and after (OR = 1.60, 95% CI: 1.39-1.83, P <0.001) adjusting for birthweight and birth order. When adjusting for birthweight within-pair difference and mean separately, the OR attenuated to 1.40 (95% CI: 1.21-1.61, P <0.001), with evidence of familial confounding (likelihood ratio test, P <0.001). We found evidence of interaction (P = 0.001) between male sex and gestational age for early neonatal death. CONCLUSIONS: After matching for gestational age and familial factors by design and controlling for birthweight and birth order, males remain at greater risk of infant death than their female co-twins. Birthweight's role as a confounder can be partially explained by familial factors.

Associations of paternal and maternal per- and polyfluoroalkyl substances exposure with cord serum reproductive hormones, placental steroidogenic enzyme and birth weight.

BACKGROUND: Maternal per- and polyfluoroalkyl substances (PFAS) exposure has been associated with placental function and fetal growth measures. However, few studies have simultaneously investigated paternal and maternal exposure effects. OBJECTIVES: We evaluated the associations of paternal or maternal PFAS levels with placental function and fetal growth measures. METHODS: We studied six PFAS measured in matched parental serums collected within 3 days before delivery in a birth cohort from LaiZhouWan, China. Outcomes evaluated include cord serum estradiol (n = 351), testosterone (n = 349), placental P450aromatase (n = 125), and birth weight (n = 369). Multiple linear regression was applied to estimate the associations for these outcomes according to paternal or maternal PFAS level after adjusting for socio-demographic confounders. Co-adjustment analysis of both paternal and maternal PFAS in the same model was performed. RESULTS: Maternal and paternal PFAS levels were correlated (Spearman's r = 0.23-0.45). Maternal PFAS were associated with increased estradiol (e.g. , PFOA: ß = 0.03, 95%CI: 0.00, 0.07), testosterone (e.g. , PFUA: ß = 0.14, 95%CI: 0.00, 0.27), and P450aromatase (e.g. , PFOA: ß = 0.13, 95%CI: 0.04, 0.22). Maternal PFAS were also associated with a lower mean of birth weight but the estimated 95% CI included the null. Paternal PFAS were not associated with any of the outcomes evaluated. CONCLUSIONS: Several maternal PFAS were associated with fetal steroid hormones and placental enzymes. Despite a correlation of PFAS level within the couples, no association was found for paternal PFAS exposure on these outcomes. The findings suggest the intrauterine PFAS exposure effect on fetal endocrine hormones and growth is unlikely to be confounded by exposure sources or familial factors shared within the couples.