Polytomous Logistic Regression Models Research Articles

Joint effects of indoor air pollution and maternal psychosocial factors during pregnancy on trajectories of early childhood psychopathology.

Prenatal indoor air pollution and maternal psychosocial factors have been associated with adverse psychopathology. We used environmental-exposure mixture methodology to investigate joint effects of both exposure classes on child behavior trajectories. For 360 children from the South African Drakenstein Child Health Study, we created trajectories of Child Behavior Checklist scores (at 24, 42, and 60 months) using latent-class linear mixed effects models. Indoor air pollutants and psychosocial factors were measured during pregnancy (second trimester). After adjusting for confounding, single-exposure effects (per natural log-1 unit increase) were assessed using polytomous logistic regression models, joint effects using self-organizing maps, and principal component analysis. Three trajectories were chosen for both internalizing and externalizing problems, with "high" (externalizing) or "increasing" (internalizing) being the most adverse trajectories. High externalizing trajectory was associated with increased exposure to particulate matter of ≤ 10 microns in diameter (PM10) (odds ratio [OR] = 1.25; 95% CI, 1.01-1.55) and self-organizing maps exposure profile most associated with smoking (OR = 2.67; 95% CI, 1.14-6.27). Medium internalizing trajectory was associated with increased emotional intimate partner violence (OR = 2.66; 95% CI, 1.17-5.57), increasing trajectory with increased benzene (OR = 1.24; 95% CI, 1.02-1.51) and toluene (1.21; 95% CI, 1.02-1.44) and the principal component most correlated with benzene and toluene (OR = 1.25; 95% CI, 1.02-1.54). Prenatal exposure to environmental pollutants and psychosocial factors was associated with internalizing and externalizing child behavior trajectories. Understanding joint effects of adverse exposure mixtures will facilitate targeted interventions to prevent childhood psychopathology. This article is part of a Special Collection on Mental Health.

Body image dissatisfaction and low adherence to the Western dietary standard among schoolchildren: a cross-sectional study

This study aimed to identify dietary patterns and to evaluate the association between body image perception and these patterns among schoolchildren. A cross-sectional study was conducted on 385 schoolchildren (both sexes, 10-17 years old) from four public schools in Salvador, Brazil. Two non-consecutive 24-hour dietary recalls were applied, and the dietary pattern was determined by Exploratory Factor Analysis after adjustment of dietary data using the Multiple Source Method. To evaluate body image perception, the Brazilian body shape silhouette was used. The associations between body image perception and dietary patterns using the polytomous logistic regression model adjusted for covariables was assessed. Three dietary patterns were obtained: "Western standard," unhealthy, "healthy pattern" and "typical dishes/ junk food." After adjustment, adolescents who wished for a slimmer silhouette (2.48 [95%CI: 1.04-6.11], 3.49 [95%CI: 1.35-9.05] and 2.87 [95%CI: 1.13-7.26]) were more likely to adhere to the quintiles with the lowest consumption of the Western standard, compared to those satisfied with their body image. No associations were identified in the other two dietary patterns. Adolescents dissatisfied with their bodies tend to adhere less to unhealthy dietary patterns.

CUMULATIVE EFFECTS OF COGNITIVE IMPAIRMENT AND FRAILTY ON GERIATRIC SYNDROMES IN A GERIATRIC OUTPATIENT SETTING

Abstract Aim With underlying geriatric syndromes, only few studies emphasized on the aspect of co-occurrence of frailty and cognitive impairment, though they are independently associated with several geriatric syndromes. Methods The participants were recruited from the outpatient clinic of geriatric department in a medical center. Frailty was assessed by a modified version of Fried’s criteria, while cognitive impairment was defined by cutoff scores of the Mini Mental State Examination. Demographic characteristics, self-assessed health status, and geriatric assessment were collected. Polytomous logistic regression model was used to evaluate factors correlated with frailty alone, cognitive impairment alone, and co-occurrence of frailty and cognitive impairment. Results The mean age of the 323 participants (60.1% women) was 80.4 ± 7.3 years. The prevalence of non-frailty, frailty alone, cognitive impairment alone and co-occurrence were 34.4%, 16.7%, 23.8% and 25.1%, respectively. When compared to the not-frail group, the odds ratios (OR) of having activity of daily living impairment (aOR 1.44, 95% CI 1.28-1.63), instrumental activities of daily living impairment (aOR 1.61, 95% CI 1.45-1.81), urinary incontinence (aOR 1.61, 95% CI 1.45-1.81) and with lower mini-nutritional assessment score (aOR 148, 95% CI 1.30-1.69) were significantly higher in the co-occurrence group after adjusted with demographic variables and Charlson comorbidity index. Conclusions Co-occurrence of frailty and cognitive impairment was associated with higher prevalence of disability, incontinence and malnutrition. It is crucial to approach frailty and cognitive impairment simultaneously in daily geriatric clinical practice due to the correlations of several geriatric syndromes to the co-occurrence of physical frailty and cognitive impairment.

Mosquito control exposures and breast cancer risk: analysis of 1071 cases and 2096 controls from the Ghana Breast Health Study

Epidemiologic data on insecticide exposures and breast cancer risk are inconclusive and mostly from high-income countries. Using data from 1071 invasive pathologically confirmed breast cancer cases and 2096 controls from the Ghana Breast Health Study conducted from 2013 to 2015, we investigated associations with mosquito control products to reduce the spread of mosquito-borne diseases, such as malaria. These mosquito control products were insecticide-treated nets, mosquito coils, repellent room sprays, and skin creams for personal protection against mosquitos. Multivariable and polytomous logistic regression models were used to estimate odds ratios (ORadj) and 95% confidence intervals (CI) with breast cancer risk-adjusted for potential confounders and known risk factors. Among controls, the reported use of mosquito control products were mosquito coils (65%), followed by insecticide-treated nets (56%), repellent room sprays (53%), and repellent skin creams (15%). Compared to a referent group of participants unexposed to mosquito control products, there was no significant association between breast cancer risk and mosquito coils. There was an association in breast cancer risk with reported use of insecticide-treated nets; however, that association was weak and not statistically significant. Participants who reported using repellent sprays were at elevated risks compared to women who did not use any mosquito control products, even after adjustment for all other mosquito control products (OR = 1.42, 95% CI=1.15–1.75). We had limited power to detect an association with repellent skin creams. Although only a few participants reported using repellent room sprays weekly/daily or < month-monthly, no trends were evident with increased frequency of use of repellent sprays, and there was no statistical evidence of heterogeneity by estrogen receptor (ER) status (p-het > 0.25). Our analysis was limited when determining if an association existed with repellent skin creams; therefore, we cannot conclude an association. We found limited evidence of risk associations with widely used mosquito coils and insecticide-treated nets, which are reassuring given their importance for malaria prevention. Our findings regarding specific breast cancer risk associations, specifically those observed between repellent sprays, require further study.

Abstract B070: Characterizing breast cancer etiologic subtypes in the Ghana Breast Health Study

Abstract Background: Breast cancer subtyping has largely been based on immunohistochemical markers of estrogen receptor status. However, RNA-based intrinsic subtypes, including luminal (hormone receptor-positive) and non-luminal (hormone receptor-negative) have been shown to be etiologically relevant, with studies also suggesting other tumor markers (e.g., TP53 status) to be important contributors to etiologic heterogeneity. Compared to luminal tumors, non-luminal tumors are more likely to be TP53 mutant than wildtype-like. TP53 mutant breast cancers are associated with more aggressive tumor features and these mutations are more common among cases of African descent. However, few RNA-based subtyping studies have described etiologic heterogeneity in indigenous African populations. Therefore, we evaluated associations between established breast cancer risk factors and risk of PAM50 intrinsic subtypes and tumor subtypes defined by RNA-based TP53 functional status in the population-based Ghana Breast Health Study (GBHS). Methods: We analyzed data from 600 invasive breast cancer cases and 2,528 controls in the GBHS. RNA was extracted from formalin-fixed paraffin-embedded tumor samples and run on the nCounter® Breast Cancer 360™ Panel. Samples were then classified for intrinsic subtype (N=285 luminal, N=315 non-luminal) using a research version of the PAM50 assay and for TP53 functional status (N=333 wildtype-like, N=300 mutant-like) using a validated RNA signature. Odds ratios (OR) and 95% confidence intervals (CI) were estimated for risk factors using polytomous logistic regression models adjusted for potential confounders. Results: About 1 in 4 tumors classified as ER-positive by immunohistochemistry were non-luminal by RNA expression. Similarly, about 1 in 5 ER-negative tumors were reclassified as luminal. Etiologic heterogeneity was observed by luminal subtype for certain reproductive factors. Higher parity (≥3 vs. &amp;lt;3 births) was associated with lower risk of luminal tumors (OR [95% CI] = 0.68 [0.49, 0.92]), but not with non-luminal tumors (0.95 [0.69, 1.30]). Likewise, later age at first birth (≥26 vs. &amp;lt;19 years) was more strongly associated with risk of luminal than non-luminal tumors (2.16 [1.37, 3.40] and 1.30 [0.84, 2.00], respectively). Consistent with previous findings, TP53 status was also an important etiologic factor and risk factor patterns mirrored those for luminal/non-luminal subtypes. Finally, we evaluated cross-classification of these two etiologic markers to identify risk factors that represent the dual action of both an estrogenic and a DNA-repair pathway. Age at first birth seems to suggest dual action. Conclusions: This study used high-quality RNA expression data to show that intrinsic subtype and TP53 status are useful breast tumor markers for describing etiologic heterogeneity, with cross-classification leading to improved understanding of the etiologic underpinnings of certain risk factors. This work has important implications for the use of gene-expression-based subtyping for breast cancer classification in this population. Citation Format: Amber N. Hurson, Ebonee N. Butler, Alina M. Hamilton, Bernard Petershie, Thomas U. Ahearn, Melissa A. Troester, Jonine Figueroa, Nicolas Titiloye, Montserrat Garcia-Closas, Mustapha Abubakar. Characterizing breast cancer etiologic subtypes in the Ghana Breast Health Study [abstract]. In: Proceedings of the 16th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2023 Sep 29-Oct 2;Orlando, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2023;32(12 Suppl):Abstract nr B070.

What can diet of migrants from the French West Indies indicate about the nutrition transition?

Abstract Background French West Indies, like the other Caribbean territories, are currently undergoing a nutrition transition. Little is known about how migration and subsequent changes in food habits could influence adherence to specific dietary patterns. Our study aimed to assess dietary intake among French West Indians who live in and/or were born in the French West Indies, compared with individuals born and living in mainland France. Methods 1,094 participants from the NutriNet-Santé e-cohort, prospectively included between 2009-2019, were categorized into 4 subgroups: born and living in the French West Indies (West Indies/West Indies WW; n = 172); born in the West Indies and living in France Mainland (West Indies/mainland, WM n = 317); born in France Mainland and living in the French West Indies (mainland/West Indies, MW n = 288) and those born and living in Mainland France (mainland/mainland, MM n = 317). We compared the four subgroups on their nutritional quality of the diet using data provided by 24h records, and sociodemographic, lifestyle and anthropometric characteristics. The association between dietary patterns and migration status was estimated by polytomous logistic regression models. Results The WM group had consumptions reflecting an intermediate nutritional quality between the WW and the MM groups, with higher consumption of plant-based foods. The WM and WW groups were more likely to adhere both to a traditional dietary pattern from the West Indies (tubers grown in the West Indies, legumes, fish and offal) and, to a lesser extent, a convenient dietary pattern (high in ultra-processed foods), compared to MM, confirming the ongoing nutrition transition. Conclusions Migrants from French West Indies may have changed their traditional dietary habits, suggesting an influence of the environment. Migrants presented both potentially protective and detrimental dietary patterns compared to their peers born in mainland France due to the maintenance of cultural habits. Key messages • Migration may influence food habits for migrants between the French West Indies and Mainland France. • The nutrition transition is on going in the French West Indies, representing a public health issue.

Social determinants of long-term reported changes in physical activity and healthy eating during the COVID-19 pandemic in Canada: multiple cross-sectional surveys analysis from the iCARE study.

The long-term consequences of COVID-19 on healthy behaviours (physical activity practice and healthy eating) among Canadians remain largely unexplored. The objectives were (i) to describe the proportion of Canadians who reported a change in healthy behaviours, 9 and 20 months since the beginning of COVID-19; and (ii) to identify the social determinants associated with healthy behaviour changes. Using two representative Canadian surveys from the International COVID-19 Awareness and Responses Evaluation study (January 2021, n=3000; November 2021, n=3002), reported changes in healthy behaviours were assessed as follows: "In general, how have the following behaviours changed since the start of COVID-19?": (1) Increase; (2) No change; and (3) Decrease. The association between individual determinants and changes in healthy behaviours was analyzed using weighted univariate polytomous logistic regression models. In January 2021, 41% and 22% of respondents reported a decline in physical activity and healthy eating, respectively, while in November 2021, 34% and 20% of respondents reported a decline in physical activity and healthy eating, respectively. The main determinants associated with changes in healthy behaviours were younger age (18-25 years), area of residency, student status, changes in bodyweight, financial concerns/insecurity, anxiety/depression, and ethnicity. Changes in healthy behaviours were also associated with household composition, presence of chronic diseases, and occupation. In sum, this study depicted long-term changes in healthy behaviours during COVID-19, with differential changes according to social determinants of health. This study highlighted the presence of health inequalities in Canada during COVID-19 and supports the implementation of personalized programs in prevention of healthy behaviour degradation.

Association of health literacy with smoking behavior among Chinese men.

Health literacy may constitute a modifiable determinant of smoking behavior and intention to quit. Little is known about the extent to which health literacy affects smoking or quitting smoking. We assessed the nationally representative cross-sectional datasets from the China Health Literacy Surveillance (CHLS) initiated in 2018. Using polytomous logistic regression models, the study investigated the association of health literacy with smoking behavior and the intention to quit smoking among men aged 15-69 in China. After confounding factors were controlled, compared with having below basic health literacy, having adequate health literacy appeared to be an independent protective factor from current smoking [current smoking vs never smoking: adjusted odds ratio [OR], 0.88; 95% confidence interval (CI), 0.81-0.96; p = 0.003; current smoking vs former smoking: adjusted OR, 0.77; 95% CI, 0.64-0.92; p = 0.003], while having intermediate health literacy was associated with current smoking vs never smoking (adjusted OR, 1.09; 95% CI, 1.02-1.17; p = 0.011) or former smoking vs never smoking (adjusted OR, 1.22; 95% CI, 1.06-1.40; p = 0.005). And having adequate health literacy was associated with intending to quit among current smokers (adjusted OR, 1.25; 95% CI, 1.10-1.42; p < 0.001). Findings provide evidence that health literacy may serve as a critical and independent protective factor for reducing poor smoking behavior or enhancing cessation intention among men. Efforts should focus on developing and evaluating intervention to control tobacco use among men with low health literacy level.

Dimensionality and invariance of ADL, IADL, BI-M2/WG-SS, and GALI in large surveys in France (2008–2014) and implications for measuring disability in epidemiology

BackgroundThe epidemiological investigation and surveillance of disability requires well-constructed, invariant, and, if possible, exchangeable measures. However, the current or recommended measures have not been thoroughly investigated with respect to these issues. Here we examined the dimensional structure and invariance of four measures across sociodemographic groups: Activities of Daily Living (ADL), Instrumental Activities of Daily Living (IADL), Budapest Initiative Mark 2 (BI-M2) and Washington Group on Disability Statistics Short Set (WG-SS), and Global Activity Limitation Indicator (GALI).MethodsWe used data from three large nationwide representative surveys conducted in France between 2008 and 2014. The surveys included these four measures and classical and modern approaches (correlations, principal component analysis, Rasch modeling) were used to assess their dimensional structure as well as their invariance through differential item functioning (DIF) for sociodemographic characteristics. Polytomous logistic regression models were used to assess gradients in health inequalities associated with these measures.ResultsFor many items of ADL, IADL, and BI-M2/WG-SS, we consistently observed disordered response thresholds, rejection of unidimensionality, and DIF evidence for sociodemographic characteristics across the survey samples. Health inequality gradients were erratic. In addition, it was impossible to identify a common continuum for GALI, ADL, IADL, and BI-M2/WG-SS or their constituent items.ConclusionThis study warns against the current practice of investigating disability in epidemiology using measures that are unsuitable for epidemiological use, incommensurable, and inadequate regarding the basic requisites of dimensionality and invariance. Developing invariant measures and equating them along a common continuum to enlarge the common bases of measurement should therefore be a priority.

Improve the model of disease subtype heterogeneity by leveraging external summary data.

Researchers are often interested in understanding the disease subtype heterogeneity by testing whether a risk exposure has the same level of effect on different disease subtypes. The polytomous logistic regression (PLR) model provides a flexible tool for such an evaluation. Disease subtype heterogeneity can also be investigated with a case-only study that uses a case-case comparison procedure to directly assess the difference between risk effects on two disease subtypes. Motivated by a large consortium project on the genetic basis of non-Hodgkin lymphoma (NHL) subtypes, we develop PolyGIM, a procedure to fit the PLR model by integrating individual-level data with summary data extracted from multiple studies under different designs. The summary data consist of coefficient estimates from working logistic regression models established by external studies. Examples of the working model include the case-case comparison model and the case-control comparison model, which compares the control group with a subtype group or a broad disease group formed by merging several subtypes. PolyGIM efficiently evaluates risk effects and provides a powerful test for disease subtype heterogeneity in situations when only summary data, instead of individual-level data, is available from external studies due to various informatics and privacy constraints. We investigate the theoretic properties of PolyGIM and use simulation studies to demonstrate its advantages. Using data from eight genome-wide association studies within the NHL consortium, we apply it to study the effect of the polygenic risk score defined by a lymphoid malignancy on the risks of four NHL subtypes. These results show that PolyGIM can be a valuable tool for pooling data from multiple sources for a more coherent evaluation of disease subtype heterogeneity.

Analysis of factors influencing USA households' choice of ethnic restaurants

PurposeThis study empirically identifies household demographic and socioeconomic characteristics as well as restaurant characteristics that affect the probability of households choosing an ethnic restaurant (American, Asian, European, Mexican and other) in the USA.Design/methodology/approachA multinomial logistic regression model is applied using the data derived from the information from the National Household Food Acquisition and Purchase Survey conducted between April 2012 and January 2013.FindingsThe empirical findings suggest that such factors as the unit cost on away-from-home food items (i.e. price), region of residence, primary respondent's ethnicity, race, education level, marital status and employment status as well as such restaurant characteristics as availability of loyalty program and presence of coupons significantly affect the probability of households choosing a particular ethnic restaurant in the USA.Research limitations/implicationsThe original dataset employed in this study does not permit the quantification of information associated with size, location, and number of years in operation for the ethnic restaurants considered. Also, the dataset does not permit the classification of the ethnic restaurants included in the “other” category.Originality/valueTo the best of the authors' knowledge, there has been no empirical micro-level analysis associated with determining factors impacting households' choice of ethnic restaurants using a polytomous logistic regression model allowing for a wide range of ethnic restaurants and covering the entire USA, based on an extensive set of household demographic and socioeconomic factors and restaurants characteristics. As such, the current study plugs this research gap, with the empirical findings furnished by this study being of importance to ethnic restaurant operators (owners) in the operators' effort to develop effective marketing strategies.

Abstract 1186: Contraception and ovarian cancer risk by tumor immune profiles

Abstract Introduction: An estimated 38 million US women currently use contraception, a critical component of women’s health. Yet, the impact of changing contraceptive types and patterns of use on ovarian cancer incidence is unclear and the biologic pathway through which contraception influences ovarian cancer risk is not well understood. Using multiplex immunofluorescence (mIF) technology, we evaluated how contraception influences the tumor immune microenvironment (e.g., T cells) to better understand the association of intrauterine device (IUD) and oral contraception (OC) use on ovarian cancer risk. Methods: We measured tumor immune profiles via mIF on tissue microarrays including tumor cores from 214 women with ovarian cancer (85% high grade serous) and 414 population-based controls frequency matched on age who participated in the New England Case Control Study. T cell types were defined by expression of CD3+ (total T cells), CD3+CD4+ (helper), CD3+CD8+ (cytotoxic), CD3+CD4+FoxP3+ (regulatory), CD3+CD4+CD69+ or CD3+CD8+CD69+ (recently activated), or CD3+Tim3+ or CD3+Lag3+(exhausted). The average immune cell density across 3 cores per tumor was used to categorize tumors as high or low abundance based on the median across all tumors. Contraception use was self-reported and included having ever used an IUD and/or OCs and duration of use. Polytomous logistic regression models were used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for ovarian cancer defined by tumor immune cell abundance adjusted for age at diagnosis or interview, parity, and family history of breast or ovarian cancer in first degree relative. Results: Overall, ever use of IUD or OC was not differentially associated with ovarian cancer risk by T cell tumor immune profiles. However, for OC use of 5 years or longer, we observed a significant reduction in risk for low-abundance total T cell (ORlow=0.44, 95% CI 0.21-0.94), cytotoxic T cell (ORlow=0.44, 95% CI:0.21-0.94), regulatory T cell (ORlow=0.31, 95% CI: 0.14-0.68), and recently activated T cell (CD3+CD4+CD69+ ORlow=0.34, 95% CI: 0.17-0.68) tumors with significant trends with longer duration. There was no significant decrease in risk of tumors with high abundance of these T cell types, although statistically significant heterogeneity was only observed in the association for duration of OC use and risk by low and high abundance of recently activated helper T cells (CD3+CD4+CD69+, p-het=0.006). Conclusion: These results suggest that ever use of IUD and OCs are not associated with risk of ovarian tumors by immune phenotype, though risk by T cell abundance may vary with longer duration. Analyses are ongoing and will be expanded to include types and patterns of contraception, other immune markers, and additional studies. Citation Format: Jennifer Mongiovi, Ana Babic, Naoko Sasamoto, Mary Townsend, Allison Vitonis, Jonathan Hecht, Jose Conejo-Garcia, Brooke Fridley, Shelley Tworoger, Kathryn Terry. Contraception and ovarian cancer risk by tumor immune profiles [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1186.

Abstract 3017: Breastfeeding and ovarian cancer risk by tumor immune profiles

Abstract Introduction: Breastfeeding is associated with decreased ovarian cancer risk, yet the biological mechanisms are not fully understood. We conducted an agnostic investigation of tumor immune profiles to better understand the role of the tumor microenvironment in the protective association between breastfeeding and ovarian cancer risk. Methods: Multiplex immunofluorescence was used to measure the abundance of T cells and B cells in ovarian tumor samples on tissue microarrays from the Nurses’ Health Study (NHS; n=337), NHSII (n=127), and New England Case Control Study (NECC; n=214). We averaged the percent of cells within the tumor epithelium that were positive for the immune cell of interest across 3 cores per tumor and categorized tumors as having high or low immune cell abundance based on the median value across all tumors. Controls were non-cases and did not provide tissue samples (n=2,045). Self-reported history of breastfeeding was used to define ever/never and total duration of breastfeeding. Polytomous logistic regression models adjusted for age, birth cohort, race, parity, oral contraception use, and study, was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for breastfeeding and ovarian cancer risk by immune cell level. Results: Overall, breastfeeding was associated with a reduced risk of tumors with low, but not high, subtypes of T cell abundance. Among parous women, history of ever breastfeeding decreased the risk of tumors with low helper T cells (CD3+CD4+; ORlow: 0.56, 95% CI: 0.41-0.77), but not high helper T cells (ORhigh: 0.85, 95% CI: 0.63-1.35; p-het=0.06). For total B cells (CD19+), having ever breastfed was associated with decreased risk of both low (ORlow: 0.64, 95% CI: 0.47-0.87) and high abundance in the tumor (ORhigh: 0.70, 95% CI: 0.62-0.96, p-het=0.67) among parous women. Similar results were observed for CD138+ plasma cells (ORlow:0.76, 95% CI: 0.56-0.99; ORhigh: 0.59, 95% CI: 0.37-0.94; p-het=0.34). Longer duration of breastfeeding was associated with decreased risk of developing tumors with low abundance of total B cells (CD19+, p-trend=0.02) and regulatory B cells (CD19+TNFR2+, p-trend=0.05), but not high abundance of CD19+ or CD19+TNFR2+ (p-het=0.01 and 0.03, respectively). Conclusion: These results indicate that among parous women breastfeeding is associated with reduced risk of ovarian tumors with low infiltration of helper T cells while no heterogeneity was observed for risk by total B cells or plasma cells. Longer breastfeeding duration was associated with reduced risk of tumors low in regulatory B cells (CD19+TNFR+) with a dose-dependent response. Together, these results suggest breastfeeding may play a role in the activation of the tumor immune response. Further efforts are ongoing to disentangle the relationships between breastfeeding and parity with the tumor immune microenvironment. Citation Format: Jennifer Mongiovi, Mary Townsend, Allison Vitonis, Ana Babic, Jonathan Hecht, Jose Conejo-Garcia, Brooke Fridley, Shelley Tworoger, Kathryn Terry, Naoko Sasamoto. Breastfeeding and ovarian cancer risk by tumor immune profiles [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3017.

Associations of overall and specific carbohydrate intake with anxiety status evolution in the prospective NutriNet-Santé population-based cohort

We investigated the association between carbohydrate intake and anxiety evolution within the general-population NutriNet-Santé cohort (N = 15,602; 73.8% female; mean age = 53.8y; mean follow-up = 5.4y). Carbohydrate intake was estimated at baseline from ≥ 2 24-h dietary records. Trait anxiety (STAI-T) was measured once at baseline (2013–2016) and once at follow-up (2020), resulting in 4 groups: “None” = absence of high anxiety (STAI-T > 40 points) at any time point; “Transient” = high anxiety only at baseline; “Onset at follow-up” = high anxiety only at follow-up; “Persistent” = high anxiety at baseline and follow-up. Polytomous logistic regression models revealed that sweetened beverage intake was associated with higher odds of “Transient” anxiety (ORQ4vsQ1 = 1.11; 95% CI 1.02–1.21). Intake of complex carbohydrates (ORQ4vsQ1 = 1.12; 1.01–1.25) was associated with higher odds of anxiety “Onset at follow-up.” The % energy from carbohydrates (ORQ4vsQ1 = 1.11; 1.03–1.19), intakes of total carbohydrates (ORQ4vsQ1 = 1.10; 1.03–1.18) and complex carbohydrates (ORQ4vsQ1 = 1.09; 1.02–1.17) were associated with higher odds of “Persistent” anxiety, whereas 100% fruit juice intake showed lower odds of “Persistent” anxiety (ORQ4vsQ1 = 0.87; 0.81–0.94). This prospective study found significant associations between dietary carbohydrate intake and anxiety status evolution among French adults. The findings could help inform dietary interventions aimed at anxiety prevention and management.

Association between Immunologic Markers and Cirrhosis in Individuals from a Prospective Chronic Hepatitis C Cohort.

Chronic hepatitis C virus (HCV) infection can affect immune response and inflammatory pathways, leading to severe liver diseases such as cirrhosis and hepatocellular carcinoma (HCC). In a prospective cohort of chronically HCV-infected individuals, we sampled 68 individuals who developed cirrhosis, 91 controls who did not develop cirrhosis, and 94 individuals who developed HCC. Unconditional odds ratios (ORs) from polytomous logistic regression models and canonical discriminant analyses (CDAs) were used to compare categorical (C) baseline plasma levels for 102 markers in individuals who developed cirrhosis vs. controls and those who developed HCC vs. cirrhosis. Leave-one-out cross validation was used to produce receiver operating characteristic curves to assess predictive ability of markers. Lastly, biological pathways were assessed in association with cirrhotic development compared to controls. After multivariable adjustment, DEFA-1 (OR: C2v.C1 = 7.73; p &lt; 0.0001), ITGAM (OR: C2v.C1 = 4.03; p = 0.0002), SCF (OR: C4v.C1 = 0.19; p-trend = 0.0001), and CCL11 (OR: C4v.C1 = 0.31; p-trend= 0.002) were all associated with development of cirrhosis compared to controls; these markers, together with clinical/demographics variables, improved prediction of cirrhosis from 55.7% (in clinical/demographic-only model) to 74.9% accuracy. A twelve-marker model based on CDA results further increased prediction of cirrhosis to 88.0%. While six biological pathways were found to be associated with cirrhosis, cell adhesion was the only pathway associated with cirrhosis after Bonferroni correction. In contrast to cirrhosis, DEFA-1 and ITGAM levels were inversely associated with HCC risk. Pending validation, these findings highlight the important role of immunological markers in predicting HCV-related cirrhosis even 11 years post-enrollment.

Modeling the longitudinal transitions of electronic cigarettes and conventional cigarettes with time-dependent covariates among adolescents

E-cigarettes may help combustible cigarette smokers switch to a less harmful alternative, or may increase the risk of subsequent initiation of cigarettes among non-smokers. Among youth, it is not clear whether both pathways occur equally, or whether one direction is more likely than the other. We used data from a prospective cohort study of youth in Southern California followed twice annually from Fall 2013 (9th grade) to Fall 2015 (11th grade) (N = 1977). A polytomous logistic regression model was used to simultaneously estimate transition rates for initiation of and abstention from e-cigarettes and cigarettes. Use of e-cigarettes was positively associated with initiation of cigarettes (OR = 7.57; 95%CI:[5.32, 10.8]) and negatively associated with cigarette abstention (OR = 0.58; 95%CI:[0.33, 0.99]) in adjusted models; cigarette use was positively associated with e-cigarette initiation (OR = 2.54; 95%CI:[1.45, 4.47]) and negatively associated with e-cigarette abstention (OR = 0.31; 95%CI:[0.17,0.57]). Uni-directional transition from e-cigarettes only to cigarettes only occurred less frequently than expected under independence (OR = 0.33; 95% CI [0.20, 0.55]), whereas simultaneously initiating both products (OR = 9.79; 95%CI:[7.22, 13.3]) and simultaneously abstaining (OR = 2.84; 95%CI:[1.50, 5.37]) were more frequent than expected. E-cigarettes were more strongly associated with subsequent cigarette initiation than the reverse, though both models indicated that use of either product seems to encourage use of the other. Models also indicated that use of either e-cigarettes or cigarettes resulted in reduced abstention of the other product. Findings suggest that prevention efforts for that continue to focus on both e-cigarettes and cigarettes are needed.

Joint Effects of Indoor Air Pollution and Maternal Psychosocial Stress During Pregnancy on Trajectories of Early Childhood Psychopathology

Background and aim: Identifying modifiable risk factors for childhood psychopathology can increase quality of life for children and prevent adult psychopathology. Previous studies have shown that indoor air pollution (IAP) and maternal psychosocial stress (MPS) during pregnancy separately influence psychopathology, but their joint effects are not well understood. Here, we use environmental mixture methodology to estimate joint effects of IAP and MPS during pregnancy on trajectories of early childhood psychopathology. Methods: We used data from the Drakenstein Child Health Study (N=360), a South African birth cohort. Exposure to IAP and MPS was measured during the second trimester of pregnancy. The Childhood Behavior Checklist (CBCL) was administered at ages 2, 3.5, and 5 years to assess child psychopathology. CBCL trajectories were created using latent class linear mixed effects models. Joint effects of IAP and MPS were summarized using principal component (PC) analysis. Individual effects of IAP and MPS and their joint effects were estimated using polytomous logistic regression models, adjusted for maternal age, sex, ancestry, and socioeconomic status. Results: Three distinct classes of CBCL trajectory were identified (class 1 (reference): lowest, class 2: highest and increases, class 3: stable at a medium-high). We consistently observed adverse associations between IAP (e.g., toluene for class 3: OR [95% CI]; 1.35 [1.00, 1.80]) and MPS (e.g., intimate partner violence for class 3: 1.44 [1.03, 1.99)) and CBCL. In the joint effects analysis, PC2 (accounted for 13% of risk factor variation) was significantly associated with class 3 (1.23 [1.03, 1.46]) and class 2 (1.27 [1.04, 1.55]) compared to class 1. Conclusions: Exposure to IAP and MPS during pregnancy was associated with CBCL trajectories indicating increased risk for psychopathology. Estimating joint effects of IAP and MPS is important to identify vulnerable subgroups for intervention to prevent childhood psychopathology. Keywords: indoor air pollution, psychopathology, exposure mixtures

Reproductive history differs by molecular subtypes of breast cancer among women aged ≤ 50 years in Scotland diagnosed 2009–2016: a cross-sectional study

BackgroundThe aetiology of breast cancers diagnosed ≤ 50 years of age remains unclear. We aimed to compare reproductive risk factors between molecular subtypes of breast cancer, thereby suggesting possible aetiologic clues, using routinely collected cancer registry and maternity data in Scotland.MethodsWe conducted a cross-sectional study of 4108 women aged ≤ 50 years with primary breast cancer diagnosed between 2009 and 2016 linked to maternity data. Molecular subtypes of breast cancer were defined using immunohistochemistry (IHC) tumour markers, oestrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor-2 (HER2), and tumour grade. Age-adjusted polytomous logistic regression models were used to estimate odds ratios (OR) and 95% confidence intervals (CI) for the association of number of births, age at first birth and time since last birth with IHC-defined breast cancer subtypes. Luminal A-like was the reference compared to luminal B-like (HER2−), luminal B-like (HER2+), HER2-overexpressed and triple-negative breast cancer (TNBC).ResultsMean (SD) for number of births, age at first birth and time since last birth was 1.4 (1.2) births, 27.2 (6.1) years and 11.0 (6.8) years, respectively. Luminal A-like was the most common subtype (40%), while HER2-overexpressed and TNBC represented 5% and 15% of cases, respectively. Larger numbers of births were recorded among women with HER2-overexpressed and TNBC compared with luminal A-like tumours (> 3 vs 0 births, OR 1.87, 95%CI 1.18–2.96; OR 1.44, 95%CI 1.07–1.94, respectively). Women with their most recent birth > 10 years compared to < 2 years were less likely to have TNBC tumours compared to luminal A-like (OR 0.63, 95%CI 0.41–0.97). We found limited evidence for differences by subtype with age at first birth.ConclusionNumber of births and time since last birth differed by molecular subtypes of breast cancer among women aged ≤ 50 years. Analyses using linked routine electronic medical records by molecularly defined tumour pathology data can be used to investigate the aetiology and prognosis of cancer.

A Generalized Integration Approach to Association Analysis with Multi-category Outcome: An Application to a Tumor Sequencing Study of Colorectal Cancer and Smoking

Cancer is a heterogeneous disease, and rapid progress in sequencing and -omics technologies has enabled researchers to characterize tumors comprehensively. This has stimulated an intensive interest in studying how risk factors are associated with various tumor heterogeneous features. The Cancer Prevention Study-II (CPS-II) cohort is one of the largest prospective studies, particularly valuable for elucidating associations between cancer and risk factors. In this paper, we investigate the association of smoking with novel colorectal tumor markers obtained from targeted sequencing. However, due to cost and logistic difficulties, only a limited number of tumors can be assayed, which limits our capability for studying these associations. Meanwhile, there are extensive studies for assessing the association of smoking with overall cancer risk and established colorectal tumor markers. Importantly, such summary information is readily available from the literature. By linking this summary information to parameters of interest with proper constraints, we develop a generalized integration approach for polytomous logistic regression model with outcome characterized by tumor features. The proposed approach gains the efficiency through maximizing the joint likelihood of individual-level tumor data and external summary information under the constraints that narrow the parameter searching space. We apply the proposed method to the CPS-II data and identify the association of smoking with colorectal cancer risk differing by the mutational status of APC and RNF43 genes, neither of which is identified by the conventional analysis of CPS-II individual data only. These results help better understand the role of smoking in the etiology of colorectal cancer.

Abstract 2255: Alu retroelement copy number alterations and non-Hodgkin lymphoma

Abstract Background: Alu retrotransposons are repetitive DNA elements that make up nearly 10% of the human genome and reflect genomic instability. Genomic instability plays an important role in the pathogenesis of non-Hodgkin lymphoma (NHL). However, the relationship between Alu retrotransposons and future risk of NHL is unclear. Methods: We conducted a nested case-control study of NHL risk and Alu retroelement copy number within the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) study, a prospective cohort with prediagnostic blood samples from 29,133 male Finnish smokers. We measured Alu copy number in 107 incident NHL cases and 1:1 matched controls (± 5 years age). DNA samples extracted from blood were assayed in triplicate using a monochrome multiplex quantitative polymerase chain reaction. We used the albumin gene for the single copy gene to control for genome content in samples. The odds ratios (OR) and 95% confidence intervals (CI) for all NHL were estimated using conditional logistic regression models, adjusting for age at randomization, number of cigarettes smoked per day and smoking duration. Polytomous logistic regression models were fit to compute ORs with Alu copy number for NHL subtypes. To determine if the association might be driven in part by elevated Alu copy number among undiagnosed cases at the time of blood sample collection, we excluded cases diagnosed within the first two years of follow-up. Results: Higher Alu copy number (using the median as the cut-point) was associated with an increased risk of NHL, both overall [OR, 95% CI=3.80 (1.76-8.17)] as well as for the diffuse large B-cell lymphoma [OR, 95% CI=4.83 (1.50-15.55)] and chronic lymphocytic leukemia/small lymphocytic lymphoma (SLL) [OR, 95% CI=2.66 (1.10-6.40)] subtypes. These associations were similar after excluding individuals diagnosed with NHL within two years after blood draw. Conclusions: These findings, which require replication, suggest that higher Alu copy number in prediagnostic leukocytes may play a role in the pathogenesis of NHL. Citation Format: Wei Hu. Alu retroelement copy number alterations and non-Hodgkin lymphoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2255.