Polychlorinated Biphenyls' Impact on Comorbidity Networks: Unveiling Epidemiological Patterns and offsetting Through Dietary Adjustments

Polychlorinated Biphenyls' Impact on Comorbidity Networks: Unveiling Epidemiological Patterns and offsetting Through Dietary Adjustments | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Polychlorinated Biphenyls' Impact on Comorbidity Networks: Unveiling Epidemiological Patterns and offsetting Through Dietary Adjustments Ying Gao, Han Lu, Huan Zhou, Jiaxing Tan This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4543285/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Polychlorinated Biphenyls (PCBs), as a major class of organic pollutants, garnered increasing attention due to their significant ability for inter-regional accumulation and migration, prolonged half-life, and relatively high toxicity. Our study aimed to assess the impact of PCBs on various diseases and mortality risks using data from the National Health and Nutrition Examination Survey (NHANES), while proposing lifestyle adjustments, particularly dietary modifications, to mitigate mortality risk. Statistical analyses employed principal component analysis (PCA), multifactorial logistic regression, multifactorial Cox regression, comorbidity network analysis, and machine learning prediction models. Results indicated significant associations between 7 types of PCBs and 12 diseases (p 1, p < 0.05), along with listing the 25 most relevant diseases, such as asthma and chronic bronchitis (OR (95%CI) = 5.85(4.37,7.83), p<0.0001), arthritis and osteoporosis (OR (95%CI) = 6.27(5.23,7.55), p<0.0001). We proposed that PCB exposure ultimately triggered adverse progression of multiple diseases and increased mortality risk, suggesting PCBs could potentially serve as specific biomarkers for certain diseases in the future. Building upon this, we further suggested that controlling dietary intake to reduce dietary inflammatory index (DII) could lower mortality and disease risks. While PCBs were independent risk factors for mortality, ample evidence suggested that adjusting DII might mitigate the adverse effects of PCBs to some extent. Further physiological mechanisms require deeper exploration through additional research. Polychlorinated Biphenyls Machine Learning Mortality Comorbidities DII Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Highlights 1 PCBs were strongly associated with an increased risk of multisystem disease in comorbid networks. 2.PCBs were an independent contributor to mortality. 3.Machine learning models validated the impact of PCBs on disease and death. 4.The group comparison confirmed the corrective effect of diet on PCBs. Introduction Increasing attention has been drawn to the potential adverse effects of human exposure to environmental pollutants in recent years. Dioxins, furans, coplanar polychlorinated biphenyls (PCBs), and organochlorine pesticides are all classified as persistent organic pollutants 1,2 . Among these, PCBs were widely used in electrical manufacturing and construction materials during the 20th century, and were commonly encountered as plasticizers and solvents 3 . The Stockholm Convention on Persistent Organic Pollutants, established in 2001, banned the production and use of PCBs due to their lipophilicity, environmental persistence, and ability to bioaccumulate in animal tissues 4 , thereby entering the human body through the food chain and external environmental exposure pathways 5 . According to pollution surveys in China, PCB contamination is particularly severe in areas of electronic waste dismantling (such as Taizhou City in Zhejiang Province, Qingyuan City and Guiyu Town in Guangdong Province), major metropolitan areas (especially in East China), and industrial clusters, primarily resulting from inadvertent generation during industrial heat treatment processes 6,7 . Traditionally, PCBs have been believed to exert toxic effects on multiple organs or systems through various mechanisms 8 . Previous studies on PCB exposure have mostly focused on the association between PCBs and individual medical conditions. Evidence suggests that PCB exposure increases the risk of hypertension 9 , diabetes 10 , pulmonary arterial hypertension 11 , and liver diseases 12 . Plasma component analysis has revealed that PCB exposure impedes hematopoietic function, leading to inflammation and oxidative stress in the body 13 . A study on elderly Americans reported a negative correlation between blood PCB levels and attention, memory, and learning performance 14,15 . Furthermore, some research indicates that PCBs are risk factors for DNA mutations, birth defects, and reproductive system diseases 16-18 . Although there are many studies on PCBs and diseases, most of them focused on one or a specific type of disease. Thus, comprehensive analyses of PCBs and comorbidity networks are still lacking 2,19 . Analyses focusing on individual diseases overlook interactions between diseases, and traditional methods for analyzing mortality risk overlook issues related to common exposure and risk confounding. With further advancements in statistics, comorbidity network analysis and machine learning-based algorithm models can effectively address these issues. Therefore, our study utilized data from over 100 million patients in the National Health and Nutrition Examination Survey (NHANES) to comprehensively investigate the integrated impact of PCBs on multiple disease categories through refined data selection and comorbidity network analysis. We also developed predictive models using machine learning algorithms to analyze the influence of PCBs on mortality risk. Additionally, we proposed for the first time a strategy to mitigate the adverse effects of PCBs by adjusting the Dietary Inflammatory Index (DII), which was validated through survival curve classification analysis and multifactorial Cox regression, offering profound clinical implications. Materials 2.1. Study design and participants As of March 1, 2024, all data used in our study are publicly accessible and sourced from the National Health and Nutrition Examination Survey (NHANES), managed by the National Center for Health Statistics (NCHS). Since 1999, data collection from participants has been conducted through questionnaire interviews, physical examinations, and laboratory tests. NHANES has obtained approval from the NCHS Institutional Review Board, and all participants have provided written informed consent. Survival status data of NHANES participants are derived from the National Death Index (NDI). Our analysis includes 69 persistent organic pollutants concurrently present in NHANES data (1999-2004), including dioxins, furans, and coplanar polychlorinated biphenyls (PCBs). Through data screening and analysis, we identified seven PCBs—PCB074, PCB170, PCB178, PCB180, PCB156, PCB157, and PCB146—as noteworthy compounds. We have provided specific nomenclature and abbreviations for these compounds: LBX074, representing PCB074(denoting 2,4,4’,5-Tetrachlorobiphenyl). LBX170, representing PCB170(2,2',3,3',4,4',5-Heptachlorobiphenyl). LBX178, representing PCB178(2,2’,3,3’,5,5’,6-Heptachlorobiphenyl). LBX180, representing PCB180(2,2’,3,4,4’,5,5’-Heptachlorobiphenyl). LBX156, representing PCB156(2,3,3',4,4',5-Hexachlorobiphenyl). LBX157, representing PCB157(2,3,3',4,4',5'-Hexachlorobiphenyl). LBX146, representing PCB146(2,3,3',4,4',5'-Hexachlorobiphenyl). We integrated NHANES data spanning from 1999 to 2004 with NDI data, meticulously selecting 30,158 participants. Participants with missing or zero NHANES weight and those with missing data for the seven PCBs relevant to our study were systematically excluded from our dataset. However, for comprehensive network analysis, we refrained from further data exclusion. This meticulous curation resulted in a final cohort of 10,961 participants. At different analysis stages, data were selected based on varying data requirements. 2.2. Methodology for Measuring Serum Levels of PCBs The analytes were quantified in serum using high-resolution gas chromatography and isotope-dilution high-resolution mass spectrometry (HRGS/ID-HRMS). Serum samples were fortified with 13C12-labeled internal standards and extracted through either C18 solid-phase extraction (SPE) or liquid-liquid extraction. Chromatographic separation occurred on a DB-5ms capillary column employing a Hewlett-Packard 6890 gas chromatograph. Quantification was achieved by ID-HRMS using selected ion monitoring (SIM) at a resolving power of 10,000 with either a Micromass AutoSpec ULTIMA or Finnigan MAT95 mass spectrometer in electron ionization (EI) mode. Detection limits were reported for each sample, accounting for sample weight and analyte recovery. From the entire persistent organic pollutant (POP) library, we ultimately identified 18 PCBs, with data for these substances in the NHANES cycles from 1999-2000, 2001-2002, and 2003-2004 accounting for over 75% of the total data. Further analysis of the 18 POPs using weighted quantile sum (WQS) analysis identified 7 PCBs with the greatest impact on mortality risk, cumulatively contributing to 95% of the total weight. As per NHANES guidelines, values falling below the limit of detection (LOD) were imputed with a value equivalent to the LOD divided by the square root of 2. 2.3. Diagnoses of medical conditions and DII As NHANES does not directly record mortality data, mortality information in this study was obtained through a probabilistic match between NHANES and the National Death Index (NDI), following procedures validated by the National Center for Health Statistics (NCHS). Mortality data from the NDI were available for analysis until December 31, 2019. All diseases available in NHANES were included in our analysis, encompassing Angina, Heart attack, Heart disease, Heart failure, Hypertension, Stroke, Alcoholic fatty liver, Hepatitis B virus (HBV), Hepatitis C virus (HCV), Hepatitis D virus (HDV), Non-alcoholic fatty liver, Hyperlipidemia, Diabetes, Osteoporosis, Hyperuricemia, Thyroid disease, Human Immunodeficiency Virus (HIV), Arthritis, Depression, Cancer, Chronic bronchitis, Emphysema, Asthma, Chronic kidney disease (CKD), and Proteinuria, totaling 25 diseases. Diabetes was defined as self-reported diabetes diagnosis, use of oral antidiabetic drugs or insulin, glycated hemoglobin (HbA1c) levels ≥6.5%, plasma glucose levels ≥200 mg/dL two hours after an oral glucose tolerance test (OGTT), or fasting plasma glucose levels ≥126 mg/dL 20 . Hypertension was determined based on self-reported hypertension or NHANES-measured data: an average systolic blood pressure ≥130 mm Hg or diastolic blood pressure ≥80 mm Hg from three measurements 21 . To assess chronic kidney disease (CKD), essential indicators including estimated glomerular filtration rate (eGFR) and urine albumin-to-creatinine ratio (UACR) were relied upon. UACR (mg/g) was calculated as the ratio of urine albumin (mg/dL) to urine creatinine (g/dL), with a UACR value exceeding 30 mg/g indicating "proteinuria." eGFR was computed using the CKD-EPI formula, expressed as: GFR = 175 × standardized serum creatinine (−1.154) × age (−0.203) × 1.212 [if Black] × 0.742 [if female], where serum creatinine is measured in mg/dL. CKD was defined as eGFR < 60 mL/min/1.73 m 2 the presence of renal damage markers (such as proteinuria), or both, persisting for at least 3 months, regardless of the underlying etiology 22 . In line with previous publications, NAFLD was defined by hepatic steatosis index (HSI) and US fatty liver index (USFLI). The formulars are as follows: HSI = 8 × (alanine aminotransferase/aspartate aminotransferase ratio) + body mass index (+2 for female; +2 for diabetes); USFLI = (e−0.8073 × Non−Hispanic Black+0.3458×Mexican American+0.0093×Age+0.6151 × loge (Gamma glutamyltransferase) +0.0249 × Waist Circumference+1.1792 × loge (Insulin)+0.8242 × loge (Glucose)−14.7812)/ (1 + e−0.8073 × Non−Hispanic Black+0.3458 ×Mexican American+0.0093 × Age+0.6151 × loge (Gamma glutamyltransferase) +0.0249 × waist circumference+1.1792 × loge (Insulin)+0.8242 × loge (Glucose) – 14.7812) × 100. USFLI cutoff value ≥ 30 or HSI value > 36 was diagnosed as NAFLD 23 . ALD was defined by a combination of an evidence of excessive alcohol consumption (≥ 210 g/week for men and ≥ 140 g/week for women) and an ALD/NAFLD index > 0, which was calculated as: −58.5 + 0.637 (Mean Corpuscular Volume) + 3.91 (Aspartate Aminotransferase [AST]/Alanine Aminotransferase [ALT]) − 0.406 (Body Mass Index) + 6.35 for Male Gender. In the subpopulation with ALD, the AST-to-platelet ratio index (APRI) and FIB-4 score were used to evaluate ALD FIB. The formula is as follows: APRI = (AST/Upper Limit of Normal/Platelet Count [109/L]) × 100, where the upper limits of normal AST levels were set at 37 IU/L for men and 29 IU/L for women; FIB-4 = Age × AST/ [Platelets in 109/L × (ALT)1/2] Cut-off values for advanced fibrosis (≥ F3) were set at 1.5 for APRI and 3.25 for FIB-4 24,25 . HBV, HCV, HDV, and HIV-positive patients were determined based on antigen measurements and quantification of relevant viral DNA or RNA levels in NHANES laboratories. Hyperlipidemia was defined as fasting triglyceride values ≥ 200 ng/dl. Smokers were defined as individuals who have smoked more than 100 cigarettes in their lifetime and currently smoke on some days or every day, while nonsmokers are those who have smoked less than 100 cigarettes in their lifetime 26 . Hyperuricemia was defined dichotomously with serum uric acid (SUA) levels ≥416μmol/L (7.0 mg/dL) for males and ≥357μmol/L (6.0 mg/dL) for females 27 . Apart from the diseases mentioned above, data on Heart attack, Heart disease, Heart failure, Stroke, Osteoporosis, Thyroid disease, Emphysema, Arthritis, Depression, Cancer, Chronic bronchitis, and Asthma were obtained from questionnaire data provided by NHANES. These indices were derived from comprehensive full blood cell count tests, the details of which can be found in the 'Questionnaire' data within the NHANES dataset. Finally, we categorized these diseases into seven major classes based on disease type: Circulatory system diseases, Digestive system diseases, Endocrine/Metabolic diseases, Immune system diseases, Respiratory system diseases, Urinary system diseases, and Others. The Dietary Inflammatory Index (DII) assesses the inflammatory effect of diet using 45 dietary parameters, normalizing individual intake of each food parameter to global intake. Standardized intake scores (Z-scores) are converted to proportions and centered. The centered proportions of these specific food intakes are multiplied by their inflammation effect scores and summed to obtain an individual's overall DII score. Participants' DII scores represent the sum of each DII score. Higher DII scores indicate a pro-inflammatory diet, while lower scores indicate an anti-inflammatory diet. In this study, 28 out of 45 food parameters were utilized for DII calculation: carbohydrates, protein, total fat, alcohol, fiber, cholesterol, saturated fatty acids, monounsaturated fatty acids, polyunsaturated fatty acids, n-3 fatty acids, n-6 fatty acids, niacin, vitamin A, thiamine, vitamin B2, vitamin B6, vitamin B12, vitamin C, vitamin D, vitamin E, iron, magnesium, zinc, selenium, folic acid, carotene, caffeine, and energy 28,29 . 2.4. Covariates We included C-reactive protein (CRP) and the systemic immune-inflammation index (SII) as covariates in our analysis. We selected and utilized data on platelet count (PC), neutrophil count (NC), and lymphocyte count (LC) in the computation, with SII calculated as SII = PC * (NC / LC) 30 . Other covariates included Age, Poverty Income Ratio (PIR), Body Mass Index (BMI), Gender, Race, Education, Smoking Exposure, and Alcohol Exposure. BMI was categorized into three groups: normal (BMI 30 kg/m²), based on participants' BMI values 31 . Alcohol consumption was assessed using data from NHANES questionnaires. Participants who had consumed fewer than 12 alcoholic drinks in their lifetime were classified as non-drinkers. Former drinkers were individuals who had consumed ≥12 drinks at any point in their lifetime but had not consumed alcohol in the past year. To minimize recall bias, smoking exposure was evaluated based on serum cotinine levels rather than relying solely on the 'smoking history questionnaire.' Current smokers were identified by serum cotinine levels > 10 ng/mg, former smokers had serum cotinine levels ≤ 10 ng/mg, and non-smokers exhibited serum cotinine levels < 0.011 ng/mg 32 . 2.5. Statistical Analysis In accordance with Centers for Disease Control and Prevention (CDC) guidelines, our statistical analyses adhered to stipulated principles. To address the complex multi-stage cluster survey design inherent to NHANES, appropriate sample weights were meticulously applied to each participant. Categorical variables were expressed as proportions, while continuous variables were presented as means (mean ± standard deviation). Descriptive statistics comprehensively summarized participants' demographic characteristics and biomarker concentrations. Specifically, for each selected PCB, analysis was conducted after stratifying into three groups based on quartiles, and for substances exhibiting highly right-skewed distributions, analysis was stratified into two groups based on the median. A total of 69 persistent organic pollutants (POPs) were initially identified in NHANES data (1999-2004), encompassing dioxins, furans, and coplanar polychlorinated biphenyls (PCBs). From these, 18 persistent organic pollutants with valid data representing 75% of the total data were selected for further analysis. Weighted quantile sum (WQS) regression, focusing on mortality risk, was performed on these 18 substances, with the top 7 substances selected based on their cumulative contribution rate to the preceding 95%, all of which were PCBs. Correlation heatmaps were employed to illustrate the interrelationships among the 7 PCBs and their associations with various diseases. Principal component analysis (PCA) was utilized to visualize the relationship between PCBs and mortality risk in two dimensions. Additionally, five diseases highly correlated with PCBs (Hyperuricemia, Hypertension, Diabetes, Chronic Kidney Disease (CKD), and Arthritis) were included in the machine learning algorithm model, while Cancer, Osteoporosis, and Hepatitis C (HCV) were excluded due to either their broad spectrum or insufficient data volume. Further, multivariate logistic regression was employed to evaluate the associations between these substances and 25 diseases, with results presented via circular plots. Adjustment for Age, Gender, Race, BMI, Social Inequality Index (SII), smoking exposure, and alcohol exposure was conducted. To illustrate the interrelationships among different diseases, logistic regression and comorbidity network analysis were performed on the 25 diseases, with the most relevant diseases determined based on Odds Ratios (ORs) and p-values, and results presented via network analysis diagrams. To further demonstrate PCBs as independent risk factors for mortality, excluding the influence of confounding factors such as diseases, we constructed five individual models based on machine learning: Support Vector Machine (SVM), Naïve Bayes, Decision Tree (Tree), Stochastic Gradient Descent (SGD), Gradient Boosting Decision Tree (GBDT), and four composite models: Random Forest, Histogram Gradient Boosting Decision Tree (hist GBDT), Bagging, and Neural Network. Additionally, a Voting algorithm was constructed for result output, with results displayed via ROC curves and confusion matrices. Furthermore, by plotting survival curves and constructing multi-factor Cox regression models, we effectively demonstrated the ability of dietary habits, moderated by DII, to mitigate the adverse effects of environmental factors such as PCBs on human health within high PCB-exposed populations. Results 3.1. Unveiling the Relationship between PCBs and Diseases and Mortality Risk Table 1 and Supplementary Table 1 present the essential characteristics of our study cohort. The mean age of participants was 40.99 years. Regarding gender distribution, females slightly outnumbered males, constituting 50.2% compared to 49.8%. Additionally, we categorized the 10,961 participants into three groups - low, moderate, and high - based on tertiles reflecting PCB concentrations. Notably, individuals in the high LBX074 group displayed distinct characteristics compared to those in the low and moderate groups (Table 1). Specifically, the high LBX074 group exhibited a higher proportion of females, a significantly greater mean age, and a substantially elevated percentage of active smokers. Similar trends were observed for other comprehensive PCB data, as outlined in Supplementary Table 1. Figures 1a and 1b depict the results of weighted quantile sum (WQS) and weighting analysis for the 18 PCBs and 25 diseases. By calculating the cumulative contribution rate, we identified the top 7 substances, ranked in descending order based on their contribution rate, accounting for 95% of the cumulative contribution rate: LBX074, representing PCB074(denoting 2,4,4',5-Tetrachlorobiphenyl). LBX170, representing PCB170(2,2',3,3',4,4',5-Heptachlorobiphenyl). LBX178, representing PCB178(2,2’,3,3’,5,5’,6-Heptachlorobiphenyl). LBX180, representing PCB180(2,2’,3,4,4’,5,5’-Heptachlorobiphenyl). LBX156, representing PCB156(2,3,3',4,4',5-Hexachlorobiphenyl). LBX157, representing PCB157(2,3,3',4,4',5'-Hexachlorobiphenyl). LBX146, representing PCB146(2,3,3',4,4',5'-Hexachlorobiphenyl). In line with this, the main text primarily focused on LBX074 data, aimed at representing this category of substances, while additional data were predominantly included in the supplementary materials. Figures 1c and 1d illustrate, in the form of heatmaps, the interrelationships among the 7 PCBs and between PCBs and the 25 diseases, respectively. Significant positive correlations were observed among all 7 PCBs (p < 0.0001), with LBX146 exhibiting the strongest correlation with the other 6 substances. The 7 PCBs showed significant positive correlations with most diseases (p < 0.0001). However, all 7 PCBs were negatively correlated with Asthma, and LBX178 and LBX167 were negatively correlated with Chronic bronchitis and Hyperlipidemia. Notably, Alcoholic fatty liver showed almost no correlation with any of the PCBs. These findings suggested that PCBs might be important environmental factors associated with the malignant progression of most diseases, affecting multiple systems in the body. However, due to their different mechanisms of action, an increase in PCBs within a certain range might partially inhibit the progression of specific diseases. Figure 2 employed principal component analysis (PCA) to illustrate the relationship between PCBs and the risk of death in a two-dimensional format. Pairwise combinations of the 7 PCBs were analyzed, and the PCA two-dimensional plot indicates that for each pair of PCBs, higher PCB levels correspond to increased risk of death. Thus, all 7 PCBs are important factors in increasing the mortality risk 3.2. Comorbidity Network Analyses: Demonstrating Interactions Between Diseases Multivariable logistic regression was employed to analyze the summarized 25 diseases, adjusting for Age, Gender, Race, BMI, SII, smoking exposure, and alcohol exposure. Among them, 12 diseases showed significant associations with PCBs (p1, p6.0, p6.0, p6, p=0.0071), Hyperlipidemia (OR>6, p=0.016), HIV (OR>6, p=0.0005), and Arthritis (OR=5.42, p=0.0065). Adjusting for covariates, it can be demonstrated that PCBs are independent influencing factors for multiple diseases. Logistic regression was used to compute the odds ratio (OR) and p-values for each pair of the 25 diseases to determine disease associations. We calculated the OR values for all disease pairs (see supplementary table 2) and presented the disease pairs most closely related to each disease (p<0.05) (see supplementary table 3). Among these pairs, Depression and Stroke emerged as the most correlated diseases (OR (95%CI): 40.21(5.83,794.00), p=0.001), while the association between Thyroid disease and Angina was the least significant (OR (95%CI): 3.60(2.76,4.66), p<0.0001). Diseases of the circulatory system tend to be interrelated, with diseases highly correlated with diabetes and proteinuria, typical of the urinary system, accounting for the majority of associations. Depression is strongly associated with some common chronic clinical conditions. Based on comorbidity network analysis, we identified 283 potential links, with the number of related links, OR values, and comorbidity network analysis results shown in Figure 3b. Each node represents a medical condition, and the thickness of the connecting lines reflects the strength of the disease pairs' association. Nodes closer to the network center have stronger centrality, indicating a greater number of connections with other diseases. Hypertension exhibits the strongest centrality and the highest number of connections, significantly influencing most other diseases. It is also notable that diseases of the circulatory system are often closely associated with other diseases. Depression, ALD, HIV, and HPL are distanced from the center, showing fewer associations with other diseases and smaller impact. 3.3. Unveiling the Association between PCBs with Mortality Using Machine Learning, introducing DII for correction. To further mitigate the influence of confounding factors on the analysis of the association between PCBs and mortality risk, we constructed nine learning models, including five individual models: Support Vector Machine (SVM), Naïve Bayes, Decision Tree (Tree), Stochastic Gradient Descent (SGD), and Gradient Boosting Decision Tree (GBDT). Additionally, we developed four ensemble models: Random Forest, Histogram Gradient Boosting Decision Tree (hist GBDT), Bagging, and Neural Network. Furthermore, we incorporated a Voting algorithm for result output. Two types of models were added to the algorithm: one containing Age, gender, race, hyperuricemia, hypertension, diabetes, CKD, and arthritis, and the other adding PCBs data. The AUC values of these two types of models, representing predictive accuracy, were compared (Figure 4a, Figure 4b). It is evident that models incorporating PCBs data exhibited varying degrees of improvement in accuracy compared to the baseline models. Overall, ensemble models outperformed individual models, with Random Forests showing significant advantages in prediction. Prior to incorporating PCBs data, the accuracy rates were as follows: SVM (0.86), SGD (0.88), Naïve Bayes (0.85), Decision Tree (0.88), GBDT (0.91), hist GBDT (0.89), Random Forests (0.98), Bagging (0.92), Neural Network (0.91), and the final Voting model (0.94). After incorporating PCBs data, the accuracy rates were: SVM (0.89), SGD (0.90), Naïve Bayes (0.86), Decision Tree (0.90), GBDT (0.94), hist GBDT (0.91), Random Forests (1.0), Bagging (0.95), Neural Network (0.95), and the final Voting model (0.96). Figures 4c and 4d illustrate the comparison of ROC curves before and after applying the Voting algorithm. Figures 4e and 4f illustrate the comparison of the confusion matrices for the Voting algorithm before and after. These results strongly demonstrate a substantive positive correlation between PCBs and mortality risk, even after controlling for baseline data and confounding factors such as diseases. The ROC curves and confusion matrices of the other algorithms are presented in Supplementary Figure 4. Furthermore, in Figure 5, we demonstrated through survival curves that we could mitigate the impact of DII by adjusting lifestyle habits. We first continued to utilize the previously built machine learning models to calculate the accuracy and errors of the predictive model containing population baseline data (Age, gender, race) and five diseases (hyperuricemia, hypertension, diabetes, CKD, arthritis), along with seven PCBs data (Figure 5a). Furthermore, we incorporated DII data of each participant into the aforementioned model for prediction and recalculated the accuracy and errors of the new model (Figure 5b). The results indicated that the DII index effectively enhanced the accuracy of the model, highlighting its significance as a contributing factor to increased mortality risk. In Figure 5c, participants were categorized into low, medium, and high groups based on the quartiles of the total amount of 7 PCBs in their bodies, and survival curves were plotted accordingly. Participants with high levels of PCBs showed a significantly increased risk of mortality, while participants with medium and low levels exhibited progressively lower risks of mortality (p<0.0001). In Figure 5d, participants with high levels of PCBs were further classified based on their DII scores: those with DII scores greater than 0 were defined as "positive," while those with DII scores less than or equal to 0 were defined as "negative." In comparison with Figure 5c, it is evident that a DII score greater than 0 (indicating reduced inflammation) effectively reduces the risk of mortality among participants with higher levels of PCBs. Figures 5e and 5f depict similar classifications for participants with medium and low levels of PCBs, respectively. However, the conclusions drawn from participants with high levels of PCBs were not statistically significant in these two groups, indicating no significant differences. In Figures 5g, h, and i, we further constructed multifactor Cox regression models to examine the association between DII index and mortality, adjusting for age, gender, race, education, BMI, smoking exposure, and SII covariates. Among the high PCBs population, a DII index less than 1 (indicating inflammation suppression) was significantly negatively correlated with mortality (OR (95% CI) = 0.7151 (0.6683, 0.7651), p < 0.05). However, in the medium PCBs and low PCBs populations, the association between a DII index less than 1 and mortality was not statistically significant (OR (95% CI) = 1.05689 (0.94773, 1.1786), p > 0.05) (OR (95% CI) = 1.125 (0.87141, 1.4525), p > 0.05). Through the above analysis, we effectively demonstrated that among the high PCBs population, modulation of dietary habits and adjustment of DII can effectively counteract the adverse effects of PCBs and similar environmental factors on human health. However, in the medium PCBs and low PCBs populations, the impact of DII adjustment on mortality is not evident. Discussion Our study revealed profound associations between PCBs and disease networks, as well as mortality risk. We delved into the interplay among various diseases and proposed, for the first time, that mitigating mortality risk and potentially alleviating the impact of environmental factors could be achieved through controlling DII. Taking LBX074 (2,4,4,5-Tetrachlorobiphenyl) as an example, LBX074 exhibited significant positive correlations with 7 diseases (OR>1, p<0.05). PCA revealed a notable increase in mortality risk with increasing levels of LBX074, a trend observed in other PCBs as well. Depression and Stroke emerged as the most relevant disease pair in network analysis (OR (95%CI): 40.21 (5.83, 794.00), p=0.001). When exploring PCBs' independent role in mortality risk through machine learning models, we controlled for confounding factors and comorbidities, further confirming PCBs as independent risk factors for mortality. The feasibility of reducing mortality risk by lowering DII was validated through survival curve plotting and multi-factor Cox regression analysis. Although the specific pathogenic mechanisms linking coplanar polychlorinated biphenyls (PCBs) to mortality rates remain unclear, our study elucidated associations between PCBs and the development of various diseases, shedding light on the factors contributing to increased mortality rates associated with coplanar PCB exposure. To the best of our knowledge, our study possesses several notable strengths. It is the first to rectify the limitations of traditional methods using artificial intelligence-based big data models to predict mortality risk associated with PCB exposure and evaluate the impact of DII on mortality risk using machine learning methods. Our findings corroborate those of traditional data analysis, with the addition of PCB data significantly enhancing the predictive accuracy of multiple models compared to those without PCB data, suggesting that various PCBs are independent influencing factors on mortality. Additionally, precise measurements of accuracy and errors through multiple iterations provided further evidence of the detrimental effects of pro-inflammatory diets on the body. Traditional statistical methods have limitations in analyzing mortality risk, as they overlook issues of shared exposure and multi-factorial risk confounding, and are unable to effectively address statistical errors. With the further development of artificial intelligence, machine learning-based algorithm models can effectively address these issues, with algorithmic results becoming increasingly accurate over time, capable of handling various data formats in dynamic, large-volume, and complex data environments. Therefore, our study, through further data screening and the construction of predictive models using machine learning algorithms, analyzed the impact of PCBs on mortality risk. Moreover, this study represents the first attempt to comprehensively investigate the combined effects of PCBs on various diseases and comorbidity networks using comorbidity network analysis. Logistic regression was employed to calculate OR values, while centrality and associated nodes were demonstrated through comorbidity network visualization. Consistent with past research, we found significant positive correlations between PCBs and various diseases such as Hyperuricemia, Diabetes, and Hyperlipidemia. Additionally, the comorbidity network analysis indicated that hypertension is a significant trigger for multiple systemic diseases, with circulatory system diseases often closely associated with various other systemic diseases, exhibiting the strongest centrality. Furthermore, through comorbidity network analysis, we first discovered significant positive correlations between PCB levels and HCV, HIV, and arthritis, likely attributable to PCB-induced inflammation, immune suppression, and apoptosis induction in cartilage cells via ROS-dependent pathways. While the specific mechanisms by which PCBs contribute to various diseases and mortality remain unclear, it is undeniable that PCBs pose significant hazards to human health, serving as independent risk factors for multiple diseases and mortality. The impact weight of PCBs is higher than that of some conventional detection substances, suggesting that PCBs may serve as specific biomarkers for certain diseases, aiding in disease prediction in the future. Moreover, further research is needed on the effects of environmental pollutants on human health. Furthermore, we have introduced for the first time the concept that adverse effects of pollutants can potentially be counteracted by altering dietary habits. Previous research indicates that higher levels of inflammation lead to an increased risk of mortality. Our study further demonstrates the significant role of DII in triggering inflammation and oxidative stress in the disease and mortality processes. Among populations with high PCB exposure, significant reductions in mortality and morbidity risks can be achieved through DII regulation. However, similar effects were not significant among populations with moderate or low PCB exposure. This suggests that PCBs may induce inflammation to a certain threshold, and DII regulation can effectively suppress their effects. Numerous studies have shown that diet, as the main source of bioactive compounds, can mediate inflammatory responses, with pro-inflammatory diets associated with increased white blood cell counts. Pro-inflammatory diets exhibit significant positive correlations with various diseases, including chronic obstructive pulmonary disease, diabetes, depression, and cardiovascular diseases, while high pro-inflammatory diets can increase the risk of mortality, possibly by increasing white blood cell and CRP levels, thereby inducing various diseases leading to mortality. One possible mechanism is the close relationship between diet and the human gut microbiota. Several animal studies have shown that high-sugar diets lead to obesity, insulin resistance, increased intestinal permeability, and low-grade inflammation. Microbial metabolites (such as SCFA butyrates or tryptophan metabolites) can control various physiological functions in the host, ranging from inflammatory responses to energy metabolism in epithelial cells. Bifidobacteria, Lactobacilli, Clostridia, Bacillus subtilis, and fragile bacilli are closely related to specific immunity via MyD88, transforming growth factor-β, IL-1, IL-6, IL-17, IL-22, γ-PgA, and PSA. Fragile bacilli, plant bifidobacteria, and bifidobacteria can regulate inflammatory responses via TLR, NF-κB, and MyD88. Inflammation is closely related to diseases, and therefore high DII can induce diseases by triggering inflammation, while low DII has the opposite effect. However, specific hypotheses cannot be tested in current studies. Therefore, future longitudinal studies could consider the potential mechanisms by which diet-driven inflammation induces mortality or disease. Similarly, future research could determine whether the use of anti-inflammatory diets (such as increasing leafy vegetables, herbs, spices, and certain fruits) can reduce WBC and CRP levels, decrease morbidity, and reduce mortality risk. However, this study also has some limitations. Firstly, due to insufficient data, more substances were not studied. Secondly, the mechanisms by which they cause multiple diseases leading to increased mortality rates remain unknown. Additionally, DII was not compared with energy-adjusted DII (E-DII), which constructs a reference database for energy-adjusted nutritional scoring based on data from the same 11 countries used to calculate DII. Without access to the unique comparison database, E-DII cannot be calculated, so we were unable to compare it in our study. Furthermore, the ubiquity and complexity of exposure not only necessitate further research on the effects of PCBs but also require further investigation into the prevention and monitoring of PCBs, which may help clinicians better understand and control exposure levels of these organic pollutants. Finally, when using the NHANES database for statistical analysis, we selected multiple variables. Indeed, when lots of variables are tested, associations flourish, most are due to chance, some are merely markers, some are due to common non-investigated factors, and just a few are causal. The non-longitudinal nature of these surveys is not helpful to discern whether the statistical associations are meaningful enough. Therefore, we cannot avoid the analysis bias caused by large databases. Conclusion After employing machine learning methods to better mitigate the impact of confounding factors, we observed a strong positive correlation between PCBs and diseases across multiple systems, indicating that PCBs may be independent risk factors for various diseases and even mortality. This suggests that PCBs may be intimately involved in the development and progression of multiple diseases. By constructing multidimensional machine learning models and conducting multiple iterations for precision and error measurement, PCBs may have the potential to become specific biomarkers for certain diseases in the future. Furthermore, we observed that in populations with high PCB exposure, significant reductions in mortality and morbidity risks can be achieved by adjusting DII, which is of great significance. The impact of PCBs on human health may be achieved through the induction of systemic inflammation, causing damage to multiple systems in the body, and improving daily dietary habits is an effective solution. However, a more comprehensive analysis is still lacking, and further research is needed to determine the specific mechanisms associated with its processes. Declarations Author Contributions: This article was written by Ying Gao, Han Lu, Huan Zhou, and Jiaxing Tan. Ying Gao, Han Lu, and Jiaxing Tan were responsible for data completion and interpretation. Jiaxing Tan and Han Lu contributed to the study design and analysis methods. Ying Gao and Huan Zhou participated in data collection and validation, and completed the writing of the entire article. Jiaxing Tan supervised the research and provided guidance throughout the study. Competing Interest Statement: The authors hereby verify the absence of any financial or personal affiliations that may have the potential to impact the credibility of the findings articulated in this paper. Declaration of generative AI and AI-assisted technologies in the writing process In the course of preparing this manuscript, the authors employed ChatGPT to enhance language usage. Subsequent to utilizing this tool, the authors meticulously reviewed and edited the content as necessary, assuming full responsibility for the final publication's content. Acknowledgments: This study was partly supported by grants from the project of the National Natural Science Foundation of China (No. 81970612 and No. 82300797) and Innovation and Entrepreneurship Training Program for College Students. Data availability ： The datasets generated and analysed during the current study are available in the NHANES repository, https://www.cdc.gov/nchs/nhanes. References Lin, Y. S. et al. Environmental exposure to dioxin-like compounds and the mortality risk in the U.S. population. Int J Hyg Environ Health 215 , 541-546 (2012). https://doi.org:10.1016/j.ijheh.2012.02.006 Wu, B. et al. Combined exposure to multiple dioxins and dioxin-like polychlorinated biphenyls on hypertension among US adults in NHANES: a cross-sectional study under three statistical models. Environ Sci Pollut Res Int 30 , 28730-28744 (2023). https://doi.org:10.1007/s11356-022-24271-3 Bouchard, M. F., Oulhote, Y., Sagiv, S. K., Saint-Amour, D. & Weuve, J. Polychlorinated biphenyl exposures and cognition in older U.S. adults: NHANES (1999-2002). Environ Health Perspect 122 , 73-78 (2014). https://doi.org:10.1289/ehp.1306532 Van Oostdam, J. et al. Human health implications of environmental contaminants in Arctic Canada: a review. Sci Total Environ 230 , 1-82 (1999). https://doi.org:10.1016/s0048-9697(99)00036-4 Helou, K., Harmouche-Karaki, M., Karake, S. & Narbonne, J. F. A review of organochlorine pesticides and polychlorinated biphenyls in Lebanon: Environmental and human contaminants. Chemosphere 231 , 357-368 (2019). https://doi.org:10.1016/j.chemosphere.2019.05.109 Chen, L. et al. Polychlorinated biphenyls in the atmosphere of an urban city: levels, distribution, and emissions. Arch Environ Contam Toxicol 57 , 437-446 (2009). https://doi.org:10.1007/s00244-009-9293-z Zhu, M. et al. Environmental contamination and human exposure of polychlorinated biphenyls (PCBs) in China: A review. Sci Total Environ 805 , 150270 (2022). https://doi.org:10.1016/j.scitotenv.2021.150270 Androutsopoulos, V. P., Hernandez, A. F., Liesivuori, J. & Tsatsakis, A. M. A mechanistic overview of health associated effects of low levels of organochlorine and organophosphorous pesticides. Toxicology 307 , 89-94 (2013). https://doi.org:10.1016/j.tox.2012.09.011 Raffetti, E., Donat-Vargas, C., Mentasti, S., Chinotti, A. & Donato, F. Association between exposure to polychlorinated biphenyls and risk of hypertension: A systematic review and meta-analysis. Chemosphere 255 , 126984 (2020). https://doi.org:10.1016/j.chemosphere.2020.126984 Silverstone, A. E. et al. Polychlorinated biphenyl (PCB) exposure and diabetes: results from the Anniston Community Health Survey. Environ Health Perspect 120 , 727-732 (2012). https://doi.org:10.1289/ehp.1104247 Assaggaf, H. et al. Polychlorinated Biphenyls and Pulmonary Hypertension. Int J Environ Res Public Health 19 (2022). https://doi.org:10.3390/ijerph19084705 Deng, P. et al. Hepatic metabolomics reveals that liver injury increases PCB 126-induced oxidative stress and metabolic dysfunction. Chemosphere 217 , 140-149 (2019). https://doi.org:10.1016/j.chemosphere.2018.10.196 Serdar, B., LeBlanc, W. G., Norris, J. M. & Dickinson, L. M. Potential effects of polychlorinated biphenyls (PCBs) and selected organochlorine pesticides (OCPs) on immune cells and blood biochemistry measures: a cross-sectional assessment of the NHANES 2003-2004 data. Environ Health 13 , 114 (2014). https://doi.org:10.1186/1476-069x-13-114 Przybyla, J., Houseman, E. A., Smit, E. & Kile, M. L. A path analysis of multiple neurotoxic chemicals and cognitive functioning in older US adults (NHANES 1999-2002). Environ Health 16 , 19 (2017). https://doi.org:10.1186/s12940-017-0227-3 Sasaki, N., Morse, G., Jones, L. & Carpenter, D. O. Effects of mixtures of polychlorinated biphenyls (PCBs) and three organochlorine pesticides on cognitive function differ between older Mohawks at Akwesasne and older adults in NHANES. Environ Res 236 , 116861 (2023). https://doi.org:10.1016/j.envres.2023.116861 Axelrad, D. A., Goodman, S. & Woodruff, T. J. PCB body burdens in US women of childbearing age 2001-2002: An evaluation of alternate summary metrics of NHANES data. Environ Res 109 , 368-378 (2009). https://doi.org:10.1016/j.envres.2009.01.003 Mitro, S. D., Birnbaum, L. S., Needham, B. L. & Zota, A. R. Cross-sectional Associations between Exposure to Persistent Organic Pollutants and Leukocyte Telomere Length among U.S. Adults in NHANES, 2001-2002. Environ Health Perspect 124 , 651-658 (2016). https://doi.org:10.1289/ehp.1510187 Morgan, M., Deoraj, A., Felty, Q. & Roy, D. Environmental estrogen-like endocrine disrupting chemicals and breast cancer. Mol Cell Endocrinol 457 , 89-102 (2017). https://doi.org:10.1016/j.mce.2016.10.003 Lan, T., Liu, B., Bao, W. & Thorne, P. S. Identification of PCB congeners and their thresholds associated with diabetes using decision tree analysis. Sci Rep 13 , 18322 (2023). https://doi.org:10.1038/s41598-023-45301-1 Casagrande, S. S., Lee, C., Stoeckel, L. E., Menke, A. & Cowie, C. C. Cognitive function among older adults with diabetes and prediabetes, NHANES 2011-2014. Diabetes Res Clin Pract 178 , 108939 (2021). https://doi.org:10.1016/j.diabres.2021.108939 Wu, M., Si, J., Liu, Y., Kang, L. & Xu, B. Association between composite dietary antioxidant index and hypertension: insights from NHANES. Clin Exp Hypertens 45 , 2233712 (2023). https://doi.org:10.1080/10641963.2023.2233712 Tan, J., Liu, N., Sun, P., Tang, Y. & Qin, W. A Proinflammatory Diet May Increase Mortality Risk in Patients with Diabetes Mellitus. Nutrients 14 (2022). https://doi.org:10.3390/nu14102011 Li, W. et al. Analysis of environmental chemical mixtures and nonalcoholic fatty liver disease: NHANES 1999-2014. Environ Pollut 311 , 119915 (2022). https://doi.org:10.1016/j.envpol.2022.119915 Li, L. et al. The Association between Non-Alcoholic Fatty Liver Disease (NAFLD) and Advanced Fibrosis with Serological Vitamin B12 Markers: Results from the NHANES 1999-2004. Nutrients 14 (2022). https://doi.org:10.3390/nu14061224 Yang, Z. et al. Association between daidzein intake and metabolic associated fatty liver disease: A cross-sectional study from NHANES 2017-2018. Front Nutr 10 , 1113789 (2023). https://doi.org:10.3389/fnut.2023.1113789 Li, Z. et al. Distribution of lipid levels and prevalence of hyperlipidemia: data from the NHANES 2007-2018. Lipids Health Dis 21 , 111 (2022). https://doi.org:10.1186/s12944-022-01721-y Tan, Y. et al. Relationship between phthalates exposures and hyperuricemia in U.S. general population, a multi-cycle study of NHANES 2007-2016. Sci Total Environ 859 , 160208 (2023). https://doi.org:10.1016/j.scitotenv.2022.160208 Liu, H. et al. Association Between Diet-Related Inflammation and COPD: Findings From NHANES III. Front Nutr 8 , 732099 (2021). https://doi.org:10.3389/fnut.2021.732099 Zhao, L., Sun, Y., Liu, Y., Yan, Z. & Peng, W. A J-shaped association between Dietary Inflammatory Index (DII) and depression: A cross-sectional study from NHANES 2007-2018. J Affect Disord 323 , 257-263 (2023). https://doi.org:10.1016/j.jad.2022.11.052 Mahemuti, N. et al. Association between Systemic Immunity-Inflammation Index and Hyperlipidemia: A Population-Based Study from the NHANES (2015-2020). Nutrients 15 (2023). https://doi.org:10.3390/nu15051177 He, K., Pang, T. & Huang, H. The relationship between depressive symptoms and BMI: 2005-2018 NHANES data. J Affect Disord 313 , 151-157 (2022). https://doi.org:10.1016/j.jad.2022.06.046 Zhang, Y. B. et al. Associations of healthy lifestyle and socioeconomic status with mortality and incident cardiovascular disease: two prospective cohort studies. Bmj 373 , n604 (2021). https://doi.org:10.1136/bmj.n604 Table 1 Table 1. Baseline demographic characteristics divided by LBX074 levels. Overall LBX074 p-Value Low Moderate High Number 10961 4015 3304 3642 Age 40.99 (21.36) 25.80 (12.77) 37.78 (18.11) 61.98 (14.25) <0.001 Gender (%) <0.001 Male 5457 (49.8) 2128 (53.0) 1847 (55.9) 1482 (40.7) Female 5504(50.2) 1887(47.0) 1457(44.1) 2160(59.3) Race (%) <0.001 Mexican American 2692 (24.6) 1351 (33.6) 824 (24.9) 517 (14.2) Non-Hispanic Black 2353 (21.5) 1021 (25.4) 679 (20.6) 653 (17.9) Non-Hispanic White 5069 (46.2) 1295 (32.3) 1484 (44.9) 2290 (62.9) Other Hispanic 438 (4.0) 176 (4.4) 161 (4.9) 101 (2.8) Other races 409 (3.7) 172 (4.3) 156 (4.7) 81 (2.2) Education (%) <0.001 Less than 9th grade 1530 (14.4) 600 (15.4) 439 (13.7) 491 (13.8) 9-11th grade 1942 (18.2) 829 (21.3) 533 (16.6) 580 (16.3) High-school graduate 2454 (23.0) 889 (22.9) 748 (23.3) 817 (23.0) College graduate or above 1995 (18.7) 581 (14.9) 673 (20.9) 741 (20.9) Some college or AA degree 2679 (25.1) 968 (24.9) 810 (25.2) 901 (25.4) others 56 (0.5) 20 (0.5) 13 (0.4) 23 (0.6) PIR 2.50 (1.60) 2.17 (1.56) 2.61 (1.63) 2.77 (1.56) <0.001 BMI stage (%) <0.001 30 2776 (26.3) 724 (18.5) 852 (26.4) 1200 (35.1) 25-30 3300 (31.2) 1035 (26.4) 1020 (31.5) 1245 (36.5) Smoking exposure (%) <0.001 Current smoker 2596 (23.8) 934 (23.4) 938 (28.6) 724 (20.0) Former smoker 6646 (61.0) 2529 (63.2) 2003 (61.1) 2114 (58.5) Non smoker 1649 (15.1) 537 (13.4) 338 (10.3) 774 (21.4) Alcohol intake (%) <0.001 Current drinker 497 (21.0) 136 (23.7) 128 (22.0) 233 (19.3) Former drinker 696 (29.4) 107 (18.6) 170 (29.2) 419 (34.7) Non drinker 1173 (49.6) 331 (57.7) 285 (48.9) 557 (46.1) SII 582.86 (367.92) 557.14 (340.40) 583.96 (395.89) 610.40 (368.91) <0.001 CRP 0.38 (0.81) 0.26 (0.56) 0.38 (0.93) 0.52 (0.89) <0.001 WBC 7.09 (2.18) 7.04 (2.03) 7.18 (2.12) 7.07 (2.39) 0.018 Lymphocyte 2.13 (1.01) 2.18 (0.62) 2.15 (0.67) 2.07 (1.49) <0.001 Monocyte 0.56 (0.18) 0.55 (0.18) 0.56 (0.18) 0.57 (0.19) <0.001 Neutrophils 4.16 (1.65) 4.08 (1.68) 4.22 (1.75) 4.18 (1.51) <0.001 Platelet count 272.08 (67.52) 277.92 (63.12) 274.24 (72.13) 263.63 (67.08) <0.001 Red blood cell 4.74 (0.50) 4.81 (0.49) 4.79 (0.48) 4.60 (0.51) <0.001 Hemoglobin 14.28 (1.51) 14.38 (1.53) 14.43 (1.55) 14.05 (1.40) <0.001 Alkaline phosphatase 90.57 (62.34) 102.04 (76.08) 94.53 (68.44) 74.22 (24.98) <0.001 Albumin 4.32 (0.33) 4.38 (0.29) 4.39 (0.36) 4.20 (0.30) <0.001 Bilirubin 0.72 (0.29) 0.74 (0.30) 0.71 (0.33) 0.71 (0.25) <0.001 Iron 88.34 (37.89) 90.12 (40.00) 90.86 (38.84) 84.06 (34.05) <0.001 The baseline table is a calculation from NHANES 1999-2000,2001-2002,2003-2004. For categorical variables, the p-value was calculated by the chi-square test. For continuous variables, the p-value was calculated by t-test. Additional Declarations No competing interests reported. Supplementary Files SupplementaryMaterialslegends.docx SupplementaryFigure1.docx SupplementaryTable1.docx SupplementaryTable2.docx SupplementaryTable3.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4543285","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":317889754,"identity":"a27be536-bb9d-4020-87eb-709b26d25086","order_by":0,"name":"Ying Gao","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Ying","middleName":"","lastName":"Gao","suffix":""},{"id":317889755,"identity":"979ad798-dc3f-48eb-b06b-a535c534ac59","order_by":1,"name":"Han Lu","email":"","orcid":"","institution":"Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Han","middleName":"","lastName":"Lu","suffix":""},{"id":317889756,"identity":"dc033942-8859-43ac-892f-f037b82336f7","order_by":2,"name":"Huan Zhou","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Huan","middleName":"","lastName":"Zhou","suffix":""},{"id":317889757,"identity":"d4785084-1a17-4745-a204-744862b97b45","order_by":3,"name":"Jiaxing Tan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1UlEQVRIiWNgGAWjYDACZgglx94AogwsiNZiYMxzAExLEG2XQWIPWAsDEVoMjvMefvHhz5/0HvYe0w0/CiQY+Nu7E/BqkWzmS7Oc2WaQ28NzLO1mD9BhEmfObsCrhZ+Zx8yYt8Egd79E8rEbPEAtBhK5+LWwgbT8+WOQziOR2HbzDzFagLYYP2ZgM0jgAdpymyhbJJt5zBh724wNQX65LWMgwUPQLwbnzxh/+PFHTp6Hvcfs5ps/NnL87b34tYC8gxIXPISUgwDzB2JUjYJRMApGwQgGANjgP6Zk4Y7YAAAAAElFTkSuQmCC","orcid":"","institution":"Pennsylvania State University College of Medicine","correspondingAuthor":true,"prefix":"","firstName":"Jiaxing","middleName":"","lastName":"Tan","suffix":""}],"badges":[],"createdAt":"2024-06-07 03:51:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4543285/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4543285/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":58966062,"identity":"a65a442b-276e-4409-8861-9d9b13526f80","added_by":"auto","created_at":"2024-06-24 18:24:08","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":4131149,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSelection of PCBs and heat map analysis of the relationship between PCBs and disease\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea. \u003c/strong\u003eProgressive weighted quantile and regression (WQS) for POPs with a valid data volume greater than 75%. Seven PCBs with the top 95% cumulative weight were selected. \u003cstrong\u003eb.\u003c/strong\u003e The prevalence of each disease was ranked among all participants. \u003cstrong\u003ec.\u003c/strong\u003e Correlation heat map analysis of 7 types of PCBs. The graph shows that the content of each PCBs in the human body is highly positively correlated. \u003cstrong\u003ed.\u003c/strong\u003e Heatmap analysis of the correlation between 7 PCBs and diseases. Lesions were expressed as 1 and 0, both using Spearman correlation analysis. The images showed that the seven PCBs were significantly positively correlated with most diseases (p\u0026lt;0.0001), however, all seven PCBs were negatively correlated with Asthma, and LBX178 and LBX167 were negatively correlated with Chronic bronchitis and Hyperlipidemia. But alcoholic fatty liver is almost unrelated to all PCBs.\u003c/p\u003e\n\u003cp\u003e* 0.001 \u0026lt; P ≤ 0.05\u003c/p\u003e\n\u003cp\u003e** 0.0001 \u0026lt; P ≤ 0.01\u003c/p\u003e\n\u003cp\u003e*** P ≤ 0.0001\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-4543285/v1/2914aefc651867660c30ede2.png"},{"id":58965743,"identity":"26fec683-4375-4a90-825a-4cfe199b009d","added_by":"auto","created_at":"2024-06-24 18:16:08","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":10793277,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePrincipal Component Analysis (PCA) of 7 PCBs and Mortality Risk\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePrincipal component analysis was used to reduce the dimensionality of the seven-dimensional data that affects the mortality risk, i.e., seven kinds of PCBs, and the seven kinds of PCBs were paired in pairs into two-dimensional graphics. The images show that the higher the PCBs, the greater the mortality risk, and all seven PCBs are important factors that increase the mortality risk.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-4543285/v1/a94ec626b73f1c0383b585f9.png"},{"id":58965498,"identity":"fcbdbd69-7754-46d3-a103-b8d5e2e83286","added_by":"auto","created_at":"2024-06-24 18:08:09","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":2884031,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLogistic regression and comorbid network analysis. PCBs are independent influencing factors for a variety of diseases.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea.\u003c/strong\u003e Taking LBX074 as the representative of PCBs, multivariate logistic regression was performed with 25 diseases, and Age, Gender, Race, BMI, SII, smoking exposure and alcohol exposure were adjusted. The color of the inner circle was derived from the p-value obtained by logistic regression, and the color of the inner circle was gray when p\u0026lt;0.05, and the color of the inner circle was white when p\u0026gt;0.05. \u003cstrong\u003eb.\u003c/strong\u003e25 disease comorbidity network diagrams. Each node represents a disease, and the width of the link represents the strength of the comorbid association. According to the comorbid network analysis, we constructed 283 possible links, performed logistic regression on the disease pairs, and the OR values and the results of the comorbid network analysis were displayed in the figure. The closer you are to the center of the network, the more central it is, and the more nodes are associated with other diseases.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-4543285/v1/ed33e9b9254467293afd987f.png"},{"id":58965490,"identity":"67c705e0-b949-453e-bab4-cf3f038a6088","added_by":"auto","created_at":"2024-06-24 18:08:08","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1412705,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of the accuracy differences between different machine learning models before and after adding PCBs.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eClassifying machine learning models into single models and composite models, as well as a Voting model for the final output, radial histograms integrate data from ROC curves across multiple models \u003cstrong\u003ea.\u003c/strong\u003e The data models including population baseline data Age, gender, race and five diseases as hyperuricemia, hypertension, diabetes, CKD, and arthritis. \u003cstrong\u003eb.\u003c/strong\u003e The data models including population baseline data Age, gender, race and five diseases as hyperuricemia, hypertension, diabetes, CKD, and arthritis, adding seven PCBs. \u003cstrong\u003ec.\u003c/strong\u003e The ROC curve before the PCBs data is included in the Voting algorithm model. \u003cstrong\u003ed.\u003c/strong\u003e The ROC curve after the PCBs data is included in the Voting algorithm model. \u003cstrong\u003ee.\u003c/strong\u003e The Confusion matrix before the PCBs data is included in the Voting algorithm model. \u003cstrong\u003ef.\u003c/strong\u003e The Confusion matrix after the PCBs data is included in the Voting algorithm model.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-4543285/v1/f430515d5d8ee6e0981febb0.png"},{"id":58965745,"identity":"d0889b19-e6c0-4e77-845f-0dc83eaa2265","added_by":"auto","created_at":"2024-06-24 18:16:09","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":3312813,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSurvival curves of different categories of participants classified according to the total amount of 7 PCBs and the DII index.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea.\u003c/strong\u003e The accuracy of the machine learning prediction model when DII data is not included. The data models including population baseline data Age, gender, race and five diseases as hyperuricemia, hypertension, diabetes, CKD, arthritis, and seven PCBs. \u003cstrong\u003eb.\u003c/strong\u003e The accuracy of the machine learning prediction model when DII data is included. The data models including population baseline data Age, gender, race and five diseases as hyperuricemia, hypertension, diabetes, CKD, arthritis, and seven PCBs, adding DII index.\u003cstrong\u003e c.\u003c/strong\u003e Survival curves for 3 types of participants classified according to PCBs. The total amount of 7 PCBs in the participants was calculated and divided into three categories: \"Low\", \"Moderate\" and \"High\" according to the tripart. \u003cstrong\u003ed.\u003c/strong\u003e Survival curves of 2 types of participants classified according to the DII index. The DII index of the participants was calculated and the DII greater than 0 was defined as “positive” and the DII less than 0 was defined as “negative”, which was classified among the participants with Total PCBs as “High”. \u003cstrong\u003ee.\u003c/strong\u003e Survival curves of 2 types of participants classified according to the DII index. The DII index of the participants was calculated and the DII greater than 0 was defined as “positive” and the DII less than 0 was defined as “negative”, which was classified among the participants with Total PCBs as “Moderate”. \u003cstrong\u003ef.\u003c/strong\u003eSurvival curves of 2 types of participants classified according to the DII index. The DII index of the participants was calculated and the DII greater than 0 was defined as “positive” and the DII less than 0 was defined as “negative”, which was classified among the participants with Total PCBs as “Low”. \u003cstrong\u003eg.\u003c/strong\u003e Multifactorial Cox regression forest plot for participants with high PCBs. DII scores below 1 are denoted as 'positive,' while DII scores above 1 are denoted as 'negative.' Factors adversely affecting health are depicted in red, while factors beneficial to health are depicted in blue. The impact of DII on mortality among the high PCBs population was explored, adjusting for age, gender, race, education, BMI, smoking exposure, and SII. \u003cstrong\u003eh.\u003c/strong\u003e Multifactorial Cox regression forest plot for participants with moderate PCBs. \u003cstrong\u003ei.\u003c/strong\u003e \u0026nbsp;Multifactorial Cox regression forest plot for participants with high PCBs.\u003c/p\u003e\n\u003cp\u003e* 0.001 \u0026lt; P ≤ 0.05\u003c/p\u003e\n\u003cp\u003e** 0.0001 \u0026lt; P ≤ 0.01\u003c/p\u003e\n\u003cp\u003e*** P ≤ 0.0001\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-4543285/v1/9cf56fbe00627a2474dbf529.png"},{"id":60470444,"identity":"85e25e5c-b1ed-460c-9a15-079104316a83","added_by":"auto","created_at":"2024-07-17 06:30:58","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":26204731,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4543285/v1/105d4839-3ee4-4ae0-b5fe-bcb341663c87.pdf"},{"id":58965486,"identity":"87c84a1a-7619-4f59-824c-b8a58a278dfd","added_by":"auto","created_at":"2024-06-24 18:08:08","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":15875,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterialslegends.docx","url":"https://assets-eu.researchsquare.com/files/rs-4543285/v1/0014fdb0cc0af211eb31b82b.docx"},{"id":58965497,"identity":"b2215a79-f999-4efb-af83-b57ad638f5de","added_by":"auto","created_at":"2024-06-24 18:08:09","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":1364035,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure1.docx","url":"https://assets-eu.researchsquare.com/files/rs-4543285/v1/f37c7e4668c678a9678854e9.docx"},{"id":58965487,"identity":"80f63029-5f61-4cc2-88ab-f93d845529e0","added_by":"auto","created_at":"2024-06-24 18:08:08","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":57625,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable1.docx","url":"https://assets-eu.researchsquare.com/files/rs-4543285/v1/b947fa154b67ec74dc2ee926.docx"},{"id":58965495,"identity":"56eda3e1-4dc6-4fc0-b87b-85bc2566d4d6","added_by":"auto","created_at":"2024-06-24 18:08:08","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":29668,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable2.docx","url":"https://assets-eu.researchsquare.com/files/rs-4543285/v1/54fcf5f5e7d77e651258e781.docx"},{"id":58965488,"identity":"c5f7bb87-15cf-4900-8fe4-c958775f54cf","added_by":"auto","created_at":"2024-06-24 18:08:08","extension":"docx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":17560,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable3.docx","url":"https://assets-eu.researchsquare.com/files/rs-4543285/v1/29b833c1380536ba37356516.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Polychlorinated Biphenyls' Impact on Comorbidity Networks: Unveiling Epidemiological Patterns and offsetting Through Dietary Adjustments","fulltext":[{"header":"Highlights","content":"\u003cp\u003e1 PCBs were strongly associated with an increased risk of multisystem disease in comorbid networks.\u003c/p\u003e\n\u003cp\u003e2.PCBs were an independent contributor to mortality.\u003c/p\u003e\n\u003cp\u003e3.Machine learning models validated the impact of PCBs on disease and death.\u003c/p\u003e\n\u003cp\u003e4.The group comparison confirmed the corrective effect of diet on PCBs.\u003c/p\u003e\n"},{"header":"Introduction","content":"\u003cp\u003eIncreasing attention has been drawn to the potential adverse effects of human exposure to environmental pollutants in recent years. Dioxins, furans, coplanar polychlorinated biphenyls (PCBs), and organochlorine pesticides are all classified as persistent organic pollutants\u003csup\u003e1,2\u003c/sup\u003e. Among these, PCBs were widely used in electrical manufacturing and construction materials during the 20th century, and were commonly encountered as plasticizers and solvents\u003csup\u003e3\u003c/sup\u003e. The Stockholm Convention on Persistent Organic Pollutants, established in 2001, banned the production and use of PCBs due to their lipophilicity, environmental persistence, and ability to bioaccumulate in animal tissues\u003csup\u003e4\u003c/sup\u003e, thereby entering the human body through the food chain and external environmental exposure pathways\u003csup\u003e5\u003c/sup\u003e. According to pollution surveys in China, PCB contamination is particularly severe in areas of electronic waste dismantling (such as Taizhou City in Zhejiang Province, Qingyuan City and Guiyu Town in Guangdong Province), major metropolitan areas (especially in East China), and industrial clusters, primarily resulting from inadvertent generation during industrial heat treatment processes\u003csup\u003e6,7\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eTraditionally, PCBs have been believed to exert toxic effects on multiple organs or systems through various mechanisms\u003csup\u003e8\u003c/sup\u003e. Previous studies on PCB exposure have mostly focused on the association between PCBs and individual medical conditions. Evidence suggests that PCB exposure increases the risk of hypertension\u003csup\u003e9\u003c/sup\u003e, diabetes\u003csup\u003e10\u003c/sup\u003e, pulmonary arterial hypertension\u003csup\u003e11\u003c/sup\u003e, and liver diseases\u003csup\u003e12\u003c/sup\u003e. Plasma component analysis has revealed that PCB exposure impedes hematopoietic function, leading to inflammation and oxidative stress in the body\u003csup\u003e13\u003c/sup\u003e. A study on elderly Americans reported a negative correlation between blood PCB levels and attention, memory, and learning performance\u003csup\u003e14,15\u003c/sup\u003e. Furthermore, some research indicates that PCBs are risk factors for DNA mutations, birth defects, and reproductive system diseases\u003csup\u003e16-18\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eAlthough there are many studies on PCBs and diseases, most of them focused on one or a specific type of disease. Thus, comprehensive analyses of PCBs and comorbidity networks are still lacking\u003csup\u003e2,19\u003c/sup\u003e. Analyses focusing on individual diseases overlook interactions between diseases, and traditional methods for analyzing mortality risk overlook issues related to common exposure and risk confounding. With further advancements in statistics, comorbidity network analysis and machine learning-based algorithm models can effectively address these issues. Therefore, our study utilized data from over 100 million patients in the National Health and Nutrition Examination Survey (NHANES) to comprehensively investigate the integrated impact of PCBs on multiple disease categories through refined data selection and comorbidity network analysis. We also developed predictive models using machine learning algorithms to analyze the influence of PCBs on mortality risk. Additionally, we proposed for the first time a strategy to mitigate the adverse effects of PCBs by adjusting the Dietary Inflammatory Index (DII), which was validated through survival curve classification analysis and multifactorial Cox regression, offering profound clinical implications.\u003c/p\u003e"},{"header":"Materials","content":"\u003cp\u003e\u003cstrong\u003e2.1. Study design and participants\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs of March 1, 2024, all data used in our study are publicly accessible and sourced from the National Health and Nutrition Examination Survey (NHANES), managed by the National Center for Health Statistics (NCHS). Since 1999, data collection from participants has been conducted through questionnaire interviews, physical examinations, and laboratory tests. NHANES has obtained approval from the NCHS Institutional Review Board, and all participants have provided written informed consent. Survival status data of NHANES participants are derived from the National Death Index (NDI).\u003c/p\u003e\n\u003cp\u003eOur analysis includes 69 persistent organic pollutants concurrently present in NHANES data (1999-2004), including dioxins, furans, and coplanar polychlorinated biphenyls (PCBs). Through data screening and analysis, we identified seven PCBs\u0026mdash;PCB074, PCB170, PCB178, PCB180, PCB156, PCB157, and PCB146\u0026mdash;as noteworthy compounds. We have provided specific nomenclature and abbreviations for these compounds:\u003c/p\u003e\n\u003cp\u003eLBX074, representing PCB074(denoting 2,4,4\u0026rsquo;,5-Tetrachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX170, representing PCB170(2,2\u0026apos;,3,3\u0026apos;,4,4\u0026apos;,5-Heptachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX178, representing PCB178(2,2\u0026rsquo;,3,3\u0026rsquo;,5,5\u0026rsquo;,6-Heptachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX180, representing PCB180(2,2\u0026rsquo;,3,4,4\u0026rsquo;,5,5\u0026rsquo;-Heptachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX156, representing PCB156(2,3,3\u0026apos;,4,4\u0026apos;,5-Hexachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX157, representing PCB157(2,3,3\u0026apos;,4,4\u0026apos;,5\u0026apos;-Hexachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX146, representing PCB146(2,3,3\u0026apos;,4,4\u0026apos;,5\u0026apos;-Hexachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eWe integrated NHANES data spanning from 1999 to 2004 with NDI data, meticulously selecting 30,158 participants. Participants with missing or zero NHANES weight and those with missing data for the seven PCBs relevant to our study were systematically excluded from our dataset. However, for comprehensive network analysis, we refrained from further data exclusion. This meticulous curation resulted in a final cohort of 10,961 participants. At different analysis stages, data were selected based on varying data requirements.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2. Methodology for Measuring Serum Levels of PCBs\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe analytes were quantified in serum using high-resolution gas chromatography and isotope-dilution high-resolution mass spectrometry (HRGS/ID-HRMS). Serum samples were fortified with 13C12-labeled internal standards and extracted through either C18 solid-phase extraction (SPE) or liquid-liquid extraction. Chromatographic separation occurred on a DB-5ms capillary column employing a Hewlett-Packard 6890 gas chromatograph. Quantification was achieved by ID-HRMS using selected ion monitoring (SIM) at a resolving power of 10,000 with either a Micromass AutoSpec ULTIMA or Finnigan MAT95 mass spectrometer in electron ionization (EI) mode. Detection limits were reported for each sample, accounting for sample weight and analyte recovery.\u003c/p\u003e\n\u003cp\u003eFrom the entire persistent organic pollutant (POP) library, we ultimately identified 18 PCBs, with data for these substances in the NHANES cycles from 1999-2000, 2001-2002, and 2003-2004 accounting for over 75% of the total data. Further analysis of the 18 POPs using weighted quantile sum (WQS) analysis identified 7 PCBs with the greatest impact on mortality risk, cumulatively contributing to 95% of the total weight. As per NHANES guidelines, values falling below the limit of detection (LOD) were imputed with a value equivalent to the LOD divided by the square root of 2.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3. Diagnoses of medical conditions and DII\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs NHANES does not directly record mortality data, mortality information in this study was obtained through a probabilistic match between NHANES and the National Death Index (NDI), following procedures validated by the National Center for Health Statistics (NCHS). Mortality data from the NDI were available for analysis until December 31, 2019.\u003c/p\u003e\n\u003cp\u003eAll diseases available in NHANES were included in our analysis, encompassing Angina, Heart attack, Heart disease, Heart failure, Hypertension, Stroke, Alcoholic fatty liver, Hepatitis B virus (HBV), Hepatitis C virus (HCV), Hepatitis D virus (HDV), Non-alcoholic fatty liver, Hyperlipidemia, Diabetes, Osteoporosis, Hyperuricemia, Thyroid disease, Human Immunodeficiency Virus (HIV), Arthritis, Depression, Cancer, Chronic bronchitis, Emphysema, Asthma, Chronic kidney disease (CKD), and Proteinuria, totaling 25 diseases. Diabetes was defined as self-reported diabetes diagnosis, use of oral antidiabetic drugs or insulin, glycated hemoglobin (HbA1c) levels \u0026ge;6.5%, plasma glucose levels \u0026ge;200 mg/dL two hours after an oral glucose tolerance test (OGTT), or fasting plasma glucose levels \u0026ge;126 mg/dL\u003csup\u003e20\u003c/sup\u003e. Hypertension was determined based on self-reported hypertension or NHANES-measured data: an average systolic blood pressure \u0026ge;130 mm Hg or diastolic blood pressure \u0026ge;80 mm Hg from three measurements\u003csup\u003e21\u003c/sup\u003e. To assess chronic kidney disease (CKD), essential indicators including estimated glomerular filtration rate (eGFR) and urine albumin-to-creatinine ratio (UACR) were relied upon. UACR (mg/g) was calculated as the ratio of urine albumin (mg/dL) to urine creatinine (g/dL), with a UACR value exceeding 30 mg/g indicating \u0026quot;proteinuria.\u0026quot; eGFR was computed using the CKD-EPI formula, expressed as:\u003c/p\u003e\n\u003cp\u003eGFR = 175 \u0026times; standardized serum creatinine\u003csup\u003e\u0026nbsp;(\u0026minus;1.154)\u003c/sup\u003e \u0026times; age\u003csup\u003e\u0026nbsp;(\u0026minus;0.203)\u003c/sup\u003e \u0026times; 1.212 [if Black] \u0026times; 0.742 [if female], where serum creatinine is measured in mg/dL.\u003c/p\u003e\n\u003cp\u003eCKD was defined as eGFR \u0026lt; 60 mL/min/1.73 m\u003csup\u003e2\u0026nbsp;\u003c/sup\u003ethe presence of renal damage markers (such as proteinuria), or both, persisting for at least 3 months, regardless of the underlying etiology\u003csup\u003e22\u003c/sup\u003e. In line with previous publications, NAFLD was defined by hepatic steatosis index (HSI) and US fatty liver index (USFLI). The formulars are as follows:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHSI = 8 \u0026times; (alanine aminotransferase/aspartate aminotransferase ratio) + body mass index (+2 for female; +2 for diabetes);\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eUSFLI = (e\u0026minus;0.8073 \u0026times; Non\u0026minus;Hispanic Black+0.3458\u0026times;Mexican American+0.0093\u0026times;Age+0.6151 \u0026times; loge (Gamma glutamyltransferase) +0.0249 \u0026times; Waist Circumference+1.1792 \u0026times; loge (Insulin)+0.8242 \u0026times; loge (Glucose)\u0026minus;14.7812)/ (1 + e\u0026minus;0.8073 \u0026times; Non\u0026minus;Hispanic Black+0.3458 \u0026times;Mexican American+0.0093 \u0026times; Age+0.6151 \u0026times; loge (Gamma glutamyltransferase) +0.0249 \u0026times; waist circumference+1.1792 \u0026times; loge (Insulin)+0.8242 \u0026times; loge (Glucose) \u0026ndash; 14.7812) \u0026times; 100.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eUSFLI cutoff value\u0026nbsp;\u0026ge;\u0026nbsp;30 or HSI value\u0026nbsp;\u0026gt;\u0026nbsp;36 was diagnosed as NAFLD\u003csup\u003e23\u003c/sup\u003e. \u0026nbsp;ALD was defined by a combination of an evidence of excessive alcohol consumption (\u0026ge; 210 g/week for men and \u0026ge; 140 g/week for women) and an ALD/NAFLD index \u0026gt; 0, which was calculated as:\u003c/p\u003e\n\u003cp\u003e\u0026minus;58.5 + 0.637 (Mean Corpuscular Volume) + 3.91 (Aspartate Aminotransferase [AST]/Alanine Aminotransferase [ALT]) \u0026minus; 0.406 (Body Mass Index) + 6.35 for Male Gender.\u003c/p\u003e\n\u003cp\u003eIn the subpopulation with ALD, the AST-to-platelet ratio index (APRI)\u0026nbsp;and FIB-4 score\u0026nbsp;were used to evaluate ALD FIB. The formula is as follows:\u003c/p\u003e\n\u003cp\u003eAPRI = (AST/Upper Limit of Normal/Platelet Count [109/L]) \u0026times; 100, where the upper limits of normal AST levels were set at 37 IU/L for men and 29 IU/L for women;\u003c/p\u003e\n\u003cp\u003eFIB-4 = Age \u0026times; AST/ [Platelets in 109/L \u0026times; (ALT)1/2]\u003c/p\u003e\n\u003cp\u003eCut-off values for advanced fibrosis (\u0026ge; F3) were set at 1.5 for APRI and 3.25 for FIB-4\u003csup\u003e24,25\u003c/sup\u003e. HBV, HCV, HDV, and HIV-positive patients were determined based on antigen measurements and quantification of relevant viral DNA or RNA levels in NHANES laboratories. Hyperlipidemia was defined as fasting triglyceride values\u0026thinsp;\u0026ge;\u0026thinsp;200 ng/dl. Smokers were defined as individuals who have smoked more than 100 cigarettes in their lifetime and currently smoke on some days or every day, while nonsmokers are those who have smoked less than 100 cigarettes in their lifetime\u003csup\u003e26\u003c/sup\u003e. Hyperuricemia was defined dichotomously with serum uric acid (SUA) levels \u0026ge;416\u0026mu;mol/L (7.0 mg/dL) for males and \u0026ge;357\u0026mu;mol/L (6.0 mg/dL) for females\u003csup\u003e27\u003c/sup\u003e. Apart from the diseases mentioned above, data on Heart attack, Heart disease, Heart failure, Stroke, Osteoporosis, Thyroid disease, Emphysema, Arthritis, Depression, Cancer, Chronic bronchitis, and Asthma were obtained from questionnaire data provided by NHANES. These indices were derived from comprehensive full blood cell count tests, the details of which can be found in the \u0026apos;Questionnaire\u0026apos; data within the NHANES dataset. Finally, we categorized these diseases into seven major classes based on disease type: Circulatory system diseases, Digestive system diseases, Endocrine/Metabolic diseases, Immune system diseases, Respiratory system diseases, Urinary system diseases, and Others.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe Dietary Inflammatory Index (DII) assesses the inflammatory effect of diet using 45 dietary parameters, normalizing individual intake of each food parameter to global intake. Standardized intake scores (Z-scores) are converted to proportions and centered. The centered proportions of these specific food intakes are multiplied by their inflammation effect scores and summed to obtain an individual\u0026apos;s overall DII score. Participants\u0026apos; DII scores represent the sum of each DII score. Higher DII scores indicate a pro-inflammatory diet, while lower scores indicate an anti-inflammatory diet. In this study, 28 out of 45 food parameters were utilized for DII calculation: carbohydrates, protein, total fat, alcohol, fiber, cholesterol, saturated fatty acids, monounsaturated fatty acids, polyunsaturated fatty acids, n-3 fatty acids, n-6 fatty acids, niacin, vitamin A, thiamine, vitamin B2, vitamin B6, vitamin B12, vitamin C, vitamin D, vitamin E, iron, magnesium, zinc, selenium, folic acid, carotene, caffeine, and energy\u003csup\u003e28,29\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4. Covariates\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe included C-reactive protein (CRP) and the systemic immune-inflammation index (SII) as covariates in our analysis. We selected and utilized data on platelet count (PC), neutrophil count (NC), and lymphocyte count (LC) in the computation, with SII calculated as SII = PC * (NC / LC)\u003csup\u003e30\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eOther covariates included Age, Poverty Income Ratio (PIR), Body Mass Index (BMI), Gender, Race, Education, Smoking Exposure, and Alcohol Exposure. BMI was categorized into three groups: normal (BMI \u0026lt; 25 kg/m\u0026sup2;), overweight (25 \u0026le; BMI \u0026le; 30 kg/m\u0026sup2;), and obese (BMI \u0026gt; 30 kg/m\u0026sup2;), based on participants\u0026apos; BMI values\u003csup\u003e31\u003c/sup\u003e. Alcohol consumption was assessed using data from NHANES questionnaires. Participants who had consumed fewer than 12 alcoholic drinks in their lifetime were classified as non-drinkers. Former drinkers were individuals who had consumed \u0026ge;12 drinks at any point in their lifetime but had not consumed alcohol in the past year. To minimize recall bias, smoking exposure was evaluated based on serum cotinine levels rather than relying solely on the \u0026apos;smoking history questionnaire.\u0026apos; Current smokers were identified by serum cotinine levels \u0026gt; 10 ng/mg, former smokers had serum cotinine levels \u0026le; 10 ng/mg, and non-smokers exhibited serum cotinine levels \u0026lt; 0.011 ng/mg\u003csup\u003e32\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003e2.5. Statistical Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn accordance with Centers for Disease Control and Prevention (CDC) guidelines, our statistical analyses adhered to stipulated principles. To address the complex multi-stage cluster survey design inherent to NHANES, appropriate sample weights were meticulously applied to each participant. Categorical variables were expressed as proportions, while continuous variables were presented as means (mean \u0026plusmn; standard deviation). Descriptive statistics comprehensively summarized participants\u0026apos; demographic characteristics and biomarker concentrations. Specifically, for each selected PCB, analysis was conducted after stratifying into three groups based on quartiles, and for substances exhibiting highly right-skewed distributions, analysis was stratified into two groups based on the median.\u003c/p\u003e\n\u003cp\u003eA total of 69 persistent organic pollutants (POPs) were initially identified in NHANES data (1999-2004), encompassing dioxins, furans, and coplanar polychlorinated biphenyls (PCBs). From these, 18 persistent organic pollutants with valid data representing 75% of the total data were selected for further analysis. Weighted quantile sum (WQS) regression, focusing on mortality risk, was performed on these 18 substances, with the top 7 substances selected based on their cumulative contribution rate to the preceding 95%, all of which were PCBs.\u003c/p\u003e\n\u003cp\u003eCorrelation heatmaps were employed to illustrate the interrelationships among the 7 PCBs and their associations with various diseases. Principal component analysis (PCA) was utilized to visualize the relationship between PCBs and mortality risk in two dimensions. Additionally, five diseases highly correlated with PCBs (Hyperuricemia, Hypertension, Diabetes, Chronic Kidney Disease (CKD), and Arthritis) were included in the machine learning algorithm model, while Cancer, Osteoporosis, and Hepatitis C (HCV) were excluded due to either their broad spectrum or insufficient data volume. Further, multivariate logistic regression was employed to evaluate the associations between these substances and 25 diseases, with results presented via circular plots. Adjustment for Age, Gender, Race, BMI, Social Inequality Index (SII), smoking exposure, and alcohol exposure was conducted. To illustrate the interrelationships among different diseases, logistic regression and comorbidity network analysis were performed on the 25 diseases, with the most relevant diseases determined based on Odds Ratios (ORs) and p-values, and results presented via network analysis diagrams.\u003c/p\u003e\n\u003cp\u003eTo further demonstrate PCBs as independent risk factors for mortality, excluding the influence of confounding factors such as diseases, we constructed five individual models based on machine learning: Support Vector Machine (SVM), Na\u0026iuml;ve Bayes, Decision Tree (Tree), Stochastic Gradient Descent (SGD), Gradient Boosting Decision Tree (GBDT), and four composite models: Random Forest, Histogram Gradient Boosting Decision Tree (hist GBDT), Bagging, and Neural Network. Additionally, a Voting algorithm was constructed for result output, with results displayed via ROC curves and confusion matrices. Furthermore, by plotting survival curves and constructing multi-factor Cox regression models, we effectively demonstrated the ability of dietary habits, moderated by DII, to mitigate the adverse effects of environmental factors such as PCBs on human health within high PCB-exposed populations.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003e3.1. Unveiling the Relationship between PCBs and Diseases and Mortality Risk\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTable 1 and Supplementary Table 1 present the essential characteristics of our study cohort. The mean age of participants was 40.99 years. Regarding gender distribution, females slightly outnumbered males, constituting 50.2% compared to 49.8%. Additionally, we categorized the 10,961 participants into three groups - low, moderate, and high - based on tertiles reflecting PCB concentrations. Notably, individuals in the high LBX074 group displayed distinct characteristics compared to those in the low and moderate groups (Table 1). Specifically, the high LBX074 group exhibited a higher proportion of females, a significantly greater mean age, and a substantially elevated percentage of active smokers. Similar trends were observed for other comprehensive PCB data, as outlined in Supplementary Table 1.\u003c/p\u003e\n\u003cp\u003eFigures 1a and 1b depict the results of weighted quantile sum (WQS) and weighting analysis for the 18 PCBs and 25 diseases. By calculating the cumulative contribution rate, we identified the top 7 substances, ranked in descending order based on their contribution rate, accounting for 95% of the cumulative contribution rate: \u0026nbsp; \u0026nbsp; \u0026nbsp; LBX074, representing PCB074(denoting 2,4,4\u0026apos;,5-Tetrachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX170, representing PCB170(2,2\u0026apos;,3,3\u0026apos;,4,4\u0026apos;,5-Heptachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX178, representing PCB178(2,2\u0026rsquo;,3,3\u0026rsquo;,5,5\u0026rsquo;,6-Heptachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX180, representing PCB180(2,2\u0026rsquo;,3,4,4\u0026rsquo;,5,5\u0026rsquo;-Heptachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX156, representing PCB156(2,3,3\u0026apos;,4,4\u0026apos;,5-Hexachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX157, representing PCB157(2,3,3\u0026apos;,4,4\u0026apos;,5\u0026apos;-Hexachlorobiphenyl).\u003c/p\u003e\n\u003cp\u003eLBX146, representing PCB146(2,3,3\u0026apos;,4,4\u0026apos;,5\u0026apos;-Hexachlorobiphenyl).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn line with this, the main text primarily focused on LBX074 data, aimed at representing this category of substances, while additional data were predominantly included in the supplementary materials.\u003c/p\u003e\n\u003cp\u003eFigures 1c and 1d illustrate, in the form of heatmaps, the interrelationships among the 7 PCBs and between PCBs and the 25 diseases, respectively. Significant positive correlations were observed among all 7 PCBs (p \u0026lt; 0.0001), with LBX146 exhibiting the strongest correlation with the other 6 substances. The 7 PCBs showed significant positive correlations with most diseases (p \u0026lt; 0.0001). However, all 7 PCBs were negatively correlated with Asthma, and LBX178 and LBX167 were negatively correlated with Chronic bronchitis and Hyperlipidemia. Notably, Alcoholic fatty liver showed almost no correlation with any of the PCBs. These findings suggested that PCBs might be important environmental factors associated with the malignant progression of most diseases, affecting multiple systems in the body. However, due to their different mechanisms of action, an increase in PCBs within a certain range might partially inhibit the progression of specific diseases.\u003c/p\u003e\n\u003cp\u003eFigure 2 employed principal component analysis (PCA) to illustrate the relationship between PCBs and the risk of death in a two-dimensional format. Pairwise combinations of the 7 PCBs were analyzed, and the PCA two-dimensional plot indicates that for each pair of PCBs, higher PCB levels correspond to increased risk of death. Thus, all 7 PCBs are important factors in increasing the mortality risk\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2. Comorbidity Network Analyses: Demonstrating Interactions Between Diseases\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMultivariable logistic regression was employed to analyze the summarized 25 diseases, adjusting for Age, Gender, Race, BMI, SII, smoking exposure, and alcohol exposure. Among them, 12 diseases showed significant associations with PCBs (p\u0026lt;0.05), with 6 diseases significantly positively correlated with PCBs (OR\u0026gt;1, p\u0026lt;0.05) (Figure 3a), including Hyperuricemia (OR\u0026gt;6.0, p\u0026lt;0.001), Diabetes (OR\u0026gt;6.0, p\u0026lt;0.0001), HCV (OR\u0026gt;6, p=0.0071), Hyperlipidemia (OR\u0026gt;6, p=0.016), HIV (OR\u0026gt;6, p=0.0005), and Arthritis (OR=5.42, p=0.0065). Adjusting for covariates, it can be demonstrated that PCBs are independent influencing factors for multiple diseases.\u003c/p\u003e\n\u003cp\u003eLogistic regression was used to compute the odds ratio (OR) and p-values for each pair of the 25 diseases to determine disease associations. We calculated the OR values for all disease pairs (see supplementary table 2) and presented the disease pairs most closely related to each disease (p\u0026lt;0.05) (see supplementary table 3). Among these pairs, Depression and Stroke emerged as the most correlated diseases (OR (95%CI): 40.21(5.83,794.00), p=0.001), while the association between Thyroid disease and Angina was the least significant (OR (95%CI): 3.60(2.76,4.66), p\u0026lt;0.0001). Diseases of the circulatory system tend to be interrelated, with diseases highly correlated with diabetes and proteinuria, typical of the urinary system, accounting for the majority of associations. Depression is strongly associated with some common chronic clinical conditions. Based on comorbidity network analysis, we identified 283 potential links, with the number of related links, OR values, and comorbidity network analysis results shown in Figure 3b. Each node represents a medical condition, and the thickness of the connecting lines reflects the strength of the disease pairs\u0026apos; association. Nodes closer to the network center have stronger centrality, indicating a greater number of connections with other diseases. Hypertension exhibits the strongest centrality and the highest number of connections, significantly influencing most other diseases. It is also notable that diseases of the circulatory system are often closely associated with other diseases. Depression, ALD, HIV, and HPL are distanced from the center, showing fewer associations with other diseases and smaller impact.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3. Unveiling the Association between PCBs with Mortality Using Machine Learning, introducing DII for correction.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further mitigate the influence of confounding factors on the analysis of the association between PCBs and mortality risk, we constructed nine learning models, including five individual models: Support Vector Machine (SVM), Na\u0026iuml;ve Bayes, Decision Tree (Tree), Stochastic Gradient Descent (SGD), and Gradient Boosting Decision Tree (GBDT). Additionally, we developed four ensemble models: Random Forest, Histogram Gradient Boosting Decision Tree (hist GBDT), Bagging, and Neural Network. Furthermore, we incorporated a Voting algorithm for result output. Two types of models were added to the algorithm: one containing Age, gender, race, hyperuricemia, hypertension, diabetes, CKD, and arthritis, and the other adding PCBs data. The AUC values of these two types of models, representing predictive accuracy, were compared (Figure 4a, Figure 4b). It is evident that models incorporating PCBs data exhibited varying degrees of improvement in accuracy compared to the baseline models. Overall, ensemble models outperformed individual models, with Random Forests showing significant advantages in prediction. Prior to incorporating PCBs data, the accuracy rates were as follows: SVM (0.86), SGD (0.88), Na\u0026iuml;ve Bayes (0.85), Decision Tree (0.88), GBDT (0.91), hist GBDT (0.89), Random Forests (0.98), Bagging (0.92), Neural Network (0.91), and the final Voting model (0.94). After incorporating PCBs data, the accuracy rates were: SVM (0.89), SGD (0.90), Na\u0026iuml;ve Bayes (0.86), Decision Tree (0.90), GBDT (0.94), hist GBDT (0.91), Random Forests (1.0), Bagging (0.95), Neural Network (0.95), and the final Voting model (0.96). Figures 4c and 4d illustrate the comparison of ROC curves before and after applying the Voting algorithm. Figures 4e and 4f illustrate the comparison of the confusion matrices for the Voting algorithm before and after. These results strongly demonstrate a substantive positive correlation between PCBs and mortality risk, even after controlling for baseline data and confounding factors such as diseases. The ROC curves and confusion matrices of the other algorithms are presented in Supplementary Figure 4.\u003c/p\u003e\n\u003cp\u003eFurthermore, in Figure 5, we demonstrated through survival curves that we could mitigate the impact of DII by adjusting lifestyle habits. We first continued to utilize the previously built machine learning models to calculate the accuracy and errors of the predictive model containing population baseline data (Age, gender, race) and five diseases (hyperuricemia, hypertension, diabetes, CKD, arthritis), along with seven PCBs data (Figure 5a). Furthermore, we incorporated DII data of each participant into the aforementioned model for prediction and recalculated the accuracy and errors of the new model (Figure 5b). The results indicated that the DII index effectively enhanced the accuracy of the model, highlighting its significance as a contributing factor to increased mortality risk. In Figure 5c, participants were categorized into low, medium, and high groups based on the quartiles of the total amount of 7 PCBs in their bodies, and survival curves were plotted accordingly. Participants with high levels of PCBs showed a significantly increased risk of mortality, while participants with medium and low levels exhibited progressively lower risks of mortality (p\u0026lt;0.0001). In Figure 5d, participants with high levels of PCBs were further classified based on their DII scores: those with DII scores greater than 0 were defined as \u0026quot;positive,\u0026quot; while those with DII scores less than or equal to 0 were defined as \u0026quot;negative.\u0026quot; In comparison with Figure 5c, it is evident that a DII score greater than 0 (indicating reduced inflammation) effectively reduces the risk of mortality among participants with higher levels of PCBs. Figures 5e and 5f depict similar classifications for participants with medium and low levels of PCBs, respectively. However, the conclusions drawn from participants with high levels of PCBs were not statistically significant in these two groups, indicating no significant differences.\u003c/p\u003e\n\u003cp\u003eIn Figures 5g, h, and i, we further constructed multifactor Cox regression models to examine the association between DII index and mortality, adjusting for age, gender, race, education, BMI, smoking exposure, and SII covariates. Among the high PCBs population, a DII index less than 1 (indicating inflammation suppression) was significantly negatively correlated with mortality (OR (95% CI) = 0.7151 (0.6683, 0.7651), p \u0026lt; 0.05). However, in the medium PCBs and low PCBs populations, the association between a DII index less than 1 and mortality was not statistically significant (OR (95% CI) = 1.05689 (0.94773, 1.1786), p \u0026gt; 0.05) (OR (95% CI) = 1.125 (0.87141, 1.4525), p \u0026gt; 0.05).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThrough the above analysis, we effectively demonstrated that among the high PCBs population, modulation of dietary habits and adjustment of DII can effectively counteract the adverse effects of PCBs and similar environmental factors on human health. However, in the medium PCBs and low PCBs populations, the impact of DII adjustment on mortality is not evident.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur study revealed profound associations between PCBs and disease networks, as well as mortality risk. We delved into the interplay among various diseases and proposed, for the first time, that mitigating mortality risk and potentially alleviating the impact of environmental factors could be achieved through controlling DII. Taking LBX074 (2,4,4,5-Tetrachlorobiphenyl) as an example, LBX074 exhibited significant positive correlations with 7 diseases (OR\u0026gt;1, p\u0026lt;0.05). PCA revealed a notable increase in mortality risk with increasing levels of LBX074, a trend observed in other PCBs as well. Depression and Stroke emerged as the most relevant disease pair in network analysis (OR (95%CI): 40.21 (5.83, 794.00), p=0.001). When exploring PCBs\u0026apos; independent role in mortality risk through machine learning models, we controlled for confounding factors and comorbidities, further confirming PCBs as independent risk factors for mortality. The feasibility of reducing mortality risk by lowering DII was validated through survival curve plotting and multi-factor Cox regression analysis. Although the specific pathogenic mechanisms linking coplanar polychlorinated biphenyls (PCBs) to mortality rates remain unclear, our study elucidated associations between PCBs and the development of various diseases, shedding light on the factors contributing to increased mortality rates associated with coplanar PCB exposure.\u003c/p\u003e\n\u003cp\u003eTo the best of our knowledge, our study possesses several notable strengths. It is the first to rectify the limitations of traditional methods using artificial intelligence-based big data models to predict mortality risk associated with PCB exposure and evaluate the impact of DII on mortality risk using machine learning methods. Our findings corroborate those of traditional data analysis, with the addition of PCB data significantly enhancing the predictive accuracy of multiple models compared to those without PCB data, suggesting that various PCBs are independent influencing factors on mortality. Additionally, precise measurements of accuracy and errors through multiple iterations provided further evidence of the detrimental effects of pro-inflammatory diets on the body. Traditional statistical methods have limitations in analyzing mortality risk, as they overlook issues of shared exposure and multi-factorial risk confounding, and are unable to effectively address statistical errors. With the further development of artificial intelligence, machine learning-based algorithm models can effectively address these issues, with algorithmic results becoming increasingly accurate over time, capable of handling various data formats in dynamic, large-volume, and complex data environments. Therefore, our study, through further data screening and the construction of predictive models using machine learning algorithms, analyzed the impact of PCBs on mortality risk.\u003c/p\u003e\n\u003cp\u003eMoreover, this study represents the first attempt to comprehensively investigate the combined effects of PCBs on various diseases and comorbidity networks using comorbidity network analysis. Logistic regression was employed to calculate OR values, while centrality and associated nodes were demonstrated through comorbidity network visualization. Consistent with past research, we found significant positive correlations between PCBs and various diseases such as Hyperuricemia, Diabetes, and Hyperlipidemia. Additionally, the comorbidity network analysis indicated that hypertension is a significant trigger for multiple systemic diseases, with circulatory system diseases often closely associated with various other systemic diseases, exhibiting the strongest centrality. Furthermore, through comorbidity network analysis, we first discovered significant positive correlations between PCB levels and HCV, HIV, and arthritis, likely attributable to PCB-induced inflammation, immune suppression, and apoptosis induction in cartilage cells via ROS-dependent pathways. While the specific mechanisms by which PCBs contribute to various diseases and mortality remain unclear, it is undeniable that PCBs pose significant hazards to human health, serving as independent risk factors for multiple diseases and mortality. The impact weight of PCBs is higher than that of some conventional detection substances, suggesting that PCBs may serve as specific biomarkers for certain diseases, aiding in disease prediction in the future. Moreover, further research is needed on the effects of environmental pollutants on human health.\u003c/p\u003e\n\u003cp\u003eFurthermore, we have introduced for the first time the concept that adverse effects of pollutants can potentially be counteracted by altering dietary habits. Previous research indicates that higher levels of inflammation lead to an increased risk of mortality. Our study further demonstrates the significant role of DII in triggering inflammation and oxidative stress in the disease and mortality processes. Among populations with high PCB exposure, significant reductions in mortality and morbidity risks can be achieved through DII regulation. However, similar effects were not significant among populations with moderate or low PCB exposure. This suggests that PCBs may induce inflammation to a certain threshold, and DII regulation can effectively suppress their effects. Numerous studies have shown that diet, as the main source of bioactive compounds, can mediate inflammatory responses, with pro-inflammatory diets associated with increased white blood cell counts. Pro-inflammatory diets exhibit significant positive correlations with various diseases, including chronic obstructive pulmonary disease, diabetes, depression, and cardiovascular diseases, while high pro-inflammatory diets can increase the risk of mortality, possibly by increasing white blood cell and CRP levels, thereby inducing various diseases leading to mortality. One possible mechanism is the close relationship between diet and the human gut microbiota. Several animal studies have shown that high-sugar diets lead to obesity, insulin resistance, increased intestinal permeability, and low-grade inflammation. Microbial metabolites (such as SCFA butyrates or tryptophan metabolites) can control various physiological functions in the host, ranging from inflammatory responses to energy metabolism in epithelial cells. Bifidobacteria, Lactobacilli, Clostridia, Bacillus subtilis, and fragile bacilli are closely related to specific immunity via MyD88, transforming growth factor-\u0026beta;, IL-1, IL-6, IL-17, IL-22, \u0026gamma;-PgA, and PSA. Fragile bacilli, plant bifidobacteria, and bifidobacteria can regulate inflammatory responses via TLR, NF-\u0026kappa;B, and MyD88. Inflammation is closely related to diseases, and therefore high DII can induce diseases by triggering inflammation, while low DII has the opposite effect. However, specific hypotheses cannot be tested in current studies. Therefore, future longitudinal studies could consider the potential mechanisms by which diet-driven inflammation induces mortality or disease. Similarly, future research could determine whether the use of anti-inflammatory diets (such as increasing leafy vegetables, herbs, spices, and certain fruits) can reduce WBC and CRP levels, decrease morbidity, and reduce mortality risk.\u003c/p\u003e\n\u003cp\u003eHowever, this study also has some limitations. Firstly, due to insufficient data, more substances were not studied. Secondly, the mechanisms by which they cause multiple diseases leading to increased mortality rates remain unknown. Additionally, DII was not compared with energy-adjusted DII (E-DII), which constructs a reference database for energy-adjusted nutritional scoring based on data from the same 11 countries used to calculate DII. Without access to the unique comparison database, E-DII cannot be calculated, so we were unable to compare it in our study. Furthermore, the ubiquity and complexity of exposure not only necessitate further research on the effects of PCBs but also require further investigation into the prevention and monitoring of PCBs, which may help clinicians better understand and control exposure levels of these organic pollutants. Finally, when using the NHANES database for statistical analysis, we selected multiple variables. Indeed, when lots of variables are tested, associations flourish, most are due to chance, some are merely markers, some are due to common non-investigated factors, and just a few are causal. The non-longitudinal nature of these surveys is not helpful to discern whether the statistical associations are meaningful enough. Therefore, we cannot avoid the analysis bias caused by large databases.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eAfter employing machine learning methods to better mitigate the impact of confounding factors, we observed a strong positive correlation between PCBs and diseases across multiple systems, indicating that PCBs may be independent risk factors for various diseases and even mortality. This suggests that PCBs may be intimately involved in the development and progression of multiple diseases. By constructing multidimensional machine learning models and conducting multiple iterations for precision and error measurement, PCBs may have the potential to become specific biomarkers for certain diseases in the future. Furthermore, we observed that in populations with high PCB exposure, significant reductions in mortality and morbidity risks can be achieved by adjusting DII, which is of great significance. The impact of PCBs on human health may be achieved through the induction of systemic inflammation, causing damage to multiple systems in the body, and improving daily dietary habits is an effective solution.\u003c/p\u003e\n\u003cp\u003eHowever, a more comprehensive analysis is still lacking, and further research is needed to determine the specific mechanisms associated with its processes.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor Contributions:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis article was written by Ying Gao, Han Lu, Huan Zhou, and Jiaxing Tan. Ying Gao, Han Lu, and Jiaxing Tan were responsible for data completion and interpretation. Jiaxing Tan and Han Lu contributed to the study design and analysis methods. Ying Gao and Huan Zhou participated in data collection and validation, and completed the writing of the entire article. Jiaxing Tan supervised the research and provided guidance throughout the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interest Statement:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors hereby verify the absence of any financial or personal affiliations that may have the potential to impact the credibility of the findings articulated in this paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of generative AI and AI-assisted technologies in the writing process\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the course of preparing this manuscript, the authors employed ChatGPT to enhance language usage. Subsequent to utilizing this tool, the authors meticulously reviewed and edited the content as necessary, assuming full responsibility for the final publication\u0026apos;s content.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was partly supported by grants from the project of the National Natural Science Foundation of China (No. 81970612 and No. 82300797) and Innovation and Entrepreneurship Training Program for College Students.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003cstrong\u003e：\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated and analysed during the current study are available in the NHANES repository, https://www.cdc.gov/nchs/nhanes.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eLin, Y. S.\u003cem\u003e et al.\u003c/em\u003e Environmental exposure to dioxin-like compounds and the mortality risk in the U.S. population. \u003cem\u003eInt J Hyg Environ Health\u003c/em\u003e \u003cstrong\u003e215\u003c/strong\u003e, 541-546 (2012). https://doi.org:10.1016/j.ijheh.2012.02.006\u003c/li\u003e\n\u003cli\u003eWu, B.\u003cem\u003e et al.\u003c/em\u003e Combined exposure to multiple dioxins and dioxin-like polychlorinated biphenyls on hypertension among US adults in NHANES: a cross-sectional study under three statistical models. \u003cem\u003eEnviron Sci Pollut Res Int\u003c/em\u003e \u003cstrong\u003e30\u003c/strong\u003e, 28730-28744 (2023). https://doi.org:10.1007/s11356-022-24271-3\u003c/li\u003e\n\u003cli\u003eBouchard, M. F., Oulhote, Y., Sagiv, S. K., Saint-Amour, D. \u0026amp; Weuve, J. Polychlorinated biphenyl exposures and cognition in older U.S. adults: NHANES (1999-2002). \u003cem\u003eEnviron Health Perspect\u003c/em\u003e \u003cstrong\u003e122\u003c/strong\u003e, 73-78 (2014). https://doi.org:10.1289/ehp.1306532\u003c/li\u003e\n\u003cli\u003eVan Oostdam, J.\u003cem\u003e et al.\u003c/em\u003e Human health implications of environmental contaminants in Arctic Canada: a review. \u003cem\u003eSci Total Environ\u003c/em\u003e \u003cstrong\u003e230\u003c/strong\u003e, 1-82 (1999). https://doi.org:10.1016/s0048-9697(99)00036-4\u003c/li\u003e\n\u003cli\u003eHelou, K., Harmouche-Karaki, M., Karake, S. \u0026amp; Narbonne, J. F. A review of organochlorine pesticides and polychlorinated biphenyls in Lebanon: Environmental and human contaminants. \u003cem\u003eChemosphere\u003c/em\u003e \u003cstrong\u003e231\u003c/strong\u003e, 357-368 (2019). https://doi.org:10.1016/j.chemosphere.2019.05.109\u003c/li\u003e\n\u003cli\u003eChen, L.\u003cem\u003e et al.\u003c/em\u003e Polychlorinated biphenyls in the atmosphere of an urban city: levels, distribution, and emissions. \u003cem\u003eArch Environ Contam Toxicol\u003c/em\u003e \u003cstrong\u003e57\u003c/strong\u003e, 437-446 (2009). https://doi.org:10.1007/s00244-009-9293-z\u003c/li\u003e\n\u003cli\u003eZhu, M.\u003cem\u003e et al.\u003c/em\u003e Environmental contamination and human exposure of polychlorinated biphenyls (PCBs) in China: A review. \u003cem\u003eSci Total Environ\u003c/em\u003e \u003cstrong\u003e805\u003c/strong\u003e, 150270 (2022). https://doi.org:10.1016/j.scitotenv.2021.150270\u003c/li\u003e\n\u003cli\u003eAndroutsopoulos, V. P., Hernandez, A. F., Liesivuori, J. \u0026amp; Tsatsakis, A. M. A mechanistic overview of health associated effects of low levels of organochlorine and organophosphorous pesticides. \u003cem\u003eToxicology\u003c/em\u003e \u003cstrong\u003e307\u003c/strong\u003e, 89-94 (2013). https://doi.org:10.1016/j.tox.2012.09.011\u003c/li\u003e\n\u003cli\u003eRaffetti, E., Donat-Vargas, C., Mentasti, S., Chinotti, A. \u0026amp; Donato, F. Association between exposure to polychlorinated biphenyls and risk of hypertension: A systematic review and meta-analysis. \u003cem\u003eChemosphere\u003c/em\u003e \u003cstrong\u003e255\u003c/strong\u003e, 126984 (2020). https://doi.org:10.1016/j.chemosphere.2020.126984\u003c/li\u003e\n\u003cli\u003eSilverstone, A. E.\u003cem\u003e et al.\u003c/em\u003e Polychlorinated biphenyl (PCB) exposure and diabetes: results from the Anniston Community Health Survey. \u003cem\u003eEnviron Health Perspect\u003c/em\u003e \u003cstrong\u003e120\u003c/strong\u003e, 727-732 (2012). https://doi.org:10.1289/ehp.1104247\u003c/li\u003e\n\u003cli\u003eAssaggaf, H.\u003cem\u003e et al.\u003c/em\u003e Polychlorinated Biphenyls and Pulmonary Hypertension. \u003cem\u003eInt J Environ Res Public Health\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e (2022). https://doi.org:10.3390/ijerph19084705\u003c/li\u003e\n\u003cli\u003eDeng, P.\u003cem\u003e et al.\u003c/em\u003e Hepatic metabolomics reveals that liver injury increases PCB 126-induced oxidative stress and metabolic dysfunction. \u003cem\u003eChemosphere\u003c/em\u003e \u003cstrong\u003e217\u003c/strong\u003e, 140-149 (2019). https://doi.org:10.1016/j.chemosphere.2018.10.196\u003c/li\u003e\n\u003cli\u003eSerdar, B., LeBlanc, W. G., Norris, J. M. \u0026amp; Dickinson, L. M. Potential effects of polychlorinated biphenyls (PCBs) and selected organochlorine pesticides (OCPs) on immune cells and blood biochemistry measures: a cross-sectional assessment of the NHANES 2003-2004 data. \u003cem\u003eEnviron Health\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 114 (2014). https://doi.org:10.1186/1476-069x-13-114\u003c/li\u003e\n\u003cli\u003ePrzybyla, J., Houseman, E. A., Smit, E. \u0026amp; Kile, M. L. A path analysis of multiple neurotoxic chemicals and cognitive functioning in older US adults (NHANES 1999-2002). \u003cem\u003eEnviron Health\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 19 (2017). https://doi.org:10.1186/s12940-017-0227-3\u003c/li\u003e\n\u003cli\u003eSasaki, N., Morse, G., Jones, L. \u0026amp; Carpenter, D. O. Effects of mixtures of polychlorinated biphenyls (PCBs) and three organochlorine pesticides on cognitive function differ between older Mohawks at Akwesasne and older adults in NHANES. \u003cem\u003eEnviron Res\u003c/em\u003e \u003cstrong\u003e236\u003c/strong\u003e, 116861 (2023). https://doi.org:10.1016/j.envres.2023.116861\u003c/li\u003e\n\u003cli\u003eAxelrad, D. A., Goodman, S. \u0026amp; Woodruff, T. J. PCB body burdens in US women of childbearing age 2001-2002: An evaluation of alternate summary metrics of NHANES data. \u003cem\u003eEnviron Res\u003c/em\u003e \u003cstrong\u003e109\u003c/strong\u003e, 368-378 (2009). https://doi.org:10.1016/j.envres.2009.01.003\u003c/li\u003e\n\u003cli\u003eMitro, S. D., Birnbaum, L. S., Needham, B. L. \u0026amp; Zota, A. R. Cross-sectional Associations between Exposure to Persistent Organic Pollutants and Leukocyte Telomere Length among U.S. Adults in NHANES, 2001-2002. \u003cem\u003eEnviron Health Perspect\u003c/em\u003e \u003cstrong\u003e124\u003c/strong\u003e, 651-658 (2016). https://doi.org:10.1289/ehp.1510187\u003c/li\u003e\n\u003cli\u003eMorgan, M., Deoraj, A., Felty, Q. \u0026amp; Roy, D. Environmental estrogen-like endocrine disrupting chemicals and breast cancer. \u003cem\u003eMol Cell Endocrinol\u003c/em\u003e \u003cstrong\u003e457\u003c/strong\u003e, 89-102 (2017). https://doi.org:10.1016/j.mce.2016.10.003\u003c/li\u003e\n\u003cli\u003eLan, T., Liu, B., Bao, W. \u0026amp; Thorne, P. S. Identification of PCB congeners and their thresholds associated with diabetes using decision tree analysis. \u003cem\u003eSci Rep\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 18322 (2023). https://doi.org:10.1038/s41598-023-45301-1\u003c/li\u003e\n\u003cli\u003eCasagrande, S. S., Lee, C., Stoeckel, L. E., Menke, A. \u0026amp; Cowie, C. C. Cognitive function among older adults with diabetes and prediabetes, NHANES 2011-2014. \u003cem\u003eDiabetes Res Clin Pract\u003c/em\u003e \u003cstrong\u003e178\u003c/strong\u003e, 108939 (2021). https://doi.org:10.1016/j.diabres.2021.108939\u003c/li\u003e\n\u003cli\u003eWu, M., Si, J., Liu, Y., Kang, L. \u0026amp; Xu, B. Association between composite dietary antioxidant index and hypertension: insights from NHANES. \u003cem\u003eClin Exp Hypertens\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 2233712 (2023). https://doi.org:10.1080/10641963.2023.2233712\u003c/li\u003e\n\u003cli\u003eTan, J., Liu, N., Sun, P., Tang, Y. \u0026amp; Qin, W. A Proinflammatory Diet May Increase Mortality Risk in Patients with Diabetes Mellitus. \u003cem\u003eNutrients\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e (2022). https://doi.org:10.3390/nu14102011\u003c/li\u003e\n\u003cli\u003eLi, W.\u003cem\u003e et al.\u003c/em\u003e Analysis of environmental chemical mixtures and nonalcoholic fatty liver disease: NHANES 1999-2014. \u003cem\u003eEnviron Pollut\u003c/em\u003e \u003cstrong\u003e311\u003c/strong\u003e, 119915 (2022). https://doi.org:10.1016/j.envpol.2022.119915\u003c/li\u003e\n\u003cli\u003eLi, L.\u003cem\u003e et al.\u003c/em\u003e The Association between Non-Alcoholic Fatty Liver Disease (NAFLD) and Advanced Fibrosis with Serological Vitamin B12 Markers: Results from the NHANES 1999-2004. \u003cem\u003eNutrients\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e (2022). https://doi.org:10.3390/nu14061224\u003c/li\u003e\n\u003cli\u003eYang, Z.\u003cem\u003e et al.\u003c/em\u003e Association between daidzein intake and metabolic associated fatty liver disease: A cross-sectional study from NHANES 2017-2018. \u003cem\u003eFront Nutr\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 1113789 (2023). https://doi.org:10.3389/fnut.2023.1113789\u003c/li\u003e\n\u003cli\u003eLi, Z.\u003cem\u003e et al.\u003c/em\u003e Distribution of lipid levels and prevalence of hyperlipidemia: data from the NHANES 2007-2018. \u003cem\u003eLipids Health Dis\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 111 (2022). https://doi.org:10.1186/s12944-022-01721-y\u003c/li\u003e\n\u003cli\u003eTan, Y.\u003cem\u003e et al.\u003c/em\u003e Relationship between phthalates exposures and hyperuricemia in U.S. general population, a multi-cycle study of NHANES 2007-2016. \u003cem\u003eSci Total Environ\u003c/em\u003e \u003cstrong\u003e859\u003c/strong\u003e, 160208 (2023). https://doi.org:10.1016/j.scitotenv.2022.160208\u003c/li\u003e\n\u003cli\u003eLiu, H.\u003cem\u003e et al.\u003c/em\u003e Association Between Diet-Related Inflammation and COPD: Findings From NHANES III. \u003cem\u003eFront Nutr\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 732099 (2021). https://doi.org:10.3389/fnut.2021.732099\u003c/li\u003e\n\u003cli\u003eZhao, L., Sun, Y., Liu, Y., Yan, Z. \u0026amp; Peng, W. A J-shaped association between Dietary Inflammatory Index (DII) and depression: A cross-sectional study from NHANES 2007-2018. \u003cem\u003eJ Affect Disord\u003c/em\u003e \u003cstrong\u003e323\u003c/strong\u003e, 257-263 (2023). https://doi.org:10.1016/j.jad.2022.11.052\u003c/li\u003e\n\u003cli\u003eMahemuti, N.\u003cem\u003e et al.\u003c/em\u003e Association between Systemic Immunity-Inflammation Index and Hyperlipidemia: A Population-Based Study from the NHANES (2015-2020). \u003cem\u003eNutrients\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e (2023). https://doi.org:10.3390/nu15051177\u003c/li\u003e\n\u003cli\u003eHe, K., Pang, T. \u0026amp; Huang, H. The relationship between depressive symptoms and BMI: 2005-2018 NHANES data. \u003cem\u003eJ Affect Disord\u003c/em\u003e \u003cstrong\u003e313\u003c/strong\u003e, 151-157 (2022). https://doi.org:10.1016/j.jad.2022.06.046\u003c/li\u003e\n\u003cli\u003eZhang, Y. B.\u003cem\u003e et al.\u003c/em\u003e Associations of healthy lifestyle and socioeconomic status with mortality and incident cardiovascular disease: two prospective cohort studies. \u003cem\u003eBmj\u003c/em\u003e \u003cstrong\u003e373\u003c/strong\u003e, n604 (2021). https://doi.org:10.1136/bmj.n604\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Table 1","content":"\u003cp\u003e\u003cstrong\u003eTable 1. Baseline demographic characteristics divided by LBX074 levels.\u003c/strong\u003e\u003c/p\u003e\n\u003cdiv\u003e\n \u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eOverall\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eLBX074\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003ep-Value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eLow\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eModerate\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eHigh\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNumber\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e10961\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e4015\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3304\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3642\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAge\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;40.99 (21.36)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;25.80 (12.77)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;37.78 (18.11)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;61.98 (14.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eGender (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Male\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 5457 (49.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2128 (53.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1847 (55.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1482 (40.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Female\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e5504(50.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1887(47.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1457(44.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2160(59.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eRace (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Mexican American\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2692 (24.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1351 (33.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;824 (24.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;517 (14.2)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Non-Hispanic Black\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2353 (21.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1021 (25.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;679 (20.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;653 (17.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Non-Hispanic White\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 5069 (46.2)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1295 (32.3)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1484 (44.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2290 (62.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Other Hispanic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;438 (4.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;176 (4.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;161 (4.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;101 (2.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Other races\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;409 (3.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;172 (4.3)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;156 (4.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 81 (2.2)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eEducation (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Less than 9th grade\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1530 (14.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;600 (15.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;439 (13.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;491 (13.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;9-11th grade\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1942 (18.2)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;829 (21.3)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;533 (16.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;580 (16.3)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;High-school graduate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2454 (23.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;889 (22.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;748 (23.3)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;817 (23.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;College graduate or above\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1995 (18.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;581 (14.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;673 (20.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;741 (20.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Some college or AA degree\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2679 (25.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;968 (24.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;810 (25.2)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;901 (25.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;others\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 56 (0.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 20 (0.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 13 (0.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 23 (0.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ePIR\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2.50 (1.60)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2.17 (1.56)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2.61 (1.63)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2.77 (1.56)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eBMI stage (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;\u0026lt;25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4488 (42.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2158 (55.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1361 (42.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;969 (28.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;\u0026gt;30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2776 (26.3)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;724 (18.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;852 (26.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1200 (35.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;25-30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 3300 (31.2)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1035 (26.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1020 (31.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1245 (36.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSmoking exposure (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Current smoker\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2596 (23.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;934 (23.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;938 (28.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;724 (20.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Former smoker\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 6646 (61.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2529 (63.2)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2003 (61.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2114 (58.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Non smoker\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1649 (15.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;537 (13.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;338 (10.3)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;774 (21.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAlcohol intake (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Current drinker\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;497 (21.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;136 (23.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;128 (22.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;233 (19.3)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Former drinker\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;696 (29.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;107 (18.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;170 (29.2)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;419 (34.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Non drinker\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 1173 (49.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;331 (57.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;285 (48.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;557 (46.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSII\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e582.86 (367.92)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e557.14 (340.40)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e583.96 (395.89)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e610.40 (368.91)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCRP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.38 (0.81)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.26 (0.56)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.38 (0.93)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.52 (0.89)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eWBC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 7.09 (2.18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 7.04 (2.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 7.18 (2.12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 7.07 (2.39)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eLymphocyte\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2.13 (1.01)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2.18 (0.62)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2.15 (0.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 2.07 (1.49)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMonocyte\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.56 (0.18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.55 (0.18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.56 (0.18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.57 (0.19)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNeutrophils\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.16 (1.65)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.08 (1.68)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.22 (1.75)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.18 (1.51)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ePlatelet count\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e272.08 (67.52)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e277.92 (63.12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e274.24 (72.13)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e263.63 (67.08)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eRed blood cell\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.74 (0.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.81 (0.49)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.79 (0.48)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.60 (0.51)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eHemoglobin\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;14.28 (1.51)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;14.38 (1.53)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;14.43 (1.55)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;14.05 (1.40)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAlkaline phosphatase\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;90.57 (62.34)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e102.04 (76.08)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;94.53 (68.44)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;74.22 (24.98)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAlbumin\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.32 (0.33)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.38 (0.29)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.39 (0.36)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 4.20 (0.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eBilirubin\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.72 (0.29)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.74 (0.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.71 (0.33)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp; 0.71 (0.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eIron\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;88.34 (37.89)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;90.12 (40.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;90.86 (38.84)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026nbsp;84.06 (34.05)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eThe baseline table is a calculation from NHANES 1999-2000,2001-2002,2003-2004. For categorical variables, the p-value was calculated by the chi-square test. For continuous variables, the p-value was calculated by t-test.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Polychlorinated Biphenyls, Machine Learning, Mortality, Comorbidities, DII","lastPublishedDoi":"10.21203/rs.3.rs-4543285/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4543285/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003ePolychlorinated Biphenyls (PCBs), as a major class of organic pollutants, garnered increasing attention due to their significant ability for inter-regional accumulation and migration, prolonged half-life, and relatively high toxicity. Our study aimed to assess the impact of PCBs on various diseases and mortality risks using data from the National Health and Nutrition Examination Survey (NHANES), while proposing lifestyle adjustments, particularly dietary modifications, to mitigate mortality risk. Statistical analyses employed principal component analysis (PCA), multifactorial logistic regression, multifactorial Cox regression, comorbidity network analysis, and machine learning prediction models. Results indicated significant associations between 7 types of PCBs and 12 diseases (p \u0026lt; 0.05), with 6 diseases showing significant positive correlations (OR \u0026gt; 1, p \u0026lt; 0.05), along with listing the 25 most relevant diseases, such as asthma and chronic bronchitis (OR (95%CI) = 5.85(4.37,7.83), p\u0026lt;0.0001), arthritis and osteoporosis (OR (95%CI) = 6.27(5.23,7.55), p\u0026lt;0.0001). We proposed that PCB exposure ultimately triggered adverse progression of multiple diseases and increased mortality risk, suggesting PCBs could potentially serve as specific biomarkers for certain diseases in the future. Building upon this, we further suggested that controlling dietary intake to reduce dietary inflammatory index (DII) could lower mortality and disease risks. While PCBs were independent risk factors for mortality, ample evidence suggested that adjusting DII might mitigate the adverse effects of PCBs to some extent. Further physiological mechanisms require deeper exploration through additional research.\u003c/p\u003e","manuscriptTitle":"Polychlorinated Biphenyls' Impact on Comorbidity Networks: Unveiling Epidemiological Patterns and offsetting Through Dietary Adjustments","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-06-24 18:08:03","doi":"10.21203/rs.3.rs-4543285/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"b3568edc-34f7-41ae-bb3c-4211393e6805","owner":[],"postedDate":"June 24th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-07-17T06:22:33+00:00","versionOfRecord":[],"versionCreatedAt":"2024-06-24 18:08:03","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4543285","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4543285","identity":"rs-4543285","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}