Association between urinary Bisphenol A & the reporting of cardiovascular diseases : case of American adults

Objective: Bisphenol A is categorized as an organic synthetic compound, which is commonly used for developing epoxy resin and polycarbonate plastics. Bisphenol A is measured as endocrine disrupter, which consists of estrogens receptors and estrogenic influence within laboratory findings. Purpose: The purpose of this study was to explore the association of urinary Bisphenol A with cardiovascular disease. Urinary BPA is considered as independent variable while CVD was treated as dependent variable. Methods: The participants were selected in the study through statistical process commenced with the data provided in 2009/10 survey by NHANES. The quantitative approach was implemented in the study to explore the association between urinary BPA and reporting of cardiovascular disease. The adult U.S. population between the age group 18 to 80 years was included as a targeted population. Results: The step 1 in binary logistic regression examined the correlation among presence of cardiovascular disease and the controls, predicting controls to be significant in forecasting the presence of cardiovascular disease. The step 2 binary logistic regression indicates the significant relation among the independent variable and controls to explore the occurrence of cardiovascular disease. Conclusion: There were no changes found by the coefficient of two-step binary regression with covariates and independent variable Bisphenol A to identify cardiovascular disease. Age was revealed as the only control variable among other variables that estimates the reporting of cardiovascular disease.


Introduction
Bisphenol A (BPA) is considered as a highly industrial chemical exercise in the manufacturing of polycarbonate plastics and epoxy resin.Epoxy resin is usually implemented as internal facing of metallic foods and drink containers to avoid deterioration, while polycarbonate plastics are met in many commodities particularly in drink and food containers.U.S. food and drug administration synchronizes the use of Bisphenol A in food contact materials (U.S. Environmental Protection Agency, 2010).Some types of thermal papers are covered by the execution of Bisphenol A, which are frequently implemented as receipts from automatic teller machines and cash registers.It is used in the factories where it is executed to make moldings and casts; however, it is also used in the poly vinyl chloride industries.
Bisphenol A is exposed through exhaustion of household dust and diet of humans.BPA is considered as endocrine disrupter, which combines estrogens receptors and is found as estrogenic influence in laboratory findings.Bisphenol A has been originated to have a subordinate similarity for nuclear estrogens receptors comparative to 17-beta estradiol, where effectiveness of estrogenic is related to 17-beta estradiol for reactions arbitrated through non-nuclear estrogens receptors (Viñas, Jeng, and Watson, 2012, 2694-2714).Endogenous androgen actions are blocked by Bisphenol A where it could directly combine with androgen receptors and it is potentially anti-androgenic (Wetherill, Akingbemi, Kanno, et al, 2007, 178-198).
ISSN 2309-3374 E-ISSN 2375-4095 Bisphenol A has been exposed to combine with thyroid receptors, which affects the functioning of thyroid through agonistic and antagonistic effects (Wetherill, Akingbemi, Kanno, et al, 2007, 178-198).Bisphenol A has become a regular environmental chemical with the wide handling of BPA-based commodities.The unbound monomer Bisphenol A molecules in the products are left in the incomplete polymerization reaction of BPA (Le, Carlson, Chua, and Belcher, 2008, 149-156).It has been observed that production process of industries discharges above 100 tons of Bisphenol A yearly into the environment (Vandenberg, Maffini, Sonnenschein, 2009, 75-95).Several human revelation assessment studies have highlighted that Bisphenol A is at demonstrable stages in more than 90% entities explored in assorted inhabitants.
The association between cardiovascular disease risk factors and BPA exposure levels has also been examined.The NHANES data was studied by Shankar, et al. in 2003Shankar, et al. in -2004, in order to analyze the diastolic and systolic blood pressure of the participants.Hypertension is considered as an important factor of risk for cardiovascular disease, as a major public health problem.There was a positive association found between the levels of urinary BPA concentration and hypertension diagnoses (Shankar et al, 2012, p. nd).In the elderly population of Korea, the association between the heart rate variability and urinary BPA levels was investigated.The (HRV) heart rate variability is a fine system of heat to regulate with changes constantly in demand of blood supply.The study of Shankar and Teppala observed a positive relationship between HRV and higher urinary BPA (Bae et al, 2012, 786-793).
The study by Lakind et al, carried out that the data by NHANES might not be enough to sketch the conclusion that the exposure of BPA was linked with the cardiovascular disease.They observed the irregularities in the earlier studies of NHANES in terms of definitions and methods, and applied reliable methods in their analysis.Their outcomes showed that there is no association found in the urinary BPA with heart attack, diabetes or coronary artery disease in the participants of NHANES (Lakind et al, 2012).It is a fact that different methods could result in conflicting outcomes raised by the problems on how consistent the data of NHANES is for establishing the relationship between the levels of BPA and CV diseases.It was proposed that the conclusions must be made with concerns using the surveys of NHANES.
The FRS (Framingham Risk Score) system was used to evaluate the risk of cardiovascular disease in elderly population.The Framingham Risk Score includes gender, age, blood pressure, history, and glucose and blood lipid levels and provides estimation of CV disease risks.No relationship between the FRS score and serum BPA were identified (Lakind et al, 2014, 121-150).The usage of specific CV endpoints and CV risk estimates might contribute to the outcomes by Olsen et al and other studies.The multiple longitudinal and cross sectional studies of epidemiology showed that BPA exposure levels are linked with CV disease risk factors in adults.These comprises of numerous independent NHANES analyses, which demonstrated that participants who reported the CV diseases were associated with the concentration of higher urinary BPA.Except diabetes, no other diseases were associated with the urine level of BPA.Alternatively, no association between CV disease and exposure of BPA was observed in other studies (Olsen et al, 2012, 179-183).
1.1 Problem statement: Plastic is interrupted when revealed to high heat with the chemical bond of BPA, which exhales Bisphenol A into the human foods (Halden, 2010).This study revealed that Bisphenol A enters into the human body by foods and drinks through oral revelation.Bisphenol A is an endocrine disrupter that belongs to the cluster of chemicals, which responds in the biological reaction equivalent to estrogen (Cantonwine, Meeker, Hu, 2010).Bisphenol A combines to the estrogen receptors of similar cells and spoils health, reproductive systems of life, and developmental phases.
1.2 Significance of the problem: According to Zoeller, Brown, Doan, (2012), Bisphenol A was related to the disrupted pancreatic functions and liver damage, which frequently results in diabetes.BPA was also associated to thyroid function disruption and obesity (Shankar, Teppala, & Sabanayagam, 2012, 1297).The association between CVD and BPA was found in the study conducted by NHANES in 2003/04.CVD were correlated with socioeconomic factor, diabetes, dyslipidemia, and unhealthy lifestyle choices.Therefore, this study was conducted to explore the association between BPA and CVD using data collected by NHANES in 2009/10.

Purpose:
The purpose of the study was to explore the association of urinary Bisphenol A with cardiovascular disease.The research introduces urinary BPA as independent variable while CVD was treated as dependent variable.CVD was defined as any reporting of heart attack, angina, coronary heart diseases, and congestive heart failure.
1.4 Research question and hypothesis: The question that stood in this study was to explore the association of urinary BPA with CVD; therefore, the research question and hypothesis was grounded in accordance to the following problem.
Is there an association of Bisphenol A and CVD, after adjusting for age, gender, race/ethnicity, arthritis, and urinary creatinine?
The null and alternate hypotheses are as follows: Null Hypothesis (H0): There is no association of Bisphenol A and CVD, after adjusting for age, gender, race/ethnicity, arthritis, and urinary creatinine.

Alternate Hypothesis (Ha):
There is an association of Bisphenol A and CVD, after adjusting for age, gender, race/ethnicity, arthritis, and urinary creatinine.
1.5 Theoretical framework: The theoretical framework of this study adopted the correlation approach to observe the association between urinary BPA and CVD through the data collected by NHANES in 2009/10.There were several techniques included when correlation approach is implemented in the study.1.6 Limitations of the study: The characteristics of BPA were studied in this research within the sample of U.S. population.Larger sample size was advantageous to maintain the reliability of the findings (Hackshaw, 2008).The width and depth of the study was depending on the study as the results of this research were based on 95% confidence interval.

Research design:
The association between concentration of urinary BPA and reporting of cardiovascular disease was studied by using the quantitative non-experimental design.The targeted sample for this study was represented by adult U.S. population between the age of 18 and 80 years.

Population:
The survey conducted in 2009/10 by NHANES included 13,272 individuals, which is considered to be the population of the study; however, 10,537 persons were surveyed out of the total population.Oversampling was suggested for persons with age 60 and older to produce reliable results.Therefore, these results play dynamically a great role in the health culture of the U.S. population.

Sampling procedure:
The participants were recruited in the study through a complex statistical procedure initiating with the census information available in 2009/10 survey.There were 15 primary sampling units created to divide the United States based on explicit features.These sampling units were then distributed into more small segments where each selected segment comprises of 30 households.The interview was then conducted by the interviewers of NHANES to collect the data such as age, sex, ethnicity, and income level.A computer algorithm was used to include households (i.e., all, one or more, none) to participate in the survey.A distinctive health profile was constructed for particular patient after the selection process was completed.15 trained representative's visits throughout the year to gather information on their health and dietary status, which were appointed through random selection.
A sample of 1465 individuals was selected for the survey.The informed approval was handed over to the selected participants.The participants of NHANES were confined by the Public Health Act that deals with endorsement for obtaining data from participants.The selected participants were also confined through the confidential information protection, Privacy Act of 1974 and Statistical Efficiency Act, which exclude disclosure of information provided by the participants without their approval.
ISSN 2309-3374 E-ISSN 2375-4095 2.4 Sample size, power analysis and precision: The chosen level of significance for this study was set at 0.05 on the fixed sample size.This level of significance was set to reduce the probability of type II errors.G*power 3.1.7was used to conduct power analysis for the study.The multivariate logistic regression was implemented in the study.The multiple binary logistic regressions were used in the study and power analysis was done through two-tail binary regression model.The multiple binary logistic regression was conducted with a confidence level of 95%, a power of 0.31, a sample size of 1465, an effect size (OR=1.10).
2.5 Study variables and covariates: The covariates, independent, and dependent variables are also listed in table 1.The dependent variable for this study was considered as cardiovascular disease reporting while urinary BPA was mentioned as independent variable.
Other self-rating scales have been generated throughout the years by Chinese scholars from Chinese medicine point of view.But these evaluation methods focus too much on Chinese medicine syndrome and hard to make judgment of qualitative and quantitative significance, and therefore cannot be used in the clinical practice effectively.

Data collection:
The data files were collected for NHANES in 2009/10 from the website of CDC in SAS file format.Statistical software for social sciences (SPSS) version 21.0 was used to enter the data.The demographics and research variable were described through descriptive analysis.Nominal data such as occurrence of CVD was calculated through frequency and percentages.Secondary data was included in the study from CDC by survey respondents.Missing data, outliers or extreme cases were screened from the data.Frequency range and descriptive statistics were carried out to verify that replies were inside the statistically satisfactory range of values and also to explore that the data was not imprecise by outliers.The study excludes those participants who did not complete the major sections.

Data analysis:
The descriptive statistics for the dependent, independent and covariate variables were constructed through frequency table.Percentages for the occurrence of study were calculated to highlight the central tendency, which includes mean, median, minimum and maximum and standard deviation.The ordinal and categorical variables were summarized and presented to show frequency distribution in frequency tables.The significant model for the logistic regression was examined through the influence of the independent variable that was presented by χ 2 coefficient.The Nagelkerke R 2 coefficient of determination was stated and implemented to estimate the variation in dependent variable by the group of independent variables.A binary multivariate logistic regression was conducted in this study.Binary logistic regression was appropriate to use because the study includes the dependent variable as CVD that was dichotomous in nature, which means that an event could be predicted through two possible outcomes (Sullivan, 2010).Several controls were measured in the study that includes age, gender, race, arthritis, and urinary creatinine levels.In the first step of regression, these controls were entered in order to recognize the contribution of CVD.BPA levels were added in the second part of the regression (Tabachnick & Fidell, 2012).

Results
From the final 1453 participants, there were 729 female and 724 male.Non-Hispanic white were in majority (661) while multiracial were in the least number (82).For CVD and arthritis, 1360 and 1444 participants respectively stated that they have not reported CVD and arthritis.Among the participants who were classified as having cardiovascular diseases, 49% participants reported that they had been classified with coronary heart disease, Heart attack (28) Angina pectoris (28) The descriptive analysis for the continuous variables, which include the age of the participants at the time of the survey classified from 20 to 69 years with mean of 43.56 years and standard deviation of 14.10 years.Creatinine levels with mean of 127.08 and standard deviation of 75.71 ranged from 4 mg/dl to 382 mg/dl.A mean of 3.29 and standard deviation of 5.71 was measured for BPA levels ranging from 0.28 ng/ml to 112.00 ng/ml.The descriptive statistics for the continuous variables are shown in table 3. The following results suggested that there was a significant association between independent variable, dependent variable, and control variable.The results for step 1 in stepwise binary logistic regression examined the association between presence of CVD and the controls, which suggested that the controls were significant in predicting the presence of CVD.The occurrence of CVD was indicated by controls, which concluded that 15% of the variance is explained by controls.The age was the only control variable that has a direct influence on CVD.The controls and BPA levels were included in the second step of the binary logistic regression.
ISSN 2309-3374 E-ISSN 2375-4095 The results for regression analysis indicated that there was a significant association between the independent variable and controls to estimate the instances of CVD.The BPA level was not found to be more predictable with CVD as coefficient of determination did not increase from 15% explained by the controls.The null hypothesis fails to be rejected and also there was no evidence found in the association between BPA levels and occurrence of CVD.Table 4 shows the results for stepwise binary logistic regression.The classification table (Table 5) was examined to describe the goodness of fit for the regression model.The classification model highlights the 1453 participants, which predicts that 93.60% participants were not having cardiovascular disease.

Discussion
4.1 Interpretation of the Findings: The results of the study estimated that there was no relation among BPA levels and instance of CVD.The results further suggested that there was an important association among independent variables, dependent variable and the controls.The binary logistic regression predicted that the control (age) was the only predictor.It was also estimated that there was no relationship found between urinary Bisphenol A levels and reporting of cardiovascular disease.

Urinary Bisphenol A and Reporting of Cardiovascular Disease:
The focal point of the study was to explore the relationship among Bisphenol A and cardiovascular disease in individuals of NHANES 2009/10.It was estimated that age was the only predictor that predicts the reporting of CVD while independent variable (BPA level) was insignificant variable to predict the instances of CVD.Therefore, we could suggest that there was no association found between urinary BPA levels and reporting of CVD based on the estimated results.NHANES has undergone wide-spreading testing and confirmed measured validity, analytical validity, internal consistency, test-retest reliability, content and construct validity to ensure validity and convenience of the data provided to researchers.Threats were also minimized and maintained by computerized results through data entry to assure the validity.
The secondary data used in this study could not be verified as it is supposed to be the limitations in study to conduct under the support of CDC.After filtering and processing the data, the population of 10,537 was reduced to 1465 participants.The reliability of the data is reflected in the research through the large sample.The sample size resulted through logistic regression suggested high level of power and therefore, the sample size was recognized to be applicable and reliable.The study could assess with higher reliability but data included in the study was limited to 1465 participants.The reporting of cardiovascular disease was also found to be limited to angina, coronary heart disease, congestive heart failure, and heart attack with elimination of several heart diseases.

Conclusion & Recommendations
There are several essential changes concerned from the findings of the present study at the assorted level of society.BPA is identified as a chemical that demands more testing and research.This study grants a diverse outlook on chronic disease especially on CVD.The factors contributing to cardiovascular disease have been recognized hence, the contribution of Bisphenol A established deprived concentration from investigators.The association between BPA and CVD is indicated through the results, which suggests that the parameters are not statistically significant.There was no change found in the coefficient of determination indicated through two step binary logistic regressions with covariates and BPA levels to indicate CVD.The analysis exposed that age was the only control variable among other variables that directly predicts the reporting of CVD.
The definition of CVD should be extended to be more comprehensive.Researchers in future could include abnormal heart rhythm, aorta diseases, heart valve diseases, and pericarditis.Wider descriptions of cardiovascular disease will include larger sample size and enhance the reliability and overview of the study.The comparison of BPA level before and after reporting CVD would illuminate the association, which might be measured dependent.The present study implements age, gender, race, arthritis and urinary creatinine as controls.This feature of the study was recommended through the inclusion of controlling diseases such as obstructive sleep apnea, severe peripheral vascular diseases, chronic renal failure, and stroke.The future research should use a larger sample size to highlight U.S. population in a better way.

Table 1 Independent, dependent, and Covariate variables of the study, NHANES 2009/2010
diagnosed with heart failure, 30% reported angina pectoris, and 30% reported with at least one past heart attack.The demographics for the percentages including frequencies and percentages are shown in table2.