Epidemiology studies

Study

Key findings/claims

FSANZ response

Association Between Bisphenol A Exposure and Risk of All-Cause and Cause-Specific Mortality in US Adults

Bao et al. (2020) JAMA Network Open 3(8):e2011620. doi:10.1001/jamanetworkopen.2020.11620

• In a cohort study of 3883 US adults enrolled in the National Health and Nutrition Examination Survey (NHANES), higher urinary BPA levels (measured in 2003-2008) were associated with an increased risk of death from all causes during approximately 10 years of observation.
• Urinary BPA levels showed some association with deaths from cardiovascular disease, although these were not statistically significant.
• There was no association between BPA and deaths from cancer.

• BPA levels were only assessed in one urine sample collected from each person. BPA is rapidly eliminated so a single measure only reflects very recent exposure and does not reliably estimate exposure over the long-term. More recent data from the NHANES survey suggest that BPA exposures may have decreased over time. Therefore caution is needed in assessing whether single BPA measures collected at one point in time are causally linked to health outcomes that develop over years.
• A relatively small number of people were studied and the number of deaths was also small. This results in a wide range of estimates for the size of the association. The lower end of this range was close to no association.
• This is the first study reporting an association between BPA and increased mortality in humans. Increased mortality has not been observed as a critical effect in animal toxicity studies, and the mechanism by which BPA might increase mortality from all causes is not known.
• Given the limitations of this study, the findings do not provide convincing evidence of a causal association between BPA exposure and all-cause mortality.
• As noted by the authors, more research is needed to see if these findings can be reproduced in other populations.

'Serum unconjugated bisphenol A concentrations in women may adversely influence oocyte quality during in vitro fertilization'

Fujimoto et al (2011) Fertility and Sterility95(5):1816-1819.

• Higher blood levels of BPA in women undergoing in vitro fertilisation (IVF) may negatively affect the maturation of ova and reduce the probability of fertilisation.
• For the nine Asian women in the study, the statistical calculations predicted that a doubling of BPA blood level would be associated with a 9% decrease in the probability of obtaining a mature ovum.
• A doubling in female serum BPA concentration was predicted to result in a 55% decrease in the probability of fertilisation, while a doubling in male serum BPA concentration was predicted to result in a 12% reduction in fertilisation probability for the five Asian men in the study.

• This publication was brief (3 pages) and lacked detail concerning the raw data used as input for the statistical calculations. For some statistical analyses, it appears that that the number of model variables was excessive with respect to the number of experimental data points available. This can lead to unreliable statistical conclusions.
• When all study participants were considered, there was no association between BPA and ova maturation.
• It was stated by the Authors that the small number of participants in the study precluded conclusive evaluation of the associations detected between BPA and ova maturation. They also stated that the results may have been biased by the possibility that some meiosis events were misclassified.
• Several factors relevant for the success of the IVF procedure were also ignored (e.g. the amount of human chorionic gonadotrophin (hCG) administered, oestrogen and progesterone levels at the time of hCG administration, whether fresh or frozen sperm were used for fertilisation, the number of oocytes retrieved from each women).
• An additional limitation of the study is that BPA was the only environmental chemical considered in the statistical analyses.
• The Authors concluded that the study was preliminary and that further studies are needed.
• It is likely that the reported associations occurred by chance and that the low levels of BPA exposure reported in the paper have no association with adverse IVF outcomes.

'Association of urinary BPA concentration with heart disease: evidence from NHANES 2003/06'

Melzer et al (2010) PLoS One5(1):e8673

• Higher urinary concentrations of BPA are associated with heart disease in the general adult population of the USA.

• Cholesterol and triglyceride levels were not taken into account as confounding factors. 
• The deficiencies described above for the Lang et al (2008) study also apply for this study.
• In contrast to Lang et al (2008), there was no association with diabetes.

'Occupational exposure to BPA and the risk of self-reported male sexual dysfunction'

Li et al (2010a) Human Reproduction25:519-527

'Relationship between urine BPA level and declining male sexual function'

 Li et al (2010b) Journal of Andrology31:500-506

'Urine BPA level in relation to semen quality'

Li et al (2011) Fertility and Sterility95(2):625-630

• Findings from this study, published in 3 separate journal articles, are that urine BPA levels in male chemical factory workers are associated with a higher risk of sexual dysfunction and decreased semen quality.

• A major shortcoming of this study is that no data were available on occupational exposure to chemicals other than BPA. The workers examined in this study were employed in manufacturing plants that produced epoxy resins and BPA. It is likely that such workers would have been exposed to a variety of different chemicals by non-oral routes (e.g. inhalation, dermal contact). However, BPA was the only chemical assayed in urine.
• A separate study found no significant associations between urinary BPA concentration and any sperm parameter in 375 men from four U.S. cities (Mendiola et al (2010) Environmental HealthPerspectives118,1286-1291).

'Prenatal BPA exposure and early childhood behavior'

Braun et al (2009) Environmental Health Perspectives 117(12):1945-1952

• Prenatal BPA exposure, as measured by maternal urinary BPA levels during pregnancy, may be associated with hyperactivity and aggression in two-year old children, especially among female children.

• Behaviour of the children was assessed at only one time point and this may not adequately reflect the overall behaviour pattern during early childhood.
• The reported statistical associations could potentially be affected by confounding factors that were not taken into account such as maternal behaviour toward the child, parental psychopathology, and alcohol/drug consumption.
• The authors themselves identified statistical limitations of the study and also stated that the results could be biased due to uncharacterised confounding factors and to inaccurate characterisation of BPA exposure.

'Association of urinary BPA concentration with medical disorders and laboratory abnormalities in adults'

Lang et al., (2008). Journal of the American Medical Association300:1303-1310

• In a study on 1455 US adults, higher urinary BPA concentrations were associated with cardiovascular diagnoses, diabetes, and clinically abnormal concentrations of two liver enzymes.

• The cause of higher urinary BPA levels could plausibly be linked to higher consumption of packaged food resulting in increased fat, sugar and salt consumption which may in turn increase the risk of diabetes and heart disease. That is, a higher BPA level may not have any causal effect but is simply a marker of a poor diet. Attributing any cause and effect relationship from this study would be unsound.
• Urinary BPA levels reported in this study represent a snapshot of short term exposure whereas longer term factors are more likely to be relevant to the development of diabetes and cardiovascular disease.
• In a subsequent issue of the journal, authors of three letters to the editor criticised several aspects of the study. These criticisms included: (i) diabetes was self-reported and there was no distinction between type I and type II diabetes (ii) the severity of the self-reported diabetes was not reported; and (iii) based on statistical considerations, the potential for false positives was considered to be substantial.