August 15, 2012

expert reaction to research into urinary bisphenol A and arterial stenosis

A study of British men and women in the journal PLoS ONE found a link between high levels of bisphenol A and the clogging up of the arteries that provide the heart with oxygen-rich blood.

Prof Richard Sharpe, MRC Centre for Reproductive Health, University of Edinburgh, said:

“This well conducted study has shown a significant association between urinary levels of BPA and the occurrence of narrowing (stenosis) of the coronary arteries in patients attending hospital for a heart angiogram (implying that they had suspected heart problems). The study findings add to a previous study from the same group showing a significant association between urinary BPA and cardiovascular disease.

“The study does not show that BPA causes narrowing of coronary arteries (usually because of plaque formation – atherosclerosis), only that the two are associated. As diet (food, drink) is established to account for ~95% of BPA intake by the general population, and as (poor) diet is an established risk factor for development of atherosclerosis, it is difficult to see how any effect of BPA exposure can be distinguished from those of diet, especially as atherosclerosis is known to develop over decades (and thus to reflect long-term diet). Moreover, virtually all BPA is inactivated in the gut and liver before it can reach the circulation, so only miniscule amounts will reach coronary blood vessels. How such small amounts of BPA could then cause atherosclerosis is unknown as no relevant mechanism has been shown for such low levels.

“Established major causes of atherosclerosis are: poor diet, obesity, high cholesterol levels, high blood pressure, age, male gender, smoking and physical inactivity. If BPA has any causal role to play, it is likely to be minimal in comparison to these established risk factors.”

Prof Warren Foster, Department of Obstetrics & Gynecology, McMaster University, said:

“The most recent paper by Melzer and colleague describes a cohort study in which the author’s suggest that urinary concentrations of Bisphenol A are higher in patients with severe coronary artery disease (CAD) compared to individuals in whom a diagnostic angiogram revealed no evidence of CAD. Thus, this study contributes to the growing literature suggesting a causal association between BPA exposure and increased risk for cardiovascular disease. The present study has a number of strong points as well as important weaknesses that deserve consideration in any discussion of potential causal relationships between BPA exposure and CAD.

“The present study adds to the ever expanding body of literature attempting to link BPA exposure with adverse health outcomes in general and CAD in particular. This study has a number of strengths including cohort study design, inclusion of a diagnostic angiogram to classify study participants, a direct measure of exposure, and appropriate statistical methods. However, on balance, the weaknesses in the study including the approach to exposure assessment, failure to measure free BPA exposure, prior evidence that dietary BPA exposure does not result in quantifiable free BPA in the serum of study subjects, and differences in the number of men across the study groups calls into question the validity of the conclusions reached in this study. Hence, any discussion about potential causal associations is premature as is any suggestion that the general population is at any risk from BPA exposure sufficient to produce the concentrations measured in the urine of this study.”

Additional information from Prof Foster: Strengths: “Among the strengths of this study include the cohort study design involving 591 patients undergoing a diagnostic angiogram. The angiograms were read both for clinical purposes and then again for classification of study participants into those with none, intermediate or severe vessel disease for this study. Good agreement was achieved between the clinical and study assessments of the angiograms. Compared to many studies in the field a relatively large sample size was employed although the analyses suggest that a larger study will be needed as the intermediate group was limited to only 86 subjects which may have impacted the ability to detect a significant increase in risk in this group. Another strength of the study is the direct measure of exposure through quantification of urinary BPA concentrations using state-of-the-art methods. Inclusion of sensitivity analyses is also seen as a strength of the study.” Weaknesses: “The present study measured exposure via a single spot urine sample and quantified BPA concentrations of total BPA. Given that the author’s point out that the primary route of exposure is from food, a more rigorous study design would have included a measure of time from last food consumption or perhaps a fasting first morning void sample. Secondly, the exposure measure does not distinguish between free and glucuronidated BPA (metabolite of BPA) but simply reports a BPA concentration (total BPA). While I agree that the use of urine concentrations of BPA is the appropriate matrix to assess exposure, it is a concern that there was no method to account for potential dilution in the urine resulting from time from last void and consumption of fluids including caffeinated beverages (diuretics). A measure of urine creatinine would have easily addressed this concern. In addition, a report of the urine specific gravity would also have been helpful. Of much greater concern is a prior pharmacokinetic study which revealed that human exposure to known amounts of BPA in the diet did not produce measureable concentrations of free BPA in the serum and thus suggests that human dietary exposure is trivial and unlikely to produce meaningful concentrations of BPA, and thus concentrations of any relevance to CAD causation or progression. A further concern is that nearly 17% of the study participants had urine concentrations of BPA that were below the limit of detection (LOD) of the assay methods used. These data support the view that the data are highly skewed and data should be presented as the geometric mean and also raise concerns about further analysis when so many samples were below the LOD which will skew the data further. “It was interesting to note that while 51.7% of the study participants with no coronary artery stenosis were female whilst the number of females in the severe CAD group was only 21.8%. Given the protective effect of estrogens on cardiovascular health it was surprising that this issue was not discussed further. Moreover, prior studies have demonstrated higher BPA concentrations in women with polycystic ovarian syndrome (PCOS; a disease characterized by androgen excess) vs. healthy controls leading to speculation that androgens may affect (attenuate) metabolism of BPA. Therefore, the results of the present study may be no more than the consequence of more men and higher overall androgen (testosterone, dihydrotestosterone, androstenedione, and DHEAS) concentrations in the participants with sever vs. the group with no evidence of CAD disease.”
‘Urinary bisphenol A concentration and angiography-defined coronary artery stenosis’ by Melzer et al., published in PLOS ONE on Wednesday 15 August 2012.