October 22, 2014

bisphenol-A and till receipts

A study published in the journal PLOS ONE has investigated the levels of exposure to bisphenol-A through the handling of till receipts, eating, and use of hand sanitiser. Roundup comments accompanied this before the headlines analysis.

Title, Date of Publication & Journal

Holding Thermal Receipt Paper and Eating Food after Using Hand Sanitizer Results in High Serum Bioactive and Urine Total Levels of Bisphenol A (BPA).   October 2014  PLOS ONE

Study’s main claims – and are they supported by the data?

The paper does not prove the claim that the touching of thermal receipt paper results in high levels of Bisphenol A in Urine or in Serum.

This paper provides no further evidence of health impacts of BPA.

The paper is composed of four sub-experiments, which look at the BPA content of till receipts and the transfer of the chemical via the hands into the bloodstream.

The only sub-experiment which examined claims outlined in title was conducted in 10 subjects (Experiment 3) using paper not representative of samples of till receipt paper previously analysed.

Experiment 1 confirmed the presence of BPA in thermal receipts in only 44% (22/50) and BPS in 52% (26/50) samples tested.  This experiment simply confirms the concentration of the chemicals in the paper.

Experiment 2 looked at the transfer of BPA to the skin surface with/without use of hand sanitizer.  3 subjects were used with hand sanitizer and each was used six times (different lengths of holding receipt ranging from 2 seconds to 4 minutes).  The 4 subjects were used without hand sanitizer each was used twice (1 minute or 4 minutes).  This is low and I would have expected more like 12 subjects in each case; such low numbers are prone to misleading results.

The receipt paper selected was the one with a higher than average concentration; 27.2mg BPA/g was higher than the highest of the 50 samples tested in Experiment 1 and so is not representative of normal till receipts.

Greater concentrations were observed in samples taken after use of hand sanitizer.  The peak time was 45 seconds with no clear explanation as to why concentrations decreased with greater holding times.  This is counter-intuitive and seems unlikely, especially as the opposite was seen with dry hands.  I would want to see this confirmed, otherwise it could just be a result of using such small sample sizes.

Experiment 3 looked at the transfer of BPA to blood after using hand sanitizer in 10 subjects  (5 males and 5 females) eating French fries after holding a receipt.

Lower levels of BPA were observed 8 minutes after the start of the experiment compared to Experiment 2, and speculation was that this was due to transdermal absorption and transfer to French fries.

No statistically significant findings were reported in this experiment.

Only 7/10 subjects had blood samples taken from the arm used to handle the French fries (one of these subjects was excluded from the main analysis).  These 7 showed no clear pattern of the effects of BPA in the blood.

The remaining 3/10 subjects had samples taken from their other arm; no explanation was given for this.  All claims about raised BPA in the blood were based on just 2 of these 3 subjects.  Both were female, and the authors speculate that the females are prone to greater absorption although the sample sizes are far too small to demonstrate this.

Overall, these are very small numbers and no firm conclusions can be drawn, despite the main claims of the study being derived from this experiment.

Experiment 4 looked at the transfer of BPA to urine and serum in 24 subjects NOT using hand sanitizer.  The only statistically significant difference in this experiment was observed between male and female subjects in Conjugated BPA in serum, which found higher levels in males.  This contradicts the gender speculation in Experiment 3.

Strengths/Limitations

Strengths

1. Attempts made to limit BPA exposure to till receipt prior to the experiment through baseline washing procedures
2. Examination of variability of sample-to-sample BPA levels between suppliers of till receipt paper.

Limitations

1. Only 44% of receipt paper contained BPA
2. The receipt paper used was not representative of receipt paper
3. Very small sample-sizes used throughout
4. Not representative of real life situation (holding a receipt for 4 minutes prior to eating)

Before The Headlines is a service provided to the SMC by volunteer statisticians: members of the Royal Statistical Society (RSS), Statisticians in the Pharmaceutical Industry (PSI) and experienced statisticians in academia and research.  A list of contributors, including affiliations, is available here.