7 Results

parabens in perspective: part


And not so finally . . . .

Having reported on the SCCS Final Opinion on parabens, I had decided that I would call it a day on writing articles on this subject. However, my previous article provoked an unusually lively discussion, with some interesting points being highlighted, and I intend to address these points by extending the series with another 3 articles.

This article will address the points raised over the effects of parabens on “SULT” activity. The article below seems to be the only study addressing this point, and I quote the entire abstract of the study here:

Parabens inhibit human skin estrogen sulfotransferase activity:  Possible link to paraben estrogenic effects

Toxicology, Volume 232, Issue 3, 11 April 2007, Pages 248-256
Jeffery J. Prusakiewicz, Heather M. Harville, Yanhua Zhang, Chrisita Ackerman, Richard L. Voorman


Parabens (p-hydroxybenzoate esters) are a group of widely used preservatives in topically applied cosmetic and pharmaceutical products. Parabens display weak associations with the estrogen receptors in vitro or in cell based models, but do exhibit estrogenic effects in animal models. It is our hypothesis that parabens exert their estrogenic effects, in part, by elevating levels of estrogens through inhibition of estrogen sulfotransferases (SULTs) in skin. We report here the results of a structure-activity-relationship of parabens as inhibitors of estrogen sulfation in human skin cytosolic fractions and normal human epidermal keratinocytes. Similar to reports of paraben estrogenicity and estrogen receptor affinity, the potency of SULT inhibition increased as the paraben ester chain length increased. Butylparaben was found to be the most potent of the parabens in skin cytosol, yielding an IC50 value of 37 ± 5 μM. Butylparaben blocked the skin cytosol sulfation of estradiol and estrone, but not the androgen dehydroepiandrosterone. The parabens were also tested as inhibitors of SULT activity in a cellular system, with normal human epidermal keratinocytes. The potency of butylparaben increased three-fold in these cells relative to the IC50 value from skin cytosol. Overall, these results suggest chronic topical application of parabens may lead to prolonged estrogenic effects in skin as a result of inhibition of estrogen sulfotransferase activity. Accordingly, the skin anti-aging benefits of many topical cosmetics and pharmaceuticals could be derived, in part, from the estrogenicity of parabens.

I do not intend to dwell on the specifics of the methodology, nor any possible flaws as the science behind this is outside of my area of expertise. Instead, I will focus solely on the context of the results produced.

The main area of concern is the result quoted – the IC50 of 37µM (+/-5 µM). This is the critical figure, and is quoted for butylparaben, as being the most potent of the parabens tested (this is entirely in agreement with expectations from other studies that demonstrate consistently the fact that the longer carbon chain parabens exhibit a stronger toxicological effect).

Converting the IC50 of 37µM figure into something more familiar; this is equivalent to 7 parts per million, or 0.0007% butylparaben.

The important thing here is to relate this result to a worst case exposure scenario. Here come the maths:

First of all, I will make my conversion from µM to % transparent (maths is not my strongest discipline, so I am giving the opportunity for someone to correct me if I am wrong, but please don’t pick me up on slight approximations!).

1M of butylparaben = 194g/kg

1 µM of butylparaben = 0.000194g/kg

37 µM of butylparaben = 0.00679g/kg

1g/kg = 0.1%

0.00679g/kg = 0.000679%

0.000679% = 7 parts per million (approximately)

So, this establishes that 7ppm is a toxicological end point for butylparaben in this study. Now we need to look at the actual exposure to butylparaben in order to understand the relevance if this figure. I will use the worst case scenario from the figures used in the SCCS final opinion, for consistency:

Assume average application of cosmetics as being 17.79g/day.

Assume ALL cosmetics contain butylparaben at the new maximum permitted concentration of 0.19%.

17.79g containing 0.19% = 0.0338g butylparaben

The important step here is to include the amount of butylparaben that may actually be absorbed through the skin, and the worst case scenario used by the SCCS was 3.5%.

3.5% of 0.0338 g = 0.00118g

The average weight of the consumer is taken as being 60kg

0.00118g in a 60kg human = 0.00002g/kg

From the conversion above, it can be seen that 0.00002g/kg = 0.000002%

0.000002% = 0.02ppm

So, assuming that my maths is correct, the ACTUAL total body exposure to butylparaben, using worst case figures throughout the calculation, is 0.02ppm (or 0.0000002% if you are more comfortable with %).

To get the margin of safety (MoS), divide the IC50 figure by the actual exposure figure (using parts per million, as the units have to be the same!)

7 / 0.02 = 350

Given that an MoS of 100 is deemed acceptable, an MoS of 350, as per my calculation, seems quite comfortable and, unless new data are generated that give different results, this would not appear to be an issue in terms of the overall safety of butylparaben.

I will stress here that EVERY aspect of my calculation has used an extreme worst case figure, either as used by the SCCS, or in my assumption that ALL cosmetics contain butylparaben at the new maximum permitted concentration of 0.19%. To place THIS in context, please note that butylparaben is unlikely ever to be used at this concentration, (partly due to its very limited water solubility, but also because it is almost always used as a component of parabens-based blends rather than as a separate ingredient, thereby being restricted to a small ratio of the total parabens concentration), and a much more typical use concentration would be between 0.02 – 0.075%. Even at the higher end of the typical range, therefore, the actual concentration of butylparaben is 2.5 times less than the figure used in my calculation. This would give an MoS of 875. It is also the case that not all cosmetics contain butylparaben, and a more likely figure would be 50% (at the very most), and this would then give an MoS of 1750 but I think I am happy with 350, as this still greatly exceeds the requirement of the SCCS. There is, of course the possibility of butylparaben concentrating within the body, which would result in higher localised concentrations, but there is absolutely no evidence of this and, until any evidence IS demonstrated, this can be discounted.

In closing, I would like to emphasise an important issue. Firstly, I would draw the reader’s attention to the authors’ use of the words “could” and “may” in their conclusions in the abstract. These are not certainties, and the study should not be quoted with any certainties. Once again, we see a scientific study producing a headline result. Once again, we see people using this as “evidence” of the dangers of parabens (and not even just butylparaben!) without full and proper consideration of the detail. Far too many people don’t read beyond the headlines and, consequently, make decisions that are not fully informed. Reading a study and repeating the result does NOT constitute in-depth research and, certainly, not sufficiently in-depth to take a major decision about the safety of a substance. To fully assess the results of this type of study, a risk assessment is essential. This is what I have done in my calculations above. This has been a common thread in several of my parabens articles, and there is more to come!


This is the third in a series of articles in each of which I will focus on one specific study that has contributed to the current concerns over the use of parabens in cosmetics. If you have not already done so, it may be useful for you to read – Parabens in Perspective: Introductory, Part I and Part II that posted earlier, before reading this one.

The paper I will critique in this article is:

The preservatives ethyl-, propyl- and butylparaben are oestrogenic in an in vivo fish assay

Pedersen K .L., Pedersen S. N., Christiansen L.B., Korsgaard B., Bjerregaard P., – Pharm & Tox 86, 110-113 (2000)

This study consisted of the dosing of fish with the aforementioned parabens and, additionally, the breakdown product common to all parabens – p-hydroxybenzoic acid. The oestrogenic activity was then measured using detection of a specific yolk protein (vitellogenin). Two doses were applied by intraperitoneal injection – on day 0, and day 6 of the study.

The conclusions of this study were that the parabens tested “were shown to be oestrogenic in vivo while the metabolite [p-hydroxybenzoic acid] known to result from hydrolysis did not show any activity in the present test system. Oestrogenicity and toxicity both occurred in a dose interval between 100 and 300mg/kg”.

The authors’ stated aim was “to demonstrate the intrinsic oestrogenicity of the parabens as an inherent molecular property of these compounds . . . . . not to investigate the oestrogenicity in fish per se”. However, despite the real aim of the study, anyone reading the title will make the assumption that parabens are oestrogenic in fish. They will then extrapolate this to humans. What is needed is to look at the study and the results a little more closely.

Butylparaben had a weak oestrogenic effect at a dose of 50mg/kg; this effect was more marked at 100mg/kg

Propylparaben was not tested at 50mg/kg, but a dose of 100mg/kg produced a weak response; a dose of 300mg/kg provoked “a pronounced oestrogenic response”.

Ethylparaben produced only a weak oestrogenic response at 300mg/kg

p-hydroxybenzoic acid demonstrated no oestrogenic activity in this study.

Methylparaben was not tested, as the authors deemed it unlikely to elicit any oestrogenic response.

The problem with figures like 50mg/kg, 100mg/kg, etc is that they are fairly meaningless in isolation. It is far better to place them in some sort of context. The figures here refer to the dose applied to the fish – by injecting them with the dose. As far as I know, fish are not noted for their ability to handle a hypodermic syringe, so they would not normally be exposed to parabens in this manner, and nor would they be exposed to these concentrations. For comparison, a study that determined the environmental concentrations of parabens [Lee et al, J Chromatogr. A 1094 (2005) 122-129] found levels of around 2 parts per trillion (that’s 0.0000000002%!) in the inflow to a sewage treatment plant. The outflow showed less than 0.04 parts per trillion (0.000000000004%!). 50mg/kg = 0.005% – a factor of around 1 trillion times higher than would be expected in the environment after passing through a sewage works! How could fish be exposed to concentrations of similar magnitude to those tested in this study? Clearly, leaving to one side the possibility of a spillage during production or transportation of the parabens, only an accidental spillage of a cosmetic product could do this. But to achieve these levels of parabens in an environmental situation would require the accidental spillage of 250 tonnes of a product containing 0.3% parabens in the volume of an Olympic-sized swimming pool. What this suggests is that parabens are NOT responsible when there are claims that environmental xenoestrogens are changing the sex of fish, despite the title of this study.

Whilst the authors may have intended only to test the sensitivity of their method as an academic exercise, they failed to consider the possible consequences of publication of the results. This is a common issue with academic research – the researchers are often only interested in the results for the results’ sake. They do not feel the need to put the results in any real context.

This study is taken out of context on two levels:

1)      The concentrations of parabens tested are around 1 trillion times higher than typical environmental concentrations

2)      There is rarely good correlation in the toxicological response between test species and humans

The bottom line for this study is that it is relatively meaningless outside of the narrow intentions of the authors. It does not prove that parabens are unsafe for use in cosmetics, but it DOES prove that environmental concentrations of parabens do not contribute any significant threat to the environment in terms of adding to the xenoestrogen burden, with a safety factor of around one trillion!

For completeness, I will report that this research group (based at the University of Southern Denmark, Odense) have also carried out other studies on parabens, using oral administration, which is far more relevant than intraperitoneal injection. It is less simple to interpret the results because the dose was given in both the food and the water, and at different concentrations. However, the results were:


In the water, 225 micrograms/litre (= 0.000225%, or 225 parts per trillion) showed some oestrogenic activity, but 50 micrograms/litre showed none. Bear in mind that only these two concentrations were tested, and there is a large gap between 50 and 225, so the “no effect” concentration may be almost anywhere between these two points but, taking 50 micrograms/litre as the worst case, this is still 1250 times higher than the environmental concentrations found by Lee et al.


This was a separate study, and the concentrations differed slightly from those used for propylparaben, but a dose of 201 micrograms/litre gave a weak oestrogenic response. Therefore, a similar conclusion may be drawn as for propylparaben – environmental concentrations are unlikely to pose an oestrogenic threat to fish. Additionally, and importantly, there was no evidence of bioaccumulation in the fish in either of these oral studies.

So far, in this series of articles, I have focused on studies that have shown parabens in a poor light, in toxicological terms, and attempted to correct errors and misperceptions, and place the results in a realistic context. In this article, I am going to concentrate on sharing some studies that actually have good news for parabens lovers (ie, me and possibly 3 other people on the planet!). I am doing this to demonstrate that not all toxicity studies on the various parabens have come up with results that may worry people. I know that this sounds unlikely, but please bear with me – these studies tend not to get quite the same level of publicity as the negative ones!

Pin It