Center for Research on Ingredient Safety

Michael Holsapple is Director of the Michigan State University (MSU) Center for Research on Ingredient Safety (CRIS) and professor in the Department of Food Science and Human Nutrition. He is also a toxicologist with over 35 years of experience. He recently spent time reviewing the issues surrounding titanium dioxide in the foods we eat.

Titanium dioxide (TiO2) is the naturally occurring oxide of titanium, and is widely used commercially as a white pigment. Commercial applications of TiO2 include use in sunscreens as protection against UVA/UVB rays, and use as a whitener in paint and cosmetics. A food-grade form of TiO2 (e.g., E171) is used as a colorant to enhance and brighten white foods, such as dairy products, candy, frosting, etc. E171 is also used to enhance the flavor of non-white foods such as dried vegetables, nuts, seeds, soups, and even beer and wine.

Two investigations recently analyzed samples of food grade TiO2. In one study, seven batches of E171 were analyzed (Peters et al., Journal of Agricultural and Food Chemistry, 62:6285-6293, 2014). The results showed that all E171 materials had similar distributions with primary particle size in the range of 60-300 nanometers (nm). Depending on the specific analytical method used, only 10-15 percent of the particles in these food-grade materials had sizes below 100 nm -- the typical size of nanoparticles. In another study, five batches of E171 were analyzed (Yang et al., Environmental Science & Technology, 48:6391-6400, 2014), and these results showed that the 5 samples contained 17-35% of nanosized primary particles.

Why are these analyses of E171 important? The safety issues associated with particles less than 100 nm are due to the high surface area to volume ratio, which can make these particles very reactive or catalytic. Additionally, nanosized particles are able to pass through cell membranes in organisms and may interact with biological systems.

My interest in TiO2 was triggered by a recent publication written by a group of French investigators (Bettini et al., Scientific Reports, 7:40373, 2017).

The study was conducted in rats that had been exposed to E171 by direct oral administration for seven days or by their drinking water for 100 days. The investigators claim that one week exposure caused inflammatory changes to the immune cells of the Peyer’s Patches that line the intestines, and in the immune cells in the spleen. Interestingly, these effects were seen with the E171 material; but not with a TiO2 nanoparticle material (NM-105) – even though both materials were detected in the immune cells of the Peyer’s Patches. In contrast, both the E171 and NM-105 caused changes in the levels of some inflammatory cytokines – proteins secreted by cells that interact with specific receptors on other cells, thereby affecting function. Subsequent studies showed that some of these effects were also observed after 100 days of exposure to E171. However, these longer-term studies did not include the comparative control (NM-105), and the effects seen with E171 were generally comparable to the short-term studies, were modest, and were possibly within background levels. Finally, they investigated whether exposure to E171 in the drinking water promoted inflammation and enhanced preneoplastic lesions in the colon. Because of numerous deficiencies in experimental design, their results can only be considered preliminary. In my opinion, no firm conclusions can be drawn until these results can be replicated using a more complete and appropriate study design relevant to TiO2 use in food.

I am not the only one to have reviewed the study by Bettini et al., and concluded that their results, while suggestive, are inconclusive. The French Agency for Food, Environmental and Occupational Health and Safety (ANSES) recent review starts by declaring that “more studies are needed to determine whether the food additive, titanium dioxide (E171), which contains nanoparticles, is safe.” ANSES also emphasized that “the study (by Bettini et al.) is not sufficient, at this stage, to challenge the European Food Safety Authority (EFSA) opinion of last year, which concluded that E171 did not cause cancer”.

Finally, the ANSES review concluded that while they “recognize the scientific value of the publication and results”, they declare that “there are limitations”. Specifically, ANSES concluded that “the study shows moderate inflammation effects but that the biological significance is not established. It calls for more studies using larger groups of animals and different doses”. The opinions expressed by ANSES are particularly noteworthy because one of the co-authors of the Bettini paper is affiliated with ANSES.

As a toxicologist, I believe in the concept of “the dose makes the poison”, and strongly echo the recommendation by ANSES to increase the number of doses of E171 used in future studies. I have also provided additional suggestions below that should be considered before further research aimed at investigating the results by Bettini et al. is initiated.

The study by Bettini and co-workers administered the E171 either by direct oral administration or in the drinking water. These routes of exposure are difficult to reconcile in light of the facts that E171 is a food-grade material, and the Joint FAO/WHO Expert Committee on Food Additives (JECFA) reported in their Monograph series that TiO2 is insoluble in water. I recognize that the results from the Bettini study showed absorption of E171, but that exposure in their studies would be difficult to assess. A much more appropriate approach for exposure assessment would be to integrate multiple doses of E171 into rodent chow and measure the food intake. Regarding studies to determine the effects of E171 on preneoplastic lesion formation, animals need to be exposed to multiple doses of E171 with and without dimethylhydrazine (DMH) pretreatment. The studies by Bettini et al. were all done after pretreatment with DMH, a potent genotoxic carcinogen, with highly variable effects on intestinal preneoplastic lesions. The findings by Bettini et al. on preneoplastic lesions could well be entirely due to the DMH, without an effect by E171.

Finally, as indicated in the ANSES announcement, “EFSA concluded that E171 was safe but could not set an Acceptable Daily Intake (ADI).” In their scientific opinion about E171 adopted in June, 2016, EFSA “considered that, on the database currently available and the considerations on the absorption of TiO2, the margins of safety (MoS) calculated from the No Observable Adverse Effect Level (NOAEL) of 2,250 mg TiO2/kg body weight (bw) per day identified in the toxicological data available and exposure data obtained from the reported use/analytical levels of TiO2 (E171) would not be of concern”.