In “The difference between hazard and risk: the dose range prevalent in toxicological studies vs real life fragrance exposure,” co-authors Kaushal Joshi, PhD, DABT, principal scientist at RIFM, and RIFM Senior Associate Scientist Arianna Bartlett, PhD, along with members of the Expert Panel for Fragrance Safety, I. Glenn Sipes, PhD, and Wolfgang Dekant, PhD, and Reproductive Toxicology Adjunct Group member Gerhard Eisenbrand, PhD, examine how toxicological dose levels compare with real-world consumer exposure to fragrance ingredients.
Using benzaldehyde and p-cymene as case studies, the researchers show that the levels associated with adverse effects in animal studies are far removed from the amounts consumers would encounter through typical fragrance use. In this Q&A with CosmeticsDesign USA’s Cassandra Stern, Dr. Joshi discusses why dose-context is central to safety science, while Dr. Bartlett explains how aggregate exposure modeling helps modernize fragrance risk assessment.
CDU: Your paper addresses a distinction that often arises in ingredient safety debates: hazard versus risk. Why was it important to address that now?
Kaushal Joshi, PhD, DABT: Because that distinction is essential to understanding what toxicology data actually mean.
Hazard describes a substance’s inherent potential to cause an adverse effect under certain conditions, often at high experimental doses. Risk asks the more relevant real-world question: what is the likelihood of harm at the levels to which people are exposed?
In safety assessment, both matter, but they are not interchangeable. Toxicological studies are designed to identify hazards, which means they often use dose levels far above normal human exposure. That is scientifically appropriate for hazard identification, but when those findings are discussed without exposure context, they can easily be misunderstood.
This paper was written to help put that into perspective. While exposure is key to many industries in making safety assessments, it is particularly impactful in fragrance safety because people’s exposure to fragrance is both extremely low and well characterized.
CDU: Why did you select benzaldehyde and p-cymene as the examples for this paper?
Dr. Joshi: We wanted case studies that were scientifically robust but also easy to explain in practical terms.
Benzaldehyde and p-cymene are relevant fragrance ingredients with published toxicity data, which allowed us to identify conservative no-observed-adverse-effect levels (NOAELs) and adverse-effect levels from in vivo studies. That gave us a solid toxicological basis for the comparison.
At the same time, both ingredients also occur naturally in foods, which made it possible to translate those dose comparisons into everyday analogies that both toxicologists and non-toxicologists can understand. That combination made them especially useful for conveying the much larger point of the paper: a hazard finding must be viewed in the context of actual human exposure.
CDU: One of the most headline-grabbing parts of the paper is the comparison between toxicological dose levels and real-life use — including almonds, raspberries, and perfume sprays. What were you trying to show with those examples?
Dr. Joshi: We were trying to make the magnitude of the dose gap visible in the existing studies and actual human exposures in real life.
Toxicologists are used to thinking in terms such as milligrams per kilogram body weight per day, but those values are abstract to most people. So, we translated them into easier-to-picture examples.
For benzaldehyde, our calculations showed that a person would need to consume about 83,000 almonds per day for life to reach the safe dose determined by studies. For p-cymene, the comparable figure was about 153,778 raspberries per day for life to reach the safe dose.
We then applied this analogy to food and fine fragrance use. Based on the assumptions used in the analysis, a person would need to apply about 138,330 sprays per day for life of a perfume containing benzaldehyde, or about 18,870 sprays per day for life of a perfume containing p-cymene, to approach the safe dose levels.
Those are clearly unrealistic scenarios, and that is exactly why they are useful. They show how far real-life exposure is from the doses used to identify hazards in toxicological studies.
CDU: The paper also underscores that RIFM has not conducted new animal toxicity studies for its human health endpoints for more than a decade. How has the scientific approach to fragrance safety evolved?
Arianna Bartlett, PhD: A major shift has been the growing ability to pair existing toxicological data with much more refined and realistic exposure science.
Historically, systemic toxicity assessments often depended on animal studies generated to meet regulatory requirements. Today, RIFM relies on existing data from robust sources and combines those data with advanced exposure modeling to assess real-world risk. That has allowed safety assessment to evolve without RIFM needing to conduct new animal toxicity studies.
A key part of that progress has been the development and continued refinement of aggregate exposure tools, including the Creme RIFM Aggregate Exposure Model. Compared to other industries, fragrance exposure is extremely well characterized: for example, we have robust data on consumer use and practices from thousands of people, allowing us to get a clear picture of exposure.
These approaches let us look beyond a single product or a single route of exposure and instead calculate how much of a fragrance material a person may encounter across many product types over time.
That is important because risk assessment becomes much more meaningful when it reflects how consumers actually use products.
CDU: Can you walk us through how the Creme RIFM Aggregate Exposure Model shaped the paper’s conclusions?
Dr. Bartlett: The model was central to the paper because it provided the real-life exposure side of the hazard-and -risk equation.
The Creme RIFM Aggregate Exposure Model uses probabilistic methods and large datasets on consumer behavior from North America, Europe, and Asia. It calculates aggregate exposure across multiple routes — oral, dermal, and inhalation — and includes cosmetic, personal care, household, and air care products.
This analysis allowed us to compare toxicological effect levels with realistic consumer exposure estimates, including high-end users represented by the 95th percentile. For benzaldehyde, the total chronic aggregate exposure from all consumer products was estimated at 0.00053 mg/kg body weight/day. For p-cymene, it was 0.00061 mg/kg body weight/day.
When translated into annual volumes of the neat ingredients, those exposures are extremely small — about 0.011 mL per year for benzaldehyde and 0.016 mL per year for p-cymene. That is roughly 0.2 drops and 0.3 drops, respectively, over the course of an entire year.
Those findings help illustrate just how low real-world exposure is, even for the most loyal users of a particular fragrance.
CDU: Your paper describes those exposure assumptions as conservative. What does that mean in practice?
Dr. Bartlett: It means the calculations were intentionally designed to overestimate exposure rather than underestimate it.
For example, we assumed 100% absorption, even though the paper notes that actual dermal absorption is lower for both benzaldehyde and p-cymene. We also assumed no evaporation, which, again, is conservative because it does not reflect what would happen under normal product-use conditions.
Beyond that, we used high-end values in several parts of the analysis, including the highest concentrations observed in food examples and conservative toxicological effect levels from robust studies. The human-equivalent dose calculations also used a body weight of 60 kg, which is below the average adult body weight in countries such as the US.
So the overall message does not depend on optimistic assumptions. Even though we used conservative estimates, the margins between real-life exposure and toxicological effect levels remain very large.
CDU: The paper’s margin of safety values are also striking. What do those numbers tell us?
Dr. Joshi: They tell us that estimated consumer exposure is far below the levels considered safe based on the available toxicology data.
The margin of safety, or MOS, compares a safe dose (or, for toxicologists, the no-observed-adverse-effect level) with the estimated human exposure. This simple calculation indicates whether there is a concern for human safety: if the MOS is greater than 100, the fragrance is generally considered safe.
In this paper, the MOS values were dramatically above that benchmark. For benzaldehyde, the MOS was 377,358. For p-cymene, it was 81,967.
Those are very large safety margins. They reinforce the same conclusion reached through the food and perfume analogies: when realistic exposure is taken into account, the risk associated with normal fragrance use is extremely low.
CDU: In beauty and personal care, ingredient discussions are often driven by hazard-based messaging, including in some “clean beauty” conversations. What do you hope this paper adds to that broader dialogue?
Dr. Joshi: What we hope to add is a scientific context.
There is understandable consumer interest in knowing more about ingredients and their safety, and transparency in those discussions is important. But hazard information alone does not tell consumers whether a material poses a real-world risk under normal conditions of use.
Our paper shows why dose and exposure matter. A substance may produce an adverse effect at high doses in a study designed to identify hazards, but that does not mean the same outcome is relevant to real-life consumer use. Risk assessment requires both sides of the equation: hazard and exposure.
I would stop short of saying this paper is about any one market movement. It is about reinforcing a core toxicological principle that applies broadly across safety science. The more those conversations are grounded in exposure-based risk assessment, the more useful and accurate they become.
CDU: What is the simplest takeaway you would want beauty industry stakeholders to understand from this study?
Dr. Bartlett: That realistic exposure assessment changes the conversation.
When we evaluate fragrance safety using modern aggregate exposure tools, the amount consumers actually encounter is extremely low — even for loyal or high-end users across multiple product categories. That is why exposure science is such an important part of modern safety assessment.
The study is a reminder that a meaningful safety evaluation cannot stop at hazard identification. It must consider how fragrance-producing ingredients are used in real-world contexts.
CDU: And what is the simplest takeaway for consumers?
Dr. Joshi: That dose matters.
A hazard finding does not automatically translate into consumer risk. When toxicological data are interpreted alongside realistic human exposure, the evidence supports the conclusion that the risk of harm from normal fragrance use is minimal, including for high-end users.
That is the central point of the paper, and it is the point we most wanted to communicate clearly.
CDU: Are there any next steps to this research?
Dr. Joshi: Yes. As next steps, we are evaluating additional fragrance ingredients with existing datasets to determine whether similar analogies can be established.
We are also examining how effects observed in high‑dose animal studies may differ from human biology, with a focus on determining the true biological relevance of those findings. If data gaps remain, we may explore targeted in vitro assays to further support our assessment and strengthen the overall weight of evidence.
