The cliff-edge of toxicological concern: highlighting the potential issues of an over-reliance on “less-than-lifetime” thresholds

Abstract

The Threshold of Toxicological Concern (TTC) is a very well-established concept in applied toxicology, and has become a key tool for the pragmatic human health risk assessment of data-poor chemicals. Within the pharmaceutical sector, regulatory guidance on genotoxins defaults to a TTC of 1.5 μg/day equating to a maximum lifetime cancer risk of 1 in 100,000. Higher doses for drug products where exposures are intermittent or otherwise “less-than-lifetime” (LTL) are also considered tolerable. This also allows substance-specific lifetime Acceptable Intakes (AIs) for known genotoxic carcinogens to be scaled up for shorter durations.

The default TTCs for assessing LTL exposures build in conservatism such that there is deviation from strict linearity. However, close to the boundaries between LTL categories there can be such a difference in the default tolerable intakes that a health risk assessment can yield conflicting results. We have presented a theoretical case study based on our recent work that illustrates this apparent “cliff-edge.” The total acceptable cumulative dose over a 56-day treatment is – in absolute terms – one third of that allowed over 28 days, despite the maximum cancer risk of the longer exposure being an order of magnitude higher. Our analysis suggests the need for careful consideration of what might represent tolerable exposures in the region of the category limits, rather than simply adopting the hardline default. Where a potential patient exposure is found to be above a default value, there is real value in refining the cancer risk estimates using the Lifetime Cumulative Dose approach.

Introduction

The Threshold of Toxicological Concern (TTC) is a very well-established concept in applied toxicology and human health risk assessment. It is based on the reasoning that the likely toxicity of untested (organic) compounds can be conservatively predicted from data on the universe of tested chemicals. In particular, the aim is to define a low level of exposure to an untested chemical below which there is “no significant risk” to human health.¹

The TTC has its origins in the Threshold of Regulation, introduced by the US Food and Drug Administration (FDA) for the pragmatic health risk assessment of “unintentional” food additives (arising as a result of migration from food-contact materials)^2,3 and has been widely used ever since. TTC values have been proposed following the statistical analysis of oral toxicity and carcinogenicity data on large numbers of compounds. For those that were not genotoxic carcinogens, a large dataset of No-Observed-Adverse-Effect Levels (NOAELs) was used to establish TTCs corresponding to the three structure-based classes established by Cramer et al.^4,5

For genotoxic carcinogens, which are assumed to have no threshold, the focus was on the lifetime daily doses estimated to be associated with “insignificant” (“negligible”) increases in cancer risk, defined initially as a maximum of one extra case in 1,000,000 exposed people. For chemicals with structural alerts for genotoxicity, a value of 0.15 μg/person/day was established by linear extrapolation from values indicating 50% tumour incidence (TD₅₀),⁶ which was initially employed for assessing potentially mutagenic contaminants in food, drinking water and the air.

Within the pharmaceutical sector, a TTC of 1.5 μg/day is considered allowable for most unavoidable genotoxic impurities in drug products for lifetime use, with the cancer risk (no more than 1 in 100,000) justified because of the health benefit these products provide. This figure is said to be generally applicable to any exposure route.⁷

Less-than-lifetime

The default TTC values discussed above are intended for risk assessment of daily, lifetime exposures. For cancer risk assessment, the use of linear extrapolation from TD₅₀ values to derive a tolerable level is generally considered to be very conservative,⁸ and particularly so when applied to exposures that are in some cases considerably shorter than lifetime or which are intermittent. These scenarios are together referred to as “less-than-lifetime” (LTL).

One approach to consideration of LTL exposures is to apply Haber’s Law, which originally stated that the acute toxicity of a gas was directly related to its concentration multiplied by the duration of exposure.⁹ This is thought to apply in a rough and ready way to other toxicological endpoints, including cancer. This theory has been used to justify that higher daily doses of a genotoxic compound could be tolerated without increasing the excess cancer risk, if the exposure duration was appropriately reduced^10–12 However, direct application of Haber’s Law is not without its drawbacks. Indeed, Felter et al. suggested that Haber himself was unlikely to have intended his work to apply to extrapolations of this nature.⁸

For drug products with LTL exposures, guidance from pharmaceutical regulators has enshrined the “lifetime cumulative dose” (LCD) approach. Here, the total, cumulative dose of a compound for which there is no threshold to its toxicity can be distributed over the total number of exposure days in a default 70-year lifetime. If applied strictly, this would imply that a dose of 1.5 μg/day for life or 38,325 μg on a single day would correspond to the same theoretical cancer risk of no more than 1 in 100,000. Further illustrative values are presented in Table 1.

Table 1.

Cumulative doses corresponding to a cancer risk <1 in 100,000 arising from a strict application of the lifetime cumulative dose (LCD) approach.

Daily dose	Duration of treatment	Treatment days	Cancer risk
1.5 μg	70 years (lifetime)	25,550
10.5 μg	10 years	3,650
105 μg	1 year	365	< 1 in 100,000
1,270 μg	1 month	30
38,325 μg	1 day	1

Felter et al. noted in their comprehensive review of risk assessment for LTL exposures that Haber’s Law may “within limits … be appropriate when applying the TTC” to obtain higher “adjusted” TTCs.⁸ In practice, however, medicines regulators have been wary of accepting the applicability of Haber’s Law, considering that the protective mechanisms that operate efficiently at lower chronic doses might not be as effective when this same cumulative amount is being experienced over a much shorter time.¹³ Thus, they have proposed an additional layer of health protection to derive LTL TTCs for three shorter-duration periods. Yet more conservatism applies for a shortest duration (up to 30 days) where the tolerable cancer risk is no more than 1 in 1,000,000, because it might be used to assess the risks posed to healthy volunteers and/or before a therapeutic benefit has been established. Table 2 illustrates the daily and cumulative intakes (of an individual mutagen) deemed tolerable by ICH for the four exposure durations.⁷

Table 2.

Acceptable intakes for individual mutagens by duration (ICH, 2023a).

Duration of treatment	<1 month	>1–12 months	>1–10 years	>10 years
Cancer risk	< 1 in 1,000,000	< 1 in 100,000
Daily intake (μg/day)	120	20	10	1.5
Cumulative total intake (μg)	3,600	7,300	36,500	38,250

Scaling for short-term exposures

The regulatory acceptance that proportionally-higher short-term exposures of tested compounds can be tolerated because the same risk levels are involved implies that the ratio between the lifetime and LTL TTC values can be used to scale up lifelong cancer AIs derived from experimental data, and this is enshrined within ICH M7.⁷ However, this highlights three possible issues.

First, while higher doses over a shorter time frame may not – using the LTL logic – present an unacceptable cancer risk, very high single doses may introduce acute systemic toxicity concerns.

Second, the higher acute/subacute doses might well exceed the capacity of DNA-repair mechanisms, which would certainly lead to a loss of linearity in cancer risk.^14,15 Indeed, this was a key consideration in EMA not accepting the use of scaled LTL thresholds for nitrosamines.¹⁵

Third, the LTL thresholds established for cancer risk assessment are themselves not strictly linear in numerical terms, being chosen to represent a conservative level of risk for each of the four exposure durations. Thus, considerably different cumulative doses of a substance would be considered tolerable depending on which side of an arbitrary cut-off the duration falls. In the theoretical case study below, we illustrate how this cliff-edge between duration categories can introduce conflicting results into a quantitative cancer risk assessment.

Case study

Consider the theoretical cancer risk to patients of epichlorohydrin, a potential drug product contaminant. Epichlorohydrin is known to be mutagenic to bacteria¹⁶ and mammalian cells,¹⁷ and to cause chromosome aberrations in the lymphocytes of exposed workers.¹⁸ An IARC Working Group has placed it in Group 2A as “probably carcinogenic to humans”.¹⁹ A sector-specific opinion was provided by ICH which, in their Addendum to M7(R2), established a lifetime acceptable intake (AI) of 3 μg/day for any exposure route²⁰ on the basis of the TD₅₀ reported in a 2-year oral cancer bioassay in rats.²¹ Thus epichlorohydrin, if present in a pharmaceutical for chronic treatment below this AI, would not be expected to pose an unacceptable cancer risk. In this case study we consider two potential LTL exposure scenarios.

First, if the exposure lasts for 28 days in a lifetime, this falls clearly within the shortest-duration category of the ICH M7 LTL values. The cancer risk assessment of a data-poor compound with this treatment duration would have relied on the TTC value of 120 μg/day for exposures up to 30 days, associated with a cancer risk of no more than 1 in 1,000,000. This daily exposure figure is 80-fold higher than the lifetime TTC of 1.5 μg/day. When there is an established AI, the ICH M7 guideline indicates that an intake 80-fold higher (i.e. 240 μg/day for epichlorohydrin) would not present an intolerable cancer risk to patients across this treatment period. As previously noted, a thorough risk assessment would also need to provide reassurance that an acute dose of this magnitude was not likely to induce any acute toxic responses.

In a second scenario, we consider how the cancer risk would differ if the maximum exposure was doubled to 56 days. At this point, the total number of treatment days moves us into the next LTL category, for 1–12 months of exposure. The appropriate TTC figure, therefore, would be 20 μg/day, a value 13.3-fold greater than the lifetime TTC. Thus, for epichlorohydrin, the lifetime cancer AI can be increased from 3 μg/day to 40 μg/day without increasing the cancer risk above the 1 in 100,000 level deemed tolerable by the regulators.

It must be remembered that the LTL values are based on the principle of total cumulative dose presenting a proportional cancer risk. It can be seen from Table 3 that for the first exposure scenario, 6,720 μg would be considered to be the acceptably health precautionary estimate of a total epichlorohydrin dose associated with a cancer risk of no more than 1 in 1,000,000. As we move to the next LTL category in the second scenario, the total acceptable epichlorohydrin dose is reduced to 2,240 μg, just one third of that applicable to the shorter treatment. Moreover, the maximum cancer risk associated with this TTC figure is potentially an order of magnitude higher.

Table 3.

Tolerable cumulative doses of epichlorohydrin for 28- and 56-day treatment periods, based on ICH M7 LTL TTC values.

Total treatment duration	Applicable LTL TTC	Ratio (LTL:Lifetime)	Adjusted cancer AI	Cumulative dose
28 days	120 μg/day	80	240 μg/day	6,720 μg
56 days	20 μg/day	13.3	40 μg/day	2,240 μg

It seems obvious that, if 6,720 μg of epichlorohydrin can be tolerated over 28 days, the same cumulative dose must be tolerable over a longer period of 56 days, if only because of a reduced risk of exceeding DNA repair capacity. Therefore, the practical value of the 40 μg epichlorohydrin/day figure for 56 days (a cumulative dose of 2,240 μg) appears doubtful. This cliff-edge between the two exposure categories introduces a fallacy that risks undermining the scientific logic underpinning the LTL approach.

A similar challenge can arise when trying to set a conservative AI for a mutagen based on the LTL TTCs. For a drug used for up to 30 days, an AI of 120 μg/day can be confidently set using the applicable LTL TTC. However, for a treatment lasting twice as long, the ICH M7 guidance indicates that the AI should be the LTL TTC of 20 μg/day. Clearly though, if 3,600 μg represents a tolerable total intake for any exposure period up to 30 days, the same cumulative dose must also be tolerable over longer durations. Thus, there seems to be a strong case for defending an AI of 60 μg/day (3,600 μg/60 days), rather than 20 μg/day, for an untested mutagen with a 60-day treatment. This illustrates that careful thought is needed when setting AIs from LTL TTC figures.

Discussion

The TTC forms the cornerstone of toxicological risk assessment for data-poor compounds. The adjustment of cancer TTC values for less-than-lifetime exposures is a well-established scientific principle, and is endorsed by regulators in the pharmaceutical and medical device sectors.

Under this approach, higher amounts are allowed for shorter-duration exposures to unavoidable impurities in drug products. Daily doses of 20 μg and 10 μg of most genotoxic carcinogens (for which there is no threshold for their toxicity) for up to 1 year and 10 years, respectively, are considered to pose a cancer risk of no more than 1 in 100,000, the same as a 70-year lifetime exposure to 1.5 μg/day. Even higher daily intakes are tolerated for up to 30 days–120 μg/day is considered to present no more than a 1 in 1,000,000 cancer risk.

However, logical fallacies are uncovered when moving between duration categories. The LTL TTC values deviate from the linear relationship implied by Haber’s Law by the introduction of additional conservatism to increase the level of health protection. This is clearest in the 1–12 month category where the total cumulative dose of 7,300 μg is around 5-fold lower than the lifetime total of 36,500 μg that would mathematically represent the same 1 in 100,000 risk level.

We have presented a theoretical case study that illustrates this apparent cliff-edge, and how it can produce inconsistent results. The acceptable cumulative dose specified by the ICH M7 guideline for a 56-day treatment is – in absolute terms – one third of the total exposure indicated for 28 days, despite the maximum cancer risk levels assigned to the longer exposure being an order of magnitude higher.

For the most robust health risk assessment it is necessary to carefully examine the total cumulative dose alongside the thresholds and benchmarks that are present in regulatory guidance for a given exposure duration. Likewise, if using the default LTL TTCs in setting an AI for a specific treatment period that is close to an M7 duration category cut-off, it is worth evaluating the values derived using TTCs from both sides of the boundary to ensure any subsequent cancer risk assessment is not unnecessarily conservative.

Other issues

Strict adherence to regulatory guidance on cancer risk assessment poses a danger of introducing additional issues. For example, the majority of the data underpinning the derivation of the TTC values was obtained by the oral route. However, there can be significant differences in absorption between clinical routes of exposure and thus cancer risks could be underestimated if route-to-route differences are not taken into account. The authors hope to address some aspects of this in a future publication.

Conclusion

Cancer risk assessment is typically carried out using data from chronic bioassays covering the full lifespan of rodents. In the absence of such data, TTC values provide an estimate of a low daily dose that will be without unacceptable risk over a 70-year human life. Application of the LCD approach allows the derivation of adjusted TTCs for LTL exposures, but can be overly conservative and introduce illogical cliff-edges between duration categories.

Careful consideration needs to be given in cancer risk assessment relying on adjusted TTCs or AIs established in regulatory guidance. Of course, if the highly conservative defaults suggest a lack of health concern, there can be a high degree of confidence in the conclusions. However, where a potential patient exposure is found to be higher than a default value, there is considerable value in revisiting the calculations using the LCD approach to obtain a more refined estimate of cancer risk.

Author contributions

All authors contributed equally to the preparation of this manuscript.

Funding

No funding was received for this work.

Conflict of interest statement. The authors declared no potential conflicts of interest with respect to the authorship and/or publication of this article.