3.1.1. Conclusions Regarding Aromatized Foods and Their Potential Alkenylbenzene Contents

Even if the amount of a specific alkenylbenzene appears to be low in a certain food category, there is a risk of dose addition depending on the dietary habits of consumers if many alkenylbenzene-containing food products are frequently ingested in a short period of time. Some alkenylbenzenes, such as elemicin and apiol, have not yet been fully assessed for their hepatotoxic and genotoxic potential and have not been sufficiently monitored in the potentially relevant food products. Since, for all existing food matrices, specific extraction, separation, and detection procedures for each of the alkenylbenzenes would have to be elaborated, standardized, and validated, this appears to be a difficult endeavor. However, given the high hepatotoxic potential of other alkenylbenzenes, there is an undoubted necessity to analytically determine all the possible alkenylbenzenes in raw and prepared food products and food preparations. Analytical techniques have advanced significantly in specificity and sensitivity in recent years, and the first promising approaches have been made to quantitate some alkenylbenzenes in foods and beverages [6,49,149–152]. Further ambitious experimental monitoring activities could build on those approaches.

#### 3.1.2. Issues Regarding Currently Available Occurrence Data for Alkenylbenzenes

The herbs and spices that we discussed above contain variable amounts of alkenylbenzenes. Depending on the family, the genus, the species and their varieties, the geographical origin (e.g., soil, humidity, solar irradiance, etc.), and the plant parts analyzed (fruits, seeds, flowers, leaves, stems, roots, in different maturation states at harvest, etc.) very different contents of alkenylbenzenes and their metabolites may prevail in foods. In addition, depending on the procedures of post-harvest treatment, sample preparation (extraction methods and duration, solvents used, etc.), and analytical methods utilized (e.g., GC, LC, etc., which are described in detail elsewhere [153]), the reliability of quantitative data may vary considerably. These circumstances call for the elaboration of standardized analytical

procedures to enable reliable quantification of alkenylbenzenes in crude spice extracts, essential oils and their oleoresins, and finished aromatized foods. Ideally, such methods would become internationally harmonized. These efforts have to be complemented with efforts to measure all known alkenylbenzenes in a representative set of well-defined finished food products that belong to food categories naturally containing alkenylbenzenes, and in those that intentionally become aromatized.

#### *3.2. Consumption of Alkenylbenzene-Containing Foodstuffs*

Besides information on occurrence, data on consumption also play an important role for the exposure assessment. In the following section, we will discuss some aspects leading to uncertainties regarding the currently available consumption data on alkenylbenzenecontaining foods

#### 3.2.1. Limited Availability of Data Regarding Consumption

Besides data regarding some few individual foods (e.g., fennel-based teas or plant food supplements) [5,12,154], the availability of consumption data—especially of current information—in the context of alkenylbenzenes is rather limited [98,109,111]. This leads to uncertainties regarding the current exposure of humans to alkenylbenzenes via the consumption of food.

Among other things, the availability of current consumption data on naturally alkenylbenzene-containing foods, such as herbs, spices, and flavored foods (e.g., baked goods or beverages) is mandatory to perform a reliable risk assessment for these substances.

Currently available consumption data regarding alkenylbenzene-containing food are not up to date, since they were largely captured and evaluated approximately 20 years ago [98,109,111]. In this context, it is important to note that the consumer behavior may have changed over the last two decades in different regions of the world [155,156]. Moreover, consumption habits may also vary between different countries in the EU as well as worldwide [157]. This may lead to differences regarding exposure of consumers to alkenylbenzenes via distinct locally favored foods or mainly regionally consumed food products, such as basil-containing pesto sauce, herbal Indonesian beverages, or herb-based Chinese medicinal teas [6,116,130].

#### 3.2.2. Lack of Biomarker Prevents Exposure Estimation

Until now, no biomarker has been identified that reliably reflects the external exposure of humans to alkenylbenzenes via food consumption under realistic conditions. However, the analytical quantification of a biomarker is a possible alternative for the estimation of the external exposure. Two types of biomarkers are commonly used for other compounds in food, i.e., mercapturic acids (MAs) in urine samples [158] or protein adducts (usually determined in hemoglobin) [159]. Amounts of a urinary MA excreted within 24 h may be used for the estimation of the daily exposure to the parent compound, if the compound ratio is known, which is converted into the MA (reverse dosimetry) [160]. Recently, *N*-acetyl-*S*-[3- - (4-methoxyphenyl)allyl]-L-cysteine was described as the main MA, which is formed from estragole and its structural congener *trans*-anethole [52]. A single controlled exposure to fennel tea (*n* = 12) resulted in the excretion of *N*-acetyl-*S*-[3- -(4-methoxyphenyl)allyl]-Lcysteine in the urine samples within 24 h. The interindividual variation of total *N*-acetyl-*S*-[3- -(4-methoxyphenyl)allyl]-L-cysteine excreted (93–1076 ng) reflected the complexity of estragole/*trans*-anethole metabolism involving different enzyme families, i.e., CYP, SULT and alcohol dehydrogenases (ADH). It hinders an accurate estimation of the external exposure for individuals from the *N*-acetyl-*S*-[3- -(4-methoxyphenyl)allyl]-L-cysteine determined in 24 h urine samples. In addition, the biomarker is not specific; one cannot distinguish between *N*-acetyl-*S*-[3- -(4-methoxyphenyl)allyl]-L-cysteine formed from estragole or *trans*-anethole. These considerations also hold true for another biomarker of exposure to these compounds, the hemoglobin adduct *N*-(isoestragole-3-yl)-valine (IES-Val) [76]. Hemoglobin adducts are considered as biomarkers of medium-term exposure, because

hemoglobin can accumulate adducts in its lifetime of ~120 d. IES-Val was shown to increase steadily when fennel tea was consumed over 28 d. However, the complexity of metabolism and the missing specificity for one compound also hinders the exposure estimation from IES-Val.

In summary, the conjugate *N*-acetyl-*S*-[3- -(4-methoxyphenyl)allyl]-L-cysteine (24 h urine) and the adduct IES-Val (hemoglobin) can only be used as biomarkers for the internal exposure to the ultimate carcinogen 1- -sulfoxyestragole or to 3- -sulfoxyisoestragole, the reactive sulfate ester metabolites of estragole and *trans*-anethole, respectively [52,76]. Similar studies have not yet been published for the other alkenylbenzenes, such as safrole or methyleugenol. The biomarkers equivalent to those described for estragole/*trans*-anethole may offer a higher specificity. However, the metabolism of safrole and methyleugenol may be as complex as that of estragole, which renders an exposure assessment for individuals at least difficult.

Together, the aforementioned data show that more research is needed regarding the exposure of humans to alkenylbenzenes via the consumption of food, especially in the context of real-life influences. In this context, the development of specific biomarkers and reliable measurement strategies is also of high importance.
