Caffeine, a Central Nervous System (CNS) stimulant of the methylxanthine class, is naturally-derived from the leaves, seeds, and fruits of many plants, including tea leaves, cocoa beans, coffee beans, guarana, and kola nuts. Caffeine has been “generally recognized as safe” or GRAS since 1959 at a limit of 0.02 percent (200 ppm) in cola-type beverages (Institute of Medicine, 2014), but the serving level has increased as well as its direct addition to a wide variety of food products outside of colas, including ice cream, shots, gum, and energy bars. To this end, total daily consumption and exposure to caffeine has slowly increased globally.

Mean caffeine intake in the United States ranges between 25 mg among two- to 11-year-old children to more than 200 mg in older adults. In United States adults 50–59 years of age, the 90th percentile intake of consumption was estimated at 450 mg/day in 2014 (Institute of Medicine, 2014). Caffeine is typically consumed in the form of coffee. It is often promoted to combat fatigue, promote energy levels (Evans et al., 2022), improve mental alertness (Chen et al., 2020), and suppress appetite (Sirotkin and Kolesárová, 2021). Despite these purported benefits, caffeine intake has also been associated with adverse effects among certain consumers sensitive to the stimulant. These effects can be explained in part due to differences in caffeine metabolism, resulting in subpopulations of individuals being classified as either slow or fast caffeine metabolizers (Cornelis et al., 2016). Differentiation into these subgroups occurs based on polymorphism of the expression of CYP1A2, the predominant enzyme involved in caffeine metabolism (Chen et al., 2015). Differences in expression of CYP1A2 determine how quickly caffeine will be metabolized and eliminated from the body, ultimately influencing potential caffeine-induced side effects when caffeine is present in excess in the body. While fast metabolizers often report few side effects following caffeine use, slow caffeine metabolizers are more likely to suffer from caffeine-induced tachycardia and insomnia due to excessive caffeine intake. Similarly, certain individuals are more likely to experience caffeine-induced anxiety, which is also based on a genetic predisposition. Studies have suggested that these effects are mediated at the level of the adenosine receptor, whereby caffeine can produce anxiogenic effects by antagonizing adenosine at A1 and A2A receptors (Alsene et al., 2003; Childs et al., 2008). Evidence of single nucleotide polymorphisms (SNPs) in genes for adenosine receptors may affect these responses to caffeine, as they have been associated with greater increases in anxiety following caffeine intake. Moreover, some studies suggest more serious health risks among slow caffeine metabolizers, such as an increased risk of hypertension and acute myocardial infarction with increased coffee consumption (Cornelis et al., 2006; Palatini et al., 2009). As such, an alternative stimulant with similar benefits and less toxicity is warranted.

Paraxanthine (1,7-dimethylxanthine) is a metabolite of caffeine that can be found naturally in small amounts in some foods including green coffee beans, cotyledonary leaves of Coffea arabica, Theobroma cacao fruits, roasted coffee beans, citrus flowers and Sicilian orange flower honey (Waller et al., 1983; Kappeler and Baumann, 1985; Kretschmar and Baumann, 1999; Zheng X.-Q. et al., 2004; Arnaud, 2011; Bothiraj et al., 2020; Garg et al., 2021). However, it is predominantly found as the major metabolite of caffeine, accounting for approximately 70%–72% of ingested caffeine (Stavric, 1988). Paraxanthine has been shown to have several of the same nootropic benefits as caffeine, including improved cognition, short-term memory, sustained attention in heathy adults (Xing et al., 2021; Yoo et al., 2021) and beneficial cognitive effects in animal models of Parkinson’s disease (Xu et al., 2010; Xing et al., 2021; Yoo et al., 2021). Moreover, the beneficial impacts on physical endurance and sports performance that have been observed with caffeine use (Guest et al., 2021) have also been suggested for paraxanthine (Jäger et al., 2022). Mechanistically, some of these effects are mediated by a similar mechanism for both caffeine and paraxanthine, which act as antagonists of adenosine receptors A1 and A2a (Snyder et al., 1981; Logan and Carney, 1984; Ferré et al., 1991; Orrú et al., 2013). Despite these similarities in mechanisms, it has been reported that there are potential differences in adenosine receptor binding affinity, as paraxanthine exhibited higher binding potencies for adenosine A1 and A2A receptors in equine forebrain (Chou and Vickroy, 2003) and a stronger locomotor activating effect in rats (Orrú et al., 2013) relative to caffeine. Furthermore, paraxanthine, but not other methylxanthines (caffeine, theobromine, theophylline) has been shown to potentiate nitric oxide neurotransmission; which has been associated to increased blood flow and subsequent changes in aerobic exercise performance and cardiovascular health (Ferré et al., 2013). Since clinical studies administering caffeine and paraxanthine to healthy subjects demonstrated that caffeine produces a greater increase in diastolic blood pressure relative to paraxanthine (Benowitz et al., 1995), it is plausible that the relatively lower blood pressure increase following paraxanthine intake compared to caffeine could be due to paraxanthine’s effects on nitric oxide production. Paraxanthine is also unique from other methylxanthine stimulants in that it seems to possess anxiolytic activity (Costentin et al., 2009; Okuro et al., 2010). Although human toxicity studies on paraxanthine are scarce, the few studies that have been conducted in humans demonstrate a lack of adverse events (Xing et al., 2021; Yoo et al., 2021). In fact, an improved safety profile for paraxanthine in comparison to caffeine has been suggested based on evidence from preclinical in vitro and in vivo studies (Weinstein et al., 1975; Renner, 1982; Nakatsuka et al., 1983; York et al., 1986; Stavric, 1988; Institute of Medicine (U.S.), 2001; Gressner et al., 2009; Monteiro et al., 2016). These studies include reports that paraxanthine has low abuse liability and does not mediate the abuse-related effects of caffeine (Lazenka et al., 2015); in addition to being less clastogenic relative to caffeine (Weinstein et al., 1975), it causes less DNA damage than caffeine (Renner, 1982); exhibits less cell toxicity compared to caffeine (Gressner et al., 2009); and results in a lower incidence of malformed fetuses and resorption rates following in utero exposure (Nakatsuka et al., 1983; York et al., 1986). Moreover, results from a preclinical safety assessment of paraxanthine (ENFINITY™) produced by Ingenious Ingredients, L.P. demonstrated a lack of mutagenicity and genetic toxicity (Purpura et al., 2021). Furthermore, paraxanthine did not result in mortality or toxicity in 28- or 90-day repeated-dose oral toxicity studies (up to 300 mg/kg bw/day) (Purpura et al., 2021). Therefore, the no observed adverse effect level (NOAEL) was determined to be ≥ 300 mg/kg bw/d, the highest dose tested, for both male and female Wistar rats.

Given that many individuals struggle with the untoward consequences of caffeine use, paraxanthine represents a potentially attractive methylxanthine substitute for achieving similar stimulant effects with an improved safety profile over caffeine, based on the observations described above from several studies (e.g., paraxanthine has anxiolytic activity, and relative to caffeine it has low abuse liability, less clastogenicity and cytotoxicity; and lower embryo toxicity). However, in order to justify the use of paraxanthine as a food ingredient or dietary ingredient for use in dietary supplements as a caffeine alternative, its safety profile must be thoroughly characterized and compared to that of caffeine. In the United States, the addition of new food ingredients into the market requires that the ingredient be either an approved food additive or GRAS for its intended use. Alternatively, a New Dietary Ingredient (NDI) entering the marketplace must be notified to FDA prior to marketing in dietary supplements. These premarket gates ensure that the novel food ingredients and new dietary ingredients have been assessed to minimize safety risks to consumers.

While a general toxicity profile for paraxanthine has supported a lack of toxicity, it remains unclear how the potential toxicity of paraxanthine compares to the caffeine molecule. As such, the aim of this study was to independently evaluate the safety of paraxanthine (Rarebird Inc.) consumption and compare it to the safety profile of caffeine consumption using the same animal model through a battery of toxicological studies conducted in accordance with international guidelines, including mutagenicity (bacterial reverse mutation and in vitro mammalian chromosomal aberration), genotoxicity (in vitro mammalian cell gene mutation) and acute, sub-acute and sub-chronic oral toxicity studies.

The battery of toxicological studies conducted on paraxanthine (supplied by Rarebird Inc.) demonstrated a lack of in vitro and in vivo toxicity. Altogether, results from the three battery genetic toxicology studies (bacterial reverse mutation, mammalian chromosomal aberration, mammalian cell gene mutation) suggest that paraxanthine is non-mutagenic at all doses tested. In the acute oral toxicity study, an LD₅₀ of 829.20 mg/kg body weight (bw) was determined for paraxanthine. There was no mortality or treatment-related adverse effects observed for the assessed parameters in the 14-day repeat dose oral toxicity study following daily oral administration of low, mid, or high doses of paraxanthine (50, 100, or 150 mg/kg bw). The same findings were observed in the subchronic repeat-dose 90-day oral toxicity study at daily doses up to 185 mg of paraxanthine/kg bw resulting in the assignment of the highest dose tested as the NOAEL. While there were no reported deaths upon administration of paraxanthine, mortality was reported in two animals in the high dose caffeine-treated group (185 mg/kg bw). Therefore, the NOAEL from the 90-day study was determined to be 150 mg/kg bw for caffeine. These findings are summarized in Figure 6.

The genetic toxicology studies provide evidence that paraxanthine is non-mutagenic. In the Ames test, paraxanthine did not result in an increased number of revertants in five strains of Salmonella typhimurium at concentrations up to 3,000 ug/plate. While the in vitro mammalian cell gene mutation test suggested that paraxanthine (up to 300 μg/ml) was mutagenic at the TK locus of cultured mouse lymphoma cells in the presence of metabolic activation, the in vivo mammalian bone marrow chromosomal aberration test demonstrated a lack of chromosomal aberration and toxicity to the bone marrow of Sprague Dawley rats (doses up to 100 mg/kg bw). When comparing the latter two studies, it is important to recognize the drawbacks of using mammalian cell culture versus an in vivo model. While cell lines are valuable research tools, the behavior of cell lines can vary in culture in comparison to an intact organism. A similar phenomenon to the findings of these genotoxicity studies has been observed with caffeine, which is mutagenic to mammalian cells in culture but not mutagenic when the mammalian cells are cultured in the presence of liver extracts containing detoxifying enzymes (Nehlig and Debry, 1994). Similarly, coffee and caffeine are devoid of mutagenic effects in vivo. Considering that paraxanthine is intended to be an ingredient in human food, the data from the in vivo Mammalian Bone Marrow Chromosomal Aberration test better represent the effects of the test item in an intact animal. Therefore, it can be concluded that paraxanthine is non-mutagenic at doses up to 100 mg/kg bw.

While these preclinical studies are important for determining the genetic toxicity potential of paraxanthine, the in vivo mammalian bone marrow chromosomal aberration test only involves the administration of a single dose of the test item. Since paraxanthine is intended to be added to foods proposed to be ingested chronically, in vivo repeat-dose studies are warranted to better understand the toxicology profile of paraxanthine. In order to establish a dose that can be used for subchronic studies, a maximum-tolerated dose study was first conducted to assess the acute toxicity of paraxanthine. Single dose administration of paraxanthine by oral gavage up to 750 mg/kg/b.w. in female Sprague Dawley rats showed no adverse effects on body weight, or body weight gain. Moreover, gross necropsy examination revealed no treatment related adverse effects. Based on these results, the calculated single dose LD₅₀ of paraxanthine was 829.20 mg/kg b. w. Therefore, the LD₅₀ of paraxanthine is 2.3-fold higher than the value that has been reported for caffeine (LD₅₀ 367 mg/kg b. w.) (Adamson, 2016). Moreover, the LD₅₀ of paraxanthine was determined to be 1,601 mg/kg bw in another study in rats, suggesting that the toxicity may be even lower than the levels reported in the present study (Purpura et al., 2021).

Repeat-dose oral toxicity was subsequently assessed in a 14-day study. Rats administered paraxanthine up to 150 mg/kg bw showed no treatment-related adverse effects on body weight, body weight gain, feed consumption, hematology, coagulation or clinical chemistry parameters, organ weights, gross pathological findings, microscopic examination of tissues or clinical signs compared to the vehicle. Although there were significant differences observed in certain parameters (feed consumption, body weight, clinical chemistry parameters), the differences were not toxicologically significant. To better assess the subchronic toxicity of paraxanthine, a repeat-dose 90-day study was subsequently conducted.

In the 90-day repeat dose oral toxicity study, there was no mortality reported in rats that received paraxanthine at doses up to 185 mg/kg bw, while mortality was observed in two rats that received the same dose of caffeine. Although a reduction in body weight gain was observed in paraxanthine-treated rats, reductions in body weight gain have also been noted in other studies where stimulant ingredients have been evaluated, including caffeine (Zheng G. et al., 2004; Greenberg et al., 2006; Tabrizi et al., 2019; Sirotkin and Kolesárová, 2021). In fact, phytochemicals, such as methylxanthines, have been reported to promote body fat reduction and weight loss in obese individuals. The role of methylxanthines on the modulation of adipose tissue function has previously been reported, and methylxanthines have been identified as potential substances for the development of novel therapeutic approaches for obesity treatment (Carrageta et al., 2018). Since the dose-dependent decrease in body weight gain was an expected pharmacological effect of the test item, and recovery of body weight gain was observed during the recovery period, the effect of both test items (caffeine and paraxanthine) on reducing body weight gain was not considered to be an adverse effect (Zheng G. et al., 2004; Kobayashi-Hattori et al., 2005; Sirotkin and Kolesárová, 2021). Moreover, comparatively, the expected/anticipated test item-related pharmacological effects of paraxanthine on body weight and body weight gain were less (less reduction and recovery) compared to caffeine-treated animals. Similar findings were observed in another study where paraxanthine was administered to Wistar rats (Purpura et al., 2021). In this study, the authors also reported reduced feed consumption, body weight and body weight gain. It was suggested that the changes could be due to increased methylxanthine-related changes in thermogenesis and appetite suppression. In addition to methylxanthines contributing to changes in body weight, behavioral changes in paraxanthine-treated animals were also likely mediated by the effects of the methylxanthine. Paraxanthine-treated animals showed clinical signs of hyperactivity and hypereflexia, while caffeine-treated rats demonstrated hyperactivity, burrowing and partial eyelid ptosis. However, the effects were not considered to be adverse events as the clinical signs were completely reversed at the next-day observation. Furthermore, all of these findings have been reported in rodent models exposed to methylxanthines such as caffeine, theophylline and theobromine (Izawa et al., 2010; Dalefield, 2017; Murray and Traylor, 2022).

Some parameters evaluated in clinical chemistry were significantly different in paraxanthine and caffeine-treated animals compared to vehicle controls. However, similar to the changes described above, the differences were considered to be non-toxicologically significant as the changes were either non-dose dependent; the values remained within a normal, healthy physiological range; or differences in values occurred in only one sex. Furthermore, several of these findings represented non-adverse changes because they were accompanied by no histopathological changes or changes in organ weights. The few observed histopathology findings were considered incidental or background to the Sprague Dawley rats used in the study as they were sporadic in nature and lacked consistency (McInnes, 2012). Significant differences in urinalysis parameters of paraxanthine and caffeine-treated rats were considered incidental, and the increase in urine volume might be due to the diuretic effect of the test item (Osswald and Schnermann, 2011).

Overall, it was concluded that were no treatment-related external or internal gross pathological changes, differences in organ weights, or effects on urinalysis, hormonal parameters, hematology, or coagulation in any of the animals at all tested dose levels of paraxanthine relative to the vehicle control groups. These results are in accordance with findings from a previous 90-day toxicity study in Wistar rats, administered paraxanthine from a different ingredient supplier. In this study, the authors reported the NOAEL to be the highest dose (300 mg/kg bw) tested since there were no treatment-related adverse effects (Purpura et al., 2021). Similarly, in the present study, the NOAEL was determined to be the highest dose tested. However, the NOAEL value in the present study was limited by the upper level studied, suggesting that the true subchronic NOAEL may exceed 185 mg/kg, as was demonstrated by Purpura and colleagues (Purpura et al., 2021).

Based on the above findings, the results demonstrate that paraxanthine was safe and well-tolerated under the conditions of these studies, as there were no treatment-related adverse effects of toxicological significance in rats exposed to all doses of paraxanthine. Comparatively, the expected/anticipated test item-related pharmacological effects of paraxanthine on body weight and body weight gain were found to be less (less reduction and recovery) compared to caffeine-treated animals. Moreover, mortality was reported in the caffeine-treated animals while all animals in the paraxanthine group survived the study. In view of the above findings, the NOAEL of paraxanthine was determined to be 185 mg/kg body weight and NOAEL for caffeine was 150 mg/kg body weight in both male and female Sprague Dawley rats following once daily administration for 90 consecutive days by oral gavage under the experimental conditions employed. These findings open future discussions about whether paraxanthine might be safer than caffeine.