3 Toxicity also depends on the frequency of use and on the susceptibility of the individual. Individual sensitivity to potentially toxic substances can vary considerably, depending on such factors as age, sex, genetic profile, nutritional status and health status. These can be explained by differences in metabolic and eliminative capacity, drug interactions, and so on (Dybing et al 2002). Thus, infants, those taking prescription medication, pregnant women, the elderly, and people with life-threatening diseases may be at greater risk. In the case of medicines, a dose has to be found at which the therapeutic benefits outweigh any adverse effects. An approximate estimate of the window of drug safety can be gained from the therapeutic index (TI), as shown in Box 3.1. The larger the value of TI, the greater is the margin of safety. For commonly used drugs, TI varies from about 2 for digoxin to about 100 for diazepam. Notably, the TI for ethanol is about 10. The TI is regarded as a very rough measure of safety because it can vary between individuals, species, end-points and routes of administration. There is also the possibility of idiosyncratic toxic effects. A more useful estimate of drug safety is the standard margin of safety (Box 3.1), which compares the lowest dose required to produce a toxic effect with the highest dose required to produce a therapeutic effect. This estimate does not rely on the slopes of the dose–response curves being similar. A value of < 1 would mean that a dose effective in 99% of a population would be toxic in more than 1% (Fleming 2003) It should be borne in mind that a xenobiotic substance may be inherently toxic or it may be metabolized into a toxic substance in the body. For further details of toxicokinetics see Chapter 4. Interactions between constituents in a mixture are notoriously difficult to predict. When two or more substances are co-administered, three outcomes are possible. The simplest is ‘additivity’, where the action and potency of the mixture are as predicted from the known actions and quantities of its constituents.1 A second possibility is ‘synergy’ (sometimes referred to as synergism or potentiation). In this case, the mixture’s action is significantly greater than would be expected on the basis of additivity. In the context of pharmacology, this would be desirable because the therapeutic dose can be reduced. However, in terms of toxicology, an enhanced effect would be undesirable. The third possible outcome is ‘antagonism’, which is the opposite of synergy. On administering two substances simultaneously, the observed action is less than anticipated. While this may be unfavorable for a therapeutic effect, it would be beneficial for toxicity. An apparently synergistic effect was seen when linalyl acetate, terpineol and (±)-camphor were tested individually or in pairs for in vitro activity against two human colon cancer cell lines. Neither camphor nor terpineol alone had any effect, linalyl acetate had a minimal effect, and linalyl acetate and terpineol together were moderately effective, causing 33% and 45% reduction in proliferation of the two cell lines. However, when all three compounds were used together, cell proliferation was reduced by 50% and 64% in the two cell lines. There was no toxic effect on normal intestinal cells (Itani et al 2008). An antagonistic effect in essential oils is exemplified by the reduced toxicity of carvacrol in the presence of thymol (Karpouhtsis et al 1998). This apparently manifests in thyme oil high in thymol and/or carvacrol which contains a combined total of 30.9–79.9% thymol plus carvacrol (see Ch. 13, p. 266–267). The rat oral LD50 values of these constituents are 980 mg/kg and 810 mg/kg, respectively. (LD50 is defined under ‘Acute oral toxicity’ below.) If we assume an average LD50 for each of 895 mg/kg, then the LD50 of a thymol/carvacrol CT thyme oil would range from a possible 1,118–2,887 mg/kg. However, the rat oral LD50 of this type of thyme oil is 4,700 mg/kg, making it about half as toxic as would be predicted from the thymol and carvacrol content. Antagonism in skin sensitization is known as quenching. (+)-Limonene had a quenching effect on cinnamaldehyde sensitization in 3 of 11 human subjects, and eugenol had a similar quenching effect in 7 of the same 11 cinnamaldehyde-sensitive subjects. It is postulated that this may be due to competitive inhibition at the receptor level (Guin et al 1984). The same may be true of thymol and carvacrol, which are isomeric. Like fruits and vegetables, essential oils contain complex mixtures of chemicals that may be harmful and/or protective. Plants are vulnerable to oxidative stress because they produce oxygen and reactive oxygen species during photosynthesis. As protection, plants biosynthesize an assortment of potent antioxidants. An antioxidant action is considered fundamental to many antitoxic effects, such as mitigating phototoxicity, allergenicity or mutagenicity. This can be seen, for example, in the antihepatotoxic actions of carvacrol, thymol and eugenol (Jiménez et al 1993; Kumaravelu et al 1995), the gastroprotective effect of 1,8-cineole (Santos & Rao 2001), the antinephrotoxic action of thymoquinone (Badary 1999) and the antimutagenic action of linalool (Beri In other cases, an effect may be seen as either countering potential toxicity, or simply as therapeutic, e.g., the anticonvulsant effects of anise oil or cumin oil (Pourgholami et al 1999; Sayyah et al 2002b), the anti-asthmatic action of turmeric oil and may chang oil (Li et al 1998; Qian et al 1980) and the anticarcinogenic action of (+)-limonene and perillyl alcohol in skin cancer (Lluria-Prevatt et al 2002; Raphael & Kuttan 2003a, 2003b). An observed effect may be enhanced or diminished by constituents in a mixture that do not express that effect directly themselves. For example, in essential oils that contain small amounts of carcinogens, the presence of large amounts of antioxidant, antimutagenic, anticarcinogenic constituents can render the oil non-carcinogenic (see Ch. 9, p. 183). The number of incidents of an adverse reaction to an essential oil depends on: • the number of people exposed to it • the degree of exposure (oil concentration and time of exposure). For example, the two most widely used essential oils in aromatherapy are lavender and tea tree, but there are more confirmed cases of both poisoning and adverse skin reactions for tea tree oil than there are for lavender oil. In this scenario, both the number of people exposed and the degree of exposure are similar, but tea tree oil is inherently more toxic. Similarly, cinnamon bark and spearmint oils are both widely used as flavoring agents in chewing gums, toothpastes and mouthwashes, but there are more reported oral adverse reactions for cinnamon than for spearmint. Virtually all cases of serious poisoning from essential oils are a consequence of oral ingestion of the undiluted oil, in amounts much higher than therapeutic doses. Judging from the large quantities ingested, a few were probably suicides. (Death from essential oil overdose is slow, and can take from 15 minutes to 3 days.) Of the many non-fatal accidents, only two cases cite long-term ill effects, one involving wintergreen oil and one wormseed oil (Kröber 1936; Heng 1987). In fatal or near-fatal cases, a variety of signs and symptoms are possible, such as convulsions, vomiting, and rapid breathing, depending on the essential oil ingested (see Table 15.1). Many cases of wintergreen oil poisoning have been recorded. In six cases of poisoning in adults, three people survived after ingesting 6 mL, 16 mL and 24 mL, and three died from ingestion of 15 mL, 30 mL and 80 mL (Stevenson 1937). None of these cases received any medical intervention. If we take an average of the non-fatal doses (15.3 mL), and an average of the fatal doses (41.7 mL), this gives a human median lethal dose of 0.2−0.6 mL/kg assuming the individuals were of average weight.2 Camphor and methyl salicylate, and the oils of clove, cinnamon and eucalyptus, are most frequently and consistently reported to cause toxicity. There are many instances of eucalyptus oil poisoning, and it is thought to be fatal to humans in oral doses between 30 mL and 60 mL (Gurr & Scroggie 1965). In recent years, cases of tea tree oil ingestion have escalated in the USA, rising from 280 incidents and 11 adverse reactions in 2001, to 966 incidents and 30 adverse reactions in 2006. For comparison, adverse reactions to clove oil were 10 and 13 in the same two years (Table 3.1). With less frequency, ingestion of citronella, hyssop, pennyroyal, sage, sassafras, thuja or wormseed oil has all caused toxicity in humans. This may not be a comprehensive list, since most essential oils will probably cause serious problems if drunk in sufficient quantity. Availability can also influence statistics, as is suggested by the increase in tea tree cases. Table 3.1 Adverse reactions to essential oils in the USA Information from: Bronstein et al 2007; Lai et al 2006; Litovitz et al 2001, 2002; Watson et al 2003, 2004, 2005. Information for specific oils not recorded before 2001. The majority of cases of essential oil poisoning involve accidents with young children, often between 1 and 3 years of age. Approximately 75% of cases in the USA are in children up to 6 years old (Table 3.2). Parents (and consumers in general) need to be aware of the risks. Perhaps contrary to expectation, young children will drink an undiluted essential oil. One report tells of a 10-month-old infant who stood up in her crib, reached for a bottle of camphorated oil, removed the cap, and drank approximately 1 oz (about 30 mL) (Jacobziner & Raybin 1962a). In an Australian report of eucalyptus oil poisoning, 78 of 109 children of 5 years or less ingested solutions intended for vaporization, and the majority were ages 1–3 years (Day & Ozanne-Smith 1997). Table 3.2 Essential oil incidents reported to the central toxicity database in the USA 1997–2006 Information from: Bronstein et al 2007; Lai et al 2006; Litovitz et al 1998, 1999, 2000, 2001, 2002; Watson et al 2003, 2004, 2005. • their natural inquisitiveness leads them to examine materials by putting them in their mouths • a liquid that is being examined by a child will probably be swallowed rather than sipped • being smaller than adults, children are more susceptible to toxic substances • metabolism in very young children is less effective than in adults. Some unfortunate infants have died because a parent administered the essential oil by mistake, thinking it was, for example, castor oil. Some died because the essential oil was intentionally administered, either by a parent or doctor, who was not aware of the toxic consequences. But in most cases, a bottle of essential oil was within reach of the child and they were able to open it. Essential oil poisoning in children is not a new problem. In 1953 Craig & Fraser (cited in Craig 1953) reported that of 502 cases of accidental poisoning in children (1931–1951, Aberdeen and Edinburgh), 74 were due to essential oils. Of 454 deaths from accidental poisoning in childhood that occurred in Britain during a similar period, 54 were caused by essential oils (Craig 1953). Statistics compiled for the USA show that in 1973 there were reports of 530 ingestions of camphor-containing products, 415 in children under the age of 5 years (Phelan 1976). The same year, doctors recommended that the sale of camphorated oil should be restricted (Bellman 1973). In all of these cases the products involved were over-the-counter (OTC) preparations, the majority being camphorated oil. Many OTC products contain camphor at 1–10% (Kauffman et al 1994). Camphorated oil contains similar quantities of camphor to several essential oils (20%), and the risk to young children from these oils is very similar. There have also been serious toxic incidents in young children who have inhaled preparations containing essential oils that have been mistakenly instilled into the nasal cavity (see Ch. 6, p. 108). Among the essential oils known to increase light sensitivity, only the citrus oils are widely used. Oils of bergamot, lemon and lime are the most commonly reported causes of photosensitivity (see Ch. 5, p. 85). At the time of writing, the latest report we could find was for bergamot oil (Kaddu et al 2001). Fresh limes, often in use during sunny weather, are also a frequent cause. The essential oils used in patch testing also tend to be those for which allergenic constituents have not been identified. Testing does not usually include, for example, cinnamon bark oil, since its major constituent, cinnamaldehyde, is routinely used in patch testing. We found four cases of cinnamon oil allergy (presumably bark oil) over a 30 year period: one caused by skin contact with the undiluted oil (Sparks 1985), two from an ointment containing cinnamon oil (Calnan 1976), and one from a cinnamon oil mud bath (Garcéa-Abujeta et al 2005). Skin allergies often follow an epidemiological pattern: • a new cosmetic ingredient is introduced • there are virtually no reports of adverse reactions • over the course of several decades, reports of adverse reactions escalate • use of the ingredient is restricted or otherwise reduced Turpentine oil was a well-known contact allergen for a long time, mainly through occupational exposure, but when the mass paint industry replaced it with petroleum-based substitutes for paint thinning, reported cases of turpentine oil allergy decreased (Schnuch et al 2004a). Similarly, cold-pressed laurel berry oil (which contains some essential oil) was widely used as a conditioner for felt hats for about 100 years (1860–1960). By the 1940s it was recognized as a major cause of dermatitis, and the felt industry ceased using laurel oil in 1962. Since 1975, there have been no reports of laurel berry oil allergy. Not every widely used essential oil causes skin problems. We found four case reports of confirmed dermatitis from citronella oil, but all of them were from contact with the undiluted oil and none of them is recent (Keil 1947; Lane 1922). Considering the extensive application of citronella oil in insect repellants for many decades, the apparent lack of skin reactions is notable. Similarly, in spite of the widespread use of lavender oil in the West since about 1990, we could only find only five confirmed instances of dermatitis from 1991–2000: two involved the undiluted oil being dripped onto pillows at night and causing facial dermatitis (Coulson & Khan 1999), two were cases with multiple sensitivities to essential oils (Schaller & Korting 1995; Selvaag et al 1995), and one was an aromatherapist with hand dermatitis (Keane et al 2000). Reported cases of tea tree oil allergy are more prevalent for this period. For the years 1991–2000, we found 29 cases, and in 21 of them the undiluted oil was used (Apted 1991; De Groot & Weyland 1993; Elliott 1993; Knight & Hausen 1994; Selvaag et al 1994, 1995; De Groot 1996; Bhushan & Beck 1997; Hackzell-Bradley et al 1997; D’Urben 1998; Rubel et al 1998; Khanna et al 2000; Varma et al 2000; Vilaplana & Romaguera 2000). One case was caused by airborne vapors, another by ingestion of the essential oil. We found two case reports of cheilitis to spearmint oil in toothpaste (Poon & Freeman 2006; Skrebova et al 1998), but none for peppermint oil. For cinnamon, not always defined as cinnamon bark oil, Allen & Blozis (1988) reported 10 cases of cheilitis, most from chewing gum, Miller et al (1992) reported 14 cases of cinnamon stomatitis, and we identified a total of 39 further cases of oral adverse reactions (Cohen & Bhattacharyya 2000; Endo & Rees 2007; Tremblay & Avon 2008). Various other adverse reactions have been reported from clinical trials. In a wound healing pilot study, a 3% concentration of lemon myrtle oil caused increased pain and inflammation in some patients, and was discontinued (Kerr 2002). In a dandruff trial using 5% tea tree oil in a shampoo base, 1 of 63 patients reported mild scalp itching; however, 3 of 62 patients who used the shampoo base only reported the same symptom; one patient in each group reported mild scalp burning (Satchell et al 2002b). In a head lice repellent trial, a formulation with 3.7% citronella oil was administered daily for four months to the heads of 103 children. One child reported a slight itching and burning sensation (Mumcuoglu et al 2004). In various trials involving the oral administration of peppermint oil or peppermint + caraway, there have been a small number of heartburn reactions, usually because patients chew capsules instead of swallowing them, and they are not meant to dissolve in the stomach (see Table 9.1). Toxicity tests approved by The Organisation for Economic Co-operation and Development (OECD) are listed in Box 3.2. This is a representative summary of currently used tests, but does not list every type of test referred to in this text. First introduced in 1927, the conventional measure of acute toxicity is the LD50: the dose that kills 50% of a group of animals, or median lethal dose. Different doses of the test substance are administered to matched groups of animals, usually rats or mice. LD50 values vary with species and route of administration. For instance, the acute oral LD50 value for wormseed oil in rats is 255 mg/kg, and 380 mg/kg in mice, and the acute dermal LD50 in rabbits is 415 mg/kg. Dermal LD50 values are generally similar to or higher than oral values, and the more invasive intraperitoneal (ip) LD50 is always lower than the oral value (Table 3.3). Table 3.3 Toxicity ranking based on rodent LD50 values, for orally administered substances aFor ease of comparison this measurement has been changed from mL/kg to g/kg. Table 3.4 gives examples of human lethal doses assuming direct extrapolation from the animal data.
Toxicity
Adverse effects
Toxic substances
Toxicity of mixtures
Interactions between compounds
et al 2007). Also see Table 9.3.
Human toxicity
Poisoning
Adverse skin reactions
Photosensitivity
Contact dermatitis
Adverse oral reactions
Other adverse reactions
Measuring toxicity
Animal toxicity
Acute oral toxicity
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
Toxicity
