9 The principal risk of administering essential oils or their constituents to any part of the GI tract is irritation and inflammation of the mucous membrane. This has occasionally been reported following overdose, for example with wormseed oil (Opdyke 1976 p. 713–715), but does not always occur, such as in a case of tea tree oil overdose (Del Beccaro 1995). The concentration of essential oil is an important determining factor, as are the total dose and frequency of exposure. Repeated exposure, therefore, increases the risk of irritation. There have been very occasional reports of oral sensitivity to peppermint oil and menthol, with burning, ulceration and/or inflammation either on contact or after excessive, prolonged use (Morton et al 1995; Rogers & Pahor 1995; Fleming & Forsyth 1998). Considering the widespread use of peppermint oil, such reactions are rare. In contrast, we found reports of 63 oral adverse reactions to cinnamon bark oil as an ingredient of toothpastes, chewing gums or food flavorings. Symptoms invariably included either erosion of mucosal tissues or allergic reaction, sometimes extending to the lips (Allen & Blozis 1988; Miller et al 1992; Cohen & Bhattacharyya 2000; Endo & Rees 2007; Tremblay & Avon 2008). In tests performed on one of us (RT) and two other volunteers, 0.02 mL of various undiluted essential oils was applied to a small area on the inside of the mouth, which was kept open to prevent dilution of the oil with saliva. Subjective observations were made for up to 3 minutes, and any irritation became apparent 15–30 seconds after application. Essential oils rich in cinnamaldehyde, eugenol, carvacrol or thymol generally produced the most severe reactions. Other constituents that are probable mucous membrane irritants include allyl isothiocyanate, ethyl acetate and p-cresol (see Constituent profiles, Chapter 14). High concentrations of perillyl alcohol may also irritate the alimentary mucosa (Hudes et al 2000). Adverse effects other than irritation are possible from essential oil ingestion, at least with very high doses. In clinical trials using perillyl alcohol in the treatment of various cancers, doses ranging from 2,400–8,400 mg/m2 per day1 (equivalent to approximately 62–218 mg/kg) produced strong GI side effects that increased with dose. These included nausea, vomiting, satiety and eructation, and may have been exacerbated by the large amount of soybean oil that the perillyl alcohol was mixed with (Ripple et al 1998, 2000; Liu et al 2003; Murren et al 2002; Bailey et al 2002, 2004, 2008). Medicines containing essential oils may be formulated in different ways depending on the desired pharmacokinetic characteristics, for example, enterically coated capsules for dyspepsia and irritable bowel syndrome. These dissolve preferentially in the relative alkalinity of the small intestine (pH about 8) but not in the acidic stomach environment. Table 9.1 shows recorded adverse effects from clinical trials. In these, the incidence of side effects is regarded as low, and equal to or better than that of more conventional medications. Table 9.1 GI reactions in clinical trials using oral doses of essential oils and constituents aAdverse reactions assessed by the researchers as being unrelated to the essential oil medication. In the other cases, causation by the test medication could not be ruled out. Since peppermint oil relaxes the lower esophageal sphincter, chewing enterically coated capsules, instead of swallowing them whole, can cause heartburn. This has happened in some trials. Mucous membrane irritation is concentration-dependent. For example, although mustard oil is a severe irritant (Leung & Foster 2003), at 0.125%, it was only slightly irritating to the human oral mucous membrane (Simons et al 2003). As mentioned in Chapter 2, phenolic compounds have a particular tendency to be irritants. Applied for 5 minutes to the tongues of dogs, undiluted eugenol caused erythema and occasionally ulcers (Lilly et al 1972); applied to rat labial mucosa and observed for up to 6 hours, it caused swelling, cell necrosis and vesicle formation (Kozam & Mantell 1978); applied to mouse mucous membrane, it progressively destroyed epithelial cells, causing an acute inflammatory response (Fujisawa et al 2001). When rats were given 150 mg eugenol po (~ 450 mg/kg), there was gastric epithelial damage, with punctate hemorrhages in the pyloric and glandular regions (Hartiala et al 1966). However, given orally at 10–100 mg/kg, it reduced the number of ulcers and the severity of lesions induced by ethanol and platelet activating factor (Capasso et al 2000), and doses of 5–50 μg/kg reduced tongue edema in mice (Dip et al 2004). Evidently, lower concentrations of eugenol have the opposite effect to higher ones on mucous membranes. Markowitz et al (1992) reported that low concentrations of eugenol exerted anti-inflammatory and local anesthetic effects on dental pulp, but high concentrations were cytotoxic, and could cause extensive tissue damage. Although thymol, another phenol, is said to irritate the gastric mucosa (Reynolds 1993), oral Lippia sidoides leaf oil, consisting of 66.7% thymol, dose-dependently inhibited alcohol-induced gastric lesions when given at 1, 5 or 10 mg/kg. This effect reduced slightly as the dose was further increased to 50 and 100 mg/kg (Monteiro et al 2007). Origanum onites oil (64.3% carvacrol, an isomer of thymol) protected rats against colonic damage induced by trinitrobenzene sulfonic acid (TNBS), used at 0.1 or 1.0 mg/kg intrarectally in olive oil. Ulceration, mucus cell depletion and inflammation were all significantly reduced (Dundar et al 2008). These phenomena may be related to the concentration-dependent, pro-oxidant/antioxidant action of phenols. In tests of gastric toxicity in rats, essential oils of Ocimum basilicum (22% eugenol), Ocimum gratissimum (30% thymol) and Cymbopogon citratus (66% citral), diluted in corn oil, were administered by gavage in doses of 50, 500, 1,000, 1,500, 2,000, 3,000 or 3,500 mg/kg/day for 14 days. NOAELs for functional damage to the stomach were 1,500, 1,000, and 1,000 mg/kg, respectively. Higher doses caused erosion of the stomach mucosa and disappearance of the surface epithelium (Fandohan et al 2008). Essential oils with non-phenolic major constituents may also be protective. Cardamon oil (26.5–44.6% 1,8-cineole) inhibited ethanol-induced gastric irritation and ulcerative lesions in rats at 40 mg/kg, although 12.5 mg/kg had no effect (Jamal et al 2006). This gastroprotective action may be related to antioxidant activity of 1,8-cineole (Santos & Rao 2001). Pre-treatment with 1,8-cineole by rectal instillation at 200 or 400 mg/kg, in an emulsion with Tween 80, attenuated TNBS-induced colonic damage in rats, and caused repletion of glutathione (Santos et al 2004). Gastroprotective and, where known, gastrotoxic doses of essential oils and constituents are shown in Table 9.2. In order to prevent GI irritation, orally dosed essential oils need to be dispersed in a vehicle, such as a gel or fixed oil. In a study to assess the efficacy of peppermint oil in reducing gastric spasm during endoscopy, a 1.6% concentration of peppermint oil emulsified in water (using a surfactant) was sprayed directly onto the pyloric ring in 50 patients; there were no side effects (Hiki et al 2003). Colpermin capsules contain 187 mg of peppermint oil in a thixotropic gel to ensure adequate dispersal in the bowel (Liu et al 1997), while Mintoil capsules contain 225 mg of peppermint oil and 45 mg of a starch that absorbs oils (Cappello et al 2007). In a study of 14 patients with enteric parasites, tablets containing 200 mg of emulsified oregano oil were used. The emulsification process was said to increase the surface area of the essential oil by six orders of magnitude (Force et al 2000). Some xenobiotic compounds cause direct injury to liver cells, while others are metabolized by CYP enzymes to toxic substances, notably highly reactive molecules called electrophiles.2 If a xenobiotic is absorbed in quantities sufficient to overwhelm detoxifying enzymes, the liver itself becomes subject to injury. This may be an acute event, or it may progress over time. Electrophiles may cause necrosis (cell death), in some cases with depletion of hepatic glutathione. The main types of liver injury caused by xenobiotics are: hepatocellular necrosis, steatosis (fatty degeneration), cholestasis (obstruction of bile flow), granulomas (small nodules) and vascular lesions. Essential oils may not be implicated in these last two. Oxidative stress, steatosis and CYP enzyme induction are closely associated with hepatic cancer. Because the liver bioactivates carcinogens, it is the primary target for cancers from DNA-reactive molecules. Essential oil constituents have been reported to increase porphyrin production, disrupt protein synthesis, interfere with metabolic enzymes, or cause various microscopic or macroscopic structural changes. A few constituents are only hepatotoxic to a subset of the population (Pharmacogenetic toxicity, below). Sufficiently high doses can be lethal. In cases of fatal toxicity from parsley apiole ingestion, considerable liver damage is generally found post mortem (Lowenstein & Ballew 1958; Amerio et al 1968; Colalillo 1974). Similarly, autopsy results indicate that fatal doses of camphor can damage the liver (Smith & Margolis 1954; Siegel & Wason 1986). Although the liver has a substantial capacity to regenerate following injury, this capacity is not unlimited, and is reduced in some individuals due to age, disease or lifestyle factors. Liver impairment typically presents as general malaise, either with or without jaundice, and is most commonly assessed by means of ‘liver function tests’. When hepatic cells are damaged, they release increased quantities of certain enzymes into the blood (Box 9.1). Liver function tests measure the levels of these, and of bilirubin. Liver damage may be categorized as hepatocellular (mainly ALT elevation) or cholestatic (mainly ALP elevation), though mixed types of liver injury are common (Bénichou 1990). An essential oil of Salvia officinalis containing 17.4% α-thujone, 3.9% β-thujone and 3.3% camphor was not toxic to rat hepatocytes at 200 nL (of essential oil dissolved in DMSO) per mL of cell suspension or less. However at 2,000 nL/mL a significant increase in LDH and decrease in glutathione occurred, indicating cell damage (Lima et al 2004). Deteriorating liver function was a feature in a case of near-fatal poisoning in a 2-year-old boy who ingested 5–10 mL clove oil (Hartnoll et al 1993). Similarly, a 15-month-old boy developed acute liver failure after ingesting 10–20 mL clove oil (Janes et al 2005). Administered orally at 600 mg/kg, eugenol, the major constituent of clove oil, caused liver damage in mice whose livers had been experimentally depleted of glutathione (Mizutani et al 1991). However, oral administration of 1,000 mg/kg eugenol to rats with normal livers did not deplete glutathione, in fact the activity of glutathione S-transferases (see Glutathione depletion, below) was induced after two weeks of treatment (Vidhya & Devaraj 1999). There is clearly a risk of hepatotoxicity from elevated doses of eugenol, but it is not known whether this is associated with glutathione depletion. Eugenol, however, is hepatoprotective in low (therapeutic) doses (see Hepatoprotection, below). The dose-dependent action of eugenol in relation to the liver is similar to its effect, and that of other phenols, in relation to the GI tract and mucous membrane irritation (discussed above). Safrole caused reversible lipid peroxidation and oxidative DNA damage in rat liver, as evidenced by dose-dependent increases in serum ALT and AST activities, on ip dosing at 250–1,000 mg/kg (Liu et al 1999). Intraperitoneal injection of 500 mg/kg of benzaldehyde or benzyl alcohol in rats caused a significant increase in the rate of reactive oxygen species formation in hepatic mitochondrial fractions (Mattia et al 1993). Benzaldehyde was subsequently found to inactivate glutathione peroxidase with a Ki value of 15 mM (Tabatabaie & Floyd 1996). Parenchymatous degeneration is the mildest form of liver degeneration. The hepatocytes become swollen, and uptake of water may cause the formation of apparently empty cytoplasmic areas (Niesink et al 1996 p. 687). Dietary methyl salicylate, fed to dogs at 150 or 350 mg/kg for 2 years, caused the formation of abnormally large liver cells, with both macroscopic and microscopic signs of hepatotoxicity. However, these adverse effects were not apparent from doses of 500 mg/kg for 9 days, 250 mg/kg for 52 days or 50 mg/kg for 2 years (Webb & Hansen 1963). Gavage doses of 200 mg/kg/day (−)-menthol in soybean oil for 28 days resulted in vacuoles appearing in the hepatocytes of 4 out of 20 rats. Lower doses were not tested, and the finding has not been replicated (Thorup et al 1983b). This is 13 times the recommended maximum oral adult dose for peppermint oil of 1.2 mL (1,080 mg) (Table 4.7). In studies on isolated rat hepatocytes, 0.1–4.0 mM (−)-menthol had a slight inhibiting effect on the release of ALT, AST and LDH, and was therefore protective at this concentration range (Manabe et al 1987). Microscopic liver lesions in rats appeared after oral administration of safrole or estragole at 650 mg/kg/day for 4 days (Taylor et al 1964). A single ip injection of 300 mg/kg estragole in mice was hepatotoxic to adult females, but not to adult males or to suckling offspring of either sex (Vasil’ev et al 2005). Dose-dependent hepatotoxicity (paler and smaller livers) was observed in pregnant female mice following savin oil administration at 15, 45 and 135 mg/kg/day sc during 10 gestational days (Pages et al 1989b). Savin oil has been reported to cause liver lesions in guinea pigs suggestive of a degenerative hepatitis (Patoir et al 1938a, 1938b). Fatty degeneration is also known as steatosis. The accumulation of fat (triglyceride) globules within liver cells results in deterioration of tissue and diminished functioning. A seriously fatty liver becomes enlarged, and can contain as much as 50% fat instead of about 5%. Safrole caused liver enlargement in rats after administration at 150 mg/kg/day for 32 weeks (Gray et al 1972). When fed to rats at 1,000 ppm or 10,000 ppm for up to 370 days, safrole caused hepatic fatty degeneration and there was an increased incidence of neoplasms at both dietary levels (Homburger et al 1962). The exposure of rat hepatocytes to micromolar quantities of α-asarone in vitro caused morphological and ultrastructural changes, fat accumulation and inhibition of protein synthesis (López et al 1993). Some individuals are not able to properly biosynthesize heme, a building block of hemoproteins including hemoglobin and myoglobin, a protein found in certain muscles. Molecules called porphyrins are used in heme biosynthesis, and heme synthesis defects are also known as ‘porphyrias’. Heme synthesis defects are a rare underlying cause of some types of anemia. (+)-Camphor, α-pinene and α-(−)-thujone (a sample containing 3% β-(+)-thujone) increased the production of porphyrin in primary chick embryo liver cell cultures at 100 μM and above (Bonkovsky et al 1992). These were the only substances tested. This suggests caution in patients with underlying defects in hepatic heme synthesis, such as acute hepatic porphyrias. Ingestion of a mouthwash caused a clinical crisis in an 18-year-old female with hereditary hepatic porphyria. The mouthwash contained 1,8-cineole, menthol, thymol and methyl salicylate, but only 1,8-cineole was active in in vitro testing. The patient admitted drinking ‘substantial quantities’ of the mouthwash over a period of several months (Bickers et al 1975). Little is known about the bioavailability of these constituents when applied topically, although peak blood levels of camphor were reported to reach a maximum of 0.2 µM after being applied undiluted to human skin (Martin et al 2004). It is therefore unlikely that topical application of essential oils will present any risk. Because camphor, pinene and cineole are ubiquitous in essential oils, and because other unknown constituents might also be active, we caution the use of all essential oils orally in people with heme synthesis defects. Glutathione depletion occurs in the presence of large amounts of a reactive substance. In this scenario, reactive molecules are free to attack and seriously damage liver and blood cells, before the glutathione is replaced. Severe damage can be fatal, and is due to cell death around the central vein, leading to liver failure and/or hemolytic anemia.3 An oral overdose of an essential oil could be absorbed in sufficient quantities to deplete hepatic glutathione, although only a few oils are known to present such a risk. When either pennyroyal oil or (1R)-(+)-β-pulegone was administered ip to mice, doses of 300 mg/kg decreased hepatic glutathione to ~ 75% of control levels in three hours, and 400 mg/kg caused extensive hepatic necrosis (Gordon et al 1982). Glutathione depletion by pulegone or pennyroyal oil at lower doses has not been investigated. Ingestion of almost 30 g pennyroyal oil by a woman was survived with few consequences other than vomiting, following treatment with N-acetylcysteine, which promotes glutathione replenishment (McCormick & Manoguerra 1988). A 22-month-old girl survived after ingesting < 20 mL of pennyroyal oil. She was given gastric lavage, followed by activated charcoal, sorbitol and N-acetylcysteine. Ten hours after ingestion, her liver function tests were normal (Mullen et al 1994). Sullivan et al (1979) reported the ingestion of 30 mL of pennyroyal oil by an 18-year-old girl resulting in massive hepatic necrosis. Liver function tests became abnormal 24 hours after ingestion; she lapsed into coma and died on the sixth day in hospital from brain stem dysfunction due to liver damage. The only intervention was the administration of plasma and platelets. Cassia bark oil depletes glutathione in experimental animals (Choi et al 2001), as do (E)-cinnamaldehyde and salicylaldehyde (see Constituent profiles, Chapter 14), two of its constituents. Cinnamaldehyde was administered ip to rats at 0.5 mL/kg (Boyland & Chasseaud 1970). Salicylaldehyde is more toxic to hepatocytes, but cinnamaldehyde depletes glutathione more extensively (Niknahad et al 2003). As an α,β-unsaturated carbonyl compound, (E)-cinnamaldehyde reacts with cellular nucleophiles such as glutathione, and there is good evidence that alkenal-mediated oxidative stress contributes to cytotoxic/genotoxic cell damage (Janzowski et al 2003). In addition, α,β-unsaturated carbonyl compounds, including (E)-cinnamaldehyde, irreversibly inhibit human glutathione S-transferase P1-1 (Van Iersel et al 1997). However, since at 25 or 50 mg/kg ip in rats, cinnamaldehyde reduced the activity of glutathione S-transferase, but did not deplete glutathione, its hepatotoxic action only manifests at very high doses. Salicylaldehyde is not found in significant concentrations in any commercially available essential oils. p-Cresol rapidly depletes hepatic intracellular glutathione in rat liver ex vivo (Thompson et al 1994, 1996). Since birch tar oil contains p-cresol as a major constituent, it probably has a potential for glutathione depletion. High doses of (E)-anethole and (E)-anethole 1’,2’-epoxide (AE), a reactive metabolite, deplete glutathione (Marshall & Caldwell 1992, 1993). The toxicity of (E)-anethole is due to AE, different amounts of which are produced in different species. The quantities of (E)-anethole-rich essential oils used in aromatherapy pose no risk (see (E)-Anethole profile, Chapter 14). Administration of 125 mg/kg ip of coumarin to male Sprague-Dawley rats depleted hepatic glutathione after 2 hours and caused hepatic centrilobular necrosis after 24 hours. A marked reduction of GSH levels was also observed in vitro. The coumarin metabolite, coumarin 3,4-epoxide is responsible for this toxicity (Lake et al 1984, 1999). Since the epoxide is the major metabolite in rats but only a minor one in humans, and is detoxified much more efficiently in humans (see Coumarin below), glutathione depletion is not an issue in humans. It would be prudent to avoid oral administration of certain essential oils in people taking drugs such as acetaminophen (paracetamol), which rapidly consumes hepatic glutathione. Box 9.2 lists essential oils high in (1R)-(+)-β-pulegone and p-cresol.
The digestive system
The gastrointestinal tract
Dose-related effects
Excipients
The liver
Hepatotoxicity
Liver function tests
Oxidative stress
Visible lesions
Fatty degeneration
Porphyrin production
Glutathione depletion
