Abyssinian tea (Catha edulis Forssk, Celastraceae)
Catha edulis Forssk belongs to the Celastraceae family and, a neighbouring species of the Evonymus, it is the sole representative of the genus Catha. The Swedish botanist Forsskal, the explorer of lower Egypt and Arabia, was the first to mention this plant, which he called gat or khat. His description of it was published posthumously in 1775 in his Flora aegyptico-arabica 8
Author: R. Paris, Madame H. Moyse
Pages: 29 to 34
Creation Date: 1958/01/01
A study of some samples of varying geographical originR. Paris
Professor, Faculty of Pharmacy, Paris Madame H. Moyse
Catha edulis Forssk belongs to the Celastraceae family and, a neighbouring species of the Evonymus, it is the sole representative of the genus Catha. The Swedish botanist Forsskal, the explorer of lower Egypt and Arabia, was the first to mention this plant, which he called gat or khat. His description of it was published posthumously in 1775 in his Flora aegyptico-arabica 
The plant has also been described under the following names: Celastrus edulis Vahl, Catha forsskalii Rich, and Trigonotheca serrata Hochst, Methyscophyllum glaucum Eckl. and Zeyher.
It is a shrub which grows to between 1-2 m in arid regions, and may reach 10 m or more in the equatorial zone; under cultivation, it may often grow to a height of 3-4 m. An evergreen in appearance, it is not unlike the spindle-tree. The leaves, which are bifarious- at least at the top of the branches- are whole, and lanceolate or ovate (8-10 cm long by 4-5 cm broad). The South African form (called Methyscophyllum glaucum Eckl. and Zeyher) is a variety with very narrow, bluey-green leaves (4.25 cm); they have a short petiole (0.5 cm) with two small caducous stipules. The limb, which at the base is smooth, has over the remainder of the leaf some thirty mossy serrations. The median vein projects on the underside and the secondary veins, running at an angle of forty-five degrees, meet in a curve near the margin of the limb. The leaves are coriaceous, smooth, shiny on the upper side, dark green when fresh, but becoming brownish or reddish-brown if kept; they have not much odour, but an aromatic, astringent taste. The flowers of type 5, arranged in axillary cymes, are small, regular in shape and greenish-white: they have a short calyx, petals longer, upright, overlapping, spreading out at the top, five alternisepalous stamens and free ovary with three loculi (Fig. 1). The fruit is a slightly trigonal, oblong capsule with three valves (Fig. 2), containing one to three seeds with a triangular, membranous wing. The albumen is pulpy and the cotyleda are foliaceous.
The anatomical characteristics of the leaf were described in 1893 by E. Collin  . Their structure is fairly ordinary: prominent veins on the underside, glabrous epidermis, phleem-lignous arc, pericyclic strands on the underside of the fascicule only, limb with asymmetrical, heterogeneous mesophyl with two palisade layers; twin crystals of calcium oxalate in the lower epidermis, under the palisade tissue and. in the phloem  . A cross-section of the lower epidermis shows stomata, usually with three attached cells; the upper epidermis, without stomata, consists of cells with sinuate walls.
Geographical distribution. According to A. Chevalier  , the plant originated in tropical East Africa. From Abyssinia, where khat grows wild along the river banks above 1500 m, it was introduced into the mountains in south Yemen and is also found in Kenya, Tanganyika, Uganda and South Africa (Transvaal, the Cape, Natal). It is cultivated in Abyssinia (in the Harar district) and in Arabia in irrigated terraces, and is often grown with the coffee plant at an altitude around 2,000 m. It is very hardy and can be grown in the Mediterranean region and in the south of the United States.
It is grown by sowing or from seedlings, and the leafy twigs are picked after three or four years. The drug is sold in bundles of stalks in full leaf weighing about 500 g and measuring 40 cm long by 8-10 cm in diameter. There are several varieties, differing in quality, the yellow leaves being the most sought after.
Chemical composition. The chemical study of khat goes back to the end of the nineteenth century; in 1887 by Flückiger and Gerock [(7)] who, in view of the stimulating properties of the drug, thought that it contained caffeine, discovered the presence of an alkaloid called cathine. Beitter, who wrote a thesis on it in 1900 [(1)] , isolated an amorphous base, katine, the sulphate and hydrochlorate of which are obtained in crystalline form: the yield is Iow-0.2% for leaves from Harar, 0.5% for those from Aden. In 1911, J. Chevalier obtained one per thousand total alkaloids [(5)] from the leaves; the content of the stalks would have been less. A little later, in 1912, Stockmann[ (22)] separated three alkaloids which he called cathine, cathinine and cathidine, the first two in a crystal form; unfortunately, he gave no constants. He pointed out that, unlike cathinine, cathine and cathidine were fairly soluble in water. Cathine and cathinine can also be distinguished by their different solubility in pure alcohol, Cathine sulphate being only soluble with difficulty in this solvent.
In 1933, Wolfes [(24)] undertook the extraction of khat alkaloid by benzine; after separation of cathidine, owing to its insolubility in water, he succeeded in obtaining cathine in a crystalline form (77°F.). The corresponding hydrochlorate, which crystallizes in pure alcohol after the addition of ether, melts at 181°F, its rotatory power being (α) 20 = + 43.3°F (water); its ultimate analysis gives the formula C 9H 13ON ClH. Through these properties, the author showed that cathine is identical with d-nor-iso-ephedrine (or d-nor-pseudo-ephedrine), a substance found in its natural state in certain Ephedras in China (Smith (21)), and of which Nagai and Kanao made the Synthesis with that of many derivatives of ephedrine [(13)] . The formula of this substance is
Twig of Catha edulis, in bloom (Beitter)
According to Manske (11), its constants are as follows:
Base, 77.5-78°F (α) D20 = + 37.9oF (methanol)
Hydrochlorate, prisms 180-181oF (α) D20 = + 43.2oF (water)
Sulphate, hexagonal flakes 295°F (α) D20 = + 48.7oF (water)
As far as we are aware, no chemical study of the constitution of the khat alkaloid later than that of Wolfe's has been made.
Beitter [(1)] , moreover, had noted the presence of mineral substances (11% ash), a tannin, a rubber-like substance similar to that of the Celastrus,glucides (manittol), and traces of essential oil. More recently, Plouvier [(20)] identified dulcitol, as with various other Celastraceae.Mustard [(12)] determined the ascorbic acid content, which is fairly high in the leaves (324 mg- 100 g) and twigs (138 mg-100 g), and which perhaps plays a role in the tonic action of the drug [(25)] . As to cholein, the presence of which is indicated in certain materia medica works, there seems to have been confusion with Caltha palustris (cf. E. Paulsen, Tidsk. Kem. Farm. Terap., 1916, 13, 237 and 262).
Use and Physiological Action
Khat leaves have been long used by the Arabs and Egyptians in order to overcome fatigue and sleep. Although there is no proof that its use goes back to classical antiquity, according to Leloup [(10)] , it was known before coffee, and it was mentioned in the fifteenth century. Its main use is for chewing, and in the Yemen it was the custom to offer it to visitors, each guest, after chewing the leaves, throwing the remains with the stalks on the ground [(2)] . Khat chewers are thus able to dance and talk all night without feeling sleepy, both fatigue and hunger being eliminated; the drug also seems to be anaphrodisiac [(10)] . With a slightly astringent taste, Khat also makes a very agreeable tea (khatis the leaf and kaftathe beverage) in the proportion of 5-15 g per litre. According to Greenway , it is also used in South Africa as a stimulant and as a remedy against diseases of the chest.
Twig of Catha edulis, with fruit (Beitter)
The fresh drug is always preferred to the dry plant, which is considered to have lost part of its effectiveness. Continued use of khat leads to addiction, and in the long run khat chewers may become auto-intoxicated, the drug inducing a state of stupor, hebetude, changes of character and incapacity for work. Greenway [(9)] reports some cases of madness and one case of poisoning. In addition to nervous diseases and heart trouble, the after-effects include gastritis, constipation and emaciation.
Cross section of the Catha edulis lea, at the median vein (E. Colin)
With the exception of certain clinical tests by Leloup [(10)] showing the beneficial effects of khat in cases of asthenia and overwork, the first physiological tests were made by Beitter [(1)] , who noted only a temporary paralysis in frogs after large doses of " katine ". But it was J. Chevalier [(5)] who made the first major experiments with cathine (the raw alkaloid, for which he does not give any constants) on different animals--frogs, guinea pigs and rabbits. This substance is not very toxic; after a period of hyper-excitation it induces muscular relaxation, motivity is gradually reduced, the heart slows down and, with heavy doses (0.01 g per frog) stops on the systole. With guinea pigs and rabbits, after intravenous injection of doses of 0.3-0.4 g, there was lack of motor co-ordination and convulsions followed by paralysis.
In 1912, Stockmann [(23)] studied separately the action of the different alkaloids that he had isolated. Cathine appeared to act both on the nervous system and on the muscles, cathinine having a less depressive effect on the encephalon, but being more stimulating to the medulla; cathidine seemed to be mainly a muscular poison but, being insoluble in water, would hardly affect the action of the drug, at least when taken as a tea. In 1942, Brücke [(3)] studied the action of pure carthine sulphate (d-nor-pseudo-ephedrine), the only clearly defined alkaloid of khat. He was particularly interested in its stimulating action on the central nervous system, which he studied in mice, noting the degree of excitation. In particular, he compared the action of cathine with that of pervitine; whereas the latter produced excitation in mice with a dose of 1 α/gr, doses a hundred times stronger of pure cathine sulphate (sub-cutaneous injection) were necessary. The 1-ephedrine had an action similar to that of cathine, the d-pseudo-ephedrine having no excitant effect with an almost fatal dose. The effect of cathine on the blood pressure of a decapitated cat was one thousand times less marked than that of adrenalin, and its action was weaker than that of ephedrine; the high blood pressure induced was lasting. The author concluded that the weak stimulating action of cathine on the nervous system, given the content of this alkaloid in the drug, cannot be the cause of the stimulating effect of the fresh plant. He considered that, at least in fresh khat, there must be some other active principle of the benzedrine type --the formula for which, very similar to that of cathine, is only one oxydril less. That, however, is merely a hypothesis.
Although earlier writers such as Dujardin Beaumetz, Leloup [(10)] and later J. Chevalier [(5)] recommended it as a general stimulant and mental excitant, khat appears to have been used only rarely in European medicine[ (25)] . The 1949 British Pharmaceutical Codex lists it only as a substitute or adulterant of tea.
Characteristic elements of Catha edulis powder
1. Upper epidermis, front view
2. Lower epidermis, front view .
3. Oxiliferous cells
4. Pericyclic strand
In view of recent French legislation on khat and having had available several samples of the drug, we have endeavoured to determine its characteristics [(16)] . This has entailed the definition of some of its chemical and physiological properties.
Parcel 1: Leaves of Ethiopian khat; an old sample kept in the Materia medica Museum of the Faculty of Pharmacy, Paris.
Parcel 2: Leaves of Amani (Tanganyika) khat, picked in 1949, by the courtesy of Mr. P. J. Greenway, Systematic Botanist.
Parcel 3: Fresh twigs of Catha edulis, picked in 1957 at the Botanical and Phytopathological Station of the Villa Thuret, Antibes, Alpes Maritimes (Director: Mr. J. Barthelet).
Parcel 4: Fresh twigs of Catha edulis, picked in July 1957 in the garden of the French Faculty of Medicine and Pharmacy, at Beirut, by Professor B. Lys.
The whole leaves of khat are fairly easily recognizable by their general morphological and anatomical characteristics (Figs. 1, 2 and 3).
The powder can be identified under the microscope, after clarifying with chloral or potassium, by the following elements: many strands isolated or in groups, colourless, tapering, with narrow lumen (Fig. 4 (4); remains of ducts, streaked or dotted; small, calcium oxalate twin crystals isolated or in series within uniserial cells; pieces of lower epidermis with stomata and slightly sinuate cells (Fig. 4 (2)) and, in particular, fragments of upper epidermis whose extremely sinuate walls are scalloped in appearance (Fig. 4 (1)). The absence of sclerites and hairs eliminate tea, and in mate the stomata are surrounded with cells with striped cuticles.
Water content: Determined by the loss of weight at 100°F; Parcel 1, 8.4%; Parcel 2, 9.2%; Parcel 3, 66%; Parcel 4, 60%.
Ash: Parcel 1, 9.4%; Parcel 2, 7.95%.
Data on infusion: With diluted ferric-chloride, it is greenish-brown in colour; on being boiled with hydrochloric acid it takes on a bright red tinge. These reactions are due to the presence of catechuical tannins and were determined, in sample No. 2, by precipitation with the Stiasny reagent (formol-hydrochloric acid), 14%. The foam colour index is approximately 100.
Essential oil: Steam distillation showed only traces of an essential oil lighter than water.
Determination of Alkaloids
This was carried out on 5 g of powder alkalinized by sodium carbonate and extracted, hot, by chloroform. The organic liquid was treated with diluted hydrochloric acid; after alkalinization by ammonia, the aqueous solutions were extracted by ether-chloroform (3-1 by volume), which was evaporated and dried.
This method gave the following contents: Parcel 1, 0.17%; Parcel 2, 0.18%; Parcel 3, 0.18%; Parcel 4, 0.2% (figures reduced to the dried plant and therefore constant in the different parcels).
By cold maceration, the figures obtained were much lower. Moreover, since certain alkaloids are soluble in water, it is advisable to undertake a kind of salting out by sodium chloride or ammonium sulphate when the alkaline aqueous solutions are extracted by ether-chloroform.
In an attempt to obtain alkaloids in crystallized form, an extraction was attempted, using a Soxhlet apparatus, with 250 g of leaves of parcel No. 2. Petroleum ether in a neutral medium carried off a small quantity of alkaloids (0.02%). After alkalinization by ammonia, the powder was successively extracted by ether and by chloroform, the process being continued as in the determination of the total alkaloids, with a resulting residue of 0.08% for ether and 0.03% for chloroform. The total yield was therefore 0.11%.
Extraction by lixiviation by means of a 3% concentration of acetic alcohol, displacement of bases by sodium carbonate and extraction by the ether-chloroform mixture, gave a lower yield of 0.07%.
The total alkaloids gave the reaction of iodoform with iodine in an alkaline medium; with a 10% concentration of copper sulphate and with soda, a blue-green precipitate insoluble in ether was obtained. A portion of the bases being dissolved in a few millilitres of pure alcohol, sulphuric acid was added until a definitely acid reaction was obtained and then, after a few drops of ether, white crystals, recovered by centrifuging and washed in pure alcohol, appeared fairly quickly (particularly if left for a few hours in a refrigerator). The fusion point (instantaneous) was 298°F (decomposition), showing that this was carthine sulphate (d-nor-iso-ephedrine). Also, by dissolving another portion in a small quantity of alcohol containing hydrochloric acid at 5% and the addition of anhydrous ether, a hydrochlorate was obtained with a fusion point (178oF, instantaneous) corresponding to that of cathine hydrochlorate.
Finally, in the analysis of khat and its preparations, paper chromatography and electrophoresis were used, both methods having been already applied by various authors and by ourselves [(14, 15, 17)] for various medicinal plants.
Using paper chromatography, we tried one-fifth and one-half tinctures, an equal weight of fluid extract, and total alkaloids. We used an ascending chromatography at a temperature of 18o± 1% on Arches 301 paper, aqueous butanol, acetic butanol [(4, 5, 1)] or the mixture petroleum ether (B.P. 80°F), amylic alcohol, acetic acid, water (1, 3, 3, 3). As detectors, the Dragendorff reagent (based on Munier's formula) and ninhydrine (0.2% solution in alcohol at 80oF) were used. Ferric perchloride and potassium were used also with the tinctures and the fluid extract, giving from the starting line a very elongated spot on the first third of the chromatogram, corresponding to the tannins. With the one-half tincture and ninhydrine, a rather dull, violet-rose spot (Rf approximately 0.5 in aqueous butanol) was obtained. With the same detector and the fluid extract six clear spots were obtained (in acetic butanol); with Rf 0.11, 0.17, 0.25, 0.28, 0.38 and 0.68, the first three and the last being violet-rose in colour and the other two yellowish.
The total alkaloids were compared with ephedrine (a 2% solution of tevogyne ephedrine hydrochlorate). This base, which showed no fluorescence in ultra-violet rays, was practically insensitive to the Dragendorff reagent (0.1 mg per sampling test was required in order to obtain a dearly defined spot), but reacted better to ninhydrine (20-40γ), giving a violet-rose spot.
This method showed the presence of three alkaloids in khat (sampling test 0.1-0.2 mg of total bases). The first two bases, not revealed by Wood's light or by the Dragendorff reagent under these conditions, gave violet-rose spots in ninhydrine with Rf values approximating to that of ephedrine. The third base, fluorescent, greenish yellow in ultra-violet rays, did not appear after detection by the Dragendorff reagent, and its Rf was much higher:
Rf of khat alkaloids
Base No. 1
Base No. 2
Base No. 3
Acetic butanol (Partridge's mixture)
Petroleum ether - iso-amylic alcohol - acetic acid-water (1,3,3,3)
As regards caffeine, this cannot possibly be confused with the khat alkaloids: it showed no colouring under ultraviolet rays or in ninhydrine, and was only detected by the Dragendorff reagent (Rf 0.68 by acetic butanol and 0.62 by the petroleum ether-iso-amylic alcohol-acetic acid-water mixture).
With paper electrophoresis, using the Machebœuf apparatus and formic acid 2N (pH 1.5), with the total alkaloids (10 mm 3of an 8% solution), in 2? hours several spots appeared fluorescent under ultra-violet rays (2 greenish yellow, travelling very little, and 2 blue). After soaking in the Dragendorff reagent, levogyrous ephedrine (2.5 mm 3of a 2% solution) gave a violet spot. The total alkaloids of khat gave 5 orange-red spots, some having travelled a little, others more or less the same or farther than ephedrine. These results suggest a greater complexity of the alkaloid composition of khat.
Using the four samples studied, experiments were made on mice by subcutaneous injection with:
One-fifth tinctures in alcohol at 60% were prepared by maceration with dry samples of the first two parcels. Just before use, these tinctures were half evaporated and made up to the original volume with distilled water. The fresh drugs (parcels 3 and 4) were stabilized with boiling ethyl alcohol, and, after drying and pulverization, extracted by boiling, using the same solvent. The alcohol was then re-distilled dry under reduced pressure and the residue recovered using a weight of boiling water corresponding to the weight of the drug used.
Toxicity for mice (sub-cutaneous injection)
Parcel No. 1: For a dose corresponding to 10 g of dried plant per kg, only temporary muscular relaxation was observed and mortality was only 10%.
It should be added that with the same dose and under the same conditions, preparations of dried tea leaves produced a mortality of 20% and those of mate 100%.
Parcel No. 2: With a dose of 5 g/kg of dried drug there was good tolerance. With 7.5 g/kg, there was a brief period of agitation with cardiac and respiratory acceleration and pedalling movements with the hind legs, followed by torpor and muscular relaxation: Mortality, 20%.
With 10 g/kg, the stage of excitation was less clearly defined and the oncoming of torpor more rapid: 60% deaths in 48 hours.
Parcel No. 3: The dose of 10 g of fresh plant per kg (corresponding to 3.3 g of dried plant) produced no reaction. With 4 g/kg (dried plant), agitation was observed. With 7.6 g/kg (dried plant) there were phenomena of paralysis and in the following few days mortality was 40%.
Parcel No. 4: After the dose of 10 g of fresh plant per kg (corresponding to 4 g of dried plant), pedalling movements with the hind legs were observed in one-third of the animals and in the others muscular relaxation : Mortality was 30% after 24 hours, 40% after 48 hours; with 25 g/kg (corresponding to 10 g of dried plant) there were the same observations (immediately after the injection mice were panting and making pedalling movements with the hind legs) and mortality was 80% after 24 hours.
It appears, therefore, as already pointed out by different authors (3, 10, 18, 19, 23), that for the same alkaloid content, the fresh plant is definitely more toxic than the dried plant. Total alkaloids were tried with a dose of 50 mg/kg and under the same conditions. With this dose- corresponding practically to 40 g of dried leaves, there was only a short stage of excitation, which was not followed by torpor or death.
The whole plant is thus relatively more toxic, particularly when fresh, than the isolated total alkaloids, which seem to represent only a part of the physiological action. It should be noted that a tincture obtained from a powder from which the majority of the alkaloids have been extracted by ether-chloroform in an ammonia medium still possesses a toxicity almost equal to that of the untreated drug. To sum up, we examined four samples of different origins (from Ethiopia, Tanganyika, the south of France and the Lebanon, the two latter being fresh), and found their total alkaloid content (compared with the dried drug) appreciably the same. Paper chromatography of the latter may be useful in the analysis of the drug. The high toxicity in sub-cutaneous injection, which is relatively higher in the case of the fresh samples, does not seem to be due solely to the alkaloids.
BEITTER, A.: Pharmacognostisch-chemische Untersuchung der Catha edulis, Thesis, Strasbourg, 190002
BOIS, D.: Les plantes alimentaires chez tous les peuples, IV - Les plantes à boissons, Lechevallier, 1937, Paris, pp. 444 445.03
BRUCKE, F. T. VON: über die zentral erregende Wirkung des Alkaloides Cathin, Archiv für die experimentelle Pathologie und Pharmakologie, 1941, 198, 10104
CHEVALIER, A.: Les Cat's d'Arabie, d'Abyssinie et d'Afrique orientale, Revue de Botanique appliquée, 1949, 29, 41305
CHEVALIER, J.: Etude pharmacodynamique du Catha edulis Forssk, Bulletin de la Société de pharmacologie, 1911, 18, 26406
COLLIN, E.: Le Catha edulis Forsskal, Journal de pharmacie et de chimie, 1893, 28, 33707
FLÜCKIGER. & GEROCK, J. E.: Contribution to the knowledge of Catha leaves, Pharmaceutical Journal of Transvaal, 1887, 18, 22108
FORSKAL, P.: Flora aegyptico-arabica, 1775, 6309
GREENWAY, P. J.: Khat. East African Agricultural Journal, 1947, 13, 9810
LELOUP, C. A.: Le Catha edulis, Thèse, Paris, 1890. Le Catha edulis et ses effets thérapeutiques, Bulletin général de théapeutique, 1890, 119, 21611
MANSKE, R. H. F. & HOLMES, H. L.: The Alcaloids, III, 348, New York, 195312
MUSTARD, M.J.: Ascorbic acid content of some miscellaneous tropical and sub-tropical plants and plant products, Food Research, 1952, 17, 3113
NAGAI, W. N. & KANAO, S.: über die Synthese der isomeren Ephedrine und ihrer Homologen -- Liebigs Annallen der Chemie, 1929, 470, 15714
PARIS, R. & FAUGERAS, G.: Application de l'électrophorèse à la séparation des alcaloides et des dérivés flavoniques chez quelques légumineuses, Annales pharmaceutiques françaises, 1955, 13, 35915
PARIS, R. & MOYSE, Mme H.: Caractérisation de la choline par chromatographie sur papier chez quelques plantes médicinales, Annales pharmaceutiques françaises, 1956, 14, 46416
PARIS, R. & VIEJO, J. P.: Identification des drogues simples et contrôle des médicaments végétaux par chromatographie sur papier, Annales pharmaceutiques françaises, 1955, 13, 42417
PERROT, Em.: Matières premières usuelles du règne végétal, Paris, Masson édit., 1942-194318
PARIS, R. & MOYSE, Mme H.: Essai de caractérisation du Kat ou Thé des Abyssins ( Catha edulis, Forssk.) (Célastracée) drogue recerement inscrite au tableau B, Annales pharmaceutiques françaises, 1957, 15, 8919
PETERS, D. W. A.: Khat: its History, Botany, Chemistry and Toxicology, The Pharmaceutical Journal, 1952, 115, 17, 3620
PLOUVIER, V.: Nouvelles recherches sur le québrachitol des Sapindacées et Hippocastanacées, le dulcitol des Célastracées et le saccharose de quelques antres families, 1949, 228, 1886. Comptes rendus de l’Académie des Sciences21
SMITH, S.: Nor-d pseudo-ephedrine, and alkaloids from Ephedra sp., Journal Chemical Society, London, 1928, 5122
STOCKMANN, R.: The active principles of Catha edulis, Pharmaceutical Journal, 1912, 89, 57723
STOCKMANN, R.: The pharmacological action of Catha edulis and its alkaloids, The Journal of Pharmacy and Experimental Therapeutics, 1912, 4, 25124
WOLFES, O.: über das Vorkommen von d. norisoephedrin in Catha edulis, Archiv der Pharmacie, 1930, 268, 8125
NATIONS UNIES, Secrétariat: Le Khat, Bulletin des stupéfiants, 1956, VIII, 4, pp. 7-1526
UNITED NATIONS, Secretariat: Khat, Bulletin on Narcotics, 1956, VIII, 4, pp. 6-13