The British Journal of Psychiatry
Hypertensive Interactions Between Monoamine Oxidase Inhibitors and Foodstuffs


(1) 4.3 per cent. of patients prescribed monoamine oxidase inhibitors for depression at the Maudsley Hospital between January 1963 and June 1964 experienced sudden hypertensive attacks accompanied by a severe throbbing headache. The features of the syndrome have been studied in 25 patients, of whom four suffered an intracerebral haemorrhage and one died. Reasons are given for the delay before this condition was more widely recognized.

(2) Attacks were not due to drug idiosyncrasy as previously suggested, but were provoked by ingestion of certain foods. 17 patients had eaten cheese, and 2 had consumed a yeast extract, Marmite within 2 hours of the attack. Tests in 2 patients confirmed the association between Marmite and hypertension.

(3) Many patients ate cheese or Marmite before and after the hypertensive attacks without ill-effects. An examination was made of factors in the individual patient, the treatment and the composition of foodstuffs which might influence susceptibility.

(a) The patient: Comparisons between subjects and a control group of 50 hospital patients taking monoamine oxidase inhibitors showed that liability to hypertensive attacks was unrelated to age, sex, personality or predisposition to headache. Isoniazid clearance tests in 7 patients showed that individual susceptibility was unconnected with rate of drug acetylation. The threshold at which headache appeared during hypertension differed between patients and rises of up to 70 mm. Hg systolic pressures were sometimes asymptomatic. The number of patients experiencing such a rise in blood pressure during treatment with monoamine oxidase inhibitors may be much higher than previously supposed.

(b) The treatment:

i. The incidence of hypertensive attacks due to food substances in hospital patients treated with tranylcypromine was five times higher than in those taking phenelzine, and the difference could not be accounted for by the extent of their use. The increased risk associated with tranylcypromine is supported by national prescribing figures and by results from animal experiments. The increased risk is probably due to the drug's efficiency at inhibiting intestinal monoamine oxidase, and not to its amphetamine-like properties.

ii. Tests on the blood pressure of two patients given yeast extracts under controlled conditions, suggest that intestinal monoamine oxidase was cumulatively inhibited but recovered rapidly. The blood pressure effects of amines in foodstuffs were probably determined by these factors rather than by raised tissue catecholamine levels, since V.M.A. excretion in three patients was normal.

(c) The foodstuff: Previous publications have shown that cheese and yeast extracts vary considerably in their tyramine content, which cannot be safely predicted by appearance, flavour or method of manufacture. Cheddar cheese is usually responsible for the hypertensive crises, but only 4 out of 14 samples tested contained sufficient tyramine. Marmite contains more tyramine than other yeast extracts, but sufficient quantities are only likely to be consumed as a beverage.

(4) A brief consideration of the historical and discarded concept of alimentary toxaemia and of other relevant facts suggests that the hypertensive crisis and its precipitation by cheese and other foodstuffs could have been anticipitated before the introduction of monoamine oxidase inhibitors into clinical medicine.