OBJECTIVE: To assess the degree of variability of transient myocardial ischaemia during daily life in patients with coronary artery disease, which could confound the interpretation of trials of the therapeutic effects of anti-ischaemic agents. DESIGN: Prospective method evaluation. SETTING: Tertiary referral centre, outpatient clinic. PATIENTS: Patients with stable angina, confirmed coronary artery disease, and a positive treadmill exercise test for ischaemia. Patients were not preselected on the basis of prior documented transient ischaemia during ambulatory ST segment monitoring. INTERVENTIONS: A simulated drug-study with 4 monitoring phases in 16 subjects. To minimise variability in ischaemic activity, patients underwent weekly 48 hour ambulatory ST segment monitoring outside hospital off all prophylactic therapy on the same weekdays for 4 weeks. MAIN OUTCOME MEASURE: Variability in the frequency and duration of transient myocardial ischaemia. RESULTS: There was marked variability in both ischaemic activity and mean duration of ischaemia in patients with confirmed ischaemia, the greatest degree of variability being between patients and from day to day within weeks within patients, with a further contribution to variability being noted between fortnights within patients. CONCLUSIONS: Despite assessment off all therapy and an adequate period of monitoring (48 hours) with small intervals between monitoring periods (5 days), marked variability in ischaemic activity was noted, and regression towards the mean was clearly shown. Ambulatory ST segment monitoring outside hospital is not a reliable method for assessing the therapeutic effects of anti-ischaemic agents.