OBJECTIVE: To define the value of tilt testing and hte additional yield of drug provocation over prolonged baseline tilt in different patient subgroups. (Many different protocols are in use for head-up tilt testing in heterogeneous groups of patients. Not all patients in reported series have recurrent syncope, and there is often a wide age range and a variable incidence of structural heart disease.) DESIGN: In a prospective study, baseline 60 degrees head-up tilt testing was undertaken for 45 minutes, initially without drug provocation. Patients who remained symptom free were given intravenous isoprenaline (isoproterenol) and further tilting or edrophonium (10 mg bolus) during tilt, in an order determined randomly before the start of the test. If they were symptom free after the first drug, they were given the other drug. A positive test was recorded when syncope or pre-syncope occurred with a rapid fall (> 30%) in blood pressure. The impact on tilt result of the type of symptoms, presence of significant structural heart disease (SHD), presence of a non-cardiovascular cause of sudden diminished consciousness (SDC), and age was then assessed by subgroup analysis. PATIENTS: 145 patients (73 female, mean age 51 (25), range 8-94) with one or more episodes of pre-syncope or syncope. RESULTS: 39 patients (27%, 21 female, age 49 (25) years) had positive tests and 106 (73%, 52 female, age 52 (25) years) negative tests. 27 (69%) had a positive test during baseline tilt at 20.5 (10.8) minutes, five (13%) with isoprenaline infusion, and seven (18%) with edrophonium bolus. Patients with recurrent syncope rather than single syncopal episodes or single or recurrent pre-syncope were more likely to have a positive tilt test (41% v 17%, P < 0.005) and patients with SHD or SDC (69/14 patients) were much less likely than patients without (16% v 42%, P < 0.0001). The yield of positive tests was similar if patients were below (26%) or above (27%) the mean age (50 years). When multiple factors were combined, the yield ranged from 0% for 21 patients under 50 years with SHD or SDC and without recurrent syncope to 73% in 11 patients over 50 years with recurrent syncope and no SHD or SDC. The additional yield in subgroups over 45 minute baseline tilt (70 (11)%) of isoprenaline (13 (10)%) was similar to that of edrophonium (17 (8)%, P = NS), but six (50% of those who were drug positive) patients required a second drug to produce a positive result (two with isoprenaline second, four with edrophonium second). CONCLUSIONS: Head-up tilt testing in a heterogeneous population has a low yield. Simple clinical characteristics define the type of patient who is likely to have a positive tilt test and the patient who is not and in whom other investigations should receive priority. The great majority of positive tests will occur during prolonged baseline testing if this is used. Isoprenaline and edrophonium produced similar additional yields of positive tests.