Design-Cross sectional study of patients prescribed thyroxine long term (n = 11), patients with thyrotoxicosis studied at presentation (n = 23), compared with controls (n = 25); longitudinal study of patients with thyrotoxicosis studied at presentation and serially after beginning antithyroid drug treatment (n = 23). Methods-24 h ambulatory monitoring of pulse and blood pressure, echocardiography, forearm plethysmography, and autonomic function tests. Results-Long-term thyroxine treatment in doses that reduced serum thyrotrophin to below normal had no effect on blood pressure, heart rate, left ventricular systolic function or stroke volume index, but was associated with an 18-4% increase in left ventricular mass index (mean (SEM) 101'9 (3.09) g/m' v controls 86-1 (4.61), P < 0.01). Thyroxine treatment, like thyrotoxicosis, had no effect on tests of autonomic function. Untreated thyrotoxicosis resulted in pronounced changes in systolic and diastolic blood pressure and an increase in heart rate during waking and sleep. Patients with thyrotoxicosis at presentation had an increase in left ventricular systolic function (ejection fraction 70 5 (1P66)% v 65 4 (1.79), P < 0 01; fractional shortening 40 4 (1.54)% v 35 6 (1.46), P < 0.01), increased stroke volume index (45.9 (2.4) mu/m2 v 36-6 (1.7), P < 0.001), and an increase in forearm blood flow, and decrease in vascular resistance. They had a similar degree of left ventricular hypertrophy to that associated with thyroxine treatment (99.3 (4.03)

In contrast to other studies we have failed to confirm cardiovascular dysfunction in GH deficient hypopituitary adults. Indeed, cardiovascular protection may be conferred on this group by the lower BP levels. Although a reduction in hydrocortisone dose was well tolerated in all patients, it appeared to confer no additional clinical benefit over the 3-month study period. In view of the conflicting data on cardiovascular function in hypopituitary patients, further prospective mortality studies are required in patients with adult GH deficiency.

A new time related method of analysing the sinoaortic baroreceptor heart rate reflex, which determines reflex latency as well as sensitivity, was used to compare the results obtained with a phenylephrine ramp method (P) with those obtained using the whole of phase IV of Valsalva (V1) and using the phase IV systolic blood pressure overshoot alone (V2). Twenty five subjects with large ranges of age and resting blood pressures were studied. Each performed two standardised Valsalva manoeuvres and received three bolus injections of phenylephrine sufficient to cause transient pressor responses of 20-30 mmHg. Mean sensitivity values with P (6.2(3.5) ms.mmHg-1) were greater than those with V1 (4.6(2.3) ms.mmHg-1, p less than 0.001) and less than V2 (7.8(4.0) ms.mmHg-1, p less than 0.001). However, linear regression analysis showed a correlation of P with V1 (r = 0.76, p less than 0.0001) and with V2 (r = 0.80, p less than 0.0001). Reflex latency with P (1084(427) ms) was less than V1 (2416(423) ms, p less than 0.0001) and V2 (1504(441) ms, p less than 0.0005). Reflex sensitivity results obtained using phase IV of Valsalva's manoeuvre are proportionately related to phenylephrine results, but large errors were introduced into the absolute values obtained when relatively small changes were made to the method of analysis.