Central arterial pressure, measured close to the heart, may be of more patho-physiological importance than conventional non-invasive cuff blood pressure. The technique of applanation tonometry using SphygmoCor has been proposed as a non-invasive method of estimating central pressure. This relies on mathematically derived generalized transfer functions, which have been previously validated using invasive peripheral pressure measurements. We compared simultaneous estimates of central aortic pressure using this technique with those measured directly during the routine diagnostic cardiac catheterization of 30 subjects (age range 27-84 years), half of whom were aged 65 years or more. This was done by applanating the left radial artery and recording the non-invasive brachial cuff blood pressure to generate a central aortic pressure estimate, using the SphygmoCor radial transfer function. The comparative results were analysed using Bland-Altman plots of mean difference. SphygmoCor, on average, underestimated systolic central arterial pressure by 13.3 mmHg and overestimated diastolic pressure by 11.5 mmHg. The results were similar in patients aged under and above 65 years. Furthermore, non-invasively measured brachial pressures were seen to give an overall closer estimate of the central arterial pressure than the SphygmoCor system. The transfer function has been validated from invasively measured arterial pressures and the current use by the system of non-invasive measures may explain the discrepancies. However, age, drugs and arterial disease would also be expected to play a role.