Twenty-three studies met the inclusion criteria.
The quality of the studies was generally poor. In particular, almost all studies suffered from selection bias. Most studies included fewer than 35 patients and were restricted to patients with aortic regurgitation, thus it was only possible to calculate the specificity for a small proportion of studies. Blinding was rarely reported and, where reported, was often not carried out. A variety of reference standards were used.
Austin Flint Murmur (12 studies).
Three studies were excluded as they dealt with patients who had both aortic regurgitation and mitral stenosis; an Austin Flint murmur cannot occur in the presence of mitral stenosis. The sensitivity ranged from 25 to 100%, and was higher in patients with severe aortic disease (57 to 100%) than among those with mild to moderate disease (0 to 50%).
Hill sign (7 studies).
The sensitivity ranged from 0 to 100%. Four studies provided data on specificity, which ranged from 71 to 100%.
Corrigan pulse (6 studies).
The sensitivity ranged from 38 to 95%. Specificity was only reported in one study, which found a specificity of 16%.
Duroziez sign (5 studies).
The sensitivity ranged from 33 to 81% and the specificity from 35 to 100% (although the 100% estimate was based on only 2 patients).
The Quinke sign (2 studies), Traube sign (2 studies), de Musset sign (1 study), Mayne sign (1 study) and Rosenbach sign (1 study) were not included in the results as they were all investigated in fewer than three studies.