Effect of contralateral severe stenosis or carotid occlusion on duplex criteria of ipsilateral stenoses: comparative study of various duplex parameters.

PURPOSE

This study compares the accuracy of various duplex parameters in grading ipsilateral carotid stenoses in patients with contralateral severe stenoses or occlusion.

METHODS

Four duplex criteria were correlated to arteriography in 356 carotid arteries in blind fashion: (1) standard criteria: a peak systolic frequency (PSF) of the internal carotid artery (ICA) of > or = 4 kHz was used to diagnose > or = 50% stenosis; (2) new criteria: a PSF of the ICA of > or = 4.5 kHz was used; (3) Fujitani criteria: a PSF of the ICA of > 4.5 kHz and an end-diastolic frequency of < 5.0 kHz was used; (4) internal carotid/common carotid artery (ICA/CCA) PSF ratio of > or = 1.5 was used.

RESULTS

The standard method overestimated 56 (16%) of 356 stenoses in contrast to 3% for the new method (p < 0.001), and this effect was most evident in the 50% to < 80% stenosis category (30%). The Fujitani method underestimated 97 (27%) of 356 stenoses, and the ICA/CCA ratio overestimated stenoses in 77 (22%) of 356. The overall exact correlation was 94%, 82%, 70%, and 75% for the new, standard, Fujitani, and ICA/CCA ratio, respectively. The kappa statistic and corresponding confidence intervals for the new method (kappa = 0.923, +/- 0.016) are significantly higher (p < 0.001) than those for the standard method (kappa = 0.760, +/- 0.027), the Fujitani method (kappa = 0.608, +/- 0.031), and the ICA/CCA ratio method (kappa = 0.642, +/- 0.051). The overall accuracy in diagnosing > or = 50% ipsilateral stenosis in the whole series was 85% for the standard method, 97% for the new method, 95% for the Fujitani method, and 81% for the ICA/CCA ratio. The new method was superior to the standard and ICA/CCA ratio methods (p < 0.001) and the Fujitani method (p = 0.024).

CONCLUSIONS

The presence of significant contralateral stenosis (> or = 50%) can lead to overestimation of ipsilateral stenosis if the standard criteria are used; however, this problem can be avoided by using a PSF of the ICA of > or = 4.5 kHz for the diagnosis of > or = 50% stenosis.