Middle cerebral artery pulsatility index: reliability at different sampling sites

Objectives To analyze the interobserver reliability of measurement of the middle cerebral artery (MCA) pulsatility index (PI) at two different sampling sites. Methods This study included 100 consecutive singleton pregnancies between 24 and 40 weeks with normal fetal growth. The PI was calculated by two independent operators at proximal and distal sampling sites of the near-field MCA. Reliability analyses were performed between observers at each sampling site by means of the intraclass correlation coefficient (ICC) for agreement. Differences between observers were explored and agreement limits calculated by means of the Bland-Altman test. Results Satisfactory flow velocity waveforms were obtained successfully in each fetus at both sampling sites. Peak systolic, end-diastolic and time-averaged maximum velocities were significantly higher at the proximal compared with the distal sampling site. Conversely, PI was significantly higher at the distal compared with the proximal site. ICCs for PI were 0.3 and 0.33 at the proximal and the distal sampling sites, respectively. The 95% interval of the PI differences between observers were +0.91 and - 1.14 at the proximal and + 1.03 and -1.08 at the distal sampling sites. In about 30% of the cases the PI difference between observers was greater than 0.5 at both sampling sites. Conclusions Moderate interobserver reliability in the measurement of end-diastolic and time-averaged maximum MCA flow velocities results in limited agreement of the PI calculation at both proximal and distal sampling sites. These results may preclude its clinical applicability. Copyright (c) 2006 ISUOG. Published by John Wiley Sons, Ltd.