Carotid Intima-Media Thickness (CIMT): A Reproducibility Study

Carotid Intima-Media Thickness (CIMT): A Reproducibility Study\ud \ud Mindy Columbus, Brian Wagner, Emma Barinas-Mitchell\ud \ud Implementation of newer ultrasound technology for measuring subclinical atherosclerosis may increase the validity of measurement of carotid artery intima-media thickness (CIMT) cross-sectionally, but may prove to be a challenge for obtaining valid and reproducible progression data in existing follow-up studies. In the Department of Epidemiology Ultrasound Research Lab (URL), participants of the ERA JUMP study are returning for a five year follow-up visit for CIMT measurements to determine progression rates of subclinical atherosclerosis. A Toshiba 140A scanner was used for the baseline measurements and the maintenance service will end soon. There is a newer scanner available for use in the laboratory, and the question is whether the follow-up images that will be taken on this scanner, a Siemens Sonoline® Antares scanner, will provide comparable results to the baseline measurements in order to predict valid progression and not introduce error due to differences in machines. A reproducibility study was conducted to determine whether a Siemens Sonoline® Antares Doppler ultrasound scanner could be used in place of a Toshiba 140A Doppler ultrasound scanner to conduct follow-up visit measurements in the ERA JUMP study. A five year progression rate of ~0.04 mm (0.043) was estimated based on the literature. An average difference of 0.03 mm in CIMT between machines was determined a priori as an acceptable difference as we wanted to detect a difference between machines that was less than the estimated progression rate in order to assess reliability. The study recruited and evaluated 20 volunteers (85% women, age range: 24-77 years) who were scanned on each machine during the same visit by the same technician for between machine reliability. The mean CIMT values for this study population ranged from 0.488 to 1.039 mm. Reproducibility was evaluated on mean CIMT with Spearman and intraclass correlations and by the absolute value of the difference between replicate values. The mean difference in mean CIMT between the machines was -0.045 and the mean absolute difference was 0.052 mm. The mean difference in CIMT values between machines ranged from 0.077 to -0.256 mm, with 95% (19/20) of the differences being negative. This demonstrates systematic bias as the images read on the Toshiba machine were thicker (higher) than the Antares machine. This is likely due to the fact that the Antares scanner produces a crisper and clearer image than the Toshiba scanner demonstrating the advancement of digital technology with the newer scanner. Although the correlation for the mean CIMT between machines was high and agreed well, according to the Spearman correlation (r=0.93, p<.0001) and the intraclass correlation (ICC = (.02534/.0282651) = 0.896), there was still a greater difference between machines than the progression rate. Since the five year mean CIMT progression rate was projected to be ~0.04 mm and the mean absolute difference was 0.052 mm, a difference in progression may be difficult to detect, and in fact, it may appear that the participant’s mean CIMT has improved. Based on these results, we conclude, that in this case, the implementation of newer ultrasound technology may adversely affect the validity of progression data for follow-up studies that utilized the older technology for baseline measurements of CIMT. \ud \u