• Nu S-Au Găsit Rezultate

Carotid intima-media thickness and hemodynamic parameters: reproducibility of manual measurements with Doppler ultrasound

N/A
N/A
Protected

Academic year: 2022

Share "Carotid intima-media thickness and hemodynamic parameters: reproducibility of manual measurements with Doppler ultrasound"

Copied!
8
0
0

Text complet

(1)

Original papers

DOI: 10.11152/mu.2013.2066.172.ci-m

Abstract

Aims: To evaluate the carotid ultrasound intra- and interobserver agreements in a common clinical scenario when making manual measurements of the intima-media thickness (IMT) and peak systolic (PSV) and end diastolic (EDV) velocities in the common (CCA) and the internal carotid (ICA) arteries. Material and methods: Three different experienced operators per- formed two time-point carotid ultrasounds in 21 patients with cardiovascular risk factors. Each operator measured freehand the CCA IMT three consecutive times in each examination. The CCA and ICA hemodynamic parameters were acquired just once.

For our purpose we took the average (IMTmean) and maximum (IMTmax) IMT values. Quantitative variables were analyzed with the t-student, and ANOVA test. Agreements were evaluated with the Intraclass Correlation Coefficient (ICC). Results:

IMTmean intraobserver agreement was better on the left (ICC: 0.930-0.851-0.916, operators 1-2-3) than on the right (ICC:

0.789-0.580-0.673, operators 1-2-3). IMTmax agreements (Left ICC: 0.821-0.723-0.853, operators 1-2-3; Right ICC: 0.669- 0.421-0.480, operators 1-2-3) were lower and more variable. Interobserver agreements for IMTmean (ICC: 0.852-0.860;

first-second ultrasound) and IMTmax (ICC: 0.859-0.835; first-second ultrasound) were excellent on the left, but fair-good and more variable on the right (IMTmean; ICC: 0.680-0.809; first-second ultrasound; IMTmax; 0.694-0.799; first-second ultra- sound). Intraobserver agreements were fair-moderate for PSVs and good-excellent for EDVs. Interobserver agreements were good-excellent for both PSVs and EDVs. Overall, 95% confidence intervals were narrower for the left IMTmean and CCA velocities. Conclusions: Intra and interobserver agreements in carotid ultrasound are variable. In order to improve carotid IMT agreements, IMTmean is preferable over IMTmax.

Keywords: ultrasonography, Doppler, carotid intima-media thickness, observer variation, reproducibility

Carotid intima-media thickness and hemodynamic parameters:

reproducibility of manual measurements with Doppler ultrasound

Juana M Plasencia Martínez, José M García Santos, Maria L Paredes Martínez, Ana Moreno Pastor

Radiology Department. University General Hospital Morales Meseguer. Murcia, Spain

Received 26.01.2015 Accepted 10.04.2015 Med Ultrason

2015, Vol. 17, No 2, 167-174

Corresponding author: Juana M Plasencia Martínez

Servicio de Radiodiagnóstico. Hospital General Universitario Morales Meseguer.

C/ Marqués de los Vélez s/n. 30008 Murcia. España

Phone: +34620402970 Fax number: +34968239975 E-mail: [email protected]

Introduction

Carotid atherosclerosis is considered to be repre- sentative for systemic atherosclerosis [1,2]. Due to the carotid accessibility, ultrasonography (US) examination is a safe technique, being recommended by the Ameri- can Society of Echocardiography (ASE) to determine the vascular age and the risk for cardiovascular events [3-

6]. Together with other imaging techniques, the results offered by US are useful for setting primary prevention therapy [7]. US could estimate the cardiovascular risk by detecting atheroma plaques and measuring intima-media thickness (IMT) at the level of the common carotid ar- tery (CCA) [7-9]. On the other hand, parameters such as peak systolic (PSV) and end diastolic (EDV) velocities assess the hemodynamic effects of stenosis and are rel- evant for management decisions [10]. However, carotid US is often performed by sonographers, in and out of Radiology departments, with different levels of training and experience and different approaches (manual, semi- automated or automated), whose reliability is many times simply assumed. But a manual subjective measurement, which is the most common approach in a clinical sce- nario, might be associated with significant variability.

The reported variability of IMT measurements has been different among studies [4,6,11,12-14] and, even more

(2)

Fig 1. B-mode (A) and Doppler (B) ultrasound in the left distal common carotid artery. Intima-media thickness. Both intima- media (empty arrowhead) and media-adventitia (arrowhead) show up as echogenic interfaces at ultrasound.

relevant, the statistical assessment has been also quite diverse, most frequently focused on correlations than on agreements [15]. That has also been the case of hemo- dynamic parameters [16,17]. Therefore, reproducibility studies are still needed, especially if carotid US is going to be used as a biomarker of cardiovascular risk. Accord- ingly, our purpose was to evaluate intra and interobserver agreements in carotid US, reproducing a usual clinical scenario. We hypothesize that IMT and hemodynamic parameters reproducibility may be not negligible.

Material and methods Patients

We wanted to build a sample made up of patients with expected carotid atherosclerosis. These subjects would account for the common clinical scenario in which man- ual measurements are performed and cardiovascular risk estimated. Their carotid walls are usually irregular and a strong subjective visual component needs to be consid- ered when manual measurements are attempted. For that purpose, we took advantage of subjects of a cardiology sample in our Institution who had undertaken computer tomography (CT) for calcium scoring and coronary ar- tery angio-CT during the last 11 months. CT had been performed in atypical chest pain and when electrocardio- gram, echocardiography, or stress tests were uncertain, non-interpretable, impossible to be performed, or dis- cordant between them or with laboratory testing. Except for one patient with right coronary artery chronic occlu- sion, the remaining subjects had no known coronary dis- ease. Carotid stenting or surgical interventions were ex- clusion criteria, but, finally, no patient was excluded. The final sample was made up of 21 patients (14 male; mean age 61.2 ± 9.4 years, range 42 - 77 years, women 44 - 77 years, men 42 - 76 years). The study was approved by the Institutional Ethics Committee. Before US, all subjects were interviewed and informed on the characteristics and objective of the study and signed the written informed consent. Each patient received a report of his/her first US examination in the second appointment.

Doppler US

All exams were performed with the same ultrasound scanner and probe, Philips IU22, L 9-3 MHz (Philips Healthcare, Eindhoven, The Netherlands), between Feb- ruary and June, 2012. In order to reproduce a common real technical scenario, we first wanted to introduce dif- ferent levels of experience between operators. Therefore, each patient was individually and independently explored by three operators (Op) with 7, 2, and 1 years experience in general and Doppler US; the first one (Op1) was a cer- tified radiologist while the other two were, respectively,

an US trained radiology resident (Op2) and a radiology resident still under US training (Op3). Each operator car- ried out two exams in each patient, two different days apart, not before 30 days from the first exam, and not be- yond 10 weeks between both exams. They were always blinded to the other operators’ results, and, when per- forming the second examination, also to their previous results. Performing US, all operators followed the rec- ommendations of the ASE [9]. Accordingly, the patient stayed in supine position, the neck slightly hyperextend- ed and rotated to the opposite side. Also looking for that common technical scenario, the operators only agreed on what the IMT limits were and how to obtain the IMT and hemodynamic measurements. Otherwise, they were free to sample Doppler signal with the incidence angles they considered more appropriate, provided it was ≤60º and not beyond 1cm from the carotid bifurcation, and to assess IMT whatever time of the cardiac cycle and with the zoom level they felt more comfortable. All data and zoom levels were recorded in our local Picture Archiving and Communication System (PACS) and archived in a Microsoft Office Excel database (Windows 2010. Micro- soft Office Corporation. Redmond, Washington, EEUU).

IMT was considered as the maximum distance be- tween the intima and adventitia internal edges, at the dis- tal centimeter of the posterior CCA wall. When detect- ed, the operator avoided the atheroma plaques for IMT measurement, but not for the hemodynamic assessment.

Each operator successively froze three CCA longitudi- nal images, identified the thickest wall, and measured the IMT freehand in each image. Along that process, the operators were instructed to avoid looking at the quanti- tative values in the screen. The averaged thickness (IMT-

(3)

mean) and the thickest width (IMTmax) were used for the analysis. Later on, they got the PSV and EDV in the distal CCA and proximal ICA on both sides, as close as possible to the carotid bifurcation, always keeping an an- gle ≤60º, and using the spectrum with the best spectral window (fig 1). The carotid bulb was avoided.

Statistical analysis

Variables in the Excel tables were exported to SPSS version 17.0 for Windows (IBM, Armonk, NY). They were presented as mean ± standard deviation. Normal distribution was checked by the Kolmogorov-Smirnov test. The two-tailed Student t test was used for any right and left comparisons while differences in IMT and hemo- dynamic parameters between the three operators were analyzed with the ANOVA test and Bonferroni correction for multiple comparisons. Finally, we assessed the intra and interobserver agreement for IMT and hemodynamic parameters. All agreement analysis was made with the Intraclass Correlation Coefficient (ICC) with 95% con- fidence intervals (95%CI). Agreement levels were clas- sified as poor (<0.20), fair (0.21-0.40), moderate (0.41- 0.60), good (0.61- 0.80) and excellent (>0.80). Statistical significance was established for a p value < 0.05.

Results

Demographic data are shown in table I.

a. Right to left differences

a.1 Intima-media thickness. The averaged IMTmean and IMTmax were significantly thicker on the left (IMT- mean 0.9271 ± 0.2593 vs 0.8404 ± 0.1736, P < 0.001; IMT- max 1.0050 ± 0.2910 vs 0.9195 ± 0.1895, P < 0.015). On an operator basis, IMTmean and IMTmax were also thicker on the left, but differences were only significant in the second examination of operators 1 and 2 (P = 0.025 and P = 0.031 for IMTmean; P = 0.041 and P = 0.016 for IMTmax).

a.2 Velocities. PSV and EDV did not show right to left differences, excluding a higher PSV in the right CCA

(operator 3; second examination), and a higher EDV in the left ICA, (operators 2 and 3; first examination).

b. Differences between operators (table II)

b.1 Intima-media thickness. All IMT magnitudes of operator 2 were different from the other two sonog- raphers measurements (fig 2). Operator 2 always used zoomed images while operators 1 and 3 zoomed images in 38% and 31% of cases, respectively. The operator 2 zoom level was also higher (≥ 1.7x in 69% of cases vs.

11% and 2% in operators 1 and 3, respectively).

b.2 Velocities. Only right ICA EDV in the second ex- amination was significantly different between the three operators, but that difference lost significance when mul- tiple comparison analysis was performed.

c. Agreement analysis (tables III and IV; 95%CI are only provided in the tables).

c.1 Intima-media thickness. The intraobserver IMT- mean agreement was good to excellent on the left (Op1, 0.930; Op2, 0.851; Op3 0.916), and moderate (Op2, 0.580) to good (Op1, 0.789; Op3, 0.673) on the right.

The IMTmax agreement was lower (Left: Op1, 0.821;

Op2, 0.723; Op3, 0.853. Right: Op1, 0.669; Op2, 0.421;

Op3, 0.480). 95%CI were only narrow for IMTmean, es- pecially on the left and for operators 1 and 3. Operator 2 nearly always obtained the lowest agreement and the widest confidence intervals (table III).

The interobserver agreement for IMTmean and IMT- max was excellent on the left in both examinations (First IMTmean: 0.852; first IMTmax: 0.859. Second IMT- mean: 0.860; second IMTmax: 0.835), and worse on the right (First IMTmean: 0.680; second IMTmean: 0.809.

First IMTmax: 0.694; second IMTmax: 0.799). 95% CI intervals were also wider on the right (table IV).

Fig 2. Mean intima-media thickness (IMTmean) measure- ments. Differences between operators (Op). In all cases, the IMTmean and maximum IMT (not shown) for the second op- erator were significantly lower. E1: first examination; E2: sec- ond examination.

Table I. Demographic characteristics of the sample (21 patients).

N (%) Previous coronary event yes 3 (14.3)

Gender male 14 (66.7)

female 7 (33.3)

Smoker yes 5 (23.8)

Diabetes yes 5 (23.8)

Total hypercholesterolemia yes 15 (71.4)

HDL-Cholesterol high 2 (11.1)

low 2 (11.1)

loss 3 (14.3)

Arterial hypertension yes 17 (81) N – number of patients

(4)

c.2 Velocities. The intraobserver trends of operators 1 and 3 were similar and differed from operator 2. For both operators, PSV agreement in the CCA was fair on the left (Op1, 0.305; Op3, 0.376) and moderate on the right (Op1, 0.499; Op3 0.485). PSV agreement was moderate in both ICA (Left: Op1, 0.455; Op3, 0.552. Right: Op1, 0.493; Op3, 0.546). Agreement improved with EDV, be- ing moderate to good in the CCA (Left: Op1, 0.683; Op3, 0.607. Right: Op1, 0.596; Op3, 0.731) and good to ex- cellent in the ICA (Left: Op1, 0.803; Op3, 0.671. Right:

Op1, 0.740; Op3, 0.812). Operator 2 agreements were more anarchic, tending to be better with EDV, but with wide 95% CI in all cases (fig. 3, table III).

Interobserver agreements with the PSV (Left, first exam: 0.885; second exam: 0.864. Right, first exam: 0.936;

second exam 0.943) and EDV (Left, first exam: 0.893; sec- ond exam: 0.887. Right, first exam: 0.906; second exam 0.872) were excellent in the CCA, and good to excellent in the ICA (Left, first exam: 0.854; second exam: 0.796.

Right, first exam: 0.842; second exam: 0.840). 95% CI were narrow in the CCA and wider in the ICAs (fig 4, table IV).

Table II. Differences between operators

Operator 1 Operator 2 Operator 3 p*

IMTmean L-CCA US 1 0.952 (0.269) 0.688 (0.136) 0.886 (0.228) 0.001

US 2 0.917 (0.198) 0.704 (0.145) 0.937 (0.286) 0.002

R-CCA US 1 0.844 (0.148) 0.664 (0.136) 0.844 (0.150) < 0.001

US 2 0.847 (0.155) 0.637 (0.129) 0.842 (0.163) < 0.001

IMTmax L-CCA US 1 1.024 (0.276) 0.766 (0.148) 0.964 (0.247) 0.002

US 2 1.020 (0.233) 0.773 (0.165) 1.020 (0.305) 0.002

R-CCA US 1 0.933 (0.168) 0.736 (0.157) 0.921 (0.163) < 0.001

US 2 0.926 (0.163) 0.686 (0.137) 0.928 (0.204) < 0.001

PSV L-CCA US 1 103.476 (18.057) 99.491 (23.140) 103.714 (27.103) NS

US 2 105.550 (17.548) 100.775 (16.956) 97.890 (17.360) NS

R-CCA US 1 109.571 (23.750) 96.291 (22.510) 103.581 (26.935) NS

US 2 99.000 (16.673) 103.853 (24.861) 106.390 (24.747) NS

L-ICA US 1 93.857 (27.623) 93.081 (28.173) 91.467 (26.899) NS

US 2 95.000 (23.149) 91.605 (25.513) 90.130 (19.585) NS

R-ICA US 1 90.381 (26.080) 94.267 (34.978) 91.576 (25.303) NS

US 2 90.737 (19.686) 93.330 (24.746) 89.250 (23.440) NS

EDV L-CCA US 1 27.095 (8.264) 26.395 (9.962) 27.000 (12.681) NS

US 2 26.850 (6.393) 24.885 (6.523) 23.800 (6.711) NS

R-CCA US 1 29.476 (10.796) 26.955 (9.251) 26.274 (8.134) NS

US 2 27.263 (5.772) 25.100 (7.713) 26.540 (6.796) NS

L-CCA US 1 28.143 (9.345) 25.491 (7.044) 24.045 (10.352) NS

US 2 28.600 (9.116) 25.380 (9.009) 28.280 (10.997) NS

R-ICA US 1 26.095 (8.166) 21.814 (7.122) 27.786 (8.637) NS

US 2 29.158 (9.179) 22.840 (6.247) 27.380 (8.553) 0.04†

Data are shown as mean (mm) ± standard deviation (between brackets).

IMTmean: averaged intima-media thickness; IMTmax: maximum intima-media thickness; PSV: Peak systolic velocity; EDV: End diastolic velocity; CCA: Common carotid artery; ICA: Internal carotid artery; R: Right; L: Left; US1: first examination; US2: second examination;

NS: non-significant.

* Bonferroni test: differences significant for operator 2.

† Difference was non-significant with Bonferroni test.

Fig 3. Intraobserver agreements for peak systolic (PSV) and end diastolic velocities (EDV) in common (CCA) and inter- nal carotid (ICA) arteries according to the intraclass correla- tion coefficient (ICC). Comparison between the first and the second examination of each operator. Agreement progressively increases from the left to the right CCA, and from PSV to EDV in operators (Op) 1 and 3. 95% confidence intervals (CI): black segments superimposed on the bars. Quantitative values of ICC and 95% CI are included in the manuscript and in table III.

(5)

Discussions

According to our results, the intra and interobserver agreements with IMT (especially IMTmean on the left) and hemodynamic parameters (especially in the CCA) went from good to excellent, but they were variations between arteries and operators, generally with wide con- fidence intervals.

Our main purpose was to assess the reproducibility of manual carotid US measurements in patients with cardio- Fig 4. Interobserver agreement for peak systolic (PSV) and end diastolic velocities (EDV) in common (CCA) and internal ca- rotid (ICA) arteries according to the intraclass correlation coef- ficient. Agreements are generally excellent but decrease from the CCA to the ICA, and from the right (R) to the left (L), both for the PSV and the EDV. Confidence intervals (black segments superimposed on bars) increase at the same time. Quantitative values of the intraclass correlation coefficient and the 95% CI are included in the manuscript and in table IV. E1: first exami- nation; E2: second examination.

Table III. Intraobserver agreement

Operator 1 Operator 2 Operator 3

L-CCA IMTmean 0.930 (0.822-0.972) 0.851 (0.622-0.941) 0.916 (0.789-0.967)

IMTmax 0.821 (0.603-0.925) 0.723 (0.423-0.881) 0.853 (0.666-0.939)

R-CCA IMTmean 0.789 (0.452-0.919) 0.580 (0.060-0.834) 0.673 (0.174-0.871)

IMTmax 0.669 (0.320-0.858) 0.421 (0.140-0.722) 0.480 (0.060-0.756)

L-CCA PSV 0.305 (-0.160-0.660) 0.305 (-0.160-0.660) 0.376 (-0.067-0.696)

EDV 0.683 (0.356-0.861) 0.648 (0.299-0.844) 0.607 (0.223-0.828)

R-CCA PSV 0.499 (0.070-0.772) 0.558 (0.165-0.798) 0.485 (0.066-0.758)

EDV 0.596 (0.206-0.822) 0.863 (0.679-0.945) 0.731 (0.413-0.890)

L-ICA PSV 0.455 (0.028-0.742) 0.710 (0.401-0.874) 0.552 (0.156-0.795)

EDV 0.803 (0.567-0.917) 0.554 (0.160-0.796) 0.671 (0.335-0.855)

R-ICA PSV 0.493 (0.063-0.769) 0.344 (-0.092-0.670) 0.546 (0.148-0.792)

EDV 0.740 (0.441-0.891) 0.504 (0.091-0.769) 0.812 (0.584-0.921)

Data correspond to intraclass correlation coefficient and 95% confidence interval (between brackets).

IMTmean: averaged intima-media thickness; IMTmax: maximum intima-media thickness; PSV: Peak systolic velocity; EDV: End diastolic velocity; CCA: Common carotid artery; ICA: Internal carotid artery; R: Right; L: Left.

Table IV. Interobserver agreement

CCI 95%CI

L-CCA IMTmean US1 0.852 0.694-0.935

US2 0.860 0.705-0.940 IMTmax US1 0.859 0.709-0.939 US2 0.835 0.653-0.930

R-CCA GIM med US1 0.680 0.338-0.860

US2 0.809 0.590-0.921 IMTmax US1 0.694 0.367-0.866 US2 0.799 0.568-0.916

L-CCA PSV US1 0.885 0.762-0.950

US2 0.864 0.713-0.942

EDV US1 0.893 0.774-0.954

US2 0.887 0.761-0.952

R-CCA PSV US1 0.936 0.868-0.972

US2 0.943 0.878-0.976

EDV US1 0.906 0.793-0.962

US2 0.872 0.725-0.947

L-ICA PSV US1 0.854 0.699-0.936

US2 0.796 0.570-0.913

EDV US1 0.850 0.689-0.934

US2 0.803 0.584-0.916

R-ICA PSV US1 0.842 0.673-0.931

US2 0.840 0.657-0.934

EDV US1 0.835 0.658-0.928

US2 0.742 0.447-0.893 Agreements correspond to intraclass correlation coefficient (ICC) with 95% confidence interval (CI).

IMTmean: averaged intima-media thickness; IMTmax: maximum intima-media thickness; PSV: Peak systolic velocity; EDV: End diastolic velocity; CCA: Common carotid artery; ICA: Internal ca- rotid artery; R: Right; L: Left; US1: first examination; US2: second examination.

(6)

vascular risk factors. In a previous study whose sample was classified in cardiovascular risk quartiles according to IMT, minimal differences in magnitude moved pa- tients to distant quartiles. In that paper, being included in the highest or the lowest quartiles depended on IMT differences of just 0.3 mm [18]. Therefore, variability of carotid ultrasound IMT magnitudes might make us- ing it as a biomarker of cardiovascular risk questionable.

Moreover, a recent large meta-analysis has questioned the prognostic utility of IMT as a risk biomarker and has argued that the extent of carotid plaque, when compared to IMT, showed a greater accuracy for the prediction of future coronary ischemic events [19]. In this scenario, the analysis of carotid US agreements becomes relevant.

A number of previous papers have analyzed the re- producibility of manual and automatized carotid IMT measurements [5,11,12,14,20,21]. However, only a few made just a correlation analysis and agreement was not actually assessed [11,12,20]. For other researchers, cor- relation based on coefficients of variation or difference of means were acceptable [14,21] or very high [5], but the interobserver agreement was not studied in the latter.

Unlike those studies, we assessed intra and interobserver agreements with the intraclass correlation coefficient and 95% CI, and agreement, although overall acceptable, was variable. IMTmean was always more reproducible than IMTmax, probably because it is just one measurement, it may be more variable than IMTmean [9]. IMTmax is the normal approach in our department and has also been the reference parameter in some studies [20]. However, IMTmean (averaged three IMT measurements in the distal CCA centimeter) is recommended by the ASE for IMT manual quantification [9] and, as our results cor- roborate, less variable [15,21].

IMT agreement in this study was always better on the left. Right-to-left thickness differences might be the first reason for agreement inconsistencies [5]. Though not always the case [22], some reports have suggested that atheromatosis might be more severe on the left CCA [23-25]. Moreover, a stiffer and less elastic wall would make it less variable during the cardiac cycle. In fact, the effect of cardiac cycle on an IMT agreement can be another possible source of variation [25]. Accordingly, a thicker and less changing left carotid wall might allow the operator to be more accurate when placing the cali- pers at the IMT edges. Further supporting that hypothe- sis, when reproducibility has been reported equal in both sides, the consistence of measurements was higher on the side where the carotid wall was thicker [5]. And lev- els of agreement were better bilaterally when IMT was also bilaterally increased [11]. However, other authors have reported a negative relation between agreement and

thickness [14,26] and taking into account that the higher agreement on the left has been also reported with semi- automated software [27], other factors than thickness could influence agreement and laterality. Accordingly, the operator’s position on the left may change the angle of insonation or the applied pressure [28,29]. And zoom- ing or experience might also have to be considered in the background of inconsistencies. In our case, the operator systematically enlarging the image also showed different degrees of agreement and lower thickness values, which is consistent with other reports [9]. On the other hand, our results do not suggest an important impact regarding the operator’s experience. Specific and fast training pro- tocols may improve IMT agreement even with inexperi- enced operators [30] suggesting that highly experienced sonographers might not significantly differ from less trained operators. However, more dedicated designs are required to achieve more reliable conclusions on zoom- ing or experience, even more considering that experience has been previously related with disagreements [29,31].

Regarding the hemodynamic parameters, our intra and interobserver agreements were slightly and clearly worse, respectively, than others previously reported [32], always with wide confidence intervals. Intraobserver concordance was higher with EDV than PSV, perhaps related to the narrower range of EDV magnitudes. How- ever, we cannot easily explain why the interobserver was better than the intraobserver agreement. Maybe, it was partially an effect of the small sample, but, once again, we cannot rule out the impact of those previously com- mented factors, related to technical, anatomical, physi- ological and pathological reasons, which, for the hemo- dynamic parameters, could be even more relevant.

Our study has several limitations. First, though we wanted to involve more patients, their special character- istics and the number of examinations in different days made it difficult to get a larger sample. As a result, the small size lowers our statistical power. However, repro- ducibility studies may be attempted with at least 20 sub- jects [33,34] and, with that sample size, we can assume a 20% error margin [35], which is acceptable for the pre- liminary objective of this work. Moreover, a small sample avoids the learning effect of successive carotid explora- tions identified in other studies [29]. Furthermore, our ef- fective sample size for a complete two time-point analy- sis was not significantly different to other two time-point agreement analysis in which the elderly with expected cardiovascular disease were included [14,36]. Second, the technical simplicity and the operator’s autonomy could have introduced unpredictable sources of variability. The foreseeable impact of subjectivity in manual quantifica- tion makes reasonable thinking that automated techniques

(7)

improve accuracy of measurements [9,19,37,38]. Several studies have shown that automated software improves re- producibility [4,5,9]. However, not all studies agree [39]

and, on the other hand, automated measurements devia- tions would have to be manually corrected, involving var- iability once again. In addition, our purpose was to repro- duce a common technical scenario in clinical practice and, in our opinion, more than a real limitation, our manual approach adds a practical value to our results.

In conclusion, when carotid ultrasound is performed in patients with cardiovascular risk factors, intra and in- terobserver agreements are variable. In order to improve carotid IMT agreements, IMTmean is preferable over IMT max, but, even in that case, variability is not negli- gible, probably depending on technical and operator fac- tors that have to be addressed in further investigations.

Conflict of interest: none Acknowledgments

We are indebted to Andrés Carrillo MD, for his valu- able help with the statistical analysis.

References

1. Nagai Y, Kitagawa K, Sakaguchi M, et al. Significance of earlier carotid atherosclerosis for stroke subtypes. Stroke 2001; 32: 1780-1785.

2. Wofford JL, Kahl FR, Howard GR, McKinney WM, Toole JF, Crouse JR 3rd. Relation of extent of extracranial carotid artery atherosclerosis as measured by B-mode ultrasound to the extent of coronary atherosclerosis. Arterioscler Thromb 1991; 11: 1786-1794.

3. O’Leary DH, Polak JF, Kronmal RA, Manolio TA, Burke GL, Wolfson SK Jr. Carotid-artery intima and media thick- ness as a risk factor for myocardial infarction and stroke in older adults. Cardiovascular Health Study Collaborative Research Group. N Engl J Med 1999; 340: 14-22.

4. Freire CM, Ribeiro AL, Barbosa FB, et al. Comparison be- tween automated and manual measurements of carotid in- tima-media thickness in clinical practice. Vasc Health Risk Manag 2009; 5: 811-817.

5. Gonzalez J, Wood JC, Dorey FJ, Wren TA, Gilsanz V. Re- producibility of carotid intima-media thickness measure- ments in young adults. Radiology 2008; 247: 465-471.

6. Polak JF, Funk LC, O’Leary DH. Inter-reader differences in common carotid artery intima-media thickness: implica- tions for cardiovascular risk assessment and vascular age determination. J Ultrasound Med 2011; 30: 915-920.

7. Naghavi M, Falk E, Hecht HS, et al. From vulnerable plaque to vulnerable patient--Part III: Executive summary of the Screening for Heart Attack Prevention and Educa-

tion (SHAPE) Task Force report. Am J Cardiol 2006; 98:

2H-15H.

8. Hodis HN, Mack WJ, LaBree L, et al. The role of carotid arterial intima-media thickness in predicting clinical coro- nary events. Ann Intern Med 1998; 128: 262-269.

9. Stein JH, Korcarz CE, Hurst RT, et al. Use of carotid ultra- sound to identify subclinical vascular disease and evaluate cardiovascular disease risk: a consensus statement from the American Society of Echocardiography Carotid Intima- Media Thickness Task Force. Endorsed by the Society for Vascular Medicine. J Am Soc Echocardiogr 2008; 21: 93- 111.

10. Carpenter JP, Lexa FJ, Davis JT. Determination of duplex Doppler ultrasound criteria appropriate to the North Ameri- can Symptomatic Carotid Endarterectomy Trial. Stroke 1996; 27: 695-699.

11. Riley WA, Barnes RW, Applegate WB, et al. Reproduc- ibility of noninvasive ultrasonic measurement of carotid atherosclerosis. The Asymptomatic Carotid Artery Plaque Study. Stroke 1992; 23: 1062-1068.

12. Latich I, Scoutt LM. Ultrasound evaluation of carotid inti- ma media thickness: The Hows and Whys. Ultrasound Clin 2011; 6: 445-461.

13. Thanassoulis G, Peloso GM, Pencina MJ, et al. A genetic risk score is associated with incident cardiovascular disease and coronary artery calcium: the Framingham Heart Study.

Circ Cardiovasc Genet 2012; 5: 113-121.

14. Stensland-Bugge E, Bønaa KH, Joakimsen O. Reproduc- ibility of ultrasonographically determined intima-media thickness is dependent on arterial wall thickness. The Tromsø Study. Stroke 1997; 28: 1972-1980.

15. Kanters SD, Algra A, van Leeuwen MS, Banga JD. Repro- ducibility of in vivo carotid intima-media thickness meas- urements: a review. Stroke 1997; 28: 665-671.

16. Lee VS, Hertzberg BS, Kliewer MA, Carroll BA. Assess- ment of stenosis: implications of variability of Doppler measurements in normal-appearing carotid arteries. Radi- ology 1999; 212: 493-498.

17. Hadlock J, Beach KW; D.E. Strandness Vascular Laborato- ry Sonographers. Velocity variability in ultrasonic Doppler examinations. Ultrasound Med Biol 2009; 35: 949-954.

18. O’Leary DH, Polak JF, Kronmal RA, et al. Distribution and correlates of sonographically detected carotid artery disease in the Cardiovascular Health Study. The CHS Col- laborative Research Group. Stroke 1992; 23: 1752–1760.

19. Inaba Yoichi, Chen JA, Bergmann SR. Carotid plaque, compared with carotid intima-media thickness, more ac- curately predicts coronary artery disease events: a meta- analysis. Atherosclerosis 2012; 220: 128-133.

20. Velázquez F, Berná JD, Abellán JL, Serrano L, Escribano A, Canteras M. Reproducibility of sonographic measure- ments of carotid intima-media thickness. Acta Radiol 2008;

49: 1162-1166.

21. Kanters SD, Elgersma OE, Banga JD, van Leeuwen MS, Algra A. Reproducibility of measurements of intima- media thickness and distensibility in the common carotid artery. Eur J Vasc Endovasc Surg 1998; 16: 28-35.

(8)

22. Ameneiro Perez S, Fernández Bolona A. Comparación de la medición bilateral del grosor íntima-media carotídeo.

Rev Cubana Angiol Cir Vasc 2008; 9. Available at: http://

www.bvs.sld.cu/revistas/ang/vol9_1_08/ang0208.htm. Ac- cessed: 2012.09.03.

23. Chaubey S, Nitsch D, Altmann D, Ebrahim S. Differing effect of modifiable cardiovascular risk factors on intima- media thickening and plaque formation at different sites of the arterial vasculature. Heart 2010; 96: 1579-1585.

24. Lee SW, Hai JJ, Kong SL, et al. Side differences of carotid intima-media thickness in predicting cardiovascular events among patients with coronary artery disease. Angiology 2011; 62: 231-236.

25. Selamet Tierney ES, Gauvreau K, Jaff MR, et al. Carotid artery intima-media thickness measurements in the youth:

reproducibility and technical considerations. J Am Soc Echocardiogr 2015; 28: 309-316.

26. Stensland-Bugge E, Bonaa KH, Joakimsen O. Reproduc- ibility of ultrasonographically determined intima-media thickness is dependent on arterial wall thickness. The Tromso Study. Stroke 1997; 28: 1972-1980.

27. Di Geso L, Zardi EM, Afeltra A, et al. Comparison between conventional and automated software-guided ultrasound assessment of bilateral common carotids intima-media thickness in patients with rheumatic diseases. Clin Rheu- matol 2012; 31: 881-884.

28. Iglesias del Sol A, Bots ML, Grobbee DE, Hofman A, Wit- teman JC. Carotid intima-media thickness at different sites:

relation to incident myocardial infarction; The Rotterdam Study. Eur Heart J 2002; 23: 934-940.

29. Nichols S, Milner M, Meijer R, Carroll S, Ingle L. Vari- ability of automated carotid intima-media thickness meas- urements by novice operators. Clin Physiol Funct Imaging 2014 Sep 12. doi: 10.1111/cpf.12189.

30. Aldridge ME, Do K, Oo T, Naqvi TZ. Carotid intima-media thickness and plaque assessment by trained medical resi- dents: validation and preliminary testing of a training proto- col. J Am Soc Echocardiogr 2013; 26: 1457-1464.

31. Vanoli D, Wiklund U, Lindqvist P, Henein M, Näslund U.

Successful novice’s training in obtaining accurate assess- ment of carotid IMT using an automated ultrasound system.

Eur Heart J Cardiovasc Imaging 2014; 15: 637-642.

32. Kohler TR, Langlois Y, Roederer GO, et al. Variability in measurement of specific parameters for carotid duplex ex- amination. Ultrasound Med Biol 1987; 13: 637-642.

33. Hopkins WG. Measures of reliability in sports medicine and science. Sports Med 2000; 30: 1-15.

34. Atkinson G, Nevill AM. Selected issues in the design and analysis of sport performance research. J Sports Sci 2001;

19: 811-827.

35. Gwet, K. Handbook of Inter-Rater Reliability: The Defini- tive Guide to Measuring the Extent of Agreement Among Multiple Raters. 3rd Edition. Maryland, USA: Advanced Analytics, LLC, 2012.

36. Lau KH, Fung YK, Cheung YT, Tsang WK, Ying M. Repeat- ability and reproducibility of ultrasonographic measurement of carotid intima thickness. J Clin Ultrasound 2012; 40: 79-84.

37. Molinari F, Zeng G, Suri JS. A state of the art review on intima-media thickness (IMT) measurement and wall seg- mentation techniques for carotid ultrasound. Comput Meth- ods Programs Biomed 2010; 100: 201-221.

38. Napel S, Xu H, Paik DS, et al. Carotid disease: automated analysis with cardiac-gated three-dimensional US technique and preliminary results. Radiology 2002; 222: 560-563.

39. Loizou CP, Pattichis CS, Nicolaides AN, Pantziaris M.

Manual and automated media and intima thickness meas- urements of the common carotid artery. IEEE Trans Ultra- son Ferroelectr Freq Control 2009; 56: 983-994.

Referințe

DOCUMENTE SIMILARE

These themes are: frequency use of New Information and Communication Technologies (NICT), use of NICT in order to find out information about church, cult

In conclusion, using a series of carotid ultrasound im- ages already acquired in a consecutive series of patients, we showed that CIMT measurement using semiautomatic

Aim: To test the ability of carotid stiffness evaluated by using ultrafast ultrasound imaging to indicate coronary athero- sclerosis and its association with the severity of

Conclusions: In cases of BCCs with thickness of ≤1 mm, there was a high correlation (r=0.870) of the tumor spread depth between micromorphological measurements and the results

Ultrasonographic evaluation of abdominal wall fat index, carotid intima-media thickness and plaque score in obstructive sleep apnea syndrome.. Nurcan Çetin 1 , İdil Güneş Tatar 1

CIMTb = carotid intima‑media thickness measured at carotid bulb level; MHO = metabolically healthy obese; MHO1 = obese with increased WC; MHO2 = obese with increased WC plus one of

The number of vacancies for the doctoral field of Medicine, Dental Medicine and Pharmacy for the academic year 2022/2023, financed from the state budget, are distributed to

Aims: The aim of this study was to calculate the normal values of renal functional reserve in healthy individuals measur- ing the resistive index and pulsatility index using