- Email: [email protected]
Common Carotid Artery Duplex for the Bubble Test to Detect Right-to-Left Shunt
Ultrasound in Med. & Biol., Vol. 36, No. 4, pp. 566–570, 2010 Copyright Ó 2010 World Federation for Ultrasound in Medicine & Biology Printed in the USA. All rights reserved 0301-5629/10/$–see front matter
doi:10.1016/j.ultrasmedbio.2010.01.009
d
Original Contribution COMMON CAROTID ARTERY DUPLEX FOR THE BUBBLE TEST TO DETECT RIGHT-TO-LEFT SHUNT BRUNO CENSORI, TANIA PARTZIGUIAN, and MARCO POLONI USC Neurologia, Ospedali Riuniti di Bergamo, Bergamo, Italy (Received 4 September 2009; revised 12 January 2010; in final form 25 January 2010)
Abstract—We prospectively compared the bubble test with agitated saline for right-to-left shunt using transcranial Doppler (TCD) of the right middle cerebral artery and second harmonic imaging duplex of the right common carotid artery (CCA) in 100 consecutive patients. Microembolic signals (MES) were counted offline. MES were classified into 6 classes: absent (class 0), 1–10 MES (class 1), 11–20 MES (class 2), 21–30 MES (class 3), 31–50 MES (class 4) and .50 MES or ‘‘curtain effect’’ (class 5). For TCD, classes 2–5 combined (i.e., ‘‘large’’ shunts), the sensitivity of duplex with the Valsalva maneuver was 95.3%, the specificity was 100%, the positive predictive 100%, the negative predictive value 96.6% and accuracy 98.0%. Second harmonic imaging duplex of the CCA may substitute TCD for the bubble test when an adequate cranial bone window is not available. This technique may also greatly increase the number of facilities where the bubble test can be carried out. However, tests with few or no MES need to be confirmed by TCD or transesophageal echocardiography. (E-mail: [email protected]) Ó 2010 World Federation for Ultrasound in Medicine & Biology. Key Words: Transcranial Doppler, Foramen ovale patent, Carotid ultrasound.
easy-to-perform technique (Di Tullio et al. 1993), 6–18% of patients do not have a suitable bone window for the examination (Marinoni et al. 1997; Wijnhoud et al. 2008) and TCD devices are usually confined to neurovascular departments, neurosurgeries and intensive care units. Insonating an extracranial artery with a duplex probe may allow the examination to be carried out in patients without a temporal bone window and would make it possible to perform the bubble test right after the study of cervico-cephalic vessels, without the need for a new instrument. In this study we compared the results of the bubble test carried out with transcranial Doppler at the level of one MCA and with a duplex probe at the level of the common carotid artery (CCA).
INTRODUCTION The search for patent foramen ovale (PFO) is becoming more and more common in different conditions, such as stroke, diving medicine, migraine with aura, and obstructive sleep apnea syndrome (Aslam et al. 2006). Different techniques can be used, such as transthoracic echocardiography, transesophageal echocardiography and transcranial Doppler (TCD) (Di Tullio et al. 1993). All are performed both in the basal condition and after one or more Valsalva maneuvers, to increase right atrial pressure and enhance right-to-left shunt. To detect venoarterial shunt with TCD, gaseous microbubbles are rapidly injected into a peripheral vein and the corresponding high-intensity signals are recorded at the level of one middle cerebral artery (MCA), although other intracranial arteries may be insonated for this purpose (Del Sette et al. 2007; Perren et al. 2008). With TCD alone, one cannot determine where the bubbles actually cross over into the arterial circulation (Peters et al. 2005; Cottin et al. 2007), but in the vast majority of cases, the shunt is caused by PFO. Although TCD is a sensitive and
PATIENTS AND METHODS We prospectively studied 100 consecutive patients with an indication to the bubble test for PFO, coming either from our outpatient neurovascular laboratory or from our Stroke Unit. The reasons for the bubble test were migraine (n 5 19), stroke (n 5 48), transient ischemic attack (n 5 16) and vertigo/transient global amnesia/paresthesias (n 5 17). There were 53 men and 47 women, with a mean age of 51.0 6 16.0 y. During
Address correspondence to: Bruno Censori, M.D., USC Neurologia, Ospedali Riuniti di Bergamo, Largo Barozzi, 1, 24100 Bergamo, Italy. E-mail: [email protected] 566
CCA duplex to detect R-L shunt d B. CENSORI et al.
the test patients were lying supine with their heads elevated at 0–20 . The antecubital vein was cannulated with an 18 F three-way stopcock. A mixture of air and microbubbles was generated by mixing 8.5 mL saline with 1 mL room air in a 10-mL syringe via to-and-fro pushing through the three-way stopcock. A small quantity of blood was aspirated within the syringe before mixing (approximately 0.5 mL), because this improves microbubble formation (Sastry et al. 2007). Three tests were carried out in each patient, one in the basal condition and two after a Valsalva maneuver (VM). For the VM, the expiratory effort was maintained for 10 s and the injection of microbubbles was started 5 s before beginning of the maneuver (Jauss and Zanette 2000). The VM was considered successful if a $30% decrease of the mean MCA velocity was observed (Jauss and Zanette 2000). Before the actual test, one or two trials of the VM were carried out to obtain a good reduction of flow velocity and to instruct the patients not to move their necks during the maneuver. Microembolic signals were counted for 30 s after microbubble injection. All patients successfully underwent all three parts of the test, although 10 VMs (5.0%) had to be repeated because of patient neck movement when the maneuver ended, which displaced the CCA from the sample volume. In some other cases, rapid refocusing of the probe over the CCA was necessary because of neck movement, but the maneuver was carried out successfully. Transcranial Doppler was carried out with a MultiDop T device (Compumedics, Singen, Germany). The 2-MHz probe was fixed at the level of the right temporal window and the sample volume was set at a depth of 50–60 mm depending on individual patients’characteristics. The duplex examination was carried out with a GE Logiq7 device (General Electric Medical Systems, Milwaukee, WI, USA), using a multifrequency linear array probe. Frequencies were 10.0–12.0 MHz for B-mode second harmonic imaging of the CCA and 5.0 MHz for pulsed-wave Doppler. For harmonic imaging, the mechanical index was 1.2 and the frame rate was 18 or 27 Hz. The right CCA was insonated in a longitudinal plane for microembolic signal (MES) detection. The sample volume was positioned at the mid- or distal segment of the vessel and was enlarged to encompass the entire diameter of the CCA. Usually, a 12- to 16-millimeter width was required. The angle steer was set at 20 to the left and the Doppler angle vector was corrected up to a value of 60 , because it is usually done for flow velocity estimations. The Doppler gain was set as low as possible to maximize the prominence of MES against the background. The pulse repetition frequency was decreased to obtain the maximum height of the Doppler spectral waveform on the screen, without signal aliasing. The duplex probe was manually held in place for the whole test. Duplex and TCD tracings were recorded and analyzed offline by an experienced
567
Fig. 1. Microembolic signals in the common carotid flow velocity spectrum.
observer (BC). The criteria for MES identification on TCD tracings are those of the literature (International Consensus 1998; Censori et al. 2000). For CCA tracings, MES were identified as short (,300 ms), unidirectional, increases in the background signal, occurring without any relationship with the cardiac cycle and usually associated with a short-lasting sound (Fig. 1). The number of signals was classified into six categories: class 0 5 no signals; class 1 5 1–10 MES; class 2 5 11–20 MES; class 3 5 21–30 MES; class 4 5 31–50 MES and class 5 5 .50 MES or ‘‘curtain’’ pattern. For the VM, the concordance between results obtained in the two tests with each method (TCD or duplex) was analyzed by means of the Wilcoxon rank-sum test for matched data. The comparison between TCD and duplex was then carried out using only the TCD series with the highest number of MES-positive results and its corresponding duplex data. MES scores were compared between the two modalities by means of the Wilcoxon rank-sum test for matched data. Furthermore, the concordance for patients with any shunt or a ‘‘large’’ shunt was compared between TCD and duplex groups by means of McNemar’s test. Finally, sensitivity, specificity, positive predictive value, negative predictive value and overall accuracy of the results in the basal condition and after the VM were calculated. The local ethics committee approved the study, and all patients gave informed consent. RESULTS One hundred basal and 200 VMs were available for MES counting. With the VM, agreement was very good both between the two TCD series and between the two duplex series (Wilcoxon rank-sum test for matched data: p 5 0.542 for TCD; p 5 0.524 for duplex). The first series of TCD results, and its corresponding duplex scores, was used for data analysis because it had the same number of
568
Ultrasound in Medicine and Biology
Table 1. TCD-Duplex agreement in the basal condition Duplex class TCD class
0
1
0 1 2 3 4 5
68 1
2 13 1
2
3
4
5
Total (n 5 100)
3
70 14 6 3 4 3
5 3 2
2
class 5 results (i.e., 29) as the second series and overall one more MES-positive result. The correspondence between TCD and duplex scores can be seen in Table 1 for the basal condition and Table 2 for the 100 VMs of the first TCD series. The Wilcoxon rank-sum test for matched data did not show significant differences between the two methods in the basal condition (p 5 0.562). There was a significant difference between the two methods for the 100 VMs (p , 0.001). Considering the 100 basal tests and the 100 VMs combined, at TCD 70 basal and 41 VM tests did not show any MES; all were also scored as class 0 at CCA duplex, apart from two, which scored class 1. At TCD, three basal tests and 29 VMs were scored as class 5; of these, 30 (93.8%) were scored as class 5 at CCA duplex, one was scored as class 4 (3.1%) and one as class 3 (3.1%). Of 89 TCD tests (basal and VM combined) with at least one MES, six did not show any MES at CCA duplex (6.7%). For all, the TCD class was 1 (i.e., they had between 1 and 10 MES, with a median MES number of 2). In 17 of 200 tests (8.5%), the TCD and duplex classes differed, but in only two cases (1.0%) the difference was greater than 1 class. Considering only TCD tests with at least 1 MES (n 5 89), TCD and duplex classes differed in 15 cases (16.8%). In all of these instances, the duplex class was lower, but the difference was greater than 1 class in two cases (2.2%) only. Considering a TCD class $1 (i.e., at least 1 MES), the sensitivity of basal duplex was 97.1%, specificity was 93.6%, positive predictive value (PPV) was 98.6% and negative predictive value (NPV) was 93.6%, for an overall accuracy of 97.0%. There was no significant difference between the two Table 2. TCD-Duplex agreement with the Valsalva maneuver Duplex class TCD class
0
0 1 2 3 4 5
41 5
1 11 2
2
3 1 1
3
4
5
Total (n 5 10)
27
41 16 5 5 4 29
4 1
3 1
Volume 36, Number 4, 2010
methods (p , 1 by McNemar’s test). With the 100 VMs, the difference between the two methods for patients with at least 1 MES approached significance (p , 0.07 by McNemar’s test). The sensitivity of duplex was 91.5%, specificity was 100%, PPV was 100% and NPV was 89.1%, for an overall accuracy of 95.0%. Considering TCD classes 2–5 combined (i.e., .10 MES), which are usually considered representative of ‘‘large’’ shunts (Jauss and Zanette 2000), the sensitivity of basal duplex was 93.8%, specificity was 100%, PPV was 100%, NPV was 98.8%, accuracy was 99.0% and the difference between the two methods was not significant (p , 1 by McNemar’s test). For the 100 VMs, there was no significant difference in the concordance for large shunts between the two methods (p , 0.50 by McNemar’s test). Sensitivity of duplex was 95.3%, specificity was 100%, PPV 100%, NPV 96.6% and overall accuracy 98.0%. With a more restrictive definition of a ‘‘large’’ shunt as one associated with .20 MES (i.e., classes 3 to 5 combined), the sensitivity of basal duplex was 100%, specificity was 100%, PPV and NPV were 100% each and overall accuracy was 100%. For the 100 VMs, again, there was no significant difference between the two methods (p , 0.50 by McNemar’s test). The sensitivity of duplex was 94.7%, specificity was 100%, PPV was 100% and NPV was 96.9%, for an overall accuracy of 98.0%. Microembolic signals of the CCA were often less prominent, both visually against the background, and acoustically, than TCD MES. DISCUSSION AND SUMMARY We have found that a modern duplex probe for neck vessel ultrasound insonating the CCA can be only slightly, although significantly, less sensitive than a TCD probe at the MCA level for the detection of MES during the bubble test for right-to-left shunt evaluation. This finding has two major implications. First, this test modality may be used when a bone window is not available, which happens in 6–18% of cases (Marinoni et al. 1997; Wijnhoud et al. 2008). Second, it may greatly increase the number of facilities where the bubble test can be carried out, because ultrasound devices with vascular software are much more diffuse than TCD devices. This may speed up the evaluation protocol of patients when such a test is indicated. After a brief training period, a cardiologist, an internist, a vascular surgeon or a vascular ultrasound technician may easily perform the investigation without removing the patient from the bed where an echocardiography or a study of neck vessels has just been carried out. The lower sensitivity of duplex primarily resulted in some patients with ,5 MES at the TCD VM being classified as without any shunt at duplex, an error without major consequences.
CCA duplex to detect R-L shunt d B. CENSORI et al.
Fig. 2. A curtain of MES in the Doppler spectrum and the corresponding shower of microbubbles within the common carotid lumen.
Others have tried to compare MCA bubble test results with those recorded at other peripheral arteries (Topcuoglu et al. 2003; Draganski et al. 2005; Daly et al. 2008). Those data are not comparable to ours because they were obtained by means of a TCD probe rather than a duplex probe (Topcuoglu et al. 2003; Daly et al. 2008), or through B-mode visualization of galactose microbubbles in axial sections of the CCA (Draganski et al. 2005). Interestingly, however, both B-mode echography of the CCA (Draganski et al. 2005) and TCD recording of MES at the CCA or common femoral artery (Daly et al. 2008) gave results comparable to ‘‘classical’’ MCA TCD recording. We chose the CCA, rather than the internal carotid artery (ICA) because after a few trials, it proved easier to insonate during the VM. Furthermore, in some patients the carotid bifurcation has a high position in the neck or is heavily calcified, and these factors may prevent an adequate visualization of the ICA even at rest. Finally, a more proximal insonification should theoretically be more sensitive to shunting, because when microbubbles are few they may theoretically all travel into the external carotid artery rather that the ICA, or into the anterior cerebral artery, and may therefore not be detectable at TCD of the MCA. Our experience, however, proved that the opposite was more often true. The duplex investigation was generally easy to carry out. The main difficulties were represented by neck movements at the start or at release of the VM, which required a rapid refocusing of the sample volume on the CCA. Although this did not happen often, a preliminary trial of the VM is indicated before the actual test, both to make sure that the maneuver is well-executed and to instruct the patients not to move too much while performing the maneuver. We have used second harmonic imaging only
569
to ensure optimal positioning of the Doppler sample volume over the CCA, and not to count microbubbles. Although showers of microbubbles could be seen very well on B-mode images (Fig. 2), single MES were often not paralleled by a corresponding microbubble within the CCA. This is probably because longitudinal B-mode images visualize only a bidimensional section of the arterial lumen, whereas the Doppler signal is derived from a tridimensional volume (Steinman et al. 2004). The main reason that a few microembolic signals visible at TCD were missed by the duplex technique probably lies in the incomplete coverage of the CCA lumen by the Doppler sample volume. Because the sample volume is positioned according to the B-mode image, it is possible that part of the lumen is actually left out of the detection volume. In some instances, on the contrary, a few MES detected in the CCA Doppler spectrum were not followed by corresponding MES in the MCA tracing. This phenomenon may have several explanations: (i) microbubbles travelling along the CCA may collapse before reaching the MCA; (ii) they may end up in the anterior cerebral artery, rather than the MCA; or (iii) they may enter the external rather than the ICA. On B-mode images we have indeed seen many microbubbles going into the external carotid artery in patients with showers of MES. Although this technique gave satisfactory results, it may be more operator-dependent than TCD, at least in regard to holding the probe in the best position at rest and especially after the VM. Furthermore, duplex probes have not been designed to facilitate MES identification, as is the case with dual-gate TCD probes for automated embolus detection. Therefore, some efforts will have to be made to test interobserver reproducibility of results and to optimize ultrasound parameters for MES detection. It may also be worthwile to test whether the delay between injection and first appearance of MES may give information about shunt site, although this strategy is of doubtful usefulness with TCD (International Consensus 1998). In conclusion, our study shows that second harmonic imaging duplex of the CCA is slightly, although significantly, less sensitive than TCD for the detection of MES. Thus, it might be used for the bubble test when there is no cranial window for TCD. Also, because of very high specificity and good sensitivity of the test to detect large shunts, institutions that do not possess a TCD device may consider the Duplex bubble test at the CCA in a first instance. However, a conventional transesophageal echocardiogram (TEE) or a TCD investigation are necessary when there are no or few MES at duplex, because some small shunts may be missed and a few large shunts may be underestimated by this method. Furthermore, TEE is necessary to confirm the cardiac site of the shunt and to properly define the anatomical characteristics of the interatrial communication.
570
Ultrasound in Medicine and Biology
REFERENCES Aslam F, Shirani J, Haque AA. Patent foramen ovale: Assessment, clinical significance and therapeutic options. South Med J 2006;99: 1367–1372. Censori B, Partziguian T, Casto L, Camerlingo M, Mamoli A. Doppler micorembolic signals predict ischemic recurrences in symptomatic carotid stenosis. Acta Neurol Scand 2000;01:327–331. Cottin V, Dupuis, Girod S, Lesca G, Cordier JF. Pulmonary vascular manifestations of hereditary hemorrhagic telangiectasia. Respiration 2007;74:361–378. Daly KJ, Hutchinson S, McCollum CN. Can a peripheral artery be used to detect venous to arterial circulation shunts? Cerebrovasc Dis 2008; 25:111–114. Del Sette M, Dinia L, Rizzi D, Sugo A, Albano B, Gandolfo C. Diagnosis of right-to-left shunt with transcranial Doppler and vertebrobasilar recording. Stroke 2007;38:2254–2256. Di Tullio M, Sacco RL, Venketasubramanian N, Sherman D, Mohr JP, Homma S. Comparison of diagnostic techniques for the detection of a patent foramen ovale in stroke patients. Stroke 1993;24:1020–1024. Draganski B, Blersch W, Holmer S, Koch H, May A, Bogdahn U, Ho¨lscher T, Schlachetzki F. Detection of cardiac right-to-left shunts by contrast-enhanced harmonic carotid duplex sonography. J Ultrasound Med 2005;24:1071–1076. International Consensus Group on Microembolus Detection. Consensus on microembolus detection by TCD. Stroke 1998;29:725–729.
Volume 36, Number 4, 2010 Jauss M, Zanette E. Detection of right-to-left shunt with ultrasound contrast agent and transcranial Doppler sonography. Cerebrovasc Dis 2000;10:490–496. Marinoni M, Ginanneschi A, Forleo P, Amaducci L. Technical limits in transcranial Doppler recordings: inadequate acoustic window. Ultrasound Med Biol 1997;23:1275–1277. Perren F, Savva E, Landis T. Trasforaminal Doppler: An alternative to transtemporal approach for right-to-left cardiac shunt assessment. J Neurol Sci 2008;273:49–50. Peters B, Ewert P, Schubert S, Abdul-Khaliq H, Lange PE. Rare case of pulmonary arteriovenous fistula simulating residual defect after transcatheter closure of patent foramen ovale for recurrent paradoxical embolism. Catheter Cardiovasc Interv 2005;64:348–351. Sastry S, Daly K, Chengodu T, McCollum C. Is transcranial Doppler for the detection of venous-to-arterial circulation shunts reproducible? Cerebrovasc Dis 2007;23:424–429. Steinman AH, Lui EYL, Johnston WK, Cobbold RSC. Sample volume shape for pulsed-flow velocity estimation using a linear array. Ultrasound Med Biol 2004;30:1409–1418. Topcuoglu MA, Palacios IF, Buonanno FS. Contrast M-mode power Doppler ultraosund in the detection of right-to-left shunts: Utility of submandibular internal carotid artery recording. J Neuroimaging 2003;13:315–323. Wijnhoud AD, Franckena M, van der Lugt A, Koudstaal PJ, Dippel ED. Inadequate acoustical temporal bone window in patients with a transient ischemic attack or minor stroke: Role of skull thickness and bone density. Ultrasound Med Biol 2008;34:923–929.
doi:10.1016/j.ultrasmedbio.2010.01.009
d
Original Contribution COMMON CAROTID ARTERY DUPLEX FOR THE BUBBLE TEST TO DETECT RIGHT-TO-LEFT SHUNT BRUNO CENSORI, TANIA PARTZIGUIAN, and MARCO POLONI USC Neurologia, Ospedali Riuniti di Bergamo, Bergamo, Italy (Received 4 September 2009; revised 12 January 2010; in final form 25 January 2010)
Abstract—We prospectively compared the bubble test with agitated saline for right-to-left shunt using transcranial Doppler (TCD) of the right middle cerebral artery and second harmonic imaging duplex of the right common carotid artery (CCA) in 100 consecutive patients. Microembolic signals (MES) were counted offline. MES were classified into 6 classes: absent (class 0), 1–10 MES (class 1), 11–20 MES (class 2), 21–30 MES (class 3), 31–50 MES (class 4) and .50 MES or ‘‘curtain effect’’ (class 5). For TCD, classes 2–5 combined (i.e., ‘‘large’’ shunts), the sensitivity of duplex with the Valsalva maneuver was 95.3%, the specificity was 100%, the positive predictive 100%, the negative predictive value 96.6% and accuracy 98.0%. Second harmonic imaging duplex of the CCA may substitute TCD for the bubble test when an adequate cranial bone window is not available. This technique may also greatly increase the number of facilities where the bubble test can be carried out. However, tests with few or no MES need to be confirmed by TCD or transesophageal echocardiography. (E-mail: [email protected]) Ó 2010 World Federation for Ultrasound in Medicine & Biology. Key Words: Transcranial Doppler, Foramen ovale patent, Carotid ultrasound.
easy-to-perform technique (Di Tullio et al. 1993), 6–18% of patients do not have a suitable bone window for the examination (Marinoni et al. 1997; Wijnhoud et al. 2008) and TCD devices are usually confined to neurovascular departments, neurosurgeries and intensive care units. Insonating an extracranial artery with a duplex probe may allow the examination to be carried out in patients without a temporal bone window and would make it possible to perform the bubble test right after the study of cervico-cephalic vessels, without the need for a new instrument. In this study we compared the results of the bubble test carried out with transcranial Doppler at the level of one MCA and with a duplex probe at the level of the common carotid artery (CCA).
INTRODUCTION The search for patent foramen ovale (PFO) is becoming more and more common in different conditions, such as stroke, diving medicine, migraine with aura, and obstructive sleep apnea syndrome (Aslam et al. 2006). Different techniques can be used, such as transthoracic echocardiography, transesophageal echocardiography and transcranial Doppler (TCD) (Di Tullio et al. 1993). All are performed both in the basal condition and after one or more Valsalva maneuvers, to increase right atrial pressure and enhance right-to-left shunt. To detect venoarterial shunt with TCD, gaseous microbubbles are rapidly injected into a peripheral vein and the corresponding high-intensity signals are recorded at the level of one middle cerebral artery (MCA), although other intracranial arteries may be insonated for this purpose (Del Sette et al. 2007; Perren et al. 2008). With TCD alone, one cannot determine where the bubbles actually cross over into the arterial circulation (Peters et al. 2005; Cottin et al. 2007), but in the vast majority of cases, the shunt is caused by PFO. Although TCD is a sensitive and
PATIENTS AND METHODS We prospectively studied 100 consecutive patients with an indication to the bubble test for PFO, coming either from our outpatient neurovascular laboratory or from our Stroke Unit. The reasons for the bubble test were migraine (n 5 19), stroke (n 5 48), transient ischemic attack (n 5 16) and vertigo/transient global amnesia/paresthesias (n 5 17). There were 53 men and 47 women, with a mean age of 51.0 6 16.0 y. During
Address correspondence to: Bruno Censori, M.D., USC Neurologia, Ospedali Riuniti di Bergamo, Largo Barozzi, 1, 24100 Bergamo, Italy. E-mail: [email protected] 566
CCA duplex to detect R-L shunt d B. CENSORI et al.
the test patients were lying supine with their heads elevated at 0–20 . The antecubital vein was cannulated with an 18 F three-way stopcock. A mixture of air and microbubbles was generated by mixing 8.5 mL saline with 1 mL room air in a 10-mL syringe via to-and-fro pushing through the three-way stopcock. A small quantity of blood was aspirated within the syringe before mixing (approximately 0.5 mL), because this improves microbubble formation (Sastry et al. 2007). Three tests were carried out in each patient, one in the basal condition and two after a Valsalva maneuver (VM). For the VM, the expiratory effort was maintained for 10 s and the injection of microbubbles was started 5 s before beginning of the maneuver (Jauss and Zanette 2000). The VM was considered successful if a $30% decrease of the mean MCA velocity was observed (Jauss and Zanette 2000). Before the actual test, one or two trials of the VM were carried out to obtain a good reduction of flow velocity and to instruct the patients not to move their necks during the maneuver. Microembolic signals were counted for 30 s after microbubble injection. All patients successfully underwent all three parts of the test, although 10 VMs (5.0%) had to be repeated because of patient neck movement when the maneuver ended, which displaced the CCA from the sample volume. In some other cases, rapid refocusing of the probe over the CCA was necessary because of neck movement, but the maneuver was carried out successfully. Transcranial Doppler was carried out with a MultiDop T device (Compumedics, Singen, Germany). The 2-MHz probe was fixed at the level of the right temporal window and the sample volume was set at a depth of 50–60 mm depending on individual patients’characteristics. The duplex examination was carried out with a GE Logiq7 device (General Electric Medical Systems, Milwaukee, WI, USA), using a multifrequency linear array probe. Frequencies were 10.0–12.0 MHz for B-mode second harmonic imaging of the CCA and 5.0 MHz for pulsed-wave Doppler. For harmonic imaging, the mechanical index was 1.2 and the frame rate was 18 or 27 Hz. The right CCA was insonated in a longitudinal plane for microembolic signal (MES) detection. The sample volume was positioned at the mid- or distal segment of the vessel and was enlarged to encompass the entire diameter of the CCA. Usually, a 12- to 16-millimeter width was required. The angle steer was set at 20 to the left and the Doppler angle vector was corrected up to a value of 60 , because it is usually done for flow velocity estimations. The Doppler gain was set as low as possible to maximize the prominence of MES against the background. The pulse repetition frequency was decreased to obtain the maximum height of the Doppler spectral waveform on the screen, without signal aliasing. The duplex probe was manually held in place for the whole test. Duplex and TCD tracings were recorded and analyzed offline by an experienced
567
Fig. 1. Microembolic signals in the common carotid flow velocity spectrum.
observer (BC). The criteria for MES identification on TCD tracings are those of the literature (International Consensus 1998; Censori et al. 2000). For CCA tracings, MES were identified as short (,300 ms), unidirectional, increases in the background signal, occurring without any relationship with the cardiac cycle and usually associated with a short-lasting sound (Fig. 1). The number of signals was classified into six categories: class 0 5 no signals; class 1 5 1–10 MES; class 2 5 11–20 MES; class 3 5 21–30 MES; class 4 5 31–50 MES and class 5 5 .50 MES or ‘‘curtain’’ pattern. For the VM, the concordance between results obtained in the two tests with each method (TCD or duplex) was analyzed by means of the Wilcoxon rank-sum test for matched data. The comparison between TCD and duplex was then carried out using only the TCD series with the highest number of MES-positive results and its corresponding duplex data. MES scores were compared between the two modalities by means of the Wilcoxon rank-sum test for matched data. Furthermore, the concordance for patients with any shunt or a ‘‘large’’ shunt was compared between TCD and duplex groups by means of McNemar’s test. Finally, sensitivity, specificity, positive predictive value, negative predictive value and overall accuracy of the results in the basal condition and after the VM were calculated. The local ethics committee approved the study, and all patients gave informed consent. RESULTS One hundred basal and 200 VMs were available for MES counting. With the VM, agreement was very good both between the two TCD series and between the two duplex series (Wilcoxon rank-sum test for matched data: p 5 0.542 for TCD; p 5 0.524 for duplex). The first series of TCD results, and its corresponding duplex scores, was used for data analysis because it had the same number of
568
Ultrasound in Medicine and Biology
Table 1. TCD-Duplex agreement in the basal condition Duplex class TCD class
0
1
0 1 2 3 4 5
68 1
2 13 1
2
3
4
5
Total (n 5 100)
3
70 14 6 3 4 3
5 3 2
2
class 5 results (i.e., 29) as the second series and overall one more MES-positive result. The correspondence between TCD and duplex scores can be seen in Table 1 for the basal condition and Table 2 for the 100 VMs of the first TCD series. The Wilcoxon rank-sum test for matched data did not show significant differences between the two methods in the basal condition (p 5 0.562). There was a significant difference between the two methods for the 100 VMs (p , 0.001). Considering the 100 basal tests and the 100 VMs combined, at TCD 70 basal and 41 VM tests did not show any MES; all were also scored as class 0 at CCA duplex, apart from two, which scored class 1. At TCD, three basal tests and 29 VMs were scored as class 5; of these, 30 (93.8%) were scored as class 5 at CCA duplex, one was scored as class 4 (3.1%) and one as class 3 (3.1%). Of 89 TCD tests (basal and VM combined) with at least one MES, six did not show any MES at CCA duplex (6.7%). For all, the TCD class was 1 (i.e., they had between 1 and 10 MES, with a median MES number of 2). In 17 of 200 tests (8.5%), the TCD and duplex classes differed, but in only two cases (1.0%) the difference was greater than 1 class. Considering only TCD tests with at least 1 MES (n 5 89), TCD and duplex classes differed in 15 cases (16.8%). In all of these instances, the duplex class was lower, but the difference was greater than 1 class in two cases (2.2%) only. Considering a TCD class $1 (i.e., at least 1 MES), the sensitivity of basal duplex was 97.1%, specificity was 93.6%, positive predictive value (PPV) was 98.6% and negative predictive value (NPV) was 93.6%, for an overall accuracy of 97.0%. There was no significant difference between the two Table 2. TCD-Duplex agreement with the Valsalva maneuver Duplex class TCD class
0
0 1 2 3 4 5
41 5
1 11 2
2
3 1 1
3
4
5
Total (n 5 10)
27
41 16 5 5 4 29
4 1
3 1
Volume 36, Number 4, 2010
methods (p , 1 by McNemar’s test). With the 100 VMs, the difference between the two methods for patients with at least 1 MES approached significance (p , 0.07 by McNemar’s test). The sensitivity of duplex was 91.5%, specificity was 100%, PPV was 100% and NPV was 89.1%, for an overall accuracy of 95.0%. Considering TCD classes 2–5 combined (i.e., .10 MES), which are usually considered representative of ‘‘large’’ shunts (Jauss and Zanette 2000), the sensitivity of basal duplex was 93.8%, specificity was 100%, PPV was 100%, NPV was 98.8%, accuracy was 99.0% and the difference between the two methods was not significant (p , 1 by McNemar’s test). For the 100 VMs, there was no significant difference in the concordance for large shunts between the two methods (p , 0.50 by McNemar’s test). Sensitivity of duplex was 95.3%, specificity was 100%, PPV 100%, NPV 96.6% and overall accuracy 98.0%. With a more restrictive definition of a ‘‘large’’ shunt as one associated with .20 MES (i.e., classes 3 to 5 combined), the sensitivity of basal duplex was 100%, specificity was 100%, PPV and NPV were 100% each and overall accuracy was 100%. For the 100 VMs, again, there was no significant difference between the two methods (p , 0.50 by McNemar’s test). The sensitivity of duplex was 94.7%, specificity was 100%, PPV was 100% and NPV was 96.9%, for an overall accuracy of 98.0%. Microembolic signals of the CCA were often less prominent, both visually against the background, and acoustically, than TCD MES. DISCUSSION AND SUMMARY We have found that a modern duplex probe for neck vessel ultrasound insonating the CCA can be only slightly, although significantly, less sensitive than a TCD probe at the MCA level for the detection of MES during the bubble test for right-to-left shunt evaluation. This finding has two major implications. First, this test modality may be used when a bone window is not available, which happens in 6–18% of cases (Marinoni et al. 1997; Wijnhoud et al. 2008). Second, it may greatly increase the number of facilities where the bubble test can be carried out, because ultrasound devices with vascular software are much more diffuse than TCD devices. This may speed up the evaluation protocol of patients when such a test is indicated. After a brief training period, a cardiologist, an internist, a vascular surgeon or a vascular ultrasound technician may easily perform the investigation without removing the patient from the bed where an echocardiography or a study of neck vessels has just been carried out. The lower sensitivity of duplex primarily resulted in some patients with ,5 MES at the TCD VM being classified as without any shunt at duplex, an error without major consequences.
CCA duplex to detect R-L shunt d B. CENSORI et al.
Fig. 2. A curtain of MES in the Doppler spectrum and the corresponding shower of microbubbles within the common carotid lumen.
Others have tried to compare MCA bubble test results with those recorded at other peripheral arteries (Topcuoglu et al. 2003; Draganski et al. 2005; Daly et al. 2008). Those data are not comparable to ours because they were obtained by means of a TCD probe rather than a duplex probe (Topcuoglu et al. 2003; Daly et al. 2008), or through B-mode visualization of galactose microbubbles in axial sections of the CCA (Draganski et al. 2005). Interestingly, however, both B-mode echography of the CCA (Draganski et al. 2005) and TCD recording of MES at the CCA or common femoral artery (Daly et al. 2008) gave results comparable to ‘‘classical’’ MCA TCD recording. We chose the CCA, rather than the internal carotid artery (ICA) because after a few trials, it proved easier to insonate during the VM. Furthermore, in some patients the carotid bifurcation has a high position in the neck or is heavily calcified, and these factors may prevent an adequate visualization of the ICA even at rest. Finally, a more proximal insonification should theoretically be more sensitive to shunting, because when microbubbles are few they may theoretically all travel into the external carotid artery rather that the ICA, or into the anterior cerebral artery, and may therefore not be detectable at TCD of the MCA. Our experience, however, proved that the opposite was more often true. The duplex investigation was generally easy to carry out. The main difficulties were represented by neck movements at the start or at release of the VM, which required a rapid refocusing of the sample volume on the CCA. Although this did not happen often, a preliminary trial of the VM is indicated before the actual test, both to make sure that the maneuver is well-executed and to instruct the patients not to move too much while performing the maneuver. We have used second harmonic imaging only
569
to ensure optimal positioning of the Doppler sample volume over the CCA, and not to count microbubbles. Although showers of microbubbles could be seen very well on B-mode images (Fig. 2), single MES were often not paralleled by a corresponding microbubble within the CCA. This is probably because longitudinal B-mode images visualize only a bidimensional section of the arterial lumen, whereas the Doppler signal is derived from a tridimensional volume (Steinman et al. 2004). The main reason that a few microembolic signals visible at TCD were missed by the duplex technique probably lies in the incomplete coverage of the CCA lumen by the Doppler sample volume. Because the sample volume is positioned according to the B-mode image, it is possible that part of the lumen is actually left out of the detection volume. In some instances, on the contrary, a few MES detected in the CCA Doppler spectrum were not followed by corresponding MES in the MCA tracing. This phenomenon may have several explanations: (i) microbubbles travelling along the CCA may collapse before reaching the MCA; (ii) they may end up in the anterior cerebral artery, rather than the MCA; or (iii) they may enter the external rather than the ICA. On B-mode images we have indeed seen many microbubbles going into the external carotid artery in patients with showers of MES. Although this technique gave satisfactory results, it may be more operator-dependent than TCD, at least in regard to holding the probe in the best position at rest and especially after the VM. Furthermore, duplex probes have not been designed to facilitate MES identification, as is the case with dual-gate TCD probes for automated embolus detection. Therefore, some efforts will have to be made to test interobserver reproducibility of results and to optimize ultrasound parameters for MES detection. It may also be worthwile to test whether the delay between injection and first appearance of MES may give information about shunt site, although this strategy is of doubtful usefulness with TCD (International Consensus 1998). In conclusion, our study shows that second harmonic imaging duplex of the CCA is slightly, although significantly, less sensitive than TCD for the detection of MES. Thus, it might be used for the bubble test when there is no cranial window for TCD. Also, because of very high specificity and good sensitivity of the test to detect large shunts, institutions that do not possess a TCD device may consider the Duplex bubble test at the CCA in a first instance. However, a conventional transesophageal echocardiogram (TEE) or a TCD investigation are necessary when there are no or few MES at duplex, because some small shunts may be missed and a few large shunts may be underestimated by this method. Furthermore, TEE is necessary to confirm the cardiac site of the shunt and to properly define the anatomical characteristics of the interatrial communication.
570
Ultrasound in Medicine and Biology
REFERENCES Aslam F, Shirani J, Haque AA. Patent foramen ovale: Assessment, clinical significance and therapeutic options. South Med J 2006;99: 1367–1372. Censori B, Partziguian T, Casto L, Camerlingo M, Mamoli A. Doppler micorembolic signals predict ischemic recurrences in symptomatic carotid stenosis. Acta Neurol Scand 2000;01:327–331. Cottin V, Dupuis, Girod S, Lesca G, Cordier JF. Pulmonary vascular manifestations of hereditary hemorrhagic telangiectasia. Respiration 2007;74:361–378. Daly KJ, Hutchinson S, McCollum CN. Can a peripheral artery be used to detect venous to arterial circulation shunts? Cerebrovasc Dis 2008; 25:111–114. Del Sette M, Dinia L, Rizzi D, Sugo A, Albano B, Gandolfo C. Diagnosis of right-to-left shunt with transcranial Doppler and vertebrobasilar recording. Stroke 2007;38:2254–2256. Di Tullio M, Sacco RL, Venketasubramanian N, Sherman D, Mohr JP, Homma S. Comparison of diagnostic techniques for the detection of a patent foramen ovale in stroke patients. Stroke 1993;24:1020–1024. Draganski B, Blersch W, Holmer S, Koch H, May A, Bogdahn U, Ho¨lscher T, Schlachetzki F. Detection of cardiac right-to-left shunts by contrast-enhanced harmonic carotid duplex sonography. J Ultrasound Med 2005;24:1071–1076. International Consensus Group on Microembolus Detection. Consensus on microembolus detection by TCD. Stroke 1998;29:725–729.
Volume 36, Number 4, 2010 Jauss M, Zanette E. Detection of right-to-left shunt with ultrasound contrast agent and transcranial Doppler sonography. Cerebrovasc Dis 2000;10:490–496. Marinoni M, Ginanneschi A, Forleo P, Amaducci L. Technical limits in transcranial Doppler recordings: inadequate acoustic window. Ultrasound Med Biol 1997;23:1275–1277. Perren F, Savva E, Landis T. Trasforaminal Doppler: An alternative to transtemporal approach for right-to-left cardiac shunt assessment. J Neurol Sci 2008;273:49–50. Peters B, Ewert P, Schubert S, Abdul-Khaliq H, Lange PE. Rare case of pulmonary arteriovenous fistula simulating residual defect after transcatheter closure of patent foramen ovale for recurrent paradoxical embolism. Catheter Cardiovasc Interv 2005;64:348–351. Sastry S, Daly K, Chengodu T, McCollum C. Is transcranial Doppler for the detection of venous-to-arterial circulation shunts reproducible? Cerebrovasc Dis 2007;23:424–429. Steinman AH, Lui EYL, Johnston WK, Cobbold RSC. Sample volume shape for pulsed-flow velocity estimation using a linear array. Ultrasound Med Biol 2004;30:1409–1418. Topcuoglu MA, Palacios IF, Buonanno FS. Contrast M-mode power Doppler ultraosund in the detection of right-to-left shunts: Utility of submandibular internal carotid artery recording. J Neuroimaging 2003;13:315–323. Wijnhoud AD, Franckena M, van der Lugt A, Koudstaal PJ, Dippel ED. Inadequate acoustical temporal bone window in patients with a transient ischemic attack or minor stroke: Role of skull thickness and bone density. Ultrasound Med Biol 2008;34:923–929.