Response to Letter by Syme
We have read with interest the comments by Dr Syme regarding our recent study investigating the potential association of knock-type-Doppler signal (KTDS) with the vessel affected by ischemia and the presence of microvascular ischemia on brain MRI.1 Dr Syme has made some provocative observations that we would like to address.
First, we did not use DWL-Compumedics Transcranial Doppler (TCD) machines. This type of a TCD unit was previously used to detect knock signals in the original case reports by Dr Syme.2,3 We believe that case reports do not offer any solid evidence to choose one TCD make over the other. In fact, for the knock signal to be truly generalizable, it has to be detected by all commercially available TCD devices because they share the same carrying frequency.
Second, Dr Syme pointed out that the filter we used for KTDS detection (100 Hz) was too high, and therefore many signals may have been missed because of the high-pass filter setting. Filters are used routinely to avoid low frequency noise. If a methodology is developed to detect knock type signals, it should then include some adjustment to make sonographers work easier and change their routine less. Besides routine diagnostic evaluations, our rationale for choosing this specific high-pass filter setting was related to the fact that this is the cut-off recommended by the consensus on microembolus detection criteria.4 In addition, Dr Syme has previously reported the knock signals to be found in the “±300 Hz of the spectrum.”2 Moreover, Dr Syme has not provided any additional details regarding the preferred, if any high-pass filter settings for detection of knock signals3 before the publication of our study. Therefore, we considered that the recommended high-pass filter setting of 100 Hz was well below the spectrum of 300 Hz at which knock signals were initially observed. If one considers the fact that most routine TCD examinations are performed with routine filter deployment, our finding of a variety of KTDS indicates that filter removal would probably result in even greater number of these signals. Thus KTDS correlation, if any, with clinical symptoms could offer even poorer positive predictive values that can be expected from our study.
There are insufficient data to recommend that all TCD exams should be done without filters because the overall value of the knock signal is unclear. We do not dismiss the possibility that a few cases described by Dr Syme offer an unprecedented insight into pathophysiology of microvascular ischemia and resulting symptoms. Cases are better evidence than just an opinion, and cases offer an opportunity to think about the problem. Despite the impression that cases could be striking and provocative, they are not sufficient to determine whether indeed a microvascular occlusion can be detected by TCD and that revascularization is induced by TCD. We need a controlled clinical trial to show or refute this. In our opinion, however, the current definition of knock signals is far from perfect, and its application to routine TCD will result in too many false-positive results that have nothing to do with patient symptoms. The Doppler signature of microscopic vessel pulsation is also far from established, and further in vitro and in vivo validation and reproducibility studies are required.
Third, Dr Syme is pointing out that the lack of association of KTDS detected by TCD and microvascular ischemia on brain MRI may be attributed to the fact that distal regions of intracranial arteries cannot be interrogated by TCD. However, the purpose of our study was to investigate the clinical relevance of KTDS detected by TCD and their potential association with symptoms of cerebral ischemia (for which MRI is routinely performed and is currently considered the most sensitive imaging to localize ischemia). Therefore, we consider that our study design was both clinically relevant and methodologically appropriate in demonstrating the absence of any relationship between KTDS and microvascular ischemia attributed to an occlusion of a small perforating artery.
Fourth, our interpretation of the study of Chung et al investigating the origins of knock signals in an in vitro model5 is different than the one proposed by Dr Syme. In our opinion, their study is the first step to uncover the complexity of modeling Doppler signals from vessels of different calibers and patency. It provides experimental evidence indicating that knock signals are not specific “flow signatures” because they can be reproduced in both obstructed and nonobstructed vessels.
In conclusion, both our report and the study by Chung and colleagues highlight the need to further determine what really constitutes a small vessel knock (rather than a common vessel motion artifact) and a proper triple-blind validation study of knock signals (with sonographers unaware of clinical localization, and physicians unaware whether or not knock was present and whether or not therapeutic insonation has been done) may be the most appropriate way to investigate with scientific rigor the clinical relevance of KTDS.
Tsivgoulis G, Man BL, Lao AY, Sharma VK, Kotsis V, Vadikolas K, Alexandrov AV. A spectrum of knock-type Doppler signals in the intracranial vessels. Stroke. 2009; 40: 644–647.
Syme PD. Detection of small vessel knock using transcranial Doppler ultrasonography. Advance Clin Neurosci Rehab. 2004; 4: 28–31.
Ringelstein EB, Droste DW, Babikian VL, Evans DH, Grosset DG, Kaps M, Markus HS, Russell D, Siebler M. Consensus on microembolus detection by TCD. International Consensus Group on Microembolus Detection. Stroke. 1998; 29: 725–729.
Chung EML, Ramnarine KV, Long CV, Udommongkol U, Chambers BR, Gittins J, Bush G, Evans DH. Doppler ultrasound detection of side-vessel occlusion: an in vitro study. Stroke. 2009; 40: 648–651.