Jeroen Brouwers

In this thesis we emphasize the importance of accurate non-invasive tools in atherosclerotic vascular diseases. Continuous improvement of these modalities is crucial in diagnostics and patient selection. Therefore, we investigate the new Doppler ultrasonography (DUS) parameter maximal systolic acceleration (ACCmax). Multiple clinical manifestations related to atherosclerosis can occur based on the location of the plaque. For example, cerebral ischemic attacks can occur as a result of a carotid artery stenosis, and an atherosclerotic stenosis in the lower extremity can lead to claudication intermittent. In accordance, in this thesis there is a distinction made in atherosclerotic vascular disease: peripheral arterial disease (PAD) of the lower extremity, carotid artery stenosis, and renal artery stenosis (RAS).

Part I Peripheral arterial disease of the lower extremity

PAD can primarily be assessed by the ankle brachial index (ABI), toe brachial index (TBI) and toe pressure (TP). However, in patients with diabetes mellitus (DM) or chronic kidney failure (CKD), medial arterial calcification (MAC) can occur leading to incompressible arteries and unreliable test results. Therefore, ABI, TBI and TP will not provide an adequate estimation of the blood flow to foot and toes in these patients. The prevalence of PAD in people with DM is 20-30%, and increases to 65% in patients with diabetic foot ulcer (DFU). MAC occurs in a third of the patients with DM, and is detected in 70% of the amputations for critical limb ischemia. It is expected that the number of patients with DM will increase to nearly 370 million people by 2030 worldwide. An accurate diagnosis of PAD in patients with MAC is important because timely recognition of critical limb ischemia is pertinent to reduce delayed wound healing, and prevent (major) lower limb amputation and mortality. Furthermore, early identification of PAD is essential to promptly start cardiovascular risk management (CVRM) and reduce the risk of cardiovascular events. Therefore, it is important to have reliable non-invasive bedside tests to diagnose PAD. We performed a systematic review (Chapter 2) to evaluate the reliability of non-invasive bedside tests compared to reference imaging techniques for diagnosing PAD in patients prone to MAC.

Remarkably, only 23 studies investigated the accuracy of bedside tests in patients prone to MAC. ABI is the most frequently used bedside test to diagnose PAD, and is accurate to diagnose PAD (PLR 8.72-17) when excluding patients with ABI >1.3. However, it is insufficient to exclude PAD (NLR 0.24-0.84). TBI or TP are often used in patients with MAC. However, these bedside tests are insufficient to diagnose or exclude PAD. Furthermore, regarding TP different thresholds were used to diagnose PAD (50, 70 and 97 mmHg) in several studies. Also, the definition of an abnormal test was not consistent for continuous waveform Doppler (CWD) analysis. When a loss of a triphasic signal was used to determine PAD, PAD could be accurately ruled out (NLR 0-0.09). However, most patients with MAC have dampened, monophasic or biphasic waveforms, making a triphasic signal (to exclude PAD) less useful in clinical practice. Furthermore, the presence of palpable pedal pulses was insufficient to diagnose or exclude PAD.

So, in general the performance of the different bedside tests was insufficient and variable between studies. Therefore, we counsel against the use of a single bedside test. Furthermore, the methodological quality of the included studies was low (in 20 of the 23 studies a risk of bias or a concern regarding applicability was present) and sample sizes were small. More methodologically well designed studies are needed to evaluate the performance of the conventional bedside test in patients prone to MAC. Also, it is important to investigate new non-invasive diagnostic parameters to diagnose PAD in this challenging group of patients.

To address the need for new non-invasive tools, we investigated the DUS parameter maximal systolic acceleration (ACCmax). In particular for patients with MAC, the ACCmax might benefit more compared to the conventional bedside test since no external pressure measurement is necessary. It measures the acceleration of blood flow by quantifying the maximal slope of the systolic doppler curve, as shown in figure 1. The ACCmax is always measured distal to the stenosis, for example in distal posterior tibial artery. Moreover, ACCmax measurements can be obtained in a very short time (less than 1 minute).

In chapter 3 fundamental research is shown in which ACCmax is investigated in an experimental in vitro setting. A circulatory flow system was used to imitate the human circulation. Furthermore, an arterial stenosis was made, which was built in the flow system, see also figure 1 in chapter 3. The following degrees of stenosis were used: 50%, 70%, 80%, and 90%. The ACCmax significantly decreased (P < .001), and the intraluminal mean arterial pressure gradient increased (P < .001) as the degree of stenosis increased. The correlation between ACCmax and the pressure gradient was strong (R2 = 0.937). Also, a very low interobserver variability was determined with respect to ACCmax for 2 independent investigators (ICC = 0.99). In chapter 4 we develop the validation of ACCmax further; in an in vivo study setting the ACCmax was compared to the conventional bedside test. In twelve healthy individuals an arterial stenosis was mimicked by compression on the common femoral artery by an ultrasounds probe, creating a local stenosis of 50%, 70% and 90%. The ACCmax showed the highest correlation with the degree of stenosis (r -0.884), compared to ABI (r -0.726), TBI (r -0.716) and TP (r -0.758). Also, in this study the interobserver variability of ACCmax was very low (ICC 0.97). Figure 1: Example analysis of Doppler spectrum. A, Normal triphasic Doppler waveform. Maximal systolic acceleration (ACCmax) is measured at inflection point at which upstroke changes from concave up to concave down. It is equal to slope of tangent line on curve at inflection point. B, Example of abnormal monophasic Doppler waveform. ACCmax, peak systolic velocity (PSV), acceleration time (AT), and systolic velocity gradient (∆Vsys) are shown. ACCmax is measured at visually judged maximum derivative of systolic phase. Combining the results of chapter 3 and 4 ACCmax proved to be superior to ABI, TBI, TP, ACCsys and PSV in experimental settings. ACCmax can assess the severity of a stenotic disease, and has an excellent interobserver variability. Although these studies form the basis of the validation of the ACCmax, prospective clinical studies are needed to determine the exact clinical value of ACCmax. Particularly in patients with MAC the ACCmax might benefit more compared to the conventional bedside test since there is no external pressure measurement needed in obtaining the ACCmax. Part II Carotid artery stenosis In carotid artery disease non-invasive measurements are not only important to diagnose an internal carotid artery (ICA) stenosis, but also crucial for decision making regarding carotid intervention. The international guidelines mentioned the option of DUS alone (without any other diagnostic tool) as acceptable strategy to make a decision regarding intervention. Therefore, accurate DUS measurement is of the utmost importance. Conventional DUS parameter (ICA PSV, optical estimation of the stenosis, PSV ratio, ICA EDV) measurements are vulnerable for local distorting factors. Vascular calcification resulting in acoustic shadowing can hamper these four measurements, and may lead to inaccurate results. To circumvent these local distorting factors ACCmax can be used since it is measured distal to the stenosis (in the distal extracranial ICA). A retrospective study was performed (chapter 5) to describe the first results of ACCmax in detecting ICA stenosis. The study population consisted of 947 different carotid arteries. By using the ACCmax a distinction can be made concerning stenosis categories since the ACCmax was significantly different between these stenosis categories (<50%, 50-69%, ≥70%). ACCmax decreased as the severity of stenosis increased. Furthermore, strong correlations between ACCmax and ICA PSV (R2 = 0.88) and PSV ratio (R2 = 0.87) were found, as shown in figure 3 of chapter 5. To investigate which DUS parameter is the most accurate one, a subgroup analysis was performed. The ACCmax, ICA PSV and PSV ratio were compared to CTA as reference test. Based on the area under the curves (AUC) in ROC analysis, the optimal cut-off values were calculated, as shown in table 1. For diagnosing a ≥70% ICA stenosis there were no significant differences in AUC between these three parameters. However, to diagnose a solitary ≥50% ICA stenosis the ACCmax was somewhat inferior as compared to conventional Duplex parameters ICA PSV and PSV ratio. However, the benefit of ACCmax is not in the accuracy of detecting a simple mild ICA stenosis, but in the ability to perform the measurement at an additional measuring point to avoid the influence of local distorting factors of a plaque, as mentioned before. Therefore, ACCmax can be used as additional measurement, particularly when DUS is the only modality before decision making regarding intervention. Our study presents the first ACCmax results in carotid artery disease. From here, future research can be initiated. In our opinion, the following studies should focus on: I) evaluating ACCmax results in a prospective design to determine the performance of the cut-off values and interobserver agreement, II) potential benefits of ACCmax in specific subgroups such as patients with proximal stenosis (diagnosed by a decreased ACCmax in CCA) and extensive calcific shadowing of the ICA. Diagnostic performance characteristics of DUS parameters to identify ICA stenosis compared to CTA as reference test: To diagnose ≥ 50% ICA stenosis: ACCmax: AUC 0.88 (0.82-0.94), Optimal cut-off 7.15, Sensitivity 82%, Specificity 88%, PLR 6.83, NLR 0.20 ICA PSV: AUC 0.94 (0.91-0.97), Optimal cut-off 143, Sensitivity 93%, Specificity 87%, PLR 7.15, NLR 0.08 PSV ratio: AUC 0.94 (0.91-0.97), Optimal cut-off 1.77, Sensitivity 88%, Specificity 89%, PLR 8.00, NLR 0.13 To diagnose ≥ 70% ICA stenosis: ACCmax: AUC 0.89 (0.82-0.95), Optimal cut-off 4.05, Sensitivity 90%, Specificity 89%, PLR 8.18, NLR 0.11 ICA PSV: AUC 0.94 (0.89-0.97), Optimal cut-off 212, Sensitivity 86%, Specificity 90%, PLR 8.60, NLR 0.16 PSV ratio: AUC 0.93 (0.89-0.97), Optimal cut-off 3.21, Sensitivity 90%, Specificity 89%, PLR 8.18, NLR 0.11 Table 1: Optimal cut-off values were calculated using Youden’s index. AUC = area under the curve, PLR = Positive likelihood ratio, NLR = negative likelihood ratio, ACCmax = maximal systolic acceleration in m/sec2, ICA PSV = internal carotid artery peak systolic velocity in cm/sec, PSV ratio = ICA PSV / CCA PSV Although the international guidelines mention the option of DUS alone to make a decision regarding intervention, many centers obtained additional imaging (computed tomographic angiography (CTA) or MR angiography (MRA)) when intervention is considered. Large trials (NASCET and ECST) determined thresholds for intervention based on diameter reduction. The diameter reduction method has therefore been established as the standard quantification method to assess the degree of ICA stenosis. The diameter reduction method is nowadays being applied to CTA and MRA images. Alternatively, on CTA cross sectional area measurements can be performed; this method also considers the asymmetric shapes of a stenosis. For irregular plaques, cross-sectional area reduction might provide a more accurate quantification of ICA stenosis. However, the European guideline did not mention the option to use the cross-sectional area reduction method. In daily clinical practice, some radiologists measure the cross-sectional area reduction method to estimate the degree of stenosis, but it is unclear to what extent this occurs. The objective of chapter 6 was therefore to determine which method is generally used. Hence, we generated a questionnaire study to investigate which method (reduction in diameter or area) radiologists typically use to assess the degree of ICA stenosis. Ninety-two questionnaires were analyzed. The respondents consist of 83 neuroradiologists, 8 neuroradiology residents, and 1 neurosurgeon. Our survey showed that the method used to assess degree of ICA stenosis varies according to the type of plaque. For a regular/non-ulcerated and calcified plaque the diameter reduction was used most often to determine the degree of stenosis, respectively 67% and 62%. However, for an irregular/ulcerated plaque the use of the area reduction method increased to 45% (use either the area reduction method exclusively or both the diameter reduction and area reduction methods). Interestingly, overall, a total of 42 respondents (46%) reported that they use the area reduction method—either exclusively or in addition to the diameter reduction method— for quantifying degree of carotid artery stenosis. Although the area reduction method is suggested to be a more accurate method and theoretically might express the true hemodynamic significance of the lesion better than the diameter stenosis method, it is not validated for selecting patients regarding intervention. In our opinion, as long as there is no consensus if measurement of the area reduction differs from measurement of the diameter reduction in a clinical setting, caution must be taken when the area reduction method is used to select patients for intervention. Part III Renal artery stenosis Renal revascularization might improve renal function and blood pressure control in patients with renal artery stenosis (RAS). However, technically successful revascularization does not guarantee positive responses in terms of blood pressure regulation or preservation of renal function. The objective of chapter 7 was to identify prognostic parameters that can select patients who will benefit from renal revascularization. We hypothesized that the resistive index (RI) and ACCmax might have this prognostic value. For diagnosing RAS there are various strategies. The ACCmax can be used to detect a significant RAS with a sensitivity of 83–94%. The RI is a measure of pulsatile blood flow and is modified by vascular resistance and vascular compliance. A high RI suggests no improvement in renal function and blood pressure after renal revascularization. However, it is not clear which patients will have a positive response from revascularization. Both measurements (ACCmax and RI) are obtained from the interlobar artery (intrarenal), in contrast with the PSV or PSV ratio which are measured in the main renal artery. Thirty-two patients who underwent a renal revascularization procedure were included. To distinguish a positive response from a non-responder the renal function (eGFR) and mean arterial pressure (MAP) were used six months after intervention. In total 13 combined positive responders and 19 combined non-responders were found in the study population. The combined positive responders had a significant lower median RI and lower median ACCmax than the combined non-responders. Furthermore, a prediction model was calculated (RI≤0.5 and ACCmax≤1.3 m/s2). The prediction model classifies patients as expected positive responders if the patient scored at or below the respective cut-off points on both variables. The performance of this prediction model is shown in table 2; an expected sensitivity of 69% and specificity of 89% for a positive response after renal revascularization was calculated. In conclusion, our prediction model is a new objective non-invasive instrument which has a contributing role in decision making regarding treatment of RAS. Although this prediction model might be promising, it should be evaluated in clinical prospective studies to ensure its validity. Table 2: Expected statistical characteristics of the prediction model (based on double-cross-validation computation) Sensitivity: 69% Specificity: 89% Positive predictive value: 82% Negative predictive value: 81%