In recent years, there have been important advances in the pathogenesis and treatment of atherosclerosis and in assessing prognosis in carotid atherosclerosis. The effect of the intestinal microbiome on atherosclerosis has revolutionized thinking about diet ¹ and about the role of renal failure in increasing cardiovascular risk. In the past, routine treatment with usual therapy for atherosclerosis has reduced cardiovascular risk in most trials by only ~9–30% ² , resulting in a residual risk of 70–80% ^{3– 6} . Recent advances in lipid-lowering therapy, based on a novel mechanism based on blocking the effects of an enzyme that destroys receptors for low-density lipoprotein (LDL) cholesterol, make it possible to lower LDL to a greater degree and in more patients than was previously possible ⁷ . Lifelong reduction of LDL resulting from a hereditary cause of low levels of LDL results in a reduction of coronary risk by ~95% ^{8, 9} , and a new approach to therapy based on “treating arteries instead of treating risk factors” ¹⁰ reduced the very high risk in patients with asymptomatic carotid stenosis by more than 80% ¹¹ . Most patients (~90%) with asymptomatic carotid stenosis would be better treated by intensive medical therapy than by carotid endarterectomy (CEA) or carotid artery stenting (CAS) ¹² . Methods to identify the few (10–15%) who could benefit from intervention are being developed ¹³ . In this review, I focus on advances in the understanding of the role of the intestinal microbiome and renal impairment on atherosclerosis, measurement of carotid plaque burden, carotid ulceration and ulcer volume, plaque texture, and detection of microemboli by transcranial Doppler (TCD).

Atherosclerosis may be thought of as a response to injury ^{14– 16} , related to flow disturbances that injure the endothelium ¹⁷ , followed by adhesion of platelets, penetration of macrophages into the subendothelium, inflammation ¹⁸ , oxidative stress, LDL oxidation, and proliferation of smooth muscle cells, similar to a scar in the artery wall. Risk factors identified in the Framingham Heart Study included hypertension, smoking, elevated LDL, diabetes, and left ventricular hypertrophy (essentially reflecting the integral of blood pressure over recent years). The importance of diet has been largely unappreciated, perhaps because lowering of fasting LDL with drugs such as statins has seemingly countered any effect of diet on fasting lipids. However, this misplaced focus on fasting lipids misses the key effects of diet, which occur during the post-prandial state ¹⁹ . A high-fat/high-cholesterol meal increases arterial inflammation ^{20, 21} and oxidative stress, and impairs endothelial function ²² , for several hours. Most of the day is spent in the post-prandial state, so diet is much more important than would be predicted by fasting LDL.

Several studies have shown that healthy lifestyle choices (not smoking, moderate alcohol intake, regular exercise, consuming a healthy diet, and maintaining a healthy weight) markedly reduce the risk of stroke. In the US Health Professionals Study and the Nurses’ Health study, persons who adopted all of these choices had an 80% reduction of stroke ²³ ; Swedish men with coronary artery disease who did so also had an 80% reduction of recurrent myocardial infarction ²⁴ . The Cretan Mediterranean diet reduced cardiovascular events by 70% in secondary prevention ²⁵ and reduced stroke by nearly half in high-risk primary prevention ²⁶ . It has been clear for many years that dietary cholesterol increases coronary risk ²⁷ . Now there is a completely new window opening on the relationship between diet and atherosclerosis, relating to the intestinal microbiome ¹ .

In recent years, it has become clear that an important interaction between diet and the intestinal microbiome brings into play additional metabolic factors that aggravate atherosclerosis beyond dietary cholesterol. This may help to explain the benefits of the Mediterranean diet. Hazen’s group from the Cleveland Clinic reported that phosphatidylcholine from egg yolk ²⁸ and carnitine ²⁹ from animal flesh (four times as much in red meat as in fish or chicken) are converted by intestinal bacteria to trimethylamine (the compound that causes uremic breath to smell fishy). Trimethylamine is oxidized in the liver to trimethylamine N-oxide (TMAO) ( Figure 1), which causes atherosclerosis in animal models. In patients referred for coronary angiography, high levels of TMAO following a test dose of two hard-boiled eggs markedly increased risk. Patients in the top quartile of TMAO had a 2.5-fold increase in the 3-year risk of stroke, death, or myocardial infarction ³⁰ .

A key issue is that vegans who consumed L-carnitine did not produce TMAO because they did not have the intestinal bacteria that produce TMA from carnitine ²⁹ ; this means that the intestinal microbiome is modifiable. A novel approach to treating atherosclerosis would be the eradication of harmful bacteria with antibiotics and recolonization by stool transplantation with beneficial bacteria. This is entirely analogous to the treatment of Clostridium difficile infection by “repoopulation” ³¹ . Our group is studying that possibility.

Patients with renal failure have high levels of total homocysteine (tHcy), but this accounts for only ~20% of the effect of impaired renal function on carotid plaque ³² . Elevated levels of thiocyanate and asymmetric dimethylarginine (a nitric oxide antagonist) add further to the very high cardiovascular risk ^{33, 34} in renal failure; this is not new. What is recently recognized is that metabolic products of the intestinal microbiome are excreted in the urine, so patients with renal failure also have high levels of TMAO, which accelerate the decline of renal function and increase cardiovascular risk ³⁵ . Besides TMAO, other metabolic products of the intestinal microbiome that probably contribute to cardiovascular risk in renal failure include indoxyl sulfate, indole-3-acetic acid, p-cresyl sulfate ³² , and phenylacetylglutamine ³⁶ . Thus patients at risk of cardiovascular disease should limit their intake of meat and egg yolk not only because of the high cholesterol content but also because of the carnitine in meat (particularly red meat) and the phosphatidylcholine in egg yolk. This is particularly important in patients with renal failure.

Although it has been reasonably clear that patients with severe symptomatic carotid stenosis benefit from CEA or CAS, the periprocedural risk of stroke or death with CAS is approximately twice that with CEA. However, the risk of asymptomatic carotid stenosis with modern medical therapy has declined markedly in recent years to ~0.5% per year ^{11, 45, 46} , so it has even been suggested ⁴⁷ that randomized trials in symptomatic stenosis should be repeated comparing intervention with intensive medical therapy. There is a major unwarranted controversy regarding the treatment of asymptomatic carotid stenosis. In the United States, ~90% of carotid intervention is for asymptomatic stenosis, even though 90% of patients would be better treated with intensive medical therapy ¹² . This is being justified by an invalid extrapolation from the medical risks in randomized trials conducted decades ago to modern risks with CEA and CAS. Although the most recent trials report that the risk of stroke or death after first deducting periprocedural risk is similar to that with modern intensive medical therapy (~0.5% per year), the risk with intervention is still much higher than the risk with medical therapy once the periprocedural risks are taken into account: ~3% with CAS and 1.5% with CEA. Furthermore, there are important caveats regarding the higher risk in real-world practice as opposed to the risks with carefully vetted interventionalists in randomized trials ^{48, 49} . There are huge discrepancies across the world in the proportion of carotid interventions performed for asymptomatic stenosis: 90% in the US, ~60% in Italy and Germany, ~15% in Canada and Australia (about right), and 0% in Denmark. These discrepancies call into question not only the advisability but also the ethics of routine intervention for asymptomatic stenosis, as practiced in the United States. The reasons for this practice do not bear scrutiny ^{50, 51} .

Approximately 10–15% of patients with asymptomatic stenosis could benefit from intervention, and fortunately there are methods available to identify them. Reduced cerebrovascular blood flow reserve is one approach that appears to be promising. In development are a number of other approaches based on imaging of vulnerable plaque, including intraplaque hemorrhage on magnetic resonance imaging (MRI) scans, neovascularity of plaques with contrast ultrasound, and plaque inflammation on positron emission tomography/computed tomography (PET/CT) scans ¹³ . Here, I will focus on the uses of ultrasound in assessing prognosis in patients with carotid atherosclerosis.

Although widely regarded as “preclinical atherosclerosis”, carotid intima-media thickness (IMT) is a different phenotype. This issue is complicated by two different approaches to the measurement of IMT—with and without plaque thickness—(the latter being a one-dimensional measurement of plaque in those participants with plaque). Then in studies in which plaque thickness is included, participants with and without plaque thickness are combined, conflating the issue ⁵² . It is increasingly clear that measuring carotid plaque burden is superior to measuring IMT, both for risk stratification and for assessment of effects of therapy ⁵³ . Plaque burden can be measured as total plaque area (TPA) ⁴² (the sum of the areas of all plaques seen in the extracranial carotid arteries) or total plaque volume (TPV) ⁵⁴ . In the High Risk Plaque study, 3D plaque burden was highly correlated with coronary calcium, whereas IMT was not ⁵⁵ , and plaque burden predicted risk of cardiovascular events to a similar extent as coronary calcium ⁵⁶ . Often plaque volume is measured not as TPV but as the volume over a defined segment of the carotids, limited to a defined distance above and below the bifurcation. This approach has advantages for such purposes as assessing effects of therapy (simplicity, potential for automatic measurement) but may lose dynamic range, a potential issue with regard to prognostic value. Besides the issue of IMT not truly representing atherosclerosis, a key reason why IMT is a weak predictor of cardiovascular risk ⁵⁷ is its narrow dynamic range: ~0.5 to 1.5 mm. The dynamic range of TPA is much greater – from 0 to ~1200 mm ²; the range of TPV would be even greater. After adjustment for age, sex, blood pressure, smoking, serum cholesterol, diabetes, homocysteine, and treatment of blood pressure and cholesterol, TPA strongly predicted risk among patients attending a vascular prevention clinic: the 5-year risk of stroke, death, or myocardial infarction, by quartile of TPA, was 5.6%, 10.7%, 13.9%, and 19.5%. Plaque progression also increased risk; patients with plaque progression in the first year of follow up had twice the risk of those with stable plaque or plaque regression. These findings were substantiated in a population-based study in Tromsø, Norway ^{58, 59} . Myocardial infarction and stroke were both strongly predicted by TPA but not by IMT in the common carotid where there was no plaque. The annual change in IMT is too small in relation to the resolution of the method to permit measurement in time frames that are clinically meaningful: the annual change is ~0.15 mm, and the resolution of the method is ~0.2–0.3 mm. A much more useful measurement in persons who do not have plaque would be to measure vessel wall volume ( Figure 3) ⁶⁰ ; this parameter amounts to a 3D measurement of the intima-media, has a much greater dynamic range, and unlike IMT, is sensitive to effects of therapy ^{61, 62} . Progression of IMT does not predict risk ⁶³ , even in large populations, whereas progression of 3D plaque volume predicted cardiovascular events in a small population of patients in our study ⁶⁴ .

Perhaps the best validated method for identifying high-risk patients with asymptomatic carotid stenosis is TCD embolus detection ^{12, 65} . Figure 4 shows a microembolus in a patient with asymptomatic stenosis. Patients with asymptomatic carotid stenosis who had two or more microemboli in 1 hour of monitoring had a 1-year risk of stroke of 15.6%, indicating that they could benefit from CEA or CAS. The cost of a TCD machine is less than the cost of two CAS, and training and certification in TCD embolus detection can be obtained in a course of 3 or fewer days. Thus TCD embolus detection or some other procedure to identify the patient as being at a higher risk than that with intervention should be considered before patients undergo CAS or CEA for asymptomatic stenosis.

In patients with asymptomatic carotid stenosis, the presence of three or more ulcers in either or both carotid arteries carried a similar risk as the presence of microemboli: an 18% 3-year risk of stroke or death. Those with two or more microemboli had a 20% 3-year risk ⁶⁶ . By combining TCD embolus detection and detection of three or more ulcers, the proportion of patients with asymptomatic stenosis who could benefit from intervention was increased from 5% to 10%. In the Asymptomatic Carotid Emboli Study (ACES) ⁶⁷ , plaque echolucency at baseline increased the risk of ipsilateral stroke (hazard ratio [HR] 6.43, 95% confidence interval [CI] 1.36–30.44, P=0.019). A combination of plaque echolucency and presence of TCD microemboli markedly increased the risk of ipsilateral stroke (HR 10.61, 95% CI 2.98–37.82, P=0.0003). This association remained significant after controlling for risk factors, degree of carotid stenosis, and antiplatelet medication. Juxtaluminal black plaque (plaque or thrombus that is so echolucent that it can be seen only by observing a gap between the wall and the Doppler flow signal) ⁶⁸ , plaque echolucency, and plaque texture analysis of ultrasound ⁶⁹ identified higher risk.

Even in patients without carotid stenosis, plaque characteristics predict cardiovascular risk. In a relatively small sample (349) of patients attending a vascular prevention clinic, followed for 5 years, the volume of ulcers ( Figure 5) predicted risk of cardiovascular events ⁷⁰ . Patients with a total ulcer volume ≥5 mm ³ experienced a significantly higher risk of stroke, transient ischemic attack, or death (P=0.009) and of stroke/transient ischemic attack/death/myocardial infarction/revascularization (P=0.017). In the same patient population, elements of plaque texture identified from radiofrequency signals in ultrasound ( Figure 6) predicted cardiovascular risk ⁷¹ .

Intraplaque hemorrhage on MRI ^{72– 74} appears promising as a method for identifying high-risk asymptomatic stenosis, as does imaging of plaque inflammation on PET/CT with fluorodeoxyglucose ⁷⁵ and imaging of active calcification with sodium fluoride ⁷⁶ ( Figure 7). It seems likely that by combining several imaging modalities, the proportion of patients with asymptomatic stenosis who could be identified as being at high enough risk to warrant intervention might be increased to ~15%. Just as it is inappropriate to perform CAS or CEA in all patients with asymptomatic stenosis, it is also inappropriate to intervene in none: ~10% of strokes are from carotid stenosis, and most patients were asymptomatic before the event. Modalities to identify the few who could benefit from intervention should be in more widespread use; indeed no patient with asymptomatic stenosis should be subjected to intervention without first being identified as being at a high enough risk of ipsilateral stroke to benefit from intervention.

Atherosclerosis is a complex process. Diet is much more important than most physicians suppose, and the intestinal microbiome has major effects on metabolic products derived from dietary constituents such as carnitine from meat and phosphatidylcholine from egg yolk. Avoidance of red meat and egg yolk is particularly important in patients with impaired renal function. New approaches to lowering LDL by blocking the effect of PCSK9, and a strategy of treating atherosclerosis directly instead of focusing on intermediate targets, show promise of reducing the residual risk that remains after current therapy. Most patients with asymptomatic carotid stenosis would be better treated with intensive medical therapy than with CEA or CAS. Microemboli on TCD, impaired cerebral blood flow reserve, juxtaluminal black plaque, echolucency, intraplaque hemorrhage on MRI, and ulceration on 3D ultrasound are ways to identify the 10–15% of patients with asymptomatic carotid stenosis who are at sufficiently high risk to benefit from intervention. Other approaches such as plaque texture on ultrasound and PET/CT imaging of inflamed plaque and early calcification will in future provide further evidence of their contribution to identifying high-risk asymptomatic carotid stenosis.

CAS, carotid artery stenting; CEA, carotid endarterectomy; CI, confidence interval; HMG CoA, hydroxymethylglutarate CoA; HR, hazard ratio; IMT, intima-media thickness; LDL, low-density lipoprotein; MRI, magnetic resonance imaging; PCSK9, proprotein convertase subtilisin-kexin type 9; PET/CT, positron emission tomography/computed tomography; TCD, transcranial Doppler; TMAO, trimethylamine N-oxide; TPA, total plaque area; TPV, total plaque volume.