Fat and cardiovascular risk: the role of Cardiac CT

The role of computed tomography in cardiovascular risk stratification is well established by means of coronary calcium score (CACS). This role is, as said, well established in asymptomatic individuals,1,2 but is more controversial in symptomatic patients; in fact, it has been more difficult to extract evidence of incremental value in patients with symptoms.3,4 At the same time, after the introduction of coronary computed tomography (CCT), CACS has shown strength in cardiovascular risk stratification on top of conventional risk factors that CCT did not improve significantly, especially in asymptomatic individuals.5–8 Eventually, it has been already demonstrated that a modified CACS can be extrapolated even for low-dose thoracic scans performed for other purposes.9–12 More recently, epicardial fat (EF) has developed into a reliable (yet not clinical) quantitative measurement of cardiovascular risk. It has been found that EF is associated with greater amounts of CACS. In this issue of the Journal, Wolff et al., provide an insight into the relationship between liver fat (LF), cardiovascular risk factors and subclinical coronary artery disease in the general population.13 The authors conclude that larger amounts of LF are related to larger volumes of EF and CACS, independent of traditional cardiovascular risk factors, providing important and novel insights into the role of LF as a marker of vascular disease. CT measurements are quite reliable, especially when we refer to volumetric assessment and to average attenuation into a sufficiently large region of interest. As compared with other imaging modalities such as echocardiography or magnetic resonance, CT is more reliable on volumetric assessment and on absolute attenuation measurements, especially in a homogeneous and non-hypervascular parenchymal organ or tissue. Wolff et al. assessed LF using a pragmatic approach, surely not the ideal one, as acknowledged by authors.13 They performed multiple sampling of liver attenuation using a small field of view images from CACS scans (i.e. low dose); hence, the whole liver was not assessed, CT scan settings were also not ideal for liver imaging (normally higher milliamperes are required) and noise was increased. However, since average attenuation was the endpoint, we can consider the method used as the best solution in this case. Wolff et al. stated that non-alcoholic fatty liver disease (NAFLD) has become an important precursor of the metabolic syndrome and related to cardiovascular events.13 However, it is not easy to have a complete picture of all the relationships that might occur in this area. In fact, some of the findings may be simple associations (having a common background metabolic alteration), whereas others may actually play a significant role as standalone risk factors and/or predisposing conditions. We also need to consider the variability of the populations that have been studied to date. Young asymptomatic individuals vs. older symptomatic patients are different both regarding the risk profile and the metabolic ‘milieu’.13 The assessment of relationship between LF and the rest of cardiovascular parameters, including EF, should focus on asymptomatic individuals especially when we search for more sensitive or specific indexes of risk that may predict outcome.14–16 Nevertheless, the assessment of symptomatic patients may bring some unexpected findings such as for the CORE320 study in which the authors found no association between EF volume and the presence/severity of coronary artery disease; they conclude that EF volume may relate to the pathophysiology of acute coronary events rather than the presence of atherosclerotic disease.16 In addition, this is exactly the problem with wide and complex syndrome with several parameters and conditions involved.17 When we look at LF, we should consider that we might just look at a satellite alteration that plays a role but that it is not influencing primarily the other factors. It is more or less similar with EF volume and ultimately also with CACS; they are phenomena that appear and probably concur in the whole risk determination, but they might be just a bit redundant. To overcome some of these issues and answer some of these questions, we need to correlate LF with major adverse cardiovascular events and lipid-lowering treatment. This might shed a bit more light into the actual role of this parameter. Another issue to consider is in the measurements. There is no standardised approach to measure LF in CT. As stated above, several parameters and different strategies may change the actual quantification. For instance, one might consider that LF associated with liver volume could improve the prognostic value of the information; this could be declined as a kind of LF volume surrogate. There could be several different approaches that have to be verified. The authors of the Rotterdam Study followed a pragmatic approach with data that were available for all the patients.13 The galaxy of cardiovascular risk assessment expands with the introduction of newer technologies and newer approaches. Hence, correlating all the parameters and finding ‘who comes first’ in terms of pathophysiology and/or risk prediction become more difficult. Larger multi-parametric outcome studies and registries should be undertaken to improve our field of view on the matter.