Response to Letters Regarding Article, “Aortic Arch Plaques and Risk of Recurrent Stroke and Death”
To the Editor:
We did not understate, as Drs Bhindi and Ormerod claim, the potential effect of statins on the risk of recurrent embolic events. We simply observed that, because patient enrollment in our study ended in 1998, statin use was not prevalent at the time the study was conducted.1 If this prevented us from taking into account the effect of statins on the results, it also provided us with an opportunity to evaluate the effect of antithrombotic treatment without the possible confounding factor of statin treatment. It should be noted that, at the time the study was conducted, lipid-lowering treatment was not routinely prescribed after a stroke, and the very relationship between serum cholesterol and stroke risk was questioned. The detection of large aortic plaques may have resulted in the initiation of lipid-lowering treatment, but such treatment would have probably led to the underestimation of the risk of recurrent events related to the presence of plaques. Therefore, the main result of our study, which is the persistence of plaque-related embolic risk despite treatment with warfarin or aspirin (rather than the comparison between these two treatments), remains valid.
Warfarin and aspirin were the sole antithrombotic treatments allowed in the study. Therefore, no patient received clopidogrel or any other antiplatelet agents in addition to aspirin. Unfortunately, data on the severity of recurrent strokes were not available, as is very often the case in multicenter studies of this type. Finally, all analyses on the risk of events related to plaque size were performed with all plaque size categories entered in the same model and were adjusted for the presence of other risk factors, thus eliminating the possibility of a significant influence of risk factors’ imbalance on the results. In the article, only the hazard ratios for plaques ≥4 mm were reported because of their clinical relevance and to allow an easy comparison with the results of previous studies.
Drs Markl and Harloff suggest that an imperfect visualization of the aortic arch by transesophageal echocardiography and the disregard of the possibility of retrograde embolization from plaques located in the proximal portion of the descending aorta may have resulted in an underestimation or misestimation of the embolic risk of the aortic plaques. Transesophageal echocardiography may indeed miss plaques located in a small segment of the proximal arch, although the magnitude of the resulting underestimation may be modest.2 In contrast, magnetic resonance imaging may overestimate the plaque size due to its inability to adequately resolve the wall layers because of limited spatial resolution, as Drs Markl and Harloff report in their article. While the diagnostic accuracy for detecting aortic atherosclerosis may further improve with advancements in imaging technology, transesophageal echocardiography remains at this time the test of choice for assessing the presence of aortic arch plaques and the one on which most studies on the associated risk of stroke have relied.
The possibility of retrograde embolization to the brain from plaques located in the proximal descending aorta has been known for some time,3 although its actual clinical relevance is unknown. In our study, large plaques (≥4 mm) in the proximal portion of the descending aorta were not independently associated with recurrent events after adjustment for other risk factors (adjusted hazard ratio 1.78, 95% confidence interval 0.79 to 4.02). This is indeed in agreement with the same article by Dr Amarenco et al that Drs Markl and Harloff cite,4 in which the odds ratio of 5.5 was in fact associated with plaques located in the distal arch (ie, proximal to the takeoff of the left subclavian artery; see Figure 1 of that article), not in the proximal descending aorta. Plaques in the descending aorta were instead associated with a nonsignificant odds ratio for stroke of 1.5, quite similar to that observed in our study. Therefore, the speculation that an underestimation of embolic risk due to retrograde embolization from the descending aorta may affect the results of our study is not supported by the data.