Reply: Abdominal aortic calcification on preoperative CT as a cardiovascular risk marker in liver transplant candidates

To the editor, We appreciate the interest of Cheema and colleagues in our study evaluating the utility of abdominal aortic calcification (AAC) in predicting major adverse cardiovascular events (MACE) following liver transplantation (LT).1 We agree with the authors that this single-center study does not definitively declare AAC to be an effective risk assessment tool in LT assessment, but rather opens a new avenue for research, which may allow us to improve our risk prediction algorithms in the future. One potential issue raised was the inclusion criteria of our study, which excluded patients with missing imaging. However, it should be noted that there were no differences in the rates of cardiovascular risk factors, including hypertension, dyslipidemia, diabetes, or smoking, between included and excluded patients. It is therefore felt unlikely that this would have introduced a significant risk of bias. Similarly, patients with known coronary artery disease were excluded. We believe this was appropriate, as these patients should undergo formal coronary assessment before LT regardless of their AAC and are therefore not the target group for this screening test. The authors also point out that our calculated AAC threshold requires external validation, a statement with which we wholeheartedly agree, and this was listed as a limitation of our study in the published manuscript. However, every new clinical tool requires an initial derivation cohort, and further assessment of this threshold in an external cohort of patients will be invaluable to help further develop our understanding of this new metric. In addition, we agree that the Revised Cardiac Risk Index (RCRI) has not been demonstrated to have good predictive power for post-transplant MACE in the LT cohort, as Cheema and colleagues point out, other traditional surgical risk scores, including the National Surgical Quality Improvement Program (NSQIP), the Charlson Comorbidity Index, or the American Society of Anesthesiologists scores.2,3 This is the very reason why this field needs to be advanced by the inclusion of novel metrics such as AAC, and while “outperforming a low benchmark” does not represent the end goal for this technology, it provides a starting point from which to build future improved risk assessment tools. Of the author’s other suggested comparators, instead of RCRI, MELD was not a significant predictor of MACE in our cohort and hence was not included in the multivariable analysis.1,4 Furthermore, we do not believe that it is realistic to expect AAC to outperform formal computed tomography coronary angiography (CTCA) as a predictor of coronary disease, and this is not the goal of this screening tool, so this comparison was not performed.5 However, we did demonstrate that AAC was a significant predictor of both coronary artery calcium score and moderate-to-severe disease on CTCA following multivariable analysis.1 Finally, with regard to Cheema and colleagues’ suggestion to provide a stratified analysis by MACE subtype, we would refer them to page 5, paragraph 3 of the original manuscript, where these data are presented. Thank you once again for your interest in our manuscript.