Markedly Elevated Aspartate Aminotransferase from Non-Hepatic Causes

Round 1

Reviewer 1 Report

The authors of this manuscript aimed to assess the non-hepatic etiologies of elevated aspartate aminotransferase (AST) levels and whether they were associated with the increased risk of mortality. The study included 430 patients with AST levels >400 U/L not related to liver disease. Etiologies of AST elevation were categorized in 3 groups: skeletal muscle damage, cardiac muscle damage, and hematologic disorder. The most common etiology was skeletal muscle damage (54.2%), followed by cardiac muscle damage (39.1%) and hematologic disorder (6.7%). The 30-day mortality in patients with skeletal muscle damage, cardiac muscle damage, and hematologic disorders was 14.2%, 19.5%, and 65.5%, respectively. The magnitude of the peak AST level significantly correlated with the 30-day mortality, with rates of 12.8%, 26.7%, and 50.0% for peak AST levels < 1000 U/L, < 3000 U/L, and ≥ 3000 U/L, respectively. In the multivariate analysis, cardiac muscle damage (odds ratio [OR] = 2.76), hematologic disorder (OR = 9.47), peak AST < 3000 U/L (OR = 2.94), and peak AST ≥ 3000 U/L (OR = 9.61) were associated with increased 30-day mortality. The authors concluded that elevated AST activity levels (>400 U/L) result from 3 main non-hepatic etiologies and the etiology and peak AST activity levels are associated with the risk of mortality.

I congratulate the authors on performing this clinically important study. I do believe that the study fills a gap in the existing knowledge with respect to the non-hepatic etiologies of AST elevation and the association between AST and short-term mortality. I have several comments and concerns that I believe are addressable by the authors:

1.      In general, this reviewer disagrees with the analysis of baseline data based on the survival status. I advise the authors to analyze the baseline data based on AST activity levels (i.e., >median versus <median). In principle, baseline data should be categorized based on a variable that is present at the baseline in all patients (mortality was not). The vast majority of similar studies do follow the latter approach. This is the most important point that should be addressed by the authors.

2.       Muscle damage is relative common but rhabdomyolysis was not. If rhabdomyolysis was defined as a creatine kinase level >1000 U/L, then not all patients had rhabdomyolysis. How many patients had rhabdomyolysis? Creatine kinase (CK) and ALT values should be added in Table 1 and Table 2.

3.      With respect to etiology, I would have expected a greater contribution of acute coronary syndromes (especially STEMI) in the etiology of AST elevation. Is cardiac etiology under-represented? AST was the first biochemical test used to diagnose myocardial infarction and it was used until substituted by more sensitive tests.

4.      The accuracy of exclusion of baseline liver disease appears to be questionable particularly in conditions characterized by multi-organ failure or metastatic cancers or hepatic infiltration during malign hematologic disorders (this was accepted as a limitation by the authors).  

5.      The authors state: “Of the 430 patients, 112 (26.0%) were admitted to the ICU and 13  (3.0%) had hepatic decompensation at enrollment.” Does this statement contradict the previous statement that patients with hepatobiliary disease were excluded? Does it mean that these patients (who developed liver decompensation) had advanced liver disease at baseline (study entry)?

6.      Please see the possibility to use the discrimination tests to see improvement by AST with respect to 30-day mortality (ROC curve analysis; C-statistic, IDI or NRI). Considering the nature of the study a more sophisticated statistical analysis may be needed.

7.      Please indicate the number of variables entered into the multivariable logistic regression model and how they were selected?

8.      The level of statistical significance should be put at a 2-sided P value of <0.05 (not 0.05).

9.      Define acute coronary syndrome. Was it STEMI, non-STEMI or unstable angina?

10.  Please instead of “predictors of mortality” or “survival” use the term “correlate of mortality” or “survival”. The term “predictor” is too strong.

11.  The title of Figure 4 may be inappropriate. I see no decision tree here. It simply shows 30-day mortality in various groups (etiologies and peak AST). The number of events/patients in each group may be inserted.

12.  ALT was also elevated. It may also have prognostic value. Can the authors compare ALT with AST in terms of association with mortality (i.e., by comparing 2 areas under the ROC curve)?

13.  The paragraph of limitations may be expanded.  

14.  Please report risk estimates (odds ratios) with 95% confidence interval throughout the material.

 

15.  The 30-day mortality may be too short. At least it should accepted as a limitation of the study.

 

Author Response

We deeply thank the editor and reviewers of the ‘Journal of Clinical Medicine’ for taking time and efforts to review our paper. Please find the revision of our manuscript entitled “Markedly Elevated Aspartate Aminotransferase from Non-hepatic Causes” and “point by point response to each of the reviewers’ comments”. All revisions to the manuscript were marked using the “Track Changes” function. The reviewers’ comments were indicated as underlined black color, while our response was indicated as red color, and all significant changes to the manuscript were indicated as highlight in yellow. In the revised paper, we changed several descriptions according to reviewers. We hope the revised manuscript will better meet the requirements of the ‘Journal of Clinical Medicine’ for publication. Here is one-by-one response to your comments.

Author Response File: Author Response.docx

Reviewer 2 Report

Review of the paper titled Markedly Elevated Aspartate Aminotransferase from Non-hepatic Causes

Ji Hee Han, M.D et al

 

Thank you for sending above paper for review.

The study has attempted to describe outcome of non-hepatic elevation of AST levels and highlight skeletal muscle, cardiac and haematological causes for such AST elevations and related 30-day mortality risk. Although the paper is well written, it’s not clear whether this is a prospective study over 10 years or a retrospective study and whether authors have focused on such a study from the beginning, or they have analysed a patient database. The ethical statement indicates this is a retrospective study “The requirement for informed consent was waived owing to the retrospective design of this study, as determined by the institutional review boards”. Lines 97-98 suggest the prospective nature of the study. 2.2 Data collection Lines 86-88: “Baseline demographics, clinical data, and laboratory data were retrieved from the patients’ medical records. Data on baseline demographics included age, sex, and underlying diseases, such as diabetes, liver cirrhosis, congestive heart failure, and end-stage renal disease”. Line 97-98: “After enrollment, all patients underwent careful history taking….”

Objective exclusion criteria of “hepatic causes” for elevation of AST should be included. This is because, for example patients with cardiac muscle involvement can have concurrent ischaemic liver releasing AST or, other systemic diseases could involve the liver to release AST, therefor it is important to know how the authors excluded hepatic AST release objectively in these patients.

Author Response

We deeply thank the editor and reviewers of the ‘Journal of Clinical Medicine’ for taking time and efforts to review our paper. Please find the revision of our manuscript entitled “Markedly Elevated Aspartate Aminotransferase from Non-hepatic Causes” and “point by point response to each of the reviewers’ comments”. All revisions to the manuscript were marked using the “Track Changes” function. The reviewers’ comments were indicated as underlined black color, while our response was indicated as red color, and all significant changes to the manuscript were indicated as highlight in yellow.

In the revised paper, we changed several descriptions according to reviewers. We hope the revised manuscript will better meet the requirements of the ‘Journal of Clinical Medicine’ for publication. Here is one-by-one response to your comments.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors have responded to all raised criticisms and have improved the manuscript. No further comments to authors.