Review Reports - Elevated Incidence and Risk of Emergent Cirrhosis Complications in Alcoholic Cirrhosis Compared with Other Etiologies

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In this manuscript, the authors provided noteworthy findings regarding the influence of cirrhosis aetiology on the incidence of complications associated with this disease. Particularly, they emphasised the importance of alcoholic aetiology in comparison to non-alcoholic aetiologies.

While I found this work to be quite intriguing, I believe the following suggestions would be beneficial in enhancing the manuscript's quality.

Abstract: I suggest adding the numeric data regarding the prevalence of GIB among ALC and NALC patients. Moreover, some abbreviations such as HE, ER and EGD need to be specified.

Main text: I recommend to carefully review the abbreviations and their application throughout the manuscript, using it after the first specification. I also suggest carefully checking the punctuation which in some cases is incorrect (for example "(OR: 15.985, 95% CI: 15.560-16.421, p<0.01). (Table 2 and Figure 3B)".

Introduction: a comment about the recent changes in the aetiology of cirrhosis and its future prospects would be interesting to include. I suggest integrating with recent publication “Huang DQ, Terrault NA, Tacke F, Gluud LL, Arrese M, Bugianesi E, Loomba R. Global epidemiology of cirrhosis - aetiology, trends and predictions. Nat Rev Gastroenterol Hepatol. 2023 Jun;20(6):388-398. doi: 10.1038/s41575-023-00759-2.”

Materials and method:

- Line 148: substitute “NACL” con “NALC”

Results: i suggest dividing result into two subparagraphs, the first concerning the demographic and comparisons data and the second regarding the multivariate analysis, and to move the tables and figure immediately after the subparagraph they are mentioned.

- Line 179: according to data presented in figure 1, the total number of patients with GIB was 67567 of which 7360 with ALC, 54449 without cirrhosis and 5758 with NALC. Are comparisons presented in table 1 referred to total ALC and NALC patients or only to patients with ALC and NALC related GIB? Please specify adding to the first line of the table the total number of patients considered and specify in the explanation which populations are considered.

- Figure 3: improve the quality of the figures and the readability of the text within them.

Since the Child-Turcotte-Pugh score is the most used method to define the compensation of the cirrhosis, it would be necessary to present results regarding the prevalence of different CP classes among ALC and NALC patients with GIB. If there is a significant difference in the prevalence of advanced CHILD B or C among ALC patients compared to NALC, the conclusion that alcoholic etiology has a greater weight than other aetiologias on the complications of cirrhosis could be biased.

Comments on the Quality of English Language

While a few inaccuracies are present, the language is fluent and easily understood.

Author Response

While I found this work to be quite intriguing, I believe the following suggestions would be beneficial in enhancing the manuscript's quality.

Response: The reviewer’s positive comments are greatly appreciated.

Abstract: I suggest adding the numeric data regarding the prevalence of GIB among ALC and NALC patients. Moreover, some abbreviations such as HE, ER and EGD need to be specified.

Response: Information based on reviewer’s comments have been added to the revised manuscript.

Response: Thank you for these comments, these issues were corrected in the revised manuscript.

Response: Thank you for this recommendation, a statement in the introduction section above has been added with reference.

Materials and method:

- Line 148: substitute “NACL” con “NALC”

Response: Thank you, this has been corrected.

Response: Thank you for this recommendation, it has revised based on reviewer’s suggestions.

Response: Additional explanations have been included in the table legend.

- Figure 3: improve the quality of the figures and the readability of the text within them.

Response: High quality images have been added in the revised manuscript.

Response: Thank you for the insightful suggestion. We fully concur with the reviewer regarding the value of the Child-Pugh (CP) score in evaluating mortality and morbidity among cirrhosis patients. Unfortunately, due to limitations in our current study, the CP score could not be collected or calculated as the necessary bilirubin, albumin, and prothrombin time (PT) data were not available in the database.

-As we proceed with our retrospective study in regional hospitals, we are closely reviewing the data of each cirrhosis patient. The CP score stands out as one of the crucial parameters in our dataset and we are ensuring its inclusion with a more comprehensive analysis.

-----------------------------------------------------------------

Reviewer 2 Report

Comments and Suggestions for Authors

This is a large retrospective study based on a large population identified from data bases on hospitalized patients. The purpose of the study, i.e the establishment of etiology as a clinical indicator of GIB in cirrhosis is very ambitious. Moreover , there are several problems with the study:

1. In the abstract there is as repetition in the two phrases in lines 17,18,19

2. The introduction is unnecessarily very long . It contains information on all etiologies of cirrhosis (with the notable exception of Autoimmune hepatitis-cirrhosis) while some information on GIB would be sufficient as this is the subject of the paper and not epidemiology of liver diseases. The lines 65-109 in particular are not very relevant to the subject.

3. The main problems are in the Methodology section.The authors use indiscriminately the terms GIB and variceal bleeding. The diagnosis should be clearly based on endoscopy findings and variceal bleeding should be examined separately from other forms of GIB such as peptic ulcer or portal gastropathy as the prognosis are quite different. How many cases in each group had indeed variceal bleeding ?

4. Diagnosis of cirrhosis is based only on hospital data. As the authors themselves recognize , the accuracy of similar data is very precarious.In particular, patients with Alcoholic cirrhosis (ALD) may have an additional aggravating factor such as HCV or NAFLD. How this possibility has been excluded?

5. Statistical corrections for age, gender and race are certainly not enough, Factors like variceal size, albumin, INR, Platelets, Child-Pugh score and MELD score are significant risk factors that shoud be taken into account.

6. Their NALC group is extremely heterogeneous. PBC, NAFLD and viral cirrhosis have different prognosis. What is the percentage of these diseases in the NALC group?

7. Most importantly, in a study like this with the large number of participants, subgroup analysis is necessary. Comparisons of ALD with the heterogeneous group of NALC is possibly misleading.

8. The discussion is unnecessarily lengthy. It should be focus on the findings of the study. Presentation of pathophysiological aspects of ALD are not relevant. In particular lines 270-320 do not add anything useful to the subject.

Author Response

In the abstract there is as repetition in the two phrases in lines 17,18,19

Response: This statement has been revised.

The introduction is unnecessarily very long It contains information on all etiologies of cirrhosis (with the notable exception of Autoimmune hepatitis-cirrhosis) while some information on GIB would be sufficient as this is the subject of the paper and not epidemiology of liver diseases. The lines 65-109 in particular are not very relevant to the subject.

Response: Thank you for this suggestion, we revised the introduction section by removing the reviewer-recommended sections.

The main problems are in the Methodology section.The authors use indiscriminately the terms GIB and variceal bleeding. The diagnosis should be clearly based on endoscopy findings and variceal bleeding should be examined separately from other forms of GIB such as peptic ulcer or portal gastropathy as the prognosis are quite different. How many cases in each group had indeed variceal bleeding ?

Response: We appreciate and agree with the comments from the reviewer. Given that this study relies on ICD-10 codes, the GI bleeding group was delineated using the following parameters: patients were identified with diagnoses of esophageal varices with bleeding, hematemesis, and gastrointestinal hemorrhage, unspecified. To ensure the exclusion of patients with GI bleeding secondary to peptic ulcers or gastritis, we specifically isolated cases of peptic ulcers with bleeding and gastritis with bleeding, excluding them from the initial groups.

-Due to limitations in the database, the ICD-10 code for portal gastropathy could not be identified.

-The methods section in the revised manuscript has been updated to incorporate your suggestions.

Diagnosis of cirrhosis is based only on hospital data. As the authors themselves recognize , the accuracy of similar data is very precarious.In particular, patients with Alcoholic cirrhosis (ALD) may have an additional aggravating factor such as HCV or NAFLD. How this possibility has been excluded?

Response: We appreciate your questions and apologize for the insufficient description of the methodology. Patients with alcoholic cirrhosis were identified by selecting individuals based on the ICD-10 code. Specifically, we focused on ICD-10 codes such as "Alcoholic fibrosis and sclerosis of the liver," "Alcoholic cirrhosis of the liver," and "Alcoholic hepatic failure." Subsequently, we identified patients with diagnoses of HCV, NAFLD, NASH, HBV, PBS or other ethology of cirrhosis using ICD-10 codes—excluding those cases from the group of patients with alcoholic cirrhosis.

-The methods section was updated accordingly.

Statistical corrections for age, gender and race are certainly not enough, Factors like variceal size, albumin, INR, Platelets, Child-Pugh score and MELD score are significant risk factors that shoud be taken into account.

Response: We fully acknowledge the parameters highlighted by the reviewer are of utmost importance for assessing the prognosis of patients with cirrhosis. Unfortunately, due to database limitations, no laboratory information could be properly identified for the patients. We are currently conducting a 10-year retrospective study involving patients in our regional hospitals. In this ongoing investigation, we can access detailed information from each patient, which will further enhance and reinforce the findings of our current study.

- Additionally, this statement is included in our limitations section.

Their NALC group is extremely heterogeneous. PBC, NAFLD and viral cirrhosis have different prognosis. What is the percentage of these diseases in the NALC group?

Response: The percentage of NASH, biliary cirrhosis and other etiology (such as HCV, HBV or other etiology induced cirrhosis in NALC: 21.4%, 3.5% and 75.1%.

Most importantly, in a study like this with the large number of participants, subgroup analysis is necessary. Comparisons of ALD with the heterogeneous group of NALC is possibly misleading.

Response: We greatly appreciate your insight on the topic. Subgroup analysis has been incorporated into the manuscript to explore the prevalence and risk factors of GIB in patients with NASH, biliary cirrhosis, or other etiologies of cirrhosis, in comparison to ALC. This information has been included in the results section of the revised manuscript.

-Our additional analysis revealed that patients with NASH (6.3%), biliary cirrhosis (5.2%), or other etiologies of cirrhosis, such as hepatitis-induced cirrhosis (6.8%), exhibited a relatively lower incidence of GIB compared to those with ALC (10.8%) (p < 0.01). Assessing the risk of GI bleeding in comparison to ALC, the odds ratio for NASH was 0.55, for biliary cirrhosis was 0.441, and for other etiologies was 0.60. These values indicate that the risk of GI bleeding in patients with NASH, biliary cirrhosis, and other etiologies of cirrhosis was 45%, 66%, and 40% lower, respectively, than in patients with ALC. These findings align with our previous observations.

The discussion is unnecessarily lengthy. It should be focus on the findings of the study. Presentation of pathophysiological aspects of ALD are not relevant. In particular lines 270-320 do not add anything useful to the subject.

Response: As highlighted, we have revised the discussion section in response to the reviewer's comments. We consider it worthwhile to discuss the potential underlying mechanisms behind our findings. However, it is important to note that additional molecular biological studies utilizing patient samples are required to fully elucidate these hypotheses.

========================================

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

most of the points of criticism have been answered. Some points can not be answered due to the inherent problems in these type of retrospective studies.