Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

Congratulations on this high-quality and scientifically robust manuscript.

I only have a few minor comments for your consideration:

1. Please provide details on the methods used for HBV-DNA and qHBsAg quantification across the different participating centers.
2. Was HBV and HDV genotyping performed centrally? If not, and genotyping was carried out locally, this should be clarified in the Methods section, especially if amplicon generation was not standardized across centers.
3. I would also suggest including a brief limitations paragraph in the Discussion to acknowledge:
   a) the multicenter design of the study, particularly the possible impact of methodological differences in HDV RNA quantification (for example, RoboGene versus home-made assays);
   b) the fact that this was not designed as a case-control study.
4. I would also suggest expanding the Discussion to address the potential clinical implications, if any, of the main findings. In particular, it would be helpful if the authors could discuss the possible role of HBsAg isoform assays in clinical practice, including their feasibility, potential added value, and whether their use in routine diagnostic laboratories could be realistically anticipated in the near future.

 

Author Response

Dear Authors,

Congratulations on this high-quality and scientifically robust manuscript.

Answer: We thank the Reviewer for his/her positive comment.

 

I only have a few minor comments for your consideration:

1. Please provide details on the methods used for HBV-DNA and qHBsAg quantification across the different participating centers.

Answer: We thank the Reviewer for the suggestion. Accordingly, we have provided the details regarding HBV-DNA and HBsAg quantification in the Methods section of the revised version of the manuscript (page 3, lines 94-98).

 

1. Was HBV and HDV genotyping performed centrally? If not, and genotyping was carried out locally, this should be clarified in the Methods section, especially if amplicon generation was not standardized across centers.

Answer: We thank the Reviewer for having highlighted this point. Viral genotyping was performed centrally at the University of Rome Tor Vergata, following the protocols described in Supplementary methods. We have added this specification in this section (page 1, lines 22-23; page 2, lines 42-44).

 

1. I would also suggest including a brief limitations paragraph in the Discussion to acknowledge:
   a) the multicenter design of the study, particularly the possible impact of methodological differences in HDV RNA quantification (for example, RoboGene versus home-made assays);
   b) the fact that this was not designed as a case-control study.

Answer: We thank the Reviewer for this suggestion. Accordingly, we have added a sentence in the Discussion section of the manuscript as follows: It should be noted that this study has some limitations: i) the study was not de-signed to be a case-control study; ii) HDV-RNA quantification was carried out also by using a home-made and not standardized assay (page 16, lines 384-386).

 

1. I would also suggest expanding the Discussion to address the potential clinical implications, if any, of the main findings. In particular, it would be helpful if the authors could discuss the possible role of HBsAg isoform assays in clinical practice, including their feasibility, potential added value, and whether their use in routine diagnostic laboratories could be realistically anticipated in the near future.

Answer: We appreciate this suggestion to discuss the role of HBsAg isoforms in clinical practice. Although this study was not specifically aimed at evaluating the clinical implications of HBsAg isoforms, our findings can pave the way for further studies aimed at better deepening the role of HBsAg isoforms in predicting different clinical phenotypes and liver disease progression. At this regard, we highlighted the correlation of S-HBsAg and M-HBsAg with altered ALT in patients with CHD, suggesting their potential capability to trigger HDV cytolytic activity. Both S-HBsAg and M-HBsAg are known to be highly immunogenic and to contain B- and T-cell epitopes, thus potentially explaining their correlation with a higher liver inflammation and disease progression [5]. Furthermore, previous studies have highlighted the association of more elevated levels of M-HBsAg with an unfavourable clinical outcome [9,10]. In particular, a previous study by our group showed that an increase in M-HBsAg levels characterizes a significant fraction of cirrhotic patients developing HCC despite prolonged HBV suppression, suggesting a potential role of M-HBsAg kinetics in identifying patients at higher HCC risk. Similarly, another study showed that M-HBsAg levels tended to be higher in patients who developed decompensation during follow-up, suggesting a role as a prognosis marker.
In a previous study by our group, we evaluated the predicting role of HBsAg isoforms in monitoring response to bulevirtide in CHD. In particular, we have identified specific HBsAg isoforms cut-off values capable to predict the achievement of HDV-RNA undetectability with or without ALT normalization after 48 and 96 weeks of bulevirtide.       
Furthermore, it should be noted that the quantification of HBsAg isoforms assays is based on the cheap and easy-to-use sandwich-ELISA technology with a low turn-around-time, supporting its feasibility in clinical practice. On this basis, although further studies are necessary, HBsAg isoforms quantification could represent a valid and affordable alternative to characterize virological profiles of CHD patients and/or to monitor treatment response. This could be particularly relevant in low/middle-income settings, where the access to HDV-RNA quantification by Real-Time PCR is still lacking due to the high costs.   
We have included these concepts in the Discussion section of the revised version of manuscript (page 15-16, lines 350-383). 

Reviewer 2 Report

Comments and Suggestions for Authors

The paper by D’Anna et al. explores the prevalence and potential pathophysiological role of HBsAg isoforms in patients with chronic hepatitis Delta (CHD), compared with HBV-monoinfected patients. The correlation between HBsAg isoforms and virological and clinical parameters was also examined.

The study is well designed and employs up-to-date laboratory methods. Data analysis appears appropriate.

I have two points for the authors’ consideration:

1. HDV patients showed higher levels of total HBsAg and of each HBsAg isoform. Multivariable analysis indicated that HDV-RNA >1,000 IU/mL was associated with higher HBsAg isoform levels.
   The HDV-RNA cut-off was set at 1,000 IU/mL according to reference 21. However, a recent Italian study (Caviglia, Liver International, 2025) reported that an HDV-RNA level <3,000 IU/mL (measured by the Robogene assay) identifies patients with inactive or mild disease. A brief discussion of this point would be valuable.
2. In the Discussion section (lines 281–287), the comment on the correlation between S-HBsAg and M-HBsAg levels and elevated ALT in CHD is intriguing. Although this correlation is shown in Figure S2, it may deserve inclusion in the main text. Its relevance is further supported by two recent studies on the safety and efficacy in CHD of drugs targeting HBsAg (Stern et al., Journal of Hepatology, 2026; Asselah et al., New England Journal of Medicine, 2026).

Author Response

The paper by D’Anna et al. explores the prevalence and potential pathophysiological role of HBsAg isoforms in patients with chronic hepatitis Delta (CHD), compared with HBV-monoinfected patients. The correlation between HBsAg isoforms and virological and clinical parameters was also examined.

The study is well designed and employs up-to-date laboratory methods. Data analysis appears appropriate.

Answer: We thank the Reviewer for having appreciated our study.

 

I have two points for the authors’ consideration:

1. HDV patients showed higher levels of total HBsAg and of each HBsAg isoform. Multivariable analysis indicated that HDV-RNA >1,000 IU/mL was associated with higher HBsAg isoform levels.
   The HDV-RNA cut-off was set at 1,000 IU/mL according to reference 21. However, a recent Italian study (Caviglia, Liver International, 2025) reported that an HDV-RNA level <3,000 IU/mL (measured by the Robogene assay) identifies patients with inactive or mild disease. A brief discussion of this point would be valuable.

Answer: We thank the Reviewer for the opportunity to comment the above-mentioned paper. In this study, an HDV-RNA level of 1,000 IU/ml was used as cut-off to differentiate patients with inactive/mild or advanced liver disease. This is in line with another recent study (Mangia et. al, Cells 2023) showing that HDV-RNA levels >1000 IU/ml at baseline predicted an unfavourable outcome in terms of increased risk of cirrhosis, transplantation and death at last follow-up visit. Overall, these studies support the correlation between the extent of HDV replication and the degree of liver damage and the role of this cut-off as a potential prognostic marker of liver disease. The choice of this cut-off has been explained in the Methods section of the revised version of the manuscript (page 4, lines 159-161).

 

1. In the Discussion section (lines 281–287), the comment on the correlation between S-HBsAg and M-HBsAg levels and elevated ALT in CHD is intriguing. Although this correlation is shown in Figure S2, it may deserve inclusion in the main text. Its relevance is further supported by two recent studies on the safety and efficacy in CHD of drugs targeting HBsAg (Stern et al., Journal of Hepatology, 2026; Asselah et al., New England Journal of Medicine, 2026).

Answer: We thank the Reviewer for this suggestion. Accordingly, we have included the Fig. S2 in the main text and now it appears as Fig. 2C. Furthermore, we have corroborated the comment regarding the association between HBsAg isoforms and ALT levels by including the new references proposed by the Reviewer. In particular, in the Discussion section of the revised version of the manuscript, the following paragraph was reported: “Overall, these results support that drugs interfering with HBsAg production/release can play a key role in therapeutic strategies aimed at curing HDV infection. In particular, two novel classes of anti-HDV drugs are under clinical evaluation: the nucleic acid polymers (specifically binding to the HBsAg domain present in all the three HBsAg isoforms) and the small interfering RNAs (specifically binding to the mRNA encoding HBsAg isoforms). These novel therapeutic strategies have proven not only to reduce HDV replicative activity but also to dampen liver inflammation, presenting a good safety profile (Stern et al., Journal of Hepatology, 2026; Asselah et al., New England Journal of Medicine, 2026)” (page 15-16, lines 365-372).

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript aims to investigate the difference in terms of HBsAg isoforms between CHD and CHB patients. The presence of higher levels of total HBsAg in the first is already known, but few data were reported comparing HBV mono and HDV/HBV co-infected individuals. The role of HBsAg isoforms in HBV and HDV infection is poorly understood, so I want to give credit to this study.

Nevertheless, I have a major concern. Although both patients’ groups were quite comparable in terms of demographic characteristics and NUC treatment, they are not such optimally paired when considering the total HBsAg levels. This is a main variable to consider since in CHD the total HBsAg is strongly correlated with the antigen isoforms. Because of that, the difference between the two groups may reflect the different levels of total HBs rather than a specific difference due to the presence of HDV. In Table 1 the author should also report the total HBsAg quantified by using the ELISA kit to see the concordance between quantifications (commercial versus ELISA). It could help the results interpretation. Moreover, the authors should consider the possibility to test a group of CHB (even with a limited N) with higher HBsAg to confirm that the observed differences are mainly due to HDV infection rather than total HBsAg per se. 

Importantly, a limitation section is lacking and the methodology used for HDV and HBV genotyping is not clear (Sanger or Illumina?).

Other minor comments in supplementary file:

The sentence in line 6 is quite awkward. In line 42 a space is lacking between By and the.

Regarding reference 21 in supplementary, I cannot find in the cited publication this connection between HDV RNA < 1000 IU/mL and CHD prognosis. 

Regarding the other minor comments, please find them in the attached file.

Comments for author File: Comments.pdf

Author Response

This manuscript aims to investigate the difference in terms of HBsAg isoforms between CHD and CHB patients. The presence of higher levels of total HBsAg in the first is already known, but few data were reported comparing HBV mono and HDV/HBV co-infected individuals. The role of HBsAg isoforms in HBV and HDV infection is poorly understood, so I want to give credit to this study.

Answer: We thank the Reviewer for his/her positive comment.

 

1. Nevertheless, I have a major concern. Although both patients’ groups were quite comparable in terms of demographic characteristics and NUC treatment, they are not such optimally paired when considering the total HBsAg levels. This is a main variable to consider since in CHD the total HBsAg is strongly correlated with the antigen isoforms. Because of that, the difference between the two groups may reflect the different levels of total HBs rather than a specific difference due to the presence of HDV.

Answer: We thank the Reviewer for the revision of our manuscript and for the insightful comments provided. The goal of this study was to characterize the levels of HBV surface glycoproteins, expressed as total HBsAg and HBsAg isoforms, in patients with CHD and CHB, matched for demographics. On this basis, the difference in total HBsAg levels observed between the two groups is part of the Results. Indeed, it should be noted that total HBsAg is the sum of all HBsAg isoforms and, consequently, each HBsAg isoform tightly correlates with total HBsAg. These correlations were confirmed with the total HBsAg obtained by both commercial and our ELISA assay: Rho=0.89, 0.82 and 0.66 for S-, M- and L-HBsAg, respectively (P<0.0001 for all) in CHD; Rho=0.77, 0.64 and 0.61 for S-, M- and L-HBsAg, respectively (P<0.0001 for all) in CHB. Furthermore, a tight and positive correlation is also observed between total HBsAg by the commercial assay and total HBsAg by our ELISA assay: Rho=0.92 and 0.77 for CHD and CHB, respectively (P<0.0001 for both). We apologize for the poor clarity on this issue, that has been better explained in the revised version of the manuscript.

 

2. In Table 1 the author should also report the total HBsAg quantified by using the ELISA kit to see the concordance between quantifications (commercial versus ELISA). It could help the results interpretation.

Answer: We have removed the datum on total HBsAg from the paragraph dedicated to patients’ characteristics and from Table 1. Total HBsAg quantified by both commercial assay and our ELISA assay has been described in the paragraph focusing on HBsAg isoforms (page 6, lines 210-212).

 

3. Moreover, the authors should consider the possibility to test a group of CHB (even with a limited N) with higher HBsAg to confirm that the observed differences are mainly due to HDV infection rather than total HBsAg per se. 

Answer: According to the Reviewer’s comment, we have evaluated the levels of each HBsAg isoform in the subset of CHD and CHB patients with total HBsAg >2000 and >5000 IU/mL, corresponding to the median and the 75^th percentile of total HBsAg observed in CHB patients. Notably, we found that the levels of M- and L-HBsAg were significantly higher in CHD than in CHB patients, showing: median (IQR) levels of 1678 (748-2763) vs 510 (247-1309) ng/mL for M-HBsAg and 4.7 (1.7-10.1) vs 0.5 (0.2-1.9) ng/mL for L-HBsAg by using an HBsAg cut-off of 2000 IU/mL (P<0.0001 for both); median (IQR) levels of 2164 (1237-3015) vs 738 (437-1974) ng/mL for M-HBsAg and 5.7 (2.3-12.6) vs 1.4 (0.5-8.6) ng/mL for L-HBsAg by using an HBsAg cut-off of 5000 IU/mL (P<0.01 for both). No significant difference was observed for S-HBsAg. This result suggests a specific association of HDV infection with these two HBsAg isoforms.

4. Importantly, a limitation section is lacking and the methodology used for HDV and HBV genotyping is not clear (Sanger or Illumina?).

Answer: According to the Reviewer’s comment, we have included limitations section in the Discussion of the revised version of the manuscript (page 16, lines 384-386). We apologize for not having properly reported the method used for viral sequencing. Specifically, we performed the sequencing of HDV full-genome on Illumina platform while the sequencing of genomic region of HBV S antigen/reverse transcriptase was based on Sanger platform. This has been carefully reported in the Methods section of main manuscript and Supplementary information (page 4, lines 137-138 of the main text).

 

Other minor comments in supplementary file:

5. The sentence in line 6 is quite awkward. In line 42 a space is lacking between By and the.

Answer: We thank the Reviewer for the comment. We have reformulated the sentence in line 6 and in line 42 of the previously submitted Supplementary information, and we have moved them in the Methods section of the revised version of the manuscript (page 3-4, lines 108-132).

 

6. Regarding reference 21 in supplementary, I cannot find in the cited publication this connection between HDV RNA < 1000 IU/mL and CHD prognosis. 

Answer: We thank the Reviewer for having highlighted this point. We have replaced ref. 21 with two new references (Caviglia et al., Liv Int 2025; Mangia et al., Cells 2023), in order to explain why the HDV-RNA cut-off value of 1000 IU/mL has been chosen. In particular, in the former, an HDV-RNA level of 1000 IU/ml was used as cut-off to differentiate patients with inactive/mild or advanced liver disease. In the latter, HDV-RNA levels >1000 IU/ml at baseline predicted an unfavourable outcome in terms of increased risk of cirrhosis, transplantation and death at last follow-up visit. This has been reported in the Methods section of the revised version of the manuscript (page 4, lines 159-161).

 

Regarding the other minor comments, please find them in the attached file.

7. of?

Answer: We have modified the text accordingly (page 2, line 64).

 

8. Add the brackets.

Answer: Thanks for the comment. We have added brackets accordingly (page 2, line 68).

 

9. Had been pts treated with IFN? How long before the study? If patients were treated, this must be introduced as another variable in all the uni and multivariate analysis.

Answer: Among 192 CHD patients, 17 have experience IFN, that suspended at least 1 year before the plasma sample collection. We have determined levels of total HBsAg and each HBsAg isoform in these 17 patients and in 106 patients receiving NUC treatment. No statistically significant differences were observed between the two groups of patients (P>0.1 for all comparisons). For this reason, IFN therapy was not included as variable for the analyses of this study.

 

10. When starting a sentence, number should be spelled out.

Answer: Thanks for the comment. We have modified the text accordingly (page 3, line 88).

 

11. In Suppl the reported methodology is Sanger not Illumina. Which technique they finally used? Did they perfom a heat-shock step for HDV retro-transcription?

Answer: As reported in Answer 4, we performed the sequencing of HDV full-genome on Illumina platform while the sequencing of genomic region of HBV S antigen/reverse transcriptase was based on Sanger platform. Moreover, heat shock of the sample was performed before HDV-RNA reverse transcription. This has been carefully reported in the revised version of Methods section of main manuscript and Supplementary information (page 4, lines 137-138 of the main manuscript; page 1, lines 6-7 of the Supplementary information).  

 

12. Specify in the manuscript that they analyzed viral genotypes/subgenotypes just in 41 over the 192 initially included patients. I have some concerns regarding the ability of Sanger sequencing to discriminate HDV sub-genotypes. Did you perform a full genome sequencing by using the same two primers used during the amplification step? Did you include any primer sequencing overlapping regions? Differently, if you used the Illumina platform, did you apply a genome fragmentation step or overlapping amplicon?

Answer: Thanks to the Reviewer for the comment. We have specified the number of samples in which HDV-RNA and HBV-DNA sequencing analyses were performed in Methods section of the revised versions of both manuscript and Supplementary information (page 4, lines 143-147 of the main manuscript; page 1, lines 22-23 and page 2, lines 42-44 of the Supplementary information).

 

13. This negative correlation may dependent of the inclusion in the analysis of HBV patients that show lower HBsAg levels and are younger. Although this result is not a main result, they should comment something on that.

Answer: We thank the Reviewer for giving us the opportunity to better explain this point. CHB patients involved in the study were matched to CHD ones for demographics, including age, that therefore is comparable between the two groups. Multivariable analysis (Table 2), carried out in overall population involving both CHD and CHB patients, showed that age is an independent factor significantly associated with lower levels of HBsAg isoforms, after adjustment for all the other variables included in this analysis. This means that the association of age with lower levels of HBsAg isoforms is confirmed in each subset of CHD and CHB patients. This result is in line with previous studies, showing that HBsAg levels undergo a progressive decrease by increasing patients’ age, as a consequence of hepatocyte turn-over and reduction of viral reservoir in both CHD and CHB patients (Taniguchi et al., Journal of Gastroenterology and Hepatology 2020; Terrault et al., Hepatology 2022). This has been included in the Discussion section of the revised manuscript (page 15, lines 325-329).

 

14. This correlation is confirmed also in HBV patients?

Answer: The correlations observed in CHD patients were fully confirmed in CHB ones. In particular, each HBsAg isoform positively correlated with total HBsAg by commercial assays, with the following correlation coefficients: Rho=0.77, 0.64 and 0.61 for S-, M- and L-HBsAg, respectively (P<0.0001 for all) in CHB.

 

15. Being part of the main results, I would add also the related graphs.

Answer: Accordingly, we have also added the graphs regarding correlations of each HBsAg isoform with total HBsAg by commercial assays in CHD patients in Supplementary material as Figure S1.

 

16. Considering that around 40% of the patients in both sub-groups are women, did you consider to stratify patients according their ALT levels and considering their sex (women >35 and man >40). Moreover, this analysis was performed for both HDV and HBV groups, or just HDV? Please specifiy.

Answer: We thank the Reviewer for the suggestion. We have repeated the analysis considering the cut-offs proposed by AASLD guidelines (35 U/L for male patients and 25 U/L for female patients) and the results were fully confirmed. This has been specified in the revised version of the manuscript (page 4, lines 157-158) and in Table 1. Moreover, the analysis reported in the manuscript was focused on CHD patients and this has been better clarified in the text. Nevertheless, following the Reviewer’s comment, we have repeated the analysis in CHB patients and notably, the levels of all HBsAg isoforms were comparable between patients with normal ALT compared to those with altered ALT. We have reported the results in the revised version of the manuscript (page 7, lines 238-242).

 

17. If you consider just CHD patients, is there any difference between patients with high ALT and patients with normal ALT?

Answer: The differences in HBsAg isoforms levels reported in the text and in fig. 2C derive from the comparison between patients with normal ALT and altered ALT.  

 

18. Consider to also test a small group of HBV patients with higher total HBsAg to ensure that the findings are confirmed independently of the HBsAg titers.

Answer: As reported in Answer 3, according to the Reviewer’s comment, we have evaluated the levels of each HBsAg isoform in the subset of CHD and CHB patients with total HBsAg >2000 and >5000 IU/mL, corresponding to the median and the 75^th percentile of total HBsAg observed in CHB patients. Notably, we found that the levels of M- and L-HBsAg were significantly higher in CHD than in CHB patients, showing: median (IQR) levels of 1678 (748-2763) vs 510 (247-1309) ng/mL for M-HBsAg and 4.7 (1.7-10.1) vs 0.5 (0.2-1.9) ng/mL for L-HBsAg by using an HBsAg cut-off of 2000 IU/mL (P<0.0001 for both); median (IQR) levels of 2164 (1237-3015) vs 738 (437-1974) ng/mL for M-HBsAg and 5.7 (2.3-12.6) vs 1.4 (0.5-8.6) ng/mL for L-HBsAg by using an HBsAg cut-off of 5000 IU/mL (P<0.01 for both). No significant difference was observed for S-HBsAg. This result suggests a specific association of HDV infection with these two HBsAg isoforms.

 

19. 40% of the CHD patients have detectable HBV. Did you try to the isoform levels in CHD patients grouped according having or not active HBV replication? It would confirm the lack of association with HBV replication.

Answer: We thank the Reviewer for this suggestion. We have repeated the analysis by stratifying CHD patients according to HBV-DNA positivity. Notably, the levels of each HBsAg isoform were comparable between patients with HBV-DNA <20 IU/mL and those with HBV-DNA >20 IU/mL (median [IQR]: 3008 [785-6183] vs 5010 [672-7442] ng/mL for S-HBsAg, 729 [119-2095] vs 1428 [201-2442] ng/mL for M-HBsAg and 3.2 [0.3-9.1] vs 2.4 [0.3-5.4] ng/mL for L-HBsAg; P>0.05 for all comparisons). This supports, as suggested by the Reviewer, the lack of association with HBV replication. This has been reported in the revised version of the manuscript (page 6, lines 221-226).

 

20. Specify the commercial technique used to quantify the total HBsAg and add the total HBsAg quantified by using the ELISA kit to compare between the two types of quantification.

Answer: We have specified commercial assays used for total HBsAg quantification in the Methods section (page 3, lines 97-98) and have removed the datum on total HBsAg from the paragraph dedicated to patients’ characteristics and from Table 1. Total HBsAg quantified by both commercial assay and ELISA assay has been described in the paragraph focusing on HBsAg isoforms (page 6, lines 194-197).

 

21. Based on this number, you considered all the patients enrolled in the study, however, based on table 1, the HBV DNA datum was available for few CHB. Has the analysis been performed excluding the HBV patients with unknown HBV DNA? If yes please explain it in the caption.

Answer: Uni- and multivariable logistic regression analyses automatically consider only patients with available data. This has been specified in the Methods section of the revised version of the manuscript (page 4, lines 167-168, 176-177).

 

22. Add also the ALT levels in the analysis to confirm the relation between HBsAg and ALT you commented in discussion.

Answer: We thank the Reviewer for the suggestion. By univariable logistic regression analyses, ALT >ULN was significantly associated with higher levels of S-HBsAg (OR [95% CI]: 2.65 [1.14-6.14]; P=0.02), while a trend of correlation was observed for M- and L-HBsAg (P>0.1 for both). The associations were not confirmed by multivariable analyses. The results have been included in Table 3.

 

23. Despite being statistical relevant, the correlation is quite moderate, which suggests that other factors may be involved in this relation. Because of that, the assumption that HBsAg isoforms parallel HDV RNA release is technically not supported by the results.

Answer: According to the Reviewer’s comment, we have softened the sentence (page 15, lines 330-336).

 

24. Again, this statement is not exactly sustained by the findings. despite CHB show low titers of HBsAg isoforms than CHD group, the low total titer per se may affect these results. Moreover, the correlation between HDV RNA and HBsAg isoforms is not so strong (intermediate Rho), which may suggest that this limited relation may be altered by the presence of other factors, potentially including the SVP.

Answer: According to the Reviewer’s comment, we have softened the sentence (page 15, lines 330-336).

 

25. Although CHD have higher HBsAg, it is quite probable that the immune response against this Ag is quite exhausted. These patients, indeed, frequently originated from patients with pre-existing HBV infection. The relation between ALT and HBsAg may be mainly related to the presence of HDV per se rather than a higher immune activation. Adding ALT in the multivariate analysis may help to understand this relation.

Answer: We agree with the Reviewer’s comment. As suggested by the Reviewer and reported in Answer 22, we included ALT in logistic regression analyses. By univariable analyses, ALT >ULN was significantly associated with higher levels of S-HBsAg (OR [95% CI]: 2.65 [1.14-6.14]; P=0.02), while a trend of correlation was observed for M- and L-HBsAg (P=0.11 and 0.18, respectively). The associations were not confirmed by multivariable analyses. For this reason, we have softened the concept in the Discussion section of the revised version of the manuscript (page 15, lines 352-358).

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript had been improved, however I still have some comments.

1. I strongly believe that the relation between HBsAg isoforms and the cytolytic or immune activity should be soften along the discussion (and potentially even the title). This assumption is based on the observation that patients with normal ALT have low HBsAg isoforms, however this finding may be a reflex of the higher viral replication.

Based on your results,  CHD patients with high viremia have more HBsAg isoforms. As generally seen in clinical practice, CHD patients with high viremia are generally characterized by having high ALT levels. Based on this assumption, the relation (it isn’t a correlation) between having normal or not ALT and HBsAg isoform levels may be the result of the natural pathogenesis of HDV infection (High viremia, so high ALT AND high HBsAg isoforms). Indeed, in multivariate analysis the relation between ALT and HBsAg isoform ranges is lost, suggesting that there is another marker (probably HDV RNA) affecting this latter (HBsAg isoform levels).

Moreover, the relationship between HBsAg isoforms and immune response, especially in the context of CHD, has not be addressed in this study.

Other minor comments:

2. line 97: substitute “while” with “and”

3. Line 194-196: Although I’m still convinced that the study needs a control group of HBV patients with very high HBsAg levels (higher than the median of CHD), the comparison that the authors reported in the answer 19 when CHD and CHB patients were stratified according their HBsAg (>2000, >5000) should be reported in the manuscript to reinforce their findings.

When reporting this data, I consider important also highlighting the N of patients included per each group (CHD or CHB, with >2000 and >5000) to show any potential biases due to samples size.

4. In Fig2C, the text in the manuscript refers to ALT>ULN whereas the figure to >40, please uniform.

5. As correctly commented by the authors, just those patients with all the records (variables) are included in a uni and multivariate analysis. Because of that, the legend should report the number of patients included per each analysis group.

6. Line 337: Technically, having more HBsAg isoforms in patients with detectable HBcrAg is not a positive correlation, but rather, a potential relation.

7. In lines 363-365: Add the reference regarding the presence of more M-HBsAg in patients who developed decompensation.

8. I appreciate that the authors corrected the references referred to HDV-RNA cut-off in the manuscript, however the wrong reference [21] is still present in supplementary.

Author Response

The manuscript had been improved, however I still have some comments.

 

1. I strongly believe that the relation between HBsAg isoforms and the cytolytic or immune activity should be soften along the discussion (and potentially even the title). This assumption is based on the observation that patients with normal ALT have low HBsAg isoforms, however this finding may be a reflex of the higher viral replication.

Answer: We thank the Reviewer for the careful revision of our manuscript and the insightful comments and suggestions aimed at improving its completeness. Accordingly, we have softened the relation between HBsAg isoforms and the cytolytic or immune activity. This has been reported in the Discussion section of the revised version of the manuscript (page 16, lines 372-378).

 

Based on your results, CHD patients with high viremia have more HBsAg isoforms. As generally seen in clinical practice, CHD patients with high viremia are generally characterized by having high ALT levels. Based on this assumption, the relation (it isn’t a correlation) between having normal or not ALT and HBsAg isoform levels may be the result of the natural pathogenesis of HDV infection (High viremia, so high ALT AND high HBsAg isoforms). Indeed, in multivariate analysis the relation between ALT and HBsAg isoform ranges is lost, suggesting that there is another marker (probably HDV RNA) affecting this latter (HBsAg isoform levels). Moreover, the relationship between HBsAg isoforms and immune response, especially in the context of CHD, has not be addressed in this study.

Answer: According to the Reviewer’s comment, we have softened the term “correlation” when referred to HBsAg isoforms and ALT (page 6, lines 249 and 251).

 

 

 

Other minor comments:

 

2. line 97: substitute “while” with “and”

Answer: We thank the Reviewer for the suggestion. We have modified the text accordingly (page 3, line 96).

 

3. Line 194-196: Although I’m still convinced that the study needs a control group of HBV patients with very high HBsAg levels (higher than the median of CHD), the comparison that the authors reported in the answer 19 when CHD and CHB patients were stratified according their HBsAg (>2000, >5000) should be reported in the manuscript to reinforce their findings. When reporting this data, I consider important also highlighting the N of patients included per each group (CHD or CHB, with >2000 and >5000) to show any potential biases due to samples size.

 

Answer: According to the Reviewer’s comment, we have included the part related to comparisons between CHD and CHB with HBsAg >2000 or >5000 IU/mL in paragraph 3.2 of Results section of the revised version of the manuscript (page 6, lines 210-219). We have also reported the number of patients analyzed in the Methods section of the revised version of the Manuscript (page 4, lines 155-159).

 

4. In Fig2C, the text in the manuscript refers to ALT>ULN whereas the figure to >40, please uniform.

Answer: We apologize for the mistake. We have now corrected and uniform Fig. 2C accordingly.

 

5. As correctly commented by the authors, just those patients with all the records (variables) are included in a uni and multivariate analysis. Because of that, the legend should report the number of patients included per each analysis group.

Answer: We have reported the number of patients considered for each analysis group in the footnotes of both Table 2 and 3.

 

6. Line 337: Technically, having more HBsAg isoforms in patients with detectable HBcrAg is not a positive correlation, but rather, a potential relation.

Answer: We thank the Reviewer for the suggestion. The term “correlation” was used according to the results shown in page 7, lines 240-243, where the correlations of HBcrAg with HBsAg isoforms and with HDV-RNA and HBV-DNA were assessed by Spearman Rho test. However, according to the Reviewer’s comment, we have modified the text in the Discussion section of the revised version of the manuscript (page 15, lines 357 and 358).

 

7. In lines 363-365: Add the reference regarding the presence of more M-HBsAg in patients who developed decompensation.

Answer: We have added the reference accordingly (ref.10, page 15, line 385).

 

8. I appreciate that the authors corrected the references referred to HDV-RNA cut-off in the manuscript, however the wrong reference [21] is still present in supplementary.

Answer: We thank the Reviewer. Previous reference 21 should be not present anymore neither in Supplementary files or in the main text of the manuscript.