Round 1

Reviewer 1 Report

This single center, observational, and nonrandomized study by Pugliese and co-workers shows that the heterologous prime–boost regimen with the mRNA-1273 (Moderna) vaccine is safe and highly immunogenic for frail patients with disorders of consciousness.

The paper is well written and has important practical relevance. You, however, should explain why you did not consider a further group of DOC patients receiving homologous vaccination, and this should be cited as a major limitation. “Community doctors recruited participants from a clinic site in S’Anna Institute, Crotone (Italy) following previously published inclusion/exclusion criteria were recruited”. This sentence is not clear, please rephrase.

Author Response

REVIEWER N°1

This single center, observational, and nonrandomized study by Pugliese and co-workers shows that the heterologous prime–boost regimen with the mRNA-1273 (Moderna) vaccine is safe and highly immunogenic for frail patients with disorders of consciousness.

The paper is well written and has important practical relevance. You, however, should explain why you did not consider a further group of DOC patients receiving homologous vaccination, and this should be cited as a major limitation.

REPLY: Unfortunately, we were unable to evaluate a further DOC group receiving homologous vaccination because this study represents the second part of a study started nine months before. The DOC group included only subjects who underwent the first vaccination cycle with the same vaccine molecule and at the same time in January/February period, and it was constituted by the same patients followed month by month. At nine months, the DOC patients group counted 24 patients and we thought it would be less significant to split them into two groups of 12 patients. Adding a new group of DOC patients, for example including patients admitted to the long-care facility unit, would have meant comparing patients who were not homogeneous anymore. We will report it as a major limitation as suggested.

“Community doctors recruited participants from a clinic site in S’Anna Institute, Crotone (Italy) following previously published inclusion/exclusion criteria were recruited”. This sentence is not clear, please rephrase.

REPLY: Done

Reviewer 2 Report

Thank the editor for giving me the opportunity to review this work on immunogenicity and reactogenicity of COVID-19 booster dose vaccination in patients with chronic disorder of consciousness (and healthy controls).  The other main focus of the study is the effect of heterologous vaccination with the two different mRNA vaccines. As reported by the authors, this manuscript is an extension of a previously published article of by the same group (JCM 2021).

The major flaws of the work are the very very small sample size, which is somehow understandable for the DOC group but less for the healthy groups, and some methodology issues in the statistical analysis.

Major comments

1. Need to tone down the conclusions of the abstract declaring the small sample size
2. Section 2.2 , great for the informed consent of the legal representative of DOC subjects, but HC too should sign an informed consent for study participation, blood withdrawal in the study, data processing and so on: it has been done?
3. Section 2.3 When did we start the vaccination campaign for the fragile DOC population in Italy? I did not expect 9 months of follow-up after second dose for them. All of the 24 DOC had the 9 months before the 3^rd dose administration? Why the authors did not evaluated the t2 for all the 32 DOC and 34 HC subjects of the first study and including the 24+12 only for the 3^rd dose response evaluation?
   Moreover was the HC groups a cohort of Health Care Workers? Otherwise I really find difficult to understand the 9 months between 2^nd and 3^rd dose for general healthy population (HCW started the primary-cycle in Jan 2021 in Italy, general population way ahead , and the 3^rd dose campaign started essentially in October2021). Please define the study period and vaccination period in 2.1 and 2.3 sections and specify how the HC groups have been enrolled.  Also add in the results section the mean or median time from 2^nd to 3^rd dose in each group.  
4. Section 2.3 The blinding of the vaccines received was done also for HC or only for DOC? Cannot understand well how in a setting of not RCT, a total blind of the vaccines received can be done, beyond the different staff for vaccine preparation and administration, the name of the vaccines was not recorded in the medical charts?
5. Section 2.3 Evaluation of adverse events in the two groups are quite different, minor AE cannot be detected in the DOC group I suppose (fatigue / pain inj. Site). The comparison should be only for major AE between DOC and HC. Authors need to better specify how the reactogenicity of vaccines has been evaluated in each group.
6. Section 2.5 I used to apply to anti-S IgG titers of SARS-CoV-2 the log2 transformation to transform to havev a normal distribution and use parametric test. I imagine that the same transformation could work with anti-RBD. I suggest trying it and in case use ANOVA and linear regression models. Need also to take account of multiple comparison (like Fig.2) with the Bonferroni correction.
7. Results 1 paragraph. I wouldn’t be so sure of the ‘perfectly matched’ study population (moreoevr ‘matched’ is not the correct term) with only 24+14+28 subjects the size is not powered to the detect statistical significance for small difference. For the age there are 10 years of difference between HC_heter and DOC_heter even if the p is >0.05 in terms of vavccine response are a lot. Any autoimmune rheumatologic diseases?   Does someone used corticosteroids or other immunesoppressor/modulant that could be associated with a different vaccine response?
8. Results of T2: please report after transformation the mean difference between T1-T2 and T2 titers among HC and DC (Factors increase are not clear to me)
9. Same approach for the response 1month after 3^rd dose, for the comparison need to adjust for some confounders like age, presence of other comorbidities that could influence the response (cancer, DM) and use of steroids or other immunomodulatory therapy

Minor comments

1. After defining for the first time the mRNA-1273 as Moderna and BNT162b2 as Pfizer-Biontech in the abstract and then in the manuscript please use only the name of the vaccine and not the brand name.
2. In the Introduction I would like to see something more on immunnocompetence of DOC, and the rationale for studing the response on this group of patients.         
3. Section 2.1 please specify also here that the second group of healthy subjects are homologous boosting strategy (otherwise cannot understand the reason for a second group of healthy subjects in the study design)    
4. Section 2.1 S’Anna à Anna
5. After the ‘Procedure’ paragraph I suggest adding a new paragraph ‘Endpoints’ that clearly specifies the endpoints of the work 1) evaluate the waning of humoral immune response to primary COVID-19 vaccine cycle with BNT162b2 at 9-months in DOC vs HC 2) evaluate the reactogenicity mrna1273 and BNT162b2 booster dose in HC and in DOC 3) evaluate the 30-days humoral response of booster vaccination comparing heterologous and homologues vaccination strategy and response in DOC vs HC.  
6. Section 2.4 Just to be clear after ‘in both groups’ please add ‘(HC_hom, DC_het) and then at the end of the paragraph add ‘(HC_heter)’
7. Table 1. Instead of ‘Clinical Data’ à ‘DOC specific Clinical Data’     
8. In Fig 1 and 2 in the labels of the y axes add ‘BAU/mL anti-RBD SARS-CoV-2 IgG titer’

Author Response

Thank the editor for giving me the opportunity to review this work on immunogenicity and reactogenicity of COVID-19 booster dose vaccination in patients with chronic disorder of consciousness (and healthy controls).  The other main focus of the study is the effect of heterologous vaccination with the two different mRNA vaccines. As reported by the authors, this manuscript is an extension of a previously published article of by the same group (JCM 2021). The major flaws of the work are the very very small sample size, which is somehow understandable for the DOC group but less for the healthy groups, and some methodology issues in the statistical analysis.

Major comments

1. Need to tone down the conclusions of the abstract declaring the small sample size

REPLY: Done

2. Section 2.2 , great for the informed consent of the legal representative of DOC subjects, but HC too should sign an informed consent for study participation, blood withdrawal in the study, data processing and so on: it has been done?

REPLY: This section has been reformulated following the reviewer’s suggestion.

3. Section 2.3 When did we start the vaccination campaign for the fragile DOC population in Italy? I did not expect 9 months of follow-up after second dose for them. All of the 24 DOC had the 9 months before the 3^rd dose administration? Why the authors did not evaluated the t2 for all the 32 DOC and 34 HC subjects of the first study and including the 24+12 only for the 3^rd dose response evaluation?
Moreover was the HC groups a cohort of Health Care Workers? Otherwise I really find difficult to understand the 9 months between 2^nd and 3^rd dose for general healthy population (HCW started the primary-cycle in Jan 2021 in Italy, general population way ahead, and the 3^rd dose campaign started essentially in October2021). Please define the study period and vaccination period in 2.1 and 2.3 sections and specify how the HC groups have been enrolled.  Also add in the results section the mean or median time from 2^nd to 3^rd dose in each group.  

REPLY: We would like to thank this reviewer for these important suggestions. The vaccination campaign started in January 2021. All subjects (both patients and HC) were vaccinated at the same time because HCs are health workers of our Institution. All individuals were vaccinated with the first dose on the 16th of January 2021, the second dose on the 6th of February and third between 11-18 of November. With respect to our previous study [13], eight DOC patients were lost because of: a) discharge; b) booster dose deferral due to clinical instability; or c) death. Moreover, ten healthy controls (HC) dropped out because of Covid infection or vaccination with a different molecule of vaccine (in relation to the local availability). Thus, 24 DOC patients and 28 HC were enrolled in this new study, together with new additional 14 HCs. The procedure was the same for all individuals enrolled in this study. Section 2.1 and 2.3 have been modified following the reviewer’s suggestions.

4. Section 2.3 The blinding of the vaccines received was done also for HC or only for DOC? Cannot understand well how in a setting of not RCT, a total blind of the vaccines received can be done, beyond the different staff for vaccine preparation and administration, the name of the vaccines was not recorded in the medical charts?

REPLY: We would like to thank this reviewer for highlighting this issue. An error occurred in the previous version. We now better explain that only the staff involved in the laboratory analysis were blinded to the treatment assignments. The blinding involved all individuals enrolled in this study

5. Section 2.3 Evaluation of adverse events in the two groups are quite different, minor AE cannot be detected in the DOC group I suppose (fatigue / pain inj. Site). The comparison should be only for major AE between DOC and HC. Authors need to better specify how the reactogenicity of vaccines has been evaluated in each group.

REPLY: In DOC patients we could only observe and report objective adverse reactions (i.e. fever) or severe adverse reactions (respiratory distress, seizures or death) and, as we reported, we only had a single case of fever and none of the severe reactions. For the HC group we asked healthy subjects to report adverse reactions 2 weeks after the booster dose, and we then reported the different prevalence in each group. This section has been re-formulated.

6. Section 2.5 I used to apply to anti-S IgG titers of SARS-CoV-2 the log2 transformation to transform to havev a normal distribution and use parametric test. I imagine that the same transformation could work with anti-RBD. I suggest trying it and in case use ANOVA and linear regression models. Need also to take account of multiple comparison (like Fig.2) with the Bonferroni correction.

REPLY: We thank the reviewer for the valuable suggestion. We applied the log2 transformation and we repeated statistical analyses after checking normal distribution through the Shapiro-Wilk test. Obtained results confirmed what we found in previous analyses.

 From the results, it is confirmed that DOC patients are characterized by a significant decrease in antibody responses with respect to controls, which persists after 9 months (t= -3.14 p-level= 0.003).  

 ANOVA

Assumption Checks

Q-Q Plot

Post Hoc Tests

 Heterologous boosting with mRNA-1273 elicited higher increase of S-RBD IgG levels than homologous boosting with BNT162b2 in DOC patients and HC who had previously received a two-dose of BNT162b2 (F = 7.42; p= 0.001). Post-hoc analysis confirmed that healthy controls with homologous vaccination are characterized by the lowest antibodies response either with respect to the other control group with heterologous vaccination (t= -2.778; p-level 0.022) or DOC patients (t= -3.544; p-level 0.002)

7. Results 1 paragraph. I wouldn’t be so sure of the ‘perfectly matched’ study population (moreoevr ‘matched’ is not the correct term) with only 24+14+28 subjects the size is not powered to the detect statistical significance for small difference. For the age there are 10 years of difference between HC_heter and DOC_heter even if the p is >0.05 in terms of vavccine response are a lot. Any autoimmune rheumatologic diseases?   Does someone used corticosteroids or other immunesoppressor/modulant that could be associated with a different vaccine response?

REPLY: Following the reviewer’s suggestion this term has been modified. Only one patient used chronic corticosteroid because of hypopituitarism. None of the HC was affected by any rheumatological disease, exception of Hashimoto hypothyroidism, included in endocrinopathies group.

8. Results of T2: please report after transformation the mean difference between T1-T2 and T2 titers among HC and DC (Factors increase are not clear to me)

REPLY: a new table has been added to the text.

9. Same approach for the response 1month after 3^rd dose, for the comparison need to adjust for some confounders like age, presence of other comorbidities that could influence the response (cancer, DM) and use of steroids or other immunomodulatory therapy

REPLY: As already done in our previous study (JCM 2021) we re-run the statistical analysis including also confounding variables. Here, we included age, sex the prevalence of hypertension (the only clinical variables differently distributed among groups) as nuisance variables. The overall pattern of findings remained the same.  This additional information has been included in the Results section.

Minor comments

10. After defining for the first time the mRNA-1273 as Moderna and BNT162b2 as Pfizer-Biontech in the abstract and then in the manuscript please use only the name of the vaccine and not the brand name.

REPLY: DONE

1. In the Introduction I would like to see something more on immunnocompetence of DOC, and the rationale for studing the response on this group of patients.     

 REPLY: The introduction has been reformulated following the reviewer’s suggestion 

2. Section 2.1 please specify also here that the second group of healthy subjects are homologous boosting strategy (otherwise cannot understand the reason for a second group of healthy subjects in the study design)    

REPLY: Please consider that the 2.3 section has already been changed following your previous suggestions

3. Section 2.1 S’Anna à Anna

REPLY: Done

4. After the ‘Procedure’ paragraph I suggest adding a new paragraph ‘Endpoints’ that clearly specifies the endpoints of the work 1) evaluate the waning of humoral immune response to primary COVID-19 vaccine cycle with BNT162b2 at 9-months in DOC vs HC 2) evaluate the reactogenicity mrna1273 and BNT162b2 booster dose in HC and in DOC 3) evaluate the 30-days humoral response of booster vaccination comparing heterologous and homologues vaccination strategy and response in DOC vs HC.  

REPLY: We would like to thank this reviewer for this important suggestion. This new paragraph has been added to the Methods Section.

5. Section 2.4 Just to be clear after ‘in both groups’ please add ‘(HC_hom, DC_het) and then at the end of the paragraph add ‘(HC_heter)’

REPLY: Done

6. Table 1. Instead of ‘Clinical Data’ à ‘DOC specific Clinical Data’     

REPLY: Done

7. In Fig 1 and 2 in the labels of the y axes add ‘BAU/mL anti-RBD SARS-CoV-2 IgG titer’

REPLY: Done

Round 2

Reviewer 2 Report

Most of the issues have been addressed by the authors,

What about using the label Healthcare workers (HCW), instead of Healthy group (HC):  there is plenty of literature on COVID-19 and HCW.  HCW have peculiar characteristics: working age (< 65 y), low burden of comorbities (not fully healthy but at least comorbidities that no lead to incapacity for work), higher risk of exposure to COVID-19 (at least in the pre omicron era), priority of vaccination and so on.

Propbabily that's my foult but still cannot understand the blinding procedure. Is it valid for both DOC and HC? Lab staff of course did not know the treatment (they just only receive a plasma sample), nurse and physicians who prepared and administred the dose of course were unblided, but you ( the 11 authors) did not know while collecting the AE if was a Moderna or Pfizer?

Do you see any different response in the single DOC patient with steroids use, if yes that should be taken into account and adjusted in the analysis (possibily the lower response in DOC is related to this one given the small sample size, is improbable but we need to take this into consideration), I will also use for these analyses the same set of covariates used in the previous JCM work.

Still see the word ‘matched’ in first paragraph of results, if you really matched the HC with DOC you have to specify the variabile of matching in the methods, if not (as I suppose) change the term with “comparable” or something like that.

Table 2 is missing the unit of measure and the column labels are a little bit messed up  in the order, and there is  a blank label in the post-booster. As previously said I’ would also report the mean difference between groups and timepoints in the text and please report if any subject was non-responder in the post-boost.

The procedure for AE reporting in HC was not standardized (as far as I can understand) need to comment in the discussion/limit.

Thank you for your great job

Author Response

What about using the label Healthcare workers (HCW), instead of Healthy group (HC):  there is plenty of literature on COVID-19 and HCW.  HCW have peculiar characteristics: working age (< 65 y), low burden of comorbities (not fully healthy but at least comorbidities that no lead to incapacity for work), higher risk of exposure to COVID-19 (at least in the pre omicron era), priority of vaccination and so on.

Reply: we perfectly agree with this reviewer

Probabily that's my foult but still cannot understand the blinding procedure. Is it valid for both DOC and HC? Lab staff of course did not know the treatment (they just only receive a plasma sample), nurse and physicians who prepared and administred the dose of course were unblided, but you ( the 11 authors) did not know while collecting the AE if was a Moderna or Pfizer?

Reply: we better specify this point.

Do you see any different response in the single DOC patient with steroids use, if yes that should be taken into account and adjusted in the analysis (possibily the lower response in DOC is related to this one given the small sample size, is improbable but we need to take this into consideration), I will also use for these analyses the same set of covariates used in the previous JCM work.

Reply: No significant response was detected in this single patient. We re-run the analysis using the same set of covariates used in our previous work but the overall pattern of findings remains the same.

Still see the word ‘matched’ in first paragraph of results, if you really matched the HC with DOC you have to specify the variabile of matching in the methods, if not (as I suppose) change the term with “comparable” or something like that.

Reply: Done

Table 2 is missing the unit of measure and the column labels are a little bit messed up  in the order, and there is  a blank label in the post-booster. As previously said I’ would also report the mean difference between groups and timepoints in the text and please report if any subject was non-responder in the post-boost.

Reply:  Table 2 has been modified accordingly

The procedure for AE reporting in HC was not standardized (as far as I can understand) need to comment in the discussion/limit.

REPLY: done