Prevalence and Specificity of Red Blood Cell Alloimmunization: Insights from Transfusion-Dependent Populations in Serbia

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors analyzed a total of 27,000 transfusions in a major hospital immunohematology lab for whom orders for red blood cell transfusion were placed, and identified newly formed antibodies in 630 patients. They characterize the identified antibodies, also with regard to warm and cold reactivity as IgG and IgM. They conclude that “the data underscore the need for targeted antigen matching and that tailored transfusion protocols and genotypic matching can improve safety for high-risk populations like thalassemia patients”.

 

1. Abstract lines 32, 33: The authors should check whether the conclusion that “the data underscore the need for targeted antigen matching and that tailored transfusion protocols and genotypic matching can improve safety for high-risk populations like thalassemia patients” is justified by the data provided, and consider to adapt the conclusions. To this, see also comments 8 - 12 below.

 

2. Intro line 63: “the study evaluates … such as thalassemia patients”: Is it known how many of the patients with antibodies in this study actually had thalassemia? Of these, was there an increased frequency of antibody formation ?

 

3. Methods, line 65: “this research aims to enhance transfusion safety”: consider rephrasing e.g. “provides a database for enhancing transfusion safety”

 

4. Methods, line 69: at which point in the analysis were the data anonymized, or pseudonymized?

 

5. Line 84: patients who already had been found to have antibodies: For clarity, add: at the beginning of the observation period.

 

6. Methods, line 82-84: patients not given antibody screening tests for platelet or plasma transfusions: is there a regulation / guideline to omit antibody screening test before platelet or plasma transfusions, which can be referenced?

 

7. Methods, line 105: It should be more clearly written when and how the tests with enzyme treated cells were applied (e.g. were they only employed when a Rhesus antibody was likely, or generally, or in selected cases (when?).

 

8. A particular value of the study lies in the testing methodology (lines 86 ff.), applying screening for antibodies and crossmatching at 37⁰C, 22⁰C and 4⁰

• Part of the antibodies in Table 2 are referenced as IgG or IgM (MNS and P1). The reviewer suggests that reference is made to the mode of identification of the tabulated antibodies and temperature (37, 22 or 4^oC) in the table, and to include more (e.g. also Le, Jk and Fy) or all antibodies with regards to the conditions with which they were identified.
• It would also be relevant to learn the diseases of the patients who were found to develop the different antibodies. E.g. line 181 ff., where we learn in the discussion that in the hematological patient population cold antibodies were found with increased frequency (which should be part of the results).

These data are likely contained in the database. Their evaluation would create significance added value and originality of the research presented.

 

9. Discussion: The discussion is a bit speculative, and the conclusions are not always clear. Some parts are recapitulating known literature e.g. in lines 198 ff. on the hemolytic capacity of certain types of antibodies, without clear reference to the results of the presented work. It should be re-written in several parts. 
10. It would be relevant to know what the consequences of the testing system are: E.g. are additional RCCs tested for certain antigens? Which antigens are tested in all donors already? Could it make sense to test for additional antigens routinely, e.g. at a certain frequency?
11. Antibodies detected with relation to disease types should be discussed in view of the literature.
12. What are the consequences in daily routine compared with a technique that screens at 37⁰C only.

Author Response

Thank you for reviewing our manuscript and giving the opportunity for revision. We are very grateful for constructive and valuable comments which are all very helpful for revising and improving our paper.
We have incorporated suggestions; Please see below, point-by-point response to the reviewers’ comments and concerns.

 

1.    Abstract lines 32, 33: The authors should check whether the conclusion that “the data underscore the need for targeted antigen matching and that tailored transfusion protocols and genotypic matching can improve safety for high-risk populations like thalassemia patients” is justified by the data provided, and consider to adapt the conclusions. To this, see also comments 8 - 12 below.
Response to comment: We appreciate the reviewer’s insight and have clarified the conclusion to ensure it is fully supported by the data. The revised text explicitly connects our findings to the potential benefits of antigen matching while acknowledging that the impact of genotypic matching is supported by prior research rather than directly assessed in our study. Line 31-35.
 „Our findings indicate that alloimmunization is predominantly associated with Rh and Kell antigens, suggesting that implementing targeted antigen matching may reduce the frequency of alloimmunization. While our study does not directly assess the impact of genotypic matching, prior literature supports its role in enhancing transfusion safety, particularly for high-risk populations like thalassemia patients.“

2. Intro line 63: “the study evaluates … such as thalassemia patients”: Is it known how many of the patients with antibodies in this study actually had thalassemia? Of these, was there an increased frequency of antibody formation ?
Response to comment: We acknowledge the importance of this aspect and have now specified the proportion of alloimmunized patients with thalassemia in the results section. This addition provides a clearer context for the conclusions drawn in the study 
Line 29-30.  “In a total of 495 patients with thalassemia, antibodies were found in 9.69%.” 
Line 172-173. “Of the 495/27530 (1.8%) patients with thalassemia, 48/495 (9.69%) were found to have antibodies detected.”

3.    Methods, line 65: “this research aims to enhance transfusion safety”: consider rephrasing e.g. “provides a database for enhancing transfusion safety 
Response to comment: Thank you for the suggestion. We have revised the sentence to more accurately describe the study’s role in contributing data rather than directly implementing safety measures: “…this study provides a database for evaluating transfusion safety and the prevalence of alloimmunization in a transfusion-dependent population. “

4.    Methods, line 69: at which point in the analysis were the data anonymized, or pseudonymized?
Response to comment: Thank you for the suggestion. We have clarified the stage at which anonymization was performed in the Methods section to ensure transparency: Line 67-69. “All patient data were anonymized before statistical analysis. Patient identifiers were removed at the stage of data extraction from hospital records, ensuring compliance with ethical guidelines. “

5.    Line 84: patients who already had been found to have antibodies: For clarity, add: at the beginning of the observation period.
Response to comment: Thank you for the suggestion. We have clarified this point in the Methods section to ensure precise understanding as observation period is stated in Line 73. We added in Line 75-81  “The study analyzed pre-transfusion test results from a total of 27530 unique patients who underwent pre-transfusion testing during the study period. If an alloantibody was detected, a retrospective review of the patient’s transfusion history was conducted to determine prior exposure to RBCs transfusions. The study assessed the prevalence rather than incidence of alloimmunization. In this manner, the study focused on all detected alloimmunization at the time of testing. The study did not monitor the development of RBC antibodies at the time after current transfusions.

6.    Methods, line 82-84: patients not given antibody screening tests for platelet or plasma transfusions: is there a regulation / guideline to omit antibody screening test before platelet or plasma transfusions, which can be referenced?
Response to comment: Thank you for the suggestion. We have clarified this point in the Methods section to ensure precise understanding. Line 101-104. “According to the current transfusion guidelines established by the National Expert Committee for Transfusion Medicine routine antibody screening is not mandated before platelet or plasma transfusions unless clinically indicated.”

7.    Methods, line 105: It should be more clearly written when and how the tests with enzyme treated cells were applied (e.g. were they only employed when a Rhesus antibody was likely, or generally, or in selected cases (when?).

Response to comment: We have added clarification regarding the specific conditions under which enzyme-treated cells were used. Line 124-127 “Enzyme-treated RBCs were employed selectively when the presence of multiple antibodies was suspected or when weak antibodies required enhanced detection. Their application was particularly emphasized in cases of suspected Rh alloimmunization.”

8.    A particular value of the study lies in the testing methodology (lines 86 ff.), applying screening for antibodies and crossmatching at 370C, 220C and 40
Part of the antibodies in Table 2 are referenced as IgG or IgM (MNS and P1). The reviewer suggests that reference is made to the mode of identification of the tabulated antibodies and temperature (37, 22 or 4oC) in the table, and to include more (e.g. also Le, Jk and Fy) or all antibodies with regards to the conditions with which they were identified.
It would also be relevant to learn the diseases of the patients who were found to develop the different antibodies. E.g. line 181 ff., where we learn in the discussion that in the hematological patient population cold antibodies were found with increased frequency (which should be part of the results).
These data are likely contained in the database. Their evaluation would create significance added value and originality of the research presented.
Response to comment: We have updated Table 2 to specify temperature conditions and expanded the discussion of disease correlations with antibody formation in the Results, in Table 2, and Discussion section.
Result Line 176-181: Out of the 132 identified autoantibodies, 46 were classified as warm-reactive (detected at 37°C), while 86 were cold-reactive (detected at 4°C). Cold-reactive autoantibodies were predominantly found in patients with hematological malignancies and other hemato-oncologic conditions. However, despite observing trends in antibody distribution among different patient groups, no statistically significant correlation was found between the specificity of the detected antibodies and the underlying diagnoses

Discussion Line 250-257: While our findings indicate that cold-reactive autoantibodies were predominantly observed in patients with hematological malignancies and other hemato-oncologic conditions, no direct correlation between antibody specificity and primary diagnosis was established. This suggests that alloimmunization risk in transfusion-dependent patients is multifactorial, likely influenced not only by the underlying disease. The absence of a disease-specific alloantibody pattern aligns with previous studies, which emphasize the complex interplay between antigenic exposure and immune response

9.    Discussion: The discussion is a bit speculative, and the conclusions are not always clear. Some parts are recapitulating known literature e.g. in lines 198 ff. on the hemolytic capacity of certain types of antibodies, without clear reference to the results of the presented work. It should be re-written in several parts. 
Response to comment: We have revised the discussion to ensure that all claims are clearly supported by our study data and reduced speculative content
Line 268-276 “In the study, Rh system antibodies (anti-D, anti-E, and anti-C) and Kell system antibodies (anti-K) were among the most frequently detected, corroborating their known immunogenicity. These findings align with previous reports that such antibodies are implicated in hemolytic transfusion reactions. In particular, anti-Kell has been associated with hemolytic disease of the fetus and newborn (HDFN), as it suppresses fetal erythropoiesis rather than causing direct hemolysis [1,9,10]. However, our study did not include an assessment of hemolytic severity in affected patients, and further investigation would be needed to establish a direct link between these antibodies and clinical hemolysis in our cohort.

10.    It would be relevant to know what the consequences of the testing system are: E.g. are additional RCCs tested for certain antigens? Which antigens are tested in all donors already? Could it make sense to test for additional antigens routinely, e.g. at a certain frequency?

Response to comment: Thank you for your insightful comment. We have addressed this point by expanding the Methods and Discussion sections to clarify our current testing strategy and its potential implication.
Methods Line 81-92: Extended red blood cell phenotyping for antigens beyond ABO and RhD (such as Kell, Duffy, Kidd, and MNS systems) is not routinely performed at our institution. However, in cases where patients have a history of alloimmunization or require frequent transfusions (e.g., thalassemia or sickle cell disease patients), extended phenotyping may be considered on a case-by-case basis to improve transfusion compatibility
Discussion Line 303-309: Targeted antigen screening for high-risk populations is helpful to prevent complications related to alloimmunization. In the long run, extended antigen matching or genotypic matching in high-risk groups, such as thalassemia patients, should be a priority for optimizing transfusion safety. However, logistical and economic constraints often limit the routine implementation of such protocols. Under good blood supply conditions, future advancements in molecular genotyping may offer a more feasible approach to personalized antigen matching, reducing the risk of alloimmunization.

 

11.    Antibodies detected with relation to disease types should be discussed in view of the literature.

Response to comment: Thank you for your insightful comment. We have addressed this point by expanding in Discussion section. Line 250-257

Discussion: “While our findings indicate that cold-reactive autoantibodies were predominantly observed in patients with hematological malignancies and other hemato-oncologic conditions, no direct correlation between antibody specificity and primary diagnosis was established. This suggests that alloimmunization risk in transfusion-dependent patients is multifactorial, likely influenced not only by the underlying disease. The absence of a disease-specific alloantibody pattern aligns with previous studies, which emphasize the complex interplay between antigenic exposure and immune response

12.    What would be the impact if only 37°C screening were performed?
Response to comment: We have included a discussion on the implications of limiting screening to 37°C and its potential impact on transfusion safety. Line 197-200.
“Our findings indicate that screening at multiple temperatures enhances antibody detec-tion, particularly for cold-reactive antibodies. A 37°C-only approach could miss clinically relevant cold agglutinins, increasing transfusion risks.”

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript describes a retrospective study on 27,530 transfused patients in Serbia with the aim to identify the prevalence of red blood cells immunization.

Some observations:

• The pre-transfusion protocol includes ABO and RhD blood typing, irregular antibody screening, and crossmatch. It would be interesting to describe in how many cases an antibody was detected only by crossmatch with negative antibody screening and, in these cases, the clinical significance of the detected antibody.
• The pre-transfusion protocol (ABO and RhD blood typing, irregular antibody screening, and crossmatch) seems applyed to all patients, included thalassaemia or sickle cell anaemia patients: in these last cases an extended red blood cells typing is not provided? Please explain.
• It would be interesting to know the meaning of “insignificant” antibodies as reported in the text.

Author Response

Thank you for reviewing our manuscript and giving the opportunity for 
revision. We are very grateful for constructive and valuable comments which are all very helpful for revising and improving our paper.
We have incorporated suggestions; Please see below, point-by-point response to the reviewers’ comments and concerns.

 

1. It would be interesting to describe in how many cases an antibody was detected only by crossmatch with negative antibody screening and, in these cases, the clinical significance of the detected antibody.
Response to comment: Thank you for the comment. We added the data in the Result section Line 160-162: " Among the 630 patients with detected antibodies, 10 had a negative antibody screening but were identified through a positive crossmatch. Of these, 4 had anti-Jka, 3 had anti-Fya, and 3 had anti-M.
Also we added paragraph in Discussion secition. Line 283-286. " In our study, 1.59% of patients showed a positive crossmatch and negative irregular antibody screening because of heterozygous reagent cells. The majority of these cases involved low-prevalence antigens such as anti-Jka and anti-Lea.

2. The pre-transfusion protocol (ABO and RhD blood typing, irregular antibody screening, and crossmatch) seems applied to all patients, including thalassemia or sickle cell anemia patients: in these last cases, is extended red blood cell typing not provided? Please explain.
Response to comment: Thank you for point it out. We claryfied this by adding paragraph in Methodology. Line 87-92
" Extended red blood cell phenotyping for antigens beyond ABO and RhD (such as Kell, Duffy, Kidd, and MNS systems) is not routinely performed at our institution. However, in cases where patients have a history of alloimmunization or require frequent transfusions (e.g., thalassemia or sickle cell disease patients), extended phenotyping may be considered on a case-by-case basis to improve transfusion compatibility."
 and in Discussion Line 221-235: " The study confirms that thalassemia patients experience a significantly higher rate (9.69%) of alloimmunization compared to the general transfused population (2.29%). Despite the recognized benefits of antigen matching, extended red blood cell phenotyping for Kell, Duffy, Kidd, and MNS systems is not routinely performed at our institution. This may contribute to the higher risk of alloimmunization observed in transfusion-dependent patients, particularly those with thalassemia who often develop multiple alloantibodies due to cumulative antigen exposure. Several studies have demonstrated that extended antigen matching reduces alloimmunization rates in these patients by minimizing exposure to immunogenic non-self-antigens. On the other side, such a practice might increase the burden on blood banks by limiting available donor units. Furthermore, the cost-effectiveness of different levels of antigen matching should be considered. Future studies should explore the long-term outcomes of alloimmunization prevention strategies, particularly in resource-limited settings. Ultimately, the findings of this study emphasize the need for tailored testing protocols that strike a balance between reducing alloimmunization risk and maintaining transfusion efficiency.."

3. It would be interesting to know the meaning of “insignificant” antibodies as reported in the text.
Response to comment: Thank you for the comment. We've added a paragraph in Discussion
Line 191-197 "We were not able to avoid detecting clinically “insignificant” antibodies related to blood group antigen. Clinically “insignificant” antibodies do not typically cause hemolytic transfusion reactions and do not necessitate antigen-negative donor units for transfusion, such as anti-Lea, anti-Leb, anti-P1, and certain anti-M or anti-N antibodies. While their detection may delay transfusion due to additional compatibility testing, they are generally not associated with significant clinical consequences."

Reviewer 3 Report

Comments and Suggestions for Authors

This study provides a comprehensive analysis of RBC alloimmunization prevalence and specificity among transfusion-dependent patients at University Clinical Center of Serbia. With a large dataset of 27,530 transfusion records, the findings describe occurrence and distribution of alloantibodies, particularly against Rh and Kell antigens. The study underscores the challenges posed by multiple alloantibodies and highlights the importance of targeted antigen matching to enhance transfusion safety.

There are several major flaws in the study, and I recommend the authors to address them.

For publication in this journal, the study did not separate thalassemia patients as a specific group of transfusion recipients, so the alloimmunization of such patients was not separately presented, but they were included and analyzed as part of the general group of transfused patients.

The description of the study design is unclear. It states that the prevalence of RBC alloimmunization is being assessed in patients who received transfusions over a one-year period. The results of the pre-transfusion tests were analyzed. However, patients who previously had alloantibodies or autoantibodies are listed as exclusion criteria, which is contradictory. Such a description of the study means that it is necessary to monitor the development of RBC antibodies for a certain period of time after transfusion, which is not the case here. On the contrary, the authors monitored previous alloimmunizations in patients.

In that case, patients who were alloimmunized but whose antibody titer fell below the level of detection by described test methodology at the time of pre-transfusion testing should also be considered and analyzed to obtain accurate results.

Furthermore, it is not described how data from patients who were transfused multiple times during the stated period were analyzed since it is not clear if 27530 patient’s individual samples were analyzed or 27530 transfused patients (the authors state both in the text), which makes the statistical analysis unreliable.

The discussion does not follow the results, so for example there is no discussion anywhere about a significant increase in the rate of alloimmunization that was observed in a study with an increase in the number of transfusions and pregnancies.

The authors state that alloimmunization rates tend to be higher due to the increased likelihood of exposure to foreign antigen, and they do not link this to the clinical condition of the patient, e.g. patients with autoimmune diseases have a higher tendency for alloimmunization, while it is reduced in patients on immunosuppressive therapy.

There is an ongoing discussion about the impact of residual WBC on the development of RBC alloimmunization. The paper does not state which type of RBC units were used, leucoreduced or not.

While the study reinforces known patterns of RBC alloimmunization, it lacks innovation and deeper clinical implications. Incorporating genotypic analysis, broader demographic factors, and a discussion on effective prevention strategies and practical solutions or policy recommendations to reduce alloimmunization rates in transfusion-dependent patients, could enhance its impact.

Author Response

Thank you for reviewing our manuscript and giving the opportunity for 
revision. We are very grateful for constructive and valuable comments which are all very helpful for revising and improving our paper.
We have incorporated suggestions; Please see below, point-by-point response to the reviewers’ comments and concerns.

 

1. The study did not separate thalassemia patients as a specific group of transfusion recipients, so the alloimmunization of such patients was not separately presented, but they were included and analyzed as part of the general group of transfused patients.
Response to comment: Thank you for the comment. We've made the corretion in Results and Discussion section.
Line  173-174 " Of the 495/27530 (1.8%) patients with thalassemia, 48/495 (9.69%) were found to have antibodies detected.."
Line 221-235" The study confirms that thalassemia patients experience a significantly higher rate (9.69%) of alloimmunization compared to the general transfused population (2.29%). Despite the recognized benefits of antigen matching, extended red blood cell phenotyping for Kell, Duffy, Kidd, and MNS systems is not routinely performed at our institution. This may contribute to the higher risk of alloimmunization observed in transfusion-dependent patients, particularly those with thalassemia who often develop multiple alloantibodies due to cumulative antigen exposure. Several studies have demonstrated that extended antigen matching reduces alloimmunization rates in these patients by minimizing exposure to immunogenic non-self-antigens. On the other side, such a practice might increase the burden on blood banks by limiting available donor units. Furthermore, the cost-effectiveness of different levels of antigen matching should be considered. Future studies should explore the long-term outcomes of alloimmunization prevention strategies, particularly in resource-limited settings. Ultimately, the findings of this study emphasize the need for tailored testing protocols that strike a balance between reducing alloimmunization risk and maintaining transfusion efficiency. “

2. The description of the study design is unclear. It states that the prevalence of RBC alloimmunization is being assessed in patients who received transfusions over a one-year period. However, patients who previously had alloantibodies or autoantibodies are listed as exclusion criteria, which is contradictory.
Response to comment: Thank you for the comment, we’ve made correction.
Line 78-81 " The study assessed the prevalence rather than incidence of alloimmunization. In this manner, the study focused on all detected alloimmunization at the time of testing. The study did not monitor the development of RBC antibodies at the time after current transfusions."

3. Patients who were alloimmunized but whose antibody titer fell below the level of detection should also be considered.
Response to comment: Thank you for the comment, we’ve made correction.
Line 213-216„Additionally, some of alloimmunized patients may have had undetectable antibody titers at the time of pre-transfusion testing. This could have led to an underestimation of the true alloimmunization rate, particularly for antibodies known to exhibit dosage effects (e.g., anti-Kidd)..“

4.    It is not described how data from patients who were transfused multiple times during the stated period were analyzed since it is not clear if 27,530 individual samples were analyzed or 27,530 transfused patients.
Response to comment: Thank you for the comment, we’ve added the paragraph in Discussion
Line 75-81 „The study analyzed pre-transfusion test results from a total of 27530 unique patients who underwent pre-transfusion testing during the study period. If an alloantibody was detected, a retrospective review of the patient’s transfusion history was conducted to determine prior exposure to RBCs transfusions. The study assessed the prevalence rather than incidence of alloimmunization. In this manner, the study focused on all detected alloimmunization at the time of testing. The study did not monitor the development of RBC antibodies at the time after current transfusions..“

5.    The discussion does not follow the results, so for example there is no discussion anywhere about a significant increase in the rate of alloimmunization that was observed in the study with an increase in the number of transfusions and pregnancies.
Response to comment: Thank you for the comment, we’ve made correction.
Line 203-208 „Our findings confirm a significant correlation between the number of transfusions and pregnancies and the rate of alloimmunization (p < 0.05). Patients with repeated transfusions have greater exposure to foreign antigens, increasing the risk of antibody formation. Similarly, pregnancy-related alloimmunization, particularly against Rh and Kell antigens, is well-documented

6. The authors do not link alloimmunization rates to clinical conditions such as autoimmune diseases or immunosuppressive therapy.
Response to comment: Thank you for the comment, we’ve made correction.
Line 210-215 „While transfusion frequency is a key factor in alloimmunization, underlying clinical conditions also play a role. Patients with autoimmune diseases may exhibit heightened immune responses, predisposing them to alloimmunization. Conversely, patients on immunosuppressive therapy may have lower alloimmunization rates due to dampened immune responses [8]. Further research based on the current findings, should explore the aforementioned impacts on the patient population described in the study. “

7. The paper does not state which type of RBC units were used, leucoreduced or not.
Response to comment: Thank you for the comment, we’ve added the paragraph in Methodology and Discussion

Methodology
Line 83-85 Leukoreduction was not systematically applied to transfused RBC units during the study period, so this variable was not accounted for in alloimmunization analysis.
Discussion
Line 298-301: Although this study provides valuable insights into RBC alloimmunization prevalence, certain limitations may have influenced the alloimmunization rates. One such factor is the lack of systematic leukoreduction which would otherwise lead to minimizing residual white blood cells in order to reduce the risk of adverse reactions. 

8. While the study reinforces known patterns of RBC alloimmunization, it lacks innovation and deeper clinical implications.
Response to comment: Thank you for the comment, we’ve added the paragraph in Discussion
Line 306-312: Targeted antigen screening for high-risk populations is helpful to prevent complications related to alloimmunization. In the long run, extended antigen matching or genotypic matching in high-risk groups, such as thalassemia patients, should be a priority for optimizing transfusion safety. However, logistical and economic constraints often limit the routine implementation of such protocols. Under good blood supply conditions, future advancements in molecular genotyping may offer a more feasible approach to personalized antigen matching, reducing the risk of alloimmunization.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The Authors provided satisfactory answers to all observations.

Author Response

We sincerely thank the reviewer for the positive assessment and valuable feedback throughout the review process. We are pleased to know that our revisions and clarifications were satisfactory. We remain grateful for your time and constructive input, which helped improve the quality of the manuscript.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have made major revisions to the manuscript, that fit the scope of the journal and, in my opinion, do not require additional changes, other than correcting the references. 

Please correct references according to the Instructions for Authors, Reference List and Citations Style Guide.

Author Response

Thank you for your thorough review and positive feedback. We greatly appreciate your acknowledgment of the relevance and improvements made to the manuscript.

As per your suggestion, we have carefully revised the reference list to fully comply with the Thalassemia Reports "Instructions for Authors" and the journal's reference and citation style guide.