Multiple Myeloma Patients Undergoing Carfilzomib: Development and Validation of a Risk Score for Cardiovascular Adverse Events Prediction

Round 1

Reviewer 1 Report

This manuscript presents good quality results of a study focused on the major predictors of cardiovascular adverse events　(CVAEs) prior to Carfilzomib(CFZ) starting, applying the European Myeloma network management protocol(EMN). The topic addressed is interesting and deserves a constructive discussion. The study provides an important contribution to identify predictors of CVAEs in MM patients who underwent a standardized baseline assessment and especially in prospectively followed-up study during CFZ treatment. The proposed "CFZ CVAEs risk score" for MM patients is quite convincing. It is scientifically sound and contains sufficient interest to merit publication. The statistics used are appropriate, the conclusions derived from these, and the allowing for a patient-tailored follow up are consistent and sound.

Author Response

Dear Reviewer 1,

please accept our pleasure to share with you the results of this study, we are glad that you appreciated this work and found it interesting for its implications in clinical practice.

Reviewer 2 Report

The article “Multiple Myeloma Patients Undergoing Carfilzomib: Development and Validation of a Risk Score for Cardiovascular Adverse Events Prediction” by Astarita A, et al. prospectively analyzed the profile, frequency and risk factors of carfilzomib (CFZ)-associated cardiovascular adverse events (CVAEs) and also investigated the effectiveness of CVAE-assessment according to the European Myeloma Network protocol in 116 patients with relapsed/refractory multiple myeloma (RRMM) in the real-life setting. Establishment of the modality for optimal patient/treatment selection with CFZ-containing therapy is one of unmet clinical needs in MM practice, and, therefore, the research topic of this study is attractive, however, there are several critical concerns with this study regarding the research design and interpretation of the results.

1. The definition of “Major CVAEs” in this study made difficult to understand the true pathophysiological association/relationship between suggested risk factors and the nature of adverse events. In this study, major CVAEs included ischemic heart disease (acute coronary syndrome), chest pain, post infusion dyspnoea, hear failure and syncope regardless of its causative, and arrhythmia. While the underlying causative for ACS could be mostly, but not always, uniform, the underlying pathologic mechanisms for chest pain, post infusion dyspnoea, heart failure and syncope, might be widely variable, and, actually, might not be always associated with cardiovascular event. For instance, chest pain could be induced by ischemic hear disease, but also by bone event, pulmonary/mediastinal event, esophageal event, or even herpes zoster reactivation, etc. The underlying mechanism for post infusion dyspnoe is apparently different from that for ischemic hear disease or syncope. The underlying mechanisms for syncope could be also variable, sometimes caused by CVAEs, such as arrhythmia or cerebral arterial infarction, but also caused by other non-cardiovascular event, such as hypoglycemia etc. Apparently, mechanisms and the impacts of pre-existing diseases for ischemic heart disease, heart failure and arrythmia are different. However, authors mixed up all these events as one category designated as “major CVAEs”, and investigated the risk factors for this category. Risk factors are needed to be assessed for individual events included in “major CVAEs” according to their pathophysiology. It is also needed to assess the underlying causatives for “chest pain”, “dyspnoea”, the nature of “heart failure” (i.e, caused by ischemic heart disease, severe arrhythmia, valve diseases, etc.), and “syncope” to know the true relationship between each CVAE and their risk factor.
2. Moreover, according to their definition, they even included patients with hypertension into major CVAEs category, in case patients showed symptoms of both “major CVAEs” and hypertension. This might miss the inclusion of patients who should be evaluated for hypertension. Considering the number of events, it is possible that the risk factors could be assesses only for hypertension with this cohort.
3. In Table 2, the ratios of each event are expected to be shown by considering whole patients’ number (i.e., N=116) as denominator, so that we are able to know the frequency of events easily.
4. In Table 3, there may be error at the top rank. Both categories were described as “No CVAEs”.
5. In Table 3, although the statistical analyses suggested the differences of mean scores of daytime SBP, daytime SD, 24h SD, night time SD, blood pressure variability, LVMi, GLS, and PWM, there were considerable overlap when compared their ranges. Authors are required to discuss about the utility of each predictor candidates in clinical setting.
6. While they separated patients into non-CVAE group, major CVAE group and hypertensive CVAE group according to the event profile, and emphasized the difference of risk factors, especially pre-existing diseases, between major CVAEs and hypertensive CVAEs (Table S4), it was curious that they then developed the risk predictive score for CVAEs by grouping major CVAEs and hypertensive CVAEs as a one category. Risk predicting models are normally individually developed for pathophysiologically different events.
7. In Figure 3, the concordance between the real frequency of CVAEs and risk prediction line was poor especially in high-risk patients (prediction event (%)>80) and in low-risk patients (prediction event (%) <30), although the risk prediction seemed to be fine in the middle range. This phenomenon needs to be discussed for conceivable reason, as this means the suggested risk prediction model is less useful for detecting truly high- and low-risk patients.
8. In total, at the moment, we are unable to know what kind of event is truly at high risk by the prediction model suggested  in this article. 

Author Response

Dear Reviewer 2,

please accept our pleasure to share with you the results of this study, we thank you for your helpful comments. We itemized the modifications in our response and highlighted the changes in the original manuscript (red color). Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The article is very well written and relevant.

Only thing I think can add to the manuscript is that CFZ is very rarely used as a single treatment, most likely the patients in the study will have recived other drugs together with CFZ, this could be added (and if this is a true single drug study, a motivation why this was the choice could also be of interest since there are quite a lot of agents available for this patient population).

Alternativly it would be interesting if some small section was used to discuss if there was any change in outcome depending on the treatment history before entering this study.

Author Response

Please see the attachment, we highlighted the changes in the original manuscript in green color. 

Author Response File: Author Response.docx

Reviewer 4 Report

This is an important study as it is one of the few prospective studies evaluating cardiovascular toxicity with carfilzomib.  It is generally well written and of interest to the myeloma community.  However, it is very confusing to follow, likely overestimates the toxicity of this agent and may cause harm in this regard if heeded by clinicians.

1. The baseline population here must be discussed more fully.  50% smokers? Over 40% with poorly controlled HTN??  This DRAMATICALLY affects the outcome, especially when grade 1 events are considered in the total of hypertensive events.
2. There is little guidance here as to the relevance of this study.  What does a clinician do when they read this?  Control BP?
3. Many of the tests done are not available to most patients.  That must be explained and cited as a weakness.
4. The validation portion of the study as listed in the title is not clear in the text.
5. Table 2 is the center piece of the paper and is incredibly confusing - the mix of numbers vs % is hard to follow.  It must be totally redone to better present the data.
6. The dosing findings are odd - that dosing did not affect CV events and only 2 patients needed dose reduction?  How is that possible?
7. When patients without pre-existing HTN are evaluated, what is the true CV risk?
8. The discussion is hard to follow when the scoring is explained. 
9. More limitations are required including the baseline population, availability of testing, etc.

Author Response

Please see the attachment, we highlighted the changes in the original manuscript in green color.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

While authors replied to most of my concerns by adding some descriptions, I still expect to convince more to the "Response 2" raised against "Point 2". Patients with both major CVAEs and hypertension should be evaluated as members of both "major CVAEs" and "hypertension", because they were actually affected by hypertension. No practical reason exists for the exclusion of those patients from "hypertension" group.  Also, again, major CVAEs inculde events with widely different  pathophysiology/pathogenesis. Again, for instance, I would emphasize the pathophysiology and pathogenesis of ischemic heart disease, arrhythmia and cardiac failure should be totally different, therefore, the underlying risks should be separately analyzed. 

I now understand that authors tried to know the risk parameter for treatment discontinuation of CFZ treatment. If this was true, in turn, why did authors need to separate patients into "major CVAE" group and "hypertension group" for risk prediction?

Author Response

Please see the attachment, we highlighted the changes in the original manuscript in green color.

Author Response File: Author Response.docx

Reviewer 4 Report

Authors have adequately addressed my concerns

Round 3

Reviewer 2 Report

I don' t think authors appropriately respond to my query about the separate evaluation of risk factors for hypertensive CVAEs and  major CVAEs (The pathophysiology and pathogenesis of ischemic heart disease, arrhythmia and cardiac failure should be totally different, therefore, the underlying risks should be separately analyzed.). I am not sure why they do not try this.