Prognostic Impact of 24-Hour Pulse Pressure Components in Treated Hypertensive Patients Older Than 65 Years

Round 1

Reviewer 1 Report

I read with interest the paper by Coccina et al. titled ”Prognostic Impact of 24-hour Pulse Pressure Components in Elderly Treated Hypertensive Patients”.

 

I thank the authors and the editor for the opportunity to review this manuscript.

 

However, I have some reservations as detailed below.

 

Major points

In this retrospective study, the authors evaluated 24-hour blood pressure in 745 elderly treated hypertensive patients and calculated pulse pressure (PP) as so-called elastic PP (elPP) and stiffening PP (stPP). The authors then evaluated the association between PP, elPP and stPP on the risk of the 284 cardiovascular events observed over about 8 years of follow up. After adjustment for covariates, PP was tended to associate with a higher risk, while elPP was associated with risk of event while stPP was not.

 

The primary outcome and a hypothesis are not stated.

 

Due to the retrospective nature of the study, and the relatively low number of patients and events, the authors should soften the wording of their conclusion. Also the authors have included both patients with and without cardiovascular diseases or risk factors apart from hypertension.

 

The associations between PP, elPP and stPP and CVD risk have been evaluated previously in larger studies.

 

The authors have not clearly stated in the title or abstract that the study is retrospective.

 

The authors had a cohort of 2264 “treated” patients aged 30-90 years and used only 745 in the study and a high number of 103 patients were lost during follow up. It is unclear how the patients were selected, I suppose that most of the 2264 were aged above 65 years but only about a third were included? The authors should clarify the patient selection and consider to use a flow chart.

 

The authors should describe the literature in more detail in the introduction.

 

The discussion should be rewritten – it reads as the results section of various previous publications and has very little actual discussion.

 

The authors state that hypertension was treated but do not specify how this was evaluated, which BP targets were used or how well blood pressure was regulated.

24 hour SBP and DBP were 133 ± 14 / 74 ± 8 mmHg whereby hypertension seems to have been controlled in some subjects and not in others whereas some subjects may not even have been hypertensive (the authors have not stated inclusion criteria, e.g. BP >140/90 and/or antihypertensive treatment). Further it is unknown whether blood pressure was regulated during the about 8 years follow up is also unknown. This should be described and also noted in the limitations.

 

The authors should consider to evaluate whether the association of PP variables were stronger than that of e.g. MAP.

 

See specific points below.

 

 

Abstract

Please state that the study was retrospective.

 

Please state whether the patients had cardiovascular diseases other than hypertension.

 

Please describe the study in the methods part i.e. association to a combined endpoint of cardiovascular events.

 

“were weakly correlated (r=0.21) and 24-hour stPP was a small fraction of PP (19.5%)” – is this an important outcome? Please omit. Was stPP a fixed fraction of PP? If so it is irrelevant.

 

”hospitalized heart failure”- ”heart failure hospitalization”

 

“were significantly associated with outcome” – “were associated with the combined outcome”.

The word “significantly” is generally redundant.

 

”marginally associated” – what do the authors mean by “maginally” please revise to state that there was a trend or better that there was no association.

 

“was to assess the prognostic impact of 24-hour PP” – please clarify, the impact on what?

 

 

Introduction

“characterized by a “curving up” at elevating pressures” – unclear, please revise.

 

To a reader unfamiliar with elPP and stPP it would be helpful to add some more descriptions, e.g. which factors are used to calculate the values, what do higher or lower values indicate? Please describe breifly the evidence on these variables on clinical outcomes.

 

Please state the primary objective and a hypothesis.

 

 

Materials and methods

The process of informed consent and ethical approval of the study is noted at the end of the manuscript, but should also be described in the methods section.

 

“This study is a retrospective examination of prospectively collected data” – it seems to be a retrospective study and the authors should make this clear.

 

Please list inclusion and exclusion criteria.

 

The authors had a cohort of 2264 patients aged 30-90 years and used only 745 in the study and 103 patients were lost during follow up. It is unclear how the patients were selected, I suppose that most of the 2264 were aged above 65 years.

 

“at least after 5 min of rest” - “after at least 5 min of rest”.

 

Was ambulatory BP measured for all 2264 patients?

 

”on a typical day” – consider to omit.

 

Why index left atrial diameter and not ml/m2?

 

Why define a low EF as <50% and not follow e.g. the common definition of HFrEF, HFmrEF and HFpEF or ≤40% vs. >40%?

 

How was “treated hypertension” defined? BP below which threshold?

 

The authors recorded variables and obtained outcomes data either from a database or by telephone interview, this should be described.

 

Please state whether the persons who evaluated the combined endpoint was blinded or not to the blood pressure data.

 

Which BP values were used if a patient underwent more than one 24 h BP measurement? The first or a mean value?

 

 

Statistical analysis

“Standard descriptive statistics were used.” – please specify or omit.

 

Please state which factors were used in the multivariate analyses.

 

Results

Age 72.5 ± 5 indicate that age was 62-82 in 95% of the patients but the age had to be above 65. Further, age is likely not normally distributed. Please use median (IQR) revise throughout the manuscript where applicable.

 

27.5 ± 4  - both numbers should have the same numbers of decimals, similar for the BP measurements. Please revise throughout the manuscript.

 

“Previous events, n (%) 75 (10)” – it would be better to state the number of patients with e.g. ischemic heart disease, peripheral arterial disease etc.

 

“ALVSD, asymptomatic left ventricular systolic dysfunction” – how was this determined? The authors used above vs below 50% LVEF regardless of symptoms in the methods section.

 

Fig 1.

Please add also PP.

 

”Pearson’s and Spearman’s correlation” – one analysis is enough.

 

What is the relevance of all the correlations? There seems to be no clinical relevance and it is not the objective of the study and can be omitted. If the authors want to keep the correlation analyses they can be moved to the end of the results section.

 

Table 3 – the evaluation of antihypertensive treatment and the time point is not described in the methods section.

Did no subject receive 4 or 5 drugs?

 

“after adjustment for various covariates, as performed for 24-hour PP components” – what do the authors mean by “as performed for 24-hour PP components”?

 

“PP was marginally associated with risk,” – the meaning of “marginally” is not clear, please add the 95% CI and P-value in the main text and replace the word marginally.

 

Fig 3

”Smoke” “family history of cardiovascular disease“ – how were these defined? This should be described in the methods section.

 

The authors should consider to add a multivariate analysis of e.g. “24-h MAP” in order to evaluate whether the PP variables are better than just the blood pressure.

 

 

Discussion

“74% of deaths in this group occurred in patients older than 58 years” – this and other details and the exact numbers etc. from many variables in previous studies that are irrelevant for the present study can be omitted.

 

L236-240 can be described in words rather than copying all the results of a previous publication. This point concerns most of the writing in the discussion.

The authors should summarize their results and discuss it in relation to the results of the literature, the differences, interpretation, limitations and clinical relevance and the perspective for future research. Not just a detailed summary of all the numbers from several papers.

 

Previous studies stratified patients according to heart rate above and below 70, this might be done in this study but the present data set is likely too small.

 

The discussion is a massive wall of text, please divide it into paragraphs.

 

“In patient group aged 60-70 years” – “In patients aged 60-70 years”.

 

”weakly correlated (r=-0.075)” – it seems that there was no correlation.

 

“different ethnic groups that were also analyzed individually or together” – “different ethnicities in the various studies”.

 

The authors should also try to describe what elPP and stPP means clinically and why higher/lower values may be associated to CVD risk.

 

The loss of patients during follow-up should be mentioned in the limitations.

 

 

Author Contributions

AMP, CC, JP, OT appear not to qualify for coauthorship according to ICMJE criteria.

Author Response

Response to the Editor

We thank the Editor for the comments and suggestions.

There are two important points that need your attention and revision:
1.Please to extend the main text of your manuscript to have more than 4000 words when revision.

Response: we have now extended the main text to more than 4000 words.

2.Section 2 (materials and methods) is with high repetition rate, please see it in the attachment file and to modify when revision.

Response: we recognize that portions of our manuscript may resemble our own and other published works in the literature. However, this depends on the fact that both standard international terminologies and terminologies from our own usual methodologies are used. We have now tried to modify some sentences as much as possible, trying not to distort the meaning.

Response to Reviewer 1

We thank the Reviewer for the comments and suggestions.

1) The primary outcome and a hypothesis are not stated.

Response: the primary outcome was to evaluate the impact of PP components on the occurrence of cardiovascular events hypothesizing that the various components could have had a different influence on cardiovascular outcome. We have now added these considerations at the end of the Introduction section.

2) Due to the retrospective nature of the study, and the relatively low number of patients and events, the authors should soften the wording of their conclusion. Also the authors have included both patients with and without cardiovascular diseases or risk factors apart from hypertension.

Response: we have now softened the wording of conclusion. All patients included were treated hypertensive patients, a minority of whom had already experienced a first cardiovascular event.

3) The associations between PP, elPP and stPP and CVD risk have been evaluated previously in larger studies.

Response: the association between PP, elPP and stPP (a novel conceptualizazion of  PP) and CVD risk has been evaluated in only 3 previous studies and different results have been obtained probably because of different ethnic groups, different ethnicities analyzed together, different population types including normotensive subjects, untreated and/or treated hypertensive patients, different age range of studied populations, different studied endpoints, different covariates used in multivariate adjustment. Given that the relevance of PP increases with aging and appears more relevant in hypertensive patients, we focused on the impact of PP components in an elderly population with  treated hypertension. We have now better described this aspect in the Introduction section.

4) The authors have not clearly stated in the title or abstract that the study is retrospective.

Response: we have now reported in the abstract that this is a retrospective study.

5) The authors had a cohort of 2264 “treated” patients aged 30-90 years and used only 745 in the study and a high number of 103 patients were lost during follow up. It is unclear how the patients were selected, I suppose that most of the 2264 were aged above 65 years but only about a third were included? The authors should clarify the patient selection and consider to use a flow chart.

Response: our database of hypertensive patients receiving drug treatment at baseline included 2264 subjects (1073 younger than 60 years; 1191 older than 60 years; 1519 younger than 65 years; 745 older than 65 years; please, see references 31-35). Given that the relevance of PP increases with aging, given that one of our interest in clinical research is geriatric cardiology, given that in western countries the most frequent age adopted today to define transition to the condition of elderly is 65 years, we focused our study to treated hypertensive patients older than 65 years. Those lost during follow-up (103) refers to the global population of 2264 subjects (< 5%) (we have now better explained this aspect).  

6) The authors should describe the literature in more detail in the introduction.

Response: we have now tried to describe the literature in more detail in the Introduction section.

7) The discussion should be rewritten – it reads as the results section of various previous publications and has very little actual discussion.

Response: we have now tried to improve the Discussion section. We have tried to better compare our results with previous studies, after their description, and try to explain any differences.

8) The authors state that hypertension was treated but do not specify how this was evaluated, which BP targets were used or how well blood pressure was regulated.

Response: our patients were on drug treatment and when they came to wear or return the ambulatory blood pressure monitor they showed the medications they were taking. As it happens in many cohorts including treated hypertensive individuals, some of these patients had controlled blood pressure, some had borderline blood pressure, and some had uncontrolled blood pressure. We have now better reported inclusion criteria. The cutoff value to define normal/controlled hypertension (140/90 mmHg) was reported in the previous version (and is reported in this version).

9) 24 hour SBP and DBP were 133 ± 14 / 74 ± 8 mmHg whereby hypertension seems to have been controlled in some subjects and not in others whereas some subjects may not even have been hypertensive (the authors have not stated inclusion criteria, e.g. BP >140/90 and/or antihypertensive treatment). Further it is unknown whether blood pressure was regulated during the about 8 years follow up is also unknown. This should be described and also noted in the limitations.

Response: As it happens in many cohorts including treated hypertensive individuals, some of these patients had controlled blood pressure, some had borderline blood pressure, and some had uncontrolled blood pressure. In the previous version we have reported that we studied 745 treated hypertensive patients older than 65 years, but we have now better reported inclusion criteria. Part of these patients were followed up in our outpatient clinic and part by their family doctors. Thus, blood pressure control was not known in all the patients. However, we do not think that this aspect may have differentially influenced the different components of PP and their prognostic impact. We have now reported this aspect in the limitation section.

10) The authors should consider to evaluate whether the association of PP variables were stronger than that of e.g. MAP.

Response: MAP was significantly associated with outcome, HR (95% CI) for cardiovascular events per 1 SD increment was 1.28 (1.12-1-48) in the model including 24-hour PP and 1.25 (1.09-1.44) in the model including 24-hour elPP and 24-hour stPP. However, 24-hour elPP was associated with cardiovascular outcome independently of MAP. We have now added these data at the end of the Results section.  

Abstract

11) Please state that the study was retrospective.

Response: we have now stated that this is a retrospective study.

12) Please state whether the patients had cardiovascular diseases other than hypertension.

Response: all these patients had treated hypertension. Only a minority (10%) also had a previous cardiovascular event. For word limitation in the abstract we have reported this aspect in the text (Results section).

13) Please describe the study in the methods part i.e. association to a combined endpoint of cardiovascular events.

Response: we have now reported the suggested description.   

14) “were weakly correlated (r=0.21) and 24-hour stPP was a small fraction of PP (19.5%)” – is this an important outcome? Please omit. Was stPP a fixed fraction of PP? If so it is irrelevant.

Response: We have now omitted this part.

15) ”hospitalized heart failure”- ”heart failure hospitalization”

Response: we have corrected the sentence.

16) “were significantly associated with outcome” – “were associated with the combined outcome”.

Response: we have now corrected the sentence.

17) The word “significantly” is generally redundant.

Response: we have now deleted this word.

18) ”marginally associated” – what do the authors mean by “maginally” please revise to state that there was a trend or better that there was no association.

Response: we have now corrected the sentence.

19) “was to assess the prognostic impact of 24-hour PP” – please clarify, the impact on what?

Response: we had previously reported “the prognostic impact”. As suggested, we have now added in the methods: “The association of these PP components to a combined endpoint of cardiovascular events was investigated.”

Introduction

20) “characterized by a “curving up” at elevating pressures” – unclear, please revise.

Response: we how tried to better express this concept.

21) To a reader unfamiliar with elPP and stPP it would be helpful to add some more descriptions, e.g. which factors are used to calculate the values, what do higher or lower values indicate? Please describe breifly the evidence on these variables on clinical outcomes.

Response: we have now tried to better express the suggested aspects. 

22) Please state the primary objective and a hypothesis.

Response: the primary outcome was to evaluate the impact of PP components on the occurrence of cardiovascular events hypothesizing that the various components could have had a different influence on cardiovascular outcome. We have now added these aspects at the end of the Introduction section.

Materials and methods

23) The process of informed consent and ethical approval of the study is noted at the end of the manuscript, but should also be described in the methods section.

Response: we used the template of the Journal that requires this information on its specific site. However, we have now reported this information also in the Methods section.

24) “This study is a retrospective examination of prospectively collected data” – it seems to be a retrospective study and the authors should make this clear.

Response: these data were prospectively collected. Now, in our database, we retrospectively analyze this new topic. This terminology has been used and published in other studies. We have now partly changed this sentence. 

25) Please list inclusion and exclusion criteria.

Response: we have now added inclusion and exclusion criteria.

26) The authors had a cohort of 2264 patients aged 30-90 years and used only 745 in the study and 103 patients were lost during follow up. It is unclear how the patients were selected, I suppose that most of the 2264 were aged above 65 years.

Response: our database of hypertensive patients receiving drug treatment at baseline included 2264 subjects (1073 younger than 60 years; 1191 older than 60 years; 1519 younger than 65 years; 745 older than 65 years; please, see references 31-35). Given that the relevance of PP increases with aging, given that one of our interest in clinical research is geriatric cardiology, given that in western countries the most frequent age adopted today to define transition to the condition of elderly is 65 years, we focused our study to treated hypertensive patients older than 65 years. Those lost during follow-up (103) refers to the global population of 2264 subjects (< 5%). We have now better reported these aspects.  

27) “at least after 5 min of rest” - “after at least 5 min of rest”.

Response: we have now changed the sentence.  

28) Was ambulatory BP measured for all 2264 patients?

Response: Yes.

29) ”on a typical day” – consider to omit.

Response:  we have now omitted this part.

30) Why index left atrial diameter and not ml/m2?

Response: Left atrial dimension may be indexed for various parameters. In our database that dates back to the early nineties we have reported the size of the left atrium indexed by body surface area.

31) Why define a low EF as <50% and not follow e.g. the common definition of HFrEF, HFmrEF and HFpEF or ≤40% vs. >40%?

Response: A cutoff value of 50% is generally used to define reduced ejection fraction independently of its sub-classification.

32) How was “treated hypertension” defined? BP below which threshold?

Response: our patients were on drug treatment and when they came to wear or return the ambulatory blood pressure monitor they showed the medications they were taking. As it happens in many cohorts including treated hypertensive patients, some of these patients had controlled blood pressure (< 140/90 mmHg, as limit of normal/controlled BP as reported in the previous version), some had borderline blood pressure, and some had uncontrolled blood pressure.  

33) The authors recorded variables and obtained outcomes data either from a database or by telephone interview, this should be described.

Response: variables were collected during the first clinic visit or when patients wore/brought back the ambulatory BP monitor (that is, at baseline; we have now better expressed this concept). Outcomes were collected at a clinic visit or by telephone interview, but patients’ records were always requested and obtained to confirm outcomes (as reported in the previous version of the manuscript). 

34) Please state whether the persons who evaluated the combined endpoint was blinded or not to the blood pressure data.

Response: in this study, the investigator (F.C) who evaluated the combined endpoint was blinded to the BP data. We have now reported this aspect. 

35) Which BP values were used if a patient underwent more than one 24 h BP measurement? The first or a mean value?

Response: we used the data from the examination carried out at the baseline. We have now better reported this aspect.

Statistical analysis

36) “Standard descriptive statistics were used.” – please specify or omit.

Response: we have now specified.  

37) Please state which factors were used in the multivariate analyses.

Response: we have now reported factors included in multivariate analyses in the statistical analysis section (in the previous version and now they are also reported in the Figure 3 legend).

Results

38) Age 72.5 ± 5 indicate that age was 62-82 in 95% of the patients but the age had to be above 65. Further, age is likely not normally distributed. Please use median (IQR) revise throughout the manuscript where applicable.

Response: we have now reported “Median age was 71 years [interquartile range 68-76 years]”. 

39) 27.5 ± 4  - both numbers should have the same numbers of decimals, similar for the BP measurements. Please revise throughout the manuscript.

 

Response: we have now revised this aspect.

 

40) “Previous events, n (%) 75 (10)” – it would be better to state the number of patients with e.g. ischemic heart disease, peripheral arterial disease etc.

 

Response: we have now reported the number of patients with the specific previous event.

 

41) “ALVSD, asymptomatic left ventricular systolic dysfunction” – how was this determined? The authors used above vs below 50% LVEF regardless of symptoms in the methods section.

 

Response: our patients underwent echocardiographic examination. Those without symptoms and EF < 50% were defined as having asymptomatic LV systolic dysfunction. We have now better reported this aspect.

 

42) Fig 1. Please add also PP.

 

Response: we have now added 24-hour PP.

 

43) ”Pearson’s and Spearman’s correlation” – one analysis is enough.

 

Response: we have now reported Pearson’s correlation only.

 

44)What is the relevance of all the correlations? There seems to be no clinical relevance and it is not the objective of the study and can be omitted. If the authors want to keep the correlation analyses they can be moved to the end of the results section.

 

Response: we have now omitted this paragraph.

 

45) Table 3 – the evaluation of antihypertensive treatment and the time point is not described in the methods section. Did no subject receive 4 or 5 drugs?

 

Response: we have now reported in the methods section: “For this specific study, the inclusion criteria were: 1) hypertensive patient; 2) use of antihypertensive drugs; 3) age older than 65 years. In Table 3 we have now reported “Triple therapy or more” and we have now reported in the text the percentage of patients using more than 3 drugs.   

 

46) “after adjustment for various covariates, as performed for 24-hour PP components” – what do the authors mean by “as performed for 24-hour PP components”?

 

Response: we have now changed the sentence.

 

47) “PP was marginally associated with risk,” – the meaning of “marginally” is not clear, please add the 95% CI and P-value in the main text and replace the word marginally.

 

Response: we have now performed the suggested changes.

 

48) Fig 3 ”Smoke” “family history of cardiovascular disease“ – how were these defined? This should be described in the methods section.

 

Response: we have now described these aspects in the methods section.

 

49) The authors should consider to add a multivariate analysis of e.g. “24-h MAP” in order to evaluate whether the PP variables are better than just the blood pressure.

 

Response: we have now reported, at the end of the Results section,  the association of MAP with cardiovascular events which was significant in both model including 24-hour PP and that including 24-hour elPP and 24-hour stPP. However, 24-hour elPP was associated with risk independently of MAP. 

Discussion

50) “74% of deaths in this group occurred in patients older than 58 years” – this and other details and the exact numbers etc. from many variables in previous studies that are irrelevant for the present study can be omitted.

Response: we have omitted this and some other aspects.

51) L236-240 can be described in words rather than copying all the results of a previous publication. This point concerns most of the writing in the discussion.

Response: we have now changed this part, and other parts, of the Discussion section.

52) The authors should summarize their results and discuss it in relation to the results of the literature, the differences, interpretation, limitations and clinical relevance and the perspective for future research. Not just a detailed summary of all the numbers from several papers.

Response: we have discussed our results in relation to previous date (after their description), reporting similarities and differences and trying to explain differences. Limitations and clinical relevance were also reported. We have now added other concepts.

53) Previous studies stratified patients according to heart rate above and below 70, this might be done in this study but the present data set is likely too small. 

Response: In the first study by Gavish and Bursztyn [24], an analysis according to median heart rate was performed and 24-hour PP and 24-hour stPP (but not 24-hour elPP) predicted all-cause mortality only in patients with heart rate less than 70 beats/min. In the second paper by Bursztyn et al. [25], evaluating the Ohasama population, an analysis according to median heart rate was performed and 24-hour elPP (but not 24-h stPP) predicted total and cardiovascular mortality only among individuals with median heart rate lower than 68.5 beats/min. In the third study by Gavish et al. [26], assessing the IDACO population, an examination according to the median heart rate was not performed in the main analysis because 24-hour rate was included as a covariate in multivariate analysis, and when focusing on patients older than 60 years, 24-hour PP and 24-hour elPP were similarly associated with increased risk of cardiovascular events, but 24-hour stPP was marginally associated with poor outcome (P=0.061 in patients aged 60-70 years and P=0.045 in those aged more than 70 years). We performed the same approach used in the third study. We included 24-hour heart rate in the multivariate analyses obtaining the reported results. Dividing the population into two groups would have reduced the number of events for each group, weakened the event-to-variable ratio in each model, and weakened statistical power. Therefore, we preferred to analyze the population as a whole and to adjust for 24-hour heart rate in multivariate Cox models.

54) The discussion is a massive wall of text, please divide it into paragraphs.

Response: we have now divided the Discussion section into paragraphs.

55) “In patient group aged 60-70 years” – “In patients aged 60-70 years”.

Response: we have now changed the text as suggested.

56) ”weakly correlated (r=-0.075)” – it seems that there was no correlation.

Response: we have now deleted this sentence.

57) “different ethnic groups that were also analyzed individually or together” – “different ethnicities in the various studies”.

Response: we have now changed the text as suggested.

58) The authors should also try to describe what elPP and stPP means clinically and why higher/lower values may be associated to CVD risk.

Response: we have now tried to better describe the aspects indicated. 

59) The loss of patients during follow-up should be mentioned in the limitations.

Response: we have now better reported that lost patients (103) refer to the entire population of 2264 patients (< 5%), and not the group aged more than 65 years. This aspect does not seem to be a limitation.

60) AMP, CC, JP, OT appear not to qualify for coauthorship according to ICMJE criteria.

Response: We apologize for this mistake. All authors contributed to the writing of the manuscript, in addition to other specific roles. We have now reported this aspect in the specific section of the manuscript.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

The aim of this study was to assess in elderly treated hypertensive patients, the prognostic impact of 24-hour pulse pressure (PP) and its “elastic” and “stiffening” components (24-hour elPP and 24-hour stPP, respectively) reflecting, respectively, constant arterial elasticity, and the tendency of the arteries to stiffen under higher pressure.  The results show that after adjustment for various covariates, 24-hour elPP was a better predictor of cardiovascular outcomes than 24-hour PP and 24-hour stPP.

The study results are clinically significant and provide important contribution to diagnostic methods aiming to associate arterial properties with outcomes.

I have the following comment:

In the Discussion the authors described results of previous studies that demonstrated different prognostic significance for lower/higher than median heart rate, suggesting that such analysis, that is missing from the present study, must be added to the results in order to compare the present results with previous studies and thus, increasing the diagnostic power of the method.

Author Response

Response to the Editor

We thank the Editor for the comments and suggestions.

There are two important points that need your attention and revision:
1.Please to extend the main text of your manuscript to have more than 4000 words when revision.

Response: we have now extended the main text to more than 4000 words.

2.Section 2 (materials and methods) is with high repetition rate, please see it in the attachment file and to modify when revision.

Response: we recognize that portions of our manuscript may resemble our own and other published works in the literature. However, this depends on the fact that both standard international terminologies and terminologies from our own usual methodologies are used. We have now tried to modify some sentences as much as possible, trying not to distort the meaning.

Response to Reviewer 2

We thank the Reviewer for the comments and suggestions.

1) In the Discussion the authors described results of previous studies that demonstrated different prognostic significance for lower/higher than median heart rate, suggesting that such analysis, that is missing from the present study, must be added to the results in order to compare the present results with previous studies and thus, increasing the diagnostic power of the method.

Response: In the first study by Gavish and Bursztyn [24], an analysis according to median heart rate was performed and 24-hour PP and 24-hour stPP (but not 24-hour elPP) predicted all-cause mortality only in patients with heart rate less than 70 beats/min. In the second paper by Bursztyn et al. [25], evaluating the Ohasama population, an analysis according to median heart rate was performed and 24-hour elPP (but not 24-h stPP) predicted total and cardiovascular mortality only among individuals with median heart rate lower than 68.5 beats/min. In the third study by Gavish et al. [26], assessing the IDACO population, an examination according to the median heart rate was not performed in the main analysis because 24-hour rate was included as a covariate in multivariate analysis, and when focusing on patients older than 60 years, 24-hour PP and 24-hour elPP were similarly associated with increased risk of cardiovascular events, but 24-hour stPP was marginally associated with poor outcome (P=0.061 in patients aged 60-70 years and P=0.045 in those aged more than 70 years). We performed the same approach used in the third study. We included 24-hour heart rate in the multivariate analyses obtaining the reported results. Dividing the population into two groups would have reduced the number of events for each group, weakened the event-to-variable ratio in each model, and weakened statistical power. Therefore, we preferred to analyze the population as a whole and to adjust for 24-hour heart rate in multivariate Cox models.

 

 

Round 2

Reviewer 1 Report

I commend the authors for their revision, the manuscript is improved.

 

I have a few more comments.

 

Particularly the authors should continue to work on the discussion.

 

Abstract

”traditional 17 parameters including” – can be omitted.

 

 

“had a trend of association with risk” – “tended to associate with increased risk”

 

 

Introduction

”not unanimous results” – try e.g. “conflicting results”.

 

 

Methods

Consider to the revise stating that you “included the 745 patients older than 65 years from the cohort of 2264”.

 

“One hundred and three patients were lost during follow-up from the global 72 population of 2264 subjects” – this information is of little relevance, better state the loss during follow-up of the included 745 patients.

 

”Study 78 population came” – ”The study population”.

 

“Clinic systolic BP <140 mmHg and clinic diastolic BP <90 mmHg were defined as 88 normal.” – Still the authors’ definition of hypertension is unclear, was hypertension defined as SBP > 140 and DBP > 90 mmHg?

 

 

Statistical analysis

”we included variables that were significantly associated 136 with outcome in univariate analysis, that is, age, smoking…” – Unclear, are these all the used variables used or were these the ones that were significant in univariate analysis and then used in multivariate analysis?

Did the authors use more variables in the univariate analysis. Please clarify and state all used variables in the univariate analysis and how variables for the multivariate analysis were chosen.

 

”asymptomatic LV systolic 140 dysfunction” – This definition is different from that stated in L115-116 which only defines LVEF as equal to and above vs. below 50% regardless of symptoms. Please use the same terminology throughout the manuscript.  

Did none of the patients with LVEF <50% have symptoms?

 

“because of the potential differences in PP between men and 141 women”

and

“because of its reported potential impact on PP, and MAP 142 to account for the steady component of BP beyond the pulsatile component represented 143 by PP”

and

“used because of their potential different influence on afterload, preload, cardiac contractility and stroke volume, and possibly on 145 arterial wall properties“

– for clarity, this can be omitted, put in parentheses or stated in a separate sentence.

 

Wasn’t MAP used in the model?

 

Results

Fig 2 “r=0.21” – consider to add the P-value.

 

“Among patients using > 3 three drug, 210 about 20% received 4-5 drugs.” This can be omitted and a line for four or more drugs can be added in Table 3.

 

Fig 3 – please use a smaller range and put more numbers on the y-axis. The figure is not very informative at the present.

 

It seems that the authors evaluated the univariate association of each of the many factors to the combined outcome but this data is not shown, only for PP, elPP and stPP. Consider to include a figure of all the univariate results. Have I misunderstood something?

This might give some insights into the risk factors in the group and point to new research and perspectives. If elPP is a better or worse predictor than other factors it might be mentioned in the discussion.

 

 

Discussion

”refine risk profile” – ”describe the risk profile” or “estimate the risk”.

 

“Moreover, for the first time,” – delete. Consider to merge this sentence with the next one.

 

”there were 103 events” – here and elsewhere there is a lot of specific data from the literature. In this instance the authors can write e.g. that about xx% of the patients had a cardiovascular event.

 

”Ohasama study population” – also irrelevant detail.

Please revise throughout and omit details that are irrelevant for the present study. The interested reader can review the literature.

For example the exact ratio of men and women of previous studies are irrelevant as the present and the previous studies seem to have found no relation to gender.

 

”1 SD increment was 1.48 269 (1.13–1.95)” – this can be simplified to an mean increase in risk by 48%.

 

“During a follow-up of up to 17 years, 368 subjects died for 275 non-cardiovascular causes, 212 died for cardiovascular causes and 290 experienced a 276 stroke.” – this can for example also be shortened.

 

“In the first study by Gavish and Bursztyn [24], 24-hour PP and 24-hour 297 stPP predicted all-cause mortality only in patients with heart rate less than 70 beats/min” – this seems to be a repetition, but is more clearly stated here than above.

 

Instead of mentioning all the previous results first and then repeat them in relation to the present results, the authors could describe each kind of evaluation / outcome and at the same time compare it to the present results.

 

Please describe that the prognostic value of MAP was similar to that of elPP.

 

“our results cannot be applied to other ethnic groups” – this might be softened to ”our results may not be applicable to other ethnic groups”.

 

“However, we do not think that this aspect may have differentially influ-336 enced the different components of PP and their prognostic impact” – this is speculative and should be omitted.

 

The percentage of loss at follow-up and the retrospective nature of the study should be noted in the limitations section.

Author Response

Response to Reviewer 1

We thank the Reviewer for the comments and suggestions.  

1) Particularly the authors should continue to work on the discussion.

Response: We have now tried to improve the Discussion section.

Abstract

2) ”traditional 17 parameters including” – can be omitted.

 Response: We have now omitted these words.

3) “had a trend of association with risk” – “tended to associate with increased risk”

 Response: After further careful scrutiny, we reviewed the data with minimal changes, as explained later. We have now reported in the abstract: “…24-hour PP had a borderline association with risk, hazard ratio (HR) 1.16, 95% confidence interval (CI) 1.00-1.34,  24-hour elPP remained associated with cardiovascular events, HR 1.20, 95% CI 1.05-1.36 , and 24-hour stPP lost its significance.  

Introduction

4) ”not unanimous results” – try e.g. “conflicting results”.

Response: We have now changed the sentence.

Methods

5) Consider to the revise stating that you “included the 745 patients older than 65 years from the cohort of 2264”.

Response: We have now changed the sentence.

6) “One hundred and three patients were lost during follow-up from the global 72 population of 2264 subjects” – this information is of little relevance, better state the loss during follow-up of the included 745 patients.

Response: None of these patients was lost during follow-up. We apologize but we probably did not express this concept well in the previous versions (both for the global population and for those older than 65 years). The global population included 2264 patients with complete follow-up (1519 younger than 65 years and 745 older than 65 years. Our initial population of treated patients included 2264 patients (with complete follow-up) + 103 patients (lost during follow up), that is, 2367 patients. Our initial population of patients younger than 65 years included 1519 subjects with complete follow up + 67 subjects lost to follow up (1586) and our initial population of patients older than 65 years included 745 patients (with complete follow-up) + 36 patients lost during follow-up (781). We have now better reported in the Methods section this aspect: “Among the initial population of 781 subjects older than 65 years, 36 were lost during follow-up” (leaving to 745 patients).

7) ”Study 78 population came” – ”The study population”.

Response: We have now changed the sentence.

8) “Clinic systolic BP <140 mmHg and clinic diastolic BP <90 mmHg were defined as 88 normal.” – Still the authors’ definition of hypertension is unclear, was hypertension defined as SBP > 140 and DBP > 90 mmHg?

Response: We have now added: “As a consequence, clinic systolic BP >140 mmHg and/or clinic diastolic BP >90 mmHg were defined as hypertension.”

Statistical analysis

9) ”we included variables that were significantly associated 136 with outcome in univariate analysis, that is, age, smoking…” – Unclear, are these all the used variables used or were these the ones that were significant in univariate analysis and then used in multivariate analysis?  and                             10) Did the authors use more variables in the univariate analysis. Please clarify and state all used variables in the univariate analysis and how variables for the multivariate analysis were chosen.

Response: We apologize, we did not express this aspect well and created some misunderstandings. We have now better reported which patient characteristics were evaluated in the univariate analysis (characteristics in Table 1) and specific BP parameters, which of them were significant (age, diabetes mellitus, previous cardiovascular events, estimated glomerular filtration rate, LV hypertrophy, LA enlargement and asymptomatic LV systolic dysfunction) and were included in the multivariate analysis (Table 4). Moreover, we reported other variables included and the reasons for their inclusion (Moreover, we decided a priori to include the following covariates in the multivariate models, regardless of their association with risk: gender (because of the potential differences in PP between men and women), 24-hour heart rate (because of its reported potential impact on PP), MAP (to account for the steady component of BP beyond the pulsatile component represented by PP), and classes of antihypertensive drugs used (because of their potential different influence on afterload, preload, cardiac contractility, stroke volume, and possibly on arterial wall properties) (Statistical analysis section, Results section, Table 4, Figure 3). However, even if we included/forced in the multivariate analysis the variables that did not reach statistical significance in the univariate analysis (body mass index, smoking habit and low-density lipoprotein cholesterol), the results were the same.

11) ”asymptomatic LV systolic 140 dysfunction” – This definition is different from that stated in L115-116 which only defines LVEF as equal to and above vs. below 50% regardless of symptoms. Please use the same terminology throughout the manuscript.  

Response: The definition reported in the Methods section defines low LVEF. Asymptomatic LV systolic dysfunction defines patients with LVEF < 50% but without symptoms. These are two different concepts. Patients with a previous HF were included in the group of patients with a “Previous event”.

12) Did none of the patients with LVEF <50% have symptoms?

Response: Asymptomatic LV systolic dysfunction defines patients with LVEF < 50% but without symptoms (footnote of Table 1). Patients with a previous HF were included in the group of patients with a “Previous event”.

13) “because of the potential differences in PP between men and 141 women, and 14) “because of its reported potential impact on PP, and MAP 142 to account for the steady component of BP beyond the pulsatile component represented 143 by PP”, and 15) “used because of their potential different influence on afterload, preload, cardiac contractility and stroke volume, and possibly on 145 arterial wall properties“ – for clarity, this can be omitted, put in parentheses or stated in a separate sentence.

Response: We have now tried to better report this part (separate sentence reporting each point with the specific explanation put in parenthesis).

16) Wasn’t MAP used in the model?

Response: Yes, MAP was included in the models.

Results

17) Fig 2 “r=0.21” – consider to add the P-value.

Response: We have now reported r value and P value in the text. Though 24-hour elPP and 24-hour stPP were significantly correlated, the correlation was weak.  

18) “Among patients using > 3 three drug, 210 about 20% received 4-5 drugs.” This can be omitted and a line for four or more drugs can be added in Table 3.

 Response: We have now changed text and Table as suggested.

19) Fig 3 – please use a smaller range and put more numbers on the y-axis. The figure is not very informative at the present.

Response: We have now changed Figure 3 reporting only the results of multivariate analyses. Global results of univariate analysis are reported in new Table 4. We have enlarged Figure 3. As far as the y axis is concerned, the one shown is the lowest range we can obtain in this log scale.

20) It seems that the authors evaluated the univariate association of each of the many factors to the combined outcome but this data is not shown, only for PP, elPP and stPP. Consider to include a figure of all the univariate results. Have I misunderstood something? and 21) This might give some insights into the risk factors in the group and point to new research and perspectives. If elPP is a better or worse predictor than other factors it might be mentioned in the discussion.

Response: We apologize, we did not express this aspect well and created some misunderstandings. We have now better reported which patient characteristics were evaluated in the univariate analysis (characteristics in Table 1) and specific BP parameters, which of them were significant (age, diabetes mellitus, previous cardiovascular events, estimated glomerular filtration rate, LV hypertrophy, LA enlargement and asymptomatic LV systolic dysfunction) and were included in the multivariate analysis (Table 4). Moreover, we reported other variables included and the reasons for their inclusion (Moreover, we decided a priori to include the following covariates in the multivariate models, regardless of their association with risk: gender (because of the potential differences in PP between men and women), 24-hour heart rate (because of its reported potential impact on PP), MAP (to account for the steady component of BP beyond the pulsatile component represented by PP), and classes of antihypertensive drugs used (because of their potential different influence on afterload, preload, cardiac contractility, stroke volume, and possibly on arterial wall properties) (Statistical analysis section, Results section, Table 4, Figure 3).

Discussion

22) ”refine risk profile” – ”describe the risk profile” or “estimate the risk”.

Response: We have now changed the sentence as suggested.

23) “Moreover, for the first time,” – delete. Consider to merge this sentence with the next one.

Response: We have now changed the text as suggested.

24) ”there were 103 events” – here and elsewhere there is a lot of specific data from the literature. In this instance the authors can write e.g. that about xx% of the patients had a cardiovascular event.

Response: We have now changed the text as suggested.

25) ”Ohasama study population” – also irrelevant detail.

Response: We have now changed the text as suggested.

26) Please revise throughout and omit details that are irrelevant for the present study. The interested reader can review the literature. and 27) For example the exact ratio of men and women of previous studies are irrelevant as the present and the previous studies seem to have found no relation to gender.

Response: We have now tried to omit irrelevant details.

28) ”1 SD increment was 1.48 269 (1.13–1.95)” – this can be simplified to an mean increase in risk by 48%.

Response: We have now changed the text as suggested.

29) “During a follow-up of up to 17 years, 368 subjects died for 275 non-cardiovascular causes, 212 died for cardiovascular causes and 290 experienced a 276 stroke.” – this can for example also be shortened.

Response: We have now tried to shorten this part.

30) “In the first study by Gavish and Bursztyn [24], 24-hour PP and 24-hour 297 stPP predicted all-cause mortality only in patients with heart rate less than 70 beats/min” – this seems to be a repetition, but is more clearly stated here than above.

Response: We have now changed this paragraph.

31) Instead of mentioning all the previous results first and then repeat them in relation to the present results, the authors could describe each kind of evaluation / outcome and at the same time compare it to the present results.

Response: We have now tried to improve this section.

32) Please describe that the prognostic value of MAP was similar to that of elPP.

 Response: We have now reported this concept.

33) “our results cannot be applied to other ethnic groups” – this might be softened to ”our results may not be applicable to other ethnic groups”.

 Response: We have now changed the sentence.

34) “However, we do not think that this aspect may have differentially influ-336 enced the different components of PP and their prognostic impact” – this is speculative and should be omitted.

 Response: We have now omitted this part.

35) The percentage of loss at follow-up and the retrospective nature of the study should be noted in the limitations section.

Response: None of these patients was lost during follow-up. We apologize but we probably did not express this concept well in the previous versions (both for the global population and for those older than 65 years). The global population included 2264 patients with complete follow-up (1519 younger than 65 years and 745 older than 65 years. Our initial population of treated patients included 2264 patients (with complete follow-up) + 103 patients (lost during follow up), that is, 2367 patients. Our initial population of patients younger than 65 years included 1519 subjects with complete follow up + 67 subjects lost to follow up (1586) and our initial population of patients older than 65 years included 745 patients (with complete follow-up) + 36 patients lost during follow-up (781). We have now better reported in the Methods section this aspect: “Among the initial population of 781 subjects older than 65 years, 36 were lost during follow-up” (leaving to 745 patients). We have now reported in the limitation section that this is a retrospective study.

 

Reviewer 2 Report

Your arguments given in response to my comment why not performing sensitivity analysis to low/high heart rate should be added as study limitation. After all, you cite papers that show such sensitivity.

Author Response

Response to Reviewer 2

We thank the Reviewer for the comments and suggestions.

1) Your arguments given in response to my comment why not performing sensitivity analysis to low/high heart rate should be added as study limitation. After all, you cite papers that show such sensitivity.

Response: We have now reported this aspect in the limitation section: “Fourth, we did not perform an analysis by dividing the patients into those with 24-hour heart rate below or above the median. Dividing the population into two groups would have reduced the number of events for each group, weakened the event-to-variable ratio in each model, and weakened statistical power. Thus, we analyzed the population as a whole and adjusted for 24-hour heart rate in multivariate Cox models.”

In the Discussion section, we have also briefly reported that in Ref. 26 when 24-hour heart rate was taken into account in a sensitivity analysis (Supplementary file), 24-hour PP and 24-hour elPP were associated with an increased risk in both patients with lower and higher heart rate among those aged more than 60 years.