Troponin Testing for Assessing Sepsis-Induced Myocardial Dysfunction in Patients with Septic Shock

Round  1

Reviewer 1 Report

major suggestions:

The manuscript is concise but I think that the work will benefit from a better discussion of the results (TnI elevation) and comparison with previous works.

I would like also to suggest some minor improvements to the text.

Page 4, Line 131.  Do you mean "17.8% vs 10.1%".

Table1 (column 1). Please indicate the units of measurements where it is applicable.

page 5, line 155. Show the units for concentration:  "0.160 vs 0.085 ng/mL"

Page 7, table 3 (column 1). Please indicate the units of measurements.

Page 7, table 3 (table legends). Show the that the data are median (interquartile).

The first raw of this Table 3 is not column headers. Could you format the column so it would be more easily see the five different table sections.

Author Response

Reviewer1:

major suggestions:

The manuscript is concise but I think that the work will benefit from a better discussion of the results (TnI elevation) and comparison with previous works.

Response>

Thank you for your suggestion. We accept your opinion and develop discussion paragraph as follow.

“Previous studies have evaluated the usefulness of different cardiac biomarkers in recognizing the early signs of myocardial injury without acute coronary syndrome [17,18]. Among these markers, a higher serum troponin level is a prognostic indicator of high mortality in sepsis cases. Its impact has had inconsistent reports however because of differences in troponin types (troponin I or troponin T), disease severity, cutoff values, and the time to measurement [10,19-21]. Wilhelm et al. reported that non-survivor of septic patients had elevated high sensitivity troponin T on admission than that of survivor [19]. Similarly, a prior study from the United States showed the troponin T level on admission was associated with a higher in-hospital (OR 1.6; p = 0.003), and 1-year mortality (OR 1.3; p=0.04) [21]. Another study reported that an elevated troponin-T not initially but on day 7 was associated with increased mortality (OR 1.38, p = 0.01) [20]. However, these studies tried to reveal correlations between troponin level and mortality. Meanwhile, Mehta et al. reported that troponin I was independent predictor of death and correlated with the lower LVEF (p < 0.001) [22]. But, they included relatively small number of patients (n = 16) and measured only LVEF via TTE.

Whether myocardial dysfunction induces troponin release remains a question of debate. Several studies correlated troponin levels and SIMD [22,23]. Røsjø and colleagues, evaluating a hs-TnT in a large cohort of patients with sepsis, found that hs-TnT levels reflected cardiomyocyte injury but did not reliably identify SIMD[24]. This is consistent with the present result that peak hs-TnI was associated with SIMD, however hs-TnI level at admission was not. Furthermore, both admission and maximum hs-TnI did not demonstrate a stepwise increase of OR for SIMD. The underlying mechanism of troponin elevation in sepsis has not yet been elucidated. Contrary to coronary artery occlusion by atherosclerosis, it has been reported that coronary blood flows are similar or even increased in cases of sepsis [25]. A rapid incline and decline in the troponin level could be interpreted to mean that SIMD is a reversible myocardial ischemia [26]. Recently, some studies have found that nitric oxide, inflammatory cells, cytokines, complement, and mitochondrial dysfunction play complex roles in the pathways to cardiac dysfunction at both the cellular and molecular level [27]. One of these effects is cell membrane permeability and functional alterations to ion channels, which contribute to the release of troponin from the intracellular space [28]. A delayed washout or increasing level of troponin may correlate with ongoing myocardial injury. Because troponin can be quite readily and cost-effectively assayed in the clinic, repeat measurements may be a valid approach to identifying cardiac dysfunction.”

I would like also to suggest some minor improvements to the text.

Page 4, Line 131.  Do you mean "17.8% vs 10.1%".

Response>Thank you for your correction. We change the sentence to “17.8% vs 10.1%”.

Table1(column 1). Please indicate the units of measurements where it is applicable.

Response>Thank you for your kind suggestion. We add the units in table 1. 

page5, line 155. Show the units for concentration: "0.160 vs 0.085 ng/mL"

Response>Thank you for your correction. We add the units for concentration and change the sentence as follow.

“The initial hs-TnI level was significantly greater in the LV diastolic dysfunction cases (0.160 vs. 0.085 ng/mL; p = 0.019)”

 Page7, table 3 (column 1). Please indicate the units of measurements.

Response>Thank you, we add the units in table 3 for high sensitivity-troponin I.

Table 3. Cardiac biomarkers for predicting SIMD on TTE

Data are presented as a number (%) or median with interquartile ranges. Number of any dysfunctions are 258, LV systolic dysfunctions are 163, LV diastolic dysfunctions are 104, RV dysfunctions are 186, and WMA are 186. Abbreviations: SIMD = sepsis-induced myocardial dysfunction; TTE = transthoracic echocardiography; hs-TnI = high-sensitivity troponin-I; LV= left ventricle; RV = right ventricle; WMA = wall motion abnormalities.

Page7, table 3 (table legends). Show the that the data are median (interquartile).

Response>Thank you for your kindness. We add a sentence.“Data are presented as median with interquartile ranges.”

Thefirst raw of this Table 3 is not column headers. Could you format the column so it would be more easily see the five different table sections.

Response>Thank you for your suggestion. We change the table format (we add table 3 because of other reviewer’s recommend and table 3 became table 4.)

Table 4. Cardiac biomarkers for predicting SIMD on TTE

Data are presented as a number (%) or median with interquartile ranges. Number of any dysfunctions are 258, LV systolic dysfunctions are 163, LV diastolic dysfunctions are 104, RV dysfunctions are 186, and WMA are 186. Abbreviations: SIMD = sepsis-induced myocardial dysfunction; TTE = transthoracic echocardiography; hs-TnI = high-sensitivity troponin-I; LV= left ventricle; RV = right ventricle; WMA = wall motion abnormalities.

Reviewer 2 Report

Introduction is too short, need more informations along with current references.

Echocardiography images  are necessary to add to meet the journal quality of publications.

There are very minimal parameters has been added in the echocardiography calculations. Needs more parameters  from echo study to confirm myocardial dysfunction in table2.

Discussion needs more references.

Author Response

Reviewer2

Introduction is too short, need more informations along with current references.

Response>Thank you. We totally accept your opinion and add sentences to explain backgrounds of our study on introduction.

“Troponin is sensitive and specific biomarkers of myocardial damage and relates with prognosis in non-acute coronary syndrome such as pulmonary embolism, trauma, stroke, and subarachnoid hemorrhage[a1] . Recent studies have demonstrated that cardiac troponins release in sepsis are associated with poor outcomes including higher mortality and longer length of stay in intensive care unit.”

Echocardiographyimages are necessary to add to meet the journal quality of publications.

Response>Thank you for your great opinion. We add an example of echocardiography images (Figure 1).

“Figure 1 showed an example of echocardiography.”

Figure 1. Example of echocardiography. A: left ventricle ejection fraction measurement via Teichholz method. B: measurement of E/e’. C: description of wall motion abnormalities. Abbreviations: PLAX = parasternal long axis; PASX = parasternal short axis; A4C = apical four chamber; A2C = apical two chamber.

Thereare very minimal parameters has been added in the echocardiography calculations. Needs more parameters from echo study to confirm myocardial dysfunction in table 2.

Response>

We agree with your opinion and add echocardiographic characteristics on Table 3. LVIDs, LVIDd, LA, LVPWs, ESV, EDV, IVSs, and E/e’ were significantly higher in SIMD (+) group. Meanwhile, LVEF and deceleration time were lower in SIMD (+) group. We describe this result to result section.

“Table 3 showed echocardiographic characteristics of cohorts. LV internal diameter, left atrium, LV posterior wall, end-systolic volume, end-diastolic volume, interventricular septum, and E/e’ were significantly higher in SIMD (+) group. Meanwhile, LVEF and deceleration time were lower in SIMD (+).”

Table 3. Echocardiographic Characteristics of Cohorts

Data are presented as median with interquartile ranges. Abbreviations: LVIDs = left ventricular internal diameter (systolic); LVIDd = left ventricular internal diameter (diastolic); LA = left atrium; LVPWs = left ventricle posterior wall (systolic); LVPWd = left ventricle posterior wall (diastolic); IVSs = interventricular septum (systolic); IVSd = interventricular septum (diastolic); LVEF = left ventricle ejection fraction; E = peak early diastolic transmitral flow velocity; A = peak late diastolic transmitral flow velocity; e’ = peak early diastolic mtral annulus velocity.

 Discussionneeds more references.

Response>Thank you for your valuable suggestion. We add more sentences and references as follow.

“Previous studies have evaluated the usefulness of different cardiac biomarkers in recognizing the early signs of myocardial injury without acute coronary syndrome [17,18].”

“Wilhelm et al. reported that non-survivor of septic patients had elevated high sensitivity troponin T on admission than that of survivor [19].”

“However, these studies tried to reveal correlations between troponin level and mortality. Meanwhile, Mehta et al. reported that troponin I was independent predictor of death and correlated with the lower LVEF (p < 0.001) [22]. But, they included relatively small number of patients (n = 16) and measured only LVEF via TTE.”

 [a1]레퍼런스추가
SHOCK, Vol. 39, No. 1, pp. 50Y54, 2013
VALUE OF CARDIAC TROPONIN I FOR PREDICTING IN-HOSPITAL

OCCURRENCE OF HYPOTENSION IN STABLE PATIENTS WITH ACUTE

PULMONARY EMBOLISM

Reviewer 3 Report

In this review manuscript, Kim et al. aimed to evaluate the role of admission and peak troponin-I testing for the identification of sepsis-induced myocardial dysfunction (SIMD) by using retrospective cohort study.

Overall, this is a focused, well designed and organized clinical data analysis study. The authors, used multiple statistical analysis model to correlate the initial and Peak hs-TnI level with various types of cardiac dysfunction based on TTE data. While the conclusion will provide additional support to the prognostic indicator role of troponin I, this manuscript need to be revised based on the following concerns:

1) As the author mentioned, using troponin T and I as cardiac injury biomarkers is still under debating. Does the current cohort have troponin T level record as well? If so, the authors should do the same analysis to have a much stronger conclusion by comparison these two markers.

2) Although, the authors showed the frequency of various types of Sepsis-induced myocardial dysfunction on TTE in Table 2, there is lack of the original data of TTE. It is intriguing that if the TnI level correlated with the severity of myocardial dysfunction levels, e.g. EF, LVPD, LVSD, etc.

3) It was good to list the limitations of this study. According to the second point, the times between the measurement of TnI and TTE exams were not consistent. The reviewer is curious that if the TnI level has any correlation with this time window? What is the determinant for this time window? And, what does the author mean “some abnormal echocardiographic parameters might have reversed to normal levels in some cases”?

4) It was better to show the hs-TnI curve as the author used AUC analysis as mentioned in 3.3 line 153.

5) In conclusion part, there was no conclusion could be made as “Cardiac dysfunction contributes to troponin elevation” based on this study but rather the later one “he peak troponin level is associated 229 with SIMD on TTE, however hs-TnI level at admission did not reliably identify SIMD”. Because no causal conclusion could be made but correlation conclusion based on these analyses.

Minor issues:

1) line 128, SCMP (+) should be SIMD (+).

2) Table2, there must be some overlap between subtypes of SIMD, data showing with a pie chart will be better.

3) line 150-153. The order of the data description in Table 3 should be consistent with the order listed in the Table. Thus, it’s better to describe LV systolic data before LV diastolic data.

4) Format issue in Table 3, lack of necessary lines between types of SIMD. In addition, since the column two from left side “Total”, the data in this column is the repeated 4 times. It’s better to merge into one.

5) Show statistic indicator in all the tables.

6) Lack of reference in line 179-180.

7) Lack of journal name in reference 24 and 26.

 Author Response

Reviewer 3

Inthis review manuscript, Kim et al. aimed to evaluate the role of admission and peak troponin-I testing for the identification of sepsis-induced myocardial dysfunction (SIMD) by using retrospective cohort study. Overall, this is a focused, well designed and organized clinical data analysis study. The authors, used multiple statistical analysis model to correlate the initial and Peak hs-TnI level with various types of cardiac dysfunction based on TTE data. While the conclusion will provide additional support to the prognostic indicator role of troponin I, this manuscript need to be revised based on the following concerns:

1) As the author mentioned, using troponin T and I as cardiac injury biomarkers is still under debating. Does the current cohort have troponin T level record as well? If so, the authors should do the same analysis to have a much stronger conclusion by comparison these two markers.

Response>We totally accept your opinion. Regretfully, our hospital cannot check cardiac troponin T in laboratory and further research about comparing troponin T and I would be good study in the future. It is true that troponin T and I is still under debating to diagnose sepsis-induced cardiac dysfunction. However, some studies reported that there was strong correlation between troponin T and I and both cardiac markers showed significantly higher in cardiac dysfunction groups [1-2]. We mentioned this in limitation.

“Third, our hospital cannot check cardiac troponin-T and there might be some difference among specific troponin types. However, previous tudy reported that there was strong correlation between troponin-T and I.”

 2) Although, the authors showed the frequency of various types of Sepsis-induced myocardial dysfunction on TTE in Table 2, there is lack of the original data of TTE. It is intriguing that if the TnI level correlated with the severity of myocardial dysfunction levels, e.g. EF, LVPD, LVSD, etc.

Response>We agreed with the reviewer’s concern about that point. Spearman correlation showed that initial hs-TnI had no correlation with all of echocardiographically driven data. However, peak hs-TnI had significantly correlate with some parameters, especially negative correlation with LVEF (LVIDs (0.122, p = 0.19), Aorta (-0.188, p < 0.001), , peak A velocity (-0.154, p = 0.007), LVEF (-0.258, p < 0.001)). This result is identical with table 4.

 3) It was good to list the limitations of this study. According to the second point, the times between the measurement of TnI and TTE exams were not consistent. The reviewer is curious that if the TnI level has any correlation with this time window? What is the determinant for this time window? And, what does the author mean “some abnormal echocardiographic parameters might have reversed to normal levels in some cases”?

Response>Thank you for your great opinion. Regretfully, there were no guideline or specific time window of troponin I measurement correlates with TTE variable. One recent study controlled time difference between troponin T measurement and echocardiography through blood samples obtained on the days of the echocardiography examinations [3]. It is ideal to check blood samples and echocardiography simultaneously. However, because of our retrospective design, we could not analyze correlations between troponin I level and echocardiographic driven data according to time. 

Previous studies reported that patients’cardiac function could be recovered fully to their premorbid state after 7-10 days [4-7]. In our study, among 35 (8.8%) patients those who were checked echocardiography after 7 days, 21 (60%) showed no cardiac dysfunctions. Therefore, we though that some of normal echocardiographic results could be recovered state and could be influence results. We add the sentences to clarify the meaning as follow.

“Previous studies reported that patients’ cardiac function could be recovered fully to their premorbid state after 7-10 days. In our study, 35 (8.8%) patients were checked echocardiography after 7 days. Among these, 21 (60%) showed no cardiac dysfunction and some of normal findings could be recovered state and could be influence results.”

 4) It was better to show the hs-TnI curve as the author used AUC analysis as mentioned in 3.3 line 153.

Response>Thank you for your great suggestion. We add the ROC curve in Figure 2.

5) In conclusion part, there was no conclusion could be made as “Cardiac dysfunction contributes to troponin elevation” based on this study but rather the later one “he peak troponin level is associated 229 with SIMD on TTE, however hs-TnI level at admission did not reliably identify SIMD”. Because no causal conclusion could be made but correlation conclusion based on these analyses.

Response>We totally agree with your opinion and change the sentence in conclusion as follow. 

“About two-thirds of patients with an elevated hs-TnI level have various cardiac dysfunctions on TTE. The peak troponin level is associated with SIMD on TTE, however hs-TnI level at admission did not reliably identify SIMD.”

 Minorissues:

1) line 128, SCMP (+) should be SIMD (+).

Response>Thank you for your kindness. We correct typo-error.

“The median age of the total cohort was 67.0 (range, 18.0 - 98.0) with slightly more men in both the SIMD (+) and (-) groups.”

2) Table2, there must be some overlap between subtypes of SIMD, data showing with a pie chart will be better.

Response>Thank you for your great suggestions. We add pie chart with table 2.

3) line 150-153. The order of the data description in Table 3 should be consistent with the order listed in the Table. Thus, it’s better to describe LV systolic data before LV diastolic data.

Response>We totally agree with your opinion and change the sentence order.

“Moreover, the peak hs-TnI level was statistically greater in the patients with LV systolic dysfunction (1.609 vs. 0.540 ng/mL; p < 0.001), RV dysfunction (2.478 vs. 0.546 ng/mL; p < 0.001), and WMA (1.950 vs. 0.455 ng/mL; p < 0.001). The initial hs-TnI level was significantly greater in the LV diastolic dysfunction cases (0.160 vs. 0.085 ng/mL; p = 0.019).”

4) Format issue in Table 3, lack of necessary lines between types of SIMD. In addition, since the column two from left side “Total”, the data in this column is the repeated 4 times. It’s better to merge into one.

Response>Thank you for your suggestion. We change the table format (we add table 3 because of other reviewer’s recommend and table 3 became table 4.)

Table 4. Cardiac biomarkers for predicting SIMD on TTE

Data are presented as a number (%) or median with interquartile ranges. Number of any dysfunctions are 258, LV systolic dysfunctions are 163, LV diastolic dysfunctions are 104, RV dysfunctions are 186, and WMA are 186. Abbreviations: SIMD = sepsis-induced myocardial dysfunction; TTE = transthoracic echocardiography; hs-TnI = high-sensitivity troponin-I; LV= left ventricle; RV = right ventricle; WMA = wall motion abnormalities.

5) Show statistic indicator in all the tables.

Response>Thank you for your correction. We add statistic indicator in all the tables (Table 2, 4).

“Data are presented as a number (%).”

6) Lack of reference in line 179-180.

Response>Thank you for exact correction. We add references in discussion section.

“Previous studies have evaluated the usefulness of different cardiac biomarkers in recognizing the early signs of myocardial injury without acute coronary syndrome.”

7) Lack of journal name in reference 24 and 26.

Response>Thank you for your kindness. We correct the reference 24 and 26 with journal name (we added more reference because of other reviewers’ opinion and now reference numbers are 32 and 34.

32. Ehrman, R. R.; Sullivan, A. N.; Favot, M. J.; Sherwin, R. L.; Reynolds, C. A.; Abidov, A.; Levy, P. D. Pathophysiology, echocardiographic evaluation, biomarker findings, and prognostic implications of septic cardiomyopathy: a review of the literature. Critical Care2018, 1–14.

34. Sanfilippo, F.; Corredor, C.; Fletcher, N.; Tritapepe, L.; Lorini, F. L.; Arcadipane, A.; Vieillard-Baron, A.; Cecconi, M. Left ventricular systolic function evaluated by strain echocardiography and relationship with mortality in patients with severe sepsis or septic shock: a systematic review and meta-analysis. Critical Care2018, 1–12.

Reference>

1. Ammann P, Maggiorini M, Bertel O, et al. Troponin as a risk factor for mortality in critically ill patients without acute coronary syndromes. J Am Coll Cardiol. 2003;41(11):2004-2009

2. Ver Elst K, Spapen H, Nguyen D, Garber C, Huyghens L, Gorus F. Cardiac tropenins I and T are biological markers of left ventricular dysfunction in septic shock. Clin Chem. 2001;46(5):650-657

3. Landesberg G, Jaffe AS, Gilon D, Levin PD, Goodman S, Abu-Baih A, beeri R, Weissman C, Sprung CL, Landesberg A. Troponin elevation in severe sepsis and septic shock: The role of left ventricular diastolic dysfunction and right ventricular dilatation. CCM. 2014:42(4):790-800

4. Parker MM, Shelhamer JH, Bacharach SL, Green MV, Natanson C, Frederick TM, Damske BA, Parrillo JE. Profound but reversible myocardial depression in patients with septic shock. Ann Intern Med. 1984;100(4):483-90

5. Kakihana Y, Ito T, Nakahara M, Yamaguchi K, Yasuda T. Sepsis-induced myocardial dysfunction: pathophysiology and management. J Intensive Care. 2016;4:22

6. Sato R, Nasu M. A review of sepsis-induced cardiomyopathy. J Intensive Care. 2015;3:48

7. Jardin F, Fourme T, Page B, Loubieres Y, Vieillard-Baron A, Beauchet A, Bourdarias JP. Persistent preload defect in severe sepsis despite fluid loading: A longitudinal echocardiographic study in patients with septic shock. Chest. 1999;116(5):1354-9

Round  2

Reviewer 1 Report

I am satisfied with the answers and corrections.

Thank you.

Author Response

Response>Thank you for your precious comment.

Reviewer 2 Report

Still the introduction is not sufficient to support the whole findings. 

The author added example of the echo images, that was no suggested in the last revision. Please add real images of your own experiments (own samples) to support the data to meet the journal quality. 

Author Response

Still the introduction is not sufficient to support the whole findings. 

 Response>Thank you for your valuable comments. We added sentences in the introduction to clarify the purpose of our study. 

“Septic shock is still a leading cause of death worldwide as it can induce multi-organ failure [1,2]. Cardiac dysfunction, referred to as sepsis-induced myocardial dysfunction (SIMD), is a loosely defined syndrome and presents in various ways, such as myocardial injury with cardiac biomarker elevation, myocardial dysfunction on echocardiography, and hemodynamic instability [3]. SIMD is a common complication (40-60%) in severe sepsis and septic shock and its presence significantly worsens the outcome [4,5]. Early recognition, aggressive supportive therapy, and appropriate catecholamine regimen are mandatory in order to improve the survival. Although echocardiography has been as a golden tool for SIMD, it is not practical for performing echocardiography for every sepsis patients during early resuscitation due to its cost and limited round-the-clock availability.

Troponin-I is the subunit of the troponin complex which is exclusive of cardiac origin. It is a sensitive and specific biomarker of myocardial damage and relates to prognosis in non-acute coronary syndromes such as pulmonary embolism, trauma, stroke, and subarachnoid hemorrhage [6]. Recent studies have demonstrated that cardiac troponins release in sepsis is associated with poor outcomes including higher mortality and longer length of stay in intensive care unit [7]. Theories for troponin elevations in septic shock include inflammation, increased myocardial wall stress by volume overload, toxicity by medications, and kidney dysfunction [8]. However, the results of the studies investigating the impact of troponin levels for SIMD in sepsis are not concordant. Although some research showed a significant association between troponin and SIMD on echocardiography such as Left ventricular (LV) systolic, diastolic dysfunction or Right ventricular (RV) dysfunction,[9,10]others did not. 

The echocardiographic findings of SIMD are still poorly defined, there is limited data regarding the relationship between troponin elevation and different type of SIMD [11-13]. We hypothesized that measurement of troponin on admission may provide information about the state of heart and may be complementary to echocardiographically derived data. Moreover, we postulated that degree of increment  in troponin level on serial measurement would be of incremental value in risk stratification. To test this hypothesis, we used a prospective registry of septic shock to evaluate the role of admission and peak troponin-I testing for the identification of SIMD by transthoracic echocardiography (TTE) in septic shock patients.”

The author added example of the echo images, that was no suggested in the last revision. Please add real images of your own experiments (own samples) to support the data to meet the journal quality. 

Response>Thank you for your kindness. We accept your opinion and change figure 1 to add real images of our own samples.

 Reviewer 3 Report

Based on the current version of this manuscript, the authors have answered most of my questions, however, there are still some mistakes or mis-interpretation.

1. The figure 1 was in poor resolution, especially panel C. One can not read the labeling when printed.

2. Although the authors have added the ROC curve in this version, it was still not clear how the cutoff value 0.668 ng/ml was calculated as shown on line 278, they author should add more detailed instruction in the method. And, the Figure No. was wrong, it should be Figure 3.

3. In section 3.3, line 268, “The peak hs-TnI level was statistically different between the SIMD (+) and SIMD (-) groups,” here, the authors should indicate more precisely that “except in patients with LV diastolic dysfunction”.

4. Although the authors have added missing lines in Table 4, it was still not satisfactory. The category title of RV was missing this time. The Peak hs-TnI for WMA was 0.539 (0.178-0.293), here 0.293 should be 2.293 according to the value in Total. The values in column “Total” were still in repeated mode, please merge into one.

5. In Discussion, line 367, please indicate clearly that the “day 7” means “day 7 after admission” or other interpretation?

6. In Discussion, line 376-378, the authors claimed “this is consistent with the present result” on line 377, however, the previous work as cited was about TnT, and it did not reliably identify SIMD. Thus, it was not appropriate to use “consistent” here.

7. Please use *p<0.05, **p<0.01, ***p<0.001 to indicate all the statistically significant group.

Author Response

Basedon the current version of this manuscript, the authors have answered most of my questions, however, there are still some mistakes or mis-interpretation.

1. The figure 1 was in poor resolution, especially panel C. One can not read the labeling when printed.

Response>We changed figure 1 as other reviewer’s comment and enhance resolution.

2. Although the authors have added the ROC curve in this version, it was still not clear how the cutoff value 0.668 ng/ml was calculated as shown on line 278, they author should add more detailed instruction in the method. And, the Figure No. was wrong, it should be Figure 3.

 Response>We agree with your concerns. We found the optimal cutoff value of hs-TnI for SIMD with Youden index and add a sentence in method (2.6. Statistics).

“The optimal cutoff value of hs-TnI for SIMD was determined with Youden index (sensitivity + specificity-1) from ROC analysis.”

3. In section 3.3, line 268, “The peak hs-TnI level was statistically different between the SIMD (+) and SIMD (-) groups,” here, the authors should indicate more precisely that “except in patients with LV diastolic dysfunction”.

Response>Thank you for the valuable comment. We totally agree with your opinion and added the sentence.

“The peak hs-TnI level was statistically different between the SIMD (+) and SIMD (-) groups except in patients with LV diastolic dysfunction.

4. Although the authors have added missing lines in Table 4, it was still not satisfactory. The category title of RV was missing this time. The Peak hs-TnI for WMA was 0.539 (0.178-0.293), here 0.293 should be 2.293 according to the value in Total. The values in column “Total” were still in repeated mode, please merge into one.

Response>We are really sorry to make a mistake. As your comments, we added “RV” category and move the total column to table 1.

Table 1.Baseline Characteristics of Patients with SIMD on TTE.

Data are presented as a number (%) or median with interquartile ranges. *p<0.05; **p<0.01; ***p<0.001. Abbreviations: SIMD = sepsis-induced myocardial dysfunction; TTE = transthoracic echocardiography; HTN = hypertension; DM = diabetes mellitus; CAD = coronary artery disease; CKD = chronic kidney disease; EKG = electrocardiogram; STTC = ST segment and T wave change; BP = blood pressure; SOFA = sequential organ failure assessment; WBC = white blood cells; Hb = hemoglobin; BUN = blood urea nitrogen; CRP = c-reactive protein; hs-TnI = high sensitivity troponin-I.

Table 4.Cardiac biomarkers for predicting the various type of SIMD on TTE.

Data are presented as a number (%) or median with interquartile ranges. Number of any dysfunctions are 258, LV systolic dysfunctions are 163, LV diastolic dysfunctions are 104, RV dysfunctions are 186, and WMA are 186. *p<0.05; **p<0.01; ***p<0.001. Abbreviations: SIMD = sepsis-induced myocardial dysfunction; TTE = transthoracic echocardiography; hs-TnI = high-sensitivity troponin-I; LV= left ventricle; RV = right ventricle; WMA = wall motion abnormalities.

5. In Discussion, line 367, please indicate clearly that the “day 7” means “day 7 after admission” or other interpretation?

Response>Thank you for your kind suggestion. We added words to clarify.

“day 7 after admission”

6. In Discussion, line 376-378, the authors claimed “this is consistent with the present result” on line 377, however, the previous work as cited was about TnT, and it did not reliably identify SIMD. Thus, it was not appropriate to use “consistent” here.

Response>Thank you. We changed the word “consistent” to “similar” for reducing misunderstand and added sentence.. 

“Although there were different types of troponins, the pattern of the result is similar with the present finding that peak hs-TnI was associated with SIMD, however hs-TnI level at admission was not.

7. Please use *p<0.05, **p<0.01, ***p<0.001 to indicate all the statistically significant group.

Response>Thank you for your kind suggestion. We added indications of all the statistically significant group in all table.

“*p<0.005, **p<0.001, ***p<0.001”< span="">