Author s response to reviews Title: Impact of perioperative administration of 6% hydroxyethyl starch 130/0.4 on serum cystatin C-derived renal function after radical prostatectomy: A single-centre retrospective study Authors: Stefan Südfeld (suedfeld.research@gmx.de) Sami R Leyh-Bannurah (s.bannurah@googlemail.com) Lars Budäus (lars.budaeus@gmail.com) Markus Graefen (graefen@uke.de) Philip C Reese (reese.research@gmx.de) Franziska von Breunig (f.breunig@uke.de) Daniel A Reuter (dreuter@uke.de) Bernd Saugel (b.saugel@uke.de) Version: 1 Date: 11 Jul 2016 Author s response to reviews: Reviewer #1: Major comments: 1. The methodology is good although a retrospective study cannot be used to find a cause-effect relationship. The respected authors pointed to some limitations in their discussion which are correct. Answer: We appreciate the reiteration of the limitations of our study design that we considered when discussing the results. We fully agree with the reviewer that the retrospective nature of our study does not allow drawing definite conclusions about a causative relation between 6% hydroxyethyl starch on serum cystatin C-derived renal function but only describes an association.
2. The English language is acceptable with some minor editorial correction to be made before publication. Answer: We re-edited our manuscript and made corrections with the help of a native Englishspeaker. We hope that this resulted in improvement of linguistic accuracy and readability for the esteemed reader. Reviewer #2: Major comments: 1) Why you chose patients with radical prostatectomy for your study? Foley's catheter irrigation? gender bias! Answer: We thank Reviewer 2 for this very thoughtful comment. In the methods section as well as the limitations paragraph, we now explain that clinical circumstances offered us the opportunity to retrospectively study patients in whom perioperative cystc measurements were available and HES is routinely administered. This gave us the unique opportunity to investigate the effects of HES in this particular study population in this particular procedural context. Further studies are needed to confirm this notion in other settings and populations. We added the limitations that come with this study design and patient selection in the limitations paragraph in the revised manuscript. 2) No control group! Answer: In our study, a control group would have been patients undergoing prostatectomy (and available cystc) without receiving HES. Only 8 patients of the 187 cases that fulfilled procedural inclusion criteria however received only crystalloid solutions and no HES. We added this information on the number of patients that primarily met inclusion criteria in terms of cystc measurements but not administration of HES to the results section. This way, we hope to bring across more clearly to the reader that it is common practice at our university centre to administrate HES perioperatively in this particular clinical setting. Moreover, since there was no increase in cystc in patients receiving HES it is unlikely that including "crystalloid-only control group" data would have let us to draw a different conclusion with respect to the hypothesis of this study, i.e. that HES does not compromise renal function in this particular setting. Including a control group would of course be an intriguing idea when planning a future prospective study on this issue.
3) What are the limitations of your study? Answer: We agree with Reviewer 2 that it is crucial to bear in mind the limitations of our study when interpreting our results. We now discuss in more detail the limitations of our study. 4) -Alot of linguistic and grammatical errors! Answer: Thank you very much. We thoroughly re-edited the text and corrected linguistic and grammatical errors with the help of a native English-speaker. We hope that this resulted in improvement of linguistic accuracy and readability for the esteemed reader. 5) What do you mean by moderate dose (1000-1000)?! What is the classification of doses you applied of HES? Answer: We thank Reviewer 2 very much for this remark. We agree with him completely, that there is no respective classification system of HES dosage in terms of effects on renal function in the literature we could refer to. As a result, we now explicitly state the median dosage instead. This way, we hope to make clearer that our conclusions are only to be applied to patients that receive this amount of HES (that we formerly felt to be adequately described as moderate as the median dose administered in our patients is actually well below the recommended maximum daily dose of 30 ml/kg HES 6%. (https://www.gov.uk/drug-safety-update/hydroxyethyl-starchintravenous-infusions [retrieved on 02/07/2016]). 6) Any conflict of interests? Answer: There are no conflicts of interest for any of the authors. That there are no competing interests is stated in the manuscript according to the instructions for authors. 7) Was you overloading your patients especially the patient with congestive heart failure? Answer: We cannot adequately address this very speculative comment. As mentioned in the limitations section data on fluid balance are not available due to practical procedural reasons; in addition, there are no data on advanced hemodynamic measurements; therefore, we cannot make a scientifically sound statement on "fluid overloading". We hope, that this limitation is adequately addressed and explained to the reader in the limitations paragraph.
8) Why you followed up your patients for only 5 days postoperatively not for 30 days? Answer: Thank you very much for this very important question. We deliberately chose to follow up our patients for 5 postoperative days because the hospital length of stay for the group of patients included in our analysis is about 5-6 days in our institution. In addition, choosing postoperative day 30 would bear the risk to evaluate effects on renal function that occur during a prolonged hospital stay (secondary infection with acute renal failure) not related to intraoperative HES administration. In addition, cystc-derived egfr-changes detect postoperative AKI as early as one day postoperatively with peak values on day three as stated in the manuscript. Therefore, a possible increase in cystc due to surgery-related AKI should have likely been detected by sequential cystc-measurements during the observed time span in this study. Reviewer #3: Major comments 1. The quality of the written English is questionable and makes it difficult to understand at times what the Authors actually mean. Answer: Thank you very much for this statement. We revised the text thoroughly with the help of a native English-speaker. We hope that this resulted in improvement of the text quality in terms of clarity and comprehensibility. 2. The retrospective nature of the study is major limitation with respect to the quality of the data retrieved and the lack of significant amount of clinical information that would allow the reader to agree (or not) with the Authors' conclusions. I am thinking for example of the Inker study in 2012 NEJM where he suggests we should actually combine scr and cystc for a more accurate diagnosis of acute renal Answer: Thank you very much for this very important aspect and for mentioning the Inker et al. paper. In our study, however, neither scr nor cystc showed a tendency towards an increase in serum concentrations. Thus, the combination of these parameters, even though an intriguing option, equally demonstrates that GFR does not increase. However, your remark will be very helpful in the conduction and interpretation of future studies.
3. Page 3 Line 2: AKI? What about using acute renal insufficiency instead of injury? Answer: Thank you very much for this question. We used the well-defined term acute renal injury according to RIFLE criteria for the sake of comparability to other studies and applicability to clinical practice. The term acute renal insufficiency is, even though commonly used in general clinical practice, indistinctly defined in terms of cut-off values for renal retention parameters. For these reasons, we chose the terminology and definition of acute renal injury according to ADQI guidelines. 4. Line 9: HES "thus far widely used"? may be in Germany, definitively not in other Answer: Thank you for sharing your impression on the use of HES with us. We removed thus far widely used from our manuscript. We meant to express that HES had been in common use before it was shown to increase the risk of acute kidney injury in critically ill patients, at least in our university hospital. However, this may not be the case for other centres, especially not in other countries. We share Reviewer 3s opinion and think that we may not generalise our former practice of HES use and fully concur with this point of criticism. 5. Line 10: "renal replacement therapy"what do they mean by that? Hemodialysis? Answer: We used the general term renal replacement therapy to describe extracorporeal techniques to replace non-endocrine kidney function in patients with renal failure including intermittent and continuous hemodialysis and hemofiltration. This is in accordance with former publications on HES in critically ill and septic patients from the Brunkhorst et al. (2008) and Myburgh et al. (2012), each published in the New England Journal of Medicine. 6. Line 22: remove "Moreover". Answer: Done. 7. Line 39: Therefore, we evaluated the influence Answer: We corrected this phrase.
8. Line 44: I am questioning the decision to combine ORP and RARP. There are so many different technical differences that could have affected their measurements. At least, they should have presented the two set of data separately, particularly because one of the criteria they later mentioned was the level of intra-abdominal pressure during RARP. Answer: We appreciate Reviewer 3s expression of concern about possible influencing factor associated with the type of prostatic surgery, especially possible differences arising from capnoperitoneum. This study was designed to test the hypothesis that HES does not lead to an increase in renal retention parameters (ie. serum creatinine or cystc) in non-critically ill patients undergoing prostate surgery. We did not aim to test influences of different operational techniques on renal function. In addition, as there is no relevant increase in cystc or serum creatinine in any of the cases included in this study (neither ORP nor RARP), separation of ORP from RARP cases would likewise show no adverse effect on these parameters. 9. Page 4 Line 26: what is an opportunity sample? They selected the patients over a specific period of time? If so, over what period of time? And or they took all comers over an precise period of time? Answer: Opportunity Sampling or Convenience Sampling are types of nonprobability sampling approaches. To not confuse the reader we avoid this term in the revised manuscript. The clinical circumstances (i.e. the presence of data on cystc in patients that had routinely received HES) gave us the opportunity to investigate our research question in a retrospective study. The study period during which patients had cystc measurements served as a basis for retrospective patient recruitment in our study and is mentioned in the appropriate paragraph within the methods section. We stated in the text that patient data were extracted from patients treated between September 2012 and April 2013. 10. Line 27: without revision of surgery? Better: who did not require a re-intervention within 5 days from the original operation? Answer: We thank Reviewer 3 for his recommendation. We changed the sentence on inclusion criteria accordingly. 11. Line 34: Last sentence of second paragraph. This shows the limitations of retrospective data. I believe that it is not proper science to pick randomly available data and then extrapolate the results to an entire population. They considered patients who only had 1 measurement and
combined with others who had more than 1 measurement in the postoperative period. They should at least tell us how many had 1-2-3 measurements in the follow-up period. Also there are no well-defined inclusion and exclusion criteria. Answer: We so far omitted information on the proportion of patients for whom postoperative cystc values were available. We fully agree with Reviewer 3 and followed his recommendation to add this information, which can now be found in the results section. This will allow the reader to evaluate the foundation of our results even better. 12. Line 49: use obtained rather than adopted Answer: We changed the wording accordingly. 13. Line 59: measurement at the time Answer: We changed the grammar accordingly. 14. Page 5 Line 4: what is in Sterofundin? Answer: We appreciate Reviewer 3s question about the contents of Sterofundin. The revised version of the manuscript now contains further information on the type of solution that was administered. 13. Line 9: the Authors make a major assumption stating that bags with fluids and blood products were all empty at the end of the operation. Answer: We thank Reviewer 3 very much for this statement. The authors can assure Reviewer 3 that, according to the experience of the authors in this setting, this assumption is actually accurate in the vast majority of cases. However, we agree that this is a limitation of our study that is due to the retrospective nature of our dataset. Hence, we added this point to the limitations section. Line 27: another assumption, everybody was Caucasian. Does it make a difference in their calculation if some of the patients were actually Asians? Answer: We appreciate this question. The MDRD formula has been validated for Caucasians and African Americans. A correction factor of 1.212 would apply for the latter. From our experience with the study populations and due to the local demographics, the vast majority of patients
treated in this setting are of Caucasian ethnicity. However, this should be noted in future prospective studies more consistently and we therefore added this point to the limitations paragraph. 14. Line 37: I am not sure why they presented the data as median and percentile range rather than mean and SD. Answer: We deliberately chose to present data as median and percentile range as these statistical measures can be used for normally and non-normally distributed data. 15. Line 49: change respective. Answer: We eliminated respective in this line. 16. Also, they eliminated from the analysis patients in whom, I suspect, they did not have a complete set of data. However, they took patient in whom there was only one measurement of GFR either at day 1-3-or 5. It is impossible to determine the weight of having only 1 measurement on the final calculation. Answer: We thank Reviewer 3 for this statement. As described in our answer to point no. 11, we now state the number of patients with 1, 2 or 3 cystatin C values available postoperatively. As an answer to Reviewer 3s suspicion regarding the elimination of cases due to incomplete data, we would like to refer the reviewer to the inclusion/exclusion criteria paragraph within the methods section, also illustrated in the patient flow chart in figure 1. Future prospective study designs may more easily be capable of delivering a more complete set of follow-up data. As mentioned, this is a major drawback of the retrospective study design as repeatedly discussed in the manuscript. 17. Line 49: correct respective Answer: Reiterating point 15 of this list, we eliminated respective in line 49. 18. Page 6: Overall there are so many missing data on the intra and postoperative management of these patients which could have affected the renal function that is difficult to come to an agreement with the Authors conclusions.
Answer: We appreciate Reviewer 3s wish for more information on factors influencing renal function. Due to the retrospective design of this study we were able to analyse data that were documented during the treatment process by the clinicians. We are sure that prospective studies addressing this issue will have the capability to collect a more detailed set of data on this topic. 19. Line 14: Median dose of HES was 1000. The range of 1000-1000 really does not say much to the reader. This is because they chose to present data with median and 25-75 percentile range Answer: We agree. We would like to refer to our answer to point no. 14. 20. Line 27: the presentation of data is confusing. Similarly, on Table 3 it is not clear what are the "Baseline egfrcystc" and the "Preoperative egfrcystc baseline" data. Why do they differ? Answer: We thank Reviewer 3 for this comment. We changed the wording of this paragraph according to this critique. We extended the caption of table 3 to further clarify that the first line of table 3 refers to the number of patients that fall into one of the groups ordered by graded baseline egfrcystc values. This gives information on pre-existing chronic kidney disease according to preoperative baseline egfrcystc measurements. There is no patient with a compromised preoperative baseline egfrcystc (i.e. <60 ml/min), which would indicate advanced chronic kidney disease. We added this information to the results section to highlight this fact. 21. Also, why cystc is increased on line 27 and then is decreased on line 38? Answer: In line 27 the numbers of patients who have increases in perioperative cystc / egfrcystc are outlined. In line 38, the decrease of the median values are stated. We changed the wording of this paragraph to state this clearer. 22. Why did they jump from >5, >10 to 25%? Does it make any clinical difference to go from >5 to >10%? Is it clinically irrelevant? Answer: We thank Reviewer 3 for this question. The reader gets a detailed report of the perioperative course of cystc and egfrcystc values. 23. Page 7 Line 1: what is moderate amount of HES? What is an excessive amount of HES?
Answer: We thank Reviewer 3 very much for this point. Please see our answer to Reviewer 2s point no. 5. 24. Line 7: what is near-normal? Answer: As described in the manuscript a near-normal renal function is defined as an egfr from 60 to 90 ml/min. 24. There is a too extensive list and description of previous studies with details that are confusing, often contradictory. They should significantly eliminate most of this section and just report the most important differences or similarities between their findings and the rest of the literature. I would also question the validity of comparing their data with patients who underwent cardiac by-pass. Answer: According to this suggestion we shortened the paragraph putting our results into perspective with previous studies. 25. Page 10 Line 15: inhibited? I am not really sure of what that means in this context Answer: We changed the word to inapplicable. 26. Line 47: potential benefits of HES? How? Which? Answer: We thank the Reviewer for these questions. We added examples on possible benefits that have not been fully investigated in the specific context of ORP and RARP, yet. To answer these questions, prospective studies are needed. 26. The list of limitations, although accurate, sort of makes the case for reviewing their study as not worth much. I would restate the entire section. Answer: We changed the content of our limitations section according to the recommendations listed by the reviewers.