Author’s response to reviews Title: Hip preserving surgery with concentrated autologous bone marrow aspirate transplantation for the treatment of asymptomatic osteonecrosis of the femoral head: retrospective review of clinical and radiological outcomes at 6 years postoperatively Authors:

of the revised manuscript (line 32). Comment 2: abstract line 51 What does this mean? Did you only have follow-up on 26 patients? Answer: We apologize for the unclear formulation in the original manuscript. What we meant was that 31 hips out of 38 (82%) were followed up for more than two years. We have clarified this aspect in the revised manuscript (lines 31-34 of the abstract; lines 90-92 of the main text). Comment 3: pg 3 line 22 They state that their patients are 'idiopathic' but later state that the patients had either corticosteroid or alcohol associations. This would not be considered idiopathic by the medical community. Some have references this as secondary osteonecrosis to distinguish this group from patients with ON from direct causes (tramua, sickle cell, for example). Answer: We agree with the reviewer that the term idiopathic was not originally used in its usual context. In Japan, alcohol and corticosteroid use often turn out to be the etiologies underlying idiopathic ONFH, which is why we used the term idiopathic in the original manuscript. Per the reviewer’s comment, we have specified that we are referring to alcoholor corticosteroidinduced secondary osteonecrosis (line 42). Comment 4: Pg 3 line 26 Please provide a reference for the sentence on incidence. Answer: Per the reviewer’s suggestion, we have added a reference to the work by Fukushima et al. (Clin Orthop Relat Res. 2010;468:2715–24) (line 50 and reference 1 in the revised manuscript). Comment 5: Pg 3 line 44 Please add 'in most patients' to the end of the sentence describing conservative treatment. Answer: We have made the change requested by the reviewer (line 57). Comment 6: Pg 3 line 51 The sentence on core decompression is very biased. I suggest adding 'longterm' outcome is unsatisfactory in a majority of patients. Answer: We thank the reviewer for pointing out the inaccurate statement. We have made the suggested change (line 59). Comment 7: Pg 3 line 51 For rotational osteotomy, again it may have a good result not all patients have a good result. Answer: Per the reviewer’s comment, we have adapted the sentence to reflect the fact that the outcomes of rotational osteotomy may or may not be good (lines 59-61). Comment 8: Pg 4 line 51 This should be 'corticosteroid' to distinguish the type of steroid that was given. Please correct throughout the manuscript. Answer: We thank the reviewer for pointing out this important detail. We have made the requested change (line 78). Comment 9: Pg 5 line 38 What is the mean & range of follow up for each of the types of ON. This is particularly important when trying to interpret the findings of the survival curves. Answer: We agree that follow-up duration represents an important aspect and have included information regarding the mean and range of follow-up duration for each type of ON, as requested by the reviewer (lines 174-176). Comment 10: Pg 5 line 48 Please see comment above about idiopathic patients. Answer: Per the reviewer’s comment, we have removed the term idiopathic when referring to our patients (line 98). Comment 11: Pg 5 line 51 What were the reasons that the remaining 15 patients received corticosteroids? Answer: We realize that this detail was not clear in the original manuscript. The remaining 15 patients received corticosteroids for various conditions, including arthritis and nephropathies. Motivated by the reviewer’s comment, we have explicitly listed the conditions for which corticosteroids were administered (lines 100-103). Comment 12: Pg 9 line 6 Please state that this is the Student's t test. Was this for matched pairs? Answer: We thank the reviewer for pointing out this omission. It was indeed the paired Student’s t test, and we have mentioned this fact in the revised manuscript (line 165). Comment 13: Pg 9 statistics Please provide information on the Survival Analysis. Also, what statistical package was used? Answer: Per the reviewer’s suggestion, we have added information on the Survival Analysis (line 167), and mentioned that the statistical analyses were performed with IBM SPSS Statistics (lines 169-170). Comment 14: Pg 9 para 2 It should be clearly mentioned what the scales are: what is best and what is worst. Also, what is the maximum number of points for each subscore. Answer: We thank the reviewer for this pertinent comment. Per the reviewer’s suggestion, we have clarified the JOA hip score scale, with detailed explanations of the maximum number of points for each category, per the original definition of Nishino et al. (J Arthroplasty. 2013;28:1736–40) (lines 149-152) and reference 12 in the revised manuscript). Additionally, motivated by the reviewer’s comment, we provided a detailed description of the mean and range of scores and subscores obtained before surgery and at the most recent follow-up (lines 181187). Comment 15: Pg 9 line 57 The first sentence of this paragraph is Materials and methods and should be deleted. Answer: We have deleted the sentence, per the reviewer’s suggestion. Comment 16: Pg 10 line 19 This paragraph needs to be expanded so that you understand that this comment is an overview. You need to state what new findings were found in this study. The paragraphs that follow are comments on the different categories of lesions. This needs to be stated in this first introductory paragraph. Furthermore, like other surgical interventions for osteonecrosis, CABMAT was not 100% successful. Therefore, you should modify your sentence to say that it showed the potential to alter the natural course of the disease for some patients. Answer: We realize that our original description was not optimal and are grateful to the reviewer for these very useful suggestions. Motivated by the reviewer’s comment, we have revised this part to clarify that we wish to provide an overall comparison between ONFH outcomes following CAMBAT and those resulting from the natural course of the disease, with emphasis on the new findings. Moreover, as the reviewer suggested, we have mentioned the scope of the following paragraphs and acknowledged the fact that CABMAT was not 100% successful (lines 197-200). Comment 17: Pg 10 line 35 Figure 3 shows your results versus the natural history. Please clearly indicate that the collapse rate for Koo's study in this comparison was specifically to his control 'natural history' group. Furthermore, it would be of particular interest to show the different survival rates for each of the Types of osteonecrosis. Answer: We thank the reviewer for these pertinent comments. We have mentioned the “untreated hip” as the reference used in the study by Koo and Kim (reference 14 in the revised manuscript). While we report lower collapse rates than those mentioned by Koo and Kim, it is worth noting that Koo and Kim defined the necrotic area by an angle, which precludes direct comparison against our present results. Per the reviewer’s suggestion, we have expanded the survival analysis to cover collapse rate for each type of osteonecrosis. We have mentioned this aspect in the revised manuscript (lines 207-211 and legend of Fig. 3). Comment 18: Pg 10 line 48 This collapse rate was for what time periods? Answer: Per the reviewer’s comment, we have mentioned the time periods for which the collapse rates were reported (i.e., 2.3, 1, and 8.3 years in the studies cited as references 15, 16, and 17, respectively; lines 215-216). We hope this provides the readers with a better context in which to assess our findings. Comment 19: Pg 11 line 6 Please be clear that the Min study was for untreated femoral heads. Answer: We thank the reviewer for pointing out this fact, and have mentioned that Min’s observations were for untreated ONFH (lines 221-222). Comment 20: Pg 11 line 16 Are their results for over the same time periods as your study? A comparison to the results of the multiple drill hole core decompression technique a procedure closer to what the authors have done should be included in the discussion. Answer: We thank the reviewer for the insightful comment and kind suggestion. We have expanded the discussion to include the work of Lee et al. (J Bone Joint Surg. 2008;90:852–7), who reported a collapse rate of 43% after core decompression (reference 20 in the revised manuscript). Additionally, we have discussed the works by Mont et al. (reference 21) and Song et al. (reference 22); however, since these authors did not report collapse rate after core decompression, we compared clinical success rates instead (lines 229-234). Comment 21: pg 11 line 29 The transition to mechanism should be distinguished by a subtitle. Answer: Per the reviewer’s suggestion, we have included a subtitle for the discussion of mechanisms (line 236). Comment 22: pg 11 line 45 This sentence needs to be revised. The mechanism is the introduction of new stem/progenitor cells to an area with deficient or dead stem/progenitor cells (i.e. the necrotic lesion). Answer: We thank the reviewer for this pertinent comment. We have revised the sentence to reflect the role of the bone marrow graft in bone formation (lines 243-245) Comment 23: pg 12 line 41 You would need to state that it is less-invasive than what. It is more invasive than a core decompression as you are obtaining cells from the iliac crest. Also you are stating that it is less expensive but you have not mentioned this in the discussion and it was not the topic of this study. Answer: We fully agree with the reviewer’s observation that CABMAT is less invasive than osteotomy, but more invasive than core decompression, and we have mentioned this fact in the discussion section. Additionally, since the study did not evaluate costs, we have deleted the sentence about expenses (lines 251-253). Comment 24: Figure 2 Please label the y axis. Please explain in the legend that a higher score is better. We also need to know what the maximum points are for each subscore. Answer: We thank the reviewer for noting our omission. We have labeled the y axis of Figure 2 and improved the legend by adding an explanation of the JOA hip score scale and how our patients fared in terms of JOA scores. Comment 25: Figures 3 and 5 Please state what the error bars are in the legend. Standard deviation? Standard Error? 95% Confidence Limits? Figure 5 There are no error bars for the last time period. Answer: We thank the reviewer for this observation. The error bars in Figures 3 and 5 represented standard errors. Nevertheless, we feel that including them in the figure did not improve the visualization of the results. Therefore, we have recreated the survival curves in Figures 3 and 5 without error bars. Joonsoon Kang (REVIEWER #2): Comment 1: There is no control group for this surgical procedure. Answer: We fully agree with the reviewer that the lack of a control group is the most important limitation of this study. Specifically, while we evaluated the efficacy of CABMAT compared to the outcomes of the natural course of ONFH (i.e., untreated hips could be considered as controls), a better approach would be to compare also against the outcomes of core decompression after matching for sex, age, and necrosis grade. Therefore, we have mentioned the lack of a control group as limitation of the study (lines 256-262). Comment 2: Size and location of ostenecrosis is very importantin AVN. This paper did not focus to this point.I do not agree to compare your data with natural history of AVN paper. Answer: We fully agree with the reviewer’s that the extent and location of osteonecrosis, are of key importance, and therefore we cannot compare our findings directly with those of past reports, since it is not possible to accurately match the study samples in terms of extent and location of necrosis. Motivated by the reviewer’s comment, we have acknowledged this limitation in the revised manuscript (lines 202-203). Nevertheless, some reports do mention the extent and location of necrosis (accounted for in the type of osteonecrosis), and therefore we believe it is reasonable to compare the findings regarding lesions with the same type (A, B, C1, and C2). The organization of the revised Discussion section reflect this approach.