Title: Acute whiplash-associated disorders (WAD) grades 1-2: Are MRI high-signal changes of alar and transverse ligaments related to outcome?

Transcription

1 Author's response to reviews Title: Acute whiplash-associated disorders (WAD) grades 1-2: Are MRI high-signal changes of alar and transverse ligaments related to outcome? Authors: Nils Vetti Jostein Kråkenes Geir E Eide (geir.egil.eide@helse-bergen.no) Jarle Rørvik (jarle.rorvik@helse-bergen.no) Nils E Gilhus (nils.gilhus@helse-bergen.no) Ansgar Espeland (ansgar.espeland@helse-bergen.no) Version: 2 Date: 13 October 2010 Author's response to reviews: see over

2 1 Editor BMC Musculoskeletal Disorders Bergen, October 13, 2010 Thank you for allowing us to submit a revised version of the manuscript: Are MRI highsignal changes of alar and transverse ligaments in acute whiplash injury related to outcome?. Below is a point-by-point response to the reviewer s comments. The comments are given in italics. All changes made are marked with yellow in the revised manuscript. Reviewer 1 Confidential comments to editor Title: Acute whiplash-associated disorders (WAD) grades 1-2: Are MRI high-signal changes of alar and transverse ligaments related to outcome? Version: 1 Date: 7 September 2010 Reviewer: Dejan Ozegovic Confidential comments to editors: 1. Is the question posed by the authors well defined? YES 2. Are the methods appropriate and well described? - How were the controls selected? What demographic information is presented re: them? Reply: The controls were recruited from responders of inviting letters sent to persons aged 18 to 80 years in the National Population Register of Bergen, Norway. All controls and patients were interviewed by the same researcher evaluating the predefined inclusion and exclusion criteria before a final study enrolment. Age and gender, but no other clinical data were registered on the included controls. Details on the controls are given in an earlier paper (Vetti et al in press Spine) and are not given here. Is there any supporting evidence for how the measurement of ligament grading was done? If not, what was the rationale for this? Pragmatic? Reply: Ligaments were graded according to an established grading system validated on several different WAD groups and controls (Kråkenes et al 2006, Myran et al 2008, Dullerud et al 2010). In all these studies the grading was based on the ratio between any high-signal part and the total cross-sectional area of the ligament. This ratio was estimated subjectively /

3 2 visually by the interpreters and no tools for objective measurements of cross-sectional areas were applied. As these ligaments are composed of continuous collagen fibres, the image with the largest cross-sectional area of high signal was used for grading. Who performed the WAD grading for the participants? Reply: First author ascertained the WAD grading by interviewing the patients and reviewing reports from the clinicians / radiologists who had evaluated the patients for neck pain and neurological findings or neck fractures / dislocations. This has now been commented in the method part (page 5, lines 2-3). The need for conventional radiography or CT in excluding neck fracture or dislocation was independently evaluated by the physician clinically responsible for each patient, and such examinations were not interpreted by our study radiologists. 3. Are the data sound? - Statistical methods appear appropriate 4. Does the manuscript adhere to the relevant standards for reporting and data deposition? YES 5. Are the discussion and conclusions well balanced and adequately supported by the data? - YES - Statement: "patients with previous neck problems probably have a poorer prognosis, and the prognosis after isolated whiplash trauma cannot be ascertained including such patients" - while this is an ideal scenario and required for determining causation, it is difficult to be able to state conclusively. Given the prevalence of neck pain in the general population (Neck Pain Task Force chapter), and poor recall by patients regarding previous pain episodes (Carragee) it seems unlikely that those stating no previous neck pain actually didn't have neck pain. Authors could discuss this point stating that they might have some misclassification but would be representative of a typical clinical population seeking care for WAD. Reply: In the method part of the revised manuscript we present our definition of prior neck problems: prior neck pain of more than 30 days in total or reported treatment for neck problems during the last 10 years. Mild or short lasting neck pain did therefore not automatically lead to exclusion. Nevertheless, due to recall bias, we agree that some of our included patients actually could have prior neck complaints according to our definition. This has now been outlined in the discussion part (page 11, lines 17-19). 6. Are limitations of the work clearly stated? - Would benefit from further discussion of limitations including dichotomization of outcomes and variables and rationale for this (expectations, etc). Reply: To be able to report on the proportions recovered at follow-up, which is common practise in WAD cohorts, we dichotomised our continuous outcome measures. The most validated cut-off values according to literature were chosen; NDI score > 8% and last week neck pain NRS-11 > 4 (Vernon et at 1991,Gabel et al 2008, Sterling et al 2003, Kongsted et al 2008, Jensen et al 1989). These cut-off values were settled before performing the crude and logistic regression analyses. However, to test the robustness of our results we also performed linear regression analyses with the same explanatory factors and NDI score and last week neck pain NRS-11 score as continuous outcome measures. We also performed the logistic regression analyses when treating the continuous explanatory variables (age, initial pain and

4 3 IES points) uncategorised. The results from these supplementary analyses confirmed our main result of no relation between ligament high-signal changes in the acute phase of injury and outcomes at 12 months follow-up. We found it inappropriate to present all these additional analyses in the manuscript. The expectation variable was originally categorised in three levels but later dichotomised as only one patient had answered that she / he to a little extent expected to get rid of the neck pain after the accident. Re: disagreements between the radiologists on presence of grade 2-3 changes, should have some discussion to why their kappa scores were only moderate? Why were their experience levels so discordant? Could this impact on the results? Reply: The MRI protocol for visualising the upper neck ligament structure and grading system for assessing ligament high-signal changes were developed by Kråkenes et al (2001). These authors also elaborated on potential explanations for disagreement which was not the purpose of this study. Our kappa values on grading ligament high-signal changes are in line with previous reports on WAD patients and controls (Kråkenes et al 2006, Myran et al 2008) and many MRI examinations in daily use. The discordant experience levels reflect the two interpreters experience in radiology at the time of study interpretation. However, both radiologists were trained in and had practised interpreting MRI images of upper neck ligaments, and the disagreement between them on presence of grade 2-3 changes was hardly caused by their different experience levels. 7. Do the authors clearly acknowledge any work upon which they are building, both published and unpublished? YES 8. Do the title and abstract accurately convey what has been found? YES 9. Is the writing acceptable? YES What next?: Accept after minor essential revisions Reviewer's report Title: Acute whiplash-associated disorders (WAD) grades 1-2: Are MRI high-signal changes of alar and transverse ligaments related to outcome? Version: 1 Date: 7 September 2010 Reviewer: Dejan Ozegovic Reviewer's report: Please see above which would all become minor essential and discretionary revisions. Level of interest: An article whose findings are important to those with closely related research interests Quality of written English: Acceptable Statistical review: Yes, but I do not feel adequately qualified to assess the statistics. Declaration of competing interests: No competing interests to declare.

5 4 Reviewer 2 Reviewer's report Title: Acute whiplash-associated disorders (WAD) grades 1-2: Are MRI high-signal changes of alar and transverse ligaments related to outcome? Version: 1 Date: 20 September 2010 Reviewer: Alice Kongsted Reviewer's report: Review comment on 'Acute whiplash-associated disorders (WAD) grades 1-2: Are MRI highsignal changes of alar and transverse ligaments related to outcome? ' This is a well-written paper on an important topic. To my knowledge, it is the first study on the predictive value of MRI signal changes in the ligaments of the upper cervical spine. The objective of the study is clearly stated and the research methods are appropriate. However, I think the background is a little too sketchy and I believe the use of the terms whiplash and WAD throughout the paper is misleading. The participants were exposed to all kinds of car accidents and not exclusively to whiplash traumas. I am not competent to judge whether the performed MRI sequences represent best standard, and I hope the paper will be reviewed by someone with expertise in MRI. Major Compulsory Revisions (which the author must respond to before a decision on publication can be reached): The cohort is described as consisting of whiplash injured subjects (WAD 1-2), but it seems that the inclusion of participants was not restricted to those exposed to a whiplash trauma. A widely accepted definition of whiplash is: an acceleration-deceleration mechanism of energy transfer to the neck (Spitzer et al. 1995). This does not include contact trauma to the head or spine, and is not in accordance with inclusion of all types of car accidents or with the fact that some of the participants sustained a head injury. Would it be more appropriate to describe this cohort by other terms than whiplash /WAD? Otherwise you should which definition of whiplash injury you made use of and report how many of the subjects were whiplash-exposed and how many sustained other types of injuries. Reply: The Quebec Task Force (QTF) (Spitzer et al 1995) defined whiplash and Whiplash- Associated disorders (WAD) as follows: Whiplash is an acceleration-deceleration mechanism of energy transfer to the neck. It may result from rearend or side-impact motor vehicle collisions, but can also occur during diving or other mishaps. The impact can result in bony or soft-tissue injuries (whiplash injury), which in turn may lead to a variety of clinical manifestations (Whiplash-Associated disorders). The critical criterion for applying the term whiplash is thus that an acceleration-deceleration mechanism of energy transfer to the neck has taken place. To confirm such mechanism after car accidents is quite difficult and practically almost impossible to do retrospectively. It seems likely that such accelerationdeceleration mechanism of energy transfer to the neck takes place at most rear-end, front-end and side impact car accidents. One of these types of car accidents was described in 94% (104/111) of the patients in our cohort. Other types of car accidents, which theoretically might less obviously involve such whiplash mechanism, concerned 6% (7/111) of our participants. However, all these patients had experienced rollover crashes or combined / complex types of car accidents which also likely induce an acceleration-deceleration mechanism of energy

6 5 transfer to the neck. As a consequence if not all, at least the vast majority of our included patients had sustained whiplash trauma according to the definition above. In this study we intended to include only those with WAD1-2 according to the WAD grades given by the QTF. Therefore, patients with neurological symptoms (WAD3), neck fractures or dislocations (WAD4), and those without symptoms (WAD0) were excluded. Contact trauma or head injury are not given as independent exclusion criteria in the definitions of whiplash or WAD and may occur concomitantly to a whiplash trauma. Nevertheless, according to literature it seems common practise not to enrol patients with head injuries in whiplash cohorts, especially those with signs of severe head injuries (ref Sterling et al 2006, Kongsted et al 2008,Borchrevink et al 1997). We registered patients subjective report of concomitant head injury (did you hit your head during the accident?). Thirteen patients (11.7%) reported such injury, but only two reported unconsciousness in relation to their car accident (< 5 minutes), and only one of these two was hospitalised. This suggests that few (if any) of the participants had sustained severe head injury due to direct contact trauma to the head. Based on these considerations we feel that it is appropriate to use the term WAD patients in the text when describing our cohort. We have now stated more clearly which definition of whiplash we have used (page 4, lines 9-10). Table 1 has been extended to specify the types of car accidents (impact directions, Table 1, page 23). I suggest that the term WAD is removed from the title. Reply: We agree that the use of the term WAD could be avoided in the title. We have now changed the title to. Are MRI high-signal changes of alar and transverse ligaments in acute whiplash injury related to outcome?. How was the sample size determined? Please, include the basis for your power calculations in the statistical analyses paragraph. Reply: When calculating the sample size for the acute WAD1-2 cohort, we focused on detecting relevant differences in the prevalence of alar and transverse ligament high-signal changes between patients at the acute phase of injury and the non-injured controls (results given in Vetti et al in press Spine). Based on the prevalence data available at that time and using a significance level of 5% and a power of 80%, we estimated prior to the study in a 2:1 design that 150 controls and 75 acute WAD1-2 patients would be needed if a difference in prevalence from 5% in the control group (Kråkenes et al 2006) to 18% in the acute WAD group should be detected as statistically significant. To account for inadequate MRI examinations and to increase the ability to detect relevant associations within the WAD group (like the association between outcomes and ligament changes), we somewhat increased the number of acute WAD patients (n = 114). The accurate number of study participants needed to find a relevant difference in outcomes at follow-up between WAD patients with and without ligament high-signal changes at the acute phase of injury could not be calculated due to lack of relevant pre-study information. We did not know the likely response rate at follow-up. The prevalence of alar and transverse ligament high-signal changes in acute WAD1-2 was unknown. Data on the distributions of our outcomes of neck disability (NDI > 8%) and neck pain (NRS-11 >4) in WAD1-2 groups were

7 6 sparse, and a clinically relevant effect size (difference in proportions recovered at 12 months follow-up) had not been established. However, by assuming that ligament high-signal changes in the acute phase of injury would be found in one third of patients it could be calculated that in a group of 100 responders a difference in the proportions recovered from 60% in those without ligament changes to 30% in those with ligament changes would be detected as statistically significant. Expecting a high response rate, inclusion of 114 acute WAD1-2 patients was thus considered adequate also for the purpose of the present study. This has now been incorporated into the statistical analyses paragraph (page 8, lines 11-15). How did you define prior neck complaints? I think this criterion must have made it very difficult to include participants, since neck complaints are extremely common in the adult population? Reply: Prior neck complaints was defined as prior neck pain of more than 30 days in total or reported treatment for neck problems during the last 10 years. This has now been added in the method Patients paragraph (page 5, lines 4-5). We also changed the term prior neck complaints to prior neck problems throughout the manuscript. It is true that excluding patients with prior neck complaints made it more difficult to include participants. However, according to the definition above, mild or very short lasting prior neck complaints did not necessarily lead to exclusion. Furthermore, we included patients both from a large primary ward and a hospital clinic and managed to complete the inclusion within a reasonable time schedule. Of 254 patients fulfilling the inclusion criteria during our study period, 121 patients were excluded due to prior neck injury or whiplash trauma and / or prior neck problems. Further details on the number of acute WAD1-2 patients excluded according to the different exclusion criteria are given in a different paper (Vetti et al in press Spine). Minor Essential Revisions (such as missing labels on figures, or the wrong use of a term, which the author can be trusted to correct) Abstract: The IES scale and measures of expectations should be mentioned in the methods section when reported on in the result section. Reply: This has been incorporated (page 2, lines 13-15). In the conclusion you state that MRI has limited value for treatment, but you did not investigate MRI findings as a potential modifier of treatment effects. It would be in line with your results to describe a limited value in relation to examination, diagnosis or prediction of prognosis. Reply: Agree. This sentence has been changed in the abstract (page 3, line 6) and in the conclusion (page 13, line 9). Background: The background is short and to the point, which I find attractive. However, I need some background information on the self-reported potential prognostic factors you include in

8 7 addition to the MRI. Further, I think new readers in the field would benefit from a little more details on what has been investigated previously in relation to upper cervical ligaments in WAD. Reply: The background part has been extended according to the comments above (page 2, lines and and page 3, lines 1-3). Methods MRI evaluation: How was the area of the ligaments measured? Please describe the tool used for that. Reply: The ratio between any high-signal part and the total cross-sectional area of the ligament was assessed subjectively / visually without using any measurement tool, as now indicated in the paper (page 6, lines 22-23). The same method was used in all prior MRI studies applying this grading system on upper neck ligaments (ref Kråkenes et al 2006, Myran et al 2008, Dullerud et al 2010, Roy et al 2004). The alar and transverse ligaments are delicate structures, and objective measurements of the area of the ligaments would be difficult. In some cases the ligament s cross section has a speckled appearance making an objective measurement of the high-signal part even more difficult. Discussion: p.9, 2nd paragraph: You argue that the signal changes are unlikely to be injury-induced since you would then expect some impact on prognosis. That is not necessarily true since you may have structural damage from an injury which heals without leading to any long-term complaints. Reply: We agree to this comment and have changed the formulation (page 10, lines 19-22). p.10, 3rd paragraph: I agree that WAD should preferable be studied in people didn t have neck pain prior to the injury. However, as mentioned above, neck pain is an extremely common complaint in the background population, and studies excluding people with previous neck pain would not only be hard to conduct they also do not reflect the reality in which the results of clinical research are to be implemented. Reply: As we defined prior neck problems as prior neck pain of more than 30 days in total or reported treatment for neck problems during the last 10 years mild or short lasting neck complaints did not necessarily lead to exclusion. About half of candidates fulfilling our inclusion criteria were excluded due to prior neck injury or whiplash trauma and / or prior neck problems (see above). We were nevertheless able to fulfil the initial inclusion of WAD1-2 patients within a 22 months period (from May 2007 until Mars 2009). We agree that excluding people from participation should raise questions regarding external validity of the study. Our results can principally be implemented only on acute WAD1-2 patients without prior neck complaints as defined above. Nevertheless, in the present study, no relation between ligament high-signal changes in the acute phase of injury and outcomes at 12 months follow-up was proven. It seems unlikely that such relations would be found in acute WAD1-2 patients with known prior neck complaints. Table 2: It would help the reader s interpretation of the results if you add 95 % CI to the percentages that are presented.

9 8 Reply: We could add 95% CI to the percentages that are presented in table 2. However, we are concerned that this could cause interpretation difficulties for some readers as overlapping 95% CIs might connect to p-values < An alternative approach could be to rather present the difference in proportions between the categories and add the 95% CIs of this difference (which would not include zero if significant differences). We feel such a table would be generally less informative and have chosen to keep the original table. However, to ease interpretation we have now stated more clearly in the footnote that the p-values are based on the difference in outcome proportions between the categories (footnote table 2, page 25). Discretionary Revisions (which are recommendations for improvement but which the author can choose to ignore) Abstract: Methods line 4. Patients reported NDI It may be more easily readable to state: Patients completed the Neck Disability Index and an 11-point numeric rating scale on last week pain intensity? Reply: Agree, changes made (page 2, lines 16-17), but parenthesis needed to define abbreviations used later in abstract. Discussion: Please discuss if the observed effect sizes in relation to signal changes would be clinically important if they were statistically significant. Reply: It seems unlikely that the observed effect sizes of 6-13% (table 2) between patients with and without high-signal changes would be clinically relevant if they were statistically significant (i.e. had been found in a larger study sample). Although a clinically relevant effect size has not been established, we feel the present study had adequate power to detect clinically important differences in proportions recovered between acute WAD1-2 patients with and without high-signal changes. We have added the basis of our sample size calculations at the statistical method paragraph (page 8, lines 11-15) and made a short comment on the study power in the discussion part (page 11, lines and page 12, line 1). Level of interest: An article of importance in its field Quality of written English: Acceptable Statistical review: No, the manuscript does not need to be seen by a statistician. Declaration of competing interests: 'I declare that I have no competing interests'

10 9 Reviewer 3 Reviewer's report Title: Acute whiplash-associated disorders (WAD) grades 1-2: Are MRI high-signal changes of alar and transverse ligaments related to outcome? Version: 1 Date: 20 September 2010 Reviewer: Paul Ivancic Reviewer's report: Major Compulsory Revisions This is a well written paper in which the authors continue their investigation of upper neck ligament injuries due to motor vehicle crashes (MVCs). The authors performed cervical spine MRIs of MVC occupants within 0-13 days following the crash. They observed high-signal changes of the alar and transverse ligaments but found that these changes did not affect outcome. They conclude that the high-signal changes were unlikely to be caused by the MVC and that upper c-spine MRI is not warranted in acute WAD1-2. In the majority of the authors earlier work, they determined upper neck ligament injuries during the late stage of whiplash injury, in which the MRI was done 6 years following the crash, on average. In their earlier studies, the authors observed that whiplash can cause injuries to the upper neck ligaments, detectable on MRI during the late stage of whiplash injury. They have also reported a higher injury risk due to rotated head posture at the time of the crash and reduced active neck range of motion in WAD patients in the late whiplash stage. 1. Please expand the Discussion section and relate your present findings to your previously reported results: a) Why were crash-induced ligament injuries detected approx 6 yrs following the crash, but not within 0-13 days? Reply: This is an appropriate question considering that the same proton-weighted MRI sequences and the same system for grading high-signal changes were used in these studies. However, the results from the initial study by Kråkenes et al, who reported a significant difference in prevalence of alar and transverse ligament high-signal changes between chronic WAD patients and non-injured controls, have not been confirmed by others (Myran et al 2008, Dullerud et al 2010). Especially, the results on the prevalence of alar ligament highsignal changes in non-injured volunteers vary between studies. For comparison to our acute WAD1-2 group, we included a separate large control group of non-injured controls and found a higher prevalence of alar ligament changes (31%, Vetti et al in press Spine) compared to Kråkenes et al (6.7%, Kråkenes et al 2006). Thus the cause of alar and transverse ligament changes during the late stage of whiplash injury is not clear. Baring these considerations in mind, there are still factors that might contribute in explaining the discrepancy between the results from the acute and the chronic WAD group. The acute morphologic changes caused by a trauma, i.e. edema or bleeding might not be detected as high-signal changes on our MRI sequences. An eventual subsequent repair process with fibrosis and scarring could nevertheless cause high-signal changes at a later stage of injury. This possibility has now been outlined in the discussion part (page 12, lines and page 13, lines 1-2). Furthermore, the chronic WAD group of Kråkenes et al consisted of patients all fulfilling the WAD2 criteria both in the acute phase and weeks later, and might have

11 10 experienced more severe traffic accidents compared to our acute WAD1-2 patients. This is discussed in an earlier paper (Vetti et al 2009) and not elaborated in the present manuscript. b) Were some of the late injuries due, for example, to accelerated degeneration caused by the crash, as compared to the controls? Reply: Our earlier work has indicated that alar and transverse ligament high-signal changes can not be explained by accelerated degeneration as such changes were not related to time since injury (40 days to 59 years, median 5 years) in a large group of clinically referred WAD1-2 patients (Vetti et al 2009). This has now been incorporated in the discussion part (page 12, lines 23-24). However, it is still possible that in some cases an acute ligament injury could induce morphological ligament changes producing high-signal changes at a later stage of injury (see above and page 12, lines 20-22). We are currently analysing follow-up MRIs from the present WAD1-2 cohort which hopefully will add valuable information on how ligament structure may alter over time after a whiplash injury. These analyses are not completed, and results are planned given in a separate paper. c) Is it possible that some individuals who develop chronic symptoms following a MVC may exhibit acute ligament injuries, detectable using MRI? Reply: We can not rule out that some individuals who develop chronic symptoms following a MVC may exhibit acute ligament injuries, detectable using MRI. However, the results from our acute WAD1-2 cohort have suggested that alar and transverse ligament high-signal changes in the acute phase of injury are not caused by the acute mechanic incident and are not related to outcome 12 months after the car accident. Our MRI protocol of proton-weighted and STIR sequences was settled after performing extensive pilot MRI examinations on 24 volunteers in order to achieve the best standards for imaging the upper neck soft tissues structures. Advance in MRI technology will possibly permit detection of subtle traumatic ligament changes that might not have been visualized by using our current MRI protocol. d) At what point with the 13-day to 6 yr period would MRI be warranted to assess upper ligament integrity? Reply: The result from the present study has suggested that high resolution upper neck MRI is not warranted in the acute phase of injury. Further studies including follow-up MRI will be needed to conclude on its value at a later stage of injury. This has now been commented in the discussion part (page 12, lines and page 13, lines 1-2). 2. In general, please provide a rationale for your study of only the upper neck ligaments. Why were the mid and lower neck ligaments not investigated? Reply: The rationale for investigating the alar and transverse ligaments is given in the background part which has now been extended with some more details on previous MRI studies of the upper neck ligaments (page 3, lines 11-17). The MRI protocol and grading of high-signal changes used in the present study was designed to evaluate upper neck ligaments and have not been validated on the mid and lower neck ligaments. It would be interesting also to evaluate the mid and lower neck ligaments, but this was beyond the purpose of this study, as now stated more clearly in the discussion part (page 12, lines 11-13).

12 11 Minor Essential Revisions 1. Pg 2 (Abstract: Methods). Suggest changing "Within 13 days after injury..." to "Within 13 days after the motor vehicle crash..." Reply: Agree, except that we prefer to use car accident in stead of motor vehicle crash, see abstract (page 2, line 11). 2. Pg 4 (Methods). Please indicate the percentage of occupants involved in each specific crash configuration (rear vs frontal vs side, etc). From Table 1, I understand that 62% were rear crashes. What was the crash configuration for the remainder? Reply: Conducted, see table 1(page 23) Discretionary Revisions None Level of interest: An article of importance in its field Quality of written English: Acceptable Statistical review: No, the manuscript does not need to be seen by a statistician. Declaration of competing interests: I declare that I have no competing interests.

13 12 Reviewer 4 Reviewer's report Title: Acute whiplash-associated disorders (WAD) grades 1-2: Are MRI high-signal changes of alar and transverse ligaments related to outcome? Version: 1 Date: 21 September 2010 Reviewer: James Elliott Reviewer's report: Discretionary Revisions These are recommendations for improvement which the author can choose to ignore. For example clarifications, data that would be useful but not essential 1. Why did the authors choose to dichotomize NDI scores (< 8% and > 8%)? Previous research has shown that those with NDI scores < 8 should be considered recovered, 8-28% are considered to have mild pain and disability and > 30% would earn membership to a moderate to severe category. This may be important, as it is only those with mod/sev pain and disability that present with a complex condition and poor recovery, presumably due to a more severe injury. Perhaps, I ve missed something here, but it seems their cut-off is too low to appreciate any significant group differences with regards to ligament damage? Reply: To be able to report on the proportions recovered at follow-up, which is common practise in WAD cohorts, we had to dichotomise our continuous outcome measures. The most validated cut-off values according to literature were chosen; NDI score > 8% and last week neck pain NRS-11 > 4 (Vernon et at 1991,Gabel et al 2008, Sterling et al 2003, Kongsted et al 2008, Jensen et al 1989). These cut-off values were settled before performing the crude and logistic regression analyses. However, to test the robustness of our results we also performed linear regression analyses with the same explanatory factors and NDI score and last week neck pain NRS-11 score as continuous outcome measures. The results from these analyses confirmed no relation between ligament high-signal changes in the acute phase of injury and outcomes at 12 months follow-up (results not given in the manuscript). The reviewer has suggested a different cut-off value for the NDI score. We have therefore performed supplementary crude analyses with the outcome measure NDI score dichotomised into NDI > 28% or NDI 28% showing no significant relations between high-signal changes and outcome (p = alar, p = transverse). We feel presenting these additional analyses in the manuscript should not be necessary. In the discussion part we have originally stated shortly that our main result was highly robust (page 10, line 8). 2. It is clear that trauma may not provide a trigger for ligament signal changes, but this reviewer wonders what, if any, influence a post-traumatic response (IES) had on the MR findings? Can the authors comment on this? Reply: It seems very unlikely that a post-traumatic response should have an independent effect on the MRI findings, and we have not analysed relations between IES and ligament high-signal changes in our acutewad1-2 cohort. However, post traumatic stress response is regarded as a potential prognostic factor in WAD. Since IES was related to outcome with a p < 0.2, this variable was included in the logistic regression models.

14 13 3. Was there any change in MRI findings over the course of the study? Did the grade 2 ligaments become grade 3 over the 12 months? Reply: To answer this interesting question, data from MRI examinations performed at 12 months follow-up would be required. We have obtained such examinations and are currently analysing the images (page 12, lines and page 13, lines 1-2). The results are planned given in a separate paper. Respectfully submitted Level of interest: An article of importance in its field Quality of written English: Acceptable Statistical review: No, the manuscript does not need to be seen by a statistician. Declaration of competing interests: I declare that I have no competing interests Yours sincerely Nils Vetti