Electronic Letters to:

Knee: D. J. Beard, H. Pandit, S. Ostlere, C. Jenkins, C. A. F. Dodd, and D. W. Murray Pre-operative clinical and radiological assessment of the patellofemoral joint in unicompartmental knee replacement and its influence on outcome J Bone Joint Surg Br 2007; 89-B: 1602-1607 [Abstract] [Full text] [PDF]

eLetters: Submit a response to this article

Electronic letters published:

Author's reply:

David J Beard, David Murray, Hemant Pandit   (8 May 2008)

Pre-operative clinical and radiological assessment of the patellofemoral joint

W. S. Khan, G. S. E. Dowd   (15 February 2008)

Author's reply: 8 May 2008
David J Beard, University Lecturer in Orthopaedics University of Oxford, David Murray, Hemant Pandit Send letter to journal: Re: Author's reply: david.beard{at}ndos.ox.ac.uk David J Beard, et al.	Sir, We thank Mr Khan and Mr Dowd for their interest in our paper. Although it could be considered minutiae, to some extent there is some merit in the comment about the cases which were excluded because of severe degeneration. As stated, total knee replacement (TKR), rather than unicompartmental knee replacement (UKR), was performed in two out of 824 knees (0.2%) which had severe lateral patellofemoral joint (PFJ) degeneration with substantial bone loss, grooving and subluxation. It was felt that the extreme level of lateral PFJ degeneration in these two patients placed them outside the indications for UKR. The basis for this decision was of clinical, rather than research, origin. As these patients were excluded the study provides no evidence-based guidance for how they should be treated in the future. However, without such evidence we still feel TKR is probably more appropriate, but further study may be warranted. This was a descriptive study with data collected prospectively. No hypothesis was stated and therefore no sample size is appropriate, nor was it performed. The point about degenerative classification is reasonable, and highlights the issue that any results (and conclusions) are highly dependent on the assessment method used, and the problems with radiographic scoring. For this reason we were careful to analyse the data with both Altman and Ahlback scores (as stated). We first analysed the data with all grades and then with various sub-classifications (not published). Despite the varied analysis the conclusions remained the same, hence we chose to report the classification which provided the safest position in terms of clinical recommendations. A classification threshold of Grade II, or score 2 ensured that knees were included if they had a relatively low level of degeneration. Two-sided t-tests were used throughout. Again, these methodological semantics do not influence the main message. The likelihood of less significance after correction (if used) does nothing to alter the main conclusion that no difference was found in outcome between those with or without degeneration - if anything it enhances it. More stringent tests would merely serve to hide the potential problem with the lateral facet. In view of the issues raised earlier in the letter, particularly the detailed concerns about the effect of 0.2% of the cohort, we find this latter unease about potential type 1 error rather odd and contradictory. The inference throughout Mr Khan and Mr Dowd's letter is that there has been some obfuscation which affects the validity of the findings, particularly with regard to the lateral facet results. This is clearly untrue. We have taken great care to ensure that the clinical recommendations are based on fact (especially with regard to the lateral facet). Where fact was absent or the issue contentious we have chosen to err on the side of caution. This is expressed clearly in the conclusion and in the abstract. D.J. Beard, University Lecturer in Orthopaedics, D. Murray, H. Pandit, University of Oxford, Oxford, UK.
Pre-operative clinical and radiological assessment of the patellofemoral joint 15 February 2008
W. S. Khan, Academic Orthopaedic Registrar Royal National Orthopaedic Registrar, Stanmore, London, UK, G. S. E. Dowd Send letter to journal: Re: Pre-operative clinical and radiological assessment of the patellofemoral joint wasimkhan{at}doctors.org.uk W. S. Khan, et al.	Sir, We read this article with interest and feel that there are a number of inconsistencies and inadequacies that cast doubt on the validity of the paper. Firstly, the authors state that "Goodfellow et al recommended that the state of the patellofemoral joint should be ignored when deciding whether or not to implant a Unilateral Knee Replacement (UKR), and we have adhered to these recommendations". Unfortunately, this statement is not supported in their study where patients found to have severe lateral osteoarthritis intra-operatively were excluded from the study and had a total knee replacement performed instead. Furthermore, we could not find any justification for this exclusion in their paper, or indeed elsewhere in the literature. Secondly, we question whether this study is indeed a prospective study. We note that the study cohort represents a subset of a group reported in another study. It is unclear whether the sample size calculation was performed before or after patient recruitment. If this is a prospective study then it is likely that some patients were lost to follow-up. Unfortunately, the authors have failed to state how many. For some reason the American Knee Society scores (AKSS) were recorded pre-operatively but not used in the analyses. Thirdly, we could not find any justification for classifying Ahlback grade II or more, and Altman score 2 or more as degenerative changes. The number of patients considered as having degenerative changes using the very different classifications was also quite varied, 9 and 54 respectively. Furthermore, the authors only compared the group with degenerative changes with the normal group for the Altman grade, and not the Ahlback score. Looking at the Ahlback score analyses that were performed, we believe that if a comparison had been made between the degenerative group as a whole with the normal group, the results would have been significantly better for the degenerative group, and in turn, difficult to explain. Lastly, the authors failed to state whether the t-tests were one-sided or two-sided. If one-sided tests were inaccurately performed then at least three out of six significant p values would not have been significant using two-sided tests. We would also argue that since a large number of tests have been performed, a Bonferroni’s Correction should have been used. This would have reduced the number of significant p values from six to one. W.S. Khan, MBChB, MSc, MRCS, PhD, G.S.E. Dowd, MD, MChOrth, FRCS, Royal National Orthopaedic Hospital, Stanmore, London, UK.