In response to the letter of Drs. M.H. Malik and M.L. Porter on our work: Meta-analysis is typically applied when a subject has been the centre of prolonged debate and many inconclusive trials have been reported. Meta-analysis increases the power to detect significant differences between the compared treatment strategies, and is especially useful when small (and thus potentially underpowered) trials exist.

All trials in our meta-analysis were randomized. Analyses per trial quality and further subgroup analyses were reported in the online supplementary material.

As stated in the manuscript, all analyses consistently favor resurfacing of the patella. We do agree that there is room for further trials that would probe the same question in selected patient subgroups e.g. the effects of patellar resurfacing in rheumatoid arthritis affected knees.

To the Editor:

I read with interest the article by Pakos, et al. They attempt to address a subject that has been the centre of prolonged debate. What is clear from all the papers that they have included in their meta-analysis is that what is required to answer the question of whether to resurface or not is a large properly constructed and executed randomised blinded prospective study given the many variables that can affect outcome following total knee arthroplasty.

By performing a meta-analysis on a variety of under-powered and often poorly performed studies we fear that they have only clouded the issue further. Given that outcome is dependent upon a multitude of variables such as implant design, surgical technique, and others that may not, as yet, be fully understood and that the process of meta-analysis itself is often affected by publication bias. The only useful method of addressing this question still remains to be performed in an adequately powered study.

To the Editor:

The article "Patellar Resurfacing in Total Knee Arthroplasty" by Emilos Pakos, et al,[1] once again revives the controversy surrounding “routine” patellar resurfacing. Although the statistics of the paper are indisputable, the article does not elaborate on a few issues which have always been the argument for proponents of not resurfacing the patella.

The incidence of complications of patellar resurfacing has ranged from 5% to 50% and such complications have accounted for as many as half of the total revisions performed[2]. Revision surgery after patellar resurfacing is even more difficult in case of failed patellar components and extensor mechanism failures[3]. Surgeon related technical factors remain a very important factor in patellar resurfacing and, as rightly pointed out by Hagena[4], the rate of patellar complications is much higher than reported in the literature, as most of our figures are from studies of knee arthroplasties performed by specialised centres with authors who have more knee arthroplasty experience than the vast majority of “average” orthopaedic surgeons with a more general practice. The high prevalence (up to 19%) of anterior knee pain is the main adverse outcome of not replacing the patella, but Levitzky, et al,(5) who reported a prevalence of 19% peripatallar knee pain in patients without patellar resurfacing reported no revisions due to this pain at 7.5 year follow up. In most cases this pain was not severe.

The Swedish Knee Arthroplasty register [6] suggests that the need for secondary patellar replacement is balanced by the need for revision of failed patellar components.

The decision To resurface or not to resurface the patella is not an easy one and we would like to congratulate Pakos, et al, for reviving this debate with their extremely well researched article.

1. Pakos EE, Ntzani E, Triklinos TA: Patellar Resurfacing in Total Knee Arthroplasty-A Meta-Analysis. JBJS American 2005; 87-A(7): 1438-1445.

2. Ayers DC Dennis D, Johanson NA, Pelligrini VA: Instructional Course Lectures, The American Academy of Orthopaedic Surgeons - Common Complications of Total Knee Arthroplasty. J. Bone Joint Surg. Am 1997; 79: 278-311.

3. Laskin R: Management of the patella during revision total knee replacement arthroplasty. Orthop Clin North Am 1998; 29: 355-360.

4. Hagena F: The patella need not be replaced during total knee replacement. In: Laskin RS, ed. Controversies in Total Knee Replacement, First ed: Oxford University Press, 2001; 187-213.

5. Levitsky KA Harris W, McManus J, Scott RD: Total knee arthroplasty without patellar resurfacing. Clinical outcomes and long-term follow-up evaluation. Clin Orthop Relat Res. Jan;(286):116-21 1993; 286: 116-121.

6. Robertsson O Knutson K, Lewold S, Lidgren L: The Swedish Knee Arthroplasty Register 1975-1997: an update with special emphasis on 41,223 knees operated on in 1988-1997. Acta Orthop Scand. 2001;72(5):503-513.

To the Editor:

The article by Pakos, et al, "Patellar resurfacing in total knee arthroplasty: A meta-analysis", was fascinating in its design, detail, and its conclusions. However, while it advanced our knowledge to a limited extent, I believe the authors missed a marvelous opportunity to help answer the controversial question of whether or not to resurface the patella at the time of total knee arthroplasty.

It does not take a study to conclude that the native patella is more likely to function successfully when articulating with an anatomically shaped trochlea (congruent articulation) than when articulating with a trochlea shaped like a segment of a toroid (incongruent articulation). The long-term success of hemi-arthroplasty of the hip in which a metal implant achieves a congruent articulation with the native acetabulum is analogous.

The lost opportunity occurred when the authors failed to gather data about the trochlear design of the implants used in the “ten independent randomized trials” they subjected to meta-analysis. In their Discussion, the authors recognized published data showing a relationship between results and “the type of prosthesis used” (1,2,3), among other factors. However, the reason they gave for failing to gather such information was that it ”would necessitate detailed data on individual patients”(italics added for emphasis). This reason does not apply to “the type of prosthesis used” because that information should be found in the “Materials and Methods” section of all reputable clinical trials about total joint replacements. The shape and design of the trochlear portion of the femoral component is of vital importance when one is trying to determine if resurfacing the patella will create more complications and failures compared to leaving the native patella intact.

The authors acknowledged as a study limitation “the cumulative sample size (1200 knees) was not very large because the majority of trials included relatively few knees”. I know of only one study with a larger sample size (4,743 knees) that could help shed some light on this problem. (4) Admittedly this study was a retrospective multicenter worldwide effort to document the long-term survivorship of a specific implant. Because this implant has an anatomic trochlear design, all participating surgeons followed the same surgical implantation technique, and the only difference was whether the patella was resurfaced or not. Only primary total knee arthroplasties were included. The 27 international participants came from 11 different countries. 4,743 TKAs were performed between 1981 and 1997. Follow-up ranged from 2 to 18 years, averaging 5.7 years. The patella was replaced in 2,838 knees (60%) and not replaced in 1,905 knees (40%). Failure was defined as re-operation for any patellofemoral problem. Thirty of the resurfaced patellas (1.1%) required subsequent patellar surgery; in contrast, two of the non-resurfaced patellas (0.1%) required subsequent resurfacing. Out of necessity, this study was supported by the manufacturer of the implant, but the huge sample size and the ten-fold difference in treatment effect between patellar resurfacing and non-resurfacing commands attention.

The authors could perform a great service to the orthopedic community by gathering data from the same 10 studies on “the type of prosthesis used” and publish their analysis concerning any effect relating to prosthesis type. I congratulate the authors on an interesting study and look forward with anticipation to learning whether or not prosthesis type makes a difference.

References

1. Andriacchi TP, Yoder D, Conley A, Rosenberg A, Sum J, Galante JO. Patellofemoral design influences function following total knee arthroplasty. J Arthroplasty. 1997;12:243-9.

2. Chew JT, Stewart NJ, Hanssen AD, Luo ZP, Rand JA, An KN. Differences in patellar tracking and knee kinematics among three different total knee designs. Clin Orthop Relat Res. 1997;345:87-98.

3. Scuderi GR, Insall JN, Scott NW. Patellofemoral pain after total knee arthroplasty. J Am Acad Orthop Surg. 1994;2:239-46.

4. Hamelynck KJ, Stiehl JB, Voorhorst PE: The LCSâ worldwide multicenter outcome study. In Hamelynck KJ, Stiehl JB (eds): LCSâ mobile bearing knee arthroplasty: a 25 years worldwide review. Berlin, Springer-Verlag, 212- 224, 2002.