Abstract
Dear Editor:
This letter represents the deliberations of a group of sports medicine fellows and faculty at the Department of Family Medicine, Jefferson Medical College, in Philadelphia. As part of a research methods course, we reviewed the article by Wojtys et al. entitled “Association Between the Menstrual Cycle and Anterior Cruciate Ligament Injuries in Female Athletes” (September/October 1998, pages 614–619). The authors reported that “… a significant statistical association was found between the stage of the menstrual cycle and the likelihood for an ACL injury (P = 0.03).”
As part of our class review of this article, we attempted to duplicate the statistical analysis shown in Table 1. To our surprise, our calculations did not coincide with those of the authors. In fact, our analysis yielded a chi-square value of 4.86, which is nonsignificant at 0.05. We would have concluded that there was no association, and so we are asking what other method the authors may have used to arrive at their conclusion.
What appears to have happened is that the observed frequency for the Stage 1 calculation was included in the denominator when it should have been the expected frequency. The calculations for Stages 2 and 3 are correct. The effect of the substitution was to inappropriately raise the obtained chi-square, such that a relationship was reported between the athlete's cycle stage and the prevalence of ACL injury, when the data do not support such a relationship.
May we also add that it is not clear why there was a P value for each row of Table 1. We thought that the analysis should be done for the entire set of data for the 28 athletes, not row by row. Dr. Wojtys and his colleagues may have had another process in mind.
However, we would like to take exception with one point: you state that there is no relationship between cycle stage and prevalence of ACL injury. We believe this is wrong. Those data remain the same. There were fewer injuries than expected in the follicular phase and more injuries than expected in the ovulatory phase. Even though the relationship is not statistically significant (P < 0.05), the data clearly support a trend (P = 0.09). As we emphasized in our paper, further investigation is needed to determine if this relationship is fact.
Thank you for your meticulous attention to this work.
Dear Editor:
I read with great interest the article by Muhle et al. entitled “Effect of a Patellar Realignment Brace on Patients with Patellar Subluxation and Dislocation. Evaluation with Kinematic Magnetic Resonance Imaging” (May/June 1999, pages 350–356) and applaud the authors’ many important contributions to the field of kinematic MRI. Our group originally created the kinematic MRI procedure for the patellofemoral joint in 1988; we developed the current, state-of-the-art active-movement, against-resistance technique; and have extensive clinical research experience with kinematic MRI applications for the patellofemoral joint. 2
Muhle et al. evaluated the effect of a patellar realignment brace on patellar subluxation and dislocation using kinematic MRI and indicated that the brace produced no stabilizing effect. Because findings from at least one of our studies 3 disputes the recent report from Muhle et al., I feel compelled to comment on their article. Notably, several other investigations performed using kinematic MRI indicated that bracing can change the position of the patella relative to the femoral trochlear groove for patients with patellofemoral malalignment.2–4
Patient Selection. Muhle et al. reported that they did not select patients with patella alta and included five patients with dysplastic trochlear grooves. However, Figures 3 and 4 of their article show kinematic MRI images at 0° and 15° of flexion that clearly demonstrate that the patients had functional patella alta (that is, patella is positioned superiorly relative to the femoral trochlear groove). We reported that bracing typically fails to alter the position of the patella in patients with osseous dysplasia or patella alta. 2 Accordingly, the overwhelming lack of an effect produced by bracing may be partially the result of inappropriate patient selection by Muhle et al.
Technique. Muhle et al. used an active-motion, kinematic MRI technique with the patient in a supine position. This kinematic MRI method is not as sensitive as the active-movement, against-resistance technique for identifying and characterizing patellofemoral malalignment.1–4 At least two of our investigations performed in patients with and without bracing used active-movement, against-resistance kinematic MRI.2,3 These studies reported that bracing improved or centralized a malaligned patella in a high percentage of the patellofemoral joints that were evaluated.
The authors need to explain why they selected this less-sensitive kinematic MRI method for their study since they and others2–4 recognize the importance of performing kinematic MRI using an active-movement, against-resistance technique. Obviously, using a less-sensitive method to evaluate the effect of bracing on patellofemoral malalignment is problematic.
Supine Versus Prone Position. Muhle et al. criticized our method of studying patients in a prone position, suggesting that “free quadriceps muscle contraction is not possible.” This statement is made without any evidence offered by the authors that a difference in patellar alignment and tracking exists with the patient in a supine versus prone position. They merely presume this to be the case when, in fact, they have no data to support this contention.
As previously reported, 2 there is no difference between performing kinematic MRI in normal subjects and patients with regard to supine or prone positioning. This fact has been verified by using kinematic MRI, as well as by studies conducted in the biomechanics laboratory at the University of Southern California. Notably, we and others have used the prone positioning technique for many kinematic MRI investigations published in the peer-reviewed literature.
Worrell et al. 4 reported that bracing changes the position of the patella, as shown by incremental passive positioning kinematic MRI of the patellofemoral joint. They performed the kinematic MRI procedure with the patient in the supine position. A recent study using a dedicated extremity MR system with the patient in a seated position, undergoing an incremental kinematic procedure, demonstrated that bracing had a beneficial effect on patellofemoral malalignment. Again, this is additional objective data obtained using kinematic MRI that bracing works for certain patients.
Considering these points, I do not think that the difference between the findings of Muhle et al. and my experience and that of other investigators with bracing can be explained by the supine versus prone position used for the kinematic MRI procedure.
Patellofemoral Assessment Criteria. Muhle et al. used quantitative criteria for assessment of the patellofemoral relationships. By comparison, our studies primarily used qualitative criteria because of the inherent deficiencies of using techniques developed predominantly for static-view plain films of the patellofemoral joint.2,3 For example, we believe that it is inappropriate to attempt to measure patellar tilt angle, lateral patellar displacement, or other such techniques when the patella is articulating above the femoral trochlear groove (see Figures 3 and 4). How can the patellofemoral relationships be reliably assessed if the patella is in a far-superior position?
The fact remains that bracing patients with patellofemoral malalignment works for certain patients. One of the mechanisms responsible for a reduction in painful symptoms is the repositioning of the patella relative to the femoral trochlear groove, as shown objectively by kinematic MRI. Similar to other treatment regimens, the patient selection criteria for the application of braces is crucial to ensure beneficial results.
We do not agree with Dr. Shellock's opinion that the kinematic MRI technique used by our group is not as sensitive as the method described by his group in the diagnosis of patellofemoral tracking abnormalities. With regard to this investigation and our previous studies, our technique is sensitive enough to differentiate patients with patellofemoral maltracking from patients with physiologic patellofemoral alignment. In addition, as this study has shown, our technique was sensitive enough to demonstrate that bracing had no beneficial effect on the patellofemoral realignment in patients with patellofemoral maltracking. Therefore, we cannot see any points for an inappropriate imaging technique.
The assertion that the bracing studies of Worrell et al. demonstrated a beneficial effect on the patellofemoral alignment seems unconvincing. Those studies, performed by an incremental, passive positioning, do not reflect the influence of the extensor mechanism during active knee motion, which is known to be most important for the patellofemoral congruence. Therefore, kinematic MR imaging studies performed without active knee motion are inappropriate to evaluate the patellofemoral relationship.
We do not share the opinion that quantitative criteria for the assessment of the patellofemoral relationship is not reliable for the diagnosis of patients with patellofemoral malalignment, especially above the femoral trochlear groove in the early degrees of knee flexion. These criteria, which are based on previous studies performed by others, were optimized by our group for the evaluation of patellofemoral malalignment in kinematic MRI studies. Although this evaluation technique needs more time for the evaluation of the patellofemoral congruence compared with qualitative criteria, we do think that this method can be reliably used for the assessment of patients with patellofemoral maltracking (see previously published reports1–3).
In addition, we do not think that several different kinematic imaging techniques can be performed for examining the patellofemoral joint as previously described.2,3 All techniques described thus far do have inherent limitations by examining the patellofemoral joint either in prone or supine position under quadriceps contraction. However, both positioning schemes are at the moment only applicable with commercial MR scanners. We do think that, in the future, active patellofemoral joint studies under weightbearing condition in upright position, as recently reported by Bergman et al., 1 will fulfill all the important criteria as described by our group for a proper diagnosis of patellofemoral malalignment.
Dear Editor:
This letter is in regard to the case report by Crockett et al. that was published in the July/August 1999 issue (“Sacral Stress Fracture in an Elite College Basketball Player After the Use of a Jumping Machine,” pages 526–528). The authors state: “To our knowledge, sacral stress fractures have not been reported in young, healthy, male athletes.” I would suggest that a proper MedLine search would have drawn their attention to an article entitled “Stress Fractures of the Sacrum Following Strenuous Activity” (Clin Orthop 243: 184–188, 1989). This article summarizes several cases and includes CT scan visualization of the actual fracture.