TO THE EDITOR:
We congratulate Noonan et al. for the depth of analysis in their article "Distraction Osteogenesis of the Lower Extremity with Use of Monolateral External Fixation. A Study of Two Hundred and Sixty-one Femora and Tibiae" (80-A: 793—806, June 1998). However, we take exception to the statement: "Reviews of the results in larger groups of patients from Europe … or the Soviet Union … have lacked sufficient detail, making it difficult to draw any valid conclusions from the data. Finally, most studies have combined the results of distraction osteogenesis of the tibia and the femur."
This statement is, at best, misleading and probably reflects the fact that Noonan et al. did not perform an in-depth search of the literature. Three years ago, we used an approach that was remarkably similar to theirs and performed a review of 281 lower-limb lengthenings in 240 patients who were managed for congenital, posttraumatic, or postinfective limb-length discrepancy at three European centers (London, England, and Naples and Rome, Italy)4. We retrospectively reviewed the hospital charts and radiographs of patients who had completed lengthening at least twelve months before the beginning of the study. The variables that were studied included age at the time of the operation, the bone segment lengthened, whether a corticotomy or an osteotomy had been performed, the amount of lengthening planned and achieved, the level or levels of the corticotomy or osteotomy, the type of external fixator used, and the difficulties encountered.
The limb-length discrepancy was reduced to within two centimeters (1.6 percent or less) of the length of the normal, contralateral limb after 249 lengthenings (89 percent) in 208 patients. The average time that the fixator was in situ was 186 days (range, ninety-four to 617 days), with an average healing index of thirty-five days per centimeter (range, twenty-six to forty-three days per centimeter). The femoral osteotomies and corticotomies healed faster than the tibial osteotomies and corticotomies. The average healing indices for the posttraumatic and postinfective limb-length discrepancies were significantly less than that for the congenital conditions (analysis of variance, p < 0.0001). The younger patients had a significantly shorter time to bone-healing and fewer complications (analysis of variance, p < 0.001). There were significantly more complications among patients in whom the lengthening exceeded 18 percent of the original length of the bone (chi-square test, p < 0.01). Bifocal lengthenings healed significantly faster than single-level ones (analysis of variance, p < 0.0001). The three external fixators used (Orthofix, Ilizarov, and Monticelli-Spinelli) were equally effective for lengthenings of less than 20 percent. The Ilizarov and the Monticelli-Spinelli circular fixators were associated with significantly fewer complications when lengthenings of more than 20 percent were performed (chi-square test for trend, p < 0.01).
Admittedly, we used three fixators (one unilateral and two circular) in our patients4. However, Noonan et al. used three different unilateral fixators (Wagner, Orthofix, and Monotube). Also, all of our patients had a limb-length discrepancy. In the study by Noonan et al., some patients had a limb-length discrepancy whereas others had congenital short stature.
It is remarkable that the statistical approach used in their study was similar to that used in ours4. We calculated descriptive statistics. Groups were compared with one or two-way analysis of variance with use of the underlying pathology, the age-range of the patients, and the apparatus as a covariate. Cross-tabulation, breakdown, regression analysis, and Pearson product-moment correlation were performed. A general linear-regression model was used, although we acknowledge that some of the variables may be correlated nonlinearly with the duration that the fixator was in situ.
Noonan et al. went one step farther and also used multivariate analysis. They should be congratulated for the quantitative approach that they used in this study and in other studies1. However, if they were looking for a quantitative interpretation of data on distraction osteogenesis from Europe, we believe that we provided that in 19964.
We fully agree that the results of femoral and tibial lengthenings should be reported separately, and we have done so for the last ten years2,3.
Finally, we caution against combining the data for patients who have short stature with those for patients who have limb-length discrepancy. The ability of the bone to regenerate is intrinsically different in these two groups5, and reporting the data together can cause spurious assumptions to be made. Indeed, in our study4, we did not include any patients who were managed for short stature.
Nicola Maffulli, M.D., M.S., Ph.D., F.R.C.S.(Orth): Department of Orthopaedic Surgery, University of Aberdeen Medical School, Polwarth Building, Foresterhill, Aberdeen AB25 2ZD, Scotland
John A. Fixsen, M.Ch., F.R.C.S.: Department of Orthopaedic Surgery, The Hospital for Sick Children, Great Ormond Street, London WC1N 3JH, England
Dr. Noonan, Dr. Leyes, Dr. Forriol, and Dr. Cañadell reply:
We thank Dr. Maffulli and Dr. Fixsen for their comments regarding our manuscript.
We agree that our failure to reference the study by Maffulli et al.4 was due to an inadequate review of the literature; however, we disagree that our statement, which was quoted in their letter, was "at best, misleading." Careful review of their study demonstrates differences for which the quoted statement is still applicable.
In their multicenter, retrospective analysis4, Maffulli et al. reported the results of distraction osteogenesis in 281 limbs that were lengthened for unilateral discrepancy due to congenital causes or growth arrest resulting from trauma or infection. The authors should be congratulated for decreasing the discrepancy to less than two centimeters in 89 percent of the limbs in patients who were three to fifty-three years old. However, one could argue that any statistical analysis of their data would be difficult because of the inherent variability in technique between different centers. For example, the authors used different osteotomy techniques; ring as well as monolateral external fixators were used; screws as well as fine wires were used; and monofocal lengthenings, bifocal lengthenings, and even ipsilateral lengthenings were performed. These variables clearly affect the outcome and they were eliminated in our report, in which all of the procedures were performed at one institution with use of a standard operative technique involving monolateral fixation with screws and lengthening at only one osteotomy site in each extremity. All of the procedures were performed in younger patients, who ranged in age from four to twenty-eight years.
We agree that the statistical analysis performed by Maffulli et al.4 is more detailed than those described in previous reports; however, there are major differences between our study and theirs. Maffulli et al. analyzed differences between femora and tibiae with regard to the rate of healing but not with regard to the rates of complications or additional procedures. They also did not attempt to analyze any differences in the prevalences of complications or additional operations according to different demographic variables. Most importantly, their study did not take into account the effect of length gained on the rate of complications or the healing index. Our study attempted to normalize the rate of complications to the percentage of length gained, and all of the analysis was done with this in mind. Additionally, we point out that use of the healing index as an outcome parameter is flawed unless the effect of the length gained is removed by statistical means. This was not done in their study and, as such, the analysis based on the healing index is invalid. Careful review of our manuscript demonstrates that statistical analysis was performed between different groups on the basis of diagnosis and that the results and analysis were not combined.
In summary, we believe that our "quantitative interpretation" adds a great deal to the available knowledge base on distraction osteogenesis. In our report on limb-lengthening in a diverse population of patients, we sought to analyze differences in the rates of complications and healing according to different variables (age, site of the osteotomy, and etiology), independent of length gained, for femora and tibiae separately. We acknowledge the contribution of Maffulli et al.4, and we believe that their study is important as it seeks to review the outcome in patients who had lengthening for limb-length discrepancy. The authors are justified in their observation that lengthenings for congenital limb-length discrepancy are more difficult than those for discrepancy following growth arrest due to infection or trauma. Because of the complementary information presented in both papers, we are sure that Dr. Maffulli and Dr. Fixsen would agree that both manuscripts provide valuable information.
Kenneth J. Noonan, M.D.: Department of Orthopaedic Surgery, Indiana University, 541 Clinical Drive, Room 600, Indianapolis, Indiana 46202-5111
Manuel Leyes, M.D.; Francisco Forriol, M.D.; Jose Cañadell, M.D.: Orthopaedic Research Laboratory, Department of Orthopaedic Surgery, Clinica Universitaria, School of Medicine, University of Navarra, Pamplona 31080, Spain