CLINICAL ORTHOPAEDICS AND RELATED RESEARCH
Number 301, pp. 10-18
Copyright 1994 J. B. Lippincott Company
Complications of Limb Lengthening
A Learning Curve
MARK T. DAHL, M.D.,* BENJAMIN GULLI, M.D.,** AND TROY BERG, M.D.**
Major complication rates during limb lengthening were plotted in a consecutive series to produce a learning curve. All unwanted events during and after treatment were considered complications, and graded as minor, serious, and severe. All serious and severe complications were considered major. A novel system was used to classify the preoperative severity of each deformity. One-hundred ten patients had 140 bone segments lengthened between 2.2 cm and 10.5 cm, with a mean of 4.4 cm. Three methods were used in lengthening: the Wagner method in 22 patients, the DeBastiani method in 34 patients, and the 11izarov method in 84 patients. Ninety-eight complications categorized as serious or severe occurred, for a total major complication rate of 72%. The percentage of major complications began to drop after 30 lengthenings to a current rate of 25%. Major complications were frequent in patients with more severe deformities, particularly in those whose cases occurred early in the series. Bone healing complications were high (72%) in the Wagner segments but were also high (80%) in the first ten patients treated with the DeBastiani technique. The first ten llizarov patients, who were treated later in the series, had a 40% rate of bone-healing complications. The current rate of major complications is 13% for those patients treated with DeBastiani's method and 33% for those patients treated with llizarov's method. This difference in complication rates appears to relate to the severity of the deformity, rather than the device used. There was a significant decrease in complications as experience was gained. Directed formal study and surgical instruction should help diminish these complications.
*From Limb Length Clinics Gillette Children~s Hospital, ST Paul, Minnesota, Fairview Riverside Medical Center, Minneapolis, Minnesota, and Shriners Hospital for Crippled Children, Minneapolis, Minnesota.
** University of Minnesota, Hospitals and Clinics, Minneapolis.
Address reprint requests to Mark T Dahl, M.D., Limb Length Clinic. Gillette Childrenís Hospital, 200 E university Ave., St. Paul, MN 55101.
Distraction osteogenesis has improved the ability to treat complex deformities by reducing the need for secondary bone grafts, decreasing the rate of late fracture, and combining the correction of multiple deformities in one treatment. (1,3,5,6) Soft-tissue complications and, to a lesser extent, bone problems continue to be prevalent in clinical reports of limb lengthening. The actual process of limb lengthening is very complex and has one of the highest complication rates of any orthopaedic procedure.(3,8,10)
The complication rate in limb-length surgery can vary considerably between investigators. DeBastiani et al.(3) report complications in 14% of their patients with lengthening, while Wagner(10) denotes 45% and Ilizarov 5%.(5) Mosley and Mosca (7) included all complications, no matter how minor, in their study of Wagner leg lengthenings. In 63 lengthenings, 142 complications were reported, for a rate of 225%. This variation in percentages is attributable to differences in techniques of lengthening, surgical experience, severity of deformity treated, and definition of complications.
To provide a frame of reference, the authors first reviewed the severity and frequency of complications in limb lengthenings as they related to the surgeon's experience, technique used, and severity of deformity. They set out to answer three questions: (1) How does a surgeon's complication rate change with experience? (2) How many cases were performed before complication rates diminished? (3) To what extent does the severity of a deformity influence the complication rate?
A composite learning curve and a deformity severity classification scheme are presented. The authors will discuss conclusions and recommendations to improve the learning curve for other surgeons before they consider caring for these complex limb deformities.
MATERIALS AND METHODS
A retrospective study and radiographic review were performed on 140 consecutive limb-lengthening cases. Data were obtained from the clinics of the senior author. All lengthening patients with greater than one year of follow-up time were included. Deformity corrections without lengthenings were excluded.
SURGICAL TECHNIQUES
The authors used lengthening methods originally described by Wagner, (10) DeBastiani et al., (3) and llizarov. (5,6) Lengthening began three to five days after the corticotomy in children, and five to seven days in adults. The Wagner technique was used in 22 patients. This technique uses a half-pin external fixation device (monolateral frame) on the intact bone that is to be lengthened. An open osteotomy is performed with immediate distraction of approximately 1.5 mm. The bone is lengthened 1.5 mm per day in graduated stages. At the completion of the distraction, the osteotomy is internally fixed with a special lengthening plate, and the lengthening gap is filled with autologous, cancellous bone. The patient is protected from full weight bearing for an extended period, allowing the bone to heal by creeping substitution. After a long period of delay (often one to three years), and when radiographic evidence of recanalization of the bone has occurred, the bone plate is removed.
The DeBastiani technique, used in 34 cases, involves applying a dynamic monolateral external fixator. An open drill-corticotomy through metaphyseal bone is performed, followed by a delay of seven to 14 days before lengthening begins. The bone segment is lengthened 1 mm per day, in four increments of one-quarter millimeter. During the lengthening phase, functional use of the limb is encouraged. The rate of lengthening is altered according to the extent of bone formation. After completion of the lengthening, the fixator is held in a neutral phase before a subsequent dynamization is performed, to allow maturation of the regenerate site. Fixators are removed on an outpatient basis after thorough corticalization is evident.
The llizarov method, applied in 84 cases, is a multiplanar system that uses circular external fixation with tensioned wires or a combination of half-pins. These fixators provide for multilevel, multiapical, and juxtaarticular corrections. Postoperative management features are similar to DeBastiani's method.
It should be understood that the method of tissue regeneration, either bone grafting or distraction osteogenesis, is not exclusive to a device, as many surgeons use different devices to achieve the same result. Methods using distraction histogenesis (i.e., llizarov and DeBastiani) have met great enthusiasm; however, frequent pitfalls exist that can cause enormous problems for patients and families. Careful preoperative planning, preassembly of the external fixator, and diligent postoperative care are required for success with any technique.
DATA ANALYSIS
For the primary analysis, the following data were recorded for all patients: age; gender; segment lengthened; length discrepancy; preoperative segment length; length achieved; mechanical axis deviation greater than 5i; etiology of deformity; hospital stay; length of time in the fixator; length of follow-up; and all complications. Measurements of length and angular deformity were made from scanograms and standing radiographs. Each case was retained in chronologic order and was also categorized by technique.
COMPLICATION CLASSIFICATION
Any unwanted event was considered a complication (Table 1). The severity of a complication was assigned a grade of minor (I), serious (II), or severe (III). Serious and severe complications were considered major because of their associated morbidity. Minor complications did not affect outcome or require extensive intervention; therefore, although they are important, they will not be emphasized in this paper. Complications that were either major and temporary, or minor and permanent, were considered to be serious. Severe complications are those that require major unplanned surgery or result in major permanent sequelae.
DEFORMITY CLASSIFICATION
A deformity severity scale was developed to correlate complication rates with the complexity of the condition. The severity of the deformity was rated according to the initial length discrepancy: Type I, < 15%; and Type 2, 16-25%; Type 3, 26-35%; Type 4, 36-50%; and Type 5, >50%. The severity type increased one level if three lesser risk factors, or two greater risk factors, were present in addition to the discrepancy (Table 2). Lesser risk factors add to the complexity of treatment but, with proper planning, usually do not compromise the end result. Greater risk factors significantly alter treatment plans, and can seriously compromise the end results (Figs. IA-IE).
TABLE 1. Complication Classification
| Complication | Minor (I) | Serious (II) | Severe (III) |
| Pin-site problems | Minor infections | Ring sequestrum | |
| Infection | Superficial Wound | Deep Wound | Osteomyelitis |
| Vascular | | | Vascular laceration or occlusion requiring repair |
| Neurologic | Hypesthesia | Neuropraxia | |
| Medical | | e.g., DVT, pneumonia | e.g., Cardiac Arrest |
| Psychological | | | Requires change in treatment |
| Premature consolidation | | Requires repeat corticotomy | |
| Delayed union/ nonunion | | LI > 2 /adult or >1.5 / child | Bone graft or retreatment necessary |
| Fracture | | Repeat fixation | Osteotomy |
| Axis deviation >5° | | 6° - 10° | >10° |
| Subluxation | | Temporary | Permanent |
| Contracture | <10° | 11°-20° | >20° and/or gait disturbance |
| Did not equalize | <2.5 cm | 2.5-5.0 cm | >5.0 cm |
LEARNING CURVES
Complication rates were defined to be the number of complications per lengthened segment. To simplify data presentation, these rates were averaged for every five consecutive lengthenings and plotted on a graph. Lines were drawn through consecutive points on the graph to produce the learning curves.
RESULTS
One-hundred forty segments were lengthened among the 110 patients in the study population. Eight segments were in the upper extremities; the remainder were in the lower extremities. Primary lengthenings were performed in 104 segments, 26 were second lengthenings, and ten were third. Sixty-four patients were male and 46 were female. Their average age was 14 years (range, three to 49 years). The average amount of length achieved was 4.4 cm (range, 2.2-10.5 cm). The average lengthening index was 1.0 months/cm for children (<18 years old) and 1.8 months/cm for adults. The initial 18 segments in the series were lengthened by Wagner's method. Surgical use of the Wagner method was discontinued in 1986, when use of the DeBastiani method began. The llizarov method was first used on the 39th segment in the series. Both the DeBastiani and llizarov devices were used in the remainder of the segments. The overall complication rate, total complications divided by segments lengthened, was 182%.
TABLE 2. Complication Risk Factors
| Deformity Length % | Type 1 <15 | Type 2 16-25 | Type 3 26-35 | Type 4 36-50 | Type5 >50 |
| Lesser Factors | Greater Factors |
| Angulation | Congenital deformity |
| Translation | Multisite deformity |
| Rotation | Multiple surgeries |
| Contracture | Previous lengthening |
| Prior infection | Nonunion |
| Anatomic location (femur, forearm, or foot) | Bone loss |
| Age (adult) | Active infection |
| Obesity | Preoperative instability |
| Poor nutrition | | | |
| Neurologic deficit | | | |
FIG. 1A- 1E. Examples of the five deformity severity types. (A) Type 1: length discrepancy only. Normal joint above and below. (B) Type 2: posttraumatic length deformity with knee DJD and limited motion. (C) Type 3: posttraumatic deformity with a 5-cm (14%) length discrepancy, and three lesser and two greater risk factors. (D) Type 4: congenital pseudarthrosis. Length discrepancy of 7 cm ( 19%) with three lesser and three greater risk factors. (E) Type 5: severe femoral bone loss from osteomyelitis treated for nonunion only, no lengthening. Length discrepancy of 30 cm (>50%), with two lesser and three greater risk factors.
TABLE 3. Wire and Pin Site Classification and Treatment
| Grade | Appearance | Treatment |
| 0 | Normal | Weekly pin care |
| 1 | Inflamed | Daily pin care |
| 2 | Serous discharge | Antibiotics |
| 3 | Purulent discharge | Antibiotics |
| 4 | Osteolysis | Remove pin |
| 5 | Ring sequestrum | Debridement |
INCIDENCE OF MINOR COMPLICATIONS
Wire and pin-site complications (inflammation, infection, loosening, or metal failure) were poorly recorded before 1989, at which time a clinical grading system was initiated (Table 3). Nearly all patients experience wire or pin infections. However, the occurrence of loosening, fracture, and ring sequestrum dropped from 10% in 1989 to a current rate of less than 5%. The exact incidence, classification, and treatment for these frequent complications is the subject of another paper. (2)
Wire and pin complications were rare in upper-extremity lengthenings. A minor loss of joint motion (<10i) occurred in 25(18%) of 140 lengthenings.
INCIDENCE OF MAJOR COMPLICATIONS
The overall major complication rate for the series was 72%. By the last year of the study, this rate had fallen to 25%; 13% for DeBastiani cases and 33% for 11izarov cases. The learning curve for major complications is shown in Figure 2. The early peak is attributable to an increase in complications when the DeBastiani method was substituted for Wagner's. Inexperience in judging the regenerate bone formation and timing of fixator removal resulted in several fractures. There was also a disproportionately high number of Type 3 and Type 4 deformities in these early DeBastiani lengthenings.
FIG. 2. The rate of major (Grades 11 and 111) complications as a function of the total number of lengthenings performed.  |
FIG. 3. The learning curves for each technique arranged temporally. |
Figure 3 shows the learning curves for each technique arranged temporally. The first segments lengthened by llizarov's method were done after experience had been gained with Wagner's and DeBastiani's methods (note the relatively low complication rate in the early llizarov patients). It is important to emphasize that tutorials and visits to experienced limb-length centers were completed before beginning the llizarov method.
The major complication rate for llizarov cases at study completion was 33%. For DeBastiani cases, major complications occurred in 13% of the patient population at study completion. When corrected for severity of deformity, the complication rates of the two devices are similar. The final severe (Grade III) complication rate approaches 10% for either technique, as illustrated in Figure 4. As experience increased, there was a significant decrease in the complication rate, falling from 55% for the first 60 cases to 12% for the subsequent 78 cases (p < 0.001). The p-values noted throughout the paper represent the observed significance levels of a one-tailed t-test.
FIG. 4. The severe complication rate versus case number. |
The correlation of deformity severity with complication rates is presented in Figure 5. The learning curve downslope in Type 3 deformities occurs between 30 and 40 cases, whereas in Type 2 deformities, the downslope took place between ten and 30 cases. Curves for Type 1 and Type 4 deformities are less meaningful because of the small number of patients in these groups.
FIG. 5. The major complication rates for deformity severity Types 2 and 3. Curves for Types 1 and 4 deformities are not shown since these groups contain too few patients to draw meaningful curves. |
Analysis on these two patient groups showed a significant drop in the complication rate as experienced was gained. The average major complication rate for the first 15 patients with Type 2 deformities was 40%. The average complication rate in the subsequent 40 patients with Type 2 deformities was 5%. Similarly, the average major complication rate for the first 40 patients with Type 3 deformities was 68%, whereas this rate dropped to 14% in the next 22 patients.
The final major (Grade II and III) complication rate in the series is approximately 80% for Type 4 deformities, 40% for Type 3 deformities, and 20% for Type 2 and for Type 1 deformities (Fig. 6). This comparison for major complication rates at study completion is significant (p < 0.01).
HIGH-RISK CIRCUMSTANCES
Major complication rates were particularly high and severe in the following instances: simultaneous lengthenings of congenital etiology; congenitally short femurs; and lengthenings where the percentage lengthened was greater then 15%.
Thirteen femoral and tibial segments were simultaneously lengthened. The results demonstrated that patients with a congenital etiology had more than twice the major complication rate than those with an acquired etiology. Complications including axis deviation, late fracture, and knee subluxation occurred most frequently in the congenitally short femoral group. Complications occurred twice as often in lower-extremity lengthenings greater than 15%. Surprisingly, relengthened segments, if less than 15%, did not have a higher complication rate.
DISCUSSION
Limb-length surgery is exceptionally challenging. The high complication rate early in the authors' series was sobering. A knowledge of the factors that predispose to complications is essential for their avoidance, when possible. Through dedicated study, careful review of the results, meticulous techniques, and the resources of a multidisciplinary team, the authors were able to bring their results into line with those of other published series.
FIG. 6. Final major complication rates for each deformity severity type. |
The definition of complication has been a subject of controversy in the limb-length literature. Paley (8) defines "true complications" as those adverse results that persist after treatment is completed. Ilizarov (6) defines a complication to be any adverse or unexpected condition or effect that alters the care plan or reduces the quality of results. Pouliquen (9) describes complications as benign, serious, or severe. In an attempt to clarify and standardize the way complications are reported, the authors use a classification scheme based on the severity and permanence of all unwanted events. They recognize that the complexity, permanence, and salvage of complications occur in a spectrum that yields poorly to a classification scheme. Nonetheless, these limitations (all complications of limb lengthening) are acknowledged for the purpose of documentation, physician education, and informed consent.
Many deformity features can variably affect the difficult process of limb lengthening. Some of these features are known to increase the risk of complications. The femur is considered more difficult to lengthen than the tibia. Lengthening a bone by 30% carries greater risk than a lengthening of 10%. In an effort to correlate the authors' complication rates with features of deformities, a classification for severity of limb deformities was developed (Table 2). The most easily measured feature, length discrepancy, is the basis of this classification. The greater or lesser risk factors were chosen based on the experience of the authors, as well as of others. The validity of this classification is substantiated by the trends in complication rates shown in Figures 5 and 6. The Type 2 deformity complication rate decreased more quickly than the Type 3 rate. Figure 6 further illustrates that the less severe the deformity, the lower the complication rate.
Limb lengthening is a very complex process with numerous pitfalls. Good judgment, accurate knowledge, meticulous technique, and relentless follow-up care are necessary to select and design devices, perform corticotomies, maximize bone regeneration, manage pin sites, maintain articular function, time fixator removal, and manage after care. Instruction in the fine points of these techniques was not available in North America until 1988 and remains limited today. Thus, the senior author, as well as many other North American limb-length surgeons are, to varying extents, self-taught.
Learning curves have the general shape of a peak, followed by a downslope, leveling out to a plateau. The peak occurs as a new technique is introduced and reflects the high rate of complications associated with inexperience. The downslope represents the experience gained as the surgeon becomes more proficient. The plateau represents maintenance of proficiency by continued practice.(4) The learning curves presented here suggest that for complex (Types 3 and 4) deformities a surgeon must perform 40 to 60 lengthenings before lower complication rates are achieved. However, for the less complex (Types 1 and 2) deformities, safety is approached after 20 lengthenings.
To demonstrate how the incidence of complications decreases with respect to experience, a comparison was made between the total number of complications per lengthening and the average number of complications. In the first 60 cases, there were 3.23 complications per lengthening. For the next 78 cases, this value was 1.81. This decrease in the average number of complications per lengthening is significant (p < 0.001). This further confirms the authorsí impression, gained from the shape of the learning curve, that the incidence of complications diminishes significantly after approximately 60 cases.
The current enthusiasm for distraction osteogenesis, combined with heavy marketing of limb-lengthening instrumentation, has encouraged more surgeons to engage in this demanding field. Most of these surgeons are not formally trained in limb-lengthening techniques, and they treat relatively few cases. As illustrated by the authorsí learning curves, this situation is likely to result in a high level of complications. It is the authorsí hope that a properly educated surgeon would be able to enter the learning curve at a more optimal point.
Attributing declining complication rates to a specific device or method is difficult because the biologic and mechanical principles continue to change and, in fact, transcend the device used. After reviewing their initial case results with limb lengthening, the authors identified several factors that contributed to the higher complication rates. The experience gained from these lengthenings has prompted several recommendations for diminishing complications. First, unless surgeons plan to dedicate a significant portion of their practice to limb lengthening, they should refer their patients to centers that specialize in such treatment. Second, if an interest in limb lengthening exists, advanced individual study should be pursued. Third, the novice limb-length surgeon should start with less complex deformities, such as acquired length discrepancies with normal joints. Progression to Type 3 and Type 4 deformities with severe length discrepancies, abnormal joints, and multiple risk factors, should be delayed until competence is achieved in the more basic aspects of limb lengthening. Above all, a cautious approach is recommended to these new techniques, allowing carefully analyzed long-term follow-up evaluation to determine their place in reconstructive surgery.
SUMMARY
The high frequency of major complications gradually diminished after approximately 60 lengthenings. Complications persist in 25% of the patients even after the authors gained substantial experience. The severity of the preexisting deformity also influenced the rate and extent of these complications.
ACKNOWLEDGMENTS
The authors thank Drs. Steven E. Koop, Gillette Children's Hospital; Lyle O. Johnson, Shriners Hospital for Crippled Children; and David A. Fisher, Riverside Medical Center, for their contributions in the care of these patients.
REFERENCES
1. Dahl, M. T., and Fischer, D. A.: Lower extremity lengthening by Wagner's method and by callus distraction. Orthop. Clin. North Am. 22:4, 1991.
2. Dahl, M. T., and Moore, D. P.: Tibial llizarov wire site sequelae, a prospective bacteriologic study. (In Press.)
3. DeBastiani, G., Aldegheri, R., Renzi-Brivio, L., and Trivella, G.: Limb lengthening by callus distraction (Callotasis). J. Pediatr. Orthop. 7:129, 1987.
4. Hughes, G. B.: The learning curve in stapes surgery. Laryngoscope 101:1280, 1991.
5. Ilizarov, G. A.: The principles of the llizarov method. Bulletin of the Hospital for Joint Diseases Orthopaedic Institute 48: 1, 1988.
6. Ilizarov, G. A.: Clinical application of the tensionstress effect for limb lengthening. Clin. Orthop. 250:8, 1990.
7. Mosley, C., and Mosca, V.: Complications of Wagner leg lengthening. Behavior of the Growth Plate. New York, Raven Press, 1988.
8. Paley, D.: Problems, obstacles, and complications of limb lengthening by the llizarov technique. Clin. Orthop. 250:81. 1990.
9. Pouliquen, J. C.: Personal communication. 1992.
10. Wagner, H.: Operative lengthening of the femur. Clin. Orthop. 136:125, 1978.
Back to Index
Bản in
