Introduction

Without intervention, gait and function in patients with spastic diplegia may deteriorate with time, especially during the pubertal growth spurt [4, 14, 22]. In one study, deterioration in gait parameters occurred in all 18 subjects after a mean followup of only 32 months [14]. In another natural history study, temporal gait parameters and gait kinematics deteriorated in all 28 patients with cerebral palsy, including 19 with spastic diplegia, seven with hemiplegia, and two with quadriplegia [4].

Management of gait dysfunction in patients with cerebral palsy may include nonoperative measures, such as physiotherapy [45], orthotics [5], casting [19], and injections of botulinum toxin type A [17]. Despite appropriate nonoperative management, however, the majority of patients with spastic diplegia will have fixed contractures and bony deformities develop which require surgical correction to maintain gait or even sitting functions. To achieve sagittal plane balance and avoid repeated episodes of surgery and rehabilitation, multiple surgical interventions may be combined in one session, known as single-event multilevel surgery [7, 23]. This approach may be particularly important to improve gait in patients with spastic diplegia [11, 12, 37].

Single-event multilevel surgery was first described by Norlin and Tkaczuk in 1985 [23] and Browne and McManus in 1987 [7]. In these studies the concept of single-event multilevel surgery was introduced and the advantages of one-session surgery such as reduction in the number of operations and, possibly, reduced chance of reoperation and complications are mentioned. Norlin and Tkaczuk reported a 5-year followup of the original cohort in 1992 [24]. They reassessed 23 subjects by means of video-analysis rather than three-dimensional (3-D) gait analysis. In their cohort none of the patients had recurrences of the deformities and they had normal growth-related development of their gaits. However, these studies predated the widespread availability of instrumented gait analysis and the development of valid functional scales. Using a combination of gait parameters and the physiologic cost index, Nene et al. [21] in 1993 reported the function of 18 patients who had single-event multilevel surgery. Improvements in a predominance of gait parameters after multilevel surgery were reported for independent [36, 40, 47] and assisted walkers [18, 46]. Thomason et al. [42] published a pilot report of the first randomized clinical trial of single-event multilevel surgery. In a subsequent report of the same study [43], Thomason et al., reported statistically significant and clinically important improvements: at 12 months after surgery they noted absence of joint contractures, correct muscle lengths and lever arms, and gait kinematics. However, functional improvements, such as Gross Motor Function Measure (GMFM66), were not seen until 2 years after surgery. These improvements in gait and function were maintained at 5 years after surgery. The most comprehensive of recent studies reporting single-event multilevel surgery is a prospective, multicenter study with a control group, performed by Gorton et al. [10]. They included a group of outcome measures, including a summary statistic of gait (the Gillette gait index) [38], a measure of gross motor function (the Gross Motor Function Classification System or GMFCS) [26], a measure of health-related quality of life, and measurements using the Pediatric Outcomes Data Collection Instrument [10]. The authors showed on the basis of a matched concurrent data set of 75 patients with spastic cerebral palsy (age 4 to 18 years), an improvement in function after 1 year for a surgical group compared with a nonsurgical group.

Instrumented clinical gait analysis consists of various combinations of kinematic, kinetic, and EMG data and is interpreted with supporting clinical data. It provides essential biomechanical data for decision-making and objective outcome assessment [9]. The movement analysis profile and gait profile score [3] were developed to provide graphic and quantitative summaries of kinematic gait data, and a single summary statistic of gait. The individual movement analysis profile domains show which gait parameters have changed and by how much. The gait profile score provides a single measure to help determine if overall gait function has improved or deteriorated. The movement analysis profile domains highlight specific gait deviations but they have not replaced detailed data interpretations required for surgical decision-making [9]. The movement analysis profile describes the magnitude of deviation of nine kinematic variables over the gait cycle. The gait profile score [3] reduces all information to one number that reflects the quality of the gait pattern with respect to normality; it is a summary statistic of gait, which quantifies the subject’s gait abnormalities relative to normative data. A higher figure indicates a more abnormal gait pattern with respect to normality, and a lower value indicates an improvement toward normal gait function. Based on an analysis of the difference in median gait profile score for patients classified at different levels of the Functional Assessment Questionnaire [25], the minimal clinically important difference for the gait profile score is 1.6° [2].

The purposes of this study were to determine whether (1) single-event multilevel surgery improved gait in patients with spastic diplegia; and (2) whether the improved function is durable.

Patients and Methods

We retrospectively studied all 14 patients with cerebral palsy and severe gait dysfunction who underwent single-event multilevel surgery between September 2004 and June 2008. Inclusion criteria were a confirmed diagnosis of spastic diplegia, GMFCS Levels I, II, or III, and ages 6 to 18 years. Prior injection of botulinum toxin A was allowed, as long as 6 months had elapsed since the last injection. Eligible subjects had to have had a preoperative gait analysis and most had short-term (ie, 1 to 3 years; n = 13) and mid-term (ie, 4 to 6 years; n = 12) followup gait studies. Exclusion criteria were a diagnosis other than cerebral palsy, dystonic or mixed movement disorder, outside the age range, and GMFCS Levels IV and V. There were four females and 10 males included in the study. They had a mean age of 13 years (median, 12.5; range, 7–18 years) at the time of preoperative 3-D gait analysis. One patient had GMFCS Level I function, 10 had Level II function, and three had Level III function. One patient had previous surgical procedures (Table 1). The mean age of the patients at the time of single-event multilevel surgery was 14 years (median, 13 years; range, 8–18 years). No patients were lost to followup. The minimum followup was 1 year (mean, 1.8 years; range, 1–3 years). No patients were recalled specifically for this study; all data were obtained from medical records. The change in the gait profile score was 10°, and this change, in relation to the minimum clinically important difference, was seven times. All patients and/or parents gave written consent for participation in the study in accordance with local ethical committee requirements. The study was performed according to the declaration of Helsinki.

Table 1 Details of the 14 patients at the first 3-D gait analysis
Full size table

Three-dimensional data were collected preoperatively for all 14 patients. One patient (Patient 10) missed the first postoperative gait analysis leaving 13 patients with short-term data at a mean followup of 2 years (range, 1–3 years). For 12 patients followup data were collected with a mean followup of 5 years (range, 4–6 years). Two patients were followed not long enough to have mid-term followup gait studies (Patients 9 and 13). The examination in our gait laboratory was performed by a physiotherapist and a human movement scientist trained in gait analysis. The clinical assessment included examining passive ROM by using a goniometer, the spasticity assessment using the Modified Ashworth Scale Instructions [6] (scale, 0–4), and the manual muscle strength test [8, 20] (scale, 0–5) of the lower limb muscles. The instrumented gait analysis included kinematics, kinetics, and dynamic EMG data using a motion capture system (6-camera Vicon 460 system™, Oxford Metrics Ltd, Oxford, UK), two force plates (Kistler Instruments AG, Winterthur, Switzerland), and an eight-channel surface EMG system (Zebris®, Tübingen, Germany). The patients walked at their self-selected speed. The Helen Hayes marker set [15] was used, and at least six trials were recorded. Anthropometric data were collected for appropriate scaling. Surface EMG readings were recorded simultaneously. Bipolar Ag/AgCl surface electrode pairs (electrode diameter 10-mm and interelectrode spacing of 22-mm) were placed bilaterally over the medial gastrocnemius, tibialis anterior, rectus femoris, and semitendinosus muscles. For electrode placement the Surface ElectroMyoGraphy for the Non-Invasive Assessment of Muscles (SENIAM) [13] recommendations for surface EMG were followed. The ground electrode was placed over the tibial tuberosity. The EMG signals were band-pass filtered (10–700 Hz) and collected at a sampling rate of 2500 Hz. All data were expressed as a percentage of gait cycle, using the Polygon software (Vicon Polygon, Oxford Metrics Ltd, Oxford, UK). From the 3-D gait data, temporal parameters (cadence, stride length, and walking speed), the movement analysis profile, and gait profile score were calculated [3]. All patients underwent at least one preoperative and one postoperative gait analysis. All data were uploaded into GaitaBase (http://gaitabase.rch.org.au) [44]. The mean interval between preoperative gait analysis and surgery was 0.8 years.

All patients had preoperative botulinum toxin test injections before muscle lengthening surgery [33]. Surgical indications were based on comprehensive biomechanical and clinical assessments, including an instrumented gait analysis and clinical and radiologic evaluations. The indications for single-event multilevel surgery were: (1) documented deterioration in gait and function during the last 12 months, (2) fixed musculoskeletal deformities (contracture, torsion, joint instability), (3) evidence from the diagnostic matrix, including 3-D gait analysis, that deformity correction at two different anatomic levels (the hip, knee, or ankle) on both sides of the body had a realistic chance of improving gait and function (Appendix 1). The contraindications were: (1) severe weakness, (2) uncontrolled spasticity or dystonia, (3) progressive neurologic disorder (eg, hereditary spastic paraparesis), (4) a patient who was unable to perform postoperative rehabilitation because of cognitive, behavioral, geographic, financial, or social factors. For the purposes of this study, single-event multilevel surgery was defined as at least one surgical procedure that was performed on two different anatomic levels (hip, knee, or ankle) on both sides of the body. The surgical procedure did not need to be symmetric and was not uniform, but individually tailored to the patient’s needs. All operations were performed with the patient receiving general anesthesia. The mean number of orthopaedic procedures was seven (median, 6.5; range, 4–15 procedures). All operative procedures were performed or supervised by the junior (ER) or senior author (RB). Perioperative antibiotics were used and postoperative epidural infusions were administered to all the patients for pain control. All patients remained as inpatients for 5 to 10 days after surgery and were discharged wearing below-the-knee plaster casts (except in case of isolated tibialis posterior lengthenings) with knee immobilizers and use of appropriate assistive devices as indicated by their GMFCS level.

The patients were first assessed 6 weeks postoperatively with radiographs to check healing of the surgical incisions and consolidation of the osteotomies and bony procedures. Custom-fitted ankle-foot orthoses were provided for all patients with surgery at the ankle level. After 6 weeks all patients followed an intensive rehabilitation program performed three to five times per week for 12 weeks or 4 to 6 weeks as inpatients in a rehabilitation center. The aim of this postoperative rehabilitation was to improve ROM, strength, balance, and function. After this time all patients had semiannual clinic visits.

Adverse events related to surgery were classified as mild if they resolved spontaneously, moderate if they resolved completely after simple treatment, or severe if there was a permanent deficit [42].

Paired data were assessed using the paired t-test and sequential data were assessed using repeated ANOVA with Bonferroni post hoc analysis (SPSS® software, Version 15.0; SSPS Inc, Chicago, IL, USA). We determined differences in walking speed, cadence, and stride length between preoperative and the two followup times using t-tests. We determined differences in each movement analysis profile and the gait profile score between the preoperative and postoperative assessments using ANOVA.

Results

After single-event multilevel surgery there was an improvement in gait: the movement analysis profile for knee flexion, ankle dorsiflexion, and foot progression improved. The temporal parameters for baseline versus short-term postoperative gait (Table 2) are shown, and the temporal parameters for baseline versus mid-term postoperative gait were compared (Table 3). The parameters for short-term versus mid-term postoperative gait are shown (Table 4). The preoperative, short-term, and mid-term kinematics (Fig. 1), and movement analysis profile are shown (Fig. 2). The mean gait profile score improved from 20° to 13° (Table 2). The change was 7° and this change, in relation to the minimum clinically important difference, was five times. There was improvement in walking speed and stride length. The movement analysis profile for hip flexion, knee flexion, ankle dorsiflexion, and foot progression showed remarkable improvements (Table 3). The mean gait profile score improved from 21° to 11°. Nine patients (64%) had additional surgery after single-event multilevel surgery for further gait correction (Table 5) combined with plate removal (when indicated). New problems were found in five patients (Patient 1, hip flexion deformity; Patient 2, unilateral hip displacement; Patient 11, severe pes planovalgus; Patient 13, over-lengthening of the tibialis anterior tendon; Patient 14, short adductors) (Table 5), and one of the procedures needed revision because of relapses in four patients (Patients 3, 4, 7, and 8). Therefore no important changes were noted between short-term and mid-term followups (Table 4). The mean gait profile score at short-term was 13° and at mid-term 11°. The change in gait profile score was 2°, and this change, in relation to minimal clinically important difference, was two times.

Table 2 Temporal parameters for baseline versus short-term postoperative gait
Full size table
Table 3 Temporal parameters for baseline to mid-term postoperative gait
Full size table
Table 4 Comparison of short- and mid-term parameters
Full size table
Fig. 1
figure 1

These graphs illustrate the preoperative sagittal plane kinematics at the short-term and mid-term followups. The gray band is the normal reference range. The dotted line is the mean kinematic for the study group. The substantial improvements in the sagittal plane kinematic scan be seen; these are more clearly and quantitatively illustrated in Figure 2.

Full size image
Fig. 2
figure 2

The movement analysis profile and gait profile score are illustrated here at preoperative, short-term followup, and mid-term followups. Decreases in the movement analysis profile and gait profile scores are an improvement toward more normal gait.

Full size image
Table 5 Surgical details of all patients
Full size table

One patient (Patient 4) had a deep infection develop after tendoachilles lengthening. The deep infection rate was 1% for all interventions. No other complications occurred and this was the only adverse event.

Discussion

The advantages of single-event multilevel surgery for gait dysfunction in patients with spastic diplegia include less time spent in the hospital and the necessity of a single period of rehabilitation. This may result in cost savings, but we are not aware of any cost benefit analyses. The first reports of single-event multilevel surgeries had a narrow focus on safety and feasibility [7, 23, 24]. Efficacy has been established in studies [10, 18, 21, 32, 34–36, 40, 43, 46, 47] which have included gait data and functional outcomes. Most studies have lacked coherent methods to determine if specific procedures have been effective and the effectiveness of the overall single-event multilevel surgery and rehabilitation package [7, 11, 21, 24, 36, 37]. The aim of our study was to investigate (1) whether single-event multilevel surgery improved gait in patients with spastic diplegia, and (2) whether the improved function is durable. A previous study by Norlin and Tkaczuk assessed 23 subjects at the 5-year followup to determine the long-term success of single-event surgery [24]. However, their study only included video recordings as a means of assessment, whereas our results are based on gait analysis. After 5 years they did not find any recurrences. In contrast, in our study 2/3 of the patients required additional surgery because of relapses or new biomechanical problems.

Our study has numerous limitations. First, we had a small sample size of 14 patients. The indications for single-event multilevel surgery are relatively uncommon; we did include all patients who had this sort of surgery. Second, we lacked control groups of untreated patients and patients treated by nonoperative methods such as orthotics and botulinum toxin injections. Third, we had incomplete followups for some patients because one patient missed the first postoperative gait analysis and two patients did not yet have the final examination. Fourth, our findings are likely relevant only to a specific group of patients with spastic diplegia for which we established specific surgical indications. Finally, we had no measure of function. However, we note most single-event multilevel surgery outcome studies have a retrospective study design and the evidence base is limited by only one published randomized clinical trial at the moment [1, 11, 12, 30, 33, 37, 42, 43].

We found severe gait dysfunction in patients with spastic diplegia can be improved for the short-term in one operative session by single-event multilevel surgery. These findings are in accordance with reported findings [10, 11, 40, 42]. In our study the preoperative gait profile score of 20° was 5.1° higher than reported by Thomason et al. [43]. References to the kinematic graphs and surgical prescription show the procedures for correction of excessive knee flexion and equinus contributed most to the overall surgical outcomes (Fig. 1). Changes in other movement analysis profile domains were made toward improved gait except for pelvic tilt, which increased only slightly (Fig. 2). Increased anterior tilt was a negative outcome of one study of single-event multilevel surgery, and more work is required to further reduce this problem [18]. It also may be that study was underpowered and could not detect noteworthy changes in some of the movement analysis profile domains. The mean age (13.6 years) of the patients at the time of the single-event multilevel surgery in our cohort was older than patients in other studies (mean, 6.4 [12], 12.0 [16], and 8.5 years [26]). The mean interval of our study (0.8 years) between the first gait analysis and surgery was comparable to the interval reported by Gough et al. (0.7 years) [12]. The surgical complication rate of 1.1% in our series was lower than that reported by Rodda et al. (10 surgical complications in four of 10 patients) [28].

The need for additional procedures after single-event multilevel surgery for treatment of spastic diplegia was first reported by Aiona and Sussman [1, 41]. After short-term followup in our gait laboratory, nine patients had additional surgeries after single-event multilevel surgery for further gait corrections (Table 5), combined with plate removal (when indicated). In our study, additional surgery was indicated for 2/3 of the patients because of relapses (n = 4) or new problems (n = 5). This number of additional surgeries is similar to that of Thomason et al. [43], who reported 22 subsequent surgeries were needed in 12 patients. A reduction in gait profile score from 20.7° to 11.1° at mid-term followup indicated 47% improvement toward normal gait; major improvements in the movement analysis profile for hip flexion, knee flexion, ankle dorsiflexion, and foot progression contributed to this. It probably is unrealistic to expect a single episode of surgery to provide lasting correction of all musculoskeletal problems and enhance and preserve gait function in all patients with spastic diplegia. It also may not be appropriate to focus on a single-event paradigm in all subjects. Staging surgery (and offering revision surgery when necessary) in this study achieved 47% improvement toward normal gait, but to preserve the early improvements 2/3 of our patients required additional surgery. In our study, the gait profile score was used to measure the overall gait disorder. The median value of the gait profile score for healthy patients is 5.2°. The improvement in our patients represents reductions of 47% (from preoperative to short-term), 62% (from preoperative to mid-term), and 31% (from short-term to mid-term) in the difference between abnormal and healthy gait patterns. The gait profile score [3] can be calculated independently of the feature analysis on which the gait deviation index is based [3, 39]. The movement analysis profile summarizes much of the complex information contained in the kinematic data and shows which parameters contributed to the change of gait. Measuring the outcomes of surgery in patients with cerebral palsy is difficult. Other factors such as the motivation of each patient and their abilities to participate in the postoperative rehabilitation program are likely to be important, although these are more difficult to quantify.

Single-event multilevel surgery, including lengthening of the contracted agonist, antagonist shortening [29], and selected bony procedures with stable fixation [30–32], improves gait dysfunction in patients with spastic diplegia. To preserve early improvements however, 2/3 of our patients required additional surgery because of relapses or new problems. It is unknown whether the improvements will last for decades.