Impact of gait analysis on correction of excessive hip internal rotation in ambulatory children with cerebral palsy: a randomized controlled trial


Correspondence to Tishya A L Wren, Children's Hospital Los Angeles, 4650 Sunset Blvd., #69, Los Angeles, CA 90027, USA. E-mail:



The aim of this study was to determine if gait analysis improves correction of excessive hip internal rotation in ambulatory children with spastic cerebral palsy (CP).


Children undergoing orthopedic surgery were randomized to receive or not receive a preoperative gait analysis report. This secondary analysis included all participants whose gait report recommended external femoral derotation osteotomy (FDRO). One-year postoperative, and pre- to postoperative change in femoral anteversion, mean hip rotation in stance, and mean foot progression in stance were compared between groups and in subgroups based on whether the recommendation for FDRO was followed.


Outcomes did not differ between the group which received a gait report (n=39; 19 males, 20 females; mean age 10y 4mo [SD 3y]; hemiplegia, 3; di/triplegia, 28; quadriplegia, 8; Gross Motor Function Classification System [GMFCS]: level I, 5; level II, 12; level III 19; level IV, 3) and the control group (n=26; 14 males, 12 females; mean age 9y 5mo [SD 2y 10mo]; hemiplegia, 1; di/triplegia, 21; quadriplegia, 4; GMFCS: level I, 4; level II, 1; level III, 9; level IV, 2; all p values >0.29), but improved more in the gait report subgroup in which the FDRO recommendation was followed (seven limbs; change in anteversion −32.9°, hip rotation −25.5°, foot progression −36.2°) than in the control group (anteversion −12.2°, hip rotation −7.6°, foot progression −12.4°; all p values ≤0.02) and the gait report subgroup in which FDRO was not performed (32 limbs; anteversion −1.0°, hip rotation 0.5°, foot progression −8.0°; all p values ≤0.003). Postoperative measures became normal only in the gait report subgroup in which the recommended FDRO was performed.


Gait analysis can improve outcomes when its recommendations are incorporated in the treatment plan.


Femoral derotation osteotomy


Randomized controlled trial

What this paper adds

  • Gait analysis can guide decision-making for femoral derotation osteotomy.
  • Femoral osteotomy is effective in correcting internal hip rotation.
  • After gait analysis, recommendations can lead to improved outcomes.

Many studies have demonstrated a significant impact of gait analysis on surgical decision-making,[1-5] but fewer have examined the effects of gait analysis on surgical outcomes.[5-8] We have recently completed the first randomized controlled trial (RCT) examining the impact of gait analysis on surgical outcomes. The overall results indicated a modest positive effect of gait analysis, which we believe underestimated the true impact of gait analysis because of unexpectedly low compliance with the gait analysis recommendations in the trial.[9] The current investigation used data from the RCT to examine the effects of gait analysis and compliance with the gait analysis recommendations on a specific clinical problem, namely correction of excessive internal hip rotation in ambulatory children with cerebral palsy (CP).

Intoeing is one of the most common gait problems in ambulatory children with CP.[10] Intoeing often causes difficulties with foot clearance and tripping, and can lead to inefficient gait owing to lever arm disease. A primary cause of intoeing is transverse plane deformity or malalignment of the pelvis, femur, tibia, and/or foot, with excessive femoral anteversion being a common contributor.[11, 12] Femoral anteversion is approximately 30° to 40° at birth and normally remodels to approximately 15° at skeletal maturity.[13-16] However, in children with CP, remodeling often fails to occur and anteversion remains high throughout development.[13, 17] The primary treatment for excessive femoral anteversion and internal hip rotation is femoral derotation osteotomy (FDRO), during which the femur is cut and the distal fragment is externally rotated before fixation of the osteotomy. Femoral derotation osteotomy is effective in correcting intoed gait if passive and dynamic hip rotation are both excessively internal as indicated by physical examination and gait analysis.[18, 19]

The purpose of this study was to examine the impact of gait analysis on the correction of excessive internal hip rotation using data from a RCT. In the RCT, ambulatory children with CP were randomized into two groups in which the referring surgeon either received or did not receive the child's gait analysis report. If the gait report was received, the surgeon incorporated it into his or her surgical decision-making, choosing which gait analysis recommendations to follow. Outcomes were compared based on whether or not the gait report was received and whether or not the recommendation for FDRO was followed.


This was a secondary analysis of data from a prospective RCT registered at" (5R01HS014169).[9] The study was approved by the Committee on Clinical Investigations at Children's Hospital Los Angeles, and written informed assent and consent were obtained from the participants and their parents. Monitoring for adverse events was performed by the study team and Committee on Clinical Investigations, but these events were not formally analyzed.

Participants and randomization

The participants in the original RCT were ambulatory children with CP, aged 3 to 18 years, who were candidates for lower extremity orthopedic surgery to improve their gait (n=186; 114 males, 72 females). Participants were recruited from the orthopedic clinics at a pediatric specialty hospital in a large metropolitan area from February 2005 through June 2009. The participants were referred by four board-certified pediatric orthopedic surgeons who routinely cared for children with CP as part of their clinical practices, but did not routinely use gait analysis. Exclusion criteria included previous surgery within the preceding year and botulinum toxin injections within the preceding 6 months.

All participants underwent preoperative gait analysis at a nearby gait laboratory in a tertiary pediatric medical center. The gait analysis study included physical examination, slow-motion videotaping, instrumented gait analysis, and dynamic electromyography, and standard clinical gait reports were produced. The participants were then randomized to one of two groups: (1) the gait report group (the referring surgeon received the participant's gait analysis report); and (2) the control group (the referring surgeon did not receive the gait report). The randomization was implemented by sealed envelopes using computer-generated random numbers balanced in groups of eight. Group assignments were generated by the study statistician and unsealed by an individual not involved in data collection or participant care. By design, the participants and personnel involved in data collection were blinded to the participants' group assignments. The referring surgeons were not blinded since they had to review the gait reports provided for half the participants.

The subset of individuals included in the current study consisted of those for whom the gait laboratory recommended external FDRO to correct excessive passive and dynamic hip internal rotation. Only individuals for whom both pre- and postoperative data were available were included.


The intervention in this study was receipt of the gait analysis report for participants randomized to the gait report group. Participants in both groups received standard clinical treatment, with surgical plans being determined by the treating physician with or without the gait analysis report, depending on the participant's randomized group assignment. Individuals were re-evaluated by the referring physician the day before surgery with physical examination, observational gait analysis, and review of the individual's medical records including the gait analysis report for participants in the gait report group. Femoral derotation osteotomy was recommended in the gait report for all participants in the current study, but it was up to the treating surgeon to decide whether FDRO would actually be performed. Treatment, including multilevel surgery with or without FDRO, was performed as determined by the referring physician. FDRO was performed distally unless varus correction of the hip was also being performed. Proximal osteotomies were routinely fixed with blade plates. Distal osteotomies were fixed with either 4.5 dynamic compression plates or K-wires, depending on surgeon preference. The amount of rotation to be performed was determined before the operation based on passive hip internal and external rotation and was confirmed intraoperatively using K-wires inserted proximal and distal to the osteotomy site.[20]

Data collection

Outcomes data were collected before the operation and approximately 1 year after the operation from February 2005 through July 2010. The outcome measures examined in the current study were obtained through physical examination and gait analysis. An experienced team of physical therapists (including SAR) performed a standardized physical evaluation, which included measurement of femoral anteversion using the trochanteric prominence angle test.[21] For this measurement, the individual lay prone with the knee flexed to 90°. The examiner internally rotated the hip until the trochanter reached its most lateral position. This position was assumed to represent the position at which the femoral neck is parallel to the floor, making the angle between the tibia and vertical equivalent to anteversion of the femur. Maximum hip internal and external rotation range of motion were also measured using standard physical therapy techniques.

Rotational alignment during gait was assessed using instrumented gait analysis. Kinematic data were acquired using an eight-camera, three-dimensional motion analysis system (Vicon Motion Systems, Oxford, UK). A set of 15 passive retroreflective markers was attached over specific bony landmarks of the pelvis and lower extremities following the conventional gait model with patella markers.[22] Use of patella markers enables accurate measurement of dynamic hip rotation, which is not always possible with older gait analysis models such as those using thigh wands.[22] Participants walked barefoot at a self-selected speed down a 15-meter walkway with or without assistive devices as needed, making multiple passes with rest breaks provided as needed. Kinematic data from at least three trials were averaged, and the averaged data were used for analysis.

Statistical methods

The outcome variables examined in this study were femoral anteversion, mean hip rotation in the stance phase of gait, and mean foot progression in the stance phase of gait. Outcomes were assessed both after the operation and in terms of pre- to postoperative change. First, outcomes were compared between the gait report and control groups using Student's t-tests. Additional analysis was then performed, subdividing limbs in the gait report group based on whether or not the recommendations for FDRO were followed. The latter analyses used analysis of variance (ANOVA) with Bonferroni post hoc tests. All tests were two-tailed, and the significance level was set at p<0.05. Statistical analysis was performed using Stata (version 12.0, StataCorp LP, College Station, TX, USA).


The selection criteria were met by 65 limbs in 44 children (Fig. 1). In the control group, unilateral FDRO was recommended in nine children with bilateral involvement and in one child with hemiplegia; bilateral FDRO was recommended in eight children with bilateral involvement. In the gait report group, unilateral FDRO was recommended in 10 children with bilateral involvement and three children with hemiplegia; bilateral FDRO was recommended in 13 children with bilaterally involvement. All participants underwent gait analysis within 5 months of their initial referral, and surgery was performed an average of 2.2 months later, with all participants undergoing surgery within 10.5 months of baseline. The mean duration of follow-up was 1 year 2 months (SD 5mo) after surgery or 1 year 5 months (SD 6mo) from baseline. Baseline demographic and clinical characteristics did not differ between the two groups, but time to follow-up was slightly longer in the control group (Table 1). No significant harms were reported.

Figure 1.

CONSORT (Consolidated Standards of Reporting Clinical Trials) flow diagram for randomized controlled trial. FDRO, femoral derotation osteomy.

Table 1. Baseline characteristics of the study participants
VariableGait report (n=39)Control (n=26)p value
  1. Statistical significance was evaluated using Student's t-tests for continuous variables and Fisher's exact tests for categorical variables. GMFCS, Gross Motor Function Classification System.

Mean age (SD)10y 4mo (3y)9y 5mo (2y 11mo)0.23
CP subtype
GMFCS level
Mean time baseline to follow-up (SD)1y 4mo (2mo)1y 7mo (7mo)0.005

Overall, there were no significant differences in outcome measures between the gait report and control groups (all p values >0.29). However, FDRO was performed in only seven out of 39 limbs in which it was recommended in the gait report group. Outcomes were significantly better for limbs in the gait report group when the recommendations for FDRO were followed. These limbs demonstrated greater improvement in femoral anteversion (−32.9° vs −12.2°; p=0.01), dynamic hip rotation (−25.5° vs −7.6°; p=0.001), and foot progression (−36.2° vs −12.4°; p=0.02) than limbs the control group and more normal postoperative values (anteversion: 23.6° vs 40.6° [p=0.01]; hip rotation: −2.8° vs 16.6° [p<0.001]; foot progression: −5.4° vs 6.4° [p=0.42]; Table 2 and Fig. 2). Similar differences were observed compared with the gait report subgroup in which the recommendation for FDRO was not followed (Table 2 and Fig. 2). There were no differences in outcome between the gait report subgroup in which the recommendation for FDRO was not followed and the control group, except for less change in femoral anteversion and slightly greater change in mean hip rotation in stance in the control group, in which internal rotation was greater before the operation.

Figure 2.

Pre- and postoperative femoral anteversion (top), mean hip rotation (middle), and mean foot progression (bottom) in the three groups. Normal values are from Fabry et al.[14] for anteversion and from our laboratory database for the kinematic variables. FDRO, femoral derotation osteomy.

Table 2. Comparison of outcome measures among control group, gait report subgroup where the gait analysis recommendation for femoral derotation osteomy (FDRO) was not followed, and gait report subgroup where the gait analysis recommendation for FDRO was followed. Pre- and postoperative values for each group and within-participant change are presented
 Control (n=26)Gait report, FDRO not done (n=32)Gait report, FDRO done (n=7)
  1. Bold indicates significant difference from control; italics indicate significant difference from gait report, FDRO not done. Significance level was p<0.05 based on ANOVA with Bonferroni post hoc comparisons. CI, confidence interval.

Femoral anteversion (°)
Mean52.540.6− 56.4 23.6 32.9
SD9.716.321.18.77.4 10.2 5.6 18.0 19.1
95% CI48.6, 56.433.9, 47.3−20.9, −3.543.9, 50.343.4, 48.84.7, 2.8 51.3, 61.6 7.0, 40.2 50.5,15.2
Mean hip rotation in stance phase (°)
Mean24.216.6−7.6 17.1 17.6 0.5 22.72.825.5
SD11.711.113.8 8.8 9.8 7.6 10.7 11.1 9.0
95% CI19.5, 28.912.1, 21.0−13.2, −2.1 13.9, 20.3 14.0, 21.12.3, 3.212.8, 32.513.1, 7.533.8,17.2
Mean foot progression in stance phase (°)
Mean18.86.4−12.412.14.1−8.0 30.8 −5.436.2
SD19. 16.1 38.1 38.5
95% CI11.1, 26.5−0.2, 12.9−18.5, −6.45.7, 18.5−0.9, 9.1−14.2, −1.7 16.0, 45.7 −40.6, 29.971.8,0.6


The results of this study suggest that gait analysis is helpful in evaluating common dynamic rotational problems such as intoeing. The gait analyses identified appropriate candidates for FDRO, and these individuals benefited when the recommended FDRO was performed. Previous investigations have found that physical evaluation is not sufficient to guide decision-making for FDRO; dynamic evaluation through gait analysis is needed to determine not only whether FDRO should be performed, but also the amount of derotation needed.[19] There was only modest improvement in passive and dynamic hip rotation when treatment was performed without the gait analysis report. The improvement that occurred was largely due to FDRO being performed in six out of 26 limbs in this group without knowledge of the gait analysis recommendations. There was little improvement when the gait analysis report was available but its recommendations were not followed. In contrast, there was clinically and statistically significant improvement in all outcome measures when the gait report was available and its recommendations were followed.

Femoral anteversion and hip rotation during gait are direct measures of femur alignment, which FDRO is intended to address. Foot progression, on the other hand, is an overall measure of transverse plane alignment combining the effects of femoral anteversion, pelvic and hip rotation, tibial torsion, and foot positioning. Foot progression showed some improvement in the control group and gait report subgroup in which FDRO was not carried out, likely as a result of external tibial derotation osteotomies (one in the control group, three in the gait report subgroup without FDRO) and correction of varus foot deformities (four in each group) in some participants. However, these procedures did not change the abnormal femoral anteversion or hip rotation and only partially corrected the internal foot progression. When the recommended FDRO was performed in the gait report group, all outcome measures were corrected to within the normal range.

The major limitation of this study is that only seven of the 39 of the recommended FDROs were performed in the gait report group, indicating poor compliance with the gait analysis recommendations. Possible reasons for this poor compliance include a high threshold for performing long bone osteotomies among the participating surgeons, as well as generally low compliance with the gait analysis recommendations in the main study. In the main trial, only 42% of the gait analysis recommendations were followed in the gait report group, which is much lower than the greater than 75% rate reported in previous studies,[4, 23] including one study examining clinical referrals from outside physicians to the same gait laboratory over the same time period.[24] This practice pattern may limit generalizability of the study findings since other surgeons may be more likely to implement the gait analysis recommendations. Previous research has shown that improvement in functional outcomes is proportional to the degree to which gait analysis recommendations are followed.[7, 9] The current study suggests that outcomes specific to a particular problem (i.e. excessive hip internal rotation) improve only when treatment recommendations addressing the cause of that problem are implemented.

Femoral derotation osteotomy was not performed in any of the children with hemiplegia. While this was a small group, it is possible that the excessive internal hip rotation was not recognized as a significant problem in these children as a result of compensatory external pelvic rotation, which helps to normalize foot progression.[25] Soft tissue surgery on the anterior tibialis, posterior tibialis, and Achilles tendons was performed in these children to correct for intoeing at the level of the foot and ankle, but the deformity at the level of the hip was not addressed.

Medical providers, patients and families, and insurance companies are increasingly demanding evidence of efficacy for medical procedures and diagnostic tests. This study provides evidence that orthopedic surgery guided by gait analysis can provide effective correction of excessive hip internal rotation, a common problem that causes tripping and gait inefficiency in ambulatory children with CP. These results demonstrate that gait analysis can improve individual outcomes, but only when the gait analysis recommendations are integrated into the overall treatment plan.


Support for this research was provided by grant number R01 HS014169 from the Agency for Healthcare Research and Quality. The Agency for Healthcare Research and Quality did not participate in the design, conduct, or publication of the study. The authors stated that they had no interests which might be perceived as posing a conflict or bias.