Secondary Logo

Journal Logo


Maintenance of Ankle Range of Motion in Children Treated for Acute Lymphoblastic Leukemia

Wright, Marilyn J. BSc PT, MEd; Hanna, Steven E. PhD; Halton, Jacqueline M. MSc, MD; Barr, Ronald D. MB, ChB, MD

Author Information
doi: 10.1097/01.PEP.0000083122.74062.1B
  • Free


Acute lymphoblastic leukemia (ALL) is the most common form of cancer in childhood. It is a malignant proliferation of the white blood cells that begins in the bone marrow and spills over into the circulation, resulting in the involvement of many organs in the process. The leukemic cells crowd out normal blood cells, resulting in anemia (reduced number of red blood cells), susceptibility to infection (from reduced numbers of normal white blood cells), and bruising (reduced number of platelets). The current treatment with intensive multiagent chemotherapy has resulted in survival rates of 80% or more. 1 The physical “costs” of cure, however, are considerable. Clinicians, including physical therapists, must now address the various late effects of cancer treatment, which may have a life-long effect on the child. Although ALL is a hematological disorder, musculoskeletal morbidity has been reported at diagnosis and during therapy, and orthopedic abnormalities are becoming recognized more frequently among the expanding population of long-term survivors of this disorder. 2–6 Problems include pain, osteopenia, osteonecrosis, fractures, and contractures. These impairments can affect gait and other gross motor skills.

In a study of children treated previously for ALL, subjects were able to perform basic gross motor functions, but musculoskeletal impairment was evident and levels of motor proficiency were significantly poorer than those of children who were without health problems and matched for age and gender. 5 The most obvious impairment observed was limitation of active and passive ankle dorsiflexion range of motion (DF-ROM). The limitation was severe enough to warrant surgical intervention in >5% of cases. 6 Analyses revealed that female gender and a younger age at diagnosis were predictors for decreased DF-ROM. Weight for age at time of follow-up and linear growth during treatment were correlated negatively with active and passive DF-ROM. 6

A certain degree of DF-ROM is necessary for normal patterns of motor activities, such as walking and stair climbing. Minimum amounts of passive ankle DF-ROM deemed necessary for normal gait vary from five to 10 degrees and measures of less than five degrees can result in musculoskeletal compensations such as knee hyperextension, abduction of the foot, increased pronation and midtarsal dorsiflexion. 7,8 Limitations can result in abnormal and inefficient gait patterns including limited advancement of the tibia in stance, premature heel rise, out toeing, decreased step length, reduced velocity, inadequate clearance in swing phase, and poor heel strike and foot lowering at the beginning of stance phase. 9 The latter three can be affected also by limitations in active DF-ROM. Various combinations of these gait deviations have been observed clinically in children receiving treatment for ALL.

Although problems with ankle DF-ROM in survivors of ALL have been identified in the literature, there is no prospective documentation of the changes. 6,10 The purpose of this study was to document changes in ankle DF-ROM in a group of children treated for ALL who underwent a program of physical therapeutic prevention, education and intervention as the standard of care and compare these children to a group of historical controls who had not received this standard of care and to a comparable group of children without health problems.


The subjects included all children diagnosed with ALL at the McMaster Children's Hospital, Hamilton Health Sciences, over a four-year period, who received a full course of treatment according to Dana Farber Cancer Institute protocols 91–01 and 95–01, and who remained in continuing complete remission in the first year following treatment. Following a month long period of intensive multiagent chemotherapy to induce remission, the children receive a two-year course of systemic multiagent chemotherapy and prophylaxis to reduce the risk of leukemic cells continuing to proliferate in the central nervous system. Treatment protocols include the use of the drugs asparaginase, corticosteroids, anthracyclines, mercaptopurine, methotrexate, and vincristine. Central nervous system prophylaxis includes intrathecal chemotherapy (injection of drugs into the spinal fluid), and cranial irradiation in some children. 1 At diagnosis, the children are assigned a risk for relapse (standard or high) according to Dana Farber Cancer Institute criteria based on age and severity of disease. 1 Children at high risk for relapse receive more intense chemotherapy, particularly with steroids and anthracyclines, placing them at greater risk for the late effects of these drugs. An earlier study provided data from a comparable group of 54 children without health problems and historical comparison data from 54 children who had been treated previously for ALL on similar protocols, but had not received a routine program of education and physical therapy. 6

Range of motion was documented throughout the two-year treatment period and up to one year after the end of treatment (mean length of follow-up 11.25 months). Active and passive DF-ROM were measured by aligning a goniometer, calibrated in one-degree increments, with the lateral plantar surface of the heel and foot. The children sat semireclined with their knees fully extended. The feet were positioned manually in subtalar neutral alignment. 11 A right angle, relative to the long axis of the fibula, was designated as neutral (zero degrees) dorsiflexion. DF-ROM measurements less than neutral were given a negative notation. Passive DF-ROM was measured at final endpoint (passive tissue extensibility) range of motion. One person (M.J.W.) performed all the measurements. Intrarater reliability coefficients were 0.760 for active and 0.927 for passive DF-ROM. Clinical variables including age at diagnosis, gender, risk for relapse, presence of lower extremity soft tissue injuries, osteonecrosis, fracture, therapeutic interventions, body weight and height while on therapy were obtained from medical records. Dual X-ray absorptiometry providing bone mineral measurements during the final year of treatment were available for 90% of the subjects.

Parent and patient education regarding the possible loss of range of motion during treatment and the importance of activity and exercise during treatment was provided to all families at their weekly clinic visits by a pediatric physical therapist. Follow up in the outpatient clinic continued throughout the two-year period of antileukemic treatment. If passive DF-ROM became compromised to <10 degrees, specific individualized interventions were provided. These included passive and active stretching programs as well as strengthening exercises and activities. Joint mobilizations were offered during clinic visits if appropriate clinically, and in some cases orthotic intervention was provided. The monitoring continued after completion of treatment for ALL.

Linear mixed-effects regression models were used to characterize change over time in ankle DF-ROM before and after the cessation of chemotherapy. The use of mixed-effects models for the analysis of individual change has been referred to as “growth curve analysis.”12,13 In addition to describing the average pattern of change over time, this method allows for orderly variations in the patterns of change; the degree of individual variations are estimated and individual growth curves are fitted for each child. To enhance the clinical meaning of the results, the data were centered 14 to estimate a segmented regression model, with parameters estimating the DF-ROM at the start of treatment, the change in DF-ROM per month during treatment, and the change in DF-ROM per month after the end of treatment. This model was applied to both active and passive DF-ROM, and then exploratory analyses were conducted by adding selected clinical variables as possible qualifiers of the average pattern of change.


The study group consisted of 40 subjects (100% of those eligible). Only one child was considered ineligible due to a coexisting diagnosis of necrotizing fasciitis. Clinical characteristics of each group of subjects are included in Table 1. All children were able to cooperate with the measurement of passive ROM; however, a few very young children were not able to reliably perform active dorsiflexion movements at certain times due to poor compliance. None of the children had limitations in hip, knee, or plantarflexion range of motion, indications of spasticity or avascular necrosis that could affect ankle measurement. Of the 36 children who had dual X-ray absorptiometry scans, 27.5% had Z-scores for lumber spine bone mineral density less than −2.0, indicating notable osteopenia. Three children had lower extremity fractures during their course of treatment.

Ankle DF-ROM of Current Cohort, Historical Controls, and Comparable Healthy Cohort

Two children had a history of idiopathic toe walking, one of whom required resting night splints to maintain adequate DF-ROM. Resting splints were supplied to two children who became nonambulatory during the final six months of their treatment due to a combination of pain, coexisting medical problems, and behavioral issues. Shoe orthotics for calcaneal valgus of >10 degrees were recommended for five children. Four children followed through with the recommendations. One child required bilateral serial casting to regain DF-ROM, which was lost after a foot injury and noncompliance with the use of resting splints and exercise. None of the children have had, nor will they require, orthopedic surgery for contracture related to treatment for ALL.

Growth Curve Analysis: Passive DF-ROM

Observed passive DF-ROM as a function of time since the start of treatment is plotted in Figure 1. The population average change in passive DF-ROM, estimated by linear mixed-effects regression, is superimposed as the solid line. The estimated parameters of change that define this line are reported in Table 2, with their 95% confidence limits. At the start of treatment, average passive DF-ROM was 16.6 degrees and did not change significantly over the 24 months of treatment. Passive DF-ROM increased significantly after treatment ended at the average rate of 0.26 degrees per month, over the 12 months of follow-up. The data in Figure 1 suggest that there is substantial variability among children at diagnosis and in the pattern of change over time. In the mixed-effects regression model, these individual differences are estimated as standard deviations in the parameters of change and have been used to calculate ranges expected to encompass 50% of parameter values in the population (see Table 2). They have been used also to plot bands on Figure 1 that encompass 50% of expected scores, after adjusting for within-subjects error. That is, there is a 50% chance that children's true passive DF-ROM will fall between the dotted lines.

Fig. 1:
Passive DF-ROM during and after chemotherapy for all subjects with fitted average change in DF-ROM (solid line) and bands encompassing 50% of expected DF-ROM (broken lines), adjusted for within-subjects error.
Estimated Parameters of Change in DF-ROM

Table 3 lists selected clinical variables that were evaluated individually as predictors of variation in the pattern of change established in the baseline model described above. Each variable is listed as a possible predictor of passive DF-ROM at the start of treatment, change in passive DF-ROM during treatment, and change in DF-ROM after treatment. Because these analyses are exploratory and are oriented to detecting possible relationships to follow up with further research, the exact p value is given up to p < 0.10.

Clinical Variables Evaluated as Predictors

Few significant (p < 0.05) predictors of the pattern of change were found among the available clinical variables. Age at diagnosis significantly predicted passive DF-ROM at the start of treatment, such that older children have lower passive DF-ROM than younger children (−0.61 degrees/year of age). Boys show significantly more positive change during treatment than do girls (+0.04 versus −0.12 degrees/month). Children categorized as having high versus low risk of relapse had significantly more positive change in passive DF-ROM after the end of treatment (+0.46 versus +0.14 degree/month). In part, this may represent a recovery from a nonsignificant trend for the high-risk subjects to lose passive DF-ROM at a greater rate during treatment (−0.13/month, p < 0.13). A similar pattern is evident for subjects with soft tissue damage of the ankle, who showed a trend for slightly faster recovery after treatment (p < 0.08), after a slightly greater decline during treatment (p < 0.20), than subjects without soft tissue damage.

Growth Curve Analysis: Active DF-ROM

Data for active DF-ROM are plotted in Figure 2 and reported in Table 2. At the start of treatment, average active DF-ROM was 11.0 degrees. Average active DF-ROM changed significantly over the 24 months of treatment at the rate of −0.10 degrees/month. After treatment ended, active DF-ROM increased significantly at the average rate of 0.26 degrees per month, over the 12 months of follow-up. As with passive DF-ROM, there is substantial variability among children in the pattern of change over time. The analyses of clinical predictors are reported in Table 3. As with passive DF-ROM, there were few significant findings. The effects of age and gender were consistent with those observed for passive DF-ROM but were estimated less reliably. The age at diagnosis was a marginally significant predictor of active DF-ROM at the start of treatment (p < 0.07) and of change during treatment (p < 0.08). Older children had lower active DF-ROM at the start of treatment (−0.56 degrees/year of age), but may have more positive change during treatment (+0.022 degrees/month/year of age). Boys showed a marginally significant (p < 0.07) tendency to change less negatively during treatment than girls (−0.02 versus −0.18 degrees/month). No other clinical variables were significant predictors.

Fig. 2:
Active DF-ROM during and after chemotherapy for all subjects with fitted average change in DF-ROM (solid line) and bands encompassing 50% of expected DF-ROM (broken line), adjusted for within-subjects error.

Prognosis: Historical and Control Comparisons

Table 1 reports the average DF-ROM at 12 months after the end of chemotherapy, predicted from the mixed-effects models, and also provides mean ankle DF-ROM for the two samples of children used for comparison. Estimated parameters from the mixed effects models were used to make predictions about the average DF-ROM expected at 12 months after the end of chemotherapy. Table 1 reports these predicted values, and also provides mean ankle DF-ROM for the two samples of children used for comparison. 6 The historical ALL group had been off chemotherapy for more than 12 months (mean = 33 months) but did not receive education or physical therapy intervention. The children who were healthy were matched to the historical ALL control group by age and gender. Confidence limits on the predicted values and means suggest that, on average, children in the ALL group who received the physical therapy standard of care are expected to have significantly higher active and passive DF-ROM than the historical ALL group. Active but not passive DF-ROM at 12 months after the end of treatment is significantly lower among the children with ALL than among the comparable group of children who were healthy. In the historical cohort, 33% of the subjects had passive DF-ROM <10 degrees (9% <5 degrees). In the current study group all children had 10 degrees or more passive DF-ROM.


Most children in the study group had sufficient DF-ROM for normal gait patterns and functional activities during treatment. Some children had significantly decreased disease-related passive DF-ROM at diagnosis, which improved. Functional active and passive ranges were restored in all children after treatment for ALL. No significant change in average passive DF-ROM was found as the children proceeded through treatment, however active DF-ROM did decrease. The etiology of the reduction of active relative to passive DF-ROM was most likely decreased force production due to the interactions or additive effects of complications associated with treatment. Antileukemic chemotherapy may affect the musculoskeletal system directly or indirectly. The most probable direct cause is vincristine-induced peripheral sensory-motor neuropathies, which present with areflexia, sensory impairment, pain and muscle weakness, and atrophy most notable in the distal extremities. 10,15–17 Studies have found significantly prolonged latencies and decreased amplitudes of sensory and motor evoked potentials in the peripheral nerves, indicating demyelination and loss of descending fibers, or loss of muscle fibers, which are partially but not totally reversible. 15,16 Indirect factors that can contribute to decreased physical activity, and therefore less weightbearing stretching force and active use of muscles, include generalized lassitude and intercurrent illness, cranial radiation, corticosteroid-induced myopathies, 18 neurotoxicity of the central nervous system, 15 obesity, pain, osteopenia, and fractures. 3,4 Improvement in DF-ROM following treatment most likely reflected discontinuation of cancer chemotherapy and reversal of the associated morbidity.

Complications varied greatly both between subjects and within subjects over the course of the treatment for ALL, a probable explanation for the large variation in both measures of DF-ROM and in the patterns of change. Our speculation, based on nonsystematic review of clinical notes, is that the varied changes reflected incidents of soft tissue ankle injury, vincristine neuropathy, and periods of inactivity, coupled with noncompliance with therapeutic suggestions. Positive changes mirrored times of increased physical activity and exercise, or other therapeutic interventions such as casting or orthotic use. Strengthening protocols have shown positive changes in similar groups receiving treatment for ALL. 19

Although individual factors, such as bone mineral density, risk of relapse, and fracture did not contribute significantly to DF-ROM in this study, it is probable that combinations of these factors affect general levels of activity and thus problems with ankle mobility. Weight showed a significant association with DF-ROM in the historical cohort but was not prognostic in the study group. The gender difference of females deteriorating more during treatment was confirmed. This could reflect a lower level of physical activity in females. A gender difference has been observed with girls being less active than boys during and after treatment for ALL. 20,21 The trend between linear growth during treatment and negative change in passive DF-ROM suggests muscle length lagging behind linear bone growth during this period. Higher DF-ROM for younger children observed at diagnosis is consistent with the normal population. Sutherland documented that the median range of passive DF-ROM among children without clinical impairments decreases from 25 degrees at one year of age to 15 degrees at seven years. 7

A limitation in this study was the use of historical controls rather than a randomly assigned concurrent control group. However, the chemotherapy administered to the two groups was very similar. Also, compliance in following the suggested activities and exercises could not be controlled rigorously in the study group. Nevertheless, in a subgroup of 10 children, parents and patients reported moderate compliance in following the suggested physical therapy programming, and, although not completely compliant with specific exercise programs, they were compliant with functional or recreational activities that promote ankle extensibility. Those with the poorest results 13 for DF-ROM were those who became inactive and were noncompliant with such therapeutic intervention.


In this study, changes in ankle range of motion associated with treatment and recovery from ALL have been documented systematically. The use of longitudinal follow-up during and after treatment provides an opportunity to explore both the average pattern of change and the degree of individual variations throughout the early recovery process. Because the sample consisted of all children who were eligible, it is highly representative of the clinical population (children with ALL). However, the degree to which these results are representative of clinical outcomes elsewhere may depend on the degree to which the cancer treatment protocol is similar. The impairment after treatment, and even while on treatment, did not show the extent of limitation apparent in the historical controls who did not have the benefit of proactive education and intervention. Patient/parent education, encouragement of activity, therapeutic exercise, continued reexamination, and appropriate orthotic use, coupled with knowledge of prognostic factors, appears to have resulted in improved DF-ROM outcome and prevention of contractures in children receiving treatment for ALL.


1. Silverman LB, Gelber RD, Dalton VK, et al. Improved outcome for children with acute lymphoblastic leukemia: results of the Dana Farber Cancer Institute consortium protocol 91–01. Blood. 2001; 97: 1211–1218.
2. Gallagher DJ, Phillips DJ, Heinrich SD. Orthopedic manifestations of acute pediatric leukaemia. Orthop Clin North Am. 1996; 27: 635–644.
3. Halton JM, Atkinson SA, Fraher L, et al. Altered mineral metabolism and bone mass during treatment for acute lymphoblastic leukemia. J Bone Min Res. 1996; 11: 1774–1783.
4. Strauss AJ, Su JT, Dalton VM, Gelber RD, Sallan SE, Silverman LB. Bony morbidity in children treated for acute lymphoblastic leukaemia. J Clin Oncol. 2001; 19: 3066–3072.
5. Wright MJ, Halton JM, Martin RF, Barr RD. Long term gross motor performance following treatment for acute lymphoblastic leukemia. Med Pediatr Oncol. 1998; 31: 86–90.
6. Wright MJ, Halton JM, Barr RD. Limitation of ankle range of motion in survivors of acute lymphoblastic leukemia: A cross-sectional study. Med Pediatr Oncol. 1999; 32: 279–282.
7. Sutherland DH, Olshen RA, Biden EN, Wyatt M. The Development of Mature Walking. London: MacKeith, 1988; 104/105.
8. Tiberio D. Evaluation of functional ankle dorsiflexion using subtalar neutral position. Phys Ther. 1987; 67: 955–957.
9. Perry J. Gait analysis: Normal and Pathological Function. Thorofare, NJ: Slack, 1992.
10. Ryan JR, Emami A. Vincristine neurotoxicity with residual equinocavus deformity in children with acute leukemia. Cancer. 1983; 51: 423–425.
11. Bohannon, RW, Tiberio D, Zito M. Selected measures of ankle dorsiflexion range of motion: differences and intercorrelations. Foot and Ankle. 1989; 10: 99–103.
12. Bryk AS, Raudenbush SW. Hierarchical Linear Models. Newbury Park, Calif: Sage, 1992.
13. Pinheiro JC, Bates DM. Mixed-effects Models in S and S-Plus. New York: Springer, 2000.
14. Draper NR, Smith H. Applied Regression Analysis. New York: Wiley, 1998.
15. Lehtinen SS, Huuskonen LIE, Harila-Saari AH, Tolonen U, Vainionpaa LK, Lanning BM. Motor nervous system impairment persists in long-term survivors of childhood acute lymphoblastic leukaemia. Cancer. 2002; 94: 2466–2473.
16. Vainionpaa L, Kovala T, Tolonen U, Lanning M. Vincristine therapy for children with acute lymphoblastic leukemia impairs conduction in the entire peripheral nerve. Pediatr Neurol. 1995; 13: 314–318.
17. Marchese VG, Chiarello LA, Lange BJ. Strength and functional mobility in children with acute lymphoblastic leukemia. Med Pediatr Oncol. 2003; 40: 230–232.
18. Khaleeli AA, Edwards RHT, Gohil K, et al. Corticosteroid myopathy: a clinical and pathological study. Clin Endocrinol (Oxf). 1983; 18: 155–166.
19. Gocha VA, Chirello LA, Lange BJ. Physical therapy intervention for children with acute lymphoblastic leukaemia. Pediatr Phys Ther. 2001; 13: 195.
20. Halton JM, Hay J, Atkinson SA, et al. Physical activity and bone mineralization in children during treatment for acute lymphoblastic leukemia. Proceedings Am Soc Pediatr Hematol/Oncol. 1993; 2: 22.
21. Matthys D, Verhaaren H, Benoit Y, Laureys G, De Naeyer A, Craen M. Gender difference in aerobic capacity in adolescents after cure from malignant disease in childhood. Acta Pediatr. 1993; 82: 459–462.

child; leukemia; lymphocytic; acute/complications; musculoskeletal diseases/etiology; ankle joint/physiopathology; range of motion/articular; physical therapy/methods

© 2003 Lippincott Williams & Wilkins, Inc.