Abstract
Background:
Temporary elbow stiffness after the treatment of a supracondylar humeral fracture in a child is often a concern of parents. However, little attention has been devoted to documenting, longitudinally, the time required for motion recovery. The purpose of the present study was to provide a prospective, longitudinal evaluation of elbow motion in a large population of pediatric patients undergoing treatment of a supracondylar humeral fracture.
Methods:
We prospectively examined 373 patients (375 fractures) who presented to our urgent care center between March 1, 2007, and September 30, 2008. On the basis of a standard protocol, patients were managed with either casting or surgery, depending on the severity of the injury, and then were followed for a minimum of seven weeks. Values of elbow flexion and extension were recorded, and the relative arc of motion was calculated as a percentage of the motion of the contralateral elbow.
Results:
In general, following a supracondylar humeral fracture, the greatest increases in flexion, extension, and the absolute and relative arcs of motion are observed within the first month after cast removal, with a progressive improvement for up to forty-eight weeks after the injury. Age had a significant effect on the recovery of elbow motion, with patients older than five years of age demonstrating a 3% to 9% lower relative arc of motion at the follow-up points in comparison with younger patients. Similarly, patients with more-severe fractures requiring surgical treatment demonstrated a decrease in relative elbow motion of 10% (with respect to the contralateral side) at the time of cast removal in comparison with those who were managed nonoperatively.
Conclusions:
The present study demonstrates that an initial rapid recovery in elbow motion can be expected after a supracondylar humeral fracture in a child, followed by a progressive improvement for up to one year after the injury. This motion recovery is slower in older patients and in those with more severe injuries.
Level of Evidence:
Prognostic Level I. See Instructions to Authors for a complete description of levels of evidence.
Supracondylar fractures of the humerus are among the most common fractures in children that are treated by orthopaedic surgeons1. The current treatment for a displaced supracondylar humeral fracture is closed reduction of the fracture with the use of percutaneous skeletal fixation2. Nondisplaced fractures are typically treated with a cast. Temporary elbow stiffness that is noted after the treatment of a supracondylar humeral fracture is often a parental concern.
Several cross-sectional studies have demonstrated that eventual recovery of motion is the norm after proper treatment of this injury3-6. However, little attention has been devoted to documenting, longitudinally, the time frame required for such recovery. To our knowledge, only three previous studies have addressed the progressive recovery of elbow motion after a supracondylar humeral fracture7-9, and all three had certain methodological limitations.
The purpose of the present study was to provide a prospective, longitudinal evaluation of elbow motion in a large population of pediatric patients undergoing treatment, both nonsurgical and surgical, for a supracondylar humeral fracture. We hypothesized that more-severe fractures would be associated with a delay in the recovery of motion. Prediction of the time required for recovery after this injury can help to set realistic expectations in the clinical setting.
We prospectively examined 512 supracondylar humeral fractures in 508 patients who presented to our urgent care center between March 1, 2007, and September 30, 2008. Patients were enrolled as part of a prospective institutional review board-approved study on pediatric elbow fractures. For the purpose of this study, multiple clinical and radiographic variables related to the elbow injury were recorded, including age, sex, side of injury, type of fracture, type of treatment, complications, and outcomes.
One hundred and thirty-seven fractures were excluded, including two fractures in two patients in whom the contralateral elbow was fractured during the recovery from the initial injury; three fractures in three patients for whom the range of motion of the normal, contralateral side was not recorded; seven flexion-type fractures in seven patients; one fracture in one patient in whom a reinjury occurred before healing was completed; and 124 fractures in 124 patients who were followed for less than seven weeks. Significant measures were employed to optimize follow-up compliance for patients who missed appointments, beginning with at least three separate calls to the patient's listed telephone number and then mailing a certified letter with rescheduled appointment information to the patient's most recent address. Therefore, a total of 375 fractures in 373 patients were included in the present study.
Of the 375 fractures, 159 (42.4%) were in girls and 216 (57.6%) were in boys. The mean age of the patients (and standard deviation) was 5.5 ± 2.4 years (range, 0.5 to 14.2 years), and the mean duration of follow-up was 19.1 ± 14.0 weeks (range, seven to eighty-nine weeks). Of the 375 fractures, 214 (57.1%) were on the left and 161 (42.9%) were on the right.
The fractures were classified by a pediatric orthopaedic surgeon (usually one of the authors [M.S.]) with use of the Wilkins modification of the Gartland system10, which has been shown to have high interobserver and intraobserver reliability11. A nondisplaced fracture was considered type I, an angulated fracture with hinging on one cortex was considered type II, and a completely displaced fracture with loss of cortical continuity was considered type III. Of the 375 fractures, 145 (38.7%) were type I, 196 (52.3%) were type II, and thirty-four (9.1%) were type III.
Patients were divided into four treatment groups (Table I). All 145 type-I fractures were treated without manipulation in a long arm cast with approximately 100° of elbow flexion (Group A). Of the 196 type-II fractures, 115 (59%) were treated with closed reduction and a long arm cast in 100° to 110° of elbow flexion (Group B). The remaining eighty-one type-II fractures (41%) underwent closed reduction and percutaneous pin fixation with the patient under general anesthesia (Group C). Surgery was indicated for patients with type-II fractures in whom the amount of swelling around the elbow area precluded the use of circumferential casting with the elbow in >90° of flexion, those with residual malrotation or extension of the distal humeral fragment following a closed reduction, and those in whom the fracture redisplaced during the follow-up period. All thirty-four type-III fractures were treated operatively (Group D) with either closed (n = 33) or open (n = 1) reduction and percutaneous pin fixation. In general, patients who were managed surgically initially received a provisional splint, which was converted to a long arm cast during the first postoperative visit. Surgery was performed at a mean of 3.4 ± 3.1 days (range, zero to fifteen days) after the injury.
Elbow immobilization was maintained for a mean of 29.0 ± 5.8 days (range, fourteen to sixty-one days). The mean duration of immobilization was 27.7 days for patients who were managed without surgery (Groups A and B), compared with 31.2 days for those who underwent surgery (Groups C and D) (p < 0.0001). For patients who underwent surgery, the pins were removed at the time of cast removal. After cast removal, active motion of the affected elbow was encouraged; strenuous activities were restricted for an additional month. No physical therapy was prescribed.
The motion of the fractured elbow as well as that of the normal, contralateral elbow was measured with a goniometer calibrated in 1° increments, with use of standard techniques. Goniometer measurements of elbow motion have been shown to be highly reproducible12. The normal, contralateral elbow had a mean flexion of 140° ± 6.8° (range, 110° to 154°), a mean hyperextension of 10° ± 5.6° (range, 28° of hyperextension to 5° of flexion contracture), and a mean absolute arc of motion of 150° ± 9.2° (range, 114° to 171°). The relative arc of motion of the affected elbow was calculated as a percentage of the arc of motion of the normal, contralateral elbow. Measurements of elbow motion were performed at each visit, beginning at the time of cast removal. The number of elbows available for follow-up at each time point is shown in a table in the Appendix.
Statistical Analysis
Initially, the Student t test was used to compare differences between groups at the various time periods, with a two-tailed level of significance of 0.05. For clarity, follow-up times were grouped into intervals ending with Weeks 6, 9, 12, 18, 24, 36, 48, and >48. Regression analyses were performed with use of a random effects mixed linear model, with the patient as the random effect. A univariate regression analysis was conducted with use of the values for the normal, contralateral elbow as a covariate adjustment and the duration of follow-up as a linear covariate. In order to determine which variables were independently related, a multivariate regression analysis was also performed.
Following preliminary evaluation of the results, the data were then analyzed with use of Kaplan-Meier survivorship curves and the Cox proportional hazards multivariate survivorship model. The end point for survivorship analysis was the achievement, in the injured elbow, of 90% of the motion in the normal, contralateral elbow (a successful result). Our survivorship analysis was designed to determine which treatment group reached the desired outcome first. Unlike a survivorship analysis with death as the outcome, a hazard ratio of >1 or a survivorship curve to the left of the others is the desired result in the present study.
Source of Funding
There was no external funding for this study.
At the time of the latest follow-up, 97% of the patients had recovered at least 80% of the arc of motion of the normal elbow, 88% had recovered 90% of the normal motion, and 38% of the elbows had recovered 100% of the normal motion.
Flexion
Overall, the mean amount of flexion at the time of cast removal was 112°. The largest increase in flexion occurred in the first month after cast removal, with mean flexion reaching 132° by Week 9. By Weeks 12, 18, 24, 36, and 48, the mean flexion reached 133°, 135°, 136°, 138°, and 140°, respectively. The values of elbow flexion according to follow-up time, for each of the groups, are presented in a table in the Appendix. After the largest improvement in the amount of elbow flexion was obtained by Week 9, there was a trend in all groups toward continued improvement through Week 48. Patients in Groups C and D, who were managed with surgery because of the severity of the injury, demonstrated a slower recovery of flexion as compared with those with less-severe injuries (Groups A and B). A significant difference in elbow flexion between patients who had had nonsurgical treatment (Groups A and B) and those who had had surgical treatment (Groups C and D) was observed through Week 18 (see Appendix).
Extension
Overall, the mean flexion contracture at time of cast removal was 47°. The largest increase in mean extension occurred in the first month after cast removal, reaching elbow hyperextension of 2° by Week 9. By Weeks 12, 18, and 24, the mean hyperextension reached 4°, 7°, and 10°, respectively, and by Weeks 36 and 48 it remained at 10°. The values of elbow extension according to follow-up time, for each of the groups, are presented in a table in the Appendix. Patients with more-severe injuries (Groups C and D) demonstrated a slower recovery of extension as compared with those with less-severe injuries (Groups A and B). A significant difference in elbow extension between patients who had had nonsurgical treatment and those who had had surgical treatment was observed at all times through Week 48, with the exception of Week 18.
Absolute Arc of Motion
The mean absolute arc of motion at the time of cast removal was 66°. This arc of motion reached 134° by Week 9. By Weeks 12, 18, 24, 36, and 48, the mean arc of motion was 138°, 142°, 147°, 147°, and 150°, respectively (see Appendix). After the greatest improvement in arc of motion was obtained by Week 9, there was a trend in all groups toward continued improvement through Week 48. Patients with more-severe injuries (Groups C and D) demonstrated a slower recovery of the arc of motion as compared with those with less-severe injuries (Groups A and B).
Relative Arc of Motion
The mean relative arc of motion was 44% of normal at the time of cast removal and reached 90% by Week 9. By Weeks 12, 18, 24, 36, and 48, the motion reached 92%, 95%, 96%, 98%, and 99% of normal, respectively (see Appendix). All groups followed a trend toward improvement in relative arc of motion through Week 48. When the patients in Groups C and D (managed surgically) were compared with those in Groups A and B (managed nonsurgically), the former group had a significantly lower relative arc of motion at the time of cast removal (36% compared with 46%; p = 0.0001), at nine weeks (85% compared with 92%; p < 0.0001), at twelve weeks (87% compared with 95%; p < 0.0001), at eighteen weeks (92% compared with 96%; p = 0.001), and at twenty-four weeks (94% compared with 98%; p = 0.005) (see Appendix).
Regression Analysis
The univariate regression analysis, adjusted for values of the normal, contralateral side and the follow-up time, demonstrated that the age of the patient, the severity of the injury as graded according to the type of fracture, and a requirement for surgery affected the longitudinal recovery of elbow motion after a supracondylar humeral fracture. In general, a decreased range of motion (both flexion and extension) was observed in association with increasing age (p < 0.001), a more severe type of fracture (p < 0.001), and the need for surgical treatment (p < 0.001). For patients who were older than five years of age, the relative arc of motion was decreased by 9.0% (p < 0.0001), 6.2% (p < 0.00001), 4.7% (p = 0.01), 5.3% (p < 0.00001), 3.4% (p = 0.01), 4.7% (p = 0.002), and 3.6% (p = 0.04) at six, nine, twelve, eighteen, twenty-four, thirty-six, and forty-eight weeks as compared with that in patients younger than five years of age. Multivariate regression analysis demonstrated that the age of the patient (p < 0.0001) and the type of fracture (p < 0.0001) were the only independent predictors of recovery of elbow motion after a supracondylar humeral fracture. These factors are probably not independent because the need for surgery was related to the fracture type. Although one might expect greater stiffness in association with an increased duration of casting, in this analysis, the length of time in a cast did not independently predict the length of time to recovery of motion, probably because the absolute difference was small (mean, 27.7 days for patients who were managed nonsurgically, compared with 31.2 days for those who were managed surgically).
The multivariate survivorship analysis included age in the model as the relationship between age and the outcome variables was highly significant. The survivorship curves for flexion (Fig. 1) and relative arc of motion (Fig. 2) demonstrated a significant difference when Group A was compared with the other three groups (p = 0.001 and p = 0.001, respectively). The value of each curve represents the probability of not yet having reached 90% of the flexion and the relative arc of motion of the normal, contralateral side. The p values represent the result of a Cox proportional hazards model with use of age and the four treatment groups as covariates (Table II).
The comparison involving Groups B and C was performed separately and allows for the evaluation of treatment independent of type. Although no significant difference was noted in the survivorship curves for flexion alone between Groups B and C (p = 0.212), a significant difference was noted for relative arc of motion between these groups (p = 0.001) (Fig. 3).
To our knowledge, this is the largest prospective longitudinal study addressing the recovery of elbow motion in children after a supracondylar humeral fracture. The results of the present study demonstrate that, in general, the largest increase in flexion, extension, and the absolute and relative arcs of motion are observed early after cast removal, with progressive improvement over time for up to forty-eight weeks after the original injury. These findings are in agreement with those of previous studies on this topic7,8.
Previous studies have addressed the progressive recovery of elbow motion after a supracondylar humeral fracture. However, those studies had methodological limitations. Wang et al.7 studied the recovery of motion following forty-five pediatric supracondylar humeral fractures. In that study, the authors indicated that 90% of elbow flexion after a supracondylar humeral fracture is recovered at a mean of thirty-nine days after cast removal. However, there was wide variability in the time required to recover flexion, with a standard deviation of twenty-six days. Moreover, in that study, patients were only followed until 90% of elbow motion was regained. No correlation was found between the recovery of motion and the severity of the fracture, the type of treatment, or the age of the patient. The authors attributed this lack of correlation to their small sample size. The results of the present study corroborate the finding that the largest increase in elbow motion is recovered early after cast removal. However, as longer follow-up times were obtained for most of our patients, we further demonstrated continued improvement of motion up to forty-eight weeks after the injury. In contrast to the results reported by Wang et al.7, the current study demonstrated that the age of the patient, the severity of injury, and the type of treatment affected the longitudinal recovery of range of motion after a supracondylar humeral fracture. It is likely that only a large sample size, such as the one in the present study, would allow for the detection of the small differences attributed to these factors.
Zionts et al.8, in a retrospective review of sixty-three surgically treated supracondylar humeral fractures, showed progressive recovery of elbow motion over a period of fifty-two weeks. In addition to the retrospective nature of their analysis and the limited sample size, they did not examine the recovery of elbow motion in patients with supracondylar humeral fractures that were treated nonsurgically. Compared with the present study, Zionts et al.8 showed a somewhat greater range of elbow motion at six weeks after the injury but showed similar results at subsequent time points. We attribute the difference in motion at the earlier time point to the fact that our initial measurements were obtained at the time of cast removal.
To our knowledge, the only other prospective, longitudinal study on the recovery of elbow motion after a supracondylar humeral fracture was performed by Keppler et al.9. That study included fifty-one children who underwent open reduction and internal fixation and were randomized to receive physical therapy. The authors demonstrated significantly greater motion at twelve and eighteen weeks in patients who received weekly physical therapy but no difference at one year after the index procedure. As open reduction and internal fixation is uncommon for the treatment of supracondylar humeral fractures in children, those results may not be applicable to most patients.
In addition to showing that the largest increase in elbow motion occurs within the first month after cast removal, with progressive improvement over time, the present study demonstrates that patients with more severe injuries, patients more than five years of age, and patients with surgically treated fractures usually require additional time to recover elbow motion. Patients with more severe fractures requiring surgical treatment demonstrated a decrease in the relative arc of elbow motion of 10% at the time of cast removal, as compared with those managed nonsurgically, and they continued to demonstrate a 4% to 8% lower relative arc of motion for up to twenty-four weeks after the original injury. Age had a significant effect on the recovery of elbow motion, with patients older than five years of age demonstrating a 3% to 9% lower relative arc of motion across the follow-up points as compared with younger patients. In this analysis, the duration of immobilization did not independently predict the recovery of motion. We attribute this lack of correlation to the small absolute difference in casting times among the treatment groups.
In our survivorship analysis, it was noted that patients reached the end point of 90% recovery of motion sooner with less severe fracture types. When we compared Groups B and C, a significant difference was seen in the recovery of the relative arc of motion, a finding that we attribute to the range of severity of fractures within the type-II fracture pattern.
Although not all patients in the present study were followed to full recovery of range of motion, 97% of the patients had recovered at least 80% of the normal arc of motion, 88% had recovered at least 90%, and 38% had recovered 100%. On the basis of the observations made on patients followed for a longer period of time, a full return of motion can also be expected in patients with shorter follow-up times. Despite this limitation, and in contrast to previous publications, the present study includes the prospective analysis of all types of pediatric supracondylar humeral fractures, with extended follow-up data, providing information that can be more widely applicable when treating these fractures. The interobserver and intraobserver reliability for the measurements of range of motion performed with the goniometer was not tested; however, 84% of the measurements were performed by the senior author (M.S.), with use of a standardized technique.
In summary, the present study demonstrates that an initial rapid recovery of elbow motion can be expected after a supracondylar humeral fracture in a child, followed by a progressive improvement for up to one year after the injury. This recovery is slower in older patients and in those with more severe fracture patterns.
Tables showing elbow motion measurements are available with the electronic version of this article on our web site at jbjs.org (go to the article citation and click on "Supporting Data").
Cheng
JC;
Lam
TP;
Shen
WY. Closed reduction and percutaneous pinning for type III displaced supracondylar fractures of the humerus in children. J Orthop Trauma.
1995;9:511-5.[PubMed][CrossRef]
Omid
R;
Choi
PD;
Skaggs
DL. Supracondylar humeral fractures in children. J Bone Joint Surg Am.
2008;90:1121-32.[PubMed][CrossRef]
Flynn
JC;
Matthews
JG;
Benoit
RL. Blind pinning of displaced supracondylar fractures of the humerus in children. Sixteen years’ experience with long-term follow-up. J Bone Joint Surg Am.
1974;56:263-72.[PubMed]
Ariño
VL;
Lluch
EE;
Ramirez
AM;
Ferrer
J;
Rodriguez
L;
Baixauli
F. Percutaneous fixation of supracondylar fractures of the humerus in children. J Bone Joint Surg Am.
1977;59:914-6.[PubMed]
Pirone
AM;
Graham
HK;
Krajbich
JI. Management of displaced extension-type supracondylar fractures of the humerus in children. J Bone Joint Surg Am.
1988;70:641-50.[PubMed]
Lee
YH;
Lee
SK;
Kim
BS;
Chung
MS;
Baek
GH;
Gong
HS;
Lee
JK. Three lateral divergent or parallel pin fixations for the treatment of displaced supracondylar humerus fractures in children. J Pediatr Orthop.
2008;28:417-22.[PubMed][CrossRef]
Wang
YL;
Chang
WN;
Hsu
CJ;
Sun
SF;
Wang
JL;
Wong
CY. The recovery of elbow range of motion after treatment of supracondylar and lateral condylar fractures of the distal humerus in children. J Orthop Trauma.
2009;23:120-5.[PubMed][CrossRef]
Zionts
LE;
Woodson
CJ;
Manjra
N;
Zalavras
C. Time of return of elbow motion after percutaneous pinning of pediatric supracondylar humerus fractures. Clin Orthop Relat Res.
2009;467:2007-10.[PubMed][CrossRef]
Keppler
P;
Salem
K;
Schwarting
B;
Kinzl
L. The effectiveness of physiotherapy after operative treatment of supracondylar humeral fractures in children. J Pediatr Orthop.
2005;25:314-6.[PubMed][CrossRef]
Wilkins
KE. Fractures and dislocations of the elbow region. : Rockwood
CA;
Wilkins
KE;
King
RE, . Fractures in children.
Philadelphia: JB Lippincott; 1984. .
Barton
KL;
Kaminsky
CK;
Green
DW;
Shean
CJ;
Kautz
SM;
Skaggs
DL. Reliability of a modified Gartland classification of supracondylar humerus fractures. J Pediatr Orthop.
2001;21:27-30.[PubMed][CrossRef]
Armstrong
AD;
MacDermid
JC;
Chinchalkar
S;
Stevens
RS;
King
GJ. Reliability of range-of-motion measurement in the elbow and forearm. J Shoulder Elbow Surg.
1998;7:573-80.[PubMed] [CrossRef]