Abstract
We performed a retrospective review of the medical records and radiographs of fifty children who had had unilateral slipped capital femoral epiphysis. Our purpose was to determine if there were any epidemiological parameters that were associated with the later development of a contralateral slip. The mean duration of follow-up was thirty-four months (range, twenty-four to sixty-eight months).Physiological maturity, determined with a modified form of the Oxford method for assessment of bone age, strongly correlated with the risk of development of a contralateral slip in patients who were initially seen with a unilateral slip (r = 0.59, p < 0.0005). The scores were determined by four raters; the variation in the predictive value of the scores among the raters was not significant (p = 0.5). The relationship between the score as determined with the modified Oxford method and the risk of contralateral involvement had a linear distribution.The four raters performed a total of 442 observations to determine the modified Oxford scores for the fifty hips. The prevalence of a subsequent slip for the hips that had been assigned a score of 16 points was 85 ± 15 per cent (95 per cent confidence limit) (seventeen of twenty observations). The risk of contralateral involvement when the score was 21 points was 11 ± 9 per cent (six of fifty-four observations). A slip did not develop in any hip with a score of 22 points or more (sixty-nine observations).For boys, the age at the time of the initial slip was predictive of a contralateral slip. A contralateral slip developed in all four boys who had been eleven years and seven months old or less at the time of the initial presentation; however, a contralateral slip developed in only nine of the twenty-two boys who had been eleven years and eight months to fourteen years and eleven months old. A contralateral slip did not develop in the three boys who had been fifteen years old or more. There was no association between age and the risk of a contralateral slip in girls.
The treatment of the radiographically normal contralateral hip in a child who has a unilateral slipped capital femoral epiphysis is controversial; the risk of development of a slip must be weighed against the risks of an additional operation. Segal et al. reported that a contralateral slip subsequently developed in nine of thirteen juvenile patients who had been seen for a unilateral slip. These authors suggested that prophylactic fixation of the normal hip be considered for boys who are 12.5 years old or less and for girls who are 10.5 years old or less.
Jensen et al. reported a subsequent contralateral slip in nine of fifty-seven children who had been seen with unilateral disease; seven of the slips were mild and two were severe. On the basis of these findings, they recommended prophylactic fixation of both hips in children who have a unilateral slip. Crawford, however, suggested that the complications that may be associated with stabilization of the apparently normal hip considerably outweigh the benefits of such fixation.
The present study was designed to define better the risk of contralateral involvement in a child who has a unilateral slip. We used a modification of the Oxford method1 to determine the bone age of our patients.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Department of Orthopaedic Surgery, Texas Tech University Health Sciences Center, 3601 Fourth Street, Lubbock, Texas 79430.
‡University of Chicago Medical Center, 5841 South Maryland Avenue, MC 3079, Chicago, Illinois 60637.
§University of Texas Medical Branch, University Boulevard, Galveston, Texas 77550.
¶South Arkansas Orthopedic Center, 1609 West 40th, Suite 501, Pine Bluff, Arkansas 71603.
We performed a retrospective review of the records of all patients in whom a unilateral slip had been treated with a single cannulated screw between 1984 and 1991. Patients who had an endocrine or metabolic disease, had bilateral involvement, or had been treated elsewhere for an initial slip were excluded. In order to be included in the study, a patient had to have had a complete set of radiographs made at the time of presentation and a minimum of twenty-four months (mean, thirty-four months; range, twenty-four to sixty-eight months) of follow-up. All patients were followed until there was evidence of partial physeal closure or until there was a slip in the initially normal hip. Fifty-six patients met the inclusion criteria, but six were excluded because they had been lost to follow-up.
The medical records provided information on the gender, age, race, and weight of the patient at presentation and any associated medical conditions. Standard growth curves were used to determine the percentile for weight in relation to age for each patient. The degree of slip was ascertained on the basis of the displacement of the epiphysis relative to the metaphysis as seen on the frog-leg lateral radiograph of both hips3.
We assessed skeletal maturity with a modification of the Oxford method1. With the Oxford method, nine distinct areas—the head of the femur, the greater trochanter, the lesser trochanter, the ilium, the ischium, the lip of the acetabulum, the junction of the ischial and pubic rami, the pubis, and the triradiate cartilage—are evaluated on the anteroposterior radiograph of the pelvis and both hips and are graded according to established standards; a point-value is then assigned for each area. The total score is a sum of the score for each distinct area.
The evaluation of all nine areas necessitates the use of serial radiographs, as some indicators are present for only a brief period during growth and development. For example, one radiograph is insufficient to determine whether ossification of the ischial apophysis has not yet begun or if the apophysis has already fused to the pelvis, as the two stages have a similar appearance. Because the decision to operate must be based on the data available from the radiographs made at the time of the initial presentation, we used the Oxford method to evaluate only the head of the femur, the greater trochanter, the lesser trochanter, the triradiate cartilage, and the ilium. Although all of the point-values allowed by the Oxford method for the areas studied were available, the raters all scored each area within a range of two or three points: 5, 6, or 7 points for the head of the femur; 4, 5, or 6 points for the greater trochanter; 3, 4, or 5 points for the lesser trochanter; 1, 2, or 3 points for the triradiate cartilage; and 3 or 4 points for the ilium (Fig. 1). As the involved hip frequently appears to be more skeletally mature and thus may be given a false point-value, we used the apparently normal hip for the scoring. A whole point was assigned for each maturity indicator, and the score for that area remained unchanged until the next indicator was plainly visible. Half points were not awarded. Acheson used only the anteroposterior radiograph for assessment with the Oxford method; however, we used the frog-leg lateral radiograph to help us to evaluate the lesser trochanter and the triradiate cartilage.
With the exception of the radiographs for Case 5, each radiograph was evaluated a total of nine times by four raters. The radiographs for Case 5 were reviewed once by rater 2 but could not be obtained for review by the other raters. All other radiographs were evaluated three times by rater 1, a pediatric orthopaedic surgeon; once by rater 2, a chief resident in orthopaedics; three times by rater 3, a pediatric orthopaedic surgeon; and twice by rater 4, a first-year medical student. Each rater viewed all forty-nine available sets of radiographs at one sitting. Second or third readings were done after an interval of at least four days. The radiographs were presented in random order such that neither the identity of the patient nor the final outcome were known to the rater.
Twenty-one girls and twenty-nine boys were initially seen for a unilateral slip. Thirty-eight children were black and twelve were white. Forty-three children were in at least the 90th percentile for weight. A slip later developed in the contralateral hip of twenty children (Table I).
To examine whether the predictive value of the score would change if the raters were varied, the number of contralateral hips that had late involvement was plotted against the mean observed score, as determined with the modified Oxford method, for each rater. The correlation coefficient (and standard deviation) was 0.68 ± 0.08 for rater 1, 0.45 ± 0.12 for rater 2, 0.53 ± 0.11 for rater 3, and 0.50 ± 0.11 for rater 4. There was no significant difference among the predictive values of the scores from the different raters (p = 0.5). It appeared that scoring with the modified Oxford method was easy to perform, as the results achieved by the first-year medical student were of generally the same quality as those achieved by the pediatric orthopaedic attending surgeons. As there was no difference in the predictive values of the total scores among the raters, the 442 observations of the fifty sets of radiographs were treated as independent Bernoulli terms in the calculation of the values (Table II).
Boys and girls were examined separately with multiple regression analysis to determine if a combination of race, age, and percentile for weight was associated with the risk of contralateral involvement. The age at the time of the initial slip was found to be significant for boys (p = 0.008). All four of the boys who were eleven years and seven months old or less had a subsequent contralateral slip. However, only nine of the twenty-two boys who were eleven years and eight months to fourteen years and eleven months old and none of the boys who were fifteen years old or more had a subsequent contralateral slip (Table III). For girls, no combination of these parameters was significantly correlated with risk (p = 0.2). A multiple regression analysis comparing the age, gender, race, percentile for weight, and mean score derived with the modified Oxford method with the outcome showed that these parameters could not be used to improve the predictive value of the modified Oxford method.
The slip in the contralateral hip was detected at a mean (and standard deviation) of 10 ± 6 months (range, one to twenty-six months) from the time of the initial presentation. Clinical and radiographic evaluations were performed in the perioperative period and then every three to four months. If a patient missed an appointment, at least one letter was routinely sent to the family offering a new appointment date. Despite all efforts to encourage routine follow-up visits, only seven of the twenty children who eventually had a contralateral slip had been seen at the specified intervals from the time of presentation until the second slip was detected. Four of the twenty patients had an asymptomatic contralateral slip that was detected on follow-up screening radiographs. The mean degree of these four contralateral slips was 10 per cent (range, 3 to 15 per cent), as determined with the test described by Bianco. One of these four children had been unreliable with regard to returning for follow-up, and a slip of 15 per cent was detected on the radiograph made nineteen months after the previous visit. In the remaining sixteen children, the contralateral slip was detected because of symptoms, and the mean degree of the slip was 15 per cent (range, 1 to 50 per cent), which is not significantly greater (p = 0.5) than that of the four slips that were detected on follow-up screening radiographs.
When the interval between follow-up visits before the detection of the contralateral slip was five months or less, the patients were classified as reliable; when the interval was more than five months, they were considered unreliable. With use of this definition, there were seven reliable patients (Cases 3, 14, 20, 29, 30, 40, and 41), in whom the mean degree of the contralateral slip was 7 per cent (range, 1 to 10 per cent), and thirteen unreliable patients, in whom the mean degree of the contralateral slip was 18 per cent (range, 1 to 50 per cent). With the numbers available, we could not detect a significant difference between the two groups with regard to severity of the contralateral slip (p = 0.1). The mean time from detection of the first slip to detection of the second slip was six months (range, one to ten months) for the reliable patients and thirteen months (range, eight to twenty-six months) for the unreliable patients (p = 0.006).
The mean degree of the index slip in all twenty children in whom a subsequent contralateral slip developed was 19 per cent (range, 2 to 45 per cent), and the mean degree of the second slip was 14 per cent (range, 1 to 50 per cent). This difference was not significant (p = 0.3).
In order to determine when or if prophylactic fixation is indicated for a radiographically normal hip in a child who is seen for a unilateral slipped capital epiphysis, the treating physician must be aware of the risks of the procedure as well as the frequency of potential outcomes if no intervention is offered. Recent articles1,18 have supported the safety and efficacy of fixation with a single screw in the treatment of slips; however, complications are still frequent enough to raise concern regarding the prophylactic use of pinning.
A higher risk for the development of a contralateral slip after unilateral presentation has been reported in younger patients. Loder et al.13 reported a mean age of twelve years for forty-one children in whom a delayed contralateral slip developed and a mean age of thirteen years for 142 children in whom one did not develop. This difference was significant (p = 0.002). Marcus and Leo believed that age was an important factor in the determination of risk for contralateral involvement and reported that boys who were less than thirteen years old were at high risk. In contrast, Jerre et al. did not detect any association between the prevalence of contralateral disease after unilateral presentation and the gender of the patient, the duration of symptoms, obesity, trauma, the side of the index slip, the severity of the index slip, or the age of the patient. One reason why those authors did not find a relationship between a younger age and the risk of a second slip may be that they included boys and girls in one group. Loder et al.13 noted that the mean age at the time of treatment was twelve years for girls and thirteen years for boys (p = 0.00001). This difference in age is very similar to the difference in age between their patients who had subsequent contralateral involvement and those who did not. Therefore, the difference between the boys and the girls could have blurred a difference due to age that would have been revealed had the sexes been examined separately.
It appears that slips occur within a relatively narrow time-period of physiological maturity during the growth of a child, although they can occur over a wide age-range8. Exner noted that the chronological ages of eight girls ranged from 10.7 to 16.3 years and those of fifteen boys, from 12.2 to 17.4 years. The bone ages of the girls ranged from 12.0 to 13.9 years and those of the boys, from 14.0 to 15.6 years. Wilcox et al. also compared bone age with chronological age in children who were seen with a unilateral slip and found that most of the patients whose bone age was younger than their chronological age were more than twelve years old. Loder et al.14 compared bone age (determined with the Oxford method1) with chronological age in thirty children at the time of presentation of a slip and found that the range of the bone ages (fifty months between the youngest and oldest bone ages) was narrower than that of the chronological ages (ninety-eight months between the youngest and oldest patients). Because these studies support the concept that slipped capital femoral epiphysis occurs in a narrow window of physiological maturity, we decided to study bone age in our patients.
Loder et al.13 found no significant difference (p = 0.12) between the severity of the index slip (mean and standard deviation, 21 ± 13 degrees) and that of the second slip (mean and standard deviation, 16 ± 12 degrees) in patients who had sequential bilateral disease. Carney et al., in a long-term follow-up study, reported increasingly poor clinical outcomes with increasing severity of the slip. The second slip does not appear to have a benign natural history. Hagglund et al.9 reported that seventeen (53 per cent) of thirty-two second slips recognized during adolescence were associated with osteoarthrotic changes during a period of long-term follow-up. Jensen et al. reported mild osteoarthrosis in one of nine hips that had a second slip detected during adolescence and severe osteoarthrosis in two.
We found, as did Jensen et al. and Hagglund et al.9, that close follow-up was not sufficient to prevent the development of a slip of the contralateral hip in children who had a unilateral slip. With the numbers available, we could detect no significant difference in degree between the index slip (mean, 19 per cent) and the second slip (mean, 14 per cent) in the twenty children in whom bilateral disease developed (p = 0.3). The degree of the slip in the children who returned regularly for follow-up examinations (mean, 7 per cent) was not significantly different (p = 0.1) from that in the children who were seen at irregular intervals (mean, 18 per cent). The mean time from the index slip to the second slip for the children who were reliable with regard to returning for follow-up was only six months, while it was thirteen months for those who were not (p = 0.006). It was no surprise that the families became less reliable with regard to follow-up as the time from the index slip increased. Therefore, depending exclusively on regular follow-up visits to decrease the severity of a second slip might not be feasible. It is possible that, if more patients had been included in this study, a decrease in the severity of the slip could have been demonstrated in children who were seen for follow-up at regular intervals. However, a certain degree of unreliability must be expected in any clinical situation. Ward et al. also noted that patients did not keep their follow-up appointments reliably. Four of their forty-nine patients were lost to follow-up in the immediate postoperative period, and five of the forty-two patients who remained in the study were seen for the first postoperative visit more than eighteen months after the index procedure.
On the basis of the development of a contralateral slip in six of six children who had endocrine disease, Wells et al. concluded that prophylactic pinning of the uninvolved hip was indicated in patients who had concomitant endocrine disease. Hagglund et al.9 recommended fixation of both hips at the time of the initial presentation in all children who have a unilateral slip. This recommendation was based on their finding that thirty-two (14 per cent) of 237 children who had a unilateral slip had a subsequent contralateral slip. Of the thirty-two contralateral slips, three were mild, thirteen were moderate, and sixteen were severe. Moderate-to-severe osteoarthrosis had developed in association with ten of the thirty-two slips by the time of the most recent follow-up examination, at sixteen to sixty-six years. Hagglund et al.9 found that 104 (51 per cent) of 205 hips that appeared normal at the time of physeal closure had a tilt deformity suggestive of a previous slip.
Ward et al. reported on forty-two patients (fifty-three hips) who had been followed for at least two years after fixation of a slip with a single screw. Complications included myositis ossificans near the screw head in two hips; penetration of the femoral head with a guide-wire, which was corrected intraoperatively, in one hip; and penetration by the screw noted postoperatively in one hip. There were no instances of chondrolysis or osteonecrosis. In three patients, the femoral head grew to such an extent that the screws no longer crossed the physis; however, the degree of slip was not found to be increased on follow-up examination. Aronson and Carlson reported no evidence of chondrolysis or osteonecrosis in forty-three of forty-four patients (fifty-eight hips) in whom a slip had been stabilized with a single cannulated screw. Osteonecrosis developed in one patient in whom an acute slip had been reduced closed. Complications included progression of one slip, loss of fixation in one hip, and a subtrochanteric fracture after removal of the implant in one hip.
Emery et al. evaluated ninety-five hips that had been treated prophylactically with multiple Crawford-Adams pins and reported that four pins had penetrated the articular cartilage of the femoral head, three pins had been implanted outside the neck or head of the femur, three pins had bent, and one pin had broken on insertion. Despite these problems with the hardware, neither chondrolysis nor osteonecrosis was detected after a mean of twenty-two months (range, two to sixty-one months). Emery et al. did not identify the criteria that they used to identify chondrolysis. Additional complications included a superficial wound infection in five patients and a keloid scar in another; in eleven patients, the pins no longer crossed the physis. Hansson followed, for a minimum of twelve months, twenty-eight asymptomatic hips that had been stabilized with a hook-pin device. Extensive radiolucent zones and elevated sedimentation rates were noted in two patients. One pin was replaced in one patient because of a draining fistula. Despite negative cultures, the fistula continued to drain until the second pin was also removed.
Probably the greatest deterrent to prophylactic stabilization of hips at risk is concern regarding unrecognized penetration of the joint by hardware and the associated risk of destruction of the joint due to chondrolysis17. However, Zionts et al. found no chondrolysis in fourteen hips in which the fixation device had entered the joint and the problem had been promptly recognized and corrected. As awareness of these risks increases and better techniques of intraoperative visualization, such as biplane fluoroscopy, are developed, it is likely that the prevalence of penetration of the joint by the pins may decline. In addition, since the asymptomatic hip is anatomically normal, prophylactic fixation may be less prone to complications than fixation of an existing slip.
In our study, none of the contralateral hips (sixty-nine observations) that had a score of 22 points or more, as determined with the modified Oxford method, were found to have a slip during the follow-up period; therefore, prophylactic fixation of such apparently normal hips seems unjustified. Hips for which the score was lower posed much more of a dilemma. No score was associated with a 100 per cent risk for the development of a contralateral slip. Therefore, the complications of fixation must be evaluated carefully before prophylactic pinning is considered.
Some problems with the present study are that it was retrospective and that several of the radiographs were suboptimum. Suboptimum radiographs could be avoided in a prospective study, thereby reducing the standard errors and making the Oxford method a reliable means for the estimation of risk. Our study included relatively few white children or girls, and it is possible that a much larger study could have demonstrated subtle differences in the risk of a second slip in these groups, although no such trend was detected in the present study. A randomized, prospective study with long-term follow-up is necessary to determine the Oxford score at which the risk-benefit ratio favors prophylactic fixation compared with close observation.
Acheson, R. M.: The Oxford method of assessing skeletal maturity. Clin. Orthop.,10: 19-39, 1957.1019
1957
[PubMed]
Aronson, D. D., and |and |Carlson, W. E.: Slipped capital femoral epiphysis. A prospective study of fixation with a single screw. J. Bone and Joint Surg.,74-A: 810-819, July 1992.74-A810
1992
Bianco, A. J., Jr.: Treatment of slipping of the capital femoral epiphysis. Clin. Orthop.,48: 103-110, 1966.48103
1966
[PubMed]
Carney, B. T.; Weinstein, S. L.; and |and |Noble, J.: Long-term follow-up of slipped capital femoral epiphysis. J. Bone and Joint Surg.,73-A: 667-674, June 1991.73-A667
1991
Crawford, A. H.: Current concepts review. Slipped capital femoral epiphysis. J. Bone and Joint Surg.,70-A: 1422-1427, Oct. 1988.70-A1422
1988
Emery, R. J.; Todd, R. C.; and |and |Dunn, D. M.: Prophylactic pinning in slipped upper femoral epiphysis. Prevention of complications. J. Bone and Joint Surg.,72-B(2): 217-219, 1990.72-B(2)217
1990
Exner, G. U.: Growth and pubertal development in slipped capital femoral epiphysis: a longitudinal study. J. Pediat. Orthop.,6: 403-409, 1986.6403
1986
[CrossRef]
Hagglund, G.; Hansson, L. I.; Hansson, V.; and |and |Karlberg, J.: Growth of children with physiolysis of the hip. Acta Orthop. Scandinavica,58: 117-120, 1987.58117
1987
[CrossRef]
Hagglund, G.; Hansson, L. I.; Ordeberg, G.; and |and |Sandstrom, S.: Bilaterality in slipped upper femoral epiphysis. J. Bone and Joint Surg.,70-B(2): 179-181, 1988.70-B(2)179
1988
Hansson, L. I.: Osteosynthesis with the hook-pin in slipped capital femoral epiphysis. Acta Orthop. Scandinavica,53: 87-96, 1982.5387
1982
[CrossRef]
Jensen, H. P.; Steinke, M. S.; Mikkelsen, S. S.; and |and |Thomsen, P. B.: Hip physiolysis. Bilaterality in 62 cases followed for 20 years. Acta Orthop. Scandinavica,61: 419-420, 1990.61419
1990
[CrossRef]
Jerre, R.; Billing, L.; Hansson, G.; and |and |Wallin, J.: The contralateral hip in patients primarily treated for unilateral slipped upper femoral epiphysis. Long-term follow-up of 61 hips. J. Bone and Joint Surg.,76-B(4): 563-567, 1994.76-B(4)563
1994
Loder, R. T.; Aronson, D. D.; and |and |Greenfield, M. L.: The epidemiology of bilateral slipped capital femoral epiphysis. A study of children in Michigan. J. Bone and Joint Surg.,75-A: 1141-1147, Aug. 1993.75-A1141
1993
Loder, R. T.; Farley, F. A.; Herzenberg, J. E.; Hensinger, R. N.; and |and |Kuhn, J. L.: Narrow window of bone age in children with slipped capital femoral epiphyses. J. Pediat. Orthop.,13: 290-293, 1993.13290
1993
[CrossRef]
Marcus, N. W., and |and |Leo, R. J.: Bilateral involvement in slipped capital femoral epiphysis. Orthop. Trans.,4: 54, 1980.454
1980
Segal, L. S.; Davidson, R. S.; Robertson, W. W., Jr.; and |and |Drummond, D. S.: Growth disturbances of the proximal femur after pinning of juvenile slipped capital femoral epiphysis. J. Pediat. Orthop.,11: 631-637, 1991.11631
1991
Walters, R., and Simon, S. R.: Joint destruction: a sequel of unrecognized pin penetration in patients with slipped capital femoral epiphyses. In The Hip. Proceedings of the Eighth Open Scientific Meeting of the Hip Society, pp. 145-164. St. Louis, C. V. Mosby, 1980.
Ward, W. T.; Stefko, J.; Wood, K. B.; and |and |Stanitski, C. L.: Fixation with a single screw for slipped capital femoral epiphysis. J. Bone and Joint Surg.,74-A: 799-809, July 1992.74-A799
1992
Wells, D.; King, J. D.; Roe, T. F.; and |and |Kaufman, F. R.: Review of slipped capital femoral epiphysis associated with endocrine disease. J. Pediat. Orthop.,13: 610-614, 1993.13610
1993
Wilcox, P. G.; Weiner, D. S.; and |and |Leighley, B.: Maturation factors in slipped capital femoral epiphysis. J. Pediat. Orthop.,8: 196-200, 1988.8196
1988
Zionts, L. E.; Simonian, P. T.; and |and |Harvey, J. P., Jr.: Transient penetration of the hip joint during in situ cannulated-screw fixation of slipped capital femoral epiphysis. J. Bone and Joint Surg.,73-A: 1054-1060, Aug. 1991.73-A1054
1991