Does the upper femoral epiphysis slip or rotate?

K. Tayton


Although much has been published on the causes of slipped upper femoral epiphysis and the results of treatment, little attention has been given to the mechanism of the slip. This study presents the results of the analysis of 13 adolescent femora, and the attempts to reproduce the radiological appearances of a typical slip. The mean age of the skeletons was 13 years (11 to 15). It was found that the internal bony architecture in the zone of the growth plate was such that a slip of the epiphysis on the metaphysis (in the normal meaning of the word slip) could not take place, largely relating to the presence of a tubercle of bone projecting down from the epiphysis. The only way that the appearance of a typical slipped upper femoral epiphysis could be reproduced was by rotating the epiphysis posteromedially on the metaphysis. The presence and size of this peg-like tubercle was shown radiologically by CT scanning in one pair of intact adolescent femurs.

Slipped upper femoral epiphysis (SUFE), was first described in 1575 by Ambroise Paré.1 However, despite a mass of literature subsequently published on this subject, and with no shortage of theories on the aetiology,2,3 it is sobering to reflect that in most patients an explanation of the cause of their condition still remains elusive.

Historically, it has usually been agreed that in chronic cases (which make up the majority) the slip is caused by shearing forces that create a fault line in the capital growth plate, probably in the hypertrophic zone, where the cells are largest and their structure is least robust.46 However, there is some evidence from work in cattle that different forces may produce failure in different parts of the growth plate: compressive loading tends to injure the zone of calcification, tensile forces the upper zone of columnation, while shear and torque forces concentrate around the junction between the columnar and hypertrophic zones.7

The relevance of this to the human hip is uncertain, as the inherent strength of a growth plate depends on several factors, including the quantity and direction of its collagen fibres, the conformation of its cartilage matrix and the general shape of the adjacent epiphyseal/ metaphyseal bone surfaces. There is no published evidence that these parameters remain the same even between two joints in the same animal, let alone in joints of different species. Slipped upper femoral epiphysis is rare in animals and the published work on bovine and ovine growth plates should be viewed with considerable caution before applying the findings to humans.

Many studies have been carried out on the shear forces normally applied to the adolescent femoral head, some using MRI scans, some in animal models, and some in human specimens.8,9 However, apart from confirming the predictable facts that abnormally vertical growth plates and excessive weight will lead to increased shear forces during normal activities, they have not improved our understanding of this condition.

This is not the case in acute slips, which are frequently related to traumatic events, and where some experimental work suggests that many, if not all, are associated with fractures.8,9

Traditionally, slips have been classified on standard radiographs,3 and because these frequently show the displaced epiphysis as a uniform crescent, it is tempting to speculate that this has led to the belief that the three-dimensional (3D) growth plate has a smooth, hemispherical appearance. Surprisingly, despite much published evidence to contradict this,2,8,10 the concept of the slipping epiphysis is still widely held, perpetuating the notion that acute and chronic slips are simply variations of the same condition, whereby one piece of bone simply slides over the uniformly-domed surface of the other until it stabilises again in a new position.

There are two problems with this classic view. The first is that the chronic instability is supposed to exist for weeks or months in the centre of a growth plate which is a dynamic zone of high biological activity, continuously growing away from the presumed fault line of the slip. Similar physeal injuries elsewhere in the juvenile skeleton tend to heal and stabilise very quickly, and do not become chronic. The second is that the epiphyseal and metaphyseal surfaces adjacent to the growth plate are not at all smooth, and certainly not conducive to easy slippage. In a previous investigation,10 the author studied the intraosseous architecture in the proximal femora of a series of calves and lambs, and found a large tuberosity projecting from the posterolateral zone of the epiphysis in all cases, associated with significant ridging on the opposing epiphyseal and metaphyseal surfaces (Fig. 1). Domestic animals rarely suffer a true chronic SUFE, and their acute slips are nearly always associated with significant trauma.11,12 It was suggested that this ‘epiphyseal tuberosity’ acted as a bone peg and effectively prevented chronic slippage from taking place.10 No reference can be found to this tuberosity in the veterinary literature. In humans, the only reference is that of Scheuer and Black,13 who briefly mention a beak-like process projecting from the underside of the capital epiphysis. Although they suggest that this may be relevant in the prevention of SUFE, they do not elaborate further.

Fig. 1

Photograph of a coronal cut through the centre of the femoral neck and capital epiphysis of a lamb, showing the well-developed epiphyseal tuberosity protruding downwards.

In order to investigate whether adolescent children had bone surfaces adjacent to their capital growth plates, which were similar to those of domestic animals, with a significant epiphyseal peg and socket, a number of skeletons of adolescents were traced and examined.

Materials and Methods

Adolescent skeletons are hard to find, but seven were traced in three collections. Only 11 femoral heads and necks were suitable for detailed analysis, and as far as is known, none had suffered a SUFE. One pair of femora, belonging to the Flowers’ Collection of the Royal College of Surgeons, was intact, but permission was granted only to examine these specimens radiologically.

Each skeleton was numbered according to its source and photographed. Being dry bones, these specimens required no prior preparation, but no growth plates remained for examination. Precise measurements were then taken from the distal surface of the epiphysis and the proximal surface of the metaphysis, where each would have abutted the growth plate, as illustrated in Figure 2.

Fig. 2

Diagram drawn to scale of the average human growth plate examined, illustrating alphabetically the various measurements taken at the epiphyseal and metaphyseal surfaces as listed in Table I. H. height, of the epiphyseal tubercle.

The mean was calculated for each measurement and a diagram was constructed to illustrate the shape and size of the average growth plate. Matching epiphyses and metaphyses were then relocated in their anatomically-aligned positions and photographed.

Attempts were made to reproduce the appearances of a typical SUFE by displacing the epiphyses into different positions with respect to the metaphyses. The traditional posteromedial displacement and various rotational displacements were tried, and the results again recorded photographically. Finally, the seventh pair of femora (belonging to the Royal College) was subjected to plain radiography and a CT scan in order to visualise the internal architecture at the level of the growth plate.


Of the six skeletons available for detailed examination, five had been salvaged from burial sites where the original graves had been turned to different uses. Inevitably, all were very old, but although none was in pristine condition they were sufficiently well preserved for the purpose of this study.

There were three male and three female skeletons, with a mean age of 13 years (11 to 15), which yielded 11 femora for study. The capital epiphyses were already separated from the metaphyses.

The dimensions of the appropriate epiphyseal and metaphyseal surfaces are shown in Table I, and it can be seen that the range of variation for each measurement is fairly small. There were too few specimens to enable statistical analysis, but the consistency was such that the findings are likely to be representative of a larger population.

View this table:
Table I.

Dimensions (mm) of the epiphyses and metaphyses in the 11 femora studied. L = length of femur. See Figure 2 for interpretation of other dimension code letters

The average dimensions of the growth plates are illustrated in Figure 2, which has been drawn to the appropriate scale, and more graphically in Figure 3, where the most prominent feature is the ‘epiphyseal tubercle’. However, despite the presence of a reciprocal notch in the metaphysis, in no example could the tubercle be described as a large projection that pegged the epiphysis down.

Fig. 3a, Fig. 3b

Drawing of the average subcapital growth-plate arrangement in a 13-year-old adolescent, a) in the anatomical position, and b) how it would appear with a minimal slip.

Photographic studies (Fig. 4) confirm that both the epiphyseal tubercle and the metaphyseal notch are proportionally smaller than their ovine and bovine counterparts.10

Fig. 4a, Fig. 4b

a) Photograph of a typical epiphysis seen from the distal aspect, illustrating the generally irregular appearance of the surface and the epiphyseal tubercle in the posterolateral area. b) Photograph of the metaphysis which matched the epiphysis shown in a) illustrating the correspondingly irregular appearance of the surface and the small notch in which the epiphyseal tubercle would locate.

Figure 5 shows the typical appearances when an attempt was made to displace the anatomically-located epiphysis posteromedially on its metaphysis, the position considered typical of a classic SUFE. Despite the age of these specimens and the fact that the epiphyseal tubercle was not large, when this manoeuvre was attempted, the anterior border of the epiphysis separated quite dramatically from that of the metaphysis as the epiphyseal tubercle disengaged from its bed and rode up over the edge of its socket. The epiphysis was so unstable in this position that balancing it satisfactorily to obtain a photograph was difficult, as shown dia-gramatically in Figure 3.

Fig. 5a, Fig. 5b

Typical a) anterior and b) lateral appearances of the epiphysis/metaphysis complex following a small amount of posteromedial displacement of the epiphysis, showing tilting of the epiphysis.

Figure 6 shows the appearance when the epiphysis was rotated medially on the metaphysis using the epiphyseal tubercle as a fulcrum. During this manoeuvre, the epiphysis could be moved through a moderate degree of rotation without much difficulty, and produced an appearance remarkably similar to the classic radiological image of SUFE. The epiphysis was relatively stable in this position, unlike when it was displaced posteromedially (Fig. 5).

Fig. 6a, Fig. 6b

Typical a) anterior and b) lateral appearances of the epiphysis/metaphysis complex following a small amount of medial rotation of the epiphysis, illustrating the remarkable similarity to the usual radiological appearance of slipped upper femoral epiphysis.

All the 11 proximal femora studied displayed similar characteristics in these two positions. Posteromedial displacement of the epiphysis was unstable in all cases with a tilted appearance, and medial rotation was reasonably stable, in all cases with no massive tilting until a significant amount of rotation had been applied.

Radiological investigation.

Radiographs of the proximal femora of a 12-year-old child, specimen 5.3 in the Flowers Collection, were of average quality, but even on the lateral projections the detail was not clear and the presence of an epiphyseal tubercle was debatable. However, Figure 7 shows a CT scanogram of the left femur and the high quality of imaging was very apparent when compared with the radiographs. This may explain why epiphyseal tubercles have received so little attention in the past when assessed only by plain films. In the CT images both left and right tubercles were well shown, each measuring approximately 4 mm in length on the axial views, pegging their respective epiphyses down into the adjacent metaphyses. From the image shown in Figure 7 it is clear that they would act as a significant block to any tendency for the epiphysis to slip.

Fig. 7

CT scanogram of the intact Royal College of Surgeons femur showing a well-defined epiphyseal tubercle.


This study investigated whether the well-developed tubercles found on the femoral capital epiphyses of cows and sheep, and which project down into the metaphyses where they act as bone pegs, were present in human adolescents. Epiphyseal tubercles were found in all cases, but were smaller than anticipated, especially when compared with their ovine counterparts. This may be because all the specimens were very old. The extent of their contribution towards the prevention of slippage in adolescents is uncertain, and it seems more likely that their principal effect is controlling the speed and direction of any displacement.

The CT of the undisturbed growth plate in Figure 7 shows the tubercle to be a well-demarcated peg. If fresh post-mortem material were readily available for study the anatomical findings might be more impressive.

From a clinical perspective, the most interesting findings concern the mechanism of displacement of the epiphysis on the metaphysis. It was shown that a pure posteromedial displacement of the epiphysis on its metaphysis cannot take place without a significant tilt, and that a very slight displacement rendered it unstable. However, a small degree of posteromedial rotation of the epiphysis about its off-centred tubercle caused it to appear to slip backwards off the curved posteromedial side of the metaphysis. In this position the epiphysis remained relatively stable and looked remarkably similar to the radiographs of an early SUFE.

Within the limitations of this study, it was seen that a rotation of more than about 30° caused the epiphysis to tilt and then slip off the metaphysis, to produce the appearance of an acute-on-chronic slip in the clinical situation. At the point of displacement, the anterior borders of the epiphysis and the metaphysis separated as the epiphyseal tilt developed, and an acute slip took place. Ballard and Cosgrove,14 in a series of 110 SUFEs, found eight cases of avascular necrosis, seven of which had the same anterior separation of the physis as described above.

Although the number of specimens in this study is small, the findings allow a hypothesis to be proposed to explain chronic SUFE. There is little published evidence regarding structural abnormalities of growth plates, but it seems likely that in most patients with a chronic SUFE the normal growth plates are simply not strong enough to withstand the forces being applied to them by excessive body weight.

The shape of the bone surfaces adjacent to the upper femoral growth plate indicates that a shear force is unlikely to cause simple displacement of the epiphysis without creating considerable instability or fracturing bone, neither of which is a regular feature of chronic SUFE. However, rotatory displacement of the epiphysis may take place as a consequence of torque forces. If this occurs, the epiphysis initially remains relatively stable because of the presence of its tubercle, which lies off-centre. Hence, as rotation progresses, the epiphysis begins to overhang the metaphysis. Relative stability is maintained until the overhang is sufficient to allow the epiphysis to begin to tilt backwards off the metaphysis posteromedially. As this happens, the anterior surfaces of the epiphysis and metaphysis begin to separate, and any further displacement will lead to a severe acute-on-chronic slip. This would explain a history lasting weeks or months in many children with chronic SUFE, most of whom seem to behave relatively normally during that period before presenting with an acute episode.


  • The authors would like to thank Dr L. Scheuer PhD, Professor S. Black MA (Oxon), FRS, Dundee University, the Vicar and staff of St Brides Church in Fleet Street, and the Curator of the Hunterian Collection in the Royal College of Surgeons, all of whom gave free access to their collection of bones, and without whom this project could not have been carried out. Thanks also go to I. Jones BSc, of the Department of Medical Illustration, Royal Gwent Hospital, for his help in the preparation of the illustrations seen here, and to N. Jones, DCR of the Department of Radiology, Royal Gwent Hospital, for her assistance with the CT scans.

  • 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.

  • Received April 16, 2007.
  • Accepted July 5, 2007.


View Abstract