ahistological study ofavascular necrosis of the …€¦ · “ avascular necrosis (whitman 1945;...

A HISTOLOGICAL STUDY OF AVASCULAR NECROSIS OF THE

FEMORAL HEAD AFTER TRANSCERVICAL FRACTURE

MARY CATTO, GLASGOW, SCOTLAND

From the University Department of Pathology, Western Infirmary, Glasgow

The association of transcervical fractures of the femur with avascular necrosis of the

capital fragment was recognised many years ago. Nevertheless several basic questions, such

as the incidence ofavascular necrosis, the extent and pattern ofbone death and revascularisation

and also the role played by avascular necrosis in non-union of these fractures, remain

unanswered. In the hope that a histological study might throw some light on these and other

problems still under discussion, all femoral heads removed at operation or necropsy from

patients with displaced transcervical fractures have been examined since January 1958.

MATERIAL AND METHODS

The material falls into the following groups: I) fifty femoral heads taken at necropsy

from patients with transcervical fractures; 2) seventy-eight femoral heads removed at primary

prosthetic arthroplasty ; 3) sixty heads excised as a secondary procedure when the fracture

became redisplaced or failed to unite. Femoral heads which were infected, had post-irradiation

osteitis or inadequate histological preparation were not included in this study. Twelve heads

in which the fracture united but secondary arthroplasty was done for late segmental collapse

of the head are discussed separately in this issue of the Journal (page 777).

For comparison the upper end of the femur was examined in necropsies selected at

random in forty-six female and four male subjects ranging from sixty-four to eighty-six years

of age with an average of seventy-four.

Preparation of specimens-Each femoral head was fixed in 10 per cent formal saline for at least twoweeks. Coronal slices about four millimetres thick were cut on a band saw and radiographs were takenon Kodak K.P.5 or Kodalith film using a Victor Raymax 50 machine at twenty-five kilovolts and fiveamperes. The bone slabs were subsequently decalcified in formic citrate buffer (Meyer 1956), washedovernight and then processed by a double embedding method (Russell 1956). After embedding inparaffin whole sections of the femoral head, or head, neck and trochanter, were cut on a Jung microtome(model K) and stained with haemalum and eosin.

In some cases the ligamentum teres was also available; it was cut into labelled serial blocks and,

after processing, was stained by haemalum and eosin. Specimens of special interest were alsostained to demonstrate elastica (Weigert and Orcein methods), mucopolysaccharides (PAS), connectivetissue (Masson’s trichrome), reticulin (Gordon and Sweet), amyloid (congo red) and by some of

Lendrum, Fraser, Slidders and Henderson’s stains (1962) for demonstrating fibrin (M.S.B., Masson

44/41 , Yellowsolve I).

BONE CHANGES IN “ NORMAL “ ELDERLY PATIENTS

It is widely recognised that from early adult life onwards some “ physiological “ bone

necrosis occurs, especially in subchondral bone and in the interstitial lamellae of the cortex,

and that this increases with advancing age and with deterioration of the vascular supply

(Jaffe and Ponieranz 1934, Rutishauser and Majno 1951, Sherman and Selakovich 1957).

It was therefore essential to examine control material to find out the pattern and degree of

osteocyte loss in “ normal “ femoral heads ofpatients ofsimilar ages to those with transcervical

fractures. Fifty upper femora were taken at random from necropsies in those over sixty-four

years of age; the only cause for exclusion from the series was radiotherapy to the pelvic

organs or infiltration by tumour. Bone death is recognised histologically by the presence of

empty bone lacunae; before the osteocytes actually disappear the whole cell may lose its

basophilia and be seen as a faint pink shadow in the lacuna. In these control cases the degree

VOL. 47 B, NO. 4, NOVEMBER 1965 749

FIG. 1 FIG. 2

Figure 1-The bone trabecula from a 74-year-old patient with a normal hip shows patchy osteocyte loss. Thereare nuclei in the fat cells of the marrow. Figure 2-There is complete loss of osteocytes in this necrotic bonetrabecula. The marrow is also necrotic and there is loss of nuclear staining. (Haemalum and eosin, x 150.)

.,,�

‘� , . . ..

4#{149},;�#{149}(�. ‘

.#{149}l:.‘ �

� 1�. . . :�. �

� ,�#{149}

‘‘4

� ,‘�. -‘ it..

;� �

THE JOURNAL OF BONE AND JOINT SURGERY

750 MARY CATTO

FIG. 3 FIG. 4

Figure 3-Living bone and marrow is in contrast to Figure 4 in which there is evidence of old necrosis in thecentral part of the trabecula devoid of osteocytes. New living bone has been laid down on the surface. Manydilated capillaries are seen in the revascularised marrow. This femoral head was removed three years after

fracture. (Haemalum and eosin, x 85.)

A HISTOLOGICAL STUDY OF AVASCULAR NECROSIS OF THE FEMORAL HEAD 751

of osteocyte loss was always greater in the Haversian bone of the inferior cortex of the neck

and in the subchondral bone plate than in spongy bone. Absence of osteocytes in trabecular

bone was essentially patchy and it was unusual to find a whole trabecula or even the area

bounded by a cement line to be completely devoid of cells (Figs. I and 2). In two femoral

heads some subchondral trabeculae showed a basophilic matrix and were almost devoid of

cells apart from small protrusions of live bone on the surface.

None of these control femoral heads showed any evidence of those marrow changes

which were found to precede or accompany loss of osteocytes in the fracture series. In practice

it was not difficult to distinguish between bone necrosis produced by sudden ischaemia after

fracture and that of “ physiological “ necrosis.

Unexpectedly, in almost two-thirds of the cases, small excrescences of fibre bone were

present at the tips or on the surfaces of trabeculae (Fig. 6). Usually few, occasionally they were

abundant. More rarely, fibre bone strands were seen in the marrow spaces (Fig. 5) and were

distributed unevenly in the head and sometimes also in the neck of the femur. In rather less

than a third of the cases and particularly in one there was bizarre bone formation. Here,

bone forming on the surface of trabeculae appeared to follow the borders of adipose cells so

that these became partly-and sometimes completely-surrounded by bone, giving a curiously

spiky appearance (Fig. 7). Both ofthese patterns in the normal femora are important because,

when seen after injury, they might wrongly be regarded as a reaction to it.

Examination of articular cartilage on femoral heads with no osteoarthritis showed

that, in normal elderly patients there was sometimes a slight patchy loss of chondrocytes from

the cartilage lacunae, especially in the deeper layer.

STUDY OF FEMORAL HEADS REMOVED WITHIN FIFTEEN DAYS

OF TRANSCERVICAL FRACTURE

Forty-nine femoral heads were removed from patients undergoing primary arthroplasty

and ten were taken from patients treated by internal fixation who died within fifteen days of

the fracture. Table I shows the details.

TABLE I

NUMBER OF FEMORAL HEADS REMOVED ON EACH OF THE FIRST FIFTEEN DAYS AFTER FRACTURE

Numberofdaysafterfracture

Numberofspecimens . .

1

� 6 �

2

8

3

8 �

4

10 �

5 � 6

8 � 4

7

3

8

� 4

9

2

10

1 �

11

I

12

� 2 �

13

I �

14

I �

15

�0 �

Total59

HISTOLOGICAL CHANGES

Fibrin and haemorrhage were present at the fracture site in all cases and, in some femoral

heads removed within twenty-four hours of injury, there was very slight proliferation of

fibroblasts in the area. This reaction often increased in subsequent days and by the fifth was

sometimes extensive and accompanied by an increase in capillaries. Foamy macrophages

appeared in damaged marrow from about the fourth day onwards and later there was formation

of oil cysts ringed by macrophages or sometimes by giant cells (Fig. 8). It was rare to see

evidence of new bone formation at the fracture site or even of plumping of osteoblasts on the

surface of trabeculae in less than five days, and in only one case at four days was one minute

area of new bone apparent. Bone formation (Fig. 9) on the surface of trabeculae at the

fracture site and in the marrow spaces was quite noticeable by the thirteenth day in some specimens.

Changes in marrow spaces-These begin to be recognisable from two days onwards. The first

evidence of ischaemia was a peculiar agglomeration of the marrow, most readily recognisable

in areas of haemopoiesis where large spaces appeared, surrounded by blood-forming cells.

VOL. 47 B, NO. 4, NOVEMBER 1965

�;; �. �

� � IFIG. 5 FIG. 6

Irregular strands of fibre bone are seen in the marrow and on the surface of trabeculae in the femoral heads ofnormal elderly patients. (Haemalum and eosin, x 100.)

p.

I


752 MARY CATTO

I __FIG. 7

A normal femoral head with bone forming on thesurface of a trabecula and enclosing within it fat cells.

(Haemalum and eosin, x 195.)

�.1

.#{248}#{149}�_.

� ; 1. ‘ � �

...‘�. �r � �, . .

FIG. 8 FIG. 9

Figure 8-An oil cyst is seen in revascularised marrow.� It is partly surrounded by giant cells. (Haemalumand eosin, . 85.) Figure 9-Active new bone formation and plumping of osteoblasts on the trabecular surfaceshas occurred at the fracture site ten days after injury. There has been much capillary and fibroblastic proliferation

in the marrow. (Haemalum and eosin, x 160.)



These cells, from four days onwards, died, lost their nuclei and hecame eosinophilic.

Often, in necrotic femoral heads removed many weeks after the fracture, these changes were

seen in a faint and ghostlike manner (Figs. 10 to 12). A similar but less striking alteration

sometimes occurred in fatty marrow and was accompanied by loss of lipocyte nuclei (Figs.

13 and 14). From three or four days onwards, necrosis of small blood vessels in marrow

(Fig. 14), capsular tags or the foveal end ofthe ligamentum teres (Fig. 15) could be recognised

by absence of nuclei, homogeneity and increased eosinophilia of the walls.

Changes in bone-Loss of osteocyte nuclei from the fragmented bone trabeculae crushed at

the fracture site was present to some extent from the fourth and fifth days, but it was notable

that it was rarely complete until the fourteenth day or even later. Osteocyte death in uncrushed

trabeculae surrounded by apparently necrotic marrow was slower and could not be discerned

until about the thirteenth or fourteenth day. It was complete, or nearly complete, at

approximately three or four weeks. There was no evidence of pre-existing trabecular death in

the femoral heads removed soon after fracture.

In assessing the state ofthe femoral head, fibroblast proliferation, plumping of osteoblasts

on bone surfaces and new bone formation were taken to indicate a blood supply in the area

or immediately adjacent to it. Agglomeration with necrosis of haemopoietic marrow, total

loss of lipocytes of fatty marrow, the presence of necrotic marrow blood vessels, complete

absence of cellular reaction at the fracture site and osteocyte loss in the uncrushed trabeculae

were thought to indicate ischaemia. In an occasional case at the fourth or fifth day in which

all the above changes except osteocyte loss were present throughout the head, it was possible

to deduce tentatively that necrosis had occurred. In general, however, and especially when

these changes were not uniform, no conclusion could be reached until the tenth day or later.

FIG. 10

#{149}�,‘:�T :;

754 MARY CATTO


At this time marrow necrosis in affected areas appeared to be complete, the boundary between

dead and living marrow was more readily defined and osteocyte loss was becoming apparent;

but the changes were even more pronounced by the sixteenth day. In practice, provided there

was ample material, there was no real problem in deciding the extent of necrosis in femoral

FIG. 11 FIG. 12

Figure 10-Normal haemopoietic marrow. Figure 1 1-Necrotic haemopoietic marrow five days after fractureshowing agglomeration and loss ofnuclei. Figure 12-Complete loss ofnuclei with the shadows of agglomerated

marrow spaces in a dead head seventeen weeks after injury. (Haemalum and eosin, x 100.)

heads removed more than sixteen days after the fracture. Usually three or four blocks of the

femoral head were available for examination. It seems foolhardy, however, to give a firm

opinion on smaller amounts of material, such as bone cores, until osteocyte loss can be

expected to be complete, often three weeks or more from the time of fracture.

ID

�

\ � ‘� ..�. �\ L

� \‘\.. �\�. . . 1 ,,�t ...-‘ I

r

/ \ - ,r,p�

� \.�. ��1

:‘ �

p.-.’

‘V. .- 7

FIG. 15

Necrotic blood vessels are seen in the ligamentum teres. (Haemalumand eosin, x 85.)

K

A HISTOLOGICAL STUDY OF AVASCULAR NECROSIS OF THE FEMORAL HEAD

DISCUSSION

755

Various interpretations have been placed on the slow disappearance of osteocytes from

bone lacunae after ischaemia. Some believe that until the osteocytes disappear the bone is

alive, and it has been suggested that if the blood supply can be restored within fifteen days

.- .

�FIG. 13 FIG. 14

Normal fatty marrow with a blood vessel is shown in Figure 13 and is in contrast to the necrotic marrow devoidof lipocyte nuclei in Figure 14. This was ten days after injury. The blood vessel wall is also necrosed. Some

osteocytes remain in the bone. (Haemalum and eosin, x 160.)

bone death will not occur (Patrick 1960). Woodhouse (l962a and b), however, produced bone

necrosis in the femoral heads of dogs after temporary occlusion of the blood supply for only

twelve hours. Sevitt (1964) suggested that the late persistence of osteocytes, for example at

eighteen days, might be the result of necrosis some days after injury. It is not, however,


756 MARY CATTO

necessary to accept this exp’anation, for osteocyte loss was not complete until three weeks

after the interruption of the blood supply to the femoral head of dogs (Bonfiglio 1954, Brown

and Catto 1964). The consensus seems to be that, in spite of the slow osteocyte loss, the

bone is dead shortly after the vascular injury (Sherman 1947, Sherman and Phemister 1947,

Phemister 1948, Bonfiglio 1954, Campbell 1961), whether this is at the time of fracture

(Crawford 1960) or at the time of manipulation and internal fixation, and whether it is caused

by actual tearing of vessels or, as has been more recently suggested (Smith 1959, Woodhouse

1962a), by their occlusion from torsion. If it is generally accepted that bone death quickly

follows the vascular injury, then a more accurate application of descriptive terms would

greatly simplify the interpretation of the voluminous literature. In spite of admonitions by

Phemister (1948, 1949a and b), Sherman (1947), Hodges (1954) and others, there is a general

tendency to write about “ late “ avascular necrosis (Whitman 1945; Christophe, Howard,

Potter and Driscoll 1953 ; Cave 1960), when in fact it is the clinical recognition of this

complication that is late. In particular, the term “ late segmental necrosis “ indicates a false

conception of the processes involved and “ late segmental collapse “ is more suitable.

Similarly there is a tendency to reserve the term avascular necrosis for its clinical

manifestations and to imply that it is in some way different from dead bone: “ Ifwe are correct

that in the vast majority of these fractures with definite displacement of the femoral heads are

dead and have to be revascularised, it is not surprising that there is a continuing high incidence

of avascular necrosis “ (Cleveland and Fielding 1954). The term “ viable head “ is sometimes

used (Compere and Wallace 1942, Brindley 1963) without clearly indicating whether this is a

femoral head which has never been dead or one which has been revascularised-an important

distinction.

STUDY OF FEMORAL HEADS REMOVED MORE THAN SIXTEEN DAYS

AFTER FRAcTURE

VIABLE FEMORAL READS

Some necrosis is known to occur in any bone immediately adjacent to a fracture

(Phemister 1930, Hatcher 1952, Ham and Leeson 1961), and an arbitrary ruling had to be made

to separate femoral heads in which the amount of bone death at the fracture site might be

considered within usual limits, from those in which it was excessive. McLean and Urist (1954)

said that at least halfa centimetre ofshaft is damaged on either side ofa closed, non-comminuted

fracture. Femoral heads which showed death of bone confined to half to one centimetre from

the fracture line have been considered to be alive. The narrow margin of dead bone which

commonly surrounded the nail track was also regarded as normal and ignored. There were

in all eighteen femoral heads judged to be alive by these criteria (Table II). Eight showed a

slight increase in the depth of dead bone at the upper edge of the fracture site and two a

similar increase at the lower edge where the inferior cortex of the neck presented a small spur.

It is not intended to imply that all the blood vessels to the live heads were intact, but

simply that sufficient viable vessels remained to prevent necrosis except at the fracture site. It

was found both from studying the operation notes and from examining the femoral heads

histologically that any remaining retinacular vessels were almost invariably attached to the

lower part of the head and were, presumably, the inferior metaphysial group. In only one

femoral head was there histological evidence of live vessels in the still attached superior part

of the retinaculum. In one specimen the vessels of the ligamentum teres alone were sufficient

to nourish the whole femoral head in which the surgeon had specifically noted this to be the

only remaining soft-tissue attachment. Similar observations were made by Schmorl (1924),

Santos (1930), Phemister (1948) and Sevitt (1964).

In the live femoral heads removed from sixteen days to ten months after injury viable

callus was invariably present at the fracture site. In the absence of good fixation the presence

of a live head was not, however, a guarantee of bony union (Charnley, Blockey and Purser



1957). In one unfixed fracture there was fibrous union eighteen weeks after the injury and in

four other patients in whom immobilisation was inadequate dense collagen and fibrocartilage

were conspicuous at the fracture site.

NECROSIS OF THE WHOLE FEMORAL HEAD

Thirty-six femoral heads were completely dead. In fifteen of these no revascularisation

had taken place and at the time ofremoval, which ranged from sixteen days to forty-two weeks

TABLE II

MATERIAL

Details of all femoral heads removed

Removed under � Removed over � T #{220}�lsixteen days � sixteen days 0

Failed nails . . 0 � 60 � 60

Primary replacement 49 29 78

Necropsy . . 10 20 � 30

Total . . 59 � 109 168

Details of femoral heads removed after sixteen days

��Lower � Foveal triangle spared � Complete necrosisAlive� head � � � � R -�

___________� spared Small Medium� Large Total � vascul�isei revascularised Total

Failed nails 2� 3 H 1 1 � 8 � 6 25 � 17 13 � 30

Primary 10 � ‘ 4 � 6 3 13 � 4 0 � 4replacement � � � I

Necropsy . � 6 � I � 3 � 4 � 4 11 � 0 2 � 2

Total . � 6 � 18 � 18 � 13 49 � 21 15 � 36

Sex and Iige� of subjects with femoral heads removed after sixteen days

� Male � Female � Total Average age

Live heads . � 4 � 14 � 18 � 78

Completely dead � 6 � 30 36 72

Partly dead . � 7 � 48 55 74

Total . . � 17 � 92 109 I

after fracture, with an average of eleven weeks, the head remained totally necrotic without

any vestige of cellular reaction (Table II). In the remaining twenty-one heads, which were

removed from three weeks to three years after fracture, with an average time of twenty-nine

weeks, some revascularisation had occurred, but in thirteen this was very slight and consisted

only of a tiny zone in the foveal region or at the inferior margin of the fracture line where a

shred of soft tissue was attached. In only eight of the thirty-six femoral heads was there

notable revascularisation, which varied from about 1 5 per cent to almost complete. In one

case there was early fibrous union between a dead and unrevascularised head and a live neck

(Case 5).


11’ . 16 FIG. l�

Figure 16-A large vascular wedge of living bone with its base on the foveal area is shown. Theremainder of the bone is necrotic. Figure 17-A small vascular fov�al wedge remains in the fernoral

head.

758 MARY CATTO


PARTIAL NECROSIS OF THE FEMORAL READ

In fifty-five femoral heads necrosis was partial. The most common pattern (forty-nine

cases) was of a wedge of living bone, with its base in the subchondral region of the fovea,

varying in size from a few trabeculae to more than half the head. The size of this wedge was

roughly classified as large (more than 33 per cent) (Fig. 16), medium (10-33 per cent) and small

(less than 10 per cent) (Fig. 17); the numbers in each group are shown in Table II. It was

clear that these areas were nourished from the ligamentum teres. In six cases the lower part

of the head was spared, the upper part being dead. In most of these the living bone included

that around the fovea but in one it barely impinged on this zone. In five, retinacular tissue

containing live blood vessels was recognised histologically and it is likely that in this group

some blood supply came from the inferior metaphysial arteries and probably, though not

invariably, also from the ligamentum teres. In all cases in which there was partial survival of

the femoral head some revascularisation of dead bone had occurred. It was difficult to reach

any conclusion about the rate of revascularisation because there was such a wide variation in

the amount of initially living bone marrow. As might be expected, in general, those femoral

heads with a large, living foveal wedge became revascularised more rapidly, the process being

virtually complete sometimes as early as four to eight weeks after the fracture, whereas those

with a small, living foveal wedge sometimes showed only small areas of revascularisation after

forty-two weeks.

There was no histological evidence in either the complete or partly necrotic femoral heads

that necrosis had occurred in more than one episode apart from slight local damage at the

fracture site.HISTOLOGICAL FEATURES OF REVASCULARISATION

Revascularisation of the marrow was recognised by proliferation of fibroblasts and leashes

ofcapillaries with groups offoamy macrophages (Figs. 18 and 19). This was usually followed

j/� {\; ��! T , ,-

p� ;: � � - �

. - .L’ �

\ ,,,t, I

r� �

�f�_ -

/, .

. �. :- �. � #{149}‘ �‘#{149}#{149}. ..

\ � ._.‘c� . �,

� .;. � � ,� � A

\�‘ : �(:.: #{149}

V

I,.

It�..�- �

.‘� I,

� � �

� .. � .1

�.

rT ‘ . F

� �.�j. ;4�

,pr , ‘�+... 1��

P,

.�-,.. �.. �

:�±iTh��I

A

FIG. 20 FIG. 21

Figure 20-Both osteoclasts and osteoblasts are seen on the surface of dead bone. The marrow has beenrevascularised. (Haemalum and eosin, x 125.) Figure 21-No cellular reaction is seen in relation to dead bone

here. The marrow is mostly fibrous and only small capillaries are present. (Haemalum and cosin, x 60.)

V�, #{149}.� . �

VOL. 47 B, NO. 4. NOVEMBER 1965


Figure 18-Foamy, fat-laden macrophages are seen at the edge of an area ofrevascularisation. (Hacmalum andeosin, � I 20.) Figure 19-The edge of an area of revascularisation is seen, still necrotic marrow remaining onthe right. Numerous thin walled capillaries are present but no osteoblast activity is seen. (Haemalum

and eosin, x 55.)

,-;

;#;;.r. �

. p�--. �

�:‘T�’�� ‘! �- 1’‘�\‘:‘: �-�. .� -,. � �

.I�.;e�.:\:J.

..,.: j

�.

�S4 . � #{149}�:

‘i.) .. �

� � � . :

� ‘, ,-�� .. . . .

�

r.’ �

,�‘ . #. �

� #{149}...-�. . �1’� �,

. �.: �-

� . ;� I ; � - ii’

�‘-- � . . .

. #{149}1

�

. ..a.., � #{149}�<‘�. � . .

FIG. 22

Twelve weeks after fracture much new bone containingosteocytes has formed on the surface of the dead bonewhich has empty lacunae. (Haemalum and eosin, < 70.)


760 MARY CATTO

by osteoclasis of dead bone and laying down of new bone on the surface of dead trabeculae

(Fig. 20), though sometimes all cellular reaction appeared to be inhibited by the formation

of poorly vascularised fibrous tissue in the marrow spaces (Fig. 21). Often osteoclasis was

slight, the striking feature being the laying down of new bone (Fig. 22).

In the early months after fracture the marrow changes were so clear cut that it was possible

to distinguish at a glance living bone surrounded by normal fatty or haemopoietic marrow

and dead bone with a few surface excrescences of new bone surrounded by very cellular,

proliferative marrow. Later the marrow cellularity decreased but central cores of unresorbed

dead bone within trabeculae persisted for several years. This made it possible, even when

revascularisation was far advanced, to deduce with a fair degree of certainty the borc.�er

between bone which, having never been dead, contained osteocytes throughout (Fig. 3), and

that which had been dead and now, after marrow revascularisation, was covered by living

bone (Fig. 4). If resorption had been un-

usually active the area of dead bone could

appear to be smaller than it had been in-

itially. It is emphasised that clearly defined

areas of bone necrosis entombed in living

bone were not seen in the femoral heads of

elderly controls but they were a feature

in bones of young adults with caisson

disease.

The patterns of revascularisation were

the same in the completely and partly dead

femoral heads. In most cases revascularisa-

tion was confined to a wedge-shaped area

around the fovea. This had its base on the

articular surface and advanced into the

marrow in a narrower spearhead (Fig. 23).

When it reached the fracture line there was

usually a shallow triangle of bone still re-

maining necrotic in the inferior part of the

head and a larger area of necrosis in the

superior part of the head. Sometimes a

small contribution to revascularisation was

made by vascular attachments at the

inferior margin of the fracture site but

these were not usually enough on their own

to supply any sizeable area and any spurs

of the inferior cortex of the neck were

slow to regain a blood supply. The upper weight-bearing part of the femoral head, and

especially the subchondral region, was almost invariably the last area to become revascularised

(Figs. 26 and 31).

In most of these capital fragments it was clear that no contribution to the restoration of

a blood supply had been made from the neck across the fracture site because there was still a

mass of dead trabeculae and fibrin at the fracture site (Fig. 24). In the primary arthroplasty

group and two of the necropsy cases this can be explained by the absence of fixation. In the

group in which pinning failed, inadequate immobilisation or continuing impaction ofthe dead

head may have destroyed attempts by granulation tissue to bridge the fracture line. In seven

of the redisplaced fractures there were fragments of dead callus (Fig. 25) lying at the fracture

site indicating attempted union and revascularisation. In the forty cases in which pieces of

the neck were available for study, although the depth of bone necrosis varied from a few

FIG. 23

Revascularisation of dead bone is occurring from aroundthe fovea. The extent of the base of the revascularised

wedge is marked by arrows.

�S�_.#{149}#{149} � � � S.

� 1’;�‘_ � ‘:� . � � , ‘-

S � � � � � � � � , . � . ,. S

� #{149}‘_)�� , S � � : � �.� � &5.,5 . �� . � . S #{149} � ‘ � : � .� � � �

I, � � � �, � � �

�..

��‘�‘�3-r

FIG. 24 FIG. 25

Figure 24-Revascularisation of the femoral head has clearly not occurred across the fracture line which stillshows on the proximal side fibrin and shattered dead trabeculae. (Haemalum and eosin, x 70.) Figure 25--A frustrated attempt at union has occurred here: dead callus can be seen at the fracture line. (Haemalum and

eosin, \/� 80.)


A HISTOLOGICAL STUDY OF AVASCULAR NECROSIS OF THE FEMORAL HEAD

millimetres to more than a centimetre there

was always very active revascularisation

and much formation of new bone on the

surface of dead trabeculae. In five corn-

pletely or partly necrotic femoral heads

there was evidence of a vascular contri-

bution from the neck with attempted union

and these are described below.

DISCUSSION

761

The normal vascular pattern of the

femoral head and neck in adults has

been widely investigated (Kolodny 1925;

Wolcott 1943; Tucker 1949; Trueta and

Harrison 1953; Judet, Judet, Lagrange and

Dunoyer 1955) and the nomenclature

adopted by Trueta and Harrison for the

blood vessels has been used in this paper.

The femoral head is supplied with blood by

cervical vessels which cross the marrow

spaces from below, by the ligarnenturn teres

(medial epiphysial artery arising from the obturator artery)and chiefly by the retinacular arteries

which run along the neck beneath the synovium. The superior retinacular group gives offfirst of

all superior metaphysial branches and then a larger lateral epiphysial artery which within the head

I’ � � ;:i;� � \\ \�4$�4� �

� ‘i-’ . . ..

� �.,�- .;,.�,. ‘ %�.. S. � � . � . � .� � ‘ .

� � � -%

� S � �#{149}� .. . ‘. �,, , . ‘I’ . � .

� .� ..‘� t’�:L.#{182}.... . . . .S’\ � � �. .� S � � � #{149}� �

iS�.

�

-.54,,

S..

S � 4P � � �

762 MARY CATTO

forms an arcade of vessels which run parallel to but above the line of the old epiphysial

cartilage plate. These vessels supply the medial and upper parts of the head while the lowest

third of the head is supplied by the inferior metaphysial arteries running in the inferior

retinaculurn. When a fracture occurs the cervical vessels are ruptured and the femoral head

then depends for its nutrition on any surviving retinacular blood vessels-though Badgley

(1960) believes that all are ruptured in displaced fractures-and on those of the ligamentum

teres. As described above it was concluded that there were usually, in these displaced fractures,

much destruction and damage of retinacular attachments whether at the time of fracture or

after manipulation but that when vessels survived they were more often in the lower than in

the upper part of the head (Harty 1953, Merle d’Aubign#{233} and Cormier 1956, Boyd 1957,

Hulth 1958a and b, Mathon 1959) although Claffey (1960) believes the inferior metaphysial

arteries are always torn. In general, the blood supply of the femoral head depends to a great

extent on the medial epiphysial arteries in the ligamentum teres and to a lesser extent on the

retinacular vessels, especially the inferior ones. It was apparent that, although in injection

studies (Trueta and Harrison 1953, Cheynel 1954, Judet et a!. 1955) anastomosis was

demonstrated between the various groups of vessels, in practice such blood supply as remained

was frequently insufficient to nourish the whole head. The amount of the head which was

supplied by the medial epiphysial arteries alone varied from a few trabeculae at the fovea to

the entire head. The contribution from any remaining inferior metaphysial arteries was

sometimes sufficient alone, or aided by the ligamentum teres vessels, to keep alive the lower

half of the femoral head. It is seen that the area of the head most vulnerable to necrosis is

the upper weight-bearing area. This histological finding is supported by many other studies

such as venography (Hulth 1958a and b), phosphorus32 injection and autoradiography

(Boyd, Zilversmit and Calandruccio 1955; Boyd and Calandruccio 1963), tetracycline maps

(Woodhouse l962b) and most recently by Sevitt’s (1964) elegant necropsy radio-opaque

injection patterns.

Some revascularisation may occur across the fracture line but when there is no significant

contribution from the foveal (medial epiphysial) vessels it is extremely slow and often of small

amount. This was seen in the twelve examples of late segmental collapse discussed in greater

detail elsewhere in this issue. In eleven of these patients with united fractures much of the

head remained dead several years after injury. It was unlikely that the delay and incompleteness

of revascularisation was brought about by necrosis of the distal femoral neck fragment because,

in the neck samples examined from redisplaced fractures, there was invariably, in spite of

varying degrees ofinitial necrosis, very active marrow revascularisation and much reossification

and callus formation (Judet et a!. 1955). In comparison with the slow revascularisation across

the fracture, the rate of revascularisation from the medial epiphysial arteries, though variable,

was quicker, and often continuing, although occasionally it was not, and then the

revascularisation front consisted of dense non-osteoblastic fibrous tissue surrounding the dead

trabeculae. Small additional contributions to revascularisation were sometimes seen at the

periphery of the head where there had been reattachment of soft tissue. It is, however, the

arteries of the ligamentum teres which play an important part in the revascularisation of the

necrotic or partly necrotic femoral heads and this is in agreement with Sevitt’s findings in 1964.

It was very striking that the last area of the femoral head to become revascularised was almost

invariably the upper subchondral region (Bonfiglio 1954, Boyd 1957). Small pockets of

necrotic bone were still present (Figs. 26 and 31) in two patients with bony union of their

fractures (Cases 1 and 2) and in one with fibrous union (Case 4).

The association of non-union and redisplacement of transcervical fractures with an

avascular femoral head is well known. Phemister (1949a) found non-union to be four times

more common in fractures with a necrotic capital fragment than in those with a live one;

and in a recent follow-up study (Brown and Abrami 1964) all of the femoral heads excised

and submitted to histology because of redisplacement were, initially, completely or partly


FIG. 28



necrotic. The interpretation of this association between non-union and necrosis is difficult.

It has been suggested (Compere and Wallace 1942) that inadequate reduction and poor

immobilisation are factors leading to death of the head and that fibrous union may result in

FIGS. 26 TO 28Case I-The gross specimen (Fig.26), slab radiograph (Fig. 27) andhistological section (Fig. 28) allshow bony union which hasoccurred in an initially almostcompletely necrotic head. Only avery small subchondral zone ofbone (marked with an arrow inFigure 26) remains dead in theupper head, and it is surrounded

by thickened trabeculae.

extension of necrosis in a partly necrotic head (Coleman and Compere 1961). The evidence,

however, is unconvincing and there were no histological features in this material to suggest

more than one episode of ischaemia. It is more likely that necrosis of the capital fragment

Case 2-Slab radiograph shows early bony union thirteen weeks afterfracture in a femoral head more than half of which had initially been dead.

Revascularisation was almost complete.

I-�_.

Figure 30-Case 3. Callus is seen bridging the fracture line. (Haemalum and eosin, � 30.) Figure 31-Case 4.Fibrous union has occurred and the head which was at the beginning half necrotic has revascularised except fora few tiny upper subchondral pockets marked with arrows. This is the only fracture in the series which was

probably, at least initially, undisplaced.


764 MARY CATTO


contributes to non-union by failure of callus formation on the head side though it is certainly

not the sole cause of non-union (Barnes l962b, 1964 ; Nicoll 1963). As is shown below, fractures

with a completely or partly necrotic capital fragment may unite.

FRACTURE HEALING IN NECROTIC FEMORAL HEADS

Although in some live or revascularised heads there was early callus formation adjacent

to the fracture site, in most of these there still remained broken trabeculae and fibrin which,

not yet organised, separated the head from the neck. In five cases femoral heads which had

been partly or totally necrotic showed progress towards union. This seems a very small number

but in many the femoral head had been removed because of redisplacement of the fracture

and most of the specimens from necropsy or primary prosthetic arthroplasty were removed

in the early weeks after fracture (Table III).

TABLE III

TIME BETWEEN FRACTURE AND REMOVAL OF FEMORAL HEAD

2-4 5-8weeks weeks

9-12weeks

13-26 � 27-52weeks weeks

12-18 � 19-24 I More thanmonths � months 24 months

Failed nails . . � 10 � 7 12 9 � 14 5 � I � 2

Primary replacement � I 2 � 12 - 4 � - I � - � -

Necropsy . . � 13 � 3 - 1 � 1 - � 1 � 1

Total . . � 35 � 22 � 12 14 � 15 6 � 2 � 3

Case I-A seventy-six-year-old woman suffered a transcervical fracture of the neck of her

right femur which was fixed with a sliding nail plate. She died of bronchopneumonia two

years and three months later following prosthetic arthroplasty for late collapse of the left

femoral head after a transcervical fracture four years before. There was sound bony union

of the right femur. On histological examination the head, which in the beginning had been

almost completely avascular, only a small foveal area being spared, had revascularised, apart

from a small superior subchondral area which was demarcated by fibrous tissue and edged

by broad trabeculae (Figs. 26 to 28).

Case 2-An eighty-three-year-old woman with osteoporosis and senile dementia sustained a

displaced subcapital fracture of her right femur, which was fixed by a sliding nail plate. She

died of bronchopneumonia thirteen weeks later. On microscopy it was seen that a large foveal

wedge ofbone (33 to 50 per cent ofthe head) had remained alive. Revascularisation was almost

complete but some tiny pockets ofdead bone remained in the subchondral region in the upper

segment of the femoral head. There was abundant vascular callus on both sides bridging the

fracture and early bony union had occurred (Fig. 29).

Case 3-This seventy-seven-year-old woman fell and fractured the neck of her femur. At

necropsy eight weeks later her death was attributed to bronchopneumonia. There had been

sparing ofa medium-sized foveal wedge and a spread ofrevascularisation to much ofthe head

although a shallow subchondral saucer of bone in the upper segment and a spur of the inferior

cortex remained necrotic. In places there was a gap between the area revascularised from the

fovea and that revascularised from the neck across the fracture line, but the fracture line

itself was already bridged by vascular callus (Fig. 30) and there seemed no reason why, given

time, bony union should not have occurred.

Case 4-A seventy-three-year-old man fell and fractured the neck of his femur twenty-two

months before his death. This fracture, unlike the others, was presumably initially impacted,


P�-w� . S :� -��. S

I � y� � � ‘� � ‘ , �1

� . ?�

. �S , � 5 � �r.�:’ ��. �..: �

.-

. .5 ‘S �

�-:�

4�1’__.;-. � S

‘ � _.‘_�‘J, .5 .....r1’�.’.

�,% :;�,�-. . S � � � . #{149}� � � 5’ - . . .

� 5% :“�‘- -.-

� �k

from the fovea. Early revascu.arisation ot deadmarrow adjacent to the fracture line is seen.

(Haemalum and eosin, x 70.)

766 MARY CATTO


but later became partially displaced. No treatment had been given and at necropsy there was

fibrous union of the fracture in its inferior part, the upper part of the fracture line remaining

ununited. Two small subchondral areas of bone necrosis were still present in the upper

segment of the head more than half of which had initially been necrotic (Fig. 3 1 ). Had this

patient’s fracture been treated by internal fixation it seems possible that more satisfactory

union might have occurred.

Case 5-This seventy-six-year-old woman sustained a displaced fracture of her femoral neck

which was manipulated and fixed by a Smith-Petersen nail. Eleven months later the nail

was extruded and, although there was no redisplacement of the fracture, the head and a small

attached trimming of the neck were removed when a prosthesis was inserted. Microscopy

showed that the entire head had been necrotic and that only about one millimetre depth of dead

marrow at the fovea had been replaced by dense fibrous tissue. Some revascularisation of the

fracture line (Figs. 32 and 33) and lower part of the head had occurred from the neck and in

places new bone had formed. Some revascularised marrow spaces contained dense collagen

and here and there along the fracture line nodules of hyaline or fibrocartilage were present,

probably indicating a barely sufficient blood supply. While it seemed possible that the fracture

might eventually unite it appeared unlikely that any substantial part of the head would be

revascularised without aid from the fovea; later the necrotic bone might have collapsed.

DISCUSSION

While it has been generally accepted that union may occur between a dead femoral head

and a live neck (Axhausen 1922, Santos 1930, Palmer 1934, Phemister 1934, Sherman and


Phemister 1947, Charnley et a!. 1957, Sevitt 1964) it is usually assumed (Nicoll 1963) that this

is later inevitably followed by collapse of the weight-bearing segment.

This certainly seems true of a completely dead head which has revascularised solely

across the fracture line with no or almost no assistance from the medial epiphysial (ligamentum

teres) arteries. It is, however, not necessarily true ofdogs (Tovee and Gendron 1954, Bonfiglio

1954, Brindley 1963) nor of human femoral heads which are initially almost completely

necrotic and have become revascularised both from across the fracture line and above from

the foveal region. These femoral heads appear capable of uniting and of being entirely

revascularised sometimes with the exception of subchondral pockets in the upper part which

are too small to allow any serious collapse of the joint surface. It seems that bony union

and a revascularised head may result from prolonged adequate fixation and revascularisation

both from across the fracture line and from the ligamentum teres vessels, especially when the initial

necrosis is only partial. When necrosis is complete and the contributions from the medial

epiphysial arteries negligible then, although union may occur, usually revascularisation is slow

and incomplete and late segmental collapse follows. With inadequate fixation, union is

unlikely, especially when the head is dead and bridging callus is only forming from the neck.

If these findings have been correctly interpreted then some difference in prognosis may

be expected between completely and partly dead heads; but unfortunately, although many

methods have been used to try and diagnose ischaemia of the femoral head at the time of

operation, they are not capable of distinguishing between complete and incomplete necrosis;

these include injection of dyes (Price 1962), radioactive substances (Tucker 1950, Arden and

Veal 1953, Boyd et a!. 1955, Arden 1958, Laing and Ferguson 1959, Boyd and Calandruccio

1963), vascular assays both venous (De Haas and McNab 1956, Hulth l958a and b, Dahlgren

1959, Harrison 1962, Johansson 1962) and arterial (Rook 1953, McGinnis, Lottes and

Reynolds 1958) and more recently oximetry (Woodhouse l962a). In some studies the assessment

of the vascular state of the femoral head depends on a bone sample sometimes taken at random

and sometimes specifically from the upper segment ofthe head (Boyd and Calandruccio 1963)

because this is the area most liable to necrosis. While this type of study is of very considerable

interest the limitations of attempting to determine the ultimate fate of the head from a small

sample is realised by these authors. As has been said above, although the upper part of the

head may be dead, if there is a live subfoveal wedge and good fixation, bony union and a

revascularised head may result. It is emphasised that initial necrosis ofonly part ofthe femoral

head does not appear to be associated with late collapse of the upper segment. An avascular

superior segment bone core alone is therefore probably insufficient evidence to justify

immediate prosthetic replacement. Indeed a bone sample taken from the subfoveal region,

though perhaps impractical because of the danger of damaging blood vessels, might give more

information on the ultimate prognosis. If a bone core from this site is ischaemic then this

almost certainly indicates a totally dead head which, though it may unite, is unlikely to become

revascularised completely without ingrowing vessels from the ligamentum teres. Collapse of

the dead bone may occur later and about half these patients develop joint symptoms severe

enough to warrant further operation (Boyd 1957).

INCIDENCE OF AVASCULAR NECROSIS AFTER TRANSCERVICAL FRACTURE

All except two of the twenty-nine patients undergoing primary prosthetic arthroplasty

more than sixteen days after fracture had this operation done because there was unavoidable

delay in treatment. This group oftwenty-seven patients, and also the necropsy group of twenty,

are unselected in respect of avascular necrosis. Of the twenty-seven femoral heads removed at

primary arthroplasty ten were thought to be alive as were six of twenty removed at necropsy,

which is a total of sixteen out of forty-seven or 34 per cent (see Table IV). This figure may be

inaccurate because the numbers are small and the delay in treatment may have caused further


768 MARY CATFO

damage to the blood vessels supplying the capital fragment. It is hoped eventually to collect

enough material to obtain a true percentage of avascular necrosis after transcervical fracture.

DISCUSSION

The incidence of avascular necrosis as assessed in clinical studies is very variable. Vv hen

patients were treated by lengthy immobilisation an apparent increase of radiological density

of the femoral head in the ununited fracture indicated its death (Santos 1930, Phemister 1934).

Phemister (1934) found an incidence of 65 per cent of necrosis in forty-nine patients, the

femoral heads of seventeen being examined histologically. Early walking has made the

radiological diagnosis of necrosis of the femoral head in ununited fractures almost impossible

TABLE IV

FINDINGS IN THE HEADS REMOVED AFTER SIXTEEN DAYS

Live----�-�-�- --_____

Number Per cent

Partly dead � Completely dead- � -�---------� �- Total

Number � Per cent � Number � Per cent

Replacements . . 10 3714 � 52 � 3 11 27

Necropsy . . 6 30 12 � 60 � 2 � 10 � 20

Total . . . 16 34 � 26 � 55 � 5 � 11 47

Failed nails . . 2 3 28 � 47 30 50 60

* Two late primary prosthetic arthroplasty cases excluded because avascular necrosis was suspected clinically.

(Boyd 1957). The lowest incidence of avascular necrosis is assessed by many authors by the

prevalence of late segmental collapse in the united fracture, Garden (1961) finding I 5 per cent,

Hargadon and Pearson 24 per cent, Cleveland and Fielding 244 per cent and Brown and

Abrami 28 per cent, in patients followed for more than a year, and Green (1960) 345 per cent

in all fractures followed for two to nine years. Linton (1944) has pointed out that the incidence

of late segmental collapse increases as the length of follow-up increases ; from 30 per cent of a

two to three-year follow-up to 56 per cent in three to seven years after fracture in his own series,

and Cauchoix and Rey (1963) found that their collapse rate increased from 255 per cent at

one year to 37�9 per cent after two years. Recently Charnley et a!. (1957) in a series of thirty-

three cases ofdisplaced fracture treated by a compression screw considered, because of extrusion

of the screw, that some degree of vascular damage was present in two-thirds of the cases.

In addition to clinical studies, Boyd and Calandruccio (1963) examined by autoradiography

femoral heads removed at primary and secondary arthroplasty after phosphorus32

administration and found that two-thirds showed loss ofsome vascularity; Woodhouse (1962a

and b) said the same on examining femoral heads removed at primary arthroplasty from

patients given tetracycline. In twenty-four speicmens removed at necropsy Sevitt (1964) found

arteriographic and histological evidence ofsome degree offemoral head necrosis in twenty-one.

In the small unselected group offemoral heads (forty-seven) in the present study it appears

that about one-third remained viable and about two-thirds were partly or completely necrotic

(Table IV).

CORRELATION OF HISTOLOGICAL AND RADIOGRAPHIC APPEARANCES

Ununited fractures-Although Santos (1930) and Phemister (1934, 1939, 1940, 1943, 1948)

found that necrotic ununited femoral heads showed, usually within six months of injury, a

relative increase in density compared with the adjacent osteoporotic pelvis and distal femur,



this was not seen in the clinical radiographs in this series since early mobilisation presumably

prevented the development of local osteoporosis. Necrosis of the femoral head may not be

attended by any change in radiological density (De Haas and MacNab 1956, Boyd 1957,

Charnley et a!. 1957, Bonfiglio and Bardenstein 1958, Bessler and Muller 1961, Hulth 1961,

Woodhouse l962a). Certainly in these patients with dead femoral heads the extrusion of the

nail and disruption ofthe fracture could not have been forecast on the radiographic appearances.

After revascularisation of dead marrow there may be osteoclastic resorption of dead bone

and laying down of new bone on the surface of dead trabeculae. The density to x-rays of

the reossifying area depends on the ratio of these two activities. While Santos (1930),

Phemister (1939, 1940, 1943, 1948, 1949) and Sherman and Phemister (1947) described a

decrease of density in the revascularised area of ununited femoral heads it was recognised by

them that this was probably because of lack of function in the hip. In the revascularising

capital fragments of this series slab radiographs showed no evidence that revascularisation

was accompanied by increased radiotranslucency; indeed there was commonly a slight increase

in the trabecular thickness except in the subfoveal zone which in the normal contains thin and

scanty trabeculae. This lack of porosis in the reossified bone was probably also the result of

relatively early walking. In some patients there was much broadening of the reossified

trabeculae and this was specially so at the fracture line when it was covered by dense collagen

and fibrocartilage and at the vascularisation front when it had become more fibrous and

avascular histologically (Figs. 34 to 36). To summarise, there was a tendency towards an

absolute increase ofdensity to x-rays in the revascularised area (Woodhouse l962a), slight and

unremarkable in clinical radiographs in most cases but more marked where the process

appeared to have stopped. This was caused by broadened trabeculae where new bone had

been laid down on the surface of unresorbed dead bone as described by Hulth (1961) and in

rabbits by Bobechko and Harris (1960). Slight marrow calcification was seen in only two

femoral heads (Fig. 37).

United fractures-No great alteration in density of the slab radiographs of the necrotic and

revascularising femoral heads was found in two of the three patients whose fractures were

progressing towards bony union, as in Figure 29, for example. In Case 1 a very small

subchondral area of bone which remained necrotic was edged by dense fibrous tissue and

broad bone trabeculae (Figs. 26 to 28). In twelve patients with bony union and late

segmental collapse reported elsewhere in this issue it was notable that alteration of the contour

of the weight-bearing surface was the first radiological evidence of bone necrosis (Bessler and

Muller 1961, Barnes l962a, Woodhouse 1962a). A zone of increased radiological density

caused by thick reossified trabeculae was often found later in the most proximal part of the

revascularised area, especially when this had become densely fibrous and apparently without

any possibility of progress. The very thick trabeculae, causing increased density to x-rays,

and found when further vascularisation is frustrated may be analogous to those seen in relation

to the bone ends in a pseudarthrosis (Judet, Judet and Roy-Camille 1958).

THE LIGAMENTUM TERES

Clinical appearances-In thirty-four of the eighty-nine patients who had excision of a femoral

head more than sixteen days after injury a note stated whether or not the ligamentum teres

bled on section. In many there was bleeding from the acetabular end when there was none

from the foveal end of the ligament. Although the ligament from four of the five completely

dead heads which had not become revascularised did not bleed and those from four of five

live heads bled briskly, in the remaining twenty-four there was no very close correlation

between the amount of bleeding found at operation and the state of vascularity or

revascularisation of the femoral head.


. S ::�: �

S � �

. -SS�S5

FIG. 34

Figure 34-Revascularisation has occurred from the fovea but is incomplete in depth. Broad trabeculae areseen at the site where revascularisation has stopped and these are visible as areas of increased density on the

slab radiograph (Fig. 35).

S..’

:� � � � � #{149}� � �

Figure 36-The broad trabeculae have central cores of dead bone covered by a mass of new, living bone.(Haemalum and eosin, x 60.) Figure 37- Marrow calcification is present at the revascularisation border in this

femoral head. There is no bony reaction. (Haemalum and eosin, � 75.)


770 MARY CATTO

Histological appearances-In only twenty-four of 109 femoral heads removed more than

sixteen days after fracture was the whole ligamentum teres available for study but in a further

forty-five a small stump was still attached at the fovea. For comparison the ligament was

examined from fifty femoral heads selected at random at necropsy from patients over the age

of sixty-four, fifteen from necropsies on infants and children under fourteen, and nine from

the intervening ages. More than half the ligaments from the elderly controls showed hyaline


I!’

,t.

I ‘

- -

:S

#{149}?

&-� 0

,,, I

S �

FIG. 38 FIG. 39Figure 38-Sclerotic blood vessels with concentr.c fibrosis and very narrow lumina are shown in the ligamentumteres. (Haemalum and eosin, x 140.) Figure 39-A completely obliterated vessel in the ligamentum teres is seen

adjacent to several vascular channels in an 18-month-old child. (Haemalum and eosin, x 225.)

FIG. 40 FIG. 41

Figure 40-A completely obliterated vessel in the ligamentum teres still shows elastica in the wall. (Weigert’selastica, >. 225.) Figure 41-A vein from the ligamentum teres of a 4-month-old infant shows early hyaline

sclerosis. (Haemalum and eosin, x 315.)


L

� ‘��I

4!

772 MARY CATTO

sclerosis chiefly affecting small veins. These appearances have been described in detail by

Elmore, Malmgren and Sokoloff (1963) in synovial blood vessels from many sites especially

in the acetabular fat pad and prepatellar fat. Sclerosis in the blood vessels of the ligament was

mentioned briefly by Chandler and Kreuscher (1932) and by Nordenson (1938). The earliest

change found in the present series was an eccentric infiltration of hyaline material between

the smooth muscle fibres of the media of venules. Later, the whole media became replaced

by hyaline material (Fig. 38) and eventually the lumen might become obliterated. The vessel

was then a solid mass of hyaline substance bounded by the external elastic lamina (Figs. 39

and 40). Perivascular fibrosis was not seen. The distribution of these lesions was patchy, a

group of obliterated vessels sometimes lying adjacent to normal ones. Special stains showed

that the hyaline material was not amyloid, fibrin or any of the material described by Lendrum

et a!. (1962) as being intermediate between fibrin and collagen, but it did, in all its stages, stain

as collagen. The hyalinisation was sometimes accompanied by reduplication of elastic laminae.

Sclerosis was found also in the vessels ofthe ligamentum teres in two ofthe fifteen children

aged four months and eighteen months respectively (Figs. 41 and 39), in the few ligaments at

intermediate ages and in the ligaments and retinaculum after fracture. The reason for describing

these patchy sclerotic vascular changes is that they might, in the absence ofcontrol material, be

ascribed to damage of the blood vessels at the time of fracture, or be thought in their later

obliterative phase to involve principally arteries and thereby influence the viability or

revascularisation of the femoral head. While the number of ligaments available after fracture

is too small to attempt any correlation with the vascular state of the head, it seems unlikely

that obliteration of small groups of veins would have any notable effect.

It was striking that there was no vascular thrombosis in any of the sixty-nine ligaments,

although occasionally a thrombosed vessel was seen in an attached retinacular tag at the

periphery of the head. The only remarkable vascular change was that the vessels of the foveal

stump of the ligamentum teres were completely necrotic in ten out of eleven necrotic femoral

heads which had failed to revascularise and in three in which revascularisation was confined

to tiny areas at the periphery of the head at the fracture line. In one of the cases in which

the fovea was necrotic the whole ligamentum teres was also available and showed necrosis of

all tissue to a depth of about half a centimetre, the more proximal parts of the vessels being

alive and patent.

CHANGES IN THE ARTICULAR CARTILAGE

Degenerative changes-The fifty normal elderly controls were examined in relation to

degenerative joint changes. The articular cartilage of none of the femoral heads was entirely

normal. The least severe changes were those of superficial flaking and of fibrous replacement

of the surface cartilage around the fovea and at the periphery of the head. Osteoarthritic

changes of fibrillation and cartilage loss were most frequent in the lower head. Three patients

showed advanced osteoarthritis with osteophyte and “ cyst “ formation. Of the 109 femoral

heads removed after fracture none showed advanced osteoarthritis but there was no other

appreciable difference in respect of degenerative changes between this and the control group.

Loss of chondrocytes-In normal elderly subjects some patchy loss of chondrocytes occurred

especially in the deeper layers of the articular cartilage. In the femoral heads removed after

fracture, changes in the cartilage covering necrotic bone were usually slow to develop and the

patchy normal loss of chondrocytes was rarely exceeded until months and sometimes years

had passed. The loss exceeded normal in twenty-five of the 109 femoral heads and was most

marked in the deep zone of the weight-bearing area of the head. In only one head, removed

more than three years after fracture, was there almost complete loss of chondrocytes. It was

remarkable that in all these specimens the cartilage kept its normal depth and contour and

showed no evidence of disintegration (Hatcher 1952, Hulth 1961).

Nicoll (1963) perhaps took too gloomy a view in suggesting that the articular cartilage

of a dead femoral head is “ doomed from the start.” Cellular loss is usually slow and Phemister

ThE JOURNAL OF BONE AND JOINT SURGERY


(1934) believed that if revascularisation of the underlying bone occurred reasonably quickly

the cartilage would survive. Even if it eventually dies it may fail to disintegrate. The minor

cartilaginous changes seen in these femoral heads with normal contours seemed unlikely to

give rise to symptoms. This is in contrast to the severe changes which may occur in late

segmental collapse once deformity of the joint surface is established. In these cases while the

cartilage covering the still unrevascularised upper segment usually retains its thickness, that

covering the revascularised bone at the periphery of the head may show severe osteoarthritis.

Vascularisation ofthe cartilage from below, resumption ofendochondral ossification, formation

of osteophytes, osteoarthritic “ cysts “ and sometimes formation of a new joint surface may

occur-changes which inevitably give rise to joint symptoms.

SUMMARY

1 . Loss of osteocytes in the bone trabeculae of the femoral heads of “ normal “ elderly

patients was patchy and distinguishable from that resulting from avascular necrosis after fracture.

2. Changes in the haemopoietic marrow were the earliest and most sensitive indicators of

ischaemia, loss of osteocytes rarely being complete until three or four weeks after fracture.

3. In 109 femoral heads removed more than sixteen days after fracture the viability could be

determined by histological means. All of these had suffered some damage to the vascular

supply but in a number the head remained alive apart from the region of the fracture line.

These heads were nourished by the blood vessels of the ligamentum teres and sometimes by

retinacular arteries, usually of the inferior group.

4. Some femoral heads became completely necrotic following fracture, others were only partly

affected. A variable amount of the subfoveal region commonly remained alive and it was

from this site that revascularisation spread into the head. The upper segment of the femoral

head least often remained alive and its subchondral region was usually the last to revascularise.

5. In a group of unselected femoral heads a third remained alive following fracture and two-

thirds were partly or completely necrotic.

6. Femoral heads which were partly necrotic appeared capable of uniting and completely

revascularising, there being invasion of the necrotic bone by vessels from across the fracture

line and from the ligamentum teres. This contrasted with the completely necrotic femoral

heads described elsewhere in this issue which united but in the absence of proliferation of

ligamenturn teres vessels failed to revascularise completely and developed late segmental collapse.

7. Avascular necrosis did not appear to be the sole cause of non-union.

8. Necrotic bone showed no alteration in radiological density. Reossifying bone in areas of

revascularisation sometimes caused an absolute increase of radiodensity especially when

associated with halted revascularisation. This increase of radiological opacity was the result

of deposition of new on dead bone with broadening of the trabeculae. Marrow calcification

was minimal.

9. Obliterative sclerosis of venules in the ligamentum teres was found in “ normal “ patients

even in infancy. No thrombosis was seen in the ligaments following fracture butwhere the femoral

heads were completely necrotic and not revascularised the ligaments were often also necrotic.

10. There appeared to be no increase in degenerative changes in the articular cartilage of the

femoral heads following fracture compared with fifty elderly controls. Some loss of chondrocytes

in the deep zone of the weight-bearing area was found in about a quarter of the femoral heads.

In only one head was the cartilage almost completely acellular. An almost normal depth and

a smooth contour of the articular cartilage were retained.

My grateful thanks are due to the orthopaedic surgeons of the Western Infirmary and Southern GeneralHospital, Glasgow, for their interest and for access to their records; to the medical staff of the Glasgow RoyalMental Hospital and to Dr Rhoda Taylor of Foresthall Hospital for autopsy material; to Dr A. M. McDonaldof the Royal Hospital for Sick Children for the samples ofjuvenile ligamentum teres; to Mr Matthew Findlayfor the histological preparations ; and to Mr George Kerr for the photographs. I am particularly indebted toProfessor Roland Barnes and Mr J. T. Brown for their help, encouragement and constructive criticism and toMr W. Sillar for many of the early arthroplasty specimens. Part of the expenses of this study were borne bya grant from the Advisory Committee on Medical Research of the Department of Health for Scotland.


774 MARY CATTO

REFERENCES

ARDEN, G. P. (1958): Modern Trends in the Treatment of the Fractured Neck of Femur. Postgraduate Medical

Journal, 34, 541.ARDEN, G. P., and VEALL, N. (1953): The Use of Radioactive Phosphorus in Early Detection of Avascular

Necrosis in the Femoral Head in Fractured Neck of Femur. Proceedings of the Royal Society of Medicine

(Section of Orthopaedics), 46, 344.

AXHAU5EN, G. (1922): Die Nekrose des proximalen Bruchst#{252}cks beim Schenkalhalsbruch und ihre Bedeutungf#{252}rdes H#{252}ftgelenk. Archivfur Klinische Chirurgie, 120, 325.

BADGLEY, C. E. (1960): Fractures of the Hip Joint-Some Causes for Failure and Suggestions for Success.Instructional Course Lectures, American Academy of Orthopaedic Surgeons, 17, 106.

BARNES, R. (1962a) : Intracapsular Fractures of the Neck of the Femur. In Modern Trends iii Orthopaedics 3.London: Butterworths.

BARNES, R. (196Th): The Diagnosis of Ischaemia of the Capital Fragment in Femoral Neck Fractures. Journal

of Bone and Joint Surgery, 44-B, 760.

BARNES, R. (1964): The Unsolved Fracture. Scottish Medical Journal, 9, 45.BESSLER, W., and MULLER, M. (1961): Le diagnostic pr#{233}coce de la n#{233}crosede Ia t#{234}tef#{233}morale. Annales de

Radiologie (Paris), 4, 21.BOBECHKO, W. P., and HARRIS, W. R. (1960): The Radiographic Density of Avascular Bone. Journal of Bone

and Joint Surgery, 42-B, 626.BONFIGLIO, M. (1954): Aseptic Necrosis of the Femoral Head in Dogs : Effect of Drilling and Bone Grafting.

Surgery, Gynecology and Obstetrics, 98, 591.BONFIGLIO, M., and BARDENSTEIN, M. B. (1958): Treatment by Bone-Grafting of Aseptic Necrosis of the

Femoral Head and Non-union of the Femoral Neck (Phemister Technique). Journal of Bone and Joint

Surgery, 40-A, 1329.

BOYD, H. B. (1957): Avascular Necrosis of the Head of the Femur. Instructional Course Lectures, American

Academy of Orthopaedic Surgeons, 14, 196.BOYD, H. B., and CALANDRUCCIO, R. A. (1963): Further Observations on the Use of Radioactive Phosphorus

(P32) to Determine the Viability of the Head of the Femur; Correlation of Clinical and ExperimentalData in 130 Patients with Fractures of the Femoral Neck. Journal of Bone and Joint Surgery, 45-A, 445.

BOYD, H. B., ZILVERSMIT, D. B., and CALANDRUCCIO, R. A. (1955): The Use of Radio-active Phosphorus (P32)to Determine the Viability of the Head of the Femur. Journal of Bone and Joint Surgery, 37-A, 260.

BRINDLEY, H. H. (1963): Avascular Necrosis of the Head of the Femur. An Experimental Study. Proceedings

of the American Academy of Orthopaedic Surgeons in Journal of Bone and Joint Surgery, 45-A, 1541.

BROWN, J. T., and ABRAMI, G. (1964) : Transcervical Femoral Fracture. Journal of Bone and Joint Surgery,

46-B, 648.BROWN, J. T., and CATrO, M. (1964): Unpublished work.

CAMPBELL, C. J. (1961) : Aseptic Necrosis of Bone. Instructional Course Lectures, America,z Academy of

Orthopaedic Surgeons, 18, 234.CAucHoIx, J., and REY, J. C. (1963): Sur le traitement des fractures r#{233}centesdu cot du femur. (Une s#{233}riede

203 ens cons#{233}cutifs.) Revue de Chirurgie Orthop#{233}dique, 49, 315.CAVE, E. F. (1960): Fractures of the Femoral Neck. Instructional Course Lectures, America,: Academy of

Orthopaedic Surgeons, 17, 79.CHANDLER, S. B., and KREUSCHER, P. H. (1932): A Study ofthe Blood Supply ofthe Ligamentum Teres and its

Relation to the Circulation of the Head of the Femur. Journal of Bone and Joint Surgery, 14, 834.

CFIARNLEY, J., BLOCKEY, N. J., and PURSER, D. W. (1957): The Treatment of Displaced Fractures of the Neckof the Femur by Compression. Journal ofBone and Joint Surgery, 39-B, 45.

CHEYNEL, J. (1954) : Introduction a la Pratique Chirurgicale de Ia Hanche. Recherches Anatomiques Personelle.

Deductions Chirurgicales Pratiques. Paris: G. Doin et cie.

CHRISTOPHE, K., HOWARD, L. 0., POrrER, T. A., and DRISCOLL, A. J. (1953): A Study ofOne Hundred and FourConsecutive Cases of Fracture of the Hip. Journal ofBone and Joint Surgery, 35-A, 729.

CLAFFEY, T. J. (1960): Avascular Necrosis of the Femorat Head. An Anatomical Study. Journal ofBone and

Joint Surgery, 42-B, 802.

CLEVELAND, M., and FIEt..rnNG, J. W. (1954): A Continuing End-Result Study of Intracapsular Fracture of theNeck of the Femur. Journal ofBone and Joint Surgery, 36-A, 1020.

COLEMAN, S. S., and COMPERE, C. L. (1961): Femoral Neck Fractures : Pathogenesis of Avascular Necrosis,Nonunion and Late Degenerative Changes. Clinical Orthopaedics, 20, 247.

COMPERE, E. L., and WALLACE, G. (1942): Etiology of Aseptic Necrosis of the Femur after TranscervicalFracture. Journal of Bone and Joint Surgery, 24, 831.

CRAWFORD, H. B. (1960): Conservative Treatment of Impacted Fractures of the Femoral Neck. A Report ofFifty Cases. Journal ofBone and Joint Surgery, 42-A, 471.

DAHLGREN, S. (1959): Venography in Fractures of the Femoral Neck. Acta Chfrurgica Scandinavica, 117, 494.




DE HAAS, W., and MACNAB, I. (1956): Fractures of the Neck of the Femur; a Method ofAssessing the Viability

of the Femoral Head. South African MedicalJournal, 30, 1005.ELMORE, S. M., MALMGREN, R. A., and SOKOLOFF, L. (1963): Sclerosis of Synovial Blood Vessels. Journal of

Bone andJoint Surgery, 45-A, 318.GARDEN, R. S. (1961): Low-Angle Fixation in Fractures of the Femoral Neck. Journal ofBone andJoint Surgery,

43-B, 647.GREEN, J. T. (1960): Management of Fresh Fractures of the Neck of the Femur. instructional Course Lectures,

American Academy of Orthopaedic Surgeons, 17, 94.HAM, A. W., and LEESON, T. S. (1961): Histology. Fourth edition, p. 329. London: Pitman Medical Publishing

Co. Ltd.HARGADON, E. J., and PEARSON, J. R. (1963): Treatment of Intracapsular Fractures of the Femoral Neck with

the Charnley Compression Screw. Journal ofBone and Joint Surgery, 45-B, 305.HARRISON, M. H. M. (1962): A Preliminary Report of Vascular Assay in Prognosis of the Fractured Femoral

Neck. Journal ofBone and Joint Surgery, 44-B, 858.HARTY, M. (1953): Blood-Supply of the Femoral Head. British Medical Journal, ii, 1236.

HATCHER, C. H. (1952): The Role of Aseptic Necrosis of Bone in Skeletal Disease and Injuries. Chapter 16.In The Musculoskeletal System. Edited by M. Ashford. New York : MacMillan.

HODGES, P. C. (1954): Normal Bone, Diseased Bone, Dead Bone. American Journal of Roentgenology,

71, 925.HULTH, A. (1958a): Femoral-Head Phlebography: a Method of Predicting Viability. Journal ofBone and Joint

Surgery, 40-A, 844.HULTH, A. (1958b): The Vessel Anatomy of the Upper Femur End with Special Regard to the Mechanism of

Origin of Different Vascular Disorders. Acta Orthopaedica Scandinavica, 27, 192.HULTH, A. (1961): Necrosis ofthe Head ofthe Femur. A Roentgenological, Microradiographic and Histological

Study. Acta Chirurgica Scandinavica, 122, 75.JAFFE, H. L., and POMERANZ, M. M. (1934): Changes in the Bones of Extremities Amputated Because of

Arteriovascular Disease. Archives of Surgery, 29, 566.JOHANSSON, S. H. (1962): Prognostic Assessment in Fractured Neck of Femur Using 131 I and Venography.

Acta Chirurgica Scandinavica, 123, 298.JUDET, J., JUDET, R., LAGRANGE, J., and DUNOYER, J. (1955): A Study of the Arterial Vascularisation of the

Femoral Neck in the Adult. Journal ofBone and Joint Surgery, 37-A, 663.

JUDET, R., JUDET, J., and ROY-CAMILLE, R. (1958) : La vascularisation des pseudarthroses des os longs d’apr#{232}sune #{233}tudeclinique et exp#{233}rimentale. Revue de Chirurgie Orthop#{233}dique, 44, 381.

KOLODNY, A. (1925): The Architecture and the Bloodsupply of the Head and Neck of the Femur and theirImportance in the Pathology of Fractures of the Neck. Journal of Bone and Joint Surgery, 7, 575.

LAING, P. G., and FERGUSON, A. B., Jun. (1959): Radiosodium Clearance Rates as Indicators of Femoral-HeadVascularity. Journal of Bone and Joint Surgery, 41-A, 1409.

LENDRUM, A. C., FRASER, D. S., SLIDDERS, W., and HENDERSON, R. (1962): Studies on the Character and

Staining of Fibrin. Journal of Clinical Pathology, 15, 401.LINTON, P. (1944) : On the Different Types of Intracapsular Fractures of the Femoral Neck. Acta Chirurgica

Scandinavica, Supplementum 86.MCGINNIS, A. E., LomS, J. 0., and REYNOLDS, F. C. (1958): Femoral Arteriograms; use to Demonstrate

Circulation of the Hip Following Neck Fractures of the Femur. Missouri Medicine, 55, 31.MCLEAN, F. C., and URIST, M. R. (1955): Bone. An Introduction to the Physiology ofSkeletal Tissue. Chicago:

University of Chicago Press.

MATHON, J. F. (1959): Etude de l’architecture et de Ia vascularisation du col et de Ia t#{234}tedu femur. Revue du

Practicien, 9, 795.MERLE D’AUBIGNE, R., and CORMIER (1956): N#{233}crose traumatique de Ia t#{234}tedu femur en dehors des

pseudarthroses. Revue de Chirurgie Orthop#{233}dique, 42, 246.MEYER, P. C. (1956): The Histological Identification of Osteoid Tissue. Journal ofPathology and Bacteriology,

71, 325.NICOLL, E. A. (1963): The Unsolved Fracture. Journal ofBone and Joint Surgery, 45-B, 239.NORDENSON, N. G. (1938): Sur Ia vascularisation de Ia t#{234}tedu femur par Ia voie du ligament rond f#{233}moral.

Lyon Chirurgical, 35, 178.PALMER, I. (1934): (Jber die Ausheilungsbedingungen der medialen SchenkelhalsbrUche nach Osteosynthese

mit einem Nagel aus rostfreiem Stahl, illustriert durch eine histologische Untersuchung. Acta Chfrurgica

Scandinavica, 75, 416.PATRICK, J. (1960): Avascular Necrosis and the Head of the Femur. Proceedings of the British Orthopaedic

Association in Journal of Bone and Joint Surgery, 42-B, 650.PHEMISTER, D. B. (1930): Repair of Bone in the Presence of Aseptic Necrosis Resulting from Fractures,

Transplantations and Vascular Obstruction. Journal of Bone and Joint Surgery, 12, 769.

776 MARY CATFO

PHEMISTER, D. B. (1934) : Fractures of Neck of Femur, Dislocations of Hip, and Obscure Vascular DisturbancesProducing Aseptic Necrosis of Head of Femur. Surgery, Gynecology and Obstetrics, 59, 415.

PHEMISTER, D. B. (1939): The Pathology of Ununited Fractures ofthe Neck ofthe Femur with Special Reference

to the Head. Journal ofBone and Joint Surgery, 21, 681.PHEMISTER, D. B. (1940): Changes in Bones and Joints Resulting from Interruption of Circulation. I. General

Considerations and Changes Resulting from Injuries. Archives ofSurgery, 41, 436.

PHEMISTER, D. B. (1943): Circulatory Disturbances in the Head of the Femur. American Academy of Orthopaedic

Surgeons Lectures, 1, 129.PHEMISTER, D. B. (1948): Lesions of Bones and Joints Arising from Interruption of the Circulation. Journal of

the Mount Sinai Hospital, 15, 55.

PHEMISTER, D. B. (1949a): Treatment of the Necrotic Head of the Femur in Adults. Journal ofBone and Joint

Surgery, 31-A, 55.PHEMISTER, D. B. (1949b) : In discussion on paper by PATRICK, J. Intracapsular Fractures of the Femur Treated

with a Combined Smith-Petersen Nail and Fibutar Graft. Journal of Bone and Joint Surgery, 31-A, 67.PRICE, E. R. (1962): The Viability of the Femoral Head after Fracture of the Neck of the Femur: a Preliminary

Communication on a Dye Clearance Technique. Journal of Bone and Joint Surgery’, 44-B, 854.RooK, F. W. (1953): Arteriography of the Hip Joint for Predicting End Results in Intracapsular and

Intertrochanteric Fractures of the Femur. American Journal of Surgery, 86, 404.

RUSSELL, N. L. (1956): A Rapid Double-embedding Method for Tissues Using an Automatic Tissue Processing

Apparatus (Histokinette-Histokine). Journal of Medical Laboratory Technology, 13, 484.

RUTISHAUSER, E., and MAJNO, G. (1951): Physiopathology of Bone Tissue. Bulletin of the Hospital for Joi,zt

Diseases, 12, 468.SANT05, J. V. (1930): Changes in the Head of the Femur After Complete Intracapsular Fracture of the Neck.

Archives ofSurgery, 21, 470.SCHMORL, G. (1924): Die pathologische Anatomie der Schenkethalsfrakturen. M#{252}nchener Medizinische

Wochenschrift, 71, 1381.SEVITT, S. (1964): Avascular Necrosis and Revascularisation ofthe Femoral Head after Intracapsular Fractures.

Journal ofBone and Joint Surgery, 46-B, 270.SHERMAN, Mary S. (1947): In discussion on paper by BoYD, H. B., and GEORGE, I. L. Complications of Fractures

of the Neck of the Femur. Journal ofBone and Joint Surgery, 29, 327.SHERMAN, Mary S., and PHEMISTER, D. B. (1947): The Pathology of Ununited Fractures of the Neck of the

Femur. Journal of Bone and Joint Surgery, 29, 19.SHERMAN, Mary S., and SELAKOVICH, W. G. (1957): Bone Changes in Chronic Circulatory Insufficiency. Journal

ofBone and Joint Surgery, 39-A, 892.SMITH, F. B. (1959): Effects of Rotatory and Valgus Malpositions on Blood Supply to the Femoral Head:

Observations at Arthroplasty. Journal ofBone andJoint Surgery, 41-A, 800.TOVEE, E. B., and GENDRON, E. (1954): The Use of Radioactive Phosphorus in the Determination of the

Viability of the Femoral Head in Dogs after Subcapitat Fractures. Journal of Bone and Joint Surgery,

36-A, 185.

TRUETA, J., and HARRISON, M. H. M. (1953): The Normal Vascular Anatomy of the Femoral Head in AdultMan. Journal of Bone and Joint Surgery, 35-B, 442.

TUCKER, F. R. (1949): Arterial Supply to the Femoral Head and its Clinical Importance. Journal of Bone and

Joint Surgery, 31-B, 82.TUCKER, F. R. (1950): The Use of Radioactive Phosphorus in the Diagnosis of Avascular Necrosis of the

Femoral Head. Journal of Bone and Joint Surgery, 32-B, 100.WHITMAN, R. (1945): A Retrospective Commentary on the Campaign for the Establishment of the Positive

Standard of Treatment for Fracture of the Neck of the Femur. Journal ofBone and Joint Surgery, 27, 334.WoLCorr, W. E. (1943): The Evolution of the Circulation in the Developing Femoral Head and Neck : an

Anatomic Study. Surgery, Gynecology and Obstetrics, 77, 61.WOODHOUSE, C. F. (1962a): Anoxia of the Femoral Head. Surgery, 52, 55.

WOODHOUSE, C. F. (196Th): Tetracycline Vascular Maps of the Femoral Head. Proceedings of Orthopaedic

Research Society in Journal of Bone and Joint Surgery, 44-A, 1029.


ahistological study ofavascular necrosis of the …€¦ · “ avascular necrosis (whitman 1945;...

Documents