Hip

The flexion gap in normal knees

AN MRI STUDY

Y. Tokuhara, Varus and valgus joint laxity of the normal living knee in flexion was assessed using MRI.
Y. Kadoya, Twenty knees were flexed to 90˚ and were imaged in neutral and under a varus-valgus stress
S. Nakagawa, in an open MRI system. The configuration of the tibiofemoral joint gap was studied in slices
A. Kobayashi, which crossed the epicondyles of the femur.
K. Takaoka When a varus stress was applied, the lateral joint gap opened by 6.7 ± 1.9 mm (mean ± SD;
2.1 to 9.2) whereas the medial joint gap opened by only by a mean of 2.1 ± 1.1 mm (0.2 to 4.2).
From Osaka City These discrepancies indicate that the tibiofemoral flexion gap in the normal knee is not
University Medical rectangular and that the lateral joint gap is significantly lax. These results may be useful for
School, Osaka City, adequate soft-tissue balancing and bone resection in total knee arthroplasty and
Japan reconstruction surgery on ligaments.

Y. Tokuhara, MD,
Orthopaedic Surgeon

S. Nakagawa, MD,
Orthopaedic Surgeon

A. Kobayashi, MD,
Lecturer

K. Takaoka, MD, Professor
Department of Orthopaedic
Surgery, Osaka City
University Medical School,
1-4-3 Asahimachi, Abeno-ku,
Osaka City, Osaka 545-8585,
Japan.

Y. Kadoya, MD, Chief of
Adult Knee Reconstruction
Department of Orthopaedic
Surgery, Osaka Rosai
Hospital, 1179-3 Nagasone-
cho, Sakai City, Osaka 591-
8025, Japan.
Correspondence should be
sent to Dr Y. Kadoya.
©2004 British Editorial
Society of Bone and
Joint Surgery
doi:10.1302/0301-620X.86B8.
15246 $2.00
J Bone Joint Surg [Br]
2004;86B1:1133-6.
Received 24 December 2003;
Accepted after revision
30 April 2004
A knowledge of the kinematics of the knee is
important for understanding the pathogenesis
and treatment of disease, particularly when
treatment involves surgical reconstruction.
Recent studies, using modern techniques, have
challenged some traditional concepts, such as
the four-bar link model of movement of the
human knee and obligatory femoral rollback.1-3
One such established concept is the prepara-
tion of a rectangular joint gap in total knee
arthroplasty (TKA). This has been regarded as
an important goal in achieving restriction of
function of the knee.4-6 It has, however, been
empirically recognised that the lateral tibio-
femoral articulation is physiologically lax and
as a consequence, the flexion gap in normal
knees may not be rectangular.7-9 Because of
technical difficulties, few quantitative data are
available on the physiological laxity of the
flexed, living knee. Such an analysis can only
be performed if the flexed knee is imaged three-
dimensionally, both in a neutral position and
under a varus-valgus stress. Recent develop-
ments in MRI have allowed the living knee to
be examined in a variety of positions and flex-
ion angles. Our aim was therefore to examine
the varus-valgus laxity of the flexed knee using
MRI, with reference to the tibiofemoral flexion
gap.
Subjects and Methods
Imaging methods. Our subjects were 20 nor-
mal volunteers, ten men and ten women with a
mean age of 27.2 years (18 to 53). They had no
symptoms in the knee and radiographs were
normal. The left knee was scanned in an open
MRI unit (Airis; Hitachi, Tokyo, Japan) using
a T1-weighted sequence (THR, 500 ms; TE, 20
ms; and the flip angle, 90˚). Consecutive coro-
nal scans of 3 mm in thickness were obtained
across the distal femur and knee. Imaging was
performed in the neutral position and under a
passive varus-valgus stress at 90˚ of flexion
(Fig. 1). Images under varus stress were
obtained with the subject lying on his/her back
in the unilateral cross-legged position (Fig. 1a).
The varus stress was applied by the weight of
the lower leg. Imaging under valgus stress was
performed in the unilateral reversed-cross leg
position (Fig. 1b) and that imaging in the neu-
tral position performed in the left lateral posi-
tion (Fig. 1c). During imaging, no special
attention was paid to the rotational position of
the tibiofemoral joint.
Methods of measurement. The slice which
crossed the lateral and medial epicondyles of
the femur and/or the longitudinal axis of the
tibia was selected. Of the 60 image sequences
obtained (20 knees in three positions), the epi-
condyles and the longitudinal axis of the tibia
featured in the same slice in 19 sequences. In
the remaining 41 image sequences the epi-
condyles and the longitudinal axis of the tibia
were not in the same slice and two different
slices were used for the measurement.
The shape and width of the flexion gap were
assessed, based upon four reference lines (A, B,
C and D) and two reference points (E and F)

VOL. 86-B, No. 8, NOVEMBER 2004 1133

1134 Y. TOKUHARA, Y. KADOYA, S. NAKAGAWA, A. KOBAYASHI, K. TAKAOKA

Fig. 1a Fig. 1b Fig. 1c

Scanning positions under a) a passive varus stress, b) a passive valgus stress and c) in a neutral position.

Medial Lateral

A A

B B
E E
Fig. 2
C F C
F Diagram and MR scan showing the reference lines
and reference points which were used to assess the
D shape and width of the flexion gap. Detailed defini-
D tions are given in the text.

(Fig. 2). Line A was the epicondylar axis. This connected
the lateral epicondylar prominence either to the medial sul-
cus (surgical epicondylar axis (SEA)) or to the medial prom-
inence (clinical epicondylar axis (CEA)). The SEA was
defined as line A if it could be seen but the CEA was used
when the SEA could not be identified. Fourteen SEAs and
20 CEAs could be reproducibly determined and 14 SEAs
and the six CEAs were used as line A. Line B was the pos-
terior condylar axis (PCA) which connected the posterior
margins of the lateral and medial femoral condyles. Line C
was the tibial articular axis (TAA) which connected the
proximal margins of the lateral and medial articular sur-

THE JOURNAL OF BONE AND JOINT SURGERY

 THE FLEXION GAP IN NORMAL KNEES
Fig. 3a
Representative MR scans in a) a neutral position, b) under a passive varus stress and c) a passive valgus stress.
faces of the tibia. Line D was the tibial osteotomy axis
(TOA) and was perpendicular to the longitudinal axis of
the tibia, which was determined by connecting the mid-
point of the width of the tibial diaphysis at two levels.
Points E and F were the posterior tips of the lateral and
medial femoral condyles.
The shape of the flexion gap in neutral and under a
varus-valgus stress was assessed using an angle between line
A and line D. This corresponded to the configuration of the
flexion gap when an ideal bone resection in the proximal
tibia and posterior femoral condyle was performed at TKA.
The angle between line B and line C was also determined
because this represented the physiological separation of the
tibiofemoral articular surfaces. The width of the lateral and
medial joint gaps was also measured as the distance
between point E and point F and line C. The flexion angle
was measured from the sagittal images of each knee.
The data were transferred to a personal computer (Apple
Macintosh; Apple Computer Inc, Cupertino, California)
and the angles and distances were measured using image
analysis software (National Institute of Health image). The
angles were expressed as a positive value when the femoral
reference lines (lines A and B) were internally rotated rela-
tive to the reference lines for the tibia (line C and D).
Results
Representative images of the knees in neutral and under a
varus-valgus stress are shown in Figure 3.
When no stress was applied, the lateral and medial fem-
oral condyles were in contact with the tibial articular sur-
face (width of the lateral and medial flexion gaps = 0 mm)
with a mean flexion angle of 87.2 ± 3.5˚ SD (81.9 to 95.0)
VOL. 86-B, No. 8, NOVEMBER 2004
1135
Fig. 3b Fig. 3c
(Fig. 3a). When a varus stress was applied, the mean lateral
flexion gap increased by 6.7 ± 1.9 mm (mean ± SD; 2.1 to
9.2) and the medial tibial and femoral joint surfaces were in
contact with a mean flexion angle of 87.9 ± 5.1˚ (75.3 to
100.2) (Fig. 3b). Under a valgus stress, the mean medial
flexion gap increased by only 2.1 ± 1.1 mm (0.2 to 4.2) and
the lateral tibial and femoral joint surfaces were in contact
with a mean flexion angle of 86.1 ± 3.4˚ (74.1 to 92.6) (Fig.
3c). The increase in the lateral flexion gap was significantly
larger than that observed for the medial flexion gap paired
t-test, p < 0.0001. The mean difference was 4.6 ± 1.8 mm
(1.1 to 9.0) and implied that, when the flexion gap was dis-
tracted, the gap was trapezoidal and not rectangular.
Although there was a tendency for women to have a larger
medial joint gap, the difference did not reach statistical sig-
nificance; men 1.8 ± 1.1 mm (0.2 to 3.2); women 2.5 ± 1.0
mm (1.1 to 4.2); unpaired t-test, p = 0.1487. On the lateral
side, there was no significant difference; men 6.5 ± 2.3 mm
(2.1 to 9.2); women 6.9 ± 1.5 mm (4.5 to 9.2); unpaired t-
test, p = 0.6745.
When the asymmetry of the flexion gap was expressed by
the angle between lines A and D, the mean value was 11.0
± 2.5˚ (6.2 to 15.0) under a varus stress and 0.1 ± 2.4˚ (-4.3
to +4.6) under a valgus stress (Table I). The mean asymme-
try of the flexion gap in the knees which we examined was
thus calculated to be 4.9˚, based upon the angle in a neutral
position of 3.1 ± 1.7˚ (1.3 to 7.5) and changes in the
stressed positions of 7.9˚ under a varus stress and 3.0˚
under a valgus stress. When the asymmetry of the flexion
gap was expressed by the angle between lines B and C, these
values were 7.9 ± 1.9˚ (4.8 to 12.6) under a varus stress and
-2.8 ± 0.9˚ (-5.0 to -1.5) under a valgus stress. The mean
1136 Y. TOKUHARA, Y. KADOYA, S. NAKAGAWA, A. KOBAYASHI, K. TAKAOKA
Table I. Measurement of the asymmetry of the flexion gap
Varus stress Valgus stress
Parameter Mean SD Range Mean
Line A and D 11.0 2.5 6.2 to 15.0 0.1
Line B and C 7.9 1.9 4.8 to 12.6 -2.8
asymmetry of the flexion gap was thus calculated to be 5.1˚
based upon the angle in the neutral position (always 0˚)
(Table I).
Discussion
Our study is the first to analyse quantitatively the discrep-
ancy of the joint gap in the medial and lateral tibiofemoral
joint for normal knees. The results indicate that the flexion
gap in a normal knee is not rectangular and demonstrates a
discrepancy of 4.6 mm and a tilt of 5˚ in the femur and
tibia. Although the relative lateral laxity in the normal knee
in flexion has been widely recognised by many orthopaedic
surgeons at arthroscopy,10 and reported by MRI,11 there
has been a paucity of quantitative data.
A discrepancy of 4.6 mm is substantial in view of the
detailed description on the symmetry of the flexion gap and
related rotation of the femoral component in TKA. By gen-
eral consensus, the femoral component is aligned to the
SEAs and aims to make the flexion gap rectangular. This is
essential for stability and kinematics during flexion.12,13 In
the light of our findings, it is reasonable to ask if it is really
necessary to make the flexion gap rectangular during TKA.
Recent biomechanical studies have shown that flexion of
the knee is associated with a significant medial-pivot inter-
nal rotation of the tibia.14-17 Thus, in rotation the medial
condyle is immobile and the lateral condyle is mobile on the
tibial articular surface. The different kinematics on the
medial and lateral sides are partly due to the more congru-
ent articular surfaces medially. However, the relative laxity
on the lateral side is also likely to be important.
We did not use a special device to apply the varus-valgus
stress, and therefore the amount of stress was not standard-
ised and was affected by several factors including the
weight of the leg and the mobility of the hip. Furthermore,
it was not always possible to obtain the same scanning
plane in the three positions (neutral, varus and valgus)
which could misleadingly suggest a gap between the sur-
faces. However, the MR scans were consecutive 3 mm thick
slices and the maximum error caused by measuring an adja-
cent slice was therefore calculated to be negligible (0.15
mm) when the radius of the femoral condyle was a sphere
with a diameter of 30 mm. We, therefore, consider that the
variation in slices had little or no impact on the main con-
clusion of our study. Another potential weakness was that
Neutral
SD Range Mean SD Range
2.4 -4.3 to +4.6 3.1 1.7 1.3 to 7.5
0.9 -5.0 to -1.5 0.0 0.0 0.0
the joint gaps were estimated by the application of a varus-
valgus stress and not by a simultaneous, bilateral distrac-
tion force. Thus, the degree of asymmetry of the joint gap
which we have shown may not be directly extrapolated to
the joint gap at TKA.
In conclusion, our results indicate that the tibiofemoral
flexion gap in the normal knee is not rectangular and that
the lateral joint gap is significantly lax. The results funda-
mentally question the concept of a rectangular flexion gap
at TKA and our understanding of tibiofemoral kinematics.
No benefits in any form have been received or will be received from a commer-
cial party related directly or indirectly to the subject of this article.
References
1. Insall JN, Scott WN, Ranawat CS. The total condylar knee prosthesis: a report of
two hundred and twenty cases. J Bone Joint Surg [Am] 1979;61-A:173-80.
2. Insall JN, Easley ME. Surgical techniques and instrumentation in total knee arthro-
plasty. In: Insall JN, Scott WN, eds. Surgery of the knee. Third ed. Philadelphia: WB
Saunders Inc, 2001:1553-620.
3. Freeman MAR, Samuelson KM, Bertin KC. Freeman-Samuelson total arthroplasty
of the knee. Clin Orthop 1985;192:46-58.
4. Laskin RS. Flexion space configuration in total knee arthroplasty. J Arthroplasty
1995;10:657-60.
5. Griffin FM, Insall JN, Scuderi GR. Accuracy of soft tissue balancing in total knee
arthroplasty. J Arthroplasty 2000;15:970-3.
6. Fehring TK. Rotational malalignment of the femoral component in total knee arthro-
plasty. Clin Orthop 2000;380:71-9.
7. Markolf KL, Mensch JS, Amstutz HC. Stiffness and laxity of the knee: the contri-
bution of the supporting structures: the contribution of the supporting structure: a
quantitative in vitro study. J Bone Joint Surg [Am] 1976;58-A:583-93.
8. Seering WP, Piziali RL, Nagel DA, Schurman DJ. The function of the primary lig-
aments of the knee in varus-valgus and axial rotation. J Biomech 1980;3:785-94.
9. Moore TM, Meyers MH, Harvey JP Jr. Collateral ligament laxity of the knee: long-
term comparison between plateau fractures and normal. J Bone Joint Surg [Am]
1976;58-A:594-8.
10. Wu WH, Richmond JC. Arthroscopy of the knee: basic setup and technique. In:
McGinty JB, Richmond JC, eds. Operative arthroscopy. Third ed. Philadelphia: Lippin-
cott, Williams and Wilkins Inc, 2003:211-17.
11. Freeman MAR, Pinskerova V. The movement of the knee studied by magnetic res-
onance imaging. Clin Orthop 2003;410:35-43.
12. Olcott CW, Scott RD. A comparison of 4 intraoperative methods to determine fem-
oral component rotation during total knee arthroplasty. J Arthroplasty 2000;15:22-6.
13. Poilvache PL, Insall JN, Scuderi GR, Font-Rodriguez DE. Rotational landmarks
and sizing of the distal femur in total knee arthroplasty. Clin Orthop 1996;331:35-46.
14. Todo S, Kadoya Y, Moilanen T, et al. Anteroposterior and rotational movement of
the femur during knee flexion. Clin Orthop 1999;362:162-70.
15. Hill PF, Vedi V, Williams A, et al. Tibiofemoral movement 2: the loaded and
unloaded living knee studied by MRI. J Bone Joint Surg [Br] 2000;82-B:1196-8.
16. Iwaki H, Pinskerova V, Freeman MAR. Tibiofemoral movement 1: the shapes and
relative movement of the femur and tibia in the unloaded cadaver knee. J Bone Joint
Surg [Br] 2000;82-B:1189-95.
17. Nakagawa S, Kadoya Y, Todo S, et al. Tibiofemoral movement 3: full flexion in the
living knee studied by MRI. J Bone Joint Surg [Br] 2000;82-B:1199-200.
THE JOURNAL OF BONE AND JOINT SURGERY