Quantitative Correlations in Degenerative Arthritis of the Knee

Document Sample
Quantitative Correlations in Degenerative Arthritis of the Knee Powered By Docstoc
					                           This is an enhanced PDF from The Journal of Bone and Joint Surgery

                               The PDF of the article you requested follows this cover page.

Quantitative Correlations in Degenerative Arthritis of the Knee
J Bone Joint Surg Am. 1973;55:956-962.

                             This information is current as of March 23, 2011

Reprints and Permissions      Click here to order reprints or request permission to use material from this
                              article, or locate the article citation on jbjs.org and click on the [Reprints and
                              Permissions] link.
Publisher Information         The Journal of Bone and Joint Surgery
                              20 Pickering Street, Needham, MA 02492-3157
             Quantitative                        Correlations                  in Degenerative
                                       Arthritis of the Knee*
           BY RAY MILLER, M.D.t,                IOWA CITY, IOWA, DONALD B. KETTELKAMP,                        M.D4,
        LITTLE ROCK, ARKANSAS, KEYRON N. LAUBENTHAL,                                   M.A.@, IOWA CITY, IOWA,
                    ATHANASE KARAGIOROOS, M.A1!,                  SAN MATEO, CALIFORNIA,                   AND
                                     GARY L. SMIDT, PH.D.@, IOWA CITY, IOWA

                    From the Department    of Orthopedic Surgery, University of Iowa,
            and the Department  ofPhysical   Therapy, University  oflowa   Hospital, Iowa City,
  and the Division of Orthopedic Surgery, University of Arkansas Medical Center, Little Rock

        ABSTRACT:          Progression           of roentgenographic           changes     over    five    to eleven         years
and the relationship of roentgenographic changes to symptoms, function and quad
riceps strength were evaluated in fifty knees (thirty-five patients). Patellofemoral
arthritis progressed more than that of the weight-bearing compartments.     Increased
degeneration of the medial compartment occurred in less than 50 per cent of the
vanis      knees.     Severity        of roentgenographic              degeneration        was    not related          to body
weight. Women had more severe patellofemoral                                  arthritis and greater limitation of
activity     than    men.     Increased          patellofemoral        arthritis   was associated          with increased
pain, increased           functional     impairment,         and decreased         quadriceps       strength.

       The quantitative   delineation  of the clinical and roentgenographic     changes
which occur with degenerative arthritis of the knee has not been well depicted despite
numbers of anatomical and roentgenographic       studies. Keyes reported that erosion of
cartilage was present in forty-seven of the seventy-four knees of cadavera that he
examined and found the erosion most frequent on the patella, and Bennett and as
sociates reported even more frequent changes with age. Ahlbäck codified the roent
genographic pattern of degenerative changes and developed a detailed digital form
for roentgenographic     evaluation.  In other studies, as of the incidence of gonar
throsis     the
           @, relationships between pain and structural alteration and to the pres
ence of osteophytes “¿ been made but do not include a comprehensive          study of
the interrelationships between the roentgenographic data, the clinical findings, and
quadriceps strength. In the present paper, we are reporting the progression of roent
genographic    changes, the relationships  of clinical to roentgenographic     findings,
and the relationship between clinical and roentgenographic    findings with quadriceps
strengthna groupof patients    withdegenerative            of

                                                  Material and Method
       Fifty knees in thirty-five individuals, nineteen men and sixteen women, the
material this      study,             a
                         representedllindividuals                     o
                                                      with a diagnosisf degenerative
arthritis of the knee seen at the University of Iowa Hospitals between 1960 and 1966,
who had available initial roentgenograms and adequate follow-up, without surgery.

        * Read   at the     Annual     Meeting      of The   American        Academy      of Orthopaedic         Surgeons,      Las
Vegas, Nevada,        February       5, 1973.
        t University of Iowa, Iowa City, Iowa 52240.
        t University of Arkansas Medical Center, Little Rock, Arkansas 72201.
        § niversity of Iowa Hospital, Iowa City, Iowa 52240.
        ¶ North San Mateo Drive, San Matco, California 94401.
956                                                                THE JOURNAL OF BONE AND JOINT SURGERY
                    QUANTITATIVE              CORRELATIONS         IN ARTHRITIS             OF THE KNEE                     957


                                                                IFpfx Dpt=(FrexOre)+ (Wgrx Dgr'sine)
                          FIG. I-A                                                          FIG. 1-B
   Fig.     1-A: Technique         of quadriceps      tension     and    torque      measurements         at 20 degrees       knee
                                                 adjusted at 90 degrees to shank.
flexion. A—anklestrap, B—tensiometer,(@—cable
    Fig. 1-B: Force vectors and other measurements                       used to determine    quadriceps    torque and
tension: A—axis of knee rotation     in sagittal plane;             Fpt—force vector of patellar tendon; Fre
force vector recorded    by tensiometer;     Wgr—force              vector due to weight of shank and passing
through shank center of gravity;O —¿angle created                 by long axis of shank and force vector Wgr;
Dre—moment     arm from knee axis to force vector                 Fre measured     as 75 per cent of shank length;
Dpt—moment     arm from knee axis to force vector                  Fpt: Dgr—distance   from knee axis to shank
center of gravity,  determined    using knee axis to                sole distance   and Drillis parameters;      Dgr
moment arm from knee axis to shank center of gravity, equal to Dgr sin 0 . The formula used to
calculate    quadriceps   tension       was: Fpt = (Fre         X Ore     + (Wgr       X Dgrsin       0

          The mean age of these patients was 57.2 years and follow-up                                      varied from five
to eleven      years with a mean of 6.9 years.                   The history         and physical         examination        were
recorded on forms suitable for computer analysis. The following items were tabu
lated: age, sex, duration of symptoms, occupation, activity level (walk, run, jump,
and so on), pain with activity, pain at rest, pain with changes in weather, locking,
catching, giving way, presence of crepitus, site of crepitus, ability to put on shoes
and socks, ability to walk on rough ground, ability to squat, ability to climb stairs,
endurance, and involvement ofotherjoints.     The physical findings tabulated included
height, weight, gait (limp, use of canes, and so forth), presence of effusion, presence
of synovial       thickening,          pain   on   patellar     compression,           patellar      abnormality,        patellar
stability,     flexion deformity,             range of motion,                instability    (medial,       lateral,    anterior
drawer,       posterior    drawer,        and rotatory),         site of tenderness,              pain with rotation,         Mc
Murray's test, presence of meniscal protrusions, and thigh circumference.
      The roentgenographic     examination at follow-up included these views: weight
bearing anteroposterior,   tunnel, patellar, and lateral. The initial and follow-up roent
genogranis were evaluated using Ahlbäck's format with the addition of measure
                                                           roentgenogram and
ment of the tibiofemoralshaft angle from the anteroposterior
measurement of the angle formed by lines tangent to the art icular surface of the tibia
and femur. Because the initial roentgenograms     were not weight-bearing,  classifica
tion of the knees into varus, neutral, and valgus was based on follow-up roentgeno
grams. A varus knee was defined as one with a tibiofemoral angle of 4 degrees of
valgus or less, a neutral knee as from 5 through 9 degrees of valgus, and a valgus
knee as one with 10 degrees or greater of valgus. This division was based on the
normal valgus of 7 degrees as reported by Steindler ± 2 degrees latitude in the
neutral group. It is recognized, however, that clinically and roentgenographically

VOL. 55-A, NO. 5, JULY          1973
958                          RAYMILLER ANDASSOCIATES

normal knees do occur in both the varus and valgus groups. To provide uniformity
all roentgenographic   evaluation was done by one examiner (R.M.) 10.
      Quadriceps torque and tension (Fpt) were determined (Figs. 1-A and I -B) for
each knee at 20 and 60 degrees of flexion using an isometric tensiometer    The
                                                                           @. ankle
strap was applied at a point three-quarters   down the length of the shank. The cable
between the ankle strap and the force table was adjusted separately for the 20 and 60
degrees tests so that the cable was perpendicular     to the shank. Tension readings
from the cable tensiometer were obtained for three maximum quadriceps contrac
tions at each position. The largest reading from each position was converted to kilo
grams from the calibration chart. The measurements necessary for the calculation of
torque and tension were the shank distance (knee axis (A), to tip of the lateral
malleolus), the distance from the knee axis (A), to the patellar tendon (Dpi), and
the distance from the knee axis to the sole of the foot. The weight ( Wgr) of the shank
and the distance from the knee axis to the center of gravity of the shank (Dgr) was
calculated from the tables of Drillis and Contini. The measurements and force vec
tors are shown in Figures 1-A and I -B.
      The Department of Biometry, University of Arkansas Medical Center, per
formed the statistical   analysis. Initial analysis eliminated the unnecessary clinical
variables. Confidence    intervals were calculated for roentgenographic    changes. De
pending on the levels    of measurement, chi square, rank correlation, or analysis of
variance methods were     used to test for statistical significance.


Roentgenographic   Changes
       The progression of roentgenographic     changes was determined separately for
the three groups of knees: varus, neutral, and valgus. Ninety-five per cent confidence
limits are given for the percentage values for worsening in the varus knees. The
number of cases in the neutral and valgus groups were so small (twelve and three)
that these two groups did not have statistical significance in general.
       Varus knees (Table I) showed a decrease in medial joint space in just less than
half of the cases. Whatever error may have been introduced in measurement of the
 medial joint space on final roentgenograms   (weight-bearing) would result in a greater
frequency of development of medial joint space narrowing than really occurred. The
decrease of the medial joint space in less than one half of the varus knees was in
sufficient in degree or frequency to have prognostic significance. In addition, knees
 in the varus groups showed a mean increase of varus angulation of 2.4 degrees; how
ever, the significance of this value is moot since the initial roentgenograms       were
 not weight-bearing “¿ The patellofemoral articulation exhibited progression of
joint space narrowing, increased facet involvement, and increased osteophyte forma
lion in more than 50 per cent of the varus knees.
       In the neutral group of twelve knees a decrease of medial joint space occurred
 in two knees and there was no loss of lateral joint space. Increased medial sclerosis
 occurred in four knees and increased lateral sclerosis in one knee. In two knees,
 cystsbecame worse. Attritionf the medial tibial           plateauand medial femoral
 condyle each increased in severity in two knees. No increased attrition occurred in
the lateral compartment. Osteophytesincreased in size on the medial tibial plateau
in three knees, the medial femoral condyle in five knees, the lateral tibial plateau in
six knees, and the lateral femoral condyle in three knees. Patellofemoral degenera
tion was more frequent with decreased joint space in five and increased facet in
volvement in four, attrition in two and osteophytes in six.
      Only three knees were classified as valgus. Decreased medial, lateral and
                                                  THE JOURNAL OF BONE AND JOINT SURGERY
                         QUANTITATIVE              CORRELATIONS                   IN ARTHRITIS                     OF THE       KNEE               959

patellar joint space occurred in one knee each. Increased medial sclerosis and cysts
also occurred in one knee each. Attrition increased in one knee each for medial
tibial plateau, lateral tibial plateau, and lateral femoral condyle. Osteophytes in
creased on the medial tibial plateau in one, medial femoral condyle in two, and on
the lateral tibial plateau and lateral femoral condyle in all three knees. Patellar
changes of loss of joint space, increased facet involvement, and attrition each oc
curred in one knee. Osteophytes increased on all three patellae. Tibiofemoral angula
tion increased an average of 6 degrees in the valgus knees but is subject to the limita
tion of non-weight-bearing      initial roentgenograms.  Two of the knees did demon
strate increased structural deterioration of the lateral compartment. One knee with a
follow-up of five years had decreased lateral joint space and one after six years
showed increased lateral tibial attrition and cyst formation.
       Roentgenographic    changes of degenerative arthritis tended to progress over the
period of follow-up and to do so more rapidly in knees with angular deformity,
both varus and valgus. In each group of knees, progression of patellofemoral         de
generation exceeded the progression in either of the weight-bearing compartments.

                                                                       TABLE           I
                                  ROENTGENOGRAPHICCHANGES FOR VARUS KNEES (N35)

                                                                   or                                                   cent
 *1. ItemsImproved                                         UnchangedWorsePer                                          WorseConfidence
    Medial joint          space                                                                                                                   66)
 2. Lateraljointspace                                             32                              3                     8.6                  (1,23)
 3.Patellofemoral                                                 15                             20                    57.1                    74)
 4. Increased facet involvement                                   15                             20                   57.1                  (39, 74)
 5. Articular surface angle                                       14                             21                   60.0                  (42, 76)
 6. Medial sclerosis                                              24                             11                   31.4                  (16, 49)
 7. Lateral sclerosis                                             24                             11                    31.4                 (16, 49)
 8. Cysts18
44)Attrition9.                                                    2617                            948.6                25.7(33,             (12,

    Medial tibial plateau                                                                                                                        44)
10. Medial femoral condyle                                        23                             12                    34.2                 (19, 52)
11. Lateral tibial plateau                                        34                              1                     2.9                   (0, 15)
12. Lateral femoral           condyle                             32                              3                     8.6                   (1, 23)
13. Patella26
15)Osteophytes                                                    349                             125.7                 2.9(12,               (0,

14. Medial       tibial plateau                                                                                                                   72)
15. Medial       femoral condyle                                   7                             28                   80.0                  (63, 92)
16. Lateral tibial plateau                                        13                             22                   62.9                  (44, 79)
17. Lateral femoral           condyle                             16                             19                   54.3                  (36, 72)
18. Patella16                                                     1119                           2454.3               68.6(36,              (50, 84)

    * 95   per    cent     confidence     limits     for    per        cent   worse.       The        confidence       limits     would   cover   the    true
mean 95 per centof the time.

Roenigenographic,                  Clinical,       and Quadriceps                      Grrelations
      The statistical significance of relationships between current roentgenographic
appearance,    clinical findings, and quadriceps strength were determined by a con
tingency coefficient, a Spearman rank correlation        coefficient, or an analysis of
      We did not try to correlate tibiofemoral      angulation (varus-valgus) with the
history and physical findings because the knees studied represented a mixture of
varus, neutral, and valgus angulation. Varus angulation, however, was associated
with narrowing of medial joint space, medial sclerosis, and loss of bone substance
VOL. 55-A, NO. 5, JULY 1973
960                              RAYMILLER ANDASSOCIATES

at the medial tibial and femoral surfaces. An increase in the severity of these roent
genographic findings of medial compartment           degeneration was associated with an
 increase in the following: pain on activity and at rest, difficulty negotiating stairs
and rough ground, limp, and use of external support. These findings were associated
with a decreasing range of motion.
       The larger the osteophytes of all five sites (medial tibia and femur, patella, and
lateral part of the tibia and femur), the greater were the abnormal physical findings
and the pain and the less the function.
        Patellofemoral degeneration, that is, loss ofjoint space, facet involvement, loss
of bone, and the presence of osteophytes merits particular attention. The degree of
 patellofemoral     arthritis, just as with the roentgenographic   changes in the medial
joint, corresponded        to the degree of clinical alterations in function. Most of the
specific items—decreased activity level (walking, running, jumping, and so on), rest
 pain, difficulty putting on shoes and stockings, climbing stairs and squatting, need
for external support, synovial thickening and loss of flexion—involve functions as
sociated with use of the knee in flexed positions which would increase patellofemoral
 pressure. In addition, there was a much closer relationship between patellofemoral
changes and quadriceps strength than between medial articular degeneration            and
strength. These findings emphasize the importance of the patellofemoral articulation
to symptoms    and   function   in the arthritic   knee.
       Increased flexion deformity was associated with increased severity of arthritic
involvement of the medial side and of the patellofemoral joint. Of interest was the
greater frequency of cysts in knees with flexion deformity. The range of flexion and
extension during the stance phase is decreased or absent in knees with flexion de
formity with corresponding        concentration     of the load. Hence, the cysts may be the
result of increased load per unit area. As would be expected, flexion deformity corn
promised the activity level, ability to squat, and the range of knee motion.
        Instability (medial, lateral, anteroposterior,       or rotatory) had few significant
relationships      with roentgenographic     abnormalities.    The anterior drawer sign was
more frequent in men and was associated with lateral tibial osteophytes. Medial in
stability was related to the frequency and size of osteophytes on the medial and
lateral aspects of the femur, on the lateral aspect of the tibia and patella, and to de
creased quadriceps tension at 60 degrees of flexion. We believe the reason for the
relative unimportance        of instability was the difficulty in differentiating    true liga
mentous instability from instability secondary to loss of articular cartilage and bone.
The distinction between ligamentous and secondary instability becomes important
in the planning of reconstructive procedures.
       Body weight was not significantly related to severity of roentgenographic
        It should be noted that the women in our patient population had significantly
more severe changes in the patellofemoral articulation than the men.
       Quadriceps strength as indicated by torque and tension were reciprocally re
lated to the degree of patellofemoral         arthritis. The sizes of the osteophytes on the
medial and particularly the lateral aspect of the femur also were inversely related
to the strength measurements.         This relationship may exist because of the involve
ment of these osteophytes at the patellofemoral joint. Those activities which require
quadriceps function or which increase patellofemoral            pressure (shoes and stockings,
stairs, and walking distances) were directly related to quadriceps strength.

    Our study, like those of Danielsson and Hernborg― and Bauer and associates,
demonstrated that degenerative arthritis of the knee is very slowly progressive. In

                                                     THE JOURNAL OF BONE AND JOINT SURGERY
                            QUANTITATIVE             CORRELATIONS               IN ARTHRITIS             OF THE    KNEE                     961

    our study the roentgenographic  features ofdegenerative arthritis, consisting of medial
    joint space narrowing, sclerosis and cyst formation, progressed in less than 50 per
    cent of the varus kneeswith a follow-up of five to elevenyears.
               Varus and valgus deformities                         were associated            with more progression                       of de
    generative        changes         than     occurred            in knees       with      neutral        angulation.         This    finding,
                                                        15, is
    although anticipated on a biomechanical basis ‘¿@ difficult to evaluate since the
    initial films were not made weight-bearing.                              and Ahlback
                                                  Danielsson and Hernborg @‘¿
    also noted more structural changes in varus than in neutral knees. Unfortunately the
    predictability  of progression of degeneration    for a given patient cannot be made
    with sufficient certainty to warrant prophylactic reconstruction. The decision to per
    form       corrective      surgery        must     be based           only on the syniptoms,                   physical         and roent
    genographic findings, and not on anticipated future problems.
          The relationship of osteophyte formation to degenerative arthritis is debatable.
    Osteophytes    have been produced experimentally     without associated degeneration
    of articular cartilage 14 and do not necessarily signify the eventual development of
    other roentgenographic evidence of degenerative arthritis @.   Conversely, increased
    osteophyte size was associated with structural changes in this study and has been
    reported by others t.4•Osteophytes were associated with increased symptoms, de
    creased       function,       abnormal           physical            findings,     and     medial         collateral       ligament           in
            Patellofenioral degeneration as evidenced by decreasing joint space and in
    creased facet involvement became significantly worse with time. Patellofemoral
    degeneration            was     associated         with          decreased         activity          level,   rest      pain,     difficulty
    putting on shoes and stockings, climbing stairs and squatting, need for external
    support, synovial thickening, loss of flexion, and decreased torque and quadriceps
    tension.       These      findings       belie    the     clinical        assumption          that     patellofemoral           symptoms
    are seldoni a problem after proximal                              tibial osteotomy            for degenerative             genu varum.
    The symptomatic   improvement                          may occur             because of realignment                     of the extensor
    mechanism  and slight anterior                        displacement             of the tibial tubercle                  which results in
    improved mechanics 8 Whether a similar mechanical advantage occurs with total
    knee arthroplasty remains unknown. The high correlation between patellofemoral
    arthritis      and      functional       disability            suggests     that     the patellofemoral                joint    is a more
    common site of disability                  than was thought to be the case.
@          We concur           with      Kellgren           that     there     is a clear     association          between          knee    pain
    and roentgenographic                 changes and with Bauer and associates that pain tends to in
    crease with increasing               deformity in the sagittal and coronal planes.
               We found,       as would          be expected,              significant       relationships          between         increased
    evidence of degeneration and increased pain with activity, decreased activity level,
    decreased quadriceps strength, decreased ability to walk on rough ground, climb
    stairs, and squat, and increased flexion deformity.
           The role of body weight and obesity in degenerative arthritis remains unclear.
    Kellgren reported an increased incidence of degenerative arthritis of the knee in
    obese individuals      We found, as others 9,15, a greater frequency of multiple joint
    involvement in heavier subjects. Roentgenographic     evidence of degenerative arthritis
    was not related to body weight, a finding also reported by Danielsson and Hem
             In addition, we found that body weight was not related to activity level, pain,
    borg “¿.
    or quadriceps strength. The only activities related to body weight were squatting and
    putting on shoes and stockings. The decreased flexibility caused by a protruding
    abdomen and heavy thighs rather than arthritis could impose this limitation. Al
    though increased body weight may make surgery more difficult, complications
    more frequent, and theoretically increase the joint force, we cannot counsel weight
    reduction to decrease pain, increase activity, or retard roentgenographic    changes.

    VOL. 55-A,NO. 5,JULY 1973
    962                                  RAYMILLER ANDASSOCIATES

@                                                a
           Studies by Ahlbäck, Kellgren @, nd Danielsson and Hemnborg showed a
    greater incidence of osteoarthritis     of the knee in women than in men. The sex in
    cidence in our study was about equal but does not represent a population survey.
    Women had more severe patellofemoral changes and a greater incidence of multiple
    joint involvement than men. These findings probably account in part for their lower
    activity level and greater difficulty with stairs and squatting. Women may also be
     more willing to follow their physician's advice to decrease their activity.
           Quadriceps strength decreased in relation to severity of patellofemoral arthritis,
     decreased activity level, difficulty climbing stairs, and abnormalities of gait require
     ing increased use of external support. These relationships support the rationale of
     using isometric quadriceps exercises to increase quadriceps strength in patients with
     degenerative    arthritis. A prospective study will be necessar.y to substantiate     the
     benefit of this regimen.

     1. AHLBACK, SvEN: Osteoarthrosis  of the Knee. A Radiographic   Investigation.    Acta Radiol.,
        Supplementum  277, 1968.
     2. BAUER, G. C. H.; INSALL, JOHN; and K05HIN0, TOMIHISA: Tibial Osteotomy      in Gonarthrosis
        (Osteo-Arthritis ofthe Knee). J. Bone and Joint Surg., 51-A: 1545-1563, Dec. 1969.
                  G.             H       a
     3. BENNETT, A.; WAINE, ANS; nd BAUER,          WALTER:    Changes in the Knee Joint at Various
        Ages; with Particular Reference to the Nature and Development     of Degenerative Joint
        Disease. New York, The Commonwealth    Fund, 1942.
     4. DANIELSSON, LARS, and HERNBORG, JERKER: Clinical   and Roentgenologic    Study of Knee
          Joints with Osteophytes. Clin. Orthop., 69: 302-312, 1970.
     5. DANIELSSON, LARS, and HERNBORG, JERKER: Morbidity                  and Mortality       of Osteoarthritis      of
          the Knee (Gonarthrosis) in Malmfl, Sweden. Clin. Orthop., 69: 224-226, 1970.
     6. DRILLIS, R., and CONTINI, R.: Body             Segment     Parameters.   Technical      Report   1 16.03,    pp.
          69-75. New York University, School of Engineering and Science. New York, 1966.
     7. KARAGIORGO5, A.: Overweight           and the Knee Joint     in Degenerative   Arthritis.    Master     of Arts
        Thesis, University of Iowa, August 1971.
     8. KAUFER,HERBERT: echanical Function of the Patella. J. Bone and Joint Surg., 53-A:
        1551-1560, Dec. 1971.
     9. KELLGREN, H.: Osteoarthrosis in Patients and Populations. British Med. J., 2: 1-6, 1961.
    10. KELLGREN, J. H., and LAWRENCE, J. S.: Radiological                Assessment   of Osteo-arthrosis.          Ann.
          Rheumat. Dis., 16: 494-502, 1957.
    11. KETTELKAMP, D. B., and CHAO, E. Y.: A Method for Quantitative      Analysis of Medial and
        Lateral Compression Forces at the Knee During Standing. Clin. Orthop., 83: 202-2 13, 1972.
    12. KEYES,E. L.: Erosions of the Articular Surfaces of the Knee Joint. J. Bone and Joint Surg.,
        15: 369-37 1, Apr. 1933.
    13. LEACH, R. E.; GREGG, THOMAS; and SIBER, F. J.: Weight-Bearing                    Radiography          in Osteo
        arthritis of the Knee. Radiology, 97: 265-268, 1970.
    14. MARSHALL, L.: Periarticular Osteophytes. Initiation and Formation in the Knee of the
          Dog. Clin.. Orthop.,   62: 37-47,   1969.
    15. RADIN, . L.; PAUL,I. L.; and ROSE,R. M.: Role of Mechanical Factors in Pathogenesis of
        Primary Osteoarthritis.    Lancet,1:5 19-522,1972.
    16. STEINDLER, ARTHUR: Kinesiology         of the Human Body Under Normal                and Pathological       Con
        ditions, p. 331. Springfield, Illinois, Charles C Thomas, 1955.

                                                                 THE JOURNAL OF BONE AND JOINT SURGERY