Variations in cross-sectional area and composition of equine tendons with regard to their mechanical function

Research in Veterinary Science /986,4/, 7-/3

Variations in cross-sectional area and composition of equine tendons with regard to their mechanical function D. J. RIEMERSMA, P. DE BRUYN, Department oj Functional Morphology, Faculty oj Veterinary Medicine, State University Utrecht, Yalelaan I, 3508 TD Utrecht, The Netherlands

properties of equine tendons (Riemersma and Schamhardt 1985) showed that these properties varied with the tendon and the site in the tendon. It was observed that strains within a loaded tendon were more or less homogeneous in spite of large variations in CSA, which indicated that the strength of a tendon site was independent of its local CSA. This was explained by the inverse relationship between CSA and the collagen content. Since the geometrical CSA was not representative for the strength of a tendon, a more extended study of the relationship between the CSA and the composition of equine tendons was undertaken in order to understand better the mechanical behaviour of the structures, which is essential when considering prevention of tendon injuries. Because collagen and water are the main components of tendons, the collagen content (cc) and the reverse of the water content, the dry weight percentage (DW), are the most important parameters of tendon composition. The tendon fibre percentage (FP) was estimated to obtain more information about the relationship between the cc and the amount of load-bearing fibres, which might clarify the relationship between load-bearing and non load-bearing collagen. The variations in CSA and tendon composition were studied in the SDFT, the deep digital flexor tendon (DDFT) and the SL of the equine hindlimb. Although the tendons of the forelimb of the horse are more frequently involved in injuries due to overload, the hindlimb was preferred since it is the object of a more extended biomechanical research programme in the authors' department (Schamhardt et al 1985, Riemersma et al 1985, Merkens et al 1986). The possible relationships between the CSA and the composition of these tendons within a tendon, between different tendons of the same horse, and between the tendons of different horses were also studied.

The cross-sectional area, collagen content, dry weight as a percentage of the wet weight and the tendon fibre percentage of the cross-sectional area of the equine hindlimb were determined in the superficial and deep digital flexor tendons and the suspensory ligament at 10, 12 and six sites between tarsus and insertion respectively. The values of each of the four parameters varied between different sites in the same tendon, between different tendons within a horse and between analogous tendons of different horses. Within a tendon the cross-sectional area was inversely proportional to the collagen content, the dry weight and the tendon fibre percentage. Within a horse and between different horses the cross-sectional area of a tendon was inversely proportional to the collagen content and the dry weight percentage. It was concluded that the cross-sectional area is not representative of the strength of an equine tendon. IT has been suggested that tendon injuries due to overload occur at tendon sites with the smallest crosssectional area (CSA), because stress (toad/cs») will be highest at these sites (Fackelman 1973, Webbon 1973). However, this would be valid only if tendons are homogeneous and isotropic and several authors have shown that this is not the case. Gillard et al (1977) and Merrilees and Flint (1980) showed that the structure and composition of rabbit flexor tendons are different for different regions of the tendon. Parry et al (1978) demonstrated that the fibrillar collagen content was different for the superficial digital flexor tendon (SDFT), the suspensory ligament (SL) and the common extensor tendon of the horse. This implies that structure and, or, composition may be different for different tendons of a horse. Sorokin and Efimov (1980) demonstrated that the morphological, histological and biochemical constitution of the tendons of humans and dogs varied along the length of the tendon. They also found an inverse relationship between the tendon mass and its tearing stress. They concluded that the variations in histological and biochemical constitution of these tendons were adaptations for taking up compression loads as well as tensile loads. Previous investigations concerning the mechanical

Materials and methods

Tendons The tendons (SDFT, DDFT and SL) were dissected from 20 hindlimbs of adult slaughtered horses of

7

8

D. J. Riemersma, P. De Bruyn included six sites, the parts of the SDFf and DDFf which curved round the fetlock joint 3, and the more distal part of the DDFf also three sites. Since the sites were chosen according to their anatomical location, the sites on small tendons were shorter than those of larger tendons. Plane transverse sections were obtained by cutting the tendons after freezing. The tendons were carefully padded with aluminium foil and plastic to prevent desiccation during the freezing process. The tendons of II limbs were used to determine eSA, ee and DW. The tendons of nine other limbs were used to determine FP.

SL

Cross-sectional area The eSA of a tendon site was determined from the quotient of volume over length. The lengths of the (still frozen) subunits were measured with sliding callipers. Volumes were then determined by water displacement at room temperature. For this, one side of an equilibrium balance was provided with a hook to which the tendon piece was mounted. The opposite side was provided with a counter weight which rested on an electronic balance (Fisher model 3OO-DR; Ainsworth). The upward force on the piece when submerged in pure water was registered on the electronic balance by weight increase on this side. The volume of a piece (mm') is then proportional to this weight increase (mg) (Archimedes). The volume of the hook was subtracted. After the eSA was determined the tendon piece was divided into three subunits by transverse section. The central subunit was used to determine ee, the other two to determine DW.

Collagen content

FIG 1: Photograph of the isolated tendons. The sites are numbered from subtarsal level to insertion. s Proximal sesamoid, p Short pastern, n Navicular bone, c Coffin bone

various breeds, which weighed between 450 and 700 kg. The SL differs from the flexor tendons because it is a tendinous muscle (synergistic median interosseus muscle). It contains residual muscle fibres and also fat, which are not present in the flexor tendons. Between tarsus and insertion the structures were subdivided into 10 SDFf, 12 DDFf and six SL sites (Fig I). The metatarsal regions of the SDFf, DDFf and

The collagen content was derived from the hydroxypyroline content. Hydroxyproline occupies 10 to II per cent of the residues of type I collagen, which corresponds to a weight percentage of about 12'5 (Block and Weiss 1956). The hydroxyproline content of the hydrolysate was determined colorimetrically (Stegemann and Stalder 1967) and multiplied by 8 to obtain the collagen content as a wet weight percentage.

Dry weight percentage The wet tendon subunits were weighed and then airdried to constant weight at 20°C and relative humidity of 30 per cent. A drying period of two days was usually sufficient.

Percentage of the

CSA

occupied by tendon fibres

The mid cross-section of a tendon piece was cut to a

,2_ 1_ 3_ 3,"" 4_ 4,-. 5_ 5'" 6""" 6. 7_ Cross-sectional area and composition of equine tendons

OOFT

SL

SOFT

2~

5 """,.

7

"..-....

.~~UI_

8

I

iJ

8~ -

.

~........

~'-'

~.

~

9~

9

10~' • s -

•

" fj~

1 •• ,}~.

10 . , '

'4"~

12"""

FIG 2: Surfaces of the mid-cross-sections of tendon sites. Trichrome Masson stained

smooth surface with a cryomicrotome. The contrast between the tendon fibres and the connective tissue (Fig 2) was accentuated by Trichrome Masson stain-

ing (Flint er al 1975). The percentage of the tendon fibres was deduced from the percentage of dark stained areas along three to six lines (depending on the

9

D. J. Riemersma, P. De Bruyn

10 N = 4 rows k = 5 columns ranking

was then defined as the percentage of the lines that covered the tendon fibres.

..

Computation and statistics

Ra

Rb

Rc

Rd

Re

Sum of ranks FIG 3: The Friedman test is based upon ranking values within rows from low to high. The different sum of ranks (R) will accentuate differences between values within the rows if they occur consistently throughout the population

size of the surface), which were drawn randomly in enlarged photographs of the stained surfaces. The sections of the lines that covered the dark stained tendon fibres were measured on a graphic tablet (Summagraphics Bitpad I), connected to an ITT 2020 computer, and expressed as a percentage of the total length of the lines that covered the surface. The FP

Superficial digital flexor tendon

Deep digital flexor tendon

Suspensory ligament

300

300

300

200

200

200

100................L.I...J............ ..L.L.L.........- I-I...J...............

100.............-I-...............-I-.................-I-.................u.

100.........................................J-I........

35 u u

To be able to compare different tendons (within a horse and between horses) with one another, the parameter values of the various sites on a tendon were averaged to give the parameter value of the whole tendon. For this, the value of each tendon site was multiplied by the length of the site. These products of adjacent single sites on a tendon were totalled and divided by the sum of lengths of the sites. Nonparametric tests were chosen in order to avoid overestimation of the large tendons or underestimation of the smaller tendons. The Friedman test (Connover 1980) was used to test the significance (P = O·05) of differences between parameter values of different sites within a tendon, and of different tendons within and between horses. The procedure of ranking and the achievement of the sums of ranks is briefly explained in Fig 3. The significance of differences between various values resulted from multiple comparison of their respective sums of ranks (Connover 1980). Kendall's T was used to test the significance of correlations (P = O·05) between different parameters of adjacent sites within a tendon, of different tendons of a horse and of analogous tendons of different horses. The correlations between analogous

30~

35~

30

25 20

25

25

20 .................L.L.L.........- I-...............- ' -....................

20 ................L-L............................................L.JL-L....................

15·~..i..JL..J....L.J..........l-l....L......

45

45

45

35~

35~

25 ....................I...J..................................-'-I...J...............

25.................................L..L..........L..L..........L..L..........L..L..........L..L

251-J-.J-I.....L..J~..l-L

w~

W

90

70

70

70~

100 200 300 Distance from tarsus (mm)

.................

100 200 300 Distance from tarsus Imm)

FIG 4: Mean values ± SEM of the CSA, cc, ow (n = 11) and FP In = 9) of different sites on the SOFT, OOFT and SL. The abcissae show the average length of a tendon and the position of the sites on these tendons

Cross-sectional area and composition of equine tendons TABLE 1: Correlations between different parameters within a tendon

CC CSA CC OW

SOFT OW

+

FP

+ +

CC

OOFT OW

+

FP

+ +

CC

Sl OW

0 0

~

E

FP

0 0 0

+ Positive correlation 0

II

Negative correlation No correlation

« Vl u

;;e u u

parameters of different tendons within a horse were also tested. Results The mean values of the CSA, CC, ow and FP of the respective tendon sites are presented in Fig 4. The abcissae of this figure represent the mean lengths of the tendons between tarsus and insertion. They are composed of the mean lengths of the subsequent sites on a tendon. The parameter values are plotted against their (relative) position on the tendons. The differences in CSA approach 100 per cent within the flexor tendons. Even the smaller differences in the region between tarsus and fetlock joint (sites I to 6), a region which is usually considered straight, were significant as were the differences in CC, ow and FP within the tendons. Within the SL the CSA, ce, ow and FP were almost constant in value except for the sixth site (sesamoid part), which differed significantly in CSA, CC, ow and FP from the other sites. The negative correlation between the eSA of a tendon site and the corresponding cc, ow and FP, which is suggested by Fig 4, was significant for the SOFT and OOFT (Table I). The CC, ow and FP were proportional to one another. Within the SL, however, the CSA was inversely proportional only to the cc and not to the ow and FP. The parameter values of the tendons as calculated from the parameter values of the various sites on the tendons, are shown in Fig 5. Each bar represents the parameter value (CSA, CC, ow or FP) of one tendon. Each group of three tendons represents the SOFT, OOFT and SL of the same horse. The parameter values of a tendon (SOFT, OOFT or SL) were different for different horses. The significance of their mutual differences is better shown in Fig 6, which presents the relative values of the parameters (sums of ranks) which were attached in the Friedman test (k = number of analogous tendons, N = number of sites on each tendon, Fig 3). Parameter values of different analogous tendons (SOFT, OOFT or SL) differed significantly if their respective sums of ranks differed by at least the critical value, which is added to the figure as a shaded bar to the right. Fig 6 shows that the

Q.

U.

FIG 5: Values of the CSA, each horse

cc, OW and FP of the SOFT,

OOFT and Sl of

parameter values of different tendons were different for different horses; not only the CSA, which may be expected to depend on the size or mass of the horse, but also the CC, ow and FP, which are relative values, differed.

"3

.s «

Ul

u

"3 ~

u u

"3 ~

~

0

a SOFT

"3

~

D ODFT

it

• Sl

12 13 14 15 16 17 18 19 20 Horse number FIG 6: Sums of ranks IN; number of tendon sites, k; number of horses) of the CSA, cc. OW and FP of the SOFT, OOFT and Sl of different horses. Critical differences are given as separate bars on the right

D. J. Riemersma, P. De Bruyn

12

TABLE 2: Correletions between analogous parameters of different tendons of a horse CSA OOFT SL SOFT OOFT

+

o o

CC OOFT SL

+

+ +

OW OOFT SL

+

+ +

FP OOFT SL

o

o o

+ Positive correlation No correlation

TABLE 3: Mean values ± SEM of the eSA, cc, OW and FP of the three tendons and the significance of their relative magnitudes SOFT

OOFT

Significances

SL

CSA (mm21 CC (% w/wl

173± 8 216± 9 239 ± 20 28·9± 1·1 31·3 ± 0·6 22·2 ± 0·8 ow(%w/wl 36·0± 1·0 39·7 ± 0·8 34·0 ± 0·9 FP (% alaI 80·1 ± 0·7 78·9 ± 0·9 63·3 ± 1·9

<

SOFT OOFT = SL OOFT> SOFT> SL OOFT> SOFT> SL OOFT = SOFT> SL

o

The correlation between the parameter values of the SDFT, DDFT and the SL within a horse are shown in Table 2. A positive correlation exists between the CSA of the SDFT and the DDFT, but the CSA of the SL was not significantly correlated to the CSA of the flexor tendons. The cc and DW of the three tendons were positively correlated, but no correlation was found between the respective FP of the three structures. The mean parameter values of the SDFT, DDFT and SL are presented in Table 3, which also contains the significance of their relative magnitude within a horse. The relatively high standard error of the mean CSA and FP of the SL should be noted. The negative correlation between the CSA of a tendon and its CC and DW within and between horses, which is suggested by Fig 6, was confirmed by statistical analysis (Kendall) for the SDFT, the DDFT and the SL. (A correlation between CSA and FP could not be estimated since the CSA and the FP were obtained from different horses.) Since the CSA, cc and DW of the SDFT, DDFT and SL individually depend on the horse and, within a horse, the values of the three tendons were related, a dependency of the mean parameter value of the three tendons of the horse could be expected. This is shown in Fig 7 in which each bar represents the relative value (sum of ranks, k = number of horses, N = number of tendons within a horse [= 31) which stands for the mean parameter value of the three tendons of one horse. Fig 7 shows that the CSA, cc and DW of the three tendons combined are different for different horses. However, this was not found for the differences in FP of the three tendons between different horses. The CSA of the tendons of a horse was inversely proportional to the cc and DW.

structures, which equine tendons are apparently not. Therefore overstress, a term often used by clinicians, should not be used to describe overload of a tendon. It would, and indeed did, lead to the wrong conclusion that the tendon sites with the smallest CSA would therefore be weak sites. It was shown (Riemersma and Schamhardt 1985)that, loading these tendons in vitro, tendon strain was independent of tendon site in spite of the differences in CSA. This indicates that the variations in CSA within a tendon are not caused by increase or decrease in the amount of load bearing fibres but by variations in non-load bearing collagen and non-collagenous substances, eg, water, glycosaminoglycans, tendon cells etc. However, the variations in FP are less than those in CSA (Fig 4) and thus the CSA of the tendon fibres is not constant along the length of a tendon. This indicates that the cc of the tendon fibre itself changes within the tendon analogous with the variation in fibrillar collagen content between different tendons (Parry et aI1978).

:;

.s


l/l U

7

8

9

10 11

Crit diff

:; ~

u

u

Discussion This investigation showed that the CSA and composition varied with the site on the tendon (Fig 4). The tendon sites with the smallest CSA contain the highest content of collagen, tendon fibres and dry substances. This variation in CSA and composition within a tendon and the relationship between them is important for understanding the mechanical behaviour of these tendons. The calculation of stress (load/csx) presumes homogeneous and isotropic

FIG 7: Sums of ranks (N = 3 150FT. OOFT. SL!. k = 11 lhorsesl of the CSA. CC and OW of the tendons of different horses. Each bar representsthe averageparameter value of the tendons of one horse. Critical differences are given as shaded bars on the right

Cross-sectional area and composition of equine tendons The increase in CSA by non-load bearing collagen and other compounds will make the tendon locally suitable for additional functions, eg, the assimilation of transverse compression forces at sites where the limb axis deviates (sites 7, 8 and 12 on the DDFT), as was previously proposed by Sorokin and Efimov (1980) for the tendons of humans and dogs. The attachment of retinacula which guide the tendon (sites 7, 8 and 9 on the SDFT) will also increase the CSA and even the amount of collagen. However, this will not increase the tearing strength since the additional fibres are not longitudinally arranged. Thus the CSA of a tendon site is not representative of the local strength of a tendon. On the analogy of this, the mean CSA of a tendon may not be representative of its strength. It has been shown (Riemersma and Schamhardt 1985) that the SL could be strained relatively easily, ie, its modulus of elasticity was low. This corresponded with a low cc relative to the cc of the flexor tendons, which is confirmed by this investigation (Fig 4, Table 3). This is explained by the presence of scanty muscle fibres and fat between the tendon fibres; substances which are not found in the flexor tendons. The higher CC, DW and FP in the sesamoid branches of the SL (site 6) relative to the gross part of the structure corresponds with the absence of muscle fibres in this region. The amount of these compounds varies with the horse, which results in a relatively high standard error of the mean CGA and FP of the SL. This may also explain the absence of a significant correlation between the CSA of the SL and those of the SDFT and the DDFT of different horses (Table 2). The cc and DW of the three tendons are positively related, as are the CSA of the SDFT and DDFT. The respective FP of the three tendons, however, are not correlated. The large variations in FP of the SL will partly account for it. The differences in collagen content of the fibres (Parry et al 1978) may also explain the absence of a significant correlation between the FP of the SDFT and DDFT in spite of a positive correlation between the cc of the two flexor tendons. Therefore the CSA of the tendon fibres may also not be representative of the strength of a tendon. It was shown in this investigation that the combined values of the CSA, CC and DW of the three tendons are different for different horses (Fig 7). It would be of interest to investigate the possible correlation between tendon(fibre) composition and the susceptibility of the tendons to overload. Although the differences in composition between different horses are rather small, as shown by the small standard error of the mean in Fig 4 and Table 3, so may be the differences in mechanical tolerance of these tendons. The origin

13

of the differences in tendon composition needs further attention. An investigation of the correlation between tendon composition and age, sex or breed was not included in this study. Such a correlation cannot be excluded and could be examined more specifically. Also, the history, in particular the state of training, of the horses was unknown. Irrespective of its origin, the obvious individual differences in tendon composition could be used to select horses on this criterion. If there is a correlation between tendon composition and tendon injury, such a selection would be valuable in the prevention of tendon injuries. It is concluded from this study that variations in the cross-sectional area of different sites on a tendon and of different tendons of the same horse and of different horses are correlated to variations in CC, DW and percentage of the CSA occupied by tendon fibres. The CSA of a tendon site is inversely proportional to the CC, the DW and the FP of the CSA. The variations in CSA and composition within a tendon are adaptations to functions additional to the receptivity of tensile loads and will not significantly influence the local tensile strength of the tendon. Since the CSA of an equine tendon is inversely proportional to the cc and the DW the thickness of a tendon on its own is not representative of its tensile properties. References BLOCK, R. J. & WEISS. K. W. (1956) Aminoacids Handbook. Springfield. Charles C. Thomas CONNOVER, W. J. (1980) Practical Nonpararnetric Statistics. New York, J. Wiley & Sons. PI' 299-302 FACKELMAN. G. E. (1973) Equine Veterinary Journal S, 141-149 FLINT, M. H., LYONS, M. F., MEANY, M. F. & REID, T. (1975) Histochemical Journal 7, 529-546 GILLARD, G. C., MERRILEES, J. M., BELL-BOOTH, P. G., REILLY, H. C. & FLINT, M. H. (1977) Biochemical Journal 163. 145-151 MERKENS, H. W., SCHAMHARDT. H. C. & KERSJES, A. W. (1986) Equine Veterinary Journal 18. 207-214 MERRILEES, M. J. & FLINT, M. H. (1980) American Journal of Anatomy 157, 87-106 PARRY, D. A. D., CRAIG, A. S. & BARNES, G. R. G. (1978) Proceedings of the Royal Society of London B 203, 293- 321 RIEMERSMA, D. J. & SCHAMHARDT, H. C. (1985) Research in Veterinary Science 39,263-270 RIEMERSMA, D. J., SCHAMHARDT, H. C. & LAMMERTlNK, J. L. M. A. (1985) Biomechanics: Current Interdisciplinary Research. Eds S. M. Perren and E. Schneider. Dordrecht, Nijhoff. PI' 731-736 SCHAMHARDT, H. c., HARTMAN, W. & LAMMERTlNK, J. L. M. A. (1985) Research in Veterinary Science 39, 139-144 SOROKIN, A. P. & EFIMOV, A. P. (1980) Archil' Anatomii, Histologii i Embriologii 9, 76-81 STEGEMANN, H. & STALDER, K. (1967) Clinica Chimica Acta 18,267-273 WEBBON, P. M. (1973) Equine Veterinary Journal S; 58-64

Accepted February 3, 1986