|Home | About | Journals | Submit | Contact Us | Français|
The microvasculature in the cortex and marrow of the adult canine tibial diaphysis was filled with the silicone elastomer Microfil, the bone was decalcified, and the water was replaced with methyl salicylate to permit three-dimensional visualization of the microvascular arrangements. The tibial nutrient artery was seen to supply the marrow and the cortex via parallel, independent sets of arterioles and terminal capillary beds. No arteriolar or capillary anastomoses were observed linking these separate beds. The major portion of the venous drainage was found to be via small venules through the cortex into periosteal veins. Many small venules draining the medullary capillaries. penetrated the cortex, and there were a few larger emissary veins, including the nutrient vein. Because the marrow and cortex have separate capillary beds in parallel, microsphere deposition should be appropriate for estimating the regional blood flows.
CLINICAL RELEVANCE: The results of this study should be of concern to surgeons who perform whole diaphyseal bone replacements, as the effluent venous vessels are important in re-establishing the circulation by microsurgical methods.
The microvascular anatomy of the tibial cortex of the immature dog has been studied by both microangiographic and histological methods. Certain studies have suggested that the arteriolar branches from the nutrient artery supply the capillaries of the cortical bone7,9.What the relationships are between the vascular bed of marrow and that of bone, however, is controversial. Whereas De Bruyn et al. stated that the centripetally oriented osteal vessels are connected in series with the sinusoids of the marrow, our own physiological5 and anatomical7 studies have suggested that the marrow and the cortical circulations are in parallel rather than in series. Antipyrine washout studies5 suggested that 70 per cent of the nutrient artery flow exchanged with cortical bone and 30 per cent, with marrow.
Anatomical data are needed in order to interpret indicator-dilution techniques in the measurement of the blood flow or transcapillary permeation to cortical bone and of marrow6, and for the estimation of regional blood flow with labeled microspheres4. Furthermore, more specific knowledge of effluent venous vessels in the tibial diaphysis will be of value to surgeons attempting whole bone replacement by microsurgical techniques. The purpose of this study was to determine the relationships between the vessels of cortical bone and those of marrow in the canine tibial diaphysis. The anatomical arrangements of the arterial and venous systems in the dog are quite similar to those in humans8.
The bone was perfused with Microfil, a silicone rubber (Canton Biomedical Products, Boulder. Colorado) that contains white pigment particles ranging in size from < 0. 1 micrometer to one or two micrometers, with occasional particles as large as five micrometers; the viscosity is fifteen to twenty-five centipoise. After addition of hardener. this substance cures with a non-exothermic reaction; an unleaded variety was used. The injectate will not take a histological stain, and it cannot be coated for studies by electron microscopy. The polymerized Microfil is stretchable, but because the vessels are fine it is fragile and therefore must be handled with care, since decalcification and dehydration do weaken the surrounding supporting structures.
Six adult dogs were used. In three. the tibial nutrient artery was cannulated with a polyethylene catheter of 0.023-inch (0.06-millimeter) internal diameter; in the other three, the popliteal artery was cannulated. A slow, continuous infusion of normal saline with heparin was performed via the cannulated artery. This was done to ensure patency of the vascular bed while the Microfil was prepared. After preparation of the perfusate, it was injected with a hand syringe at a pressure of 100 to 150 millimeters of mercury. The injection was pulsating in that it varied between 100 and I50 millimeters of mercury at a rate of seventy pulsations per minute. Pressures were recorded with an anaeroid manometer. Injections were made until the Microfil was flowing freely in the venous system and then continued for three to five additional minutes.
The dogs were killed by an overdose of sodium thiopental and the hind limbs were disarticulated at the hip. The limbs were placed overnight in a refrigerator to allow the Microfil to set firmly. The following day the tibiae were dissected, care being taken to preserve the periosteum and the muscle attachments. The diaphysis was cut in one-centimeter cross sections or two-centimeter longitudinal sections, and the sections were placed in 10 per cent formalin in decalcifying solution (850 milliliters of 10 per cent formalin and 150 milliliters of formic acid) until complete decalcification could be observed by roentgenographic study. They were then dehydrated in increasing concentrations of alcohol, ending with absolute alcohol. The process of decalcification usually required between two and three weeks; the alcohol changes were done every twenty-four hours. When dehydration had been completed, all specimens were immersed in methylsalicylate. Clearing of the specimens required twenty-four to thirty-six hours.
The specimens were viewed with a Wild Heerbrugg M74 dissecting microscope. This permitted excellent stereoscopic viewing, which unfortunately cannot be reproduced photographically. Photographs were made with HC-135 Kodak film with a Nikon F2 camera. The films were developed with Prostar processing.
The specimens were photographed while they were immersed in methylsalicy late, and the source of light was two fiberoptic units at different angles. This arrangement made it possible to change the position of a specimen at will in order to observe specific structures.
There was no difference in filling of the specimens whether the injection was via the popliteal artery or the nutrient artery. The perfusate appeared on periosteal surfaces and at muscle attachments before it appeared in the nutrient vein that accompanied the nutrient artery. This is not surprising, as the studies by Cofield et al. indicated that only 5 to 10 per cent of the effluent venous diaphyseal blood leaves via the nutrient vein. For this reason, it is important to prolong injection of the perfusate for several minutes beyond the appearance of the Microfil in the nutrient vein in order to fill medullary structures. The injection of the tibia can be accomplished more easily by injection into the popliteal artery.
Our concept is summarized in diagrammatic form in Figure 1 and in the cross section in Figure 2. The nutrient artery enters at about a 45 to 60-degree angle without branching during passage through the cortex; within the marrow it divides into two major branches, ascending and descending the shaft. Arteriolar sub-branches (lateral branches) lead directly to the cortical bone; others go to the marrow. The cortical arterioles are directed radially and enter the cortex singly or in bundles of two to six arterioles, and within the cortex they give rise to branches directed parallel to the tibial axis as well as radially (Fig. 2). The drainage of the cortical capillaries is principally into the periosteal venous complex and almost none drain into endosteal veins.
The arterioles supplying capillaries for the marrow are short and profusely branched, with short distances between the branches. The venous drainage of marrow capillaries is of two types: the first is to small venules that traverse the cortex directly, often entering the cortex in the company of one or more arterioles; the second is into the central medullary sinus, which is in continuity with the nutrient vein. It is likely that there is some confluence within the outer layers of the cortex of venules draining marrow sinusoids and those draining the cortical capillaries in the haversian canals.
The major feature is the degree of independence of the capillary supplies to marrow and cortical bone. In the diaphyseal region the nutrient artery supplies both within a few millimeters; blockage of the nutrient artery will leave both the marrow and the cortex completely dependent on blood from metaphyseal collaterals in the medulla. There does not appear to be other than a minor supply from periosteal arterioles to the external cortex. But the capillary beds of marrow and cortex are totally independent, and there is no supply from cortical capillaries to medullary sinusoids — that is, there is no portal system linking these two capillary beds.
Figure 2 illustrates the pattern of radially oriented arterial branches from the nutrient artery. These are conduit vessels that traverse the marrow to enter the endosteal surface of the cortex . Figure 3 , a cross section taken somewhat at an angle, reveals the conduit vessels — lateral branches of the nutrient artery — traversing the cortex and anastomosing with the periosteal network of vessels. An arteriolar conduit vessel traverses the marrow from a branch of the nutrient artery and enters the cortex endosteally. These lateral branches of the nutrient artery divide further, and each major branch takes its own path. Figure 4 reveals this pattern more specifically: the arterioles give off multiple branches within the cortex which run in the longitudinal axis of the bone and obliquely toward the periosteal surface, to anastomose with arterioles in the substance of the periosteum.
Figure 5-A shows an opening of a canal on the endosteal surface of the cortex . Just before entering the osteal canal, this large, radially directed arteriole divides into smaller arterioles to supply the capillaries of the haversian canals. Figure 5-B illustrates a large vascular canal containing venules and arterioles. The vessels that branch off and run longitudinally in the haversian canals are capillaries or exchange vessels, not conduit vessels.
Within the marrow, or the medullary cavity, the branches of the lateral nutrient artery (conduit vessels) may travel to the cortex without dividing (Fig. 3). Other branching arterioles supply marrow sinusoids (Fig. 6). Most of the branching of the arterial supply to the marrow sinusoids occurs close to the endosteal surface. The sinusoidal network is very dense, even in these adult dogs with fatty marrow (Figs. 2, ,6,6, and and7).7). The termination of the arterioles into the sinusoids is either in a funnel-like transition or at sharp angles . The arterioles that enter the marrow sinusoids originate from the branches of the lateral nutrient artery before they penetrate the endosteal surface of the cortex.
Figure 7 shows a transcortical emissary vein, which serves to connect the veins within the medullary cavity to the venous network on the periosteal surface. The size of this emissary vein suggests that periosteal emissary veins are an important avenue for drainage of venous blood from the tibial diaphysis.
The endosteal sinusoids lead to collecting sinusoids. These collecting sinuses may be located centrally in the bulk of the marrow substance or at the periphery near the endosteal surface. The former converge and drain into the large central medullary sinus, partially via the nutrient vein but principally via the metaphyseal and epiphyseal veins at both ends of the bone. The peripherally located collecting sinuses provide a second pathway for drainage of venous blood, which occurs from the endosteal sinusoids through the cortical bone to the periosteal surface as illustrated in Figure 8 . The venous vessel may travel in the same canals in which the radially flowing cortical arterioles are located (Figs. 5-B and and8).8). After traversing the cortex, the vein joins the periosteal network of veins (Fig. 8). These venous vessels are distinguishable from the arterioles because they are larger than the cortical arterioles and they may be more irregular in diameter or shape, probably in part because the distending pressure of the Microfil is not so high (Fig. 9). This type of vessel is not to be confused with an emissary vein (Fig. 7), which, although originating in the sinusoids at the endosteal surface, follows a more direct course through the cortex without branching and drains into a large collecting periosteal vein. These emissary veins are similar to the veins seen in the human tibia8.
The periosteal complex is predominantly venous, as illustrated in Figure 10. The great density in this region is understandable in view of the fact that much of the venous drainage from the diaphysis enters this complex (and only about 10 per cent enters the nutrient vein) and that, in addition, the periosteal surface can be an area of active bone formation.
Previous studies from this laboratory, utilizing microangiography, histological techniques, and clearing of decalcified sections, have outlined the extra-osseous and intra-osseous distribution of vessels to the canine and human tibiae7,8. However, the present study has allowed more complete examination of the circulation of marrow and of cortical bone . The marrow of the adult dog is an adipose tissue, and the sinusoids of fat marrow appear to be structurally sufficiently well enclosed so that the Microfil does not leak out, as one would expect if dog capillaries were similar to the closed, continuous capillaries observed in the human by Trueta and Harrison. The arteriolar branches of the lateral branches of the nutrient artery (Figs. 1 through through4)4) are conduit vessels, whereas the capillaries in the haversian canals are exchange vessels. The sinusoids located near the endosteal surface are supplied from small arterioles that branch off before the major conduit vessel enters the bone. It is possible that it is these vessels that have led some previous workers (for example, Dc Bruyn et al.) to the view that there might be a system of cortical and marrow vessels in series.
An additional source of confusion is the great density of the venous vascular network. Many of these observations would be difficult to make without a dissecting microscope and the careful teasing-away of sinusoidal vessels (Fig. 6) which allows demonstration of a separate marrow arteriole to the endosteal sinusoids.
The anatomical studies are consistent with past studies5 on relative tissue-perfusion rates of cortex and marrow, which revealed that approximately 70 per cent of nutrient-artery perfusion is distributed to cortical bone and 30 per cent, to the marrow.
The important observation is that marrow and cortex have parallel, separate circulations; each is supplied by terminal arterioles . The concept of parallel circulations in bone does not exclude the concept of centrifugal flow, which is physiologically sound. Since the pressure in the medullary cavity is greater than the pressure registered in the extra-osseous soft tissues11, centrifugal flow is the natural direction of flow for both systems.
In measuring blood flow with microspheres, it is as sumed that the number of spheres entering each tissue will be proportional to the blood flow, a situation that can occur only if the arterial supplies are parallel. The present anatomical study is therefore an important validation of this basic premise in the application of the microsphere technique in the dog.
Blood may flow to marrow and cortex at different rates; the separateness of the two systems may explain why perfusion at physiological pressure does not fill the venous system within a few minutes. Furthermore, it may explain why materials may reach the periosteal veins and return to the popliteal vein several minutes before they appear in the nutrient vein.
Such observations would suggest that the cortical bed draining into the low-pressure periosteal venous system has a lower vascular resistance than the marrow circulation, draining as it does into a high-pressure intra medullary vein and to an exit via the nutrient vein. This description is consistent with Brånemark's observation that the flow in the sinusoidal network is sluggish and sometimes stops completely.
De Bruyn et al . conducted their experiments in guinea pigs and rabbits; thus the possibility of species differences from larger mammals should be considered. It is also important to realize that in spite of good stereoscopic vision, structures in close proximity, such as those shown in Figure 5-B , would appear as a single vessel if observed with transmitted-light microscopy. The same holds true for microangiography, in which all superimposed images would appear as a single shadow.
Complete filling of the marrow structures would be undesirable, as it would obscure details of vascular architecture, particularly the relationship of the arteriole to the marrow sinusoid. Improved filling can be obtained by raising the pressure of the extra-osseous soft tissue to make it equal to or slightly higher than that of the intramedullary pressure as measured by Wilkes and Visscher; this would force the perfusate to find intramedullary pathways of diversion and drainage via the metaphyseal systems. This, however, departs from normal physiology, and probably the same result would be obtained simply by prolonging the injection as long as necessary to produce perfusion of the slow-flowing sinusoidal network.
*This investigation was supported in part by Research Grants AM-l5980 and HL-l9139 from the National Institutes of Health, Public Health Service.