The systematic study of the orbital veins began in the 18th century.
2 Almost up to the present, anatomic records show two different views of the organization of the orbital venous system. One of these is that the SOV corresponds fairly accurately to the ophthalmic artery in its course and branching pattern.
3 This view is challenged, however, by a second, completely opposing view that is supported by most investigators and appears today to be generally accepted. Among these are early anatomists, including Soemmering
4 and later investigators,
5 6 7 8 9 10 11 who stated that in contrast to veins elsewhere in the body, orbital veins do not accompany the arterial system. In regard to the degree of variability of the orbital venous system, most investigators agree. The SOV presents an especially constant course amid a number of variations.
12 In particular, the muscular branches appear to be far less constant than the principal veins,
13 and of the principal veins, the SOV is more constant than the IOV.
14
The SOV is convincingly the most frequently described vein. Most investigators have described it as formed by the joining of two contributing veins, the superior (supraorbital vein) and inferior (angular vein) roots of the SOV.
5 15 Nevertheless, some investigators have stated that the SOV has only a single origin, arising predominantly or even exclusively from one trunk anteriorly.
16 As Brismar
10 suggested, the finding of a single origin on phlebograms may be due to the technique used (i.e., angular vein approach without sufficient compression of the angular and frontal veins). An accurate account of the topography of the roots of the SOV is given by Murakami et al.
17 The superior root enters the orbit, running slightly backward, medially and cranially, under the orbital roof. It then unites with the inferior root at the cranial aspect of the medial extension of the levator palpebrae superioris muscle, a few millimeters behind the reflected tendon of the superior oblique muscle. The inferior root arrives in the orbit through an opening in the orbital septum, perforating the orbicularis oculi muscle before joining the superior root at the anteromedial orbit. As most have described,
5 17 from the point of union of the two roots, the SOV runs through the anterosuperior part of the orbital cone in adipose tissue to the medial border of the superior rectus muscle. It is surrounded by elastic tissue, which connects it with neighboring structures, such as muscles. From its entry into the muscle cone, the SOV runs backward and laterally under the superior rectus muscle and above the optic nerve before crossing to reach the medial border of the superior rectus muscle. Koornneef
18 19 20 described the SOV as being suspended in a hammock of connective tissue septa beneath the superior rectus muscle. Next, it follows the lateral border of the superior rectus muscle, descending dorsomedially into the tight connective tissue of the superior orbital fissure.
5 Finally, the SOV drains into the cavernous sinus.
In contrast to the SOV, the IOV is often not clearly described. Three different veins are described. One of these is the IOV, and the other is either the infraorbital vein
21 or, in more recent reports, the veine ophthalmique moyenne (middle ophthalmic vein).
5 Some investigators explicitly deny the existence of an infraorbital vein and consider the IOV comparable with the infraorbital artery in course and position.
22 Huber
23 stated that the inferior venous system exists only as a venous network or plexus. Most investigators agree that an IOV originates from a plexus.
Henry
13 was the first to describe a veine ophthalmique moyenne. He considered the vein to be a posterior muscular vein, arising within the muscle cone from one of several origins: the lateral side of the medial rectus muscle at its caudal end, the medial collateral vein, the lateral collateral vein, or the inferior aspect of the lateral rectus muscle. According to Brismar,
11 the veine ophthalmique moyenne was present in approximately 20% of his phlebograms. Jo and Trauzettel
15 were able to identify the vein in only 1% of their dissections.
Brismar
11 was also one of the few investigators who saw the medial ophthalmic vein. Forty percent of his phlebograms demonstrated such a vein. The medial ophthalmic vein arises from either the angular vein or the anterior segment of the SOV. It runs posteriorly along the orbital roof close to the medial orbital wall to either descend into the cavernous sinus directly or rejoin the SOV in the posterior orbit.
1
There are many tributaries from the SOV and IOV. Among these, the most clinically significant is the central retinal vein. The vein leaves the optic sheath at a variable distance from the globe, anterior to the central retinal artery. It has a large anastomosis with the SOV.
24
Two ethmoidal veins have been observed by most investigators. The anterior ethmoidal vein normally drains into the inferior root of the SOV, whereas the posterior ethmoidal vein often drains into a venous network under the orbital roof.
5
Another group of veins, mostly four but sometimes five, is called the vortex (or vorticose) veins. They obliquely perforate the sclera of the globe just posterior to its equator and drain the choroidal network.
14
According to Bergen,
5 Sesemann
22 was one of the first to note the presence of collateral veins connecting the SOV with the IOV. They are the anterior, medial, lateral and posterior collateral veins. Brismar
11 was able to identify the anterior and medial collateral veins in more than 90% of his phlebograms. The lateral collateral vein was visible in approximately 70% and the posterior collateral vein in only approximately 20%.
Description of another functionally important vein, the lacrimal vein, varies greatly between different reports depending on the method the investigators used. Brismar,
11 who used phlebograms, was unable to identify the vein at all in 75% of his cases. The lacrimal vein is better described by those investigating through dissection. The origin of the vein is formed by the union of a principle lacrimal vein and an accessory branch.
14 17 It also has abundant communications with extraorbital and intraorbital veins. For instance, it frequently communicates with the palpebral and conjunctival veins.
13
The muscular veins appear to be far less constant in their distribution and trajectory than the principal veins. Because of their relative variability, most investigators provide only a rough outline of the course of these veins. Basically, muscular veins for the superior oblique, superior rectus, medial rectus, and lateral rectus join the SOV, whereas for inferior rectus and inferior oblique muscles, the veins drain into the inferior venous plexus.
13
According to the literature, the orbital venous system also has abundant interconnections with extraorbital venous structures. The cavernous sinus is the most important posterior communication.
5 Another frequently mentioned communicating vein is that which travels through the inferior orbital fissure and connects the IOV with the pterygoid plexus.
14 24 Anteriorly, the orbital venous system communicates with the facial venous system through two connections. One is the angular vein, which directly bridges the inferior root of the SOV to the facial vein.
5 This communication is often used by radiologists for contrast-filling of the orbital veins.
25 A second communication to the facial venous system is the supraorbital vein, which links the frontal vein.
5
In the present study, there were several findings that reveal previously undescribed aspects of the orbital venous system, whereas others confirm the results of past studies. The SOV in this study showed some significant variations. One is the variant anatomy of its roots. A single origin was seen in more than 10% of the dissections. Although it appears to be an uncommon variation, it may have clinical implications, as, for example, in the SOV approach to CCSFs, where localization of the roots of the SOV is important.
26 Thus, in cases in which a search for the superior root (supraorbital vein) is difficult, the clinician should be aware of the existence of a possible single origin, which is invariably found inferior to the trochlea.
Another clinically relevant finding in this study is the variant anatomy of the medial ophthalmic vein, which was present in 25% of the dissections. As illustrated in
Figures 3 and 4 , it forms a circular loop bridging the SOV with other surrounding veins. This may promote development of venous thrombosis due to the potential increase in turbulent flow in that area. In 2 of 12 cases of embolization of the CCSF through the SOV (McNab A, personal communication, 2002), partial thrombosis of the SOV was evident from a midpoint of the orbit extending anteriorly. This partial thrombosis could have arisen at a point of sluggish or turbulent flow, as might occur in a venous system
(Figs. 3 4 8 9) . The connective tissue hammock suspending the SOV under the superior rectus, as described by Koornneef,
18 could also provide a point of kinking and possible thrombosis anterior to this point.
In many reports, it is emphasized that the SOV proceeds in a course completely separate to that of the ophthalmic artery.
10 17 However, the anterior portion of the first part of the SOV is sometimes found to be accompanied by the ophthalmic artery. Although the ophthalmic artery has an observably convoluted course, its anterior part is usually fairly straight, running just inferior to the first part of the SOV. Moreover, according to the results, the vortex veins were observed to be very tortuous compared with other orbital veins. Their tortuosity, which has been commented on infrequently in past studies, probably allows free rotation of the eyeball without stretching the veins. Another major contradictory finding is the anatomy of the medial ophthalmic vein. As this study shows, the medial ophthalmic vein has a very short course, running within the muscle cone and draining invariably into the SOV
(Figs. 3 4 8 9) . This is, however, inconsistent with some textbooks,
1 and the results from most of the past studies.
5 10 For instance, Dutton and Waldrop
1 state that the medial ophthalmic vein runs along the medial orbital wall in the extraconal space, and their schematic illustration shows that the vein proceeds in a long course, draining directly into the cavernous sinus.
The dissections of the inferior orbit also demonstrated variable anatomy. Only one IOV was evident in the inferior orbit, originating from the inferior ophthalmic venous plexus
(Fig. 1) . This finding is contradictory to those in many previous studies. Despite much effort, the veine ophthalmique moyenne controversially described by many investigators
5 13 was not found in any of the dissections. Because the inferior medial and lateral vortex veins are also closely associated with the IOV, it is possible for them to be mistaken for an additional IOV.
In conclusion, the orbital venous system has a complex structural organization. It can be separated into two systems: the superior orbital venous system and the inferior orbital venous system. The principal vein in the superior orbital venous system is the SOV, which arises from the joining of two extraorbital veins: the supraorbital and angular veins. The SOV is divided into three parts, and each part proceeds in a different course and communicates with different tributaries. The course of the SOV is generally constant, running backward medially to laterally in relation to the superior rectus muscle. It terminates in the cavernous sinus after passing through the superior orbital fissure. The IOV is the main vein in the inferior orbital venous system. It originates as a venous plexus between the inferior rectus muscle and the optic nerve. Running inferior to the optic nerve and with a very small caliber, the IOV is often very difficult to locate. In the posterior orbit, the IOV usually joins the SOV, but it occasionally drains directly into the cavernous sinus. Between the two venous systems, there are abundant communications through medial and lateral collateral veins.
Both of the systems exhibit several variations, but the inferior orbital venous system appears to be less consistent than the superior orbital venous system. Nevertheless, the variations of the SOV have more relevant clinical implications.
In this study, the anterior approach used to dissect the fresh specimen had one significant limitation. The veins communicating with the orbital floor and the pterygoid plexus had to be incised to allow the dissection to be performed. Therefore, details of these connections could not be identified.