Secondary Logo

Journal Logo


The Origin of the Spinal Subdural Space: Ultrastructure Findings

Angel Reina, Miguel, MD*,; De Leon Casasola, Oscar, MD†,; López, Andrés, MD*,; Antonio De Andrés, José, MD‡,; Mora, Miguel, MD§,; Fernández, Agustín, MD

Author Information
doi: 10.1097/00000539-200204000-00040
  • Free


The subdural space is described by textbooks as “a potential cavity between the dura and arachnoid mater”(1–3); it has been visualized by contrast injection, and epidural catheters have been introduced using radiological techniques (4–8). However, when Blomberg (9) attempted to evaluate the subdural space by endoscopic methods in cadavers, he could not visualize it in all cases.

During the last decades, different authors (10–16) have studied the ultrastructure of the cranial dura-arachnoid interface in animal and human meninges and have doubted the existence of a subdural space. Vandenabeele et al. (17) studied samples of the spinal meninges, and they could not find it. In all these studies, the dura-arachnoid interface has been identified by different names: the medial border of the spinal dura mater (10), dural border cell layer (18), subdural mesothelium (13), or subdural compartment (14). We observed the ultrastructure of the dura-arachnoid interface under transmission and scanning electron microscopy from samples of spinal meninges to evaluate the presence or absence of a subdural space.


The dural sac and its neural content at the lumbar level were removed from three donors immediately after their death. Approval from the research ethics committee and family consent for the donation of the organs and the procedures included in this research was obtained. The subjects were 48, 55, and 60 yr of age and had been admitted to the intensive care unit because of cerebral hemorrhage and subsequently diagnosed with brain death. The laminectomy and dissection were performed with the help of an orthopedic surgeon. The dural sac was removed in a block from T11–L5 section. Extracted specimens containing the dural sac were studied without further manipulation by scanning and transmission electron microscopy. After subsequent dissection, they were also studied by surgical optical microscopy. Care was exercised to observe specimens where the dura mater was in continuity with the trabeculated portion of the arachnoid as well as specimens where the laminar arachnoid was macroscopically separated from the dura mater as a result of the dissection.

Preparation of Samples

Transmission Electron Microscope.

The specimens were fixed for 4 h in a solution of glutaraldehyde 2.5% and a buffer phosphate solution to a pH value of 7.2–7.3. They were later fixated with a solution of 1% osmium tetroxide for 1 h. The specimens were dehydrated with acetone and were soaked in resin epoxy Epon 812. Control group slides were dyed with Richardson’s methylene blue dye. Ultra thin slides of 70 nm of thickness were made with an ultramicrotome and treated with 2% acetate of uranilo solution and with Reynolds lead citrate solution. The specimens were observed under a Zeitz transmission electron microscope (EM 902; Carl Zeitz, Oberkochen, Germany).

Scanning Electron Microscope.

The samples were fixed by immersion for 4 h in 2.5% glutaraldehyde and phosphate solution buffered to a pH value of 7.2–7.3. They were later dehydrated in acetone and pressurized under CO2 to reach the critical point. A carbon layer was then deposited on the samples to a thickness of <200 Armstrong and covered with a gold microfilm. The samples were studied with a JEOL JSM 6400 scanning electron microscope (JEOL, Tokyo, Japan).


Observations During the Dissection.

We observed a translucent membrane that detached from the internal coating of the dural sac when light pressure or light friction was applied on its internal surface. This membrane wrapped around all the nerve roots at their exit from the medullary cone and the spinal cord, joining the internal surface of the dura mater only in the areas where the nerve roots come out from the subarachnoid space to cross the dural sleeves and in some isolated randomly located points within the dural sac (Fig. 1). As a result, this membrane formed a cylinder with lateral short sleeves and edges that joined the internal surface of the dural sleeves.

Figure 1
Figure 1:
Dissection of the dural sac at the level of cauda equina. The arachnoid mater appears as a translucent membrane. This has been separated from the internal surface of the dura mater. Caudal nerve roots and epidural fat are also seen.

Electron Microscopy.

The results of the findings obtained by transmission electron microscopy (TEM) or scanning electron microscopy (SEM), referring to the same structure, will be described by indicating the method in parentheses. The arachnoid membrane was seen as a compact laminar portion covering the dural sac in its inner surface and a trabeculated portion that extended like a spider web to the pia mater of the spinal cord and the medullary roots (SEM). Between the laminar arachnoid portion and the most internal layer of the dura mater, there was a cellular interface (TEM) that we called the dura-arachnoid interface (Fig. 2).

Figure 2
Figure 2:
Dura-arachnoid interface is seen below the dural lamina (most internal aspect of the dura). The dura-arachnoid interface is filled with neurothelial cells and amorphous material (transmission electron microscopy [TEM], magnification ×5000, bar = 1 μm).

The subdural space, as classically described, was not present (TEM) in the specimens where the dural sac had its entire thickness and the dura mater was macroscopically in continuity with the arachnoid trabeculae. The dura-arachnoid interface was occupied by neurothelial cells, and there was not a real space between the arachnoid and dura mater membranes (TEM, SEM;Fig. 2). These cells were arranged concentrically below the dura mater, and their thickness varied between 5 and 7 μm, distributed in 4–8 parallel cellular planes (TEM). The length of each neurothelial cell was longer than 100 μm, and their width varied between 0.5 and 1 μm (TEM, SEM;Fig. 3). They were oriented in different directions, and their morphology presented numerous ramifications (SEM;Fig. 3). The union of elongated and flat cells formed each plane with multiple intercellular “digitlike” cytoplasmic prolongations (TEM). Depending on the cells studied, an intercellular space or cells joined by scarce desmosomes or other specialized bridging were found. When the intercellular space was present, its thickness varied. Sometimes, this space resembled lacunar spaces occupied by amorphous material where the collagen or elastic fibers were almost absent (TEM;Fig. 2).

Figure 3
Figure 3:
Dura-arachnoid interface. Neurothelial cells have a syncytial structure and show ramifications. They present an irregular diameter (approximately 1 μm) and a smooth surface. Collagen fibers can also be seen, and they belong to the laminar portion of the dura mater. However, they are about 10× thinner, and they do not have ramifications. Between the neurothelial cells, collagen fibers may be seen from the last dural lamina of the dura mater (scanning electron microscopy [SEM], magnification ×3000, bar = 10 μm).

The lack of fibers between these cells makes them different from adjacent cellular planes in the dura mater and the arachnoid membrane (TEM). The neurothelial cells presented a nucleus with disperse chromatin, few mitochondria, a poorly developed endoplasmic reticulum, and pinocytotic vesicles (TEM). In those areas with fissures, the tearing extended mainly through the amorphous material. In larger planes, a fissure was generated and expanded towards the lateral zones where a significant amount of amorphous material was usually present (TEM). Once the epicenter of the fissure is established, it enlarges towards areas of low mechanical resistance where the amorphous substance is located between the neurothelial cells. There were lacunar zones with large amounts of amorphous substance. In these zones, the coherence and mechanical resistance appeared to be less because the fissure developed easily. When the tips of a new fissure hit a more resistant zone, like joints between neurothelial cells, its advancement causes one of these cells to break or, if the generated forces are too weak, the fissure expands in another direction (SEM). This explains the presence of folded cellular fragments at the surface of the generated subdural fissure (SEM).

In those specimens where the dura mater was macroscopically separated from the arachnoid, we could observe two laminae with a thickness of 300 and 40 μm, respectively (SEM;Fig. 4). Both laminae were separated by a space that corresponds to the subdural space (SEM), and both surfaces in contact with the subdural space were formed by neurothelial cells (Fig. 4). Parallel to the subdural space, and inside the dura-arachnoid interface thickness, we encountered other subdural fissures (two or three small fissures parallel to the longitudinal axis) (SEM;Fig. 4). We called these spaces the secondary subdural spaces. The surfaces adjacent to the subdural space appeared smooth, shiny, and had some cytoplasmic fragments originated from broken neurothelial cells (SEM). The shiny appearance observed via electron microscopy was produced by the intercellular amorphous substance (SEM).

Figure 4
Figure 4:
Subdural space seen by scanning electron microscopy, limited by dura mater (thickness of 300 μm) and laminar arachnoid (thickness of 40 μm). The internal surface of both membranes is covered with neurothelial cells. Inside the dura-arachnoid interface and parallel to the subdural space, there are other secondary subdural fissures (magnification ×200, bar = 100 μm).


We could not find the subdural space in those samples of human spinal meninges where surgical manipulation was avoided. The subdural space was not seen under transmission electron microscopy. Instead, we found a compartment between the dura and arachnoid mater filled with neurothelial cells that we named the dura-arachnoid interface.

Although the number of subjects studied was very small, the absence of a subdural space in all of them creates doubts about the existence of a virtual subdural space, believed until now to be present in all individuals at birth. The use of transmission electron microscopy allowed us to identify the ultrastructure of the dura-arachnoid interface. The presence of neurothelial cells filling up the dura-arachnoid interface helped us to consider that the subdural space would not be comparable to other virtual spaces such as the interpleural space. We did not observe within the interface other structures such as collagen fibers; there were few intercellular joints that made this cellular plane softer with regard to the dura mater and arachnoid (neighboring cellular planes). There were large intercellular lacunar spaces separated by thin bridges of cytoplasm. These spaces were filled with an amorphous material of low resistance and unknown composition. Therefore, the cohesion forces at the level of the interface would be significantly lower. Using scanning electron microscopy, we observed the neurothelial cells tridimensional structure (Fig. 3).

A subdural space can appear if neurothelial cells break up because of pressure exerted by mechanical forces, air or fluid injection, creating fissures within the interface. Fissures could grow larger towards weaker areas, giving up a subdural space that could expand according to the pressure exerted.

We saw a number of parallel fissures under scanning electron microscopy (Fig. 5). These fissures can be produced within the dura-arachnoid interface in relation to specific tissue characteristics, depending on the distribution of the applied forces. Some fissures are produced in an incomplete form, whereas others expand to a greater extent, creating a real subdural space.

Figure 5
Figure 5:
Dura-arachnoid interface model. From top downwards there is the dural lamina (dura mater’s most internal portion) followed by the dura-arachnoid interface; the interface seems filled up with neurothelial cells, showing the formation of the subdural space. Below are the laminar and trabecular portions of the arachnoid mater.

We took samples from the cadavers immediately after death and avoided surgical manipulation of a known number of specimens to prevent the results from being altered by artifacts, for these could be responsible for the appearance of a false subdural space. In our study, we applied light pressure over a number of samples, and it was possible to observe how the arachnoid layer separated from the dural sac, creating an extensive subdural space.

These findings confirm the view of other authors, in particular, the results of Vandenabeele et al. 17, who obtained human specimens of the dura mater and arachnoid at spinal level during surgical procedures on the spine. In contrast to Vandenabeele et al., we studied the complete lumbar segment of the dural sac, avoiding surgical manipulation of the samples. This new concept may have clinical implications because it could better explain the variability of the onset and extension of the neural blockade caused by a nonintentional subdural injection.

The authors wish to thank F. Sellers, MD, from the Department of Orthopedic Surgery of Mostoles Hospital and F. Maches, MD, from the Department of Anesthesiology of Madrid Montepríncipe Hospital, Spain.


1. Bridenbaugh PO, Greene NM, Brull SJ. Spinal (subarachnoid) neural blockade. In: Cousins MJ, Bridenbaugh PO, eds. Neural blockade in clinical anesthesia and management of pain. 3rd ed. Philadelphia: Lippincott-Raven, 1998: 206–7.
2. Bernards CHM. Anestesia epidural y raquídea. In: Barash PG, Cullen BF, Stoelting RK, eds. Anestesia clínica. 3rd ed. México: McGraw-Hill Interamericana, 1999: 761.
3. Brown DL. Spinal, epidural and caudal anesthesia: anatomy, physiology and technique. In: Chesnut DH, ed. Obstetric anesthesia: principles and practice. 2nd ed. St Louis: Mosby, 1999: 189.
4. Smith GB, Barton FL, Watt JH. Extensive spread of local anaesthetic solution following subdural insertion of an epidural catheter during labour. Anaesthesia 1984; 39: 355–8.
5. Hartrick CT, Pither CE, Umeshraya P, et al. Subdural migration of an epidural catheter. Anesth Analg 1985; 64: 175–8.
6. Boys JE, Norman PF. Accidental subdural analgesia. Br J Anaesth 1975; 47: 1111–3.
7. Manchanda VN, Murad SHN, Shilyansky G, Mehringer M. Unusual clinical course of accidental subdural local anesthetic injection. Anesth Analg 1983; 62: 1124–6.
8. Ralph CJ, Williams MP. Subdural or epidural? Confirmation with magnetic resonance imaging. Anaesthesia 1996; 51: 175–7.
9. Blomberg RG. The lumbar subdural extraarachnoid space of humans: an anatomical study using spinaloscopy in autopsy cases. Anesth Analg 1987; 66: 177–80.
10. Waggener JD, Beggs J. The membranous covering of neural tissue: an electron microscopy study. J Neuropathol Exp Neurol 1967; 26: 412–26.
11. Schachenmair W, Friede RL. The origin of subdural neomembranes. I. Fine structure of dura-arachnoid interface in man. Am J Pathol 1978; 92: 53–68.
12. Yamashima T, Friede RL. Light and electron microscopic studies on the subdural space, the subarachnoid space and the arachnoid membrane. Neurol Med Chir (Tokyo) 1984; 24: 737–46.
13. Alcolado R, Weller RO, Parrish EP, Garrod D. The cranial arachnoid and pia mater in man: an anatomical and ultrastructural observations. Neuropathol Appl Neurobiol 1988; 14: 1–17.
14. Orlin JR, Osen K, Hovig T. Subdural compartment in pig: a morphologic study with blood and horseradish peroxidase infused subdurally. Anat Rec 1991; 230: 22–37.
15. Frederickson RG. The subdural space interpreted as a cellular layer of meninges. Anat Rec 1991; 230: 38–51.
16. Haines DE. On the question of a subdural space. Anat Rec 1991; 230: 3–21.
17. Vandenabeele F, Creemers J, Lambrichts I. Ultrastructure of the human arachnoid mater and dura mater. J Anat 1996; 189: 417–30.
18. Nabeshima S, Reese TS, Landis DMD, Brightma MW. Junctions in the meninges marginal glia. J Comp Neurol 1975; 164: 127–70.
© 2002 International Anesthesia Research Society