FormalPara Key Points

Perispinal injection is a novel method of drug delivery to the CNS.

Perispinal injection utilizes the cerebrospinal venous system (CSVS) to facilitate drug delivery to the CNS by retrograde venous flow.

Macromolecules delivered posterior to the spine are absorbed into the CSVS.

1 Introduction

“It seems incredible that a great functional complex of veins would escape recognition as a system until 1940… In the first four decades of the last century, our knowledge of the vertebral veins was developed and then almost forgotten.” Batson, 1940 [1].

Physiological barriers, including the blood–brain barrier (BBB) and the blood–cerebrospinal fluid barrier (BCSFB), prevent large molecules (those with a molecular weight [MW] over 600) and many smaller molecules from efficiently penetrating into the central nervous system (CNS) after systemic administration [24]. Specialized methods of drug delivery are required to maximize the therapeutic potential of fusion proteins, monoclonal antibodies, and other macromolecules for CNS indications [25].

One such specialized method, perispinalFootnote 1 administration, has been used to deliver the fusion protein etanercept (MW 150,000) to the CNS [616]. Perispinal administration involves delivery into the anatomic region posterior to the ligamentum flavum and the spinal canal, and is therefore less complicated than epidural or intrathecal injection [616]. Perispinal administration delivers a substance into the anatomic region drained by the external vertebral venous plexus (EVVP), a division of the cerebrospinal venous system (CSVS) [1, 620]. Blood within the EVVP communicates with the deeper, valveless, bidirectional venous plexuses comprising the remainder of the CSVS [1, 625]. Bidirectional and retrograde blood flow between the spinal and cerebral divisions of the CSVS is made possible by the absence of venous valves within the internal vertebral venous plexus [1, 1931]. The anatomical basis for perispinal administration of etanercept has its origins in the detailed studies of the CSVS published in 1819 and 1829 by the French anatomist Gilbert Breschet [21, 22] (Figs. 1, 2, 3). Breschet’s findings were nearly forgotten after the turn of the century, until they were rediscovered and confirmed by American anatomist Oscar Batson in 1940 [1, 20].

Fig. 1
figure 1

Cerebrospinal venous system. Detail of plate 5 from Breschet G, Recherches anatomiques physiologiques et pathologiques sur le systáeme veineux. Paris: Rouen fráeres; 1829. Courtesy of the Sidney Tobinick collection

Fig. 2
figure 2

Cerebrospinal venous system. Detail of plate from Breschet G, Essai sur les veines du rachis. Paris: Faculte de Medecine de Paris; 1819. Courtesy of the Sidney Tobinick collection

Fig. 3
figure 3

Cerebrospinal venous system. Detail of plate III from Breschet G, Recherches anatomiques physiologiques et pathologiques sur le systáeme veineux. Paris: Rouen fráeres; 1829. Courtesy of the Sidney Tobinick collection

Fig. 4
figure 4

Cerebrospinal venous system. Detail from Todd RB (ed), The Cyclopaedia of Anatomy and Physiology. 1847, page 630, Fig. 360; after Breschet (1829)

Fig. 5
figure 5

Spinal veins. From Gray, 1858. Figure 222, page 416 from Gray H and Carter HV, Anatomy, descriptive and surgical. 1st ed. London: John W. Parker and Son.; 1858; after Breschet (1829)

Fig. 6
figure 6

Spinal veins. From Gray, 1858. Figure 222, page 416 from Gray H and Carter HV, Anatomy, descriptive and surgical. 1st ed. London: John W. Parker and Son.; 1858; after Breschet (1829)

Fig. 7
figure 7

External vertebral venous plexus in the cervical subcutaneous space. Magnetic resonance image. Courtesy of the Sidney Tobinick Collection

Fig. 8
figure 8

Figures from Corning’s 1888 article, Corning JL, XXIA Further Contribution on Local Medication of the Spinal Cord, with Cases. Transactions of the Medical Society of the State of New York for the Year 1888. 1888: pages 260–269. Figure 1 depicts the solid needle, 3 inches in length, utilized in the apparatus depicted in Fig. 2. Figure 2 depicts the spinal canal, the ‘foramen vertebrae’, and the method Corning utilized to estimate the depth of the posterior border of same. Figure 3 depicts the hollow needle Corning utilized to deliver cocaine using the 6.2 ml syringe depicted in Fig. 4

Recently, additional supporting evidence, previously unrecognized, was discovered in the publications of American neurologist James Leonard Corning [3236]. In 1885, the same year that Paul Ehrlich provided the first experimental evidence of the BBB, Corning reported rapid onset of spinal anesthesia after perispinal injection of cocaine ‘between the spinous processes’ (i.e. by interspinous injection), likely as a result of Corning’s familiarity with Breschet’s anatomical findings [3236].

2 Breschet and the Cerebrospinal Venous System (CSVS)

“… blood is poured by the dorsi-spinal, the basi-vertebral and the spinal-medulli veins, and by the spinal plexus, depositing it to all parts along these veins….” Gilbert Breschet 1819 [21] (Fig. 2)

Gilbert Breschet (1783–1845), Professor of Anatomy at the University of Paris, surgeon to the Hotel Dieu in Paris, consulting surgeon to King Louis Phillipe, and member of the Royal Swedish Academy of Sciences, accurately detailed the anatomy and physiology of the spinal venous plexuses and their interconnections and drainage patterns, forming the basis for the modern conception of the CSVS [1, 17, 1923, 25, 30] (Figs. 1, 2, 3). In the second half of the 19th century, Breschet’s findings regarding the anatomy and physiology of the spinal veins were widely known in Europe and were detailed in major anatomy texts of the time, including Curveilhier’s Anatomy (1844); volumes III (1847) and IV (1852) of Todd’s Cyclopaedia of Anatomy and Physiology (Fig. 4); the first (London 1858) and later editions of Gray’s Anatomy (Figs. 5, 6); and the 1867 and later editions of Quain’s Anatomy [3741].

Fig. 9
figure 9

Positron emission tomography image, transverse section, of a living rat brain following perispinal extrathecal administration of 64Cu-DOTA-etanercept, imaged 5–10 min following the administration of etanercept. Note enhanced signal in the choroid plexus. Reproduced from Tobinick et al. [18]

The first of Breschet’s major works on the spinal veins, his “Essai sur les veines du rachis [Essay on the Veins of the Spine]”, was published in 1819 [21] (Fig. 2). This work secured for Breschet the highly sought-after post of Inspector-General of Anatomy at the Faculty of Medicine in Paris [21, 38, 42]. At the Faculty of Medicine, Breschet worked with an esteemed group, including the anatomists Jean Cruveilhier and Guillaume Dupytren [20, 21, 38, 43]. The treatise is divided into nine sections derived from Breschet’s “careful study of the sources, the path, the connections and the endings of the veins of the spine … lengthy, often repeated trials and errors [of specialized injections] … [and] dissection of parts of this venous system” [21]. Sections of the treatise discuss all of the interconnected venous plexuses of the spine, specifically including sections on the EVVP (the ‘dorsi-spinal veins’), the veins of the spinal cord, and the free communication of the blood flow within and between these venous plexuses and the cerebral veins (Figs. 1, 2, 3, 4). The treatise gives details regarding the specialized injection methods that Breschet utilized, including using ichtyocolle and “wax, soft turpentine, and a body of resin, colored with iron cyanide (Prussian blue)” to meticulously map out the anatomy, venous blood flow patterns and connectivity of the veins in and around the spine, the spinal cord, and the brain [21, 22, 37, 39, 40].

Breschet found that (1) the EVVP drains the anatomic area posterior to the spine, including the skin and muscles posterior to the spine (Figs. 1, 2, 3, 4, 5, 6, 7); (2) veins comprising the EVVP perforate the ligamentum flavum to join the internal vertebral venous plexus (Figs. 1, 2, 3, 4, 5, 6, 7); and (3) all of the spinal venous plexuses, including those of the spinal cord, were interconnected and that blood within the venous plexuses, constituting the CSVS, intercommunicates [21, 22, 37, 3941, 4446] (Figs. 1, 2, 3, 4, 5, 6, 7). Breschet and colleagues depicted the anatomic continuity of the spinal and cerebral venous systems (Figs. 1, 2, 3, 4). As Cruveilhier wrote, cerebral venous drainage could take place via outflow through the vertebral venous plexus [17, 1922, 37, 39, 45] (Figs. 1, 2, 3, 4).Footnote 2

The accuracy of Breschet’s pioneering descriptions has been confirmed by a multitude of subsequent anatomic and clinical studies [1, 17, 1923, 25, 27, 28, 30, 31, 47, 48] (review [17]). The First Edition of Gray’s Anatomy (1858) states:

“The veins of the spine are described and illustrated from the well-known work of Breschet (Gray 1858, Preface [39])… . The Dorsi-Spinal Veins [EVVP] commence by small branches, which receive their blood from the integument of the back of the spine, and from the muscles in the vertebral grooves. They form a complicated net-work, which surrounds the spinous processes, laminae, and the transverse and articular processes of all the vertebrae. At the bases of the transverse processes, they communicate, by means of ascending and descending branches, with the veins surrounding the contiguous vertebrae, and with the veins in the interior of the spine, in the intervals between the arches of the vertebrae, perforating the ligamenta subflava [ligamentum flavum]… .” Gray 1858, p. 415 [39].

Today, we owe Breschet a great debt for his detailed and accurate description of the spinal venous plexuses, long before the availability of radiologic methods of imaging vascular pathways [2022, 37]. Modern anatomical texts and reviews, including current editions of Gray’s Anatomy and Netter’s anatomical atlases, confirm Breschet’s findings [19, 25, 30, 4852]. Magnetic resonance imaging of the spine clearly depicts the EVVP in sagittal images of the spine (Fig. 7).

3 Corning and Perispinal Injection of Cocaine

James Leonard Corning (1855–1923), although born in the US, received his Doctor of Medicine degree from the University of Wurzburg in Bavaria, Germany, in 1878, prior to returning to the US to practice neurology in New York City [53, 54]. A year after Sigmund Freud and Carl Kollar published their 1884 papers noting the local anesthetic properties of cocaine, Corning began experimenting with neurological applications of cocaine [33, 55, 56]. He developed novel methods enabling the effective use of cocaine as a local anesthetic in lower concentration, thereby limiting its systemic toxicity [33, 57].

By 1885, Corning’s clinical experience with cocaine and knowledge of spinal venous anatomy led to the first demonstration of spinal anesthesia [3236]. In view of the place and details of his education, as well as his publications, Corning’s famous first use of cocaine for spinal anesthesia in 1885 was in fact likely based on Breschet’s findings regarding spinal venous anatomy and accomplished by perispinal injection [38, 42, 53, 54].Footnote 3, Footnote 4 Corning explained the scientific rationale in his famous 1885 paper [32]:

“… in order to obtain the most immediate, direct, and powerful effects upon the cord with a minimum quantity of a medicinal substance, it is by no means necessary to bring the substance into direct contact with the cord; it is not necessary to inject the same beneath the membranes, as in the case of the frog, since the effects are entirely due to the absorption of the fluid by the minute vessels. On the other hand, in order to obtain these local effects, it is first necessary to inject the solution in the vicinity of the cord, and secondly, to select such a spot as will insure the most direct possible entry of the fluid into the circulation about the cord … .

Protocol of Experiments … As the introduction of a hypodermic needle beneath the membranes of the medulla spinalis is not practicable without removal of the arches of the vertebrae (on account of the danger of wounding the cord), I decided to inject the anesthetic between the spinous processes of the lower dorsal vertebrae. I was led to resort to this expedient from a knowledge of the fact that in the human subject numerous small veins (venae spinosae) run down between the spinous processes of the vertebrae, and, entering the spinal canal, join the more considerable vessels of the plexus spinalis interna. From these theoretical considerations I reasoned that it was highly probable that, if the anaesthetic was placed between the spinous processes of the vertebrae, it (the anaesthetic) would be rapidly absorbed by the minute ramifications of the veins referred to, and, being transported by the blood to the substance of the cord, would give rise to anesthesia of the sensory and perhaps also of the motor tracts of the same… .”

Corning’s injections of cocaine, “placed between the spinous processes” using a hollow needle that he later depicted (Fig. 8), resulted in spinal anesthesia, as he detailed:

“Experiment I.—This was performed on a young dog … . I injected … cocaine into the space situated between the spinous processes of two of the inferior dorsal vertebrae. Five minutes after the injection there were evidences of marked inco-ordination in the posterior extremities … . A few minutes later there was marked evidence of weakness in the hind legs, but there were no signs whatever of feebleness in the anterior extremities. I now tested the condition of sensibility by means of a powerful faradaic battery, one of the conducting cords of which was attached to a fine wire brush. When the wire brush was applied to the hind-legs, there was no reflex action whatever … .

Experiment II.—This was performed on a man … . I injected … cocaine into the space situated between the spinous processes of the eleventh and twelfth dorsal vertebrae … [there was no effect and I repeated the injection] … . About ten minutes later the patient complained that his legs ‘felt sleepy’; and, on making a careful examination with the wire brush, I found that sensibility was greatly impaired … . The impairment of sensibility was principally limited to the lower extremities, the lumbar regions, the penis, and the scrotum … .” Corning, 1885 [32].

The details included in his publications suggest that Corning utilized perispinal administration of cocaine delivered by interspinous injection into the interspinous space, posterior to the spinal canal i.e. superficial (posterior) to the ligamentum flavum [3234, 36, 58] (see electronic supplementary material).

In later publications, Corning described a more invasive method of delivering cocaine, namely intrathecal injection, which required needle penetration through both the ligamentum flavum and the dura mater [35, 36]. Corning distinguished the clinical results obtained with these two methods, apparently favoring the perispinal method: “As a rule, the paraesthesia and anaesthesia are more irregularly distributed [with intrathecal injection at L 2-3] than when the posterior columns of the cord are anaesthetized in the manner first described [by perispinal injection at T10-11].” [36].

Intrathecal delivery of local anesthetics was rapidly adopted as a method of achieving spinal anesthesia, despite potential complications related to intrathecal needle delivery [5961]. By the early 20th century Corning’s perispinal method had been misconstrued, particularly in the US [62, 63].Footnote 5

4 Batson and the CSVS

Oscar Vivian Batson (1894–1979), Professor of Anatomy at the University of Pennsylvania, is best known for establishing the role of the vertebral venous plexuses (‘Batson’s plexus’) in the dissemination of cancer [1, 20]. Batson’s finding that the vertebral veins were valveless and communicated freely with intracranial veins provided an explanation for patterns of cerebral metastasis via the CSVS that had otherwise been unexplained [1, 20]. In fact, the extent and significance of Batson’s anatomical and physiological studies of the vertebral veins have significance well beyond Batson’s plexus as a route of cancer metastasis [9, 11, 17, 19, 6466]. Batson re-discovered the work of Breschet and produced further evidence of the anatomic continuity of the spinal and cerebral venous plexuses by injection and radiologic experiments [1, 20, 27]. Batson showed that radio-opaque dye injected in peripheral regions drained by the vertebral veins could produce visible delivery of dye into the cerebral veins of human cadavers via the vertebral venous plexus (see Figs. 5, 6, 7 in Batson [1]). Subsequent animal and human experiments by Batson and Anderson confirmed these findings of retrograde venous flow cephalad through the vertebral venous plexuses into the cerebral venous sinuses [1, 20, 27]. Batson’s experiments validated Breschet’s findings and established the unique, linear, bidirectional nature of blood flow within the CSVS [1, 17, 2022, 2628]. The linear, bidirectional blood flow between the valveless cerebral and spinal divisions of the CSVS contrasts with the circular nature of blood flow within the systemic circulation, as established by Harvey centuries earlier [1, 17, 19, 20, 26].

Batson’s finding that dye injected into a peripheral catchment area of Batson’s plexus could reach the cerebral veins is supported by clinical and pathological findings [19, 31, 52, 67, 68], including observations of bilateral blurred vision occurring within 1 min of injection of a 3 mL test dose of lidocaine meant for epidural analgesia [31]Footnote 6 :

“We believe her blurred vision was a result of direct test dose intracranial venous system dissemination via Batson’s vertebral venous plexus … . Batson’s vertebral venous plexus … communicates directly with … the intracranial venous system.” Vallejo et al. [31].

Although Batson made no mention of the use of the vertebral venous plexus for drug delivery, nor of Corning, carriage of cocaine through Batson’s plexus to the spinal cord, after perispinal interspinous injection, provides an anatomic explanation for the rapid spinal analgesic and anesthetic effects of cocaine reported by Corning after perispinal injection [3234, 69].

5 Perispinal Administration of Etanercept

Clinical experience with perispinal administration of etanercept began with its use for spinal disorders, first reported in 2001 [6, 12, 7072]. The therapeutic potential of etanercept for the treatment of spinal disorders is today supported by independent studies, including multiple randomized clinical studiesFootnote 7 [7376]. After CNS improvements were noted in multiple patients treated for intractable intervertebral disc-related pain using perispinal etanercept (PSE; 25 mg) [6, 12, 70], an institutional review board-approved clinical trial of PSE 25–50 mg weekly administered open-label in 15 subjects with mild-to-severe Alzheimer’s disease over a period of 6 months was performed [13].Footnote 8 Additional clinical experience suggests the therapeutic potential of PSE for additional forms of dementia [15, 77]. More recently, PSE has been successfully utilized in more than 1000 patients for treatment of chronic intractable neurological dysfunction after stroke or brain injury [10, 11, 16, 78]. The scientific rationale supporting the use of etanercept for stroke or brain injury includes multiple, independent studies and reviews [7996]. Rapid neurological improvement, beginning within minutes of perispinal injection, is characteristically seen following PSE injection, suggesting novel patterns of etanercept distribution to the CNS after perispinal administration [6, 7, 1016, 65, 70, 77, 78, 97].

For the treatment of brain disorders, Trendelenburg positioning for several minutes is used immediately after PSE is administered [10, 11, 1315]. This head-down tilt positioning is used to attempt to facilitate delivery of etanercept into the choroid plexus and cerebrospinal fluid (CSF) after its absorption into the CSVS [9, 11, 65, 98] as it has been demonstrated in basic science models that head-down tilt can increase intracerebral venous pressure and facilitate passage of plasma proteins into the CSF [18, 99, 100].

In vivo drug distribution after perispinal administration has been investigated by independent academic scientists in collaboration with this author [9, 18]. In 2007, enhanced delivery of radiolabeled diethylene triamine pentaacetic acid (DTPA) into the cerebral venous system after perispinal (compared with antecubital) injection followed by Trendelenburg positioning was observed in a human subject [9]. Following this human result, in collaboration with scientists at Stanford, the in vivo distribution of radiolabeled etanercept after perispinal administration and head-down tilt in a rat was investigated [18]. Positron emission tomography (PET) imaging suggested rapid penetration of radiolabeled etanercept into the CSF within the cerebral ventricles, with accentuation of signal within the choroid plexus within the ventricles [9, 18] (Fig. 9, transverse image). Coronal and sagittal PET images acquired at the same time as the transverse image (Fig. 10) confirm the pattern of delivery suggested by the transverse image.Footnote 9

Fig. 10
figure 10

Positron emission tomography image, transverse, coronal and sagittal sections of a living rat brain following perispinal extrathecal administration of 64Cu-DOTA-etanercept, imaged 5–10 min following the administration of etanercept. Reproduced in part from Tobinick et al. [18]

In 2014–2015, six basic science studies were published providing independent support for the therapeutic potential of etanercept in stroke models [79, 84, 86, 93, 94, 101]. These studies join the increasing evidence supporting the therapeutic potential of etanercept for multiple brain and spinal cord disorders [616, 18, 24, 6466, 7076, 7881, 8394, 97, 102116]. Emerging findings regarding movement of molecules through the spinal veins, CSF, brain interstitial fluid, lymphatics and transport of macromolecules through physiological CNS barriers suggest the existence of previously unappreciated anatomic pathways that may facilitate delivery of large molecules into the CSF and the brain after perispinal delivery [4, 25, 50, 52, 98, 117, 118123]. These new findings and accumulating evidence of the central involvement of neuroinflammatory mechanisms in brain disorders challenge existing dogma regarding brain physiology, support the rapid neurological effects seen following PSE administration, and underscore the importance of further research exploring CNS delivery of etanercept and other large molecules after perispinal injection [6, 7, 916, 24, 25, 50, 52, 70, 7785, 8797, 106, 114, 117, 119123].

6 Therapeutic Implications, Potential Advantages and Limitations of Perispinal Delivery

As Corning first demonstrated in 1885, the ability to selectively deliver a drug to the CNS by perispinal injection can have significant therapeutic implications [32, 124]. Corning’s first perispinal injection of cocaine produced spinal anesthesia without the risk of dural puncture [32].Footnote 10 A century later, emerging evidence suggests that perispinal delivery has the potential to enable treatment of intractable CNS disorders with drugs that, if administered systemically, would have difficulty reaching the CNS in a therapeutic concentration [6, 7, 916, 70, 77]; cf. epidural or intrathecal deliveryFootnote 11 [12, 7375, 125127].

Further research is needed, but current evidence suggests the potential advantages of perispinal delivery include the following:

  1. 1.

    Less invasive and less complex administration compared with neuraxial (epidural or intrathecal) delivery.

  2. 2.

    Elimination of two of the procedural risks of neuraxial delivery: post-dural puncture headache and needle injury to the spinal cord.

  3. 3.

    Suitability for use by the primary care physician in the primary care setting, without the requirement for specialized imaging, such as fluoroscopy, to facilitate needle placement.

  4. 4.

    Rapid and sustained brain effects, suggesting a different and more rapid pattern of brain delivery than that produced by spinal epidural or intrathecal delivery.

  5. 5.

    Delivery of CNS-active drugs, including macromolecules, to the CSF and CNS in therapeutically effective quantity, which drugs would otherwise have difficulty penetrating the BBB and/or the BCSFB in an optimal therapeutic amount.

The limitations of perispinal delivery, at present, include the following:

  1. 1.

    The need for basic science and clinical investigation to provide additional data regarding drug distribution, pharmacodynamics, pharmacokinetics, and safety and efficacy data for each drug candidate delivered by perispinal administration.

  2. 2.

    The need for funding and completion of randomized, double-blind, controlled clinical trials (RCT) to provide the additional data regarding safety and efficacy necessary for regulatory approval of drugs delivered by perispinal administration.

  3. 3.

    Limited knowledge of the functionality of the BBB and the BCSFB, as well as the distribution, pharmacokinetics, and pharmacodynamics of drugs after perispinal delivery and limited knowledge of the extent or function of drug efflux transporters in the CSVS vasculature.

  4. 4.

    Limited familiarity of the scientific and medical communities with cerebrospinal venous anatomy and physiology.

Further research addressing all of the limitations enumerated above is needed. At present, RCT data for perispinal delivery for the treatment of brain disorders is not yet available, although a randomized, double-blind clinical trial of PSE for the treatment of Alzheimer’s disease is currently underway at Griffith University in Australia (Australian New Zealand Clinical Trials Registry ID ACTRN12612000876897), with an RCT of PSE for post-stroke neurological dysfunction having received Institutional Review Board approval and scheduled to begin in 2016 (ACTRN12615001377527).

7 Conclusions

Perispinal injection is a novel emerging method of drug delivery to the CNS. Corning utilized knowledge of this anatomical pathway to successfully achieve spinal anesthesia by perispinal injection of cocaine in 1885. More than a century later, the therapeutic potential of perispinal injection for CNS disorders is highlighted by the rapid neurological improvement in patients with otherwise intractable neuroinflammatory disorders that may ensue following the administration of PSE. Double-blind, placebo-controlled clinical trials are necessary to fully characterize the efficacy of perispinal injection for drug delivery to the CNS and to obtain regulatory approval. More studies are needed in order to standardize methods of perispinal delivery. Perispinal delivery merits intense basic science and clinical investigation as a new method for enhancing delivery of macromolecules to the CNS and related structures.