Experimental Hydrocephalus

Published on 26/03/2015 by admin

Last modified 26/03/2015

Print this page

This article have been viewed 1407 times

CHAPTER 189 Experimental Hydrocephalus

In the past 20 years our understanding of the pathophysiology of hydrocephalus has advanced significantly, yet critical gaps remain. A recent position paper identified 10 major areas that require immediate attention.¹ Questions that are particularly amenable to experimental study include the following: How is cerebrospinal fluid (CSF) absorbed normally, and what are the causes of CSF malabsorption in hydrocephalus? Why do the ventricles dilate in communicating hydrocephalus? What happens to the structure and function of the brain when it is compressed and stretched by the expanding ventricles? What is the role of cerebrovenous pressure in hydrocephalus? What causes normal-pressure hydrocephalus? What causes low-pressure hydrocephalus? What is the pathophysiology of slit ventricle syndrome? What is the pathophysiologic basis for neurological impairment in hydrocephalus, and to what extent is it reversible? How is the brain of a child with hydrocephalus different from that of a young or elderly adult? It is difficult to address any of these important questions exclusively with clinical research, especially because of the wide variations in cause, onset, duration, and treatment complications that accompany hydrocephalus. Fortunately, animal models are available that approximate clinical conditions, and experimental hydrocephalus research is fundamental to promoting better treatments for this disorder. The recent white paper from a workshop sponsored by the National Institutes of Health bears witness to the importance of basic and clinical research in hydrocephalus.²

Unlike some neurological diseases or disorders, many mechanisms contribute to the pathophysiology of hydrocephalus. In fact, it is difficult to investigate these mechanisms individually because hydrocephalus is such a multifactorial disorder. Nevertheless, a clear distinction can be made between primary or causative mechanisms and secondary responses to ventriculomegaly. Primary mechanisms consist largely of developmental disorders that cause congenital hydrocephalus or pathologies such as intraventricular, subarachnoid, and intraparenchymal hemorrhage, meningitis and other infections, and tumors. Secondary mechanisms are far-reaching and include axonal damage, demyelination, cell death, gliosis and inflammation, biomechanical compression and stretch, edema, metabolic impairment, cerebrovascular effects and hypoxia-ischemia, synaptic and dendritic deterioration, neurotrophic changes, alterations of neurotransmitters and neuromodulators, and impaired clearance of toxins and metabolites. Often these secondary mechanisms overlap, making it difficult if not impossible to define the precise role of each. In spite of these obstacles, an emerging body of evidence is beginning to demonstrate that hydrocephalus (1) manifests, at least initially, as a disorder of periventricular white matter; (2) has a major impact on the structure and function of the cerebral vasculature; (3) involves metabolic and molecular changes that may not produce clinical symptoms but can have protracted effects on neurological function; (4) injures neurons in many ways but does not cause significant neuronal cell death unless ventriculomegaly is severe; (5) often follows a slowly progressive pathophysiologic pattern, which may allow considerable plasticity; and (6) is more effectively (but not completely) treated when intervention begins relatively soon after onset. A review by Del Bigio ³ and several other summaries⁴^–⁹ provide excellent introductions to the pathophysiology of hydrocephalus, and the reader is encouraged to consult these publications for details.

This chapter summarizes the basic science approaches to experimental hydrocephalus, focusing on new findings produced by such research and identifying promising new areas of research.

Experimental Models

Practically all the key questions listed earlier are best investigated, at least initially, in animals, and many in vivo models of hydrocephalus have been developed over the past 50 years.^4,^10,¹¹ In addition, there is renewed interest in the use of mathematical models that can be tested in animals. It is appropriate to review these models here because they all have advantages and disadvantages as well as specific relevance to different clinical applications.

Animal Models of Congenital and Transgenic Hydrocephalus

The rat congenital model most often used in experimental research is the H-Tx strain, which develops obstructive hydrocephalus from aqueductal stenosis in the perinatal period.¹¹^–¹⁴ From four chromosomes within a heterozygous background^13,^15,¹⁶ and incomplete penetrance,¹⁷ these animals develop hydrocephalus within several days of birth, which in rats corresponds to the third trimester of human brain maturation. In this model, ventriculomegaly becomes severe by the second postnatal week (Fig. 189-1), and the animals usually expire by 20 to 25 days of age if intracranial pressure (which does not rise until postnatal day 12) is not reduced by shunting. Drainage of CSF can be accomplished in H-Tx rats with either ventriculoperitoneal or ventriculosubcutaneous shunts, which are more effective when placed early (3 to 5 days of age) rather than late (12 to 14 days of age). Inbred strains of Wistar-Lewis (LEW/Jms) rats also develop aqueductal stenosis through nonmendelian mechanisms as early as day 17 in a 21-day gestational period.¹⁸^–²¹ The frequency of hydrocephalus and the ratio of affected males to females are significantly higher when the LEW/Jms parent is male. These rat models are excellent for studies of neonatal and juvenile hydrocephalus, especially that caused by aqueductal stenosis, because ventriculomegaly occurs naturally, the brain is large enough for customized shunting, the rats are amenable to behavioral testing, the cost is relatively low, and a wealth of data is available for correlation. Nevertheless, they are not ideal for long-term experiments unless shunting is performed, and their size restricts the use of clinical shunt systems and pressure probes.

FIGURE 189-1 Examples of the H-Tx neonatal rat model of congenital hydrocephalus (A and B) and the adult rat model of basal cistern–obstructed communicating hydrocephalus (C-M). A, T2-weighted magnetic resonance imaging (MRI) scans illustrating the progression of hydrocephalus at 4, 11, and 21 days of age. B, Skull changes correlating to the same ages in A; note the conspicuous doming of the head beginning at 11 days, when intracranial pressure begins to rise. C-F, T2-weighted images at four coronal locations in the brain of a hydrocephalic animal 4 weeks after kaolin injection. G-J, Corresponding images in a saline-injected control animal. Dilation of all portions of the ventricular system are characteristic of all severely hydrocephalic animals. K, Cresyl violet–stained coronal section illustrating the location of kaolin deposits in the basal cistern. L, T2-weighted MRI at the same level as K showing patent foramina of Luschka (FL). M, Midsagittal T1-weighted MRI after gadolinium injection into the frontal horn of the lateral ventricle. Note tracer throughout the subarachnoid space, except at the location of the kaolin deposits in the basal cistern (asterisks). CA, cerebral aqueduct; FM, foramen of Monro.

(A and B, Modified from Jones HC, Harris NG, Rocca JR, Andersohn RW. Progressive tissue injury in infantile hydrocephalus and prevention/reversal with shunt treatment. Neurol Res. 2000;22: 89.)

Several interesting mouse models of hydrocephalus have provided valuable insights into the causes of ventriculomegaly.^11,²²^–³³ The most widely used models include the SUMS/NP,²⁸^,³² hy3,^31,³⁴^–⁴⁰ transforming growth factor (TGF)-β1 overexpression,^26,^30,⁴¹^–⁴⁶ hyh with point mutation in α-SNAP and ependymal denudation that precedes aqueductal stenosis,⁴⁷^–⁵¹ fibroblast growth factor (FGF)-2,⁵² L1-cell adhesion molecule deficient,^27,⁵³^–⁵⁵ aquaporin deficient,⁵⁶^–⁵⁹ hpy,⁶⁰ members of the conserved forkhead–winged helix transcription factor gene (previously Mf1),^29,⁶¹^–⁶³ heparin-binding epidermal growth factor,⁶⁴ and collagen deficienct.⁶⁵ Most recently, Sweger and colleagues⁶⁶ developed a double transgenic mouse model that allows expression of the G1-coupled Ro1 receptor exclusive to astrocytes. By controlling Ro1 expression with a tetracycline-on promoter in drinking water, these mice develop enlarged ventricles, partial ependymal denudation, morphologic changes in the subcommissural organ, and obliteration of the cerebral aqueduct at designated times. This represents a powerful experimental model because the pathogenesis of hydrocephalus can be studied with neonatal, juvenile, and adult onset. The main disadvantage of these mouse models is that their size limits the use of CSF shunts and invasive physiologic sensors.

Animal Models of Acquired Hydrocephalus

Animal models of acquired hydrocephalus offer the advantages of precise timing, greater efficiency, and sometimes reduced costs. Kaolin, an inert silicate, is a well-accepted agent for inducing hydrocephalus in a wide variety of infant and adult animals (mouse, rat, rabbit, hamster, cat, dog) via injections into the CSF, with no evidence of direct pathologic effects on structures distant to the injection site. Although this induction method is mechanical (surgical) and produces hydrocephalus abruptly, which is not always the case clinically, it is useful when more “natural” models are not available.

Usually obstructive (noncommunicating or intraventricular) hydrocephalus is induced by kaolin injections into the cisterna magna or fourth ventricle, or both. In adult animals (cat, dog, rat, guinea pig, nonhuman primate), such intracisternal obstructions consistently produce a moderate degree of ventriculomegaly, presumably because the fused cranial sutures create a nonexpandable skull. In addition, the more mature brain, with well-developed myelin and less intercellular space, probably influences the extent of ventriculomegaly. In contrast, intracisternal kaolin injections in neonatal and juvenile animals consistently produce severe, rapidly progressing types of hydrocephalus. Skull enlargement always accompanies ventriculomegaly in these animals, and the severity of hydrocephalus is represented by Evans ratios of 0.5 to 0.9 (compared with 0.3 in nonhydrocephalic controls). The neonatal rodent models usually do not survive beyond 3 weeks of age unless treated by CSF drainage; those animals that do survive after this time typically exhibit mild forms of ventriculomegaly. A transitional model of obstructive hydrocephalus has also been used extensively by injecting kaolin into the cisterna magna of 3-week-old juvenile rats. This model remains “pediatric” because brain maturation is only about 50% complete (in rodents, neurogenesis and migration have concluded, myelination has peaked but not finished, and synaptogenesis continues for another 3 weeks), but the skull does not expand.

Communicating extraventricular hydrocephalus has been more difficult to achieve with acquired approaches. Attempts to produce communicating hydrocephalus with kaolin injections into the cortical subarachnoid space of adult rats ⁶⁷^,⁶⁸ and dogs⁶⁹^–⁷⁴ or with silicone injections into the subarachnoid space⁷⁵^,⁷⁶ have consistently produced only moderate ventriculomegaly and usually require a long time (several months) to develop. Likewise, Silastic has been infused into the basal cisterns of monkeys to produce mild forms of communicating hydrocephalus.⁷⁷ Recently, Li and colleagues⁷⁸ developed a novel model of communicating hydrocephalus by injecting kaolin into the basal cisterns of adult rats (see Fig. 189-1); these animals develop moderate ventriculomegaly within a week of induction and exhibit impairments in CSF absorption and pulsatility (see later).

Nonmechanical induction methods, such as viral ⁷⁹ and bacterial inoculations,⁸⁰ have also been used to produce communicating hydrocephalus, but these procedures involve the additional influence of the induction substances on pathophysiology and thus are not true representations of the effects of hydrocephalus alone on the brain. Growth factors such as TGF-β and FGF-1 and -2,^30,^41,⁵² as well as neurotoxins,³^,⁸¹ have all been successful to varying degrees.

An important advance in experimental studies has been the application of CSF shunting techniques.⁸²^–⁸⁹ Initially, larger models such as feline,⁸³^,⁸⁸ lagomorph,⁹⁰ and canine⁹¹^–⁹⁷ were employed, and these still have the advantage of accommodating commercially available shunt systems. With improved neurosurgical techniques, smaller neonatal, infantile, and juvenile rats have been shunted successfully,⁹⁸^–¹⁰⁴ but the catheters are usually custom made, and valves are seldom used. It is also important to realize that shunted animals, especially the youngest and smallest ones, develop shunt malfunctions at approximately the same rate as human patients. Nevertheless, it is surprising that no studies have attempted to evaluate the effects of repetitive shunt malfunction.

Mathematical Models of Hydrocephalus

Several mathematical models have revealed possible mechanisms associated with hydrocephalus.¹⁰⁵^–¹²² One of the prevailing hypotheses is that the mechanical properties of hydrocephalic brain tissue are fundamentally changed compared with healthy brain tissue. However, this change in material properties has not yet been quantified, so these models and others like them are not useful as predictive models. One particular mathematical model that has been reliable in matching patient outcomes describes the brain as a quasi-linear viscoelastic tissue,¹²³ which essentially means that there is an initial elastic response to ventriculomegaly, followed by a different long-term viscoelastic response. The flaw that prevents this model and all others from being predictive is that the material properties of the brain are not known precisely enough to be useful. CSF infusion tests are available to measure compliance (stiffness) of the brain by monitoring intracranial pressure changes during the injection of artificial CSF into the ventricles or the lumbar subarachnoid space.¹²⁴ These tests have shown that compliance decreases in the hydrocephalic brain.¹²⁴^–¹³³ However, compliance is an extrinsic property of the entire contents of the cranial cavity and can be influenced by the amount of tissue, the amount and flow velocity of blood, the tension of the meninges, and the possible expansion of the skull. In contrast, the material properties needed to construct a predictive mathematical model, such as shear or elastic modulus, are intrinsic properties of brain tissue independent of the amount of tissue. Current tissue testing methods to acquire intrinsic properties typically involve the removal of brain tissue for mechanical tests, but this approach eliminates many structural and physiologic factors of the “living” brain, especially the mechanical support offered by blood volume and flow, that tend to hydraulically stiffen the tissue. Fortunately, a mechanical testing method has been developed that allows direct measurement of the initial elastic response as well as the long-term viscoelastic response of “living” brain tissue.¹³⁴^,¹³⁵ This method uses a mechanical probe that displaces the surface of the cortical tissue inward slightly and holds the displacement for a period of time, which is an ideal way to investigate tissue properties without damaging the brain.

Magnetic resonance elastography ¹³⁶^,¹³⁷ measures the elastic properties of brain tissue in vivo, based on the propagation of waves through deformed tissue detected with magnetic resonance imaging (MRI). Tissues with different material properties conduct these waves differently. Although this method holds great potential for noninvasive tissue property measurement, the true material properties of the brain (healthy or hydrocephalic) have yet to be measured directly, so the accuracy of this method has not been verified. In addition, the device used to perform the mechanical perturbation of the tissue requires a flat surface,¹³⁷ so more developed gyral hemispheres are problematic, and the device is too large for rodent brains.

Nevertheless, once they are measured accurately, the intrinsic material properties of the hydrocephalic brain would be useful in the treatment of this disorder. Clinical decision making could benefit from knowing whether the relative stiffness of the brain increases with repeated shunt malfunction or at certain ages.

Pathophysiologic Mechanisms and Treatment Possibilities

Several previous reviews have provided excellent summaries of the pathophysiology of hydrocephalus.^2,^4,^5,¹³⁸^–¹⁴⁰ In general, mechanisms of injury include morphologic and functional changes in the cerebral vasculature, hypoxia-ischemia, gliosis and neuroinflammation, edema, axonal degeneration and impaired intracellular transport, demyelination and oligodendrocyte death, dendritic and synaptic degeneration and plasticity, altered levels of neurotransmitters and neuromodulators, and impaired protein clearance¹⁴¹ and lymphatic absorption. Exciting new data have revealed several novel mechanisms, dramatically broadening the view of the pathogenesis and secondary pathophysiology of hydrocephalus. These new findings are presented here rather than providing an exhaustive review of experimental studies.

Gliosis and Inflammation

Gliosis is a consistent finding in hydrocephalus, and inflammation and glial scar formation could play a major role in creating the chronic problems that plague hydrocephalic patients, but the time course and permanence of the reaction are not completely known.^28,^138,¹⁴²^–¹⁵³ It has been suggested by many investigators that scar formation is a permanent fixture in hydrocephalic brains,^89,^144,¹⁵² even those that have been shunted successfully. Previous studies in both congenital and acquired models of hydrocephalus have shown that glial fibrillary acidic protein (GFAP) RNA and protein levels increase with the progression of hydrocephalus. Additionally, Mangano and colleagues¹⁵² illustrated that microglial cell proliferation and activation increased in regions distant from the cortical “lesion,” suggesting that neuroinflammation is related to damage throughout the cortical pathways. Furthermore, GFAP-labeled reactive astrocytes surround cystic lesions in severely hydrocephalic H-Tx animals but are not present in the white matter surrounding the ventricles.¹⁵⁴^,¹⁵⁵ Experimental models of hydrocephalus have demonstrated that shunting can reduce the amount of GFAP protein and RNA present in the cerebral cortex, but these levels begin to rise over time,⁸⁹ suggesting that reactive astrocytosis is highly sensitive to suboptimal CSF drainage. Clinically, increased levels of GFAP protein have been found in the CSF of patients with normal-pressure hydrocephalus and those who developed secondary hydrocephalus due to subarachnoid hemorrhage.¹⁵⁶^–¹⁶⁰ The possibility of using GFAP protein levels as a diagnostic tool for hydrocephalus is currently being explored.¹⁵⁸^,¹⁶¹

It is likely that gliosis may dramatically change the mechanical properties of the brain so that it becomes more rigid (less compliant) and resistant to increases in CSF pressure and flow.^126,^127,^130,^131,¹⁶²^–¹⁷¹ The importance of these properties in hydrocephalus is illustrated by the finding that reduced compliance, measured using the pressure-volume index, is a better predictor of shunt success than measurements of ventricular size.^162,^172,¹⁷³ Unfortunately, no studies have directly examined the relationship between glial alterations and intracranial compliance. Compliance can be measured in animals using CSF infusion tests,¹⁷⁴^–¹⁷⁶ and in hydrocephalus, resistance to CSF outflow is increased over a wide range of intracranial pressures.¹⁷⁶ Compliance is usually lower in hydrocephalic animals, but it varies depending on intracranial pressure.

Most recently, an interesting and potentially powerful relationship has been suggested between astrocytes and aquaporin channels, which can have major impact on CSF absorption. Aquaporins are cell membrane proteins, and most are permeable only to water.¹⁷⁷^–¹⁸⁰ Aquaporin 4 (AQP4) is found primarily on the endfeet of astrocytes that contact microvessels in the periventricular white matter and the subpial region of the cerebral cortex,¹⁸¹ as well as in ependymal cells lining the lateral ventricles. These locations place astrocyte aquaporin channels in a good position to transport CSF from the brain to the vascular system, and several investigators have suggested that they may play a major role in CSF absorption.¹⁸² Aquaporin channels require connections to a basal lamina that specifically contains collagen XVIII molecules; if these connections are not present, hydrocephalus ensues.^65,^183,¹⁸⁴ One structural response of endothelial cells to chronic hypoperfusion is an abnormal basal lamina,¹⁸⁵ so it is not surprising that aquaporin changes have been reported in experimental hydrocephalus.^56,¹⁸⁶^–¹⁸⁸ More directly, expression of AQP4 (but not AQP1 or AQP9) increases in the cerebral cortex and hippocampus of rats with mechanically induced hydrocephalus, suggesting that this water channel plays a compensatory role in CSF absorption during ventriculomegaly.¹⁸⁶ Fibrosis in the form of reactive astrocytosis could therefore have an important impact on CSF absorption, which might not recover with ventricular shunting alone.

Minocycline has recently shown promise as a specific inhibitor of microglial cells,¹⁸⁹^–¹⁹² one of the main elements of glial scar formation in hydrocephalus.^89,^152,^193,¹⁹⁴ Although the mechanism of action is still unknown, minocycline has multiple benefits in brain injury,^191,¹⁹⁵^–²²⁷ and its promise as a neuroprotective agent is illustrated by the recent initiation of clinical trials in Parkinson’s disease,²²⁸^,²²⁹ Huntington’s disease,²³⁰^–²³³ amyotrophic lateral sclerosis,²³⁴ multiple sclerosis,²³⁵^,²³⁶ and schizophrenia.²³⁷^,²³⁸ Based on previous studies showing positive effects in white matter,^195,^203,^215,²³⁹^–²⁴² Miller and McAllister²⁴³ initiated tests of minocycline’s ability to inhibit gliosis in the H-Tx rat model of congenital hydrocephalus. Preliminary results are encouraging, in that both numbers of glia (astrocytes and microglia) and cortical mantle thickness were significantly reduced when minocycline was administered after hydrocephalus had progressed considerably (Fig. 189-2).

FIGURE 189-2 Stereologic analyses depicting changes in astrocytes in H-Tx rats with congenital hydrocephalus. The graph on the left shows the effects of progressive hydrocephalus on glial fibrillary acidic protein (GFAP) levels in the occipital cortex of control, untreated hydrocephalic, and shunted hydrocephalic H-Tx rats (n = 5 for each group). In untreated animals, increases began at 5 days (2.46-fold; P = .08), progressed at 12 days (1.81-fold; P = .17), and became significant by 21 days (5.26-fold; P = .0008). Ventricular shunting at 15 days of age significantly reduced GFAP levels within 6 days (5.23-fold; P = .0003). However, shunted animals with longer recovery periods exhibited higher GFAP levels than normal, suggesting that residual astrogliosis persists. P < .05 relative to control (*) and untreated hydrocephalic (**) groups. The graph on the right shows the effects of minocycline administered to H-Tx rats at 15 to 21 days of age with severe ventriculomegaly. Note that minocycline significantly reduced the density of both astrocytes and microglia.

It should be noted that to date only a few other studies have attempted to protect the hydrocephalic brain by infusing nimodipine, a calcium channel antagonist, into the ventricles of juvenile hydrocephalic rats to reduce white matter damage, but these interventions produced only limited short-term success. Nevertheless, in the near future pharmacologic successes achieved in animal models of hydrocephalus have a strong likelihood of leading to supplemental clinical treatments for humans.

Intracranial Pulsatility

Thirty years ago, Di Rocco and coworkers ²⁴⁴ showed that ventricular dilation can be caused by abnormal intracranial pulsatility in the absence of a physical obstruction to CSF flow. Abnormal intracranial pulsations are also clearly related to clinical hydrocephalus; intraventricular pulsatility, detected as CSF pulsations in the aqueduct on cine MRI, are often dramatically elevated in hydrocephalus²⁴⁵^–²⁴⁸ and can be a useful diagnostic criterion in normal-pressure hydrocephalus. Although authors have posited theories on the cause of these elevated pulsations,²⁴⁹^–²⁵¹ no study has provided a clear link between abnormal pulsations and the underlying pathophysiology of hydrocephalus.

Furthermore, the dissipation of cerebral arterial pulsatility, resulting in minimal (homeostatic) capillary and venous pulse pressure, is believed to be critical for normal cerebrovascular function.²⁵¹^–²⁵⁴ Although dissipation is easily accomplished in other organs, the closed cranium requires a more complex system for this function; CSF and venous blood serve this purpose. The major arteries entering the cranium are located within the CSF-filled subarachnoid spaces, allowing the efficient coupling and transfer of pulsation. Pulsations can be transferred out of the cranium either directly via CSF flow through the foramen magnum into the compliant spinal subarachnoid space or indirectly via coupling to the sagittal sinus within the convexity. An important component of the pathophysiology of hydrocephalus is a change in intracranial compliance, which may lead to a redistribution of the pulsation dissipation mechanism.²⁵⁰^,²⁵¹ One potential pathway is directly into parenchymal capillaries. Thus, hydrocephalus may be caused by a breakdown in this pulse dissipation mechanism (or a failed windkessel),^251,^253,²⁵⁵^–²⁵⁸ which normally protects the capillary bed from excessive pulsatile shear forces.

Increased capillary pulsatility may have several pathophysiologic consequences. Structural responses may lead to the loss of parenchymal microvessels, and in fact, decreased capillary density has been shown in experimental hydrocephalus.^4,²⁵⁹^–²⁶⁴ In addition, the integrity of the blood-brain barrier is compromised in hydrocephalus.²⁶⁵^–²⁶⁷ Both these effects might explain the marked loss of cerebral perfusion that has been well documented in clinical²⁶⁸^–²⁸² and experimental hydrocephalus.^34,^67,^276,^278,²⁸³^–²⁸⁹ Importantly, it was recently shown that excessive pulsatile stress can impair hemodynamics through changes in endothelial cell homeostasis²⁹⁰^–²⁹⁵ mediated by nitric oxide.²⁹⁶ A marked increase in nitric oxide synthase immunoreactivity has also been reported in a kaolin hydrocephalus rat model during the first few weeks of ventriculomegaly.²⁸⁶ These examples all provide a compelling case for the importance of pulsation dissipation in maintaining normal capillary function.

Zou and colleagues ²⁹⁷ and Wagshul and associates,²⁹⁸ using two different types of analyses, recently showed that adult dogs normally exhibit a pulse dissipation mechanism termed a notch (because when intracranial pressure is graphed against frequency, it appears as a trough) and that this notch changes as pressure is raised. Further studies are needed to determine whether this change in intracranial hydrodynamics is causative or a secondary response to ventriculomegaly. One series of investigations that addressed this issue used the novel model of communicating hydrocephalus in adult rats described earlier.⁷⁸ Using high-field-flow MRI in this model, Wagshul and coworkers²⁹⁹ demonstrated that aqueductal pulsatility (stroke volume) increased within 24 hours of basal cistern obstruction, and the severity of ventriculomegaly was associated with the maintenance of high stroke volume. In contrast, only mild ventriculomegaly developed in animals that exhibited an early but transient increase in aqueductal pulsatility. These experiments clearly indicate that pulsatility plays a role in the pathophysiology of experimental hydrocephalus (Fig. 189-3).

FIGURE 189-3 Aqueductal pulsatility (stroke volume [SV]) in adult rats with communicating hydrocephalus induced by kaolin obstruction of the basal cisterns. Based on the size of the pulsations and their temporal course, the animals clustered into three distinct groups: group 1 (n = 5) exhibited marked increases in aqueductal pulsation, rising mostly within the first 1 to 2 weeks and persisting throughout the entire study; group 2 (n = 3) exhibited modest initial increases in pulsations that persisted for the first week and then returned to near normal levels; and group 3 (n = 7) exhibited no changes in aqueductal pulsations compared with baseline values or intact controls. vv, ventricular volume.

Cerebrospinal Fluid Absorption—Lymphatic, Arachnoid, Microvascular

The traditional view of CSF absorption exclusively via arachnoid granulations into the superior sagittal sinus or the veins adjacent to the spinal roots has been challenged by a series of experiments.³⁰⁰^–³¹⁰ Initially, the results of these experiments were challenged because they were performed on animals or human cadavers and consisted of latex Microfil infusions to reveal potential CSF pathways from the subarachnoid space. However, the latest in vivo studies clearly demonstrate that cranial lymphatics, most notably the nasal pathways surrounding the olfactory nerves, are capable of transporting large volumes of CSF. Most important, these pathways are impaired in adult rats with communicating hydrocephalus (Fig. 189-4).