The Monro–Kellie hypothesis

The syndrome known as aliquorrhea or spontaneous intracranial hypotension is almost invariably a consequence of spontaneous CSF leaks.1 Head and spine MRI have revolutionized the diagnosis of CSF leaks leading to the identification of more patients than ever before. An increasing interest in the subject of CSF volume depletion attributable to CSF shunt overdrainage and particularly to spontaneous CSF leaks has been rekindled. Many of the MRI abnormalities seen in CSF volume depletion have been explained by the Monro–Kellie hypothesis.

In this article, the initial concepts championed by Monro and Kellie are described and are followed by contributions of others to what is now referred to as the Monro–Kellie hypothesis or doctrine.

Alexander Monro (1733–1817), a Scottish anatomist from a well-known medical family, like his father and his son, was a professor of anatomy at the medical school of Edinburgh.2 He is the most famous of the three and is credited not only for his contribution to the hypothesis in discussion, but also for his description of the interventricular foramina bearing his name.

In a monograph published in 1783 on certain aspects of structure and function of the nervous system,3 applying some of the principles of physics to the intracranial contents, Monro made the following points that 1) the brain was enclosed in a nonexpandable case of bone; 2) the substance of the brain was nearly incompressible; 3) the volume of the blood in the cranial cavity was therefore constant or nearly constant; and 4) a continuous outflow of venous blood from the cranial cavity was required to make room for the continuous incoming arterial blood. Thus appeared in the literature for the first time the hypothesis that the blood circulating in the cranium was of constant volume at all times.

George Kellie of Leith, another Scot and himself a former pupil of Monro, studied the amount of venous blood in the brain of humans and animals that had died of various causes, including drowning, hanging, and exsanguination.4-6⇓⇓ He noted that there was no significant difference in the amount of the brain venous blood in these various circumstances, a finding that supported Monro’s observations. In his experiments on animals that had been bled to death, Kellie noted that the brain had maintained its usual appearance and was not pale or drained of blood. He also noted that the amount of blood in the cerebral vessels of dogs given lethal doses of prussic acid was not affected by gravitation, regardless of whether the animal was hanging by its ears or hind legs. Kellie was fascinated that congestion of cerebral vessels is not noted in some instances in which it might be most suspected. He remarked on postmortem observations on the bodies of two pirates who were hanged at Leith and were dissected by Monro and himself. Although extracranial tissues were congested and engorged, the intracranial vessels were not. Kellie similarly noted that 1) the brain was contained in, and exactly filled, an unyielding case of bone; 2) the brain itself was only minutely compressible; and 3) it was unlikely that any circulating fluid could be withdrawn from or any overflow could be introduced within the cranial cavity without simultaneous equivalent replacement or displacement.

Many of Kellie’s findings were also noted by John Abercrombie, a pathologist at the medical school in Edinburgh, and hence the occasional citations in the literature of the “Monro–Abercrombie Doctrine.”7 Abercrombie’s opinions were mostly based on experiments on animals that had been bled to death. He noted that although other organs had been exsanguinated, the brain had maintained its usual appearance. Repetition of the same experiments on animals with a small trephine opening of the skull, however, resulted in exsanguination of the brain similar to that of other organs. Abercrombie, a highly respected authority in his time, in a classical monograph in 1928,8 supported the findings of Monro and Kellie and was instrumental in attracting wide acceptance to the hypothesis.6

Monro, Kellie, and Abercrombie essentially ignored the existence of CSF and assumed that the volume of intracranial blood seesawed between the arterial and venous side while its overall volume remained essentially constant. This is not entirely surprising. The second-century concept of Galen that “spiritus animalis,” or vital spirit, filled the cerebral ventricles had survived for many centuries. Even when Vasalius, an anatomist of Padua in the 16th century, described fluid in the ventricles, he was not believed. Indeed, as late as the early 19th century, Romberg, the well-known German neurologist, thought that the brain ventricles were filled with “humid gas.” Francois Magendie, a French physiologist of the 19th century, deserves the credit for convincing his contemporaries of the presence of CSF. He tapped the cisterna magna in animals, started CSF analysis, and showed communication between the subarachnoid space and the fourth ventricle through the foramen that now bears his name.9

Recognition of CSF as a vital intracranial component added another dimension to the equation of intracranial contents and had a profound effect on the understanding of intracranial pressure and intracranial fluid content. Soon came a due revision of the Monro–Kellie hypothesis.

George Burrows, an English physician, in 1846 questioned for the first time the concept of fixed intracranial blood volume, which was then the basis of the Monro–Kellie hypothesis.6 Using rabbits, Burrows repeated many of Kellie’s experiments on gravitation, strangulation, and exsanguination. He was also cognizant of the work of Magendie and the experiments by Ecker of Stuttgart on CSF and its dynamics. Burrows declared a reciprocal relationship between intracranial blood volume and CSF volume and noted that an increase in one required a decrease in the other and vice versa. Therefore, the CSF volume was introduced into the equation of the Monro–Kellie hypothesis. Research on the CSF volume–pressure relationships flourished in the latter part of the 19th century.

Cushing offered a precise formula for the Monro–Kellie hypothesis and indicated that with an intact skull, the sum of the volume of the brain plus the CSF volume plus the intracranial blood volume is constant. Therefore an increase in one should cause a reduction in one or both of the remaining two.10

Initially, the emphasis was understandably directed mainly toward increased intracranial pressure with its enormous clinical implications. The less common and typically far less serious issue of decreased intracranial pressure, therefore, did not receive an early emphasis.

In the past decade, the discovery of MRI abnormalities in CSF leaks has resulted in the identification of more patients, and it has also called for pathophysiologic explanations for these abnormalities. Fishman pioneered the reintroduction of the concept of the Monro–Kellie hypothesis in the context of a decrease in intracranial pressure and reduced CSF volume.11

Considering that the cranium remains a rigid structure after the closure of the fontanels, in the presence of an intact skull, a decrease in CSF volume (whether caused by CSF leak or CSF shunt overdrainage) would require volume compensation. Because the brain volume remains nearly constant, it is the intracranial blood volume that will be affected, and consequently, a compensatory intracranial hyperemia occurs. This hyperemia is reflected primarily on the venous system leading to diffuse meningeal venous hyperemia, engorgement of venous sinuses, and hyperemia of the pituitary gland. With meningeal venous hyperemia, as addressed by Fishman, because leptomeninges have blood–brain barriers and pachymeninges do not, it is the pachymeninges that enhance on MRI, likely as the result of pressure gradient–driven extravasation of contrast material that appears as diffuse pachymeningeal gadolinium enhancement. Venous engorgement is also responsible for enlargement of the pituitary gland. In CSF volume depletions, subdural fluid collection also participates in intracranial volume compensation.

Changes also take place at the spinal level. The vertebral canal is encased by bony and fibrous components that are tough and only slightly elastic. While in the cranial cavity, the dura is closely applied to the inner surface of cranial bones; in the spinal canal, there is an epidural space between the dura and the fibroosteal canal, which is filled with fatty areolar tissue and epidural venous plexus. Therefore, a mild degree of dural collapse in response to CSF volume depletion can occur and is compensated by engorgement of the epidural venous plexus. Other MRI changes related to volume compensation at the spinal level include pachymeningeal gadolinium enhancement and subdural fluid collections, similar to but less noticeable than the intracranial changes.

Thus evolved into its current form the hypothesis that existed in the core of the doctrine championed by Alexander Monro more than 2 centuries ago. Its modern applications have been essential in explaining many of the MRI abnormalities seen in CSF volume depletions whether caused by CSF leak or CSF shunt overdrainage.