Abstract
Cerebrospinal fluid (CSF) leak is recognized to cause headaches that typically but not invariably have orthostatic features (present in upright posture, relieved by recumbency). Head magnetic resonance imaging (MRI) typically shows diffuse pachymeningeal enhancement. A 24-year-old woman, after resection of a right temporoparietal glioma, developed CSF leak from the craniotomy site, resulting in subgaleal fluid collection and associated with diffuse pachymeningeal gadolinium enhancement as well as posture-related headaches. The headaches, however, were present in a recumbent position and relieved after several minutes of being in an upright position. It is postulated that CSF leak took place when the patient was recumbent and ceased when she was upright, a position in which there is decrease in intracranial pressure. After cessation of the leak, along with disappearance of subgaleal collection of CSF, both the headaches and the pachymeningeal enhancement resolved.
Introduction
Cerebrospinal fluid (CSF) volume depletion is known to cause headaches. These may or may not be associated with other clinical manifestations, including pain and stiff feeling of the neck, nausea, emesis, horizontal diplopia (usually due to unilateral or bilateral sixth cranial nerve palsy), dizziness, change in hearing, visual blurring, photophobia, interscapular pain, radicular upper limb symptoms, and facial numbness or weakness (1). The typical headaches are orthostatic (present in an upright position, relieved by being in a recumbent positon). However, there is considerable variability. Sometimes, particularly when the headaches become chronic, the orthostatic features may dampen, and the headache may evolve into lingering chronic daily headaches. Less frequently the opposite may be noted, in that there is initially a lingering headache for some days or weeks before the typical orthostatic features appear (2, 3). Sometimes neck or interscapular pain may precede the headaches by days or weeks. Uncommonly, the headache may have a thunderclap onset (4).
Although headaches of CSF volume depletion frequently can be aggravated by Valsalva-type manoeuvres, this author has seen a few patients with documented spontaneous CSF leaks whose headaches were actually triggered by Valsalva-type manoeuvres (cough, sneeze, straining, etc.), with each headache lasting for variable periods ranging from a few minutes to a few hours. In intermittent CSF leaks, the headaches, with whatever features that they might have, may appear and disappear for variable periods. Head MR imaging abnormalities in CSF volume depletion include diffuse pachymeningeal gadolinium enhancement, imaging evidence of sinking of the brain (descent of cerebellar tonsils, decrease in size of prepontine and perichiasmatic cisterns, flattening of the optic chiasm, and crowding of the posterior fossa), subdural fluid collections, decrease in size of the ventricles (‘ventricular collapse’), enlargement of the pituitary gland, and engorgement of the cerebral venous sinuses (3).
Over-draining CSF shunts are well-recognized causes of CSF volume depletion. Spontaneous CSF leaks have been increasingly recognized in the past decade and are probably the sole cause of spontaneous intracranial hypotension. The overwhelming majority of these leaks occur at the level of the spine and only rarely at the skull base. Post-surgical CSF leaks can occur after spine or cranial base surgeries.
In this communication, a patient with CSF leak from a craniotomy site is reported. The patient had a subgaleal hygroma and diffuse pachymeningeal gadolinium enhancement as well as posture-related headaches. Both the hygroma and pachymeningeal enhancement resolved after cessation of the leak along with disappearance of the headaches.
Report of the case
A 24-year-old right-handed woman had a history of migraine headaches since her teens. These occurred about once or twice per month. In spring 1998 the headaches increased in frequency to one headache each day or every other day. These were associated with nausea, emesis and photophobia. Many of the headaches began during sleeping hours, awakening the patient. She had obtained good relief from sumatriptan injections. In June 1998 she developed sensory seizures involving her left side. Head MRI showed a 5 × 5 cm right temporoparietal mass that on biopsy was found to be a grade II oligodendroglioma. Radiation therapy was subsequently started (5400 cGy in 30 fractions) and was completed by mid-August 1998. The migraine headaches persisted. She continued to use oral analgesics and sometimes also administered sumatriptan injections. Phenytoin was used to control the sensory seizures.
By early 1999 the seizures were more difficult to control. Sequential MRIs showed appearance of areas of enhancement within the tumour and an overall mild increase in tumour size (Fig. 1, A and B, left). In September 1999 gross total resection of the tumour was carried out and it was found to have dedifferentiated into a grade III mixed oligoastrocytoma. After surgery there was a subgaleal hygroma at the craniotomy site. The migraine headaches ceased after this surgical resection, but the patient developed a different type of intermittent headache that had definite postural features. These particularly occurred when the patient was recumbent, sometimes associated with nausea and emesis. After being upright for several minutes the headaches would ease and would subsequently subside. Head MRIs, in addition to the postoperative changes, showed diffuse pachymeningeal enhancement, small subdural fluid collections, and large extracranial subgaleal collection of CSF (Fig. 1, A and B, middle). With cessation of the leak and disappearance of extracranial fluid (CSF), the pachymeningeal enhancement (Fig. 1, A and B, right) and the headaches also resolved. The patient, however, died in April 2000 with recurrence and further dedifferentiation of the tumour into a glioblastoma multiforme.

T2-weighted axial (A) and T1-weighted gadolinium-enhanced axial (B) and coronal (C) MRIs. The images of 11/19/98 show zone of T2 signal abnormality related to the previously biopsied and radiated grade 2 oligodendroglioma, as well as foci of enhancement at the deep margins of the lesion. MRIs of 9/13/99, 3 days after the resection of the recurrent dedifferentiated tumour (to a grade 3 mixed oligo-astrocytoma) show extracranial subgaleal fluid collection, diffuse pachymeningeal gadolinium enhancement and also bifrontal subdural fluid collections. MRIs of 10/18/99 (about 5 weeks later) show resolution of subgaleal fluid and also disappearance of diffuse pachymeningeal enhancement and subdural fluid collections.
Discussion
CSF, a circulating body fluid sometimes referred to as the ‘third circulation’ (5) is formed by the choroid plexus at a rate of about 0.35 ml per minute. The mean volume of CSF in adults is about 210 ml. There is, however, considerable variability. The volume of cranial CSF is greater for men and smaller for women, and much smaller for younger persons compared with older ones who have more generous ventricular volumes and subarachnoid cisterns (6). The volume of spinal CSF is less in obese than in non-obese persons (7).
In the horizontal position, the CSF pressure in the lumbar, cisternal and the intracranial or vertex is equal, measuring approximately 60–180 mm of water (8). When an individual assumes an erect posture, these pressures diverge. The vertex CSF pressure becomes negative, while lumbar pressure rises to between 375 and 550 mm of water. The zero point lies somewhat above the level of cardiac atria and somewhat below the level of cisterna magna. In a series of experiments aimed at the study of headaches resulting from experimental drainage of cerebrospinal fluid, Kunkle et al. (8) noted that headaches in the upright posture could be induced when approximately 10% of the CSF volume was withdrawn. This 10% decrease in CSF volume caused approximately 40% decrease in the already negative vertex CSF pressure, an observation that was in agreement with the exponential relationship between CSF pressure and volume (9).
With recognition of MRI abnormalities in CSF volume depletion and identification of a much larger number of patients than before, a broader clinical and imaging spectrum of the disorder is recognized, including patients with CSF pressures that are consistently within normal limits (10), or those who may not show abnormal pachymeningeal enhancement (11), or patients with documented CSF leaks, low CSF pressures, and diffuse pachymeningeal gadolinium enhancement, who have no headaches (12). Furthermore, the headaches in CSF leaks may show considerable variability. Although some patients may have headaches that lack orthostatic features, in the majority the headaches are orthostatic, and even when the headache is lingering and steady, often it is worse when the patient is upright and decreases with recumbency.
The case reported here is featured by a posture-dependent headache with features paradoxical to the usual headaches of CSF leaks. In this patient, the headaches were invariably produced by lying down for several minutes and were relieved after several minutes of assuming an upright posture. Since the level of the leak was at the craniotomy site, it is likely that CSF loss took place when the patient was reclined, a position in which there is an increase in intracranial CSF pressure. Loss of CSF volume in this position probably somehow led to stress on some of the intracranial pain-sensitive structures. On the other hand, in the upright posture, craniospinal divergence of CSF pressures that result in negative intracranial pressure probably led to cessation of the leak and disappearance of the headache. The headache resolution, however, was not immediate and appeared after several minutes of delay, enough for some of the depleted CSF to be replenished.
Although unexplained, some have noted that the patients with brain tumours are more likely to present with headaches if they have a prior history of headaches (13). The patient reported here had a long-standing history of migraine that substantially increased in frequency with her brain tumour. Interestingly, although the migraine headaches were not effected by biopsy or radiation therapy, they completely resolved after gross total resection of the tumour.
It might be argued that the patient's posture-related headaches were due to the brain tumour. However, these developed only after resection of the tumour, when there was no mass effect or shift of the midline. Imaging data suggest correlation between brain tumour headaches and size of the tumour, as well as shift of the midline structures. Furthermore, headaches of brain tumours do not possess such very reproducible posture-related features. The ‘classic’ view that brain tumour headaches are severe, worse in the morning, and associated with nausea and emesis, does not hold in a substantial majority. Only in one-third of the patients are the headaches worse in the morning, while in another third there is no particular timing, and the remaining third is roughly split between headaches that are worse during the day and those that are worse at night (13).
It could also be argued that meningeal enhancement might have been post-surgical. Postoperative meningeal enhancements (14) are typically focal and often, although not always, unilateral. The meningeal enhancement in our patient involved only the pachymeninges, was bilateral, diffuse, uninterrupted, non-nodular, and involved both supratentorial and infratentorial pachymeninges. Resolution of the patient's posture-related headaches, along with disappearance of the subgaleal fluid (cessation of the leak), in tandem with normalization of the pachymeningeal enhancement, strongly suggest that these were related. This unusual posture-related headache resulting from CSF leak from a site that is less frequently encountered by neurologists than neurosurgeons, is yet another mode of the still broadening clinical presentation of CSF leaks.
