Spontaneous Csf Leaks Mimicking Benign Exertional Headaches

Abstract

Spontaneous CSF leaks are increasingly recognized, and a broader clinical and imaging spectrum of the disorder is emerging. The headaches of CSF leaks are typically orthostatic, but sometimes especially with chronicity the orthostatic features are blurred into lingering chronic daily headaches. Additional types of headache are also increasingly recognized. Two patients with spontaneous CSF leaks presented with intermittent transient severe headaches provoked by Valsalva-type manoeuvres. Orthostatic features were absent and the patients were asymptomatic if they avoided the provoking manoeuvres. One patient had been treated for 6 years for benign exertional headaches and had failed many medical treatments, including courses of indomethacin. He was found to have a leak from cribriform plate. The second patient had been symptomatic for several months, had diffuse pachymeningeal gadolinium enhancement on head magnetic resonance imaging, spinal meningeal diverticula, and CSF leak at the thoracic spine level. Headaches that mimic benign exertional headaches are yet another mode of the still broadening clinical presentation of spontaneous CSF leaks.

Introduction

The clinical and imaging spectrum of CSF leaks is broadening (1, 2). Although occasional patients may not display headaches—the so-called acephalgic form of CSF leaks (3)—the headache is by far the most common clinical manifestation of the CSF leaks. These headaches are typically orthostatic (present in the upright position, relieved by recumbency). These may or may not be accompanied or precede or followed by interscapular or posterior neck pain, they may be throbbing, but often are non-throbbing (4, 5). Sometimes, especially with chronicity, the headaches may lose their orthostatic features and transform into chronic daily headaches or steady lingering headaches (6). Sometimes the headaches of CSF leaks may begin as lingering chronic daily type headaches and after days or weeks the typical orthostatic features may appear. Sometimes the headaches remain without any orthostatic features. Rarely the headaches may have thunderclap onset (7) or display paradoxical orthostatic features (present when recumbent and relieved when upright) (8). Occasional patients may report onset of headaches after having been up and about for some hours and therefore report headaches that typically appear in the afternoon. These ‘second half of the day’ headaches (B Mokri, unpublished data) are often improved or relieved by recumbency, although not quickly. The headaches of CSF leaks are often noticeably intensified by Valsalva-type manoeuvres. In this study, two patients with spontaneous CSF leaks are reported in whom the only clinical manifestation consisted of headaches provoked by Valsalva-type manoeuvres that clinically mimicked typical benign exertional headaches.

Patient 1

A previously healthy, right-handed man with a distant history of nasal polypectomy and partial turbinate removal, in 1997 at age 42 years began to note occipital headaches provoked by Valsalva-type manoeuvres including coughing, sneezing, bending, lifting, and straining. These were somewhat throbbing, each lasting 20–40 min. A head magnetic resonance imaging (MRI) in March 1998 showed diffuse pachymeningeal gadolinium enhancement (Fig. 1A) and descent of the cerebellar tonsils below the foramen magnum, crowding of the posterior fossa, and obliteration of perichiasmatic cistern. A left frontal meningeal biopsy showed only normal dura with reactive meningothelial cells. Microbiology studies were negative, and antibiotic treatments brought no relief. A computed tomography (CT) myelogram in late October 1998 showed a 5-mm meningeal diverticulum at T12 and an 8-mm diverticulum at T8, both on the left. An Indium-111 radioisotope cisternography showed slight asymmetry of uptake over the cerebral convexities.

Figure 1

(Patient 1) (A) Coronal section T1-weighted gadolinium-enhanced MRI during the initial symptomatic phase demonstrates diffuse pachymeningeal gadolinium enhancement. (B) After resolution of symptoms, the diffuse pachymeningeal enhancement has also resolved. When (A) and (B) are compared, in retrospect it becomes obvious that during the symptomatic phase there were additional subtle abnormalities, including ‘ventricular collapse’ (ventricles are smaller in (A) than in (B)), and enlargement of the pituitary gland (larger in (A) than in (B)). After recurrence of the symptoms (C), all of the previously noted MRI abnormalities have reoccurred (diffuse pachymeningeal gadolinium enhancement, enlarged pituitary, and ventricular collapse). At the same time, unenhanced sagittal MRI (D) shows descent of the cerebellar tonsils below the foramen magnum and the engorgement of sagittal sinus (best seen in the posterior one-half of the sinus where there is also engorgement of the adjacent veins).

In January 1999, the patient was seen at the Mayo Clinic Rochester. By this time, the headaches had nearly resolved. The neurological examination was normal. A repeat head MRI showed decrease in descent of cerebellar tonsils and complete resolution of meningeal enhancement (Fig. 1B). The recommendation was to further observe rather than proceed with invasive therapeutic procedures. Subsequently, headaches completely resolved over the next several weeks. Several months later there was recurrence of headaches and reappearance of meningeal enhancement, descent of cerebellar tonsils, enlargement of pituitary, decrease in size of perichiasmatic cistern, and enlargement of cerebral venous sinuses (Fig. 1C,D). The headaches were less severe and were shorter in duration (5–10 min). He had learned how to avoid provoking factors and how to cope with the headaches, and decided not to proceed with further diagnostic or therapeutic measures.

Patient 2

In May 1986 at age 54, this right-handed man began to note headaches provoked by Valsalva-type manoeuvres such as lifting, coughing, sneezing, blowing nose, straining, etc. The headache which was primarily right frontal tended to extend to the entire head. There were no associated manifestations. Each headache would last for about 20 min. If he avoided the provoking manoeuvres, he was pain-free. A head CT scan was normal. Trial of over-the-counter analgesics brought no relief. The pain gradually became more prominent, although maintaining its original features. In July 1986 the patient was seen by a neurologist, and neurological examination was normal. An unenhanced head MRI was normal. The headaches were diagnosed as benign exertional headaches. A trial of indomethacin 25 mg b.i.d. and later a trial of long-acting propranolol 80 mg per day and then 120 mg per day offered no relief. Indomethacin was tried again with a dose of 25 mg t.i.d. and again there was no benefit. The patient attempted to avoid the provoking factors as much as possible and lived with the intermittent headaches when they occurred. In late 1980s and early 1990s, he began to make the observation that sometimes with Valsalva-type manoeuvres a small amount of clear fluid leaked from the nose. Indium-111 radioisotope cisternography demonstrated abnormal tracer distribution with little flow over the cerebral convexities even at 24 h. Furthermore, radioactivity count of the nasal pledgets showed increased activity in the right nasal pledgets, suggesting a right-sided CSF leak. CSF examination at the time of cisternography showed an opening pressure slightly below 100 mm of water. Protein and glucose concentrations and cell count were normal. A high-resolution head CT scan with 3-mm coronal sections through the paranasal sinuses performed before and after injection of water-soluble contrast in the subarachnoid space for cisternography failed to show the site of the leak. As of December 2000, patient's intermittent headaches had persisted, although he had learned how to avoid provoking factors as much as possible. His neurological examination was normal. The patient did not wish to pursue with further diagnostic studies as he thought that he had learned to cope with the problem.

Discussion

Valsalva-type manoeuvres (such as coughing, sneezing, straining, heavy lifting, bending, pushing, etc.) can aggravate any type of headache. Exertional headaches, however, are referred to transient headaches that are actually provoked by such manoeuvres. These headaches are more frequently associated with intracranial lesions than not (63% vs. 37%) (9). When no structural cause is detected, these headaches are referred to as benign exertional headaches. Jule Tinel, the French neurologist best known for the sign bearing his name, in 1932 originally described the exertional headaches under the term ‘la céphalée a l'effort' (10). Sir Charles Symonds in 1956 reported his experience on 27 patients with ‘cough headaches.’ Six of these patients had documented structural (organic) aetiology (11). Rooke and his colleagues at the Mayo Clinic by 1950 had come across several successive patients with exertional headaches who had demonstrated no evidence of intracranial pathology despite vigilant search. In 1968 Rooke reported the results of the study he had undertaken in 1950, and he used the term ‘benign exertional headaches’ for all acute transient headaches provoked by exertion and Valsalva-type manoeuvres, if they were not associated with structural intracranial pathology (9).

The classification of the International Headache Society (IHS) (12) chose to split what essentially Rooke had lumped under benign exertional headaches into: (i) benign cough headaches, (ii) benign exertional headaches, and (iii) benign sexual headaches. The latter per se was subclassified into the dull headache type, explosive headache type (the most common one), and postural headache type.

It is worth noting that six of the 27 patients with cough headaches reported by Symonds were found to have organic disease and on long-term follow-up complete relief had occurred in nine, improvement in six, while the remaining six patients had either continued without change or had died of unrelated cause. In Rooke's study of a much larger number of patients with exertional headaches, a considerable majority showed structural organic disease. Rooke's series consisted of 303 patients with symptomatic exertional headaches (there was evidence of intracranial structural disease) and 103 patients with benign exertional headaches (with no evidence of intracranial disease). Even in this group of 103 patients, on long-term follow-up, 10 patients disclosed presence of intracranial disease. It is likely that the technology of the time would not have allowed Symonds or Rooke to study the patients as extensively and accurately as can be done today. Therefore, it is probable that some of the patients that were considered to have had no structural disease may indeed have had problems that could have been identified by today's technology. As an example, it is likely that our first case may not have been diagnosed with organic structural disease, not only at the time of Rooke and Symonds but in the pre-MRI era, even with the availability of head CT scanning. Furthermore, it is possible that the same patient might had been mistakenly diagnosed as Chiari malformation in the earlier years of MRI without the advantage of gadolinium enhancement. It might be argued that the headaches in our two patients might have been exertional headache variants. This would appear unlikely. In the first patient, when he was symptomatic, head MRI showed typical findings that are seen in CSF leaks including diffuse pachymeningeal enhancement and MR evidence of descent of brain. When the symptoms markedly improved, the pachymeningeal enhancement completely resolved, and there was definite decrease in descent of the brain. When the symptoms increased again, the same MRI abnormalities that had improved or resolved again appeared. Furthermore, when this patient was initially symptomatic, he underwent a meningeal biopsy in another institution. The biopsy findings were identical to those reported in CSF leaks (13). In the second patient, not only did nasal pledget show unilateral increase in radioactivity, more importantly there was very little activity over the cerebral convexities at 24 h. This finding in the absence of meningeal disease is highly suggestive of CSF leak (14). Headaches that closely mimic benign exertional headaches are yet another mode of the still broadening clinical presentation of spontaneous CSF leaks.