Secondary Logo

Journal Logo


Neurosarcoidosis: How Good Are the Diagnostic Tests?

Lynch , Joseph P. III MD

Author Information
Journal of Neuro-Ophthalmology: September 2003 - Volume 23 - Issue 3 - p 187-189
  • Free

In this issue of the Journal, Frohman et al. (1) report 24 patients with anterior visual pathway (AVP) sarcoidosis and emphasize that sarcoidosis is frequently not recognized at presentation. They contend that a strong presumptive diagnosis of sarcoidosis usually can be established without having to perform biopsies of optic nerves or intracranial sites. Here is their evidence:

Among the 17 patients in whom the diagnosis of sarcoidosis was not known at the time of presentation, all had an abnormality on at least one of five diagnostic tests: serum angiotensin converting enzyme (SACE), chest radiograph, gallium-67 citrate (Ga-67 ) scan, brain magnetic resonance imaging (MRI), and lumbar puncture (see their Table 4). All but two patients had abnormal results on more than one of these tests. In one patient (Patient #5), only SACE was abnormal; in patient #15, only the Ga-67 scan was abnormal; lumbar puncture was not performed in either patient. Seven patients had abnormalities on three tests; 10 had abnormalities on four.

Based on my experience and the reports of other series dealing with neurosarcoidosis, their yield from noninvasive diagnostic tests appears relatively high. But when these tests are combined with bronchoscopy or mediastinoscopy, it should be possible to reach a diagnosis in most cases without having to resort to intracranial biopsy.

AVP sarcoidosis is a subset of neurosarcoidosis, a variant that afflicts 5% to 18% of patients with sarcoidosis (2–6). Seventh nerve palsy is the most common focal manifestation, occurring in up to 50% of patients with neurosarcoidosis (3,4,6,7). Eighth nerve palsy may be next most common, but it is often unrecognized by the patient. AVP involvement probably ranks third, comprising 15% of cases of neurosarcoidosis. Other neurological intracranial manifestations include headache, encephalopathy, stroke, seizures, hydrocephalus, focal mass lesions, involvement of the pituitary or hypothalamus, and myelopathy (3,4,6,7).

Histologic diagnosis of the intracranial lesions is difficult due to their relative inaccessibility and the potential morbidity associated with biopsy. Fortunately, more than 80% of patients with neurosarcoidosis have manifestations outside the cranial cavity, but they may be occult. In three series of neurosarcoidosis, non-neurologic involvement was noted in 94% (67 of 71) (7), 83% (29 of 35) (8), and 100% (24 of 24) (9) of patients. In two large series, intracranial biopsy was necessary in only 8% (6 of 71) (7) and 18% (12 of 68) of patients (4). Unfortunately, studies assessing the sensitivity and specificity of various clinical, radiographic, or serological studies in patients with suspected ocular or CNS sarcoidosis are rudimentary. Here is a comparison of the results of Frohman et al. (1) and other investigations.

  1. Serum ACE concentrations were elevated in 16 of 21 (76%) patients in this series (1). In two other series of neurosarcoidosis, elevated SACE levels varied from 24% (12 of 51) (4) to 65% (11 of 17) (8) of patients.
  2. Ga67scans demonstrated increased uptake in salivary glands or lung in 93% of patients in this series (1). Importantly, Ga67 scans were abnormal in all 10 patients with no known history of sarcoidosis at the time of the visual loss. But in two other series of neurosarcoidosis, the reported sensitivity of Ga67 scans was considerably lower: 57% (4 of 7) (8) and 45% (14 of 31) (4).
  3. Chest radiographs were consistent with sarcoidosis in 72% of patients in this series (1). Other investigators have found abnormal chest radiographs consistent with sarcoidosis in as many as 71% (25 of 35) (8) and as few as 31% (21 of 68) of patients (4). Chest computed tomographic (CT) scans are more sensitive for sarcoidosis and may be helpful in patients with normal chest radiographs (10).
  4. Lung and mediastinal biopsy were infrequently performed in this series (1) but are often rewarding in patients with neurosarcoidosis (11). Bronchoscopy has a diagnostic yield of 60% to 97%, according to the number of biopsies and radiographic stage of the disease (12). One study of patients with neurosarcoidosis cited a positive yield in 59% (11 of 17) (8). When hilar lymphadenopathy is present without lung parenchymal infiltrates, yields are 60% to 85%. Higher yields may be expected when lung infiltrates are present (12); they should be lower when patients have normal chest radiographs, but there are no published data on this point. Mediastinoscopy has a higher yield than bronchoscopy in patients with hilar or mediastinal adenopathy, but it is more invasive, expensive, and has increased morbidity (including keloid formation). Therefore, mediastinoscopy should be reserved for patients with intrathoracic lymph node enlargement and negative bronchoscopy.
  5. Brain MRI was positive in 70% of patients in the current series, showing thickening and enhancement of the optic nerve and cranial base meninges. These findings are similar to previous reports (6,9), but they are nonspecific and may occur in myriad infectious and neoplastic diseases (6).
  6. Lumbar puncture revealed a lymphocytic pleocytosis or elevated protein in 88% of patients in this series (1). Others have reported elevations in CSF protein or leukocytes in 70% to 73% (4). However, these CSF abnormalities tend to be mild and utterly nonspecific (4,6,8) and do not reflect the course or severity of neurologic involvement.
  7. Electromyography/nerve conduction tests (EMG) (8), the Kveim skin test (4), and biopsies of skin, conjunctiva, lacrimal glands, upper respiratory tract mucosa, liver, or muscle (4,8) often are applied to the diagnosis of occult sarcoidosis. EMG is uncomfortable and its sensitivity is likely to be low in the absence of localizing findings. The Kveim test involves intradermal inoculation of sarcoidal splenic tissue and biopsying the site of induration 5 to 6 weeks later to document granulomas (4). In neurosarcoidosis series, positive Kveim tests have been noted in 85% (41 of 45) (4) and 40% (2 of 5) of patients (8). However, the Kveim test is available in only a few centers worldwide, is not standardized, and results are not available for several weeks. Given these limitations, it has been supplanted by other techniques. Biopsies of conjunctiva or lacrimal glands have substantiated the diagnosis of sarcoidosis in 10% to 55% of patients (5,13), but the yield is low in the absence of visible nodules. Muscle biopsies have shown non-necrotizing granulomas (NNG) in 25% to 50% of patients with sarcoidosis, particularly when arthralgias, erythema nodosum, or fever are present (5). However, muscle biopsies are uncomfortable and the yield is not known in patients without muscle symptoms or signs. NNG have been noted in 21% to 97% of patients with sarcoidosis who had percutaneous liver biopsies (5). However, yields are likely to be low in the absence of specific liver findings. Because liver biopsies have potential morbidity, they should be considered only in patients with abnormal liver function tests or hepatic enlargement or low attenuation hepatic lesions on CT.

Mindful of these data, I recommend the following approach to the diagnosis of suspected AVP sarcoidosis. When clinically evident superficial lesions are present, such as conjunctival or cutaneous nodules or enlarged lacrimal glands, biopsy of these sites is warranted. When no obvious or visible lesion is identified, Ga67 scan or chest CT scan should be done, aiming to pinpoint sites to biopsy. If the Ga67 scan is negative, transbronchial lung biopsy should be performed, irrespective of whether chest CT scan shows specific aberrations. If bronchoscopy is negative and chest CT scan demonstrates intrathoracic adenopathy, mediastinoscopy should be performed. If these maneuvers are negative, I do not recommend “blind” biopsies of conjunctiva, muscle, or liver, as the yields are too low. If an extracranial histologic diagnosis cannot be reached, an empiric trial of high dose corticosteroid therapy should be initiated, provided the clinical course is consistent with sarcoidosis and appropriate follow-up can be assured.


1. Frohman L, Guirgis M, Turban R, et al. Sarcoidosis of the anterior visual pathway: 24 new cases. J Neuro-Ophthalmol 2003; 23:190–197.
2. Gullapalli D, Phillips 2nd. LH Neurologic manifestations of sarcoidosis. Neurol Clin 2002; 20:59–83.
3. Stern BJ, Krumholz A, Johns C, et al. Sarcoidosis and its neurological manifestations. Arch Neurol 1985; 42:909–17.
4. Zajicek JP, Scolding NJ, Foster O, et al. Central nervous system sarcoidosis—diagnosis and management. QJM 1999; 92:103–17.
5. Lynch 3rd, JP Sharma OP, Baughman RP. Extrapulmonary sarcoidosis. Semin Respir Infect 1998; 13:229–54.
6. Tierstein A. Neuromuscular sarcoidosis. Semin Respir Crit Care Med 2002:505–12.
7. Lower EE, Broderick JP, Brott TG, et al. Diagnosis and management of neurological sarcoidosis. Arch Intern Med 1997; 157:1864–8.
8. Chapelon C, Ziza JM, Piette JC, et al. Neurosarcoidosis: signs, course and treatment in 35 confirmed cases. Medicine (Baltimore) 1990; 69:261–76.
9. Dumas JL, Valeyre D, Chapelon-Abric C, et al. Central nervous system sarcoidosis: follow-up at MR imaging during steroid therapy. Radiology 2000; 214:411–20.
10. Kosmorsky GS, Meisler DM, Rice TW, et al. Chest computed tomography and mediastinoscopy in the diagnosis of sarcoidosis-associated uveitis. Am J Ophthalmol 1998; 126:132–4.
11. Ohara K, Okubo A, Kamata K, et al. Transbronchial lung biopsy in the diagnosis of suspected ocular sarcoidosis. Arch Ophthalmol 1993; 111:642–4.
12. Lynch 3rd, JP, Kazerooni EA, Gay SE. Pulmonary sarcoidosis. Clin Chest Med 1997; 18:755–85.
13. Nichols CW, Eagle Jr, RC Yanoff M, et al. Conjunctival biopsy as an aid in the evaluation of the patient with suspected sarcoidosis. Ophthalmology 1980; 87:287–91.
© 2003 Lippincott Williams & Wilkins, Inc.