Isolated CNS Sarcoidosis Versus Systemic Sarcoidosis With CNS Involvement: A Same Disease?

Abstract

Neurological involvement occurs in 5 to 15% of patients with sarcoidosis. It rarely represents the sole manifestation of the disease, a condition called isolated neurosarcoidosis.

Objectives:

To describe patients with definite isolated central neurosarcoidosis. To compare their characteristics to a group of systemic sarcoidosis with central neurologic involvement.

Methods:

Monocentric retrospective study of all patients presenting with central neurosarcoidosis (NS) over a 10 year period, subsequently divided into 2 groups: isolated neurosarcoidosis (INS) and systemic neurosarcoidosis (SNS).

Results:

We report 10 cases of INS and subsequently, we compared their characteristics to a group of 30 patients with SNS. INS patients exhibited brain parenchymal involvement (8/10), meningeal disease (8/10), myelitis (3/10), cranial neuropathy (3/10), neuroendocrine impairment (1/10). Cerebro-spinal fluid (CSF) analysis was conducted in 8/10 patients and showed pleocytosis in 6/8 (75%), elevated protein level in (4/8) 50%, oligoclonal intrathecal synthesis in 1/5 (20%). All patients received steroids, 7/10 (70%) required associated immunosuppressive therapy, 5 of which TNFα inhibitors. When compared to patients with SNS, INS patients were more likely to experience seizures (60% vs 23.3%); display encephalic parenchymal enhancing lesions (80% vs 39.3%) or encephalic leptomeningeal involvement (80% vs 35.7%). Serum angiotensin converting enzyme (ACE) was elevated in a third of patients with SNS but none of those with INS.

Conclusion:

The phenotypes of patients with INS are similar to the ones described in SNS. Serum ACE should not be regarded as a diagnostic test in patients with isolated neurosarcoidosis but could be useful in detecting subclinical extra neurologic involvement during follow up.

Introduction

Sarcoidosis is a multisystemic granulomatous disorder of unknown cause. Even though lungs, skin, eyes and their respective draining lymph nodes are the most commonly affected organs, nervous system involvement occurs clinically in approximately 5 to 15% of patients. ¹ The disease is known to affect any part of the nervous system but most frequently manifests as meningeal disease, cranial nerve palsy, parenchymal mass and spinal cord lesions. ² Despite extensive workup, up to one fifth of patients show no evidence of extra neurological involvement, a condition called isolated neurosarcoidosis ³ (INS), which represents a diagnostic challenge for the practitioner. This condition, for which pathology keeps a central role, has received few attention in the past, and the updated criteria for neurosarcoidosis ³ have questioned the existence of an “organ-specific granulomatous disorder” distinct from sarcoidosis.

The objectives of our study were first, to describe a case-series of patients presenting with isolated central nervous system (CNS) sarcoidosis, lacking any extra-neurological involvement, and secondly, to compare their characteristics to a group of patients with systemic sarcoidosis with CNS involvement (systemic neurosarcoidosis or SNS).

Methods

We retrospectively reviewed the medical records of all patients diagnosed with sarcoidosis associated with neurological disorder (using International Classification of Diseases-10 codes) identified at Bordeaux University teaching Hospital (tertiary center) during the period between January 2009 and January 2019. Concomitantly we looked up in the Pathology laboratory Database (ADICAP) all the results of nervous system biopsy consistent with non-caseating granuloma. Patients were included if they fulfilled the criteria for definite or probable neurosarcoidosis ³ and exhibited CNS involvement, which was mandatory. Patients with possible neurosarcoidosis were excluded. Data collected included demographic details, clinical features, ancillary studies, pathology, treatment and outcomes. This study was approved by the local ethic committee of the University Hospital of Bordeaux. Statistical analysis was performed with XLSTAT® software to compare differences between the 2 groups, using non parametric X2 test for qualitative values and Mann-Whitney tests for quantitative values. A p-value < 0.05 was considered significant. The data that support the findings of this study are available from the corresponding author, upon reasonable request.

Results

352 records were identified in our analysis, of which 40 fulfilled the criteria for definite (n = 16) or probable (n = 24) neurosarcoidosis with CNS involvement. Among them, 10/40 patients (25%) had no extra-neurological involvement detected clinically and on morphologic imaging studies, fulfilling the definition of INS. Conversely, 30/40 patients (75%) had at least one extra neurological involvement of sarcoidosis, and were classified as systemic neurosarcoidosis (SNS).

Clinical characteristics and ancillary studies of patients with INS are detailed in Table 1.

Table 1.

Characteristics and Disease Course in Patients with Isolated Definite Neurosarcoidosis.

CASE	AGE AT ONSET	SEX	CLINICAL PRESENTATION	CSF	MRI	BIOPSY	TREATMENT AND OUTCOMES
N°1	64	M	Subacute cervical myelitis with severe tetraparesis and sphincter dysfunction	–	Cervical spondylosis and cervical longitudinally extensive myelitis with dorsal subpial enhancement	Cervical meninges and spinal cord parenchyma	Laminectomy, Steroid monotherapy Monophasic
N°2	35	F	Unilateral facial palsy	Lymphocytic meningitis and elevated protein	Multiple parenchymal hyperintense nodules with enhancement and diffuse perivascular enhancement	Left temporal parenchyma	Steroid monotherapy Multiphasic requiring escalation therapy and TNFα inhibitors
N°3	42	M	Generalized tonic-clonic seizure	Lymphocytic meningitis and elevated protein	Left frontal cortical hyperintensity with restricted diffusion, pachymeningeal enhancement	Left frontal meninges	Steroid monotherapy Monophasic
N°4	59	M	Focal left hemisensory seizure	Lymphocytic meningitis and elevated protein	Right frontal hyperintensity with leptomeningeal enhancement	Right centrum semiovale	Steroid monotherapy Multiphasic Suicide
N°5	65	F	None	–	Multiple parenchymal hyperintensities with diffuse leptomeningeal enhancement	Right parieto-occipital parenchyma	Steroid monotherapy Multiphasic requiring escalation therapy and TNFα inhibitors
N°6	32	F	Focal left occipital seizure with secondary generalization	Lymphocytic meningitis	Left occipital parenchymal and leptomeningeal enhancement with vasogenic edema	Left parieto-occipital parenchyma and meninges	Steroids and TNFα inhibitors Monophasic
N°7	38	M	Bilateral optic neuritis	Lymphocytic meningitis and elevated protein	Bilateral optic nerve hyperintensity with multiple parenchymal leptomeningeal enhancement	Left temporal parenchyma	Steroids and cyclophosphamide Monophasic
N°8	46	F	Intracranial hypertension	Normal	Infratentorial leptomeningeal enhancement with hydrocephalus	Cerebellar tonsil	Cerebro-spinal fluid derivation, steroids and cyclophosphamide Monophasic
N°9	50	F	Subacute diabetes insipidus and nodular myositis	Normal	Normal	Muscle	Steroid monotherapy Multiphasic requiring escalation therapy and TNFα inhibitors
N°10	59	F	Left lower leg paresis	Lymphocytic meningitis and 10 oligoclonal bands	Small conus medullaris myelitis with enhancement	Right frontal parenchyma	Steroids and cyclophosphamide Multiphasic requiring escalation therapy and TNFα inhibitors

The INS group consisted of 6 females and 4 males, with an average age of 49.1 years old [32-65]. They all fulfilled the criteria for definite neurosarcoidosis.

All patients except one were diagnosed after subacute development of neurological symptoms. Clinical presentation consisted of seizures (n = 3), myelitis (n = 2), unilateral facial palsy (n = 1), bilateral optic neuritis (n = 1), intracranial hypertension (n = 1), and diabetes insipidus (n = 1). CNS involvement in patient #5 was discovered fortuitously during systematic brain magnetic resonance imaging (MRI) follow-up after aneurysm embolization. Initial brain MRI (Figure 1) showed both meningeal and parenchymal lesions in 6/10 (60%) patients and isolated leptomeningeal involvement in 1/10 patient (10%). Leptomeningeal enhancement occurred in 7/10 patients (70%) and was more common than pachymeningeal enhancement occurring in 2/10 patients (20%), with only one patient featuring both. One patient had normal imaging despite severe central diabetes insipidus. The 2 patients with initial myelopathy had spinal cord MRI showing either longitudinally extensive myelitis with dorsal subpial involvement or small tumefactive myelitis.

Figure 1.

MRI results of 2 patients with isolated neurosarcoidosis. (A) patient #5: FLAIR - axial – large right occipito-temporal hyperintensity (B) patient #5: contrast enhanced T1FS - axial - multiple right occipito-temporal parenchymal and leptomeningeal enhancement (C) patient #6: FLAIR - axial - left occipital cortical and subcortical hyperintensity (D) patient #6: contrast enhanced T1 - axial – left occipital parenchymal and leptomeningeal enhancement.

All of these 10 patients had normal thoraco-abdominal CT scan. 6/10 patients (60%) underwent 18-Fluorodeoxyglucose (FDG) positron emission tomography (PET)-scan, which revealed no abnormalities in 5 and allowed the diagnosis of ovarian neoplasm recurrence in one (#5). Minor salivary gland biopsy was performed in 9/10 patients and none showed evidence for granuloma. Cerebro-spinal fluid (CSF) analysis was conducted on 8/10 patients and revealed pleocytosis for 6/8 (75%), elevated protein level for 4/8 (50%), and oligoclonal synthesis for 1/5 patient (20%). Blood lymphopenia was observed in 4/10 patients (40%) and slight elevation of CRP in 5/10 patients (50%). Serum Angiotensin-converting enzyme (ACE) level was normal in all patients.

Pre-biopsy presumed diagnosis was mostly neoplasm or infectious disease, especially tuberculosis.

All patients underwent biopsy of brain parenchyma (n = 8), spinal cord (n = 1) or muscle (n = 1) exhibiting non-caseating granuloma. Only one patient displayed slight necrosis on biopsy but thorough work up for infectious differential diagnosis was negative. There was no evidence of CNS spread of ovarian neoplasm in patient #5. They all fulfilled the criteria for “definite neurosarcoidosis” defined by Stern and al. (3).

First line therapy consisted of steroids for all patients, combined with immunosuppressive therapy for 4 of them: cyclophosphamide (n = 3) and infliximab (n = 1). Additionally, 3 patients required immunosuppressive therapy during follow-up, and 5 patients were treated with infliximab. One patient died by suicide before escalation therapy and 2 patients were long lost after respectively 8 and 20 months.

During a median follow-up of 25.5 months, 5/10 patients (50%) had relapsing remitting course and the other 5/10 patients (50%) had monophasic illness. Brain parenchymal involvement (8) and meningeal disease (8) were the most commonly observed lesions, followed by cranial neuropathy ³ and myelitis. ³ Only one patient (#9) developed PNS involvement in the form of bilateral lower limbs nodular myositis with positive muscle biopsy.

Under treatment, none of our 10 patients exhibited clinical features of extra-neurologic sarcoidosis during a mean follow-up of 56 months, 60% of them benefited repeated chest CT without any evidence of mediastinal or pulmonary involvement.

We compared the 10 patients from the INS group to 30 patients with SNS characterized by both CNS and extra neurological involvement. Data comparison can be found in Table 2.

Table 2.

Comparison of Isolated and Systemic Neurosarcoidosis (NS) Patients.

	ISOLATED NS	SYSTEMIC NS	p-value
Characteristics	n = 10	n = 30
Female	6 (60%)	16 (53.3%)	0.741
Male	4 (40%)	14 (46.7%)
Age	49.1 (12.4)	46.9 (15.2%)	0.7
Clinical symptoms	n = 10	n = 30
Optic neuritis	1 (10%)	0 (0%)	0.079
Other cranial neuropathy	2 (20%)	14 (46.7%)	0.14
Seizures	6 (60%)	7 (23.3%)	0.032
Myelitis	3 (30%)	6 (20%)	0.51
Neuro-endocrine	1 (10%)	4 (13.3%)	0.78
Peripheral nervous system	1 (10%)	10 (33.3%)	0.15
Systemic features i	0 (0%)	9 (30%)	0.049
Brain MRI	n = 10	n = 28
Pseudo tumoral lesions	4 (40%)	5 (17.9%)	0.16
Parenchymal enhancement	8 (80%)	11 (39.3%)	0.027
Leptomeningeal enhancement	8 (80%)	10 (35.7%)	0.016
Pachymeningeal enhancement	2 (20%)	3 (10.7%)	0.46
Hydrocephalus	1 (10%)	1 (3.6%)	0.43
Spine MRI	n = 5	n = 13
Myelitis	3 (60%)	5 (38.5%)	0.41
Spinal cord enhancement	2 (40%)	4 (30.8%)	0.71
Meningeal enhancement	2 (40%)	4 (30.8%)	0.71
Normal	2 (40%)	8 (61.5%)	0.41
Serum	n = 10	n = 30
Elevated Calcium (%)	1 (10%)	3 (10%)	1
Elevated CRP (%)	5 (50%)	12 (40%)	0.58
Elevated ACE (%)	0 (0%)	10 (33.3%)	0.035
Lymphocytopenia (%)	4 (40%)	6 (66.6%)	0.14
CSF	n = 8	n = 17
Pleocytosis	6/8 (75%)	11/17 (65%)	0.607
Elevated protein level	4/8 (50%)	11/17 (65%)	0.484
Oligoclonal synthesis	1/5 (20%)	2/11 (18.2%)	0.931
Evolution	n = 10	n = 30
Monophasic	5 (50%)	18 (60%)	0.580
Multiphasic	5 (50%)	12 (40%)	0.580

CSF cerebrospinal fluid; ACE angiotensin conversion enzyme; CRP C-reactive protein.

The SNS group included 16 females and 14 males with an average age of 46.9 years old [22-75]. All SNS patients had associated mediastinal involvement and 20% were diagnosed with sarcoidosis before onset of neurologic symptoms. Apart from mediastinal lymphadenopathy, most common extra neurologic lesions involved lungs (50%), joints (30%) and skin (13%). 20% of patients in this group fulfilled the criteria for definite neurosarcoidosis and 80% for probable neurosarcoidosis.

INS patients were more likely to experience seizures (60% vs 23.3%, p = 0.032), while SNS patients more often suffered from systemic features (fever and/or fatigue) at onset (30% vs 0%, p = 0.049). INS patients more often displayed parenchymal enhancing brain lesions (80% vs 39.3%, p = 0.027) and encephalic leptomeningeal involvement (80% vs 35.7%, p = 0.016). Serum ACE was elevated in a third of SNS but none of the INS patients (p = 0.035).

Discussion

INS is a rare and challenging condition with heterogeneous clinico-radiological presentations and numerous differential diagnoses. Nevertheless, it is important to highlight that neurological involvement in sarcoidosis is the initial manifestation of the disease in 50% of patients with neurosarcoidosis and that the absence of extra neurological features in those patients is reported in 10 to 19% of cases. ⁴ We focused our study on patients with CNS involvement because of its critical implication regarding the decision to undergo biopsy. Moreover, the spectrum of granulomatous diseases is distinct in patients with isolated PNS involvement. Whilst the vast majority of neurosarcoidosis phenotypes have been described in case reports of isolated neurosarcoidosis, grouped data relative to this specific entity are scarce, ⁴ and when mentioned in broader studies often lacking pathologic confirmation. The recently updated criteria for neurosarcoidosis published by the international consortium in 2018 ³ consider a specific place for INS and brought-up several interesting questions regarding this particular phenotype, notably the potential existence of a neurological organ-specific granulomatous disorder distinct from sarcoidosis.

We described here a study of 10 patients diagnosed with definite INS, only partially representative of the large phenotypic spectrum of CNS sarcoidosis. In our study, patients with INS included only those without any evidence of systemic involvement either initially or during the course of the disease. Pathological evidence of granulomatous involvement was also required. This criteria might have been too restrictive, and our findings cannot be easily generalized to patient with “probable neurosarcoidosis”. The small number of patients described certainly gives statistical comparison between the groups little power, and the absence of 18 FDG-PET in 4 patients might have missed extra neurologic involvement in a few of those. However it represents one of the main studies dedicated to this peculiar entity, with the one reported by Nozaki and al. ⁴

Our study does not support the existence of a specific entity distinct from sarcoidosis. Indeed, the before mentioned phenotypes of patients with INS encompass those described in SNS, and few collected data significantly differed between the 2 groups. In our study, INS patients were more likely to display parenchymal and/or leptomeningeal enhancement on brain MRI, as reported previously, ⁴ but this feature is often considered a pre-requisite before undergoing biopsy for cryptogenic CNS lesions. Serum ACE should not be regarded as a diagnostic test for neurosarcoidosis, especially in its isolated forms, consistent with the results reported by Nozaki and al. ⁴ it might however be useful in detecting systemic involvement during follow up. Systemic features were always absent in our patients with INS, and their presence should prompt the consideration for either systemic involvement or differential diagnosis.

Facing a patient with an isolated form of neurosarcoidosis despite systemic work-up, the highest remaining concern should be the large differentials of CNS granulomas, both infectious and non-infectious. To our knowledge no clear guidelines have been published to clarify the following steps to undertake after a CNS pathology exhibiting epithelioid granuloma. Despite the absence of reliable biomarkers for neurosarcoidosis, interesting newly available techniques, and notably CSF metagenomic analysis, ⁵ might assist the clinician in this difficult path.