LAUSR

Sarcoidosis has developed an almost legendary mystique among neurologists that is rooted in our limited understanding of the disease. The unknowns are daunting: the precise etiology of sarcoidosis remains unresolved, the factors that drive neurologic organ system involvement remain unclear, the diagnosis can be challenging to confirm in the nervous system, and the potential clinical manifestations of neurosarcoidosis (NS) are often alleged to be protean. Neurosarcoidosis is commonly considered in the differential diagnosis of patients with unexplained neurologic symptoms and clinical deficits, so much so that a tired aphorism is that no differential diagnosis is complete without consideration of sarcoidosis. This situation is not without consequence. Among neuroinflammatory disorders, diagnostic workups for NS account for a disproportionately large share of total inpatient costs,1 and tremendous variability in clinical practice exists regarding application of various proposed diagnostic criteria.2,3 Particularly in the absence of histopathologic findings, this can lead in some cases to premature or unwarranted diagnosis of NS. Despite the mystery that surrounds NS, the biology of sarcoidosis has become clearer during the past few decades.4,5 We now recognize a usually delimited set of neurologic syndromes most commonly associated with sarcoidosis,6,7 and treatment alternatives to glucocorticoids have suggested benefit.8,9 However, little longer-term data from the current era are available to inform counseling of patients with NS regarding clinical course and anticipated prognosis. Sarcoidosis is a multisystem inflammatory disorder characterized by nonnecrotizing granulomas. Clinical expression of the disease tends to relate to the anatomical locations of granulomatous infiltration, together in some cases with systemic symptoms. Distinctive, well-described clinical and imaging findings are associated with sarcoidosis, especially pulmonary manifestations.10 Infiltration of the nervous system by inflammatory granulomas—NS—is thought to affect more than 5% of people with sarcoidosis and is an important cause of morbidity and mortality associated with the disease.6,7 Although substantial regional variability remains in incidence and prevalence, sarcoidosis has been described worldwide11 and has a remarkably high prevalence among some populations in the United States (including African American individuals)12,13 and Europe (including Scandinavians).14 Sarcoidosis exhibits complex genetic heritability, and nearly all susceptibility genes identified to date relate to well-known immunologically relevant pathways.15 In this issue of JAMA Neurology, Joubert and colleagues16 present a unique data set to help inform NS prognosis by analyzing survival, relapse risk, and functional status in 234 patients undergoing evaluation at a single referral center (Hôpital La Pitié-Salpêtrière, in Paris, France) from 1990 through 2015. The median age at NS diagnosis was 42 years (interquartile range, 32 to 53 years), 117 patients (50.0%) were women, 125 (53.4%) were white, 58 (24.8%) were African or Caribbean, and 45 (19.2%) were North African; the median follow-up was 8 years (range, 2 months to 23 years). Although most patients in the cohort had central nervous system (CNS) sarcoidosis, 24 (10.3%) had peripheral nervous system (PNS) disease and 19 (8.1%) had muscle manifestations. Only 18 patients (7.8%) in the cohort had sarcoidosis isolated to the nervous system, whereas 55 (23.5%) had evidence of sarcoidosis in more than 3 organ systems, findings suggesting the multiorgan potential and systemic breadth of the disease. The overall 10-year survival among patients with neurosarcoidosis was 89% (95% CI, 84%-94%), with death attributed directly to NS in 6 cases (2.6%) and sepsis in 3 cases (1.3%). Five years after diagnosis, 17 of 160 patients (10.6%) exhibited worsened neurologic disability as defined by the Expanded Disability Status Scale (EDSS) score compared with their baseline at cohort entry. Sarcoidosis relapses in any organ system, which the investigators defined permissively as “the occurrence of any new neurologic or nonneurologic symptom or biological abnormality (eg, urinary sediment or abnormal liver function test result) attributed to sarcoidosis by the patient’s referring physician,”16 were common, occurring in 136 patients (58.1%) during follow-up, with a 10-year relapse-free survival of only 14% (95% CI, 9%-22%). Treatment with immunosuppression was associated with a lower hazard of sarcoidosis relapse of any kind. With glucocorticoids (254 therapeutic sequences), the hazard ratio (HR) of relapse was 0.59 (95% CI, 0.39-0.90); weekly oral methotrexate sodium (125 therapeutic sequences), 0.48 (95% CI, 0.290.79); intravenous pulse cyclophosphamide (120 therapeutic sequences), 0.18 (95% CI, 0.09-0.37); hydroxychloroquine (59 therapeutic sequences), 0.29 (95% CI, 0.14-0.63); and infliximab (28 therapeutic sequences), 0.31 (95% CI, 0.11-0.82). Treatment with mycophenolic acid (also known as mycophenolate; 64 therapeutic sequences) was not associated with reduced relapse risk (HR, 0.67; 95% CI, 0.37-1.23). With only 14 therapeutic sequences, the analysis of azathioprine sodium was underpowered to reach any meaningful conclusion about efficacy. Any comparison of relative efficacy of these different agents may be confounded by indication. When restricting the analysis to neurologic relapses, the 10-year neurologic relapse-free survival was low at 28% (95% CI, 20%-38%), and at least 1 neurologic relapse (or death) occurred in 72% of all patients with NS. Treatment with intraveRelated article Opinion