Characteristics, outcome and treatments with cranial pachymeningitis

Abstract The aim of this study was to determine the characteristics, treatment, and outcome according to each etiology of pachymeningitis. We conducted a retrospective multicenter French nationwide study between 2000 and 2016 to describe the characteristics, outcome, and treatment of pachymeningitis. We included 60 patients (median age 55.5 years; interquartile range [IQR] 30–80, female/male ratio 0.43). Neurologic signs were present in 59 patients (98%) and consisted of headache in 43 (72%), cranial nerve palsy in 33 (55%), confusion in 10 (17%), seizures in 7 (12%), and focal neurologic signs in 9 (15%). Fever and weight loss were present in 8 (13%) and 13 cases (22%), respectively. Cerebral venous thrombosis was present in 8 cases (13%). Analysis of cerebrospinal fluid showed moderate hyperproteinorachia (median 0.68 g/L; IQR 0.46–3.2) with or without pleiocytosis. Diagnosis included idiopathic pachymeningitis (n = 18; 30%); granulomatosis with polyangiitis (n = 13; 17%); Erdheim-Chester disease (n = 10; 17%); IgG4-related disease and tuberculosis (n = 3; 5% each); Rosai-Dofman disease, microscopic polyangiitis, and sarcoidosis (n = 2, 3% each); cryptococcal meningitis, Lyme disease, ear-nose-throat infection, postlumbar puncture, low spinal-fluid pressure syndrome, and lymphoma (n = 1 each). We found no difference in demographics and neurologic presentation among idiopathic pachymeningitis, Erdheim-Chester disease, and granulomatosis with polyangiitis. In contrast, frequencies were lower with idiopathic pachymeningitis than Erdheim-Chester disease for general signs (6% and 40%, respectively, P = .041) and complete neurologic response (0% vs 39%, P = .045). The detection of extraneurologic signs and routine screening are needed to classify the pachymeningitis origin. Prospective studies are warranted to determine the best treatment in each case.


Introduction
Pachymeningitis is a rare inflammatory disease characterized by local or diffuse spinal dura thickening, which could be a feature of various conditions. [1] Clinical manifestations depend on the location of inflammatory lesions and compression of adjacent nervous structures. Clinical and radiologic signs are not specific to pachymeningitis and the presence of extraneurologic features as well as immunologic, histologic, and infectious work-up could help to determine the underlying etiology. Among the main associated diseases are the infectious diseases such as tuberculosis, syphilis, cryptococcal infection, and Lyme disease; autoimmune or inflammatory diseases such as granulomatous with polyangiitis (GPA), sarcoidosis, and recently described IgG4related disease; and malignancies, in particular lymphoma. [1,2] Several case series of pachymeningitis have been reported, but large series comparing characteristics by etiologies and outcome are still lacking. [1,[3][4][5] Early diagnosis and treatment are crucial in preventing neurologic damage. Treatment mostly includes steroids or immunosuppressive therapy, but the best treatment regimen and outcome related to the various underlying situations are not determined.
In this nationwide study in France, we report the characteristics, treatment, and outcome of 60 patients with pachymeningitis and compare the profile and outcome of pachymeningitis with different etiologies.

Patients and methods
We conducted a retrospective study of patients from 17 referral centers in the "Société National Francaise de Médecine Interne" (SNFMI) network in France between March 2010 and April 2016. Physicians were asked by SNFMI to declare all cases with pachymeningitis during 12 months and data were collected retrospectively from the physicians in charge of the patients. Pachymeningitis was defined as focal or diffuse thickening of dura mater visualized on magnetic resonance imaging (MRI) of the brain and/or histologic analysis of dura mater consistent with persistent inflammation.
Clinical, laboratory, and imaging data were collected, as were data on treatments at baseline, at 6, 12 months, and at last available visit. Clinical data recorded included the presence of general symptoms; noninfectious fever; impaired lung, ear-nosethroat (ENT), kidney, nervous system, skin, joint, eye, and heart; and presence of venous/arterial thrombosis. Neurologic impairment data collected included headaches, localized neurologic signs, cranial nerve involvement, confusion, epilepsy, and vestibular and cerebellar impairment. Laboratory data recorded included proteinorachia, cerebrospinal fluid (CSF) cell numbers and type, antinuclear antibodies, antineutrophil cytoplasmic antibodies (ANCAs), ECA, syphilis serology, BK cultures from different sites if available, cryptococcal infection analysis, Borrelia analysis, HIV and human T-cell lymphotropic virus-1 serologies. Histology findings from biopsies were recorded if available, as were IgG4 immunostaining data and blood levels. Data for routine laboratory variables of disease activity, including erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) levels were collected. The presence of parenchymal impairment and ventricular dilatation were recorded on MRI.
The underlying disease was diagnosed by using international criteria (ie, American College of Rheumatology criteria for GPA), and idiopathic pachymeningitis was considered in the absence of associated disease. Steroid amount was analyzed at the initiation of each new treatment regimen and during follow-up. Neurologic treatment response was defined as complete with disappearance of all signs (clinical and/or CSF if realized) present at baseline, partial with more than 50% improvement, and nonresponse for all remaining cases. Treatment response was analyzed for extraneurologic impairments. Radiologic response was defined as complete disappearance of pachymeningitis, partial response with at least 50% decrease of thickening, and no response for all remaining cases. Steroid dependence was defined as prednisoneequivalent amount >20 mg/d for at least 2 months. Relapse was defined as reappearance of neurologic and/or extraneurologic signs after at least 3 months' remission. The follow-up was considered from the diagnosis time to the last available news.
The study was performed in accordance with the ethical standards of the Helsinki Declaration and was approved by an institutional review board (Comité de Protection des Personnes, Aulnay sous Bois, Ile de France X).

Statistical analysis
Data are described with mean ± standard deviation and median (interquartile range [IQR]) for continuous variables and frequencies (%) for categorical variables. Results are expressed as observed data (missing data not replaced) to account for nonavailable data and percentages are calculated take into account the overall number of available cases. Chi-squared or Fisher exact tests were used to compare categorical variables and Mann-Whitney U or Student t test to compare continuous variables. All tests were 2 sided and P < .05 was considered statistically significant. Statistical analyses involved use of GraphPad v3.1.0.
With Erdheim-Chester disease, at 6 and 12 months, neurologic partial or complete response was found for 1 (20%) and 0 and 1 (20%) and 0 patients, respectively. MRI partial response was noted in 3/5 patients (60%). Three of 5 patients at the last visit showed neurologic partial or complete response, with MRI partial response in 4/5 (80%).

Mekinian et al. Medicine (2018) 97:30 Medicine
Relapse-free survival significantly differed among these 3 diseases, with median time to relapse to 2.7 years in patients with GPA and 10.2 years in idiopathic pachymeningitis and was not reached in Erdheim-Chester disease (log rank P < .05) (Fig. 1).

Discussion
We report here one of the largest series of pachymeningitis including descriptions and for the first time the comparison of features, response rates, and outcome of several etiologies of pachymeningitis. Idiopathic pachymeningitis represented the most frequent situation, and several extraneurologic signs, particularly ENT and ocular involvements, should be screened and can help pinpoint the etiology. Acute-phase reactants, immunologic data, serum IgG4 dosage, and infectious work-up are needed to prevent misdiagnosing the underlying disease, in particular with only isolated neurologic signs. The outcome is still challenging, because 30% of patients have persistent signs despite treatment. In this study, we show the particular different relapsefree survival according the underlying disease.
In our case series, one of more common entities remained idiopathic pachymeningitis (30%), which is similar to previous data. [3,4] A survey in Japan revealed idiopathic pachymeningitis in 44% of cases, followed by ANCA-associated vasculitis and in particular GPA. [4] Idiopathic pachymeningitis is a challenging diagnosis, and its characteristics are similar to an emerging inflammatory condition, the IgG4-related syndrome. Only a few data are available to determine the part of IgG4-related syndrome among "idiopathic pachymeningitis." [2,6,7] Wallace et al performed IgG4 immunostaining for 14 cases of idiopathic pachymeningitis and reclassified 4 cases (29%) as IgG4 syndrome, [7] which emphazises the importance of this analysis in all cases of idiopathic pachymeningitis, even isolated cerebral forms.
Few data are available concerning pachymeningitis in GPA and other ANCA-associated vasculitis. [3,[8][9][10][11][12] One of the recent important series analyzed central nervous system involvement in GPA and found a 46% prevalence of pachymeningitis, defining a granulomatous pattern. [11] These patients commonly have headaches and more frequently myeloperoxidase-ANCA, which differs from our series, 64% with PR3 specificity. In our GPA patients, pachymeningitis was frequently associated with orbital pseudotumor, tracheal stenosis, and rhinosinusitis and less frequently, a kidney and lung pattern. In Erdheim-Chester disease, 10% to 25% of patients had neurologic involvement, a factor associated with disease severity. [13][14][15] The neurologic involvement is rarely isolated, and the presence of extraneurologic features could help in the diagnosis.
Few studies have compared the clinical features and outcome of patients with idiopathic pachymeningitis and other pachymeningitis diseases. Among 28 patients, 20 had idiopathic pachymeningitis and no specific neurologic, CSF data or MRI features in comparison to secondary pachymeningitis. [16] Another study comparing idiopathic pachymeninigitis to diseases related to ANCA-associated vasculitis and IgG4 syndrome showed a predominant male ratio in idiopathic forms and older age in GPA patients. [4] In our study, the presence of general signs, extraneurologic signs, high CSF cell number and acute-phase reactants allowed for discriminating idiopathic disease from other etiologies.
Our study contains several limitations. All consecutive patients could not be included in this study, and thus the frequency of various pathologies is not representative of their real prevalence. As well, many patients were selected from internal medicine departments, which could underrepresent some etiologies such as infectious disease. The few cases of pathologies such as sarcoidosis, paraneoplastic, and infectious diseases did not allow us to evaluate the features and outcome of pachymeningitis associated with these disorders. In idiopathic pachymeningitis, IgG4 serum levels and IgG4+ plasmocyte analysis were not always available, because some cases were diagnosed before this analysis could be performed and thus could not be excluded as misdiagnosis. The absence of IgG4 screening in all cases could underestimate the real prevalence of this entity in the subset of pachymeningitis.

Conclusion
The characteristics and the etiologies of pachymeningitis need to be carefully analyzed. Routine screening and detection of extraneurologic signs are needed to classify the pachymeningitis origin. Prospective studies are needed to determine the best treatment in each case.