Neurologic Manifestations in Primary Sjögren Syndrome

A Study of 82 Patients

Abstract

Neurologic involvement occurs in approximately 20% of patients with primary Sjögren syndrome (SS). However, the diagnosis of SS with neurologic involvement is sometimes difficult, and central nervous system (CNS) manifestations have been described rarely. We conducted the current study to describe the clinical and laboratory features of SS patients with neurologic manifestations and to report their clinical outcome.

We retrospectively studied 82 patients (65 women and 17 men) with neurologic manifestations associated with primary SS, as defined by the 2002 American-European criteria. The mean age at neurologic onset was 53 years. Neurologic involvement frequently preceded the diagnosis of SS (81% of patients). Fifty-six patients had CNS disorders, which were mostly focal or multifocal. Twenty-nine patients had spinal cord involvement (acute myelopathy [n = 12], chronic myelopathy [n = 16], or motor neuron disease [n = 1]). Thirty-three patients had brain involvement and 13 patients had optic neuropathy. The disease mimicked relapsing-remitting multiple sclerosis (MS) in 10 patients and primary progressive MS in 13 patients. We also recorded diffuse CNS symptoms: some of the patients presented seizures (n = 7), cognitive dysfunction (n = 9), and encephalopathy (n = 2). Fifty-one patients had peripheral nervous system involvement (PNS). Symmetric axonal sensorimotor polyneuropathy with a predominance of sensory symptoms or pure sensory neuropathy occurred most frequently (n = 28), followed by cranial nerve involvement affecting trigeminal, facial, or cochlear nerves (n = 16). Multiple mononeuropathy (n = 7), myositis (n = 2), and polyradiculoneuropathy (n = 1) were also observed. Thirty percent of patients (all with CNS involvement) had oligoclonal bands. Visual evoked potentials were abnormal in 61% of the patients tested. Fifty-eight patients had magnetic resonance imaging (MRI) of the brain. Of these, 70% presented white matter lesions and 40% met the radiologic criteria for MS. Thirty-nine patients had a spinal cord MRI. Abnormalities were observed only in patients with spinal cord involvement. Among the 29 patients with myelopathy, 75% had T2-weighted hyperintensities. Patients with PNS manifestations had frequent extraglandular complications of SS. Anti-Ro/SSA or anti-La/SSB antibodies were detected in 21% of patients at the diagnosis of SS and in 43% of patients during the follow-up (mean follow-up, 10 yr). Biologic abnormalities were more frequently observed in patients with PNS involvement than in those with CNS involvement (p < 0.01). Fifty-two percent of patients had severe disability, and were more likely to have CNS involvement than PNS involvement (p < 0.001). Treatment by cyclophosphamide allowed a partial recovery or stabilization in patients with myelopathy (92%) or multiple mononeuropathy (100%).

The current study underlines the diversity of neurologic complications of SS. The frequency of neurologic manifestations revealing SS and of negative biologic features, especially in the event of CNS involvement, could explain why SS is frequently misdiagnosed. Screening for SS should be systematically performed in cases of acute or chronic myelopathy, axonal sensorimotor neuropathy, or cranial nerve involvement. The outcome is frequently severe, especially in patients with CNS involvement. Our study also underlines the efficacy of cyclophosphamide in myelopathy and multiple neuropathy occurring during SS.

INTRODUCTION

Sjögren syndrome (SS) is an autoimmune disease that may be primary or secondary to another connective tissue disease (mainly systemic lupus erythematosus, rheumatoid arthritis, or scleroderma). This syndrome is characterized by mononuclear infiltration and destruction of salivary and lachrymal glands leading to xerostomia and xerophthalmia. Similar mononuclear infiltrates invading visceral organs or vasculitic lesions can give extraglandular manifestations¹⁸. Neurologic involvement has been reported in primary SS since its initial clinical description by Sjögren in 1935⁴⁵. After 1980, Alexander and coworkers^1,6 suggested in a series of publications that neurologic manifestations could be observed in approximately 20%-25% of cases of SS. However, the exact prevalence remains controversial²¹. Neurologic symptoms can affect the peripheral nervous system (PNS) and the central nervous system (CNS)³⁰. Although PNS involvement is relatively well documented and is reported in about 10%-20% of SS patients, CNS manifestations are still a matter of discussion. The differences between the results reported in previous studies could be due to the use of heterogeneous diagnostic criteria and to variability in the origin of the patients. In the present study, we report 82 consecutive patients with neurologic manifestations associated with primary SS as defined by the American-European criteria⁵² who were referred to the neurologic and internal medicine departments of our hospital. The aim of the study was to describe the neurologic, laboratory, and magnetic resonance imaging (MRI) patterns of these patients and their clinical outcome.

PATIENTS AND METHODS

We retrospectively studied 82 consecutive patients with primary SS referred to the neurologic or internal medicine departments of Lille University Hospital between January 1993 and December 2001. All patients fulfilled the American-European criteria for SS⁵² and had neurologic manifestations during the follow-up. Our neurologic department specializes in inflammatory (MS, optic neuritis, myelopathy) and neuromuscular disorders (peripheral neuropathy and myopathy). Apart from a specific study¹³ on the primary progressive forms of MS, only those patients with sicca syndrome or with neurologic symptoms suggesting SS are screened for SS.

Diagnosis of Primary Sjögren Syndrome

All patients with secondary SS were excluded (for example, patients with systemic lupus erythematosus or rheumatoid polyarthritis). Clinical symptoms of sicca complex, namely xerostomia, xerophthalmia or recurrent salivary enlargement, were evaluated. Ocular involvement was documented by Schirmer test (abnormal if less than 5 mm of the filter paper was moistened in 5 minutes) or Rose Bengal score (keratoconjunctivitis if score was >4 according to the Van Bijsterveld scoring system)⁵⁰. Xerostomia was confirmed by an abnormal salivary scintigraphy⁴⁴ or unstimulated salivary flow. Biopsy samples of the minor salivary glands were suggestive of SS if the lymphocytic focus score was equal to or greater than an aggregate of 50 mononuclear cells in 4 mm² of glandular tissue or if there was an abnormal Chisholm score (higher than 2)¹². Screening for autoantibodies to Ro/SSA and La/SSB was systematically performed by Ouchterlony double gel immunodiffusion and by western blotting technique in some cases. Screening for other connective tissue diseases was systematically performed to eliminate all but primary SS. Exclusion criteria were the following: past head and neck radiation treatment, hepatitis C infection, acquired immunodeficiency disease, preexisting lymphoma, sarcoidosis, graft versus host disease, use of anticholinergic drugs or other drugs including sicca symptoms (within a period of less than 4 times the half-life of the drug). All patients fulfilled at least 4 of the 6 consensus criteria, including minor salivary gland histopathology abnormalities or anti SSA/SSB, or 3 of the 4 objective criteria⁵².

Clinical Findings

Demographic characteristics and the following clinical data concerning SS were studied: age at onset of SS, time to diagnosis of SS, extraglandular manifestations (lung, articular, skin, hematologic, or renal involvement, Raynaud phenomenon, autoimmune thyroiditis) at diagnosis and during follow-up, age at neurologic involvement, and delay before onset of sicca symptoms. Neurologic data were also described: symptoms at onset, type of manifestations, course, treatment, and outcome.

Neurologic Outcome

Clinical outcome was assessed using the Modified Oxford Handicap Scale (MOHS)⁹:

• 0 = No symptoms.
• 1 = Minor symptoms not interfering with lifestyle.
• 2 = Minor handicap with symptoms leading to some restriction in lifestyle but not interfering with the patient's capacity to look after himself or herself.
• 3 = Moderate handicap with symptoms that significantly restrict lifestyle and prevent totally independent existence.
• 4 = Moderately severe handicap with symptoms that clearly prevent independent existence though not needing constant attention.
• 5 = Severe handicap leading to total dependence and requiring constant attention during night and day.
• 6 = Death.

We evaluated patients on this scale at the onset and at the end of follow-up, at the onset and at the end of immunosuppressive treatment, and at the time of most severe disability. At the end of the follow-up, patients were classified into 2 groups corresponding to "Minor disability" (MOHS score ≤2) and "Severe disability" (MOHS >2), respectively, to look for prognostic factors.

For the patients with spinal cord involvement, clinical outcome was also evaluated using the Lipton scale³². These patients were classified into 3 subgroups, corresponding to good, fair, or poor prognosis. The good prognosis subgroup comprised patients with normal walking with no, or only mild, neurologic examination abnormalities or mild sphincter dysfunction. The fair prognosis subgroup comprised patients who required unilateral or bilateral help to walk more than 100 meters and had persistent mild sensory or sphincter dysfunction. The poor prognosis subgroup comprised patients who were unable to walk 100 meters even with help and who had severe sensory or sphincter dysfunction.

Other Findings

Depending on the clinical findings, patients underwent brain or spinal cord magnetic resonance imaging (MRI), visual evoked potentials (VEP), cerebrospinal fluid (CSF) analysis, nerve conduction studies, electromyogram (EMG), nerve or muscular biopsy, or electroencephalogram.

Laboratory Data

The following parameters were evaluated: complete blood count, erythrocyte sedimentation rate, serum cryoglobulins, total serum gammaglobulins, serum protein immunoelectrophoresis, C-reactive protein, total hemolytic complement and complement factors (C3 and C4), antinuclear antibodies (ANA), precipitating antibodies to the extractable antigens (ENA), Ro/SSA, La/SSB, RNP and Sm, antinative DNA, antineutrophil cytoplasmic antibodies (ANCA), rheumatoid factor (RF), anticardiolipin antibodies, antiprothrombinase, human immunodeficiency virus (HIV), hepatitis C and hepatitis B serology. ANA were tested by indirect immunofluorescence on HEP-2 cell line as substrate, and were considered positive at a dilution higher than or equal to 1:80. The presence of anti-Ro/SSA, anti-La/SSB, and ENA antibodies was evaluated by Ouchterlony double gel immunodiffusion (all patients) and by western blotting in the most recent cases (39 patients).

Statistical Analysis

Values of quantitative variables were expressed as mean and range. Values of qualitative variables were expressed as a percentage. Statistical analysis was performed using the SAS program (version 8.2, SAS Institute Inc., Cary, NC). The chi-squared test was applied to analyze qualitative differences. To compare quantitative parameters, the Student t test was used for large samples of similar variance and the nonparametric Wilcoxon or Kruskall-Wallis tests were used for other samples. Statistical significance was established at p < 0.05. When several variables appeared to have statistical significance using univariate analysis, an unconditional logistic regression analysis was performed by multivariable analysis to rule out possible confounding variables. For small samples, a simple description was preferred.

RESULTS

Demographic Features

The cohort consisted of 65 (79%) women and 17 (21%) men (ratio of women to men: 3.8:1). The mean age at onset of SS symptoms was 48.6 years (range, 15-79 yr). The mean age at the time of diagnosis of SS was 55.1 years (range, 28-80 yr). The mean age at onset of neurologic manifestations was 53.1 years (range, 19-79 yr). The mean duration of the disease at the time of the study was 10.5 years.

Glandular Features (Table 1

TABLE 1: Demographic and Glandular Features of 82 Patients With Sjögren Syndrome

Seventy-eight (95%) patients presented clinical xerostomia. For 35 patients it was the first symptom of SS. For the remaining 43 patients, the mean delay between onset of SS and xerostomia was 6.2 years. Twenty patients (24%) had recurrent salivary gland enlargement. Objective xerostomia was observed in 43 (67%) of the 64 patients tested (abnormal salivary scintigraphy or unstimulated salivary flow). At the time of diagnosis of SS, minor salivary gland biopsies showed lymphocytic infiltration in 78 (95%) of the 82 patients, fulfilled grade 3 of Chisholm in 57 patients, and grade 4 in 21 patients. Seventy (85%) patients had complaints of xerophthalmia, 24 of these at the onset of the disease. Objective xerophthalmia was observed in 70 (90%) of the 78 patients tested (Schirmer test and/or Rose Bengal). Xerophthalmia was asymptomatic in 8 cases. Decreased tear production was documented by Schirmer test in 64 (84%) of the 76 tested patients. The Rose Bengal test was characteristic of keratoconjunctivitis in 17 (57%) of the 30 tested patients. Four patients had abnormal Rose Bengal test and normal Schirmer test results.

First Clinical Symptoms at the Onset of SS (Table 2)

TABLE 2: Clinical Manifestations at the Onset of SS

The different presenting symptoms of SS are shown in Table 2. Only 36 (44%) patients had sicca symptoms as the first manifestation. Neurologic complications were the first symptom in 47 (57%) patients, and occurred in isolation in 26 patients. For the remaining 35 patients, the mean delay before neurologic involvement was 8.3 years.

Temporal Relationship Between Onset of Sicca Symptoms and Onset of Neurologic Features

Sicca symptoms occurred before the onset of neurologic features in 31 (38%) patients. Twelve (15%) patients developed neurologic complications at the same time as sicca symptoms. Thirty-nine (47%) patients had neurologic features before the onset of sicca symptoms (mean delay before xerostomia and/or xerophthalmia, 6 yr).

Diagnosis of SS

The mean duration between onset of SS symptoms and diagnosis of SS was 6 years. SS was diagnosed after the onset of neurologic features in 66 (81%) patients, with a mean delay of 4.6 years. In 61 of these patients, SS was suggested following neurologic symptoms. The delay between onset of neurologic symptoms and SS diagnosis was significantly shorter when sicca symptoms were present at the time of neurologic involvement than in patients without sicca symptoms (3 and 5.2 yr, respectively, p = 0.04).

Extraglandular Features (Table 3

TABLE 3: Extraglandular Manifestations During Follow-Up of 82 Patients With Neurologic Complications of SS

Seventy-one (86%) patients had extraglandular features other than neurologic manifestations. Fifty-two (63%) patients had more than 1 extraneurologic complication. The main extraglandular features are shown in Table 3. Extraneurologic complications were significantly more frequent in patients with PNS involvement than in those with CNS involvement (Figure 1).

FIGURE 1: Systemic extraneurologic complications of SS. *p < 0.05; **p < 0.005; CNS, central nervous system; PNS, peripheral nervous system.

Neurologic Features (Table 4

TABLE 4: Neurologic Symptoms*

The main neurologic features are presented in Table 4. Fifty-six (68%) patients had CNS involvement and 51 (62%) patients had PNS involvement. During the follow-up, 25 (30%) patients had both CNS and PNS involvement, 31 (38%) had isolated CNS complications, and 26 (32%) had isolated PNS features. There was no significant difference between age at onset of PNS complications and age at onset of CNS complications.

Neurologic symptoms occurring at onset of SS involved the CNS more frequently than the PNS (p = 0.005). PNS manifestations preceded sicca symptoms far less frequently than did CNS manifestations (p = 0.02). Therefore, the delay between onset of neurologic features and diagnosis of SS was shorter in patients with PNS involvement than in CNS involvement, but the difference did not reach significance.

CNS Manifestations

Fifty-six (68%) patients had CNS disorders, which were mostly focal or multifocal. Central neurologic involvement preceded the diagnosis of SS in 45 patients.

Spinal Cord Involvement

Twenty-nine patients (22 women and 7 men) had spinal cord involvement, 1 of whom had motor neuron disease. For 17 (61%) of these patients, spinal cord involvement was the first symptom of SS. Twelve (41%) patients had acute myelopathy. Five of them had severe transverse myelitis resulting in tetraparesis or paraparesis and sphincter dysfunction, 2 had moderate subacute transverse myelitis, 1 had lateral cervical myelitis resulting in hemiplegia, and 4 had mild posterior myelopathy with a predominance of sensory symptoms and a recurrent course. Sixteen (55%) had chronic myelopathy, which was frequently revealed by progressive paraparesia. Two patients had no sensory symptoms during follow-up, with a course suggesting lateral primitive sclerosis. One patient had progressive myeloradiculitis. The course of the disease mimicked primary progressive multiple sclerosis in 13 patients. One patient had clinical signs of motor neuron disease, which was confirmed by electrophysiologic studies. The course was similar to that of amyotrophic lateral sclerosis except for the occurrence of bladder dysfunction.

Brain Involvement

Thirty-eight (46%) patients had diffuse or focal brain involvement. Focal involvement occurred in 33 patients. The various focal neurologic symptoms are shown in Figure 2. The course was acute (stroke-like episodes) in 13 patients. Ten patients (women, mean age at onset: 40 yr) had recurrent deficits mimicking relapsing-remitting multiple sclerosis, including brain, spinal cord, and optic features. The delay between onset and diagnosis of SS was significantly longer (9.8 yr, p = 0.002) for these 10 patients, and sicca symptoms were often absent at the onset of neurologic features (60%, mean delay before onset of sicca symptoms: 10 yr).

FIGURE 2: Focal brain symptoms.

Nine patients had severe cognitive impairment. Two of them had a whole psychometric evaluation that confirmed subcortical dementia. Two patients complained of memory deficits but were normal on psychometric evaluation. Eight patients also had focal brain deficits, 2 of whom had vascular lesions resulting in apraxia. Seven patients had seizures, which were partial in 4 patients (3 patients with temporal and 1 with parietal seizures) and generalized in 3 patients. Finally, aseptic meningoencephalitis occurred in 1 patient associated with cranial nerve palsy (cochlear nerve) and cerebellar involvement. One patient had recurrent encephalitis.

Optic Neuritis

Thirteen patients had visual loss consecutive to optic neuritis. Myelopathy occurred at the same time in 2 patients, suggesting neuromyelitis optica. One patient had severe optic neuritis leading to blindness.

PNS Manifestations (Figure 3)

FIGURE 3: Peripheral nervous system involvement.

Fifty-one patients had PNS involvement, as described below.

Peripheral Neuropathies

Peripheral neuropathies were the most frequent neurologic finding (36 patients). Sixty-seven (82%) patients in our study had an EMG, which detected neuropathy in 54% of them. In 5 patients the neuropathy was asymptomatic. Based on the clinical and EMG features, 19 (53%) patients had a distal axonal sensory-motor neuropathy, 9 (25%) had pure sensory neuropathy, 7 (19%) had multiple mononeuropathy, and 1 (3%) had chronic polyradiculoneuropathy.

Sensory motor neuropathies were thus the most frequent type of neuropathies, with predominantly sensory symptoms. Autonomic symptoms were observed in only 3 patients (bladder dysfunction) but screening for them was not systematically performed. A sural nerve biopsy was performed in 5 patients, revealing lymphocytic infiltration in 1 patient.

Among the pure sensory neuropathies, 4 patients had clinical and electromyographic features of ganglionopathy with severe ataxia. The remaining 5 patients had symmetric distal symptoms predominantly in the legs with a pure axonal sensory neuropathy on EMG. Four of them had autonomic nervous system involvement (sphincter involvement in 3 patients and blood pressure dysregulation in 1 patient). A sural nerve biopsy (performed in 2 cases) showed lymphocytic nonnecrotizing vasculitis.

Multiple mononeuropathy frequently occurred in patients with severe systemic complications of SS (for example, cutaneous vasculitis in 5 patients, Raynaud phenomenon in 6 patients, arthralgia). The onset was mainly acute or subacute with recurrent symptoms. Sural nerve biopsies were performed in 2 patients and revealed lymphocytic infiltrates without necrotizing vasculitis.

Cranial Nerve Involvement

The cochlear nerve was impaired in 6 patients, revealed by hearing loss and vestibular symptoms in 2 patients. Five patients had sensory trigeminal nerve involvement, and 4 patients had regressive facial nerve involvement. Two patients had a loss of olfaction, probably due to first cranial nerve involvement.

Other Neuromuscular Disorders

Two patients had EMG-confirmed muscular involvement, mimicking myositis, with normal muscular biopsy. One patient had myeloradiculitis and 1 patient had bilateral L5 involvement.

CSF, VEP, and MRI Results

CSF and VEP Analysis

Fifty-four (66%) patients underwent CSF analysis. The cell count was increased in 16 (30%) patients, revealing aseptic lymphoid meningitis. Thirty percent of patients (all with CNS involvement) had oligoclonal bands. Only 1 (14%) of the 7 patients tested with isolated PNS involvement had CSF abnormalities (increased cell count). VEP were abnormal in 61% of the patients tested. They confirmed optic neuritis in the 13 patients with a history of visual loss and defined 12 additional patients with subclinical optic neuritis.

Brain MRI

Fifty-eight (71%) patients had a brain MRI. We observed white matter lesions in 70% of these patients. Radiologic findings suggesting MS were observed in 40% of patients (Figure 4). We also observed gray matter lesions in the basal ganglia (17% of patients). Corpus callosum lesions were rarely observed (14%). Lesions were found more frequently in patients with CNS involvement (80%) than in patients without CNS involvement (25%; p = 0.008).

FIGURE 4: Brain MRI on FLAIR showing white matter lesions and severe atrophy suggestive of but not specific to multiple sclerosis in a patient with relapsing-remitting symptoms.

Spinal Cord MRI

Thirty-nine (47.5%) patients had spinal cord MRI, which showed T2-weighted hyperintensities in 49% of cases, all with clinical spinal cord involvement. MRI was always normal in patients without spinal cord involvement. Among the 29 patients with myelopathy, 75% had T2-weighted hyperintensities (cervical in 82%, dorsal in 47%, and lumbar in 12%), 35% with extended lesions, 65% with a single lesion, and 40% with centromedullar lesions. In cases of acute myelopathy we frequently observed an extended T2-weighted hyperintensity involving a large part of the cord (Figure 5).

FIGURE 5: Spinal cord MRI on T2-weighted images showing an extended hypersignal in the cord in a patient with acute myelopathy.

Laboratory Findings

The main laboratory findings during follow-up are shown in Table 5. Several abnormalities (for example, lymphopenia, hypergammaglobulinemia, ANA, cryoglobulins) were more frequent in cases with PNS involvement than in those with CNS involvement. None of the patients was positive for anticardiolipin, hepatitis C, or HIV antibodies.

TABLE 5: Laboratory Findings

ANA

ANA were detected by immunofluorescence with significant titer levels in 44 patients (53.7%). Anti-Ro/SSA or anti-La/SSB antibodies were detected in 17 (21%) patients at the time of diagnosis of SS and in 35 (43%) patients during follow-up. Ten (12%) patients had anti-Ro/SSA and anti-La/SSB antibodies, 1 patient had only anti-La/SSB antibodies, and 24 patients had only anti-Ro/SSA antibodies. Immunoblotting was more sensitive (46%) than Ouchterlony immunodiffusion (26%). Anti-Ro/SSA antibodies were not more frequent in patients with PNS involvement than in those with CNS involvement, but there was a difference in sensitivity depending on the technique used (see Figure 6), suggesting that immunoblotting is more effective, especially in patients with CNS involvement. None of the patients was positive for double-stranded DNA or Sm antibodies.

FIGURE 6: Detection of antiRo/SSA according to the technique used. *p < 0.05; SSAWB, positive detection of anti-Ro/SSA by immunoblotting; SSAOuch, positive detection of anti-Ro/SSA by Ouchterlony immunodiffusion; SSA WB, Ouc-, positive detection of anti-Ro/SSA by immunoblotting, negative detection by Ouchterlony immunodiffusion; PNS, peripheral nervous system; CNS, central nervous system; All, all patients in the study (n = 82).

Cryoglobulinemia

Cryoglobulinemia was detected in 30 (36.6%) patients. It was oligoclonal (type 2) in 12 of these patients, all with PNS involvement. Table 6 shows the frequency of cryoglobulinemia according to the type of neuropathy.

TABLE 6: Cryoglobulins and Type of Neuropathy

Outcome Profiles

The mean neurologic follow-up was 7 years. Eight patients died, 6 of them from causes unrelated to SS. Twenty-six patients had a relapsing disease course. At the end of the follow-up, the mean MOHS disability was 2.8. Fifty-two percent of patients had severe disability. Severe disability was more frequent in cases of CNS involvement (78%) than in PNS involvement (26%, p < 0.001). We did not find any other significant prognostic factor.

Seventy-three patients were treated with corticosteroids, resulting in a durable neurologic amelioration or stabilization in 29 (45%) cases. This treatment was totally ineffective in 13 patients, mostly those with polyneuropathies. Thirty-four patients received immunosuppressive therapy, including cyclophosphamide in 20 patients. Cyclophosphamide was administered monthly (700 mg/m² intravenous) for 6 or 12 months. It resulted in a partial recovery or stabilization in 11 of the 12 (92%) treated patients with spinal cord involvement and in all of the cases with multiple mononeuropathy.

DISCUSSION

To our knowledge, this study is the largest to date on SS patients with neurologic manifestations. While it does not allow us to determine the frequency of neurologic complications, it underlines the diversity of neurologic manifestations of SS and suggests that the frequency of CNS involvement is similar to that of PNS involvement. This result is quite different from that of previous studies^{10,15,18,20,26,27,35,36,39,40,43,51} (Table 7), where PNS involvement was mainly described. Indeed, the prevalence of CNS involvement during SS is highly heterogeneous depending on the different diagnostic criteria of SS, the definition of neurologic involvement, and the selection of patients from neurologic, internal medicine, or rheumatology departments^21,30. The twin sources of our patients (neurologic and internal medicine departments) may well have helped to reduce a potential bias in the proportion of SS patients with PNS involvement and those with CNS involvement.

TABLE 7: Prevalence of Neurologic Manifestations in SS, Previous Studies*

Our study also underlines the variety of the CNS manifestations that affected 56 of our patients. Previous studies on CNS features in SS usually involved fewer than 20 patients^2,16,39. In our cohort, CNS involvement was most frequently focal (brain, spinal cord, and optic nerve), as previously described^1,39. Spinal cord involvement can be acute with severe symptoms (for example, acute transverse myelitis) or chronic with progressive myelopathy^14,21,31,34. CNS symptoms occurring in SS may mimic relapsing-remitting multiple sclerosis (10 of our patients) because of possible multifocal involvement and the remitting-relapsing course; some chronic myelopathies could also mimic primary progressive multiple sclerosis (13 of our patients), as previously described^3,13. On MRI, we observed white matter changes compatible with MS in 40% of our patients, but we also found some differences compared with MS patients in the localization of the lesions (rare corpus callosum lesions, involvement of basal ganglia). Furthermore, SS patients were older and the prevalence of oligoclonal bands was markedly lower (30%). Lastly, the presence of clinical features that are atypical of MS (association of PNS involvement and extraglandular features) may help in the diagnosis. Table 8 summarizes the main clinical and imaging differences between SS and MS. In our study we observed a high proportion of patients with a clinical feature suggesting MS (28% of patients, including 13 patients with chronic progressive myelopathy and 10 patients with relapsing-remitting symptoms). This high percentage may be due in part to a recruitment bias, as our neurologic department specializes in MS. However, we are also the regional center for myelopathy, optic neuritis, and peripheral neuropathy, the main features of SS patients with neurologic symptoms. Lastly, we also included patients from the internal medicine department in order to reduce this bias. A coincidental occurrence of MS and SS could also be suggested but, except for primary progressive MS¹³, this association remains rare (between 0 and 3.1%)^37,42.

TABLE 8: Comparison of Features of SS and MS

Acute focal symptoms could also mimic stroke¹⁶, suggesting the influence of an ischemic mechanism. It is noteworthy that we found a high frequency of optic nerve involvement (abnormal VEP in 61% of patients tested), suggesting the need for systematic screening for SS in cases of optic neuropathy^21,48. We also observed a case of motor neuron disease mimicking amyotrophic lateral sclerosis, as previously described³¹.

Lafitte et al³¹ found a high frequency of cognitive dysfunction during SS. We cannot confirm this finding because we did not systematically screen patients for cognitive impairment. However, in the 2 patients who underwent whole neuropsychologic assessment, we confirmed the previous cognitive profile with subcortical dysfunction. Furthermore, we excluded from our study psychiatric features included in several previous studies^15,33,35.

Our study confirms that peripheral neuropathy is the most common PNS complication in SS, affecting 36 of our 51 patients with PNS involvement. As previously described, the most common symptoms of these neuropathies were sensory features^17,30. Our series included the main types of neuropathy already described in SS, including sensory-motor axonal polyneuropathy, pure sensory axonal neuropathy, sensory neuronopathy, and multiple mononeuropathy^19,22,28,36. We did not observe T2-weighted hyperintensities on spinal cord MRI in cases of neuronopathy, in contrast to Mori et al⁴¹, who suggested that areas of high intensity in the spinal dorsal columns could reflect the degree of neuronal involvement.

We also found a high frequency of cranial nerve involvement^2,49. Whereas the trigeminal nerve seems to have been more frequently involved in previous studies^{17,28,31,35,46}, we found a more frequent involvement of the cochlear nerve.

Table 9 shows the clinical and laboratory features of SS in our study compared with those in previous studies^{1,10,16,18,43,51}. We did not find any differences in the prevalence of extraglandular features, except for cutaneous involvement. Cutaneous involvement (vasculitis) seems to be particularly frequent in cases of neurologic SS according to Alexander et al^1,2,4 and Molina et al³⁸, who suggested a relationship between vasculitic lesions and neurologic complications. We found a low prevalence of anti-La/SSB antibodies in our study compared with other results in the literature on SS. We also found that immunoblotting was more sensitive than the Ouchterlony method, especially in cases of CNS involvement, suggesting that this method should be used systematically. Cryoglobulinemia was frequent especially in cases of neuropathy, suggesting that it may have a pathologic role in these cases. An important finding of our study is the higher prevalence of extraglandular and biologic features in patients with PNS involvement than in patients with CNS involvement. This result may help to explain why, in the past, the frequency of CNS involvement in SS has probably been underestimated.

TABLE 9: Characteristic Features Observed in SS Patients, Previous Studies (%)

The current study agrees with those of Mellgren et al³⁶ and Lafitte et al³¹, who described many cases of SS with neurologic manifestations as presenting symptoms. The absence of sicca symptoms in 47% of our patients at the onset of neurologic signs, especially in cases of CNS involvement, emphasizes the difficulty of diagnosis and the utility of systematically screening for SS in patients with neurologic symptoms such as sensory motor axonal polyneuropathy, pure sensory neuropathy, cranial nerve involvement (especially cochlear and trigeminal nerves), acute or chronic myelopathy, or multifocal CNS involvement (MS-like) with onset after 50 years of age.

The pathophysiologic mechanisms of neurologic involvement in SS are still unclear. They seem to differ according to the clinical features. In neuronopathies, lymphocytic infiltration of the dorsal ganglia seems to be the main phenomenon involved. Griffin et al²³ have also described lymphocytic (T-cell) infiltration of the ganglia with neuronal degeneration, determined by a biopsy of the sensory ganglia in 3 patients. In multiple mononeuropathies, nerve biopsy often reveals vasculitis²², which could explain the efficacy of corticosteroids and immunosuppressive drugs in these cases. In symmetric axonal polyneuropathies, vasculitis is probably not the only mechanism, except in a few reported cases with necrotizing vasculitis^8,22,36. We found a high frequency of cryoglobulins in sensory-motor neuropathy, demonstrating their possible pathophysiologic role, as previously suggested^24,36. Other studies have proposed that a specific abnormality of humoral immunity could target neural tissues. Indeed, the similarity between ganglionopathies during SS and Denny Brown neuropathy suggests a role of antineuronal antibodies. Moll et al⁴⁰ found antineuronal antibodies by Western blotting in the sera of 6 of 12 patients with SS and neuropathy. The underlying pathogenesis of CNS complications in SS remains unknown. Many studies have suggested an ischemic mechanism^1,2,4,39. However, the high prevalence of features mimicking multiple sclerosis is not consistent with this suggestion. Mononuclear cell infiltration in the CNS is another hypothesis, supported by Bakchine et al⁸, who described a pathologic case. Additional mechanisms such as immunologically mediated CNS vascular damage, the action of antineuronal antibodies³⁵, or a direct role of anti-Ro/SSA antibodies have been suggested^6,38.

Another important finding in our study is the neurologic outcome and the description of treatments used after a mean neurologic follow-up of 7 years. While we cannot reach a definite conclusion in the absence of double-blind, placebo-controlled studies, we observed several cases of prolonged improvement with cyclophosphamide. In our study, severe disability was more frequent in cases of CNS involvement than in cases of PNS involvement, suggesting the need for intensive and early treatment in cases of CNS involvement. However, our incomplete knowledge of the pathogenic mechanisms coupled with the heterogeneity of CNS involvement makes it difficult to propose any therapeutic guidelines. Cyclophosphamide was particularly effective in cases of myelopathy, with improvement or stabilization of disability. Similar findings have been reported^2,14,53. Other immunosuppressive treatments (azathioprine, chlorambucil) have been tried with variable efficacy^5,25,54. Konttinen et al²⁹ described the efficacy of plasmapheresis and prednisone, but in a sporadic case of acute transverse myelopathy. As previously reported, the efficacy of corticosteroids seems to be variable in cases with CNS manifestations¹¹.

With regard to neuropathies, we observed different therapeutic responses: poor efficacy of corticosteroids in axonal polyneuropathies and neuronopathies, and efficacy of cyclophosphamide and corticosteroids in multiple mononeuropathies, suggesting that they should be systematically administered. In severe cases, intravenous immunoglobulins have been used successfully in a small sample of patients with ganglionopathy⁴⁷. As in our experience, the results of using corticosteroid treatments are not conclusive; it would be useful to conduct a prospective randomized study using this drug to treat neuropathies associated with SS.

CONCLUSION

The current study underlines the diversity of neurologic involvement of SS. The frequency of neurologic manifestations revealing SS and the high level of negative biologic markers, especially in patients with CNS involvement, could explain the high number of misdiagnosed cases.