Analysis of risk factors associated with central neuropathy in patients with primary Sj?gren′s syndrome

， , ， ， ，

Nanfang Hospital, Southern Medical University, 1.Department of Neurology; 2.Department of Rheumatology; 3.First Clinical Medicine College, Southern Medical University, Guangzhou 510015，China

[Abstract] Objective: To explore the clinical manifestations and risk factors of primary Sj?gren′s syndrome associated central nervous system(pSS-CNS). Methods:A total of 291 patients with primary Sj?gren′s syndrome(pSS)in Nanfang Hospital of Southern Medical University from January 2015 to December 2020 were selected and divided into the pSS-CNS group and pSS-non-CNS group.The general conditions, laboratory data were collected to analyze the clinical characteristics between the two groups. Then multivariate Logistic regression analysis was performed on the risk factors of pSS-CNS, and ROC curve was drawn to evaluate the predictive value of each risk factor. Results:① pSS-CNS patients were mainly neuromyelitis optic spectrum disorders (NMOSD,15 cases, 33.3%), the other 30 cases included leukoaraiosis, stroke, epilepsy, dementia and other central lesions. ② pSS-CNS group had a higher European League Against Rheumatism Sj?gren′s Syndrome Disease Activity Index (ESSDAI) than pSS-nCNS group [(11.14±2.97) vs (5.39±2.06), P<0.01]. ③ The independent risk factors for pSS-CNS were IgG (OR: 0.890; 95% CI: 0.826～0.959; P=0.002), IgM (OR: 0.367; 95% CI: 0.182～0.739;P=0.005), CH50 (OR: 0.935; 95% CI: 0.901～0.970;P<0.01), PLT (OR: 1.005; 95% CI: 1.001～1.009; P=0.010), the following curves of the four area: 0.890(0.826～0.959), 0.367 (0.182～0.739), 0.935(0.901～0.970), 1.005(1.001～1.009), and the area under the curve of the combined application of the four was 0.790 (0.761～0.858). Conclusions：The clinical manifestations of pSS-CNS are varied. Some patients may present the symptoms of central nervous system as their first or only clinical symptom. In our study, NMOSD was the most common one. IgG, IgM, CH50 and PLT are independent risk factors for pSS-CNS, among which, the higher the PLT, the lower the IgG, IgM and CH50, and the higher the possibility of central neuropathy.

[Keywords] primary Sj?gren′s syndrome； central nervous system； neuromyelitis optic spectrum disorders; risk factors analysis

1 Background

Primary Sj?gren′s syndrome (pSS) is an autoimmune disease that can affect multiple systems. It is characterized by progressive decline in salivary gland and lacrimal gland function, with high heterogeneity[1-2]. Neuropathy is one of the most common extraglandular manifestations of pSS. Literature data estimate the presence of neurological symptoms in about 8.5%～70% of patients diagnosed with pSS, and the most common neurological complication is peripheral neuropathy[3]. In contrast, central nervous system (CNS) involvement is relatively much less common, with prevalence between 2.5% and 60%, which may be from the fact of lacking a unified definition and diagnostic criteria[4]. A Spanish study of pSS patients followed up for 13 years found that 13% of patients will eventually develop life-threatening complications, of which central nervous system disease is second only to lymphoma, and the mortality rate is close to 20%[5], which is worth attach great importance to it clinically. Moreover, some patients with pSS have the nervous system as the first manifestation, usually lead to clinical misdiagnosis and missed diagnosis. Presently, there have been some clinical studies on pSS combined with neurological diseases in China and abroad, but the number of cases is relatively small. Our study collected the clinical data of 291 patients with pSS, 45 of which were pSS-CNS. We analyzed their clinical characteristics and independent risk factors in order to provide help for the clinical diagnosis and treatment of primary Sj?gren′s syndrome.

2 Methods

2.1 Participants

Inpatients in the Department of Neurology and Rheumatology and Immunology of Nanfang Hospital, Southern Medical University from January 2015 to December 2020 were screened.The diagnostic criteria for pSS were based on the 2016 pSS diagnostic criteria of the American College of Rheumatology and the European League Against Rheumatism[1]. Patients with other connective tissue diseases such as systemic lupus erythematosus, rheumatoid arthritis, and systemic sclerosis were excluded.The neurologist and the rheumatologists combined with the clinical manifestations and impact data of the patients, and screened out the qualified pSS-CNS patients under their joint determination. Exclusion criteria were the patients with central nervous system lesions before the diagnosis of pSS and neurological damage caused by diabetes, hyperlipidemia, atherosclerosis, malignant tumors and other diseases[6]. 291 patients with pSS were included in the study.

2.2 Study Design

The clinical data of 291 patients were collected, including gender, age, age of onset, course of disease, clinical manifestations, European League Against Rheumatism Sj?gren′s Syndrome Disease Activity Index (ESSDAI), laboratory inspections. According to whether the patients had central nervous system damage, 291 pSS patients were divided into pSS-CNS (45 cases, 15.5%) and pSS-nPNS (246 cases, 84.5%), and the clinical characteristics and laboratory data were compared between the two groups to analysis the possible risk factors for central nervous system damage. The pSS-CNS group was subdivided into 2 groups, in which with a diagnosis of myelitis as one group (15 cases, 33.3%), and the non-myelitis patients were divided into the other group (30 cases, 66.7%). Analyze the statistical differences between the two groups (Table 1).

2.3 Statistical Analysis

SPSS 22.0 was used for data analysis. Enumeration data were expressed by rate, and its comparison between groups was by2test; measurement data was expressed by mean±SD, and comparison between groups was by independent sample t-test. The statistically significant variables were analyzed by multivariate logistic regression and ROC curve was drawn to assess the predictive value of each risk factor.

Table 1 Comparison between pSS-CNS and pSS-nCNS

3 Results

3.1 Baseline data of pSS-CNS and pSS-nCNS patients

There were 45 patients with pSS-CNS, 43 females (95.6%) and 2 males (4.4%). The disease duration was 10 hours to 18 years, and the average was 25.25 months. In pSS-nCNS group, 213 (86.6%) were females and 33 (13.4%) were males. The longest disease duration was 23 years, and the shortest was 0.2 months (6 days), with a mean of 37.88 months. There was no statistical significance in gender, age, age of onset and disease duration between the two groups (Table 1).

3.2 Clinical manifestations of the pSS-CNS group and pSS-nCNS group

Compared with pSS-nCNS, headaches (24% vs 7%,P=0.010), limb numbness (33% vs 5%,P<0.01) and limb weakness (33% vs 14%,P=0.002) were more common in pSS-CNS, while concurrent hematological damage (4% vs 17%,P=0.038) and fever (24% vs 11%,P=0.014) were relatively rare. There was no significant difference between the two groups in terms of dental caries, parotid gland enlargement, Raynaud′s phenomenon, joint swelling and pain, and interstitial pneumonia (Table 1).

3.3 The laboratory results of the pSS-CNS group and the pSS-nCNS group

Through independent sample T test, we found that compared with the pSS-nCNS group, the neutrophil count (5.20±3.39 vs 4.16±2.78,P=0.031), platelets (239.16±81.84 vs 195.93±88.78,P=0.010), albumin/globulin ratio (1.37±0.30 vs 1.18±0.32,P=0.001) in the pSS-CNS group were higher, while globulin (28.36±6.43 vs 72.57±7.79,P=0.043), IgG (12.87±4.77 vs 17.02±6.98,P<0.01), IgM (0.98±0.50 vs 1.55 ± 1.50), and CH50 (51.58±10.62 vs 54.22±11.54) were lower. In addition, we found that the disease activity score (ESSDAI) between the two groups was significantly different (11.14±2.97 vs 5.39±2.06,P<0.01). The disease activity coefficient of the pSS-PNS group was significantly higher than that of the pSS-nCNS group (Table 1).

3.4 The treatment in pSS-CNS group and pSS-nCNS group

There were significant differences in treatment between the two groups.All patients in the pSS-CNS group,while 80% of patients in the pSS-nCNS group,received medication. In the pSS-CNS group, 33 patients had central nervous system lesions as the first symptom and they had not received any drug therapy before, the others who received low-dose steroid and hydroxychloroquine therapy had central nervous system lesions during the course of pSS. Forty-one patients (91%)with pSS-CNS were treated with steroids at the time of the study, a significantly higher proportion than those in the pSS-nCNS group, with 17 patients receiving steroid pulse therapy. Compared with pSS-nCNS group, more patients in pSS-CNS group were treated with cyclophosphamide (18% vs 3%,P<0.01) and azathioprine (9% vs 0.4%,P=0.001), and the difference was statistically significant (Table 1).

3.5 Analysis of independent risk factors of pSS-CNS

Logistic regression model was used to analyze the independent risk factors of pSS-PNS. ESSDAI (EULAR Sj?gren′s syndrome disease activity index) contains information on clinical manifestations, humoral immune indicators, hematological system, lungs and other systems. Considering that ESSDAI will repeat and interfere with other indicators in this study, the ESSDAI indicator was excluded when performing Logistic regression analysis. In addition, the various clinical symptoms of pSS patients were highly subjective, so we finally removed them during logistic regression. Logistic multivariate regression analysis was performed forP<0.05 indicators: neutrophils count, platelet count, albumin/globulin ratio, globulin, IgG, IgM and CH50 (Table 5). We found that PLT, CH50, IgM, and IgG were independent risk factors for pSS-CNS, and the AUC (95%CI) of the above independent variables were 1.005 (1.001～1.009), 0.935 (0.901～0.970), 0.367 (0.182～0.739) and 0.890 (0.826～0.959) (P<0.05), respectively. Notably, the area under the curve AUC for the combined application of the four indicators was 0.790 (0.761～0.858) (P<0.01),

suggesting good predictability for the development of pSS-CNS (Table 2, Figure 1).

Table 2 The independent risk factors of pSS-CNS

Figure 1 ROC curve for four risk factors of primary Sj?gren′s syndrome associated central nervous system (pSS-CNS).

3.6 Comparison of baseline data and laboratory tests between the pSS-NMOSD group and the pSS-nNMOSD group

Among the 45 pSS-CNS patients, 43 were female and 2 were male, with a ratio was 11.5 ∶1, a mean age of 52.53±15.48 years old. Among them, 15 patients were neuromyelitis optica spectrum disorders (NMOSD), including 5 acute transverse myelitis patients and 10 neuromyelitis optica (NMO) patients. The remaining 30 pSS-CNS patients were leukoaraiosis (5 cases, 16.7%), stroke (6 cases, 20.0%), leukoaraiosis combined with stroke (6 cases, 20.0%), Parkinson′s disease (2 cases, 6.7%), anxiety and depression (3 cases, 10%), multiple sclerosis (1 case, 3.3%), epileptic seizures (2 cases, 6.7%), encephalitis (3 cases, 10%) and demyelination (2 cases, 6.7%). 45 patients were divided into pSS-NMOSD group and pSS-nNMOSD group, we compared the differences in general conditions and laboratory data of pSS-CNS.

Through the independent-sample t-test between the two groups, it could be found that the age (43.00±12.78 vs 57.23±14.55,P= 0.002) and age of onset (41.00±13.99 vs 55.33±14.62,P=0.003) were lower in the pSS-NMOSD group than that in the pSS-nNMOSD group, while immunoglobulin G (IgG) (15.03±6.82 vs 12.03±3.05,P=0.046), glomerular filtration rate (eGFR) (120.55±39.13 vs 91.51±24.18,P=0.004) were contrary. And the difference between the two groups was statistically significant (Table 3).

Table 3 Comparison of general conditions and laboratory data between pSS-NMOSD group and pSS-nNMOSD group

4 Discussion

Primary Sj?gren′s syndrome is a chronic inflammatory autoimmune disease with lymphocyte proliferation and progressive exocrine gland damage, and a variety of autoantibodies are present in the serum, which may present with multiple organs and systems involvement in addition to impaired salivary and lacrimal gland function. Central nervous system disease is one of the serious complications. Its clinical manifestations vary and can be divided into two categories. One is focal manifestations, including stroke, neuromyelitis optica, multiple sclerosis-like manifestations, and amyotrophic lateral sclerosis-like manifestations, etc. The other is diffuse manifestations, including cognitive disorders, dementia, mental disorders, and aseptic meningitis, etc.[7]Recent studies have found that pSS-related central neuropathy is not rare, and even some central nervous system symptoms have appeared early before the diagnosis of primary Sj?gren′s syndrome[8]. Therefore, exploring the clinical characteristics and risk factors of pSS-related central nervous system damage plays a key role in guidng early clinical identification of patients and early intervention for improving the prognosis of patients.

At present, domestic and foreign literature reports on pSS-CNS mainly focus on case reports or descriptive studies, and there are only a few retrospective case-control studies, which mainly analyze the risk factors of neurological diseases in pSS, and do not distinguish between the central nervous system and the peripheral nervous system lesions. Li et al[9]found by analyzing a retrospective case-control study of 142 cases of pSS that age ≤45 years, disease course ≤4 years, RF elevation ≤3 times, and ESSDAI>3 points were risk factors for pSS-related neuropathy. This study was also a retrospective case-control study. Different from previous studies, we considered that the disease activity coefficient score ESSDAI was duplicated with other clinical and laboratory indicators in this study, so the ESSDAI indicator was excluded when performing logistic regression analysis. Besides, the various clinical symptoms appeared in pSS patients are highly subjective, so we finally also exclude them during Logistic regression analysis. Therefore, we conclud that the independent risk factors for pSS-CNS are CH50, IgG, IgM and platelet count, which were inconsistent with previous conclusions.

Up to now, the pathogenesis of pSS-CNS has not been fully clarified. Our study found that the CH50, IgG, and IgM in the serum of pSS-CNS patients are lower than those of pSS-nCNS patients, suggesting that the complement system and immunoglobulin IgG and IgM may play an important role in it. Sanders et al[10]found that SC5a-9 in the complement activation system was present in the cerebrospinal fluid of 16 patients with pSS-CNS by detecting the cerebrospinal fluid, which indicates that intrathecally activated terminal complement plays a role in central nervous system diseases. In our study, the total CH50 in peripheral blood in patients with pSS-CNS was reduced, which may be related to the complement consumption in the cerebrospinal fluid. Cui et al[6]found that patients with extensive involvement in the pSS-CNS had higher blood IgG levels than those without extensive involvement. Pars et al[11]believed that oligoclonal bands of IgG synthesis can be detected in the serum and cerebrospinal fluid of patients with pSS-CNS, and cerebrospinal fluid examination can be used as a diagnostic test to rule out other inflammatory or autoimmune diseases. However, in this study, we find that the serum levels of IgG and IgM in pSS-CNS patients are lower than those in pSS-nCNS patients. Due to the limitation of retrospective studies, we are unable to collect the cerebrospinal fluid of pSS-CNS patients to verify whether there is a corresponding increase in various immunoglobulins in the cerebrospinal fluid of these patients. More studies are needed to confirm it.

As we all know, platelets have hemostatic effect. And platelets have received increasing attention in recent years as immune and inflammatory effector cells[12]. Sarah et al[13]found that in the early stages of multiple sclerosis or experimental autoimmune encephalitis, platelets stimulate the proliferation and differentiation of pathogenic Th1, Th17 and IFN-γ/IL-17 double-positive CD4+T cells by secreting soluble factors 5-HT, platelet-activating factor (PAF) and PF4, showing distinct proinflammatory properties. Our study show that in pSS-CNS patients, blood platelet count is higher than those in pSS-nCNS patients, suggesting that platelets may act as an inflammatory factor in pSS-CNS patients, and it may be a cause in the development of central nervous system-related complications. Of course, this needs to be confirmed by further research.

NMOSD is an autoimmune central nervous system demyelinating disease with optic neuritis and myelitis as the main clinical manifestations. 10% to 40% of NMOSD patients may be accompanied by autoimmune diseases, such as thyroiditis, systemic lupus erythematosus or Sj?gren′s syndrome[14]. Studies have found that the incidence of NMOSD combined with Sj?gren′s syndrome is approximately 2%～30%[15], and NMOSD is the most common form of central neuropathy in pSS patients. In this study, pSS-CNS is mainly pSS-NMOSD (proportion of 33.3%), consistent with previous researches.

Patients with pSS-NMOSD have more severe condition and worse prognosis[16]. At present, most of the studies on pSS-related myelitis are clinical reports. AQP4 is the main aquaporin of the central nervous system. It has been found that the appearance of anti-AQP4 antibody may predict an increase in the possibility of NMOSD complicated by Sj?gren′s syndrome, and anti-AQP4 antibody may participate in the concurrent pathogenesis of the two[16]. Some studies have also show that pSS-NMOSD patients are mainly female, and the positive rates of anti-AQP4 antibody and anti-SSA antibody are very high[17], thus the pathogenesis of pSS-NMOSD has not yet been clarified and further research is needed.

In summary, central nervous system disease is a kind of common complication of pSS, of which neuromyelitis optica spectrum diseases has high morbidity. This study shows that compared with pSS-nCNS patients, pSS-CNS patients have lower CH50, IgG, and IgM levels, and higher blood platelet counts. The above indicators are independent risk factors for pSS-CNS which should be observed in clinical practice to predict the possibility of CNS complications, achieve early detection, diagnosis, and treatment to improve the prognosis.