Neutrophil-related indicators as potential peripheral blood biomarkers for the diagnosis of neurosyphilis: a cross-sectional study

Abstract

Background

Neurosyphilis (NS) is the most serious complication elicited by the invasion of Treponema pallidum (T. pallidum) into the central nervous system. Identifying the neuro-invasion of T. pallidum as early as possible is important. Looking for indicators with good performance and convenient to operate is in great demand for the diagnosis of NS.

Methods

A cross-sectional study was conducted on 428 human immunodeficiency virus (HIV)-negative syphilis patients who underwent therapy in Shanghai Skin Disease Hospital from April 2022 to June 2023. The syphilis patients were further grouped into non-neurosyphilis (NNS), asymptomatic neurosyphilis (ANS) and symptomatic syphilis (SNS). Laboratory examination results of these patients were analyzed.

Results

Significantly elevated neutrophil CD64 (nCD64) index and neutrophil to lymphocyte ratio (NLR) were observed in ANS and SNS patients compared with those of NNS patients. nCD64 index and NLR were strongly correlated with serum and cerebrospinal fluid (CSF) TRUST titer, CSF WBC and CSF protein levels in NS patients. nCD64 index and NLR were independent factors influencing NS, especially ANS. Logistic regression models based on nCD64 index, NLR and serum TRUST titer exhibited better diagnostic performances to the currently used markers in NS diagnosis. Particularly, for NS patients with negative CSF TRUST, diagnostic performance of the combined nCD64 and NLR was obviously superior to CSF protein and CSF WBC.

Conclusions

Neutrophil-related indicators-nCD64 index and NLR have potential application value in the diagnosis of NS, especially for the CSF TRUST negative patients.

Highlights

• 1. Neutrophil-related indicators increased in neurosyphilis.

• 2. Neutrophil-related indicators were strongly correlated with CSF neurosyphilis markers.

• 3. Logistic regression model utilizing neutrophil-related indicators demonstrated good diagnostic advantages, especially in asymptomatic neurosyphilis cases.

Introduction

Syphilis is a sexually transmitted disease (STD) caused by the spirochete Treponema pallidum (T. pallidum) subspecies pallidum, which can cause multi-system lesions and remains a worldwide public health concern [1, 2]. According to the data released by the World Health Organization (WHO), it was estimated that 7.1 million adults between 15 and 49 years acquired syphilis in 2020 (https://www.who.int/news-room/fact-sheets/detail/syphilis). In China, the number of new cases came to 497,934 in 2022, with an increase of 17,914 cases compared to the number reported in 2021 (https://data.stats.gov.cn). Typical clinical progression of syphilis comprises a primary chancre at the point of inoculation, followed by the non-itchy rash on the palms and soles as the secondary syphilis weeks later, a latent period with no symptoms, and the tertiary syphilis with the involvement of the eyes, central nervous system (CNS), and cardiovascular systems etc. [3] Neurosyphilis (NS), elicited by the invasion of T. pallidum into the CNS, is the most serious and common complication [4]. It can occur at any stage of the infection, leading to meningitis, meningovascular disease, or parenchymal syphilis (including tabes dorsalis and general paresis) etc. [2] According to the appearance of related clinical manifestations, NS can be divided into asymptomatic NS (ANS) and symptomatic NS (SNS) [5]. About 26.6% of latent syphilis patients appeared to have ANS, and these patients were more likely to progress to SNS [6]. Intervention in time may effectively prevent the occurrence of SNS. Therefore, to identify the neuro-invasion of T. pallidum as early as possible is of great importance.

Currently, the laboratory diagnosis of NS in syphilis patients is mainly based on the examination of cerebrospinal fluid (CSF) with the treponemal and nontreponemal tests, along with the CSF white blood cells (WBC) counts, and protein (PRO) levels [4]. CSF treponemal test includes the T. pallidum particle assay (TPPA) and the fluorescent treponemal antibody–absorption assay (FTA–ABS). The recommended nontreponemal test in CSF is the venereal disease research laboratory (VDRL) test, while recent studies have proved that rapid plasma reagin (RPR) test and toluidine red unheated serum test (TRUST) have comparable performances with VDRL test [7]. However, no single test can definitively confirm or exclude the diagnosis of NS due to the defects in specificity and sensitivity [8]. Moreover, on account of the asymptomatic nature of early stage NS and the invasive operation, some patients refuse to undergo the lumbar puncture, which may lead to the missed detection of ANS [9, 10]. Hence, looking for indicators with good performance and convenient to operate is in great demand for the diagnosis of NS.

Based on our previous work, in which we observed a significant association between a neutrophil activation marker with the occurrence of NS [11], we went further to get a comprehensive knowledge of the potential role of peripheral neutrophil-related markers in human immunodeficiency virus (HIV)-negative NS population. We reviewed the clinical data and found that neutrophil CD64 (nCD64) index and neutrophil-to-lymphocyte ratio (NLR) were positively correlated with the serum and CSF TRUST titers of HIV-negative NS patients. Moreover, nCD64 index and NLR were also positively correlated with CSF protein levels and WBC counts. Logistic regression analysis indicated that nCD64 index and NLR were independent risk factors in predicting NS, ANS and NNS. Logistic models established based on NLR and nCD64 index showed better or at least comparable diagnostic performances compared to the currently CSF markers in NS diagnosis, especially for patients with negative CSF TRUST. These results suggest the potential value of neutrophil-related indicators in the diagnosis of NS as convenient peripheral blood indicators.

Methods

Clinical cohort

A total of 428 patients diagnosed with syphilis and hospitalized in Shanghai Skin Disease Hospital from April 2022 to June 2023 were included in this study. The enrollment of patients was shown in Fig. 1.The research was approved by the ethics committee of Shanghai Skin Disease Hospital (SSDH-IEC-SG-080–4.0). To ensure complete clinical information, all included patients underwent cerebrospinal fluid and serology testing. A propensity score matching (PSM) was performed to ensure the comparability of clinical endpoints among NNS, ANS, and SNS patients and the balance of gender. We used a 1:1:1.5 ratio with the nearest neighbor matching procedure without replacement. The diagnosis of syphilis was based on the patients’ medical history, physical and neurological examinations, and laboratory tests of serum and CSF. NS diagnosis referred to the “Guidelines for diagnosis and treatment of syphilis, gonorrhea and genital Chlamydia trachomatis infection (2020)” [12]. The inclusion criteria for non-neurosyphilis (NNS) were as followed: (1) serological TPPA and TRUST positive; (2) cerebrospinal fluid TPPA and TRUST negative; and (3) no clinical manifestations of NS. Patients with ANS should meet the following conditions: (1) serological TPPA and TRUST positive; (2) CSF TPPA positive; (3) meet one of the three: CSF WBC ≥ 5 cells/μL; CSF protein ≥ 450 mg/L; CSF TRUST titer > 0; and (4) no clinical manifestations of NS. For SNS patients, (1) serological and CSF inclusion criteria referred to ANS patients; meanwhile, (2) they should have clinical symptoms of neurologic disorder without other causes. The symptoms included paralytic dementia, tabes, dementia, syphilitic meningitis, and psychiatric symptoms such as personality changes, cognitive decline etc. Patients with (1) gonorrhea, mycoplasmas, chlamydia, herpes simplex virus (HSV), human papillomavirus (HPV) and HIV infection; (2) autoimmune diseases; and (3) those who had used antibiotics within 2 weeks prior to hospitalization were excluded.

Fig. 1

Flowchart of inclusion and exclusion of the cohort

Laboratory tests for peripheral whole blood

Blood routine test was performed on the automatic blood analyzer (Sysmex XN-2800, Kobe, Japan). NLR was calculated by neutrophil counts/lymphocyte counts. Neutrophil CD64 expression was determined by the flow cytometer according to the instruction of manufacturer (BD Pharmingen, San Diego, USA). To exclude the influence of monocytes, which constitutively express CD64, the neutrophil CD64 (nCD64) index was included. The calculation formula was as followed: nCD64 index = $\frac{\text{CD64MFI(PMN)/CD}\text{64MFI(LYM)}}{\text{CD64MFI(MO)/CD64MFI(PMN)}}$, MFI refers to mean fluorescent intensity. Serum nontreponemal test was executed using the toluidine red unheated serum test (TRUST) kit (Rongsheng, Shanghai, China). Serum treponemal test was performed on an automatic electro-chemiluminescence analysis platform (Roche Cobas e801, Mannheim, Germany).

CSF examination

WBC counts in CSF were performed under a microscope with the Neubauer’s counting chamber (Shanghai, China). Nontreponemal test in CSF was performed with the TRUST kit (Rongsheng). CSF TPPA was carried out using the Treponema pallidum antibody detection kit (Fujirebio Inc, Tokyo, Japan). CSF protein was detected by an automatic biochemical analyzer (Roche Cobas e702, Mannheim, Germany).

Statistical analysis

Categorical variables were expressed as case numbers and percentages (%), and continuous variables were presented as medians and interquartile ranges. The χ² test was used for comparison between categorical variables. For two group comparisons, the two tailed Mann–Whitney test was used. For multiple group comparisons, we used Kruskal–Wallis test followed by Dunn’s multiple comparisons test. Spearman rank correlation was used to analyze the correlation among all parameters. The receiver operating characteristic (ROC) curve and area under the ROC curve (AUC) were used to evaluate the clinical diagnostic performance of the indicators. Logistic regression was used for predictive analysis. Mediating effect of age and gender was analyzed by structural equation modeling (SEM). P < 0.05 was considered statistically significant. Sensitivity analysis was used to correct for confounding factors including age and sex. All statistical analyses were conducted using the MedCalc v20.218 and GraphPad Prism 8.0.

Results

Characteristics of the enrolled patients

According to the criteria outlined in the Methods section, a total of 110 NNS, 134 ANS, and 184 SNS were enrolled in this study. The flow diagram of patient enrollment was summarized in Fig. 1. Basic information of the enrolled patients before and after PSM was summarized in Supplementary Table 1 and 2. The cohort was evenly distributed between men and women (P = 0.57) after PSM. NS patients were older than NNS patients (P < 0.0001), which was consistent with disease progression. For serological detection, all enrolled patients had positive serum TPPA and TRUST results. Serum TRUST titers of NNS patients were lower than those of the NS patients, and SNS patients usually had higher TRUST titers than the ANS patients (P < 0.0001). In terms of CSF examination, all NNS patients had negative CSF TPPA and TRUST results. All NS patients were positive in CSF TPPA test, while some patients in the ANS and SNS group showed negative CSF TRUST results (Supplementary Table 2). For those with positive CSF TRUST results, titers of the SNS patients were significantly higher than those of the ANS patients (P < 0.0001). Clearly higher CSF protein levels and WBC counts were found in NS patients than those of the NNS patients, with the most significantly increased levels in the SNS patients (P < 0.0001).

Changes of neutrophil-related indicators in syphilis patients

As the first line of innate immune system, WBC in blood play important roles in infection and inflammation. Thus, we analyzed the blood routine examination results of the enrolled patients. Significant differences in counts and percentages of WBC, eosinophil (EO), basophil (BASO), lymphocyte (LYM) and monocyte (MONO) were found across the NNS, ANS and SNS patients (Fig. 2A, B). Notably, neutrophil counts showed more pronounced difference compared with other indicators, which were much higher in ANS (M = 4.23, P < 0.0001) and SNS (M = 3.79, P < 0.0001) than those in NNS (M = 3.19). The same pattern was observed for neutrophil percentages (ANS M = 61.75, SNS M = 62.80, and NNS M = 57.50, P < 0.0001). We further investigated neutrophil-related indicators. As a result, NLR, derived from the ratio of neutrophil and lymphocyte counts, was obviously different in ANS (M = 1.97, P < 0.0001) and SNS (M = 2.36, P < 0.0001) patients compared with those of the NNS (M = 1.52) patients (Fig. 2D). Moreover, the activation marker of neutrophil, nCD64 index, was also clearly increased in NS patients (Fig. 2C). These data indicated that changes of nCD64 index and NLR might be correlated with the disease progression of syphilis patients.

Fig. 2

Distribution of neutrophil-related indicators in syphilis patients. A, B Comparison of blood routine indicators across ANS, SNS and NNS patients. C, D Distribution of nCD64 index (C) and NLR (D) in NS patients was shown separately. (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001)

Correlation of nCD64 index and NLR with serum TRUST titer

The serum nontreponemal test result is recognized to be correlated with syphilis activity and is usually recommended for assessing the relapse and treatment efficacy of syphilis patients. Studies reported that patients with a serum RPR titer ≥ 1:32 were more likely to have NS [3]. Thus, correlations of nCD64 index and NLR with serum TRUST titer in NNS and NS patients were analyzed. Results indicated that neither nCD64 index nor NLR was correlated serum TRUST titer in NNS patients (Supplementary Fig. 1). However, distinct correlation was observed between nCD64 index (r = 0.235, P < 0.0001) and NLR (r = 0.160, P = 0.0040) with serum TRUST titer in NS patients (Supplementary Fig. 1).

Correlation of nCD64 index and NLR with CSF indicators

The invasion of T. pallidum into the CNS is usually indicated by CSF TRUST, CSF protein (CSF PRO), and CSF WBC. Thus, correlations of nCD64 index and NLR with CSF indicators in NS were analyzed. In overall NS patients, markedly elevated nCD64 index and NLR were observed in patients with higher CSF TRUST titer (Fig. 3A, B). Consistently, nCD64 index was significantly correlated with CSF PRO levels (r = 0.39, P < 0.0001) and CSF WBC (r = 0.32, P < 0.0001), while NLR was significantly correlated with CSF protein levels (r = 0.36, P < 0.0001) but not CSF WBC (Fig. 3G).

Fig. 3

Correlation analyses of nCD64 index and NLR with CSF NS indicators. The nCD64 index (A) and NLR (B) increased with CSF TRUST titers in NS patients. Similar patterns were found in ANS (C, D) and SNS (E, F) patients. Correlation matrix of nCD64 index and NLR with CSF PRO and CSF WBC in NS (G), ANS (H) and SNS (I). Each circle represented the correlation analysis result of the indicators corresponding to the vertical and horizontal axes. The Spearman correlation r and P values were marked on each circle. (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001)

According to the clinic symptoms, NS patients were divided into the ANS and SNS patients. In both SNS and ANS patients, the consistent upward trend of nCD64 index and NLR in patients with increasing CSF TRUST was also noticed (Fig. 3C–F). Similar correlation trend of the two indicators with CSF PRO, CSF WBC was observed (Fig. 2H, I). It should be noted that, both nCD64 index (CSF PRO r = 0.36, P < 0.0001; CSF WBC r = 0.32, P < 0.0001) and NLR (CSF PRO r = 0.25, P < 0.0001; CSF WBC r = 0.21, P < 0.001) were distinctly correlated with CSF protein and CSF WBC in SNS patients (Fig. 2I). These results further indicated that peripheral neutrophil-related indicators–nCD64 index and NLR might reflect the CNS involvement in syphilis patients.

Correlations of neutrophil-related indicators with CNS involvement in syphilis patients

To further confirm whether changes of neutrophil-related indicators reflect the CNS involvement, binary logistic regression analyses were performed in different syphilis patients. Serum TRUST titer (adjusted OR = 1.09, 95% CI 1.02–1.16), nCD64 index (adjusted OR = 4.75, 95% CI 2.62–8.61) and NLR (adjusted OR = 5.51, 95% CI 2.61–11.65) were found to be independent predicting factors of NS after adjusted for age and sex (Fig. 4A, Supplementary Table 3). Furthermore, nCD64 index (adjusted OR = 4.55, 95% CI 2.04–8.14), NLR (adjusted OR = 4.75, 95% CI 2.06–10.93) and serum TRUST titer (adjusted OR = 1.08, 95% CI 1.04–1.14) were also independent predicting factors of ANS (ANS = 1, NNS = 0) (Fig. 4D, Supplementary Table 3). SEM was performed to further explore the potential mediating effect of gender and age. The results showed that neither gender nor age had significant mediating effect on nCD64 (gender P = 0.432, age P = 0.241), while both significantly mediated the effect of NLR (P < 0.001).

Clinically, there are NS patients with negative CSF TRUST. In our cohort, 43 ANS and 78 SNS patients were CSF TRUST negative. Logistic regression showed that nCD64 index (adjusted OR = 5.48, 95% CI 2.30–13.04) and NLR (OR = 2.51, 95% CI 1.04–5.08) were independent predictors of NS with negative CSF TRUST, while serum TRUST titers were not (adjusted OR = 1.02, 95% CI 0.99–1.04) (Fig. 4G, Supplementary Table 3).

Diagnostic performance of neutrophil-related indicators in NS patients

To figure out the efficacy of nCD64 index and NLR in NS diagnosis, the ROC curves were plotted. AUC values of serum TRUST, nCD64 index and NLR were 0.726, 0.798 and 0.797, respectively (Table 1). Logistic regression model 1 (LRM1, Y = −4.11 + 0.08*serum TRUST titer + 1.42*nCD64 index + 1.24*NLR) was developed by incorporating these three parameters. Interestingly, the AUC value of LRM1 (AUC = 0.869) was remarkably better than serum TRUST (AUC = 0.726, P < 0.001), CSF WBC (AUC = 0.752, P < 0.001) and CSF TURST (AUC = 0.813, P = 0.008), and was parity with CSF PRO (AUC = 0.839, P = 0.239) (Fig. 4B, C, Table 1).

Table 1 Diagnostic performance nCD64 index and NLR in discriminating NNS and ANS and CSF negative patients

We further checked the predictive efficacy of these indicators in ANS diagnosis. Another logistic regression model–LRM2 was established (Fig. 4E, F, Table 1). The AUC value of LRM2 (Y = − 4.39 + 0.09*serum TRUST titer + 1.44*nCD64 index + 0.88*NLR) was 0.840, which was much better than CSF WBC (AUC = 0.753, P = 0.02) and was comparable to CSF PRO (AUC = 0.843, P = 0.932) and CSF TURST (AUC = 0.840, P = 0.974).

For CSF TRUST negative NS patients, AUC values of nCD64 index (AUC = 0.708, P < 0.001) and NLR (AUC = 0.729, P < 0.001) were much higher than that of serum TRUST (AUC = 0.627) (Fig. 4H, I, Table 1). The AUC of diagnostic model LRM3 (Y = − 8.58 + 1.71*nCD64 index + 0.95*NLR) for NS in CSF negative patients reached 0.777, which was far better than CSF PRO (AUC = 0.624, P = 0.004) and CSF WBC (AUC = 0.650, P = 0.006). Thus, these data indicated that neutrophil-related–nCD64 index and NLR had great potential in the diagnosis of ANS and CSF TRUST negative NS patients.

Discussion

NS is one of the most severe complications of syphilis due to its irreversible and serious damages to the central nervous system [2]. Early diagnosis and treatment are crucial to the prognosis of patients. In this study, we focused on peripheral blood indicators and found that neutrophil-related indicators nCD64 index and NLR were significantly increased in NS patients. Further investigation indicated that nCD64 index and NLR were distinctly correlated with serum and CSF TRUST titers, CSF protein levels and CSF WBC counts, suggesting the potential value of these two indicators in NS diagnosis. Logistic regression and ROC analysis demonstrated that LRMs incorporating serum TRUST titer, nCD64 index and NLR had better or at least comparable performances to the currently CSF markers in NS and ANS diagnosing. Moreover, combined analysis of nCD64 index and NLR showed remarkably superiority in diagnosing NS patients with negative CSF TRUST.

The diagnosis of NS mainly depends on the neurological manifestations, positive CSF treponemal/nontreponemal test and abnormal CSF indicators in syphilis patients [13, 14]. However, the diversity of symptoms in NS make the diagnosis difficult or even overlooked [4, 15]. ANS is the most common form of NS. Studies have shown that about 13.5% of patients with latent syphilis have ANS, who are more likely to refuse lumbar puncture [16]. Moreover, there are suspicious NS patients with negative CSF TRUST/VDRL [17], which cannot be precisely defined. Therefore, the diagnosis of NS remains challenging and many researchers are dedicated to exploring new indicators, especially for ANS patients. Chen et al. reported that serum and CSF ubiquitin C-terminal hydrolase-L1 (UCH-L1), glial fibrillary acidic protein (GFAP) and neurofilament light protein (NF-L) could be used as novel markers for the diagnosis of NS with pleasant performances [18, 19]. Marra et al. found that NS patients had higher CSF and serum chemokine (C–X–C motif) ligand 13 (CXCL13) concentrations compared to patients with syphilis but not neurosyphilis. Besides, a cutoff of 10 pg/mL CSF CXCL13 and a cutoff of 250 pg/mL or evidence of intrathecal synthesis of CXCL13 had high specificity for diagnosis of both SNS and ANS [20]. With the development of technologies like polymerase chain reaction (PCR) testing, proteomic and metabonomics analysis, more and more potential markers have been reported in NS diagnosis [21]. However, the defect in sensitivity or specificity, lack of international consensus and mature commercial detecting systems limit their clinical application. Thus, we tried to figure out whether there were certain indicators which were clinically in use could help NS diagnosis.

As an important part of the innate immune system, neutrophils play key roles in the defense against infectious pathogens and inflammation inducing agents [22]. The main pathogen-eliminating mechanisms of neutrophils include phagocytosis, granule release, reactive oxygen species production, and neutrophil extracellular trap formation [23]. Immediate release of neutrophils into circulation after pathogen invasion can cause a rapid rise in neutrophil counts, NEU%, and NLR. Meanwhile, strong expression of the high affinity immunoglobulin G Fc receptor I (Fcγ RI, CD64) is reported to be quickly induced on neutrophils in response to infections or proinflammatory cytokines [24]. Therefore, these neutrophil-related indicators have been demonstrated to be reliable markers for infectious diseases such as sepsis [24], mycobacterium tuberculosis infection [25], community-acquired pneumonia [26, 27], and COVID pneumonia [28, 29] etc. Interaction of T. pallidum and human neutrophils has been observed and T. pallidum is reported to inhibit the apoptosis of neutrophils through intrinsic and extrinsic pathways [30]. However, the application of these indicators in the diagnosis of syphilis remains undetermined. By reviewing the clinical data of 428 syphilis patients, we found that both nCD64 index was significantly correlated with the serum and CSF TRUST titers and CSF profiles, as described in our previous study [11]. More importantly, a similar trend was found in neutrophil counts and NRL. Further combination of nCD64 index and NLR with serum TRUST showed better or at least comparable diagnostic performances to the currently CSF markers in NS and ANS diagnosis. Therefore, the LRMs established in this study might help the diagnosis of NS in patients without lumbar puncture (see Fig. 4).

Fig. 4

Performance of nCD64 index and NLR in NS diagnosis. A, D, E Forest maps for risk assessment of NS, ANS, CSF TRUST negative NS patients. B, C Comparison of diagnostic performances of LRM1 with serum and CSF indicators in NS diagnosis. E, F Comparison of diagnostic performances of LRM2 with serum and CSF indicators in ANS diagnosis. H, I Comparison of diagnostic performances of LRM3 with serum and CSF indicators for CSF TRUST negative NS patients. (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001)

Specially, we checked the diagnostic performances of nCD64 and NLR in NS patients with negative CSF TRUST. As sensitivity of nontreponemal tests for neurosyphilis varies from 30 to 70%, there is false-negative result which may interfere the diagnosis of NS [13]. Distinctly increased nCD64 and NLR were observed in CSF TRUST negative NS patients, and LRM based on these two indicators showed greatly improved performance (AUC = 0.777, specificity of 75.90% and sensitivity of 79.25%) in diagnosing NS patients with negative CSF TRUST, compared to the CSF PRO (AUC = 0.624, specificity of 36.90% and sensitivity of 87.74%) and WBC (AUC = 0.650, specificity of 79.76% and sensitivity of 50.94%). Hence, nCD64 and NLR might help diagnosing the CSF TRUST negative NS patients.

In this study, we also found the discrepancy of monocytes, eosinophils, and basophils among the patients with syphilis infection. However, we did not go further to investigate their potential in NS diagnosis by combining them with nCD64 index and NLR. Besides, a previous study established an LRM through several blood indicators, the comparison of the two model was not compared in this study [8]. Whether the combination of these parameters like different WBC, patients’ age, weight etc. with nCD64 index and NLR would further improve the diagnostic performance is worth studying and we will keep on it in our future work. Another limitation of this study was that HIV-positive syphilis patients were not included due to insufficient data. Given the rising number of HIV co-infected syphilis patients, further investigation was needed to ascertain whether our present conclusion was applicable to the diagnosis of NS in these patients.

In conclusion, our current study demonstrated for the first time that neutrophil related indicators–nCD64 index and NLR may benefit the diagnosis of different NS patients, particularly in asymptomatic patients where diagnostic tests have limited sensitivity. Considering the potential risks of lumbar puncture and limitations of CSF examination, the LRMs established by peripheral blood indicators in this study would provide great help to diagnose the NS patients, especially those with negative CSF TRUST.