Отличительные прогностические последствия нервных и астроглиальных биомаркеров крови в расстройствах спектра нейромелита по сравнению с множественным склерозом.

Introduction

Neuromyelitis optica spectrum disorders (NMOSD) and multiple sclerosis (MS) are autoimmune demyelinating diseases of the central nervous system (CNS), but they differ significantly in their underlying mechanisms and pathology 1 – 4 . NMOSD is primarily associated with autoantibodies targeting the aquaporin-4 (AQP4) protein expressed on astrocytes, leading to astrocyte damage and secondary demyelination 5 , 6 . Meanwhile, MS is characterized by immune-mediated damage to the myelin sheath and oligodendrocytes, although the specific antigen driving the immune response remains unidentified 7 – 9 . Despite these differences, both conditions are characterized by relapses and deterioration throughout life, necessitating the development of biomarkers to facilitate long-term prognostication 10 .

Recent advancements have highlighted the potential of serum glial fibrillary acidic protein (sGFAP) and serum neurofilament light chain (sNfL) as promising biomarkers for assessing disease activity and predicting disability in neurological diseases including NMOSD and MS 11 – 13 . These markers reflect different types of CNS damage, with sGFAP indicating astroglial injury and sNfL reflecting neuronal damage. In MS, studies have established a strong correlation between elevated levels of sGFAP and sNfL with future disease progression, including brain atrophy and increasing disability, suggesting their utility in long-term prognostication 14 – 20 . Similarly, in NMOSD, these biomarkers—particularly sGFAP—have been associated with current disease disability, relapse activity, and the future occurrence of relapses 21 – 26 . However, there have been limited studies investigating the long-term prognostic value of these biomarkers simultaneously in cohorts of both MS and NMOSD patients.

Given the distinct pathogenic mechanisms underlying NMOSD and MS, we hypothesized that the prognostic significance of sGFAP and sNfL would differ between these diseases. To test this hypothesis, we aimed to prospectively investigate the prognostic value of sGFAP and sNfL in predicting future outcomes, including neurological disability, cognitive function, clinical relapses, and disability progression, in NMOSD compared to MS.

Results

Baseline characteristics of participants

During the study period, a total of 260 patients were initially considered, including 192 with MS and 68 with NMOSD. Of these, 92 patients with relapsing-remitting MS and 41 with NMOSD consented to participate and met the criteria for enrollment, as they were not experiencing a recent clinical attack within 2 months prior to enrollment at the time (Fig.  1 ) 21 . NMOSD patients were significantly older and more likely to be female compared to MS patients (Table  1 ). They also had fewer years of education and a shorter duration of the disease than those with MS. After adjusting for age, baseline Expanded Disability Status Scale (EDSS) and Processing Speed Test (PST) scores were comparable between the groups. As for blood biomarkers, while sNfL levels did not differ between the groups, sGFAP levels were significantly higher in patients with NMOSD.

Fig. 1 Flow chart of the study participants.

Flow chart of the study participants.

Table 1 Baseline characteristics of patients with NMOSD and MS. NMOSD MS
p
p ’ ( n  = 41) ( n  = 92) Age, median [IQR] 50.0 [45.5–60.5] 44.0 [35.0–53.8] < 0.001 Age at onset, median [IQR] 43.0 [33.5–52.0] 32.0 [23.0–39.8] < 0.001 Female, n (%) 37 (90.2) 65 (70.7) 0.014 Education, median [IQR] 12.0 [12.0–16.0] 16.0 [12.0–16.0] 0.049 Disease duration, years, median [IQR] 9.0 [4.5–13.0] 11.0 [7.3–14.0] 0.030 Last Clinical Attack location, n (%) 0.076 ON 8 (19.5) 16 (17.4) TM 26 (63.4) 36 (39.1) Brain 5 (12.2) 31 (33.7) ON + TM 0 2 (2.2) ON + Brain 1 (2.4) 2 (2.2) TM + Brain 1 (2.4) 5 (5.4) EDSS, median [IQR] 2.5 [2.0–4.0] 2.0 [1.0–3.5] 0.041 0.912 PST score, median [IQR] 41.0 [19.0–53.5] 47.0 [36.3–63.0] 0.019 0.723 Immune modulatory agent a , n (%) 40 (97.6) 85 (92.4) 0.247 B-cell depleting treatment (BCDT) 12 (29.3) Non-B-cell depleting treatment 28 (68.3) High-efficacy treatment 20 (21.7) Low-efficacy treatment 65 (70.7) Blood biomarker sGFAP, pg/mL, median [IQR] 127.2 [102.6–229.3] 90.0 [68.1–129.0] < 0.001 0.036 sNfL, pg/mL, median [IQR] 12.9 [9.1–20.3] 9.8 [7.6–14.7] 0.033 0.929 EDSS = Expanded disability status scale; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorder; ON = Optic neuritis; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein; sNfL = Serum neurofilament light chain; TM = Transverse myelitis. a NMOSD patients were treated with Azathioprine ( n  = 19), Rituximab ( n  = 12), Mycophenolate mofetil ( n  = 9). MS patients were treated with Interferon-β ( n  = 26), Teriflunomide ( n  = 22), Fingolimod ( n  = 11), Dimethyl fumarate ( n  = 11), Alemtuzumab ( n  = 9), Glatiramer acetate ( n  = 4) and others ( n  = 2).

Table 1 Baseline characteristics of patients with NMOSD and MS. NMOSD MS
p
p ’ ( n  = 41) ( n  = 92) Age, median [IQR] 50.0 [45.5–60.5] 44.0 [35.0–53.8] < 0.001 Age at onset, median [IQR] 43.0 [33.5–52.0] 32.0 [23.0–39.8] < 0.001 Female, n (%) 37 (90.2) 65 (70.7) 0.014 Education, median [IQR] 12.0 [12.0–16.0] 16.0 [12.0–16.0] 0.049 Disease duration, years, median [IQR] 9.0 [4.5–13.0] 11.0 [7.3–14.0] 0.030 Last Clinical Attack location, n (%) 0.076 ON 8 (19.5) 16 (17.4) TM 26 (63.4) 36 (39.1) Brain 5 (12.2) 31 (33.7) ON + TM 0 2 (2.2) ON + Brain 1 (2.4) 2 (2.2) TM + Brain 1 (2.4) 5 (5.4) EDSS, median [IQR] 2.5 [2.0–4.0] 2.0 [1.0–3.5] 0.041 0.912 PST score, median [IQR] 41.0 [19.0–53.5] 47.0 [36.3–63.0] 0.019 0.723 Immune modulatory agent a , n (%) 40 (97.6) 85 (92.4) 0.247 B-cell depleting treatment (BCDT) 12 (29.3) Non-B-cell depleting treatment 28 (68.3) High-efficacy treatment 20 (21.7) Low-efficacy treatment 65 (70.7) Blood biomarker sGFAP, pg/mL, median [IQR] 127.2 [102.6–229.3] 90.0 [68.1–129.0] < 0.001 0.036 sNfL, pg/mL, median [IQR] 12.9 [9.1–20.3] 9.8 [7.6–14.7] 0.033 0.929 EDSS = Expanded disability status scale; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorder; ON = Optic neuritis; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein; sNfL = Serum neurofilament light chain; TM = Transverse myelitis. a NMOSD patients were treated with Azathioprine ( n  = 19), Rituximab ( n  = 12), Mycophenolate mofetil ( n  = 9). MS patients were treated with Interferon-β ( n  = 26), Teriflunomide ( n  = 22), Fingolimod ( n  = 11), Dimethyl fumarate ( n  = 11), Alemtuzumab ( n  = 9), Glatiramer acetate ( n  = 4) and others ( n  = 2).

Baseline characteristics of patients with NMOSD and MS.

EDSS = Expanded disability status scale; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorder; ON = Optic neuritis; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein; sNfL = Serum neurofilament light chain; TM = Transverse myelitis.

a NMOSD patients were treated with Azathioprine ( n  = 19), Rituximab ( n  = 12), Mycophenolate mofetil ( n  = 9). MS patients were treated with Interferon-β ( n  = 26), Teriflunomide ( n  = 22), Fingolimod ( n  = 11), Dimethyl fumarate ( n  = 11), Alemtuzumab ( n  = 9), Glatiramer acetate ( n  = 4) and others ( n  = 2).

Correlation of blood biomarkers with clinical outcomes

A total of 37 patients with NMOSD and 87 with MS completed the follow-up EDSS evaluation, while 31 with NMOSD and 61 with MS completed the follow-up PST evaluation (Fig.  2 ). Over the 2-year follow-up period, EDSS scores in patients with MS increased (baseline vs. 2-year, median (IQR): 2.0 (1.0–3.5) vs. 2.0 (1.0–3.5), p  = 0.024), indicating a decline in disability, while those in patients with NMOSD did not exhibit significant changes (2.5 (2.0–4.3) vs. 3.0 (2.0–3.5), p  = 0.362). Meanwhile, PST scores improved in both NMOSD (36.0 (19.0–53.0) vs. 41.0 (23.0–54.0), p  = 0.009) and MS (44.0 (30.5–60.0) vs. 49.0 (38.5–64.5), p  < 0.001).

Fig. 2 Comparisons between initial and follow-up clinical outcomes. ( a ) Changes in EDSS scores, ( b ) Changes in PST scores. EDSS = Expanded disability status scale; PST = Processing speed test. Statistical comparisons were conducted using the Wilcoxon signed-rank test.

Comparisons between initial and follow-up clinical outcomes. ( a ) Changes in EDSS scores, ( b ) Changes in PST scores. EDSS = Expanded disability status scale; PST = Processing speed test. Statistical comparisons were conducted using the Wilcoxon signed-rank test.

Correlations between blood biomarkers and clinical outcomes were assessed. Before the analysis, we evaluated the influence of treatment type on baseline blood biomarker levels. No significant differences were observed between B-cell depleting treatment (BCDT) and non-BCDT in patients with NMOSD (sGFAP: 125.9 vs. 126.8 pg/mL, p  = 0.850; sNfL: 11.56 vs. 14.89 pg/mL, p  = 0.172), nor between high-efficacy and low-efficacy treatments in patients with MS (sGFAP: 106.90 vs. 84.49 pg/mL, p  = 0.101; sNfL: 9.93 vs. 9.58 pg/mL, p  = 0.534). In patients with NMOSD, sGFAP was consistently positively correlated with both baseline EDSS scores ( r  = 0.43, p  = 0.005) and future EDSS scores ( r  = 0.44, p  = 0.006). In contrast, sNfL showed a positive association with EDSS scores that did not reach statistical significance ( r  = 0.31, p  = 0.066) (Fig.  3 a). Conversely, in patients with MS, sNfL was positively correlated with both baseline EDSS scores ( r  = 0.25, p  = 0.018) and future EDSS scores ( r  = 0.32, p  = 0.003). Meanwhile, sGFAP did not show a significant association with baseline EDSS scores but exhibited a trend toward a positive correlation ( r  = 0.19, p  = 0.079) and demonstrated a significant correlation with future EDSS scores ( r  = 0.28, p  = 0.010) (Fig.  3 b).

Fig. 3 Correlations between blood biomarkers and clinical outcomes at baseline and 2-year later. ( a ) sGFAP, sNfL–EDSS correlations in NMOSD, ( b ) sGFAP, sNfL–EDSS correlations in MS, ( c ) sGFAP, sNfL–PST correlations in NMOSD, ( d ) sGFAP, sNfL–PST correlations in MS. P values for comparisons of correlation coefficients between baseline and 2-year outcomes were as follows: (a) sGFAP: p  = 0.958, sNfL: p  = 0.682, (b) sGFAP: p  = 0.532, sNfL: p  = 0.617, (c) sGFAP: p  = 0.754, sNfL: p  = 0.757, (d) sGFAP: p  = 0.180, sNfL: p  = 0.005. ** p  < 0.01 for correlation coefficients, baseline vs. 2-year outcomes. EDSS = Expanded disability status scale; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorder; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein, sNfL = Serum neurofilament light chain.

Correlations between blood biomarkers and clinical outcomes at baseline and 2-year later. ( a ) sGFAP, sNfL–EDSS correlations in NMOSD, ( b ) sGFAP, sNfL–EDSS correlations in MS, ( c ) sGFAP, sNfL–PST correlations in NMOSD, ( d ) sGFAP, sNfL–PST correlations in MS. P values for comparisons of correlation coefficients between baseline and 2-year outcomes were as follows: (a) sGFAP: p  = 0.958, sNfL: p  = 0.682, (b) sGFAP: p  = 0.532, sNfL: p  = 0.617, (c) sGFAP: p  = 0.754, sNfL: p  = 0.757, (d) sGFAP: p  = 0.180, sNfL: p  = 0.005. ** p  < 0.01 for correlation coefficients, baseline vs. 2-year outcomes. EDSS = Expanded disability status scale; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorder; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein, sNfL = Serum neurofilament light chain.

For PST scores in NMOSD patients, sGFAP correlated negatively with baseline ( r = −0.51, p  < 0.001) and future ( r = −0.45, p  = 0.010) scores. Notably, sNfL also showed significant negative associations with baseline ( r = −0.50, p  < 0.001) and future ( r = −0.44, p  = 0.012) scores (Fig.  3 c). For PST scores in MS patients, sGFAP correlated with both baseline PST scores ( r = −0.30, p  = 0.004) and future scores ( r = −0.49, p  < 0.001). Conversely, sNfL did not show a significant correlation with baseline scores, but did correlate with future scores ( r = −0.51, p  < 0.001) (Fig.  3 d). No significant differences were found between baseline and future correlations of serum biomarkers and clinical outcomes in NMOSD patients. However, in MS patients, significant differences were observed in the correlations between sNfL and PST scores from baseline to future values ( p  = 0.005) (Table  2 ).

Table 2 Correlations and comparison of correlation coefficient by using fisher transformation. EDSS NMOSD MS Baseline 2-year p * Baseline 2-year p * ( n  = 41) ( n  = 37) ( n  = 91) ( n  = 87) r p r p r p r p
vs. sGFAP
0.43 0.005 0.44 0.006 0.958 0.19 0.079 0.28 0.010 0.532
vs. sNfL
0.22 0.166 0.31 0.066 0.682 0.25 0.018 0.32 0.003 0.617 PST NMOSD MS Baseline 2-year p * Baseline 2-year p * ( n  = 41) ( n  = 31) ( n  = 91) ( n  =  61) r p r p r p r p
vs. sGFAP
−0.51 < 0.001 −0.45 0.010 0.754 −0.30 0.004 −0.49 < 0.001 0.180
vs. sNfL
−0.50 < 0.001 −0.44 0.012 0.757 −0.09 0.371 −0.51 < 0.001 0.005 EDSS = Expanded disability status scale; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorder; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein; sNfL = Serum neurofilament light chain. * p for interaction between baseline and 2-year correlations.

Table 2 Correlations and comparison of correlation coefficient by using fisher transformation. EDSS NMOSD MS Baseline 2-year p * Baseline 2-year p * ( n  = 41) ( n  = 37) ( n  = 91) ( n  = 87) r p r p r p r p
vs. sGFAP
0.43 0.005 0.44 0.006 0.958 0.19 0.079 0.28 0.010 0.532
vs. sNfL
0.22 0.166 0.31 0.066 0.682 0.25 0.018 0.32 0.003 0.617 PST NMOSD MS Baseline 2-year p * Baseline 2-year p * ( n  = 41) ( n  = 31) ( n  = 91) ( n  =  61) r p r p r p r p
vs. sGFAP
−0.51 < 0.001 −0.45 0.010 0.754 −0.30 0.004 −0.49 < 0.001 0.180
vs. sNfL
−0.50 < 0.001 −0.44 0.012 0.757 −0.09 0.371 −0.51 < 0.001 0.005 EDSS = Expanded disability status scale; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorder; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein; sNfL = Serum neurofilament light chain. * p for interaction between baseline and 2-year correlations.

Correlations and comparison of correlation coefficient by using fisher transformation.

EDSS = Expanded disability status scale; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorder; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein; sNfL = Serum neurofilament light chain.

* p for interaction between baseline and 2-year correlations.

Finally, we assessed whether baseline sGFAP and sNfL levels could predict changes in EDSS and PST scores during the study period (ΔEDSS and ΔPST). However, no significant associations were identified in these predictive analyses (Fig.  4 ).

Fig. 4 Correlations between blood biomarkers and changes in clinical parameters. ( a ) sGFAP, sNfL–ΔEDSS correlations in NMOSD, ( b ) sGFAP, sNfL–ΔEDSS correlations in MS, ( c ) sGFAP, sNfL–ΔPST correlations in NMOSD, (d) sGFAP, sNfL–ΔPST correlations in MS. EDSS = Expanded disability status scale; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorder; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein, sNfL = Serum neurofilament light chain.

Correlations between blood biomarkers and changes in clinical parameters. ( a ) sGFAP, sNfL–ΔEDSS correlations in NMOSD, ( b ) sGFAP, sNfL–ΔEDSS correlations in MS, ( c ) sGFAP, sNfL–ΔPST correlations in NMOSD, (d) sGFAP, sNfL–ΔPST correlations in MS. EDSS = Expanded disability status scale; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorder; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein, sNfL = Serum neurofilament light chain.

Clinical attacks and disease progression

During the follow-up period, a clinical attack occurred in four NMOSD patients (9.8%) and six MS patients (6.5%). Baseline sGFAP and sNfL levels showed no significant difference between patients with and without attacks in either the NMOSD (sGFAP: 121.9 vs. 128.2 pg/mL, p  = 0.911; sNfL: 12.22 vs. 16.49 pg/mL, p  = 0.608) or MS (sGFAP: 90.66 vs. 82.71 pg/mL, p  = 0.849; sNfL: 9.83 vs. 8.42 pg/mL, p  = 0.194) groups.

Additionally, six-month confirmed disability progression (6 m-CDP) was observed in two NMOSD patients (4.9%) and six MS patients (6.5%). The ages of NMOSD patients with 6 m-CDP were 69 and 64 years, while the median age [interquartile range] of MS patients with 6 m-CDP was 49.5 [43–59] years (Table  3 ). In NMOSD, all patients with 6 m-CDP experienced worsening ambulation as the primary manifestation of disease progression. Notably, one patient (patient #1) also showed a decline in PST scores during this period. However, MRI evaluations were not conducted at the time of progression to confirm CNS pathology changes. Baseline sGFAP and sNfL levels did not differ between patients with and without 6 m-CDP in either group, with NMOSD showing sGFAP levels of 229.3 pg/mL versus 121.9 pg/mL ( p  = 0.243) and sNfL levels of 32.75 pg/mL versus 12.22 pg/mL ( p  = 0.108), while in MS, sGFAP was 70.34 pg/mL versus 90.66 pg/mL ( p  = 0.502), and sNfL was 9.66 pg/mL versus 9.83 pg/mL ( p  = 0.772).

Table 3 Clinical information of patients with 6 m-CDP. NMOSD MS #1 #2 #1 #2 #3 #4 #5 #6 Age 69 64 70 51 48 48 28 55 Sex F F F F F M F F Disease duration, years 16 9 14 18 23 11 7 13 Age at Last Clinical Attack 67 57 58 33 48 40 25 53 Last Clinical Attack location TM TM ON TM TM Brain Brain ON Education, years 6 9 9 12 12 16 16 16 EDSS scores Baseline 4.5 5.5 3.5 2 3.5 3.5 0 6 Two years later 5.5 6 6 3 5.5 6 2.5 6.5 PST scores Baseline 22 25 7 46 31 53 61 22 Two years later 15 24 13 44 - 55 65 30 Immune modulatory agent MMF AZA INF-b TERI FTY - GA FTY Blood biomarker, baseline sGFAP, pg/mL 232 226.6 467 75.2 62.6 65.5 45.2 145.3 sNfL, pg/mL 46 19.5 36.5 9.3 12.6 8.4 4 10 6 m-CDP = Six-month confirmed disability progression; AZA = Azathioprine; Brain = Brain lesions; EDSS = Expanded disability status scale; FTY = Fingolimod; GA = Glatiramer acetate; INF-b = Interferon beta; MMF = Mycophenolate mofetil; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorders; ON = Optic neuritis; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein; sNfL = Serum neurofilament light chain; TERI = Teriflunomide; TM = Transverse myelitis.

Table 3 Clinical information of patients with 6 m-CDP. NMOSD MS #1 #2 #1 #2 #3 #4 #5 #6 Age 69 64 70 51 48 48 28 55 Sex F F F F F M F F Disease duration, years 16 9 14 18 23 11 7 13 Age at Last Clinical Attack 67 57 58 33 48 40 25 53 Last Clinical Attack location TM TM ON TM TM Brain Brain ON Education, years 6 9 9 12 12 16 16 16 EDSS scores Baseline 4.5 5.5 3.5 2 3.5 3.5 0 6 Two years later 5.5 6 6 3 5.5 6 2.5 6.5 PST scores Baseline 22 25 7 46 31 53 61 22 Two years later 15 24 13 44 - 55 65 30 Immune modulatory agent MMF AZA INF-b TERI FTY - GA FTY Blood biomarker, baseline sGFAP, pg/mL 232 226.6 467 75.2 62.6 65.5 45.2 145.3 sNfL, pg/mL 46 19.5 36.5 9.3 12.6 8.4 4 10 6 m-CDP = Six-month confirmed disability progression; AZA = Azathioprine; Brain = Brain lesions; EDSS = Expanded disability status scale; FTY = Fingolimod; GA = Glatiramer acetate; INF-b = Interferon beta; MMF = Mycophenolate mofetil; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorders; ON = Optic neuritis; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein; sNfL = Serum neurofilament light chain; TERI = Teriflunomide; TM = Transverse myelitis.

Clinical information of patients with 6 m-CDP.

6 m-CDP = Six-month confirmed disability progression; AZA = Azathioprine; Brain = Brain lesions; EDSS = Expanded disability status scale; FTY = Fingolimod; GA = Glatiramer acetate; INF-b = Interferon beta; MMF = Mycophenolate mofetil; MS = Multiple sclerosis; NMOSD = Neuromyelitis optica spectrum disorders; ON = Optic neuritis; PST = Processing speed test; sGFAP = Serum glial fibrillary acidic protein; sNfL = Serum neurofilament light chain; TERI = Teriflunomide; TM = Transverse myelitis.

Discussion

This prospective longitudinal study assessed the prognostic value of sGFAP and sNfL in predicting two-year clinical outcomes in patients with NMOSD and MS. The baseline levels of these biomarkers correlated with both concurrent and future outcomes, showing distinct patterns between the two conditions.

Both sNfL and sGFAP have been recognized in recent years as disease biomarkers for MS and NMOSD. In MS, sNfL levels are valuable for predicting subclinical disease activity, assessing relapse risk, and identifying the development of gadolinium-enhancing lesions 20 , 27 – 31 . Conversely, sGFAP offers additional insights into disease progression, particularly due to its enhanced ability to detect disability progression independent of relapse activity (PIRA) 16 , 19 , 32 . The differential associations of these biomarkers in MS have been attributed to their specific roles: higher baseline sNfL levels are indicative of white matter volume loss, while baseline sGFAP levels correlate specifically with gray matter (GM) loss 19 , 33 , 34 . sGFAP is considered more reflective of a diffuse disease process in the brain, as its elevation is associated with astrocyte proliferation, activation, and potential injury 35 , 36 . Meanwhile, sNfL levels increase to a lesser extent in progressive MS, with these elevations attributed to neuronal loss as part of the progression pathogenesis, subclinical neuroinflammation in chronic active lesions, and changes in normal-appearing white matter 19 , 37 . In NMOSD, sGFAP, as a marker of astrocyte damage, is more closely associated with disease status, including disease disability, relapse activity, treatment response, and future relapse risk 21 – 26 . While sNfL also correlates with disease disability, this association is relatively weaker compared to sGFAP 21 , 23 , 25 .

In this study, simultaneous analysis of these biomarkers in NMOSD and MS patients revealed distinct correlations with clinical parameters, highlighting their differing pathogenesis 1 – 4 , 10 . In MS, sNfL showed a strong correlation with EDSS scores, consistent with its prognostic ability for white matter volume loss. In contrast, in NMOSD, EDSS scores were predominantly associated with sGFAP, reflecting its astrocytopathic nature of the disease, while this was not observed in MS. Aligning its role as a marker for disease progression and potential GM loss 16 , 19 , 32 , sGFAP correlated with both concurrent and future cognitive function in MS and NMOSD, although the association patterns differed between the two diseases. While patients with NMOSD exhibited comparable correlations between blood biomarkers and PST scores, patients with MS demonstrated significantly stronger NfL–PST correlations at future time points than at baseline. This suggests that neurodegeneration in MS may progress more rapidly in patients with severe underlying neuronal damage.

Although NMOSD and MS are both associated with chronic autoimmune inflammation, their progression patterns differ significantly 4 . NMOSD primarily involves attack-related progression with limited inter-attack neurodegeneration 38 , 39 , unlike MS, which involves ongoing, silent neurodegenerative processes that contribute to worsening clinical outcomes over time, independent of clinical relapses. These differing progression patterns may be attributed to distinct brain pathology and pathogenic mechanisms 4 . Cortical GM lesions, a hallmark of neurodegeneration in MS, are also present in NMOSD but differ significantly in structure and composition. In NMOSD, cortical GM lesions are defined by scattered loss of cortical neurons in layers II‒IV, loss of AQP4 immunoreactivity on astrocytes in layer I without complement deposition, microglial activation in layer II, and lymphocytic meningeal infiltration 40 , 41 . Importantly, NMOSD lesions lack cortical demyelination and meningeal B-cell lymphoid follicle-like infiltrates, which are characteristic features of MS brains 40 , 41 . Moreover, astrocyte loss may also explain the relatively limited neurodegeneration in NMOSD. While astrocyte loss initially causes excitotoxic neuronal and axonal damage, activated astrocytes in MS promote inflammation, tissue injury, and chronic neurodegeneration 42 , 43 . In NMOSD, acute astrocyte loss contributes to focal neurodegeneration in active lesions 44 , whereas chronic astrocyte activation in MS drives progressive neurodegeneration.

Clinical progression outcomes in real-world settings also differ between NMOSD and MS 38 . Recent studies report a PIRA prevalence of less than 3% in NMOSD 39 , 45 , significantly lower than the approximately 5% annual PIRA rate in MS, where PIRA accounts for over 50% of disability accrual events 46 . In our cohort, the prevalence of 6 m-CDP was lower in NMOSD patients (4.9%, 2 patients) than in MS patients (6.5%, 6 patients) but exceeded rates reported in prior studies 39 , 45 . This elevated rate should be interpreted with caution, given the small sample size and potential confounding factors, including age-related changes such as muscle disuse atrophy. Notably, NMOSD patients who developed 6 m-CDP exhibited higher baseline levels of sGFAP (229.3 pg/mL vs. 121.9 pg/mL) and sNfL (32.75 pg/mL vs. 12.22 pg/mL) compared to those without 6 m-CDP. Further research with larger cohorts and extended follow-up is needed to validate these observations.

The correlation of PST scores with both sGFAP and sNfL in NMOSD is noteworthy. Cognitive dysfunction in NMOSD has received less attention, and the usefulness of PST or symbol-digit modality tests in NMOSD is not well established as in MS 47 . However, the consistent associations of PST scores with neuronal and astroglial damage markers emphasize the importance of incorporating these tests in NMOSD care. The observed improvement in PST scores over time may be attributed to a practice effect from repeated testing 48 , 49 , rather than genuine cognitive improvement. Longer-term studies with larger sample sizes are needed to clarify the clinical significance and implications of these findings.

There are several limitations to this study. The relatively small number of participants may have influenced the strength and reliability of the correlations between baseline sGFAP and sNfL with the occurrence of clinical relapses or 6 m-CDP in both NMOSD and MS patients. The limited number of participants and the infrequent occurrence of significant clinical deterioration may have contributed to the absence of significant associations between baseline blood biomarkers and clinical events or changes in clinical parameters (ΔEDSS and ΔPST) during the study period. Nevertheless, distinct biomarker correlations based on disease type and specific time points were still identified. Another limitation is the exclusion of patients with progressive MS, which may have reduced the ability to detect stronger associations between the biomarkers and future clinical outcomes in the MS group. An age imbalance between two disease groups should also be noted. Considering that age can impact neurological status, cognitive function tests, and the relationship between disease biomarkers and clinical parameters 10 , the 6-year median age difference between the groups may have influenced the observed differences between NMOSD and MS, warranting a cautious interpretation. Finally, the lack of analysis correlating MRI findings with blood biomarkers represents another constraint of the study.

In conclusion, sGFAP and sNfL were associated with both concurrent and future outcomes in NMOSD and MS patients, but the patterns differed, suggesting distinct pathogenesis and neurodegeneration in each disease. These findings provide further insights into the mechanisms of CNS demyelinating diseases and underscore the importance of disease-specific interpretations of biomarkers.

Methods

Participants and study design

From October 2019 to July 2021, we prospectively enrolled patients with either MS or NMOSD who visited the department of Neurology at a tertiary medical center (Asan Medical Center, Seoul, Korea). The study adhered to the principles of the Declaration of Helsinki and was approved by the institutional review boards and ethics committee (approval No. 2020 − 1678). Written informed consent was obtained from all participants before enrollment.

The diagnoses of NMOSD and MS were made in accordance with established consensus criteria 50 , 51 . Specifically, NMOSD diagnosis required the presence of aquaporin-4 (AQP4) antibodies detected through a cell-based assay. Patients who had experienced a clinical relapse within 2 months prior to enrolment or had neurological impairments that could interfere with cognitive testing, such as severe visual impairment (20/200 or worse) or significant limb paralysis, were excluded from the study.

Blood samples were collected at enrolment, and neurological disability was evaluated using the Expanded Disability Status Scale (EDSS). The Processing Speed Test (PST), an iPad-based tool simulating the Symbol Digit Modalities Test was also conducted with additional informed consents 52 . EDSS and PST evaluations occurred at 6-month intervals. Clinical relapses and 6-months confirmed disability progression (6mCDP), defined as a sustained increase in the EDSS score for at least 180 days since the previous visit of ≥ 1.0 point from an EDSS score of 1.0–5.0 or ≥ 0.5 point from an EDSS score of ≥ 5.5 53 , were also investigated.

Blood biomarkers

Serum samples were gathered and preserved at a temperature of −80 °C following standardized protocols. Only samples that had been thawed once were used in the experiments, and they were carefully thawed immediately before analysis. NfL and GFAP serum levels were assessed in duplicate using a Simoa™ HD-1 Analyzer (Quanterix, MA, USA) at PrismCDX (Gyeonggi-do, South Korea) by a clinical data blinded investigator.

Statistical analysis

We depicted the clinical traits of the patients using the median and interquartile range for continuous variables and frequency (%) for categorical variables. We compared clinical variables between MS and NMOSD patients using the Mann-Whitney U test for continuous variables and the Chi-square test for categorical variables. EDSS, PST scores, and serum biomarker levels for each group were compared using ANCOVA with age as a covariate. Longitudinal analyses were performed on patients who underwent evaluation for both EDSS and PST at baseline and approximately two years later. Paired values of EDSS and PST scores (baseline and two-year) were compared using the Wilcoxon signed-rank test due to their non-parametric distributions. Pearson’s correlation test was used to examine the relationships between log-transformed serum biomarker levels and EDSS or PST scores. We compared the statistical differences between baseline correlations and those at the two-year follow-up using Fisher’s Z transformation. We also evaluated the correlations between log-transformed serum biomarker levels and changes in clinical parameters (EDSS and PST scores) over the study period. All statistical analyses were performed using GraphPad Prism version 10.1 and SPSS Statistics version 28.0.