Apheresis treatment in autoimmune neurological diseases: Predictors of good clinical outcome and success of follow-up therapy with B-cell depletion

Objective Apheresis treatment (AT) is an established standard of treatment in various neurological autoimmune diseases. Since not all patients equally benefit from AT, we saw the need to investigate the effect of different clinical, paraclinical and technical-apparative factors on the clinical outcome. Additionally, we wanted to find out whether patients who improved due to AT continue to be clinically stable under B-cell depletion (BCD). Methods We screened all patients (n=358) with neurological diseases who received AT at the Medical center of the University of the Saarland in the past 20 years. Different factors (e.g., age, sex, duration until onset of AT, type of AT, number of cycles, csf parameters) were analyzed retrospectively. Clinical disability was measured using the modified Rankin scale (mRS), visual acuity and the Expanded Disability Status Scale (EDSS).


Introduction
Apheresis treatment (AT) is an established standard of treatment for autoimmune, mostly antibody-mediated diseases [1].The debate of a potential benefit of apheresis procedures in long-COVID syndromes [2,3] has renewed the focus of the scientific community on the as of now poorly understood factors predicting outcome of AT in other autoimmune neurologic diseases in which autoimmune processes, parainfectious or paraneoplastic mechanisms are discussed as underlying causes.
Therapeutic plasma exchange (TPE) is to be distinguished from immunoadsorption (IA) [4].In TPE, plasma volume is separated from other blood components and replaced by substitute solutions (usually human albumin or Fresh Frozen Plasma (FFP)).This procedure unselectively removes (exchanges) multiple proteins such as antibodies as well as clotting factors, hormones, or various kinds of drugs.In contrast, in IA, immunoglobulins, complementary factors or immune complexes of plasma are selectively bound in an adsorber column.Both procedures are to be understood as a short-term intervention into the immune system and are used in severe neurological -mostly steroid-refractory -autoimmune diseases, primarily used for disease stabilisation.The expected extent of success of AT in various autoimmune diagnoses is currently the subject of research projects.The decrease in immunoglobulins due to the procedure depends on the type of immunoglobulins considered as well as intravascular baseline concentration [5], re-synthesis, redistribution from extracorporeal compartments, exchanged plasma volume and possibly the length of a therapeutic cycle and the time interval between cycles [6].
In TPE proteins have to be replaced (with corresponding allergic risk) and there is a greater circulatory burden.Neither is the case with IA.TPE is equivalent to a non-selective AT, whereas IA is more selective, and a larger TPV can be treated per session.
With TPE 1 to 1.5 times the patient's total plasma volume can be treated per session, with IA with tryptophan filters 2 to 2.5 liters can be treated.A standard of about 5-7 sessions per AT cycle is established [1,7].
Established and modified indications for evidence-based use of AT are made considering American Society for Apheresis (ASFA) guidelines which are continuously updated [1].The

J o u r n a l P r e -p r o o f
Journal Pre-proof guideline serves as a guide in the treatment of patients, sorting the disease into four categories (CI, CII, CIII, CIV): CI and CII include evidence-based firstline (CI) or secondline therapy (CII) via AT (in each case combined with other therapeutic modalities, if appropriate), whereas in CIII the role of AT remains unclear and in CIV it has been shown to be ineffective or even harmful [1].In some diseases (e.g., MG), long-term treatment with AT is necessary if other treatment options fail [8,9].However, AT is generally used in cases of (sub)acute clinical deteriorations.In a vast variety of immunologically mediated neurological diseases [10][11][12] such as Guillain-Barré syndrome [13][14][15][16], chronic demyelinating inflammatory polyneuropathy (CIDP) [17][18][19], relapsing multiple sclerosis (RMS) [20,21], myasthenia gravis (MG), autoimmune encephalitis (AE) or NMOSD [22,23] indication for AT can be given as first-line therapy or as a second-line treatment after unsuccessful treatment with steroid pulses [1].
We saw the need to investigate the effect of clinical, paraclinical, and technical-apparative factors in patients with different neurological diseases, because not all individuals benefit from AT.The influence of different variables [20] such as patient age, gender, duration until onset of AT, number of cycles, type of relapse has not coherently been investigated in the largely retrospective studies or case series.Existing studies usually only provided results for individual diseases.However, in large neurological departments the treatment of rare diseases must also be taken into account.Therefore, we decided to examine the entirety of patients treated with AT in accordance with the existing diseases, the paraclinical and ATrelated parameters.In addition to the parameters mentioned above the response to the different types of AT was of greater interest, as IA does not require plasma substitution and there are particular cases in which IA is more expensive than TPE.Additionally, the role of seropositivity or seronegativity for autoimmune antibodies has not yet been sufficiently investigated when it comes to AT due to small case sizes.
To assess the outcome for all subjects, functional disability was determined via the modified Rankin scale (mRS).Additionally, in some subpopulations the Expanded Disability Status Scale (EDSS) and visual acuity were examined.

J o u r n a l P r e -p r o o f
Journal Pre-proof Since many neurological autoimmune diseases present with a chronic disease course after AT, there often is a need for follow-up therapy with immunotherapeutics after the initial use of AT.We were interested to investigate whether patients who had received AT benefit from B-cell depletion (BCD), since ultimately both therapeutic concepts aim at reducing antibodymediated autoimmunity.Pharmacological destruction of B cells is accompanied by several monoclonal antibodies directed against CD19 or CD20.The original substance is Rituximab, which is used in hemato-oncology.Rituximab which is directed against CD20 is also used as an off-label medication in neurology (for RMS, PMS, MG, vasculitis, NMOSD, MOGAD, AE and other diseases as an individual healing attempt).Approved follow-on products against CD20 are ocrelizumab (for RMS and PPMS), ofatumumab (for RMS), ublituximab (for RMS) and inebilizumab against CD19 (for Aquaporin-4 antibody positive NMOSD) [24][25][26][27][28].The substances directed against CD20 deplete a large part of B cells, while inebilizumab also destroys early forms of B cells and some types of plasma cells [24,28].

J o u r n a l P r e -p r o o f
Journal Pre-proof

Study design and participants
Every Patient who was in the Department of Neurology from January 2004 to September 2023 and received AT (TPE or IA) was analyzed retrospectively and included in this study.
Both methods are based on prior plasma separation by means of a plasma filter or centrifuge.
Plasma exchange was performed via a large central venous catheter.In both cases, tryptophan was used for the binding before the residual blood plasma is re-infused [4,5,29] In most cases anticoagulation with heparin or citrates was administered during AT.
Calculation of the individual total plasma volume (TPV) was achieved by specifying hematocrit, body height and weight by use of formulas and nomograms [30].
The study was anonymized and conducted with approval of the local ethics committee (Ethikkommission der Ärztekammer des Saarlandes, ethics vote No. 158/14).Active consent was not required.The analysis was based on data documented by the department of neurology and nephrology.
Clinical parameters (diagnoses, age, sex, examination findings, visual acuity, symptom duration) and supportive instrumental examinations as well as common disability scores before and after AT were evaluated (mRS for all, EDSS for myelitis, ON, RMS, PMS and NMOSD/MOGAD, visual acuity for ON of different diseases).
The seven-item modified Rankin scale (mRS), which was originally developed to evaluate outcome post stroke and is utilized in many stroke studies, reflects the extent of disability using a standardized measure [31][32][33].This score is beneficial when it comes to determining functional disability in everyday life, which is the decisive factor for all patients, before and after AT regardless of the disease.In the meantime, mRS has been used for a variety of other neurological diseases [24,[34][35][36][37].An mRS of 0-2 points reflects a lack of disability or moderate disability, whereas patients with an mRS of 3-5 points permanently depend on assistance (mRS of 6 points corresponds to death due to illness).
Individual diagnoses were combined into 11 diagnostic groups, as shown in Table 1.We also investigated the proportion of ineffective previous steroid therapy.Cerebrospinal fluid (CSF)

J o u r n a l P r e -p r o o f
Journal Pre-proof findings were assessed if available (cell count, pleocytosis, oligoclonal bands (OCB)).
Laboratory results (CRP, antibody levels for IgG and IgM in serum) and the presence of pathological autoantibodies were also included in the analysis.AT specific data (type of AT, number of cycles, exchange volume) was collected.The AT was performed in the Department of Nephrology.Tryptophan filters by the company Asahi Kasei Medical (Japan) were used.In TPE, exchange volume was calculated with 40 ml/kg body weight.The exchange volume in IA was 2.5 liters for all patients.

Statistical analysis
To compare unpaired non-parametric data between the different groups (diagnosis group, sex, presence of ocb/pleocytosis/autoantibodies, type of apheresis, stability under bcd) Mann-Whitney-U-and Kruskal-Wallis-test were performed.In case of multiple testing, Bonferroni correction was used as post hoc-test.Wilcoxon-rank-test was used for statistical analysis of paired non-parametric data (single group benefit in mRS due to AT).Correlations between age, time to treatment, number of sessions, plasma exchange volume on the one hand and ΔmRS on the other hand were analyzed using Spearman's correlation coefficient.A statistical analysis was considered significant if the p-value was < 0.05.

Software
SPSS statistics (version 28.0.1.1)was used for statistical analysis, GraphPad Prism (Version 10.0.2),Microsoft Excel and Power Point (Microsoft 365) were used to create the graphics, and Microsoft Word and Excel (Microsoft 365) were used for word processing.

J o u r n a l P r e -p r o o f
Journal Pre-proof

Results
(K1) Screening procedures and final selection 358 patients received AT in the Department of Neurology from January 2004 to September 2023.After consideration of the final diagnoses, 23 patients were assigned to category IV according to the current American Society for Apheresis (ASFA) guidelines.As no autoimmune pathology and no therapeutic response is assumed in these patients, they were excluded from the sub-analysis below.
Fig. 1 Flowchart with the screened patients.Patients with ASFA category IV diagnoses were excluded from the sub-analysis if no therapeutic response was expected.The remaining patients were assigned to ASFA categories I, II and III.The distribution of the various AT procedures is shown below.

(K2) Clinical characteristics of the included population
A total of 335 patients, 178 female (53,1%), were included in this study.Table 1 provides an overview of patient characteristics, laboratory findings, and apheresis treatment parameters.
The mean age of all patients was 53.6 years (±17.4) with a minimum of 18 and maximum of 86 years.Patients with RMS were of the youngest age (37.8 years (±10.4)),patients with autoimmune neuropathy (AN) had the highest average age of 61.9 years (±9.4).Patients included in this study with RMS, PMS, ON, Myelitis, NMOSD/MOGAD or AE suffered from a steroid-refractory worsening of the disease.Over 2/3 disability-causing events were due to AIDP, CIDP, multiple sclerosis, and myasthenia gravis.The events optic neuritis (ON) and

J o u r n a l P r e -p r o o f
Journal Pre-proof myelitis could be assigned to different diseases.In almost 45% of the patients, the neurological manifestation represented the initial diagnosis of the disease.Pathological autoantibodies were found in about 25% of the reported cases (Table 2 contains the name and frequency distribution of the autoantibodies).CSF pleocytosis and/or the detection of oligoclonal bands, both expressing possible inflammation in the CSF space, were found in 16.1 and 14.9 % respectively.A total of 2669 cycles of AT (of which 1461 cycles of TPE and 1208 cycles of IA) were evaluated.
Overall the AT was well tolerated.The most common side effects were allergic reactions after substitution of FFP (8.6%) and hypotension (4.2%).Three patients died of their disease during the hospital observation phase; however this was not a consequence of AT.Two patients experienced an air embolism after removal of their central venous catheter.No other serious adverse events or deaths occurred.Plasma substitution was performed using albumin and FFP.A slight majority of patients received therapeutic TPE.Combination therapy was mostly initiated by switching to IA due to allergic reactions to FFP during TPE. 5 cycles were usually planed with the option of discontinuation in case of deterioration or additional cycles in case of a favourable response; on average, 8 cycles of AT (± 6.4) were performed.

J o u r n a l P r e -p r o o f
Journal Pre-proof   Includes primary progressive multiple sclerosis (PPMS) and secondary progressive multiple sclerosis (SPMS) with progression independent of relapse activity (PIRA). 4includes the diagnoses thrombotic thrombocytopenic purpura, Lambert-Eaton myasthenic syndrome, paraneoplastic cerebellar degeneration and neuropathy, stiff person syndrome, neurosarcoidosis.However, our evaluation takes into account the final diagnoses (Table 1 and Fig. 2).Each disease was assigned to a category (I to IV) according to the current American Society for Apheresis (ASFA) guidelines [1]; notes in table S1 in the supplement.The diagnoses polyradiculitis, vasculitic and other autoimmune polyneuropathy as well as ganglionitis and neurosarcoidosis were assigned to category CIII due to their autoimmune pathophysiology and the therefore possible but not yet sufficiently investigated efficiency of apheresis therapy.
The assignment allows an estimation of the probability of response to the AT.
A total of 358 patients were screened (Figure 1), the Categorization was made according to the ASFA guideline: CI 197 patients, CII 104 patients, CIII 34 and CIV 23 patients.Figure 2  Patients in green (dark green ΔmRS>1, light green ΔmRS=1) benefited best, gray (ΔmRS=0) and red (ΔmRS<0) symbolize unchanged or worsened condition.In line with good clinical practice, the categories with the best therapeutic response according to ASFA (CI and CII) contained the most patients.Deterioration under AT was found in the CI and CIV categories.
No patient with improvement was found in CIV.With ascending category, the proportion of patients with response (green) decreased with CI 66.5%, CII 62.5%, and CIII 38.2%.The proportion of patients with a very good outcome (dark green) was highest in CII with 22.1% and lowest in CIII with 5.9%.As expected, all 23 patients of CIV showed a lack of improvement or even a worsening of the mRS.Patients in category IV with non-autoimmune-mediated diseases were excluded for further analysis in the absence of an expected response to AT.With the comparison of the ASFA group.The PMS group showed a significant result, but the group size was too small for a meaningful statistical analysis.S2 and S3 in the Supplement.

J o u r n a l P r e -p r o o f
Furthermore, we investigated the influence of the type of AT on the extent of functional improvement (mRS).We found no significant difference between patients treated with TPE (n=175), IA (n=138) or with a combination of both procedures (n=22

Alternative disability scores: EDSS and visual acuity
In the observed patient population of patients diagnosed with RMS, PMS, NMOSD/MOGAD, ON or Myelitis, we found a highly significant correlation (Spearman's r=0.852, p<0.001) between the change in mRS and the change in the EDSS (EDSS) due to the AT.EDSS after AT was significantly lower than before (median before and after AT=3, mean 4.1 resp.3.4, p<0.001,Wilcoxon rank test).
However, with the (para-)clinic and apheresis related parameters (age of patient, time to treatment, pleocytosis, OCB, type of AT, number of cycles, and exchange volume) better correlations were shown using mRS as the clinical outcome parameter for the abovementioned diagnoses.
As various diseases can lead to optic neuritis, we used the parameter of visual acuity determined by ophthalmologists.With regards to the parameter of visual acuity change due to AT ( visual acuity), we also found significant improvement after undergoing therapy (median 0.1 resp.0.25, mean 0.18 resp.0.36, p < 0.001, Wilcoxon rank test).No Group differences were found in the Kruskal-Wallis test, therefore the group sizes were too small.Similar to the EDSS, the change in the visual acuity correlated significantly with the mRS as well (Spearman's r=0.907; p<0.001).Time point of prior steroid therapy did not influence the improvement of visual acuity in the given cohort (Spearman's r=-0.09;p=0.81), neither did the dose of it (Spearman's r=-0.35;p=0.32).

J o u r n a l P r e -p r o o f
Journal Pre-proof

(K5) Benefit of apheresis treatment as an indicator of stability under B cell depletion
After initial AT, 76 of 335 patients received long term B cell depletion (BCD), using rituximab, ocrelizumab or ofatumumab.Concerning these patients, we analyzed if initial profit through AT correlates with later disease stability under BCD, which is defined as a constant or improved mRS one year after starting therapy.We had follow up data on 50 of the 76 patients undergoing b cell depletion.The patient characteristics can be found in Table S4 in the Supplement.
Figure 4 shows the course of the mRS values of the 50 patients in the color beam.Initially, there is a clear improvement in disability due to AT (decrease in the mRS sum score by 43 points) and stabilization during the variable time between the end of AT and the start of BCD (mean time between end of AT and start of BCD was 14.7 days).In the follow-up after one year of BCD, however, there was again a worsening of disability (increase in the mRS sum score by 17 points).
Patients who initially showed response to AT had a significantly more stable course through later BCD in the statistical analysis using the Mann-Whitney U test (p=0.039)(Tab.5).
In particular, patients with RMS and AE who benefited from AT appear to be stable under BCD.However, patients with NMOSD/MOGAD or MG/LEMS who benefited from AT show a variable response to BCD (Tab.S4).In the post hoc analysis, the AE subgroup was significantly more stable than the NMOSD/MOGAD and MG/LEMS subgroups (p<0.001 in each case).J o u r n a l P r e -p r o o f Journal Pre-proof

Discussion
The complication rate of AT was very low in our study, most frequently allergic reaction to FFP.The data were congruent with the known good tolerability [38].
In our analysis, we were able to reproduce the benefit of AT for the diagnoses that could be assigned to ASFA categories I and II.Likewise, the demonstration of unchanged or worsened mRS for all individuals in ASFA category IV, for whom no therapeutic response was expected according to ASFA criteria, allowed the exclusion of these patients from the detailed followup analysis [1].While there was no difference between groups CI and CII in terms of mRS outcome, both groups were significantly superior to CIII.This is probably due to the heterogeneous composition of this group, in which the proportion of patients with a very good response was lowest, but on the other hand no patients deteriorating under AT were included.
The epidemiological data mean age of 53 years and predominance of the female sex is consistent with other AT studies in neurological diseases [11,39] and were not significantly influencing variables.In contrast, younger age appears to be advantageous in other studies for individual subgroups, e.g.AE [37].We could show a similar trend for the RMS subgroup.The CSF parameters presence of pleocytosis or detection of intrathecal OCB did not play a relevant role for the outcome and is consistent with the fact that inflammatory signs cannot be detected in CSF in all diseases.
While smaller studies have reported varying response rates of more than 50% to almost 90% [11,12,39].Our patients from ASFA CI, II, III showed a similar improvement but the different diagnoses and heterogeneous composition of the collective and requires more detailed data analysis: In the disease-related subgroup analysis, patients with AIDP, RMS, CIDP, MG/LEMS and NMOSD/MOGAD (each p<0.001) and AE (p=0.002) and AN (p=0.011)showed the most significant improvement (even if before and after data on instrumental diagnostics (e.g.MRI) were unfortunately not available).This reflects the results of other studies [11,[39][40][41][42], some of which, in contrast, have shown no therapeutic response in CIDP [11,12].In this disease group and MG/LEMS, a faster onset of AT (time to treatment) was associated with a better

J o u r n a l P r e -p r o o f
Journal Pre-proof outcome.It should be noted here that the time of deterioration tends to be recorded by patients as a rough time period.Diagnosis groups "myelitis of other kinds" and "other" did not appear to benefit from AT, which is probably due to the heterogeneous composition.The CIDP group was the only group in which a higher number of AT sessions or a higher exchanged plasma volume was beneficial for the functional outcome (mRS).
Optic neuritis occurs in various diseases: isolated, in RMS [21], in MOGAD, in NMOSD, and double-negative NMOSD.Especially in the latter group, TPE does not seem to lead to a reliable improvement of vision, whereas a combination of rapid onset of TPE and few previous ON episodes in Aquaporin-4 antibody positive NMOSD patients show the best chances of improvement [43].In our study, visual acuity was significantly better after AT, but there were no statistically significant differences between the subgroups.
We saw no difference in outcome when using TPE, IA or a combination of both (p=0.29).
However, smaller studies may suggest an advantage for IA (better effect and/or fewer complications) for individual disease groups [18].There is evidence that TPE and IA lead to similar clinical improvements in MS and NMOSD patients [1], with levels of proinflammatory cytokines (Interferon-y, Tumor-necrosis-factor-α, Interleukin-12 and others) appearing to be lowered during IA [44].
We were able to show that the subgroup of patients with pathological autoantibodies (26.9%), regardless of whether the cell function is impaired, benefited more from AT than patients with no successful antibody detection.The most common antibody detected was anti-acetylcholine-Ab in MG patients (10.4%).It is already known that MG patients like in our collective show a rapid regression of symptoms and a decrease in acetylcholine antibody titres due to IA [45].
Unfortunately, the number of autoantibody-positive patients was too small to be able to make a statement on the improved effectiveness of TPE, IA or combination.
With regard to diseases triggered by autoantibodies, there is an important follow-up aspect for treatment: AT is mostly aimed at the treatment of autoimmune diseases in which pathological autoantibodies are detected or suspected.However, B cells, which mature into antibody-producing plasma cells, are not affected by AT.Thus, from a pathophysiological J o u r n a l P r e -p r o o f Journal Pre-proof point of view, it appears conclusive that patients who respond to AT should also stabilize under BCD treatment.For the first time, we investigated the relationship between response to AT and subsequent stability under BCD in neurological diseases: We were able to show that the majority of patients who clinically benefit from AT (mRS≥1) are also stable or improve under BCD (mRS≥0).This fact must be interpreted with caution due to the limited long-term data of only one year of BCD to date and the small patient population.Nevertheless, the combination of short-term antibody reduction (through AT) and the long-term targeted reduction of pathological antibodies through BCD appears to be a promising therapeutic concept.
Of course, this study has some weaknesses.For example, it is a purely retrospective data analysis, which can lead to distortions in the interpretation.The same applies to the fact that the respective diagnoses were made by different physicians over a long period of time, not by a single physician in a standardised manner.Larger populations need to be studied in order to take into account the heterogeneity of the different diagnoses and also the different BCD therapies, which influence the B-cell populations and antibody-levels differently [24,26].
A prospective design and the additional use of disease-specific scores would be desirable, even though the mRS correlated well with functional disability across diseases [12,37,39], and with EDSS and visual acuity in our own cohort.For now, we were able to show that EDSS and visual acuity are good parameters for detecting AT-related improvement in suitable patients.

Conclusion
Our real-world data from the large cohort of 335 included patients and more than 2500 cycles of TPE and/or IA reflects the benefit of AT in many neurological autoimmune diseases.
In particular, functional disability, measured via the mRS, improves in patients with AIDP, RMS, CIDP, MG/LEMS, NMOSD/MOGAD, AE, but to a lesser extent also in ON, AN and PMS.The presence of pathological autoantibodies in the serum is a predictor for a better response to TPE, IA or a combination, in contrast to the detection of inflammatory findings in the CSF.
Patients with a response to AT are significantly more stable in long-term treatment with BCD.
J o u r n a l P r e -p r o o f

2 without a diagnosis
of multiple sclerosis, NMOSD or MOGAD.
is a multidimensional diagram and represents an alternative visualization to a flow chart.It symbolizes each individual's response to AT.The same numbers are assigned to the same diseases.The outcome was colored by the improvement or worsening of mRS (ΔmRS).

Fig. 4
Fig. 4 Representation of the distribution of 50 patients on the mRS values 0-5 in the color scale (green, yellow, orange, red) over the course of the study at four points in time.Low mRS values (green) reflect low disability and high values reflect severe functional disability (orange/red).The absolute number of patients per mRS value is shown; the percentage distribution is presented below.The values are measured at the time before AT (mRS pre-AT), after completion of AT (mRS post-AT), before the start of BCD (mRS pre-BCD) and after one year of therapy with BCD (mRS 1Y-post-BCD).Decrease (mRS total score) due to AT and slight increase in total disability in the follow-up after one year of BCD.
Correlation of the time interval from symptom onset to the start of AT (time in days (d)) with mRS per disease group.Expressed as mean ±SD.A negative correlation factor Spearman's r means that a shortened time interval leads to an improved outcome. 3ab.3