Results of Cerebrospinal Fluid Analysis in 108 Patients With Progressive Multifocal Leukoencephalopathy

Background: Progressive multifocal leukoencephalopathy (PML) is caused by an opportunistic infection with JC polyoma virus (JCPyV) and mainly affects immunocompromised patients. It leads to pronounced demyelination of the central nervous system (CNS) resulting in severe disability or even death. Detection of JCPyV DNA in the cerebrospinal uid (CSF) is usually accepted as proof for the diagnosis of PML. Values from routine CSF parameters, like CSF cell count, protein concentration, Qalbumin levels, or intrathecal immunoglobulin synthesis are mostly considered as normal; however, this has not been investigated systematically. Methods: We therefore analyzed those standard CSF parameters in a cohort of 108 PML patients that were treated at four different neurological centers in Germany. The patients exhibited different underlying conditions with natalizumab treatment in multiple sclerosis (n=54) and human immunodeciency virus (HIV)-infection (n=25) being the most frequent. The data were collected at the respective centers in accordance with local requirements and then jointly analyzed. The results of the total PML cohort were compared with a control group of patients with normal pressure hydrocephalus (NPH) and idiopathic intracranial hypertension (IIH) or an HIV group without PML, respectively. Results: The PML group showed an elevated cell count (p<0.001) compared to the control group, however, this effect was mainly driven by HIV-PML patients. This subgroup also demonstrated a signicantly higher proportion of patients with a disturbed blood-CSF-barrier function. Immune reconstitution syndrome (IRIS) occurred in 41/108 patients and was characterized by a trend for an increase in CSF cell count (p=0.052), CSF lactate (p=0.052), and an augmented intrathecal immunoglobulin synthesis. Conclusions: This comprehensive, retrospective study on diagnostic results in PML patients provides insight into the CSF ndings in patients with PML. It demonstrates that CSF changes in PML patients may be specic for the underlying condition that predisposes for the development of PML and thus data

early diagnosis and consequently a prompt start of therapy/cessation of causal immune suppression could improve the prognosis. Unfortunately, the diagnosis of PML is often delayed. Potential reasons are the lack of speci c radiographic features in brain magnetic resonance imaging (MRI) and the fact that CSF JCPyV-PCR may be negative, particularly at an early stage of the disease. The search for a potential blood-or CSF-biomarker for PML, especially in CSF negative patients, has not been successful so far (Möhn et

Study design and setting
Patients were treated at the Department of Neurology at Hannover Medical School (n=50), St. Josef Hospital Bochum (n=38), University Hospital Münster (n=19), and University Hospital Cologne (n=12). Five patients who suffered from HIV infection were excluded for further analysis because three of them were diagnosed with additional cerebral toxoplasmosis and in two patients PML diagnosis was only suspected. Another six patients had to be excluded due to insu cient data quality, resulting in n=108 patients available for nal analysis. In all patients that were diagnosed with PML other viral CNS infections such as varicella zoster virus (VZV)-, herpes simplex virus (HSV)-, Epstein-Barr virus (EBV)-, or cytomegalovirus (CMV) encephalitis had been ruled out. Results of the PML cohort were compared with an age-matched control group consisting of patients with normal pressure hydrocephalus (NPH) (n=8) and idiopathic intracranial hypertension (IIH) (n=13) (suppl. table 2). In addition, the data of the HIV subgroup were compared with an HIV control group (n=37) whose patients had undergone a lumbar puncture for other reasons, for example cognitive de cits, suspected encephalitis, suspected vasculitis, unexplained encephalopathy, or seizures (suppl. Diagnostic procedures CSF and serum were analyzed by routine methods (Skripuletz et al, Dis Markers 2014). CSF cell count, total protein, and lactate were analyzed immediately after CSF withdrawal by lumbar puncture. CSF cells were counted manually with a Fuchs-Rosenthal counting chamber. For further analyses the residual CSF was centrifuged (145 g for 15 min) and the supernatant frozen at −70 °C. The function of the blood-CSF barrier was estimated as CSF/serum albumin quotients (Qalbumin). The age-adjusted upper reference limit of Qalbumin (Qalb) was calculated using the formula Qalb = 4 + (age in years/15) (Reiber, Mult Scler 1998). CSF oligoclonal bands were determined by isoelectric focusing in polyacrylamide gels with consecutive silver staining. Five patterns of oligoclonal bands were distinguished following the recommendations of the rst European consensus on CSF analysis in multiple sclerosis (Andersson et al, J Neurol Neurosurg Psychiatry 1994).

Statistical analysis
Statistical analysis was performed using GraphPad Prism software version 8.0 (GraphPad Software, San Diego, USA). The Mann-Whitney U-test as the nonparametric equivalent of the t-test for independent samples was used to compare the individual groups. For this test the data does not need to be normally distributed, the variables only need to be ordinally scaled. Data are displayed as means and standard error of the mean (SEM) or standard deviation (SD), respectively. P < 0.05 was considered statistically signi cant.

Patients characteristics and standard CSF parameters
In total, the CSF results of 108 patients from four centers were analyzed. As many patients received several lumbar punctures, a total of 316 CSF analyses were considered. The patients' mean age was 48 years, with a range from 19 to 81 years. The underlying diagnoses included multiple sclerosis treated with natalizumab (MS/NTZ, n = 54), HIV infection (n = 25), and hematological diseases such as B-cell non-Hodgkin lymphoma (n = 10), multiple myeloma (n = 2), chronic lymphocytic leukemia (n = 6), and acute myeloid leukemia (n = 1) (referred to collectively as lymphoma group). Ten patients of the lymphoma group were treated with rituximab mono-or combination-therapy. Other chemotherapy regimens included bendamustine, melphalane, mitoxantrone, or methotrexate. In four of the patients an organ transplant had been performed and they received immunosuppressive therapy with tacrolimus, mycophenolate mofetil, or ciclosporin. One patient each suffered from bronchial carcinoma, sarcoidosis, microscopic polyangiitis and common variable immunode ciency (CVID). In two cases no explanatory underlying disease was found. Detailed information about the individual patients can be found in supplementary table 1. Regarding the standard CSF parameter of the rst lumbar puncture (Table 1), 24/108 patients (22%) had an elevated cell count and 35/108 patients showed an elevated Qalbumin indicating a disturbed blood-CSF-barrier. The mean lactate content was 1.64 mmol/l (range: 0.99-2.8 mmol/l), whereby 6/108 (6%) subjects presented with increased lactate levels.  (Fig. 3.

Analysis of standard CSF parameters in subsequent lumbar punctures
Several patients received subsequent lumbar punctures after PML was initially diagnosed. The maximum number of lumbar punctures in one patient was 26. To investigate whether the results of CSF analysis changed over time, the average time between the rst lumbar puncture and subsequent punctures was calculated. In total, up to the rst 9 punctures were considered. The results of the further punctures (10 to max. 26) were not considered representative for the total cohort because of too few patients. None of the parameters changed signi cantly over time or in the course of subsequent punctures (Fig. 4).

CSF cell distribution in PML patients at rst lumbar puncture
In 85 patients of the total cohort a differentiation of cell distribution was performed at the rst lumbar puncture. Sixty-seven patients (79%) showed a lymphocytic predominance while in 10 patients (12%) the majority of cells was monocytic. Six patients (7%) exhibited a mixed cell distribution and only two patients (2%) demonstrated mainly granulocytes within the CSF. The latter is best explained by arti cial blood admixture. Considering the individual subgroups, patients of the MS/NTZ-, the HIV-, the lymphomaand the transplant-group all showed a lymphocytic predominance (Table 2). Regarding both control groups, the clear majority of patients showed a lymphocytic predominance. A certain percentage of HIV patients showed a mixed cell distribution, whereby in all of those lumbar punctures an arti cial blood admixture could be observed.

Discussion
Here, we present our results regarding the analysis of standard CSF parameters in a cohort of 108 PML patients. Compared with a control group consisting of patients diagnosed with NPH or IIH our cohort exhibited a signi cantly higher CSF cell count. However, this was mainly due to the HIV-PML group, while the CSF cell counts were normal in the other PML subgroups (Fig. 2). CSF protein and Qalbumin as an indicator for the integrity of the blood-CSF barrier function were similar in both PML and control groups ( Fig. 1). Compared with the control group and the MS/NTZ-PML subgroup the HIV-PML patients had a higher CSF protein, Qalbumin, and CSF lactate (Fig. 2). It is likely that these in ammatory changes in HIV patients are due to the HIV infection itself. However, cell count of HIV-patients with validated PML was elevated in comparison with patients of the HIV control group (Fig. 3). Previous studies have shown that HIV infection is frequently accompanied by CSF pleocytosis occurring early in the infection and often this. In addition to the cell count, protein and Qalbumin levels of the HIV-PML group were signi cantly elevated compared to the respective control group as well (Fig. 3). This indicates that the presence of PML in HIV patients leads to an increased dysfunction of the blood-CSF barrier. In summary, the majority of PML patients show inconspicuous results on CSF testing. In the subgroup of HIV patients, however, increased cell counts as well as elevated CSF protein and Qalbumin levels are regularly observed.
Pleocytosis may be an effect of the CNS involvement of the HIV infection itself, but just like CSF protein concentrations and Qalbumin levels, CSF cell count in HIV patients might also increase due to PML.
In the few patients of the cohort in whom IRIS was documented, a trend, albeit not statistically signi cant, for an increase in cell count and CSF lactate concentration was observed, while the other parameters remained unchanged (table 3). In addition, the immune response during IRIS seems to lead to augmented intrathecal immunoglobulin production (table 5) The detection of OCB at the initial lumbar puncture was highly dependent on the underlying disease.
Thus, the majority of patients with intrathecal immunoglobulin synthesis belonged to the MS/NTZsubgroup (Fig. 5). However, some patients of the other groups, especially HIV infected patients, exhibited positive OCB as well. This is consistent with the observation that HIV infection induces a humoral immune response in the CNS, as measured by an increased intrathecal IgG production, in both neurologically asymptomatic and symptomatic patients (Gisslén et  Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This study was partly funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany's Excellence Strategy -EXC 2155 -Projektnummer 390874280 (to MS).

Figure 1
Comparison of CSF cell count, lactate, Qalbumin, and CSF protein of PML patients with control patients.
PML: progressive multifocal leukoencephalopathy. Data is presented as mean ± standard deviation. Levels of signi cance: ***p<0.001, two-tailed p-value, Mann-Whitney-test with comparison of median values was applied. ns: not signi cant.