Laboratory Diagnosis in Neurology, 1 Ed.

19 CSF Analysis Report

H. Reiber

Integrated CSF Report

Objectives. Although no final diagnosis can be made on the basis of CSF analysis without clinical information, it is nevertheless the case that an integrated CSF report (Fig. 19.1), unlike an isolated listing of individual parameters, can make an important contribution to the differential diagnosis of neurological diseases and to the internal quality control of CSF analysis. The presentation of all CSF data in the form of a single integrated report (Fig. 19.1) portraying a disease-related data pattern permits:

• Clinical plausibility control.

• Plausibility control of the accuracy, i. e., the reliability of individual measurements.

It also helps decision-making about the next analytical step or steps (Reiber and Peter, 2001) by:

• Giving plausibility to a suspected diagnosis.

• Making an initially suspected diagnosis appear implausible.

• Suggesting an unsuspected diagnosis.

Components. A competent CSF report (Fig. 19.1) always includes the following components (Reiber and Peter, 2001; Reiber et al., 2001; Reiber et al., 2006):

• The physician's statement regarding the differential diagnosis or suspected diagnosis.

• Visual inspection.

• Cell count and cytology.

• Protein analysis, including individual data, calculated quotients, and intrathecal fractions.

• Graphical representation of immunoglobulin response patterns in the quotient diagrams.

• Oligoclonal IgG.

• Antibody indices.

• Tumor markers.

• Destruction markers.

• Lactate.

• Ferritin.

• Interpretation as normal vs. pathological relative to reference ranges (Chap. 21).

• Comments providing analysis-related information or indicating any particular significance of CSF data for the clinical (differential diagnostic) interpretation (Reiber et al., 2001; Reiber, 2005 a; Reiber et al., 2006).

CSF Analysis Procedure

Flow chart. The physician's specifications regarding the diagnostic question or suspected diagnosis enable the clinical chemist to design a cost-effective flow chart for the analysis (Fig. 19.2Table 19.1). Worked through step by step, the flow chart helps decision-making about whether to stop or to continue to a more detailed analysis.

Visual inspection. The CSF is visually inspected using the following criteria:

• Clear/colorless (= normal).

• Turbid (cell count higher than about 1000/μL) or purulent (cell count higher than about 10 000/μL; or there is a massive increase in protein, such as in tuberculosis with spiderweb clot).

• Bloody/artificially contaminated with blood.

• Xanthochromic (old hemorrhage).

If the CSF is bloody, artificial bleeding should be excluded (clouds of blood, decreasing erythrocyte counts in the three-tubes test). The mathematical correction of protein data for blood contamination or hemorrhage is shown in Chap. 21, “Cells” (see Table 21.10).

Cell count and cytology. For the detection of acute inflammatory diseases, for differential diagnosis, and for follow-up studies, the cell count (leukocytes) in CSF is still one of the most important parts of CSF analysis (Kölmel, 2003; Linke et al., 2003; Kluge et al., 2005). The cell count should therefore never be omitted. As a matter of principle, the erythrocyte count (artificial or pathological) should be recorded as well, together with semiquantitative information on CSF hemoglobin. This makes it easier to recognize unreliable quotients (see below). The differential cell count may provide essential information on tumors, hemorrhage, or various causes of inflammation (Kölmel, 2003; Linke et al., 2003; Kluge et al., 2005). Automated differential cell counts are not sufficiently sensitive, and they are unsuitable for recognizing the various cell types and bacteria relevant to diagnosis. The variable dynamics of cell invasion and cell differentiation characteristic of different causes of disease (Reiber et al., 2010) should be taken into account.

Protein analysis. Albumin, IgG, IgA, and IgM: the CSF/serum quotients (Q) are calculated from the respective CSF and serum values and printed out in quotient diagrams (Chap. 5, “Proteins”; Fig. 19.1). Intrathecal fractions (in%) are reported numerically, although they can be read directly from the quotient diagrams as well (Figs. 5.21 and 19.1). Protein concentrations in CSF and serum are biologically interconnected, and the calculated ratio (CSF/serum quotient) represents an endogenous reference for the individual patient. To improve the analytical sensitivity, CSF and serum samples should be analyzed in the same analytical run (Reiber et al., 2003 a).

image

Fig. 19.1 CSF report from the Neurochemistry Laboratory, University of Göttingen, Germany, which formed the basis for a variety of commercial software developments adaptable to local analytical profiles. A report that integrates a patient's CSF/serum data makes it possible to recognize a typical disease-related pattern. Together with possible analytical and clinical plausibility controls, it also forms part of a qualified quality assessment of CSF analysis. The data presented here originate from a patient with definitive multiple sclerosis. At the time of the first “diagnostic” puncture, the question was whether there were any signs of inflammation. The combination of a normal cell count, a humoral immune response (oligoclonal IgG; IgGIF = 74%; IgMIF = 57%), and a polyspecific immune response with increased antibody indices combined for measles, rubella, and varicella-zoster indicates a chronic inflammatory process. Differential diagnosis: multiple sclerosis or autoimmune disease with CNS involvement.

image

Fig. 19.2 Flow chart of CSF analysis.

Table 19.1 Levels of CSF analysis. The extent of CSF analysis depends on the differential diagnostic question. Because of the laboratory methods and the costs involved, it is recommended that decisions regarding further special analyses are delayed until the basic program has been carried out. There are essentially three levels, from emergency analysis to the full array of tests (see also flow chart in Fig. 19.2)

A. Emergency program

• Cell count

• Total protein (may be only semiquantitative determination)

• Lactate

• May include a screening test for detection and differentiation of bacterial pathogens

B. Basic program and more detailed analysis (depending on clinical indication for analysis on and protein data)

• Cytology (cell count, differential cell count, detection of bacteria); if possible, always make a cell preparation (even when cell count is normal) for evaluation when needed.

• Total protein (quantitative determination) (used as reference value for dilutions used in subsequent automated nephelometric analysis of single proteins)

• Albumin, IgG, IgA, and IgM

• Qualitative, sensitive determination of oligoclonal IgG (if IgG quotient is not clearly larger than the albumin quotient)

• Detection of microorganism-specific antibodies

• Lactate (determine in all cases of barrier dysfunction with moderately increased cell counts).

C. Special parameters with confirmed clinical relevance

• Specific microorganism detection by PCR (acute inflammation)

• MRZ antibody response (chronic inflammation, MS)

• Tau protein, β-amyloid1–42, 14–3-3 protein, NSE, or S 100B (dementia, degeneration)

• CEA, IgM (tumors)

• β-Trace protein (CSF fistula, posttraumatic leak)

• Tumor cytology (differentiation of tumors)

• Antineuronal antibodies (neurological symptoms of systemic tumors)

• Ferritin in cases of hemorrhage

Table 19.2 Response types and the diseases they indicate

Response type

Diseases

No IgG, IgA, IgM

• Early bacterial meningitis and viral encephalitis

• Guillain-Barré syndrome (polyradiculitis)

IgG dominance

• Multiple sclerosis (IgMIF occurs in 50%, and IgAIF in 20% of cases)

• Neurosyphilis (2-class response, IgMIF occasionally dominant, IgAIF is very rare)

• HIV encephalitis (1-class response)

IgA dominance

• Neurotuberculosis (IgAIF in isolation or combined with weak IgG response)

• Brain abscess

• Adrenoleukodystrophy!

IgM dominance

• Lyme neuroborreliosis (IgMIF > IgAIF > IgGIF)

• Mumps meningoencephalitis (3-class response)

• Non-Hodgkin lymphoma with CNS involvement (1-class response, isolated IgMIF > 0)

• Neurotrypanosomiasis (3-class response, frequency of IgMIF > 0 in 95% of cases)

IgG + IgA + IgM without dominance

Opportunistic infections in immunodeficiency (CMV, toxoplasmosis)

Evaluation

Quotient Diagram (Reibergram)

Characteristics of quotient diagrams. The quotient diagrams with their hyperbolic reference ranges (also known as Reibergrams, Fig. 5.21) were discussed in detail in Chap. 5, “Quotient Diagrams (Reibergrams).” They are well suited to detecting intrathecal synthesis of IgG, IgA, and IgM based on the analysis of lumbar, cisternal, or ventricular CSF from patients of any age (including newborns). What does differ (between ventricular, cisternal, and lumbar CSF), however, is the age-related reference range of the albumin quotient (Chap. 21, “Proteins”), which, as supported by theory (Reiber, 1994, 2003), has various limits for a blood–CSF barrier dysfunction. As a normal expression of a hyperbolic function we find QIgG > QAlb when QAlb < 0.6 × 10−3, even if the CSF is normal (e. g., ventricular CSF). This means that the definition of an intrathecal humoral immune response as QIgM > QIgAQIgA > QIgGQIgG > QAlb (which is supported by both theory and practice) should not be applied when QAlb < 1.0 × 10−3, not least because of analytical imprecision.

The following should also be taken into account when evaluating quotient diagrams:

• Exchange equilibrium: The quotient diagrams may be used only when the exchange between blood and CSF is at equilibrium. Infusion of plasma expanders after major loss of blood or after plasmapheresis may mimic IgG synthesis, or the IgG quotient may be too low after therapeutic administration of immunoglobulins. Depending on the size of the molecules involved, it may take 8–48 hours until a new equilibrium is established between the compartments (for references see Reiber, 1994).

• Paired analysis: An important prerequisite for the most sensitive use of quotient diagrams is paired analysis of CSF and serum samples in the same analytical run in a corresponding concentration range, which will guarantee quotients to be unbiased by different calibration curves of different accuracies and slope-related variations of the concentrations (Reiber et al., 2003 a).

• Contamination with blood: Heavy contamination of the CSF with blood makes the evaluation of quotients difficult or even impossible (for corrective calculation, see Chap. 21, “Cells”). This applies particularly to QIgM associated with low albumin quotients, which could easily lead to a false interpretation as intrathecal IgM synthesis. Basically, however, even in this situation it should always be remembered that, as supported by theory, despite any contamination with blood, a numerically larger quotient for the larger molecule (e. g., QIgG > QAlb or QIgM > QIgA) indicates intrathecal synthesis. The quotients should always be calculated, even when the erythrocyte count is high (> 7000 erythrocytes/μL), but should be reported as numbers only, without evaluation in the diagram or calculation of an intrathecal fraction.

Typical disease-related patterns. The real significance of IgG, IgA, and IgM quotient diagrams is that they facilitate the recognition of disease-related patterns (Table 19.2Fig.19.3 ff.).

The response types listed in Table 19.2 show a one-, two-, or three-class response and varying dominance of one particular immunoglobulin class; the tabular presentation, although clear, is of far less practical value than Figs. 19.3and 19.4, which allow the complex response patterns of the bacterial and viral diseases to be recognized at a glance. It is precisely this possibility of recognizing a specific immunoglobulin pattern at a glance that makes the graphical presentation in the report so attractive (Chap. 19, “Integrated CSF Report”). The diagnostically relevant patterns relate to the first, “differential diagnostic” puncture.

For an adequate interpretation of the intrathecal immune response, however, knowledge of special neuroimmunological response conditions (no isotype switch, Chap. 5, “Immunoglobulins” and Fig. 5.22) and of the temporal dynamics of the intrathecal humoral immune response are both of special relevance (Fig. 19.5).

In looking for disease-specific data patterns, it should not be forgotten that in some diseases typical “pictures” are more the exception than the rule. This may have to do with the localization of the disease process in the brain: only processes that are close to the CSF space can show up clearly in the CSF analysis (Chap. 2).

Pointers toward further analysis. Often the patterns in the quotient diagrams in the context of the differential diagnosis offer the basis for further, more detailed analysis.

Facial nerve palsy. In the case of germ-induced facial nerve palsies, Borrelia-induced disease shows a characteristic immunoglobulin pattern with IgM dominance (Fig. 19.3 d; see also Fig. 5.21) which should be taken as a pointer to further analysis of the Borrelia-specific antibody index. If, on the other hand, the facial nerve palsy is caused by zoster virus (Fig. 19.4 a), the CSF proteins look mostly normal (in about 50% also without oligoclonal IgG). In this case a VZV PCR might perhaps be recommended.

Opportunistic infections in HIV-positive patients. The diagnosis of opportunistic infections in HIV-positive patients is often not easy. The quotient diagram always gives a clear indication of an opportunistic infection by a two- or threeclass immune response (possibly with barrier dysfunction), which is not observed in pure HIV encephalopathy (Fig. 19.4 c). Here, PCR should be considered.

Oligoclonal IgG

When there is doubt about whether intrathecal IgG synthesis is occurring (QIgG ≤ QAlb; IgGIF < 10%), detection of oligoclonal IgG bands by means of isoelectric focusing is the most sensitive way of demonstrating that it is. The interpretation is based on the five types of band patterns (see Fig. 4.13 legend) according to the international consensus (Andersson et al., 1994). In acute diseases, the detection of an immune response by oligoclonal IgG is clinically less sensitive than the antibody index used for detecting the causative antigen (Felgenhauer and Reiber, 1992) (Chap. 5, “Antibody Index”).

Because of the relative sensitivities of the detection methods (Chap. 5, “Antibody Index”), the presence of an intrathecal IgG fraction (IgGIF > 10%) without oligoclonal IgG represents a contradiction—but not the other way around.

Microorganism-Specific Antibodies in CSF and Serum

Antibodies Against the Causative Antigen

In many differential diagnoses, detecting the causative antigen is the decisive diagnostic step. Direct germ detection (by PCR) is often not sensitive enough; antibody detection via the antibody index (see Chap. 5, “Antibody index”) is much more sensitive (Table 19.3).

Subacute sclerosing panencephalitis. A relative high antibody index is often found in this (now disappearing) CNS infection caused by measles, but this does not constitute definite confirmation of SSPE because a polyspecific response (e. g., in MS) can produce similarly high antibody index values (e. g., up to 45) for antibodies with the same high affinity. Only by means of the specific fraction (F s) is it possible to identify antibody synthesis as being directed against a causative antigen (Quentin and Reiber, 2004; Table 19.4).

Zoster ganglionitis (facial nerve palsy). This is caused by the varicella-zoster virus (VZV). With otherwise completely normal CSF findings (Fig. 19.4 aFig. 5.26 a and Chap. 5, “Antibody Index”)—even the absence of oligoclonal IgG—the only sign is an increased VZV antibody index (VZV AI = 2.1, Table 19.3).

HSV encephalitis. No earlier than 1–2 weeks after the onset of the disease, a humoral immune response is recognizable (Fig. 19.4 b) that becomes manifest either through the presence of oligoclonal IgG or an increased antibody index. In addition to the dominating herpes simplex antibody (e. g., HSV AI = 37.9, Table 19.3), concomitant VZV antibody synthesis is also observed (VZV AI = 13.0, Table 19.3). Here, too, the level of the absolute concentration (serum titer, or Fs in Table 19.4) helps to decide which of the two parameters represents the causative antigen.

HIV infection. Oligoclonal IgG is found in both stages of pure HIV encephalopathy (Table 19.3) and with opportunistic Toxoplasma infection of the CNS. Whereas a two- or three-class immune response with barrier dysfunction is found in the patients with opportunistic infection (Fig. 19.4 c), the IgG quotient in the patient with the HIV encephalopathy (Fig. 19.4 c) is below the discrimination line (QLim).

MRZ antibody reaction. Detection of intrathecal synthesis of antibodies against measles, rubella, and VZV (MRZ) in chronic diseases (Table 19.3), such as multiple sclerosis (MS) (Fig. 5.26 bChap. 5, “Antibody index”), lupus ery-thematosus with CNS involvement, or chronic neuroborreliosis (Table 19.3), is of special diagnostic relevance. Chronic neuroborreliosis, which does not develop from the acute form, can be clearly distinguished from the acute form by a dominant IgG class response and an MRZ reaction with an increased Borrelia antibody index (IgG). Of course, since chronic borreliosis is a rare disease, MS—where the Borrelia AI may be increased due to a polyspecific immune response—should be ruled out.

image

image

Fig. 19.3 a–d Bacterial infections of the CNS.

a Meningococcal meningitis. Dynamics of CSF protein data from a patient with meningococcal meningitis. Lumbar punctures were performed on days 1 (■), 3, 6, and 13 (•). The cell counts were 7250/μL, 2730/μL, 213/μL, and 2/μL, respectively. The square (■) indicates the diagnostically relevant initial puncture. Treatment started on the day of admission, and the clinical course was without complications.

b Neurotuberculosis. Protein data in the quotient diagrams of patients with neurotuberculosis. Lumbar puncture was performed 1 week after treatment started. In patient 1 (•), dominant IgA synthesis was directly recognizable by an intrathecal IgA fraction (IgAIF = 35%). In patient 2 (○), intrathecal IgA synthesis was recognized unambiguously only by the relation QIgA > QIgG, with the QIgA value still within the threshold area of the discrimination line (QAlb = 24 × 10−3QIgG = 13.9 × 10−3QIgA = 17.1 × 10−3QIgM = 8.0 × 10−3). In patient 3 (■), additional intrathecal synthesis of IgM was detected 1 week after the start of treatment.

c Neurosyphilis. The data from two patients with neurosyphilis are representative of the meningovascular form (•) with pure IgG synthesis (IgGIF = 30%), and the parenchymatous form (■) (progressive paralysis) with dominant intrathecal IgM synthesis (IgMIF > 80%) in addition to intrathecal IgG synthesis which is also intensive (IgGIF = 80%). In both cases, IgA synthesis was not observed, a very typical result.

d Neuroborreliosis. Characteristic data pattern of neuroborreliosis with dominant IgM class response (humoral and cellular), blood–CSF barrier dysfunction, increased cell count (336/μL), and normal lactate (< 2.1 mmol/L). This pattern has a clinical sensitivity of 70% and a specificity of 96%, even before intrathecal antibodies to Borrelia have been detected (Tumani et al., 1995); see also Figs. 5.21 and 5.22.

Polyspecific Concomitant Immune Response

Definition of a Polyspecific Concomitant Immune Response

A polyspecific immune response is the synthesis of antibodies against antigens which are not the causative antigen of the disease. The corresponding antigens do not need to persist in the organism to maintain the antibody synthesis.

General concomitant immune response. The polyspecific immune response can only be understood on the basis of the immunological network functions. The polyspecific immune response is not a particularity of the immune response in the brain. It is just that classic serology, which focuses on the immune reaction against a causative antigen, can ignore the fluctuations in normal titers of infection in serum by setting the cut-off values high enough. This habit, dictated by practical needs, means that the general immune response of many other B cell clones of variable specificities becomes lost from sight. For example, in Guillain-Barré polyradiculitis, the titers of up to 20 out of 25 blood parameters examined may rise concomitantly (Terryberry et al., 1995). Thus, one has a polyspecific immune response with antibodies against many antigen species, which may not persist in the body and have nothing to do with the causative antigen. This is immediately evident in the case of increased autoantibody synthesis.

The example of MS. Whereas in SSPE or in herpes encephalitis, detecting the respective antibodies provides evidence of the causative antigen, in MS the presence of MRZ antibodies manifests a polyspecific immune response. Also found in MS are intrathecally synthesized Toxoplasma antibodies (10% of cases) or Borrelia antibodies (depending on the geographic region, up to 25% of cases); they may distract at best, and at worst may lead to the wrong diagnosis and treatment. It has been shown for SSPE, herpes simplex encephalitis, and also Fuchs’ heterochromic iridocyclitis (Quentin and Reiber, 2004) that the immune response to the causative, persisting antigen reaches an intensity 20–60 times higher than the polyspecific response with the same antibodies in MS.

image

image

Fig. 19.4 a–d Viral infections of the central nervous system.

a Zoster ganglionitis. Data from patients with facial paralysis (caused by VZV) are often normal in the quotient diagram. In the presented case, only the VZV antibody index was increased (VZV-AI = 2.4, HSV-AI = 1.0); oligoclonal IgG was not detectable. The cell count was normal. For comparison, data from a patient with facial paralysis caused by Borrelia are shown in Fig. 5.21.

b Herpes simplex encephalitis. Dynamics of protein data from a patient with herpes simplex encephalitis who underwent lumbar puncture on days 1, 7, and 30 after admission. On day 1 (■), no humoral immune response had yet occurred; the cell count was 57/μL; there was no oligoclonal IgG from CNS but identical bands in CSF and serum; HSV-AI = 0.7, VZV-AI = 1.0; the HSVPCR was positive. The second puncture on day 7 (•) revealed the following: cell count, 280/μL; oligoclonal IgG in CNS was detected in addition to identical bands in CSF and serum; HSV-AI = 10.5, VZV-AI = 1.6; the HSV-PCR was positive. The third puncture on day 30 yielded the following data: cell count, 30/μL; three-class immune response (IgGIF, IgAIF, IgMIF > 0) and oligoclonal IgG; HSV-AI = 97, VZV-AI = 65.

c HIV infection and opportunistic toxoplasmosis. Case 1 (•) shows HIV encephalitis in a 30-year-old patient in an early stage: cell count, 22/μL; no oligoclonal IgG; HIVAI = 1.0; Toxoplasma AI = 0.9. Case 2 (○) represents another patient with opportunistic toxoplasmosis: cell count, 140/μL; increased albumin quotient and humoral three-class immune response; Toxoplasma AI = 9.2; HIVAI = 5.7; CMV-AI = 1.0. The other two cases represent a patient with opportunistic toxoplasmosis and barrier dysfunction (Δ), and a patient with opportunistic toxoplasmosis and two-class immune response (□).

d Intrathecal lymphoma in a patient with AIDS stage C 3: cell count, 18/μL; no oligoclonal IgG; HIV-AI = 4.5; IgMIF = 65%.

Intensity and Affinity of Intrathecal Antibodies

Table 19.4 shows examples of a polyspecific immune response and an immune response to the causative antigen:

• In acute or subacute inflammation of the CNS associated with a persistent antigen (HSV, measles, or rubella), the amount of antibodies directed specifically against the causative antigen (specific fraction, Fs, in Table 19.4) lies between 3% and 20% of all intrathecal antibodies produced. The rest of the intrathecally synthesized antibodies (i. e., up to 97%) in an acute immune reaction are therefore directed nonspecifically against numerous other antigens.

• In chronic inflammation of the CNS—such as MS, for which no causative antigen is known and persistence of an antigen (e. g., measles) is not detectable—the mean fraction of a single intrathecally produced antibody species in the CSF (the specific fraction, F s) is very much smaller than in acute inflammation: in MS, 0.52% for measles and 0.14% for HSV (Table 19.4). Direct comparison (Jacobi et al., 2007, Reiber et al., 1998; Quentin and Reiber, 2004) clearly shows that the amount of antibodies produced in response to a causative antigen is 20–60 times higher than the amount produced in a polyspecific immune response.

• The affinity of the intrathecal polyspecific IgG class antibodies (MRZ reaction, etc.) is as high as for the antibodies in cases of a causative antigen (SSPE, HSV-encephalitis, Fuchs’ heterochromic cyclitis, etc.). This is consistent with the observation that the polyspecific antibodies are synthesized by perivascular, affinity-maturated B memory cells (Reiber et al., 2010).

By quantification but not by affinity of the intrathecally synthesized antibodies, it is possible to distinguish between antibody production against a causative antigen and a polyspecific concomitant response, such as Toxoplasmaantibodies in MS.

Clinical Interpretation of the Humoral Immune Response

Basically, on detection of a humoral immune response (whether by intrathecal immunoglobulin fraction, or oligoclonal IgG, or antibody index), three possible interpretations should be considered:

image Acute inflammation.

image Chronic inflammation.

image “Scar” not requiring treatment.

It is not always possible to distinguish between these on the basis of CSF analysis alone, for various reasons as follows.

Isotype switch. The classic isotype switch from IgM to IgGclass reaction observed during acute inflammation in blood does not occur in CSF. An example of a serological isotype switch without isotype switch in the CSF recorded during long-term observation of acute neuroborreliosis is shown in Fig. 5.22 (Chap. 5, “Quotient Diagrams (Reibergrams)”). Intrathecal IgM synthesis is therefore not an indication of the acuteness of a CNS process, unlike in serology.

image

image

Duration of the intrathecal immune response. In some infectious diseases the humoral immune response in the CNS can sometimes subside very slowly (Fig. 19.5), even when the patient has been treated successfully and the disease has been overcome. For example, intrathecal antibodies with pathological antibody index values have been found 19 years after herpes encephalitis and 22 years after successful treatment of progressive paralysis (neurosyphilis). Provided the cell count is normal, renewed treatment is not required in such cases. In the case of neuroborreliosis presented in Fig. 5.22 (Chap. 5, “Quotient Diagrams (Reibergrams)”), a humoral immune response with oligoclonal IgG and increased Borrelia antibody index was likewise found at follow-up puncture almost 2 years later, despite complete recovery (Tumani et al., 1995). This exponential decline in the concentration of antibodies produced against a causative antigen at the end of the disease phase is easier to observe in the CSF than in the blood only because the new intrathecal fraction is observed on the basis of the low total antibody concentration in CSF compared to in blood in an inflammatory process. This advantage applies for both the qualitative (oligoclonal IgG) and the quantitative method (antibody index).

Polyspecific immune response. To evaluate a surprisingly increased antibody index, it is important to know that, in addition to specific antibodies to the causative antigen, every humoral immune response produces predominantly polyspecific antibodies. Quantification of the specific fraction (Fs in Table 19.4), or PCR, enables a distinction between the polyspecific immune response and antibody production directed against the causative antigen (see above).

Tumor Markers

The fundamental distinctions (see also Chap. 16) are between:

• A primary brain tumor.

• Metastases from systemic tumors.

• Secondary brain damage by systemic tumors associated with antineural antibodies.

To identify metastases, determination of carcinoembryonic antigen (CEA), for example, can be helpful (Fig. 19.6) (Jacobi et al., 1986), or of the production of IgM in non-Hodgkin lymphoma (Reiber and Peter, 2001; Reiber, 2005 b). Especially with CEA, it is important to emphasize that CSF analysis represents only disease processes in the brain regions near the CSF space (Chap. 1).

The presence of antineural antibodies (Chap. 7, “Antineural Antibodies”) often allows early diagnosis before there are secondary neurological defects due to systemic tumors.

Brain Proteins

Destruction markers. Depending on the differential diagnostic question, surrogate markers for destruction should be included in any complete CSF report (Reiber et al., 2001). They include brain proteins that contribute to the differential diagnosis of dementia or degradation processes in the brain (e. g., NSE, S 100B, tau protein, β-amyloid1–42, protein 14–3-3) (Chap. 11). It must be emphasized that these “dementia markers” are by no means specific, but, if the differential diagnostic question is clear, a combination of these markers (e. g., reduced β-amyloid together with increased tau protein in Alzheimer-type dementia) may perhaps permit distinction between different causes of disease (Alzheimer's disease and multi-infarct dementia; see also Chap. 11).

β-Trace protein. In cases of CSF fistula or postoperative CSF leakage, β-trace protein is analyzed in the discharge as a means of detecting CSF in the discharge (Reiber et al., 2003 b; Reiber, 2004). Detailed reference ranges are given in Chap. 14, “External Ventricular Drainage, Ventriculoatrial and Ventriculoperitoneal Shunts”

PCR

Direct detection of microbial genomes in the CSF by means of PCR (Chap. 5, “Detection of Pathogen-specific Genome Sequences Using Nucleic Acid Amplification Techniques”) is worth doing because of the high clinical sensitivity of this test for HSV, VZV, CMV, and tuberculosis (Wick et al., 1988; Cinque et al., 1996). For many other pathogens, such as Borrelia (Wilske, 2004) or Toxoplasma gondii, the frequency of PCR detection in CSF is very low (< 50%), and antibody detection is usually much more sensitive, so PCR is not recommended here (Chap. 5, “Diagnosis of Pathogens”).

Additional Parameters

Lactate. An important parameter in the differential diagnosis of bacterial vs. viral infections, lactate should always be determined when there is a barrier dysfunction and inflammation is suspected. Lactate is preferred to glucose determination (Chap. 5, “Lactate and Glucose”). Lactate values in the range of 2.1–3.4 mmol/L have too many possible causes to contribute to differential diagnosis. The increased lactate levels associated with inflammation continue to be detectable over a certain period even after treatment has started, in contrast to glucose (Chap. 5, “Lactate”). The reference range for lactate values in ventricular CSF is higher than that for lumbar CSF.

Ferritin. Ferritin is increased in CSF (> 18 μg/L) if there is subarachnoid hemorrhage, and this can be regarded as a supplementary test to cytology (Wick et al., 1988). Increased ferritin levels have also been found in herpes simplex encephalitis (microhemorrhage). Since the CSF ferritin concentration is independent of the serum level, serum analysis may be omitted.

image

Fig. 19.5 Delayed normalization of the intrathecal immune response (IgGLoc) in seven patients after neurosyphilis. CSF was obtained at different times after the onset of acute illness; treatment was successful in each case. Examples of patients in the acute phase of neurosyphilis are shown in Fig. 19.3 c.

image

Fig. 19.6 Meningeal carcinomatosis. Detection of intrathecal release of carcinoembryonic antigen (CEA) is documented in the quotient diagram (QCEA/QAlb), which is identical to the quotient diagram for IgA because of the similarity in molecular size. The data represent four patients with brain metastases from primary breast cancer, colon cancer, cancer of the urinary bladder, and adenocarcinoma. One patient with systemic carcinoma and pathologically increased CEA value in the blood (67.4 ng/mL) shows a normal CSF pattern (QAlb = 2.4; QIgG = 1.1; QCEA = 0.4; 2 cells/μL) and no signs of brain metastases. However, this would also be the finding if there were metastases in the frontal brain, i. e., far from the CSF space, since CSF analysis characterizes only the area of the brain that is close to the CSF.

Other low molecular weight analytes. Neuropeptides, cytokines, neurotransmitters, amino acids, or biogenic amines (Chap. 21) are relatively small molecules and therefore barely evaluable diagnostically in lumbar CSF, owing to rapid exchange processes in the CNS, cellular metabolism, and in some cases also because of their short half-life in the blood. Lipids in CSF (Chap. 21, “Amino Acids, Lipids, and Vitamins”) are not important diagnostically either, since disorders of lipid metabolism are identified by blood analysis. A particular exception are the peptides which are released directly from the choroid plexus into ventricular CSF. In case of normal CSF flow rate their changes may be still represented to some extent in lumbar CSF.

Blood Analysis in Neurological Diseases

Clinical neurochemistry increasingly uses the analysis of brain-specific proteins (e. g., NSE, S 100B) in peripheral blood (Schaarschmidt et al., 1994; Reiber, 2003). Serial blood analysis, for example, enables one to make a prognosis after cerebral hypoxia or cerebral infarction (Schaarschmidt et al., 1994; Reiber, 2003).

Interpretation

Standard Interpretations

Table 19.5 lists standard interpretations as they are used in the context of knowledge-based interpretation programs.

Table 19.5 Integrated evaluation

Evaluation

Meaning

Normal CSF report

Includes normal CSF cytology findings

Normal CSF protein report

All test results except cytology are normal

Blood–CSF barrier dysfunction

QAlb > QAlb (ref.), the age-related reference

Increased cell count

> 4/μL

Inflammatory process

• Intrathecal fractions of IgG, IgA, and IgM > 10%

• QIgG > QAlb (QAlb > 0.6 × 10−3)

• QIgA > QIgG

• QIgM > QIgA

• Oligoclonal IgG in CSF

• Microorganism-specific antibody synthesis

• Cell count > 20/μL

Intrathecal synthesis of specific antibodies

AI ≥ 1.5

Hemorrhage

• Subarachnoid hemorrhage vs. artificial blood contamination

• Cytology (siderophages), ferritin > 18 μg/L

Tumor, tumor metastases

• QCEA > QLim (IgA)

• Cytology

• Antineuronal antibodies

Disease-Related Data Patterns—Summary

The suspected diagnosis based on the patient's history and the basic CSF analysis program (Table 19.1) often reveals very characteristic patterns of immunoglobulin class responses in the CNS (Figs. 19.319.4). These patterns may then provide clues to a more detailed analysis (Table 19.6).

Table 19.6 shows the frequency of pathological values in neurological diseases at the time of diagnostic lumbar puncture. To obtain a disease-specific data pattern that has sufficient value for the differential diagnosis, a number of data (e. g., specific antibodies, lactate level) or special optional tests are necessary in addition to the basic information such as cell count, barrier dysfunction, and intrathecal fraction.

No diagnosis can be established only on the basis of individual pathological values, not even in the form of several parameters taken in combination. Only the clinician can make the diagnosis, against the background of the patient's history, the clinical findings, and supplementary analytical tests (e. g., imaging procedures).

List of Particular Analytical and Diagnostic Comments

Software programs (e. g., www.wormek.com) allow the storage of frequently recurring comments that can be retrieved using a shortcut. Analysis-related comments include all explanations and indications for carrying out the analysis and all problems occurring during the analysis. Diagnosisrelated comments are interpretations that may contribute to the differential diagnosis, but they must never be understood as the diagnosis itself.

The following examples from the Neurochemical Laboratory in Göttingen, Germany were all added individually to the respective CSF data reports (Fig. 19.1), not printed out automatically as part of a knowledge-based evaluation program:

image

Analysis-related comments:

– Analysis of “parameter X” was repeated.

– Further analysis was not possible due to small sample volume.

– Analysis of oligoclonal IgG was not performed because QIgG > QAlb.

– Intrathecal IgA synthesis was detected because QIgA > QIgG (even if IgAIF = 0%).

– Total cell count was corrected for blood contamination (reduction of 1 leukocyte/μL per 1000 erythrocytes/μL in CSF).

– Differential cell count could not be interpreted because of heavy blood contamination.

– Correction of CSF albumin and Ig concentrations for blood contamination (in the range of 1000–7000 erythrocytes/μL) resulted in IgIF ≤ 0%.

– Due to heavy blood contamination (> 7000 erythrocytes/μL), interpretation of CSF/serum quotients is unreliable.

– Despite heavy blood contamination, there is evidence (QIgG > QAlb—or QIgA > QIgG or QIgM > QIgA) of a humoral immune response in the CNS. The absolute values of the quotients, however, are not reliable enough for an intrathecal fraction to be calculated.

• Diagnosis-related comments:

– Positive MRZ response: chronic inflammation (autoimmune type). Differential diagnosis: MS or autoimmune disease with CNS involvement.

– A high albumin quotient (> 20 × 10−3) or a high cell count (> 90/μL) or absence of oligoclonal IgG do not support the suspected diagnosis of MS.

– Markedly raised albumin quotient, dominant intrathecal IgA synthesis, increased CSF lactate (> 3.4 mmol/L), and moderate pleocytosis suggest tuberculous meningitis with high probability (PCR recommended).

– Lactate > 3.4 mmol/L together with a cell count > 500/μL indicates bacterial infection.

– Intrathecal three-class reaction with HIV AI ≥ 1.5 indicates opportunistic infection (antibody analysis or PCR for a relevant microorganism recommended).

– Isolated intrathecal IgM synthesis without other inflammatory signs (normal cell count, no oligoclonal IgG) may indicate lymphoma.

– Intrathecal synthesis of carcinoembryonic antigen (CEA) with QCEA > QLim (IgA) suggests tumor metastases in the brain.

– The combined findings of decreased β-amyloid1–42 and increased tau protein are compatible with the suspected diagnosis of Alzheimer's disease.

– Tau protein values > 1300 pg/mL are observed in Creutzfeldt-Jakob disease (analysis of protein 14–3-3 and neuron-specific enolase in CSF recommended).

References

Andersson M, Alvarez-Cermeño J, Bernardi G, et al. Cerebrospinal fluid in the diagnosis of multiple sclerosis: a consensus report. J Neurol Neurosurg Psychiatry 1994;57:897–902

Bechter K, Reiber H, Herzog S, Fuchs D, Tumani H, Maxeiner HG. Cerebrospinal fluid analysis in affective and schizophrenic spectrum disorders: Identification of subgroups with immune responses and blood–CSF barrier dysfunction. Journal of Psychiatric Research 2010, in press

Cinque P, Cleator GM, Weber T, et al. The role of laboratory investigation in the diagnosis and management of patients with suspected herpes encephalitis: a consensus report. J Neurol Neurosurg Psychiatry 1996;61:339–345

Felgenhauer K, Reiber H. The diagnostic significance of antibody specificity indices in multiple sclerosis and herpes virus induced diseases of the nervous system. Clin Invest 1992;70:28–37

Jacobi C, Reiber H, Felgenhauer K. The clinical relevance of locally produced carcinoembryonic antigen in cerebrospinal fluid. J Neurol 1986;233:358–361

Jacobi C, Lange P, Reiber H. Quantitation of intrathecal antibodies in cerebrospinal fluid of subacute sclerosing panencephalitis, herpes simplex encephalitis and multiple sclerosis: discrimination between microorganism-driven and polyspecific immune response. J Neuroimmunol 2007;187:139–146

Kölmel HW. Liquorzytologie. In: Zettl UK, Lehmitz R, Mix E, eds. Klinische Liquordiagnostik. Berlin: de Gruyter; 2003: 135–160

Kluge H, Wieczorek V, Linke E, et al. Atlas der praktischen Liquorzytologie. Stuttgart: Thieme; 2005

Linke E, Wieczorek V, Zimmermann K. Qualitätskontrolle in der Liquorzytodiagnostik. In: Zettl UK, Lehmitz R, Mix E, eds. Klinische Liquordiagnostik. Berlin: de Gruyter; 2003: 366–377

Quentin CD, Reiber H. Fuchs heterochromic cyclitis: Rubella virus antibodies and genome in aqueous humor. Am J Ophthalmol 2004; 138:46–54

Reiber H. Flow rate of cerebrospinal fluid (CSF)—a concept common to normal blood-CSF barrier function and to dysfunction in neurological diseases. J Neurol Sci 1994;122:189–203

Reiber H. Proteins in cerebrospinal fluid and blood: barriers, CSF flow rate and source-related dynamics. Restor Neurol Neurosci 2003;21:79–96

Reiber H. Beta-trace protein concentration in nasal secretion: discrepancies and flaws in recent publications. Acta Neurol Scand 2004;110:339–341

Reiber H, Uhr M. Liquordiagnostik. In: Berlit P, ed. Klinische Neurologie. 3rd ed. Heidelberg: Springer; 2010

Reiber H. Liquordiagnostik. In: Thomas L, ed. Labor und Diagnose, 6th ed. Frankfurt: TH Books; 2005 b

Reiber H, Peter JB. Cerebrospinal fluid analysis—disease-related data patterns and evaluation programs. J Neurol Sci 2001;184:101–122

Reiber H, Ungefehr St, Jacobi C. The intrathecal, polyspecific and oligoclonal immune response in multiple sclerosis. Mult Scler 1998;4:111–117

Reiber H, Otto M, Trendelenburg C, Wormek A. Reporting cerebrospinal fluid data—knowledge base and interpretation software. Clin Chem Lab Med 2001;39:324–332

Reiber H, Thompson EJ, Grimsley G, et al. Quality assurance for cerebrospinal fluid protein analysis: international consensus by an internetbased group discussion. Clin Chem Lab Med 2003a;41:331–337

Reiber H, Walther K, Althaus H. Beta-trace protein as sensitive marker for CSF rhinorhea and CSF otorhea. Acta Neurol Scand 2003b;108:359–362

Reiber H, Teut M, Pohl D, Rostasy KM, Hanefeld F. Paediatric and adult multiple sclerosis: Age-related differences and time course of the neuroimmunological response in cerebrospinal fluid. Multiple Sclerosis 2009;15:1466–1480

Reiber H, Kruse-Sautter H, Quentin CD. Different B cell repertoires in brain and eye of the individual multiple sclerosis patient. Patterns of antibodies, oligoclonal IgG, and isotypes in cerebrospinal fluid and aqueous humor. J Neurimmunol 2010, submitted

Schaarschmidt H, Prange H, Reiber H. Neuron-specific enolase concentrations in blood as a prognostic parameter in cerebrovascular diseases. Stroke 1994;24:558–565

Terryberry J, Sutjita M, Shoenfeld Y, et al. Myelin- and microbe-specific antibodies in Guillain-Barré syndrome. J Clin Lab Anal 1995;9:308–319

Tumani H, Nölker G, Reiber H. Relevance of cerebrospinal fluid parameters for early diagnosis in neuroborreliosis. Neurology 1995;45:1663–1670

Wick M, Fink W, Pfister W, et al. Ferritin in cerebrospinal fluid differentiation between central nervous system haemorrhage and traumatic spinal puncture. J Clin Pathol 1988;41:809–814

Wilske B. Diagnosis of lyme borreliosis in Europe. Vector Borne Zoonotic Dis 2004;3:215–227