| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Clinical Immunology |
1 Institute of Clinical Chemistry and Laboratory Medicine, and2 Department of Neurology, Johannes Gutenberg University, Mainz, Germany.
aAddress correspondence to this author at: Institute of Clinical Chemistry and Laboratory Medicine, Medical Centre of the Johannes Gutenberg University, D-55101 Mainz, Germany. Fax 49-6131-176627; e-mail lackner{at}zentrallabor.klinik.uni-mainz.de.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
free light chains (FLC) in cerebrospinal fluid (CSF) and serum, and compared it with traditional tests for intrathecal immunoglobulin synthesis.
Methods: FLCs were measured by nephelometry in CSF/serum pairs from 112 patients. Samples were excluded if blood contamination of CSF as a result of traumatic lumbar puncture (n = 12) or monoclonal bands in both CSF and serum (n = 5) were present. The remaining sample pairs were grouped according to the presence (n = 71) or absence (n = 24) of oligoclonal bands. Data were analyzed as FLC concentrations in CSF, as FLC CSF/serum ratios, and by use of the quotient diagram described previously for immunoglobulins.
Results: Both FLC concentrations in CSF and the FLC CSF/serum ratio identified patients with oligoclonal bands with high specificity and sensitivity. The areas under the ROC curves were 0.991 (95% confidence interval, 0.944–0.998) and 0.978 (0.924–0.996), respectively. Exclusion of patients with impaired blood–CSF barrier function further improved diagnostic accuracy. To account for patients with impaired blood–CSF barrier function, data were also analyzed in a quotient diagram. Only two patients without detectable oligoclonal bands would have been misclassified by this approach.
Conclusions: Our data indicate that the nephelometric assay for FLCs in CSF reliably detects intrathecal immunoglobulin synthesis. This automated and quantitative method could simplify the diagnostic procedure for CSF analysis in the routine laboratory.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
In theory, the determination of free light chains (FLCs) might be a sensitive alternative to the above-mentioned approaches. Immunoglobulin light chains are typically secreted together with intact immunoglobulins by plasma cells. Although the half-life of FLCs in serum is very short (2–4 h) because of rapid renal elimination, this clearance pathway is not available from CSF; therefore, one can assume that the half-life of FLCs in CSF is comparable to that of other proteins. Thus, even small amounts of intrathecal immunoglobulin synthesis with concomitant FLC secretion will disproportionately increase the CSF concentration of FLCs, making them a potentially sensitive marker of intrathecal immunoglobulin synthesis.
Attempts to determine FLCs in CSF have been made previously (5)(6). The methods described to date for the detection or quantification of FLCs are laborious and time-consuming, however; therefore, they have never been incorporated into routine CSF analysis. Recently, an automated nephelometric assay for the detection of monoclonal gammopathy based on specific polyclonal antibodies against epitopes of FLCs that are hidden in intact immunoglobulins was introduced (7). We applied this test to CSF/serum pairs and evaluated its diagnostic accuracy in the detection of intrathecal immunoglobulin synthesis in comparison with the determination of OCBs, which is considered to be the most sensitive procedure and was taken as the reference standard for our study.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
FLCs. All patients had undergone an elective spinal puncture in the Mainz University Hospital between August 2002 and January 2004; a request for OCBs had been submitted, and sufficient surplus CSF and serum were available for additional tests. Patients were divided into two groups. Group 1 (n = 24) had no detectable intrathecal fraction and no OCBs; group 2 (n = 71) had positive OCBs. Forty-nine patients in group 2 had a diagnosis of multiple sclerosis (MS) according to the McDonald criteria (8) (group 2A), whereas 22 patients had other neurologic diseases (group 2B). CSF samples from 12 patients were excluded because of blood contamination resulting from traumatic lumbar puncture. Of these patients, nine would have been assigned to group 1 (no OCBs; no detectable intrathecal fraction) and three to group 2 (positive OCBs). Another five patients were excluded because they had monoclonal bands in both the CSF and serum. The characteristics of the patients are shown in Table 1. The anonymous use of surplus patient material is approved by the local ethics committee, and informed consent was obtained from the patients.
methods
Immunoglobulin and albumin concentrations were measured by nephelometry (BN II; Dade Behring) in fresh CSF and serum samples. Intrathecal production of immunoglobulin was assessed according to a quotient diagram described previously (2).
OCBs were determined by immunofixation-peroxidase (Hydragel 6 CSF; Sebia) on a semiautomated agarose electrophoresis system (Sebia Hydrasys LC). This method performs comparably to other analytical methods, such as immunofixation electrophoresis, high-resolution agarose gel electrophoresis (9), and isoelectric focusing (10). The system can detect OCBs in a concentration range of 30–125 µg/L. Analysis was performed by laboratory technicians, and the results were read by an experienced staff physician (B.A. or J.L.) and given as positive or negative. After analysis of immunoglobulin concentrations and OCBs, samples were aliquoted and stored at –20 °C until further use.
Nephelometric measurement of FLCs was performed with the FreeliteTM Human Kappa Free Kit (The Binding Site Ltd.). Briefly, a monospecific antibody against FLCs was coated on polystyrene latex. Antibodies were generated from sheep immunized with FLCs, and subsequent antisera were adsorbed extensively against intact immunoglobulins until they were specific for FLCs (7). FLCs in CSF and serum samples were measured according to the manufacturer’s protocol on the BN II automated analyzer. According to the manufacturer, the lower detection limit for FLCs is 0.06 mg/L. The intraassay CV (10 replicate determinations of a serum-based control) was 4.0% at 0.54 mg/L. The interassay CV (10 determinations on separate days of the same control material) was 14% at 0.55 mg/L. Calibrators and controls were provided by the manufacturer and consisted of stabilized human sera containing polyclonal FLCs; calibrators and controls were diluted to the appropriate concentrations for serum and CSF determinations.
Statistical analysis of diagnostic accuracy, including confidence intervals, was done with MedCalc 4.31.010.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
FLCs are shown in Table 1
. Although the groups were not specifically matched, there were no major differences in age and sex distribution between the groups. The median albumin ratio was higher in group 2B (7.1 x 10–3) than in group 1 (5.3 x 10–3) and group 2A (4.8 x 10 –3), consistent with impaired blood–CSF barrier function as expected for this group, which included patients with infectious etiologies of intrathecal inflammation. The higher relationship between IgA/IgM and combined positivity in quotient diagrams in group 2B reflects the heterogeneity of neurologic diagnoses in this group.
|
detection of FLCS
In 33% of group 1 patients, the FLC concentration in CSF was at or below the lower detection limit of 0.06 mg/L. The median CSF concentration of FLCs in group 1 was 0.09 mg/L (range, 0.06–0.47 mg/L). Four patients had FLC concentrations in CSF >0.3 mg/L: a 74-year-old man with rectal carcinoma metastatic to the brain (0.47 mg/L); a 44-year-old man with cerebral infarction (0.41 mg/L); and two women, 66 and 41 years of age, respectively, with carcinomatous meningeosis in the course of breast cancer (0.38 and 0.31 mg/L, respectively). These patients also had relatively high FLC CSF/serum ratios, ranging from of 40 x 10–3 to 83 x 10–3, and mild to moderate impairment of the blood–CSF barrier with albumin ratios ranging from 7.9 x 10–3 to 18.6 x 10–3. CSF/serum ratios were calculated for all patients in group 1 and ranged from 5 x 10–3 to 83 x 10–3, but it should be mentioned that these ratios are an overestimation for the cases with FLCs in CSF below the lower detection limit (see above).
FLCs were detectable in groups 2A and 2B, with medians of 2.7 mg/L (range, 0.29–27.1 mg/L) and 3.6 mg/L (range, 0.43–28.3 mg/L), respectively. The upper outlier in group 2B was a patient with tuberculous meningitis and strongly impaired blood–CSF barrier function (Qalb = 32.5 x 10–3, where Qalb is the CSF/serum ratio of albumin) and a FLC concentration of 28.3 mg/L. The CSF/serum ratio for FLCs was 1.14 for this patient. In contrast, Qalb was within the reference interval for all three upper outliers of group 2A. All three had FLC CSF/serum ratios >1.2. Three patients in group 2A and one patient in group 2B had FLC concentrations in CSF <0.5 mg/L, which is above the highest value in group 1 (Fig. 1
). When patients with impaired blood–CSF barrier (Qalb >7 x 10–3) were removed from the analysis, there was no overlap between FLC concentrations in groups 1 and 2. This indicates that determination of FLC concentrations in CSF provides information similar to that provided by measurement of OCBs provided that patients with impaired blood–CSF barrier function are excluded.
|
ROC analysis was performed for the FLC concentration in CSF and for FLC CSF/serum quotients. The presence of OCBs was taken as the reference for intrathecal immunoglobulin synthesis. The areas under the curves were 0.991 (95% confidence interval, 0.944–0.998) and 0.978 (0.924–0.996), respectively (Fig. 2
). As expected, removal of patients with disturbance of the blood–CSF barrier, as indicated by Qalb >7 x 10–3, from the analysis increased the area under the ROC curve.
|
csf/serum ratios of FLCS and albumin
Because neither of the above-mentioned criteria takes into account blood–CSF barrier function, CSF/serum ratios of FLCs were plotted against the respective ratios of albumin as is typically done for immunoglobulins. Again, a clear separation of patients with and without intrathecal immunoglobulin synthesis could be achieved with little overlap between group 1 and groups 2A or 2B (Fig. 3
).
|
The hyperbolic function commonly used for the immunoglobulin quotient diagrams:
 |
FLC data. To obtain an optimal fit, the following parameters were set empirically: a/b = 6.5; b2 x 106 = 5; c x 103 = 1. With these parameters, two patients from group 1 were classified as having intrathecal FLC production. One patient from group 1 had Borrelia infection, but no Borrelia-specific antibodies could be detected in CSF. A diagnosis of CNS involvement was therefore discarded. The second patient from group 1 had cerebral infarction with mild impairment of blood–CSF barrier function (Qalb = 7.88 x 10–3). Neither patient had positive OCBs or evidence of intrathecal immunoglobulin synthesis according to analysis of quotient diagrams. Both patients had CSF leukocyte (<5/µL) and erythrocyte (negative) counts within the appropriate reference intervals. |
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
FLCs to detect intrathecal immunoglobulin synthesis was reported previously by other groups (5)(11)(12), but the test has not been incorporated into diagnostic use. The determination of free immunoglobulin light chains has been technically difficult in the past and not feasible in routine clinical diagnostics. Here we show that a novel, automated assay for the detection of FLCs is suitable for routine detection of intrathecal immunoglobulin synthesis.
We reanalyzed 112 CSF/serum pairs from patients who underwent analysis of OCBs in CSF during their diagnostic workup. We assumed that the measurement of OCBs represents the most sensitive procedure to detect intrathecal immunoglobulin synthesis. Thus, any alternative procedure must be evaluated against this "gold standard". We excluded 17 CSF/serum pairs because of blood contamination resulting from traumatic lumbar puncture or the presence of monoclonal bands. Blood contamination of CSF is known to bias quotient diagrams in favor of intrathecal synthesis. Inclusion of these patients in the study would not have changed the results appreciably (data not shown); all patients would have been classified properly. The number of patients with monoclonal bands was too small to draw conclusions about the diagnostic value of FLC concentrations in CSF in this group.
The absolute concentration of FLCs in CSF as well as the CSF/serum ratio of FLCs identified intrathecal immunoglobulin synthesis with high sensitivity and specificity. Because group 1 did not include patients with severe impairment of blood–CSF barrier function, who might present with increased total FLC concentrations or CSF/serum ratios, the high diagnostic sensitivity and specificity are limited to patients with normal or almost-normal blood–CSF barrier function.
The use of a quotient diagram plotting CSF/serum ratios of FLCs vs albumin again provided good separation of patients with and without intrathecal immunoglobulin synthesis. The separating line between the patient groups was empirical, but the data show that there was a clear separation into two major populations. Thus, with larger numbers of patients with and without intrathecal immunoglobulin synthesis, it should be possible to generate an improved calculation that can be used for differentiation.
Two patients were misclassified according to this procedure (see the Results). One of these suffered from a Borrelia infection presumably without intrathecal immunoglobulin synthesis. Although in this case no evidence of CNS involvement was found by conventional methods, it cannot be excluded with absolute certainty that the FLC result is in fact correct and that the other methods missed intrathecal immunoglobulin synthesis. Only for the sake of clarity did we presume that the presence of OCBs is 100% sensitive for the detection of intrathecal immunoglobulin synthesis. However, this is not undisputed.
The predominance of FLCs in intrathecal inflammation is widely accepted in the literature (6)(12)(13)(14), but the diagnostic use of FLC quantification is not yet clear. One group noted that the presence of FLC banding in CSF was highly correlated with the concentration of FLCs in CSF (15). Furthermore, the use of FLC quantification to monitor the time course of MS has been documented in the Avonex study (16). These investigators found a positive correlation between FLC, quantified by ELISA (17), and cerebral lesion volume after 2 years. A previous report from the same group pointed to the predictive value of FLC concentrations in CSF for the progression of MS, and they even proposed FLCs as a target for intervention in therapeutic trials (18). The automated quantification of FLCs in CSF that we describe here could facilitate the use of this marker in the routine laboratory.
In summary, we describe here a novel approach to quantitatively detect intrathecal immunoglobulin synthesis by measurement of FLCs. This assay is rapid and reliable, reducing tedious manual procedures and uncertainty in result interpretation to a minimum. In addition, it may be useful to monitor the progression of inflammatory neurologic diseases, e.g., MS. One major advantage, in our opinion, is the ready availability of the assay for use on a routine nephelometric analyzer. Thus, daily assays would be easily integrated into laboratory processes. Furthermore, the turnaround time is 
1.5–2 h for this assay, compared with at least 4–5 h for OCB determinations. In addition, the analysis of FLCs is more sensitive than the determination of intrathecal immunoglobulin synthesis by quotient diagrams for intact immunoglobulins. The sensitivity of the FLC assay is similar to that for OCB determinations. Different from OCBs, the FLC assay is quantitative and provides a means to follow changes in intrathecal immunoglobulin synthesis. Clearly, experience with larger patient numbers is needed to come to a final evaluation of this method for FLCs, but our data indicate that nephelometric determination of FLCs in CSF and serum may be a useful diagnostic procedure for detection and potentially for monitoring of intrathecal immunoglobulin synthesis.
|
Acknowledgments |
|---|
FLCs (FreeliteTM Human Kappa Free Kit) were kindly provided by The Binding Site GmbH (Schwetzingen, Germany). |
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  A. R. Bradwell Serum Free Light Chain Measurements Move to Center Stage Clin. Chem., May 1, 2005; 51(5): 805 - 807. [Full Text] [PDF]
|
|
|
|
|
|
E. J. Thompson Quality versus Quantity: Which Is Better for Cerebrospinal Fluid IgG? Clin. Chem., October 1, 2004; 50(10): 1721 - 1722. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, group 2A; 
, group 2B