| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Editorials |
1 Department of Neuroimmunology, National Hospital for Neurology & Neurosurgery, Institute of Neurology, Queen Square, London WC1N 3BG, United Kingdom, Fax 44-207-837-8553, E-mail e.thompson{at}ion.ucl.ac.uk
An article in this issue (1) reopens an old question: Is qualitative or quantitative testing of immunoglobulins in cerebrospinal fluid (CSF) more diagnostically useful? The short answer is that each has pros and cons. In the diagnosis of multiple sclerosis, it can be misleading to perform only quantitative analysis, but it is much less problematic to perform only qualitative analysis (2). To detect IgG synthesis within the brain/CSF, a recent International Consensus has reaffirmed that qualitative is better than quantitative analysis and has gone further to state that there must be isoelectric focusing followed by immunofixation (Freedman et al., submitted for publication).
The study of "free" light chains (Bence Jones proteins) has been rekindled by the report by Fischer et al. (1) in this issue, in which the authors examine commercially available antibodies in the hope that they will provide a better test for intrathecal antibody synthesis. Several specific issues can be considered as individual questions, although some may be rhetorical. I will examine clinical aspects before chemical issues.
Is the assertion by Consensus neurologists, that isoelectric focusing is equivalent to the IgG index (3), correct? Yes and no. The latter (quantitative) test is not sufficient to make the diagnosis because it is only 
75% sensitive vs >95% for the focusing (qualitative) test (3).
How often should we repeat the lumbar puncture to get the correct diagnosis? It would be best to get the right answer on the first (and thus the only) occasion. However, there are some circumstances when a second puncture may be clinically indicated, and these relate to 
/
immunofixation after IgG isoelectric focusing (4). If a single band is found (although this is not an oligoclonal pattern), more than one-half these patients will progress to a full oligoclonal pattern (5).
Is there a 2-h "rush" to get an answer? Probably not, as the authors point out (1) that their method takes 1.5–2 h vs 4–5 h for isoelectric focusing. Multiple sclerosis is compatible with normal longevity; therefore, a clinician has ample time to get a better answer.
issues on separation
How can one reconcile use of polyclonal standards for free light chains when interpreting individual samples, which are similar to monoclonal bands? This is a conceptual conundrum as samples from patients with multiple sclerosis typically display multiple "mini" monoclonal bands. Thus, it is true that one is not strictly comparing like with like.
Which has better resolution, electrophoresis or focusing? Clearly the latter, which is why the Consensus neurologists also "prefer" this method of separation for antibody molecules (2)(3).
There are currently two Food and Drug Administration-cleared methods; therefore, is there a comparison? Yes, this was originally an abstract in an AACC Oak Ridge Meeting (April 2003) reporting a comparison between electrophoresis (with the present technique) vs isoelectric focusing, but it has been slightly modified, using Sebia instead of Helena (6). For focusing, the International Consensus [Ref. (2) and Freedman et al., submitted for publication] criterion for a positive result is for the presence of two bands in CSF that are not present in the parallel serum sample.
How useful is electrophoresis (vs focusing) in ROC analysis, given that light chains cover an area under the curve of 0.994? This is a hypothetical question because they did not use focusing. To address the positive aspects, at least Fischer et al. (1) used immunofixation and thus avoided the many pitfalls associated with concentrating CSF.
issues in qualitative/quantitative analysis
What is the clearest advantage of quantitative analysis of light chains? As Fischer et al. rightly state, this is to "follow changes" (1), e.g., in response to therapy and possibly prognosis (7). To reiterate, however: for diagnosis, qualitative analysis is much preferred. In terms of basic pathophysiology, there are clearly quantitative changes during the course of multiple sclerosis, whereas in isoelectric focusing, patients have a so-called "fingerprint" pattern that typically does not change (8).
Who else has noticed this dichotomy between diagnosis and therapy? This has been debated for many years by hematologists dealing with myeloma, and many accept that qualitative analysis is better for diagnosis, whereas quantitative analysis is better for monitoring therapy. This has been confirmed in large survey studies (9). One could go further and say that for many other serum proteins, qualitative analysis is a necessary adjunct to a quantitative analysis (10).
What is the best way to keep "uncertainty in result interpretation to a minimum" (1)? The answer must be reliable quality control. The IgG patterns after focusing should be classified according to one of the five recognized types (2), but in most people’s experience, 80% are normal (polyclonal). There is an additional trade-off between qualitative analysis with "uncertain" interpretation vs quantitative analysis, which is certainly less sensitive/specific.
issues in sensitivity
What is the basic technique involved in the production of antibodies against free light chains? With extensive adsorption, one ends up with not only lower affinity but also, as is clearly demonstrated yet again in this study (1), with lower sensitivity. An alternative approach is to use "unadsorbed" monoclonal antibodies (11). These can match the specificity of the current antibodies but without the heavy cost of poor sensitivity.
How useful is a quantitative test in which one third of normal patients have values that are indistinguishable from the background values? Again, further work is required to improve sensitivity.
How close is the sensitivity for the quotient (CSF concentrations divided by serum concentrations) of free light chains (94%) vs the "gold standard" (electrophoresis) chosen by Fischer et al. (1), which the authors set by definition to 100%? Beyond the need for focusing as the Consensus gold standard, one would raise the obvious supplementary question: What would ROC analysis show, using not the quotient but the absolute concentrations of CSF free light chains?
How useful is a hyperbolic curve for quotients in these circumstances? In multiple sclerosis, the barrier function for the majority of patients is normal or shows only minor degrees of change (2). Therefore, a hyperbolic curve is relevant when there are much higher concentrations of CSF total protein.
issues in specificity
Have others noticed "false" positives as well? The best examples are probably Sindic and Laterre (12) with similar findings by Lamers et al. (13). They each showed abnormalities in noninflammatory brain diseases, e.g., amyotrophic lateral sclerosis.
What are the pros and cons of "in-gel" immunofixation vs immunoblotting on nitrocellulose (or other fiber) membranes? This is a protracted story, but the fundamental step is that proteins bind directly to membranes. However, for in-gel fixation to work, the antibody must first precipitate the relevant protein before it can be "visualized" by any staining (14). Another advantage of in-gel immunofixation is that it avoids the use of nitrocellulose, but the problem of "pro-zone" (IgG concentrations too high) remains.
Last, but certainly not least, can one presume that there is no additional benefit by analyzing free light chain bands [by isoelectric focusing] compared with the nephelometric quantification used in this study (1)? Yet again, there is not only much additional information from the qualitative technique (13), but this also loops back to the first general point above, namely, that qualitative analysis of CSF IgG is better than quantitative analysis.
I apologize if I have raised more questions than answers, but then that is what makes the future more interesting. We should probably have an International Consensus on measuring free light chains and their calibrators: perhaps a topic for a future Editorial (15)
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
