ClinicSpeak StatsSpeak: JCV antibody index before PML

Should I be monitoring by JCV antibody index? #ClinicSpeak #MSBlog #MSResearch #StatsSpeak

"The case reports below are very important; they show that in 4 MSers who developed PML on natalizumab their JCV-index was high and stayed persistently high before they were diagnosed with PML. The paper reports the median anti-JCV antibody index before and during natalizumab therapy in these four MSers was 3.04 with a range of 2.04–3.59. Why is the paper important? It is important because there is some preliminary data suggesting that a rising JCV-index is another risk factor for the development of PML. This paper would suggest that this may only apply to MSers with a low index as once you have a high index the assay is saturated and can't detect a rising level of antibody. 

This paper reinforces the risks associated with staying on natalizumab if you are JCV positive, particularly if you have a high JCV antibody index. How good is the index at stratifying risk? It depends which side of the fence you sitting on. As far as lab assays go it not very good; we have a statistical test called a receiver operating curve, or ROC, analysis that we apply to diagnostic or predictive tests. When you apply a ROC analysis, setting the cutoffs for sensitivity and specificity at 80%, the assay does not pass muster. Although the cut-offs are quite sensitive (>90%) they are not specific (<45%). Why an 80% cut-off? This is the traditional thresholds that has been set by laboratry scientists for allowing an assay into routine clinical practice. Can I make this simpler to understand? Yes, 1 in 10 cases of PML developed in MSers with an index less than 1.5, 1 in 13 with an index less than 1.3, 1 in 23 with an index less 1.1, 1 in 59 with an index less than 0.9 and 1 in 167 with an index less than 0.7. Even simpler, if you are JCV positive on natalizumab you should seriously consider other options. Hopefully we will be able to derisk natalizumab therapy in the future.

StatsSpeak: Please note that a median value is used in this paper instead of an average or mean to describe the JCV-index; the median describes the value above and below which 50% of the population values lie. It is often used instead of an average or mean when there a very few samples or if the samples are not normally or symmetrically distributed. The mode refers to the commonest value and is not relevant to this paper. I will do a more detailed post on ROC analysis in the future."


Epub: Vennegoor et al. Longitudinal JCV serology in multiple sclerosis patients preceding natalizumab-associated progressive multifocal leukoencephalopathy. Mult Scler. 2015 Feb pii: 1352458514567728.

Background: The presence of anti-John Cunningham Virus (JCV) antibodies is a risk factor for the development of progressive multifocal leukoencephalopathy (PML) in MSers treated with natalizumab. It has been suggested that an increase in serum anti-JCV antibody index precedes the development of PML. 

Case reports: We here describe extensive longitudinal serum anti-JCV antibody indexes of four MSers who developed PML. Anti-JCV antibodies were measured using the STRATIFY JCV™DxSelect™ test. All four MSers had rather stable high anti-JCV antibody indexes in all samples obtained before developing PML. Possibly caused by reaching the saturation level of the assay, no increase in anti-JCV antibody indexes was seen just before the diagnosis of PML. 

Conclusion: This study confirms that high serum anti-JCV antibody indexes precede natalizumab-associated PML.

CoI: multiple

Labels: , , , ,