Diagnosis of central nervous system vasculitis caused by varicella zoster virus (VZV) infection requires detection of intrathecally synthesized anti-VZV antibodies. One way of doing this is to test paired serum and cerebrospinal fluid samples and to calculate the antibody index using the Reiber formula. This method is not commonly used in the USA but can improve diagnostic accuracy. CLI chatted to Dr Elitza Theel and Dane Granger (Mayo Clinic, Rochester, MN, USA) about their work to initiate this method.

Can you start by giving us some background about varicella zoster virus, please?

Elitza Theel (ET): Varicella zoster virus (VZ) is a member of the Orthoherpesviridae family of herpes viruses, similar to herpes simplex virus (HSV), Epstein–Barr virus (EBV) and cytomegalovirus (CMV). These are enveloped viruses with a DNA genome. They’re quite ubiquitous and they’re highly contagious, spreading through aerosols. Typically, an individual will inhale an infected aerosol into their respiratory tract; the virus then replicates in lymphoid tissue (such as the tonsils) before it spreads and disseminates to other organs, including most notably the skin.

After infection, there’s an incubation period anywhere from 10 to 21 days, after which an otherwise naive non-immune individual will typically develop primary infection that most commonly manifests as varicella (commonly known as chickenpox), which is really the dominant manifestation. Classically, patients develop this vesicular rash on their body. Some of the symptoms that appear before the rash develops and continue into the first week of illness include fever, malaise and abdominal pain. Most cases of chicken pox will resolve; however, there can be significant complications, such as secondary bacterial infections of the lesions to VZV pneumonia as well as very severe consequences like VZV encephalitis or meningitis. Typically, those more severe consequences occur in elderly individuals as well as those that are significantly immunosuppressed. Additionally, naive pregnant women are at risk of potentially passing that infection to the fetus, which can develop a congenital varicella infection and that can have pretty severe consequences.

Most people will recover after the primary VZV infection; however, the virus will establish latency in those individuals, typically in the sensory trigeminal nerves or the dorsal root ganglia. The virus will remain there ‘dormant’ until, for whatever reason, it reactivates and then spreads via the neurons, typically to the skin, leading to herpes zoster (or what we refer to as shingles). Shingles, again manifests as a vesicular rash, but in this case it’s typically restricted to a specific dermatome or specific area of the body rather than being disseminated and it can be quite painful. The best way to avoid chicken pox and to avoid the potential complications of chickenpox and shingles is to get vaccinated, both having your primary vaccine series with MMRV [mumps, measles, rubella, varicella] as a child and then getting the shingles vaccine later in life. The age of availability of the shingles vaccine varies depending on the country you live in: for example, 50 years and older in the USA, Canada and Europe, and 65 and older in the UK.

How does VZV infection of the central nervous system (CNS) arise?

ET: VZV infection of the CNS is one of the more severe complications. Although it is not very common, CNS infection with VZV can lead to the development of different clinical manifestations, which are important to understand because the diagnostic approaches differ. For example, VZV CNS infection can lead to acute meningitis, myelitis and encephalitis and the highest risk is in elderly and immunocompromised individuals. Only around 1.5% of individuals with zoster will go on to develop some sort of CNS infection; however, CNS infections can develop after primary VZV infection as well. CNS infection is thought to occur via spread of VZV along the nerves and then direct invasion and infection of the CNS, including cerebral arteries. Interestingly, as far as I can see, VZV is one of the only viruses that’s known to specifically infect cerebral arteries, which is what leads to VZV vasculopathy.

The diagnostic approach is fairly straightforward for VZV encephalitis and meningitis; however, VZV vasculopathy is much trickier to diagnose. And that is because it often develops weeks to months after resolution of the herpes zoster infection and is largely the result of inflammation of the cerebral arteries that leads to their remodelling and weakening, which then typically results in strokes, aneurysms and the development of what’s referred to as giant cell arteritis. So, unlike the other manifestations of VZV CNS infection, vasculopathy occurs after VZV DNA has declined and disease is really an immune mediated response. So in those patients, testing of spinal fluid by PCR to detect VZV DNA is not really beneficial and we really have to rely on a serologic response at that point.

Why is testing necessary for VZV?

ET: It is difficult to distinguish between viral or even in some cases bacterial pathogens that are causing the CNS infection. Typically, these days, patients with acute meningitis/encephalitis will be tested for different bacterial, fungal and viral pathogens using a molecular assay panel, that also includes many of the herpes viruses including VZV. For those acute presentations, the cerebrospinal fluid (CSF) will usually be positive for VZV DNA. It is important to pinpoint the causative agent of those infections because you can have much more targeted antiviral, antibacterial or antifungal therapy, as required. So testing is definitely important in those cases.

Typically, for a classic presentation of chickenpox and shingles, a clinical diagnosis is sufficient – especially in a patient that has not been vaccinated before. However, in some cases it might be necessary to confirm the diagnosis, for example if the rash is not classic or in suspected shingles if the vesicles aren’t in a typical clustered pattern. Often, clinicians will confirm the diagnosis by swabbing the vesicle and submitting the sample for PCR testing to detect the VZV DNA. Again, it is important to definitively determine what the cause of the rash is for a number of reasons, including helping to control the infection and limiting spread to naive individuals and more at risk patients. Additionally, in some cases where patients may be at risk for consequences, they might receive antiviral treatment with acyclovir or valacyclovir. So there are multiple reasons to confirm the diagnosis.

How is testing for CNS-VZV usually done and what are the limitations of these methods?

ET: As mentioned above, for acute manifestations we collect CSF and test it using a molecular approach to detect VZV DNA. However, because of the delayed manifestation of – in particular – VZV vasculopathy, detection of VZV nucleic acid is very insensitive. One of the best studies to show this was by Maria Nagel in 2007 [1] where they identified 14 patients with VZV vasculopathy. Testing the patients’ CSF by PCR showed that only four of those 14 were positive, whereas 93% (13 out of the 14) were positive using an antibody index test. Dane, I’ll let you talk about what that is.

Dane Granger (DG): Thank you, yes, the antibody index test is considered the gold standard for intrathecal antibody detection. We have to keep in mind though, that it is normal for VZV IgG to be present and detected in the CSF if your serologic test is sensitive enough given that many people have been vaccinated for VZV or have had the infection, because IgG antibodies can cross the blood brain barrier. So, it is important to discriminate between passively diffused versus intrathecally, or locally produced, antibodies. The way to do this is to establish a VZV IgG antibody index. There are a couple methods for attaining an antibody index, of which the Reiber method is more widely used in Europe [5] and consists of determining two quotients.

The first quotient is the CSF to serum ratio of IgG against your pathogen of interest. In this case any of the serologic tests where you could get an accurate measurement of the anti-VZV IgG concentrations can be used – enzyme immunoassays (EIAs) etc. Typically, an EIA with a curve relating to relative units is used to test paired serum and CSF samples collected within 24 hours or so of each other. The resulting CSF and serum values are each multiplied by the dilution at which they were tested and the resulting CSF value is divided by serum value to get your pathogen-specific IgG quotient.

The second quotient can be attained using one of two methods, regardless, the goal is to baseline this against total IgG levels that passively diffuse across the blood-brain barrier. What we do in our lab is to measure total IgG by nephelometry for both CSF and serum and create a total IgG quotient by dividing the CSF measurement by the serum measurement. To get the VZV antibody index, you then divide the pathogen-specific quotient by the total IgG quotient.

The Reiber method recommends testing for albumin in CSF and serum. And that case arises when your measured total IgG levels are significantly elevated due to high intrathecally (locally) produced IgG, which can occur for example, if the patient has multiple sclerosis. Now the baseline, the denominator total IgG quotient, has been compromised and artificially raised because you have an excess of CSF IgG, which could result in a false negative pathogen-specific antibody index. In these scenarios, Reiber recommends measuring albumin concentrations in both CSF and serum as an alternative. Again, we generate a CSF to serum ratio of albumin and that quotient is then used in a larger equation that defines a diffusion constant that uses albumin levels to determine what your theoretical limit of total IgG would be in the CSF due to passive transfusion. Albumin is only produced in the liver, so it’s a perfect surrogate. If you can measure how much albumin is in the CSF, you would know your total IgG quotient limit (QLim). If your measured total IgG is higher than your quotient limit, you would then use your quotient limit value. If your measured IgG is lower than your quotient limit, you go ahead and use your measured total IgG for the denominator. Ultimately, regardless of which approach is taken, an antibody index value of 0.7 to 1.3 is normal (i.e. negative), whereas 1.5 or greater would be indicative of intrathecally produced anti-VZV IgG (in this case). These cut-off levels can be verified or established by each individual laboratory through their own validation, but those are the typically agreed values for the classic Reiber formula, which is what we do in our lab.

The second method that can be used, I would call a pre-dilution method where again you would measure your paired serum and CSF for total IgG and take that ratio, dilute your samples to that ratio and then test those on your ELISA. I think a limitation there would be that you are really not verifying that all of your total IgG is passively diffused because you’re not bringing in that albumin component.

So those are the two approaches. I just want to add in that there are some limitations, the antibodies that we are testing for – for VZV or other analytes – in disease states can be persistent, so you may still have a positive antibody index possibly years later. This is a limitation no matter which antibody index method you would use.

Another limitation, from a pre-analytic standpoint, is with a traumatic lumbar puncture you will get CSF blood contamination. So, it’s important to always do your due diligence at the pre-analytic stage of screening those out and not testing them as the calculation cannot correct for that, and also being very careful to catch xanthochromatic samples which can negatively impact results as well.

ET: Yes, we don’t want to test CSF that’s contaminated with blood because if the point is to ensure detection of intrathecal antibody synthesis, you can’t guarantee that is what you are measuring if you have blood in the sample, so those samples should not be tested.

Basic Reiber formula

What are your thoughts about tips for these tests or potential improvements that could be?

DG: I think some areas for improvement would be using assays that have higher analytical measuring range or dynamic range so that you can reduce the number of dilutions that you have to perform. Also, I think another way, possibly, would be determining not only the antibody index but the antibody titers over time – so doing longitudinal measurements to see if you see a drop of delta over time, maybe that could help to clarify the situation and demonstrate that even though you have a positive antibody index, are you seeing a drop in titre that would suggest disease resolution?

ET: I think it’s important to understand that with this method, it’s not like an ELISA that you can do pretty quickly, there’s a lot of calculations, and a lot of collaboration with other laboratories, because, for example, our lab does not do total IgG and albumin testing. So there’s a number of obstacles to overcome and these are difficult tests to implement, and in the USA at least, there’s only a couple of labs that do this, including ours.

Also, for the antibody index test, as Dane indicated, we need paired CSF and serum. And for our test at least, we require those two to be collected within about 24 hours of each other. Because there is routine turnover of the CSF, the concern is that if you collect CSF on Monday and then serum on Saturday, as you’re comparing antibody levels in those two sources, there might be quite a bit of variability in those samples that could affect the results. So timing of paired specimen collection is really important for test accuracy as well.

DG: The devil in the details with these tests. Not all assays will have ideal linearity. Having good ELISA measurements is very important, so what we end up doing is an endpoint dilution and we always select the serum and corresponding CSF results when they fall into a very similar range in the standard curve of the assay to avoid inaccuracies due to non-optimal linearity if using a different part of the curve. A lot of those details really add up to make sure you’re giving a really accurate measurement.

ET: If a lab is interested in bringing this sort of assay up, I highly recommend contacting a laboratory that already has these tests up and running to talk through some of the details – some of the failures that those labs have seen, advice for how to implement it, and also for help with validation and verification. One of the biggest challenges we had was finding paired CSF and serum samples – if you don’t do that routinely, it’s really difficult to acquire specimens for assay verification.

I would also suggest working/pairing up with a clinician proponent, be it in infectious diseases or in neurology. Having that support from the clinical side is really important for when you bring in this sort of test to replace a classic technique.

DG: Another thing I would recommend (and this is something we’re going to add in) is a little more validation data to our VZV method, because now we’re testing for antibodies to a virus for which upwards of 80 to 90% of the general population is seropositive. We now raise the risk of false positive results because nearly every single patient is going to have a positive antibody test. So, we really want to work to minimize those false positives, maybe by doing some extra work with the normal values which then applies some rigor to the Reiber recommended greater than 1.5 index, and make sure that fits with what you’re seeing in our “healthy population”. Obviously these aren’t healthy people if they’re getting a lumbar tap, but maybe you can find a population that isn’t needing microbiology/infectious disease work-up. Also, you may need to play around with what you’re starting dilution level will be before you start calling VZV IgG EIA results positive. Alternatively, it could be at the other end, where you might increase your final antibody index value from 1.5 to a little higher if you’re really concerned about mitigating the adverse outcomes of false positives.

Initially a test like this is very much for the labs that have the mindset of doing lab-developed tests – it would be a challenge for a lab that doesn’t have that background but works more with routine larger analysers, random access analysers where they’re really using a lot of FDA cleared tests.

Other applications

ET: Antibody index testing is also a component of the European and the Infectious Diseases Society of America (IDSA) guidelines for how to diagnose neuroborreliosis caused by Borrelia burgdorferi. So there is precedent for using this sort of approach to detect true intrathecal synthesis of antibodies for diagnosis of neuroinvasive disease. In our laboratory, we also recently started performing antibody index testing for diagnosis of neurosyphilis. To my knowledge, nobody else in the in North America uses that for neurosyphilis. In the USA, for testing for syphilis we rely on the Venereal Disease Research Laboratory (VDRL) test, which was developed in the 1940s and is not even specific to syphilis. So having a specific anti-treponemal antibody and being able to say yes, we detect true intrathecal synthesis versus not as important. I know we digress, but my point is there is precedence for doing this sort of testing for sure.

The interviewees

Dr Elitza S. Theel PhD, D(ABMM), F(ADLM), F(AAM)
Director, Infectious Diseases Serology Laboratory | Co-Director, Vector-Borne Diseases Laboratory Service Line | Director, STI Testing | Professor, Laboratory Medicine and Pathology | Division of Clinical Microbiology

Email address: Theel.Elitza@Mayo.edu

Dane Granger BS
Principal Developer at Mayo Clinic
Assistant Professor of Laboratory Medicine and Pathology Division of Clinical Microbiology, Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN, USA

Email address: Granger.Dane@Mayo.edu

Bibliography
1. Nagel MA, Forghani B, Mahalingam R, Wellish MC, Cohrs RJ, Russman AN et al. The value of detecting anti-VZV IgG antibody in CSF to diagnose VZV vasculopathy. Neurology 2007;68(13):1069–1073 (https://doi.org/10.1212/01.wnl.0000258549.13334.16).
2. Nagel MA, Cohrs RJ, Mahalingam R, Wellish MC, Forghani B, Schiller A et al. The varicella zoster virus vasculopathies: clinical, CSF, imaging, and virologic features. Neurology 2008;70(11):853–860 (https://doi.org/10.1212/01.wnl.0000304747.38502.e8).
3. Nagel MA, Jones D, Wyborny A. Varicella zoster virus vasculopathy: the expanding clinical spectrum and pathogenesis. J Neuroimmunol. 2017;308:112–117 (https://doi.org/10.1016/j.jneuroim.2017.03.014).
4. Chapter 102: Varicella-Zoster Virus. In: Carroll KC, Pfaller MA, Landry ML, McAdam AJ, Patel R, Richter SS, Warnock DW (eds) Manual of Clinical Microbiology, 12th edn. ASM Press 2019. ISBN 978-1555819835.
5. Reiber H, Lange P. Quantification of virus-specific antibodies in cerebrospinal fluid and serum: sensitive and specific detection of antibody synthesis in brain. Clin Chem. 1991;37(7):1153–1160. PMID: 1855284.
6. 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(2):189–203 (https://doi.org/10.1016/0022-510x(94)90298-4).