Lyme algorithms: stick to standard, move to modified?

Charna Albert

June 2023—For Lyme disease testing, immuno­blots became optional in 2019 when the FDA cleared enzyme immunoassays for use as part of a modified two-tiered testing algorithm. “It was a historic event in the world of Lyme diagnostics,” says Elitza Theel, PhD, D(ABMM), director of the infectious diseases serology laboratory and co-director of the vector-borne pathogens service line at Mayo Clinic.

But the standard two-tiered testing algorithm with its immunoblots hangs on.

Dr. Theel, a member of the CAP Microbiology Committee, which monitors, through the CAP Tick-Transmitted Diseases Survey, how many labs have made the switch to the modified algorithm since the FDA’s approval of multiple ELISAs for use in confirmatory testing, says the number is smaller than expected.

“I suspect it has to do with a number of things: the SARS-CoV-2 pandemic and then mpox, and all the challenges we are experiencing with a limited workforce. In microbiology, like other areas, we continue to experience a shortage of technologists. So you prioritize things that have to get done with the resources you have, and I’m assuming that transitioning to the modified Lyme algorithm is not a priority with everything else going on.”

“So there have been constant pressures and stressors on laboratories over the past few years,” she continues. “And many laboratories rely on reference labs to do their immunoblots rather than run them themselves, and the reference labs continue to offer Lyme blot testing, so if it’s not broken, why fix it?”

A limited understanding of the modified algorithm’s advantages may play a part too.

But immunoblots are immuno­blots. Nobody really wants to do them, she says. “They’re clunky.” Although most labs have transitioned to using automated processors to perform blot testing, “those instruments are typically dedicated to just Lyme blots; they take up precious space and require maintenance,” Dr. Theel says. “And even though we now have optical densitometry readers to help us interpret blot results, some can still be difficult to finalize, despite those readers. So they’re still challenging.”

Under the standard two-tiered testing algorithm, a reactive sensitive enzyme immunoassay is confirmed by supplemental immuno­blot testing for IgM and IgG antibodies to Borrelia burgdorferi.

The modified two-tiered testing algorithm is entirely EIA or chemiluminescent immunoassay-based. “There are still two tiers, but the second-tier assays can be immunoassays rather than blots, with the key requirement that the Borrelia antigens used in the two immunoassay tiers differ,” Dr. Theel says.

With the modified algorithm, a laboratory with an EIA processor can perform first- and second-tier testing on the same instrument and in-house, with a quicker turnaround time to result. When it comes to interpretation, “with the modified algorithm you just get positive or negative, which can be easier to interpret than remembering the number of bands you need to see on immunoblots for them to be considered as either positive or negative.” The Western blot assay is considered IgM-positive if reactivity is observed in at least two of three B. burgdorferi antigens, and IgG blots are positive if reactivity is seen in at least five of 10 antigens. While all laboratories report qualitative result interpretations, some laboratories may also report the bands that are detected for qualitatively negative immuno­blots. The reasoning for this varies, Dr. Theel says, with some clinicians wanting to know what antigens are reactive, regardless of the qualitative interpretation.

“This, however, can lead to both clinician and patient confusion, unfortunately. Patients have a hard time understanding this concept—that even though you have bands there, it doesn’t necessarily mean infection. It could be cross-reactivity or other nonspecific binding. So one of the benefits of the modified algorithm is that it eliminates the challenges of blot banding pattern interpretation and visibility on both the clinician and patient side, which I think is an improvement, particularly for those individuals who are not as familiar with the caveats associated with this sort of testing.”

Though Mayo Clinic has fully transitioned to the modified algorithm (“with a lot of notification and education,” she says), Mayo Clinic Laboratories’ reference lab clients still have the option to order the standard algorithm. “The majority of our Lyme testing is still using the standard algorithm with the immunoblots.” And the lab does report out which specific bands are detected, even if the qualitative result for the blot is negative. “Back in the day, we did not report bands and we got pushback from clinicians who wanted to see how negative was a negative and how positive was a positive, so we decided to report bands out. But now we hear from health care workers who call in and say, ‘Well, you called it negative, but you also list out detected bands, so what does that mean?’” The challenge is ongoing and it is difficult to please everyone, she notes. It’s something that the lab tries to address in its result comments.

“For the Lyme-savvy clinicians, they can look at the banding pattern to see which antibodies are present against which antigens, which in some cases can help provide some information on whether this is a new recent infection versus a more remote infection or possibly cross-reactivity, because you’ll see different antibodies develop to different antigens at different stages of infection. That’s at least partly the reasoning for including bands on negative test results.” Not everyone is aware of those nuances of testing, however. “Ultimately, it causes more confusion than it helps,” in her experience.

Dr. Theel and coauthors evaluated two distinct modified two-tiered testing algorithms using Zeus Scientific Lyme disease EIAs (Sfeir MM, et al. J Clin Microbiol. 2022;60[5]:e02528-21). “Our study,” she says, “was one of the first to look at those assays that had received FDA clearance for use as part of a modified algorithm to see how they performed.” One of the algorithms they evaluated used a VIsE1/pepC10 polyvalent EIA, followed by a whole-cell sonicate polyvalent EIA. The second algorithm used the same first-tier EIA followed by separate IgM and IgG whole-cell sonicate EIAs. In a retrospective phase, the authors compared each modified algorithm to the standard algorithm using archived samples from Lyme disease patients or control subjects.

The first and second modified algorithms were more sensitive (56 percent and 74 percent) than the standard algorithm (41 percent) among 61 patients with acute erythema migrans. In 75 Lyme disease patients with neuroborreliosis, carditis, or arthritis, sensitivity was comparable between algorithms. Compared with the standard algorithm, both modified algorithms provided increased sensitivity in erythema migrans patients, comparable sensitivity in later disease, and non-inferior specificity.

Says Dr. Theel, “We showed improved sensitivity early on, no change in sensitivity at later stages, and equivocal specificity relative to the standard algorithm.” The sensitivity of the modified algorithm in patients with acute disease (73 to 77 percent, according to the literature, compared with 35 to 50 percent by the standard algorithm) is its primary advantage, she says (Fig. 1), though its sensitivity remains imperfect. “Early on in disease a negative result still does not rule out infection.” Another advantage of the modified algorithm: Antibodies to a wider range of Borrelia species can be detected. “The blots we use in North America are specific for B. burgdorferi strain B31,” she says. “They will not effectively detect antibodies to other species of Borrelia.”

The modified algorithm is susceptible to the same causes of cross-reactivity as the standard algorithm, and, like the standard algorithm, the modified algorithm can’t be used to monitor response to therapy. “And it’s still going to be challenging to diagnose reinfection,” she says. As a qualitative method, the ELISA provides an intensity signal rather than a quantitative antibodies value. “So we do not have the ability to look at IgG antibody expansion over time, like we could with the blot assays, and that can be helpful when trying to determine a potential Lyme reinfection case.”

Suppose a patient had Lyme disease previously and presents a year later with similar symptoms, she says. “Initially they have six bands, but then you retest them and they’ve moved up to 10 bands over a two-week period. That would be highly suggestive that this is a new recent reinfection.” With the two-step EIA sequence, “if you test them now they’re going to be positive because of that prior infection, and if you test them in two weeks they’re still going to be positive. You aren’t able to tell if there is a progression or expansion of the immune response.” The index values from the ELISA readouts can provide information about whether the result is a “high-positive,” which would be a high number, or a “low-positive,” which would be a value near the threshold. “But it’s not a true quantitative assay, so clinicians shouldn’t be using this to monitor values or levels over time. Plus, we do not report index values on our patient results.”

At Mayo Clinic, the first-tier test is a VIsE/pepC10/OspC total antibody EIA. “If those are reactive, we automatically reflex to two separate EIAs looking for IgM and IgG antibodies to whole-cell sonicate material from Borrelia burgdorferi,” she says. If either of the supplemental assays is positive, the patient is considered positive.

Another option for the second-tier test is a single whole-cell sonicate IgM/IgG ELISA, which does not differentiate between the classes of antibody present. “It’s potentially beneficial from a laboratory perspective because you’re only running a single EIA, but it doesn’t give enough information from a clinical interpretation perspective,” Dr. Theel says.

Guidance from the CDC and others says IgM-specific results for Lyme should not be interpreted in patients with more than 30 days of symptoms. “That’s because that IgM might be there due to a past infection or it may be present due to cross-reactivity leading to a false-positive.” With the single IgG/IgM ELISA, “you can’t tell if it’s reactive or positive because you’re detecting IgM, in which case you’d want to know if the patient has more or less than 30 days of symptoms, or if you’re detecting just IgG.”

Another option for the lab would be to provide two different orders: a first-tier test that automatically reflexes to two separate IgG and IgM EIAs, to be ordered only for patients with fewer than 30 days of symptoms, and a first-tier test that automatically reflexes to a single IgG EIA, to be ordered only for patients with more than 30 days of symptoms. “But that would likely be challenging for clinicians, so we offer a single orderable with a reflex to IgM and IgG and include in comments that the IgM should not be considered for patient management if he or she has more than 30 days of symptoms.”

Result reporting for the modified algorithm at Mayo Clinic is sequential rather than simultaneous, with the laboratory reporting a positive first-tier test, owing to a limitation of the laboratory information system. “My preference would be that the results would be reported after all testing is completed, with the final interpretive comment, because when you report a positive result for the first-tier test—which is what the package inserts and guidance documents recommend—there are still providers who are confused and think that this is the final result, and that may lead to additional or potentially unnecessary treatment.”

At Mayo Clinic, when the initial test is positive, the lab says in the comment that it is not diagnostic and that supplemental testing has been ordered by reflex. “But that of course depends on the clinician reading the comment, which does not always occur,” she says.

The Association of Public Health Laboratories released a Lyme disease algorithm reporting aid, to which Dr. Theel contributed and which she recommends laboratories refer to when putting together their Lyme testing algorithm. In Fig. 2 is Table 4 from the document (the full document is at https://shorturl.at/jqt01).

Neuroinvasive Lyme disease, which occurs in up to 15 percent of untreated Lyme disease patients, requires a different testing strategy. “We should not be using the modified or standard two-tiered testing algorithms on spinal fluid,” which were validated strictly on serum, Dr. Theel says. The recommendation for neuroinvasive Lyme, alongside clinical presentation, is to perform a Lyme antibody index test. “You compare the level of Lyme antibodies in cerebrospinal fluid versus serum relative to total immunoglobulin in those two sample types. It’s a much more specialized testing process to ensure that any antibodies detected in spinal fluid are there due to intrathecal synthesis versus passive diffusion of the antibodies from serum into the spinal fluid.” Mayo Clinic implemented the assay in 2018.

Used with permission; copyright Association of Public Health Laboratories. All rights reserved. Full document at https://shorturl.at/jqt01.

Neuroinvasive Lyme disease has both peripheral and central nervous system manifestations: cranial neuritis, radiculoneuritis, and plexopathies, or meningitis, encephalitis, and myelitis. “And typically you see a lymphocytic predominance in spinal fluid testing,” she says. Per the Infectious Diseases Society of America, a patient with one or more of these symptoms and an epidemiologically plausible exposure to ticks infected with B. burgdorferi would meet the criteria for testing.

The CDC doesn’t differentiate between systemic and neuroinvasive Lyme disease, Dr. Theel says. “Whether you have neuroinvasive Lyme or just Lyme, you’re counted as a Lyme disease case. So we don’t have a good sense of the prevalence of neuroinvasive Lyme disease in the U.S.” Mayo Clinic sees a peak of neuroinvasive Lyme every summer and fall, at about five percent during August and September annually. “So I think it’s much more prevalent than we think.”

Neuroinvasive Lyme is often thought of in connection with B. garinii, she says, a Borrelia species found in Europe. “So we think that’s a European manifestation of Lyme disease. But it’s also a manifestation of infection with Borrelia burgdorferi, and I don’t think we’re as cognizant of that fact or consider it as frequently as we perhaps should.”

Charna Albert is CAP TODAY associate contributing editor.