All in one and one for all: multiplex testing

March 2007
Feature Story

Anne Paxton

In the movie world, it’s a theatre with multiple screens.

In signal processing, it’s combining many signals into a single channel. In juggling, it’s throwing many balls at one time with the same hand.

It’s multiplexing. And the clinical laboratory version of it—conducting multiple tests simultaneously on a single sample—is transforming diagnostic testing for infectious diseases, autoimmune disorders and allergies, toxicology, and genetically inherited diseases.

“The beauty of multiplexing is it allows you to do so many tests at one time. Using flow cytometry technology and being able to do multiplexing has opened a whole new vista in our ability to treat patients more accurately,” says rheumatologist Steven Mendelsohn, MD, PhD.

Multiplex testing screens for multiple targets, typically more than 10 pathogens or disease markers, in a single test. Positive results can be reported in qualitative and semiquantitative forms. The developer of the basic xMAP technology on which suspension-bead array multiplex testing is based, Luminex Corp. of Austin, Tex., explains that some other technologies such as “lab on a chip” or microarrays are sometimes described by the term “multiplexing,” but are fundamentally different. Unlike xMAP, they do not technically measure different things at the exact same time; they may require multi-step processes that aren’t significantly different from traditional ELISA technology.

Multiplexing’s main attraction: significant savings in reagents, consumables, and time. “Just being able to get multiple results on a single sample at a single time can improve your efficiency drastically,” says Denise Facaros, CLS (ASCP), laboratory manager for HealthCare Clinical Laboratories, Stockton, Calif.

Four features of multiplexing are making the laboratory world sit up and take notice:

• It permits the performance of many tests on a much smaller patient sample.
• It conducts every individual test exactly the same way, eliminating the effect of different methods, timing, or conditions on results.
• Control tests are more reliable because they are performed under the exact same conditions as the unknown—that is, simultaneously and in the same well.
• It is more likely to determine cause of disease or targets of drugs. Rather than being ordered based on a hypothesis, it tests for all known markers in “shotgun” fashion.

The Luminex multiplex technology is currently used in the Tm Bioscience ID-Tag RVP (Respiratory Virus Panel) assay. The RVP multiplex panel detects 19 respiratory viruses, including influenza A, subtyped to H1, H3, or H5; influenza B; parainfluenza viruses 1, 2, 3, and 4; adenovirus; respiratory syncytial virus (RSV); four different coronaviruses (229E, OC43, NL63, HKU1); human metapneumovirus (hMPV); and group rhino- and enteroviruses. The European CE marked version, available only outside the U.S., will also detect the SARS coronavirus.

Viral nucleic acids are extracted from respiratory samples; a one-step RT-PCR using 16 different primer sets is performed, followed by target-specific primary extension. “The amplified products are then hybridized to the Luminex beads that have special tags that recognize the different respiratory virus nucleic acid targets,” says Christine C. Ginocchio, PhD, director of microbiology, virology, and molecular diagnostics for North Shore-Long Island Jewish Health System Laboratories, New York. “The Luminex xMAP system, instrument and software, identify which beads have a fluorescent signal that relates to a specific viral target.” In addition to the 19 targets, the assay detects internal and external controls. “The internal control is used to monitor the entire performance of the assay from nucleic acid extraction through the Luminex xMap detection process, and the external control is used to monitor the integrity of the reagents,” she says.

The North Shore-LIJ laboratory, which performs more than 4 million tests a year for 11 hospitals in the Long Island area, conducted clinical trial studies using about 300 patient samples and the Tm Bioscience ID-Tag RVP multiplex panel. Dr. Ginocchio presented a poster at the recent Seasonal and Pandemic Influenza meeting in Washington, DC, that emphasized the assay’s reproducibility and its ability to detect mixed infections. “Overall, we detected a variety of respiratory viruses in 29 percent of specimens tested by direct immunofluorescence [DFA], in 49.9 percent of the specimens by rapid viral culture using R-Mix cells [Diagnostic Hybrids, Athens, Ohio], and in 64 percent of the specimens by the RVP assay,” she says. “The increase in identifying specimens positive for a respiratory virus was due to the fact that we are detecting viruses we normally do not culture for or grow routinely in the laboratory—the rhinoviruses, parainfluenza 4, and coronaviruses.”

Dr. Ginocchio says the Luminex instrument is good for many types of multiplexing. “The current system can use a hundred different beads to detect a hundred different targets, with the capacity for many more.”

She believes multiplex testing is going to play a significant role in testing for infectious diseases. “First of all, it gives laboratories the ability to detect viruses that cause severe respiratory disease, that we previously had no simple way of rapidly detecting in the laboratory.” This will greatly expand the number of times a diagnosis can actually be made, she says, especially when the causative agent is a virus that may die quickly (for example, RSV) or is difficult to grow in routine culture (for example, hMPV).

“We also identified a number of mixed infections we do not routinely pick up by routine viral culture,” Dr. Ginocchio says. These data may shed light on why there are different clinical outcomes to certain respiratory infections—for example, why some children are seen in the emergency department with RSV infections and are discharged while other children with RSV end up in the intensive care units. “Studies from our laboratory have shown that a more severe clinical course is often related to a mixed infection—RSV and another respiratory pathogen such as hMPV or a coronavirus,” she says. Identifying mixed infections may help in understanding clinical outcomes, disease progression, and the correlates of increased severity of disease, she adds.

Also important is the RVP assay’s ability to differentiate viral subtypes, particularly for influenza A, H1, H3, and H5. “If we ever have to deal with an avian influenza outbreak, we will be able to not only identify influenza A but subtype the influenza A strains at the same time, with the results basically available in a day,” she says, comparing that with traditional viral culture, which may take five to 14 days to identify certain viruses.

“R-Mix culture is relatively fast and detects approximately 90 percent of respiratory viruses in 24 to 48 hours, but it is limited to eight or 10 viruses, not the comprehensive 19 viruses detected by the RVP assay,” she points out.

Culture also cannot differentiate subtypes of influenza A, which would require supplemental testing that the majority of clinical labs generally don’t perform. In addition, the rapid tests, the 15- to 30-minute test for influenza A and B, are not able to differentiate between the subtypes of influenza A and have inferior sensitivity to detect influenza A and B when compared with both culture and molecular detection. Currently, the rapid tests are also limited to detecting just influenza A and B and RSV.

The time frame for running the test, including extraction, is six to 6.5 hours, Dr. Ginocchio says. “This is a significant improvement over culture and also in the sense that you can detect so many viruses at one time.” She says they have real-time molecular assays in the laboratory that are more rapid—2.5 hours—and as sensitive as the RVP assay for the detection of RSV and hMPV. “These assays are excellent,” she adds, “but to cover the scope of the RVP assay you would have to set up many individual assays to get the same result.”

The RVP assay is also a fraction of the cost. “If we had to run 19 individual real-time tests or even six to 10 real-time mutiplex tests, that probably would cost the laboratory just in reagents alone between $250 and $500, depending on whether or not the tests were developed in-house or use analyte-specific reagents.” In contrast, the RVP test would be one-fifth to one-tenth of that cost. Though a real-time assay requires less hands-on time and fewer technical steps, “this would have to be multiplied over many assays to get the same 19 results as the RVP assay,” Dr. Ginocchio says. “The time to results may be a little longer, but results are still available within a day, at a fraction of the cost, and you only need to set up a single test per sample.”

Under Food and Drug Administration (FDA) guidelines, manufacturers are not allowed to sell in the U.S. analyte-specific reagents that are multiplexed, thus limiting the number of multiplex assays available to only those that have received FDA approval. “Currently there are no multiplex tests approved by the FDA for the detection of respiratory viruses, but we anticipate that the RVP assay will be through the FDA within the next several months,” she says.

One of the assay’s drawbacks is working with amplicons in the laboratory. “You do have to open up trays that have amplified products,” Dr. Ginocchio says. “Many laboratories like ours have transitioned our traditional PCR assays to real-time platforms because we did not want to expose the laboratory to amplicons with the chance of cross-contamination of samples or contaminating the laboratory.” Using this assay requires the laboratory to revert to conventional PCR and to make sure it has in place the appropriate containment and cleanup so samples do not become contaminated in the laboratories. Therefore, for laboratories in particular that perform real-time PCR using enclosed systems only, it is essential to retrain technologists in the necessary precautions. “If you do not enforce the proper procedures, you could certainly have a problem in the laboratory with contamination if you’re not careful. This is probably the biggest criticism people have with the current format of the assay,” Dr. Ginocchio says.

The ability to rule in and to rule out certain diseases will be important, she emphasizes. “Within a single day, you can provide results with a broad-spectrum panel that covers the major respiratory viruses.” This information, she says, should promote the rapid initiation of appropriate antiviral therapy and the discontinuation of inappropriate antibiotic therapy, and provide insight into how respiratory viruses interact with each other.

The RVP assay will rapidly identify viruses circulating in the community and therefore can be used for epidemiologic and surveillance studies. “Currently we have these respiratory outbreaks caused by something we know is not influenza but we do not know what it is, and multiplexing will be very helpful in detecting the responsible respiratory virus,” she says. In her area last spring there was an outbreak of human metapneumovirus within a six-week period. “Without a molecular assay we would never have understood what caused the outbreak.”

Dr. Mendelsohn, a solo practitioner at Mountain Regional Medical Center and a clinical associate of the University of North Carolina, Chapel Hill, has been using the Athena Multi-Lyte system (manufactured by Zeus Scientific and marketed by Inverness Medical Innovations) in his rheumatology practice for three and a half years.

“I’ve been a rheumatologist for 25 years, and right from the get-go we were doing immunologic testing in my office with immunofluorescent technology, but I’ve always been disappointed in the results; they were so subjective and the test is so difficult to do. It also varies so much from laboratory to laboratory—there’s very little consistency,” he says.

He was intrigued when he heard about multiplex testing in flow cytometry. “We tried the Luminex machine for a little while, running parallels with a hundred patients doing analysis using traditional technology and the multiplex, and I was thrilled with the results. For us, it turned out to be more clinically accurate, and the correlations run 80 percent compared to any other technology. This actually gives you a number rather than a titer. You’re not just looking at a pattern and guessing—you’re actually getting an individual amount and measuring each antibody.”

He has an advantage, he points out, because he sees the laboratory side and the clinical side. “These results are correlating more closely with what we were seeing in the clinic. As a rheumatologist, I see lots of folks with autoimmune disease making antibodies against their own tissues, and the key is to actually measure exactly what they are making antibodies against.”

“For instance, a patient with lupus tends to make anti-DNA or anti-SM antibodies. Not only can you measure them individually with the multiplex test, and it’s very accurate and reproducible, but it also tells us whether the antibody levels are dropping or coming down, so you can monitor the patient as well as diagnose.”

“It sometimes even gives you a little bit of a glimpse into what might be happening with the patient in the next weeks and months.”

Since the test actually gives an ANA qualitative result plus quantitative measures of anti-DNA and eight separate antibodies, “you’re basically getting 10 pieces of information,” he notes.

Another advantage of multiplexing is that it doesn’t take more than a drop of serum. “An original ANA test would involve a single sample, then you’d have to come back with more sample for each individual test, and it would often take three or four vials of blood to get each one of those tests done.”

Looking through a fluorescent microscope and doing stains was more of an art form than a laboratory test, he recalls. “It took a good technologist maybe three to six months to get good at it, whereas this test is basically automated. You do the titrations, you do the mix, then the machine does the titration and sets up the beads and counts them.”

“Because you’re doing 100 assays for each one of the autoantibodies, compared to just one assay for more traditional methods, the technologist can do 100 different patients within a couple of hours—something that normally would take a week or two.” His clinic, which is relatively small, does its average of 20 tests a week in one batch.

He estimates that the financial outlays required are roughly comparable. “If you compare, for example, having to go out and buy a fluorescent microscope, that could be $4,000 or $5,000 a year to lease, so over five years with materials you had to buy, you would be spending about the same. Where you save money from the laboratory point of view is with technologist time.”

He is looking forward to the addition of more tests so the laboratory can take a small sample and measure up to 200 separate tests. “I know they are working feverishly to get rheumatoid factor and other autoantibodies available on this multiplexing technology.” But in addition, he would like to see more physician education on the availability and advantages of multiplexing in toxicology and infectious disease testing. “I think it will become a standard eventually. It’s so powerful having these microbeads quickly getting actual numbers rather than just doing guesswork.”

The manual ANA fluorescent test has never been very easy to do well, agrees Stephen Apfelroth, MD, PhD, director of the clinical laboratory for the North Bronx Healthcare Network, Bronx, NY. “We’ve been thinking for a long while about what to do with the ANA. Some laboratories had gone to putting a cocktail of antigens on an ELISA plate as the preliminary screening test, but the sensitivity and specificity have never been satisfactory to people, so they never really did away with the manual fluorescence test.”

Used most frequently for diagnosing lupus, the traditional test has several downsides. “It has to be a trained microscopist who does the reading, and usually there are only two or three people in the laboratory who do it regularly, because it’s a specialty item.”

“Then Athena came up with a system that could do a multiplex assay and get a number on each element of the cocktail instead of one number from the whole mix. That has the advantage of roughly correlating to what the patterns used to tell you. For example, antigens that come up positive give different fingerprints—they could be diffused patterns, say, or speckled patterns in ANA, and that’s something rheumatologists are looking at anyway.”

After a positive ANA, the rheumatologist would go back and order each separate antigen to get more information about which antigens were positive and how strongly. “But the multiplex test basically allows you to skip that test and get the information right away.”

“Since the multiplex test involves only one well, you’re handling one sample instead of having to do separate wells or separate tubes for different antigens; it does the whole array,” Dr. Apfelroth says. “So it’s much more convenient and it lends itself to assays that are done in multiplex routinely, particularly measles, mumps, varicella, the ToRCH assay [toxoplasmosis, rubella, CMV, HSV]. Those are the majority along with the EBV [Epstein-Barr virus] antibodies.”

Multiplexing is more suited to screening, he points out. “You could test for those antigens before, but who was going to test for nine separate antigens to get a screening result?”

However, as with most things, the unexpected cropped up. Luminex, he says, had to adjust the assay because the technology is unique and didn’t work particularly well with the artificial CAP proficiency testing material. “They were using a high-dilution no-wash method originally for their assays, and they seemed to have a problem with the CAP matrix effect with the artificial serum they were using, so they had to go to a lower dilution in order to get satisfactory results. But that seemed to be an issue strictly related to the artificial specimens and not with the patient specimens,” Dr. Apfelroth says.

Once a laboratory has obtained buy-in from rheumatologists on the multiplexing technology, the cost becomes a secondary issue, in his view. “If you compare the cost of an individual test to do nine individual analytes, that would be significantly more expensive. So it’s a little more expensive than traditional ANA, but you get a lot more information out of it.”

Luminex has licensed its multiplex technology to several manufacturers, but they’re not all employing it the same way. “We all use the same core xMap technology from Luminex,” says Michael Barcellos, global marketing head for the BioPlex 2200 Division of Bio-Rad Laboratories, Hercules, Calif. “We, however, use a proprietary version of that technology based on magnetic beads, which has allowed us to develop the first and only fully automated random-access system using their technology.”

“By adapting the Luminex technology to magnetic beads, we can develop reagents in a heterogeneous assay format on a fully automated system so labs can get improved sensitivity, specificity, and overall better performance by being able to perform all wash steps onboard.”

The advantage of random access is that it permits routine laboratories to bring in multiplexing technology just as high-throughput labs do.

“The technology, in a fully automated, random-access format, allows the laboratory to manage workflow as samples come in, so they no longer have to do batch testing, they’re able to load the system with multiple reagent packs, then load the samples being tested for any variety of tests on the menu,” Barcellos says.

Launched in 2006, the BioPlex 2200 has four U.S. customers, and an additional 10 systems under evaluation. “We’ve had a single reagent kit available for the system, the ANA screen kit, so all of our customers are running that. But we’ve just launched our EBV IgG and IgM panels and have a syphilis product under review at the FDA.”

Now in development at Bio-Rad are four other multiplex panels for autoimmune disease markers and six panels for infectious diseases, with others in development in areas such as toxicology, cardiac, and diabetes.

HealthCare Clinical Laboratories,a department of St. Joseph’s Medical Center in Stockton, Calif., was performing typical standard slide ANA testing in 2004, when laboratory manager Denise Facaros attended an educational seminar and heard about Luminex technology. “I was really interested in what they had to offer, and we decided it was something we wanted to pursue along with our medical director.”

At first, it was a challenge to get the physician clients up to speed, she says, though the laboratory issued a newsletter explaining the change. “Some doctors would call and say, ‘Gee, where’s my pattern and titer?’ but when we explained what we were doing and how to interpret the results, they were satisfied.”

However, the technology didn’t work so well on the first go-around. “We did correlation studies with the selected samples we ran here, and for the most part concordance was fine when looking at positive patients. It turned out when we started testing all comers we began to see a lot of false-positives. The method was revised to a heterogeneous assay, which we thought would improve the situation, but actually things became far worse: There were all sorts of unusual reactions and it made our lives miserable. We couldn’t feel comfortable with our test results, so we retested all the positives by EIA and found a lot of these patients were actually negative.”

That’s when Bio-Rad introduced its instrument, she says. “We explained the difficulty we were experiencing and were initially very skeptical. We spoke on numerous occasions with their technical specialists, who said they had technical enhancements where we wouldn’t see those particular problems. So we gave it a chance.”

Using Bio-Rad’s BioPlex 2200, they ran more than 500 specimens. “We additionally screened numerous ‘false-positives’ that were clinically questionable, that had been previously reflexed by EIA and found to be negative. We did not get any positive reactions with those specimens, so we felt comfortable that a negative was a true negative,” Facaros says.

She is not sure why the early multiplexing experiment was unsuccessful. “We had a feeling, but we couldn’t prove it, that when a patient had some sort of a high immunoglobulin response, for example with hepatitis G virus or multiple myeloma, or some kind of illness, we would see these unusual multiple antibodies present that did not clinically correlate.” However, her laboratory is happy with the BioPlex 2200 since it went live in January 2006.

“We do about 60 to 80 specimens a day, and we’re done within an hour and a half. Everything is completed, resulted, and reported—whereas before we were struggling with a nearly all-day process, setting up the test manually, getting results and having to reflex our positives.” The laboratory is confident in its results now. “The patients who clearly have lupus or other connective tissue diseases have correlated extremely well.”

Downtime has been minimal and related mostly to the newness of the analyzer. “We discovered some salts building up in syringes, and found we needed to do more frequent maintenance, but it was of no impact to our lab.”

To acquire the system, the laboratory had to go through the usual acquisition process with the chief financial office. “The company gave us mostly ‘soft’ savings—helping physicians so patients don’t have to come back, getting results that make clinical sense, not following up false or weak positives.” But she ran a financial analysis based on the product they were using, and even though it is more expensive, Facaros thought it was well worth it. “When we acquired our BioPlex they only had autoimmune testing available, but we knew they are working hard on bringing up new assays through the FDA. EBV was just released, and we’re looking forward to it as well as to the syphilis assay.”

Testing for inherited disease markers is another area in which multiplexing is flourishing. The Canadian province of Ontario, for example, already has an ambitious newborn screening program that tests each infant for 27 genetic mutations. But a new initiative hints at potential applications of multiplex testing in that realm.

Illumina of San Diego announced in January that it has a research collaboration agreement with the Children’s Hospital of Eastern Ontario to screen newborns for spinal muscular atrophy and hemoglobinopathies such as sickle-cell anemia and thalessemia. Later this year the multiplex testing will include a test for cystic fibrosis.

Illumina’s VeraCode technology, which uses digital holographic codes to provide a detection method for its multiplex assays, and its BeadXPress reader are scheduled for market availability by the end of March, according to the company.

Jean Amos Wilson, PhD, CGMB, is scientific director for human genetics at Focus Diagnostics, Cypress, Calif. One of the company’s specialties is providing interpretation of genetic tests through its GenomEx Genetic Laboratory Interpretive Services. “Our clients are laboratories at hospitals and small laboratories that have internalized genetic tests that are simple to do analytically but are difficult to interpret,” she says.

Focus Diagnostics’ first product is cystic fibrosis interpretation, as that assay is in the process of migrating to the “middle market,” says Dr. Amos Wilson, who designed the Tm Bioscience CF70 assay several years ago when she was with Specialty Laboratories. “When a large-volume assay becomes analytically easy to do and volume rises, our clients want to do it in their own laboratory.”

She has done thousands of CF assays on a variety of Luminex instruments and reports that the commercially available multiplex CF platforms are reliable. “But just because some multiplex tests have become analytically simple, that doesn’t mean the laboratory has the wherewithal to interpret results and their impact for a specific patient.”

Focus offers interpretation for several cystic fibrosis multiplex analysis platforms through partnerships with the companies that have produced commercial CF tests. “We have partnerships with a number of these platform vendors, and, as they place their technology with laboratories that want to use our services, we have platform-specific reports for the laboratories,” Dr. Wilson says.

The multiplexing concept has been around for a number of years and has held a lot of promise for clinical diagnostics, Bio-Rad’s Barcellos says, but only recently has it become available in the more routine areas of the laboratory.

Could multiplex testing replace microplate methods? No, says Barcellos. “But we will succeed in replacing testing areas where markers are frequently or always tested in a panel. There’s no mystery why Zeus, Inverness, Inova and Bio-Rad are going after many of the same markets.”

“Any laboratory that has a desire to gain control of laboratory costs in the areas of labor and management of supplies will be able to generate a lot more results with a lot less labor and fewer reagents using a multiplex product,” he says. “Right now, mainly mid- to high-volume laboratories will benefit the most. But within a year, even smaller to mid-size laboratories will be able to justify the purchase of multiplex systems.”