Real-time PCR for the rest of us

June 2006
Feature Story

William Check, PhD

When it comes to molecular testing for infectious diseases, the operative word is “simplicity,” says Washington C. Winn Jr., MD. “Unless you’re a research laboratory, even if you’re a sophisticated academic laboratory, you are in the business of producing results in a timely fashion. So anything that speeds up and simplifies the work is desirable,” says Dr. Winn, who is professor of pathology, University of Vermont College of Medicine, and director of the microbiology laboratory in the Department of Pathology and Laboratory Medicine, Fletcher Allen Health Care, Burlington, Vt.

Companies that make PCR instruments have promised, with the regularity of politicians promising tax cuts, to deliver this simplicity. Tax cuts have come, but laboratorians are still waiting for simplified PCR machines.

“The first big advantage after the introduction of PCR was real-time PCR,” Dr. Winn says, which eliminated post-amplification processing and appreciably shortened the time to results, making same-day clinical decisionmaking possible. However, Dr. Winn notes, substantial complexity remained, with the need for a demanding pre-amplification extraction. And, he adds, “There have been very few FDA-cleared kits, meaning that people have had to put together their own assays.”

In Dr. Winn’s view, the big change in infectious diseases molecular pathology will come with an increase in the number of FDA-approved kits and, ultimately, automated machines that are “distressingly” like chemistry analyzers. “Microbiologists are ambivalent about those,” he notes. But that’s what’s needed for molecular pathology to become widespread in infectious diseases.

After years of wandering in the desert of complexity, Dr. Winn has finally seen the Promised Land: the first truly simple real-time PCR analyzer. “Cepheid’s GeneXpert is the first of those integrated systems that will simplify molecular microbiology,” he says. GeneXpert combines nucleic acid extraction, amplification, and detection in one closed cartridge. And it allows random access. “It does what real-time PCR is especially good for,” Dr. Winn says—assays like enterovirus meningitis, HSV encephalitis, and Bordetella pertussis that are “one-off” and needed rapidly.

“At the beginning at least, I don’t see us doing high-volume assays such as HIV or HCV or bacteria on these platforms,” Dr. Winn says. “I see them as more for special problems, which tend to be lower volume.”

Other molecular microbiologists share Dr. Winn’s vision. “We are moving toward a situation where high-efficiency extraction is linked to high-efficiency real-time PCR,” says Gary W. Procop, MD, section head of clinical microbiology at the Cleveland Clinic. “We are early in that process and there is still a lot of progress that needs to be made.”

Nonetheless, Dr. Procop agrees that GeneXpert heralds the first step into that new territory. “GeneXpert is a unique instrument that will have a niche in rapid molecular diagnostics, particularly for low-volume tests that provide a medically critical result,” he says. Further, he predicts that GeneXpert’s simplicity will make it possible for PCR testing to be done in many settings that don’t do it now. “That was the original intention of real-time PCR instruments, such as Roche’s LightCycler,” Dr. Procop says. “But that didn’t happen, since users still need a decent amount of molecular savvy to run them. That is not true with GeneXpert.” As Cepheid puts together FDA-approved kits, GeneXpert and its progeny could be the way that people not classically trained in molecular diagnostics can employ such testing, in Dr. Procop’s estimation.

Beverly B. Rogers, MD, professor of pathology at the University of Texas Southwestern Medical School in Dallas and chief of pathology at Children’s Medical Center, has been working with PCR since 1987. “I’ve seen it through its conception and adolescence. Real-time PCR made a profound change for the clinical laboratory.” Having amplification and detection take place simultaneously in the same tube greatly reduced the risk of amplicon cross-contamination and made the technology much more amenable to clinical microbiology laboratories. However, Dr. Rogers notes, “You still have extraction, which needs to be done offline; then the sample needs to be transferred for amplification.”

Now Dr. Rogers is seeing the early adulthood of PCR. Her laboratory was one of three sites in the clinical trial of GeneXpert’s assay for enterovirus in the CSF of meningitis patients. Data from this trial are now before the FDA. “What this instrument will do is to bring PCR into the realm of moderate-complexity testing,” she says. “It allows PCR to be done by technologists without special training.” This will make it possible for laboratories in small-to-medium hospitals to offer PCR tests on site, and for large laboratories to offer tests on a 24/7 basis. Dr. Rogers considers this the greatest achievement to date in molecular testing in microbiology. Thinking back to 1987, she says, “I thought, When will the day come when you can do this testing in any clinical laboratory in any community hospital? That day is now.”

GeneXpert is the first real-time PCR system that can be used to provide “real-time service,” in the words of Yi-Wei Tang, MD, PhD, associate professor of medicine and pathology and director of the molecular infectious diseases laboratory, Vanderbilt University Medical Center. “It is the first real-time PCR to put everything together for you,” he says. “All you need to do is add the specimen and some diluents, push a button, and let it go.”

Having simple, rapid PCR available is “critical” in certain contexts, Dr. Tang says. “A lot of PCR tests, although they are very complicated, are very special—patients and physicians want the result rapidly.” As an example, he cites the detection of enterovirus in a young child who comes to the ED with meningitis.

“The parents are scared and need something very quickly to tell them what is going on,” he says. “With GeneXpert you can provide that real-time service.” Specimens can be tested immediately. The GeneXpert PCR results will be available at the same time other clinical and laboratory data arrive “so the physician can make a fully informed medical management decision,” Dr. Tang says.

Another situation in which a rapid and sensitive test will be helpful is HSV encephalitis. “We have a good therapy, acyclovir,” Dr. Tang says. “But it must be given early and by IV. So it is good in CNS infection to have a quick test for HSV.” A patient with suspected HSV infection of the central nervous system is put on IV acyclovir. The test result determines whether acyclovir is continued or discontinued.

Of course, not all tests need to be done quickly, Dr. Tang notes. Quantitation of CMV in transplant patients, for example, can wait. “It is okay to batch them,” he says.

At this time, GeneXpert is “unique,” agrees Karen Kaul, MD, PhD, director of molecular diagnostics, Evanston (Ill.) Northwestern Healthcare and professor of pathology and urology, Northwestern University Feinberg School of Medicine, Chicago. “This is a long-awaited instrument that does nucleic acid preparation and amplification in one platform without human intervention between steps.”

Dr. Kaul sees GeneXpert fitting into hospitals that have a need for rapid turnaround time but do not have a full molecular laboratory, as well as to extend the service capabilities to times when the molecular lab may not be open, such as evenings and weekends. When a blood culture bottle turns positive in the middle of the night in a patient with septicemia, for example, the GeneXpert test could save a few hours in diagnosing methicillin-resistant Staphylococcus aureus, or MRSA, compared with waiting until the molecular lab opens the next morning.

She would tend to use GeneXpert for lower-volume tests that need rapid results, such as enterovirus and HSV. For high-volume tests, it might be less attractive. With only four chambers per block, it would be slow going. “If I were doing 100 or more tests daily for a particular analyte, unless they came in a slow trickle and truly could not wait to be run in a batch, I would batch them and put them in a larger-throughput instrument,” Dr. Kaul says. She notes Cepheid’s plans for a 16-module instrument, which will increase throughput considerably for higher-volume tests.

In her laboratory, Dr. Kaul has designed HSV, CMV, and other assays with one set of temperature parameters, so they can be batched on the LightCycler. Assays for MRSA (targeting sequences for S. aureus and methicillin-resistance) screening also are “pseudo-batched” in her laboratory. But not all hospitals have the volume and expertise to make this feasible.

Some have objected that the GeneXpert extraction doesn’t provide as pure DNA as a biorobot. “I don’t worry too much about purity of DNA,” Dr. Kaul says. She points out that many assays such as Roche’s chlamydia/gonococcus assay do not use pure DNA either and amplification is still fine.

However, Dr. Kaul does note that GeneXpert is “likely to be expensive.” Her hunch is correct: The four-cartridge block lists for about $60,000. Cartridges will cost between $35 and $80 each.

Cepheid’s first FDA submission for GeneXpert is an assay for Group B streptococcus, or GBS, in pregnant women. Jeanne A. Jordan, PhD, HCLD, medical director of the microbiology, virology, immunology, and molecular diagnostics laboratories at Magee Hospital of the University of Pittsburgh Medical Center, tenured associate professor in the Department of Pathology at the University of Pittsburgh, and associate director of Magee Women’s Research Institute, has just finished a clinical trial of the GBS kit. Her data are being used as part of Cepheid’s FDA submission. The trial compared results for GBS on GeneXpert relative to the gold standard of the Centers for Disease Control and Prevention—broth-based enrichment with culture—and to the FDA-cleared IDI Strep B assay, the first real-time assay. “Our conclusion was that GeneXpert compared very favorably to the other assays,” Dr. Jordan says.

Cepheid is requesting moderate-complexity designation for the GBS assay, which would permit it to be run in near-patient settings—stat and emergency department laboratories and labor and delivery suites. In qualifying trials, the FDA required that nurses do the testing on the GeneXpert. (Research technologists did culture and IDI assays.) For training, nurses were required first to watch technologists do the entire process, then to perform proficiency testing on the instrument, consisting of a well-characterized panel of four samples run three times per day for 10 days. “They were proficient after the first go-around,” Dr. Jordan says of the nurses. “The protocol is extremely easy.” A swab is inserted into a hole in the cartridge and broken off at a prescored line, two buffers are completely squeezed out of self-contained plastic squeeze bottles into the cartridge, the lid is closed, and the cartridge is placed inside the instrument. “From there it is a few computer keystrokes, and in about 75 minutes you get the result,” Dr. Jordan says. (Other assays may take longer. Assay for enterovirus, for example, requires a reverse transcriptase step to make cDNA.)

Though GBS on the GeneXpert is optimal for point-of-care testing or near-patient sites, Dr. Jordan says, in her hospital she envisions the instrument in the clinical laboratory because they have pneumatic tube systems from many of their intensive-care sites, including the labor and delivery suites. Also, her laboratory is open 24/7. Dr. Jordan has been doing GBS on all three shifts with the IDI Strep B assay on Cepheid’s SmartCycler, a standard real-time PCR instrument. At the American Society for Microbiology meeting in May she presented a year’s worth of data showing that this approach is feasible with trained general technologists.

“GeneXpert would be much more user-friendly,” she says. “For those low-personnel shifts it would be great. My wish would be to be able to do intrapartum GBS samples on the GeneXpert.”

She says she would need an FDA-cleared assay. And she notes that “capital is always an issue.” Her volume would be right for this system, with five to 10 stat intrapartum GBS samples per day.

However, not everyone sees this application as a gimme. Says Dr. Procop, “I have had this argument with Cepheid about using GeneXpert for Group B strep in pregnant women.” The CDC recommends a GBS screen at 36 weeks of pregnancy. “With a good screening program and high patient compliance, rapid results are not usually necesssary,” Dr. Procop says. “And even patients without a screening result can be treated if they have symptoms, such as if they present with prolonged rupture of membranes or fever.” At the Cleveland Clinic, the incidence of infection is extremely low, he says. “Most of our obstetricians did not screen prior to the current CDC recommendation, and have never screened at labor, but treat based on symptoms with successful results. That has been an effective strategy.” However, Dr. Procop notes this catch-22: “Rapid screening may be preferred in patients who don’t get prenatal care. However, these patients often don’t have money to pay for this test, and, although it may render an important result, is yet another burden to the health care system. It is significantly more expensive than culture, and one must consider its value when contrasted with treating based on symptoms.”

Other sites in the GBS clinical trial used GeneXpert on samples from women in the third trimester, obviously geared toward placing the instrument in the offices of OB/GYNs. Dr. Jordan is not fully comfortable with that. “My hospital does a significant amount of POC testing,” she says. “When those tests are not done under daily supervision of the central laboratory, the non-laboratory people are not as mindful of QA/QC as the laboratory personnel, and we are constantly having to remind them of those responsibilities. So if you are putting something like this into doctors’ offices and they have to have a moderate-complexity license to do this testing, you obviously must have someone who appreciates the QA/QC requirements, things that are near and dear to the heart of any laboratory person but may not be on the radar of a non-laboratory person.” She also sees that prompt troubleshooting will be mandatory in office settings. “That will require very good technical support from the company.”

Thus far, the FDA has not approved any PCR test for a POC setting, Dr. Tang notes. “So Cepheid has a long way to go to get this,” he says. “It will take a change of concept.”

GeneXpert will be only as useful as the number of kits available for it. Typically, molecular laboratories do many homebrew real-time PCR assays. “To do homebrews [on the GeneXpert] would require closely working with the company to design assay parameters and choosing the cartridge,” Dr. Jordan says. In Dr. Rogers’ view, the greatest limitation on GeneXpert is that homebrews will not be easily done on this instrument. “Say I want a new test for a virus. In my current situation I just order the primers and do it. But GeneXpert is not meant for development. Cepheid will partner with you to do development, but you can’t do it without a partnership.” Dr. Rogers is working with Cepheid to develop a Bordetella assay for the GeneXpert.

She will next be involved in trials of a MRSA assay, which starts by looking for the spa gene (an indicator of S. aureus) and the mec gene (which specifies methicillin resistance). If both are negative, or if mec alone is positive, a negative result is reported in about 50 minutes. If only spa is positive, methicillin-sensitive staph is reported in the same time frame. If both are positive, GeneXpert carries out a completely separate MRSA-specific confirmatory reaction on an aliquot of sample extract that is pulled from a holding area.

Cepheid intends to seek regulatory approval on developed products, rather than to offer them as analyte-specific reagents, a strategy suited to its goal of targeting laboratories without molecular expertise.

In line with the other pathologists interviewed for this article, Dr. Jordan doesn’t think GeneXpert will be the optimal choice for higher workloads—tests with 100 samples per day, for example. However, she notes that Cepheid unveiled a 16-cartridge GeneXpert model at the ASM meeting last month. For a 1.5-hour assay, that offers the potential to do 256 samples in a 24-hour day. List price for this instrument, which is also random-access and will accept any kit cleared for the four-slot GeneXpert, is $125,000. Availability is expected in late 2006.

Dr. Jordan’s laboratory runs chlamydia/gonococcus on the Gen-Probe Aptima system. As volume for this test increases, she and her colleagues are thinking about bringing in Gen-Probe’s Tigris, which also features integrated, fully automated operation. (Gen-Probe uses an amplification reaction called TMA, for transcription-mediated amplification, which avoids the patents on PCR.)

Dr. Winn also does chlamydia/gonococcus on a Gen-Probe instrument. With an annual volume approaching 15,000 specimens, he is also, he says, “right on the cusp of the volume that would justify using Tigris. As soon as we reach that volume we will switch.”

What is attractive about the Tigris, Dr. Procop explains, is that “you wand the tube right into the machine and out the other end comes the result. It is truly automated, like a chemistry instrument, probably one of the most automated systems out there. And in all the reviews in the peer-reviewed literature it is very good on accuracy.” His laboratory will probably switch to that method. (They now use two MagnaPure extraction robots coupled to Cobas analyzers, with manual transfer of samples.) Tigris operates in batch mode, Dr. Procop notes, but that is acceptable for this test.

A critical perspective on automated real-time PCR options comes from Daniel H. Farkas, PhD, HCLD, vice president of clinical diagnostics at Chondrogene, Toronto. “When the Holy Grail—black boxes where you don’t need highly specialized, molecularly trained technologists—hits the market, then you will have a democratizing of molecular diagnostics and the ability to decentralize it,” Dr. Farkas says. At this point, in his view, “we are taking baby steps in that direction.”

He calls Tigris “a great system” but “very large and very capital-intensive” and perhaps not right for modest-volume laboratories.

“I love Cepheid’s GeneXpert,” he says. “It is the first real manifestation of the black box—sample in, answer out in a reasonable time. It has the potential to play a large role in POC testing, such as the delivery room for Group B strep or the ED for HSV.” In his opinion, however, “GeneXpert’s low throughput does not make it an appropriate platform for large reference laboratories. And it is not necessarily the ideal platform for medium-volume hospital laboratories, such as The Methodist Hospital in Houston,” where he was director of molecular diagnostics until recently.

Dr. Farkas thinks that Roche’s Cobas Ampliprep fits between the other two systems. But it is a closed system. “I’m a photography enthusiast,” he says. “I have Canon bodies, which stops me from buying Nikon lenses. Why should Roche do any differently?”

One drawback of the Cobas Ampliprep is that it is for research use only and not available in the United States. In addition, says Dr. Tang, based on his own experience, “it’s too large and expensive.” Adds Dr. Rogers, “I guarantee it will never be in a compact cartridge format suitable for POC testing.”

For nonmolecular laboratories, Dr. Farkas suggests an alternative route to obtain PCR capability. Start with a low-throughput robot, such as a Qiagen BioRobot EZ1, which in 20 minutes generates DNA from six specimens and costs about $25,000, or an eight-sample MagnaPure Compact for $35,000. Add a real-time PCR instrument, such as a LightCycler or SmartCycler, for $40,000 to $50,000. “For $75,000 or so you have two small-footprint instruments that don’t necessarily require molecular expertise and have well-written protocols that any trained technologist can follow,” he says.

Not too long ago, Dr. Procop says, nucleic acid extraction had to be done manually. “Now there are many options to get a high-quality DNA or RNA product using automated and semiautomated methods.” Choices include the Roche MagnaPure, BioRobots from Qiagen, and BioMérieux’s EasyMag, among others. Pluses of these machines, Dr. Procop says, is that they save labor, provide nucleic acid that is probably slightly superior to manual kits, and decrease the chance of human error. The chief negative is cost: Large robots cost in the $70,000 to $80,000 range. All current molecular technologies also entail a human cost, says Franklin Cockerill III, MD, professor of microbiology and medicine, director of bacteriology, and chair of microbiology at the Mayo Clinic and Mayo College of Medicine. “One of the biggest challenges we see as more and more PCR is done is ergonomic injury,” Dr. Cockerill says, “particularly for some of our aging technologists in the 40+ age range. Ergonomic injuries can be quite significant.” Dr. Cockerill sees the ever-increasing use of swabs as a major contributor to ergonomic injury, with swabs being used to provide specimens for VRE, MRSA, Bordetella, respiratory panels, Group A and Group B strep, chlamydia/gonococcus, HSV, and other viruses. “I would like to see absolutely no manipulation of the specimen and the specimen container never opened,” Dr. Cockerill says. These are “lofty goals,” he concedes, but he believes that, ultimately, vendors will provide automation for real-time PCR that will eliminate the need to snap swabs.

About the GeneXpert, Dr. Procop raises the “contentious issue” of external controls. This is a problem resulting from GeneXpert’s ability to do one test at a time. “For every laboratory-validated PCR run, we have to have external positive and negative controls,” he says. “And some regulations even say we have to have external high and low positive controls.” Obviously the cost of an individual test would become prohibitive if three control cartridges were needed for every run. “Regulations will have to be changed for GeneXpert or specifically addressed during the FDA submission process,” Dr. Procop says. “Regulatory agencies will have to keep pace with newer technology.” He envisions a situation similar to that with rapid EIA devices for RSV or flu. “In such instances, it has been deemed acceptable to validate each lot. Perhaps such an approach will be taken with the GeneXpert; then we would use an internal amplification control to demonstrate that amplification was not inhibited.” However, Cepheid notes that, for FDA-approved tests, the abbreviated checklist on the microbiology checklist applies, which recommends daily controls.

Dr. Rogers is working with Cepheid on a similar issue related to assays for viruses that are quantitated, such as EBV, CMV, HIV, and HCV. “Typically, we do a standard curve and use a three-point standard on a daily run,” she says. “That approach will not be transferable to GeneXpert. You would need a separate cartridge for each point on the curve. It would be way too costly.” Dr. Rogers would like to design an assay with an internal standard in the same cartridge as the sample.

Even under the best of circumstances, cost—or rather reimbursement—will be an issue, Dr. Winn says. “Medicare in New England just about covers reagents and extraction, but not controls,” he says. “How can you do a test where you can’t pay for the whole package of reagents, not to say labor?”

Still, financial considerations notwithstanding, laboratorians are strongly attracted to functional innovation. “When I’m looking for options, I don’t look for who is the market leader at that moment,” Dr. Winn says. “I look for a company that is moving in the right direction in the most innnovative and nimble fashion.” In his opinion, Gen-Probe and Cepheid are now leading in the dance.

“[Industry leaders] have a lot more hungry competition,” Dr. Kaul agrees. “And that’s good for us laboratorians.”