William Check, PhD
September 2013—Rapid and accurate molecular assays have gradually infiltrated the field of bacterial diagnosis. For several potentially lethal nosocomial pathogens—Clostridium difficile, vancomycin-resistant enterococcus, Mycobacterium tuberculosis, and methicillin-resistant Staphylococcus aureus—FDA-approved molecular assays are making a difference. Not only have they improved the accuracy of diagnosis, benefiting patients and clinicians, they have also been a boon to infection control practitioners.
“For the longest time molecular technology was focused mostly on viruses, because they took so long to grow. Now there has been a major change and molecular assays for nosocomial bacteria are being approved as well,” says
Esther Babady, PhD, D(ABMM), assistant attending microbiologist in the Department of Laboratory Medicine and assistant director of the microbiology service, Memorial Sloan-Kettering Cancer Center.
Rapid molecular assays are also making a difference in terms of the algorithms that infection control practitioners use for bacterial diagnosis, Dr. Babady says. “Take mycobacteria, for example. When a patient comes in to get a specimen for mycobacteria, we first do a smear. If the smear is positive, the patient must go into isolation until we know whether it is TB or not. That can take a long time.” Rapid molecular diagnosis for M. tuberculosis (MTB) can dramatically shorten isolation time.
Phyllis Della-Latta, PhD, D(ABMM), director of the clinical microbiology service at New York-Presbyterian Hospital, Columbia University Medical Center, and professor of clinical pathology at Columbia University College of Physicians and Surgeons, calls real-time nucleic acid amplification tests the new gold standard for detecting many infectious agents, thanks to their simplified technology, rapid turnaround time to results, and enhanced sensitivity and specificity. However, there are caveats to consider. “NAATs may need to be used as an adjunct to conventional culture methods when drug-susceptibility testing or genotyping is required,” she says. In addition, “the NAATs cannot predict microbial viability, limiting the ability to measure infectivity.”
Dr. Babady’s first, eye-opening exposure to the extensive impact of molecular assays on managing nosocomial infections came with the implementation of a NAAT for C. difficile infection. She calls it a game-changer, noting it’s a one-hour test and a big change from having to wait two to three days. “We used to do the cytotoxicity assay, which can take up to 48 hours. Implementing something that took an hour changed the algorithm that Infection Control had for putting patients in isolation.” Since then she and colleagues have been implementing rapid molecular assays. “Every one has had an impact,” she says, “not only on patient care but also on infection control.”
Interested in how other hospitals handled these changes now that molecular assays are becoming easier to implement, Dr. Babady organized a symposium on the impact of rapid molecular diagnostic tools on infection control practices at this year’s annual meeting of the American Society for Microbiology.
[dropcap]R[/dropcap]aymond Widen, PhD, scientific director for esoteric testing/R&D at Tampa (Fla.) General Hospital, spoke on molecular assays for detecting C. difficile infection. “We switched from a rapid enzyme immunoassay antigen test for the bacterial A and B toxin to a molecular assay, which detected 30 percent more positive patients. If we had just used the antigen screen, those patients would have been called negative.” That equates to 30 percent more people being treated properly, he says, and receiving appropriate management and isolation measures.
Before the switch, physicians did not trust negative rapid EIA antigen results, leading to orders for three or more tests. “And they admitted treating based on symptoms even with a negative result,” Dr. Widen says. Physicians are convinced of the accuracy of the molecular result, so they order only one test. “If it is negative, they look for another cause for the patient’s diarrhea.”
Simple culture is not helpful for diagnosis because it does not specify toxigenic strains. Toxigenic culture—growing the organism in culture and then testing for cytotoxin production—is useful but takes 96 or more hours. Some laboratories test for glutamate dehydrogenase followed by testing for toxin. By this algorithm, up to 20 percent of positive cases may be missed.
In the past few years several molecular amplification assays for detecting the tcdB or tcdA genes that code for toxins B or A directly from stool samples have been developed, many now FDA-cleared. Molecular assays for the toxin A and B genes are now considered the gold standard for diagnosing C. diff diarrhea. As one seminal publication concluded, “This review considers the role of NAATs for diagnosing CDI and explores their potential advantages over two-step algorithms, including shorter time to results, while providing comparable, if not superior, accuracy” (Tenover FC, et al. J Mol Diagn. 2011;13: 573–582).
In his talk, Dr. Widen showed a slide from that publication comparing sensitivity, specificity, and cost for several testing approaches. A popular algorithm in which samples negative by EIA and positive by GDH are reflexed to cytotoxin testing missed 23 out of 100 positive patients, whereas direct nucleic acid amplification testing missed only five. “Doing molecular testing on every patient did cost more,” Dr. Widen says—$35 per patient versus $18.32. “But with the algorithm, 23 infected people were not being treated and not put into isolation.” An algorithm reflexing to NAAT was worse than NAAT alone—it had delayed results and more hands-on time and was no more sensitive than molecular only. “Results with a molecular assay are almost as quick as with the rapid antigen test,” Dr. Widen says. Turnaround time for rapid antigen is 30 minutes; a NAAT provides an answer in 1.5 to 2.5 or three hours, depending on the system.
After comparing results using four FDA-cleared molecular assays—three PCR-based and one isothermal amplification assay—with their existing approach, Dr. Widen adopted the molecular technique. “True positive” was defined as the consensus of the molecular assays. “Since molecular tests are more sensitive than our previous standard method, we took the consensus among the four kits for true positive.” He called overall agreement among the four molecular methods “excellent”—greater than 95 percent. “Molecular was the best method for us,” he says.
EIA is no longer offered and education was needed to explain the change. “One challenge was to ensure total understanding that a single PCR result had better predictive value than the toxin EIA. The idea that three tests were needed for the EIA had to be replaced with trust in a single PCR. Fortunately, this change happened after the whole influenza story came out and we learned the antigen test for that virus was fast but not sensitive.” So clinicians had been primed to accept the facts about EIA and PCR.
Additional education was needed to enforce the need for liquid stool. The molecular assay for C. diff is not a screening test. It is approved to detect toxin in patients with C. diff disease. “We wanted to make sure the patients had true disease, not just colonization,” Dr. Widen says. “We wanted to avoid unnecessary treatment of patients without symptoms of C. diff disease, which might occur if nondiarrheal samples were tested since it’s clear that individuals may be asymptomatically colonized.”
Finally, it was necessary to stress that test of cure is not an acceptable use of the C. diff PCR. “It was not meant for that,” Dr. Widen says. “We don’t know how long DNA persists after infection has been cleared.” A patient with a positive molecular test for C. diff infection is put into isolation for the duration of the inpatient stay. Upon discharge, the contents of the room are terminally cleaned with bleach solution; privacy curtains are discarded and routine cleaning is done. Dr. Widen is now evaluating a UV-generating system designed for terminal decontamination of rooms in which C. diff-infected patients have resided.
The authors of a recent publication from Mayo Clinic raised the possibility of using the C. diff assay for a form of screening, which is not now recommended. “They didn’t say everybody should screen for C. diff like for MRSA,” Dr. Widen says. The Mayo investigators concluded that their data “could provide the basis for designing studies of targeted surveillance for C difficile” (Leekha S, et al. Am J Infect Control 2013;41:390–393). Dr. Widen suggests that the rationale would be to stop shedding even in patients who don’t have disease. “Perhaps colonized patients are reservoirs for spread,” he says. “Could we reduce nosocomial incidence of C. diff even lower if we screened people who have some risk factors?” Right now, he says, that idea is “hypothetical.”
[dropcap]A[/dropcap]ctive surveillance with a molecular assay is appropriate for vancomycin-resistant enterococcus (VRE) in some health care settings. The University of Arizona Medical Center decided in 2007 to institute this policy, says Donna M. Wolk, MHA, PhD, D(ABMM), who spoke at the ASM meeting on the impact of rapid detection of VRE. Dr. Wolk is now system director for clinical microbiology, Geisinger Health System.
One analysis found 37 percent mortality for VRE infections, she says, twice as high as for infections with vancomycin-sensitive enterococcus (MMWR 1993;42:597–599). Most VRE infections occur in at-risk populations, such as the elderly in long-term care facilities and in hospitalized patients, particularly immunocompromised patients with granulocytopenia, liver transplant recipients, and those in the ICU.
High-level vancomycin resistance (>64 μg/mL) is conferred mainly by two genetic clusters, vanA and vanB. Hospital outbreaks are more often associated with vanA, which is the genetic target of most PCR assays. Both genetic clusters can be carried by conjugative transposons that can transfer resistance to S. aureus.
Decisions about whether to actively screen for VRE are facility-specific, Dr. Wolk says, citing a CDC recommendation that could be helpful (MMWR 1995; 44[RR12]:1–13).
The aim of active surveillance is to reduce hospital-acquired infections with VRE by increasing VRE detection and speed of detection. Early detection enables early isolation practices that limit exposure. Screening can be performed with culture or PCR; it can involve all hospital admissions or just those considered at high risk, such as intensive care, oncology, and surgery patients. Testing can be done on various specimens: perirectal, perianal, anal or stool. Some specimen types have inhibitors for PCR assays.
When molecular assays with high sensitivity for the vanA gene became available, Dr. Wolk, then at UA, did what she calls an “example intervention.”
Starting in 2007, she collected data on the number of VRE hospital-acquired infections per 1,000 hospital bed days and the number of VRE tests performed. Active surveillance via culture was performed between 2008 and early 2010, followed by PCR (Cepheid GenXpert) in late 2010–2012.
Active surveillance was performed in all patients admitted to ICU and all transplant patients. As the number of tests for VRE rose substantially from 2009 through 2012, the rate of hospital-acquired infections dropped from 0.57 to 0.22 per 1,000 bed days.
During this same period, calculated costs associated with hospital-acquired infections decreased substantially, despite the added cost of active surveillance testing. “If you want the big picture, you can’t just look at the cost of laboratory testing,” Dr. Wolk notes. “You must look at the overall hospital costs associated with additional length of stay, laboratory tests, and diagnostic procedures for those who acquire infections while in the hospital.”
Dr. Wolk is cautious about drawing conclusions from these data. “All we’re showing is an association,” she says. “We can’t prove statistically that the hospital-acquired infection rate dropped because of active surveillance with PCR. We didn’t have the luxury to set up the study that way. We can only say that in our organization there were three associated events—when we stopped doing any kind of surveillance, rates went up; when we re-started active surveillance, rates went down. During that time our hand hygiene didn’t substantially change, so in our setting, the association was enough to keep us testing.”
Some experts think VRE screening is not worth the costs. “From our data, I don’t see that as a viable argument,” Dr. Wolk says. In her view, properly designed studies would be worth doing in at-risk populations, and each organization needs to make its decision based on infection rates, resources, workforce, and bed management needs.
[dropcap]U[/dropcap]se of molecular assays to diagnose Mycobacterium tuberculosis is at the other end of the spectrum from the active surveillance model. The power of NAAT in the TB diagnostic arena is its ability to improve timeliness and accuracy of results and to guide infection control decisions. Dr. Della-Latta calls early MTB detection “the essential link to breaking the chain of TB transmission.”
“Prior to the advent of NAATs, the lab had been totally dependent on insensitive and non-specific AFB [acid-fast bacillus] microscopy and conventional culture with long turnaround time to results. Thanks to NAATs, labs can now fast track TB detection,” she says.
Culture is still considered the gold standard for TB diagnosis, but there are many limitations, among them the prolonged time to isolate MTB from culture (two to six weeks) and the potential for culture-negative TB cases (24 percent in New York City in 2012). This underscores the clinical utility of the ultrasensitive NAATs. The algorithm for incorporating current NAATs is complicated by variations in sensitivity based on the bacillary burden of the specimen; they are less sensitive with AFB smear-negative and extrapulmonary specimens due to their paucibacillary nature.
Early in the 1990s, Dr. Della-Latta and colleagues at NYP/CU were among the first to evaluate and implement the only NAAT for MTB detection at the time—Gen-Probe’s MTD assay. “To date, its use in the routine clinical microbiology lab has been limited because there are many manual steps and it is not a closed system, creating a risk of amplicon contamination,” she says. Recently, Cepheid introduced its Xpert MTB/RIF assay, “which is a real-time PCR assay that is extremely easy to use and offers the added advantage of detecting rifampin resistance.”
Dr. Della-Latta explained the algorithm she uses. All AFB smear-positive specimens are routinely reflexed to NAAT. AFB smear-negative specimens require her consultation with an infectious disease or pulmonary physician to determine the clinical index of suspicion. She also requires two to three respiratory specimens within eight to 24 hours or, if clinically indicated, an induced or bronchoalveolar lavage specimen to optimize accuracy. “The algorithm for smear-negative specimens underscores the urgent need to enhance the sensitivity of molecular assays,” she says.
The impact of the test result on isolation procedures illustrates the value of NAAT for MTB. All patients with a clinical suspicion of having TB are promptly placed on airborne isolation in a negative-pressure room. If the NAAT results are positive, patients remain on airborne isolation either for two weeks of treatment and clinical improvement or until they are discharged on directly observed therapy. With a negative NAAT result, discontinuation of isolation is more conservative. “There must be definitive evidence that the patient does not have TB because he or she has another disease, which responds to therapy, or if they have a history of nontuberculous disease. In making these decisions, the molecular result trumps the AFB smear result,” Dr. Della-Latta explains.
Although TB has been on the decline in the U.S., nontuberculous mycobacterial (NTM) infections are on the rise. “If specimens from patients with NTM pulmonary disease are AFB smear-positive, patients might be inappropriately placed in airborne isolation precautions and treated with a TB drug regimen,” Dr. Della-Latta says. “Therefore, the lab must perform NAATs promptly in order to distinguish MTB from NTM.” In this regard, a new generation of molecular assays is needed to identify NTM disease.
Of interest is the often unrecognized importance of extrapulmonary TB disease and its impact on patient management and infection control. In New York City, the proportion of TB cases with pulmonary disease only was 64 percent, compared with 23 percent for extrapulmonary disease only and 13 percent for cases with both, Dr. Della-Latta says. “Extrapulmonary TB infections with draining abscesses, lymph nodes, soft tissue or visceral lesions can be potential sources of airborne infection. Procedures that cause aerosolization, such as irrigation of abscesses, can also cause TB transmission.” Therefore, early detection of extrapulmonary TB by NAATs has clinical utility.
A recent publication concluded that NAAT has “good potential for the diagnosis of extrapulmonary TB and that its ease of use makes it applicable for countries where TB is endemic” (Vadwai V, et al. J Clin Microbiol 2011;49: 2540–2545). “We have experienced a high success rate in detecting extrapulmonary disease, such as lymphadenitis, when using Gen-Probe’s MTD assay,” Dr. Della-Latta says. However, the effect of NAAT results on MTB decisionmaking in extrapulmonary TB remains problematic as indicated in her collaborative paper (Weiner, et al. Chest 2005;128:102–107).
[dropcap]D[/dropcap]r. Babady, convener of the ASM session, was the only speaker who discussed viral pathogens, in a talk on the impact of rapid molecular respiratory panels on infection control practice. Implementing a rapid respiratory virus molecular panel greatly changed infection control practices at Memorial Sloan-Kettering Cancer Center.
Before 2011 Dr. Babady’s methods for detecting influenza viruses were rapid antigen tests, direct fluorescent antibody, and viral culture. DFA can detect eight viruses—influenza A and B, parainfluenza 1–3, adenovirus, human metapneumovirus, and respiratory syncytial virus (RSV); the test takes two hours to perform. It could take three to 14 days for a final diagnosis, starting with DFA and proceeding to culture, which was considered the gold standard.
Isolation, on the other hand, was based on symptoms only. To discontinue precautions, the patient needed to be DFA negative and culture negative at three days (except for RSV).
The value of PCR for rapid diagnosis of influenza virus, in particular, became apparent during the 2009 swine flu pandemic. During her talk Dr. Babady showed performance characteristics of 10 assays, almost all real-time PCR. One group is rapid (about one hour), simple, and random access; the others are high-throughput batch instruments requiring prior nucleic acid extraction and taking four to eight hours.
Dr. Babady and her colleagues performed a validation study of one multiplex assay from each group: BioFire Diagnostics FilmArray RVP and Luminex xTAG RVP Fast (Babady NE, et al. J Clin Microbiol 2012;50:2282–2288). Dr. Babady chose the FilmArray because of its significantly higher sensitivity. “The sensitivity, simplicity, and random-access platform make FilmArray RVP an excellent choice for laboratory on-demand service with low to medium volume,” she and her colleagues concluded. “In 2011, when we started to implement molecular assay technology, our technologists were not as experienced. We thought it would be easier to start with simple assays,” she told CAP TODAY.
One challenge with the new assay, she says, is that there are no standardized guidelines for infection control policy based on PCR results. Another is the question Dr. Della-Latta raised: Does the nucleic acid detected represent live or dead virus? A related question: How long does virus shed? One study showed higher rates of influenza viral shedding with PCR compared with culture as long as seven days after symptom onset (Leekha S, et al. Infect Control Hosp Epidemiol 2007;28:1071–1076). “That’s one of the challenges,” Dr. Babady says. “No one knows for sure what shedding means.”
Currently, a positive PCR result is the basis for implementing isolation precautions at Memorial Sloan-Kettering. Discontinuing precautions requires test of cure—a negative PCR after seven days. (The exceptions are rhinovirus and coronavirus, for which precautions are discontinued after four days.) Still, Dr. Babady says, “Infection Control has to make decisions based on the clinical know-how of physicians rather than knowledge of the sensitivity of all viruses on the platform.”
Because of the changes in infection control practice associated with the adoption of the molecular assay, the number of isolation days remained the same after the molecular assay was adopted, even though three times as many patients were swabbed and three times as many viruses were identified.
Patients who arrive at the urgent care center with respiratory symptoms are now handled differently. “We used to have to swab the patient and put them into isolation until we got a clear result,” Dr. Babady says. “Now the patient stays in the urgent care center for a couple of hours while we do the PCR. So this change affects even whether these patients are admitted.”
[dropcap]W[/dropcap]ith Medicare about to add hospital-acquired infections to the list of hospital-acquired conditions for which it will not pay, hospitals have an additional incentive to lower the rate of nosocomial infection. Molecular assays increase the accuracy and speed with which nosocomial infections can be detected. Working from the premise that early detection allows early treatment and isolation, and that early therapeutic and preventive measures can lower the risk of nosocomial transmission, microbiology directors have a powerful argument to present to hospital administrators for the broad adoption of molecular testing.
William Check is a writer in Ft. Lauderdale, Fla. As part of this session, Christine C. Ginocchio, PhD, MT, senior medical director and chief of the Division of Infectious Disease Diagnostics at North Shore-LIJ Laboratories, New York, gave the ASM Sepsis Award Lecture on “Rapid Detection and Identification of Bloodstream Pathogens: Clinical and Infection Control Implications.” Dr. Ginocchio addressed existing and future molecular methods for coagulase-negative staphylococci; S. aureus (including MRSA); Enterococcus, Candida, and Klebsiella species; E. coli; and P. aeruginosa. We report on her presentation in an upcoming issue.