February 2009
Feature Story

Anne Ford

How fitting that the shorthand name of one of the wiliest health care-associated infections—C. difficile—sounds so much like “c’est difficile,” French for “it is difficult.” With an increasing number of patients developing Clostridium difficile-related diarrhea, not to mention a hypervirulent strain on the loose, no one’s exactly murmuring a pleased “Ooh la la.” More like “Sacre bleu!”

“It looks like we haven’t reached the peak of the U.S. epidemic as yet,” says C. difficile epidemiology expert Dale Gerding, MD, who types C. diff isolates sent to him by hospitals around the country. Dr. Gerding is associate chief of staff for research and development at Hines Veterans Affairs Hospital, Hines, Ill., and professor of medicine at Loyola University School of Medicine, Maywood, Ill. “We’re still battling the same problem.”

What problem is that? Take your pick: The C.-difficile-associated, severe, often recurring diarrhea in patients. The potential for toxic megacolon. The ability of spores to survive alcohol-based hand hygiene gels. At least one strain’s resistance to fluoroquinolone antibiotics. The insensitivity of some enzyme immunoassay tests commonly used in C. difficile testing. Not to mention a recent national prevalence study by the Association for Professionals in Infection Control and Epidemiology that found that of 110,550 inpatients in 648 participating hospitals, 13 out of every 1,000 were either infected or colonized with C. difficile—a number APIC calls “6.5 to 20 times greater than previous incidence estimates.” Oh, and Dr. Gerding’s statement that U.S. infection rates are increasing by roughly 20 percent each year. “It just keeps on going at almost a linear rate,” he says. Sort of like the Energizer Bunny, except much, much less adorable.

In other words, the picture isn’t much brighter than the last time CAP TODAY reported on the C. difficile epidemic (“C. difficile back with a ven­geance,” May 2006). However, the news isn’t all bad. Several manufacturers are developing or introducing new, more sensitive tests to determine if a patient is infected with the bacteria. Potential new treatments are on the horizon. And with luck, the U.S. Department of Health and Human Services’ “HHS Action Plan to Prevent Healthcare-Associated Infections” (released last month), which calls for a 30 percent reduction in C. difficile case rate per patient days and administrative/discharge data for ICD-9-CM-coded C. difficile infections in the next five years, will spur more hospitals to action. Still, how much hope is there for stemming the C. diff tide?

Certainly the continuing prevalence of the hypervirulent NAP1 strain isn’t soothing any worries. To refresh your memory: “This is a specific strain of C. difficile that emerged first in North America, in Pennsylvania. This NAP1 strain has a genetic change that results in literally 16 to 23 times more toxin production in vitro,” explains William Jarvis, MD, principal investigator of the APIC study and president and co-founder of Jason and Jarvis Associates, a private consulting firm in health care epidemiology with offices in South Carolina, Oregon, and San Francisco.

“There’s no question that this strain is more virulent than other strains. It seems to be playing a big role” in the epidemic, says L. Clifford McDonald, MD, chief of the Prevention and Response Branch of the Centers for Disease Control and Prevention’s Division of Healthcare Quality Promotion. “We’re continuing to see it as the most common single strain in outbreaks in severe cases. That’s been better documented since 2006, with some studies showing that in humans who have more severe disease that that strain is responsible. But the exact reason for the hypervirulence has never been really determined.”

One ongoing question with regard to the NAP1 strain is how much significance to assign the fact that one of its characteristics is a deletion in the tcdC-negative regulator gene of toxin expression. One theory holds that this deletion is responsible for the strain’s increased virulence, and that that increased virulence causes more severe disease in humans. But, Dr. McDonald says: “It’s clear that it’s not just that deletion. Probably more important is an upstream point mutation. The point mutation does lead to early truncation of the protein and probable disregulation of toxin production.”

Unfortunately, the APIC survey was not able to determine the precise prevalence of the NAP1 strain. “The fact that [only] two percent [of the hospitals surveyed] are doing cultures means there’s no isolate” to send for testing, Dr. Jarvis points out. “So, not surprisingly, we had no facility report a NAP1 strain, even though we know from Dr. Gerding’s data that it has been seen in a very large number of U.S. states.”

“Just like the NAP1 strain emerged, there may be another C. diff strain that emerges today, tomorrow, next year, 10 years from now, that spreads across the United States, the world,” he continues ominously, “and we’re going to be delayed in even recognizing that such a strain exists because no one has the capability of doing culture and identifying what it really is.”

Indeed, Dr. Gerding is already seeing new strains arise. “We’re still continuing to see this epidemic [NAP1] strain as the dominant strain, but we’re also starting to see other types come up now,” he says. “This is what we’ve been expecting.” Thus far most of those new strains appear to be toxinotype Os, that is, “they don’t have any major changes in their pathogenicity locus,” he says. “But none of them has undergone enough analysis to say much more than that.”

Another question: whether C. difficile will become resistant to the antibiotics used to treat it—vancomycin and metronidazole—the way that the NAP1 strain has already become resistant to fluoroquinolones. At the moment, “there really isn’t any resistance to speak of,” Dr. Gerding says. That said, he adds, recent studies have shown that metronidazole is significantly less effective than vancomycin in treating patients with severe C. diff infection. “But it has not been correlated with any resistance by the organism in vitro to the drug. It’s something else. It may have to do with the fact that we don’t have as much metronidazole staying in the gut as we do with vancomycin.”

By some accounts, the number of patients acquiring C. difficile infections who do not have the typical risk factor—exposure to antibiotics—is growing. Ciarán Kelly, MD, associate professor of medicine at Harvard Medical School, chief of the Herrman L. Blumgart Internal Medicine Firm, and director of gastroenterology training at Beth Israel Deaconess Medical Center, Boston, goes so far as to say that previously C. difficile acquisition without prior anti­biotic exposure was considered a “freak occurrence.”

Dr. Gerding says, “I’m not too concerned. In the hospital, there seem to be almost no patients who haven’t had antibiotic exposure who get C. diff. But in the community, we’re finding 30, 40 percent” who do not report having been exposed to antibiotics. “It may be that they don’t have any, or it may be that we just don’t know how to look for it, or how to interview the patients. There are more and more of those kinds of patients being reported. Whether this is a significant part of the C. diff problem or not, I’m not certain. I still think the vast majority of cases are in the hospital, and they’re associated with antibiotics.”

Another still unknown aspect of C. difficile is its incubation period. Previous studies have suggested that period is seven days or less, but, Dr. McDonald points out, “we certainly do see a lot of C. difficile right after people are being discharged from the hospital. Whether that’s because they’re going out on antibiotics and getting it in the community, or actually they’re acquiring it in the hospital and developing disease in the community, we don’t know, but that makes us wonder whether the incubation period couldn’t be a little bit more prolonged.”

In addition, some studies have revealed the presence of C. difficile in retail meats. “It’s been known since around 2000 that it was starting to show up in food-producing animals more—actually causing disease in neonatal pigs and being often found to colonize cattle,” Dr. McDonald says. “It’s now been noticed that some of the strains, predominantly one group of strains that were very prevalent in food-producing animals in the early part of this decade, are increasingly starting to be seen in humans. But at least as of this date, there is no conclusive evidence that you get C. difficile from food.”

Then, too, the role of proton pump inhibitors, if any, in the spread of C. diff infection has yet to be nailed down. Karen Carroll, MD, professor of pathology at Johns Hopkins University School of Medicine and director of the Division of Medical Microbiology at Johns Hopkins Hospital, says the topic “deserves a lot more study” and that “there’s some evolving literature to show that perhaps those are overprescribed and could be contributing to one’s risk, particularly in the community among healthy populations.”

But Dr. Gerding says the correlation, if any, is far from clear: “You’ll find about as many studies that indicate a risk for proton pump inhibitors of getting C. diff as you will that find no risk.” And if PPIs do increase one’s risk of acquiring C. diff infection, it’s not clear how. By neutralizing stomach acid, they might make it easier for the bacteria to survive in the body—“but the spores are pretty resistant to acid anyway,” he says. “The other possibility is that these drugs [the PPIs] have activity against some bacteria. So they could be acting like an antibiotic to upset the normal bacteria flora, and in the process of doing that, could make the patients susceptible to C. diff infection just like an antibiotic does.”

Unfortunately, one C. diff treatment that had looked especially promising—a toxin-­binding agent from Genzyme Corp. called tolevamer—has been taken off the table. “A lot of us in the field had significant hopes” for it, Dr. Kelly confesses. “Many of us believe that antibiotic treatment is intrinsically not the best way to approach the disease, and predisposes the patient to the most common complications, which is recurrence. A phase II study indicated that tolevamer was as effective as antibiotics and associated with lower rates of recurrence. We thought we were going to get the first new treatment for C. diff in more than two decades.” The idea was that tolevamer would bind the toxins released by C. difficile and remove them from the body. Instead, it ended up failing two large phase III studies. “It was very disappointing,” he says.

On the bright side, another investigational treatment, a narrow-spectrum antibiotic from Optimer Pharmaceuticals called fidaxomicin (also called OPT-80), remains promising. A recent phase III clinical trial showed that while its response rate was similar to that of vancomycin, fidaxomicin had significantly lower recurrence rates (15 percent compared with 25 percent for vancomycin). “The theory behind this antibiotic was that it would affect the healthy normal bacteria flora to a lesser extent,” Dr. Kelly says. “From what we know about that study so far, it seems as though that may be the case, but there isn’t a full peer-reviewed publication” yet.

In addition, Massachusetts Biologic Laboratory and Medarex Inc. are collaborating on two injectable human monoclonal antibodies called CDA-1 and CDB-1, which, it’s hoped, will reduce relapse rates when given in conjunction with vancomycin or metronidazole. “That is a totally new approach, based on the observation that many people who become infected with C. diff don’t develop the disease, but are symptomless carriers,” says Dr. Kelly. “The treatment is based on the idea that a faulty antibody response makes individuals susceptible.” A phase II study just yielded the report that the recurrence rates in study participants who received the active antibodies were 70 percent lower than those who received a placebo. So, Dr. Gerding says, between fidaxomicin and the new monoclonal antibodies, “it looks like we have at least two drugs or approaches that will result in lower relapse rates.”

Meanwhile, a more esoteric treatment, fecal transplant, has yet to be widely embraced. The idea is that by administering a filtered stool sample (usually donated by a family member) to a patient with multiple recurrences of C. diff, normal gut flora can be reestablished. The “eww” factor aside, “all the case series that have been published at least have suggested almost 100 percent success rates,” Dr. McDonald says. “That really suggests that it works. But at the same time, they have not taken off, in part because a lot of people don’t feel comfortable doing them. There’s some sense of liability, the sense that it’s not standardized, you don’t really know what you’re giving them.” Besides, Dr. Gerding says: “I think it’s too desperate an approach. If we’re going to do something like this, we need to figure out how to construct a synthetic fecal transplant where we can control the constituents and make sure we don’t have anything in it that doesn’t belong there.”

On the prevention front, experts point to antibiotic stewardship programs in hospitals as a key way to control the spread of C. difficile. Those programs might include such measures as automatic stop orders, so that, for example, a patient who receives antibiotic prophylaxis for a surgical procedure will receive only the appropriate number of doses. They might also include pharmacy formularies that restrict which antibiotics physicians can access, or online physician ordering systems that provide educational prompts to physicians ordering antibiotics. Just less than half (47 percent) of the institutions surveyed in the APIC study reported having implemented an antibiotic stewardship program, and “they were all doing a variety of different things” within their programs, Dr. Jarvis says. “Given that antibiotic controls and antibiotic stewardship are such a critical element of the comprehensive program for preventing C. difficile-associated infection, we feel that more emphasis needs to be placed on that, and perhaps even having a person full or part time dedicated to that effort.”

With so many unknowns about the nature and future of the C. difficile epidemic, what can laboratories do in the meantime? Says Dr. Gerding bluntly: “What they can do is better testing.”

If enzyme immunoassay testing for C. difficile were a schoolchild, it would be the awkward kid with too few friends, rarely the first pick for the dream team—in this case of testing solutions. So why does just about everyone let it hang around? For one, it’s much faster than the much more reliable cytotoxin assays. It’s convenient and cost-effective. And though the sensitivities of some of the EIAs are reported to be only about 73 to 75 percent, there are those, like Meridian’s Premier and Immuno­card, for which the sensitivity has been found to be higher (Planche T, et al. Diagnosis of Clostridium difficile infection by toxin detection kits: a systematic review. Lancet Infect Dis. 2008;8:777–784).

Still, sensitivity is a well-known problem, and the common physician practice of ordering an EIA multiple times on the same patient in an attempt to compensate for the problem can clog laboratory workflows and lead to unnecessary and costly patient isolation. However, “the biggest clinical consequence of the low sensitivity is the high level of falsely negative results, which of course can lead to delay in treatment,” Beth Israel’s Dr. Kelly says. “It’s a situation that none of us are content with, ordering a test and then ignoring the result because the test isn’t sensitive enough.”

“We’ve all gotten into the habit of testing two or three or four stool samples for C. diff until we get a positive,” agrees Lance R. Peterson, MD, associate epidemiologist and director of microbiology and infectious diseases research in the Department of Pathology and Laboratory Medicine of north suburban Chicago’s NorthShore University Health­System, and professor of pathology and medicine at Northwestern University’s Feinberg School of Medicine, Chicago. And that’s not good. “When you’re using indirect assays such as enzyme immunoassays, there’s been no validation that repeating a test in a person on the same kind of sample within a few days that the test is even reliable to repeat. Even with the second test, your positive predictive value is less than 50 percent. So if you have a negative first test, and then a day or two later you do another stool sample and you have a positive, you have a 50–50 chance of having a true positive or a false positive.” Not only that, he adds, “if you have a 70-percent-sensitivity test and a 10 percent positive prevalence of disease in the samples you’re testing, it takes you five tests in order to detect all 100, or 10 percent of the 1,000 patients you’re sampling and by that time you have more false-positives than true positives.”

In the APIC study, the C. difficile infection of nearly 95 percent of the patients reported was detected via EIA. “They [laboratories] are taking the easy way out, which is to buy a test off the shelf with poor sensitivity, but one that they can turn around quickly and without requiring a huge amount of technical skill,” Dr. Gerding says. “If they really wanted to get on top of this problem, they would back up their current enzyme immunoassay testing with something like culture and actually try to get the organism out of the stool. But most labs don’t have anaerobic culturing capability anymore, so that isn’t very practical. And they don’t want to do cell cytotoxin assays because that requires maintaining a cell line and also requires a fairly high skill level by the people who are looking at the cells to interpret the tests. It’s laborintensive, and, of course, they’ve had their budgets cut and they’re really struggling to be able to provide any kind of testing with the budgets they have.”

Taking EIA off the table, however, would leave nearly all laboratories without a fast, affordable way to detect C. difficile. That’s where some new molecular tests come in. Several manufacturers, including Cepheid, Prodesse, and BD, are developing or introducing real-time PCR tests for the detection of C. diff, while others are developing molecular tests that use other means.

“We think that molecular testing will provide the right technology at the right time for meeting the demands of the C. diff diagnosis,” says Cepheid’s chief medical and technology officer, David Persing, MD, PhD. Cepheid’s Xpert PCR test for C. difficile, which was released in Europe in November, detects the toxigenic form of the C. diff genome and specifically identifies the NAP1 epidemic strain in less than an hour. Run on the company’s GeneXpert system, it detects three targets: toxin B (“the most reliable marker of toxigenic C. diff,” Dr. Persing says); binary toxin, which has been implicated in some of the more virulent strains; and a single base deletion causing a frameshift mutation in the tcdC gene. “When you detect all three of those targets, what we found in our clinical trials is that it is an excellent indicator of the presence of the NAP1 epidemic strain,” says Dr. Persing. Specific detection of the NAP1 epidemic strain may be of benefit in tracking hospital outbreaks in real time, he says, thus aiding infection-control personnel in what he calls their “search-and-destroy mission.”

He’s particularly proud of the fact that the Xpert test, which generates results in about 45 minutes, doesn’t require batching and thus can be run on a stat basis. “All the other molecular tests out there run specimens in batches,” he says. “In fact, most of the laboratory-based enzyme immunoassays are run in batches as well. That limits their turnaround time.” In addition, he says, the Xpert’s sensitivity and specificity are about 96 percent compared with toxigenic culture. The company hopes to receive FDA approval for the test in the second half of this year.

Meanwhile, BD received FDA clearance in December for its real-time PCR test, the BD GeneOhm Cdiff assay, which targets the toxin B gene, detects the NAP1 strain, and produces results in about two hours. Medical affairs director Tobi Karchmer, MD, MS, says it is the first molecular C. difficile assay cleared in the United States. In clinical trials, she says, the GeneOhm test was compared with cell culture cytotoxicity assay. “We had a set of data that came from fresh stool specimens, and in that case sensitivity was 93.8 percent compared to the cytotoxicity reference method, and the specificity was 95.5 percent. We had a second set of data from frozen stool specimens, and in that case the sensitivity was 100 percent and the specificity was 97.7 percent.”

One of BD’s clinical trial sites was Wishard Memorial Hospital, Indianapolis, which is affiliated with the Indiana University School of Medicine. “There hardly is a comparison” between the BD PCR test and the hospital’s existing C. diff detection method, says Deanna Fuller, clinical research manager for the Department of Pathology and Laboratory Medicine. The clinical trial laboratory used a three-tier testing algorithm that called for common antigen and EIA; cytotoxin assays were then run on any positive results. In comparison to that method, the BD test achieved 95.8 percent sensitivity and 98.7 percent specificity, after discord­ant results were resolved, Fuller says.

Johns Hopkins Hospital also served as a clinical trial site for the GeneOhm assay, allowing Dr. Carroll to familiarize herself with it. Her laboratory has been using a two-step algorithm for detecting C. diff—screening with an EIA antigen test for glutamate dehydrogenase, then testing antigen-positive samples with cell culture cytotoxin—but she was so impressed with the BD test that she plans to switch to it in the next six months or so. “While I don’t think it is as sensitive as toxigenic culture, I think it’s definitely better than cyto­toxin testing and allows for same-day testing,” she says. “That said, PCR is expensive, and given the volume of testing that we do, what I’m thinking I might do is continue to screen with the antigen test and then confirm the positive antigens by PCR.”

Thomas Davis, MD, professor of pathology at Indiana University and director of Wishard’s clinical microbiology laboratory, also plans to institute the Gene­Ohm assay in his lab, but is similarly concerned about cost. “I don’t think there’s any way that this technology is going to be cheap­er in the laboratory,” he says. “It’s probably going to cost twice or three times as much to do this technology as the EIA technology, but that’s not where you save the money. You save the money with patient management.”

Speaking of money, one potential barrier to going molecular for many laboratories is the need to purchase a thermal cycler, plus the costs associated with training workers to perform real-time PCR assays. That’s why Meridian Bioscience is touting its forthcoming Illumigene C. difficile test—which, at CAP TODAY press time, was scheduled to begin clinical trials this month—as “molecular simplified.” The assay uses loop-mediated isothermal amplification technology, or LAMP, rather than polymerase chain reaction technology, Rich Connors, associate director of molecular marketing, explains. “It does with chemistry what PCR does with temperature. Is it any better or any worse [than PCR]? No, it’s just different. It’s a little bit quicker, it’s a little bit simpler, and it doesn’t require large amounts of capital equipment, because you don’t need a thermocycler.” He expects it to be the first FDA-approved assay based on LAMP technology.

And those much-maligned enzyme immunoassays? Connors points to the December 2008 Lancet Infectious Diseases article that reported median sensitivities for Meridian’s Premier and Immunocard assays for toxins A and B of 95 percent and 90 percent, respectively, based on a systematic review of the literature. Manufacturers’ instructions must be followed, he says. “EIAs are designed to evaluate samples from clearly symptomatic patients, and in these cases will give PPVs [positive predictive values] that are quite high. It’s also important to ensure that samples are collected and transported correctly.”

A new EIA is awaiting FDA clearance: Inverness Medical’s Techlab C. Diff Quik Chek Complete, which is a combined membrane EIA for glutamate dehydrogenase and the A and B toxins in a single assay. It can be completed in 30 minutes and does not require additional equipment or batching of specimens. “The beauty of the new GDH assay is that it detects most all of the negatives. It had a sensitivity of 91.7 percent compared to bacterial culture,” says clinical trials participant Wallace Greene, PhD, director of the diagnostic virology laboratory and associate professor of pathology, Penn State College of Medicine, Hershey. When compared with cytotoxic cell culture, all toxin-positive strains were detected by the GDH assay. However, the toxin EIA was only 95.2 percent sensitive. Specimens that are positive for GDH but negative for the toxins should be tested by more sensitive assays—either bacterial culture, cytotoxin, or molecular assay, if available, Dr. Greene says. “This allows a laboratory to report the negative specimens and most of the positive specimens in less than an hour.” Only a very few specimens would require further testing with the more expensive and labor-­intensive molecular assays, he says.

With the C. Diff Quik Chek Complete, “instead of taking two or three days to report a negative result using the cytotoxin assay, thus allowing a patient to be removed from contact isolation, we can do it in less than an hour. In most cases, patients won’t even be placed in isolation. That’s a huge cost savings for the hospital.” Still, Dr. Greene wonders, can labs get away with a simple, rapid, and relatively cheap GDH assay to reliably detect negative specimens and report results essentially as they come into the lab, “or must we use a more expensive, time-consuming molecular assay that would only be run once or twice a day for all specimens to significantly improve sensitivity?

“Until we do a head-to-head comparison with our patient population, we just can’t answer that question,” he says.