Guest commentary: Diagnostic and therapeutic approaches to C. difficile infection

In this guest commentary, Infectious Disease News Editorial Board member Ellie J.C. Goldstein, MD, clinical professor of medicine at the University of California, Los Angeles School of Medicine and director of the R.M. Alden Research Laboratory in Santa Monica, California, discusses the latest research on the diagnosis and treatment of Clostridium difficile infection, as well as transmission of the disease in the health care setting.

Clostridium difficile infection, or CDI, is a major concern in health care institutions and in the community setting, and has vast economic consequences as well. CDI affects about 1% of all hospitalized adults in the United States, and is linked to increased length of stay, morbidity and mortality. The excess cost attributed to CDI is approximately $1.2 to $6 billion annually. Hospital-acquired CDI is now more common than MRSA and is set to become a quality care measure that will affect Medicare reimbursement in 2017.

Disruption of the gut microbiota through exposure to systemic antibiotics can lay the groundwork for C. difficile spores to cause disease. Despite evidence-based interventions to prevent CDI, rates have not declined.

One factor for the persistent rate of CDI may be a result of overdiagnosis due to the use of the overly sensitive polymerase chain reaction (PCR) testing. Other factors include antibiotic overuse and the use of proton pump inhibitors. The diagnosis of CDI is somewhat challenging and controversial, and is further complicated by host immunity (inability of the host to make immunoglobulin G antibodies against the toxins), long-term C. difficile colonization, and the failure to recognize non-CDI diarrhea. According to the current (published 2010) but soon-to-be-updated clinical practice guidelines from the Infectious Diseases Society of America (IDSA), a diagnosis of CDI requires both a positive laboratory test and clinical symptoms of CDI. As there are many causes of diarrhea, a positive laboratory test alone might include asymptomatic patients with C. difficile colonization, which does not require treatment. Studies have suggested that the presence of detectable toxin — albeit by somewhat less sensitive enzyme immunoassay (EIA) toxin testing — occurs in patients who are likely to have CDI complications.

Standard-of-care antibiotic treatment — metronidazole, vancomycin and fidaxomicin — is often successful in promoting resolution of primary CDI during therapy but the risk of relapse remains unacceptably high, between 15% to 30%. A study by Johnson and colleagues has noted that metronidazole is inferior to vancomycin therapy in even mild and moderate CDI.

Fidaxomicin is a macrocyclic, bactericidal, more narrow-spectrum antibiotic, and thereby potentially has less effect on the indigenous fecal microbiota compared with vancomycin. After successful standard-of-care treatment, about one-quarter of patients will experience CDI recurrence after completing standard-of-care antibiotic treatment. After a first recurrence, the probability of a second recurrence is about 45%, and this risk percentage increases with subsequent recurrences.

The epidemiology of CDI has changed since metronidazole was considered a treatment option, with evidence that binary toxin-positive strains such as BI/NAP1/027 are epidemic, hypervirulent and less amenable to treatment with metronidazole. The 027 strain is associated with more toxin production, a higher recurrence rate, and excess mortality compared with other strains. Toxin production is the mediator of CDI, as toxins induce inflammation and epithelial injury.

Zinplava (bezlotoxumab, Merck) is a monoclonal antibody directed against C. difficile toxin B, and has been shown in two large, international, double blind, placebo-controlled, randomized clinical trials — MODIFY I and MODIFY II — to reduce CDI recurrence in patients receiving standard-of-care antibiotic treatment for CDI.

Finally, we are still gaining knowledge about the use of fecal microbiota transplant (FMT) to reduce the risk for recurrent CDI, the timing of FMT, the risks involved, and identification of suitable donors.

IDWeek 2016 featured several educational symposia, oral abstract sessions, and poster presentations related to current and potentially new approaches to diagnose and treat CDI. The following are highlights from some key presentations.

Diagnostic challenges

Christopher R. Polage, MD, at the University of California, Davis, reviewed the current approach to diagnosis of CDI. He expressed that a second phase of a CDI epidemic has emerged since the introduction of PCR–based assays and nucleic acid amplification tests, after which both toxin-positive and toxin-negative strains of C. difficile were reported as causing CDI.

“I have also become increasingly convinced that most toxin-negative/PCR–positive patients don’t need treatment and probably don’t have CDI,” he said during his presentation.

PCR tests are sensitive but nonspecific; a positive result does not imply active toxin production but rather that an individual is colonized with a strain of C. difficile that is capable of producing toxins.

Although toxin tests are less sensitive than molecular tests, the clinical utility of toxin tests is supported in the literature, Polage said. For instance, drugs used to treat CDI have better efficacy in toxin-positive vs. toxin-negative patients in clinical trials. His own study revealed that toxin-positive patients had a longer duration of diarrhea than toxin-negative/PCR–positive patients. More importantly, toxin-negative/PCR–positive patients had shorter duration of symptoms, few or no CDI–related complications, and similar outcomes to toxin-negative/PCR–negative patients with little or no treatment in most cases. “Toxin-negative complicated CDI is quite rare,” he said.

The confusion over tests for CDI stems from a lack of a reference test for clinical disease and the higher prevalence of non-CDI diarrhea and C. difficile colonization compared with true CDI. Although toxin tests have sometimes been criticized as being insensitive, this insensitivity has been exaggerated by the assumption that all toxin-negative/PCR–positive patients have CDI.

Overdiagnosis and unnecessary CDI treatment has adverse consequences. Treatment to eradicate colonization is ineffective because the gut microbiota is disrupted by treatment. Unnecessary treatment may induce symptomatic disease in susceptible patients, possibly increase spore production/shedding, leading to a “super shedder” state that could increase transmission, and may prolong the carrier state.

Until new IDSA/Society for Healthcare Epidemiology of America diagnostic guidelines are finalized, the best testing approach is a two-step algorithm that starts with a sensitive screen with a high negative predictive value test, such as glutamate dehydrogenase (GDH) EIA or PCR, Polage believes. If either GDH EIA or PCR is negative, report it as negative with no additional testing. It is rare to miss a case with these tests, he said. CDI should be confirmed in GDH/PCR–positive samples using toxin EIA. If toxin EIA is positive, CDI is likely. If toxin EIA is negative, clinical evaluation is needed to determine CDI vs. colonization. Cytotoxic assay is used by some centers to further test PCR–positive but EIA–negative specimens to determine if CDI is present.

FMT update

Elizabeth L. Hohmann, MD, from Harvard Medical School, updated the current thinking on the use of FMT. The consensus is that FMT should be reserved for CDI patients who have had at least three episodes or two hospitalizations. Use as first-line therapy is not current practice, she said during her presentation. She prefers using FMT in patients who have improved on a course of standard treatment, typically vancomycin. The colonic route seems most effective on the endpoint of diarrhea resolution, with effectiveness as high as 93% reported with the cecum/ascending colon as the site of the infusion. Other routes have effectiveness in the range of 80%.

A possible explanation for the effectiveness of FMT is that microbial diversity of the gastrointestinal (GI) tract increases after the transplant and approaches the diversity of the donor.

FMT donors are carefully screened; some of the criteria to be a donor are age 18 to 50 years, no antibiotic use during the previous 6 months, no history of GI illnesses, a normal BMI, and qualifying as a blood donor. Because of the strict criteria and time commitment involved, only about one-quarter of those screened ultimately serve as donors.

Oral capsules are being used as a noninvasive method of delivery at Massachusetts General Hospital (MGH). Its protocol is as follows:

• 15 capsules on each of 2 successive days;
• NPO every 4 hours before and 1 hour after capsule consumption;
• standard phone follow-up only, unless patients relapse; and
• use of a commercially available capsule.

The protocol is offered as billed care at MGH for recurrent CDI. Hohmann shared some of the results obtained at MGH in more than 240 patients treated thus far. Of 183 FMT recipients who reached 8 weeks of follow-up, the first treatment achieved a cure rate of 82%. Retreatment upon relapse resulted in a 91% cure rate. The cure rate improved to 93% after a third treatment if needed.

“I’ve had patients aged 7 to 95 take the capsules successfully,” she said.

The strength of the data is follow-up to 6 months, and the majority of recipients are doing “very well.” The course of recovery is generally rapid. Spontaneous late relapses are rare. Cumulative adverse events were as follows: transient fever (2%), hospitalizations (14%), deaths (8%), diarrhea (62%), nausea/bloating (25%), and ulcerative colitis flare (1%). No deaths were deemed related to FMT.

“I tell people, particularly if they’re younger, that we don’t know all of the effects of FMT,” Hohmann said. “There may be inflammatory problems or conditions related to the GI microbiome that we don’t know about. And that’s what FDA is most concerned about in regulating FMT. I know about two cases of obesity after FMT.”

Original research: Diagnosis

A retrospective cohort study of CDI from October 2014 to September 2015 across the South Texas Veterans Health Care System found that overdiagnosis of CDI appears to be a significant health care problem. Among the patients diagnosed with CDI, one-quarter did not meet the criteria for CDI testing. Two-thirds of patients tested had potential causes of loose stools other than CDI, such as use of stool softeners and laxatives. Inadequate indications at the time of testing patients, together with a very sensitive PCR test, can overestimate the disease burden and detect colonized patients, the authors concluded.

More than one-third (37%) of C. difficile tests that diagnosed hospital-onset-CDI were sent inappropriately, ultimately leading to overdiagnosis, according to a chart review of hospital-onset CDI incident cases at Mount Sinai Beth Israel Hospital in New York. In the patients inappropriately tested, 65% had no documented diarrhea, and almost half of those with diarrhea had used laxatives within 48 hours of the test. Recent antibiotic use was more than six times more common in patients with colonization rather than CDI, although this finding was not significant.

Original research: Transmission

Hospitals in Oregon and Washington that are highly connected to other health care facilities through patient sharing were found to have significantly higher rates of incident facilitywide CDI, and thus represent important prevention targets to reduce the overall CDI burden, reported Rachel B. Slayton, PhD, MPH, at the CDC.

Toxin EIA–negative patients are not a significant source of C. difficile transmission at the University of California, Davis. Polage and colleagues performed a genetic-relatedness analysis of all isolates recovered from symptomatic patients who had a stool sample submitted for C. difficile testing from 11 high-risk units (six ICUs and five wards) during a 4-month period in 2014. They applied National Healthcare Safety Network laboratory-identified, or LabID, rules to classify all positive patients regardless of whether they were toxin-negative or toxin-positive.

The genetic relatedness analysis was carried using capillary gel PCR ribotyping and whole genome sequence (WGS) characterization of single nucleotide polymorphisms (SNPs). Capillary gel PCR ribotyping was used to identify patients who had unique ribotypes and those who had similar ribotypes. For patients in whom the ribotypes were identical and therefore indistinguishable by PCR ribotype, WGS was performed.

A cutoff of two SNPs or less was used to define a transmission event. Directionality was inferred based on the admit discharge transfer date; an overlapping stay inferred a direct transmission, and a nonoverlapping stay inferred an indirect transmission.

Among 398 tests from 273 patients, 98 CDI events (93 patients) were identified. Isolates were recovered from 85 CDI events (81 patients), and 101 isolates from 78 patients were ribotyped. WGS was performed on 65 isolates from 48 unique patients. Using the cutoff of two SNPs or less, Polage and colleagues found that there were only nine transmission events involving 13 patients.

“So, of these 85 CDI events, you only had 10% that appeared to meet a relatedness criteria to some other patient,” Polage said. Further, only two of 39 (5.1%) of hospital-onset CDI events analyzed were related to another CDI event in the cohort.

Despite the inclusion of toxin-negative patients, few hospital-onset CDI events are related to a prior symptomatic patient. The data suggest that horizontal transmission from symptomatic patients accounts for a minority of hospital-onset CDI in hospitals with active inpatient prevention programs.

Original research: Therapeutics (monoclonal antibodies)

Options to reduce the risk of recurrence in patients with recurrent CDI who require concomitant antibiotics are lacking. Among CDI patients in the MODIFY I and MODIFY II trials who received standard-of-care therapy for CDI while also receiving concomitant antibiotics during the follow-up period following standard-of-care antibiotic treatment, bezlotoxumab was associated with a reduction in recurrent CDI of about one-third compared with placebo. In patients receiving concomitant antibiotics during standard-of-care treatment, the rate of recurrent CDI was 17.3% in the bezlotoxumab arm compared with 26.7% in the placebo arm. The rate of recurrent CDI was also lower with bezlotoxumab compared with placebo among the patients who received concomitant antibiotics during the 90 days of follow-up (14.9% vs. 21%).

Another pooled analysis from MODIFY I and MODIFY II, presented by Eric R. Dubberke, MD, at Washington University School of Medicine, showed that bezlotoxumab reduced the proportion of patients with CDI recurrence regardless of the standard-of-care antibiotic used for the presenting episode. The absolute reduction in CDI recurrence in the bezlotoxumab arm compared with placebo was 9.7% in patients treated with metronidazole, 15.4% in those treated with vancomycin, and 11.9% in patients treated with fidaxomicin. The overall absolute reduction in CDI recurrence with bezlotoxumab was 12.2%.

A single dose of bezlotoxumab in patients receiving standard-of-care antibiotics reduced CDI recurrence in a subgroup infected with the BI strain, similar to the overall effect in the primary population, according to another analysis of MODIFY I and MODIFY II. In the group with the BI strain, the rate of CDI recurrence was 40.5% in the placebo arm of the study, which was reduced to 28.2% in the bezlotoxumab arm. By comparison, in patients infected with other strains, CDI recurrence rates were 36.7% in the placebo arm and 20.3% in the bezlotoxumab arm.

Original research: Therapeutics (FMT)

In a severely ill population with many comorbidities (n = 8) at the University of Alabama at Birmingham, FMT had a 75% success rate in patients refractory to standard-of-care antibiotics. Six of the eight patients treated with FMT were experiencing their first case of CDI refractory to initial therapy; four of these experienced cure. Half of the transplants performed were by nasogastric tube; three of these four were cured by FMT (two after relapses) and two patients died of multiorgan failure.

In a retrospective case-control study of 109 patients at the Mayo Clinic, a tapered course of vancomycin preceding FMT reduced the risk for a relapse following FMT by 85% compared with those not receiving tapered vancomycin. The delivery site of the transplanted stool, the volume of the stool, the weight of the stool, and the endoscopist performing the procedure did not affect outcomes.

Data from 482 health care facilities that administer FMT revealed that individual donors from a universal stool bank (OpenBiome) have little impact on the effectiveness of FMT. The findings come from a 1,999-patient cohort, for whom the overall clinical cure rate of FMT was 83.9%. Effectiveness was examined across the 28 donors whose stool was used in more than 10 CDI patients, and researchers found no significant difference in clinical cure rates between donors. The donors were healthy, young, and lean: their mean age was 28.5 years and their mean BMI was 24 kg/m². Study investigator Majdi Osman, MD, MPH, from Harvard Medical School, noted that 99.7% of the population is located within a 4-hour drive of a universal stool bank provider, and 83.2% are within 1 hour.

Conclusion

There is both progress and pitfalls in the diagnosis and therapy of CDI. The risk for CDI and recurrent CDI is related to the host, microbiome (antimicrobial exposure), and the strain of C. difficile. Overdiagnosis of CDI may occur with the use of inappropriate testing. Patients can have diarrhea but not CDI because of excessive antibiotic use (a side effect of such use) or laxative use, or excessive use of proton pump inhibitors. The highly sensitive PCR assays can also result in a significant number of asymptomatic colonized patients being categorized as having CDI. The monoclonal antibody bezlotoxumab reduces the proportion of patients with CDI recurrence, regardless of the standard-of-care antibiotic used for the initial episode and despite the presence of epidemic (027 and related) strains. More knowledge is being gained about FMT for the treatment of recurrent CDI, including delivery methods and donor suitability, but more research is needed on its long-term effects.

References:

Cammarota G, et al. J Clin Gastroenterol. 2014;doi:10.1097/MCG.0000000000000046.

Cepeda V, et al. Abstract 2096. Presented at: IDWeek; Oct. 26-30, 2016; New Orleans.

Cohen SH, et al. Infect Control Hosp Epidemiol. 2010;doi:10.1086/651706.

Dubberke ER, et al. Abstract 838. Presented at IDWeek; Oct. 26-30, 2016; New Orleans.

Gerding DN, et al. Phase 3 double-blind study of bezlotoxumab (BEZ) alone & with actoxumab (ACT) for prevention of recurrent C. difficile infection (rCDI) in patients on standard of care (SoC) antibiotics (MODIFY II). Presented at: ICAAC; Sept. 17-21, 2015; San Diego.

Hartmann C, et al. Abstract 2117. Presented at: IDWeek; Oct. 26-30, 2016; New Orleans.

Hohmann E. Abstract 2397. Presented at: IDWeek; Oct. 26-30, 2016; New Orleans.

Johnson S, et al. Abstract 2114. Presented at: IDWeek; Oct. 26-30, 2016; New Orleans.

Johnson S, et al. Clin Infect Dis. 2007;doi:10.1086/511870.

Johnson S, et al. Clin Infect Dis. 2014;doi:10.1093/cid/ciu313.

Mullane K, et al. Abstract 2115. Presented at: IDWeek; Oct. 26-30, 2016; New Orleans.

Osman M, et al. Abstract 841. Presented at: IDWeek; Oct. 26-30, 2016; New Orleans.

Polage C. Abstract 2396. Presented at: IDWeek; Oct. 26-30, 2016; New Orleans.

Polage C, et al. Abstract 839. Presented at: IDWeek; Oct. 26-30, 2016; New Orleans.

Polage CR, et al. JAMA Intern Med. 2015;doi:10.1001/jamainternmed.2015.4114.

Slayton R, et al. Abstract 840. Presented at: IDWeek; Oct. 26-30, 2016; New Orleans.

Walker J, et al. Abstract 2116. Presented at: IDWeek; Oct. 26-30, 2016; New Orleans.

Wilcox MH, et al. Phase 3 double-blind study of actoxumab (ACT) & bezlotoxumab (BEZ) for prevention of recurrent C. difficile infection (rCDI) in patients on standard of care (SoC) antibiotics (MODIFY I). Presented at: ICAAC; Sept. 17-21, 2015; San Diego.

Yeung M, et al. Abstract 2102. Presented at: IDWeek; Oct. 26-30, 2016; New Orleans.

Disclosure: Goldstein has served on advisory boards for Bayer Pharmaceuticals, Bio-K+, Kindred Healthcare, Novartis, Merck, Daiichi Sankyo, Sanofi-Aventis, Summit and Rempex; served on the speakers bureau for Bayer and Merck; and received research grants from several pharmaceutical companies.