Commentary on "High frequency of nonadherence to Clostridium difficile treatment guidelines".

P. Froom

DOI: https://doi.org/10.14423/SMJ.0000000000000168

2014-09-01

Abstract:The diagnosis of aClostrium difficile infection (CDI) is seen usually in hospitalized patients with diarrhea who are most often being treated with antibiotics. CDI is rarely noted in patients with severe abdominal pain without diarrhea. As pointed out in the study by McEllistrem and colleagues in this issue of the Southern Medical Journal, recommended treatment is metronidazole for mild CDI and vancomycin for more severe disease. The clinical utility of treatment may include the shortening of the symptomatic period, prevention of severe complications, and prevention of relapses, as well as the spread of infection to other patients. Before implementing consensus guidelines, we need to consider the evidence. A major problem is defining CDI in hospitalized patients. Iv et al concluded that the state of the art for laboratory testing is confusing, controversial, and in flux. If cultures are taken systematically upon admission and during hospitalization, then most patients with positive cultures will be asymptomatic carriers. Symptomatic infections are the consequence of toxins identified by various tests. There is, however, no gold standard and there is considerable interlaboratory variation in the choice of tests and in the test results themselves. Misclassification will lead to inappropriate treatment and will bias randomized controlled trials toward the absence of an observed effect. The enzyme immunoassay for toxin A/B is the primary test used in up to 90% of clinical laboratories; however, it is insensitive and nonspecific. Adding to or replacing the test with other tests such as the polymerase chain reaction test to detect the gene for toxin production (tcdB gene) still results in a large proportion of patients being treated unnecessarily; most antibiotic-associated diarrhea is not caused by CDI, and the use of this test and others alone or in various combinations leads to a considerable but imprecise risk of overdiagnosis and inappropriate treatment. Not only is there considerable diagnostic uncertainty, but a Cochrane review also summarized the evidence and concluded that there is uncertainty whether mild CDI should be treated at all. Furthermore, there are various ways to divide patients into those with mild or severe disease, and the evidence that differential treatment has clinical utility is supported by only a single randomized controlled study that had considerable limitations. Zar et al randomized patients to be treated with either metranidizol or vancomycin. According to their definitions, 46% (69/150) of the patients had severe disease, and those patients fared better with vancomycin, with a cure rate of 97% (30/31) compared with 76% (29/38) in patients treated with metranidizol. Treatment failure was defined as the persistence of diarrhea, the need for colectomy, or death after 5 days of therapy; however, the study also included a surrogate and a questionable endpoint, a positive result of a C. difficile toxin A assay after 6 days of treatment. The number of individual outcome measures was not detailed, making it difficult to determine the clinical utility of the differences in outcomes, and there was no significant difference in the relapse rate. It was remarkable that Zar et al recruited patients from a single teaching hospital during an 8-year period but the interpretation of the data is difficult because the clinical presentation of the patients, the selection process, and the randomization practices were unclear. Another major concern is the advisability of extrapolation of these findings to other settings with fewer patients with severe disease, and to patients without low albumin concentrations; overall, eight of the nine patients not responding to metronidizol in their study had albumin Invited Commentary