8/5/09 - NorthShore University HealthSystem (NorthShore) researchers advocate a renewed approach to accurately diagnose Clostridium difficile infection (CDI or C. diff) because of what they believe are forgotten principles that may lead to widespread misdiagnoses. Their paper, “Does My Patient Have Clostridium difficile Infection?” appears in the August 4, 2009 issue of the Annals of Internal Medicine, the flagship journal of the American College of Physicians. 

CDI is a powerful and potentially fatal, antibiotic-resistant bacterial infection that severely impacts the human digestive system. “It is more virulent and more frequent in the United States than it used to be and the reason is unknown,” said Lance R. Peterson, MD, FASCP, associate epidemiologist at NorthShore and the paper’s co-author. “In a time of changing disease characteristics, accurate diagnosis of this infection is critical.” 

Peterson and co-author, Ari Robicsek, MD, epidemiologist at NorthShore, believe clinicians should reexamine two very important diagnostic principles critical to accurately diagnosing CDI. First, the patient must have clinically significant diarrhea (three or more loose stools per day) when tested. Second, laboratories must perform only one high-sensitivity lab test to detect CDI because multiple testing is often misleading. 

“CDI has always relied on using a sensitive test to diagnose the presence of the organism combined with testing only those patients who have the proper symptoms,” said Peterson. “If an inferior test is used, then patients with the disease are missed. If testing is done on people without the proper symptoms, then money is wasted for testing people who really are not sick with this infection, and a false-positive test in such a setting can cause a person to be given an antibiotic they do not need. 

“Current testing generally relies on lab tests with poor sensitivity, so we have forgotten that the basic principle for detecting CDI is to use a test with high sensitivity. Further, clinicians often test patients without enough diarrhea to make the clinical diagnosis of CDI and have, thus, forgotten the principle that only the proper patients should be tested.” 

“As for multiple testing,” added Robicsek, “if the first test has a negative result, then the clinician should not perform another test. If a second test is conducted, odds are if that test is positive it is most likely a false-positive result. In other words, the second test – if positive – is more likely to be wrong than right, and that leads to widespread misdiagnosis.” 

Previous research indicates that commonly used laboratory tests to detect CDI such as glutamate dehydrogenase (GDH) showed false-positive and false-negative results repeatedly and had too low a specificity rate to be used as a stand-alone diagnostic test. The easy-to-use, rapid enzyme immunoassay (EIA) test also provided false-positive results, especially after multiple testing. 

“The best rapid test is the real-time polymerase chain reaction (qPCR) because it has comparable sensitivity to culture of toxigenic C. difficile organisms and can produce a reliable result very quickly, usually in one or two hours,” said Peterson. 

Peterson and Robicsek conducted an in-house study and found that the qPCR test had 93.3% sensitivity and 97.4% specificity versus the EIA test that had 73.3% sensitivity and 97.6% specificity. In addition, the qPCR test targets the more virulent toxin B strain found in CDI. 

“You don’t have to repeat the qPCR test. So the first test is a true indication that the person has CDI because the test is highly sensitive and specific. An accurate test result leads to a proper diagnosis which means the patient receives the appropriate medical treatment and care,” concluded Robicsek. 

The authors hope their report stimulates discussion and better understanding that when a proper test is done for a patient with the clinical findings of CDI, it only needs to be done once a majority of the time; and that if a highly sensitive test is negative then another reason for diarrhea should be considered.

× Alternate Text