Diagnostic value of repeated enzyme immunoassays in Clostridium difficile infection

Abstract

Objectives: There has been a significant increase in the prevalence, severity, and mortality of Clostridium difficile infection (CDI), with an estimated three million new cases per year in the United States. Yet diagnosing CDI remains problematic. The most commonly used test is stool enzyme immunoassay (EIA) detecting toxin A and/or B, but there are no clear guidelines specifying the optimal number of tests to be ordered in the diagnostic workup, although multiple tests are frequently ordered. Thus, we designed a study with the primary objective of evaluating the diagnostic utility of repeat second and third tests of stool EIA detecting both toxins A and B (EIA (A&B)) in cases with negative initial samples, and sought to describe the physicians' patterns of ordering this test in the workup of suspected CDI.

Methods: A retrospective study was carried out using a database of all stool EIA (A&B) tests ordered for a presumptive diagnosis of CDI. All patients were adults admitted to a major teaching hospital over a three-and-a-half-year period (tests completed within 5 days of ordering the first test were grouped into a single episode, and only the first three samples per episode were analyzed). Age, gender, and results of stool EIA were tabulated. In addition, physicians' ordering patterns and proportion of positive stools relative to the number of tests ordered were also analyzed. A single positive EIA result was interpreted as evidence for the clinical presence of CDI.

Results: A total of 3,712 patients contributed to 5,865 separate diarrhea episodes (total stool EIA (A&B)=9,178), and 1,165 (19.9%) of these episodes were positive for CDI. Of the positive patients, 73.2% were over the age of 65 years and 54.2% of them were females. The most frequent ordering pattern for presumptive CDI was a single stool test (60.1%), followed by two more tests (23.2%). Three tests were still ordered in 16.6% of the cases. Of the 1,165 positive cases, 1,046 (89.8%) were diagnosed in the very first test, 95 (8.2%) in the second, and only 24 (2.0%) in the third test. In 1,934 instances, a second test was ordered after an initial negative result, of which 95 (4.91%) became positive. In 793 episodes, a third test was ordered after two negative samples, of which only 24 (3.03%) became positive.

Conclusions: This study highlights the low diagnostic yield of repeat stool EIA (A&B) testing. Findings strongly support the utility of limiting the workup of suspected CDI to a single stool test with only one repeat test in cases of high clinical suspicion, and avoiding the routine ordering of multiple stool samples. As Clostridium difficile is becoming an endemic health-care problem resulting in major financial burdens for the US health-care system, clear guidelines specifying the optimal number of stool EIA (A&B) tests to be ordered in the diagnostic workup of suspected CDI must be established to assist physicians in the practice of evidence-based medicine.