Objectives: To compare the performance of two rapid systems for the diagnosis of cholera with the culture method, and to propose a strategy for improving the specificity and sensitivity of these systems and reducing the costs involved in making a diagnosis.
Methods: The following institutions participated in the study: the National Bacteriology Referral Center (Centro Nacional de Referencia en Bacteriologia, CNRB) of the Costa Rican Institute for Research and Teaching in Nutrition and Health (Instituto Costarricense de Investigacion y Ensenanza en Nutricion y Salud, INCIENSA) and various hospitals in the provinces of Alajuela, Guanacaste and San Jose, in Costa Rica. A total of 237 feces samples were used to asses the performance of two tests for the rapid detection of Vibrio cholerae 01: the Pathogen Detection Kit (PDK, Intelligent Monitoring Systems, Gainesville, Florida, USA) and Cholera-SMART (New Horizons Diagnostics Corp., Columbia, Maryland, USA), both when applied directly (direct SMART and direct PDK) and when applied to specimens cultured in broth-enriched medium for 6 hours (SMART-6 and CPK-6) and for 18 hours (SMART-18 and PDK-18) at 37 degrees C in alkaline peptone water. Liquid and partially formed stools were cultured and examined by means of the rapid direct test; when the initial result was negative, the tests were repeated after culture for periods of 6 and 18 hours. Rectal and fecal swabs were obtained from feces cultured in enriched-broth medium for 6 and 18 hours. In addition, we studied the sensitivity of the rapid testing systems by using pure cultures of V. cholerae 01 (strain SOS-833, CNRB, Costa Rica) that were incubated for 18 to 24 hours, and we assessed the usefulness of observing motility under the microscope in order to rationalize the use of rapid methods.
Results: The sensitivity of the direct SMART test and of the direct PDK test was 100% when samples obtained from liquid and partially formed stools and from the intestinal contents of dead bodies were used. With these samples, the direct SMART procedure showed a specificity of 100%, whereas the direct PDK procedure showed a specificity that ranged from 85.7% to 77.4%, depending on the type of sample. False positives obtained with the direct PDK method turned out to be negative with PDK-6 and PDK-18. Among the rectal and fecal swabs of persons with and without diarrhea or who had received prior treatment with antibiotics, three results that were negative with the SMART-6 procedure and two that were negative with the PDK-6 procedure turned out to be positive with the SMART-18 and PDK-18 procedures, respectively. Both systems showed excellent concordance (kappa index above 0.9) throughout. Both systems were sensitive to 6 x 10(7) colony-forming units per milliliter (cfu/mL), which was concordant with the microscopic observation of 10 microorganisms or more per field with the type of motility that characterizes vibrios (at 1000 x magnification). Samples having fewer than 10 microorganisms with the motility that characterizes vibrios had concentrations between 6 x 10(3) and 6 x 10(6) cfu/mL and became positive only after incubation in enriched-broth medium for 6 to 18 hours. We propose a strategy for diagnosing the presence of V. cholerae 01 infection in less time than it takes with traditional methods, with positive and negative predictive values of 100%.
Conclusions: The SMART and PDK systems make it possible to accurately diagnose cholera quickly, don't require sophisticated equipment or highly qualified technical personnel, and perform satisfactorily in field conditions. Through the proposed strategy, it becomes possible to improve the specificity and sensitivity of these systems and to reduce the cost of making a diagnosis, thus making them suitable for use in cholera surveillance in low-income settings where this disease is a serious public health problem.