Value of Tropheryma whipplei quantitative polymerase chain reaction assay for the diagnosis of Whipple disease: usefulness of saliva and stool specimens for first-line screening.

BACKGROUND Whipple disease (WD) is a chronic infectious disease caused by Tropheryma whipplei. WD DNA has been found in stool and saliva specimens from patients and asymptomatic carriers. METHODS A total of 4418 samples that were sent to our center for determination of WD were tested by a T. whipplei-specific quantitative polymerase chain reaction (PCR) based on repetitive sequences. Definite WD was diagnosed in 71 patients, including 55 patients with classic WD (defined by positive results of periodic acid-Schiff staining and/or specific immunohistochemistry of small-bowel biopsy specimens) and 16 patients with localized WD (including patients with endocarditis, neurologic infection, and uveitis). RESULTS Of the persons without WD, 2.3% had stool specimens positive for T. whipplei by PCR and 0.2% had saliva specimens positive for T. whipplei by PCR. Diagnosis of WD was likely in patients with positive results of both PCR of saliva specimens and PCR of stool specimens (positive predictive value, 95.2%). When the bacterial load was >10(4) colony-forming units per g of stool, the positive predictive value was 100%. A negative result of PCR of a saliva or stool specimen had a negative predictive value of 99.2% for classic WD. For localized WD, positive results of both PCR of saliva specimens and PCR of stool specimens had a sensitivity of 58% (compared with 94% for classic WD). The positive predictive value of testing of blood, cerebrospinal fluid, and urine specimens was 100% for each, and the positive predictive value for testing of duodenal biopsy specimens was 97.5%. CONCLUSIONS T. whipplei-specific quantitative PCR of saliva and stool specimens should be performed as first-line noninvasive screening for WD. When the results for both types of specimens are positive, diagnosis of classic WD should be highly suspected, especially if a high bacterial load is detected. Because PCR of saliva and stool specimens lacks sensitivity for determination of localized WD, invasive samples should be tested on the basis of clinical manifestations.