STRAIN RELATEDNESS IN PERSISTENT AND RECURRENT CANDIDURIA

Purpose: We assessed the value of determining strain relatedness in di fferentiating persistent from recurrent candiduria. Materials and Meth ods: Prospective monitoring of patients with candiduria (10(4) or grea ter colony forming units per ml.) during a 5-month period. All patient s with persistent or recurrent infection after documented clearance we re selected. Pair isolates were typed using restriction endonuclease a nalysis of genomic deoxyribonucleic acid with SfiI. Isolates were cons idered related if all deoxyribonucleic acid bands matched. Results: We encountered 22 and 5 patients with persistent and recurrent infection , respectively. The isolates were recovered 1 to 140 days apart (21.56 +/- 28.97). Most patients were women (85.2%) with a mean age of 66.41 +/- 18.11 years. Risk factors included antibiotics (100%), indwelling catheter (88.9%) and diabetes mellitus (40.7%). Of 15 individuals who received antifungal therapy candiduria persisted in 10 and resolved b ut recurred within 4 to 26 days (13.00 +/- 9.08) after treatment in 5. Candida albicans accounted for 34 of 58 isolates (58.6%), and it was mixed with other species in 4 cultures. Paired strains were geneticall y identical in 26 of 27 patients. Strain persistence was documented in 21 of 22 cases with persistent infection and in all 5 patients with r ecurrent disease. Conclusions: These findings show that strain persist ence is exceedingly frequent in candiduria. These results imply that d etermining strain relatedness of Candida urinary isolates may not be r eliable in differentiating persistent from recurrent infection.