The point prevalence of Clostridium difficile stool shedding in hospitalized infants from two neonatal intensive care units (NICUs) was examined utilizing standard clinical testing compared with duplex PCR to identify toxigenic and non-toxigenic C. difficile strains.
All infants from the two NICUs affiliated with a single academic medical center were eligible for inclusion. Stool collection was blinded to patient characteristics and occurred during a one week period at each NICU and repeated with a second weeklong collection 6 months later to increase sample size. Stools were tested for C. difficile using EIA (GDH/toxin A/B) with samples testing +/+ or +/− subsequently evaluated by Loop-Mediated Isothermal Amplification (LAMP) and by duplex PCR amplification of tcdB and tpi (housekeeping) genes. Cytotoxicity assays were performed on all samples positive for C. difficile by any modality.
Eighty-four stools were collected from unique infants for evaluation. EIA results showed 6+/+ [7.1%], 7 +/− [8.3%], and 71 −/− [84.5%] samples. All 6 EIA +/+ were confirmed as toxigenic C. difficile by LAMP; 6/7 EIA +/− were negative by LAMP with one identified as invalid. Duplex PCR concurred with LAMP in all 6 stools positive for toxigenic C. difficile. PCR identified 2 EIA −/− stools positive for tpi, indicating shedding of non-toxigenic C. difficile. Cytotoxicity assay was positive in 4/6 duplex PCR positive samples and negative for all stools that were EIA +/− but negative by molecular testing.
C. difficile blinded point prevalence in infants from two NICUs was 7.1% by molecular methods; and lower than expected based on historical incidence estimates. In house duplex PCR had excellent concordance with clinically available LAMP and EIA tests, and added detection of non-toxigenic C. difficile strain shedding. Evolving NICU care practices may be influencing the composition of infant gut microbiota and reducing the point prevalence of C. difficile shedding in NICU patient stools.