Evaluation of methods for typing coagulase-negative staphylococci

One hundred and forty-two coagulase-negative staphylococci (CNS) isolated from dialysate effluent or skin of patients receiving continuous ambulatory peritoneal dialysis (CAPD) were typed by extended antibiogram (16 antibiotics) and biotype (26 reactions). These isolates were then typed by supplementary methods to determine the most suitable typing method for an epidemiological study of antibiotic resistance. These included phage typing, reverse phage typing, plasmid typing, whole-cell protein typing by SDS-PAGE with analysis by densitometry, and immunoblotting. The percentage of isolates typed successfully by the supplementary methods were: phage typing 20%, reverse phage typing 0%, plasmid typing 66%, SDS-PAGE 100%, immunoblotting 100%. The discrimination of each method was: phage typing 20%, plasmid typing 37%, SDS-PAGE 69%, immunoblotting 57%. Reproducibility was 88% for phage typing and 97% for plasmid typing. The reproducibility of the whole-cell protein typing was 83% if the same extracts were used but only 43% when separate protein extracts were analysed on separate occasions. However, strain relatedness was highly reproducible. The determination of an antibiogram-biotype profile was not a sufficiently accurate typing method for an epidemiological study of antibiotic resistance. Whole-cell protein typing by SDS-PAGE or immunoblotting was technically demanding but was the most effective of the supplementary methods for detecting erroneous discrimination and false matching produced by antibiogram-biotype combinations.