Cumulative survival profiling: a new PAP-based method for detecting heteroresistance in staphylococcal clinical isolates

Abstract The modified area under the curve of the population analysis profile (PAP-AUC) developed by Wootton et al. (J Antimicrob Chemother 47(4) 2001) is considered the gold standard for detecting heteroresistance in staphylococci. However, the PAP-AUC method does not explicitly control for the variation in the amount of initial inoculum between tested isolates and the comparator reference heterogeneous vancomycin-intermediate S. aureus (hVISA) strain (Mu3), which has been reported to influence classification results. Further, the method requires batching the isolates and utilizing single Mu3 PAP for each, which does not capture the heterogeneity of Mu3 and adds to the labor intensity of the method. We confirmed that the area under the PAP curve for Mu3 is strongly dependent on the initial inoculum. After interpreting this strong relationship, we derived the cumulative survival: a new metric that lacks the initial inoculum influence and measures the experience under the vancomycin concentration gradient. We devised a new method, termed cumulative survival profiling (CSP), to infer a distribution for the cumulative survival metric for isolates; CSP requires conducting multiple PAPs with a broad range of initial inoculum values. We used CSP to profile Mu3 and one isolate (CoNSB18) from 67 clinical isolates of coagulase-negative staphylococci associated with central-line bloodstream infection (CoNS CLABSI). Mu3 and CoNSB 18 were the respective references for hVISA and hVISA that is associated with poor clinical response. Their CSPs reveal two overlapping distributions of cumulative survival, reflecting the heterogeneity of hVISA and indicating that the most resistant in the hVISA spectrum are likely to be associated with a poor clinical response. The 2.5% quantiles of their distributions successfully cross-validated >80% of their PAPs (accuracy). To detect similar isolates to these two reference strains among the CoNS CLABSI isolates, rather than conducting CSP for each isolate, we retrospectively estimated the cumulative survivals using point estimate profiling (PEP) utilizing single PAPs. We estimated that two-thirds of the CoNS CLABSI isolated from unique subjects are heteroresistant and that about 80% of those may be associated with a poor clinical response. Standardization of PAP-based detection of heteroresistance and/or poor clinical response is within reach by using CSP and PEP. Our results should raise the profile of heteroresistance in CoNS CLABSI.