External validation of the INCREMENT-CPE mortality score in a carbapenem-resistant Klebsiella pneumoniae bacteraemia cohort: the prognostic significance of colistin resistance.

External validation of the INCREMENT-CPE risk score (ICS) for 30-day all-cause mortality is needed. There is also scarce information about whether colistin resistance influences the prognosis of carbapenem-resistant Klebsiella pneumoniae (CRKp) bacteraemia. In this study, the ability of ICS to predict all-cause mortality in the KAPECOR cohort was calculated using the area under the receiver operating characteristic (AUROC) curve. The association of colistin resistance with mortality was studied. The ICS showed an AUROC curve of 0.77 (95% CI 0.68-0.86). A cut-off of 8 points showed 96.8% sensitivity and 50.7% specificity. Mortality of low-risk patients was not different in those treated with monotherapy versus combination therapy. However, mortality of high-risk patients treated with combination therapy (37.8%) was significantly lower than in those treated with monotherapy (68.4%) (P = 0.008). To study the prognostic significance of colistin resistance, 83 selected cases of bacteraemia due to colistin-susceptible CRKp were obtained from the INCREMENT cohort for comparison. Colistin resistance could not be shown to be associated with higher mortality in either the high-risk ICS group [adjusted odds ratio (aOR) = 1.56, 95% CI 0.69-3.33; P = 0.29] or in 37 ICS-matched pairs (aOR = 1.38, 95% CI 0.55-3.42; P = 0.49), or in a sensitivity analysis including only KPC isolates (aOR = 1.81, 95% CI 0.73-4.57; P = 0.20), but the precision of estimates was low. These results validate ICS for all-cause mortality and to optimise targeted therapy for CRKp bacteraemia. Colistin resistance was not clearly associated with increased mortality.