Bloodstream infections caused by Klebsiella pneumoniae: prevalence of bla(KPC), virulence factors and their impacts on clinical outcome

BACKGROUND: Klebsiella pneumoniae bloodstream infections (BSIs) occur with significant prevalence and high mortality worldwide. Antimicrobial resistance and virulence are two main factors participating in the pathogenicity of K. pneumoniae. Here we investigated the prevalence of bla(KPC) and virulence factors in K. pneumoniae isolated from patients with BSIs and their association with clinical outcome. METHODS: The clinical data of 285 K. pneumoniae BSI cases diagnosed from January 2013 to December 2015 in a Chinese university hospital were retrospectively evaluated. The “string test” was performed to identify hypermucoviscous K. pneumoniae (HMKP). bla(KPC), rmpA, magA and serotype-specific genes were detected by PCR amplification. Finally, a Cox proportional hazards model was employed to determine the predictors of 14-day mortality. RESULTS: Of these isolates, the prevalence of bla(KPC) and rmpA were 33.3% (95/285) and 31.6% (90/285) respectively. 69 isolates (24.2%, 69/285) were HMKP. rmpA was strongly associated with HM phenotype. The KPC-producing KP and HMKP were almost non-overlapping and only three HMKP isolates harbored bla(KPC). K1 (28, 40.6%) and K2 (22, 31.9%) were the most common serotypes in HMKP. 44.9% of HMKP BSIs had origin of biliary tract infection or liver abscess. The 14-day mortality was 100% in bla(KPC)(+)/HM(+) subgroup (3/3), followed by bla(KPC)(+)/HM(−) (39/92, 42.4%), bla(KPC)(−)/HM(+) (5/66, 7.6%) and bla(KPC)(−)/HM(−) (7/124, 5.6%). The 14-day cumulative survival was significantly different between bla(KPC)(+) and bla(KPC)(−) subgroup (Log-rank p < 0.001) but almost equal between bla(KPC)(−)/HM(+) and bla(KPC)(−)/HM(−) subgroup (Log-rank p = 0.578) under the condition of comparable illness severity between bla(KPC)(−)/HM(+) and bla(KPC)(−)/HM(−) subgroup. Independent risk factors for 14-day mortality were Pitt bacteremia score (HR 1.24, CI 95% 1.13–1.36, p < 0.001), Charlson comorbidity index (HR 1.24, CI 95% 1.09–1.41, p = 0.001), septic shock (HR 2.61, CI 95% 1.28–5.35, p = 0.009) and bla(KPC) (HR 2.20, CI 95% 1.06–4.54, p = 0.034). CONCLUSIONS: Most of HMKP were antibiotic-susceptible and people infected received appropriate antimicrobial therapy, which may explain the favorable outcome of HMKP BSIs. The KPC-producing HMKP BSIs were scarce but life-threatening. bla(KPC) was valuable in predicting 14-day mortality. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s12879-018-3263-x) contains supplementary material, which is available to authorized users.