Klebsiella pneumoniae is a Gram-negative, encapsulated, facultatively anaerobic bacteria which is one of the most important opportunistic pathogens in hospitals all over the world (Wyres et al., 2020; Chang et al., 2021). With the widespread around the world, the infection of K. pneumoniae has caused huge health and economic burden to mankind in the past decade. In 2017, the World Health Organization included it in the list of dominant pathogens, emphasizing its extremely important position in public health issues (World Health Organization, 2017).

In humans, K. pneumoniae often colonizes the nasal and digestive tract without causing any symptomatic disease. However, the colonization can turn into a bloodstream infection of K. pneumoniae (BSI-KP) when the host immunity fails to control the pathogen growth. As a large-scale multicenter epidemiological study and in-depth genomic analysis in China demonstrated, BSI-KP has a considerable prevalence and high mortality worldwide (Jia et al., 2021). It is reported that the case fatality rate of BSI is between 21% and 69% (Zhang et al., 2020). Clinicians give great attention to BSI-KP because it can easily cause sepsis/septic shock in patients that often lead to multiple organ dysfunction syndromes (MODS) and death, especially encountered with hyper-virulent strains. Furthermore, BSI-KP can affect the patient's prognosis by increasing the length of hospital stay, treatment cost, and complications after discharge. Therefore, it is an urgent and valuable task for us to explore the risk predictors in patients with BSI-KP.

Statin is a kind of commonly used drug in clinical practice and have both lipid-lowering and anti-inflammatory properties. Nearly twenty years ago, Almog et al. (Almog, 2003) hypothesized that statins have a strong protective effect against sepsis by virtue of diverse anti-inflammatory effects that are independent of their lipid-lowering ability. Five years ago, Lee et al. (Lee et al., 2018) found that statins potentially decreased the 30-day mortality in sepsis patients from a large population.Recently, it was found that prior statin treatment take the positive effect in patients with COVID-19 and type 2 diabetes infection (Cheng et al., 2019; Davoudi et al., 2021). During the past two decades, a few researches was focused on the evidence of the statins use improving the outcome of patients with different infections, but are still need to be strengthened.

Patients with BSI-KP have a relatively high risk of poor outcomes, and the situation of CRKP infection often emerged after treating with antibiotics (Fair and Tor, 2014). There were a series of studies on the risk factors of mortality in patients infected with carbapenem-resistant K. pneumoniae (CRKP) but few was focused on the susceptible strains which accounted for majority episodes in clinical settings (Jia et al., 2021). The database of MIMIC IV (Medical Information Mart for Intensive Care IV) includes clinical data of patients with BSI-KP during an eleven-year period, and most of the K. pneumoniae strains were susceptible to carbapenems. In this study, based on the MIMIC IV database, we aimed to find effective factors especially the medication factors such as antibiotics and statins, to predict the risk of death and severity, which would provide the evidence for clinical practice in reducing the mortality and severity in BSI-KP patients.

It was the first time for us to investigate the risk factors of mortality in patients with BSI-KP based on the MIMIC IV database. Since a number of improvements have been made, MIMIC IV released a series of new tables, including table microbiologyevents that comprised of variety culture results. This study collected data from MIMIC IV v1.0, which was released on March 16th, 2021. Compared to recent research results, the 28-day mortality in our study was relatively lower (11.8%) partly because most BSI-KP isolates were carbapenem-susceptible, and the antimicrobial treatment guided by the culture result was effective. According to the data, isolates from 1/33 patients of the non-survival group were meropenem resistant. Only 3/279 patients were positive with meropenem-resistant KP in blood culture.

As results from Table 2 , patients with hyperlipidemia and those with statins use had a lower risk of death within 28 days than those without hyperlipidemia and statins use. It was also found that hyperlipidemia patients are more likely to use statins than non-hyperlipidemia patients ( Table S2 ). Then we focused on the effect of statins used in 28-day mortality. The result suggested that statins use presented no statistical difference of mortality analysis neither in hyperlipidemia patients nor in non-hyperlipidemia patients ( Table S1 ), which probably attributed to the sample size and the intergroup heterogeneity. A multivariate analysis was performed to confirm the potential correlation between statins and 28-day mortality. As shown in Table 2 , patients with statins use had a lower risk of death within 28 days than those without statin use. Therefore, statins use was an independent protective factor for the death of patients with BSI-KP.

Statins have lipid-lowering properties by inhibiting 3-hydroxy-3 methylglutaryl coenzyme A reductase, a rate-limiting enzyme in cholesterol biosynthesis. They also have anti-inflammatory and immunomodulatory effects resulting in improved endothelial function, reduced thrombogenicity, and plaque stabilization (Dobesh and Olsen, 2014). Fifteen years ago, a meta-analysis published on Lancet (Lewington et al., 2007) demonstrated that statins use could reduce coronary event rates and total stroke rates in patients with a wide range of ages and blood pressures. Researchers had focused on the functions of statins earlier, especially in patients with sepsis (Almog, 2003; Novack et al., 2006). Since then, a series of studies have been performed and showed that statin use improved the outcome of BSI patients with transplantation and cancer (Sun et al., 2009; Hsu et al., 2009; Yu et al., 2014). However, there found no evidence indicating that statins use is associated with a reduction in mortality of sepsis patients (Pasin et al., 2013; Wan et al., 2014). In our study, it was observed that statins use had a significant difference in the risk of septic shock and the 28-day mortality of patients admitted into the intensive care unit (ICU) ( Table S3 ). Besides, Mehl et al. (Mehl et al., 2015) proposed that prior statin use is associated with a lower 90-day total mortality in Gram-negative BSI but not in Gram-positive BSI. However, two studies showed statin treatment in patients with Staphylococcus aureus bacteremia was associated with lower mortality (Lopez-Cortes et al., 2013; Smit et al., 2017). This study also found that statin use could decrease the 28-day mortality in patients with BSI-KP ( Table 2 ).

The effect of statins in patients may be different for individual statins. Lee et al. (Lee et al., 2018) found that simvastatin (hazard ratio [HR], 0.72; 95% CI, 0.58-0.90) and atorvastatin (HR, 0.78; 95% CI, 0.68-0.90) were associated with improved 30-day survival, whereas rosuvastatin was not. Quellette et al. (Ouellette et al., 2015) pointed out that atorvastatin use was associated with improved mortality in septic patients compared with pre-hospital simvastatin use. In addition, a study base on patients with chronic obstructive pulmonary disease (COPD) also observed that Fluvastatin and Atorvastatin are more effective in reducing C-reactive protein and pulmonary hypertension (Lu et al., 2019). It inferred that the drug-specific effect of statins on BSI is not correlated to their lipid-lowering potency. Statins have pleiotropic effects such as anti-inflammatory, antithrombotic, and antioxidant effects and their lipid-lowering effects. In sepsis and septic shock, cytokine release from endothelial cells, procoagulant molecules, and thrombocyte production are encouraged. Statins act on sepsis by providing vascular relaxation in the endothelium and reducing the expression of adhesion molecules and cytokines (Goncuoglu et al., 2021).

On the other hand, Pawar et al. (Pawar et al., 2018) found that continued statins use for at least two days after admission provided a survival benefit among bacteremia patients. Similar to our study, about 77% (21/91) of patients continued using statins for more than two days. Thus, more laboratory and clinical data are encouraged, and randomized controlled trials will be needed to define the role of statins in BSI patients.

Based on the binary logistic regression results, we built two ROC curves to predict the risk of 28-day mortality and septic shock in BSI-KP patients. In the model for 28-day mortality, the variable age was considered as a confounding factor because of no clinical significance and was excluded ( Figure 4A ). Quinolones use was regarded as a protective factor mainly attributed to the majority isolates were sensitive to ciprofloxacin ( Figure 2 ). There were two main reasons for the protective effect of quinolones. One is the combination of drugs in the process of anti-infection treatment, and the other is that the utilization rate of carbapenems (20.8%) is lower than quinolones (57.7%). By contrast, the effectiveness of variable gender in the model for septic shock was still preserved since it was not determined whether males were less likely to suffer from septic shock ( Figure 4B ). As a result, both of the models had good value for prediction.

This study has some limitations. First, the data were obtained from a single institution during a limited study period, and thus, the results may not be widely representative or generalizable. Second, the data was insufficient to detail the effect of statins use on the 28-day mortality of patients. Third, since it was a retrospective study, the confounders still exist, such as patients’ comorbidities, and the selection bias may influence the results. With the current knowledge, more studies are needed to confirm the use of statins to benefit from their pleiotropic effects in the treatment of BSI-KP patients.

The use of statins could decrease the risk of 28-day mortality in patients of BSI-KP. The risk factor-based prediction model provided evidence for drug treatment in BSI-KP patients. Paying more attention to the treatment strategy will be an optimal way to improve outcomes in clinical practice.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

QX designed the study and extracted the data. BZ and QX analyzed the data and drafted the manuscript. YX and BZ supervised the study and revised the manuscript. PS contributed to the data analyzation and the figure presentation. All authors contributed to the article and approved the submitted version.