Bacterial infection is one of the major causes of human disease, and antibiotics have been developed for killing bacterial pathogens[1]. However, the number of bacteria resistant to broad-spectrum antibiotics is increasing because of antibiotic misuse and abuse, posing a great threat to human health[2]. Rapid and accurate characterization of bacterial susceptibility to antibiotics is therefore urgently needed for guiding correct and effective antibiotic therapy in patients with serious infections.
Klebsiella pneumoniae, a Gram-negative bacterium, is one of the most important causes of nosocomial infections[3]. Klebsiella pneumoniae is also the most common multi-drug-resistant bacteria[4]. This bacterium can cause pneumonia, urinary system infection, peritonitis, sepsis, and other diseases in patients[5]. Especially for patients in the ICU ward, due to the serious disease of these patients, the immune system is often compromised. If a patient infected with Klebsiella pneumoniae does not receive effective treatment, it can affect the patient’s life[6]. Thus, the faster information on bacterial susceptibility to antibiotics is available, the faster the patients’ treatment can be optimized, minimizing the risk of developing antibiotic resistance as well as use of costly pharmaceuticals. This potentially saves the patients’ life and reduces their irreversible lone-term side effects.
Heavy water (D2O)-based single-cell Raman micro-spectroscopy (D2O-Raman) combining Raman microspectroscopy with heavy water (D2O) has been demonstrated for quantitative assessment of the metabolism-inhibiting effects of drugs[7]. Raman shift at the C-D (carbon–deuterium vibration) band around 2040–2300 cm− 1, which has been considered a universal biomarker for general metabolic activity, has been used to probe bacterial responses to different drugs by simple adding heavy water (D2O) to samples[8]. In addition, based on the metabolic activity, the minimum inhibitory concentration (MIC-MA) has been proposed for evaluating the metabolism-inhibiting efficacy of drugs. Furthermore, this method is able to detect “nongrowing but metabolically active” (NGMA) cells, which can lead to many latent or recurring infections[9]. Despite this progress, D2O-Raman spectroscopy has not been established as a reliable and rapid AST tool for clinical applications. Here, we tested 29 Klebsiella pneumoniae isolates from ICU patients with antibiotics resistance in conditions as close as possible to a real clinic using commercially available antibiotics at the same concentration and preparation as those used in clinical treatments. The results showed that, compared with the Vitek 2, the results of D2O-Raman spectroscopy showed good agreement in terms of susceptibility and resistance (S/R), with a concordance of 90%. Moreover, discriminant factor analysis (DFA) of these D2O-Raman results were successful for discriminating the type of resistance. These promising results highlight the potential of this technique for faster determination of bacterial susceptibility to certain antibiotics.