Phage-antibiotic combinations against Klebsiella pneumoniae: impact of methodological approaches on effect evaluation
RB Gorodnichev, Anastasiia O. Krivulia, Maria Kornienko, Narina K. Abdraimova, Maja V. Malakhova, Marina V. Zaychikova, Dmitry Bespiatykh, Valentin A. Manuvera, Egor Shitikov
Abstract
Background The combined use of bacteriophages and antibiotics represents a promising strategy for combating multidrug-resistant bacterial pathogens. However, the lack of uniformity in methods for assessing combination effects and experimental protocols has resulted in inconsistent findings across studies. This study aimed to evaluate the effects of interactions between phages and antibiotics on Klebsiella pneumoniae strains using various statistical approaches to formalize combination effects. Methods Effects were assessed for four antibiotics from distinct classes (gentamicin, levofloxacin, meropenem, chloramphenicol), three phages from different genera (Dlv622, Seu621, FRZ284), and a depolymerase (Dep622) on three K. pneumoniae strains of the KL23 capsule type. Antibiotics were used at C max concentrations, and phages at sublethal levels. A modified t -test, Bliss independence model, two-way ANOVA, and checkerboard assay were employed to evaluate the results. Results Among 48 combinations, 33 effects were statistically significant, including 26 cases of synergy and 7 of antagonism. All statistical methods showed consistency in identifying effects; however, the t -test and Bliss method detected a greater number of effects. The strongest synergy was observed with levofloxacin in combination with Seu621 or Dep622 across all bacterial strains. Checkerboard assays confirmed synergy in selected cases but indicated that combined effects could vary with antimicrobial concentrations. Conclusion The choice of analytical method substantially impacts the detection of phage-antibiotic effects. The t -test and Bliss method, due to their simplicity and sensitivity, may be optimal for clinical application, while two-way ANOVA for confirming strong interactions. These results emphasize the need to consider interaction characteristics when designing therapeutic strategies.