Severe bacterial infections remain a leading cause of death worldwide, and antibiotics are the cornerstone of treatment. Yet, their use presents a paradox: while timely administration is essential to save lives, especially in sepsis and pneumonia, excessive or indiscriminate use contributes to adverse outcomes, microbiome disruption, and the growing crisis of antimicrobial resistance (AMR). This review highlights the importance of striking a careful balance between urgency and harm. 

Epidemiology varies across settings. Community-acquired infections are typically less severe, but hospital- and intensive care unit (ICU)-acquired infections carry greater mortality and are more often linked to resistant pathogens. The global burden of AMR is highest in sub-Saharan Africa, South Asia, and Eastern Europe, where limited access to diagnostics and healthcare infrastructure further worsens outcomes. Risk factors include frequent antibiotic exposure, hospitalisation, comorbidities, and poor sanitary conditions. 

A central clinical challenge is determining when to start antibiotics. In septic shock or when bacterial infection is highly suspected, immediate initiation is critical. However, in stable patients, a brief diagnostic window—allowing up to three hours for tests—can reduce unnecessary treatment. Importantly, antibiotics should never be prescribed for colonisation or non-infectious inflammation. Source control, when feasible, is a fundamental component of therapy. 

Special populations require tailored approaches. Travellers and migrants may present with unusual infections such as malaria or rickettsioses, while immunocompromised patients face higher risks from both classical and opportunistic pathogens. In these groups, empiric regimens should be guided by disease severity and local resistance patterns rather than immune status alone. 

Diagnostics are evolving rapidly. Traditional cultures remain essential, but new tools—including matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS), multiplex polymerase chain reaction (PCR) panels, and metagenomics—enable faster identification of pathogens and resistance genes. When paired with antimicrobial stewardship, these tests shorten the time to optimal therapy, though consistent mortality benefits have yet to be shown. 

Therapeutic strategies emphasise appropriateness, de-escalation, and pharmacokinetic optimisation. Excessively broad empiric therapy is harmful, but delays in appropriate treatment are equally dangerous. Continuous or prolonged beta-lactam infusion, supported by therapeutic drug monitoring, helps achieve effective concentrations in critically ill patients. Pharmacists play a pivotal role in managing dosing and drug–drug interactions. 

Antibiotics also exert collateral damage. Beyond classical toxicities such as nephrotoxicity or neurotoxicity, they disrupt the microbiota, facilitating Clostridioides difficile infection, secondary bloodstream infections, and prolonged dysbiosis. Evidence supports shorter courses—five to seven days in most cases—provided adequate source control is achieved. Biomarkers like procalcitonin may assist in limiting duration. 

Looking ahead, precision medicine offers promise. Genomic diagnostics, host-response profiling, and artificial intelligence may enable truly individualised therapy. Alternative strategies such as bacteriophages, anti-virulence drugs, and microbiome-preserving approaches are under development. 

The future of antibiotic therapy lies in precision—starting early when needed, tailoring regimens to the patient and pathogen, minimising collateral damage, and stopping treatment as soon as possible. Stewardship across human, animal, and environmental health remains the cornerstone of preserving antibiotic effectiveness. 

Reference 

Timsit, JF., Ling, L., de Montmollin, E. et al. Antibiotic therapy for severe bacterial infections. Intensive Care Med (2025). https://doi.org/10.1007/s00134-025-08063-0