01Purpose and principlesWhat the treatment does and how it fits into care.
Antibiotic choice begins with syndrome and severity. Community exposure, hospital contact, prior resistant isolates and anatomical source determine the initial spectrum. The same drug may be appropriate for uncomplicated infection and inadequate for meningitis, endocarditis, abscess or prosthetic infection because dose, penetration, bactericidal requirement and duration differ.
Targeted treatment uses organism identity, susceptibility, source and patient response. A laboratory 'sensitive' result does not guarantee clinical suitability when the drug cannot reach the compartment, is unsafe for the patient or has no evidence for the syndrome. Conversely, broad treatment should not continue merely because the patient improved before cultures returned.
Prescribing safety is dynamic. Acute kidney injury, obesity, critical illness, dialysis, drug interactions and absorption alter exposure. Therapeutic drug monitoring is part of treatment for selected aminoglycosides and glycopeptides. Every prescription needs a documented indication, dose, route, review date, duration and contingency if results or physiology change.
Key points
- Empirical therapy is a reasoned bridge to better information: record syndrome, likely source, severity, organisms covered, route, duration plan and review time.
- Use previous microbiology, recent antibiotics, healthcare or travel exposure, colonisation, immune status and local resistance data to estimate resistant-organism risk.
- Choose the narrowest regimen that safely covers the dangerous plausible pathogens and penetrates the infected compartment; more drugs do not automatically mean better treatment.
- Dose for indication, weight, renal and hepatic function, age, pregnancy, allergy phenotype and altered pharmacokinetics; severe infection may require a full loading dose despite renal impairment.
- At 48 to 72 hours review diagnosis, cultures, source control, response, route and duration: stop, narrow, switch, continue with an end date, or broaden only for a documented reason.
- Combine antimicrobial treatment with drainage, debridement, device removal or obstruction relief whenever anatomy prevents cure by medicine alone.
02Indications, selection and cautionsWho may benefit, who needs urgent treatment and important alternatives.
Shock, meningitis, neutropenic sepsis or rapidly progressive invasive infection requires prompt syndrome-specific empirical intravenous therapy after feasible cultures.
Recent broad antibiotics, hospitalisation, overseas healthcare, prior resistant isolate, indwelling devices and structural disease increase the chance standard first-line therapy will fail.
Persistent bacteraemia, collection, obstruction, necrotic tissue or prosthetic infection indicates a mechanical focus and should not trigger serial antibiotic substitutions alone.
New kidney injury, cytopenia, liver-test abnormality, rash, diarrhoea, QT prolongation or neurotoxicity may represent antimicrobial harm requiring urgent review.
03Assessment before treatmentTests and checks that guide safe selection.
Read from the initial assessment onwards. Tests may run in parallel in urgent care; first-line, preferred, confirmatory, definitive and gold-standard labels appear only when the chapter explicitly states them.
- 01
Pretreatment cultures from blood and sourceFirst step - Why
- Identify the pathogen and susceptibility while preserving the opportunity to narrow.
- Interpretation and limitations
- Sampling quality and prior exposure determine sensitivity. Negative culture does not exclude infection, but repeated sterile results should prompt diagnosis and duration review.
- 02
Renal and hepatic function with weight - Why
- Select loading, maintenance and interval and establish a toxicity baseline.
- Interpretation and limitations
- Do not reduce an indicated loading dose reflexively for renal impairment; subsequent dosing and interval should follow current product, BNF, pharmacy and level guidance.
- 03
Susceptibility report and minimum inhibitory concentration context - Why
- Determine whether an organism is treatable with the drug at achievable exposure.
- Interpretation and limitations
- Interpret S, I and R categories using current laboratory definitions, dose exposure and infection site with microbiology; avoid reading the report as an automatic menu.
- 04
Imaging for source and penetration - Why
- Find collections, obstruction, bone, valve, central nervous system or prosthetic involvement that changes treatment.
- Interpretation and limitations
- Anatomical diagnosis can change agent, route and duration more than a small difference in inflammatory markers.
- 05
Therapeutic drug monitoring - Why
- Measure exposure for selected medicines with narrow therapeutic windows or variable clearance.
- Interpretation and limitations
- Timing must match the dosing strategy. Interpret concentrations with dose time, renal trend, infection target and pharmacy or microbiology advice.
04Treatment approachPreparation, options, escalation and aftercare.
01Empirical startWrite a defendable initial prescriptionFirst stepA bacterial syndrome requires treatment before organism and susceptibility are known.+
- 1Define source, severity and dangerous pathogens; collect high-value cultures and review allergies, weight, organ function, pregnancy, interactions and previous isolates.
- 2Select route, dose and spectrum from a current national or local syndrome guideline, give time-critical treatment promptly and document indication and review time.
- 3EscalationArrange source imaging or drainage, monitor response and toxicity, and communicate what result or clinical change will trigger escalation.
02Microbiology availableNarrow to a targeted regimenOrganism identification or susceptibility improves diagnostic certainty after empirical treatment has begun.+
- 1Decide whether the isolate is a pathogen, contaminant or coloniser by matching specimen quality and syndrome.
- 2Choose the safest narrow active agent with evidence and penetration for the site; adjust dose to current organ function and discontinue redundant components.
- 3Define total duration from source control and clinical response, arrange therapeutic monitoring and provide an oral plan when bioavailability and absorption are adequate.
03Apparent treatment failureDiagnose before broadeningFever, inflammation, bacteraemia or organ dysfunction persists beyond the expected response.+
- 1Verify administration, dose, access, adherence, absorption, renal clearance and culture interpretation, and repeat examination for deterioration.
- 2Search for an undrained focus, obstruction, foreign material, resistant pathogen, superinfection or a non-infectious mimic using targeted imaging and sampling.
- 3Change spectrum only for a supported hypothesis with microbiology input, and set a new review point and source-control plan.
05Regimens, contraindications and interactionsTreatment details and the circumstances that modify them.
Piperacillin with tazobactam
Common severe-infection adult regimen is 4.5 g intravenously every eight hours; some protocols use every six hours.Avoid after serious immediate beta-lactam hypersensitivity and reduce maintenance in renal impairment. Review sodium and potassium, liver tests and blood counts during longer exposure; piperacillin can delay methotrexate elimination, so check the full interaction plan and de-escalate promptly.
Gentamicin
Use a weight-based once-daily regimen, commonly 5 to 7 mg/kg intravenously, following the local nomogram and levels.Use actual or adjusted weight and current renal function, avoid unmonitored accumulation and minimise other nephrotoxic or ototoxic drugs. Measure correctly timed concentrations and assess hearing and vestibular symptoms; pregnancy, myasthenia gravis or neuromuscular blockade requires specialist risk assessment.
Vancomycin
Give a weight-based intravenous loading dose then individualise maintenance to renal function and measured exposure under local guidance.Individualise the loading and maintenance plan to weight, renal function, dialysis and locally adopted concentration or exposure targets. Infuse slowly to reduce infusion reaction and reassess renal and hearing toxicity, especially with aminoglycosides, piperacillin-tazobactam or other nephrotoxins.
06Complications, monitoring and follow-upAdverse effects, response and longer-term review.
- Review temperature, physiology, focal symptoms and organ function against an expected response time defined for the infection rather than waiting for CRP alone.
- Check administration records, vascular access and missed or delayed doses; a correct prescription that is not delivered cannot be judged microbiological failure.
- Recalculate renal dosing during acute change, dialysis or recovery, and arrange correctly timed levels for aminoglycosides, vancomycin and other monitored agents.
- Assess rash, diarrhoea, cytopenia, liver injury, neurological change, QT risk and drug interactions throughout treatment and after relevant prolonged courses.
- Document a 48-to-72-hour decision: stop, narrow, switch route, continue to a stated end date or revise after a new diagnosis or source-control finding.
07Special situationsVariants, exceptions and circumstances that change the usual approach.
Loading and maintenance differ
A full loading exposure may be needed in severe infection even with renal impairment, while later interval and maintenance must reflect clearance and measured levels.
Susceptible is not sufficient
Laboratory activity must be combined with site penetration, dose, infection burden, source control and clinical response before a targeted regimen is accepted.
Broadening can conceal the problem
Persistent fever from an abscess, thrombosis, drug reaction or malignancy will not improve simply because another antibiotic is added.
Duration starts with control
For many deep infections, the meaningful treatment clock depends on adequate drainage, debridement or bloodstream clearance rather than the first empirical dose.
08Common pitfallsFrequent interpretation and management errors.
- 01
Using the same empirical regimen for every source of severe infection.
- 02
Reducing all doses for renal impairment without considering the need for a loading dose.
- 03
Reading a susceptibility report without checking specimen quality and anatomical site.
- 04
Adding agents for persistent CRP while ignoring source control and clinical trajectory.
- 05
Allowing an empirical prescription to continue without an indication, review date and end plan.