Extended antibiotic prophylaxis in primary total knee arthroplasty: A narrative review of current evidence and controversies

Table 1 Characteristics of included observational cohort studies evaluating antibiotic prophylaxis beyond 24 hours in primary total knee arthroplasty and related arthroplasty cohorts (2000 to 2025)

Ref.	Design	Setting/population	Arthroplasty type	Sample size	High-risk definition	Extended prophylaxis regimen	Comparator	Follow-up/endpoint	Primary outcome measure	Key finding	Limitations
Inabathula et al[13]	Retrospective cohort	Single system or multi-site cohort; high-risk primary arthroplasty patients	Primary TKA and THA	2181 primary TKAs	Composite: Obesity, diabetes, immunosuppression, prior infection, other comorbidities	Oral cephalexin 7 days after standard perioperative IV prophylaxis	Standard prophylaxis; no extended oral course	90 days	90-day PJI	Lower 90-day PJI reported in high-risk patients receiving 7-day EOA	Retrospective; confounding by indication; risk definition not universally validated; applicability to routine-risk patients uncertain
Veltman et al[14]	Registry-based observational cohort	Dutch Arthroplasty Register; hospital-level protocol comparison	Primary THA and TKA	242179 hip and knee arthroplasties	Not risk-stratified (population-level analysis)	Compared single-dose vs 24-hour vs multiple-dose IV perioperative protocols	Other duration protocols	1 year (revision for infection in registry)	Revision for infection (proxy for PJI)	Similar revision-for-infection risk across single-dose and multiple-dose strategies	Registry endpoint is revision, not all PJIs; limited microbiology and adherence detail; confounding by hospital-level practice
Kheir et al[15]	Prospective cohort (American Association of Hip and Knee Surgeons award study)	High-risk primary arthroplasty cohort	Primary TKA and THA	3855 total (TKA + THA)	High-risk composite (study-specific criteria)	Oral cefadroxil 500 mg twice daily for 7 days after standard perioperative IV prophylaxis	Standard prophylaxis; no EOA	1 year	1-year PJI	PJI decreased in high-risk patients receiving 7-day EOA (as reported)	Non-randomized; potential selection bias; generalizability depends on local baseline infection rate and perioperative pathway
Carender et al[16]	Retrospective cohort	Morbidly obese primary arthroplasty patients	Primary TJA (445 TKAs reported)	445 TKAs	Morbid obesity (body mass index ≥ 40)	Oral antibiotics 7-14 days (agent not standardised across all centres)	Standard prophylaxis; no EOA	Early postoperative period	Early PJI; wound complications	No significant reduction in early PJI or wound complications in morbidly obese subgroup receiving EOA	Limited power for rare events; heterogeneity in antibiotic selection and adherence; retrospective confounding
Kelly et al[12]	Cohort study (reimplantation context)	Two-stage exchange for PJI; not primary TKA prophylaxis	Reimplantation after PJI	Not applicable to primary TKA	Not applicable	Prolonged oral antibiotics after reimplantation	No prolonged antibiotics	Recurrence window per study	Recurrent PJI; resistance profile	Higher proportion of resistant organisms among recurrent PJIs in prolonged-antibiotic group (as reported)	Not primary TKA; included for resistance plausibility and stewardship implications only
Bundschuh et al[17]	Retrospective cohort (single-surgeon experience)	Universal extended prophylaxis in primary and aseptic revision cases; single-surgeon protocol	Primary and aseptic revision TJA (approximately 1250 primary TKAs)	Approximately 1250 primary TKAs	Not risk-stratified (universal protocol)	Oral cefdinir 7 days; concomitant intra-wound vancomycin powder	Historical or contemporary non-EOA comparator	3 months	3-month PJI	Lower short-term PJI reported with protocol; attribution to EOA vs vancomycin powder unclear	Strong confounding by co-intervention (vancomycin powder); single-surgeon external validity limits; before-after design vulnerable to secular trend
Flynn et al[9]	Retrospective institutional policy-change cohort	High-volume institutional transition to universal EOA	Primary TJA (> 4500 TJAs including TKA)	> 4500 primary TJAs	High-risk subgroup analyses performed (definition study-specific)	Oral cephalexin 10 days for all primary arthroplasty cases	Pre-policy standard prophylaxis (≤ 24 hours)	90 days and 1 year	PJI at 90 days and 1 year	No reduction in PJI overall or in high-risk subgroups (as reported)	Before-after design; adherence and outpatient completion difficult to verify; baseline PJI already low limits absolute benefit
Dhodapkar et al[18]	Administrative database cohort	United States claims/administrative dataset	Primary TKA	> 860000 TKAs	Not specified	Postoperative day-0 only vs day-0 plus day-1 IV dosing	Shorter duration	30 days	30-day SSI	No difference in 30-day SSI with extension to postoperative day 1; antibiotic choice influenced outcomes	Addresses intraoperative antibiotic choice rather than post-discharge duration; administrative dataset; limited clinical granularity

EOA: Extended oral antibiotic; IV: Intravenous; PJI: Periprosthetic joint infection; SSI: Surgical site infection; TJA: Total joint arthroplasty; TKA: Total knee arthroplasty; THA: Total hip arthroplasty.

Table 2 Systematic reviews and meta-analyses evaluating prophylaxis duration in primary hip and knee arthroplasty (2000-2025)

Ref.	Scope and database	Arthroplasty types covered	Studies/patients included	Prophylaxis strategies compared	Infection outcome (PJI/SSI)	Certainty of evidence and key limitations
Siddiqi et al[20]	Systematic review and meta-analysis of total joint arthroplasty literature	TKA and THA	Multiple RCTs and observational studies; specific n per protocol in paper	Prolonged prophylaxis vs shorter protocols (varied definitions across included studies)	No reduction in SSI or PJI beyond initial dosing; certainty limited by low event rates	Low-certainty; heterogeneous protocols and eras; underpowered RCTs; rare-event bias; variable reporting of harms
Fernandes et al[19]	Systematic review and meta-analysis of primary hip and knee arthroplasty literature	TKA and THA	Nearly 300000 TKAs as stated in manuscript	Single-dose vs extended prophylaxis (definition of extended varies across included cohorts)	Lower PJI odds with single-dose vs extended (reported odds ratio = 0.78; 95%CI: 0.63-0.98); SSI trend favoured single-dose but did not reach significance	Meta-analysis inherits confounding from observational studies; definition of extended differs across cohorts; harms and resistance variably captured across included studies

PJI: Periprosthetic joint infection; RCT: Randomised controlled trial; SSI: Surgical site infection; TKA: Total knee arthroplasty; THA: Total hip arthroplasty.

Table 3 Narrative reviews and non-primary evidence sources [narrative reviews included for contextual synthesis (not primary evidence)]

Ref.	Scope	Arthroplasty focus	Principal themes addressed	Contribution to evidence base and key caveats
Hong et al[8]	Narrative review	Total knee arthroplasty (primary or revision focus)	EOA strategies; controversy around post-discharge prophylaxis; antibiotic stewardship concerns	Highlights lack of randomised evidence for EOAs; draws attention to stewardship implications; not primary evidence and dependent on quality of included studies reviewed

EOA: Extended oral antibiotic.

Table 4 Evidence synthesis: Prophylaxis strategies, efficacy, and stewardship implications in primary total knee arthroplasty (restructured evidence synthesis comparing prophylaxis duration strategies – effect on infection outcomes and stewardship implications)

Ref.	Prophylaxis strategy	Typical regimen	Evidence sources in this review	Efficacy: Standard-risk primary TKA	Efficacy: High-risk subgroups	Harms and stewardship concerns	Practice interpretation (based on cited evidence)
Veltman et al[14]	Single pre-incision dose only	Cefazolin or equivalent within 60 minutes pre-incision; no postoperative doses	Registry comparisons; meta-analyses; guideline frameworks	No worse outcomes than longer courses in large registry and pooled analyses; single-dose performed as well or better than extended protocols in one meta-analysis (odds ratio = 0.78 favouring single-dose vs extended)	High-risk effects not established by randomised data; not specifically tested in defined high-risk subgroups in these analyses	Minimises antibiotic exposure; lowest selective pressure for resistance; lowest drug-related adverse event burden	Reasonable default in many guideline frameworks; strongest alignment with stewardship goals; supported by large-scale observational and registry evidence
Fernandes et al[19]	Standard prophylaxis ≤ 24 hours	Pre-incision dose plus limited postoperative dosing to complete 24 hours	Large observational datasets; orthopaedic society positions (American Academy of Orthopaedic Surgeons and American Association of Hip and Knee Surgeons-type frameworks as referenced in manuscript)	No consistent reduction beyond this window compared with single-dose in large observational datasets	No consistent evidence that extending past 24 hours adds benefit beyond perioperative coverage	Low but non-zero adverse event risk; still limited total exposure; modestly greater selective pressure than single-dose	Dominant orthopaedic standard; supported by society positions; appropriate for most primary TKA patients
Veltman et al[14]; Dhodapkar et al[18]	IV prophylaxis extended beyond 24 hours (multi-day IV)	Continued IV dosing for multiple postoperative days	Large registry and administrative analyses	No reduction in infection outcomes in population-level analyses	Not clearly beneficial; no data specifically supporting this approach in high-risk groups	Line-related risks if IV access prolonged; selection pressure; nursing and cost burden	Not supported for routine primary TKA; duration does not appear to be the modifiable driver relative to antibiotic choice and perioperative pathway optimisation
Inabathula et al[13]; Kheir et al[15]	Selective EOAs in high-risk patients	7 days of oral cephalexin or cefadroxil (typically 500 mg twice daily) after standard perioperative IV prophylaxis; applied only to defined high-risk patients	Two influential comparative cohort studies	Not intended for standard-risk patients in these studies; no data support routine use outside high-risk criteria	Reported reduction in early or 1-year PJI in study-defined high-risk cohorts; causality uncertain given non-randomised designs and confounding by indication	Potential for resistance selection; outpatient adherence variability; Clostridioides difficile risk biologically plausible even if not observed in these cohorts	Hypothesis-generating only; if used, should be restricted to protocolised, clearly defined high-risk criteria with stewardship oversight and robust surveillance; randomised evidence is absent
Flynn et al[9]; Bundschuh et al[17]	Universal EOA for all primary arthroplasty patients	7-10 days of oral antibiotics post-discharge for all patients regardless of risk	Large institutional policy-change cohort); universal single-surgeon protocol	No consistent reduction in 90-day or 1-year PJI when applied broadly	No benefit detected even in high-risk subgroup analyses in at least one large cohort; attribution problem compounded by co-interventions	Ecologic resistance risk at population scale; overtreatment of genuinely low-risk patients; adds unnecessary antibiotic exposure across the majority who would not benefit	Not supported by current evidence; routine universal EOA is difficult to justify given low baseline PJI rates and stewardship principles; should not be extrapolated from single-surgeon bundle studies
Bundschuh et al[17]	EOA plus co-interventions (e.g., intra-wound vancomycin powder)	Oral antibiotics paired with local antibiotic adjuncts; protocol varies by institution	Single-surgeon protocol combining cefdinir with vancomycin powder	Any observed reduction in infection rate cannot be attributed to EOA alone given the bundled design	Same attribution problem; high-risk-specific effect cannot be isolated from the bundle	Risk of confounding and inappropriate extrapolation to EOA as an independent intervention	Results should be interpreted as bundle effects only; not proof of EOA efficacy; co-intervention designs are insufficient to inform isolated EOA policy decisions
Kelly et al[12]	Prolonged antibiotics in non-primary TKA settings (resistance signal)	EOAs after two-stage exchange reimplantation	Reimplantation cohort; included for stewardship implications not primary prophylaxis efficacy	Not applicable to primary TKA prophylaxis efficacy	Not applicable	Strong signal for selection of resistant organisms among failures in prolonged-antibiotic group; clinically meaningful change in microbiology of recurrent PJI	Supports stewardship caution: Prolonged antibiotic exposure can meaningfully alter the microbial landscape; findings are generalisable as a harm signal even if not directly applicable to primary prophylaxis

EOA: Extended oral antibiotic; IV: Intravenous; PJI: Periprosthetic joint infection; TKA: Total knee arthroplasty.