A scoping review was undertaken to systematically map and characterise the published literature on AMR in KSD across both procedural and non-procedural clinical contexts, and to identify AMS-relevant signals within this evidence base. A protocol was developed a priori to define the review objectives, research questions, eligibility criteria, clinical context categories, and data-charting framework. Formal prospective registration was not pursued because the review was exploratory in scope and not designed as a systematic review of intervention effectiveness; moreover, scoping reviews are not routinely eligible for registration in all review registries. Nevertheless, the protocol, screening framework, and review log were retained to support methodological transparency and reproducibility, and are available from the authors upon reasonable request. Reporting was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR). The study selection process is summarised in Figure 1.

Figure 1. PRISMA-ScR flow diagram of study selection for the scoping review. PubMed records identified (n = 71), duplicate records removed (n = 8), and inaccessible records removed before screening (n = 5) resulted in records screened (n = 58). Following eligibility assessment, studies assessed for full-article screening (n = 35), ineligible studies excluded (n = 4), and studies included in the review (n = 31) are shown.

A structured literature search was conducted in PubMed to identify peer-reviewed studies reporting AMR in the context of KSD and its medical or procedural management. PubMed was selected as the primary database because of its strong coverage of biomedical, clinical, and microbiological literature relevant to urolithiasis, urinary tract infection, and antimicrobial resistance. The final PubMed search was conducted on 12 January 2026. The search was restricted to studies published between 1 January 2016 and 31 December 2025 and to articles published in English.

The search strategy combined Medical Subject Headings (MeSH) and free-text terms relating to three core concepts: kidney stone disease, stone-related procedures or therapies, and bacterial antimicrobial resistance. The final PubMed search string was as follows:

(“Urolithiasis” [Mesh] OR “Kidney Calculi” [Mesh] OR nephrolithiasis [tiab] OR “kidney stone*” [tiab] OR “renal stone*” [tiab]) AND (“Nephrostomy, Percutaneous” [Mesh] OR “Lithotripsy” [Mesh] OR “Ureteroscopy” [Mesh] OR “Drug Therapy” [Mesh] OR “Anti-Bacterial Agents” [Mesh] OR PCNL [tiab] OR SWL [tiab] OR RIRS [tiab] OR ureteroscop* [tiab] OR “medical expulsive therapy” [tiab] OR tamsulosin [tiab]) AND (“Drug Resistance, Bacterial” [Mesh] OR “antimicrobial resistance” [tiab] OR “antibiotic resistance” [tiab] OR MDR [tiab] OR “multidrug-resistant” [tiab]).

The search was designed to capture studies addressing both direct microbiological resistance outcomes and clinically relevant AMR-related findings arising within stone-associated pathways of care.

Studies were considered eligible if they were peer-reviewed primary or review publications reporting AMR-relevant microbiological findings in the context of KSD, stone-associated urinary tract infection (UTI), or the clinical management of patients with urinary stones. Eligible AMR-related outcomes included antimicrobial susceptibility testing results, defined resistance phenotypes such as extended-spectrum beta-lactamase (ESBL)-producing Enterobacterales, fluoroquinolone resistance, multidrug resistance (MDR), carbapenem resistance, or other clinically relevant resistance patterns; resistance-associated risk factors in stone populations; and stewardship-relevant management findings linked to resistance risk.

Studies were excluded if they did not report extractable AMR-related microbiological or susceptibility data, were not relevant to KSD or stone-associated infection pathways, involved non-human populations, were published outside the predefined date range, were not available in English, or lacked accessible full text such that key data could not be reliably extracted. Preprints and non-peer-reviewed materials were also excluded.

Titles and abstracts were screened against the predefined eligibility criteria, and potentially relevant records were subsequently assessed at full-text level. Screening and full-text review were undertaken independently by two reviewers. Discrepancies were resolved through discussion and consensus. Where full-text articles were not accessible through institutional subscriptions, attempts were made to obtain them through interlibrary request and, where feasible, by contacting the corresponding authors. Records that remained unobtainable after these efforts were excluded and documented as full-text unavailable.

Data were charted using a standardised extraction form developed a priori and piloted before full extraction. The extraction framework captured bibliographic details (author, year, journal), study design, population characteristics, clinical context, procedure or exposure category, microbiological sampling source, organism profile, reported resistance phenotype(s) and/or susceptibility findings, and stewardship-relevant management observations. Sampling sources included, where applicable, midstream urine, bladder urine, renal pelvic urine, stone culture, blood culture, irrigation fluid, and device-associated samples. When more than one sampling source was reported within a single study, findings were charted separately to preserve within-study microbiological discordance.

Each included study was assigned to a primary clinical context category according to predefined operational definitions. These categories were: percutaneous nephrolithotomy (PCNL); ureteroscopy (URS), including ureteroscopy-based interventions such as flexible ureteroscopy and ureteroscopic lithotripsy; retrograde intrarenal surgery (RIRS); decompression, defined as urgent drainage procedures such as ureteric stent placement and/or percutaneous nephrostomy when these constituted the primary clinical context; device-centred pathways, defined as studies in which indwelling urinary devices represented the principal exposure or microbiological context; and non-procedural contexts, including stone populations studied outside a specific intervention pathway, such as observational epidemiological studies, microbiological investigations, genomic studies, and narrative reviews. Where a single dominant exposure or procedural context could not be determined, studies were classified as mixed or unclear and were summarised descriptively.

In keeping with scoping review methodology, formal risk-of-bias assessment was not used as a basis for study exclusion. However, methodological limitations and internal validity concerns were documented during data charting in order to support cautious interpretation of the evidence base. Where appropriate, design-specific critical appraisal tools were applied to contextualise study quality: the Newcastle-Ottawa Scale (NOS) for observational studies, RoB 2 for randomised trials, and the Joanna Briggs Institute (JBI) Critical Appraisal Checklist for case reports. Appraisal findings were summarised descriptively and were used to inform interpretation rather than to determine eligibility. Data charting was performed independently by two reviewers using the predefined extraction form. Any discrepancies in extraction or classification were resolved through discussion and consensus.

The evidence was synthesised using structured narrative methods supported by tabulation. The distribution of included studies across clinical context categories was summarised descriptively, and key microbiological and AMR-related findings were mapped according to procedural or non-procedural context, sampling source, organism profile, and reported resistance phenotype. Quantitative meta-analysis was not undertaken because of substantial heterogeneity in study design, populations, microbiological methods, sampling frameworks, and outcome definitions.

Stewardship-related content was coded using an a priori analytical framework. Studies were classified as demonstrating explicit AMS framing when they directly used stewardship language or made clearly stewardship-oriented recommendations, such as antibiotic-sparing strategies, minimisation of unnecessary treatment duration, avoidance of routine broad-spectrum prescribing, or explicit calls for stewardship-guided practice. Studies were classified as demonstrating implicit AMS framing when they contained stewardship-aligned recommendations without explicitly using AMS terminology, for example, culture-guided therapy, risk-stratified antimicrobial selection, avoidance of agents with high local resistance rates, or recommendations to tailor prophylaxis according to patient or procedural risk. Where relevant, the dominant stewardship message type, including duration minimisation, culture-guided selection, or prophylaxis individualisation, was also recorded to support structured interpretation of stewardship signals across the included literature. AMS coding was performed independently by two reviewers using the a priori framework. Any disagreements in coding were resolved through discussion and consensus.

A total of 31 studies met the eligibility criteria and were included in the review. The included studies were published between 2016 and 2025, with one study published in 2016 and one in 2017, two studies in 2018 and 2019, five studies in 2020, four in 2021, five in each of 2022, 2023, and 2024, and one study in 2025. Study-level characteristics, clinical or procedural context, and extracted microbiological findings are presented in supplementary data (Table S1). Patient subgroup and clinical scenario data extracted from selected studies are presented in Supplementary data (Table S2), and stewardship-related extraction fields are presented in Supplementary data (Table S3).

Across the 31 included studies, clinical and procedural contexts were grouped into five categories for descriptive synthesis: non-procedural/epidemiological contexts, PCNL, drainage/device-related pathways, URS/RIRS/FURS-related pathways, and TUL. Non-procedural or epidemiological contexts accounted for 15 studies, PCNL for 9 studies, drainage/device-related contexts for 4 studies, URS/RIRS/FURS-related contexts for 2 studies, and TUL for 1 study. These distributions are shown in Figure 2, while study-level context classifications and extracted microbiological observations are presented in Supplementary data (Table S1).

Figure 2. Publication-year distribution of the 31 included studies according to primary kidney stone disease management context. Each stacked bar represents the number of studies published in a given year (2016-2025), stratified by the predefined clinical or procedural context used for descriptive synthesis: non-procedural/epidemiology, percutaneous nephrolithotomy (PCNL), drainage/device-related pathways, ureteroscopy/retrograde intrarenal surgery/flexible ureteroscopy (URS/RIRS/FURS), and transureteral lithotripsy (TUL). The figure illustrates the predominance of non-procedural/epidemiological and PCNL-focused studies within the evidence base, with comparatively limited representation of other endourological contexts.

Non-procedural and epidemiological studies were represented throughout the review period and constituted the entirety of the 2022 publication set. By contrast, PCNL-focused studies were concentrated in the middle of the review window and were most prominent in 2020. Drainage- and device-related studies appeared intermittently across 2020, 2021, 2023, and 2025. URS/RIRS/FURS-related contexts were represented in 2018 and 2021, while the sole TUL-context study was published in 2023. These patterns indicate that the published AMR literature in KSD has been weighted towards observational and PCNL-associated contexts, with more limited representation of other endourological pathways.

Resistance phenotypes were categorised at study level and are summarised in Figure 3. Across the included studies, MDR was reported in 10 studies and ESBL-associated resistance in 9 studies. Fluoroquinolone-resistant phenotypes were reported in 3 studies, XDR in 1 study, and CRE-associated resistance in 1 study. In addition, 7 studies were classified within an “other resistance/stewardship signal” category, reflecting resistance findings that did not fall within the predefined MDR, ESBL, fluoroquinolone-resistant, XDR, or CRE groupings, or studies in which resistance-related findings were reported primarily through stewardship-relevant clinical signals. These phenotype distributions are further detailed in Figure 3, and organism-level and susceptibility-related findings are presented in Supplementary data (Table S1 and Table S2).

Figure 3. Publication-year distribution of reported antimicrobial resistance phenotypes across the included kidney stone disease studies. Each stacked bar represents the number of studies published in a given year (2016-2025) reporting multidrug resistance (MDR), extended-spectrum beta-lactamase (ESBL)-associated resistance, fluoroquinolone-resistant phenotypes, extensively drug-resistant (XDR) phenotypes, carbapenem-resistant Enterobacterales (CRE)-associated resistance, or other resistance/stewardship-related signals. The figure demonstrates that MDR and ESBL-associated resistance dominated the AMR literature in kidney stone disease, while more advanced or less frequently reported phenotypes, including XDR and CRE, appeared only sporadically within the review window.

Across the evidence base, resistant Gram-negative organisms were prominent, particularly Escherichiacoli, Klebsiellapneumoniae, Proteusmirabilis, and other Enterobacterales. Several studies also reported clinically important Gram-positive and mixed microbiological findings, including Enterococcusfaecalis, as well as fungal isolates in selected contexts. MDR and ESBL-producing organisms were the most frequently recurring resistance patterns, while fluoroquinolone resistance, XDR, and carbapenem resistance were reported less commonly but remained clinically important where identified.

Temporally, MDR was reported in studies published in 2016, 2018, 2019, 2020, 2022, 2024, and 2025, while ESBL-associated resistance was reported from 2018 onwards and appeared repeatedly in studies published between 2020 and 2024. Fluoroquinolone-resistant phenotypes were reported in 2021, 2023, and 2024, XDR in 2020, and CRE-associated resistance in 2023. Overall, the phenotype profile suggests that the AMR literature in KSD has been dominated by MDR- and ESBL-related concerns, with smaller but clinically significant signals relating to more advanced resistance phenotypes. Although some individual studies reported study-specific proportions, including 33% MDR isolates in one microbiological cohort, 44% MDR isolates in a mixed stone-procedure cohort, and approximately 19.6% ESBL Enterobacterales in a febrile upper-tract stone cohort, heterogeneity in patient populations, sampling sources, microbiological methods, and resistance definitions was too substantial to support a single meaningful cross-study range estimate for resistance phenotype burden.

Patient subgroup characteristics and clinical scenarios were variably reported across the included literature. Extracted examples are presented in Supplementary data (Table S2) and include sex distribution, age structure, clinical setting, urolithiasis with UTI or other complicated infection presentations, prior nephrolithiasis history, prior antibiotic exposure, obstructive presentations requiring urgent decompression, and the presence and duration of indwelling urinary devices. Because Supplementary data (Table S2) contains selected extracted examples rather than complete subgroup extraction for all 31 studies, these findings should be interpreted as illustrative rather than exhaustive.

Within the studies for which subgroup-level information was available, higher AMR burden or more resistant phenotypes were reported in several recurring scenarios. These included older age and male sex in some observational cohorts, prior kidney stone history, recurrent UTI, prior antibiotic exposure, indwelling ureteric stents or percutaneous nephrostomy catheters, obstruction requiring urgent decompression, and large or complex stone burden in PCNL contexts. In selected studies, recurrent or device-associated infection pathways were linked to ESBL-producing organisms, carbapenem-resistant organisms, XDR phenotypes, or reduced susceptibility to commonly used empirical agents. These subgroup and scenario-level resistance signals are summarised in Supplementary data (Table S2).

Microbiological sampling strategies varied across the included studies and are presented in Supplementary data (Table S1). Sampling sources included pre-operative midstream urine and bladder urine cultures and, in procedural contexts, intra-operative or device-associated sources such as renal pelvic urine, stone cultures, irrigation-fluid isolates, and nephrostomy- or stent-related specimens. Where more than one sampling source was reported within a single study, findings from each source were extracted separately in order to preserve within-study microbiological discordance.

Several studies, particularly in PCNL-related contexts, reported discordance between conventional urine cultures and renal pelvic urine or stone cultures. In these studies, intra-operative or stone-associated cultures identified additional or different organisms and, in some cases, were described as better predictors of post-procedural infection or sepsis than pre-operative urine cultures alone. These findings support the view that reliance on bladder or midstream urine alone may underestimate microbiological burden in selected high-risk KSD pathways. Sampling contexts and extracted microbiological observations are presented in Supplementary data (Table S1) and summarised in Supplementary data (Table S1 and Table S2).

Stewardship-related content varied across the included studies. Explicit AMS framing was identified in 5 of 31 studies, while the remaining 26 studies were classified as containing implicit or non-explicit stewardship-relevant reporting. The studies with explicit AMS framing addressed antibiotic-sparing aminoglycoside strategies, duration minimisation after urgent decompression, prophylaxis individualisation in RIRS, short-course culture-directed prophylaxis in PCNL, and non-antibiotic adjunctive approaches within UTI-prone populations that included nephrolithiasis. These stewardship-related findings are presented in Supplementary data (Table S3).

Although most studies did not use explicit AMS terminology, many nevertheless contained stewardship-relevant observations, including culture-guided antimicrobial selection, resistance-aware empirical prescribing, avoidance of unnecessary prolonged antibiotic exposure, and consideration of local susceptibility patterns. Accordingly, stewardship signals were not confined to the explicitly framed studies alone, but were distributed more broadly across the literature. These broader stewardship-related findings are presented in Supplementary data (Table S3) and summarised alongside resistance-related findings in Supplementary data (Tables S1-S3).

This scoping review synthesised evidence published between 2016 and 2025 on AMR in KSD across procedural and non-procedural care pathways. Taken together, the findings presented in Supplementary data (Tables S1-S3) support a clear overarching interpretation: KSD care frequently operates within an AMR-prone clinical ecology in which recurrent urinary tract instrumentation, repeated healthcare exposure, intermittent or prolonged device use, and cumulative antibiotic exposure interact with the capacity of stones to harbour microorganisms and biofilm-associated communities. Within this framework, resistance phenotypes do not appear as isolated microbiological observations, but as recurring features of a broader and clinically important infection ecology.

Across the included studies, resistant Gram-negative uropathogens were prominent, with repeated reporting of MDR and ESBL-producing Enterobacterales, alongside less frequent but clinically consequential signals relating to carbapenem resistance and XDR phenotypes. These findings, as presented in Supplementary data (Table S1 and Table S2), suggest that AMR in stone-associated populations is not merely descriptive, but likely to influence empirical adequacy, therapeutic escalation, and the likelihood that post-procedural infection becomes difficult to manage. At the same time, stewardship framing was limited, with only a minority of studies explicitly presenting their findings through an AMS lens, as presented in Supplementary data (Table S3). This is important because where antibiotic indication, timing, duration, and de-escalation are poorly reported, it becomes difficult to distinguish clinically necessary therapy from avoidable exposure that may amplify selection pressure.

When interpreted alongside contemporary guidance, the overall direction of the evidence is broadly concordant with the core principles of current European Association of Urology recommendations, including urine assessment before intervention, treatment of active UTI before elective stone surgery, and procedure-appropriate peri-operative prophylaxis, alongside avoidance of unnecessary antibiotic use in lower-risk contexts such as SWL when baseline urine is sterile [4]. However, the present review also highlights two clinically important implementation tensions. First, guideline-concordant prophylaxis may still be inadequate where local resistance prevalence is high or where patient-level risk factors, such as prior colonisation, recent antibiotic exposure, or recurrent instrumentation, are not sufficiently incorporated into decision-making. Second, reliance on midstream or bladder urine culture alone may underestimate upper-tract or stone-associated microbial reservoirs, thereby creating a mismatch between apparent pre-operative reassurance and actual post-procedural risk. These tensions position resistance in stone pathways as both a patient safety issue, when inadequate antimicrobial coverage contributes to infectious deterioration, and a stewardship issue, when broad or prolonged antibiotic exposure is used without clear additional benefit. Interpretation is further complicated by the historical evolution of severe infection definitions, with increasing standardisation under Sepsis-3 criteria over time [14].

The distribution of evidence presented in Supplementary data (Table S1) indicates that resistance burden and infection risk are not evenly distributed across KSD pathways. PCNL was a major focus of the included literature and appears consistently associated with comparatively higher infectious risk than many other common stone procedures. This is biologically plausible, given that renal puncture and tract dilation, larger irrigation volumes, longer procedural duration, and episodes of raised intrarenal pressure may facilitate bacterial mobilisation, translocation, and systemic inflammatory responses [15]. Within multiple cohorts, positive pre-operative urine culture has emerged as a reproducible predictor of post-PCNL infectious complications, including systemic inflammatory response syndrome and sepsis [16][17]. Importantly, susceptibility profile also appears to matter: MDR bacteriuria before PCNL has been associated with greater odds of post-operative infectious complications, suggesting that culture positivity alone is insufficient without corresponding attention to resistance phenotype [8]. Procedural complexity, including larger or staghorn stone burden, multiple access tracts, longer operative duration, and higher intrapelvic pressures, has also been linked to post-procedural infection, although confounding by underlying case complexity remains difficult to exclude fully [15][16].

By contrast, URS and RIRS generally appear to carry lower absolute rates of severe infectious complications than PCNL, but their stewardship relevance may still be substantial because these procedures are performed at high volume. Infectious morbidity after these procedures may be influenced by obstruction, pre-existing stents, prolonged operative time, and high-pressure irrigation, while indwelling stents may provide surfaces for biofilm persistence and may complicate interpretation of urine culture findings [5]. Even when the per-case probability of severe infection is relatively low, high procedural frequency combined with repeated peri-operative or post-procedural prescribing may produce meaningful antibiotic selection pressure at population level.

Device-associated pathways represent a further concentration point for risk. Ureteric stents, nephrostomy tubes, and related urinary devices may promote colonisation, recurrent infection, and repeated antibiotic exposure, thereby reinforcing selective pressure over time [5]. The evidence presented in this review warrants cautious interpretation because device exposure is often correlated with more complex disease, greater obstruction burden, and higher baseline morbidity. Nonetheless, the repeated convergence of indwelling devices, recurrent antibiotic use, and resistant microbiology across multiple study contexts is sufficiently consistent to justify explicit pathway-level stewardship attention.

Obstruction warrants particular emphasis because it alters both the clinical urgency and the stewardship time horizon. In obstructed infected systems, urgent decompression and antimicrobial therapy are essential for patient safety [4][7]. However, empirical treatment decisions are often made before complete susceptibility data are available, and antibiotic duration may extend unnecessarily when definitive stone management is delayed. Although the present review does not permit quantitative estimation of optimal treatment duration, owing to heterogeneity and limited prospective evidence, the mapped findings support a cautious clinical principle: once source control has been achieved and the patient is clinically improving, antibiotic exposure should be actively reviewed, narrowed where microbiology permits, and not continued by default as an indefinite bridge to definitive treatment.

One of the most consistent themes across the included studies is that microbiological sampling strategy materially influences both risk estimation and downstream antibiotic decisions. Discordance between pre-operative midstream or bladder urine cultures and intra-operative cultures obtained from renal pelvic urine or stone fragments has been repeatedly documented, particularly in PCNL-related cohorts [11][12]. This finding is clinically important because it supports the concept that stones may act as protected microbial reservoirs, including reservoirs of resistant organisms that are not adequately captured by conventional lower-tract urine sampling [11][12].

The biological plausibility of this phenomenon is considerable. Bacteria may persist within the stone matrix or within biofilm-like communities adherent to stone surfaces or adjacent devices, thereby reducing antimicrobial penetration and enabling persistence despite apparently appropriate therapy [11]-[13]. In practical terms, this means that reliance on bladder or midstream urine alone may underestimate microbiological burden before higher-risk procedures [11][12]. It also provides a potential explanation for situations in which post-operative deterioration prompts reactive broadening of antibiotics despite apparently reassuring pre-operative urine results.

From a stewardship perspective, the value of expanded microbiological sampling lies not in sampling intensity alone, but in whether additional cultures are linked to an actionable management plan. A risk-aligned approach would reserve renal pelvic urine or stone culture sampling for patients and procedures in which the incremental diagnostic yield is most likely to alter treatment, such as PCNL in the context of suspected infection stones, recurrent infection, prior severe infectious complications, indwelling devices, or complex obstruction [11][12]. In such settings, intra-operative culture results may support more targeted therapy and earlier de-escalation, potentially reducing unnecessary broad-spectrum exposure.

A further interpretive challenge is that the current literature rarely provides consistent diagnostic criteria for distinguishing colonisation from true infection in stone-associated pathways. Across many included studies, microbiological positivity was reported without uniform reference to colony count thresholds, symptom burden, inflammatory markers, or clear clinical correlation, particularly in the setting of indwelling devices and upper-tract sampling. This definitional ambiguity is important for stewardship because microbiological detection alone may not reliably indicate clinically significant infection, yet may still prompt escalation, prolongation, or unnecessary broadening of antibiotic therapy. More standardised reporting that integrates microbiological findings with symptomatology and clinical severity would therefore improve both interpretation of the evidence base and the stewardship relevance of future studies.

Nevertheless, the present review also highlights important limitations in the underlying evidence base. Definitions of sampling source, timing, laboratory methods, and contamination control vary across studies, and infection endpoints are not uniformly defined. In addition, older and newer studies may not be directly comparable because severe infectious outcomes have been framed using evolving sepsis definitions [14]. These sources of heterogeneity reduce the precision with which the incremental predictive value of stone or renal pelvic cultures can be estimated, even where the biological rationale for culture discordance is compelling.

The findings presented in Supplementary data (Tables S1-S3) suggest that stewardship opportunities in KSD are largely pathway-based rather than confined to antibiotic selection alone. Several practical leverage points emerge from the mapped evidence.

First, microbiological risk stratification should be made more explicit within stone care pathways. The review indicates that higher resistance burden is repeatedly associated with recurrent stone disease, prior antibiotic exposure, indwelling devices, obstruction, and more complex interventions such as PCNL, as summarised in Supplementary data (Table S2). These variables are generally available during routine pre-operative assessment and could be incorporated into structured risk-assessment tools to guide culture acquisition, review of prior microbiology, and alignment of prophylaxis with local susceptibility patterns.

Second, prophylaxis should remain procedure-appropriate but increasingly informed by local resistance epidemiology and patient-specific risk. Current guidance appropriately discourages unnecessary antibiotics for SWL in patients with sterile urine [4]. However, the phenotype distributions mapped in this review indicate that ESBL and MDR signals may materially affect empirical reliability in selected settings [7][8]. This creates a clear tension: routine broad-spectrum prophylaxis for all patients would be stewardship-poor and likely harmful, yet failure to incorporate local epidemiology and patient risk may lead to inadequate coverage in high-risk subgroups. A defensible position is therefore one of stratified prophylaxis, integrating procedure type, patient-level risk factors, and local susceptibility data, with explicit expectations for narrowing and discontinuation once culture results and clinical response permit.

Third, antibiotic duration should itself be treated as a stewardship endpoint. In stone pathways, avoidable exposure often accumulates through extended post-operative courses, repeated attempts to sterilise urine, or continuation of antibiotics during delays to definitive source control. Evidence relating to PCNL prophylaxis suggests that extended regimens beyond short peri-operative strategies may offer limited additional benefit while increasing exposure burden [6]. Where prolonged therapy is used, the rationale should be clearly linked to microbiology, source control status, and clinical trajectory rather than applied routinely.

Fourth, device management should be understood as a stewardship intervention. Indwelling devices may sustain colonisation and biofilm persistence and may trigger repeated antibiotic prescribing [5]. Reducing device dwell time where feasible, expediting definitive stone clearance, and avoiding inappropriate treatment of asymptomatic colonisation may therefore reduce antibiotic exposure without compromising patient safety, provided that microbiology is interpreted within the broader clinical context.

Finally, the present review suggests that the KSD literature remains relatively underdeveloped in relation to stewardship evaluation. As presented in Supplementary data (Table S3), only a minority of included studies used explicit stewardship framing, and many did not report antibiotic indication, duration, or de-escalation with sufficient clarity to permit meaningful assessment. Future studies should therefore treat antibiotic exposure as an outcome in its own right, alongside clinical endpoints and resistance phenotypes, and should report microbiological sampling strategies transparently to enable comparison of discordance and reservoir-related findings across settings.

Several limitations constrain interpretation of the present evidence base and should be regarded as priorities for future work. First, much of the available literature is retrospective and single-centre, which limits generalisability and increases susceptibility to bias and confounding. Second, substantial heterogeneity exists in procedure categorisation, microbiological sampling strategy, laboratory methodology, timing of cultures, and definitions of infectious outcomes. This heterogeneity complicates cross-study comparison and precludes robust quantitative estimation of procedure-specific risk or the incremental value of specific sampling sources.

An additional methodological limitation is that the current evidence base is derived almost exclusively from conventional aerobic culture methods. As a result, the possible contribution of anaerobic bacteria, fastidious organisms, and other atypical pathogens to stone-associated infection ecology remains poorly characterised. This represents an important gap because reliance on standard aerobic culture alone may underestimate the full microbiological complexity of infected stones, biofilm-associated reservoirs, and discordant culture states. Future studies should therefore consider broader microbiological detection frameworks, including methods capable of identifying anaerobic and non-conventionally recovered organisms.

Confounding also remains a major challenge. Prior antibiotic exposure, indwelling devices, obstruction severity, stone burden, and comorbidity often cluster together and are not consistently adjusted for in the included studies. As a result, it remains difficult to disentangle whether resistant microbiology is driven primarily by healthcare exposure and cumulative antibiotic use, by particular procedural characteristics, or by underlying patient phenotype. Resistance phenotyping is also reported inconsistently, and longitudinal follow-up across recurrent episodes of care is uncommon despite the chronic and relapsing nature of KSD.

A further gap lies in the limited prospective evaluation of stewardship interventions within stone pathways. Although many studies describe resistance patterns, comparatively few examine pathway modifications such as structured microbiological risk stratification, risk-aligned culture acquisition, shortened antibiotic durations, or device dwell-time governance. Fewer still link such interventions to both patient outcomes and downstream resistance burden. High-value future work would include multicentre prospective cohorts using harmonised infection definitions, standardised sampling frameworks, structured capture of antibiotic exposure, and longitudinal follow-up across recurrent interventions. Such designs would improve causal inference and generate evidence that is more directly translatable to clinical stewardship policy and service redesign.