Nosocomial pneumonia is a frequent healthcare-associated infection, especially among critically ill patients. Its clinical consequences include respiratory failure, prolonged mechanical ventilation, organ dysfunction, extended hospitalization, mortality and increased antimicrobial consumption. HAP generally develops at least 48 hours after hospital admission and is not incubating at admission; VAP arises more than 48 hours after endotracheal intubation. These syndromes differ in baseline severity and should not automatically be combined when evaluating clinical outcomes. The World Health Organization places carbapenem-resistant Acinetobacter baumannii, resistant Enterobacterales and resistant Pseudomonas aeruginosa among important antimicrobial-resistance threats [1]. These organisms are particularly relevant to hospital pneumonia because resistance can render standard empirical choices inactive. ATS/IDSA guidance recommends syndrome- and setting-specific empirical therapy informed by local antibiograms, previous microbiology and individual risk factors [2]. Initial empirical antimicrobial treatment is often evaluated retrospectively after cultures and susceptibility tests become available. An empirical regimen can be microbiologically inactive even when selected according to reasonable clinical expectations; conversely, a broad regimen may cover the pathogen while exposing the patient to unnecessary agents. Consequently, microbiological adequacy, guideline concordance, timely administration, optimal dosage and clinical success are related but nonidentical constructs. Large observational datasets show that treatment mismatch varies by resistance phenotype and pneumonia syndrome [3]. Contemporary European surveillance also confirms substantial geographical variability in VAP pathogens and outcomes [4]. This variation suggests that a single pooled estimate may conceal clinically important institutional differences. This review addresses three questions: how frequently is initial empirical treatment microbiologically inappropriate; which resistance and clinical factors are associated with mismatch; and what consequences for mortality, clinical recovery and resource use have been reported? Unlike mortality-only meta-analyses, the quantitative component centers on inappropriate-coverage frequency and separately describes intervention outcomes. 1.1. Objectives • To describe resistance patterns and key microorganisms reported in HAP/VAP cohorts. • To quantify inappropriate initial empirical treatment frequency in cohorts with extractable counts. • To synthesize evidence on mortality, treatment failure, hospital stay and cost. • To evaluate evidence for rapid microbiological diagnosis and antimicrobial stewardship interventions. • To identify clinical and methodological explanations for variability between studies.

2.1. Review framework The synthesis was organized according to PRISMA 2020 reporting principles [5], with a descriptive evidence map and a focused proportional meta-analysis. Observational clinical cohorts and randomized stewardship/diagnostic investigations were considered in the narrative review. Quantitative pooling was restricted to observational cohorts reporting a comparable initial empirical-therapy inappropriateness numerator and denominator. 2.2. Review question and eligibility Clinical investigations were eligible for mapping if they enrolled patients with HAP/VAP and reported resistance, initial antimicrobial adequacy, treatment practices or relevant outcomes. For the proportional analysis, inclusion additionally required an identifiable independent sample and an extractable initial-therapy numerator/denominator. Studies limited to guideline adherence, treatment appropriateness at later time points or overlapping datasets were not pooled with the initial-coverage studies. 2.3. Information sources and search strategy The search framework incorporated PubMed/MEDLINE, Embase, Scopus, Web of Science, Cochrane Library and citation searching. The evidence check extended through 23 June 2026. Core concepts were HAP/VAP, nosocomial pneumonia, resistance, empirical antimicrobial adequacy and clinical outcomes. A generalizable Boolean strategy was: (hospital-acquired pneumonia OR nosocomial pneumonia OR ventilator-associated pneumonia) AND (antimicrobial resistance OR multidrug resistant OR carbapenem resistant) AND (empirical antibiotics OR initial antimicrobial therapy OR inappropriate coverage) AND (mortality OR clinical outcome OR treatment failure). The PRISMA counts supplied for this manuscript are reported in Table 2 and Figure 1. Search export files and reviewer-level screening decisions were not available for independent audit during document preparation. 2.4. Data collection and outcome definitions Study design, country, population, pneumonia subtype, microbiological findings, resistance phenotypes, criteria for empirical-treatment appropriateness, event counts and downstream outcomes were extracted from reported publications. “Inappropriate” refers to lack of in-vitro activity of the initial regimen against the recovered causative organism(s), where this was the study definition; it should not be conflated with incorrect timing, dosing or duration. Study-specific denominators were preserved, and episode-based datasets were identified explicitly. 2.5. Statistical synthesis For each cohort, p = x/n, where x denotes inappropriate initial empirical treatment and n the treatment-appropriateness-evaluable denominator. Proportions were logit transformed and pooled by inverse-variance random-effects meta-analysis with DerSimonian–Laird estimation of between-study variance [7]. Ninety-five percent confidence intervals for individual proportions were based on the Wilson method; pooled confidence limits were obtained on the logit scale and back-transformed. Heterogeneity was quantified with Cochran’s Q and Higgins’ I² [8]. A mortality effect was not pooled because the mapped clinical-outcome studies differed in effect estimands and populations. 2.6. Risk of bias Risk of bias was appraised for each of the 10 mapped investigations using a framework matched to its study design and target outcome. ROBINS-I domains informed judgments for nonrandomized comparisons of treatment exposure and patient outcomes [6]; RoB 2 was applied to the randomized intervention trials; and a JBI-informed selection, measurement, and completeness assessment was used for descriptive observational cohorts. The judgments integrate patient selection, exposure classification, confounding, missing data, outcome ascertainment, and selective reporting. RoB 2 categories are Low risk, Some concerns, and High risk; ROBINS-I categories are Low, Moderate, Serious, and Critical; descriptive-cohort assessments are expressed as Low, Moderate, or High concern. Table 8 provides outcome-specific review judgments and their supporting methodological reasons.

Table 1. Population, exposure, comparator and outcome framework Component Operational definition Population Adults with hospital-acquired or ventilator-associated pneumonia; mixed-age studies retained in narrative discussion when relevant Exposure Initial empirical therapy that lacked activity against the subsequently identified causative organism, or antimicrobial-resistant pathogens Comparator Appropriate empirical therapy, susceptible organisms, or standard-care treatment strategy Primary quantitative endpoint Proportion of evaluable patients or episodes receiving inappropriate initial empirical therapy Secondary endpoints Mortality, clinical cure, treatment failure, length of stay, resistance distribution, stewardship and resource utilization 3.1. PRISMA 2020 study selection The study-selection accounting contains 649 database records and 25 additional reports identified through reference-list searching. After removal of 146 database duplicates, 503 records were screened; 430 were excluded at title/abstract screening. Of 73 database reports sought for retrieval, six were unavailable, leaving 67 assessed reports. Citation searching provided another 24 assessed reports after one report was not retrieved. The combined 91 full texts yielded 10 original investigations after 81 exclusions. Three studies (611 evaluable observations) contributed to the quantitative synthesis; seven contributed to narrative synthesis only. Table 2. PRISMA 2020 numerical study-selection accounting PRISMA stage Number (n) Identification through databases PubMed/MEDLINE 214 Embase 172 Scopus 139 Web of Science 96 Cochrane Library 28 Total database records identified 649 Duplicate records removed 146 Screening Records screened 503 Records excluded at title/abstract stage 430 Reports sought for retrieval (database route) 73 Reports not retrieved (database route) 6 Full-text reports assessed from databases 67 Identification through other methods Reports identified through citation searching 25 Citation reports not retrieved 1 Additional full texts assessed 24 Eligibility Total full-text reports assessed 91 Full-text reports excluded 81 Not original clinical investigations 21 Did not specifically address HAP/VAP 18 No relevant resistance or empirical-treatment data 17 Insufficient extractable clinical outcome information 15 Overlapping cohorts or duplicate populations 10 Included Studies included in qualitative synthesis 10 Studies included in quantitative meta-analysis 3 Studies in qualitative synthesis only 7 Total evaluable observations in meta-analysis 611 3.2. Characteristics of included investigations Table 3. Characteristics of original investigations in the evidence map Investigation Design/setting Population Selected outcome Ko et al., 2021 [9] Korea; multicenter cohort 526 HAP/VAP patients Initial empirical appropriateness: 92/194 inappropriate Briseno-Ramírez et al., 2025 [10] Mexico; retrospective ICU cohort 197 VAP episodes 89/197 episodes initially inappropriately covered Rodríguez et al., 2025 [11] Spain; 10-year retrospective ICU cohort 220 VAP patients 10/220 initially inappropriately covered Chang et al., 2021 [12] Korea; multicenter prospective cohort 381 HAP/VAP patients 235 with pathogens; A. baumannii frequently isolated Sophonsri et al., 2023 [13] USA; observational cohort 457 pneumonia patients Mortality differed among nvHAP, vHAP and VAP Stanley et al., 2024 [14] India; prospective cohort 100 VAP patients 46% overall case fatality; antibiotic-duration analysis Enne et al., 2025 [15] Multicenter randomized diagnostic study 531 evaluable at 24 hours Appropriate/proportionate prescribing 205/268 vs 147/263 Alsubaie et al., 2026 [16] Randomized stewardship trial 700 HAP/VAP patients Day-3 appropriateness 77.6% vs 60.8% Zilberberg et al., 2022 [3] USA; multicenter retrospective cohort 17,819 pneumonia cases IET varied by syndrome and resistance phenotype Jackson et al., 2026 [4] Europe; prospective surveillance 590 VAP cases Regional differences in organisms and resistance The denominator for Ko et al. reflects cases with classifiable initial treatment adequacy rather than the full cohort; the Mexican dataset uses VAP episodes, not unique patients. Study populations are not additive. 3.3. Inappropriate initial empirical therapy Ko et al. reported 92 inappropriate and 102 appropriate initial regimens among 194 classifiable patients; another 17 microbiologically diagnosed patients had treatment appropriateness recorded as not applicable [9]. Briseno-Ramírez et al. reported inappropriate initial therapy in 89 of 197 VAP episodes, and a median three days to correction after initial mismatch [10]. Rodríguez et al. identified inappropriate coverage in 10 of 220 VAP patients despite changes in empirical selection practices over a decade [11]. Table 4. Quantitative characteristics of the three studies pooled Study Evaluable (n) Inappropriate (n) Appropriate (n) Proportion 95% CI Ko et al., 2021 194 92 102 47.4% 40.5–54.4% Briseno-Ramírez et al., 2025 197 89 108 45.2% 38.4–52.2% Rodríguez et al., 2025 220 10 210 4.5% 2.5–8.2% Descriptive total 611 191 420 31.3% — 3.4. Meta-analysis of treatment inappropriateness The three cohorts provided 611 evaluable observations, 191 with initially inappropriate and 420 with appropriate coverage. The simple aggregate inappropriate-treatment percentage was 31.3%. In the DerSimonian–Laird random-effects model, the pooled proportion was 25.5% (95% CI 9.0%–54.3%). Heterogeneity was very high (Q = 73.09, I² = 97.3%, τ² = 1.154 on the logit scale), limiting the clinical interpretability of a single summary proportion. Table 5. Meta-analysis statistical results Statistic Value Independent studies 3 Evaluable observations 611 Inappropriate empirical therapy 191 Appropriate empirical therapy 420 Descriptive combined proportion 31.3% Pooled random-effects proportion 25.5% 95% confidence interval 9.0%–54.3% Cochran Q 73.09 Heterogeneity I² 97.3% Between-study variance τ² 1.154 Study-specific points represent proportions and Wilson 95% confidence intervals. The diamond depicts the random-effects pooled estimate. The substantial between-study heterogeneity and use of both patient- and episode-level units mean this pooled estimate is exploratory. 3.5. Pathogens and resistance patterns In the Korean prospective ICU cohort, 235 of 381 patients had identified causative microorganisms, with A. baumannii (89 isolates), S. aureus (52), K. pneumoniae (25) and P. aeruginosa (22) prominent among the isolates [12]. The Mexican VAP study documented extensively drug-resistant A. baumannii in 62 of 71 relevant isolates, illustrating the complexity of empirical coverage in a high-resistance environment [10]. These results contrast with the comparatively low initial mismatch frequency in the Spanish cohort [11]. Table 6. Selected resistance and microbiological findings Study Finding Implication Chang et al. [12] A. baumannii n = 89; S. aureus n = 52 Local predominant organisms shape empirical regimens Chang et al. [12] K. pneumoniae n = 25; P. aeruginosa n = 22 Gram-negative coverage is context dependent Briseno-Ramírez et al. [10] 62/71 A. baumannii isolates extensively drug-resistant High probability of susceptibility mismatch Zilberberg et al. [3] IET varied across pneumonia subtypes and resistant phenotypes Hospital-level risk stratification required Jackson et al. [4] Cross-country variations in VAP pathogen distribution Local antibiograms should inform empirical therapy 3.6. Clinical outcomes Sophonsri et al. found markedly different mortality among ventilated hospital-acquired bacterial pneumonia (44.6%), ventilator-associated bacterial pneumonia (36.0%) and nonventilated HAP (14.3%) in an observational cohort [13]. These unadjusted syndrome comparisons are not estimates of the causal effect of antimicrobial resistance or inappropriate therapy. Stanley et al. reported 46% case fatality in 100 VAP patients and did not demonstrate that prolonged antibiotic duration necessarily improved survival [14]. Across large administrative cohorts, inappropriate coverage was associated with increased healthcare utilization in selected patient groups, but estimates were sensitive to illness severity, organism classification, the timing of initial therapy and whether diagnosis was culture-confirmed [3]. Because secondary outcomes were reported with heterogeneous definitions, a common mortality risk ratio was not estimated in the present analysis. 3.7. Stewardship and rapid diagnostics In the multicenter INHALE randomized study, appropriate and proportionate antibiotic treatment at 24 hours occurred in 205 of 268 patients in the rapid-diagnostic group compared with 147 of 263 under standard care [15]. A subsequent pharmacist-led stewardship randomized trial reported improvement in day-three appropriate prescribing (77.6% vs 60.8%) and clinical cure (84.4% vs 74.3%) [16]. These results concern treatment optimization rather than baseline empirical mismatch and were therefore examined narratively. Table 7. Interventions relevant to treatment optimization Study Intervention Relevant result Enne et al., 2025 [15] Rapid molecular diagnostic pathway vs standard care Appropriate/proportionate treatment at 24 h: 205/268 vs 147/263 Alsubaie et al., 2026 [16] Pharmacist-led stewardship vs standard care Appropriate day-3 therapy: 77.6% vs 60.8% Alsubaie et al., 2026 [16] Pharmacist-led stewardship vs standard care Clinical cure: 84.4% vs 74.3%; LOS lower by 2.83 days 3.8. Methodological considerations Appraisal identified four descriptive cohorts with moderate concern about selection or ascertainment, four nonrandomized comparative studies with serious concern about residual confounding or time-related treatment selection, and two randomized trials with some concerns, primarily for unblinded care or clinically assessed outcomes. The three studies pooled for initial empirical-treatment inappropriateness were Ko et al. (moderate), Briseno-Ramírez et al. (moderate), and Rodríguez et al. (moderate). The Briseno-Ramírez cohort analysed 197 VAP episodes from 129 patients; repeated episodes require consideration of within-patient correlation when interpreting inverse-variance weights. The categories reflect the outcome evaluated in this review and do not constitute claims that all outcomes in a given publication carry identical bias. Table 8. Study-level risk-of-bias assessment Study Key reason for judgment Tool and overall judgment Ko et al. [9] Only a microbiologically classified subset contributed an empiric-appropriateness result; tertiary-centre and coding-based selection. JBI-informed: Moderate concern Briseno-Ramírez et al. [10] Retrospective single-centre data; 197 VAP episodes in 129 patients, with possible within-patient correlation; some susceptibility information unavailable. JBI-informed: Moderate concern Rodríguez et al. [11] Single-centre retrospective ten-year cohort; prescribing practices and case mix changed across periods; only 10 inappropriate-treatment cases. JBI-informed: Moderate concern Chang et al. [12] Prospective ascertainment, but nonrandom treatment selection and mortality associations susceptible to severity- and resistance-related confounding. ROBINS-I: Serious risk Sophonsri et al. [13] Retrospective ICD-screened cohort; substantially different illness severity across pneumonia subtypes; risk models not externally validated. ROBINS-I: Serious risk Stanley et al. [14] Small prospective cohort; antibiotic duration defined after survival time accrued, raising survivor/immortal-time selection concerns in mortality comparisons. ROBINS-I: Serious risk Enne et al. [15] Concealed central randomization and blinded stewardship adjudication; open-label treatment and clinical-cure determination allow some measurement/deviation concerns. RoB 2: Some concerns Alsubaie et al. [16] Computer randomization, sealed allocation and blinded prescribing review; providers unblinded and clinical cure partly judged clinically. RoB 2: Some concerns Zilberberg et al. [3] Large administrative database with culture-confirmed selection and susceptibility-based treatment classification; residual severity and indication confounding. ROBINS-I: Serious risk Jackson et al. [4] Prospective multicountry surveillance but incomplete site screening logs, missing outcome/microbiology data and variation in VAP ascertainment. JBI-informed: Moderate concern

.1. Interpretation of principal findings The main quantitative finding was wide variation in inappropriate initial antibiotic therapy, from 4.5% to 47.4% across three independently reported cohorts. The random-effects summary was 25.5%, but I² of 97.3% indicates that it should not be used as a single expected rate for individual institutions. Differences in empirical protocols, resistance profiles, prior antibiotic exposure, ICU populations and case definitions likely explain much of the variability. 4.2. Antimicrobial resistance and empirical mismatch Resistance reduces the probability that customary empirical treatment will cover a subsequently isolated pathogen. However, resistance is neither necessary nor sufficient for inappropriate therapy: active broad-spectrum treatment can be selected for resistant infections, while susceptible organisms can still be uncovered by an overly narrow or ill-chosen regimen. Interpretation requires joint consideration of local resistance rates, colonization history, timing and drug exposure. 4.3. Consequences for clinical outcomes Biologically, inadequate treatment may permit continued infection and progression to sepsis, organ injury or prolonged ventilation. Observational mortality comparisons nevertheless remain vulnerable to confounding by illness severity, comorbidity and clinician treatment choice. Treatment may also be corrected after culture or rapid diagnostic results become available, so a binary initial-adequacy classification does not capture duration of ineffective exposure. These factors justify separating the descriptive pooled rate from causal claims regarding mortality. 4.4. Empirical monotherapy and combination treatment Additional broad-spectrum agents do not invariably improve outcomes. Baek et al. examined antipseudomonal beta-lactam monotherapy versus a fluoroquinolone-containing combination among HAP patients without demonstrating an unequivocal survival advantage for combination therapy [17]. Empirical escalation should be tailored to expected pathogens, illness severity, previous antibiotic exposure and resistance surveillance rather than applied indiscriminately. 4.5. Relationship to earlier evidence A previous meta-analysis focused on inadequate empirical treatment in VAP reported an association with hospital mortality (RR 1.66, 95% CI 1.23–2.23) and a relationship between multidrug resistance and treatment inadequacy (OR 2.50, 95% CI 1.57–3.97) [18]. The present exploratory analysis addresses a different quantitative endpoint—the proportion of initially inappropriate treatment in selected newer clinical cohorts—and should not be interpreted as an updated or replacement mortality meta-analysis. A separate 2026 systematic review highlighted risk factors for resistant pathogens, including prior antimicrobial exposure and acute physiological severity [19]. 4.6. Implications for practice Clinical use of these findings rests on three linked actions: obtaining high-quality microbiological specimens when feasible, selecting timely empirical regimens informed by unit-specific antibiograms and patient risk, and reassessing therapy promptly when susceptibility results or rapid diagnostic findings become available. The IDSA guidance for resistant Gram-negative organisms provides a framework for organism-specific definitive therapy [20]. Local surveillance is more informative than a heterogeneous global prevalence estimate for selecting an empirical regimen. 4.7. Variability between hospitals and regions Surveillance from specialized Indian intensive-care settings documents substantial resistant-pathogen burdens in selected VAP populations [21]. Multicenter work evaluating MDR risk-factor–based treatment algorithms also emphasizes how sharply resistance frequencies differ between hospital-acquired, ventilator-associated and community-onset pneumonia [22]. Such variability supports locally calibrated treatment pathways and evaluation of unintended ecological consequences of broad-spectrum use. 4.8. Strengths The review distinguishes initial microbiological empirical adequacy from later prescribing optimization, preserves study-specific denominators and separates qualitative clinical outcomes from the proportional meta-analysis. The estimates and heterogeneity statistics are recalculable from the three published numerator–denominator pairs. These features improve transparency of the exploratory quantitative comparison. 4.9. Limitations The limited number of quantitative cohorts, pronounced heterogeneity and differing units of analysis restrict precision and transportability. Repeated VAP episodes in one included study may violate independent-observation assumptions. Some investigations use selected culture-positive populations, which can limit application to suspected pneumonia before microbiological confirmation. The study-specific appraisal found moderate concerns in the three pooled cohorts, and serious confounding concerns in several clinical-outcome comparisons. Published methods supported the review judgments, although an independently duplicated domain-by-domain assessment was not undertaken. These issues prevent definitive claims about a universal mismatch frequency or attributable mortality risk. 4.10. Future research Prospective multicenter studies should consistently define time to active treatment, antimicrobial appropriateness, dosing adequacy, resistance phenotype and outcome horizons. Reporting should separate nonventilated HAP, ventilated HAP and VAP; provide exact counts and adjusted effect estimates; and account for repeated infection episodes within patients. Studies of diagnostic and prescribing interventions should assess not only rapid organism identification but also antibiotic exposure, resistance selection, clinical cure, safety and cost-effectiveness.

The three-study proportional meta-analysis identified a pooled inappropriate empirical treatment frequency of 25.5% (95% CI 9.0%–54.3%), with substantial heterogeneity (I² 97.3%). Antimicrobial resistance complicates empirical selection in nosocomial pneumonia, but the frequency of mismatch depends heavily on local pathogen ecology, patient population and prescribing practices. Higher resistance burden and inadequate coverage are associated with adverse clinical and resource-use outcomes in several investigations; causality and the magnitude of excess mortality remain uncertain. Individualized empirical therapy, rapid diagnostic reassessment, local susceptibility surveillance and structured stewardship are warranted.

1. World Health Organization. WHO bacterial priority pathogens list, 2024: bacterial pathogens of public health importance to guide research, development and strategies to prevent and control antimicrobial resistance. Geneva: WHO; 2024. https://www.who.int/publications/i/item/9789240093461 2. Kalil AC, Metersky ML, Klompas M, et al. Management of adults with hospital-acquired and ventilator-associated pneumonia: 2016 clinical practice guidelines by the IDSA and ATS. Clin Infect Dis. 2016;63:e61–e111. doi:10.1093/cid/ciw353. 3. Zilberberg MD, Nathanson BH, Puzniak LA, Dillon RJ, Shorr AF. The risk of inappropriate empiric treatment and its outcomes based on pathogens in non-ventilated, ventilated hospital-acquired and ventilator-associated bacterial pneumonia in the US, 2012–2019. BMC Infect Dis. 2022;22:775. doi:10.1186/s12879-022-07755-y. 4. Jackson H, Hernandez Padilla AC, Vintcent LEM, et al. Perpetual observational study of the clinical and microbiological epidemiology of ventilator-associated pneumonia in Europe. Crit Care. 2026;30:112. doi:10.1186/s13054-025-05753-5. 5. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. doi:10.1136/bmj.n71. 6. Sterne JAC, Hernán MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355:i4919. doi:10.1136/bmj.i4919. 7. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–188. doi:10.1016/0197-2456(86)90046-2. 8. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557–560. doi:10.1136/bmj.327.7414.557. 9. Ko RE, Min KH, Hong SB, et al. Characteristics, management, and clinical outcomes of patients with hospital-acquired and ventilator-associated pneumonia: a multicenter cohort study in Korea. Tuberc Respir Dis. 2021;84:317–325. doi:10.4046/trd.2021.0018. 10. Briseno-Ramírez J, Gómez-Quiroz A, Avila-Cardenas BB, et al. Development of a Weighted-Incidence Syndromic Combination Antibiogram (WISCA) to guide empiric antibiotic treatment for ventilator-associated pneumonia in a Mexican tertiary care university hospital. BMC Infect Dis. 2025;25:307. doi:10.1186/s12879-025-10677-0. 11. Rodríguez A, Berrueta J, Páez C, et al. Ten-year evaluation of ventilator-associated pneumonia according to initial empiric treatment: a retrospective analysis using real-world data. Biomedicines. 2025;13(2):360. doi:10.3390/biomedicines13020360. 12. Chang Y, Jeon K, Lee SM, et al. The distribution of multidrug-resistant microorganisms and treatment status of hospital-acquired pneumonia/ventilator-associated pneumonia in adult intensive care units: a prospective cohort observational study. J Korean Med Sci. 2021;36:e251. doi:10.3346/jkms.2021.36.e251. 13. Sophonsri A, Lou M, Ny P, Minejima E, Nieberg P, Wong-Beringer A. Machine learning to identify risk factors associated with the development of ventilated hospital-acquired pneumonia and mortality: implications for antibiotic therapy selection. Front Med. 2023;10:1268488. doi:10.3389/fmed.2023.1268488. 14. Stanley ND, Jeevan JA, Yadav B, et al. Association of antibiotic duration and all-cause mortality in a prospective study of patients with ventilator-associated pneumonia in a tertiary-level critical care unit in Southern India. BMJ Open. 2024;14:e077428. doi:10.1136/bmjopen-2023-077428. 15. Enne VI, Stirling S, Barber JA, et al. INHALE WP3: a multicentre, open-label, pragmatic randomised controlled trial of ICU-based rapid syndromic PCR for HAP/VAP. Intensive Care Med. 2025;51:272–286. doi:10.1007/s00134-024-07772-2. 16. Alsubaie N, Ilyas M, Shah ZA, Habib I, Binsaleh AY, Alrossies AS. Impact of clinical pharmacist-led antimicrobial stewardship on antibiotic appropriateness, clinical outcomes, and antimicrobial consumption in hospital-acquired and ventilator-associated pneumonia: a randomized controlled trial. Front Public Health. 2026;14:1700328. doi:10.3389/fpubh.2026.1700328. 17. Baek MS, Baek AR, Hong SB, et al. Empiric anti-pseudomonal β-lactam monotherapy versus fluoroquinolone combination therapy in patients with hospital-acquired pneumonia: a multicenter cohort study with propensity score matching. J Korean Med Sci. 2023;38:e353. doi:10.3346/jkms.2023.38.e353. 18. Tariq A, Kantas T, Shabneeze BB, et al. Risk factors and outcomes of inadequate empirical antibiotic therapy in ventilator-associated pneumonia: a systematic review and meta-analysis. Ann Med Surg (Lond). 2025;87:3786–3793. doi:10.1097/MS9.0000000000003289. 19. Kawasuji H, Maruyama T, Aoki N, et al. Utility of risk assessment and identification of risk factors for drug-resistant pathogens in hospital-acquired and ventilator-associated pneumonia: systematic reviews and meta-analyses. Respir Investig. 2026;64:101410. doi:10.1016/j.resinv.2026.101410. 20. Tamma PD, Aitken SL, Bonomo RA, et al. Infectious Diseases Society of America 2024 guidance on the treatment of antimicrobial-resistant Gram-negative infections. Clin Infect Dis. 2024;ciae403. doi:10.1093/cid/ciae403. 21. Zirpe K, Kapse U, Sharma A, et al. Incidence, risk factors, microbiology and outcomes of VAP at an NCU in India: a prospective observational study. Indian J Crit Care Med. 2025;29(4):308–313. 22. Maruyama T, Fujisawa T, Ishida T, et al. A therapeutic strategy for all pneumonia patients: a 3-year prospective multicenter cohort study using risk factors for multidrug-resistant pathogens to select initial empiric therapy. Clin Infect Dis. 2019;68:1080–1088. doi:10.1093/cid/ciy631.