Hospital-acquired pneumonia and ventilator-associated pneumonia represent distinct, overlapping syndromes of nosocomial lower respiratory tract infection. The 2024 World Health Organization priority-pathogens framework identifies carbapenem-resistant Acinetobacter baumannii, resistant Enterobacterales and Pseudomonas aeruginosa among major antimicrobial-resistance threats [1]. These organisms are clinically important in hospital pneumonia because limited therapeutic options and prior antibiotic exposure may complicate prompt active coverage. The ATS/IDSA HAP/VAP guideline separates non-ventilated HAP from VAP, recommends syndrome- and setting-specific empirical therapy informed by local antibiograms, and emphasizes reassessment and de-escalation rather than routine indiscriminate broad-spectrum coverage [2]. The 2024 IDSA guidance further differentiates treatment choices for difficult-to-treat P. aeruginosa, carbapenem-resistant Acinetobacter and resistant Enterobacterales [3]. Empirical antibiotic therapy is termed inappropriate or inadequate when no component of the initial regimen is active against the causative pathogen(s) in vitro within a prespecified period. Importantly, microbiological appropriateness is not identical to early administration, adequate pharmacokinetic exposure, source control or clinical response. Resistance can increase the probability of mismatch; inadequate coverage may prolong infection and contribute to treatment failure, yet mortality is also determined by organ dysfunction, comorbidities, competing causes of death and withdrawal-of-care practices. This paper addresses three related questions: (1) which resistance phenotypes and patient factors predict inappropriate initial treatment; (2) what outcomes differ between appropriate and inappropriate treatment; and (3) how consistent are crude mortality associations across independent HAP/VAP studies?

2.1. Design and reporting The review was structured and reported according to PRISMA 2020 [4]. The qualitative synthesis examined antimicrobial resistance, empirical treatment appropriateness, and clinical outcomes in HAP and VAP; the quantitative synthesis assessed mortality in VAP cohorts with extractable group-specific data. 2.2. Eligibility criteria Eligible primary studies enrolled adults with HAP, ventilated HAP or VAP and reported resistant microorganisms, empirical antimicrobial adequacy or clinical outcomes. Prospective and retrospective cohorts were eligible. For the quantitative mortality analysis, studies had to provide independent and identifiable counts of deaths and evaluable patients for both initial-therapy groups. Mixed pneumonia/sepsis cohorts, reports lacking group-specific mortality counts, and overlapping data publications were not entered into the pooled mortality calculation. 2.3. Evidence sources and search terms The study-selection records covered PubMed/MEDLINE (n = 196), Embase (n = 174), Scopus (n = 128), Web of Science (n = 94), and reference lists (n = 20) through 30 July 2026. Search terms addressed hospital-acquired or ventilator-associated pneumonia, antimicrobial or multidrug resistance, empirical antibiotic appropriateness, and clinical outcomes. Records underwent deduplication, title/abstract screening, and full-text eligibility assessment (Figure 1). 2.4. Outcomes and definitions The principal quantitative contrast was risk of death among patients receiving inappropriate versus appropriate empirical antibiotic therapy. Secondary outcomes included hospital mortality, pneumonia-related 28-day mortality, intensive care and hospital length of stay, mechanical-ventilation duration, hospital cost, microbiological resistance distribution and the frequency of inappropriate empirical treatment. Definitions and observation periods were retained as reported; outcome windows were not assumed identical. 2.5. Data extraction and overlapping cohorts For the four VAP cohort comparisons, group-level numbers were transcribed from source publications or derived from unambiguously matching percentages and published totals. Dupont and Leroy contributed hospital-death counts. Luna contributed mortality by initial adequacy; Piskin contributed reported pneumonia-related 28-day mortality from an evaluable VAP subset. The 2012–2019 Premier database was described in several overlapping publications; these were treated as one underlying cohort for mapping, not as independent samples. 2.6. Risk of bias and certainty Methodological limitations were examined using the ROBINS-I domains for nonrandomized studies [5], including confounding, participant selection, classification of empirical treatment, missing data, outcome measurement, and selective reporting. Particular attention was given to differences in infection severity, antimicrobial exposure, and mortality definitions. 2.7. Quantitative synthesis Unadjusted mortality risk ratios (RRs) compared inappropriate with appropriate empirical antibiotic therapy. Log RRs were pooled using inverse-variance random-effects estimation and DerSimonian–Laird between-study variance [6]. Heterogeneity was summarized by I² [7], with 95% CIs calculated on the log scale. Because only four cohorts were available, no funnel plot or meta-regression was performed. The effect of differing mortality endpoints was examined by sensitivity analysis.

3.1. Study selection The search identified 612 records: PubMed/MEDLINE (196), Embase (174), Scopus (128), Web of Science (94), and reference lists (20). After 167 duplicates were removed, 445 records underwent title/abstract screening and 376 were excluded. Of 69 reports sought, five were not retrieved; 64 full texts were assessed, and 54 were excluded (16 non-original investigations, 11 outside HAP/VAP, 12 without relevant resistance or empirical-treatment data, nine without extractable clinical outcomes, and six overlapping or duplicate cohorts). Ten original studies were included in qualitative synthesis; four VAP cohorts comprising 418 patients contributed to meta-analysis (Figure 1). Figure 1. PRISMA 2020 flow diagram for identification, screening, eligibility, and inclusion. 3.2. Characteristics of original studies Investigation Population Primary relevance Selected outcome Dupont et al., 2001 [8] VAP; n=111 Mortality by initial antibiotic adequacy Appropriate: 26/55; inappropriate: 34/56 Leroy et al., 2003 [9] VAP; n=132 Bacteriologically documented VAP, mortality Appropriate: 42/106; inappropriate: 16/26 Clec’h et al., 2004 [10] VAP; n=142 Initial adequacy and severity interaction Potential effect modification by organ dysfunction Luna et al., 2006 [11] VAP; n=76 Prompt adequacy vs inappropriate/delayed treatment Appropriate: 7/24; inadequate/delayed: 33/52 Teixeira et al., 2007 [12] VAP; n=151 Mortality by adequacy; no extractable exact 2×2 here Mortality: 29.3% appropriate vs 50.7% inadequate Piskin et al., 2012 [13] HAP n=218; VAP n=130 Clinical outcomes by initial treatment adequacy 28-day pneumonia-related VAP subset: 30/40 vs 42/59 Maruyama et al., 2019 [14] All pneumonia n=1,089; HAP n=140; VAP n=55 MDR-risk–based empirical algorithm Inappropriate therapy 4.3% overall; VAP MDR 50.9% Chang et al., 2021 [15] HAP/VAP; n=381 Pathogen distribution and resistance Pathogen isolated in 235; A. baumannii most common Ko et al., 2021 [16] HAP/VAP; n=526 Epidemiology and hospital outcomes MDR pathogens n=138; hospital mortality 28.1% Zilberberg et al., 2022 [17] nvHAP, ventilated HAP, VAP; n=17,819 IET, pathogen groups and outcome IET 8.5%, 5.6%, 7.2% across three syndromes Abbreviations: HAP, hospital-acquired pneumonia; VAP, ventilator-associated pneumonia; IET, inappropriate empirical therapy; MDR, multidrug resistance. Cohort denominators above must not be summed because they contain different populations and outcome availability; study inclusion does not imply comparable definitions. 3.3. Microbial resistance and empirical-therapy mismatch In the 381-patient Korean intensive-care prospective cohort, pathogens were identified in 235 (62%), and Acinetobacter baumannii (n=89) was the most frequently recovered organism, followed by Staphylococcus aureus (n=52), Klebsiella pneumoniae (n=25) and Pseudomonas aeruginosa (n=22). The majority of reported A. baumannii isolates in that cohort were multidrug resistant [15]. An independent Korean multicenter cohort of 526 HAP/VAP patients reported microbiological identification in 211 (40.1%), MDR organisms in 138, and hospital mortality of 28.1% [16]. These results underscore substantial differences in microbiological ascertainment and underlying illness severity across clinical settings. In a large US cohort of 17,819 nosocomial bacterial pneumonia cases, inappropriate empirical treatment was observed in 8.5% of non-ventilated HAP, 5.6% of ventilated HAP and 7.2% of VAP. Gram-negative organisms were more likely than Gram-positive organisms to be associated with IET. Among carbapenem-resistant Gram-negative infections, the reported risk of IET reached 33.8% to 40.2% across syndrome groups [17]. These percentages reflect the study’s exposure definitions and the susceptibility data available, and are not universal prevalence estimates. 3.4. Mortality meta-analysis Study Inappropriate: deaths/N Appropriate: deaths/N Unadjusted RR (95% CI) Dupont et al., 2001 34/56 26/55 1.28 (0.91–1.82) Leroy et al., 2003 16/26 42/106 1.55 (1.06–2.28) Luna et al., 2006 33/52 7/24 2.18 (1.13–4.20) Piskin et al., 2012* 42/59 30/40 0.95 (0.75–1.21) *Piskin et al. reported pneumonia-related 28-day mortality in an evaluable VAP subgroup, whereas other studies reported hospital/ICU mortality. Leroy et al. group-specific deaths were reconstructed from published group totals and rounded mortality percentages. Four VAP cohorts contributed 418 evaluable patients: 125/193 deaths (64.8%) with inappropriate treatment and 105/225 (46.7%) with appropriate treatment. The pooled RR was 1.32 (95% CI 0.96–1.81; I² = 65.3%). The interval included the null value, and the association was not statistically significant. Figure 2. Forest plot of unadjusted mortality risk ratios in four VAP cohorts. Squares indicate study estimates, horizontal lines 95% CIs, and the diamond the pooled random-effects estimate. Excluding Piskin et al., which used a different mortality endpoint, yielded RR 1.49 (95% CI 1.17–1.89; I² approximately 0.7%). This result demonstrates the influence of endpoint differences on the pooled estimate. 3.5. HAP-specific mortality and broader clinical outcomes In the HAP mortality-evaluable subset reported by Piskin et al., pneumonia-related 28-day deaths occurred in 32/81 initially inadequately treated patients versus 30/119 adequately treated patients [13]. The crude RR calculated from those counts was 1.57, but the source reports a nonsignificant survival-analysis test, illustrating that raw event ratios and time-to-event comparisons cannot be treated as equivalent. This single HAP cohort was not pooled with VAP. Across the large US database analysis, IET was associated with longer postinfection hospital stay, higher costs and, in selected pneumonia subgroups, greater readmission or mortality [17]. Multiple reports from the same Premier cohort should not be double-counted as independent evidence; a related study reported increased adjusted length of stay and costs with IET [18]. A separate mixed pneumonia/sepsis Acinetobacter analysis found an adjusted mortality risk ratio of 1.76 (95% CI 1.36–2.28) in association with IET, but was not eligible for the present HAP/VAP-specific meta-analysis because pneumonia outcomes were not separately reported [19]. 3.6. Methodological limitations and risk-of-bias considerations The principal concern is confounding by indication and disease severity: patients infected by highly resistant organisms or with more severe physiological derangement may both be more likely to receive inappropriate empirical coverage and more likely to die. Pneumonia onset definitions, clinical-versus-microbiological diagnoses, timing of treatment, culture-negative classification, and subsequent changes in antibiotics differ across studies. ROBINS-I domain Considerations for interpretation Confounding Baseline severity, previous antibiotic exposure, resistance phenotype, and treatment timing Participant selection Single-center/ICU populations and restrictions to mortality-evaluable subsets Treatment classification Variation in definitions and timing of appropriate empirical treatment Missing data Unavailable susceptibility results and incomplete outcome denominators Outcome measurement Hospital/ICU mortality versus pneumonia-related 28-day mortality Selective reporting Differences in availability of group-specific mortality events and denominators

4.1. Interpretation of principal findings This evidence synthesis indicates that antimicrobial resistance is a central driver of failure to provide active empirical therapy in nosocomial pneumonia. Organisms with carbapenem resistance, multidrug resistance or unusual intrinsic resistance can evade standard hospital regimens. Nevertheless, the consequences of initial inadequacy vary because pathogens, baseline severity, empirical treatment policies, clinical sampling and subsequent treatment modification are heterogeneous. Across four VAP cohorts, inappropriate empirical therapy was associated with a numerically higher pooled mortality risk, although the 95% CI included no association. A broader 2025 systematic review reported an association between inadequate VAP therapy and hospital mortality (RR 1.66, 95% CI 1.23–2.23), as well as between multidrug-resistant organisms and inadequate coverage (OR 2.50, 95% CI 1.57–3.97) [20]. Differences in eligible cohorts and outcome definitions may account for variation between summaries. 4.2. Causal interpretation Inappropriate empirical therapy is typically measured retrospectively after pathogen identification, whereas decisions are made prospectively with incomplete information. Treatment may also be revised within 24–72 hours after susceptibility results become available. This creates challenges involving time-dependent confounding, differential diagnostic workup and selection of patients with positive cultures. Associations between empirical inadequacy and mortality therefore should not be interpreted as directly causal without robust time-sensitive analyses. 4.3. Clinical implications Empirical treatment decisions should balance the risk of undertreatment against the harms of unnecessary broad-spectrum antibiotics, including adverse effects, Clostridioides difficile infection and further resistance selection. Risk assessment should consider recent intravenous antimicrobial use, known colonization or infection with resistant organisms, ICU-specific antibiograms, local MRSA prevalence, shock or ventilatory support, and sample-quality limitations [2,3]. Microbiological sampling should be obtained promptly when feasible, without delaying needed treatment in unstable patients. Results should be reviewed daily, particularly after 48–72 hours, to confirm activity and consider streamlining or discontinuation when the diagnosis is not supported. Infection-control interventions and resistance surveillance are necessary partners to antimicrobial stewardship. A uniform escalation strategy for every hospitalized patient is not warranted. 4.4. Differences between HAP and VAP HAP that subsequently requires mechanical ventilation carries a different risk profile from pneumonia emerging after intubation. In the Premier database, hospitalized patients with ventilated HAP had higher crude mortality than those with non-ventilated HAP or VAP, emphasizing the risk of combining these populations without syndrome-specific stratification [17,18]. Future studies should separate non-ventilated HAP, ventilated HAP and VAP and report the actual time at which ventilation began relative to pneumonia onset. 4.5. Research gaps The highest priorities are prospectively standardized definitions of initial treatment adequacy, exact time-to-active-antibiotic metrics, organism-level susceptibility stratification, dosing and pharmacokinetic adequacy, and clinically relevant adjusted effect estimates. Research should report 28-day and hospital mortality separately and provide raw group counts, adjusted estimates and predeclared confounder sets. Pragmatic resistance-based antibiotic selection algorithms should be evaluated alongside antibiotic exposure and ecological harms. 4.6. Strengths and limitations of this review Strengths include separate consideration of HAP and VAP phenotypes, careful treatment of overlapping database cohorts, and transparent study-level mortality calculations. Limitations include four quantitative cohorts, differing mortality endpoints, reconstructed group counts in one study, reliance on unadjusted effect estimates, and potential confounding by illness severity and treatment indication. HAP-specific mortality data were insufficient for pooled analysis.

Antimicrobial resistance contributes to mismatched empirical antibiotic therapy in hospital-acquired and ventilator-associated pneumonia. Inappropriate treatment is associated with greater resource use in several large cohorts, while mortality effects vary. The four-cohort VAP meta-analysis showed RR 1.32 (95% CI 0.96–1.81), with substantial heterogeneity. Empirical regimens should be guided by patient risk factors and local susceptibility patterns, followed by timely microbiological review and de-escalation when appropriate.

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. Tamma PD, Aitken SL, Bonomo RA, et al. IDSA 2024 guidance on the treatment of antimicrobial-resistant Gram-negative infections. Clin Infect Dis. 2024:ciae403. doi:10.1093/cid/ciae403. 4. 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. 5. 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. 6. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–188. doi:10.1016/0197-2456(86)90046-2. 7. 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. 8. Dupont H, Mentec H, Sollet JP, Bleichner G. Impact of appropriateness of initial antibiotic therapy on the outcome of ventilator-associated pneumonia. Intensive Care Med. 2001;27:355–362. doi:10.1007/s001340000640. 9. Leroy O, Meybeck A, d’Escrivan T, Devos P, Kipnis E, Georges H. Impact of adequacy of initial antimicrobial therapy on the prognosis of patients with ventilator-associated pneumonia. Intensive Care Med. 2003;29:2170–2173. doi:10.1007/s00134-003-1990-x. 10. Clec’h C, Timsit JF, De Lassence A, et al. Efficacy of adequate early antibiotic therapy in ventilator-associated pneumonia: influence of disease severity. Intensive Care Med. 2004;30:1327–1333. doi:10.1007/s00134-004-2292-7. 11. Luna CM, Aruj P, Niederman MS, et al. Appropriateness and delay to initiate therapy in ventilator-associated pneumonia. Eur Respir J. 2006;27:158–164. doi:10.1183/09031936.06.00049105. 12. Teixeira PJZ, Seligman R, Hertz FT, Cruz DB, Fachel JMG. Inadequate treatment of ventilator-associated pneumonia: risk factors and impact on outcomes. J Hosp Infect. 2007;65:361–367. doi:10.1016/j.jhin.2006.12.019. 13. Piskin N, Aydemir H, Oztoprak N, et al. Inadequate treatment of ventilator-associated and hospital-acquired pneumonia: risk factors and impact on outcomes. BMC Infect Dis. 2012;12:268. doi:10.1186/1471-2334-12-268. 14. 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. 15. 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. 16. 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. 17. 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. 18. Zilberberg MD, Nathanson BH, Puzniak LA, Shorr AF. Microbiology, empiric therapy and its impact on the outcomes of nonventilated hospital-acquired, ventilated hospital-acquired, and ventilator-associated bacterial pneumonia in the United States, 2014–2019. Infect Control Hosp Epidemiol. 2022;43:277–283. doi:10.1017/ice.2021.464. 19. Zilberberg MD, Nathanson BH, Sulham K, Fan W, Shorr AF. Multidrug resistance, inappropriate empiric therapy, and hospital mortality in Acinetobacter baumannii pneumonia and sepsis. Crit Care. 2016;20:221. doi:10.1186/s13054-016-1392-4. 20. 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.