Trial of antibiotic restraint in presumed pneumonia: A Surgical Infection Society multicenter pilot : Journal of Trauma and Acute Care Surgery

Secondary Logo

Journal Logo


Trial of antibiotic restraint in presumed pneumonia: A Surgical Infection Society multicenter pilot

Guidry, Christopher A. MD; Beyene, Robel T. MD; Watson, Christopher M. MD; Sawyer, Robert G. MD; Chollet-Hinton, Lynn PhD; Simpson, Steven Q. MD; Atchison, Leanne PharmD; Derickson, Michael MD; Cooper, Lindsey C. PharmD; Pennington, G. Patton II MD; VandenBerg, Sheri RN; Halimeh, Bachar N. MBBS; O'Dell, Jacob C. MD

Author Information
Journal of Trauma and Acute Care Surgery 94(2):p 232-240, February 2023. | DOI: 10.1097/TA.0000000000003839
  • Free
  • SDC
  • CME Test
  • Infographic
  • Best Of

Pneumonia is the most common intensive care unit (ICU)–acquired infection and carries a high associated mortality.1 The diagnosis of pneumonia with or without sepsis can be difficult, particularly in the surgical and trauma population, where numerous noninfectious causes of fever, leukocytosis, or organ dysfunction can obscure the clinical diagnosis.2–5

Decisions regarding when to start antibiotics in the case of a potential pneumonia generally must balance two opposing narratives. On one side, multiple societal guidelines recommend a clinical and culture-based approach to antibiotic initiation.6,7 On the other, the Surviving Sepsis Campaign (SSC) has historically urged for rapid initiation of antibiotics, although this stance has been softened in the most recent guidelines.8,9 These SSC recommendations are tied to hospital reimbursement via the Centers for Medicare and Medicaid Services Severe Sepsis and Septic Shock Early Management bundle. Recently, the SSC and Severe Sepsis and Septic Shock Early Management bundle have been criticized for urging antibiotic overuse without demonstration of improved outcomes.10–14 Despite the evidence against rapid antibiotic administration in many cases, there remain no randomized studies of antibiotic initiation strategies in the surgical and trauma population.

The purpose of this study was to compare an immediate antibiotic administration protocol contrasted to a specimen-initiated protocol based on objective evidence of infection in intubated surgical and trauma patients with suspected hospital-acquired or ventilator-associated pneumonia. We hypothesized that a protocol of specimen-initiated antibiotic initiation would have similar compliance and outcomes to an immediate initiation protocol for patients who were suspected of having a pneumonia but did not require vasopressors.



This study was designed as a pragmatic multicenter cluster-randomized crossover pilot trial.

All participating investigators received independent full institutional review board approval at their respective institutions. Each site was independently granted a minimal risk determination and waiver of informed consent by their respective institutional review boards. The study received funding from the Surgical Infection Society Foundation, which was not involved in the design or completion of the study. The trial was registered with (NCT04438187). The equator guideline was used to ensure proper reporting of methods, results, and discussion.

As a pilot trial, no sample size calculation was conducted. Since this is a study of patients with suspected but not yet confirmed pneumonia, we estimated a roughly 50% rate of pneumonia among all patients receiving a culture.15,16 We limited the duration of each arm of the study to 4 months or 100 patients (whichever occurred first), allowing for a maximum enrollment of 200 patients with suspected pneumonia.

Inclusion / Exclusion Criteria

Adult patients 18 to 88 years of age who were suspected of having pneumonia were included in this study. All patients needed to be admitted to a surgical intensive care unit (SICU) or trauma intensive care unit (TICU) for at least 48 hours for a primary surgical diagnosis before collecting a culture. Patients needed to be intubated, but not necessarily for 48 hours, at the time of culture, ensuring that our population included health care–associated but not necessarily ventilator-associated pneumonia. As a pragmatic study, we did not control the clinical parameters used by each intensivist to determine which patient should receive a culture.17 For study purposes, suspicion for pneumonia was defined as collection of a respiratory culture based on the assessment of the clinical care team. Cultures were obtained via bronchoscopy or mini-bronchoalveolar lavage (BAL). If patients had multiple cultures obtained, then only the initial culture was considered for this analysis.

Patients were excluded if they were not admitted to a SICU or TICU or did not have concern for pneumonia requiring culture. Patients who were pregnant, incarcerated, had a primary diagnosis of burns, had an absolute neutrophil count of <500 cells/mL, or were immunosuppressed were also excluded from the study.

Immediate Initiation Protocol

In the immediate initiation protocol, patients were started on broad spectrum antibiotics immediately after collection of an appropriate respiratory culture regardless of clinical status. Antibiotics were to be stopped after 72 hours if cultures were negative and no other source of infection was identified. Choice of antibiotics was left to intensivist discretion. If a patient in the immediate initiation protocol was already on antibiotics for another indication, antibiotics were immediately broadened.

Specimen-Initiated Protocol

In the specimen-initiated protocol, if the patient did not require vasopressors, antibiotics were held until there was objective evidence of infection. Patients who were considered by the intensivist to be in septic shock or required vasopressors were started on antibiotics immediately. For the purposes of this analysis, vasopressor use served as a surrogate for shock as is common in the literature on this subject.18,19 Objective evidence of potential pneumonia was defined by a Gram stain listing “2+,” “moderate,” or “many” bacteria, or >10,000 colony-forming units/mL of bacteria on a BAL specimen, mini-BAL specimen, or quantitative endotracheal suction specimen. If a Gram stain reported bacteria but did not report either quantitative or qualitative findings, then it was considered adequate evidence of potential infection.20,21 If a patient in the specimen-initiated protocol was already receiving antibiotics for another indication, no changes in the antibiotic regimen were made unless there was some objective evidence of pneumonia. If there was objective evidence, then antibiotics were broadened as indicated. Choice of antibiotics was left to intensivist discretion.

The intensivist could override the study to start or stop antibiotics as deemed necessary if it was felt that further participation in the study was harmful to the patient, although this would be listed as noncompliance.


The result of each culture and any antibiotic resistance was recorded. Patients whose final culture grew contaminants, “normal oropharyngeal flora,” or a pathogen below the colony-forming units threshold were considered to have had growth on culture (“positive” culture) but were not considered a clinical pneumonia. For patients in the specimen-initiated arm, this may result in initiation of antibiotics for what is ultimately normal flora. Antibiotics would be expected to stop in both protocols if the final cultures did not demonstrate a clinical pneumonia.

Randomization and Crossover

Randomization was conducted by the primary investigator. Each site was randomly assigned a number between 0 and 100 using the RANDBETWEEN function in Microsoft Excel (Microsoft Corporation, Redmond, WA). The sites receiving the lowest two random numbers were assigned to begin with the immediate initiation protocol, while the others were assigned to the specimen-initiated protocol. Enrollment for the first phase of the study ran from February 1, 2021, to May 31, 2021. During the washout period, an interim data analysis was conducted and reviewed by the Data Safety and Monitoring Board. After approval from the Data Safety and Monitoring Board, each site crossed over to the opposite arm of the study and continued enrollment from September 6, 2021, to January 5, 2022.


The primary outcome of the study was protocol compliance. Compliance education was provided to the ICU teams by each site investigator. Noncompliance in the immediate initiation arm included starting antibiotics before culture (except for patients in shock) and failure to stop antibiotics after 72 hours if cultures were negative. Noncompliance in the specimen-initiated arm included initiation of antibiotics in the absence of objective evidence (except for patients in shock), failure to initiate antibiotics when objective evidence was present, and failure to stop antibiotics if final cultures were negative. Secondary outcomes included all-cause in-hospital mortality at 30 days and ventilator-free-alive days at 30 days following culture.

Statistical Analysis

Continuous variables were compared using Wilcoxon rank sum, while categorical variables were compared using χ2 or Fisher's exact tests as appropriate. Outcomes were compared both as an intention-to-treat and per-protocol analysis. Subgroup analysis was further conducted for patients on vasopressors, patients with an increase in sequential organ failure assessment (SOFA) scores of at least 2 from the time of admission to culture, patients with pneumonia, and excluding patients already on antibiotics for another indication. Potential clustering effects by site were evaluated and did not reveal evidence of site-level differences in the associations between treatment arm and any trial outcome. Given minimal statistical evidence of cluster effects, no change to study conclusions, and convergence complications with model estimation due to underpowered analyses, standard statistical techniques for individual-level comparisons were used. Significance was set at p < 0.05. The analysis was conducted with SAS software, version 9.4 (SAS Institute, Cary, NC) and R (version 4.1.1). Study data were stored in an online Research Electronic Data Capture database.22,23


A total of 186 patients were included with 50% in each protocol (Fig. 1). The population consisted of 94.6% trauma patients. A total of 244 respiratory cultures were obtained, and there was no difference in the median number of cultures between groups (1 vs. 1; p = 0.88). The overall pneumonia rate per the initial cultures was 65.6%. Demographics and comorbidities are listed in Table 1. The median age was 50.5 years, and 21% of the overall cohort was female. Fewer patients in the specimen-initiated arm required surgery during their admission. Other than the operative rate, there were no statistical differences between the two groups. Descriptive statistics per site are listed in Supplemental Digital Content (Appendix Tables A–D,

Figure 1:
CONSORT diagram.
TABLE 1 - Demographics and Comorbidities
Immediate Specimen Initiated
Variable 93 93 p
Age, y 45 (30–61) 56 (36–66) 0.07
Female sex 20 21.5% 19 20.4% 0.86
Race* 0.57
 White 69 74.2% 67 72.0%
 Black 14 15.1% 18 19.4%
 Other 9 9.7% 6 6.5%
 Unknown 1 1.1% 2 2.2%
Ethnicity* 0.43
 Hispanic or Latino 9 9.7% 5 5.4%
 Non-Hispanic/non-Latino 84 90.3% 86 92.5%
 Unknown 0 0.0% 2 2.2%
Body mass index 29.3 (24.2–34.5) 28.3 (23.9–33.7) 0.48
 Cardiac disease 14 15.1% 23 24.7% 0.1
 Renal disease 1 1.1% 1 1.1% 1
 Hemodialysis 2 2.2% 1 1.1% 1
 Inflammatory bowel disease 0 0.0% 1 1.1% 1
 Liver disease 0 0.0% 3 3.2% 0.25
 Malignancy 3 3.2% 1 1.1% 0.62
 Peripheral vascular disease 1 1.1% 4 4.3% 0.38
 Pulmonary disease 4 4.3% 12 12.9% 0.06
 Other 12 12.9% 22 23.7% 0.06
 Transfusion during admission 63 67.7% 55 59.1% 0.22
ICU admission diagnosis
 Trauma: blunt 79 84.9% 78 83.9% 0.84
 Trauma: penetrating 9 9.7% 10 10.8% 0.81
 EGS: IAI 1 1.1% 1 1.1% 1
 EGS: non-IAI 1 1.1% 3 3.2% 0.62
 Vascular surgery 0 0.0% 1 1.1% 1
 Scheduled postsurgical admit 1 1.1% 0 0.0% 1
 Decompensation on ward 1 1.1% 0 0.0% 1
 Other 1 1.1% 0 0.0% 1
Operation during admission 35 37.6% 20 21.5% 0.02
Active infection on admission 3 3.2% 5 5.4% 0.72
COVID-19 during admission 5 5.4% 3 3.2% 0.72
Admission scores
 APACHE II 15 (11–20) 16 (11–20) 0.95
 SOFA 6 (4–7) 6 (3–7) 0.92
 Injury Severity Score** 25 (16.5–34.5) 25 (16–34) 0.9
 Glasgow Coma Scale 8 (5–13) 7 (3–14) 0.68
 qSOFA 0.56
 0 6 6.5% 9 9.7%
 1 39 41.9% 42 45.2%
 2 36 38.7% 35 37.6%
 3 12 12.9% 7 7.5%
Site enrollment 0.27
 A 11 11.8% 18 19.4%
 B 18 19.4% 15 16.1%
 C 30 32.3% 21 22.6%
 D 34 36.6% 39 41.9%
Data are presented as n (%) or median (IQR).
*p Values are calculated exclusive of patients with “unknown” race or ethnicity.
**For trauma patients only.
IAI, intra-abdominal infection; APACHE II, Acute Physiology and Chronic Health Evaluation II; IQR, interquartile range; qSOFA, Quick Sequential Organ Failure Assessment.

Characteristics of each culture episode are listed in Table 2. There were no differences in the days from admission to culture, the number of patients requiring vasopressors, and the number of patients on antibiotics for another indication. Sequential organ failure assessment scores did not differ between groups at either the time of admission or the time of culture. Overall, 44.1% of patients had an increase in their SOFA score of at least 2 between admission and the time of culture. Antibiotics were started at a median of 0 hours after culture in the immediate initiation protocol and a median of 9.3 hours after culture in the specimen-initiated protocol. Antibiotics were avoided completely in 19.4% of patients in the specimen-initiated protocol. There were no differences in number of the antibiotic days prescribed for pneumonia or total antibiotic days between groups.

TABLE 2 - Culture Episode Data
Immediate Specimen Initiated
Variable 93 93 p
Days from ICU admission to culture 6 (5–9) 7 (4–9) 0.98
Vasopressors within 24 h of antibiotics 34 36.6% 30 32.3% 0.69
On antibiotics for other infection 10 10.8% 17 18.3% 0.15
Maximum temperature (°C) within 24 h of culture 38.6 (37.9–39.3) 38.8 (38.1–39.3) 0.9
Maximum WBCs (1,000 cells/mL) within 24 h of culture 15.1 (11.5–18.6) 15.8 (11.8–19.4) 0.18
Maximum SOFA within 24 h of culture 6 (5–9) 6 (5–8) 0.85
Maximum SOFA within 24 h of antibiotics 7 (5–9) 6 (4–8.5) 0.34
Change in SOFA from admission to culture 1 (−1 to 4) 1 (−1 to 3) 0.7
Change in SOFA from culture to antibiotics 0 (0–0) 0 (−1 to 1) 0.47
Increase in SOFA of ≥2 from admission to culture 40 43.0% 42 45.2% 0.77
Increase in SOFA of ≥2 from culture to antibiotics 11 11.8% 11 11.8% 1
Any growth on culture* 77 82.8% 80 86.0% 0.54
Clinical pneumonia based on culture results 66 71.0% 56 60.2% 0.12
Hours from culture to antibiotics 0 (0–2) 9.3 (0–26) <0.0001
Never started antibiotics 0 0.0% 18 19.4% <0.0001
Total days of antibiotics 14 (8–21) 13 (8–20) 0.45
Antibiotics for pneumonia, median (IQR) 7 (6–8) 7 (0–8) 0.05
Antibiotics for pneumonia, mean (SD) 7.3 3.49 5.9 4.31
Data are presented as n/% or median (IQR).
*Includes any growth at all on reported on culture data including contaminants, normal oropharyngeal flora, or growth of pathogenic organism less than the 10,000 colony-forming unit threshold.
IQR, interquartile range; WBC, white blood cell.

Outcomes are listed in Table 3. Overall protocol adherence was 78.5%. There was no statistical difference in the protocol adherence rate between groups. The listed causes of noncompliance were as follows: continuation of antibiotics despite negative cultures (72.5%), initiation of antibiotics before culture (17.5%), inappropriate continuation of antibiotics after transfer out of unit (5%), failure to start antibiotics immediately in the immediate initiation arm (2.5%), and 2.5% where the causes were unlisted. There were no statistical differences between the listed causes of noncompliance between groups. When analyzed on an intention-to-treat basis, there were no differences in the number of ventilator-free alive days, rate of mortality due to pneumonia, ICU mortality, or all-cause 30-day mortality between groups. Similarly, there were no statistical differences between groups when analyzed on a per-protocol basis. There were no differences in outcomes for any of the other subgroup analyses (Supplemental Digital Content, Appendix Tables E–I, Notably, there was no difference in the timing of antibiotic initiation for patients who were on vasopressors (median, 0.38 vs. 1 hours; p = 0.38; Supplemental Digital Content, Appendix Table F, and treated according to protocol, which was expected.

TABLE 3 - Outcomes
Immediate Specimen Initiated
Variable 93 93 p
Protocol adherence 69 74.2% 77 82.8% 0.15
Ventilator-free alive days at 30 d 7 (0–15) 8 (0–17) 0.6
Mortality due to pneumonia 2 2.2% 1 1.1% 1
ICU mortality 15 16.1% 14 15.1% 0.84
All-cause 30-d mortality 17 18.3% 17 18.3% 1
Data are presented as n (%) or median (IQR).
IQR, interquartile range.

Isolated pathogens for patients with pneumonia are listed in Table 4. The most common organisms were methicillin-sensitive Staphylococcus aureus and methicillin-resistant S. aureus followed by Escherichia coli and Klebsiella pneumoniae. There were more isolates of Klebsiella oxytoca in the specimen-initiated group but otherwise no difference in pathogens between groups. The overall resistance rate was 25.4% with no difference between the immediate initiation and specimen-initiated protocols (24.2% vs. 26.8%; p = 0.75).

TABLE 4 - Pathogenic Organisms
Total Immediate Specimen Initiated
Pathogens 122 66 56 p
Methicillin-sensitive Staphylococcus aureus 45 36.9% 23 34.8% 22 39.3% 0.62
Methicillin-resistant S. aureus 20 16.4% 13 19.7% 7 12.5% 0.28
Escherichia coli 10 8.2% 7 10.6% 3 5.4% 0.34
Klebsiella pneumoniae 9 7.4% 7 10.6% 2 3.6% 0.18
Pseudomonas aeruginosa 9 7.4% 4 6.1% 5 8.9% 0.73
Enterobacter cloacae 8 6.6% 6 9.1% 2 3.6% 0.29
Haemophilus influenzae 8 6.6% 5 7.6% 3 5.4% 0.73
Serratia marcescens 7 5.7% 3 4.5% 4 7.1% 0.7
Enterobacter aerogenes 6 4.9% 4 6.1% 2 3.6% 0.69
Stenotrophomonas maltophilia 6 4.9% 1 1.5% 5 8.9% 0.09
Streptococcus spp. 6 4.9% 1 1.5% 5 8.9% 0.09
Acinetobacter baumannii 4 3.3% 2 3.0% 2 3.6% 1
K. oxytoca 4 3.3% 0 0.0% 4 7.1% 0.04
Proteus mirabilis 4 3.3% 3 4.5% 1 1.8% 0.62
S. pneumoniae 4 3.3% 3 4.5% 1 1.8% 0.62
Citrobacter spp. 3 2.5% 2 3.0% 1 1.8% 1
Corynebacterium spp. 3 2.5% 1 1.5% 2 3.6% 0.59
Hafnia alvei 1 0.8% 0 0.0% 1 1.8% 0.46
Data are presented as n/%.
Organisms add up to >100% since multiple organisms could be present in a single sample.


In this limited pilot trial, we have demonstrated that a protocol of specimen-initiated antibiotic administration for suspected hospital-acquired or ventilator-associated pneumonia not requiring vasopressors can be reliably and safely applied when compared with an immediate initiation protocol. The overall 78.5% protocol adherence rate was higher than we initially anticipated and did not differ significantly between groups. While overall antibiotic durations were similar between two groups, we were able to avoid antibiotics entirely in 19.4% of patients in the specimen-initiated arm.

Most data surrounding the timing of antibiotic initiation are focused on patients with potential sepsis who present in the emergency department. Levy et al.24 found that rapid initiation of antibiotics was associated with decreased mortality. However, this study primarily evaluated overall sepsis bundle compliance and did not evaluate patients in shock or on vasopressors separately. A larger study by Seymour et al.18 including 49,331 patients found that, while overall mortality was increased with delays in antibiotic initiation (odds ratio, 1.04; 95% confidence interval, 1.03–1.06), this was primarily driven by patients requiring vasopressors. Patients who did not need vasopressors did not have an association between the timing of antibiotics and mortality.18 In our study, patients requiring vasopressors were started on antibiotics immediately regardless of intervention group. Accordingly, we found no statistical differences in the timing of antibiotics for patients on vasopressors with no difference in mortality compared with those patients not on vasopressors. Most recently, Bisarya et al.19 demonstrated no association between the timing of antibiotics and mortality in a study of 74,114 patients presenting to the emergency department with suspected sepsis. In our study, while underpowered, patients with an increase in their SOFA scores of ≥2, consistent with the SEPSIS-3 definition, similarly had no differences in outcomes.25 The PHANTASi (prehospital antibiotics in the ambulance for sepsis) trial by Alam et al.,26 the only randomized trial in this population, also demonstrated no association between earlier initiation of antibiotics and mortality. A 2021 machine learning analysis of the PHANTASi trial data by Schinkel et al.27 suggested that patients 76 years or older may benefit from rapid antibiotic administration. Unfortunately, we are underpowered to evaluate this subset of patients. The SSC has recently endorsed a brief period to reasonably assure that infection is the cause of organ dysfunction before initiating antimicrobial therapy.9

Our study is most directly comparable with the 2012 before-and-after, quasi-experimental study by Hranjec et al.28 In that study, the authors demonstrated increased odds of mortality for ICU patients treated under what they deemed an aggressive initiation protocol, comparable with our immediate initiation protocol (odds ratio, 2.5; 95% confidence interval, 1.5–4.0). That study also demonstrated no difference in pneumonia-specific mortality.28 A similar study published in 2021 by Le Terrier et al.29 demonstrated lower overall and ICU-related mortality for patients treated under a restrictive antibiotic initiation protocol. In contrast to these studies, we failed to identify a difference in outcomes between our immediate and specimen-initiated protocols. However, we are not powered to detect small differences in our outcomes. The studies of Hranjec et al.28 and Le Terrier et al.29 are single-center nonrandomized and suffer from inherent bias because of their quasi-experimental design.30 Both these studies are also significantly smaller than their emergency department–based counterparts and may be underpowered. Other observational studies of ICU-acquired infections have also demonstrated no association between antibiotics administration and outcomes.31–33

Bloos et al.34 performed a cluster-randomized educational sepsis intervention focused on the timing and management of empiric antibiotics but ultimately failed to demonstrate a difference in the timing of antibiotics. Our study demonstrated an overall >9-hour median difference in the time to antibiotics between the two protocols with no difference in outcomes. This difference in antibiotic timing is significantly larger than that observed in the PHANTASi trial and is much longer than that suggested by the current SSC guidelines.9,26

While our limited findings did not demonstrate a difference in outcomes, we do not advocate changing practice based on this unpowered pilot, nor do we advise purposely withholding antibiotics in cases where an infection is clear. Surgical ICU and TICU patients are subject to numerous stressors that predispose them to infections from numerous sources including wounds, invasive catheters, and pneumonia. However, many of these same stressors make it difficult to diagnose pneumonia specifically or infections in general. Indeed, one study found that many critically ill patients with suspected sepsis were unlikely to have ever had an infection at all.16 Diagnostic uncertainty is an inherent, but frequently ignored, aspect of the treatment of pneumonia and sepsis.13 Our limited findings suggest, as the current SSC guidelines advise, that a short period of additional diagnostic workup may be safe in selected low-risk patients who do not require vasopressors.9 In cases where the diagnosis of infection is certain, antibiotics should be started immediately. A recent study found that, while the timing of antibiotics was not associated with mortality, even for those in shock, delays in antibiotics were associated with an increased risk of progression to septic shock in some patients.19 Avoiding antibiotic delays for those patients in whom antibiotics may be lifesaving, while constraining their use in those who cannot benefit, may require the application of biomarkers with the capability of specifically detecting the host response to infection, such as monocyte distribution width measurements or gene expression diagnostic assays.35–40

Our study is strengthened by its inclusion of multiple centers and the cluster-randomized crossover design. We also demonstrate applicability in a real-world setting. Our study also has several limitations worth considering. As a pilot trial, we did not conduct a formal power analysis and are unable to demonstrate small differences in outcomes. In addition, we did not control the choice of empiric antibiotics, the duration of antibiotics in the setting of pneumonia, or which patients should have been cultured. It is possible that practice variability between centers regarding the aggressiveness of obtaining cultures as well as variation in antibiotic choice or duration may introduce bias, although this may be mitigated somewhat by the crossover design. We also did not record data on specific antibiotics used and are unable to analyze outcomes based on antibiotic choice. However, when the antibiotic choices themselves are not the subject of the study, there is precedent for not recording these data.41 We also do not have outcomes based on specific diagnostic sampling method (BAL, bronchial wash, or mini-BAL). In addition, the participating centers are not similarly sized, resulting in significant variability in enrollment per center. One center has a large proportion of the overall enrollment. This enrollment variability may also be a source of bias, but the influence of this bias on our ability to assess our intervention should be mitigated to a degree with the crossover design. Also, we did not include median days from intubation to culture and so are unable to provide a breakdown of outcomes for hospital-acquired versus ventilator-associated pneumonia. Finally, our population consists overwhelmingly of trauma patients and may not be broadly generalizable to a nontrauma population (Supplemental Digital Content, Supplementary Data 1,


This multicenter cluster-randomized crossover pilot trial demonstrates that a protocol of specimen-initiated antibiotic administration can be successfully applied to a population of intubated TICU and SICU patients who are suspected of having pneumonia but do not require vasopressors. We have demonstrated a high degree of protocol adherence and no differences in outcomes compared with a protocol of immediate antibiotic administration despite significant differences in the timing of antibiotics. We further demonstrated successful avoidance of antibiotics in a small subset of patients in the specimen-initiated protocol. Although in terms of outcomes we are unable to assert equivalence between our two protocols based on sample size, our data do suggest adequate equipoise to allow additional larger studies to verify our results in a wider population of patients.


C.A.G. and R.G.S. contributed in the literature review, concept, and study design. C.A.G., R.T.B., C.M.W., R.G.S., L.A., M.D., L.C.C., G.P.P., S.V., B.N.H., and J.C.O. contributed in the data collection. C.A.G. and L.C.H. contributed in the data analysis. C.A.G., R.T.B., C.M.W., R.G.S., L.C.H., and S.Q.S. contributed in the interpretation of results. C.A.G. contributed in the manuscript writing. All authors contributed in the critical review of manuscript.


C.A.G. received a grant from the Surgical Infection Society Foundation for this study. The remaining authors declare no conflicts of interest.


1. Vincent JL, Sakr Y, Singer M, Martin-Loeches I, Machado FR, Marshall JC, et al. Prevalence and outcomes of infection among patients in intensive care units in 2017. JAMA. 2020;323(15):1478–1487.
2. Eguia E, Cobb AN, Baker MS, Joyce C, Gilbert E, Gonzalez R, et al. Risk factors for infection and evaluation of Sepsis-3 in patients with trauma. Am J Surg. 2019;218:851–857.
3. Leonard KL, Borst GM, Davies SW, Coogan M, Waibel BH, Poulin NR, et al. Ventilator-associated pneumonia in trauma patients: different criteria, different rates. Surg Infect (Larchmt). 2016;17(3):363–368.
4. Krebs ED, Hassinger TE, Guidry CA, Berry PS, Elwood NR, Sawyer RG. Non-utility of sepsis scores for identifying infection in surgical intensive care unit patients. Am J Surg. 2019;218(2):243–247.
5. Piriyapatsom A, Lin H, Pirrone M, De Pascale G, Corona De Lapuerta J, Bittner EA, et al. Evaluation of the infection-related ventilator-associated events algorithm for ventilator-associated pneumonia surveillance in a trauma population. Respir Care. 2016;61(3):269–276.
6. Kalil AC, Metersky ML, Klompas M, Muscedere J, Sweeney DA, Palmer LB, et al. Management of adults with hospital-acquired and ventilator-associated pneumonia: 2016 clinical practice guidelines by the Infectious Diseases Society of America and the American Thoracic Society. Clin Infect Dis. 2016;63(5):e61–e111.
7. Torres A, Niederman MS, Chastre J, Ewig S, Fernandez-Vandellos P, Hanberger H, et al. International ERS/ESICM/ESCMID/ALAT guidelines for the management of hospital-acquired pneumonia and ventilator-associated pneumonia: Guidelines for the management of hospital-acquired pneumonia (HAP)/ventilator-associated pneumonia (VAP) of the European Respiratory Society (ERS), European Society of Intensive Care Medicine (ESICM), European Society of Clinical Microbiology and Infectious Diseases (ESCMID) and Asociacion Latinoamericana del Torax (ALAT). Eur Respir J. 2017;50(3):1700582.
8. Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, et al. Surviving Sepsis Campaign: international guidelines for management of sepsis and septic shock: 2016. Crit Care Med. 2017;45(3):486–552.
9. Evans L, Rhodes A, Alhazzani W, Antonelli M, Coopersmith CM, French C, et al. Surviving Sepsis Campaign: international guidelines for management of sepsis and septic shock 2021. Crit Care Med. 2021;49(11):e1063–e1143.
10. Nicks BA, Manthey DE, Fitch MT. The Centers for Medicare and Medicaid Services (CMS) community-acquired pneumonia core measures lead to unnecessary antibiotic administration by emergency physicians. Acad Emerg Med. 2009;16(2):184–187.
11. Rhee C, Strich JR, Klompas M, Yealy DM, Masur H. SEP-1 has brought much needed attention to improving sepsis care...but now is the time to improve SEP-1. Crit Care Med. 2020;48(6):779–782.
12. Rhee C, Chiotos K, Cosgrove SE, Heil EL, Kadri SS, Kalil AC, et al. Infectious Diseases Society of America position paper: recommended revisions to the national severe sepsis and septic shock early management bundle (SEP-1) sepsis quality measure. Clin Infect Dis. 2021;72(4):541–552.
13. Prescott HC, Iwashyna TJ. Improving sepsis treatment by embracing diagnostic uncertainty. Ann Am Thorac Soc. 2019;16(4):426–429.
14. Guidry CA, Sawyer RG, Winfield RD. Challenging the dogma of aggressive initiation of antibiotics in sepsis. Surg Infect (Larchmt). 2021;22(5):473–476.
15. Croce MA, Swanson JM, Magnotti LJ, Claridge JA, Weinberg JA, Wood GC, et al. The futility of the clinical pulmonary infection score in trauma patients. J Trauma. 2006;60(3):523–527 discussion 7–8.
16. Klein Klouwenberg PM, Cremer OL, van Vught LA, Ong DS, Frencken JF, Schultz MJ, et al. Likelihood of infection in patients with presumed sepsis at the time of intensive care unit admission: a cohort study. Crit Care. 2015;19:319.
17. Fernando SM, Tran A, Cheng W, Klompas M, Kyeremanteng K, Mehta S, et al. Diagnosis of ventilator-associated pneumonia in critically ill adult patients—a systematic review and meta-analysis. Intensive Care Med. 2020;46(6):1170–1179.
18. Seymour CW, Gesten F, Prescott HC, Friedrich ME, Iwashyna TJ, Phillips GS, et al. Time to treatment and mortality during mandated emergency care for sepsis. N Engl J Med. 2017;376(23):2235–2244.
19. Bisarya R, Song X, Salle J, Liu M, Patel A, Simpson SQ. Antibiotic timing and progression to septic shock among patients in the ED with suspected infection. Chest. 2022;161(1):112–120.
20. Allaouchiche B, Jaumain H, Chassard D, Bouletreau P. Gram stain of bronchoalveolar lavage fluid in the early diagnosis of ventilator-associated pneumonia. Br J Anaesth. 1999;83(6):845–849.
21. Duflo F, Allaouchiche B, Debon R, Bordet F, Chassard D. An evaluation of the Gram stain in protected bronchoalveolar lavage fluid for the early diagnosis of ventilator-associated pneumonia. Anesth Analg. 2001;92(2):442–447.
22. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377–381.
23. Harris PA, Taylor R, Minor BL, Elliott V, Fernandez M, O’Neal L, et al. The REDCap consortium: building an international community of software platform partners. J Biomed Inform. 2019;95:103208.
24. Levy MM, Rhodes A, Phillips GS, Townsend SR, Schorr CA, Beale R, et al. Surviving Sepsis Campaign: association between performance metrics and outcomes in a 7.5-year study. Crit Care Med. 2015;43(1):3–12.
25. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA. 2016;315(8):801–810.
26. Alam N, Oskam E, Stassen PM, Exter PV, van de Ven PM, Haak HR, et al. Prehospital antibiotics in the ambulance for sepsis: a multicentre, open label, randomised trial. Lancet Respir Med. 2018;6(1):40–50.
27. Schinkel M, Paranjape K, Kundert J, Nannan Panday RS, Alam N, Nanayakkara PWB. Towards understanding the effective use of antibiotics for sepsis. Chest. 2021;160(4):1211–1221.
28. Hranjec T, Rosenberger LH, Swenson B, Metzger R, Flohr TR, Politano AD, et al. Aggressive versus conservative initiation of antimicrobial treatment in critically ill surgical patients with suspected intensive-care-unit-acquired infection: a quasi-experimental, before and after observational cohort study. Lancet Infect Dis. 2012;12(10):774–780.
29. Le Terrier C, Vinetti M, Bonjean P, Richard R, Jarrige B, Pons B, et al. Impact of a restrictive antibiotic policy on the acquisition of extended-spectrum beta-lactamase-producing Enterobacteriaceae in an endemic region: a before-and-after, propensity-matched cohort study in a Caribbean intensive care unit. Crit Care. 2021;25(1):261.
30. Ho AMH, Phelan R, Mizubuti GB, Murdoch JAC, Wickett S, Ho AK, et al. Bias in before-after studies: narrative overview for anesthesiologists. Anesth Analg. 2018;126(5):1755–1762.
31. Bloos F, Thomas-Ruddel D, Ruddel H, Engel C, Schwarzkopf D, Marshall JC, et al. Impact of compliance with infection management guidelines on outcome in patients with severe sepsis: a prospective observational multi-center study. Crit Care. 2014;18(2):R42.
32. Abe T, Ogura H, Shiraishi A, Kushimoto S, Saitoh D, Fujishima S, et al. Characteristics, management, and in-hospital mortality among patients with severe sepsis in intensive care units in Japan: the FORECAST study. Crit Care. 2018;22(1):322.
33. van Zanten AR, Brinkman S, Arbous MS, Abu-Hanna A, Levy MM, de Keizer NF, et al. Guideline bundles adherence and mortality in severe sepsis and septic shock. Crit Care Med. 2014;42(8):1890–1898.
34. Bloos F, Ruddel H, Thomas-Ruddel D, Schwarzkopf D, Pausch C, Harbarth S, et al. Effect of a multifaceted educational intervention for anti-infectious measures on sepsis mortality: a cluster randomized trial. Intensive Care Med. 2017;43(11):1602–1612.
35. Agnello L, Sasso BL, Giglio RV, Bivona G, Gambino CM, Cortegiani A, et al. Monocyte distribution width as a biomarker of sepsis in the intensive care unit: a pilot study. Ann Clin Biochem. 2021;58(1):70–73.
36. Polilli E, Frattari A, Esposito JE, Stanziale A, Giurdanella G, Di Iorio G, et al. Monocyte distribution width (MDW) as a new tool for the prediction of sepsis in critically ill patients: a preliminary investigation in an intensive care unit. BMC Emerg Med. 2021;21(1):147.
37. Marcos-Morales A, Barea-Mendoza JA, Garcia-Fuentes C, Cueto-Felgueroso C, Lopez-Jimenez A, Martin-Loeches I, et al. Elevated monocyte distribution width in trauma: an early cellular biomarker of organ dysfunction. Injury. 2022;53(3):959–965.
38. Sweeney TE, Khatri P. Benchmarking sepsis gene expression diagnostics using public data. Crit Care Med. 2017;45(1):1–10.
39. Miller RR 3rd, Lopansri BK, Burke JP, Levy M, Opal S, Rothman RE, et al. Validation of a host response assay, SeptiCyte LAB, for discriminating sepsis from systemic inflammatory response syndrome in the ICU. Am J Respir Crit Care Med. 2018;198(7):903–913.
40. Verboom DM, Koster-Brouwer ME, Ruurda JP, van Hillegersberg R, van Berge Henegouwen MI, Gisbertz SS, et al. A pilot study of a novel molecular host response assay to diagnose infection in patients after high-risk gastro-intestinal surgery. J Crit Care. 2019;54:83–87.
41. Sawyer RG, Claridge JA, Nathens AB, Rotstein OD, Duane TM, Evans HL, et al. Trial of short-course antimicrobial therapy for intraabdominal infection. N Engl J Med. 2015;372(21):1996–2005.

Antibiotic timing; pneumonia; sepsis; randomized clinical trial

Supplemental Digital Content

Copyright © 2022 American Association for the Surgery of Trauma.