Pre-Hospital Aspiration Is Associated with Increased Pulmonary Complications

Abstract

Background:

Rates of ventilator-associated pneumonia (VAP) are highest among patients intubated on an emergency basis following trauma. We reported previously a retrospective analysis demonstrating an association between subjective aspiration and VAP after pre-hospital intubation. We hypothesize that by directing paramedics to note features of aspiration at intubation, we will confirm prospectively the association between pre-hospital aspiration and subsequent pneumonia in trauma patients.

Methods:

Paramedics collected data regarding aspiration at the time of intubation. All intubated patients admitted to a level 1 trauma center intensive care unit (ICU) were included. Data comprised a clinical impression of pre-hospital aspiration, as well as the presence and timing of blood and emesis in the airway. Injury severity, co-morbidities, and outcomes were collected from the trauma registry. Healthcare-associated pneumonia (HAP) was identified by medical record review of both bronchoalveolar lavage culture results and discharge diagnosis. Descriptive statistics and univariate analysis of outcomes by aspiration status, as well as covariable adjustment using propensity scores, were performed.

Results:

Of the 228 patients, 89 (39%) were determined by paramedics to have aspirated. The majority of those who aspirated (84 [94%]) did so prior to intubation. Patients who aspirated had higher Injury Severity Scores than those who did not aspirate (25.0±1.7 vs. 21.9±1.5 points; p=0.04) and lower preintubation Glasgow Coma Scale scores (8.2±0.50 vs. 9.6±0.40; p=0.02). Of the 89 patients who aspirated around the time of intubation, 14 (16%) developed HAP vs. five (3.6%) of those who did not aspirate (p<0.01). We observed non-significant increases in mortality rate, ICU length of stay (LOS) and duration of mechanical ventilation after aspiration (deaths: 21 [23.6%] vs. 23 [16.6%]; p=0.19; ICU LOS: 5.3±0.9 vs. 4.1±0.5 days; p=0.13; duration of mechanical ventilation: 5.3±1.2 vs. 3.2±0.5 days; p=0.10).

Conclusions:

Aspiration prior to intubation was reported commonly by paramedics and was associated with a higher risk of HAP.

Compared with other critically ill populations, trauma patients are at higher risk of developing pneumonia during their hospitalizations [1]. The regular need for emergency intubation is one factor that may contribute to this elevated risk [2,3]. Patients who are intubated on an emergency basis in the pre-hospital setting develop ventilator-associated pneumonia (VAP) more frequently than do those intubated on an emergency basis in the emergency department [4,5]. However, there is substantial variability in the provision of pre-hospital care for trauma patients, specifically in the frequency and success of intubation [6,7]. In the current study community, there is a high rate of pre-hospital first-attempt intubation success, with similar rates of VAP in patients intubated in the pre-hospital and emergency department settings [8 –10].

The partnership between emergency medical services (EMS) providers and trauma care leaders in Seattle and King County has enabled rigorous investigation of pre-hospital intubation in trauma patients. We reported previously the outcomes of trauma patients who underwent pre-hospital intubation by paramedics. Patients who were noted in pre-hospital records to have aspirated gastric contents around the time of intubation had higher subsequent rates of VAP than did those patients who had clear airways around the time of intubation. However, there was a substantial quantity of missing data, which may have introduced selection bias [11].

Therefore, we undertook a prospective investigation to delineate further the characteristics of pre-hospital aspiration around the time of intubation, as well as to confirm the correlation observed previously between pre-hospital aspiration and pneumonia.

Patients and Methods

Paramedics collected data prospectively on all patients who were intubated prior to admission to our Level I trauma center between May 2010 and December 2011. In addition to standard clinical documentation, the Emergency Medical Services (EMS) responders recorded the presence of blood and emesis in the patients' oropharynx prior to intubation; the responders' perception of aspiration during intubation; and confirmation of tracheal aspiration via the presence of blood or emesis suctioned from the endotracheal tube. Because there is no standard way currently to confirm aspiration, we defined aspiration by any or all of the following criteria: (1) Blood or emesis in the patient's oropharynx prior to intubation; (2) blood or emesis in the oropharynx during intubation; or (3) blood or emesis suctioned from the endotracheal tube.

All adult trauma patients (age ≥18 y) who were admitted subsequently to the intensive care unit (ICU) were included in the study. Baseline characteristics, injury information, physiologic variables, and outcome data were determined from the hospital's trauma registry and medical record review. The primary outcomes were the development of healthcare-associated pneumonia (HAP), ventilator-associated pneumonia (VAP), and in-hospital death. The U.S. Centers for Disease Control and Prevention definitions were employed [12]. Healthcare-associated pneumonia was defined as pneumonia developing at least 48 h after admission, whereas VAP required a minimum of 48 h of mechanical ventilation beforehand. It is standard protocol in our ICU to use bronchoalveolar lavage (BAL) to diagnose VAP [12]. Secondary outcomes were hospital and ICU length of stay (LOS), number of days of mechanical ventilation, and discharge to home from the hospital.

Patients who had clinical evidence of aspiration were compared with those who did not using univariate analysis, assessing for differences in baseline characteristics, injury severity, and primary and secondary outcomes. Chi-square tests, Fisher exact tests, Student t-tests, and Wilcoxon rank sum tests were utilized as appropriate. Additionally, we used an exact logistic regression method to estimate the unadjusted odds ratio (OR) of developing HAP for each type of aspirate (blood, emesis, or both).

Propensity scores are a statistical method used to reduce the selection bias that is inherent in observational studies, whereby patient characteristics influence whether they are exposed to an event of interest [13,14]. As the event of interest in this study was aspiration, we calculated propensity scores to adjust for the baseline likelihood that a patient would aspirate. Based on available literature and expert opinion, characteristics were selected a priori that have been postulated to affect a person's risk of aspiration (gender, age, use of etomidate for intubation, blood alcohol concentration, Glasgow Coma Scale (GCS) score, maximum Abbreviated Injury Scale score for head injury, diagnosis of diabetes mellitus, and presence of blunt trauma). Propensity scores for aspiration were estimated using these factors as independent variables in a logistic regression model. The propensity score was then used as a covariate adjustor in a logistic regression model of the association between potential aspiration and HAP [13]. The Hosmer and Lemeshow goodness-of-fit test was performed for the two models that included the propensity score. All statistical analyses were performed using Stata/SE version 11.0 (StataCorp, College Station, TX).

Results

There were 280 patients who were intubated in the field by paramedics after traumatic injury. We excluded 52 who either were not locatable in the trauma registry (29 patients), were under 18 years of age (one patient), or who died or were extubated in the emergency department (22 patients). Two hundred twenty-eight subjects remained.

Eighty-nine patients, or 39% of the population, had clinical evidence of aspiration around the time of pre-hospital intubation (Table 1). Of the 78 patients who were believed to have aspirated blood, 74 (95%) did so prior to intubation. Similarly, of the 22 patients who likely aspirated gastric contents, 17 (77%) did so prior to intubation

Table 1.

Baseline Characteristics and Severity of Injury

	No Aspiration (n=139)	Aspiration (n=89)	p
Percentage	61	39
Age (mean±standard deviation)	40.9±1.5	41.4±1.9	0.75
Male (%)	102 (73.4)	68 (76.4)	0.61
Mean blood alcohol concentration (mmol/L)	105.8±11.2	63.6±11.0	0.009
Comorbidities (%)
Obesity	4 (2.9)	3 (3.4)	1.0
Hypertension	15 (10.8)	16 (18.0)	0.12
Diabetes mellitus	3 (2.2)	8 (9.0)	0.02
Smoker	16 (11.5)	5 (5.6)	0.16
Respiratory disease	5 (3.6)	3 (3.4)	1.0
Neurological deficit	2 (1.4)	2 (2.2)	0.64
Psychiatric condition	25 (18.0)	8 (9.0)	0.06
Alcoholism	9 (6.5)	3 (3.4)	0.38
Drug abuse	15 (10.8)	4 (4.5)	0.14
History of trauma	13 (9.4)	11 (12.4)	0.47
ISS (mean)	21.9±1.5	25.0±1.7	0.04
GCS at scene (mean)	9.6±0.42	8.2±0.46	0.02
Max head AIS (mean)	1.6±0.16	2.8±0.21	<0.001

AIS=Abbreviated Injury Scale score; GCS=Glasgow Coma Scale score; ISS=Injury Severity Score; Max=maximum.

Patients who had clinical evidence of aspiration were similar to those who did not in terms of age, gender, and comorbidities, except for diabetes mellitus (Table 1). However, aspiration around the time of pre-hospital intubation was associated with higher severe injury, as demonstrated by higher Injury Severity Scores (ISSs), higher maximum head Abbreviated Injury Scale scores, and lower Glasgow Coma Scale (GCS) scores. Interestingly, blood alcohol concentrations were on average lower in those patients with clinical evidence of aspiration (Table 1).

Clinical evidence of aspiration was associated with a four-fold increase in the risk of HAP and more than three times the risk of VAP (Table 2). Aspiration was associated with a non-significant increase in the risk of in-hospital death. Compared with patients who did not aspirate, those with clinical evidence of aspiration showed a trend toward longer ICU LOS, more days of mechanical ventilation, and decreased likelihood of discharge to home. However, these differences were not statistically significant. There were no differences in hospital LOS (Table 2).

Table 2.

Primary and Secondary Outcomes

	No Aspiration (n=139)	Aspiration (n=89)	p
HAP (%)	5 (3.6)	14 (15.7)	0.002
VAP (%)	4 (2.9)	10 (11.2)	0.01
In-hospital death (%)	23 (16.6)	21 (23.6)	0.19
ICU LOS (mean days)	4.1±0.5	5.3±0.9	0.13
Ventilator days (mean)	5.3±1.2	3.2±0.5	0.10
Disposition to home (%)	101 (72.7)	58 (65.2)	0.15
Hospital LOS (mean days)	9.1±1.0	12.5±2.5	0.46

HAP=healthcare-associated pneumonia; ICU=intensive care unit; LOS=length of stay; VAP=ventilator-associated pneumonia.

Of the 19 cases of HAP that were diagnosed in the study population, four were considered early pneumonias (diagnosed on or before hospital day 4), whereas the remaining 15 cases were late pneumonias. Twelve cases were diagnosed using qualitative bacterial cultures, and 16 organisms were isolated. The predominant organism cultured was Staphylococcus aureus, although gram-negative bacilli also were common (Table 3).

Table 3.

Timing and Bacterial Etiology of Healthcare-Associated Pneumonias Diagnosed by Qualitative Cultures

Organism	Early (≤4 d)	Late (>4 d)
Staphylococcus aureus	1	5^a
Streptococcus pneumoniae	0	2^a
Enterobacter spp.	0	3^b
Haemophilus influenzae	0	2
Escherichia coli	0	1
Serratia marcescens	1	0
Stenotrophomonas maltophilia	0	1
All gram-negative bacilli	1	7

Note: Twelve cases of pneumonia grew 16 organisms.

One culture in this group was from an uncontaminated endotracheal aspirate sample.

One culture from this group was from an uncontaminated sputum sample.

Note: Seven cases were diagnosed clinically using the U.S. Centers for Disease Control and Prevention criteria and therefore are not represented here.

Patients who aspirated had a higher risk of developing pneumonia, regardless of the material that was aspirated (blood, emesis, or both). This association was strongest for those who were presumed to have aspirated emesis alone (Table 4).

Table 4.

Development of Pneumonia as a Function of the Nature of the Aspirate

	No HAP (n=209)	HAP (n=19)	OR (CI)	P value
No aspiration (%)	134 (64.1)	5 (26.3)	–	–
Aspiration of blood (%)	59 (28.2)	8 (42.1)	3.6 (1.0–14.6)	0.051
Aspiration of emesis (%)	7 (3.4)	4 (21.0)	14.6 (2.4–87.4)	0.003
Aspiration of blood+emesis (%)	9 (4.3)	2 (10.5)	5.8 (0.5–42.4)	0.168

All odds ratios are calculated relative to the No aspiration group.

CI=confidence interval; HAP=healthcare-associated pneumonia, OR=odds ratio.

An unadjusted logistic regression model found a five-fold increase in the risk of developing HAP among those people who had clinical evidence of aspiration (Table 5). When the propensity for aspiration was used subsequently as a covariate adjustor in the HAP model, the odds ratio for developing HAP, given potential aspiration, remained elevated; as was the case when the ISS was added to the model (Table 5). Both adjusted models were well fit to the data.

Table 5.

Logistic Regression Models for Healthcare-Associated Pneumonia (HAP) Given the Possible Presence of Aspiration

Model of HAP	OR (95 % CI)	p	Goodness of fit (χ²)	p
Unadjusted model	5.0 (1.7–14.4)	0.003	–	–
Adjusted for propensity score	3.4 (1.0–11.2)	0.048	4.99	0.76
Adjusted for propensity score and ISS	4.7 (1.2–16.8)	0.017	4.70	0.79

CI=confidence interval; HAP=healthcare-associated pneumonia; ISS=Injury Severity Score; OR=odds ratio.

Discussion

Although aspiration around the time of intubation has been studied in the operating room and emergency department, little systematic information is available from the pre-hospital setting. To our knowledge, this study describes the most detailed data regarding aspiration around the time of pre-hospital intubation among seriously injured individuals. We found that more than one-third of patients with serious trauma had clinical evidence of aspiration prior to hospitalization, and that the majority of such patients were judged to have aspirated prior to intubation. Furthermore, we confirmed that those patients who aspirated were at higher risk of developing both healthcare- and ventilator-associated pneumonia.

Aspiration may lead to a variety of clinical outcomes, depending on factors such as the patient's susceptibility, the timing of aspiration, and the volume and nature of the aspirate [15]. The majority of pneumonias that developed in this study were considered late-onset, being diagnosed five or more days after admission. However, this does not preclude the possibility that aspiration was a predisposing factor. In some cases, inoculation secondary to aspiration may take several days to present as pneumonia [16]. Alternatively, elements of the aspirate other than bacteria (such as digestive enzymes and gastric acid) may play a role in predisposing patients to pulmonary infections.

Bacterial colonization of the oropharynx, whether from oral or gastrointestinal sources, is a risk factor for the development of HAP. Concordance has been demonstrated between oropharyngeal flora and causative organisms of HAP [17]. Early colonization of the oropharynx in hospitalized patients tends to be predominated by “community acquired” organisms such as S. aureus, Streptococcus pneumoniae, and Haemophilus influenzae. On the other hand, as time passes, an increasing number of gram-negative species colonize the upper airways, originating from both the oral cavity and the gastrointestinal tract [18,19]. This study demonstrated the propensity for gram-negative bacteria to cause late-onset pneumonias, with such organisms being cultured almost as frequently as those that are considered “community acquired.” Unfortunately, there were very few cases of early pneumonia, prohibiting a robust comparison of the organisms and the timing of infection.

Currently, the Society for Healthcare Epidemiology of America/Infectious Diseases Society of America recommendations for prevention of pneumonia include maintaining the patient in a semi-recumbent position, performing daily sedation interruption and ventilator weaning assessments, and regular administration of antiseptic oral care [20]. The U.S. Centers for Disease Control and Prevention have advocated the use of ventilator bundles, including these elements, to achieve a high degree of compliance with such prevention measures, and the rate of bundle adoption as a quality-improvement measure has been demonstrated to be associated with a reduction in VAP rates [21]. However, in critically ill trauma patients, the impact of the ventilator bundle recently was called into question [22]. The results of the current study raise the possibility that in at least some trauma patients, early aspiration may initiate a disease process that cannot be stopped by the interventions that are part of the current ventilator bundle. This perspective is supported by the only other published study in which paramedics judged the timing of peri-intubation aspiration. In the San Diego Paramedic Rapid Sequences Induction (RSI) Trial involving severely head-injured trauma patients, the majority of those who were believed to have aspirated did so before the arrival of rescue personnel [23].

Pneumonias that develop as a result of aspiration therefore may be considered part of the spectrum of traumatic injury, attributable to patient disease rather than healthcare interventions. There are potential implications for reimbursement that emerge from this proposition. Moreover, if a substantial proportion of trauma patients are aspirating prior to care by EMS providers, interventions are needed that focus on mitigating the impact of such events. For instance, Minshall et al. performed early (≤48 h after intubation) screening nonbronchoscopic BAL on all trauma and emergency general surgery patients admitted to the surgical ICU. They found that such screening was positive for meaningful bacterial growth in almost one-half of their patients. In addition, not only did they demonstrate that pneumonia developed more frequently in patients who had a positive compared with a negative screening BAL, but in those patients who developed pneumonia subsequently, the organism cultured from screening BAL was highly predictive of the organism causing the pneumonia [16]. Although it may not be feasible to perform screening BALs in all intubated trauma patients, a selective strategy may enable early identification of inoculated organisms in patients at high risk (such as those with witnessed aspiration). Such a practice could facilitate timely and appropriate antibiotic treatment if infection were to develop subsequently.

Several limitations of this study must be noted. The first is the inherent difficulty in assessing whether aspiration has occurred [24]. We acknowledge that the presence of foreign material in the oropharynx or airway around the time of intubation is not proof of pulmonary aspiration. However, the insertion of a foreign body (endotracheal tube) through a grossly contaminated oropharynx may facilitate the introduction of potentially pathogenic bacteria and injurious oral and gastric secretions into the respiratory tract, thus conferring a higher risk for subsequent pneumonia. Other groups have confirmed pulmonary aspiration by obtaining bronchoalveolar lavage samples containing gastrointestinal enzymes such as amylase [25]. In order to address this issue further, we plan to assess whether radiographic evidence of aspiration on initial chest computed tomography scan correlates with clinical evidence of aspiration.

We also recognize the challenge of performing research in emergency situations [26 –28]. Healthcare providers are rightly focused on clinical care rather than documentation for research purposes. As we saw in our previous study involving paramedics and emergency intubation, such circumstances can lead to a substantial problem with missing data. In partnership with the Seattle and King County EMS programs, we were able to assemble more comprehensive paramedic data regarding field intubation. For the variables used in this study, data completion rates ranged from 90%–100% (205–228 observations). Despite the challenges, research in emergency settings is feasible. In the future, electronic transmission of real-time data, including video assessment of the scene or intubation, may provide additional unencumbered data collection for quality improvement and research.

Finally, our study population was small. We chose to focus on trauma patients, rather than including all patients having emergency intubation in the pre-hospital setting because of their high risk of developing pneumonia while hospitalized [1]. This also was an attempt to exclude patients requiring pre-hospital intubation for respiratory failure caused by pneumonia. Further characterization of the factors that place trauma patients at such high risk is necessary to decrease the incidence of these complications.

In the future, we hope to gather corroboration for the results of this study in larger populations, including examining patients outside the local pre-hospital setting. We plan to explore innovative ways in which to assess the presence or absence of aspiration in trauma patients with the examination of early computed tomography scans. The introduction of novel interventions to decrease the morbidity associated with pre-hospital aspiration is essential, particularly in cases where aspiration may not be preventable.

Conclusions

We found that pre-hospital aspiration around the time of intubation is common in trauma patients, occurring in as many as 40% of cases. In a large majority of patients, clinical evidence of aspiration was present prior to interventions by paramedics; such aspiration was associated with a substantial increase in HAP. Management that focuses on decreasing the morbidity of aspiration, such as early recognition and treatment of infection, may be necessary in such patients.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Cook

, Norwood

, Berne

. Ventilator-associated pneumonia is more common and of less consequence in trauma patients compared with other critically ill patients. J Trauma Injury Infect Crit Care, 2010; 69:1083–1091.

Rodriguez

, Gibbons

, Bitzer

, et al. Pneumonia: Incidence, risk factors, and outcome in injured patients. J Trauma Injury Infect Crit Care, 1991; 31:907–914.

Cavalcanti

, Ferrer

, et al. Risk and prognostic factors for ventilator-associated pneumonia in trauma patients. Crit Care Med, 2006; 34:1067–1072.

Sloane

, Vilke

, Chan

, et al. Rapid sequence intubation in the field versus hospital in trauma patients. J Emerg Med, 2000; 19:259–264.

Eckert

, Davis

, Reed II

, et al. Urgent airways after trauma: Who gets pneumonia. J Trauma Injury Infect Crit Care, 2004; 57:750–755.

Bulger

, Nathens

, Rivara

, et al. National variability in out-of-hospital treatment after traumatic injury. Ann Emerg Med, 2007; 49:293–301.

Hubble

, Brown

, Wilfong

, et al. A meta-analysis of pre-hospital airway control techniques 1: Orotracheal and nasotracheal intubation success rates. Pre-hospital Emerg Care, 2010; 14:377–401.

Bulger

, Copass

, Maier

, et al. An analysis of advanced pre-hospital airway management. J Emerg Med, 2002; 23:183–189.

Warner

, Sharar

, Copass

, Bulger

. Pre-hospital management of the difficult airway: A prospective cohort study. J Emerg Med, 2009; 36:257–265.

10.

Evans

, Zonies

, Warner

, et al. Timing of intubation and ventilator-associated pneumonia following injury. Arch Surg, 2010; 145:1041–1046.

11.

Evans

, Warner

, Bulger

, et al. Pre-hospital intubation factors and pneumonia in trauma patients. Surg Infect, 2011; 12:1–6.

12.

Surveillance definition of healthcare-associated infection and criteria for specific types of infections in the acute care setting 2013. Available at www.cdc.gov/nhsn/PDFs/pscManual/17pscNosInfDef_current.pdf. Accessed June 26, 2013.

13.

Austin

. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivar Behav Res, 2011; 46:399–424.

14.

Rosenbaum

, Rubin

. Reducing bias in observational studies using subclassification on the propensity score. J Am Stat Assoc, 1984; 79:516–524.

15.

Marik

. Aspiration pneumonitis and aspiration pneumonia. N Engl J Med, 2001; 344:665–671.

16.

Minshall

, Eriksson

, Hawkins

, et al. Early nonbronchoscopic bronchoalveolar lavage: Predictor of ventilator-associated pneumonia?. J Trauma Acute Care Surg, 2013; 74:448–453.

17.

Torres

, El-Ebiary

, Gonzalez

, et al. Gastric and pharyngeal flora in nosocomial pneumonia acquired during mechanical ventilation. Am Rev Respir Dis, 1993; 148:352–357.

18.

Ewig

, Torres

, El-Ebiary

, et al. Bacterial colonization patterns in mechanically ventilated patients with traumatic and medical head injury: Incidence, risk factors, and association with ventilator-associated pneumonia. Am J Respir Crit Care Med, 1999; 159:188–198.

19.

Wahl

, Zalewski

, Hemmila

. Pneumonia in the surgical intensive care unit: Is every one preventable?. Surgery, 2011; 150:665–672.

20.

Coffin

, Klompas

, Classen

, et al. Strategies to prevent ventilator-associated pneumonia in acute care hospitals. Infect Control Hosp Epidemiol, 2008; 29(Suppl 1):S31–S40.

21.

Resar

, Pronovost

, Haraden

, et al. Using a bundle approach to improve ventilator care processes and reduce ventilator-associated pneumonia. J Qual Pat Safety, 2005; 31:243–248.

22.

Croce

, Brasel

, Coimbra

, et al. National Trauma Institute prospective evaluation of the ventilator bundle in trauma patients: Does it really work?. J Trauma Acute Care Surg, 2013; 74:3543–3560.

23.

Vadeboncoeur

, Davis

, Ochs

, et al. The ability of paramedics to predict aspiration in patients undergoing pre-hospital rapid sequence intubation. J Emerg Med, 2006; 30:131–136.

24.

Raghavendran

, Nemzek

, Napolitano

, Knight

. Aspiration-induced lung injury. Crit Care Med, 2011; 39:818–826.

25.

Weiss

, Moazed

, DiBardino

, et al. Bronchoalveolar lavage amylase is associated with risk factors for aspiration and predicts bacterial pneumonia. Crit Care Med, 2013; 41:765–773.

26.

Brice

, Friend

, Delbridge

. Accuracy of EMS-recorded patient demographic data. Prehosp Emerg Care, 2008; 12:187–191.

27.

Thompson

. Ethical challenges of informed consent in pre-hospital research. Can J Emerg Med, 2003; 5:108–114.

28.

Leonard

, Scharff

, Koors

, et al. A qualitative assessment of factors that influence emergency medical services partnerships in pre-hospital research. Acad Emerg Med, 2012; 19:161–173.