Ventilator-associated pneumonia in trauma patients with open tracheotomy: Predictive factors and prognosis impact

Abstract

Objective:

To assess the predictive factors of ventilator associated pneumonia (VAP) occurrence following open tracheotomy in trauma patients.

Materials and Methods:

We conducted an observational, prospective study over 15 months, between 01/08/2010 and 30/11/2011. All trauma patients (except those with cervical spine trauma), older than 15 years, undergoing open tracheotomy during their ICU stay were included. All episode of VAP following tracheotomy were recorded. Predictive factor of VAP onset were studied.

Results:

We included 106 patients. Mean age was 37.9 ± 15.5 years. Mean Glasgow coma Scale (GCS) was 8.5 ± 3.7 and mean Injury Severity Score (ISS) was 53.1 ± 23.8. Tracheotomy was performed for 53 patients (50%) because of prolonged ventilation whereas 83 patients (78.3%) had tracheotomy because of projected long mechanical ventilation. Tracheotomy was performed within 8.6 ± 5.3 days. Immediate complications were bleeding events (22.6%) and barotrauma (0.9%). Late complications were stomal infection (28.3%) and VAP (52.8%). In multivariate analysis, independent factors predicting VAP onset were delayed tracheotomy (OR = 0.041; CI95% [1.02-7.87]; P = 0.041) and stomal infection (OR = 3.04; CI95% [1.02-9.93]; P = 0.045). Thirty three patients died in ICU (31.1%) without significant impact of VAP on mortality.

Conclusion:

Late tracheotomy and stomal infection are independent factors predicting VAP onset after open tracheotomy in trauma patients. The occurrence of VAP prolongers mechanical ventilation duration and intensive care unit (ICU) length of stay (LOS) but doesn’t increase mortality.

INTRODUCTION

Tracheotomy is one of the oldest and the most conducted procedures in critically ill patients.[1,2] This technique facilitates airway management and offers numerous advantages such as shortening the duration of mechanical ventilation, easier weaning from mechanical ventilation, avoiding translaryngeal complications…[3,4,5,6] Trauma patients represent a specific population who can potentially benefits of these advantages. In fact, during, the first hours of trauma, mechanical ventilation is always required for severe cases because of consciousness impairment, acute respiratory failure (chest trauma, aspiration…) or uncontrolled shock.[7] Secondly, even though satisfactory weaning parameters are reached, extubation may be delayed because of the inability of these patients to protect their airway or persistent impaired level of consciousness.[8] For all these reasons, tracheotomy was proposed as an alternative modality for airway management in trauma patients and gave encouraging results concerning reduction of mechanical ventilation duration and ICU length of stay.[9,10] Several studies have even suggested a decreased incidence of ventilator-associated pneumonia (VAP) with this procedure.[9,11] However, available data in the literature dealing with this infectious complication are conflicting because of heterogeneity of studied population: Different level of trauma severity, different modalities of tracheotomy procedure (open tracheotomy versus percutaneous tracheotomy) and different delay (early versus late tracheotomy).

The aim of the present study is to identify independent factors correlated with VAP onset in severe trauma patients undergoing open tracheotomy during their ICU stay.

MATERIALS AND METHODS

Study design

This prospective observational study was conducted over 15 months, between 01/08/2010 and 30/11/2011 in our medical surgical intensive care unit (HabibBourguiba University Hospital – Sfax – Tunisia). Our study was approved by our local ethic board.

Patients

All trauma patients older than 15 years and undergoing open tracheotomy during their ICU stay were eligible for our study. We excluded patients meeting at least one of the following criteria:

• Patients who developed VAP prior to tracheotomy.

• Tracheotomy performed for urgent airway management.

• Patients with cervical spine trauma.

• Patients who undergone percutaneous tracheotomy.

During the study period, 386 adult trauma patients were admitted in our ICU. All of them were ventilated either on the scene of the accident or at their arrival to our emergency department. During their ICU stay, 106 (27.5%) needed tracheotomy and were included in our study. For each included patient, the following data were recorded on admission:

• Demographic parameters: Age, sex, previous medical history.

• Clinical parameters: Blood pressure, heart rate, respiratory rate, neurological state (the consciousness was judged by the Glasgow Coma Scale (GCS)[12] calculated on the scene of the accident before sedation infusion, localization signs, motor deficiency…).

• Exhaustive chart of trauma lesions were obtained by imaging test. All our patients had undergone cranial, thoracic, abdominal and pelvic computed tomography. The trauma severity was assessed by the Injury Severity Score (ISS).[13]

• The clinical severity was assessed by the Simplified Acute Physiology Score (SAPS II)[14] and the Sequential Organ Failure Assessment (SOFA) score.[15]

• Therapeutic measures (mechanical ventilation, sedation, surgery, osmotherapy).

Ventilation and weaning modalities

Patients were kept under sedation and mechanical ventilation for at least 48 hours. Sedation withdrawal was decided if intracranial hypertension was controlled and if the patients had a good level of consciousness, a positive end expiratory pressure (PEEP) of 5 cm H₂O and 50% of oxygen. Then, A T-piece test is tried and extubation is performed within 1 hour if the test is successful. However, the trial was stopped if the patient developed a respiratory rate >30 per minute, SpO2 <95%, sweating, hypertension, shock or impairment of consciousness level. Extubated patients were considered as definitely weaned if off mechanical ventilation for more than 48 hours.

The decision regarding tracheotomy was made primary by the ICU team. Main indications of tracheotomy were prolonged intubation defined by a delay over 7 days, intubation projected to be longer than 14 days (according to clinical and radiological findings), difficult weaning from mechanical ventilation, failure of extubation and swallowing disorders. Patients who were projected to have prolonged ventilation but who were tracheotomized after more than 7 days of mechanical ventilation were considered to have 2 criteria for tracheotomy.

Diagnosis of ventilator-associated pneumonia

The diagnosis of ventilator-associated pneumonia was considered if the chest X-ray shows evidence of new and persistent (>48 hours) pulmonary infiltrates with at least two of the following criteria: Fever >38°C or hypothermia <35°C, White blood cells >12000/mm³ or <4000/mm³ , purulent respiratory secretion and/or worsening of respiratory insufficiency.[16,17] The bacteriological documentation was done by a broncho-alveolar lavage. A threshold of 10⁴ was required for considering the diagnosis of VAP.

Follow up

For each patient, we recorded the delay of successful weaning from mechanical ventilation regarding to tracheotomy (if off mechanical ventilation for more than 7 days) and the duration of mechanical ventilation free days. We also recorded the ICU length of stay (LOS) and late complications due to tracheotomy (stomal infection and VAP). For some patients, decanulation was tried when possible. The delay of successful decanulation was mentioned. ICU mortality rate was also recorded.

Statistical analysis

Included patients were divided into two groups: Those who developed VAP after tracheotomy (VAP (+) group) Vs those who didn’t experience this nosocomial event (VAP (−) group). Qualitative variables were expressed as percentages whereas quantitative variables were expressed as means ± standard deviation or medians. Univariate analysis comparing the two populations was conducted in order to check any baseline difference and to identify factors correlated to VAP onset. Qualitative variables were compared by using the Chi2 test or Fischer's exact test as appropriate whereas all quantitative variables were compared by the t-test. The normal distribution of quantitative variables was checked by the Kolmogorov-Smirnov test. All tests were two-sided. The level of significance was set at P < 0.05. Secondly, variables were subjected to a multivariate analysis with a logistic regression procedure and forward stepwise selection of P < 0.10 in order to identify independent factors predicting VAP onset in trauma patients with open tracheotomy. Odds ratios were calculated with 95% confidence intervals. SPSS version 18 was used for statistical analyses.

RESULTS

Mean age of our patients was 37.9 ± 15.5 years. Sex ratio was at 7.8. Ninety eight patients (89.6%) had no previous medical or surgical history. Mean SAPSII score was 37.8 ± 10.6. Mean SOFA score was 7.2 ± 2.6 and mean ISS score was 53.1 ± 23.8. On admission, 56 patients (52.8%) were hypertensive. Mean systolic blood pressure was 111.6 ± 19.5 mm Hg whereas diastolic blood pressure was 67.5 ± 15 mm Hg. Mean SpO₂ was 98.3 ± 3.7%. Mean GCS was 8.5 ± 3.7. Generalized seizures were reported in 3 patients (2.8%). Anisocoria was reported in 23 patients (21.7%) and agitation was reported in 4 patients (3.8%). According to clinical examination and imaging tests, 91 patients (85.8%) had cranial trauma, 60 patients (56.6%) had thoracic trauma, 21 patients (19.8%) had abdominal trauma whereas 10 patients (9.4%) had pelvic trauma. Biological findings on admission are summarized in [Table 1].

Table 1.

Clinical finding on admission

All our patients were sedated and intubated within the hours following trauma. Main indications for mechanical ventilation were consciousness impairment for 91 patients (95.8%), acute respiratory failure for 30 patients (28.3%) and uncontrolled shock for 8 patients (7.5%). All of them were ventilated according to volume controlled mode. Vasopressor support was required for 57 patients (53.8%). Fluid loading was performed in 65 patients (61.3%). Transfusion was indicated in 48 patients (45.3%). Amoxicilline + clavulanic acid was the most used antibiotic (99 patients) during the acute phase of trauma. Emergent surgery was performed for 57 patients (53.8%).

Tracheotomy indications

Difficult weaning – extubation failure

A T-piece test was performed only for 31 patients (29.2%) within 5.9 ± 4 days regarding mechanical ventilation (MV) initiation. The first trial was considered as successful only for 18 patients (17%) whereas 13 patients (12.3%) were resumed to mechanical ventilation. Extubation for at least one time was attempted in 21 patients (19.8%) within 6.24 ± 4.1 days and all of these attempts failed. For these patients, main causes of weaning and/or extubation failure were chest trauma in 13 cases (41.9%), neurological status worsening in 19 cases (61.3%), pulmonary embolism in 4 cases (12.9%), swallowing disorders in 4 patients (12.9%) and left ventricular failure in one case (3.2%).

For the remaining 75 patients, T-piece test was not performed because weaning criteria were not met.

Prolonged ventilation or projected prolonged ventilation

Tracheotomy was performed for 53 patients (50%) because of prolonged ventilation whereas 83 patients (78.3%) had tracheotomy because mechanical ventilation was projected to be long.

Tracheotomy: Complications and follow up

Open tracheotomy was performed within 8.6 ± 5.3 days (median at 7.5 days), ranging from 1 to 34 days. Fifty three patients (50%) had early tracheotomy whereas the remaining 53 patients (50%) had late tracheotomy. Immediate complications due to tracheotomy were bleeding in 24 patients (22.6%) and barotrauma in one patient (0.9%). After tracheotomy, VAP were diagnosed in 56 patients (52.8%) within 5.2 ± 5.9 days regarding tracheotomy (ranging from 1 to 25 days). Successful definite weaning from mechanical ventilation was obtained in 78 patients (73.6%): 38 patients of the VAP (+) group (67.9%) and 40 patients of the VAP (−) group (80%) (P = 0.157). Mean delay of successful weaning was significantly shorter for the VAP (−) group (5.9 ± 6.9 days Vs 10.8 ± 12.5 days; P = 0.037).Late complications were stomal infection in 30 patients (28.3%), catheter infection in 10 patients (9.4%) and VAP in 56 patients (52.8%).

Mean ICU length of stay (LOS) was 28.1 ± 16.3 days. LOS was significantly longer in patients who developed VAP (+) group compared to VAP (−) group with respective mean durations of 31.9 ± 18.5 days and 23.3 ± 12.5 days (P = 0.01).

Ventilator-associated pneumonia: Characteristics and predictive factors

For patients who developed VAP after tracheotomy, identified microorganisms were Acinetobacterbaumannii in 16 cases (15.1%), Pseudomonas aeruginosa in 15 patients (14.2%), enterobacteriaceaespp in 14 patients (13.2%) and SARM in one patient (0.9%). VAP was not documented for 10 patients (9.4%).

The epidemiological and clinical characteristics on admission were similar between the two groups [Table 2]. No significant difference was found between VAP (+) and VAP (−) group concerning core temperature on the day of tracheotomy with mean temperature at 38.55 ± 1.07°C and 38.31 ± 1.06°C respectively (P = 0.252). Except significantly lower bicarbonate plasmatic level in VAP (+) patients, biological results recorded on the day of tracheotomy were also comparable between the two groups [Table 3].

Table 2.

Epidemiological and clinical characteristics on admission

Table 3.

Biological findings the day of tracheotomy

Delay of tracheotomy regarding the initiation of mechanical ventilation was 8 ± 5.1 days in VAP (+) patients and 9.3 ±5.5 days in VAP (−) patients (P = 0.201). Early tracheotomy was performed in 30 patients among those who developed VAP (53.6%) and in 23 patients among those who didn’t experience this complication (46%) (P = 0.436). The whole duration of mechanical ventilation was significantly longer in patients who developed VAP after tracheotomy (21.8 ± 14 days Vs 16 ± 8.5 days; P = 0.01).

Bleeding events were significantly more frequent in VAP (+) group (19 patients of the VAP (+) group (33.9%) Vs 5 patients in the VAP (−) group (10%); P = 0.003). Similarly Stomal infections were more frequent in the VAP (+) group as they were reported in 23 patients in this group (41%) Vs 7 patients (14%) in the VAP (–) group (P = 0.002).

In multivariate analysis, factors identified as independently predictive of VAP onset were delayed tracheotomy (OR = 0.041; CI95% [1.02-7.87]; P = 0.041) and stomal infection (OR = 3.04; CI95% [1.02-9.93]; P = 0,045) [Table 4].

Table 4.

Multivariate analysis of factors predicting VAP onset after open tracheotomy

Prognosis impact

Thirty three patients died in ICU (31.1%). Mortality was not significantly different between VAP (+) and VAP (−) groups with respective rates at 35.7% and 26% (P = 0.281).

DISCUSSION

Severe trauma is one of the leading causes of death and morbidity in the world.[18] During the first hours following trauma, intubation and mechanical ventilation can be required for several reasons: Coma, acute respiratory failure, uncontrolled shock…[19] However, after the acute phase, most of these patients need intubation mainly for airway protection, which may considerably delay extubation, prolonger sedation and mechanical ventilation.[20,21] For all these reasons, tracheotomy represents an attractive procedure for airway management in this specific group of patients.[2,19] On the other hand, tracheotomy is not without risk and serious complications such as bleeding events, canula misplacement, barotraumas or stomal infections can complicate the immediate or the late course of tracheotomized patients.[2,4,22] VAP is one of the most threatening complications that can considerably worsen trauma patient's prognosis.[23] Its incidence reaches 40 to 60% in ventilated patients.[23,24,25] This high incidence can be explained by impaired consciousness, immunodepression and the need for emergency tracheal intubation and invasive therapeutics.[23,26,27] Available data dealing with the relationship between tracheotomy and ventilator-associated pneumonia are conflicting. Several studies didn’t find any difference between tracheotomized patients and those undergoing prolonged intubation in terms of VAP incidence.[20,28] However, Lesnik et al.[9] reported in a retrospective study including 101 patients with blunt and multiple organ trauma, that early tracheostomy significantly reduces the incidence of VAP when compared to prolonged endotracheal intubation. The same conclusions were highlighted by Kluger et al.[11] who retrospectively studied 118 cases of trauma patients requiring tracheostomy during their ICU stay: Early tracheostomy was associated with a significant reduction of VAP incidence. Similarly, Rodriuez et al.[29] found in their prospective randomized trial, a significantly decreased frequency of VAP in patients who underwent early tracheostomy (within 5 days of intubation). In our study, multivariate analysis shows also that delayed tracheotomy (> 7 days regarding the initiation of mechanical ventilation) was an independent factor predicting VAP onset (OR = 0.041; CI95% [1.02-7.87]; P = 0.041). Several hypotheses can be advanced to explain reduced VAP incidence with early tracheotomy: First, tracheotomy can reduce mechanical ventilation duration which represents the main risk factor for VAP onset.[9,10,20,30] Second, this procedure facilitates nursing care which may avoid secretions pooling above the endotracheal cuff and aspiration through the vocal cords, kept open by the tube.[6] Third, tracheotomy is associated with less endotracheal damage such as ulcerations and edema which may causes repetitive aspiration episodes[31] and finally, tracheotomized patients have the opportunity to have an oral nutrition sooner than intubated patients which reduce the need of orogastric tubes used for enteral feeding.

To the best of our knowledge, there is no previous study which aimed to identify independent factors predicting VAP onset in trauma patients undergoing tracheotomy during their ICU stay. In the multivariate analysis, our study shows that independent factors predicting VAP onset after tracheotomy were delayed tracheotomy and stomal infection. We hypothesize that stomal infection results mainly from secretions pooling above the endotracheal cuff. The risk of stomal infection could be also enhanced by bleeding events following tracheotomy. In fact, our univariate analysis revealed that these events were significantly higher in VAP (+) group. In a systematic review including 17 randomized controlled trials comparing percutaneous dialatationaltracheostmoy (PDT) and surgical tracheosotmy (ST), Delaney et al. reported that that PDT was associated to a significant reduction of wound infection and bleeding events regarding patients who underwent ST in the operating theatre.[32] Thus, causality relation between both complications is highly suspected. The higher incidence of VAP in patients with stomal infection may be explained by oropharyngeal colonization with virulent microorganisms which can secondly invade airways and disseminate through the lungs.

Previous studies reported that early weaning from mechanical ventilation was a major advantage that can be reached with tracheotomy.[9,10,20,28] In our study, patients who didn’t develop ventilator associated pneumonia were weaned faster than those who developed nosocomial pneumonia (5.9 ± 6.9 days Vs 10.8 ± 12.5 days; P = 0.037). This is not surprising as similar conclusions were found in trauma patients who were ventilated with endotracheal intubation.[33] As a consequence, the onset of VAP was also associated with longer ICU LOS (31.9 ± 18.5 days Vs 23.3 ± 12.5 days; P = 0.01). These findings are particularly important. In fact, previous studies highlighted that early tracheotomy in trauma patients improves resource utilization with reduced mechanical ventilation duration and ICU LOS.[10] Our study suggests that theses benefices seem to be lost with the occurrence of ventilator associated pneumonia.

CONCLUSION

Our study shows that late tracheotomy and stomal infections are independent factors predicting VAP onset in tracheotomized patients. VAP occurrence is associated with significantly longer mechanical ventilation and ICU LOS. These findings should incite physicians to avoid such complication by identifying patients who would need prolonged ventilation and thus, may benefit of early tracheotomy. Stomal infection should be prevented by multiple local aseptic cares. Secretions pooling above the canula cuff should be also avoided by repetitive suctioning.