Continuous Aspiration of Subglottic Secretions in Preventing Ventilator-Associated Pneumonia

Methods

Patients

All patients admitted to the intensive care unit of the Sabadell Hospital from June 1990 to March 1993 who required intubation and were expected to receive mechanical ventilation for at least 72 hours were eligible for study. The intubation could be done either in the intensive care unit or in the emergency department. Patients were excluded if they were intubated in other areas of the hospital, if they carried a tracheostomy tube, or if they developed a pneumonia or died during the first 72 hours of mechanical ventilation. Study was considered complete when a patient was extubated, when a tracheostomy was done, when a patient died, or when ventilator-associated pneumonia was diagnosed. The follow-up period consisted of the patient's remaining stay in the intensive care unit.

All patients were intubated with the same type of endotracheal tube (Hi-Lo Evac; Mallinckrodt Laboratories, Athlone, Ireland), which incorporates a dorsal separate lumen ending into the subglottic area by creating a large elliptical dorsal opening above the cuff for aspiration of subglottic secretions (Figure 1). The size of each endotracheal tube was selected by the attending physician; after intubation, the correct position of endotracheal tube was verified by a roentgenogram of the chest.

Patients were randomly assigned to receive continuous aspiration through the additional lumen from the endotracheal tube (case-patients) or to receive standard treatment in which the additional lumen remained closed (control patients). Intra-cuff pressure was monitored every 4 hours (Mallinckrodt GmbH, Mallinckrodt Laboratories, Neunkirchen-Seelscheid, Germany) and was kept above 20 mm Hg. Subglottic drainage was continuous, and secretions were collected in a mucous collector (Mocstrap; Proclinics, Barcelona, Spain). The amount of secretions obtained daily was also recorded. If subglottic drainage was negative, the permeability was checked every 4 hours by injecting sterile saline serum into the evacuation lumen. All patients received stress ulcer prophylaxis with sucralfate. We did not use a selective decontamination regimen or antibiotic prophylaxis. The study protocol was reviewed and approved by the hospital's institutional review board for human studies.

Data Collection and Definitions

We prospectively recorded each patient's demographic characteristics, diagnosis at admission, and underlying diseases. The Acute Physiology and Chronic Health Evaluation (APACHE) II scoring system of Knaus and colleagues [22] was used to assess the severity of an acute illness. Several risk factors for ventilator-associated pneumonia were recorded, such as continuous sedation, previous surgery, multiple trauma, structural or pharmacologic coma (score of ≤ 8 according to the Glasgow coma scale [23]), and use of muscle relaxants and antibiotic treatment during the intensive care unit stay.

Ventilator-associated pneumonia was suspected in patients who met the following criteria after 72 hours of mechanical ventilation: fever (body temperature ≥ 38.3 °C), leukocytosis (>12 000 leukocytes/mm³) or leukopenia (<4000 leukocytes/mm³), purulent secretions, and the presence of new and persistent pulmonary infiltrates. The diagnosis of pneumonia was confirmed by a positive protected specimen brush culture containing 10³ colony-forming units (CFU)/mL or more, a positive bronchoalveolar lavage culture with 10⁴ CFU/mL or more, or by a good clinical response to antibiotic agents. Additional criteria were the absence of a diagnosis other than pneumonia and pathologic findings consistent with pneumonia in patients who died and in whom autopsy was authorized. Ventilator-associated pneumonia was histologically diagnosed when foci of consolidation with intense polymorphonuclear leukocyte accumulation in the bronchioles and alveolar spaces were observed. Roentgenograms of the chest were interpreted by a radiologist who had no knowledge of patients' treatment groups.

Crude mortality rates included all deaths that occurred in the intensive care unit in patients with ventilator-associated pneumonia. Death was considered attributable to the pulmonary infection if the patient died before having any objective response to antimicrobial therapy or if the pulmonary infection was considered a contributing factor to death in patients with additional conditions.

We used the definitions described by Knaus and colleagues [22] to define the underlying diseases. We considered previous surgery when a surgical procedure had been done within the present admission to the hospital and considered previous antibiotic treatment when a patient was receiving antibiotic agents at randomization.

Bacteriologic Examination

In patients in whom ventilator-associated pneumonia was suspected, bronchoscopy was done while they were being ventilated with an FIO₂ of 1.0 and without positive end-expiratory pressure. All patients received midazolam as a sedative and atracurium as a relaxant while the bronchoscopy was done. The fiberoptic bronchoscope (Olympus BF 20; Olympus Corp., Tokyo, Japan) was passed into the trachea through the endotracheal tube by a special connector (Unimed Ltd., Shaftesbury, United Kingdom) and was advanced under visual control to the bronchial orifice of the abnormal lobe. The telescoping plugged catheter (TAG Medical, Bobigny Cedex, France) was inserted through the inner suction channel and was advanced to a wedged peripheral position. Airway secretions were obtained using a previously described technique [24]. For a bronchoalveolar lavage, the bronchoscope was sustained in a wedged position, and lavage was done with three 50-mL aliquots of sterile isotonic saline. The bronchoscope was then removed, and ventilation with an FIO₂ of 1.0 was continued for 15 minutes.

After the protected specimen brush was transected into a sterile vial containing 1 mL of sterile lactate Ringer's solution, the vial was vigorously agitated for at least 60 seconds to suspend all the material from the brush. Specimens were immediately sent to the laboratory for quantitative cultures. Aliquots of 0.01 mL were then taken from the original suspension and inoculated into blood agar, chocolate agar, anaerobic kanamicin blood agar, anaerobic blood agar, MacConkey agar, buffered charcoal yeast extract agar, and Sabouraud media. One 0.001-mL aliquot was also inoculated into chocolate agar media. Culture plates were incubated at 37 °C under adequate aerobic and anaerobic conditions; all plates except for the Sabouraud plates were evaluated for growth at 24 and 48 hours. For the protected specimen brush, bacterial counts of 10³ CFU/mL or greater were used as the cutoff point to diagnose pneumonia. Two serial 10-fold dilutions were then done on the recovered bronchoalveolar lavage fluid, and 0.01-mL aliquots of the original suspension and each dilution were placed onto plates in the same way as for the protected specimen brush sample. All protected specimen brush and bronchoalveolar lavage fluid isolates were identified by standard laboratory techniques [25]. Two blood cultures were done simultaneously in all patients, as were pleural fluid cultures if present. Surveillance cultures for aerobic microorganisms in the subglottic secretions were obtained from patients in the continuous aspiration group and were repeated every 5 days until ventilator-associated pneumonia developed or until patients were extubated or died.

Statistical Analysis

We compared the characteristics of the two groups using the Student t-test or Mann-Whitney test for continuous variables and the chi-square test or Fisher exact test for categorical variables. We defined the cumulative incidence as the number of events divided by the number of patients and estimated risk ratios and 95% CIs [26]. The incidence rate was defined as the number of events divided by the number of days the patient was at risk because of the presence of an endotracheal tube. We compared incidence rates of pneumonia in the two groups in terms of relative risk with 95% CIs [26]. We used the Kaplan-Meier survival analysis to calculate the probability of the development of nosocomial pneumonia during mechanical ventilation [27]. The log-rank test was used for comparisons. Using the same methods, we compared the median time patients with or without pneumonia received mechanical ventilation, the duration of stay in the intensive care unit, and the duration of mechanical ventilation in the two groups. The mortality rate in each group was treated as a censoring event. We calculated the sample size to ensure that the study had a power of 80% to detect a 20% reduction in the cumulative incidence of pneumonia (a decrease from an expected rate of 30%, observed from previous surveillance studies in our intensive care unit, to 10%), with an α-error of 0.05. All P values were two-tailed, and P values of 0.05 or less were considered to indicate statistical significance.

Results

Patients

One hundred ninety patients were enrolled in the study. Nineteen patients receiving continuous aspiration of subglottic secretions were excluded before 72 hours of mechanical ventilation because they were extubated (8 patients), died (9 patients), or developed pneumonia (2 patients). Eighteen patients from the control group were also excluded: Eleven were extubated and 7 died before 72 hours of mechanical ventilation. Of the 153 remaining patients, 76 received continuous aspiration of subglottic secretions and 77 carried the same type of endotracheal tube but did not receive aspiration of subglottic secretions. The two groups were similar in demographic characteristics, severity of illness on admission, and underlying diseases (Table 1). The distribution of indications for intubation and risk factors for ventilator-associated pneumonia were also similar in both groups (Table 1). Some of the patients had more than one risk factor for ventilator-associated pneumonia during their stay in the intensive care unit. In both groups, the mean intracuff pressure, episodes in which intracuff pressure was less than 20 cm H₂O, and the number of times the endotracheal tube was changed were also similar (Table 1).

Incidence of Ventilator-Associated Pneumonia

Ventilator-associated pneumonia developed in 39 of the 153 patients (cumulative incidence, 25.5%): 14 (18.4%) receiving continuous aspiration of subglottic secretions and 25 (32.5%) in the control group (relative risk, 1.76; 95% CI, 0.99 to 3.12) (Table 2). The incidence rate was 19.9 episodes of ventilator-associated pneumonia/1000 ventilator days in the patients receiving continuous aspiration of subglottic secretions and 39.6 episodes/1000 ventilator days in the control patients (relative risk, 1.98; CI, 1.03 to 3.82) (Table 2).

The 39 patients with pneumonia were ventilated for 8.1 ± 5.4 days (mean ± SD) before ventilator-associated pneumonia was diagnosed. Episodes of ventilator-associated pneumonia developed later in patients receiving continuous aspiration (12.0 ± 7.1 days) than in the control patients (5.9 ± 2.1 days) (P < 0.001). Moreover, when we analyzed only the first week of mechanical ventilation, we observed that 21 cases of pneumonia developed in the control group and only 3 developed in the group receiving continuous aspiration (P < 0.001). During the second week of ventilation, we diagnosed 7 episodes of ventilator-associated pneumonia in the subglottic aspiration group and 4 episodes in the control group (P > 0.05); the remaining 4 cases of ventilator-associated pneumonia were diagnosed in the patients receiving continuous aspiration of subglottic secretions after the second week of ventilation. Survival analysis and log-rank test results indicated that the control group had a significantly higher probability (P = 0.02) of developing nosocomial pneumonia during mechanical ventilation (Figure 2).

We also analyzed the effect of concomitant antibiotic treatment. Ventilator-associated pneumonia developed in 18 patients not receiving antibiotic treatment at randomization: 13 patients in the control group (33.5 episodes/1000 ventilator days) and 5 patients receiving continuous aspiration (13.2 episodes/1000 ventilator days). Ventilator-associated pneumonia developed in 21 of the patients treated with antibiotic agents: 12 patients in the control group (22.7 episodes/1000 ventilator days) and in 9 in the group receiving continuous aspiration (15.3 episodes/1000 ventilator days). More detailed information on the effect of antibiotic treatment is shown in Table 3.

Cause of Ventilator-Associated Pneumonia

The microorganisms isolated in significant concentration from the protected specimen brush or bronchoalveolar lavage fluid in patients with ventilator-associated pneumonia are shown in Table 4. In two cases of Pseudomonas aeruginosa pneumonia (one episode in each study group), the blood cultures were positive. In five cases the cause was not established, and the diagnosis of ventilator-associated pneumonia was confirmed by good clinical response to antibiotic treatment (three episodes) or by histologic findings (two episodes). Fourteen bacterial organisms were isolated in 15 episodes in the subglottic aspiration group compared with 24 pathogens isolated in 25 episodes among the 77 patients in the control group. This difference was caused by a significant (P < 0.03) reduction in the number of gram-positive cocci and Haemophilus influenzae organisms in the patients receiving continuous aspiration of subglottic secretions. However, we observed no difference in the number of P. aeruginosa or Enterobacteriaceae organisms. Pseudomonas aeruginosa represented 81% of isolates in patients with ventilator-associated pneumonia receiving antibiotic agents at randomization and only 38.8% of isolates in the group not receiving them (Table 3).

The mean volume of subglottic secretions aspirated daily was higher in patients without pneumonia (18.4 ± 14.8 mL compared with 12.3 ± 11.2 mL, respectively), although this difference was not significant (P = 0.09). Cultures of bacteria from subglottic aspirates were done in all patients receiving continuous aspiration of subglottic secretions. Most cultures were polymicrobial. Many gram-positive cocci, Enterobacteriaceae, and commensal buccal flora were obtained during the first 5 days of intubation. We observed lower concentrations of these microorganisms in sequential cultures, whereas we found that the number of P. aeruginosa organisms in the subglottic secretion samples progressively increased. The number of fungi remained constant. More information on the microorganisms isolated in subglottic secretions is shown in Table 5.

We observed a close association between microorganisms isolated from subglottic secretions and microorganisms that cause ventilator-associated pneumonia. Indeed, 12 of the 14 bacteria identified in significant concentration in episodes of pneumonia were previously identified in the subglottic secretion culture, although all were isolated in polymicrobial culture and the specificity was low. Microorganisms were detected in subglottic secretion cultures a mean of 4.8 ± 2.9 days before ventilator-associated pneumonia was diagnosed. This correlation and the time between the identification of the microorganism in the subglottic secretion culture and the development of pneumonia are shown in Table 6.

Outcome

The Kaplan-Meier survival analysis indicates that the median duration of ventilation was higher in patients with ventilator-associated pneumonia than in those without ventilator-associated pneumonia (22 ± 5 days compared with 9 ± 1 days, respectively; P < 0.001). Patients with ventilator-associated pneumonia also stayed longer in the intensive care unit (32 ± 8 days compared with 16 ± 2 days; P < 0.001). However, the two groups did not significantly differ in the duration of ventilation (13 ± 1 days compared with 11 ± 1 days; P > 0.2) or in intensive care unit stay (22 ± 2 days compared with 19 ± 4 days; P > 0.2).

Fifty-eight patients died, representing a crude mortality rate of 37.9%. In the patients receiving continuous aspiration of subglottic secretions, the crude mortality rate was 39.5%; in the control patients, the rate was 36.4% (P > 0.05). The mean APACHE II scores were 20.5 ± 7 and 18.9 ± 7.1, respectively (P > 0.05). In patients with ventilator-associated pneumonia, the mortality rate was 51.3% (without differences between the two groups) compared with 33.3% in patients without pneumonia. Of the 14 patients with pneumonia who received continuous aspiration, 7 died, whereas 13 of the 25 patients with ventilator-associated pneumonia in the control group died (P > 0.05). We considered death to be attributable to the pulmonary infection in 14 of 153 patients (9.1%); most attributable deaths (71.4%) occurred in the control group (relative risk, 2.46; CI, 0.77 to 7.84).

Discussion

We found that the use of continuous aspiration of subglottic secretions in intubated patients reduced the incidence of ventilator-associated pneumonia by 43.4%. This decrease was caused by a significant reduction in the incidence of pneumonia during the first days of mechanical ventilation. Our results agree with the findings reported by Mahul and colleagues [21], who used manual intermittent aspiration of subglottic secretions and found a decrease in the incidence of ventilator-associated pneumonia and a delay in the emergence of pneumonia during mechanical ventilation. These investigators used a similar endotracheal tube, but they aspirated the subglottic secretions hourly and included different treatments for prophylaxis of gastric bleeding. This technique represents a simple, inexpensive, and useful approach in the prevention of nosocomial pneumonia; indeed, it primarily affects cases of primary endogenous pneumonia, which are mainly caused by indigenous flora already present in the oral cavity of patients at the time of intubation. After the first week of mechanical ventilation, we did not find significant differences between the incidence of pneumonia in the two groups. However, episodes of ventilator-associated pneumonia caused by P. aeruginosa after the first week were delayed by 3.9 ± 4.9 days in the patients receiving continuous aspiration of subglottic secretions.

The virulence of the bacterial species, the size of inoculum, and the capacity of the pulmonary defense mechanisms are decisive in the development of nosocomial pneumonia. With suction of subglottic secretions, the volume of oropharyngeal secretions aspirated into the bronchial tract should decrease and the size of inoculum should be lower. The minimum inoculum needed for development of experimental gram-positive cocci pneumonia is higher than the inoculum needed for Enterobacteriaceae or P. aeruginosa pneumonias [28, 29]. Consequently, the reduction in the size of inoculum could explain the delay in the development of P. aeruginosa pneumonia and the decrease of gram-positive or H. influenzae pneumonia found in the patients receiving subglottic aspiration. Interestingly, the volume of subglottic secretions aspirated in patients with pneumonia was lower than that in patients without pneumonia; this finding may indicate an increase in the inoculum of oropharyngeal secretions aspirated into the bronchial tract in patients with pneumonia due to the reduction in the effectiveness of subglottic aspiration. In addition, this finding emphasizes the importance of maintaining adequate cuff pressure to prevent microaspirations of subglottic secretions to the lung.

The two groups were similar in their demographic characteristics, underlying disease, and severity of acute illness. More patients in the control group received surgical procedures, but this difference was not significant. In studies of patients with and without mechanical ventilation [30, 31], previous abdominal or thoracic surgery has been shown to be a risk factor for the development of nosocomial pneumonia; thus, the fact that more patients in the control group had had surgery may cause concern. However, of the 28 patients who had a surgical procedure, we observed only one episode (11%) of ventilator-associated pneumonia in 9 patients receiving continuous aspiration of subglottic secretion (diagnosed after 25 days of mechanical ventilation) compared with four (21%) episodes in the 19 control patients. These findings confirm that the subglottic aspiration technique also decreases the incidence of ventilator-associated pneumonia in patients who have had surgery and delays the emergence of pneumonia.

For most types of nosocomial pneumonia, the colonization of the upper respiratory tract and subsequent aspiration of oropharyngeal secretions represents the most common route of inoculation [8, 9, 14]. In mechanically ventilated patients, even those with an endotracheal tube that has an adequately inflated cuff, aspiration around the tube from the colonized oropharynx into the tracheobronchial tree may occur [32]. Indeed, we found that 85.7% of microorganisms causing infection were previously isolated in cultures of subglottic secretions. However, in two episodes in which P. aeruginosa was involved, the organism could not have been previously isolated in subglottic cultures, which suggests that Pseudomonas species can colonize the lower respiratory tract independent of upper airway colonization, as previously reported [33-35]. In these cases, organisms may be directly inoculated into the lung through the endotracheal tube, from hands of medical or nursing staff, or from the respiratory equipment; this method of prevention has no effect on this situation.

The underlying disease, length of intubation, and the antibiotic treatment can influence the cause and the incidence of ventilator-associated pneumonia. Consequently, other intensive care units serving different patient populations may obtain different results. For example, the effect of continuous aspiration of subglottic secretions may be more marked in patients with a high proportion of gram-positive cocci or H. influenzae pneumonias, as is the case in patients with cranioencephalic trauma [36]. On the other hand, the effect may be less significant in patients who have had colonization by P. aeruginosa, such as those with chronic obstructive pulmonary disease [37]. In our study, the presence of antibiotic treatment did not modify the influence of continuous aspiration of subglottic secretions on the incidence of ventilator-associated pneumonia: We observed a reduction in the incidence of pneumonia in subgroups of the patients receiving continuous aspiration of subglottic secretions, independent of whether the patients were receiving antibiotic treatment at randomization (Table 3). However, we observed a different cause of ventilator-associated pneumonia in patients who received antibiotic agents. Indeed, the microorganism most frequently isolated as responsible for pneumonia in patients receiving antibiotic agents was P. aeruginosa, as has been previously described [38, 39]. Interestingly, despite the reduction in the number of cases of ventilator-associated pneumonia in the continuous aspiration group and the longer intubation period associated with such types of pneumonia, the duration of ventilation was similar in the two groups. We believe that this was due to the different virulences of microorganisms. Indeed, the greater incidence of ventilator-associated pneumonia in the control group was caused by H. influenzae and gram-positive cocci, and it has been recently reported that the effect of these microorganisms on morbidity and mortality is dramatically lower than that of P. aeruginosa[39].

The crude mortality rate in our study was 38%, and we did not find differences between the two groups of patients studied. This finding agrees with those of most studies of selective digestive decontamination, which did not find a significant reduction in mortality rate and in length of stay in intensive care units [18, 40-42]. However, only 14 patients (9.1%) died as a direct result of ventilator-associated pneumonia; P. aeruginosa was present in 8 (57.1%) of the patients who died. Previous studies have reported a high mortality rate in pulmonary infections with Pseudomonas and Acinetobacter species, and this was attributed to the higher virulence of these microorganisms [38]. Interestingly, most deaths directly related to infection occurred in the control group. In contrast to studies of selective decontamination of the digestive tract, the reduction in the incidence of ventilator-associated pneumonia represents a reduction in the total amount of antibiotic agents used; this is of great importance not only because it affects costs but also because it may be a major factor in the development of antibiotic resistance and in the selection of flora with greater virulence.

In summary, we show that the incidence of nosocomial pneumonia in intubated patients can be significantly reduced by using a simple method that decreases the long-term microaspirations through the cuff of endotracheal tubes. This system of prevention is inexpensive and does not add cost to treating patients with a conventional artificial airway. Furthermore, this measure helps reduce the antibiotic dosage and may be combined with other methods of prevention. Although further studies should examine the usefulness of this technique in specific patient populations (such as patients with trauma), we recommend that it be incorporated into the management of intubated patients.