Abstract
Rationale: Ventilator-associated pneumonia (VAP) causes substantial morbidity and mortality. The influence of subglottic secretion drainage (SSD) in preventing VAP remains controversial.
Objectives: To determine whether SSD reduces the overall incidence of microbiologically confirmed VAP.
Methods: Randomized controlled clinical trial conducted at four French centers. A total of 333 adult patients intubated with a tracheal tube allowing drainage of subglottic secretions and expected to require mechanical ventilation for ≥48 hours was included. Patients were randomly assigned to undergo intermittent SSD (n = 169) or not (n = 164).
Measurements and Main Results: Primary outcome was the overall incidence of VAP based on quantitative culture of distal pulmonary samplings performed after each clinical suspicion. Other outcomes included incidence of early- and late-onset VAP, duration of mechanical ventilation, and hospital mortality. Microbiologically confirmed VAP occurred in 67 patients, 25 of 169 (14.8%) in the SSD group and 42 of 164 (25.6%) in the control group (P = 0.02), yielding a relative risk reduction of 42.2% (95% confidential interval, 10.4–63.1%). Using the Day 5 threshold, the beneficial effect of SSD in reducing VAP was observed in both early-onset VAP (2 of 169 [1.2%] patients undergoing SSD vs. 10 of 164 [6.1%] control patients; P = 0.02) and late-onset VAP (23 of 126 [18.6%] patients undergoing SSD vs. 32 of 97 [33.0%] control patients; P = 0.01). VAP was clinically suspected at least once in 51 of 169 (30.2%) patients undergoing SSD and 60 of 164 (36.6%) control patients (P = 0.25). No significant between-group differences were observed in duration of mechanical ventilation and hospital mortality.
Conclusions: Subglottic secretion drainage during mechanical ventilation results in a significant reduction in VAP, including late-onset VAP.
Clinical trial registered with www.clinicaltrials.gov (NCT00219661).
The impact of subglottic secretion drainage in preventing ventilator-associated pneumonia (VAP), especially late-onset VAP, remains controversial.
The drainage of subglottic secretions is effective in preventing both early-onset and late-onset VAP.
VAP results from microbial invasion of the normally sterile lower respiratory tract and lung parenchyma, which can then overwhelm the host's defense to establish infection. Aspiration of oropharyngeal pathogens and leakage of subglottic secretions containing bacteria around the endotracheal tube cuff are the primary routes of bacterial entry into the lower respiratory tract (3–6). Subglottic secretion drainage (SSD), using a specially designed endotracheal tube with a separate dorsal lumen that opens immediately above the endotracheal cuff, has been developed to prevent the occurrence of VAP. In two single-center randomized controlled trials (7, 8) and one observational study (9), SSD reduced the incidence of VAP. However, these positive results were not confirmed in three other single-center trials (10–12). A meta-analysis of these studies concluded that SSD was effective in reducing the rate of early-onset VAP (13). SSD was only suggested as a measure to prevent VAP in the U.S. and Canadian guidelines in 2003 and 2004 (14, 15). In the more recent American Thoracic Society and Canadian guidelines, SSD was recommended—instead of suggested—by the experts (3, 16). However, the authors of the latter guidelines acknowledged that this stronger level of recommendation was attributable to the growing availability of SSD tracheal tubes rather than to stronger evidence of efficacy (16). Whether SSD is an effective measure to prevent VAP, especially late-onset VAP, remains controversial and may account for the low rate at which mechanically ventilated ICU patients benefit from this preventive measure in routine practice (17, 18). Furthermore, concerns about the safety of continuous SSD were raised in an experimental study in sheep, which found widespread injuries to tracheal mucosa and/or submucosa (19).
To shed light on these persisting controversies, we designed a multicenter randomized controlled trial with the main objective of assessing the impact of SSD on the incidence of VAP in a large unselected population of mechanically ventilated ICU patients. Diagnosis of VAP was based on distal respiratory secretion sampling, using strict microbiological criteria. The secondary objectives were to assess the impact of SSD on the incidence of early- and late-onset VAP, duration of mechanical ventilation, length of ICU stay, and incidence of postextubation laryngeal dyspnea. Some of the results of this study have been previously reported in the form of an abstract (20).
The study was conducted in four French medicosurgical intensive care units (ICUs). During the study period, all consecutive patients admitted to the ICU were screened to identify those older than 18 years who were intubated with a specific endotracheal tube allowing the drainage of subglottic secretions (Hi-Lo Evac tube; Mallinckrodt Medical, Athlone, Ireland) and were expected to require mechanical ventilation for at least 48 hours.
Patients admitted in ICUs with prior tracheostomy or already intubated with a tube other than the Hi-Lo Evac tube were excluded. Patients admitted with psychotropic drug overdose or acute drunkenness, whose mechanical ventilation is frequently shorter than 48 hours (21), were excluded. Patients admitted after cardiac arrest, whose overall mortality rate 48 hours after ICU admission was high (22, 23) and from whom care was frequently withheld or withdrawn, were also excluded, as in another preceding study on VAP prevention (24). Patients who had already undergone mechanical ventilation for more than 24 hours at the time of screening for eligibility or patients already enrolled in another trial were also excluded.
During the study period, for financial reasons, it was possible to replace all the conventional tracheal tubes with Hi-Lo Evac tubes (more costly) only in a few units where the patients could be intubated before being transferred to the participating ICUs. In the other units, where both conventional and Hi-lo Evac tubes were available, the Hi-Lo Evac tube was used only for intubations of patients with an anticipated positive screening for potential study inclusion.
The study was approved by the Poissy/St-Germain en Laye Ethics Committee (St-Germain en Laye, France). Written consent was obtained from the patients or their proxies.
Patients were randomly assigned to undergo subglottic secretion drainage (SSD group) or not (control group). A randomization list in blocks of six, stratified by each participating ICU, was generated from a computer, and treatment allocation was concealed, using sequentially numbered opaque sealed envelopes. All caregivers and other research personnel were blinded to the randomization schedule and the block size.
In the SSD group, subglottic secretions were suctioned manually with a 10-ml syringe at an intended frequency of one suction per hour. The number of SSD attempts and the amount of subglottic secretions collected in each patient in the SSD group were prospectively recorded. SSD was not performed in the control group. Because of the nature of the intervention, physicians and nurses could not be blinded to the randomization arm.
All randomized patients were screened daily for the occurrence of clinical suspicion of VAP. Screening for VAP was maintained until the first episode of microbiologically confirmed VAP, or 48 hours after weaning from the ventilator, or death, or Day 28 of mechanical ventilation, whichever occurred first. Patients requiring invasive mechanical ventilation within 48 hours of an attempt at extubation were maintained in the study as described previously.
Clinical suspicion of VAP was based on the presence of a recent and persistent infiltrate on chest X-ray and at least two of the following criteria: fever (>38.3°C) or hypothermia (<36°C), leukocytosis (>10.109/L) or leukopenia (< 4.109/L), and purulent tracheal secretions. Confirmation of pneumonia required a positive quantitative culture of either a protected telescoping catheter sample (25) or bronchoalveolar lavage fluid (26). Samples were obtained before any change in antibiotic therapy or the introduction of new antibiotics. VAP was confirmed if quantitative culture of the protected telescoping catheter sample or bronchoalveolar lavage fluid grew at least 103 or at least 104 colony-forming units (CFU)/ml, respectively, of at least one microorganism. Microbiologists at each participating center were blinded to the randomization arm.
Episodes of pneumonia diagnosed within 48 hours of ventilation onset were not considered ventilator-associated (27). A threshold of 5 days after initiation of mechanical ventilation was used to distinguish early-onset (<5 d) VAP and late-onset (≥5 d) VAP (3). The rate of early-onset VAP was assessed in all randomized patients, whereas the rate of late-onset VAP was assessed in patients with at least 5 days of mechanical ventilation without early-onset VAP.
Compliance with preventive measures routinely used in each participating center, including oral route of insertion of the tracheal and gastric tubes, enteral delivery of nutritional support, cuff pressure maintained between 20 and 30 cm H2O, and semirecumbent body position, was prospectively recorded. In two of the participating ICUs, trunk position was available directly from an angle measurement scale attached to the ICU beds. In the two remaining ICUs, trunk position was determined according to the judgment of the attending nurses. Trunk position and cuff pressure were recorded every 3 hours.
The following variables potentially influencing the occurrence of VAP were also prospectively recorded: antibiotic therapy, stress ulcer prophylaxis, need for reintubation, use of sedative drugs and paralytic agents, transfusion of erythrocytes, transfer out of the ICU, and type of device used to heat and humidify gases. No selective digestive decontamination was used.
The primary end point was the incidence of VAP in the SSD and control groups. Secondary outcome variables included the rate of early- and late-onset VAP, duration of mechanical ventilation before the occurrence of VAP, total duration of mechanical ventilation, ICU and hospital mortality, ICU stay, and tracheostomy rate. The safety of SSD was assessed by recording episodes of postextubation laryngeal dyspnea (inspiratory dyspnea associated with an inspiratory noise such as a stridor) and the rate of reintubation.
From the incidence rates of VAP observed in patients requiring mechanical ventilation for at least 48 hours in the participating centers before study start and results of previous studies (7, 8, 12), we estimated an incidence of VAP of 20% in the control group and 10% in the SSD group. Randomly assigning 220 patients to each group would allow detection of this difference with 80% power and a two-tailed significance level of 0.05. A 3-year study period was planned to include 440 patients. At the end of the study period, only 333 patients were included.
Continuous variables were not normally distributed and, therefore, are presented as median (interquartile range [IQR]) values. The SSD and control groups were compared by Mann-Whitney U test for continuous variables and by chi-square or Fisher's exact test for categorical variables. Relative risks and absolute risk reductions were estimated with their 95% confidence interval (CI).
The cumulative rates of remaining free of VAP in the two groups were examined by the Kaplan-Meier method and compared by log-rank test. All P values were two-sided; significance was set at P < 0.05. No interim analysis was performed. All patients were analyzed in the group to which they were randomly assigned, according to the intention-to-treat principle.
All statistical tests were performed with Stata software (release 8.0, 2003; StataCorp, College Station, TX).
Between June 2003 and September 2006, 2,159 patients required mechanical ventilation in the participating centers (Figure 1). At ICU admission, 945 patients were not intubated with the Hi-Lo Evac device. Among the 1,214 remaining patients, 881 met exclusion criteria (Figure 1). Finally, 333 patients were included, 164 in the control arm and 169 in the SSD arm.
Figure 1. Study flow diagram. *Seventeen of these 117 patients finally required mechanical ventilation more than 48 hours.
[More] [Minimize]Among the included patients, 17 (5.1%) required mechanical ventilation for less than 48 hours (9 in the control group and 8 in the SSD group). These patients were maintained in the analysis. Within the first 48 hours of mechanical ventilation, 24 patients developed pneumonia (9 in the SSD group and 15 in the control group; P = 0.18).
There were no significant differences in baseline characteristics between patients undergoing SSD and control patients (Table 1), except for a history of cirrhosis, which was more frequent in the SSD group (14.8%) than in the control group (7.3%; P = 0.04). Subglottic secretion drainage was performed at a median of 18 times per day (15–19) in the SSD group, corresponding to a frequency of one suction every 90 minutes. The median daily volume of subglottic secretions suctioned was equal to 14 ml (8–28 ml), with a minimal value of 0 and a maximal value of 197 ml.
SSD (n = 169) | Control (n = 164) | |||
|---|---|---|---|---|
| Age, median [IQR] | 67 [53–77] | 70 [56–78] | ||
| Male sex, no. (%) | 101 (59.8) | 96 (58.5) | ||
| Previous comorbidity, no. (%) | ||||
| At least one comorbidity | 70 (41.4) | 59 (36.0) | ||
| Chronic heart insufficiency | 6 (3.6) | 8 (4.9) | ||
| Chronic respiratory insufficiency | 13 (7.9) | 12 (7.3) | ||
| Chronic renal insufficiency | 4 (2.4) | 4 (2.4) | ||
| Cirrhosis | 25 (14.8) | 12 (7.3) | ||
| Solid malignancy | 7 (4.1) | 7 (4.3) | ||
| Hemopathy | 5 (3.0) | 3 (1.8) | ||
| History of diabetes treated with insulin | 5 (3.0) | 7 (4.3) | ||
| Immunosuppressive treatment | 13 (7.7) | 11 (6.7) | ||
| HIV infection | 3 (1.8) | 3 (1.8) | ||
| Diagnostic category, no. (%) | ||||
| Medical | 142 (84.0) | 142 (86.6) | ||
| Scheduled/emergency post-op admission | 27 (16.0) | 22 (13.4) | ||
| SAPS II, median [IQR] | 51 [40–64] | 52 [39–63] | ||
| SOFA score, median [IQR] | 8 [6–11] | 8 [6–11] | ||
| Reason for ICU admission, no. (%) | ||||
| Pneumonia | 62 (36.7) | 62 (37.8) | ||
| Septic shock | 23 (13.6) | 18 (11.0) | ||
| Urinary tract infection | 4 (2.4) | 4 (2.4) | ||
| Abdominal infection | 11 (6.5) | 8 (4.9) | ||
| Meningo-encephalitis | 8 (4.7) | 2 (1.2) | ||
| Endocarditis | 1 (0.6) | 1 (0.6) | ||
| Pancreatitis | 5 (3.0) | 8 (4.9) | ||
| Necrotizing fasciitis | 1 (0.6) | 2 (1.2) | ||
| Acute asthma/COPD exacerbation | 9 (5.3) | 10 (6.1) | ||
| Gastrointestinal bleeding | 8 (4.7) | 8 (4.9) | ||
| Drug overdose | 4 (2.4) | 4 (2.4) | ||
| Acute stroke | 3 (1.8) | 8 (4.9) | ||
| Status epilepticus | 3 (1.8) | 5 (3.1) | ||
| Myocardial infarction/heart failure | 11 (6.5) | 9 (5.5) | ||
| Trauma | 4 (2.4) | 4 (2.4) | ||
| Other pulmonary disease | 4 (2.4) | 2 (1.2) | ||
| Other neurologic disease | 3 (1.8) | 2 (1.2) | ||
| Miscellaneous | 5 (3.0) | 7 (4.3) | ||
| Noninvasive ventilation before intubation, no. (%) | 30 (17.8) | 32 (19.5) | ||
At least one episode of VAP was clinically suspected at least once in 111 patients (51 [30.2%] patients in the SSD group and 60 [36.6%] in the control group), representing a total of 150 suspicions of VAP and subsequent collection of microbiological samples (SDD, n = 66; control, n = 84). Forty microbiological samples in the SSD group (60.1%) and 59 in the control group (70%) were bronchoalveolar lavage samples. The remaining 51 microbiological samples were protected telescoping catheter samples (SDD, n = 26; control, n = 25).
VAP was microbiologically confirmed in 67 patients. VAP was significantly less frequent in the SSD group (25 [14.8%]) compared with the control group (42 [25.6%]; P = 0.02), yielding a relative risk of VAP of 0.58 (95% CI, 0.37–0.90) (Table 2). Use of SSD resulted in an absolute event reduction of 10.8% (95% CI, 2.3–19.4%), indicating that 11 occurrences of VAP could be avoided for every 100 patients treated with SSD. Incidence density was 17 and 34 episodes of VAP per 1,000 ventilator days in the SSD and control groups, respectively (P = 0.002) (Table 2). The probability of remaining free of VAP over the duration of mechanical ventilation was significantly higher in the SSD group compared with the control group (log-rank test, 9.55; P = 0.002) (Figure 2). Adjustment for a prior history of cirrhosis (a baseline characteristic more frequent in patients randomized to the SSD group) did not alter the results (data not shown). The median duration of mechanical ventilation before the occurrence of VAP was similar in the two groups (8.0 d). The rate of occurrence of patients with clinically suspected VAP not microbiologically confirmed was similar in the two groups (26 [15.4%] in the SSD group vs. 18 [11.0%] in the control group; P = 0.26).
Figure 2. Cumulative rates of patients remaining free of ventilator-associated pneumonia (VAP) in the subglottic secretion drainage (SSD) and control groups, using the Kaplan-Meier method.
[More] [Minimize]SSD (n = 169) | Control (n = 164) | ARR (95% CI)* | RR (95% CI) | P Value | |
|---|---|---|---|---|---|
| VAP, no. (%) [95% CI] | 25/169 (14.8) [9.8 to 21.1] | 42/164 (25.6) [19.1 to 33.0] | 10.8 (2.3 to 19.4) | 0.58 (0.37–0.90) | 0.02 |
| Early-onset VAP (<5 d) | 2/169 (1.2) [0.1 to 4.2] | 10/164 (6.1) [3.0 to 10.9] | 4.9 (0.9 to 8.9) | 0.19 (0.04–0.87) | 0.02 |
| Late-onset VAP (>5 d)† | 23/126 (18.3) [11.9 to 26.1] | 32/97 (33.0) [23.8 to 43.3] | 14.7 (3.2 to 26.3) | 0.55 (0.35–0.88) | 0.01 |
| Incidence density per 1,000 MV d, no. [95% CI] | 17.0 [11.0–25.0] | 34.0 [24.6–45.7] | 0.002 | ||
| Prior duration of MV (d), median [IQR] | 8 [6–10] | 8 [5–13] | 0.90 |
Among the 67 episodes of VAP, 56 (83.6%) were monomicrobial (Table 3). Subglottic secretion drainage was associated with a significant reduction in VAP due to gram-positive cocci and with a trend toward lower incidence of VAP due to Pseudomonas aeruginosa (Table 3).
SSD (n = 169) | Control (n = 164) | P Value | ||||
|---|---|---|---|---|---|---|
| Monomicrobial VAP, no. (%) | 20 (11.8) | 36 (22.0) | ||||
| Polymicrobial VAP, no. (%) | 5 (3.0) | 6 (3.7) | ||||
| Organisms, no. (%) | ||||||
| Gram-negative | 26 (15.4) | 32 (19.5) | ||||
| Pseudomonas aeruginosa | 9 (5.3) | 16 (9.8) | 0.15 | |||
| Acinetobacter spp. | 2 (1.2) | 2 (1.2) | ||||
| Stenotrophomonas maltophilia | 2 (1.2) | 3 (1.8) | ||||
| Enterobacteriaceae | 10 (5.9) | 8 (4.9) | ||||
| Haemophilus spp. | 2 (1.2) | 2 (1.2) | ||||
| Branhamella catarrhalis | 1 (0.6) | 1 (0.6) | ||||
| Gram-positive | 4 (2.4) | 12 (7.3) | 0.04 | |||
| Staphylococcus aureus | 2 (1.2) | 8 (4.9) | ||||
| MSSA | 2 (1.2) | 5 (3.1) | ||||
| MRSA | 0 | 3 (1.8) | ||||
| Streptococcus pneumoniae | 2 (1.2) | 4 (2.4) | ||||
| Anaerobics* | 0 | 1 (0.6) | ||||
| Mixed oropharyngeal flora | 0 | 1 (0.6) | ||||
Both early- and late-onset VAP were significantly reduced in the patients undergoing SSD compared with the control patients (early onset, 1.2 vs. 6.1%, P = 0.02; late onset, 18.3 vs. 33.0%, P = 0.01) (Table 2).
The tracheostomy rate, duration of mechanical ventilation, and length of ICU stay were not significantly different between the two groups (Table 4). Hospital mortality was 47.3% in the SSD group and 51.2% in the control group (P = 0.51).
SSD (n = 169) | Control (n = 164) | P Value | |
|---|---|---|---|
| Postextubation laryngeal dyspnea, no. (%) [95% CI] | 8/79 (10.1) [4.5–19.0] | 4/89 (4.5) [1.2–11.1] | 0.25 |
| Tracheostomy, no. (%) [95% CI] | 25 (14.8) [9.8 to 21.1] | 19 (11.6) [7.1–17.5] | 0.42 |
| Duration of MV (d), median [IQR] | 8 [5–13] | 7 [3–15] | 0.42 |
| Length of ICU stay (d), median [IQR] | 11 [7–19] | 11 [6–20] | 0.33 |
| ICU deaths, no. (%) [95% CI] | 71 (42.0) [34.5–49.8] | 65 (39.6) [32.1–47.6] | 0.74 |
| Hospital deaths, no. (%) [95% CI] | 80 (47.3) [39.6–55.2] | 84 (51.2) [43.3–59.1] | 0.51 |
Among the patients with at least one attempt at extubation, postextubation laryngeal dyspnea occurred in 12 patients: 8 of 79 (10.1%) in the SSD group and 4 of 89 (4.5%) in the control group (P = 0.25). Reintubation was required in six patients (four in the SSD group and two in the control group) without apparent clinical consequences.
Compliance with concomitant VAP prevention strategies was similar between the two groups (Table 5). These data were available for 99.4% of the patients included in the study, except for data on the semirecumbent body position and intracuff pressure, which were available for only 96.0% of the patients. The overall use of antibiotic therapy before the onset of mechanical ventilation was similar in the two groups (36.7% in the SSD group vs. 36.0% in control group; P = 0.91). In addition, the number of patients receiving antibiotics during the VAP screening period, type of antibiotics administered, and percentage of time each type of antibiotic was administered were similar in the two groups (Table 6).
SSD (n = 169) | Control (n = 164) | P Value | ||||
|---|---|---|---|---|---|---|
| Stress ulcer prophylaxis* | ||||||
| Proton pump inhibitor, no. (%) | 141 (83.9) | 141 (86.5) | 0.54 | |||
| H2 receptor antagonists, no. (%) | 7 (4.2) | 9 (5.5) | 0.62 | |||
| Sucralfate, no. (%) | 8 (4.8) | 7 (4.3) | 1 | |||
| No prophylaxis for at least 1 d, no. (%) | 89 (53.0) | 82 (50.3) | 0.66 | |||
| % of time without ulcer prophylaxis, median [IQR]† | 66.7 [33.3–92.3] | 66.7 [33.3–90.0] | 0.60 | |||
| Enteral feeding,* no. (%) | 137 (81.6) | 133 (81.6) | 1 | |||
| % of time with enteral feeding, median [IQR]‡ | 90.5 [66.7–100] | 87.5 [62.5–100] | 0.32 | |||
| Intravenous sedation,* no. (%) | 164 (97.6) | 152 (93.3) | 0.07 | |||
| % of time with intravenous sedation, median [IQR]† | 69.8 [42.9–100] | 73.9 [46.0–100] | 0.55 | |||
| Paralytic agents,* no. (%) | 35 (20.8) | 36 (22.1) | 0.79 | |||
| % of time with paralytic agent infusion, median [IQR]† | 28.6 [11.3–53.6] | 33.3 [20.0–50.0] | 0.71 | |||
| Oral introduction of tracheal tube,* no. (%) | 167 (99.4) | 163 (100.0) | 1 | |||
| Nasal introduction of gastric tube,* no. (%) | 39 (23.2) | 47 (28.3) | 0.26 | |||
| Reintubation,* no. (%) | 21 (12.5) | 20 (12.3) | 1 | |||
| Transfer out of ICU,* no. (%) | 82 (48.8) | 66 (40.5) | 0.15 | |||
| Airway humidification with HH,* no. (%) | 21 (12.5) | 23 (14.1) | 0.74 | |||
| % of time with heated humidifiers, median [IQR]† | 65.2 [44.4–100] | 83.3 [53.8–100] | 0.65 | |||
| Transfusion, no. (%) | 64 (37.9) | 59 (36.2) | 0.82 | |||
| Trunk position‡ | ||||||
| Mean trunk position (°), median [IQR] | 29.2 [27.8–30.0] | 29.5 [27.5–30.3] | 0.68 | |||
| Patients with trunk position < 30° at least once, no. (%) | 138 (83.1) | 119 (77.3) | 0.21 | |||
| Duration of trunk position < 30° (h), median [IQR] | 27 [9–70] | 31 [10–57] | 0.71 | |||
| Intracuff pressure‡ | ||||||
| Mean intracuff pressure (cm H2O), median [IQR] | 28.3 [26.8–29.4] | 27.9 [25.9–29.7] | 0.46 | |||
| Patients with intracuff pressure < 20 at least once, no. (%) | 98 (59.4) | 87 (56.1) | 0.57 | |||
| Duration of intracuff pressure < 20 cm H2O (h), median [IQR] | 10 [4–18] | 11 [6–25] | 0.93 | |||
SSD (n = 169) | Control (n = 164) | P Value | |
|---|---|---|---|
| Antibiotic daily dose per patient, median [IQR] | 1.3 [0.88–2] | 1.33 [1–2] | 0.84 |
| Antibiotic therapy during follow-up, no. (%) | 158 (93.5) | 150 (91.5) | 0.54 |
| % of time with antibiotic therapy, median [IQR]* | 100 [77.8–100] | 100 [83.3–100] | 0.71 |
| Group 1 β-lactams,† no. (%) | 131 (77.5) | 126 (76.8) | 0.90 |
| % of time with group 1 β-lactams, median [IQR]* | 92.3 [54.5–100] | 100 [62.1–100] | 0.71 |
| Group 2 β-lactams,‡ no. (%) | 52 (30.8) | 44 (26.8) | 0.47 |
| % of time with group 2 β-lactams, median [IQR]* | 47.2 [21.1–100] | 60.6 [33.3–100] | 0.55 |
| Aminoglycosides, no. (%) | 60 (35.5) | 51 (31.1) | 0.42 |
| % of time with aminoglycosides, median [IQR]* | 25.8 [17.1–50.0] | 33.3 [14.3–61.3] | 0.44 |
| Quinolones, no. (%) | 36 (21.3) | 34 (20.7) | 1 |
| % of time with quinolones, median [IQR]* | 50.0 [29.9–93.5] | 60.0 [33.3–1] | 0.50 |
| Macrolides, no. (%) | 42 (24.9) | 49 (29.9) | 0.32 |
| % of time with macrolides, median [IQR]* | 42.9 [23.1–100] | 66.7 [25.0–100] | 0.27 |
| Glycopeptides, no. (%) | 26 (15.4) | 20 (12.2) | 0.43 |
| % of time with glycopeptides, median [IQR]* | 31.7 [20.0–75.0] | 48.3 [21.1–87.8] | 0.44 |
| Imidazoles, no. (%) | 18 (10.7) | 17 (10.4) | 1 |
| % of time with imidazoles, median [IQR]* | 69.7 [39.2–100] | 84.6 [33.3–100] | 0.91 |
In our study, intermittent SSD resulted in a significant reduction in the overall incidence of VAP, and in the incidence of early- and late-onset VAP in an unselected population of ICU patients. Moreover, the absolute risk reduction in our patients indicates that 11 occurrences of VAP could be avoided for every 100 patients treated by SSD.
One strength of our study is its multicenter design with a relatively large sample size, equivalent to 60% of the unselected ICU patients included in the latest meta-analysis on VAP prevention (13). Other strengths include the concealment of the allocation process, complete follow-up of included patients, the intention-to-treat analysis, and high compliance with the planned subglottic secretion drainage procedure. We also carefully ensured that potentially confounding factors, including compliance with preventive measures for VAP other than SSD in the participating centers, use of antibiotics, and tracheal cuff pressure were equally distributed among the two groups. Reporting of these potential confounding factors was limited in previously published studies (7, 8, 12). Last, each subjective clinical suspicion of VAP resulted in an objective diagnostic pulmonary sample testing with largely accepted thresholds to confirm or infirm the diagnosis of VAP (25, 26).
It could be expected that an intervention reducing the rate of microbiologically confirmed VAP, such as SSD in the present study, would also reduce the rate of clinically suspected VAP. Several conditions, not expected to be prevented by the use of SSD, could mimic a suspicion of VAP, including atelectasis, pulmonary edema, and pulmonary embolism. In our study, the rate of patients with nonmicrobiologically confirmed VAP is not significantly different between the SSD group (15.4%) and the control group (11.0%). Moreover, patients with microbiologically confirmed VAP represent only about 50% of all the patients with clinically suspected VAP. Therefore, the significant reduction in microbiologically confirmed VAP translates into only a trend to reduction (30.2 vs. 36.6%; P = 0.25), not a significant reduction, in clinically suspected VAP, especially in a study specifically powered to detect a difference in the rate of microbiologically confirmed VAP only.
The meta-analysis on VAP prevention, published while the present study was being performed, concluded that SSD appeared to be effective in preventing early-onset VAP only (13). Despite an overall low rate of early-onset VAP in the control group, potentially attributable to the high rate of antibiotic use in our study patients, our results confirm the protective effect of SSD on early-onset pneumonia in mechanically ventilated patients. Furthermore, unlike the previous studies included in the meta-analysis (13), SSD was also associated with a significantly lower incidence of late-onset VAP in our study. The large sample size of patients with prolonged (≥5 d) mechanical ventilation in our study might have contributed to reveal this beneficial effect. In addition, because of the low rate of early-VAP, few patients were censored after an early episode of VAP, leaving more patients assessable for late-onset VAP. Our finding of a protective role for SSD in late-onset VAP is consistent with a randomized trial assessing the simultaneous use of SSD and a new polyurethane cuff (28). However, the respective contribution of SSD and the new cuff material to the reduction in late-onset VAP was unclear (28). More generally, our finding supports the orotracheal aspiration pathway as an important pathophysiologic mechanism for both early- and late-onset VAP.
Large variations in the volume of retrieved subglottic secretions have been previously reported. In one observational study, secretions were retrieved in less than 50% of collection attempts with suctioned volume ranging from 0.3 to 15.0 ml (29). This variability was confirmed in our study and might be partly explained by the wide variability in the volume of secretions present above the endotracheal cuff, as shown in an experimental and clinical imaging study (30). Other factors including secretion viscosity, suction pressure, and difficulties in maintaining patency of the subglottic suction line and cuff inflation have been proposed to explain this variability (31). In our study and a previous one (7), manual intermittent suction was used to aspirate subglottic secretions. Although other suctioning techniques including automatic intermittent (8) or continuous suctioning (10–12) have also been proposed, to date, the most effective technique to remove subglottic secretions remains unclear.
Despite the marked decrease in the incidence of microbiologically confirmed VAP with the use of SSD in our study, secondary end points, that is, mortality, and duration of mechanical ventilation and ICU stay, were not influenced by the use of SSD. The real impact of SSD on important patient outcomes could therefore be questioned. However, many previous studies of other measures used to prevent VAP, including endotracheal tube with a polyurethane cuff (28), oral decontamination with chlorhexidine (32), or silver-coated endotracheal tubes (33), failed to demonstrate a significant reduction in mortality associated with the reduction in VAP. Considering that VAP could be associated with an absolute excess mortality of 10 to 30% (1), a prospective randomized trial of a preventive intervention resulting in an absolute 10% reduction in the rate of VAP (as in our study) could be expected to result in an ICU mortality reduction of only 1 to 3%. For such a reduction to be statistically significant, and depending on the expected baseline mortality rate in the control group (from 20 to 40%), a large number of patients (several thousands) would have to be included in each arm. It is therefore highly likely that future randomized controlled trials on VAP prevention will remain underpowered to detect a beneficial effect on mortality.
Previous studies on VAP prevention, and specifically SSD, also failed to show an impact on durations of mechanical ventilation and stay in ICU. However, the meta-analysis by Dezfulian and colleagues, which pooled the results of the first five trials of SSD, demonstrated a significant reduction in these two variables (13), suggesting that our study, taken individually, as well as the previous studies conducted on SSD, was insufficiently powered to detect such a beneficial effect.
In a trial of VAP prevention such as our study, several factors including the rate of antibiotic use and the rate of VAP in the control group are known to potentially affect the generalizability of the results. The rate of 1.31 antibiotic drugs administered per patient and per day during the VAP screening period in our ICU population was high but in accordance to that reported in other populations of critically ill patients, such as in large national databases of ICU patients in Germany (1.33) and Sweden (1.25) (34, 35). The 25.6% VAP rate in our control group was relatively high, but close to the 22.1% VAP rate in the control patients (also intubated with the Hi-Lo Evac tube as in our study) in a randomized trial assessing the simultaneous use of SSD and a new polyurethane cuff in patients (28) despite lower severity score on admission and ICU mortality than in our study patients. Another factor that might influence the generalizability of our study results is that almost 50% of the patients who required mechanical ventilation in the participating ICUs during the study period were not intubated with the Hi-Lo Evac tube. This resulted from the heterogeneous availability of the Hi-Lo Evac tube through all the pre- and in-hospital structures involved in emergency intubation of critically ill patients before ICU admission. Because many mechanically ventilated patients in the ICU worldwide are intubated before their admission to the ICU, exclusion of these patients from our study, as in previous trials of SSD (7, 8, 12), would only poorly reflect daily life practice. Generalizability of our results would have been affected if the use of a Hi-Lo Evac tube for pre-ICU intubation had resulted from systematic patient selection, which was not the case in our study.
Our study has several limitations. First, as in previous randomized trials on SSD, physicians and nurses could not be blinded to the randomization arm. However, the potential effect of lack of blinding was minimized by several factors. The use of other interventions with a potentially preventive effect on VAP was similar in the two groups. Clinical suspicion of VAP in the SSD group (30.2%) was only slightly lower than in the control group (36.6%), suggesting that the lower rate of VAP in the SSD group could not be attributed to a difference in physician behavior in detecting VAP. Distal microbiological samples and strict biological criteria were used to confirm clinically suspected VAP. In addition, microbiologists were blinded to the randomization arm. Second, only three-quarters of the planned number of patients had been included at the end of the planned study period. However, although positive results of trials that include a lower-than-expected number of patients are often attributed to a higher-than-expected treatment effect (36), the observed absolute risk reduction in VAP incidence in our study (10.8%) was close to that expected (10%). Third, as discussed previously, the number of patients excluded because of the use of another device than the Hi-Lo Evac tube for intubation before ICU admission was high. Fourth, an economic evaluation was not performed. However, in a cost–effectiveness analysis using a decision-model approach, regular use of the Hi-Lo Evac tube was shown to produce significant savings due to the reduction in VAP incidence, irrespective of the higher cost of the device (37).
Although continuous SSD has been associated with tracheal wall injury in one experimental study (19), only one observational study has reported a high incidence of postextubation laryngeal dyspnea (38). However, the small sample size (n = 8) precluded any firm conclusions about the safety of SSD. In our study, the incidence of postextubation events, including laryngeal dyspnea with or without the need for immediate reintubation, as well as the overall rate of reintubation, did not differ significantly between the two groups. This finding suggests that clinicians should not refrain from using intermittent SSD because of safety concerns.
In conclusion, the results of this randomized, multicenter study demonstrated that intermittent subglottic secretion drainage significantly reduces the incidence of microbiologically confirmed VAP, including late-onset VAP, without any notable adverse events. These results should encourage ICU physicians to progressively integrate SSD into their VAP preventive measures and prompt physicians involved in pre-ICU care to use endotracheal tubes permitting SSD.
The authors thank the nursing staff and the staff of the department of microbiology at each participating center for their assistance. The authors dedicate this article to the memory of their friend and colleague, Jérome Baudot, who was behind the participation of the Hospital of Avignon in this trial and who did not live to see this manuscript.
| 1. | Chastre J, Fagon JY. Ventilator-associated pneumonia. Am J Respir Crit Care Med 2002;165:867–903. |
| 2. | Klompas M. Does this patient have ventilator-associated pneumonia? JAMA 2007;297:1583–1593. |
| 3. | American Thoracic Society, Infectious Diseases Society of America. Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med 2005;171:388–416. |
| 4. | Oikkonen M, Aromaa U. Leakage of fluid around low-pressure tracheal tube cuffs. Anaesthesia 1997;52:567–569. |
| 5. | Petring OU, Adelhoj B, Jensen BN, Pedersen NO, Lomholt N. Prevention of silent aspiration due to leaks around cuffs of endotracheal tubes. Anesth Analg 1986;65:777–780. |
| 6. | Young PJ, Rollinson M, Downward G, Henderson S. Leakage of fluid past the tracheal tube cuff in a benchtop model. Br J Anaesth 1997;78:557–562. |
| 7. | Mahul P, Auboyer C, Jospe R, Ros A, Guerin C, el Khouri Z, Galliez M, Dumont A, Gaudin O. Prevention of nosocomial pneumonia in intubated patients: respective role of mechanical subglottic secretions drainage and stress ulcer prophylaxis. Intensive Care Med 1992;18:20–25. |
| 8. | Smulders K, van der Hoeven H, Weers-Pothoff I, Vandenbroucke-Grauls C. A randomized clinical trial of intermittent subglottic secretion drainage in patients receiving mechanical ventilation. Chest 2002;121:858–862. |
| 9. | Rello J, Sonora R, Jubert P, Artigas A, Rue M, Valles J. Pneumonia in intubated patients: role of respiratory airway care. Am J Respir Crit Care Med 1996;154:111–115. |
| 10. | Bouza E, Jesus Perez M, Munoz P, Rincon C, Maria Barrio J, Hortal J. Continuous aspiration of subglottic secretions (CASS) in the prevention of ventilator-associated pneumonia in the postoperative period of major heart surgery. Chest 2008;134:938–946. |
| 11. | Kollef MH, Skubas NJ, Sundt TM. A randomized clinical trial of continuous aspiration of subglottic secretions in cardiac surgery patients. Chest 1999;116:1339–1346. |
| 12. | Valles J, Artigas A, Rello J, Bonsoms N, Fontanals D, Blanch L, Fernandez R, Baigorri F, Mestre J. Continuous aspiration of subglottic secretions in preventing ventilator-associated pneumonia. Ann Intern Med 1995;122:179–186. |
| 13. | Dezfulian C, Shojania K, Collard HR, Kim HM, Matthay MA, Saint S. Subglottic secretion drainage for preventing ventilator-associated pneumonia: a meta-analysis. Am J Med 2005;118:11–18. |
| 14. | Dodek P, Keenan S, Cook D, Heyland D, Jacka M, Hand L, Muscedere J, Foster D, Mehta N, Hall R, et al. Evidence-based clinical practice guideline for the prevention of ventilator-associated pneumonia. Ann Intern Med 2004;141:305–313. |
| 15. | Tablan OC, Anderson LJ, Besser R, Bridges C, Hajjeh R. Guidelines for preventing health-care–associated pneumonia, 2003: recommendations of CDC and the Healthcare Infection Control Practices Advisory Committee. MMWR Recomm Rep 2004;53:1–36. |
| 16. | Muscedere J, Dodek P, Keenan S, Fowler R, Cook D, Heyland D. Comprehensive evidence-based clinical practice guidelines for ventilator-associated pneumonia: prevention. J Crit Care 2008;23:126–137. |
| 17. | Coffin SE, Klompas M, Classen D, Arias K, Podgorny K, Anderson DJ, Burstin H, Calfee DP. Strategies to prevent ventilator-associated pneumonia in acute care hospitals. Infect Control Hosp Epidemiol 2008;29:S31–S40. |
| 18. | Cook D, Ricard JD, Reeve B, Randall J, Wigg M, Brochard L, Dreyfuss D. Ventilator circuit and secretion management strategies: a Franco-Canadian survey. Crit Care Med 2000;28:3547–3554. |
| 19. | Berra L, De Marchi L, Panigada M, Yu ZX, Baccarelli A, Kolobow T. Evaluation of continuous aspiration of subglottic secretion in an in vivo study. Crit Care Med 2004;32:2071–2078. |
| 20. | Lacherade JC, Guezennec P, Debbat K, Hayon J, Monsel A, De Jonghe B, Fangio P, Appere de Vecchi C, Outin H, Bastuji-Garin S. Impact of subglottic secretions drainage on ventilator-associated pneumonia: a randomized multicentre trial [abstract]. Intensive Care Med 2008;34:S102. |
| 21. | Yanagawa Y, Sakamoto T, Okada Y. Recovery from a psychotropic drug overdose tends to depend on the time from ingestion to arrival, the Glasgow Coma Scale, and a sign of circulatory insufficiency on arrival. Am J Emerg Med 2007;25:757–761. |
| 22. | Keenan SP, Dodek P, Martin C, Priestap F, Norena M, Wong H. Variation in length of intensive care unit stay after cardiac arrest: where you are is as important as who you are. Crit Care Med 2007;35:836–841. |
| 23. | Nolan JP, Laver SR, Welch CA, Harrison DA, Gupta V, Rowan K. Outcome following admission to UK intensive care units after cardiac arrest: a secondary analysis of the ICNARC Case Mix Programme Database. Anaesthesia 2007;62:1207–1216. |
| 24. | Lacherade JC, Auburtin M, Cerf C, Van de Louw A, Soufir L, Rebufat Y, Rezaiguia S, Ricard JD, Lellouche F, Brun-Buisson C, et al. Impact of humidification systems on ventilator-associated pneumonia: a randomized multicenter trial. Am J Respir Crit Care Med 2005;172:1276–1282. |
| 25. | Trouillet JL, Chastre J, Vuagnat A, Joly-Guillou ML, Combaux D, Dombret MC, Gibert C. Ventilator-associated pneumonia caused by potentially drug-resistant bacteria. Am J Respir Crit Care Med 1998;157:531–539. |
| 26. | Pham LH, Brun-Buisson C, Legrand P, Rauss A, Verra F, Brochard L, Lemaire F. Diagnosis of nosocomial pneumonia in mechanically ventilated patients: comparison of a plugged telescoping catheter with the protected specimen brush. Am Rev Respir Dis 1991;143:1055–1061. |
| 27. | Fartoukh M, Maitre B, Honore S, Cerf C, Zahar JR, Brun-Buisson C. Diagnosing pneumonia during mechanical ventilation: the clinical pulmonary infection score revisited. Am J Respir Crit Care Med 2003;168:173–179. |
| 28. | Lorente L, Lecuona M, Jimenez A, Mora ML, Sierra A. Influence of an endotracheal tube with polyurethane cuff and subglottic secretion drainage on pneumonia. Am J Respir Crit Care Med 2007;176:1079–1083. |
| 29. | O'Neal PV, Munro CL, Grap MJ, Rausch SM. Subglottic secretion viscosity and evacuation efficiency. Biol Res Nurs 2007;8:202–209. |
| 30. | Greene R, Thompson S, Jantsch HS, Teplick R, Cullen DJ, Greene EM, Whitman GJ, Hulka CA, Llewellyn HJ. Detection of pooled secretions above endotracheal-tube cuffs: value of plain radiographs in sheep cadavers and patients. AJR Am J Roentgenol 1994;163:1333–1337. |
| 31. | Depew CL, McCarthy MS. Subglottic secretion drainage: a literature review. AACN Adv Crit Care 2007;18:366–379. |
| 32. | Koeman M, van der Ven AJ, Hak E, Joore HC, Kaasjager K, de Smet AG, Ramsay G, Dormans TP, Aarts LP, de Bel EE, et al. Oral decontamination with chlorhexidine reduces the incidence of ventilator-associated pneumonia. Am J Respir Crit Care Med 2006;173:1348–1355. |
| 33. | Kollef MH, Afessa B, Anzueto A, Veremakis C, Kerr KM, Margolis BD, Craven DE, Roberts PR, Arroliga AC, Hubmayr RD, et al. Silver-coated endotracheal tubes and incidence of ventilator-associated pneumonia: the NASCENT randomized trial. JAMA 2008;300:805–813. |
| 34. | Meyer E, Schwab F, Jonas D, Rueden H, Gastmeier P, Daschner FD. Surveillance of antimicrobial use and antimicrobial resistance in intensive care units (SARI). 1. Antimicrobial use in German intensive care units. Intensive Care Med 2004;30:1089–1096. |
| 35. | Walther SM, Erlandsson M, Burman LG, Cars O, Gill H, Hoffman M, Isaksson B, Kahlmeter G, Lindgren S, Nilsson L, et al. Antibiotic prescription practices, consumption and bacterial resistance in a cross section of Swedish intensive care units. Acta Anaesthesiol Scand 2002;46:1075–1081. |
| 36. | Montori VM, Devereaux PJ, Adhikari NK, Burns KE, Eggert CH, Briel M, Lacchetti C, Leung TW, Darling E, Bryant DM, et al. Randomized trials stopped early for benefit: a systematic review. JAMA 2005;294:2203–2209. |
| 37. | Shorr A, O'Malley P. Continuous subglottic suctioning for the prevention of ventilator-associated pneumonia: potential economic implications. Chest 2001;119:228–235. |
| 38. | Girou E, Buu-Hoi A, Stephan F, Novara A, Gutmann L, Safar M, Fagon JY. Airway colonisation in long-term mechanically ventilated patients: effect of semi-recumbent position and continuous subglottic suctioning. Intensive Care Med 2004;30:225–233. |
