Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Continuous subglottic suction is effective for prevention of ventilator associated pneumonia

Three Part Question

In patients undergoing [mechanical ventilation] does [subglottic suction] reduce the incidence of [Ventilator associated pneumonia]?

Clinical Scenario

You performed a difficult Aortic Valve replacement and triple-coronary arterial-bypass-graft on a 77-year-old man, with a 30-year history of smoking. The operation proceeded uneventfully, but in the Intensive care it was not possible to extubate him on the first night due to basal collapse, and over the next few days he develops a ventilator-associated-pneumonia (VAP).
You search the internet for manoeuvres that may avoid this frustrating complication and find that continuous subglottic suction would avoid pooling of secretions around the endotracheal tube and thus perhaps reduce VAP. Thus you resolve to search for evidence for this simple intervention.

Search Strategy

Medline 1966–Sept 2004 using the Ovid interface.
[ OR OR] AND [exp pneumonia/OR OR secreti$.mp OR ventilat$.mp OR aspirat$.mp] Limit to human studies.

Search Outcome

Four hundred and fifty seven papers were found from the reported search and cross-checking reference lists, of which 13 were deemed to be relevant. This included 7 RCTs, one cohort study and several reviews, from which 3 were selected. The papers are presented in the table

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Girou et al,
18 critically ill patients requiring mechanical ventilation for >5 days. Randomised to continuous subglottic suction and semi-recumbent body position (n=8) or to receive standard care in spine position (n=10)PRCT (level 2b)Daily sampling and culturing of oropharyngeal and tracheal secretionsMedian bacterial count in trachea were 6.6 log 10. CFU/ml (interquartile range, IQR, 4.4–8.3) in patients who received continuous suction and 5.1log 10 CFU/ml (IQR 3.6– 5.5) in control patients. Study found no significant difference in the bacterial count between the 2 groups studied.Study period was only ten days from start of mechanical ventilation. Sample size was small (n=18) 9 patients were excluded from study but no satisfactory explanation was provided. Most patients had heavily colonised tracheal secretions from day 1 of study and therefore sampling tracheal secretions in this case was an inappropriate indicator of VAP
Valles et al,
190 critically ill general patients requiring mechanical ventilation for >3 days. Randomised to receive continuous aspiration of subglottic secretions (CASS)(n=76) or to receive usual care (n=77)PRCT (level 1b)Duration of ventilationSubglottic suction (CASS) 13±1 day. Control group 11±1 day p> CASS group and 58.4% of control group received an antibiotic agent at the time of randomisation Of 190 patients entered into the study, 15 were extubated and 16 died before the end of the study.
Incidence of VAPSubglottic suction (CASS) 14/76(18.4%) and 19.9 episodes/1000 ventilator days in the patients. Control patients 39/77(32.5%) and 39.6 episodes/1000 ventilator days (RR=1.98;CI 95%:1.03 to 3.82)
Time to VAPEpisodes of VAP occurred later in patients receiving CASS (12.0±7.1 days) than in control patients (5.9± 2.1 days) (P= 0.003)
Smulders et al,
150 patients admitted to a general ICU, expected to receive ventilation >72h. Intermittent suction ET tube with intermittent secretion drainage every 20s, for 8s (n=75). Control group standard ET tube (n=75)PRCT (level 1b)Incidence of ventilator associated pneumoniaIntermittent suction 3/75(4%). Control patients 12/75(16%). P=0.014Clinical diagnosis of VAP without quantitative cultures of LRTI Chest radiograph interpreted by one radiologist only- Possiblilty of bias cannot be excluded
Duration of mechanical ventilationIntermittent suction 5.8±4.4 days. Control patients 7.1±5.4 days p=NS
Rello, Valles et al,
All patients intubated in the ICU or the emergency department (n=83) All patients from 1993 to 1994 were intubated using the HI-Low Evac ET tube, Mallinckrodt Laboratories, Ireland. Intracuff pressure monitored 4 hourly, subglottic failure defined as no secretions for 24 h.Cohort study (level 2b)Risk factors for VAPFailure of CASS increases risk of VAP RR=5.29 (95% CI=1.29 to 22.64). Non pneumonia pts 30% failure, pneumonia patients 43% failure. Increased risk in patients with cuff pressures <20 cmH2O RR=2.57, (95% CI=0.78 to 8.03)Single centre study- Possiblilty of institutional bias in patient selection or institutional practices Study limited to only first 8 days of ventilation 7(8.43%) of the patients presented to the hospital with community acquired pneumonia
Incidence of VAPPneumonia occurred in 12/83 patients
Kollerf et al,
Systematic review of a wide range of non-pharmacological and pharmacological preventative strategies against VAPSystematic review (level 1a)Guide for the development of a programme to prevent VAPContinuous subglottic suction recommended for clinical use as non-pharmacological measure to prevent VAP (grade A recommendation). Widely used techniques such as chest physiotherapy were identified as ineffective in preventing VAPHealth status of patients in each study reviewed was not taken into consideration Only 2 RCTs identified Author bias in grading of recommendations
Kollef et al,
343 patients undergoing cardiac surgery and requiring mechanical ventilation in the cardiothoracic ITU (CTITU) 160 patients were randomly assigned to receive CASS using Hi-Low- Evac ET tube, and 183 patients were assigned to receive routine post-operative care without CASSPRCT(level 2b)Incidence of VAPCASS patients 8/160(5.0%). Routine care patients 15/183(8.2%) RR=0.61%; (CI 95% 0.27 to 1.40) p=0.238Diagnosis of VAP was made clinically and not confirmed by examination of bronchoscopically obtained specimens. No differences in duration of ventilation, hospital stay, or mortality Flawed randomization technique using patient birth year.
Onset of VAPCASS patients mean 5.6±2.3 days. Routine care patients mean 2.9±1.2 days; (p=0.006)
Pneumatikos et al,
61 patients admitted to the ICU who were predicted to need ventilation for >5 days. Patients were randomly assigned to receive Selective Decontamination of Subglottic Area (SDSA) using suction and antibiotics (n=30) or placebo (n=31) SDSA was administered by continuous infusion of a suspension containing 73mg polymyxin E, 73mg tobramycin and 500mg amphotericin B in 500 ml 0.9 saline solution at an infusion rate of 2 ml/h in the subglottic area, and intermittent suctionPRCT(level 1b)Incidence of VAPSDSA patients 5/30(16%). Control patients 16/31 (53%) p<0.01Randomisation process not described Patients receiving ventilation between days 2 and 5 excluded
Gastric fluid and tracheal secretion culturesNegative bronchial cultures in 14 of 31 (45%) SDSA patients vs 3 (10%) control patients p<0.001. Overall 46% reduced tracheal colonisation and 70% decreased incidence of VAP in patients receiving SDSA.
Collard et al,
Systematic review of all papers, using Medline, Cochrane Library, reference checking, DARE 1966-2001 3 RCTs foundSystematic Review (level 1a)RecommendationAspiration of subglottic secretions is a promising new strategy for the prevention of ventilator-associated pneumonia but cannot be recommended for general use because of the mixed results in the literature (grade IIa). It may be most effective in patients requiring prolonged (<3 days mechanical ventilationOnly english language articles searched
Mahul et al,
145 patients with probable intubation for more than 3 days admitted to a general ICU (46% Medical, 54% surgical) Hourly suction groups. Hourly suction and Hi-Lo Evac ET tube vs standard ET tube. Stress ulcer prophylaxis Either aluminium hydroxide (20 ml/6 h) or Sucralfate 1g/6 h given. Patients randomized to one of 4 groups in combinations of above protocolsSingle blind PRCT (level 1b)Incidence of nosocomial pneumoniaSubglottic suction groups 9/70(13%). Standard intubation groups 21/75 (29%) p=0.04Infilatrates on CXR after day 2 and positive bronchoalveolar lavage was considered positive for nosocomial pneumonia
Incidence of nosocomial pneumoniaSucralfate groups 13/73 (18%). Aluminium Hydroxide gps 17/72 (24%) p=NS
Dodek et al,
Systematic review of RCTs that presented evidence on prevention of ventilator associated pneumonia Searched Medline, Embase, and Cochrane databases up to April 2003Systematic Review (level 1a)Drainage of subglottic secretions5 studies (labelled as level 2 trials) found that shows that subglottic suction decreases VAP. Recommend that clinicians consider the use of subglottic suction secretion drainageMissed the study by Girou et al and Pneumatikos et al.
Semi-recumbent positioningRecommend 45 degree semi-recumbent positioning on the basis of one level 1 trial
Shorr et al,
Theoretical economic analysis of 100 patients requiring non-elective ventilation and ICU care. From literature review incidence of VAP was estimated as 25% and the benefit of subglottic suction was estimated as a 30% relative risk reductionEconomic analysis (level 3b)Cost benefitSubglottic suction would save $4,992 per case of VAP preventedOnly 3 RCTs identified using subglottic suction to estimate costs
Cost of subglottic suctionConventional ET tubes cost $1 per patient. Continuous suction tubes cost $15 per patient. A case of VAP requires 6-9 extra days in ICU
Cooke et al

Survey of 84 French and Canadian university affilitated ICUs into the use of ventilator circuit and secretion management strategiesSurvey (level 2b)Use of subglottic secretionFrance 4.2% of units. Canada 3.2% of unitsThe response rate was 72/84 (85.7%) for French and 31/32 (96.9%) for Canadian ICUs.
Reason for non use of subglottic suctionLack of convincing benefit was cited as the main reason for non use by 52.7% of respondents. 30% cited cost and non-availability by 33% of respondents
Metz et al,
39 criticall ill patients with an expected ventilation time of more than 3 days Randomized to 3 groups. All patients received the Hi-Lo Evac II ET tube subglottic lavage gp n=10, subglottic flushing every 3 h with 20ml saline Pharyngeal lavage gp n=15, above protocol plus oropharyngeal lavage of 500ml saline every 6 h Control gp n=14, Hi-lo Evac tube, only with once daily 2ml lavageUnblinded PRCT (level 2b)Subglottic colonisationNo differences between groupsNon-blinded Results not clearly displayed, very small sample sizes with sample size calculations
Mean day of pneumonia onsetSubglottic lavage gp. Median day 4 (range 3-6). Pharyngeal lavage gp. Median day 4 (range 2-6). Control gp. Median day 5 (range 4-6)
Changes in lavage induced changes in floraLavage groups showed a significantly increased number of changes of subglottic flora compared to controls


Valles et al. in 1995 performed a large randomized study in 190 patients who were likely to be ventilated for more than 3 days. They found a relative reduction of 43% in ventilator-associated-pneumonia (VAP) and continuous suction delayed the time to the onset of VAP from a mean of 5 days to 12 days (NNT of 5). The same authors then performed a cohort study in 83 patients intubated in their general ICU or emergency department, where all patients received continuous subglottic suction. They found that 43% of patients who developed pneumonia suffered failure of the suction compared to 30% of those who did not. In addition low cuff pressure was also significantly associated with pneumonia. Smulders et al. in 2002 performed a large PRCT in 150 general ICU patients with predicted ventilation of over 3 days. They used subglottic suction, but in order to avoid possible tracheal wall damage they instituted intermittent suction with 8 seconds of suction every 20 seconds. They found a significantly reduced incidence of pneumonia in the suction group, reducing the incidence of VAP from 16% to 4%. This is a number needed to treat of 8. Pneumatikos et al performed a slightly different study where they randomised 61 patients to use of the Hi-Low Evac subglottic suction tubing or controls. However, instead of simple suction they used a continuous infusion of antibiotics down the tube with intermittent suction. They found a marked reduction of VAP from 53% down to 16% Mahul performed a randomized trial in 145 patients with a predicted intubation time of over 3 days. A significant reduction in nosocomial pneumonia was found with hourly subglottic suction. 29% of controls suffered pneumonia compared to only 13% in the suction group. Kollef et al performed the only study in patients post cardiac surgery. They randomized 343 patients using their birth years to either continuous subglottic suction or normal ET-tube. They found a non-significant reduction of VAP from 8.2% in controls to 5% in the subglottic suction group, P=0.238. They did, however, find a significant delay in the onset of VAP, with a mean time of 2.9 days in the control group compared to 5.6 days in the subglottic suction group (P=0.006). They concluded that 1006 patients would have been required to achieve significance for the difference that they found, but that if their findings were significant, the number needed to treat to prevent one pneumonia in all cardiac surgical patients would be 32. Marin Kollef also performed a systematic review for the New England Journal of Medicine in the same year and concluded that there was grade A evidence to support the use of continuous subglottic suction routinely. Among the many reviews in the literature Collard performed one of the most recent and well performed. They stated that the evidence was in fact quite mixed, grading it at IIa and stated that continuous suction has not convincingly been shown to reduce VAP in all patients but should perhaps be considered in all patients who may require more than 3 days of ventilation. In addition a recent systematic review performed by Dodek et al. in 2004 for the Canadian Critical Care trials Group recommended that clinicians consider the use of subglottic secretion drainage in all their patients. A survey of practise was performed in 2002 in France and Canada into protocols used to reduce VAP in university affiliated ICUs. They found that less than 5% of units used subglottic suction. The primary reason cited was lack of evidence for benefit, with cost and lack of availability also cited. In contrast an interesting cost analysis was performed in 2003. The cost of subglottic suction is $15 per tube compared to $1 per conventional ET tube, and the cost of one episode of VAP was estimated to be $5,365. With a 30% reduction assumed for the suction strategy they estimated the cost benefit to be $4,992 per case of VAP saved. They also reported that the cost of VAP would have to be as low as $330 for the strategy to be non-cost effective. Thus, in summary, clinical benefits have been shown for subglottic suction in the highest-risk patients. Only 8 patients being ventilated for more than 3–5 days need to be treated to prevent one episode of pneumonia. The benefits markedly reduce when you consider lower risk patients such as patients post cardiac surgery where 32-patients must be treated to prevent one pneumonia. Subglottic suction has also been shown to delay the onset of VAP but no benefits in terms of ventilation time, hospital stay or mortality benefit have ever been shown. However, it has been shown that even if the benefits of subglottic suction are marginal, the cost benefit of this cheap intervention is likely to substantial.

Clinical Bottom Line

Subglottic suction significantly reduces the incidence of VAP in high-risk patients (NNT of 8 if ventilated over 3 days), although the benefit is lower in elective cardiac patients. Subglottic suction is currently not commonly used, but even with marginal benefits, its use is likely to be highly cost effective.


  1. 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.
  2. 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.
  3. Smulders K, van der HH, 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.
  4. 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.
  5. Kollef MH. Current Concepts: The Prevention of Ventilator-Associated Pneumonia New Eng J Med 1999;340:627–634.
  6. Kollef MH, Skubas NJ, Sundt TM. A randomized clinical trial of continuous aspiration of subglottic secretions in cardiac surgery patients. [see comment]. Chest 1999;116:1339–1346.
  7. Pneumatikos I, Koulouras V, Nathanail C, Goe D, Nakos G. Selective decontamination of subglottic area in mechanically ventilated patients with multiple trauma. Intensive Care Med 2002;28:432–437.
  8. Collard HR, Saint S, Matthay MA. Prevention of ventilator-associated pneumonia: an evidence-based systematic review.[see comment][Review][69 refs]. Ann Intern Med 2003;138:494–501.
  9. Mahul P, Auboyer C, Josepe R, Ros R, 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.
  10. Dodek P, Keenan S, Cook D, Heyland D et al. Canadian Critical Care Trials Group, Canadian Critical Care Society. Evidence-based clinical practice guideline for the prevention of ventilator-associated pneumonia. Ann Intern Med 2004;141:305–313.
  11. Shorr AF, O'Malley PG. Continuous subglottic suctioning for the prevention of ventilator-associated pneumonia: potential economic implications. Chest 2001;119:228–235.
  12. Cook D, Ricard JD, Reeve B, Randall J, Wigg M, Brochard L, Dreyfuss D. Ventilator circuit and secretion management strategies: a Franco-Canadian survey.[see comment]. Crit Care Med 2000;28:3547–3554.
  13. Metz C, Linde H-J, Gobel L, Taeger K. Influence of intermittent subglottic lavage on subglottic colonisation and ventilator-associated pneumonia. Clinical Intensive Care 1998;9:20–24.