Indian Journal of Medical Sciences Medknow Publications on behalf of Indian Journal of Medical Sciences Trust ISSN: 0019-5359 EISSN: 1998-3654 Vol. 61, Num. 4, 2007, pp. 201-211

Indian Journal of Medical Sciences, Vol. 61, No. 4, April, 2007, pp. 201-211

Original Contributions

Comparison of closed endotracheal suction versus open endotracheal suction in the development of ventilator-associated pneumonia in intensive care patients: An evaluation using meta-analytic techniques

Peter JohnVictor, Chacko Binila, Moran JohnL

Department of Medical Intensive Care, Christian Medical College and Hospital, Vellore
Correspondence Address:Medical Intensive Care Unit, Christian Medical College and Hospital, Vellore - 632 004, peterjohnvictor@yahoo.com.au

Code Number: ms07033

Abstract

Keywords: Endotracheal suctioning, meta-analysis, publication bias, random effects

Ventilator-associated pneumonia (VAP) is a common nosocomial infection in the intensive care unit (ICU) with an incidence ranging from 6.8 to 44%.[1],[2],[3],[4] VAP increases costs,[5],[6] length of hospital stay[7] and mortality,[8],[9] and any strategy to reduce its occurrence is worthy of consideration. Endotracheal suctioning is performed in intubated mechanically ventilated patients as a routine essential part of care to clear endotracheal secretions. Two methods of endotracheal suctioning are in practice - the open endotracheal suctioning (OES) system, where suctioning is performed after disconnecting the respiratory circuits and using sterile single-use suction catheters. This technique of suctioning has been reported to be associated with arterial desaturation, inability to maintain PEEP and cardiac arrhythmias, particularly in patients with cardiorespiratory instability.[10],[11] In the closed endotracheal suctioning (CES) system, which was developed to minimize these complications, suctioning is performed without disconnecting the respiratory circuit and uses multi-use in-line catheters that are enclosed in a sheath along with the respiratory circuit. This mode of suctioning has comparatively fewer physiological disturbances and consequences during suctioning[10],[12] and provides ease of use, given that only one operator is required for suctioning.[13] Further, CES is postulated to reduce VAP rates by decreasing environmental contamination during suctioning.[14] These potential advantages have led to the conduct of several randomized controlled trials (RCTs) that compared CES and OES. These individual trials failed to show a superiority of one type of suctioning over the other. This evaluation was undertaken to assess if there was any advantage of CES over OES with respect to the development of VAP. The hypothesis was that there would be no difference in the incidence of VAP between CES and OES suctioning. Secondary outcomes assessed in the study were mortality, duration of ventilatory support, as well as hospital and ICU length of stay (LOS).

Materials and Methods

Selection of trials

RCTs comparing CES with OES in mechanically ventilated patients were considered for inclusion. CES was defined as endotracheal suctioning performed without disconnection from the respiratory circuit, employing a multi-use in-line suctioning catheter. OES was defined as endotracheal suctioning done after disconnection of the respiratory circuit and employing a single-use-suctioning catheter under aseptic precautions. Only trials reporting VAP rates and/or mortality were considered for inclusion. Studies reporting only physiological endpoints, those performed in children or infants and non-English articles were excluded.

Search strategy

A computerized literature search was performed using PubMed® and OVID® Medline for the period 1966-July 2006. The search was restricted to studies on adult human population and was carried out using the search terms: suction or suctioning or endotracheal suctioning or tracheal suctioning or open suctioning or closed suctioning AND randomised or randomized trials or clinical trials or controlled trials . Abstracts of trials generated by electronic search were reviewed, and trials pertaining to open and closed suctioning were retrieved for detailed evaluation. The references of identified articles were reviewed to identify other relevant articles. The Cochrane Central Register of Controlled trials was also searched to identify other trials on this topic. A systematic review was identified at the completion of this study.[15] A second meta-analysis on the same subject was also published following submission of this article for journal review.[16] Personal correspondence with authors was not sought.

Quality assessment

Quality assessment was performed in an unblinded fashion by two investigators using a quality score[17] modified for this study and adapted for ventilated patients. This score assessed composite aspects of study quality (10 aspects in total, with scores 0 or 1; minimum total score 0 and maximum total score 10). Differences in opinion were settled by consensus.

Data abstraction

Two investigators independently abstracted data using standardized data collection forms. The extracted variables were predefined and differences in data abstraction were settled by consensus.

Outcome measures

The primary outcome assessed was number of patients developing VAP. This was defined in various studies as - new onset of purulent bronchial secretions, body temperature of> 38°C or < 35.5°C (definitions of body temperature were variable), white cell count (WCC) of> 10,000 or < 4,000/cmm (cut off values for WCC were variable), chest radiography showing new or progressive infiltrates and a (quantitative in some studies) culture of respiratory secretions (endotracheal or lavage or protected brush catheter) suggesting infection. Secondary outcomes included:

(i) Hospital and ICU length of stay in days
(ii) Mortality defined as deceased when discharged from hospital
(iii) Duration of MV in the two groups

Statistical methods

Preliminary analyses of baseline characteristics of the two groups were performed using inverse variance weighted differences. The effect of the suctioning type on VAP rates and mortality was expressed with 95% CI and ' P ' values as risk difference (RD) and hospital and ICU LOS and duration of ventilation as mean weighted difference (MWD) in days. Treatment effect was assessed using the methods of Mantel Haenszel (fixed-effects); and where heterogeneity of treatment effects was present, the DerSimonian and Laird (random-effects) estimator was used. Alternative estimates of treatment effects were obtained using quality scores as weights, as provided by the METAN program,[18] using StataÒ Release 8 (College Station, Texas, 2003).

Heterogeneity of treatment effects was assessed as (i) the extent, diagnosed by means of the Q statistic, considered significant at P £ 0.1[19],[20] and (ii) the impact (upon the variation of pooled treatment effect) by means of the I ² measure, where an I ² of < 30% indicates mild heterogeneity, 30-50% moderate, and substantially> 50%, severe heterogeneity.[20] Meta-regression analysis[21] was undertaken to assess the (potential) effect of average age; percent male patients; average percent of medical, surgical and trauma patients on the development of VAP and mortality.

Publication bias was not reported, as the ability to adequately detect bias is limited when the number of trials included is less than 10 and when the sample size in the individual studies is small.[22] A cumulative meta-analysis was also performed[23] to study possible time trends in treatment effects.

Results

Preliminary search identified 731 trials on suctioning in the adult population. A single investigator reviewed these abstracts, and 48 articles were identified for detailed evaluation by two authors. Forty-two articles were excluded [Figure - 1]. Six articles fulfilled criteria for inclusion, and a further three articles were identified by hand-search and review of other articles. The study cohort thus consisted of 9 RCTs from 1966-2006 that compared CES with OES.[5],[6],[12],[24],[25],[26],[27],[28],[29] [Table - 1] summarizes the study characteristics as well as the quality scores. A breakdown of the quality scores for the different studies is provided in the Appendix.

The CES group consisted of 644 patients; and the OES group, 648 patients. The two groups were matched [Table - 2] for age; sex; APACHE II score; and the number of medical, surgical or trauma patients (inverse variance weighted differences, P ³ 0.26).

VAP rates

All studies reported VAP rates. VAP developed in 19.2% of patients (248/1,292) and ranged from 0 to 53% in the individual studies [Table - 3]. There was no differential treatment effect [Figure - 2],[Table - 4] of the type of suctioning on VAP rates (fixed effect, RD - 0.01, 95% CI - 0.05, 0.03, P = 0.63, heterogeneity P = 0.2). Heterogeneity was not observed.

Mortality rate

The overall mortality was 22.9% (266/1,164) and reported in only five studies and ranged from 13 to 68% in the individual studies [Table - 3]. The mortality rates of 23.2% (135/581) in the CES group and 22.5% (131/583) in the OES group [Figure - 3],[Table - 4] were similar (fixed effect, RD - 0.01, 95% CI - 0.04, 0.05, P = 0.8, heterogeneity P = 0.9). There was no heterogeneity.

Other outcomes

Hospital LOS was not reported in the studies. The duration of ICU stay was evaluated using the random-effects estimator due to heterogeneity of treatment effects; this was not different in the two studies reporting ICU LOS [Table - 4]. The duration of mechanical ventilation, in four studies, assessed by the fixed-effects estimator was significantly lower in the OES group [Figure - 4]; however, one study[28] contributed substantially (93%) to the estimates.

The cost of suctioning

The cost of suctioning was provided in four studies. Two studies[5],[6] reported higher costs with closed suctioning, one study reported similar cost[24] and in one study cost was marginally higher with open suctioning.[12]

Meta-regression analysis

Meta-regression analysis showed that none of the predefined clinical variables (average age; percent male patients; average percent of medical, surgical and trauma patients) predicted VAP or mortality ( P ³ 0.34). Small numbers precluded multi-variate analysis.

Cumulative meta-analysis

Cumulative meta-analysis [Figure - 5] did not demonstrate any substantial variation in the point estimates for VAP, although there was a retraction in the confidence intervals over time with the addition of the more recently published studies.

The effect of trial quality score, assessed as 'quality-weights,' on treatment outcomes (VAP and mortality) was minimal with no change in the point estimates or level of significance. Similarly, there was no substantial effect of the quality score on the duration of ventilation ( P value remaining significant at 0.001).

Discussion

This meta-analysis has not demonstrated an advantage of CES over OES in the primary outcome (VAP rates). Given that no difference was established between the two methods of endotracheal suctioning (CES and OES) on primary outcome, the expectation would have been a translation of this 'lack of difference' to secondary outcomes. No difference was established with respect to mortality or ICU length of stay; but a statistically significant reduction in the duration of ventilation, favoring OES, was found. However, the number of studies reporting these secondary outcomes varied from two to five [Table - 4], reflecting intra-study publication bias[30]; and the estimates of these outcomes are therefore uncertain. However, if one were to speculate a reason for the observed reduction in the duration of ventilation with OES, more effective clearance of respiratory secretions with OES compared with CES might have effected this difference.

CES is postulated to have several advantages over OES, which include lower gasometrical and hemodynamic impairment,[12] lower risk of contamination of the endotracheal system due to a protective sheath[25] and decreased environmental exposure,[14] as well as ease of use and reduced nursing time.[12] These potential advantages have not been demonstrably translated into improvements in clinically meaningful outcomes (incidence of VAP, mortality and ICU LOS). The disadvantages of CES - higher costs[5],[6] and reduced effectiveness in clearing secretions[31] - may actually favor the use of OES in the routine suctioning of patients in different ICU environments.

Despite the theoretical advantages of CES, failure to demonstrate a clinical benefit may be due to several reasons - a true effect, the meta-analysis being under-powered to detect a difference or a relatively low incidence of VAP in the cohorts included in these studies. The VAP rates of 18.6% (120/644) and 19.8% (128/648) in the CES and OES groups respectively are considerably lower than other studies, where incidences of up to 44% have been recorded.[2] With the use of conventional power calculations, as suggested by Flather and colleagues,[32] we observed that this meta-analysis was actually under-powered (power 63%) to detect a difference between the groups. In order to demonstrate a 5% difference in VAP assuming a baseline VAP rate of 20%, 1,252 patients would need to be evaluated in each treatment arm for 80% power. It is also interesting to note, in the cumulative meta-analysis [Figure - 5], that the effect-line favored CES suctioning, albeit the 95% CI always extended beyond unity.

Although a similar meta-analysis was published at the time of completion of this meta-analysis,[15] as well as a second one during the review process of this article,[16] the current meta-analysis is more comprehensive: the systematic process of trial identification and data abstraction is formally presented; other clinically relevant endpoints (mortality, ICU length of stay and duration of MV) have been canvassed and the potential effect of differential trial quality has also been formally addressed. More importantly, the cumulative meta-analysis, power analysis and the presumed intra-study publication bias suggest that the information required to definitively address these questions is incomplete.

Given the null results of the meta-analyses, the question of applicability of this meta-analysis to all ICU patients needs to be addressed. Only two studies[5],[24] explicitly stated that all patients admitted to the ICU requiring mechanical ventilation were included to the trial. The majority of trials included only a small number of patients (and the reasons for exclusion were not stated), and this may limit generalizability. Further, there was a wide variation in VAP rates as well as mortality in the individual studies included in the meta-analysis [Table - 3]. We also did not specifically look at physiological endpoints, and the question of use of CES in patients on high respiratory supports was not addressed in the current meta-analysis. Several other studies[10],[11],[12] have suggested that there is less desaturation with CES in patients on high respiratory support, in whom the CES may be preferred. The favorable effect of CES on arterial de (saturation) in the smaller studies has not translated to a benefit in the recent meta-analysis.[16] The same meta-analysis also reported significant differences in heart rate (6 beats/min) and mean arterial pressure (3-5 mmHg) in favor of CES.[16] These differences are however unlikely to be clinically significant or meaningful.

Costs are also vital, more so in developing countries such as ours, where any increase in cost without a definitive improvement in clinical endpoints cannot justify the use of CES. Until definitive clinical benefit is demonstrated in further trials, OES may continue to be favored in these nations.

Other issues may impact upon the question of CES versus OES. In developing countries where the space allocated to individual beds may be restricted (picture), close proximity of beds may lead to environmental contamination of the respiratory tract. The high incidence of pulmonary tuberculosis in developing countries such as ours poses greater risk to the health personnel, and the mode of suctioning assumes greater importance. Thus, studies comparing CES with OES may be more relevant in the developing world. These issues may not be paramount in countries where occupational health and safety concerns preclude the use of OES and where the ease of use as well as reduced nursing time with CES may override cost concerns.

We submit therefore that the question of open or closed endotracheal suctioning in mechanically ventilated patients is still very much an 'open' issue.

Acknowledgments

References

Copyright 2007 - Indian Journal of Medical Sciences

The following images related to this document are available:

Photo images