Journal of Critical Care (2012) 27, 647–654

Changes in heart rate, mean arterial pressure, and oxygensaturation after open and closed endotracheal suctioning:A prospective observational study☆,☆☆,★

Irene P. Jongerden RN, PhDa,⁎, Jozef Kesecioglu MD, PhDa, Ben Speelberg MD, PhDb,Anton G. Buiting MD, PhD c, Maurine A. Leverstein-van Hall MD, PhDd,Marc J. Bonten MD, PhDd,e

aDepartment of Intensive Care Medicine, University Medical Center Utrecht, PO Box 85500, 3508 GA Utrecht, The NetherlandsbDepartment of Intensive Care, St Anna Hospital, PO Box 90, 5660 AB Geldrop, The NetherlandscDepartment of Medical Microbiology, St Elisabeth Hospital, PO Box 90151, 5000 LC Tilburg, The NetherlandsdDepartment of Medical Microbiology, University Medical Center Utrecht, PO Box 85500, 3508 GA Utrecht, The NetherlandseJulius Center for Health Sciences and Primary Care, University Medical Center Utrecht, P.O. Box 85500, 3508 GA Utrecht,The Netherlands

0h

Keywords:Endotracheal suctioning;Mechanical ventilation;Intensive care;Complications

AbstractPurpose: It is widely assumed that closed suction systems (CSSs), as compared with open suctionsystems (OSSs), better guarantee optimal oxygenation with less disturbance of physiologic parametersin mechanically ventilated intensive care patients. We, therefore, quantified changes in heart rate (HR),mean arterial pressure (MAP), and peripheral oxygen saturation (SpO2) in patients undergoingendotracheal suctioning (ES) with CSS and OSS.Materials and Methods: We performed a prospective observational study nested within a crossover trialin 4 intensive care units between January 2007 and February 2008. Per unit, 50 ES procedures wereselected at random, and HR, MAP, and SpO2 were measured before and after ES.Results: In total, 197 complete ES procedures (103 OSS and 94 CSS) were monitored. Mean HR, MAP,and SpO2 changed directly after ES and returned to baseline after 5 minutes. Changes in HR and MAPwere comparable after using CSS and OSS, whereas in SpO2, slightly better values were monitored 3and 5 minutes after OSS, these differences being rather small (0.3%-0.7%) and clinically not relevant.Conclusions: Changes in HR, MAP, and SpO2 were comparable and mild during and after CSS andOSS. Both systems can be considered equally safe.© 2012 Elsevier Inc. All rights reserved.

☆ Competing interests The authors declare that they have no competing interests.☆☆ Institution in which work was done: University Medical Center Utrecht and St Elisabeth Hospital Tilburg, The Netherlands.★ Trial registration: ISRCTN75875670.⁎ Corresponding author. Tel.: +31 88 7561140; fax: +31 88 7567760.E-mail address: i.p.jongerden@umcutrecht.nl (I.P. Jongerden).

883-9441/$ – see front matter © 2012 Elsevier Inc. All rights reserved.ttp://dx.doi.org/10.1016/j.jcrc.2012.02.016

648 I.P. Jongerden et al.

1. Introduction

Endotracheal suctioning (ES) is an essential and fre-quently performed procedure in mechanically ventilatedpatients in the intensive care unit (ICU). By ES, secretionsare cleared from the tracheobronchial tree, guaranteeingoptimal oxygenation and avoiding accumulation of secre-tions, tube occlusion, increased work of breathing, atelecta-sis, and pulmonary infections [1,2]. Yet, ES may also haveadverse effects, such as disturbances in cardiac rhythm andhypoxemia [3].

Nowadays, 2 systems are available to perform ES: thesingle-use open suction system (OSS) and the multiple-useclosed suction system (CSS). The latter does not requiredisconnection from the ventilator and can remain connectedto the patient for at least 24 hours, depending on hospitalprotocol and CSS type. In contrast, the OSS requiresdisconnection from the mechanical ventilator, either bycomplete disconnection or through opening a valve of aswivel (semiopen). It is suggested that, by interruption, thepatient is predisposed to physiologic disturbances because ofa decay of intrathoracic pressure, such as hypoxemia, alteredmean arterial pressure (MAP), and alterations in heart rate(HR) (brachycardia, arrhythmias) [4-6]. Several studiescompared both systems with regard to physiologic distur-bances, mostly with both procedures performed in a singlepatient [1,6-11]. Results from most studies favored CSS,although differences between ES systems were rather smalland clinically not relevant [12]. Furthermore, results aredifficult to generalize because of small sample sizes (9-35patients) and differences in performance of ES (duration, useof preoxygenation and hyperoxygenation). It remained,therefore, inconclusive whether 1 system should be preferredover the other, and we aimed to determine whether CSS, ascompared with OSS with or without use of a swivelconnector, changes cardiorespiratory parameters after ES inmechanically ventilated ICU patients.

2. Materials and methods

Between January 2007 and February 2008, a prospectiveobservational study was performed with a focus onunintended side effects (physiologic disturbances) duringand after ES with OSS and CSS. The study was nested in alarger crossover trial in which, during periods of 6 months,either CSS or OSS was used for all mechanically ventilatedpatients to determine occurrence of cross-transmission withGram-negative bacteria in both systems (described else-where) [13]. Four ICUs participated in the trial: 2 ICUs froma university hospital with 10 beds (4 single rooms and 6 onthe ward) and 8 beds (1 single room and 7 on the ward) andtwo 8-bed units from a teaching hospital (all single rooms).

Because inclusion was planned within the larger crossovertrial in which use of CSS or OSS was dictated by study

protocol, it was not possible to randomize individual patients toeither of both systems. Therefore, per study period and per unit,25 bed numbers were selected at random by using a computerprogram (Research Randomizer, Urbaniak & Plous, SocialPsychology Network, Wesleyan University, Middletown CT,USA), accumulating to a total of 200 observations of ESprocedures. Based on previous studies, with 100 patients ineach study arm, we expected to demonstrate a difference of8 mm Hg on MAP, 11 beats per minute on HR, and 3% onperipheral oxygen saturation (SpO2), with α = .05 and β = .80.

Three times a week, from Monday to Thursday, researchnurses checked whether the indicated bed number wasoccupied and whether the patient was on ventilation. If not,the “neighbor” bed was selected.

Both CSS and OSS were not considered experimentaltreatments (because they both are frequently used), andtherefore, the institutional review board of both hospitalswaived the requirement for ethical approval and informedconsent. The study was designed in line with the Strength-ening the Reporting of Observational Studies in Epidemiol-ogy statement [14].

All ES procedures were performed on indication by ICUnurses and according to protocol, according to which OSSwasperformed through a rubber-sealed swivel connector placedbetween the tube and the Y-piece of the ventilator circuitry(hospital 1) or by disconnection (hospital 2). The catheter wasinserted into the endotracheal tube or tracheostomy untilresistance was met and withdrawn 0.5 cm. A negative pressureof maximum 20 (hospital 2) or 30 kPa (hospital 1) was set, andthe catheter was withdrawn while gently rotating. The swivelconnector was closed again, and the procedure lasted for a totalof 10 to 15 seconds. For CSS, the procedure was similar exceptthat the patient remained connected to the ventilator. Closedsuction system was replaced every 24 hours (Ballard TrachCare Double Swivel Elbow, Kimberly Clark, Draper, UT,USA; hospital 1) or every 72 hours (Ballard Trach Care 72Hour, Kimberly Clark, Draper, UT, USA; hospital 2).

Preoxygenation and postoxygenation with 100% oxy-gen were applied in both procedures when consideredindicated (ie, in patients who were respiratory unstable), asjudged by attending nurses, as was the use of protectivemasks and glasses. Nonsterile gloves were to be usedduring all procedures.

2.1. Outcome

The primary outcome was the change in cardiorespira-tory parameters, and we used measurements that werecontinuously available during routine monitoring: HR,MAP, and SpO2.

The 3 parameters were monitored before ES (baseline),immediately after ES, and subsequently 3 and 5 minutes afterES. Data were registered by research nurses as recorded bythe bedside monitor (Spacelabs Monitor 90387; SpacelabsHealthcare, Issaquah, Wash, and Philips HP Merlin; Philips

Table 1 Baseline characteristics

CSS OSS

Patients observed, n 80 85Patient characteristicsSex (% male) 73 71Age (y), a mean (SD) 59 (16.5) 61 (16.7)APACHE II score,mean (SD)

20 (6.1) 20 (6.8)

Diagnosis (% surgical) 38 35Primary diagnosis, n (%)Pulmonary/respiratory 21 (26) 27 (32)Trauma 11 (14) 18 (21)Gastrointestinal 14 (18) 13 (15)Cardiovascular 12 (15) 5 (6)Neurosurgical 6 (8) 10 (12)Neurologic 8 (10) 7 (8)Sepsis 6 (8) 2 (2)Renal 2 (3) 1 (1)Thorax surgical 0 (0) 2 (2)Total duration MV,median (IQR)

19 (9-29) 17 (10-31)

Length of ICU stay,median (IQR)

17.5 (9.3-28.0) 18.0 (10.0-31.5)

ICU mortality (%) 30 20Hospital mortality (%) 35 27No. of observations 95 105Characteristics ofobservationsPEEP, median cmH2O (IQR)

8 (5-10) 5 (5-6)

Preoxygenation(% of observations)

21 27

Postoxygenation(% of observations)

13 22

649Changes in HR, MAP, and oxygen saturation after OSS and CSS

Healthcare, Eindhoven, The Netherlands). Peripheral oxygensaturation was measured continuously by pulse goniometry(DS-100A; Spacelabs, and M1020A; Philips). The electro-cardiographic tracing was continuously monitoring HR;MAP was measured with an indwelling arterial catheter or, ifsuch a catheter had not been inserted, a noninvasive bloodpressure cuff (Spacelabs, and M1008A; Philips). Further-more, clinical data were collected through a registration format the time of monitoring: admission date to ICU, age,positive end-expiratory pressure (PEEP), ES system, use ofpreoxygenation and postoxygenation, and whether thepatient was disconnected during CSS.

2.2. Statistical analysis

For univariate analysis, continuous variables were testedwith Kolmogorov-Smirnov tests for normal distribution;t tests were used when data were normally distributed,otherwise, nonparametric Mann-Whitney U tests. Dichoto-mous variables were analyzed by using χ2 tests.

Changes in SpO2, HR, and MAP for each suction systemwere evaluated by using a repeated-measures analysis ofvariance (ANOVA) including 4 levels of time (before ES,directly after ES, and 3 and 5 minutes after ES) as within-subject variables and ES system (CSS or OSS) as between-subject factor. In addition, we added preoxygenation (yes/no),ventilation route (tube/tracheostoma), and PEEP (continuous)as covariates in the analysis and calculated estimated marginalmeans, representing mean value for each level of time,adjusted for these covariates. Because of differences inperformance and materials used for ES between hospitals,analyses were stratified according to center.

Imputation was used for baseline missing data using anexpectation-maximization analysis with the Impute functionin SPSS software (version 15.0; SPSS, Chicago, Ill), withinclusion of study period, age, sex, diagnosis, and mechani-cal ventilation (MV) as key variables in the imputationmodel. A total of 4 Acute Physiology and Chronic HealthEvaluation II scores (2%) were missing. Expectation-maximization analysis revealed that data were missing atrandom, meaning that differences in missing data are relatedto the observed data and that missing values were replacedby imputed values. Apart from increasing the sample size,imputation corrects for possible bias because of selectivemissing values [15,16].

All statistical analyses were performed with the StatisticalPackage for the Social Sciences version 15.0. P b .05 wasconsidered statistically significant.

Duration MV beforeobservation, mediandays (IQR)

6 (3-13.3) 5 (3-14)

Tube/tracheotomy(% tube)

93.7 72.4

APACHE, Acute Physiology and Chronic Health Evaluation; IQR,interquartile range.

a Age at time of ICU admission.

3. Results

During the 12-month study period, 200 ES procedureswere monitored in 165 patients (118 men; mean age, 60 ± 17years). In 29 patients, more than 1 procedure was observed,

all on different days. There were no significant differences inbaseline characteristics of the observed patients in whom ESwas performed with CSS or OSS (Table 1).

In 38 (19%) of 200 observations, the randomized bedwas notoccupied or the patient was not on MV at that moment, and thepatient in the neighbor bedwas selected for observation (Fig. 1).

Although OSS and CSS were dictated during 6-monthstudy periods, nonadherence to the prescribed systemoccurred in 7% of the MV days, most often because ofprone positioning (reason for using CSS during OSS) and ofweaning (reason for using OSS during CSS) [13]. As a result,3 CSS procedures were observed during OSS period, and8 OSS procedures, during CSS, counting up to a total of 95bedside observations during CSS and 105 during OSS

Fig. 1 Flowchart. ⁎Nonadherence to the prescribed ES system occurred in 7% of mechanical ventilation days.

650 I.P. Jongerden et al.

procedures. Because of random selection, more patients inwhom ES was performed with OSS had a tracheostomy, andPEEP levels were lower as compared with CSS (Table 1).Besides, the baseline difference in PEEP was also due to thepreference of some physicians to use CSS in patients withPEEP values of 10 cm H2O or higher. In 3 patients (1 CSSand 2 OSS), observational data were not complete becausethe patient was transported (ie, to computed tomographicscan) before the last observation had been performed.Furthermore, in 2 patients (1 CSS and 1 OSS), pulseoxymetry was not connected correctly, and these measure-ments were excluded from analysis.

3.1. Changes in HR

There were 103 complete measurements of OSS and 94for CSS. There were differences in mean HR before and afterES (P b .01), with largest differences directly after ES, as

Table 2 Heart rate, MAP, and SpO2 before and after ES

Before ES After ES

CSS OSS CSS OSS

HR 89.9 (17.5) 89.3 (18.0) 93.6 (18.3) 93.3 (18.1)MAP 85.5 (16.6) 87.8 (18.8) 88.8 (15.6) 92.1 (19.7)SpO2 97.2 (2.7) 97.6 (2.8) 97.7 (2.3) 97.6 (3.0)

P value was calculated with repeated-measures ANOVA, test of within-subject

compared with baseline (Table 2). Repeated-measuresanalysis revealed that differences in time were not relatedto PEEP level, preoxygenation, or ventilation route.Furthermore, there was no significant difference betweenCSS and OSS groups (P = .44) nor between hospitals (P =.40) (Fig. 2A and B). Restricting the repeated-measuresANOVA to patients who were orally intubated (162observations) resulted in comparable changes within time(P b .05) and no difference between CSS and OSS (P = .52).

3.2. Changes in MAP

For this analysis, there were 102 and 93 completeobservations for OSS and CSS, respectively. Mean MAPwas higher directly after ES as compared with other timepoints (P b .05) (Table 2), this difference being related toPEEP level (P b .05). There was no significant differencebetween CSS and OSS (P = .61). However, there was a

After 3 min After 5 min P

CSS OSS CSS OSS

90.6 (17.4) 89.8 (17.1) 89.7 (16.4) 89.3 (17.0) .9786.4 (15.5) 89.2 (19.5) 85.8 (15.5) 87.0 (17.2) .3497.6 (2.5) 97.9 (2.2) 97.5 (2.6) 98.2 (2.0) .04

effects for interaction effect. Values are presented as mean (SD).

96

94

92

Est

imat

ed M

arg

inal

Mea

ns

90

88

1 2 3 4Time

96

System

94

92

Est

imat

ed M

arg

inal

Mea

ns

90

88

1 2 3 4

OSSCSS

Time

H1 H2

Fig. 2 Mean HR before, during, and after ES. Estimated marginal means = means for HR adjusted for PEEP, preoxygenation, andventilation route.

651Changes in HR, MAP, and oxygen saturation after OSS and CSS

difference between hospitals, with slightly higher meanMAP values in hospital 2 as compared with hospital 1 on alltime points (P b .05) (Fig. 3A and B). Restricting analysesto patients who were orally intubated (160 observations)revealed no significant difference in mean MAP in time(P = .07) and no differences between CSS and OSS (P = .80).

3.3. Changes in SpO2

For this analysis, there were 102 and 93 competemeasurements for SpO2 after OSS and CSS, respectively.There were small differences in mean SpO2 after 3 and 5minutes, as compared with baseline (P b .05) (Table 2).Differences in time were related to ES system (P b .05),ventilation route (P b .05), and preoxygenation (P b .001).Positive end-expiratory pressure level was correlated tosaturation after 3 and 5 minutes, with higher PEEP levelsbeing associated with slightly lower saturation after 3minutes (P b .01) and after 5 minutes (P b .05). However,associations were very weak (R2 = 0.05 and R2 = 0.03 for 3and 5 minutes, respectively), and differences in mean SpO2 intime were not related to PEEP level (within-group analysis,

94

92

Est

imat

ed M

arg

inal

Mea

ns

90

88

1 2 3 4Time

Est

imat

ed M

arg

inal

Mea

ns

H1

86

84

82

94

92

90

88

86

84

82

Fig. 3 Mean arterial pressure (mean) before, during, and after ES. Epreoxygenation, and ventilation route.

P = .37). There was a small interaction effect between ESsystem and the covariates (P b .05; effect size, .03), andpreoxygenation was associated with higher SpO2 values afterOSS, as compared with CSS (P b .05). For the othercovariates, no significant differences were found. Further-more, there was no main effect for ES system (between-group analysis, P = .48). Stratifying centers revealed asignificant difference between hospitals, with slightly highervalues of mean SpO2 in hospital 1 as compared with hospital2 on all time points (P b .001) (Fig. 4A and B). In patientswho were orally intubated (160 observations), there weresmall differences in mean SpO2 values in time (P b .05), andthese differences were related to ES system (P b .05) andpreoxygenation (P b .001). However, there was nosignificant difference between CSS and OSS (P = .76).

4. Discussion

The main feature of this study is that changes in HR,MAP, and SpO2 induced by ES are minor and comparablewhen using either CSS or OSS.

System

1 2 3 4

OSSCSS

Time

H2

stimated marginal means = means for MAP adjusted for PEEP,

Est

imat

ed M

arg

inal

Mea

ns

1 2 3 4Time

H1 H2

100100

99

98

97

96

95

Est

imat

ed M

arg

inal

Mea

ns

1 2 3 4Time

SystemOSSCSS

99

98

97

96

95

Fig. 4 Mean SpO2 before, during, and after ES. Estimated marginal means = means for SpO2 adjusted for PEEP, preoxygenation, andventilation route.

652 I.P. Jongerden et al.

We performed a pragmatic study, evaluating ES as it wasperformed during standard care and when clinicallyindicated according to international guidelines [17]. Mostprocedures only lasted 10 to 15 seconds and were notassociated with clinically relevant disturbances in physio-logic parameters.

Although it has been suggested to apply preoxygenationbefore ES to minimize desaturation, especially in hypoxemicpatients [17], it was applied in only 24% of ES procedures inthe current study, and although preoxygenation wasassociated with higher SpO2 values after OSS, as comparedwith CSS, differences were rather small (97.5% after CSSand 98.3% after OSS) and, therefore, not consideredclinically relevant. Similarly small changes in SpO2 afterpreoxygenation have been reported before [18,19].

The mean HR in the current study (89 and 90 beats perminute before ES in OSS and CSS, respectively) was slightlylower than that reported in other studies (range, 91-108 beatsper minute before ES) [1,5,6,9,11,18]. Yet, the mean HRonly increased—on average—with 4 beats per minute afterES, with no difference between both suction systems, whichconfirms the findings of an earlier meta-analysis (6 beats perminute in favor of CSS) [12].

Mean MAP values as found in our study before ES (86-88mm Hg in CSS and OSS, respectively) were in the range ofvalues reported in other studies (74-93 mm Hg)[1,6,11,18,20]. In the current study, changes in MAP werenot related to the suction system, which contrasts the findingsof a recent meta-analysis in which the mean MAP decreasedafter using CSS (pooled standardized mean difference, −0.43mm Hg) as compared with OSS [12].

A decrease in SpO2 is assumed common after OSS andrare after CSS [21]. Mean values for SpO2 before ES in ourstudy were high (97% in CSS and 98% in OSS) andcomparable with those reported in other studies (rangingfrom 95% to 98%) [1,6,9-11]. However, where previousstudies favored use of CSS with slightly higher SpO2 values,only small differences were found both in adults as inneonates [1,22-25]. In addition, a study in spontaneous-

breathing patients resulted in a maximum median deviationof 1% from baseline after CSS [21].

In 1 hospital, the negative pressure for ES was slightlyhigher (30 kPa) than recommended (20 kPa) [17]. The lattervalue has been recommended to prevent side effects whilestill effectively removing secretions, but a clear-cutoptimum value has not been defined yet. No differences inside effects were observed with a negative pressure of −400cm H2O as compared with −200 cm H2O (approximately40-20 kPa) in a single small-sized (n = 9) study [8]. In ourstudy, there was no evidence of any side effects of thenegative pressure as used.

Closed suction system has been advocated as a techniquethat may limit cardiorespiratory instability because ofpressure loss during disconnection [6,26]. However, CSSalso interferes with intratracheal pressure [21,27]. It has alsobeen suggested that catheter size is of greater influence thansuction method in this regard [28]. Yet, both aspects have notbeen addressed in our study.

Our study has several limitations. First, we did notrandomize patients to either intervention because the studywas nested within a cross-over trial. All patients received ESwith a predefined suction system according to a studyprotocol. To avoid selection bias, we have used a predefinedschedule to include patients for this substudy, whichincluded the selection of another patient if the selectedpatient was not ventilated. There were no differences inpatient characteristics between study groups; however, therewere differences in observations, with higher PEEP levels inthe CSS group and more patients orally intubated ascompared with the OSS group.

A second limitation is the preference of some physiciansto use CSS in patients with PEEP values of 10 cm H2O orhigher. Our study protocol did not dictate this practice,although it has been advised for adult patients needing highvalues of PEEP [17]. The preference of some physicianscontributed to the baseline difference in PEEP. IncludingPEEP as a covariate in repeated-measures analysis did notreveal differences between CSS and OSS.

653Changes in HR, MAP, and oxygen saturation after OSS and CSS

We also did not record secretion volumes beingremoved after using CSS or OSS. The latter was moreeffective in removing secretions than CSS in animalmodels and in vitro [29,30]. If true in vivo, we wouldhave expected more procedures of ES with CSS, whichwas not observed (both mean and median, 6 times perday) [13].

In this pragmatic study, CSS and OSS were performedwhen indicated and according to hospital protocol. Allpatients on MV requiring ES were eligible, and with 200 ESprocedures monitored, we aimed to determine a small butdistinct difference between CSS and OSS (of 8, 11, and 3units for HR, MAP, and SpO2, respectively). However, inboth groups, baseline values were within reference ranges,and values returned to baseline within 5 minutes withoutlarge differences.

Because both systems can be considered equally safe, thechoice of systems to be used will depend on user friendliness,personal preference, and costs. As in most places in theworld, OSS was less expensive than CSS in our setting (priceof OSS, €0.38 per catheter and €2.70 per swivel connecter;price per CSS, €11,20; price level in the Netherlands, 2009)[3,12]. Therefore, performing ES with OSS, as comparedwith CSS, might save money without inducing changes inHR, MAP, and SpO2.

In conclusion, the results of our study do not support theassumption that CSS, as compared with OSS, differentlyaffects HR, MAP, and SpO2 during and after ES. Bothsystems can be considered equally safe in mechanicallyventilated ICU patients.

Acknowledgments

The authors thank the following persons for theirassistance in monitoring ES procedures and collecting data:Fieke Kloosterman, Piet Vos, Twan Verhoeven, and Fritsvan Beers. The authors received a grant from the NetherlandsOrganisation for health research and development (ZonMwproject number 62300037); MB has received researchfunding from the Netherlands Organization of ScientificResearch (NWO Vici 918.76.611).

References

[1] Johnson KL, Kearney PA, Johnson SB, et al. Closed versus openendotracheal suctioning: costs and physiologic consequences. CritCare Med 1994;22:658-66.

[2] Zeitoun SS, Botura Leite de Barros AB, Diccini S. A prospective,randomized study of ventilator-associated pneumonia in patientsusing a closed vs. open suction system. J Clin Nurs 2003;12:484-9.

[3] Subirana M, Sola I, Benito S. Closed tracheal suction systems versusopen tracheal suction systems for mechanically ventilated adultpatients. Cochrane Database Syst Rev 2007:CD004581.

[4] Gunderson LP, Stone KS, Hamlin RL. Endotracheal suctioning–induced heart-rate alterations. Nurs Res 1991;40:139-43.

[5] Fernandez MD, Piacentini E, Blanch L, et al. Changes in lung volumewith three systems of endotracheal suctioning with and without pre-oxygenation in patients with mild-to-moderate lung failure. IntensiveCare Med 2004;30:2210-5.

[6] Cereda M, Villa F, Colombo E, et al. Closed system endotrachealsuctioning maintains lung volume during volume-controlled mechanicalventilation. Intensive Care Med 2001;27:648-54.

[7] Bourgault AM, Brown CA, Hains SMJ, et al. Effects of endotrachealtube suctioning on arterial oxygen tension and heart rate variability.Biol Res Nurs 2006;7:268-78.

[8] Lasocki S, Lu Q, Sartorius A, et al. Open and closed-circuitendotracheal suctioning in acute lung injury—efficiency and effectson gas exchange. Anesthesiology 2006;104:39-47.

[9] Valderas Castilla D, Bravo Paramo C, Torres Gonzales JI, et al.Repercussion on respiratory and hemodynamic parameters with aclosed system of aspiration of secretion. Enferm Intensiva 2004;15:3-10.

[10] Rabitsch W, Kostler WJ, Fiebiger W, et al. Closed suctioning systemreduces cross-contamination between bronchial system and gastricjuices. Anesth Analg 2004;99:886-92.

[11] Lee CK, Ng KS, Tan SG, et al. Effect of different endotrachealsuctioning systems on cardiorespiratory parameters of ventilatedpatients. Ann Acad Med Singapore 2001;30:239-44.

[12] Jongerden IP, Rovers MM, Grypdonck MH, et al. Open andclosed endotracheal suction systems in mechanically ventilatedintensive care patients: a meta-analysis. Crit Care Med 2007;35:260-70.

[13] Jongerden IP, Buiting AG, Leverstein-van Hall MA, et al. Effect ofopen and closed endotracheal suctioning on cross-transmission withGram-negative bacteria: a prospective crossover study. Crit CareMed 2011;39:1313-21.

[14] von EE, Altman DG, Egger M, et al. The Strengthening the Reportingof Observational Studies in Epidemiology (STROBE) statement:guidelines for reporting observational studies. Ann Intern Med2007;147:573-7.

[15] Sterne JA, White IR, Carlin JB, et al. Multiple imputation for missingdata in epidemiological and clinical research: potential and pitfalls.BMJ 2009;338:b2393.

[16] Little RJA, Rubin DB. Statistical analysis with missing data. NewYork 2002.

[17] AARC Clinical Practice Guidelines. Endotracheal suctioning ofmechanically ventilated patients with artificial airways 2010. RespirCare 2010;55:758-64.

[18] Demir F, Dramali A. Requirement for 100% oxygen before and afterclosed suction. J Adv Nurs 2005;51:245-51.

[19] Oh H, SeoW. A meta-analysis of the effects of various interventions inpreventing endotracheal suction-induced hypoxemia. J Clin Nurs2003;12:912-24.

[20] Cardenosa Cendrero JA, Sole-Violan J, Bordes BA, et al. Role ofdifferent routes of tracheal colonization in the development ofpneumonia in patients receiving mechanical ventilation. Chest1999;116:462-70.

[21] Seymour CW, Cross BJ, Cooke CR, et al. Physiologic impact ofclosed-system endotracheal suctioning in spontaneously breathingpatients receiving mechanical ventilation. Respir Care 2009;54:367-74.

[22] Hoellering AB, Copnell B, Dargaville PA, et al. Lung volume andcardiorespiratory changes during open and closed endotracheal suctionin ventilated newborn infants. Arch Dis Child Fetal Neonatal Ed2008;93:F436-41.

[23] Paul-Allen J, Ostrow CL. Survey of nursing practices with closed-system suctioning. Am J Crit Care 2002;9:9-19.

[24] Ackerman MH, Mick DJ. Instillation of normal saline beforesuctioning in patients with pulmonary infections: a prospectiverandomized controlled trial. Am J Crit Care 1998;7:261-6.

654 I.P. Jongerden et al.

[25] Lee ES, Kim SH, Kim JS. Effects of a closed endotrachealsuction system on oxygen saturation, ventilator-associatedpneumonia, and nursing efficacy. Taehan Kanho Hakhoe Chi2004;34:1315-25.

[26] Maggiore SM, Iacobone E, Zito G, et al. Closed versusopen suctioning techniques. Minerva Anestesiol 2002;68:360-4.

[27] Kiraly NJ, Tingay DG, Mills JF, et al. Volume not guaranteed: closedendotracheal suction compromises ventilation in volume-targetedmode. Neonatology 2010;99:78-82.

[28] Tingay DG, Copnell B, Grant CA, et al. The effect of endotrachealsuction on regional tidal ventilation and end-expiratory lung volume.Intensive Care Med 2010;36:888-96.

[29] Copnell B, Tingay DG, Kiraly NJ, et al. A comparison of theeffectiveness of open and closed endotracheal suction. Intensive CareMed 2007;33:1655-62.

[30] Lindgren S, Odenstedt H, Olegard C, et al. Regional lung derecruit-ment after endotracheal suction during volume- or pressure-controlledventilation: a study using electric impedance tomography. IntensiveCare Med 2007;33:172-80.