A Bedside Decision Tree for Use of Saline With Endotracheal Tube Suctioning in Children

This prospective observational study explores the use of a bedside decision tree in endotracheally intubated infants and children admitted to the Just For Kids Critical Care Center (JFKCCC) at Kosair Children’s Hospital in Louisville, Kentucky. The JFKCCC is a 26-bed intensive care unit located within a free-standing tertiary care pediatric hospital. Patients met inclusion criteria if they were no more than 18 years of age, admitted to the PICU, and required mechanical ventilation through an endotracheal tube for at least 48 hours.

Demographic information obtained for each patient included age, sex, endotracheal tube size, the presence of a cuffed endotracheal tube, duration of mechanical ventilation, length of PICU stay, Pediatric Index of Mortality 2 (PIM2) score,¹¹  primary reason for intubation (hemodynamic instability, trauma, respiratory disease, neurological disease, or out-of-hospital cardiac arrest), location of intubation (PICU, other unit within the hospital, or outlying hospital), and whether the patient had a pulmonary comorbid condition such as asthma or chronic lung disease.

Upon study entry, the bedside caregiver continued to suction the endotracheal tube per unit protocol and used the saline decision tree to determine if saline was warranted for each suctioning pass. In our unit, nurses have the ability to suction outside of the decision tree in extenuating circumstances such as suspected occlusion of the endotracheal tube or as a rescue maneuver in the context of acute hypoxemia or hypercarbia. Impending occlusion of the endotracheal tube was suspected if there was an abrupt decrease in breath sounds, a change in the end-tidal carbon dioxide waveform, a loss of inspired tidal volume, or an increase in peak pressure on the ventilator.

If criteria were met for suctioning with saline (Figure 1), then generally 1 to 2 mL of normal saline, from a prepackaged vial, was instilled into the endotracheal tube and suctioning was performed. The suctioning protocol includes hyperoxygenation at a minimum of 10% above the patient’s baseline unless contraindicated, suctioning pressures of 80 to 120 mm Hg depending on the patient’s age, and brief (10–15 second) shallow suctioning passes. The standard method of suctioning in our PICU is a closed technique that allows the patient to remain connected to the ventilator while an in-line suctioning catheter (Kimberly-Clark Ballard) is passed through the endotracheal tube. Saline can then be instilled via a side port if warranted. An open technique is occasionally used. This involves disconnecting the patient from the ventilator, manually ventilating the patient with either an anesthesia or bag valve mask attached to the endotracheal tube, and then suctioning the endotracheal tube with a separate sterile catheter. Saline is instilled directly into the endotracheal tube if necessary.

Following each suctioning pass that occurred after the patient had been intubated for 48 hours, bedside caregivers recorded the time, date, and primary reason for saline instillation: either extrapolated through the “high risk for plugging” branch point of the decision tree (subjectively thick secretions or the inability to remove secretions without saline), physician request, or suspected occlusion of the endotracheal tube. Use of saline and any adverse events that occurred during or within a 10-minute period immediately following the suctioning episode were recorded. Four types of adverse events were captured: hemodynamic instability, bronchospasm requiring unexpected bronchodilator use, oxygen desaturation defined as oxygen saturation shown by pulse oximetry (Spo₂) less than or equal to 90% (or 5% below patient’s baseline), and failure of Spo₂ to return to baseline within the 10-minute window after suctioning. Hemodynamic instability was determined at the nursing level. If the patient had changes in hemodynamic status related to the suctioning episode that prompted the nurse to engage the physician at the bedside, an adverse event was recorded. The study period concluded when the patient was extubated. Patients reentered the study (as a new participant) if they were extubated for more than 1 hour before reintubation and remained intubated for at least 48 additional hours.

Complications associated with mechanical ventilation were noted for each patient and included VAP and ventilator-associated lower respiratory tract infection as defined by the Centers for Disease Control and Prevention,¹²  endotracheal tube occlusion defined as suspected obstruction requiring removal of the tube, and unplanned extubation as tracked by research nurses for all intubated patients.

This study was reviewed and approved by the institutional review board at the University of Louisville. It was approved as an exempt quality improvement project that did not require informed consent for enrollment.

Descriptive analysis of the analytic sample was performed for the individual patients (n = 69) and at the level of suctioning episode (n = 1986). Means and standard deviations were calculated for continuous variables, whereas frequencies and proportions were calculated for categorical data. Spearman correlation coefficients were calculated to explore relationships between continuous variables. Chi-squared tests were used to test associations between categorical variables. Mann-Whitney tests were used to test associations between categorical and continuous variables.

Two-level logistic regression modeling was used to further explore the association of use of saline for suctioning episodes with adverse events to account for potential nonindependence of the suctioning episodes at the patient level. This analysis was conducted by using the Complex Samples module of SPSS version 22 (IBM SPSS, IBM Corp), where the variable patient was set as the unit of clustering.

During the study period, 98 patient ventilation episodes met inclusion criteria; 74 episodes (67 patients) were enrolled in the study and 69 episodes (62 patients) had at least 1 suctioning pass after a minimum of 48 hours of mechanical ventilation (Figure 2). Five patients were enrolled more than once (3 patients twice; 2 patients 3 times). Table 1 lists the demographic information for study participants. The mean patient age was 4.2 (SD, 5.2) years. For all ventilator patient episodes, 74% involved intubation with a cuffed endotracheal tube, more than half (51%) were intubated outside of the PICU, hemodynamic instability was the most common indication for intubation (42%), and approximately one-third of patients had underlying lung disease.

Mean length of stay in the PICU was 16 (SD, 10) days with a mean of 7.5 (SD, 5.9) days of mechanical ventilation. Mean probability of death based on PIM2 severity of illness scores (probability of death = exp(PIM2 score)/[1+exp(PIM2 score)])¹¹  was 19% (SD, 27%).

In 1986, a total of 1990 documented suctioning episodes during this study had complete bedside caregiver documentation. The rate of completed documentation did not differ between suctioning episodes that used saline (1003, 50.5%) and those that did not (983, 49.5%). The mean number of suctioning episodes per enrollment was 29 suctioning passes, with approximately half performed with saline (Table 1). Indications given for saline instillation were as follows: thick secretions (87%), followed by the inability to remove secretions without saline (6%), suspected occlusion of the endotracheal tube (2%), and physician request (<2%). Data on the indication for saline use were missing for 3%. A total of 53 suctioning episodes (3%) were performed as a rescue maneuver in the patient, with saline used 34 times (64%) for this indication. An adverse event occurred with 26% of suctioning passes. Desaturation was the most common adverse event (18%), followed by bronchospasm (11%) and hemodynamic instability (0.5%; Table 1).

One or more adverse events (hemodynamic instability, bronchospasm, oxygen desaturation, and failure of Spo₂ to return to baseline) were noted in 586 suctioning passes and were more frequent in the saline group than in the dry suctioning group (39.3% vs 19.5%, P < .001; Table 2). Hemodynamic instability (1.2% vs 0.3% of episodes, P = .04), bronchospasm (17.6% vs 7.1% of episodes, P < .001) and oxygen desaturation (26% vs 13.5% of episodes, P < .001) were more frequent when saline was used than when it was not. Failure of oxygen saturation recovery within 10 minutes after suctioning was higher in the saline group, but the difference was not statistically significant (P = .27, Table 2).

Patient factors associated with adverse events during suctioning are shown in Table 3. Patient age was lower among suctioning episodes with an adverse event (mean of 2.4 vs 3.5 years, P < .001). The presence of a cuffed endotracheal tube (P = .001) and an endotracheal tube diameter of 4.0 mm or less (P < .001) were both more common among suctioning episodes with an adverse event. The primary indication for intubation also was associated with adverse events during suctioning. For example, if a patient was intubated for respiratory illness, an adverse event occurred during approximately 30% of suctioning episodes as compared with a patient intubated for trauma—only 2% of suctioning passes had an adverse event noted (P < .001). Underlying respiratory disease was present more often among suctioning episodes with an adverse event (42% vs 33%, P < .001). Patients with suctioning episodes associated with an adverse event had a longer duration of mechanical ventilation (mean of 14.7 vs 12.4 days, P < .001).

The frequency of adverse events also varied by indication for saline use within this group of suctioning events (Table 3). Saline use for presence of thick secretions in the endotracheal tube (the most common indication) was somewhat underrepresented (83% vs 90%) and the indication no results without saline was more frequent (8.6% vs 4.4%) in the adverse event group (P < .001).

Two-level logistic regression modeling was performed to account for clustering of data at the ventilation episode level (69 episodes in 62 patients) to further evaluate whether saline instillation was independently associated with adverse events (Table 4). Three factors remained associated with adverse events: saline instillation (odds ratio, 2.78; 95% CI, 1.79–4.32; P < .001), an endotracheal tube size of 4.0 mm or less (odds ratio, 2.90; 95% CI, 1.13–7.45; P = .03), and the presence of a comorbid respiratory disease (odds ratio, 1.78; 95% CI, 1.06–2.99; P = .03). Sex, age, the presence of a cuffed endotracheal tube, reason for intubation, and duration of mechanical ventilation did not remain associated with adverse events. Because thick secretions were the indication for use of saline in 87% of the suctioning events in this group, saline indication subsets were not included in the multivariable analysis.

We found a concerning frequency of adverse events associated with all suctioning episodes as well as with saline suctioning events in our sample of pediatric patients. We believe ours is the first pediatric study to demonstrate a significant increase in the risk of clinically significant hemodynamic deterioration requiring physician intervention after a saline suctioning pass. This study is also the first to report a significant increase in the unexpected use of bronchodilators when saline is used, suggesting that saline may be an airway irritant causing bronchospasm immediately following a suctioning pass. Our study also identified potential risk factors for suctioning-related adverse events including saline instillation, endotracheal tube diameter of 4.0 mm or less, and the presence of a comorbid respiratory disease.

This study is one of the largest to date in children, and we believe it is the first to quantify the number of suctioning passes per day along with the frequency of saline administration in intubated patients when a bedside decision tree is used. As expected, suctioning of the endotracheal tube continues to be a significant component of patient care within our PICU, with approximately 5 suctioning passes per patient per day that an endotracheal tube is in place. Furthermore, saline is used in approximately 50% of all suctioning passes, with thick secretions being the most common indication for the use of suctioning.

In the 2010 clinical practice guidelines of the American Association for Respiratory Care,²  the authors warn of hazards associated with endotracheal tube suctioning, including changes in lung compliance, capacity, and volume; hypoxia; trauma; bronchospasm; and hemodynamic instability. Our study confirms that endotracheal tube suctioning is indeed associated with risk for deterioration in the patient’s condition. On average, a patient experienced an adverse event with 26% of suctioning passes whether or not saline was instilled. Although not all possible complications were evaluated, we did note that oxygen desaturation was the most likely adverse event, followed by bronchospasm and hemodynamic deterioration. Although most of these events were limited, they could prove catastrophic in patients intubated for hemodynamic instability where equilibrium is already tenuous.

Our study further demonstrates a significant increase in adverse events (oxygen desaturation, hemodynamic instability, and bronchospasm) when suctioning with saline was chosen over dry suctioning, which holds true even with logistic regression modeling. Many adult studies have demonstrated significant decreases in oxygenation when saline is administered just before a suctioning pass. In the one randomized controlled pediatric study, Ridling et al⁸  noted a significant decrease in Spo₂ (approximately 5% vs 1%) at 1 and 2 minutes after a suctioning pass with saline compared with dry suctioning passes; however, no difference in Spo₂ was noted at 10 minutes after suctioning. Our findings support these data in that significantly more suctioning passes were associated with oxygen desaturation when saline was used, but there was no sustained difference in oxygen saturation 10 minutes after the suctioning episode occurred, suggesting a very limited effect.

We also identified 2 additional patient factors after multivariate adjustment (including clustering at the patient level) that may contribute to suctioning-related adverse events. Only a small endotracheal tube and the presence of a comorbid respiratory disease continued to be associated with adverse events, although that association was not as significant as that for saline instillation.

It is unclear why a smaller endotracheal tube would increase the risk of adverse events. Although smaller endotracheal tubes would be associated with younger patients, age does not appear to be an independent risk factor for adverse events. Perhaps because tubes with smaller diameters also have shorter lengths and therefore a narrower window of accurate measured suctioning before the carina is stimulated, patients with these tubes would be at higher risk for agitation leading to desaturation, bronchospasm, and possible hemodynamic instability. Further studies are needed to address this.

It is not surprising that patients with an underlying comorbid respiratory disease at the time of intubation would be at higher risk for adverse events with suctioning. Patients with underlying asthma and chronic lung disease are at high risk for bronchospasm and hypoxia with very little stimulus.

Previous studies have had mixed results concerning the impact that saline instillation has on the rate of complications associated with mechanical ventilation such as ventilator-associated infections and occlusion of the endotracheal tube. In vitro studies have suggested that instillation of saline may actually increase the risk of ventilator-associated infections by dislodging significantly more viable bacteria¹³  and exposure to contaminated saline vials.¹⁴  One recent study¹⁵  in adults showed fewer cases of microbiologically proven VAP when saline instillation occurred before suctioning (relative risk reduction, 54%) and no association between saline use and endotracheal tube obstruction. Other respiratory interventions such as physiotherapy may have contributed to a reduction in their VAP rate.¹⁶  In the only pediatric study⁹  to date in which rates of endotracheal tube occlusion were measured as a primary outcome, when a very small (2.5-mm) endotracheal tube was used, saline instillation was associated with an increased mean time to endotracheal tube occlusion (13.5 vs 77.6 hours, P < .05).

During our study period, very few ventilator-associated complications occurred, consistent with our unit’s low rates for VAP (2.26/1000 ventilator days in 2012) and unplanned extubation (0.52/100 ventilator days in 2012) with only 5 endotracheal tube changes per year for suspected tube occlusion. With so few events during the study period, we were unable to evaluate the potential role that instillation of saline might play in the development of these ventilator-associated complications.

This study has several limitations. First, this study design is observational and the suctioning episodes were not independent. There is also potential for confounding by indication, in that the use of saline for suctioning episodes is specified for certain clinical indications that may themselves increase the risk of adverse events with suctioning. This limitation was most likely partially offset by the inherent subjectivity of several of the algorithm-based indications. Nurses and respiratory therapists were not monitored when administering saline, and adherence to the algorithm was not tracked. Nonetheless, the large number of suctioning events in the study, the number of other covariates assessed, the 2-level logistic regression design to adjust for cluster impacts at the ventilation episode (patient) level, and the relatively large signal magnitude of the association of saline with adverse events with suctioning lend credence to our findings. With only 69 patient episodes analyzed in the study representing 545 ventilator days and no power analysis, only 20% of all ventilator days (up to 2500) reported in our PICU were evaluated, and we cannot exclude a degree of selection bias.

Saline use nearly doubled (50% vs 25% in an unpublished unit survey performed less than a year prior) within our unit during the study period (after the algorithm was instituted). Training about suctioning practice did not change during the study period, and it is unlikely that admitting diagnoses or severity of illness changed drastically in this period to account for a perceived increase in the need for saline. It is unclear if caregivers perceived an expectation to use saline more frequently during the study period.

We demonstrate that saline use for endotracheal tube suctioning is associated with adverse events such as hemodynamic instability, bronchospasm, and transient hypoxemia. Additionally our findings revealed that endotracheal tubes of 4.0 mm or less and comorbid respiratory diseases increased a patient’s risk of experiencing an adverse event during suctioning. Clinical decision trees and algorithms are commonplace in PICUs, and practitioners rely on them to inform safe medical management. Our findings suggest that if the saline arm of the decision tree is chosen, patients are at increased risk of an adverse event during suctioning. Perhaps other techniques such as humidified gases and chest physiotherapy should be used before administration of saline. If saline is administered, it should be used cautiously, and the bedside caregiver should be vigilant in monitoring for adverse events. Further studies are needed to delineate the benefit versus harm of saline instillation and alternative approaches to airway clearance for patients undergoing mechanical ventilation. CCN