Although the need for intensive care has often been defined by the need for ventilation, and there are literally thousands of publications on techniques, principles, complications, and challenges of ventilation, there is a surprising lack of evidence for best practice regarding a fundamental technique in ventilation: suctioning of the airway. Since tracheostomy or endotracheal intubation was first undertaken, potential obstruction of the endotracheal tube by mucus has been a consistent and life-threatening problem. That has been particularly true for infants and children, especially those with increased respiratory secretions. The obvious (not always so easy) solution is adequate humidification and suctioning. Thus, endotracheal suctioning is probably the most common procedure in pediatric and neonatal intensive care practice.

The ideal suctioning technique would be pain- and discomfort-free, safe (with no adverse events such as loss of lung volume, desaturation, cardiovascular changes, CNS changes, damage to the respiratory system at any level, introduction of infection, etc.), and effective (removing all excessive secretions, keeping the endotracheal tube clear and unobstructed).

The reality is that suctioning has been associated with a plethora of adverse events and unpleasant side effects. In preterm infants, it has been associated with changes in cerebral oxygenation (1,2) and pressures (3) and hemodynamics (2–4); in infants, with atelectasis (5), transient bacteremia (6), hypoxia, and cardiovascular changes (7,8); and in children with hypoxia (8) and upper lobe atelectasis (9). For obvious reasons, we do not have the patient's perspective on endotracheal suctioning in infancy. However, in adult studies endotracheal suctioning is clearly remembered as unpleasant and in a recent study, pain on endotracheal suctioning was rated as moderate to severe by more than half the patients (10).

Although detailed recommendations for suctioning technique are available in most pediatric intensive care textbooks, the underlying evidence for the recommendations is often limited and is based on adult data. Although preoxygenation has been widely recommended as a means of decreasing complications after endotracheal suctioning, recent reviews concluded that there was not adequate evidence to fully support the practice in preterm infants (11,12). Similarly there was inadequate evidence to support the practice of nondisconnection of the ventilator during suctioning (13). Recently, a reviewer was unable to find any evidence to address the question of whether endotracheal suctioning in neonates should be limited to keeping the suction catheter within the endotracheal tube or whether it should be extended into the trachea beyond (14). An adult study showed that minimally invasive suctioning (limited to endotracheal tube) was associated with fewer adverse events, no deleterious effects (15), and less subsequent recall of endotracheal suctioning (16). Within the published pediatric literature, there is a wide range of techniques reported (Table 1).

Table 1 Some recent publications on endotracheal suctioning in infants and children
Full size table

In 1991, Singh et al. (17) were among the first to examine detailed techniques in pediatric suctioning (Table 1). Since then a number of studies have focused on the process of endotracheal suctioning. Initial studies considered the theoretical aspects of flow within the endotracheal tube during suctioning (18) and moved on to data obtained with a simple model. Those data were expanded with some studies considering lung mechanics after endotracheal suctioning (19,20). Some elegant theoretical (20,21) and practical studies followed, which highlighted the complexity of flows within suction catheters and endotracheal tubes during endotracheal suction.

The effects of endotracheal suctioning probably depend on many issues including underlying lung pathology, patient sedation and use of paralysis, the details of the suctioning technique, particular ventilatory techniques such as pressure control or volume control modes (22), the use of PEEP (23), and potentially whether recruitment maneuvers are used after the procedure.

A number of studies have focused on the issue of whether “open” or “closed” systems make a significant difference (24–26). Hoellering et al. (27) have recently reported on their studies on endotracheal suctioning in 20 infants [mean gestational age 34.5 wk (24–40 wk), chronological age 18.5 d (3–61 d), and weight 1.93 kg (0.57–5.68 kg) kg] on conventional ventilation. There was no difference in the drop in lung volume (as measured by respiratory impedance tomography) after open or closed suctioning. By contrast, there was a trend toward a drop in lung volume after open suctioning for a group of 10 infants [mean gestational age 40 wk (23–42 wk), chronological age 3 d (1–38 d), and weight 3.28 kg (0.83–3.70 kg)], who were on high-frequency ventilation.

In this edition of the journal, Copnell et al. (28) have presented data on “the effect of suction method (open, closed in-line and closed with a side-port adaptor), catheter size and suction pressure on lung volume changes during endotracheal suction” during both conventional and high-frequency oscillatory ventilation. They provided a very carefully standardized model of animals with lung injury analogous to surfactant deficiency (newborn piglets after multiple saline lavage), standardized the ventilatory approach to these animals in line with current recommendations for the ventilation of infants (children and adults) with ARDS, standardized the lung volume of the subjects at all test points, and applied a standardized suction technique with a single pass of the suction catheter to the end of the ETT and 6 s of applied suction (at different pressures). In this study, closed systems did seem to be advantageous with regard to maintenance of lung volume, but only in certain circumstances, and even then not at a clinically significant level.

How do these findings relate to current practice in intensive care? First, the specific condition that has been modeled is analogous to hyaline membrane disease and ARDS in older infants and children. In pediatric practice, ARDS is a relatively uncommon reason for ventilation, and much more work will be needed to optimize suctioning in patients with more common conditions such as bronchiolitis and viral or bacterial bronchopneumonia. In a population of patients with variable pathology Choong et al. (26) demonstrated an increased loss of lung volume in patients with “noncompliant lungs” (compliance <0.8 mL cm H2O−1 kg−1 and fraction of inspired oxygen requirements ≥0.4).

The issue of preoxygenation and preparation for suctioning needs to be addressed (11,12). In this particular study, animals were maintained on fraction of inspired oxygen of 1.0 throughout, and the animals were paralyzed and sedated (unlike current practice in most neonatal and pediatric intensive care units).

In both this article (28) and previous studies from the same group (21,29), it was notable that, when using the two smaller catheters (6 and 7FG) at the highest pressure, volume loss was less than or similar to that generated by an 8-FG catheter at the lowest pressure. Thus, recommendations for suctioning need to address both catheter size and suctioning pressure. Furthermore, there was a wide variation in changes in lung function measurements with suctioning—a feature of much the pediatric work related to suctioning and chest physiotherapy techniques (30,31)—despite the standardization of the model.

This article has not addressed the issue of whether there may be regional changes in lung volume (with or without overall changes in lung volume). Lindgren et al. (32) in an animal model of acute lung injury (saline lavage) demonstrated that the lung volume loss was predominantly in dorsal regions of the lung (and not from the overall lung), with almost complete deaeration of these areas during open suctioning. They applied suctioning for 10 s with vacuum level −20 kPa (∼−150 mm Hg, −200 cm H2O) and a 14-F catheter.

The article was not directed at the question of what pattern or method of suctioning is most effective at removing secretions (as pointed out by the authors in the Discussion). In a study of 18 adult patients with acute lung injury (33), it was noticeable that open suctioning removed significantly more secretions than closed suctioning (despite worse desaturation associated with open suctioning). Similarly, more secretions were removed with a suction pressure of −400 cm H2O than with −200 cm H2O (32). In previous animal studies (20,23), open suction techniques removed more secretions than closed systems.

Clearly, there is much work to be done to understand the influence of different suctioning systems, different methods and techniques of suctioning, the underlying respiratory pathophysiology of the child, the best possible ways of timing the need for suctioning, and the best techniques for removal of troublesome secretions.

As we learn more about appropriate suctioning techniques, it is probably as important that we address ways of implementing this research at the clinical level. Recently, Kelleher and Andrews (34) studied the practice of open endotracheal suction in two adult intensive care units and found substantial variation in practice and poor adherence to best practice suctioning recommendations. They reported significant discrepancies in practices regarding respiratory assessment techniques, hyperoxygenation and infection control practices, patient reassurance, and the level of negative pressure used to clear secretions. Encouragingly, Day et al. (35) demonstrated improvement in both knowledge and practice in a group of nurses who were provided with focused teaching on endotracheal suction techniques.