Introduction

Endotracheal tube (ETT) suction is the most frequent invasive procedure performed on ventilated infants [1]. It is associated with hypoxia and haemodynamic instability [2, 3, 4], complications generally attributed to loss of lung volume [5, 6, 7]. In paediatric and adult populations open ETT suction, involving disconnection from the ventilator, results in greater lung volume loss than closed suction, in which the ventilator circuit remains intact [5, 6, 7]. The magnitude of lung volume change during open ETT suction has never been quantified in ventilated newborn infants, or in any patient receiving high-frequency oscillatory ventilation (HFOV). HFOV is frequently used to treat newborn infants with severe lung disease [8], who are likely to be at the greatest risk of complications from lung derecruitment, including chronic lung disease and intraventricular haemorrhage.

Measuring change in lung volume during HFOV is difficult, especially in the neonatal population. Respiratory inductive plethysmography (RIP) is a non-invasive method of measuring relative changes in lung volume [9, 10], which has recently been validated during HFOV [11, 12, 13].

An understanding of the effects of open ETT suction on newborn infants receiving HFOV would assist in optimising suction practices. The aim of this study was to describe the pattern and magnitude of change in lung volume during open ETT suction in infants receiving HFOV.

Methods

Study population

The study was performed in the Neonatal Unit, Royal Children's Hospital, Melbourne. The Ethics in Human Research Committee approved the study, and informed parental consent was obtained in each case. Muscle-relaxed infants receiving HFOV (Sensormedics 3100A high-frequency oscillator, Sensormedics, Yorba Linda, Calif.) and requiring ETT suction at least every 4 h were eligible for enrolment. Infants were ventilated through the largest appropriate uncuffed ETT for their weight, and in each case an ETT leak of < 10% was confirmed using a Florian respiratory mechanics monitor (Acutronic Medical Systems, Zug, Switzerland). Infants were not studied if they had congenital heart disease, a known chromosomal abnormality, refractory hypotension, or an FIO2 of greater than 0.9.

Measurements

Change in lung volume was measured with a low-pass filtered, DC-coupled respiratory inductive plethysmograph (Respitrace 200, Non-invasive Monitoring Systems, North Bay Village, Florida) sampling at 200 Hz. After thermal and signal stabilisation [10], the voltage outputs were zeroed and then calibrated to the tidal volume measured at the airway opening using a Florian respiratory monitor. Calibration was performed during 15 positive pressure inflations (peak pressure 26 cm H2O and end-expiratory pressure 6 cm H2O) delivered with a Neopuff infant resuscitator (Fisher and Paykel Healthcare, East Tamaki, New Zealand). A calibrated volume signal was then derived from the sum of the chest and abdominal RIP voltages, allowing changes in lung volume to be quantified from the observed changes in RIP voltage [9].

Suction protocol

Prior to ETT suction, the lungs were recruited with a lung volume optimisation manoeuvre as part of a separate study [13], utilising an open lung approach, such that ventilation was being applied upon the deflation limb of the pressure-volume relationship above the critical closing pressure. A single episode of ETT suction was performed using a pre-measured 6-F disposable catheter (Mallinckrodt, Rowville, Victoria, Australia). After disconnection from ventilation the catheter was inserted to the tip of the ETT, and suction applied at a pressure of –100 mmHg (13.3 kPa) for 6 s while simultaneously withdrawing the catheter. Saline was not instilled during or prior to ETT suction. Infants were not handled and there were no changes to ventilator settings during the study.

Data Collection and Analysis

Lung volume data were digitised and continuously recorded from 30 s before until 60 s after ETT suction using a custom-built data acquisition program designed with LabVIEW 6.0 (National Instruments, Austin, Texas). Each recording was divided into four phases: (a) baseline (pre-suction); (b) disconnection (between disconnection from HFOV and application of suction); (c) suction (from the application of negative pressure until reconnection to HFOV); and (d) Post-suction (a period of 60 s after recommencing HFOV). The change in lung volume (ΔV L) at the end of each phase was determined and expressed relative to the value at baseline. For the baseline and post-suction phases the mean of oscillatory wave troughs (the end-expiratory volumes) over the final 5 s of RIP recording was used to calculate ΔV L; for the disconnection and suction phases a point estimate was used. The point of maximum lung volume loss in the whole recording was also determined. ΔV L values for each phase were compared using repeated-measures analysis of variance (ANOVA) and Tukey post-test. Statistical analysis was performed using GraphPad Prism version 4.02 for Windows (GraphPad Software, San Diego, Calif.).

Results

Patient population

Seven infants were studied and completed the protocol without complications. Their demographic characteristics and clinical details are shown in Table 1.

Table 1 Infant characteristics at study commencement. PPHN persistent pulmonary hypertension of the newborn requiring inhaled nitric oxide, MAS meconium aspiration syndrome, NEC post-operative laparotomy for necrotising enterocolitis, HMD hyaline membrane disease, GA gestational age, P aw mean airway pressure, Δ P pressure amplitude, Fr HFOV frequency, AaDO 2 alveolar-arterial oxygen difference
Full size table

Change in lung volume during suction

Table 2 shows the ΔV L during each phase of the study for each infant. In each infant open ETT suction resulted in a consistent pattern of change in lung volume (Fig. 1) with a distinct loss in lung volume after disconnection, and further volume loss with suction (p < 0.0001, repeated-measures ANOVA). In all cases ΔV L had stabilised before reconnection to HFOV. The point of greatest volume loss occurred during the application of suction in all infants, with a mean ΔV L of −13 ml/kg (SD 4 ml/kg; p < 0.0001).

Fig. 1
figure 1

Representative plot of ΔV L against time 10 s before until 60 s after the suction procedure during open endotracheal tube suction (Infant 7). Grey bar: disconnection phase; hatched bar: suction phase

Full size image
Table 2 Change in lung volume from baseline during each phase of the study
Full size table

Figure 2 shows ΔV L at each phase of the study. Disconnection caused a 10 ml/kg (SD 4 ml/kg) volume loss (p < 0.001 vs. baseline; Tukey post-test), accounting for 76.5% (SD 14.1%) of the total. An additional 3 ml/kg (SD 2 ml/kg) loss occurred during suction (p > 0.05 vs. disconnection; Tukey post-test). After recommencing HFOV, lung volume was rapidly restored. The final post-suction ΔV L 60 s after suction was not significantly different from baseline [mean ΔV L −1 ml/kg (SD 4 ml/kg), p > 0.05; Tukey post-test].

Fig. 2
figure 2

ΔV L during the disconnection, suction and post-suction phases of open ETT suction: combined data from all infants. The minimum ΔV L occurred during the suction phase in all cases. All data expressed as mean and error bars represent standard deviation. ΔV L referenced to baseline. There was a significant difference in ΔV L compared with baseline during the disconnection and suction phases (asterisk: p < 0.001). The final ΔV L did not differ significantly from baseline (cross: p > 0.05). All p -values are repeated-measures ANOVA and Tukey post-test

Full size image

Discussion

In this study we found that open ETT suction in muscle-relaxed newborn infants ventilated with HFOV resulted in significant, albeit transient, loss of lung volume. A striking finding was the consistent pattern and magnitude of lung volume change occurring during open ETT suction, as shown in Fig. 1. This suggests that, even in a heterogeneous population of ventilated newborn infants on HFOV, open ETT suction exerts a considerable effect on lung volume during mechanical ventilation.

Concerns about loss of lung volume during ETT suction relate to the subsequent detrimental effect this has on cardiorespiratory stability [5, 7, 14]. An understanding of this relationship in newborn infants has been hampered by a lack of reliable, and practical, methods of measuring lung volume at the bedside, especially during HFOV. This study, the first to our knowledge exclusively in newborn infants, was able to show that RIP could be used to describe and quantify, in detail, change in lung volume during and after ETT suction. Further research in a larger cohort is warranted to determine the relationship between lung volume change, measured with RIP, and cardiorespiratory stability during ETT suction in newborn infants.

This is the first study to quantify lung volume loss due to open ETT suction during HFOV. Significant derecruitment during open suction has also been described in the paediatric [7] and adult populations [5, 6], during conventional ventilation. The entire suctioning procedure resulted in a lung volume loss of 13 ml/kg, predominantly due to disconnection. Closed ETT suction, which does not require disconnection from ventilation, has been advocated as a method of preventing lung volume loss during ETT suction in adult [5, 6] and paediatric populations [7]. In these populations closed ETT suction results in less hypoxia and cardiovascular instability than open ETT suction, possibly at the expense of poorer effectiveness [15]. A comparison of lung volume changes during open and closed ETT suction in neonates has not been studied and is required.

Equally important as the minimum lung volume is the duration in which the lung remains at a low volume state. In this study all infants, except infant 3, regained the lung volume lost during ETT suction by 60 s post-suction, without the need of any recruitment manoeuvres. This is a somewhat surprising result given that alveoli once collapsed require a higher pressure to re-open [16]. In infant 3 at 60 s post-suction lung volume remained 9 ml/kg below baseline, indicating that a recruitment manoeuvre would have been beneficial to avoid atelectasis. Conversely, post-suction lung volume was 4 ml/kg above baseline in infant 6. The application of a recruitment manoeuvre in this infant would place the lung at risk of overdistension. The practice of using recruitment manoeuvres after open ETT suction is often advocated [6, 17, 18] but has not been validated in newborn infants. The RIP may aid clinicians in determining when recruitment is warranted.

This study aimed to standardise ventilation in each subject to the region above the closing pressure on the deflation limb of the pressure-volume relationship. This has been a limitation of previous studies describing lung volume changes during ETT suction in which the volume state of the lung was unknown [5, 6, 7].

This small study has some limitations. Firstly, all the infants in this study were muscle-relaxed and would not be able to sustain lung volume when disconnected to atmosphere. Most newborn infants receiving ETT suction are spontaneously breathing and may maintain lung volume better during disconnection. The study population involved predominantly term infants with a diverse range of respiratory diseases; it cannot be assumed that results would be the same in the preterm infant with hyaline membrane disease, or in the paediatric population.

Another limitation refers to the limitations of RIP to measure changes in lung volume. The RIP cannot discriminate between gas and fluid changes. Changes in intra-thoracic blood volume cannot be excluded. Furthermore, in this study the RIP ΔV L signal was derived from the sum of equally weighted chest and abdominal voltage signals, and calibrated against a known volume. This method is simple and clinically applicable but may be less accurate than methods that take into account the degree of lung injury in determining a calibration factor using qualitative diagnostic calibration [19]. The RIP is a validated tool for measuring change in global lung volume during HFOV [11, 12, 13], but neonatal respiratory disease is rarely characterised by uniform alveolar volume. The RIP is unable to determine the effect of ETT suction in the dependent and non-dependent regions of the lung. Electrical impedance tomography, which measures regional impedance changes, may be an alternative to RIP in the future [20].

Conclusion

In this study of muscle-relaxed and ventilated newborn infants, open ETT suction caused a significant immediate loss of lung volume, principally due to disconnection. This was brief and recovery was rapid.