Cuff Pressure of Endotracheal Tubes After Changes in Body Position in Critically Ill Patients Treated With Mechanical Ventilation

A prospective, interventional study was done with ICU patients who were sedated and receiving mechanical ventilation. Each patient was his or her own control. The study was conducted between February and May 2011 in the 12-bed ICU of the General Hospital Sint Blasius, Dendermonde (Belgium). The study was approved by the appropriate ethics committees. Informed consent was obtained from each patient or the patient’s family.

A convenience sample of 12 patients was selected. Inclusion criteria were age older than 18 years, oral intubation with a TaperGuard Evac endotracheal tube (Covidien), conventional mechanical ventilation, and adequate sedation (score of −5 on the Richmond Agitation-Sedation Scale)¹⁸  and analgesia (score of 3–4 on the Behavior Pain Scale).¹⁹  Exclusion criteria were factors that might influence cuff pressure or compromise the safety or well-being of the patient: pregnancy, palliative care, difficult intubation, decreased mobility of the neck, history of neck surgery, core temperature less than 35ºC or greater than 37.5ºC, morbid obesity (body mass index >35, calculated as weight in kilograms divided by height in meters squared), and potential contraindications for changes in body position such as unstable spinal cord injury or unstable hemodynamic and/or respiratory status.

The anesthesiologist determined the size of the endotracheal tube. Intubation was performed according to local practices. Correct cuff positioning was confirmed by using laryngoscopy, auscultation revealing bilateral breathing sounds, and chest radiography. Endotracheal tubes were secured with tape in a corner of the mouth. The position of the endotracheal tube was not changed during the procedure.

Each patient was brought into a neutral starting position: backrest with a 30º head-of-bed elevation and the head in a neutral position. Each bed had a protractor for measuring the elevation of the head of the bed. After the starting position was established, the suctioning system of the endotracheal tube was disconnected to avoid any influence of suctioning on the cuff pressure. Neuromuscular blockade with intravenous cisatracurium (0.18 mg/kg) was used to prevent patients from breathing against the ventilator and thus causing increases in cuff pressure during the study intervention. The pilot balloon of the endotracheal tube was connected with a tube leading to a calibrated universal pressure monitor (Ultramedic LTD, BioTech Instruments Inc). This monitor measures pressures from −949 to 1055 cm H₂O with an accuracy of ±1%. The monitor was zeroed before each study patient or procedure. A manifold was connected in the middle of the length of the tube, and a syringe filled with air was placed on the manifold to control cuff inflation. The syringe connection was sealed after cuff inflation. At the starting position and during an end-expiratory ventilatory hold, the cuff pressure was set at 25 cm H₂O. During the procedure, the pressure monitor was continuously attached to the pilot balloon of the endotracheal tube because reconnecting results in air loss.²⁰ 

Each patient was brought into 16 different body positions from the starting position (in the same sequence): anteflexion of the head; hyperextension of the head; left lateral flexion of the head; right lateral flexion of the head; rotation of the head to the left; rotation of the head to the right; semirecumbent position with 45º elevation of the head of the bed, recumbent position with 10º elevation of the head of the bed; horizontal backrest (ie, supine position); Trendelenburg position 10°; left lateral position at 30º, 45º, and 90º; and right lateral position at 30º, 45º, and 90º. These positions were chosen because they are part of daily ICU routine. For the lateral positions, foam blocks with a slopes of 30°, 45°, and 90° were used. Changes in position were executed one after the other. Patients were not returned to the starting position after each measurement to check cuff pressure because this maneuver would have substantially lengthened the time required for the procedure. Moreover, reaching a target position often requires going through a series of previous positions (eg, to achieve a 90° lateral position starting from a semirecumbent position, a patient must go from a recumbent position with 10° head-of-bed elevation to supine to a lateral position of 30°, and a lateral position of 45°. Therefore, returning to the start position to (eventually) adjust the cuff pressure was deemed pointless. In every position, traction on the endotracheal tube was avoided.

Once a patient was brought in a predetermined position, an end-expiratory ventilatory hold was performed for a 4 to 5 seconds and cuff pressure was recorded. The procedure required a collaboration of 4 health care workers: an intensivist (W.S.) guaranteed the sedation, neuromuscular blockade, and end-expiratory ventilator holds; 2 nurses placed the patient in the correct position; and 1 researcher (C.L.) was responsible for monitoring the cuff pressure and avoiding interference with measurement (eg, traction on the endotracheal tube). The total procedure took about 20 to 25 minutes.

Variables recorded included age, sex, length, weight, body mass index, and diagnostic category. Variables of the endotracheal tube included size, depth of the endotracheal tube to the teeth, location of the tube in the mouth, mode of mechanical ventilation, and positive end-expiratory pressure. Other outcome variables were cuff pressures in the predetermined positions.

Study power was assessed. Because a deviation of 5 cm H₂O would be needed to reach the target cuff pressure limits, the power analysis was based on an anticipated difference of 5 or more cm H₂O, with a standard deviation of 7 cm H₂O. For this difference to be significant (β= 0.80; α= 0.05), a sample of 12 patients was needed.

Variables are reported as number and percentage, median and first through third quartile, or mean and standard deviation. The Wilcoxon paired sample test was used to compare cuff pressure in the starting position (25 cm H₂O) with cuff pressure in a defined position. Besides absolute deviations (in centimeters of water), the number of measurements outside the recommended cuff pressure (<20 cm H₂O or >30 cm H₂O) was noted. These observations are considered clinically relevant. The Fisher exact test was used for comparisons between categorical variables. Within-patient variability in cuff pressure across the 16 positions was evaluated by using repeated-measures analysis of variance. The Greenhouse-Geisser method was used to correct for nonsphericity, with P > .05 indicating no within-patient variability. SPSS 19 software (IBM SPSS Statistics) was used for analyses. All tests were 2-tailed. Statistical significance was defined as P less than .05.

The characteristics of the 12 patients in the study are given in Table 1. The data included all measurements for each patient. No adverse events occurred. In total, 192 measurements were made (16 positions × 12 patients). Table 2 gives the mean cuff pressure per body position and the number of measurements that were outside the target range. In every position, the cuff pressure differed significantly from the pressure at the starting position. Among the 192 measurements, 78 (40.6%) were greater than the upper target limit of 30 cm H₂O. In contrast, no measurements were less than the lower target limit of 20 cm H₂O. Only 17 measurements (9%) were greater than 25 cm H₂O. Figure 1 illustrates cuff pressures according to body position. In each position, the upper target pressure limit was exceeded at least once.

Within-patient variability was substantial (Figure 2). Repeated-measures analysis of variance indicated significant variability in patients’ cuff pressure across the 16 positions (F = 3.63; P = .02). Cuff pressure remained within the target interval in all 16 positions for only 1 patient. The upper target limit was exceeded 2 times for 1 patient, 3 times for another patient, 5 times for another patient, and 9 times for another patient. In 2 patients, the target limit was exceeded in 6 positions, and in 2 other patients, it was exceeded in 7 positions. In 3 patients, the upper target limit was exceeded 11 times.

Our results indicate that simple and frequently executed changes in body position can have a significant effect on the cuff pressure of endotracheal tubes in patients receiving mechanical ventilation. No cuff pressures were less than the 20 cm H₂O lower threshold; 40.6% of measurements exceeded the 30 cm H₂O upper target limit. Additionally, because of high within-patient variability (Figure 2), we could not predict which patient would have high cuff pressure in particular positions. We found no relationship between cuff pressures and patients’ characteristics. We think that anatomical differences between patients may be responsible for this observation. Autopsy studies^14,21  indicated that tracheal cartilage rings can differ substantially from the classic horseshoe shape.

Brimacombe et al²²  also described increased cuff pressure after changes in head and neck positions in 10 anesthetized patients. Starting from a 40.8 cm H₂O pressure in a neutral position (30 mm Hg), the mean cuff pressures in anteflexion, extension, and rotation of the head were 51.7, 43.5, and 47.6 cm H₂O, respectively. As in our study, all deviations were significant. Of note, cuff pressure in the starting position in the study by Brimacombe et al was 30 mm Hg (40.8 cm H₂O), which is higher than the currently recommended target pressure. In a crossover study, de Godoy et al²³  evaluated the extent that changes in body position in patients treated with mechanical ventilation resulted in changes in cuff pressure. They looked for the effect of changes from 35° semirecumbent positioning to lateral positioning and back and used a slightly different target cuff pressure interval (24.5–29.9 cm H₂O) than we did. Nevertheless, their findings that 50.7% of measurements resulted in relevant increases in cuff pressure (defined as an increase from 27.2 cm H₂O to >29.9 cm H₂O) are similar to our results. As in our study, decreases in cuff pressure (from 27.2 cm H₂O to <24.5 cm H₂O) were less common (5% of measurements).

Cuff pressures greater than 30 cm H₂O are clinically important because they compromise the mucosal capillary perfusion and may lead to tracheal injury. In turn, these acute injuries may evolve into tracheal stenosis or formation of a fistula. We hypothesize that ICU clinicians’ awareness of the deleterious effects of excessive cuff pressure is low, because the consequences of such pressure do not become obvious until weeks or months after the patient is discharged from the ICU; thus ICU staff rarely observe the effects. Nevertheless, the incidence of clinically relevant injuries after endotracheal intubation is high. In a prospective study,²⁴  the estimated incidence of tracheal stenosis in patients intubated for 5 to 10 days was 4%, and up to 12% among patients intubated for 11 to 24 days. In addition, intubation-related tracheal injury contributes to health care costs via increased costs of care and length of hospitalization during the index admission and future costs for treatment and repair.²⁵  In a recent study,²⁶  83% of patients included in an randomized trial had at least 1 tracheal ischemic lesion. Yet, findings on fiber-optic tracheoscopy 2 weeks after intubation were normal in all except 1 patient, who experienced tracheal rupture.

Despite the obvious risks associated with excessive pressure on the tracheal wall, in daily practice, clinicians rarely evaluate cuff pressure to be sure it is correct. Indeed, cuff pressures outside the target range are common,^27–29  and the frequency with which cuff pressure is measured and adjusted varies from never to at most every 8 hours.^30–33  Additionally, cuff pressure is often evaluated by using finger palpation.³⁰  Several investigators^34–36  have reported the inability of clinicians to adequately estimate cuff pressure by using finger palpation, and this practice generally results in excessive pressures. Monitoring cuff pressure via a manometer results in fewer complications after incubation.³⁷ 

Our data indicate that cuff pressure should be measured after each change in a patient’s body position and supports use of continuous monitoring of cuff pressure with automatic adaptation to a preset pressure. When all other factors that may cause alterations in cuff pressure are considered, the question is whether or not monitoring the pressure by using intermittent manual evaluation remains an option. Care procedures that require monitoring of cuff pressure appear to be multiple and may lead to a further increase in nurses’ workload. Devices to provide automated monitoring and adjustments of cuff pressure have been developed. Such devices can maintain the pressure within the target range^38,39  and reduce the risk of ventilator-associated pneumonia.³⁸ 

Our study has limitations. First, cuff pressure was measured once, soon after correct positioning of the patient, and was not followed for a more prolonged period. Sole and colleagues^40,41  found that alterations in cuff pressure after changes in patients’ body position often were transient and normalized within 15 minutes. While interpreting the present results, one must realize that cuff pressures were observed for only a very short period of time: although temporary overinflation might be harmless, a temporary decrease in cuff pressure might result in marked microaspiration of subglottic secretions, leading to ventilator-associated pneumonia. Second, we used an endotracheal tube with a tapered cuff. The extent, if any, that cuff shape influences cuff pressure after changes in body position is unknown. Third, no clinical outcomes were measured. The focus of our study was deviations in cuff pressure outside the target interval, because such changes can be considered clinically relevant. Because the deleterious effect of cuff pressures not in the target range has been demonstrated before, observing how excessive pressures evolve over time and whether or not they result in clinically important complications would be unethical. Finally, we did not measure intra-abdominal pressure although it might have influenced cuff pressure. Still, because one of the requirements for inclusion in the study was stable clinical status, most likely none of our patients had intra-abdominal hypertension.

Our study also has some strengths. First, having each patient serve as his or her own control counters the problem of nonequivalence between an experimental and a control group. Second, factors that might influence the cuff pressure were eliminated as much as possible by the selection process (exclusion criteria), the neuromuscular blockade, and measurement of the cuff pressure during an end-expiratory ventilator hold. Finally, we evaluated the effect of changes in body position over a broad range of positions that are often used in daily practice.

Changes in a patient’s body position resulted in significant deviations in the cuff pressure of endotracheal tubes. In 40.6% of the measurements, the cuff pressure exceeded the upper target level, and thus theoretically was clinically relevant. Because simple and frequent changes in a patient’s body position may result in potentially harmful cuff pressure, our observations suggest a need for a strict monitoring of this pressure.

Christelle Lizy and Walter Swinnen contributed equally to the manuscript. We thank Mr Chris Parmentier, head nurse, and his team of ICU nurses at General Hospital Sint Blasius, Dendermonde (Belgium), for their valuable support and cooperation during the study. We also thank Mr Filip Vanthyune, biomedical engineer at Ghent University Hospital, for his technical advice and support. The results of a pilot test of this research were presented at the 24th Congress of the European Society of Intensive Care Medicine, Berlin, Germany, October 2–5, 2011 (see Intensive Care Med. 2011;37(suppl 1):S108; abstract 0411), and parts of the final results were presented at the 33rd International Symposium on Intensive Care and Emergency Medicine, Brussels, Belgium, March 19–22, 2013.