Mask ventilation is an essential component of airway management for apnoeic subjects during either resuscitation or general anaesthesia. However, upper airway obstruction occurs frequently and may cause difficult, and sometimes impossible, mask ventilation.1–4 The overall incidence of difficult mask ventilation during adult general anaesthesia is reported as 2.2% with impossible mask ventilation occurring 0.15% of the time.4 It is also known that difficult mask ventilation often coexists with difficult tracheal intubation.2,4 The latter can lead to a ‘can’t intubate, can’t ventilate’ scenario, which will have a catastrophic outcome if effective spontaneous respiration, or a surgical airway, cannot be established in a timely manner. Therefore, effective mask ventilation is as important as successful tracheal intubation: in short, they are rescue techniques for each other. Unfortunately, the importance of effective mask ventilation has not been appreciated to the extent that successful tracheal intubation is appreciated.5 Clinicians may unintentionally compromise a patient's safety by repeated attempts at tracheal intubation without optimising mask ventilation.
It has long been recognised that head and body positions influence upper airway patency. Head extension6–8 and a lateral recumbent position7,9 reduce upper airway obstruction, whereas head flexion reduces the pharyngeal space.7,10 However, the effect of axial head rotation on airway patency has not been evaluated systematically. Head rotation increases the cross-sectional area of the upper airway in normal awake subjects11–13 and, during drug-induced sleep endoscopy, improves the upper airway patency in patients with obstructive sleep apnoea (OSA).14,15 However, others have found that head rotation did not decrease pharyngeal critical closing pressure in normal sleeping subjects.7 Thus, it remains uncertain whether head rotation improves airway patency and efficiency of mask ventilation in anesthetised individuals. We hypothesised that head rotation applied during face mask ventilation would reduce upper airway obstruction and improve tidal volume in anaesthetised, paralysed apnoeic adults.
The aim of this study was to test this hypothesis and to determine the effect of 45° head rotation on the efficiency of mask ventilation. Secondarily, we set out to clarify which factors were associated with changes in the efficiency of mask ventilation induced by head rotation.
Materials and methods
Ethical approval for this study (protocol number 2013P002571/MGH) was provided by the Partners Human Research Committee, Boston, Massachusetts, USA (Director Rosalyn Gray) on 6 March 2014. Written informed consent was obtained from all patients.
Subject enrolment and randomisation
All subjects, aged 18 to 75 years, with a BMI in the range of 18.5 to 35.0 kg m−2, who met the criteria for ASA physical status classification I to III, and required general anaesthesia with tracheal intubation for elective surgery at Massachusetts General Hospital (Boston, Massachusetts, USA) were recruited. We excluded patients with limited head rotation or neck extension, as well as those with gastro-oesophageal reflux, a ‘full stomach’, known sleep apnoea on continuous positive airway pressure (CPAP) therapy and those requiring awake intubation. Given the paucity of evidence to support the use of head rotation as a manoeuver to optimise face mask ventilation, we limited our study to a patient population where we believed adding time to the induction period would not present a significant risk to the subject. The process of subject selection and randomisation is shown in Fig. 1. For each patient, a standard airway examination was performed.
Patients were randomised into two equal groups. Randomisation was accomplished by creation of a series of envelopes containing the group assignments (T.I.). Either ‘Group A’ or ‘Group B’ was written on a sheet of paper, which was then placed in a sealed envelope. There were an equal number for each group. The envelopes were then shuffled and numbered. To determine group assignment, the envelopes were opened in sequence, as each patient was recruited to the study. The envelope was opened immediately before the induction of anaesthesia. Thus, the staff involved in the study did not know the group assignment at the time of recruitment, but were not blinded during the clinical portion of the study, or during data analysis.
Standard ASA monitors (electrocardiogram, noninvasive blood pressure and pulse oximetry) were applied. To measure and record respiratory parameters [airway flow, airway pressure, tidal volume, and end-tidal CO2 (ETCO2)], a carbon dioxide/flow sensor (NICO; noninvasive cardiopulmonary management system, model 7300; Respironics Corp., Murrysville, Pennsylvania, USA) was placed between the mask and breathing circuit. In addition, to measure the expiratory tidal volume (VTE) during mask ventilation, thoracoabdominal movements were recorded by respiratory inductive plethysmography (RIP) (Respitrace Calibrator; Ambulatory Monitoring Inc., Ardsley, New York, USA). As previously described,16 transducers, held in place by expandable bands, were placed around the thorax and abdomen before induction. The experimental setup is shown in Fig. 2. For the study, the patient's occiput was elevated about 10 cm above the operating table with a gel head rest and folded blankets, as is standard practice at our institution.
The study procedure began with preoxygenation using 100% oxygen given via a circle system and standard face mask (Sure Seal Mask; Hudson RCI/Teleflex, Research Triangle Park, North Carolina, USA) until the fraction of expired oxygen reached 0.8 or greater. The oxygen flow rate was not prespecified, but in our normal practice, typical flow rates are in the region of 10 l min−1. For induction of anaesthesia, all patients were given fentanyl (1 to 2 μg kg−1), propofol (1 to 2 mg kg−1) followed by either cisatracurium (0.2 mg kg−1) or rocuronium (0.6 mg kg−1) intravenously. General anaesthesia was maintained with inhaled sevoflurane or intermittent boluses of propofol. Upon achieving apnoea, as evidenced by no flow and no respiratory effort, mask ventilation was carried out with pressure control ventilation using an anaesthesia ventilator (Apollo; Dräger Medical, Telford, Pennsylvania, USA) at 10 breaths per minute, inspiratory to expiratory ratio 1 : 2, a positive inspiratory pressure (PIP) of 15 cmH2O, and a positive end expiratory pressure (PEEP) of 0 cmH2O. We set the PIP at 15 cmH2O to minimise gastric insufflation.17 The mask was held in place with two hands and the jaw was positioned to optimise the airway. In Group A, the mask ventilation was performed with a neutral head position for 1 min (total 10 breaths) (Step 1), followed by an axial head position rotated 45° to the right for 1 min (Step 2) and, finally, the head was returned to the neutral position for another 1 min (Step 3). In Group B, all the ventilator settings were identical to those of Group A, but the sequence of head positioning was rotated → neutral → rotated. Head rotation was estimated with the aid of a protractor (Fig. 2). If no ventilation was possible for at least four breaths, as evidenced by a lack of chest wall movement, the absence of an ETCO2 trace and no observed inductance changes of RIP measurements, the next head position was applied. Then, if no ventilation was possible over 4 breaths in the next head position, the study was terminated and routine airway management was initiated. If ventilation was detected at the next head position, the study continued as planned. At any time, if the SpO2 decreased to 92% or the ETCO2 increased to greater than 6.7 kPa (50 mmHg), the study was terminated and routine airway management was resumed.
Upon completion of the study, an endotracheal tube was inserted. Once the airway was secured, breaths were delivered through the endotracheal tube at various PIPs including 15 cmH2O. The calibration curve for the RIP was then generated as previously described.18 For each patient, the VTE during each head position was then calculated retrospectively by using the patient's own calibration curve.
At each head position, up to 10 breaths were collected and the last 3 breaths were analysed. All other ventilation parameters, airway flow, tidal volume, actual PIP/PEEP and ETCO2 were recorded by the NICO monitor. Data from the same head positions (Step 1 and Step 3) were averaged and the mean value was used as a single datum point for final analysis. When the study was terminated during Step 2, the data at Step 1 were included in the analysis. Zero volumes were not included.
We defined ‘airway obstruction’ as occurring when the ratio VTE during mask ventilation in the neutral head position/VTE after tracheal intubation at 15 cmH2O was less than 1.0, and ‘low VTE’ as VTE of less than 6 ml kg−1.
We planned a study of a continuous response variable utilising a prospective crossover design. Our primary outcome of interest was absolute VTE in ml. We based our sample size on a previous study that measured PIP.16 In pressure control ventilation, PIP is the major determinant of VTE. Assuming a mask-ventilation failure rate of 20%, we enrolled 40 subjects to ensure the desired sample size of 32 participants. No additional power calculation was conducted.
We first assessed individual variable distributions through density plots and created Table 1, which includes overall descriptive statistics as well as the standardised mean difference (SMD) for each variable by assigned group. Next, we used a series of linear subject level random intercept models to distinguish which associated factors were statistically significant in this prospective crossover study. Methodological checks were performed to examine the effect of head position order, and measurement occasion (first, second or third assessment). To identify difference in VTE by head position, a univariate mixed model was run for each effect moderator. Differences between continuous variables were assessed using t tests or Mann–Whitney tests and categorical variables were assessed with Chi-squared tests or Fisher exact tests, as appropriate. All P values, as well as adjusted P values using a Bonferroni correction for multiple testing, were reported. All statistical tests were two tailed with alpha set to 0.05. All analyses were performed with statistical software (RStudio, version 3.2.3; R Foundation for Statistical Computing, Vienna, Austria).
A total of 40 patients were randomly assigned to group A or B (Fig. 1). Patient enrolment started on 1 July 2014 and terminated on 5 January 2016. Enrolment was terminated when we reached our enrolment target. Two patients were excluded from group A due to protocol violations, resulting in 38 patients included in the final analysis. For two patients in group B, we were compelled to terminate the study because of severe airway obstruction. One patient showed no evidence of ventilation in Step 2 (neutral) and in the next head position (Step 3, rotated), although a small VTE (1.1 ml kg−1) was measured during Step 1 (rotated). The other patient could not be ventilated either in Step 1 (rotated) nor Step 2 (neutral). No patient experienced a SpO2 less than 92%, or an ETCO2 greater than 6.7 kPa. We did not identify any specific harm associated with treatment. The mean age of the patients was 62 (range, 40 to 75) years, the mean BMI was 28 (range, 20 to 35) kg m−2 and 27 patients (71%) were male. Demographic data are summarised in Table 1. Methodological checks confirmed that group order was not statistically significant, but head position and measurement occasion were significant in relation to VTE. If measurement occasion was included in the models, VTE estimates changed by less than 0.05 ml.
Figure 3a shows VTE of each patient during mask ventilation in neutral position and head rotation. The VTE was significantly higher with head rotation (612.6 ml) than in the neutral position (544.0 ml): difference [95% confidence interval, 95% CI], 68.6 [46.8 to 90.4] ml, P < 0.0001). The average VTE was not significantly different between group A and group B during either the neutral position (576.4 vs. 523.7 ml, P = 0.315) or head rotation (675.9 vs. 564.4 ml, P = 0.110). Forty-five degrees head rotation significantly increased VTE in both groups: mean difference [95% CI] group A, 99.5 [65.3 to 133.7] ml, P < 0.001; group B 40.7 [13.7 to 67.7] ml; P < 0.005. The VTE improvement was not significantly different between the groups (P = 0.296) (Fig. 3b).
We evaluated the effects of head rotation on VTE in the context of several variables. Of these variables, age, height, Mallampati classification, airway obstruction and low VTE significantly changed the relationship between head rotation and VTE (Table 2). The effects of these five moderators are illustrated in Fig. 4. Overall, all patients benefited from the 45° head rotation, but certain patients experienced a greater benefit. For example, patients who experienced airway obstruction experienced a higher increase in VTE when changing from the neutral to 45° head rotation than patients who did not experience airway obstruction [difference in slopes (95% CI): −69.2 (−113.1 to −25.5) ml, adjusted P < 0.05, Fig. 4a]. Younger patients appeared to benefit significantly more from the 45° rotation than older patients [difference in slopes (95% CI): -5.0 (2.5 to 7.6) ml, adjusted P < 0.005, Fig. 4b]. Patients with Mallampati score I benefitted more from the 45° head rotation than patients with Mallampati score II [difference in slopes (95% CI): 104.7 (52.4 to 156.9) ml, adjusted P = 0.002] or III [difference in slopes (95% CI): 95.7 (38.5 to 152.6) ml, adjusted P = 0.021, Fig. 4c]. Individuals with low VTE experienced greater benefit from the head rotation than patients that did not have low VTE [difference in slopes (95% CI): -69.9 (-122.6 to -17.3) ml, adjusted P = 0.230, Fig. 4d]. Finally, taller patients were able to maintain a higher VTE with the assistance of the 45° head rotation [difference in slopes (95% CI): -2.8 (-4.8 to -0.6) ml, adjusted P = 0.253, Fig. 4e].
The most important findings of our study are 45° head rotation significantly increased expiratory tidal volume compared with the neutral head position in apnoeic adult patients, and the effect of head rotation was independently associated with airway obstruction, younger age and Mallampati classification I.
Effect of head position on airway patency
The mechanisms of upper airway obstruction during general anaesthesia are multifactorial and include a reduction of pharyngeal dilator muscle activity and gravitational effects on anterior upper airway structures in the supine position.19 It has been reported that lateral positioning9,11–13 and reverse Trendelenburg position8,20 improve the efficiency of mask ventilation, and a change in the direction of gravity by a position change was considered to be the mechanism most responsible for reducing upper airway collapsibility. It has been proposed, but never systematically determined, that head rotation improves the efficiency of mask ventilation. Thus, to our knowledge, this is the first in-depth study to demonstrate that head rotation can indeed improve VTE, probably by reduction in airway obstruction.
The influences of head rotation on upper airway collapsibility have been visually studied by imaging11–13 and endoscopic14,15 approaches. In MRI studies of normal awake individuals, both the head rotated and lateral recumbent position induced significant increases in the antero-posterior dimensions of the retroglossal11 and retropalatal region.12 Walsh et al.13 observed that awake patients, with and without OSA, had increased circularity of the upper airway shape in the lateral position and concluded that this change in shape might decrease the propensity to upper airway collapse. In a study of OSA patients under complete muscle paralysis, Isono et al.9 observed similar enlarged airways in both retropalatal and retroglossal regions in the lateral position. In terms of muscle relaxation, their findings support the hypothesis of our study. Although it remains unknown if head rotation leads to a reduction in airway obstruction, Ono et al.11 observed that head rotation induced a marked increase in the upper airway diameter in the neck in addition to an increase in the cross-sectional area in the retroglossal region. As reduction in the length of the pharyngeal tube is believed to cause airway obstruction,21 head rotation possibly leads to shift of the soft tissue out of the submandibular space and improves mask ventilation by the reduction of airway obstruction.
Strengths and limitations of this study
Instead of measuring tidal volume directly during mask ventilation, we chose RIP because of the potential for mask leak. With RIP, we could calculate the inspiratory and expiratory tidal volume regardless of mask leaks.18 We designed the study such that we could both determine and compare the tidal volumes generated during mask ventilation in two head positions for each patient. This made it possible for within patient comparisons, rather than between group comparisons. The latter would have required a much larger sample size. Moreover, as the PIP was kept at a constant level during mask ventilation, the VTE directly reflects airway patency and inversely reflects the degree of airway obstruction. Thus, even though the degree of airway obstruction during each head position was not measured directly, it could be estimated by comparing the VTE obtained at a constant PIP during mask ventilation with that obtained at the same PIP after tracheal intubation where airway obstruction was absent.
The crossover design of the study allowed us to minimise the time effect due to uncertainty of reaching a steady-state of anaesthesia and muscle relaxation during anaesthesia induction, and the sequence effect of two consecutive manoeuvres. When the mean VTE obtained during Step 1 and Step 3 were compared with that obtained during Step 2, the time effect was controlled. The randomised sequence of the intervention eliminated the sequence effect. Therefore, head position itself, theoretically, contributed to the VTE.
There are several limitations to this study. First, we studied only adults. An early study in naturally sleeping infants demonstrated no significant improvements in tidal breathing parameters by head rotation.22 Another study in anaesthetised and paralysed infants observed increased upper airway collapsibility with head rotation in the prone position.23 Given the structural differences between infant and adult airways, with different proportions of central mobile elements to their enclosing chambers,24 it is obviously impossible to apply our findings to all age groups. Second, we studied only paralysed patients, and so, the results of this study cannot be applied to nonparalysed or partially paralysed patients, even though head rotation improves upper airway patency in awake subjects.11–13 Third, we evaluated VTE in an optimised neck and jaw position during both head positions because this is the first step in airway management when we are confronted with difficult mask ventilation.25,26 Walsh et al.13 demonstrated a nonsignificant decrease in pharyngeal critical closing pressure during head rotation when care was taken to avoid flexion or extension. Thus, our findings cannot be applied to patients with difficulties in neck extension and jaw thrust. Fourth, we used a two-handed technique for mask ventilation and such is not a primary technique,27 but it is a method for difficult (grade 3) mask ventilation.3 We focused on the direct effect of head rotation on mask ventilation by achieving a complete seal between the face mask and the patient's face. Fifth, the fixed height of the head elevation may have a variable effect on upper airway collapsibility in each patient. However, we optimised neck and jaw position, thus we do not think that the head elevation caused airway obstruction due to neck flexion in smaller patients or patients with shorter necks. Finally, we only tested the effect of head rotation to the right but not to the left. Because airway obstruction for the majority of individuals is symmetric, we would not expect rotation in the opposite direction to alter our observation.
In this study, we observed the effect of head rotation on VTE was greatest in patients with airway obstruction and low VTE during the neutral head position (Fig. 4a and 4d). We originally defined these two subgroups because these populations are those most in need of some rescue manoeuvres during mask ventilation. As normal tidal volume is 5 to 7 ml kg−1 during spontaneous breathing at rest,28 we selected 6 ml kg−1 as the cut-off value for low VTE, although this value may not represent sufficient mask ventilation to the extent that adequate preoxygenation is obtained when difficult tracheal intubation is not suspected.25 Indeed, none of our patients with low VTE had a SpO2 less than 92% during the study period because they were all undergoing elective surgery and were well preoxygenated. As, difficult mask ventilation often coexists with difficult tracheal intubation,2,4 this population, in particular, requires sufficient ventilation to allow for multiple or prolonged attempts at intubation. Airway obstruction was not always accompanied by low VTE. Our finding indicates that head rotation has the potential to reduce peak airway pressure during mask ventilation and this is important as increased inflation pressures during mask ventilation may lead to gastric insufflation and aspiration.17
Although it is hard to know how many study patients have been classified into grade 3 mask ventilation (inadequate, unstable, or requiring two providers) as we always held the mask with two hands,3 the effect of head rotation was not associated with any predictors for grade 3 mask ventilation (BMI >30 kg m−2, presence of a beard, Mallampati classification III or IV, age ≥57 years, jaw protrusion–severely limited and snoring).2 On the contrary, we observed the increased VTE effect of head rotation mostly in younger patients and patients with Mallampati classification I. Therefore, we might only observe the beneficial effects of head rotation in patients with easier mask ventilation: for example, grade 1 mask ventilation (ventilated by mask only) or grade 2 mask ventilation (ventilated by mask and oral airway/adjuvant with or without muscle relaxant). This study suggests that head rotation is a readily available option to increase VTE when airway obstruction is encountered, especially in younger adults with Mallampati classification I. Future studies including obese patients may provide greater information as to which patients will benefit most from head rotation.
In conclusion, in anaesthetised, apnoeic adult patients, 45° head rotation significantly improved the efficiency of mask ventilation compared with the neutral head position. The effect was independently associated with younger age, Mallampati classification I and airway obstruction.
Acknowledgements relating to this article
Assistance with this study: the authors give special thanks to Timothy T. Houle from Massachusetts General Hospital for his assistance with statistical analysis.
Financial support and sponsorship: this work was supported by the Department of Anesthesia, Critical Care and Pain Medicine, Massachusetts General Hospital, Boston, Massachusetts, USA.
Conflicts of interest: RMK received research grants from Covidien and Venner Medical. RMK is a consultant for Covidien and Orange Medical. The other authors disclose no other conflicts of interest.
Presentation: preliminary data from this study were included as a poster presentation at the International Anesthesia Research Society (IARS) 2016 Annual Meeting and International Science Symposium, 21 to 24 May 2016, San Francisco.
1. Practice guidelines for management of the difficult airway: an updated report by the American Society of Anesthesiologists Task Force on Management of the Difficult Airway. Anesthesiology
2. Kheterpal S, Han R, Tremper KK, et al. Incidence and predictors of difficult and impossible mask ventilation. Anesthesiology
3. Han R, Tremper KK, Kheterpal S, O’Reilly M. Grading scale for mask ventilation. Anesthesiology
4. Kheterpal S, Martin L, Shanks AM, Tremper KK. Prediction and outcomes of impossible mask ventilation: a review of 50,000 anesthetics. Anesthesiology
5. Frerk C, Mitchell VS, McNarry AF, et al. Difficult Airway Society 2015 guidelines for management of unanticipated difficult intubation in adults. Br J Anaesth
6. Safar P, Escarraga LA, Chang F. Upper airway obstruction in the unconscious patient. J Appl Physiol
7. Walsh JH, Maddison KJ, Platt PR, et al. Influence of head extension, flexion, and rotation on collapsibility of the passive upper airway. Sleep
8. Ikeda H, Ayuse T, Oi K. The effects of head and body positioning on upper airway collapsibility in normal subjects who received midazolam sedation. J Clin Anesth
9. Isono S, Tanaka A, Nishino T. Lateral position decreases collapsibility of the passive pharynx in patients with obstructive sleep apnea. Anesthesiology
10. Liistro G, Stanescu D, Dooms G, et al. Head position modifies upper airway resistance in men. J Appl Physiol
11. Ono T, Otsuka R, Kuroda T, et al. Effects of head and body position on two- and three- dimensional configurations of the upper airway. J Dent Res
12. Zhang W, Song X, Masumi SI, et al. Effects of head and body positions on 2- and 3-dimensional configuration of the oropharynx with jaw protruded: a magnetic resonance imaging study. Oral Surg Oral Med Oral Pathol Oral Radiol Endod
13. Walsh JH, Leigh MS, Paduch A, et al. Effect of body posture on pharyngeal shape and size in adults with and without obstructive sleep apnea. Sleep
14. Safiruddin F, Koutsourelakis I, de Vries N. Analysis of the influence of head rotation during drug-induced sleep endoscopy in obstructive sleep apnea. Laryngoscope
15. Safiruddin F, Koutsourelakis I, de Vries N. Upper airway collapse during drug induced sleep endoscopy: head rotation in supine position compared with lateral head and trunk position. Eur Arch Otorhinolaryngol
16. Oto J, Li Q, Kimball WR, et al. Continuous positive airway pressure and ventilation are more effective with a nasal mask than a full face mask in unconscious subjects: a randomized controlled trial. Crit Care
17. von Goedecke A, Voelckel WG, Wenzel V, et al. Mechanical versus manual ventilation via a face mask during the induction of anesthesia: a prospective, randomized, crossover study. Anesth Analg
18. Banzett RB, Mahan ST, Garner DM, et al. A simple and reliable method to calibrate respiratory magnetometers and respitrace. J App Physiol
19. Hillman DR, Platt PR, Eastwood PR. The upper airway during anaesthesia. Br J Anaesth
20. Neill AM, Angus SM, Sajkov D, McEvoy D. Effects of sleep posture on upper airway stability in patients with obstructive sleep apnea. Am J Respir Crit Care Med
21. Isono S. Obesity and obstructive sleep apnoea: mechanisms for increased collapsibility of the passive pharyngeal airway. Respirology
22. Downs JA, Stocks J. Effect of neck rotation on the timing and pattern of infant tidal breathing. Pediatr Pulmonol
23. Ishikawa T, Isono S, Aiba J, et al. Prone position increases collapsibility of the passive pharynx in infants and small children. Am J Respir Crit Care Med
24. Sunder RA, Haile DT, Farrell PT, Sharma A. Pediatric airway management: current practices and future directions. Pediatr Anesth
25. El-Orbany M, Woehlck HJ. Difficult mask ventilation. Anesth Analg
26. Davies JD, Costa BK, Asciutto AJ. Approaches to manual ventilation. Respir Care
2014; 59:810–822. discussion 822-824.
27. Joffe AM, Hetzel S, Liew EC. A two-handed jaw-thrust technique is superior to the one-handed ‘EC-clamp’ technique for mask ventilation in the apneic unconscious person. Anesthesiology
28. Malhotra A. Low-tidal-volume ventilation in the acute respiratory distress syndrome. N Engl J Med