Glossopharyngeal Insufflation and Pulmonary Hemodynamics in Elite Breath Hold Divers : Medicine & Science in Sports & Exercise

Journal Logo


Glossopharyngeal Insufflation and Pulmonary Hemodynamics in Elite Breath Hold Divers


Author Information
Medicine & Science in Sports & Exercise 42(9):p 1688-1695, September 2010. | DOI: 10.1249/MSS.0b013e3181d85dc3

Glossopharyngeal insufflation (GI), a technique with repetitive compression of mouthfuls of air by the oropharyngeal muscles into the trachea (15), was first described in 1951 in patients with postpoliomyelitic syndrome (5). Application of this technique allowed patients with respiratory muscle insufficiency to increase vital capacity (VC) and thereby prolong the "off-respirator" time.

Breath hold divers (BHD) use the GI technique after maximal inspiration to total lung capacity (TLC) to increase lung volume beyond TLC (28). This increase in thoracic gas volume is needed for ear pressure equalization on descent and augments available oxygen stores, and this might protect the lungs from possible barotraumas caused by compression (8,16). Applying this technique allowed breath hold divers to compete more successfully for depth, distance, and breath hold time (1,8,17,21). Recently, it has also been reported that elite swimmers can increase their lung volume and buoyancy by using GI and possibly enhance performance (23).

Acute lung hyperinflation induced by GI leads to possibly deleterious changes in pulmonary volumes and mechanics (17) and elicits significant hemodynamic effects, including a decrease in systemic arterial blood pressure, elevation of HR, and acute biventricular systolic dysfunction (9,16,22). Furthermore, severe bradycardia with syncope (3), pneumomediastinum (10), and possibly excessive ventricular wall stress (24) causing cardiac arrhythmias were reported. However, there is still only scarce information on pulmonary hemodynamic changes during GI. In Swedish elite breath hold divers, emptying of the thoracic vessels was observed by static magnetic resonance imaging (MRI), but no quantification of these changes was available (16). An echocardiographic study by Scherhag et al. (27) proposed that long-term effects of breath hold diving might carry the risk of developing chronic pulmonary hypertension.

MR flow measurements using velocity-encoding cine (VEC) sequences are a well-established, noninvasive, and radiation-free imaging modality to assess the pulmonary hemodynamics (14,18,20).

The aim of this study was to investigate and to quantify the acute pulmonary hemodynamics during GI by VEC-MRI of the pulmonary artery (PA) in experienced elite breath hold divers, with concomitant measurement of lung volume changes by MR-compatible spirometry (MRc-spirometry) (6).



Nine male healthy elite breath hold divers (Table 1) participated and signed their informed consent after full explanation of the study. The study was conducted in conformity with the principles of the Declaration of Helsinki from 2000 and was approved by the local ethics committee.

Supine lung function, maximal GI volume in supine position, and training status/performance data of the divers.


MRc-spirometry was used to assess baseline spirometry in the supine position as described previously (6) and to simultaneously record respiratory volumes during the GI maneuver and MR flow measurements. Divers were equipped with a nose clip and the handheld of the MRc-spirometer. Initially, the maximum GI volume (GImax) for each diver was established. The divers were asked to breathe through the spirometer up to their maximal TLC while inspiratory VC (VCinGImax) was measured. Then they left the mouthpiece of the spirometer to perform a maximal GI maneuver, defined as the level where each diver reached his personal "sensation of fullness." At the end of the maneuver, the divers expired maximally into the MRc-spirometer to obtain expiratory VC (VCexGImax), thus enabling the calculation of GImax by subtracting VCinGImax from VCexGImax. All measurements were conducted in a reclined position.

For VEC-MR measurements, the breath hold time had to be approximately 2 min, the MR sequence time. To obtain the GI volume during MR flow measurement (GIVEC-MRI), the procedure started as described for GImax. After inspiration to TLC and recording of VC (VCinVEC-MRI), the divers left the mouthpiece, performed a submaximal GI maneuver, and held their breath at this level of hyperinflation for the entire time of the flow measurement. After the termination of the MR sequence, the divers expired into the MRc-spirometer to obtain expiratory VC (VCexVEC-MRI). GIVEC-MRI was calculated by subtracting VCinVEC-MRI from VCexVEC-MRI.

Study protocol.

Morphological sequences were obtained in all divers for the morphological assessment of the lung parenchyma (for MRI acquisition parameters and evaluation of VEC-MRI, see Supplemental Digital Content 1, The sequence of the study protocol is shown in Figure 1.

Flowchart showing the course of the study protocol.

Free breathing flow measurements were performed before (FB1) and subsequently after (FB2) the GI maneuver on the basis of mutiplanar fast imaging with steady precession (trueFISP) localizer acquired during tidal breathing to obtain the level of the main PA. At voluntary hyperinflation, MR measurements started first with an inspiratory localizer followed immediately by the VEC-MR measurement. GIVEC-MRI volume was recorded simultaneously as shown in previous paragraphs. Hemodynamic parameters were calculated from VEC-MRI (for details, see Supplemental Digital Content 1,, and these are listed in Table 2. All flow and volume measurements were performed in a reclined position.

Parameters derived from VEC-MRI at baseline, during GI, and subsequently after GI.

Statistical analysis.

Statistical analysis was performed by a statistician. P < 0.05 was considered statistically significant. Because VEC-MRI parameters were measured repeatedly under different conditions, changes in these parameters were evaluated by calculating the mean differences with their respective 95% confidence intervals (CI). Because of the small number of subjects, we did not rely on the assumption of normal distribution and used the signed rank test to test whether the differences between parameters under different conditions were different from zero. Spearman rank correlation coefficients (29) were calculated between the GI volume during flow measurement (GIVEC-MRI) and VEC-MRI parameters during GI.


MRI found no morphological pathologies in the diver's lungs. All divers performed the GI maneuvers and flow measurements successfully without clinical problems. For technical reasons, VEC-MRI could not be evaluated in one diver during FB1. The GI volume during flow measurements (GIVEC-MRI) achieved a mean of 95% of GImax (SD = 35%, median = 92%, P < 0.003).

Comparison of VEC-MRI parameters between baseline (FB1) and GI.

Hemodynamic VEC-MRI parameters showed significant changes during GI compared with those during baseline (Tables 2 and 3). The mean flow decreased by 45% (P < 0.007; Fig. 2A), and the right ventricular output (flow per minute) decreased by 41% (P < 0.007). Forward volume decreased by 43% (P < 0.02). The right ventricular cardiac index decreased by 40% (P < 0.007; Fig. 2B). Mean and peak velocity both decreased by 42% and 37%, respectively (P < 0.007; Fig. 2C). Maximum area and distensibility of the PA decreased from 8.1 to 6.8 cm2 (−14%, P < 0.02) and from 34% to 15% (−54%, P < 0.007; Fig. 2D), respectively.

Mean differences (FB1−GI) and ratios (GI/FB1) as well as SD, 95% CI, and P values for VEC-MRI parameters between GI and baseline.
VEC-MRI parameters in each diver at baseline, during GI, and at subsequent free breathing (A-F).

Acceleration time decreased significantly by 36% during GI (P < 0.007; Fig. 2E). Mean HR increased by 13% during GI, but the difference from baseline was not statistically significant (Fig. 2F). The individual values of the divers show that five divers had a pronounced increase in HR of 72% and that four divers had a nearly constant or even decreasing HR during GI.

Comparison of VEC-MRI parameters between GI and FB2.

Subsequently, after GI, the majority of the parameters returned to baseline, and no statistical difference could be obtained between free breathing before and after GI. However, acceleration time remained significantly low (P < 0.03) for the difference and the ratio between baseline and FB2.

Visual inspection of the PA flow curves identified two profiles: seven of the divers had an abolished diastolic flow in the PA (Fig. 3A), whereas two divers showed a nearly preserved diastolic flow in the PA (Fig. 3B). The diastolic recoil, however, was less pronounced even in those divers with a similar flow profile during GI and free breathing. Furthermore, all except one diver showed an end-diastolic dip of the flow and velocity curve during GI that was not present during free breathing.

Characteristic flow profile and pulse wave velocity of the PA during baseline free breathing and GI in two divers. A, Narrowing of the flow curve during GI is observed besides a decrease in velocity and flow. B, Narrowing of the flow curve and a preserved early diastolic shape of the flow and velocity curve during GI are observed besides a decrease in velocity and flow, although the diastolic recoil is less pronounced compared with free breathing. The end-diastolic dip of the flow and velocity curve not present at free breathing was found in eight of nine divers during GI.

Correlation to GI volume.

No statistically significant correlation was found between GIVEC-MRI volume and the calculated VEC-MRI parameters. Spearman correlation coefficients were −0.61 for mean velocity, −0.58 for peak velocity, −0.51 for mean flow, −0.56 for forward volume, −0.55 for the right ventricular output, and −0.51 for acceleration time. Spearman correlation coefficients for HR, distensibility, and PA area with GIVEC-MRI volume were <0.4. Divers with increasing HR during GI had a mean GIVEC-MRI volume of 2.1 L, and those with unchanged or decreasing HR had a GIVEC-MRI volume of 1.2 L.

Scatter plots showing the association between GIVEC-MRI volume and the respective VEC-MRI parameters are shown in Figure 4. A significant difference was found between the divers with a GIVEC-MRI > 2 L and those with GIVEC-MRI < 2 L for the VEC-MRI parameters: acceleration time (P < 0.02), forward volume (P < 0.02), right ventricular cardiac output (P < 0.02), mean flow (P < 0.04), and mean velocity (P < 0.04).

Scatter plots showing the association between GI volume during flow measurements (GIVEC-MRI) and VEC-MRI parameters. VEC-MRI parameters differ significantly between divers with GIVEC-MRI < 2 L and those with GIVEC-MRI > 2 L. Numbers indicate single subjects.


For the first time, this study characterized the effects of acute voluntary lung hyperinflation on pulmonary hemodynamics in elite experienced breath hold divers using VEC-MRI. Although at baseline normal hemodynamic values were found in the healthy elite breath hold divers, acute impairment of pulmonary circulation similar to pulmonary arterial hypertension (PAH) could be detected during the hyperinflation maneuver. Indices of pulmonary hemodynamics, such as right ventricular output, cardiac index, forward volume, and pulmonary blood flow, all decreased substantially and were highly significant during GI (all P > 0.01) from baseline values.

A decrease in the maximal area of the PA indicates the volume reduction in the pulmonary circulation during GI, being the result of right ventricular depletion as demonstrated by transthoracic echocardiography. In a study by Potkin et al. (24), GI induced acute biventricular systolic dysfunction consistent with an acute pressure overload. The findings suggest that the acutely increased alveolar pressure during GI is transmitted to the pulmonary vasculature, thereby eliciting a right ventricular pressure overload and left ventricular dysfunction after both impaired left ventricular filling and decreased contractility. From another study of four breath hold divers using finger photoplethysmography, rapid decreases in mean arterial pressure and pulse pressure during GI were reported (22). Calculated cardiac output and stroke volume both decreased by more than 80%. Scherhag et al. (27) proposed in an echocardiographic study that repetitive breath hold diving might carry the risk of developing manifest pulmonary hypertension within 2-3 yr. In our clinical and morphological studies during free breathing in these well-experienced elite subjects, we did not find evidence of any manifest changes that support the assumption of Scherhag et al.

The present MRI study of pulmonary hemodynamics during GI thus complements the hypothesis that GI induces an acute right ventricular pressure overload. A decrease in acceleration time, flow, and velocity as obtained by the present MRI study during acute lung hyperinflation by GI is characteristic for the findings seen in PAH (2,12,25,26). Mean velocity correlates best with invasively measured mean PA pressure (26). Mean velocity in the presented diver group was 10.5 cm·s−1 during GI, which is lower than the normal low cutoff value of 11.7 cm·s−1 found in the study by Sanz et al. (26) in patients with PAH. The flow and velocity profiles of the PA during GI resemble the one found in patients with PAH characterized by a loss of elasticity (32). The end-diastolic dip of the flow profile might also be related to a suspected loss of diastolic recoil as a result of the decreased compliance and distensibility of PA during GI (32). Nevertheless, the PA area behaves opposite during GI than in patients with PAH, where it is significantly increased, revealing the major pathophysiological difference between the two conditions. Although PAH is a pathological finding with pulmonary vasoconstriction and vascular remodeling resulting in high afterload for the right ventricle, GI causes an increase in right ventricular afterload because of the high alveolar pressure and collapsing arteries.

Normalization of the majority of the parameters shortly after the maneuver suggests that the described hemodynamic effects are transient and fully reversible shortly after cessation of voluntary hyperinflation, like diaphragm and thoracic wall deformation (7). Acceleration time, however, remained significantly low compared with baseline. This indicates that its maximum is only reached later in time, although the flow in the PA normalized immediately after GI. It is known from echocardiographic studies that acceleration time is a surrogate parameter of PA compliance and one of the initial indicators for pulmonary hypertension (4,19). The decreased acceleration time across the duration of GI might indicate a time lag in the recovery between flow and compliance of the PA, although mean distensibility, as defined and measured in this study, recovered immediately after GI (13). A transient alteration of lung recoil pressure at high lung volumes has been reported recently (28). Accordingly, static lung compliance remained elevated for about 3 min even when the subjects had terminated GI and returned to baseline tidal breathing (30). Whether the decrease in distensibility may result from a temporal loss of wall elasticity of the PA, consecutively to a substantial rise in transpulmonary pressure, remains speculative.

Mean HR during GI as measured in this study increased by 13%. Looking at the individual divers (Fig. 2F), it becomes evident that this number is misleading because five divers showed a pronounced increase in HR of 70%, which is in concordance with previous studies (3,22,24) but is counterbalanced in the mean value of four divers showing a nearly constant or even decreasing HR during GI. This difference may be explained by the different level of voluntary lung hyperinflation. Divers with increasing HR had a higher mean GIVEC-MRI. The degree of hypotension resulting from rising transpulmonary pressure may cause an HR response via sympathetic baroreceptor activation similar to a Valsalva maneuver (11).

The association found between hemodynamic indices and acute voluntary lung hyperinflation during GI supports the view that the cardiodepressive effects are related to the degree of increased VC. The correlation was significant only for forward volume/body surface area (BSA), but a trend (P < 0.1) was obvious for mean and peak velocity, too. However, breath hold divers with a GIVEC-MRI volume ≥ 2 L differed significantly concerning their VEC-MRI parameters to those with a GIVEC-MRI < 2 L. Interestingly, two of the three divers with higher GI volumes showed a preserved early diastolic profile of the flow and velocity curve in the PA. The third diver with a GIVEC-MRI of 2 L who showed a flow and velocity curve with abolished diastolic recoil was more than 10 yr older, with an age of 43 yr. Speculatively, this may indicate that age, and thus altered distensibility of the pulmonary vasculature, may account for a difference in diastolic recoil (31). It is a limitation of our study that we did not measure pressures during GI maneuvers, and therefore, we can only assume that these effects are due to an increased intrathoracic pressure reflected by GI volumes. This should be the subject of future studies.

This study did not explicitly investigate the possible long-term effects of GI on the cardiovascular system; however, no persistent pathological morphological changes were detected in free breathing measurements, and most hemodynamic indices returned to baseline values after cessation of voluntary lung hyperinflation. It may be argued, however, that persistent changes should have been detectable in a sample of experienced BHD using GI since almost 6 yr on average. We cannot rule out that long-term use of GI with increasing age might induce changes to the pulmonary vasculature, affecting the diastolic recoil, as seen in our eldest subject. The effect of increasing age and subsequent damage to the pulmonary vascular system and the effect on the right side of the heart should be the subject of future studies.

In conclusion, VEC-MRI measurements of the PA showed a substantial decrease in pulmonary hemodynamic parameters during GI, which tends to be associated with the increase in lung volume. The hemodynamic changes mimic the changes seen in PAH but, unlike the latter, are mostly reversible after the cessation of voluntary lung hyperinflation except for acceleration time in the PA. Long-term effects on the cardiopulmonary system with increasing age cannot be ruled out for sure and ought to be subject for further studies. In younger subjects, repeated performance of the GI maneuver is not associated with lasting effects on the pulmonary cardiovascular system.

There was no special funding supporting this study. All authors declare to have no conflict of interest regarding this study.

The authors thank all the divers for their participation.

The results of the present study do not constitute endorsement by the American College of Sports Medicine.


1. AIDA. Worldwide Federation for breath-hold diving [Internet]. Lausanne (Switzerland): AIDA; [cited 2009 Jan 4]. Available from:
2. Abolmaali N, Seitz U, Esmaeili A, et al. Evaluation of a resistance-based model for the quantification of pulmonary arterial hypertension using MR flow measurements. J Magn Reson Imaging. 2007;26(3):646-53.
3. Andersson J, Schagatay E, Gustafsson P, Örnhagen H. Cardiovascular effects of "buccal pumping" in breath-hold divers. In: Gennser M, editor. XXIV Annual Scientific Meeting of the European Underwater and Baromedical Society. Stockholm (Sweden): National Defence Research Establishment; 1998. p. 103-6.
4. Dabestani A, Mahan G, Gardin JM, et al. Evaluation of pulmonary artery pressure and resistance by pulsed Doppler echocardiography. Am J Cardiol. 1987;59(6):662-8.
5. Dail CW. 'Glossopharyngeal breathing'; by paralyzed patients; a preliminary report. Calif Med. 1951;75(3):217-8.
6. Eichinger M, Puderbach M, Smith H, et al. Magnetic resonance-compatible-spirometry: principle, technical evaluation and application. Eur Respir J. 2007;30(5):972-9.
7. Eichinger M, Walterspacher S, Scholz T, et al. Lung hyperinflation: foe or friend? Eur Respir J. 2008;32(4):1113-6.
8. Ferrigno M, Lundgren CEG. Human breath-hold diving. In: Lundgren CEG, editor. The lung at Depth. New York (NY): Dekker; 1999. p. 529-85.
9. Fitz-Clarke JR. Computer simulation of human breath-hold diving: cardiovascular adjustments. Eur J Appl Physiol. 2007;100(2):207-24.
10. Jacobson FL, Loring SH, Ferrigno M. Pneumomediastinum after lung packing. Undersea Hyperb Med. 2006;33(5):313-6.
11. Junqueira LFJ. Teaching cardiac autonomic function dynamics employing the Valsalva (Valsalva-Weber) maneuver. Adv Physiol Educ. 2008;32(1):100-6.
12. Laffon E, Laurent F, Bernard V, De Boucaud L, Ducassou D, Marthan R. Noninvasive assessment of pulmonary arterial hypertension by MR phase-mapping method. J Appl Physiol. 2001;90(6):2197-202.
13. Ley S, Fink C, Puderbach M, et al. MRI measurement of the hemodynamics of the pulmonary and systemic arterial circulation: influence of breathing maneuvers. AJR Am J Roentgenol. 2006;187(2):439-44.
14. Ley S, Unterhinninghofen R, Ley-Zaporozhan J, Schenk J, Kauczor H, Szabo G. Validation of magnetic resonance phase-contrast flow measurements in the main pulmonary artery and aorta using perivascular ultrasound in a large animal model. Invest Radiol. 2008;43(6):421-6.
15. Lindholm P, Norris CMJ, Braver JM, Jacobson F, Ferrigno M. A fluoroscopic and laryngoscopic study of glossopharyngeal insufflation and exsufflation. Respir Physiol Neurobiol. 2009;167(2):189-94.
16. Lindholm P, Nyren S. Studies on inspiratory and expiratory glossopharyngeal breathing in breath-hold divers employing magnetic resonance imaging and spirometry. Eur J Appl Physiol. 2005;94(5-6):646-51.
17. Loring SH, O'Donnell CR, Butler JP, Lindholm P, Jacobson F, Ferrigno M. Transpulmonary pressures and lung mechanics with glossopharyngeal insufflation and exsufflation beyond normal lung volumes in competitive breath-hold divers. J Appl Physiol. 2007;102(3):841-6.
18. Lotz J, Meier C, Leppert A, Galanski M. Cardiovascular flow measurement with phase-contrast MR imaging: basic facts and implementation. Radiographics. 2002;22(3):651-71.
19. Maeba H, Nakatani S, Sugawara M, et al. Different time course of changes in tricuspid regurgitant pressure gradient and pulmonary artery flow acceleration after pulmonary thromboendarterectomy: implications for discordant recovery of pulmonary artery pressure and compliance. Circ J. 2007;71(11):1771-5.
20. Mousseaux E, Tasu JP, Jolivet O, Simonneau G, Bittoun J, Gaux JC. Pulmonary arterial resistance: noninvasive measurement with indexes of pulmonary flow estimated at velocity-encoded MR imaging-preliminary experience. Radiology. 1999;212(3):896-902.
21. Muth C, Ehrmann U, Radermacher P. Physiological and clinical aspects of apnea diving. Clin Chest Med. 2005;26(3):381-94, v.
22. Novalija J, Lindholm P, Loring SH, Diaz E, Fox JA, Ferrigno M. Cardiovascular aspects of glossopharyngeal insufflation and exsufflation. Undersea Hyperb Med. 2007;34(6):415-23.
23. Nygren-Bonnier M, Gullstrand L, Klefbeck B, Lindholm P. Effects of glossopharyngeal pistoning for lung insufflation in elite swimmers. Med Sci Sports Exerc. 2007;39(5):836-41.
24. Potkin R, Cheng V, Siegel R. Effects of glossopharyngeal insufflation on cardiac function: an echocardiographic study in elite breath-hold divers. J Appl Physiol. 2007;103(3):823-7.
25. Roeleveld RJ, Marcus JT, Boonstra A, et al. A comparison of noninvasive MRI-based methods of estimating pulmonary artery pressure in pulmonary hypertension. J Magn Reson Imaging. 2005;22(1):67-72.
26. Sanz J, Kuschnir P, Rius T, et al. Pulmonary arterial hypertension: noninvasive detection with phase-contrast MR imaging. Radiology. 2007;243(1):70-9.
27. Scherhag A, Pfleger S, Grosselfinger R, Borggrefe M. Does competitive apnea diving have a long-term risk? Cardiopulmonary findings in breath-hold divers. Clin J Sport Med. 2005;15(2):95-7.
28. Seccombe LM, Rogers PG, Mai N, Wong CK, Kritharides L, Jenkins CR. Features of glossopharyngeal breathing in breath-hold divers. J Appl Physiol. 2006;101(3):799-801.
29. Spearman C. The proof and measurement of association between two things. By C. Spearman, 1904. Am J Psychol. 1987;100(3-4):441-71.
30. Tetzlaff K, Scholz T, Walterspacher S, et al. Characteristics of the respiratory mechanical and muscle function of competitive breath-hold divers. Eur J Appl Physiol. 2008;4(103):469-75.
31. Warnock ML, Kunzmann A. Changes with age in muscular pulmonary arteries. Arch Pathol Lab Med. 1977;101(4):175-9.
32. Zuckerman BD, Orton EC, Stenmark KR, et al. Alteration of the pulsatile load in the high-altitude calf model of pulmonary hypertension. J Appl Physiol. 1991;70(2):859-68.


Supplemental Digital Content

©2010The American College of Sports Medicine