Journal Logo

Basic Sciences: Original Investigations

Arterial Oxygen Desaturation Kinetics during Apnea


Author Information
Medicine & Science in Sports & Exercise: November 2005 - Volume 37 - Issue 11 - p 1871-1876
doi: 10.1249/01.mss.0000176305.51360.7e
  • Free


The response to apnea during underwater diving in humans is believed to be a defensive reflex designed to reduce peripheral oxygen consumption, while maintaining oxygenation to the heart and brain. The cardiovascular changes associated with the diving response are well established (1–6,8,9,11,15,16,22–29,33) and include bradycardia, decreased cardiac output, increased arterial blood pressure, redistribution of peripheral blood flow, and splenic release of erythrocytes. Although these cardiovascular responses have recently been shown to be effective in conserving oxygen in humans during apneic exercise (1,2,22,23), their effectiveness during apnea at rest has still to be proved (3,24,33).

Analysis of pre- and postapneic expired gas and correlations between the magnitude of the associated cardiovascular changes and apnea time have been used to associate an oxygen-conserving role of the human diving response to apnea at rest (24) and during exercise (1,23). Neither approach, however, has provided quantification of oxygen consumption kinetics during the apneic period. The ability to measure whole body oxygen consumption kinetics during apnea would enable accurate description of the oxygen-conserving potential of the diving response; however, indirect calorimetry requires the subject to be breathing, and direct calorimetry does not have a sufficient sampling frequency to measure the change.

The rate of arterial oxygen desaturation has been used to reflect the rate of oxygen consumption kinetics during apnea (2,23). A smaller reduction in arterial oxygen saturation (2) and the correlation of the rate of arterial oxygen desaturation with cardiovascular changes (23) have been associated with an oxygen-conserving potential of the diving response during apnea in an exercising human. No quantifiable evidence, however, has been produced of an attenuated reduction in arterial oxygen desaturation during resting face immersion apnea. Studies have reported an oxygen-conserving effect, with reductions in arterial oxygen saturation of less than 4% (3,6,33). The association of a biologically significant oxygen-conserving effect with minimal oxygen reduction is debatable.

The lack of evidence for an oxygen-conserving effect during resting face immersion apnea may result from the inability of untrained humans to prolong apnea long enough for an effect to be observed. The aim of this investigation was to study the diving response and the rate of arterial oxygen desaturation during face immersion apnea in divers capable of extended apnea durations (i.e., >4 min). We hypothesized that these elite apnea divers would display an augmented diving response compared with untrained subjects, which would result in an attenuation of the rate of arterial oxygen desaturation.



Ten trained freedivers (17 yr average training with experience in national and international competition) and ten controls received a verbal description of the experiment and were required to complete a written informed consent, which was approved by the university’s human research ethics committee.


Following completion of the informed consent, subjects had an intravenous catheter inserted into a prominent vein within the antecubital fossa of their nondominant arm. Anthropometric measures and biographical information were recorded before pulmonary function was assessed. The subjects then assumed a prone position on a plinth with their head resting on the cover of a container filled with cold water (10°C). Following a 15-min rest, baseline levels of all hematological and cardiovascular variables were determined. Subjects were then instructed to undertake their normal predive breathing routine before face immersion. The trained freedivers all used a series of hyperventilation maneuvers before apnea designed to reduce carbon dioxide levels in the body and thus postpone the apnea breaking point (8). Normalization of the predive breathing was not undertaken because the ritual is idiosyncratic; however, all subjects were instructed to undertake the same procedure before each apnea. Five maximal apneas, with face immersion, were separated by 2-min recovery intervals as previously described (27).

Pulmonary function.

Forced (FVC) and slow vital capacity (SVC) maneuvers and maximal voluntary ventilation were determined using a Medical Graphics CPX-D Metabolic cart with BREEZE Ex v3.06 pulmonary function software (Medical graphics, St. Paul, MN).

Cardiovascular measurements.

Arterial oxygen saturation (SaO2) was assessed by pulse oximetry (RAD 9, Masimo, Irvine, CA). A topical vasodilator cream (Finalgon, Boehringer Mannheim, Australia) was applied to the ear lobe to increase local perfusion before placement of the pulse oximeter sensor. Blood pressure was noninvasively measured by a pressure sensor (NIBP100A, SDR Clinical Technology, NSW, Australia) placed distal to the styloid process over the radial artery. Heart rate was measured by three-lead electrocardiography (ECG) (Q4500, Quinton, Seattle, WA). Calf blood flow (CBF) was assessed immediately before and at 30-s intervals during each apnea by venous occlusion strain gauge plethysmography (Hokanson, Bellevue, WA). SaO2, blood pressure, ECG, and CBF were recorded at 200 Hz per channel on a data-collection system (MP100A-CE, Biopac Systems, Goleta, CA). Calibrated analog signals were converted to digital, stored, and analyzed with an AcqKowledge 3.5.7 software package (Biopac Systems, Goleta, CA).

SaO2 analysis.

SaO2 data from each apnea was individually fitted to a sigmoid function using MATLAB (Mathworks Inc., Natick, MA). The sigmoid function [SaO2(t)] was fitted using a nonlinear, least-squares regression, as shown in the following equation: SaO2(t) = d + a/(1 + eb(t − c)). Where a represents the difference between the minimal and maximal SaO2 values; b is an indicator of the steepness of the slope; c is the time point at the maximal rate of change in SaO2; and d is the minimal SaO2 value. The maximal slope was calculated from the first derivative of the logistic function (19), using the following equation: maximum slope (%·s−1) = −a·b/4. The SaO2 data were also numerically analyzed with a continuous two-slope fit using MATLAB (Mathworks Inc., Natick, MA), to determine the line of best fit for the slope of the desaturation (ΔSaO2/Δt).

Hematological measurements.

Hemoglobin (Hb), hematocrit (Hct), and lactate were measured in triplicate at baseline and immediately following apneas 1, 3, and 5. A 7-mL blood sample was collected into sterile vacutainer tubes containing EDTA, at baseline and within 10 s of the completion of the apneas, and stored at 4°C until analysis (within 24 h). Hct was assessed with a microhematocrit centrifuge and corrected for 4% trapped plasma within the packed red cells. Hb was measured by the cyanmethemoglobin method (7) and lactate by the ultraviolet endpoint method using the spectrophotometric assay procedure (12).

Statistical analysis.

All variables are described as means ± standard deviations. Repeated measures factorial analyses of variance (ANOVA) were used to compare the apnea duration and hematological and cardiovascular variables. When significant F-ratios were observed, Scheffé’s tests were applied post hoc to determine where the differences occurred. Anthropometric, pulmonary, and curve-fitting coefficients were compared with paired samples t-test. A multiple regression analysis (utilizing the backward elimination variable selection method) was used to analyze the correlation between the ΔSaO2/Δt and apnea-induced changes in heart rate, mean arterial pressure (MAP), and calf blood flow (CBF), and preapnea slow vital capacity and Hb concentration. The level of significance was set at P < 0.05 for all ANOVA and correlation procedures.


The trained freedivers and the group of untrained controls did not differ significantly in age, height, or weight (Table 1). Baseline respiratory function revealed significantly greater vital capacities for the trained freedivers, but no difference in the dynamic measures of forced expired volume in 1 s or the maximal voluntary ventilation (Table 1).

Anthropometric and pulmonary function information for the freedivers (trained) and the control subjects (untrained).

Apnea duration increased significantly with each successive apnea for both groups, except between the third and fourth apnea (Table 2). The freedivers had significantly longer maximal apnea durations overall compared with the control group (trained 246 ± 44.0 s, untrained 129 ± 39.4 s, P < 0.001). Minimal heart rate obtained during sinus rhythm, minimal CBF, and maximal MAP recorded during each apnea is presented in Table 2. The change in heart rate (ΔHR) induced by the apnea was significantly different between the groups (trained −27.2 ± 9.5 bpm, untrained −19.7 ± 9.3 bpm, P < 0.001), as was the change in MAP (ΔMAP) (trained 48 ± 20.7 mm Hg, untrained 37 ± 10.0 mm Hg, P = 0.002). The bradycardia was also associated with an incidence of dysrhythmic beats in four of the trained freedivers during maximal apneas.

Apnea duration and cardiovascular responses with repeated apnea.

Analysis of the blood results indicated no significant difference between groups for Hb (trained 14.9 ± 1.2 gm·dL−1, untrained 14.5 ± 1.4 gm·dL−1, P = 0.26), or Hct (trained 45.1 ± 1.8, untrained 41.3 ± 1.5, P = 0.24); however, a significant increase in both variables was noted with repeated apneas (Table 3). Blood lactate was also not significantly different between the groups (trained 1.34 ± 0.40 mmol·L−1, untrained 1.19 ± 0.30 mmol·L−1, P = 0.11) (Table 3). When the results were normalized for apnea duration, however, a trend was noted toward the untrained control group accumulating more lactate (trained 0.02 ± 0.04 mmol·L−1·min−1, untrained 0.12 ± 0.05, mmol·L−1·min−1, P = 0.08).

Hematologic responses with repeated apnea.

The SaO2 nadir achieved during apnea was significantly different between the groups (trained 78 ± 9.3 % (range 61–93%), untrained 90 ± 5.7 % (range 75–96%), P < 0.001). The numerical treatment of the SaO2 data produced a sigmoid curve with an error, relative to the fitted data, across all apneas of 0.36 ± 0.10% and a continuous two-slope function, where the second slope had a correlation of 0.98 ± 0.02 with the absolute data. The sigmoidal curve fitting of the SaO2 data resulted in significant group effects for all coefficients, but the maximal slope was not significantly different between the freedivers and control subjects (Table 4).

Coefficients obtained from numerically fitting the arterial oxygen saturation data with a four-parameter sigmoidal logistic function.

The continuous two-slope data fit resulted in the freedivers having a longer initial lag phase (trained 65 ± 18.4 s, untrained 42 ± 13.3 s, P < 0.001) and the average lag phase across the five repeated apneas was related to each diver’s forced vital capacity (FVC r2 = 0.18, P = 0.045; SVC r2 = 0.20, P = 0.034). The second line of the two-slope data-fitting method represents the slope of the arterial desaturation (ΔSaO2/Δt) (13,14,20,23,30). A significant group effect was noted for the ΔSaO2/Δt (trained = −0.14 ± 0.04 %·s−1, untrained = −0.11 ± 0.05%·s−1, P = 0.015); however, this straight line fitting underestimates the rate of decline in untrained subjects because the ΔSaO2/Δt never reaches a consistent rate of change as observed in trained apnea divers. Multiple regression analysis to determine independent predictors of ΔSaO2/Δt across repeated apneas was therefore performed only on the freedivers. Univariate correlates of ΔSaO2/Δt (ΔHR, ΔMAP, Δblood flow, SVC, and [Hb]) were entered into the model. The ΔMAP and Δblood flow did not significantly contribute to the ΔSaO2/Δt and were removed from the model. The ΔHR, SVC, and [Hb] accounted for approximately 85% of the variation in the ΔSaO2/Δt (ΔSaO2/Δt = 0.513 * [Hb] – 0.504 * SVC − 0.96 * ΔHR, adjusted r2 = 0.85, P = 0.014).


The ability to quantify the rate of arterial desaturation (ΔSaO2/Δt) as an indirect marker of oxygen consumption in a closed system has enabled an oxygen-conserving potential to be attributed to the diving response in humans during exercise (1,2,22,23). Previous attempts at attributing an oxygen conservation effect to the diving response in resting humans have been difficult because of the limited breath-hold capabilities of untrained subjects. The application of a four-parameter logistic equation to the apnea-induced arterial desaturation in the current study enabled the comparison of apneas of different durations and SaO2 nadirs, and indicated that the ΔSaO2/Δt did not differ between the trained freedivers and untrained control subjects, despite apnea durations twice as long. Within the trained freedivers, however, those with the largest bradycardia and initial oxygen stores were able to slow the ΔSaO2/Δt two to three times that of the least marked response, reflecting an oxygen-conserving potential of the diving response in trained freedivers during resting face immersion apnea.

The maximal duration of the apnea reported using a similar protocol in members of the Croatian national apnea diving team (5), on average, were a minute less than in the present study (Table 2). Similar apnea durations have been achieved by trained divers (180–440 s, N = 9) while seated (17) or by elite freedivers (240–300 s, N = 3) in the supine position (10) during single apnea attempts. The maximal duration of the apnea achieved by the current freedivers (182–305 s) in the prone position, which restricts thoracic expansion, entitles them to be classified as elite. Indeed, our subjects included previous freediving world record holders.

Apnea duration has previously been subdivided into its easy-going and struggle phases (21,25). The phases are differentiated by monitoring thoracic movements as an indication of the initiation of involuntary diaphragmatic contractions (25). The contractions occur because of an elevation in arterial carbon dioxide content, which is directly related to the oxygen utilization (V̇O2) during the apnea (21). The primary interest in the present study was a quantification of the ΔSaO2/Δt as an indirect marker of V̇O2 in a closed system (2,23). As V̇O2 is not suspended with the initiation of the struggle phase and the increased oxygen cost of the involuntary diaphragmatic contractions have been shown not to affect the rate of arterial desaturation (30), the apneas were not separated into the two phases.

The ΔSaO2/Δt has previously been calculated by manually drawing a line adjacent to the falling portion of the SaO2 curve (13,20), or by drawing a line between two prechosen time points, irrespective of intermediary data points (23). These methods are appropriate if a significant linear falling portion of the SaO2 curve is observable. This straight line fitting, however, underestimates the rate of decline in untrained subjects because the ΔSaO2/Δt never reaches a consistent rate of change (i.e., linear slope) as occurs in trained freedivers. Indeed, it has been identified that a curvilinear equation would provide a more accurate fit of the SaO2 curve (14). To our knowledge this is the first study to fit a logistic function to the SaO2 data, thus enabling the calculation of the maximal slope of the curve and the comparison of apneas with different SaO2 nadirs.

Comparison of the SaO2 kinetics between trained divers and untrained control subjects during maximal face immersion apneas has been limited to the work of Ferretti et al. (10), who indicated that at any given time point, the SaO2 tended to be higher in divers. No statistical comparison was attempted because the diving group was limited to three subjects. In the current study, the freedivers had a significantly longer initial plateau period before any arterial desaturation (Fig. 1) and a longer time until the maximal rate of arterial desaturation (coefficient c, Table 3), which supports the observations of Ferretti et al. (10) and reflects the freedivers’ larger lung volumes (Table 1) and greater relative filling preapnea; however, no significant difference was seen in the maximal slope of the desaturation curves between the groups (Table 4).

FIGURE 1— Arterial oxygen saturation (SaO2) curve as numerically derived from the four-parameter sigmoidal logistic function, averaged across all apneas, for both trained and untrained divers. Curves are the mean response for the groups.

The diving response is characterized by bradycardia (Table 2). The bradycardia was also associated with an incidence of dysrhythmic beats, primarily premature ventricular contractions, noticeable during the final few seconds of the extended apneas (>290 s). These dysrhythmic beats are common during apnea (4,11,23) and may be responsible for preventing syncope if all premature beats are able to maintain cardiac output (9). This study showed that the bradycardia in combination with the preapnea measurements of vital capacity and Hb explained 85% of the individual variance in the (freedivers) ΔSaO2/Δt during face immersion apnea. In comparison, others have reported bradycardia to explain 58% of the ΔSaO2/Δt during apnea in combination with exercise, and the addition of the change in MAP increased the variance explanation to 88% (23). The inclusion of apnea-induced changes in MAP and CBF in the present model, however, did not improve the regression coefficient. The difference between the studies may be explained by the magnitude of the peripheral vasoconstriction induced by apnea that results in the changes to CBF and MAP. During exercise, the lower limb vasculature would be vasodilated, enabling a larger magnitude of change, whereas at rest, the predominant vasoconstrictor tone would minimize the magnitude of change, initiated by the face immersion apnea.

The increase in Hct and Hb observed in our trained freedivers (4.7 and 4.9%, respectively) were lower than those observed in Korean (ama) divers (10.5 and 9.5%, respectively) (16). The differences observed could be indicative of the differing experimental protocols used. The ama were studied under a normal diving shift of approximately 3-h duration involving repetitive depth dives, whereas the five face immersion apneas in the current study were completed within 45 min. The increased energy expenditure of depth dives (8) over a longer time period would result in an increased plasma volume loss because of the maintained elevations in arterial blood pressure. The arterial blood pressures obtained in a number of the freedivers in this present study (∼260/160 mm Hg) reached levels previously not seen during face immersion (5,26,28), but were of a similar magnitude to those observed during depth dives in three elite freedivers (11). The elevated blood pressure could cause hemoconcentration as a result of increased filtration of fluid from the vascular compartment. The hemoconcentration observed in both trained and untrained subjects (Table 3) could also be partially explained by contraction of the spleen releasing stored erythrocytes. The splenic contraction hypothesis has been supported indirectly by increases in Hct and Hb observed in subjects with spleens, but not in those without (25), and directly by ultrasonographic measurement of the reduction in spleen volume with apnea (5,16,26). The contraction of the human spleen, mediated through stimulation of α-adrenoreceptors (29), increases the circulating erythrocyte mass gradually in the initial apnea and maintains this enhanced oxygen-carrying capability throughout the repeated apneas (5,25). The enhanced oxygen-carrying capability, and increased carbon dioxide-buffering potential of the increased erthyrocyte mass (26), could contribute to the extended apnea duration with successive attempts (Table 2).

The reduction in peripheral blood flow associated with the diving response (Table 2), accentuated by cold water face immersion (28), requires the peripheral tissue to rely on anaerobic metabolism and, theoretically, should result in significant elevations in postapnea blood lactate measurements. The attenuated blood lactate responses (Table 3) relative to apnea duration of the trained divers, however, are reflective of changes observed in divers (18) and in controls following 3 months of apnea training (17). The reduced blood lactate accumulation with extended apneas has been explained by a reduction in the availability of blood glucose increasing the reliance on fat metabolism (18) or the downregulation of energy demand by the peripheral tissue (15). Alternatively, increased catabolism by other tissues, specifically the heart (31), could be responsible for the attenuated blood lactate in the trained divers. It is likely that the reduced peripheral blood flow (Table 2) would also play a contributory role in the observed MAP rise (8). These changes would increase the timing and intensity of arterial wave reflections, thus significantly boosting myocardial afterload (32). If chronically sustained, these changes are known to be detrimental to cardiovascular health, and whereas the elevations in MAP were maintained above baseline during the apnea recovery intervals (27), all participants’ blood pressure had returned to baseline levels before leaving the laboratory. It is not known whether prolonged involvement in the sport of freediving results in similar pathologic conditions as those associated with chronic hypertension.

In conclusion, numerically fitting the SaO2 data to a four-parameter sigmoidal logistic function provided no quantifiable difference in the rate of arterial oxygen desaturation, between the trained freedivers and control subjects. The two-slope continuous method, however, did indicate that the trained freedivers who had larger initial oxygen stores and produced the largest bradycardia were able to slow the rate of arterial oxygen desaturation two to three times that of the least marked response.


1. Andersson, J. P. A., M.H. Liner, A. Fredsted, and E. K. A. Schagatay. Cardiovascular and respiratory responses to apneas with and without face immersion in exercising humans. J. Appl. Physiol. 96:1005–1010, 2004.
2. Andersson, J. P. A., M. H. Liner, E. Runow, and E. K. A. Schagatay. Diving response and arterial oxygen saturation during apnea and exercise in breath-hold divers. J. Appl. Physiol. 93:882–886, 2002.
3. Andersson, J. P. A., and E. Schagatay. Arterial oxygen desaturation during apnea in humans. Undersea Hyperb. Med. 25:21–25, 1998.
4. Arnold, R. Extremes in human breath hold, facial immersion bradycardia. Undersea Biomed. Res. 12:183–190, 1985.
5. Bakovic, D., Z. Valic, D. Eterovic, et al. Spleen volume and blood flow response to repeated breath-hold apneas. J. Appl. Physiol. 95:1460–1466, 2003.
6. Delahoche, J., P. Delapille, F. Lemaitre, E. Verin, and C. Tourny-Chollet. Arterial oxygen saturation and heart rate variation during breath-holding: comparison between breath-hold divers and controls. Int. J. Sports Med. 26:177–181, 2005.
7. Eilers, R. Notification of final adoption of an international method and standard solution for hemoglobinometry specifications for preparation of standard solution. Am. J. Clin. Pathol. 47:212–214, 1967.
8. Ferretti, G. Extreme human breath-hold diving. Eur. J. Appl. Physiol. 84:254–271, 2001.
9. Ferretti, G., and M. Costa. Diversity in and adaptation to breath-hold diving in humans. Comp. Biochem. Physiol. 136:205–213, 2003.
10. Ferretti, G., M. Costa, M. Ferrigno, et al. Alveolar gas composition and exchange during deep breath-hold diving and dry breath holds in elite divers. J. Appl. Physiol. 70:794–802, 1991.
11. Ferrigno, M., G. Ferretti, A. Ellis, et al. Cardiovascular changes during deep breath-hold dives in a pressure chamber. J. Appl. Physiol. 83:1282–1290, 1997.
12. Fink, W. J., and D. L. Costill. Analytical methods for the measurement of human performance: human performance laboratory. Muncie, IN: Ball State University, 1990, 38–43.
13. Fletcher, E., C. Costarangos, and T. Miller. The rate of fall of arterial oxyhemoglobin saturation in obstructive sleep apnea. Chest 96:717–722, 1989.
14. Fletcher, E. C., R. Kass, J. I. Thornby, J. Rosborough, and T. Miller. Central venous O2 saturation and rate of arterial desaturation during obstructive apnea. J. Appl. Physiol. 66:1477–1485, 1989.
15. Hochachka, P. Defense strategies against hypoxia and hypothermia. Science 231:234–241, 1986.
16. Hurford, W. E., S. K. Hong, Y. S. Park, et al. Splenic contraction during breath-hold diving in the Korean ama. J. Appl. Physiol. 69:932–936, 1990.
17. Joulia, F., G. J. Steinberg, F. Wolff, O. Gavarry, and Y. Jammes. Reduced oxidative stress and blood lactic acidosis in trained breath hold divers. Respir. Physiol. Neurobiol. 133:121–130, 2002.
18. Joulia, F., J. G. Steinberg, M. Faucher, et al. Breath-hold training of humans reduces oxidative stress and blood acidosis after static and dynamic apnea. Respir. Physiol. Neurobiol. 137:19–27, 2003.
19. Kent, B., J. Drane, B. Blumenstein, and J. Manning. A mathematical model to assess changes in the baroreceptor reflex. Cardiology 57:295–310, 1972.
20. Krachman, S. L., J. Crocetti, T. J. Berger, W. Chatila, H. J. Eisen, and G. E. D’Alonzo. Effects of nasal continuous positive airway pressure on oxygen body stores in patients with Cheyne-Stokes respiration and congestive heart failure. Chest 123:59–66, 2003.
21. Lin, Y. C., D. A. Lally, T. O. Moore, and S. K. Hong. Physiological and conventional breath-hold breaking points. J. Appl. Physiol. 37:291–296, 1974.
22. Lindholm, P., and D. Linnarsson. Pulmonary gas exchange during apnoea in exercising men. Eur. J. Appl. Physiol. 86:487–491, 2002.
23. Lindholm, P., P. Sundblad, and D. Linnarsson. Oxygen-conserving effects of apnea in exercising men. J. Appl. Physiol. 87:2122–2127, 1999.
24. Schagatay, E., and J. Andersson. Diving response and apneic time in humans. Undersea Hyperb. Med. 25:13–19, 1998.
25. Schagatay, E., J. P. A. Andersson, M. Hallen, and B. Palsson. Role of spleen emptying in prolonging apneas in humans. J. Appl. Physiol. 90:1623–1629, 2001.
26. Schagatay, E., H. Haughey, and J. Reimers. Speed of spleen volume changes evoked by serial apneas. Eur. J. Appl. Physiol. 93:447–452, 2005.
27. Schagatay, E., M. van Kampen, and J. Andersson. Effects of repeated apneas on apneic time and diving response in non-divers. Undersea Hyperb. Med. 26:143–149, 1999.
28. Sterba, J. A., and C. E. Lundgren. Breath-hold duration in man and the diving response induced by face immersion. Undersea Biomed. Res. 15:361–375, 1988.
29. Stewart, I., and D. McKenzie. The human spleen during physiological stress. Sports Med. 32:361–369, 2002.
30. Strohl, K., and M. Altose. Oxygen saturation during breath-holding and during apneas in sleep. Chest 85:181–186, 1984.
31. Vary, T., D. Reibel, and J. Neely. Control of energy metabolism of heart muscle. Annu. Rev. Physiol. 43:419–430, 1981.
32. Westerhof, N., and M. O’Rourke. Haemodynamic basis for the development of left ventricular failure in systolic hypertension and for its logical therapy. J. Hypertens. 13:943–952, 1995.
33. Wolf, S., R. Schneider, and M. Groover. Further studies on the circulatory and metabolic alterations of the oxygen-conserving (diving) reflex in man. Trans. Assoc. Am. Physicians 78:242–254, 1965.


©2005The American College of Sports Medicine