The effect of various breath‐hold techniques on the cardiorespiratory response to facial immersion in humans

Abstract Repeated maximal breath‐holds have been demonstrated to induce bradycardia, increase haematocrit and haemoglobin and prolong subsequent breath‐hold duration by 20%. Freedivers use non‐maximal breath‐hold techniques (BHTs) to improve breath‐hold duration. The aim of this study was to investigate the cardiorespiratory and haematological responses to various BHTs. Ten healthy men (34.5 ± 1.9 years) attended five randomized experimental trials and performed a 40 min period of quiet rest or one of three BHTs followed by a maximal breath‐hold challenge during facial immersion in water at 30 or 10°C. Cardiovascular and respiratory parameters were measured continuously using finger plethysmography and breath‐by‐breath gas analysis, respectively, and venous blood samples were collected throughout. Facial immersion in cold water caused marked bradycardia (74.1 vs. 50.2 beats/min after 40 s) but did not increase breath‐hold duration compared with warm water control conditions. Facial immersion breath‐hold duration was 30.8–43.3% greater than the control duration when preceded by BHTs that involved repeated breath‐holds of constant duration (P = 0.021), increasing duration (P < 0.001) or increasing frequency (P < 0.001), with no difference observed between BHTs. The increased duration of apnoea across all three BHT protocols was associated with a 6.8% increase in end‐tidal O2 and a 13.1% decrease in end‐tidal CO2 immediately before facial immersion. There were no differences in blood pressure, cardiac output, heart rate, haematocrit or haemoglobin between each BHT and control conditions (P > 0.05). In conclusion, the duration of apnoea can be extended by manipulating blood gases through repeated prior breath‐holds, but changes in cardiac output and red blood cell mass do not appear essential.


INTRODUCTION
Three to five repeated maximal breath-holds have been demonstrated to induce bradycardia, increase haematocrit (Hct) and haemoglobin (Hb) and increase the subsequent breath-hold time by ≤20% (Baković et al., 2003;M. Richardson et al., 2005).These physiological responses are consistent with the mammalian dive reflex (MDR) on facial immersion in cold water (Foster & Sheel, 2005;Nepal et al., 2014;Schagatay et al., 2001).Peripheral vasoconstriction is also characteristic of the dive reflex and, together with bradycardia, is exacerbated with a reduction in water temperature during facial immersion, probably via stimulation of the trigeminal nerve (Nepal et al., 2014;Paulev et al., 1990).However, there is limited evidence to show that a lower water temperature prolongs breath-hold time in humans (Foster & Sheel, 2005), which would suggest that pronounced bradycardia and vasoconstriction of selective vascular beds are perhaps not obligatory for prolonging breath-hold duration.
Erythrocyte release from the spleen is considered to be the cause of increased Hct and Hb, because splenic contraction and a 14% reduction in the splenic volume are observed during breath-holding in humans and are more pronounced in trained 'freedivers' (deep water diving without the use of breathing apparatus; Baković et al., 2003).Indeed, studies in splenectomized individuals demonstrate a reduced breath-hold time in some (Baković et al., 2003), but not all cases (Schagatay et al., 2001), regardless of the dive reflex.
Thus, it is still unclear whether these physiological responses to repeated maximal breath-holds contribute to prolonged breath-hold duration.
It is common practice for human freedivers to use repeated breath-hold techniques (BHTs) before a dive to manipulate blood gases and prolong subsequent dive duration.It is well established that breath-hold time is almost doubled by breath-holding with hyperoxic gas mixtures to increase the partial pressure of arterial oxygen (P aO 2 ) (Ferris et al., 1946;Gross et al., 1976) or by preceding breath-holding by voluntary or mechanical hyperventilation to lower the partial pressure of arterial carbon dioxide (P aCO 2 ) (Klocke & Rahn, 1959).However, to our knowledge, whether repeated breath-holds manipulate P aO 2 and P aCO 2 and contribute to the prolongation of subsequent breath-hold duration has not been investigated.Thus, the aim of the study was to investigate the effect of three different repeated BHTs routinely used by freedivers, which are thought to manipulate P aO 2 and P aCO 2 to varying degrees, on the cardiorespiratory and haematological responses to breath-holding during facial immersion.
By comparing these responses observed during facial immersion in warm and cold water, it would also be possible to provide insight into the contribution, if any, of the MDR to breath-hold prolongation.
We hypothesized that a BHT that uses repeated and progressive near-maximal breath-holds would create the largest cardiorespiratory perturbation and induce the most prolonged breath-hold duration.

Highlights
• What is the central question of this study?
What is the effect of three repeated breathhold techniques routinely used by freedivers, thought to manipulate arterial partial pressures of O 2 and CO 2 , on the cardiorespiratory and haematological response to breath-holding during facial immersion?
• What is the main finding and its importance?
All three techniques increased breath-hold by a similar duration, probably owing to the similar marked increase in end-tidal O 2 and decrease in end-tidal CO 2 observed in all three trials before facial immersion.These were the only cardiorespiratory changes that were consistently manipulated before the maximal breath-hold.This would suggest that pronounced bradycardia and vasoconstriction of selective vascular beds are probably not obligatory for prolonging breath-hold duration.All subjects were verbally informed about the nature and risks of the experiment in conjunction with a participant information sheet and consent form before written informed consent was obtained.

Participants
Ten healthy males (34.5 ± 1.9 years of age, 179.5 ± 1.2 cm in height and weighing 85.5 ± 1.7 kg) participated in this study.All participants were non-smokers and were not suffering from any cardiovascular, neurological, respiratory or metabolic diseases.Furthermore, none of the participants had been splenectomized, and they were not taking any prescribed medication.All participants were asked to report to the human physiology laboratories at the University of Exeter on five occasions in a randomized order, each separated by ≥1 week, having abstained from strenuous physical activity and alcohol for the previous 24 h and from food for the previous 2 h (Ghiani et al., 2016;Schagatay & Lodin-Sundström, 2014).

Protocol
On arrival in the laboratory, participants' height and weight were measured.Participants were seated in a comfortable upright position (Cicolini et al., 2011) in a temperature-controlled laboratory (21 total breath-hold duration 20:30 min), beginning at 20 s and adding 10 s to the breath-hold duration after every three breath-holds until reaching 1:20 min, followed by two final breath-holds of 1:30 min, separated by increasing recovery periods (nine sets at 40 s, three sets at 50 s, 11 sets at 60 s).Ten seconds before each breath-hold, the participant was instructed to perform a 'breath-up' consisting of a deep inhalation (∼80% of self-predicted maximal inspiratory volume), a complete exhalation and a final full inhalation before holding.
After the BHT period was complete, the face mask was removed, and participants were asked to rest for 2 min for recovery of spleen size (Espersen et al., 2002).A nose clip was fitted, followed by the 'breath-up' manoeuvre 10 s before facial immersion (Andersson et al., 2000) in warm water (30 • C), during which participants were asked to hold their breath for as long as comfortably possible.To exacerbate the MDR, one of the control protocols (MDR) required facial immersion in cold water (10 • C) (Fagius & Sundlof, 1986;Tipton, 2003).Participants responded to a shoulder tap with a hand signal as a safety control measure.At the breakpoint, the face mask was replaced once the face had been dried, and participants rested quietly for a further 60 min.

Sample collection and analysis
Heart rate (in beats per minute), MABP (in milllimetres of mercury), SV (in millilitres per beat), Q (in litres per minute), TPR (in milllimetres of mercury) and pre-breath-hold ET CO 2 (ETCO 2 ; in milllimetres of mercury) and O 2 (ETO 2 ; in milllimetres of mercury) measurements were collected continuously, with the exception of ETCO 2 and ETO 2 during facial immersion, owing to face mask removal.A total of 2 ml of venous blood was obtained every 20 min from the beginning of the protocol and immediately after facial immersion until completion.
Blood samples were analysed immediately for Hb concentration (in grams per litre) using a microcuvette system (HemoCue Hb 201+; HemoCue, Sweden) and Hct (as a percentage) using a 75 mm microfuge tube (SciQuip), with centrifugation for 2 min at 4,000g and measurement with a micro-haematocrit reader (Hawksley).The remainding whole blood was transferred to a vacutainer containing 10.2 mg of potassium EDTA centrifuged for 8 min at 3,250g, and plasma was stored at −80 • C for future analysis.Noradrenaline was analysed by an enzyme-linked immunosorbent assay (ELISA) method using a commercially available kit (noradrenaline ELISA; Abnova, Taipei City, Taiwan).Absorbance was measured and quantified against known standards using an EnSpire 2300 plate reader (Perkin Elmer, MA, USA).

Data and statistical calculations
Cardiorespiratory data were reported every minute during CON and MDR or at time points that matched ET gas measurements immediately before the 'breath-up' manoeuvre (Pre-BH) and at the stipulated breakpoint of each breath-hold (Post-BH) during the three BHT protocols.
Maximal breath-hold duration was analysed using a one-way ANOVA.All other data were measured using a two-way ANOVA with the factors of condition and time.The assumptions to conduct When a significant main effect was observed for interaction, a post-hoc Sidak test was used with Geisser-Greenhouse correction for sphericity where necessary.Data were analysed using commercially available software (Prism v.8; GraphPad Software, San Diego, CA, USA).All variables are presented as the mean ± SD.A P-value of <0.05 was considered significant.

Maximal breath-hold duration during facial immersion
The mean maximal breath-hold duration for participants was 149.7 ± 27.8, 147.2 ± 40.6, 214.5 ± 53.3, 205.7 ± 25.2 and 195.8 ± 24.6 s for CON, MDR, RBH, PBH and IBH, respectively (RBH P = 0.021, PBH P < 0.001 and IBH P < 0.001, all vs. CON; Figure 2).All the participants increased breath-hold duration from CON to PBH and IBH compared with eight participants in RBH.No difference in maximal breath-hold duration was observed between MDR and CON (Figures 1a and 2).
During Pre-BH MABP, only RBH was higher than CON (P = 0.004; Figure 4a) both overall and immediately before facial immersion (RBH 102.7 ± 16.9 mmHg, P = 0.0476), whilst no differences were observed in PBH (93.2 ± 17.4 mmHg; Figure 4b), IBH (92.8 ± 8.3 mmHg; Figure 4c) or MDR.Post-BH BHT overall results were all higher than CON (RBH P = 0.023, PBH P < 0.001 and IBH P < 0.001; Figure 4a-c); however, no differences were observed immediately before facial immersion or MDR when compared with CON.During facial immersion, all BHTs were significantly different from CON (RBH P = 0.007, PBH P = 0.032 and IBH; P = 0.016; Figure 4a-c), whilst no difference was observed in MDR (Figure 1b).All BHTs returned to their baseline at the 48 min time point, with no changes throughout the recovery period and no differences compared with CON.
Observing the overall recovery period, RBH (Hb P = 0.038 and Hct P = 0.035) and PBH (Hb P = 0.019) displayed differences from CON.
No differences were observed across all protocols at the 65 min time point compared with CON; however, IBH (Hb P = 0.031) was different from CON at the 85 min time point.
During the BHT phase, overall lactate increased during PBH (P = 0.024;

DISCUSSION
The aim of the present study was to investigate the effect of three different repeated BHTs routinely used by freedivers, thought to manipulate P aO 2 and P aCO 2 to varying degrees, on the cardiorespiratory and haematological responses to breath-holding during facial immersion in warm water.We hypothesized that a BHT that uses repeated near-maximal breath-holds would create the largest cardiorespiratory perturbation and induce the longest breath-hold duration.
However, contrary to this hypothesis, all BHTs increased breathhold duration to a similar degree.These observations were likely attributable to the similar increase in ETO 2 and decrease in ETCO 2 observed in all three trials before facial immersion compared with the control conditions.Indeed, these were the only cardiorespiratory changes that were consistently manipulated before the maximal breath-hold.Thus, previous reports that an increase in circulating Hct and Hb and a reduction in HR are required for breath-hold prolongation with prior breath-holding, where blood gases have been manipulated, perhaps require further investigation.For example, HR was elevated before performing a maximal breath-hold after the three breath-hold test trials and, although facial immersion in cold water induced a marked bradycardia commonly associated with the MDR, the breath-hold duration was not greater than for facial immersion in warm water.
Performing a maximal breath-hold during facial immersion in water at 30 • C in the present study caused an increase in MABP and TPR, a decrease in Q, SV and HR and an increase in circulating Hct, Hb, noradrenaline and lactate.These findings are in line with other reports (Foster & Sheel, 2005;Parkes, 2006;Schagatay, 2014) and with the theory that the increase in MABP is attributable to a sympathetically mediated peripheral vasoconstriction, whereas the decline in Q could be a subsequent baroreceptor response or independently chemoreceptor mediated (Paulev et al., 1990;Perini et al., 1998).Consistent with what would be expected from stimulating the MDR, facial immersion in cold water at 10 • C caused a further reduction in HR.However, this marked bradycardia did not prolong maximal breath-hold duration and would perhaps indicate that the dive reflex is advantageous only when accompanied by other physiological circumstances that occur during deep diving, such as hyperbaria.
Indeed, HR and Q were either similar or higher than control values before and during facial immersion for each of the BHTs, which is in agreement with other studies (Heusser et al., 2009).Thus, other cardiorespiratory or haematological responses observed during a maximal breath-hold might confer a milieu favourable to prolonging subsequent breath-hold duration.
The BHTs used in the present study resulted in a 30.8-43.3% prolongation of breath-hold duration during a subsequent breath-hold test.To our knowledge, this is the largest increase in breath-hold demonstrated that a 20% increase in breath-hold duration after three to five repeated maximal breath-holds was associated with a 1.4% increase in Hb concentration, probably attributable to a 14% reduction in splenic volume during the prior breath-holds.The increase in Hb, Hct and splenic contraction, which appears to be more pronounced under hypercapnia (M.X. Richardson et al., 2012), accounts for ∼60% of breath-hold prolongation (Schagatay et al., 2001), and greater increases in Hb concentration are associated with long-term breathhold training (Lemaitre et al., 2009;Schagatay, 2014;Zoretic et al., 2014).The present study provided a unique scenario to provide further insight into the relationship between Hb concentration and breathhold duration.Although comparatively larger increases in Hb were observed before (5.5%) and during (2.1%) the maximal breath-hold in the PBH compared with CON conditions in the present study, there was no difference in Hb concentration in the IBH conditions compared with CON conditions, despite an increase (30.8-43.3%) in maximal breath-hold duration during both sets of conditions.Furthermore, the increase in Hct before and during the maximal breath-hold followed a similar pattern in both PBH and IBH.Thus, although an increase in circulating Hb and Hct was observed during maximal breath-holds, as found in other studies (Baković et al., 2003;Espersen et al., 2002;M. Richardson et al., 2005), the magnitude of increase before or during apnoea does not appear to influence breath-hold duration.This observation suggests that other factors could be more important; for example, changes in blood gases.
It is well established that breath-hold duration is almost doubled by breath-holding with hyperoxic gas mixtures to increase P aO 2 (Ferris et al., 1946;Gross et al., 1976) or by preceding breath-holding by voluntary or mechanical hyperventilation to lower P aCO 2 (Klocke & Rahn, 1959).However, whether repeated breath-holds manipulate P aO 2 and P aCO 2 , contributing to prolongation of subsequent breathhold duration, has not been investigated.Throughout all three BHTs, there was a marked increase in ETO 2 and decrease in ETCO 2 observed before facial immersion and maximal breath-hold compared with control conditions, which appeared to be the only cardiorespiratory changes consistently manipulated before the maximal breath-hold.
The reason for the consistent changes observed in ETO 2 and ETCO 2 after all three techniques is not clear but might be attributable to recovery hyperventilation immediately after every breath-hold of each technique.Hence, when compared with RBH, the increase in ETO 2 and decrease in ETCO 2 occurred earlier during PBH and IBH; both had much shorter recovery periods during the initial breath-holds and, presumably, less time to normalize blood gases.In line with this theory, the increase in ETO 2 and decrease in ETCO 2 in RBH eventually occurred when the recovery periods reduced to a similar duration to those in PBH and IBH (i.e.∼60 s or less).Given the incremental nature of the increase in ETO 2 and decrease in ETCO 2 , it would appear, perhaps counterintuitively, that repeated short recovery periods between breath-holds, in addition to prolonged breath-hold periods, might have a greater effect on blood gas manipulation and maximal breath-hold duration.More research on this practice is clearly warranted, but it is interesting to note that PBH, which had longer initial breathholds than IBH and short (60 s) recovery periods, appeared to have the earliest changes in circulating Hb, Hct, noradrenaline and lactate, perhaps reflecting greater cardiorespiratory 'stress' and recovery hyperventilation.
In conclusion, this is the first study to demonstrate that progressive repeated breath-holds will manipulate ETO 2 and ETCO 2 and, presumably, the associated blood gases and will prolong breath-hold duration during facial immersion.The increase in maximal breath-hold duration was similar across all three BHTs and is consistent with that observed anecdotally in freedivers.In contrast, there did not appear to be a clear association with maximal breath-hold duration and any other cardiovascular or haematological parameters measured across the three techniques, indicating that they do not have a major effect on breath-hold prolongation.However, it is important to note that the present study was not performed in trained divers, who might have different physiological responses to a maximal breath-hold after repeated breath-holds.Indeed, splenic contraction observed during breath-holding in humans is more pronounced in trained freedivers (Baković et al., 2003).Thus, further research on the physiological effects of these BHTs in trained individuals is clearly required.
Maximal breath-hold duration during facial immersion in warm water (30 • C) after 40 min of quiet rest (CON) or a 40 min repeated breath-hold technique in which each breath-hold duration was kept constant (RBH) or gradually prolonged (PBH) or the number of breath-holds was increased (IBH).Values are the mean ± SD for n = 10.*P < 0.05, ***P < 0.01 and ****P < 0.001 significantly greater than CON Figure 3d-f).During Pre-BH, overall ETCO 2 decreased (RBH P = 0.024 and IBH P = 0.003), with further increases observed immediately before facial immersion (RBH 32.8 ± 6.5 mmHg, P = 0.016; and IBH 33.3 ± 5.1 mmHg, n.s.; Figure 3d,f) compared with CON.No differences were observed when comparing PBH (34.7 ± 2.9 mmHg, n.s.; Figure 3e) or MDR with CON.Conversely, during Post-BH, only PBH increased overall ETCO 2 compared with CON (P < 0.001), and no differences were observed immediately before facial immersion in all protocols.At the 48 min time point, PBH ETCO 2 was acutely lower than CON (34.6 ± 4.1 mmHg, P = 0.006), and no difference was observed 1 min later and thereafter.All other protocols had returned to baseline by the 48 min time point, and no differences were observed during the remaining recovery period.
Figure 5a-c) and MDR displayed no differences compared with CON.During Post-BH, all BHTs were different compared with CON (RBH P < 0.001, PBH P < 0.001 and IBH P = 0.009), whilst CON and MDR elicited no change.Although change over time occurred in all protocols, no differences were observed before facial immersion during BHTs

F I G U R E 7
Plasma noradrenaline (a-c) and lactate (d-f) concentrations during a 40 min repeated breath-hold technique in which each breath-hold duration was kept constant (RBH; a,d) or gradually prolonged (PBH; b,e) or the number of breath-holds was increased (IBH; c,f).Each breath-hold technique measurement is compared with a corresponding value during 40 min of quiet rest (CON).After a 2 min 'rest' period, the response to a maximal breath-hold during facial immersion in warm water (30 • C) is presented on the right-hand x-axis.Values are the mean ± SD for n = 10.Main effects of repeated-measures ANOVA are presented in each panel duration recorded in the literature after prior voluntary breath-holding and the first documented use of these techniques in a laboratory experiment that has measured cardiorespiratory and haematological responses.Baković et al. (2003) and M.Richardson et al. (2005)