Eur J Appl Physiol (2005) DOI 10.

1007/s00421-005-0083-7

O R I GI N A L A R T IC L E

Jeremy Cornolo Jean-Pierre Fouillot Laurent Schmitt Camillo Povea Paul Robach Jean-Paul Richalet

Interactions between exposure to hypoxia and the training-induced autonomic adaptations in a live hightrain low session

Accepted: 22 September 2005 Springer-Verlag 2005

Abstract The autonomic and cardiovascular adaptations to hypoxia are opposite to those resulting from aerobic training. We investigated (1) whether exposure to hypoxia in a live hightrain low (LHTL) session limits the autonomic and cardiovascular adaptations to training, and (2) whether such interactions remain 15 days after the end of the LHTL. Eighteen national swimmers trained for 13 days at 1,200 m, living (16 h day1) either at 1,200 m (live lowtrain low, LLTL) or at a simulated height of 2,5003,000 m (LHTL). Subjects were investigated at 1,200 m before and at the end of the training session, and after the following 15 days of sea-level training. Cardiovascular parameters and the autonomic control assessed by spectral analysis of RR and diastolic blood pressure (DBP) variability were obtained in the resting supine position and in response to an orthostatic test. At the end of the 13-day training, resting heart rate (HR) and sympathetic modulation on heart decreased in LLTL (10.1% and 25.4%, P<0.01, respectively) but not in LHTL (5.8, 15.5%, respectively). Total peripheral resistance (TPR) and DBP became higher in LHTL than in LLTL (P<0.05). Stroke index decreased in both groups during the tilt test, counteracted by an increase in HR and sympathetic modulation to the heart and vasculature, and a decrease in vagal modulation to the heart. After the following 15-day sea-level training, dierences in TPR and DBP between groups disappeared. During an LHTL session, adaptations to hypoxia interacted with the autonomic
J. Cornolo J.-P. Fouillot C. Povea J.-P. Richalet (&) ` Laboratoire, Reponses cellulaires et fonctionnelles a lhypoxie, Universite Paris 13, EA2363, ARPE, 93017 Bobigny, France E-mail: richalet@smbh.univ-paris13.fr Tel.: +33-148-387757 Fax: +33-148-387777 L. Schmitt Centre National de Ski Nordique, ID Jacobeys, 39200 Premanon, France P. Robach Ecole Nationale de Ski et dAlpinisme, 74401 Chamonix, France

and cardiovascular adaptations to training. However, these interactions did not limit the adaptations to the following sea-level training. Keywords Heart rate variability Live hightrain low Hypoxia Training Orthostatic test

Introduction
When subjects are exposed to acute or sub-acute hypoxia, resting heart rate (HR), mean blood pressure (BP) and cardiac output (CO) increase secondary to an increase in sympathetic activity (Bernardi et al. 1998; Cornolo et al. 2004) and a fall in vagal tone (Bernardi et al. 1998; Cornolo et al. 2004). The live hightrain low model (LHTL, Levine and Stray-Gundersen 1992), which improves endurance performance of athletes (Rusko et al. 1999), is characterised by a continuous exposure to hypoxia at rest interrupted by training workloads in normoxia. As LHTL implies training in normoxia it should provoke chronic resting autonomic adaptations, such as a decrease in HR due to a fall in sympathetic and an increase in vagal activity (Bonaduce et al. 1998; Iellamo et al. 2002; Yamamoto et al. 2001) and a decrease in intrinsic HR (Bonaduce et al. 1998). Thus, the autonomic and cardiovascular responses to hypoxia are opposite to those resulting from aerobic training, while exposure to hypoxia induces similar autonomic and cardiovascular eects as acute exercise (Bernardi et al. 1998; Cornolo et al. 2004; Nakamura et al. 1993) which could persist after exercise (Cleroux et al. 1992; Mourot et al. 2004; Terziotti et al. 2001; Howard et al. 2000; Furlan et al. 1993) or post-hypoxic exposure (Arabi et al. 1999; Xie et al. 2001). The objective of this study was to evaluate if the autonomic cardiovascular adaptations to training by highly trained subjects could be blunted by the discontinuous hypoxic stress characterising the LHTL model. The training regimen was predominantly aerobic, and hypoxic stress was moderate, and we hypothesised that

the discontinuous hypoxic exposure would not modify the training-induced cardiovascular and autonomic adaptations in elite athletes. Moreover we evaluated whether possible changes remained after a following 15day training at sea level. A passive head-up tilt (HUT) test assessed the cardiovascular and autonomic adaptations to LHTL and the following training (Beitzke et al. 2002; Fu et al. 2004; Gratze et al. 1998).

Methods
The study was conducted on healthy, sea-level, native competitive swimmers (males, n=16; females, n=2): 400 m front crawl (n=16) or long-distance (n=2) specialists, recruited for their participation in a national championship. After having given their written informed consent to participate in the study, a medical examination and echocardiography were performed. Then, the subjects were matched by gender and maxi _ mal oxygen uptake V O2 max to either a live low train low group (LLTL, n=9, aged 17.00.1 years, height 180.32.3 cm, body weight 67.52.1 kg, _ V O2max =58.51.9 ml min1 kg1; mean SEM) or an LHTL group (n=9, aged 19.91.0 years, height 178.91.6 cm, body weight 70.82.9 kg, _ V O2max =57.91.8 ml min1 kg1). The procedures, training sessions and residence were at 1,200 m above sea level (Ecole Nationale de Ski Nordique, Premanon, France, barometric pressure, Pb = 674 mmHg). The LLTL lived and slept at ambient Pb (1,200 m asl), while the LHTL lived and slept (16 h/day) at a simulated altitude obtained through an oxygen extraction system (OBS, Husysund, Norway): 5 days at 2,500 m (inspired O2 fraction, FiO2=17.4%) followed by 8 days at 3,000 m (FiO2=16.4%). All the subjects trained for 13 days at 1,200 m in a 25-m pool. Subjects gave their informed written consent and the protocol was approved by the ethics committee of the Necker Hospital, Paris. Protocol The present study is part of an evaluation of the eects of an LHTL session. Results concerning performance and haematology have been presented (Brugniaux et al. 2005; Robach et al. 2005). The subjects were investigated before (PRE), at the end of the training camp (POST1), and after 2 weeks of training following the training camp (POST2). A HUT test was performed in the morning (08:30 13:00) after a light breakfast without coee, before the daily workouts in the training camp. Subjects performed the HUT in the same order. At POST1, the interval between the last continuous nocturnal hypoxic exposure and recordings was 14.5 h. To avoid shortterm eects of exercise on HR variability (HRV) (Furlan et al. 1993), the last training session was on the

previous day or earlier. Subjects sat quietly for 30 min. Then, they were instrumented in the upright position and strapped on a manually operated tilt bed with foot support and tilted supine. After 10 min, values were collected for 3 min. When these measurements were completed, the subjects were exposed to 60 HUT. After 6 min of standing, values were collected for 4 min. Finally, the subjects were returned to the supine position. The HUT was terminated in case of presyncope (Task Force on Syncope, European Society of Cardiology 2001). Since respiration modulates autonomic activity, breathing patterns aects the spectral power of the HRV (Akselrod 1995). Subjects with a spontaneous breathing frequency below 9 b min1 in the supine position were instructed to breath at 12 b min1 for 2 min before being exposed to HUT. If HRV parameters of the 12 b min1 period were dierent from those of the spontaneous breathing frequency period, the subjects were excluded from the study. However, audio instruction could have inuenced the HRV recordings (Bernardi et al. 2000). The study started in December during the preparation period for competitive swimming, after a period of relative detraining. In the training camp 2 daily workouts of 2 h each, 7 days per week, were supervised by a national coach. According to Avalos et al. (2003) improvement of the swimming aerobic capacity requires a large amount of aerobic training. For that purpose, the predominant training intensity was close to the blood lactate accumulation threshold (OBLA), while two to three high load workouts were performed to maximise the impact of the LHTL. The training regimen was reduced between POST1 and POST2, similar to the 2-week training period preceding a major competition (Avalos et al. 2003; Mujika et al. 1996). Training intensity and volume were quantied according to Mujika et al. (1996). Progressive swimming until exhaustion was performed before the training camp, and the blood lactate concentration was determined. Then the training load was ranked into ve intensities (IV), corresponding to ve lactate concentrations, and weighted. The training stimulus was quantied by multiplying the workout with the training duration. Between POST1 and POST2, the subjects reported their training program in a book.

Parameters Cardiovascular system Beat-by-beat time series of RR intervals were recorded from a standard 2-canal ECG using four ECG chest electrodes. The RR intervals were instantaneously obtained as the dierence between successive RR peaks. Continuous beat-to-beat arterial BP was recorded by photoplethysmography using the vascular unloading

technique (Parati et al. 2003). The Task Force Monitor system (TFM) (CNSystems, Graz, Austria) provides non-invasive estimations of changes in the cardiovascular system and its autonomic control (Beitzke et al. 2002; Eckert et al. 2003; Parati et al. 2003). The sensor was placed on the volar surface of the right middle nger, while the right arm was maintained horizontal. Continuous BP was corrected to oscillometric BP obtained every minute at the brachial artery on the contralateral arm. Continuous and oscillometric BP signals were transmitted to a device monitor detecting uctuations in a range of 50250 mmHg (5 mmHg). Recordings of RR intervals and systolic BP (SBP) were analysed using the sequence method to estimate the spontaneous baroreex sensitivity (SBS) (Tank et al. 2000). From beat-to-beat time series of intervals the TFM system identied spontaneous sequences of three consecutive beats in which RR intervals and SBP either increased (+RR/+SBP) or decreased (RR/SBP) over four cardiac beats. The minimum change for a rise or fall in SBP was 1 mmHg and for RR it was 4 ms beat1 (Eckert et al. 2003; Gratze et al. 1998). The slope of the regression line was calculated for each sequence of RR/SBP relationship and episodes with correlation coecients r>0.95 were evaluated. The mean of the individual slopes was taken as a measure of the SBS. Stroke volume (SV) was determined by transthoracic impedance cardiography (Beitzke et al. 2002; Gratze et al. 1998). A circular band electrode was placed around the neck while two other electrodes were placed on the medial line under the xyphoid at a distance of at least 3 cm from the outer electrode. A current of 400 lA and 40 kHz passed through the thorax from the electrode placed around the neck to the two others. The voltage u(t) is proportional to the thorax impedance Z (Z(t) = u(t)I0). The rst derivative (dZ/dt) of the impedance signal Z(t) is supplied by the cardiograph. Algorithms calculate SV using the Kubiceks formula, eliminating the inuence of breathing and detecting the maxima of the dZ/dt signal (C-point), the aortic opening (B-point) and the aortic closing points (X-point). The monitor has a high reproducibility for consecutive SV measures (Gratze et al. 1998) and a good correlation with a thermodilution method. Stroke index (SI), HR (1/RR), cardiac index CI = (SV HR)/body surface area and total peripheral resistance index TPRI = (mean BP/ CO)/body surface area) were calculated. Power spectral analysis of the beat-by-beat variability of HR and diastolic BP (DBP) was obtained by the autoregressive method. Specic characteristics of the power spectrum of HR and DBP variability quantify sympathetic and parasympathetic control of the cardiovascular system (Akselrod 1995; Pagani et al. 1997; Zhang et al. 2002; Task Force 1996). Two frequencies were considered: the low (LF, 0.040.17 Hz) and the high frequency (HF, 0.170.40 Hz) band. The RR-LF and DBP-LF were considered as markers of sympathetic activity respectively on heart and

vasculature (Akselrod 1995; Pagani et al. 1997; Zhang et al. 2002). Because the RR-HF is a result of respiratory sinus arrhythmia mediated by the vagus, its amplitude reects the respiratory modulation of cardiac vagal outow. The RR-LF/HF ratio is an index of the sympatho-vagal balance (Akselrod 1995; Task Force 1996). Both LF and HF inuences were expressed in absolute (au, ms2) and in normalised units (nu): HF nu = (HF ms2)/((LF ms2+HF ms2)100). Normalised units minimise the eect of the changes in TP on LF and HF components. Data analysis After being converted from analog to digital with a 12-bit resolution and sampled at 1 KHz per channel, data were transmitted to the TFM system connected to a PC and stored for further analysis. After visual inspection, values were corrected by omitting artefacts. Finally 512 RR and 256 DBP intervals and corresponding variables were kept and averaged for each condition. Statistical analysis Results were expressed as mean SEM. A Shapiro Wilk test was performed to ensure that the data satised assumptions consistent with a normal distribution. Then, a two-way analysis of variance was performed to assess the eects of both time (PRE, POST1 and POST2) and group (LLTL and LHTL). When main eects or interactions reached a statistically signicant level, the Tukey post hoc test was applied to locate the dierence. Adaptations to the standing position were expressed as a percentage of the baseline resting supine values. Linear regression analyses were used to evaluate interactions between parameters in response to tilt. Data were analysed using STATISTICA (5.1) statistical package. A P value <0.05 was taken as evidence of signicance and a Cohens D test was performed to ensure the statistical power of the results and minimise a statistical type II-error.

Results
In the supine position two subjects of the LHTL group breathed at 6 b min1. This breathing rate shifted the HRV power spectrum towards low frequencies compared to the controlled breathing frequency and these subjects were excluded from the statistical analysis. Both groups performed the same volume and intensity of training during the 13-day camp and during the following 2-week training (Table 1). In the camp, the training load included 85% at or below OBLA (intensity I and II), 13% above OBLA (intensity III and IV) and 2% of speed/weight-lifting exercises (intensity V).

Table 1 Quantication of training LLTL Training camp TW1 TW2 Routine training RTW1 RTW2 Time (h wk1) TU wk1 Time (h wk1) TU wk1 Time (h wk1) TU wk1 Time (h wk1) TU wk1 24.41.0 833.3 21.21.2 724.0 15.30.6* 435.0* 15.90.6* 453.0* LHTL 25.31.3 772.6 21.11.0 755.0 11.40.8* 373.0* 17.11.0* 444.6*

Training units per week (TU wk1) calculated from the method of Mujika et al. (1996) and time of training in hours per week (h wk1); during the rst (TW1) and the second (TW2) week of the training camp; during the rst (RTW1) and the second (RTW2) week of training following the training camp; for both living lowtraining low (LLTL) and living hightraining low (LHTL) groups. Values are mean SEM for LLTL (n=9) and LHTL (n=7) *Training camp versus routine training: P<0.05

LHTL and LLTL training session Supine rest At the end of the training camp HR decreased in the LLTL (6.9 bpm) but not in the LHTL (3.7 bpm) group (Fig. 1a). The RR-LF nu decreased in the LLTL but not in the LHTL group (Fig. 1b), while the RR-HF nu and also the RRLF/HF remained constant (Table 2). Mean and systolic BP and TPR did not change while the diastolic BP tended to increase in the LHTL group (P=0.06) (Fig. 1c, Table 2). The diastolic BP and TPR were higher in the LHTL than in the LLTL group and tended to be higher for mean and systolic BP (P=0.06). Sympathetic tone on the vasculature (DBPLF nu) fell in both groups during the training camp (Table 2). Head-up tilt Responses were similar in the LLTL and LHTL groups before and after the training camp. Stroke index decreased during HUT at PRE and POST1, inducing an increase in HR secondary to a rise in cardiac sympathetic modulation (RR-LF nu, RR-LF/HF) and a decrease in vagal activity (RR-HF au, RR-HF nu) (Table 3). In both groups the increase in sympathetic modulation of the heart was negatively correlated with its resting supine level (LLTL, r = 0.75, P=0.01; LHTL, r = 0.83, P=0.01). The decrease in SI in response to the HUT was enhanced between PRE and POST1 in the LLTL group concomitantly to a rise in the increase in HR. The TPR did not signicantly change with HUT while DBP-LF nu increased, except in the LHTL group at PRE. The TPR response was negatively correlated with the increase in HR (PRE, r = 0.60, P=0.01; POST1, r = 0.64, P=0.006). Cardiac index, mean, systolic and diastolic BP were maintained with HUT. SBS decreased in both groups during the tilt, except in the LHTL group at POST1.

Post-LHTL and post-LLTL 15-day normoxic training Supine rest The HR was lower after the 15-day post-hypoxic training period than at the beginning of the training session (LLTL, 5.6 bpm; LHTL, 7.8 bpm) (Fig. 1a). The RR-LF nu was lower than at the beginning of the training session in both groups (Fig. 1b). Mean BP decreased in both groups during the training (Fig. 1c), while the systolic and diastolic BP decreased only in the LHTL group (Table 2). The DBP-LF nu was lower than before the training session in the LHTL group (Table 2). Head-up tilt The decrease in SI was reduced at POST2 compared to POST1 in the LLTL group, and CI was lower than in the LHTL group (Table 3). The increase in RR-LF au and RR-LF nu with the tilt was higher at POST2 than at PRE in the LLTL group, while the rise in HR remained in both groups (Table 3). The decrease in RR-HF au and RR-HF nu and the increase in RR-LF/ HF in response to the tilt remained in both groups. The TPRI did not change with the tilt to standing while DBP-LF nu increased. The decrease in SBS remained constant in both groups.

Discussion
The main nding from this study is that during an LHTL session, discontinuous hypoxia (13 days, 16 h day1 at a simulated height of 2,5003,000 m) interacts with the autonomic and cardiovascular adaptations to endurance training in highly trained athletes. The physiological consequences of this interaction disappeared after a 15-day training performed at sea level after LHTL.

Fig. 1 Mean heart rate (HR) (a), normalised low-frequency power (LF nu) of beat-by-beat variability of RR intervals as a marker of the cardiac sympathetic activity (b) and mean blood pressure (MBP) (c) in the basal resting supine position in LLTL and LHTL groups before (PRE) and at the end of the 13-day training camp (POST1), after 15 days of training following the training camp (POST2). Bars and lines show mean SEM for LLTL (n=9) and LHTL (n=7) *P<0.05, POST1 and POST2 versus PRE. P<0.05, LHTL versus LLTL

LHTL and LLTL training session Supine rest The training session took place at the beginning of the competitive season. The resting HR was decreased by training (Bonaduce et al. 1998; Iellamo et al. 2002; Yamamoto et al. 2001) and this fall paralleled the one

observed in cardiac sympathetic modulation assessed by RR-LF nu. This fall in HR was not associated with an increase in vagal activity (RR-HF nu), as observed in high-level athletes training at 75% of their maximum load (Iellamo et al. 2002). The training load was higher and close to a strenuous endurance training program (Avalos et al. 2003; Mujika et al. 1996), which may modify the vagal adaptations (Iellamo et al. 2002). A decrease in the intrinsic HR could also have occurred (Bonaduce et al. 1998). The decrease in HR and sympathetic modulation on the heart observed in LLTL was not signicant in the LHTL group. These results could represent an eect of hypoxia, reducing the autonomic and cardiovascular adaptations to training. Moreover, exercise and hypoxia have similar autonomic and cardiovascular eects opposite to those resulting from the endurance training (Nakamura et al. 1993; Bernardi et al. 1998; Cornolo et al. 2004). It is unlikely that residual postexercise eects occurred during the LHTL session because: (1) post-exercise residual sympathetic tone is mainly observed in hypertensive men (Cleroux et al. 1992) or after maximal exercise (Mourot et al. 2004; Terziotti et al. 2001), while 85% of our training was at or below the OBLA; (2) residual post-exercise eects decrease with the increase in training status while our subjects were national swimmers (Yamamoto et al. 2001); (3) subjects returned to their hypoxic rooms 1 h after their workouts while residual eects of exercise were evidenced for up to 2 h (Mourot et al. 2004; Terziotti et al. 2001; Howard et al. 2000). In fact, residual post-hypoxia eects may have occurred during the LHTL session since the altitude used (2,500 3,000 m) corresponded to the upper limit recommended by Levine and Stray-Gundersen (1992) for an LHTL session. Residual eects of hypoxia could have persisted until the beginning of the training workouts, limiting the training adaptations. Finally, these interactions between hypoxia and training were small: the decrease in HR (5.8%) and RR-LF nu (15.5%) in the LHTL group was not signicant. Sympathetic control of the vasculature decreased in both groups at the end of the training session while mean and diastolic BP and TPR remained stable. This discrepancy could have been due to the training- and hypoxia-induced vascular remodelling (Iwasaki et al. 2003). The higher BP and TPR after LHTL than after LLTL may have represented hypoxic after-eects (Arabi et al. 1999; Bernardi et al. 1998; Xie et al. 2001). The increase in diastolic BP observed in the LHTL group (9 mmHg or 14.6%) was higher than after an acute (Xie et al. 2001) or 2-night (Arabi et al. 1999) hypoxic exposure and was mainly mediated by TPR. Although BP and TPR did not change signicantly after training, the dierence between the two groups became signicant after training. The interactions between exposure to hypoxia and training adaptations were limited.

Table 2 Basal resting supine values of heart rate and blood pressure variability and of the cardiovascular parameters LLTL PRE Heart rate and blood pressure variability 2,340490 TP (ms2) RR-LF (ms2) 880220 2 850270 RR-HF (ms ) RR-HF nu 31.93.9 RR-LF/HF 1.90.3 DBP-LF nu 41.93.4 Cardiovascular parameters Stroke index (ml m2) 47.72.3 3.30.1 Cardiac index (l min1m2) Systolic BP (mmHg) 118.72.5 Diastolic BP (mmHg) 64.52.3 TPRI (ml mmHg1 min1 m2) 27.41.5 1 18.02.5 SBS (ms mmHg ) POST1 2,210420 750230 760260 34.44.9 1.60.5 32.34.8 # 57.52.3 3.50.2 120.22.5 64.32.0 25.11.1 16.84.2 POST2 2,340590 530100 840290 38.97.1 1.40.4 34.85.6 48.42.3 3.00.1 115.61.6 60.92.1 27.41.6 18.82.7 LHTL PRE 3,7401140 1,700560 1,440800 38.15.6 1.80.7 53.28.1 50.74.8 3.20.2 123.73.1 65.23.9 29.02.4 19.24.0 POST1 2,880990 740260 1,100550 39.57.4 2.01.1 39.65.2# 52.14.5 3.10.3 126.92.9 74.54.9 31.01.8 16.32.7 POST2 3,6701040 1,200290 1,710550 42.13.1 0.80.1 33.35.0# 53.13.7 3.00.2 120.02.5$ 64.72.4$ 30.02.4 23.23.8

Living lowtraining low (LLTL) and living hightraining low (LHTL) groups before (PRE) and at the end of the 13-day training camp (POST1), after the 15 days of training following the training camp (POST2) in both supine and standing positions. Total power (TP); low (RR-LF) and high (RR-HF) frequency power of RR intervals; RR-LF/HF ratio; low frequency power of diastolic blood pressure (DBPLF). Total peripheral resistance index (TPRI), spontaneous baroreex sensitivity (SBS). Values are mean SEM for LLTL (n=9) and LHTL (n=7) # Dierent with corresponding PRE (P<0.05) Dierent between LLTL and LHTL (, P<0.05; , P<0.01) $ Dierent with corresponding POST1 ($, P<0.05)

Head-up tilt Passive HUT was used to strengthen the assessment of the cardiovascular and autonomic function (Beitzke et al. 2002; Fu et al. 2004; Gratze et al. 1998). Responses of both groups of subjects appeared normal (Beitzke

et al. 2002; Fu et al. 2004; Task Force 2001): tilt to standing decreased SI and increased HR in parallel with a rise in sympathetic modulation and a fall in vagal activity, maintaining a constant CI and BP. The mean slope of SBS decreased due to activation of the sympathetic nervous system: for the same shift of systolic BP,

Table 3 Percentage of changes of heart rate and blood pressure variability and of the cardiovascular parameters during the tilt test LLTL PRE POST1 POST2 38.46.4*** 16.922.3 165.5104.4## 116.428.6***; # 72.37.8* 65.67.7** 992.3265.8* 182.3107.9*** 19.34.0**; $$$ 11.77.8$ 5.42.4 1.81.3**; $ 10.12.8 1.67.2 59.52.8** LHTL PRE 36.74.3*** 10.532.8 47.278.3 79.625.4** 80.74.9* 71.85.3** 1292.1681.4** 38.919.2 23.24.7* 5.37.1 0.421.9 0.71.1 3.02.5 0.59.5 73.14.4*** POST1 43.94.9*** 38.473.4 85.984.5 77.132.9** 35.9109.9 59.010.9* 1256.3861.7* 109.353.4*** 25.55.8** 8.011.1 0.72.0 0.31.0 1.73.6 1.29.0 42.521.8 POST2 39.86.2*** 82.9107.5 324.5285.7 95.824.1** 68.114.5* 67.810.8** 1336.9494.1** 90.821.4** 29.53.7**; 2.56.7 3.72.6 0.11.5* 6.52.7 9.68.5 60.28.5*

Heart rate and blood pressure variability HR 29.64.5*** 41.44.5***; # TP 43.018.8 * 28.013.5 10.324.0 28.437.4 RR-LF ms2 RR-LF nu 22.711.5 81.819.2*** 2 76.86.8* 76.76.8* RR-HF ms RR-HF nu 51.014.8** 54.112.6** RR-LF/HF 446.0188.2* 610.8146.4** DBP-LF nu 60.913.0*** 80.816.7** Cardiovascular parameters Stroke index 23.44.1*** 33.33.5***; # Cardiac index 0.76.5 4.65.8 Mean BP 3.22.0 0.21.9 Systolic BP 2.21.0* 2.01.4# Diastolic BP 5.22.3 5.32.7 TPRI 6.85.8 7.96.1 SBS 60.88.4* 40.322.1*

Living lowtraining low (LLTL) and living hightraining low (LHTL) groups before (PRE) and at the end of the 13-day training camp (POST1), after the 15 days of training following the training camp (POST2) in both supine and standing positions. Heart rate (HR); Total power (TP); low (RR-LF) and high (RR-HF) frequency power of RR intervals; RR-LF/HF ratio; low frequency power of diastolic blood pressure (DBP-LF). Total peripheral resistance index (TPRI), spontaneous baroreex sensitivity (SBS). Values are mean SEM for LLTL (n=9) and LHTL (n=7) *Dierent between supine and standing (%) (*P<0.05; **P<0.01; ***P<0.001) # Dierent with corresponding PRE (#P<0.05; ##P<0.01) $ Dierent with corresponding POST1 ($P<0.05; $$P<0.01; $$$P<0.001) Dierent between LLTL and LHTL (P<0.05)

the increase in RR interval was reduced (Beitzke et al. 2002). For both groups, the lower the resting supine level of sympathetic modulation to the heart, the higher was the increase in response to tilt. Thus, the training-induced decrease in sympathetic modulation may have facilitated the cardiovascular response to the orthostatic stress. Whether the test was performed before or after LLTL and LHTL, the tilt from supine to standing did not induce an increase in TPR. This lack of response in TPR was compensated by a stronger increase in HR and was neither due to exposure to discontinuous hypoxia nor to training. This dissociation between the increase in sympathetic tone on vasculature and the lack of increase in TPR (Fu et al. 2004) may be due to the long-term training and is independent of the specic modalities of swimming (exercise in supine position) (Franke et al. 2003). Finally the hypoxia-induced increase and dominant sympathetic tone on heart and vasculature did not limit the further adaptations to a tilt test performed 14.5 h after the last nocturnal hypoxic exposure (Bernardi et al. 1998; Cornolo et al. 2004). In order to obtain the higher training-induced benets, the training load was similar to a strenuous endurance program of elite swimmers (Avalos et al. 2003; Mujika et al. 1996). Since training was conducted before the competitive period, we did not increase the training intensity to avoid the risk of overreaching. However, the total power of RR intervals tended to decrease in the LHTL group which could have represented the initiation of a fatigue process (Pichot et al. 2002). The hypoxic stimulus could have been too intense or additive with training, perturbing the recovery process of the athletes. The present study was associated with autonomic cardiovascular adaptations to training in both groups and increase in red cell volume after LHTL, adding to the slight improvement of swimming performance after LHTL while it remained constant after LLTL (Robach et al. 2005). Our results concerning performance, haematology and acclimatisation (Brugniaux et al. 2005; Robach et al. 2005) are in agreement with previous reports (Levine and Stray-Gundersen 1992; Rusko et al. 1999). The post-LHTL and post-LLTL 15-day normoxic training Supine rest A 2-week delay is often used between an LHTL session and a major competition. Thus, we conducted the 2 weeks of sea-level training, performed after the LHTL session, according to the period of training preceding a swimming competition, characterised by its decreased load (Mujika et al. 1996). Cardiovascular and autonomic benets of either LLTL or LHTL were maintained during the training: the decrease in resting supine

HR and sympathetic modulation persisted while the increase in vagal modulation on heart was reinforced. Similar benets of this training on HR and the autonomic control to the heart, and disappearance of the dierence in BP and TPR between groups, argue for an interaction between discontinuous hypoxia and training during the LHTL session. Responses to head-up tilt were similar to those observed at the end of the LHTL training session. Interindividual variability in response to standing tended to be lower compared to the posttraining session recordings. Conclusions In order to investigate whether the exposure to discontinuous hypoxia interacts with training-induced autonomic and cardiovascular adaptations during an LHTL session performed by highly trained subjects, national elite swimmers were exposed 16 h day1 to a 2,500 3,000 m simulated altitude during a 13-day aerobic training camp. In the supine resting position, but not in response to an orthostatic test, exposure to discontinuous hypoxia interacted with the autonomic and cardiovascular adaptations to training. This interaction disappeared after a following training session and did not limit the autonomic and cardiovascular adaptations.
Acknowledgements The author thanks the Federation Francaise de Natation and all the volunteers for their participation in this study at the Centre National de Ski Nordique (Premanon, Jura). This project was supported by the International Olympic Committee and the French Ministry of Sports.

References
Akselrod S (1995) Components of heart rate variability: basic studies. In: Malik M, Camm AJ (eds) Heart rate variability. Futura Publishing Company, Armonk, pp 147163 Arabi Y, Morgan BJ, Goodman B, Puleo DS, Xie A, Skatrud JB (1999) Daytime blood pressure elevation after nocturnal hypoxia. J Appl Physiol 87:689698 Avalos M, Hellard P, Chatard JC (2003) Modeling the trainingperformance relationship using a mixed model in elite swimmers. Med Sci Sports Exerc 35:838846 Beitzke M, Pster P, Fortin J, Skrabal F (2002) Autonomic dysfunction and hemodynamics in vitamin B12 deciency. Auton Neurosci 97:4554 Bernardi L, Passino C, Spadacini G, Calciati A, Robergs R, Greene R, Martignoni E, Anand I, Appenzeller O (1998) Cardiovascular autonomic modulation and activity of carotid baroreceptors at altitude. Clin Sci (Lond) 95:565573 Bernardi L, Wdowczyk-Szulc J, Valenti C, Castoldi S, Passino C, Spadacini G, Sleight P (2000) Eects of controlled breathing, mental activity and mental stress with or without verbalisation on heart rate variability. J Am Coll Cardiol 35:14621469 Bonaduce D, Petretta M, Cavallaro V, Apicella C, Ianniciello A, Romano M, Breglio R, Marciano F (1998) Intensive training and cardiac autonomic control in high level athletes. Med Sci Sports Exerc 30:691696 Brugniaux JV, Schmitt L, Robach P, Jeanvoine H, Zimmermann H, Nicolet G, Duvallet A, Fouillot JP, Richalet JP (2005) Living hightraining low: tolerance and acclimatization in elite endurance athletes. Eur J Appl Physiol (in press)

Cleroux J, Kouame N, Nadeau A, Coulombe D, Lacourciere Y (1992) Aftereects of exercise on regional and systemic hemodynamics in hypertension. Hypertension 19:18391 Cornolo J, Mollard P, Brugniaux JV, Robach P, Richalet JP (2004) Autonomic control of the cardiovascular system during acclimatization to high altitude: eects of sildenal. J Appl Physiol 97:935940 Eckert S, Lotz N, Bergemann C, Kemper B, Petzoldt R, Tschope D, Horstkotte D (2003) Autonomic dysfunction indicated by reduced baroex sensitivity in patients with diabetes mellitus. In: Thirteenth European Meeting on Hypertension. Milan, Italy Franke WD, Mills KK, Lee K, Hernandez JP (2003) Training mode does not aect orthostatic tolerance in chronically exercising subjects. Eur J Appl Physiol 89:263270 Fu Q, Witkowski S, Levine BD (2004) Vasoconstrictor reserve and sympathetic neural control of orthostasis. Circulation 110:29312937 Furlan R, Piazza S, DellOrto S, Gentile E, Cerutti S, Pagani M, Malliani A (1993) Early and late eects of exercise and athletic training on neural mechanisms controlling heart rate. Cardiovasc Res 27:482488 Gratze G, Fortin J, Holler A, Grasenick K, Pfurtscheller G, Wach P, Schonegger J, Kotanko P, Skrabal F (1998) A software package for non-invasive, real-time beat-to-beat monitoring of stroke volume, blood pressure, total peripheral resistance and for assessment of autonomic function. Comput Biol Med 28:121142 Howard MG, Collins HL, Dicarlo SE (2000) Post-exercise elevations in sympathetic nerve activity and baroreex function in normotensive rabbits. Clin Exp Hypertens 22:431444 Iellamo F, Legramante JM, Pigozzi F, Spataro A, Norbiato G, Lucini D, Pagani M (2002) Conversion from vagal to sympathetic predominance with strenuous training in high-performance world class athletes. Circulation 105:27192724 Iwasaki KI, Zhang R, Zuckerman JH, Levine BD (2003) Doseresponse relationship of the cardiovascular adaptation to endurance training in healthy adults: how much training for what benet? J Appl Physiol 95:15751583 Levine BD, Stray-Gundersen J (1992) A practical approach to altitude training: where to leave and train for optimal performance enhancement. Int J Sports Med 13:S209S212 Mourot L, Bouhaddi M, Tordi N, Rouillon JD, Regnard J (2004) Short- and long-term eects of a single bout of exercise on heart rate variability: comparison between constant and interval training exercises. Eur J Appl Physiol 92:508517 Mujika I, Busso T, Lacoste L, Barale F, Geyssant A, Chatard JC (1996) Modeled responses to training and taper in competitive swimmers. Med Sci Sports Exerc 28:251258

Nakamura Y, Yamamoto Y, Muraoka I (1993) Autonomic control of heart rate during physical exercise and fractal dimension of heart rate variability. J Appl Physiol 74:875881 Pagani M, Montano N, Porta A, Malliani A, Abboud F, Birkett C, Somers VK (1997) Relationship between spectral components of cardiovascular variabilities and direct measures of muscle sympathetic nerve activity in humans. Circulation 95:14411448 Parati G, Ongaro G, Bilot G, Glavina F, Castiglioni P, Di Rienzo M, Mancia G (2003) Non-invasive beat-to-beat blood pressure monitoring: new developments. Blood Press Monit 8:3136 Pichot V, Busso T, Roche F, Garet M, Costes F, Duverney D, Lacour JR, Barthelemy JC. (2002) Autonomic adaptations to intensive and overload training periods: a laboratory study. Med Sci Sports Exerc 34:16601666 Robach P, Schmitt L, Brugniaux JV, Roels B, Millet G, Hellard P, Nicolet G, Duvallet A, Fouillot JP, Moutereau S, Lasne F, Pialoux V, Olsen NV, Richalet JP (2005) Living hightraining low: eect on erythropoiesis and aerobic performance in highlytrained swimmers. Eur J Appl Physiol (in press) Rusko HK, Tikkanen H, Paavolainen L, Hamalainen I, Kalliokoski K, Puranen A (1999) Eect of living in hypoxia and _ training in normoxia on sea-level V O2max and red cell mass. Med Sci Sports Exerc 31:S86 Tank J, Baevski RM, Fender A, Baevski AR, Graves KF, Ploewka K, Weck M (2000) Reference values of indices of spontaneous baroreceptor reex sensitivity. Am J Hypertens 13:268275 Task Force of the European Society of Cardiology, the North American Society of pacing and Electrophysiology (1996) Heart rate variability. Standards of measurement, physiological interpretation, and clinical use. Circulation 93:10431065 Task Force on Syncope, European Society of Cardiology (2001) Guidelines on management (diagnosis and treatment) of syncope. Eur Heart J 22:12561306 Terziotti P, Schena F, Gulli G, Cevese A (2001) Post-exercise recovery of autonomic cardiovascular control: a study by spectrum and cross-spectrum analysis in humans. Eur J Appl Physiol 84:187194 Xie A, Skatrud JB, Puleo DS, Morgan BJ (2001) Exposure to hypoxia produces long-lasting sympathetic activation in humans. J Appl Physiol 91:15551562 Yamamoto K, Miyachi M, Saitoh T, Yoshioka A, Onodera S (2001) Eects of endurance training on resting and post-exercise cardiac autonomic control. Med Sci Sports Exerc 33:14961502 Zhang R, Iwasaki K, Zuckerman JH, Behbehani K, Crandall CG, Levine BD (2002) Mechanism of blood pressure and RR variability: insights from ganglion blockade in humans. J Physiol (Lond) 543:337348