Patients

Echocardiography

The patients were monitored by electrocardiography throughout the observed sessions and examined whilst standing in a slightly tilted position both on land and in a swimming pool for 20–30 min with the water level up to the sternal notch. Transthoracic echocardiography examinations were performed using Siemens Sequoia 512 with a 3v2c transducer (Mountain View, CA, USA), and data were stored digitally on magnetic optical disks. To protect the probe from water, the transducer was placed in a latex stocking. In accordance with the recommendations of the American Society of Echocardiography, LVEF was calculated using the method of discs, modified Simpson rule [11]. Estimation of pulmonary capillary wedge pressure was calculated by a formula using transmitral Doppler and tissue Doppler, as described by Nagueh [12]. Mean arterial pressure was estimated by the following equation: pressure_diastolic+1/3(pressure_systolic-pressure_diastolic). Systemic vascular resistance was calculated as: mean arterial pressure/cardiac output.

Ventricular long axis function

Ventricular long axis (LAX) function was assessed with M-mode [13] with the beam positioned from the apex to the insertion of the septal and lateral portion of the mitral valve and from the right ventricular (RV) lateral free wall. Pulsed-wave tissue Doppler imaging (TDI) was assessed from the same positions with additional registration from the anterior and inferior walls. Systolic velocities, early diastolic velocities and atrial velocities were obtained [14]. Annular systolic motion was also measured by calculation of the integral (area) of the systolic velocity curves – tissue velocity time integral (TVTI) [15, 16].

Pulsed wave Doppler

Pulsed wave Doppler was carried out for calculation of stroke volume and cardiac output [17] by registration of velocity time integral from left ventricular outflow tract (LVOT) in the five-chamber view. Cardiac output was calculated from the formula: stroke volume = (LVOT area × VTI) × heart rate, where LVOT area was calculated as: 3.14 × (D/2)²[18].

With exception of LVEF, measurements were averaged from at least three heart beats.

Brain natriuretic peptide

For analysis of BNP, blood samples were collected from patients after 30 min in a recumbent resting position. The samples were placed in chilled tubes, centrifuged and frozen at -70°C until plasma concentration was measured using a immunoradiometric method (Shionogi, Osaka, Japan) [19].

Statistical methods

Data was analysed using SPSS 11 for Windows (Chicago, IL). The differences between the three sessions were analysed with one way ANOVA (analysis of variance). Paired data (land vs WWI) were analysed with the Wilcoxon signed rank test. Relations of paired data were analysed with paired sample T-test. Coefficient of variation (CV) was calculated according to the formula: CV% = s*100/mean, where s (intra-patients error) = SD/√2 and mean = pooled mean value. Data are expressed as mean ± SD. A p-value < 0.05 was considered as statistically significant.

Acute effect of WWI (n = 18)

Systolic function

There was a significant hemodynamic response during WWI as reflected in several variables: cardiac output improved from 3.1 ± 0.8 to 4.2 ± 0.9 L/min, and stroke volume from 43.9 ± 13.6 to 64.4 ± 16.5 mL (both p < 0.0001); LVEF improved from 31 ± 9 to 35 ± 8% (p < 0.05); in addition, we found improved LAX function as LV TVTI increased from 1.2 ± 0.4 to 1.7 ± 0.5 cm and RV TVTI from 1.6 ± 0.6 to 2.5 ± 0.8 cm (both p < 0.0001, Figure 1A–B); left atrioventricular plane displacement (AVPD) increased from 5.5 ± 2.1 to 8.2 ± 2.7 mm (p < 0.01).

Volumes and filling pressure

Left ventricular end diastolic volume increased from 123 ± 36 to153 ± 56 mL, LV end systolic volume from 85 ± 32 to 100 ± 40 mL and estimated pulmonary capillary wedge pressure from 10.0 ± 5.8 to12.4 ± 5.2 mmHg (p < 0.01, respectively).

In three patients with left bundle branch block, we observed pronounced post-systolic contraction when the patients were investigated on land. During WWI, this abnormal contraction was abolished (Figure 2A–B). Furthermore, a more regular heart rate was observed (Figure 3A–B).

Eight weeks without and with hydrotherapy (n = 12)

The effect of WWI was similar during all three sessions with no significant differences after 8 weeks of hydrotherapy (Table 3 and Figure 4). There were no significant differences in BNP during the study period. The levels at baseline, after the control period and after 8 weeks of hydrotherapy were 169 ± 158, 134 ± 102 and 147 ± 125 ng/L, respectively. In addition, exercise and cardiopulmonary function were unchanged (data not shown).

	A	B	C
LV TVTI s (cm) land	1.3. ± 0.4	1.3 ± 0.4	1.3 ± 0.4
LV TVTI s (cm) WWI	2.0 ± 0.3**	2.0 ± 0.3**	1.9 ± 0.4**
LVEDV (mL) land	122 ± 40	114 ± 40	110 ± 41
LVEDV (mL) WWI	151 ± 59*	145 ± 56*	142 ± 41**
RV TVTI s (cm) land	1.8. ± 0.5	1.8 ± 0.5	1.9 ± 0.5
RV TVTI s (cm) WWI	2.8 ± 0.6**	2.9 ± 0.7**	2.8 ± 0.8**
PCWP (mmHg) land	9.5 ± 3.6	9.7 ± 4.0	10.1 ± 4.2
PCWP (mmHg) WWI	12.3 ± 6.0**	11.7 ± 4.7*	12.2 ± 4.4*
SVR (RU) land	30 ± 7	30 ± 5	30 ± 7
SVR (RU) WWI	21 ± 5**	23 ± 5**	21 ± 5**

Data are mean ± SD

*p < 0.05; **p < 0.01, land vs. warm water immersion.

There were no significant differences between the sessions.

A: Baseline, B: After 8 weeks of control period without changes in daily lifestyle,

C: After 8 weeks of hydrotherapy twice weekly.

Bpm, beat per minute; LV, left ventricular; LVEDV, left ventricular enddiastolic volume; PCWP, pulmonary capillary wedge pressure; RU, resistant unit; RV, right ventricle; SV, stroke volume; SVR, systemic vascular resistance; TVTI, tissue velocity time integral; WWI; warm water immersion.

Heart rate and ventricular load

The increase in LV volume was caused by hydrostatic pressure and, according to the Bainbridge reflex, would be expected to result in an increase in heart rate [20]. However, we observed a reduced heart rate. Considering the short time of WWI (20–30 min), this was most likely caused by activation of the autonomic nervous system through an enhanced parasympathetic activity [10]. Intravascular volume is controlled by complex systems that involve renal, autonomic, and neurohormonal activities [21–23]. The estimated decrease in afterload was accompanied by a decrease in MAP which is most likely caused by thermally mediated dilatation of skin vessels. A decrease in afterload and unloading of LV ejection is a beneficial aspect of WWI.

Long axis and diastolic function

We showed that WWI improved several echocardiographic indices of systolic and diastolic function. A lower heart rate alleviates ventricular filling and prolonged diastasis eases myocardial perfusion and is associated with improved systolic function [24]. In our study, effects on heart rate and diastolic function are important to explain why an increase in preload did not cause adverse cardiac effects in heart failure patients.

For many years, the M-mode technique was the established method to study LAX function. However, with the development of newer and relatively preload-independent techniques such as TDI and strain rate, the interest in measuring LAX function has increased dramatically. We have recently shown that LAX function is of major importance for long-term survival [25]. In this present study, we investigated LAX function using: 1) M-mode, which measures amplitudes of atrioventricular plane motion, and 2) TDI, which measures myocardial velocities. Although the M-mode technique showed a significant increase in systolic amplitude, the systolic velocity was unaffected when measured by TDI. This could be explained as TDI is less dependent on preload [26, 27]. Nevertheless, both RV and LV TVTI increased, which demonstrate enhanced LAX function. When TVTI is applied, amplitudes of myocardial movements are recorded, giving similar information as obtained with M-mode echocardiography. M-mode and TVTI are therefore closely connected. Compared with myocardial velocities, we found that AVPD with M-mode and TVTI increased significantly during WWI.

An abnormal post-systolic contraction in patients who suffered from left bundle branch block was shown to be normalized by WWI. Post-systolic contraction is a delayed ejection motion of the myocardium [28], a phenomenon related to ischemia [29] and intra-ventricular dyssynchrony [30].

Echocardiographic protocol

Our experimental protocol was comfortable for the patients compared with other studies where patients have been exposed to heart catheterization [31] or investigated with transesophagal echocardiography [6]. Echocardiography is not normally performed in a standing position, but this allowed us to study cardiac effects in a position that is commonly adopted during hydrotherapy.

Hydrotherapy

Exercise is an important complement to medical treatment, to prevent and reduce symptoms of heart disease, however, regular exercise is required to maintain a positive result [1]. Controlled trials have shown favourable effects after exercise in CHF, leading to improved physiological and psychological effects [2]. Nevertheless, in this present study we did not observe any improvement in cardiac function, exercise tolerance or peak VO₂ after 8 weeks of hydrotherapy. One explanation for this could be insufficient exercise intensity (40–70% of maximal heart rate reserve). It has been shown that hydrotherapy three times weekly improves functional capacity compared with a deterioration in patients randomized to a control group [4]. Unpredictably, patients in our study maintained their physical capacity during the control phase. One reason could be that they unconsciously became more active ahead of the training regime because of the extra attention they received.

Brain natriuretic peptide

Plasma concentrations of natriuretic peptides have been shown to be powerful predictors of mortality and disease severity in patients with CHF [32]. Long-term treatment with angiotensin-converting enzyme inhibitors and beta-blockers have been reported to reduce plasma levels of BNP [33, 34]. Despite optimal medical treatment, the patients in our study had raised BNP levels. In agreement with the measured unaltered cardiac function BNP remained unchanged after the hydrotherapy