Methods

Results

Table 1 shows the demographic characteristics for the 11 healthy women who completed the study. The mean cervical dilation for participants on admission was 4 cm. Labor duration ranged from 126 to 1294 min. No participant had analgesia during the study period, and no maternal or fetal infections or abnormal fetal heart rates were attributed to hydrotherapy. Of the 11 participants, 1 nulliparous woman received no analgesia and delivered within minutes after leaving the tub, 9 women received IV analgesia at approximately 10, 15, 17, 30, 60, 90, 120, 270, and 480 min after leaving the tub; 5 women subsequently received pitocin augmentation or epidural analgesia, and 4 received both; 9 women experienced spontaneous vaginal delivery and 2 had cesarean sections for failure to progress with cephalopelvic disproportion. Neonatal mean APGAR scores were 7.8 (range 6–9) at 1 min and 9 (range 9–10) at 5 min.

Table 1

Demographic Characteristics for Participants (N = 11)

When compared with pre-immersion levels (baseline), hydrotherapy was associated with a significant decrease in parturient anxiety at 15 min and 45 min. In the group as a whole there were no significant differences between pre-immersion and immersion pain levels (Table 2). When participants were separated into subgroups according to pre-immersion pain level, however, there were significant differences, with pain decreasing more for women with high levels of baseline pain (≥ 70 mm) than for women with low baseline levels (≤ 50 mm) after 15 and 45 min of immersion (Table 3).

Table 2

Mean (SD) for Anxiety, Pain, Oxytocin, Vasopressin, β-endorphin and Cortisol Levels at Baseline (0) and at 15 and 45 min of Hydrotherapy (Wilcoxon Signed Ranks Test)

Table 3

Comparison of Mean (SD) Changes in Pain and Cortisol Levels from Baseline (0) to 15 and 45 Min of Hydrotherapy in Low and High pre-Immersion (Baseline) Pain Groups (Mann-Whitney U Test)

We were not able to collect data on all neuroendocrine variables for all 11 women at all time points due to our inability to draw blood specimens. There were no differences between pre-immersion and immersion cortisol levels. However, cortisol levels decreased twice as much after 15 min of immersion for women with high baseline pain as for those with low baseline pain (Table 3). βE levels increased for the group as a whole at 15 min of bathing but did not significantly differ between baseline and 45 min. Hydrotherapy was associated with a significant decrease in O from baseline to 15 and 45 min of immersion. Plasma V also decreased significantly from baseline to 15 and 45 min of immersion (Table 2). Neither E nor NE levels changed significantly with hydrotherapy.

CX frequency decreased significantly (p = .04) with bathing. All women had a positive mean PVS% at 15 min (+4.8, n = 8) and at 45 min (+5.2%, n = 7). PVS% at 15 min of hydrotherapy was significantly positively correlated with CX duration (r = .74, p = .04). This was the only statistically significant relationship between study variables.

Discussion

Our findings that anxiety significantly decreased throughout hydrotherapy for the group as a whole and that pain decreased more for women with high baseline levels than for women with low baseline levels at both measurements are congruent with our previous report in a separate group of laboring women (Benfield et al., 2001) as well as the work of others (Cammu et al., 1994; da Silva et al., 2009; Eldor et al., 1992; Lenstrup et al., 1987). However, a floor effect may have existed in the low-pain subgroup because the lower limit of the pain scale was set at zero.

Decreases in pain were mirrored by decreases in C levels for women with high baseline pain levels. At 15 min into hydrotherapy, plasma C levels had decreased twofold in women who reported high pain prior to immersion. Decreases in C during immersion, at rest, and with exercise at 60% VO_2max have been noted in late pregnancy (McMurray, Katz, Berry, & Cefalo, 1988). Although C levels in this sample of parturients decreased during hydrotherapy, they remained higher throughout the study than those of college-age men with 4 years of resistance training who were subjected to high-intensity resistance exercise to evoke a maximal anticipatory and psychophysiological stress response (Kraemer et al., 2005). Their mean C level at 30 min pre exercise was approximately 525nmol/L, or 19.02ug/dl, and their peak level, at 10 min post exercise, was 935nmol/L, or 33.51ug/dl. The mean C levels of women in our study were 46.23 ug/dl pre immersion, 41.77 ug/dl at 15 min of immersion, and 40.30 ug/dl at 45 min of immersion. Even with the known hypercortisolism of pregnancy and labor, this comparison suggests that labor stress in women with high subjective pain is relatively severe and that the hydrotherapy intervention may be a highly effective way to provide relief from that stress.

While hydrotherapy was associated with a decrease in C levels, it failed to reduce either plasma E or NE levels in either the total group or either of the two pain subgroups. In a prior study with a separate sample of women in labor, we found that urine catecholamine levels did not decrease significantly with a hydrotherapy intervention (Benfield et al., 2001). While some researchers have hypothesized that a decrease in catecholamine levels is the mechanism underlying the effects of hydrotherapy (Odent, 1983), our current and previous findings as well as those of other researchers suggest that dampening of the sympathetic nervous system with resulting decreases in plasma catecholamines may not be the underlying mechanism for these effects (Robertson et al., 1998; Rush et al., 1996). While the sympathetic response is beneficial for coping with brief periods of stress (seconds to 30 min), it is likely not the best indicator for the prolonged stress of labor. Still, our results must be interpreted with caution due to the small sample size and the large individual variation in catecholamine response.

βE levels increased, though not significantly, with 15 min of immersion then decreased at 45 min to levels again not significantly different from those at pre immersion. Similarly, in one study, βE levels in late pregnancy increased during immersed exercise at 60% VO_2max above both dry and immersed resting levels (McMurray, Berry, & Katz, 1990). However, it is uncertain which aspects of labor stress, if any, are analogous to the stress of exercise. Few previous studies have specifically evaluated maternal pain reports and plasma βE levels in labor or the effect of pain-reduction interventions on these variables. Raisanen, Paatero, Salminen, and Laatikainen (1984) found that maternal pain and plasma βE levels increased concurrently across unmedicated labor increase and were positively, though not significantly, correlated. Bacigalupo, Riese, Rosendahl, and Saling (1990) found subjective pain and plasma βE levels to be highly correlated (p < .001) in parturients whose cervical dilation was near 6 cm. Other researchers have found that βE and subjective pain levels both decreased after epidural analgesia (McLean, Thompson, Zhang, Brinsmead, & Smith, 1994; Raisanen et al., 1984; Scull et al., 1998) and that changes in pain scores before and during epidural analgesia were correlated with βE levels (r = 0.84, p < .001; Raisanen et al., 1984).

The increasing plasma βE levels known to exist with advancing labor may reflect the physiological stress associated with effective labor progress in addition to parturient pain. Proopiomelanocortin- (POMC) derived peptides, including βE, are present in human myometrium (Clifton et al., 1998), and the POMC gene has been identified in the human placenta at term gestation (Grigorakis, Anastasiou, Dai, Souvatzoglou, & Alevizaki, 2000). These findings point toward mechanisms of action for βE that may be separate from pain control. Specifically, βE may play a local role in labor progress. Thus, it cannot be determined whether peripheral levels of βE during labor are reflective of its role in pain control or represent locally released βE that has participated more directly in labor. Thus, while βE may have a role in pain control, it is unclear what that role is or whether, in the current study, the higher concentration of βE at15 min was related to O inhibition or decreases in uterine contractility.

As expected based on previous research (Hammerum et al., 1998), V levels decreased significantly and remained decreased throughout hydrotherapy. Unexpectedly, O levels followed the same pattern, and the frequency of uterine contractions also decreased significantly. In vitro, human myometrium is significantly more sensitive to V than to O after the onset of labor (Thornton et al., 2002). In vivo, the exogenous infusion of synthetic arginine-vasopressin (1 unit intravenous) was found to “augment an expected contraction.” (Embrey & Moir 1967, p. 649). Because V acts at both O and V receptors on the myometrium, both hormones may be involved in the regulation of labor (Bossmar et al., 1994). We can, thus, speculate that the decrease in V and O concentrations during hydrotherapy was connected with the significant decrease in contraction frequency. We did not measure contraction intensity because of the infection risks associated with use of an intrauterine pressure catheter in water. Therefore, we do not know whether hydrotherapy-associated decreases in V and O levels had any effect on contraction intensity. Additionally, given the limited frequency of our sampling protocol, we may have missed the pulsatile release of O in labor (Fuchs et al., 1991; Otsuki, Yamaji, Fujita, Takagi, & Tanizawa, 1983). While the changes in V and O mimicked each other, they were not significantly correlated with each other or with subjective measures of anxiety or pain.

A positive PVS% with immersion to the chest has been shown to occur in as little as 15 min with immersed parturients (Benfield et al., 2001). As anticipated, all the women in our study had a positive PVS%, which increased slightly throughout the intervention. This finding is consistent with the findings of Katz and colleagues (1988), who investigated immersion in pregnancy, and with our prior work on hydrotherapy and labor (Benfield et al., 2001). However, the percent shift in the current study was less than that reported in our earlier study at 15 min (7.8%) and 60 min (9.6%) of hydrotherapy (Benfield et al., 2001). We can only hypothesize that slower tub filling (in the prior study they were immersed to chest upon entering the tub) and less immediate hydrostatic pressure may have resulted in the smaller amount of PVS%. In any case, our findings are congruent with the expected onset and positive direction of plasma volume shift, which are opposite the negative shift often seen with intense exercise on land.

The positive PVS% during hydrotherapy may have been associated with improved uteroplacental perfusion. Interestingly, we found that PVS% at 15 min of hydrotherapy was strongly and positively correlated with contraction duration. An inverse relationship between blood flow and contractility indicates that uterine smooth muscle is dependent on blood supply for normal metabolite levels and force production (Larcombe-McDouall, Harrison, & Wray, 1998). Perhaps as contraction frequency decreased, the perfusion between contractions increased, resulting in longer contraction duration. Theoretically, intensity should also increase in this case, but this remains to be studied.

Hydrotherapy has been reported to improve uterine dystocia and decrease the need for augmentation and epidural use (Cluett et al., 2004). Dystocia and dysfunctional labor are thought to be associated with more severe pain and to increase the likelihood of a request for epidural analgesia (Hess, Pratt, Soni, Sarna, & Oriol, 2000; Panni & Segal, 2003). Severe pain and lack of cervical dilatation beyond 5 cm are reported to improve with hydrotherapy (Odent, 1983). It seems reasonable to speculate that decreased contraction frequency and improved uterine profusion could facilitate a return to normal contraction synchrony, intensity and frequency, thereby decreasing pain and dysfunction. Thus, positive PVS% may merit further investigation as an intervention for labor dysfunction.

Finally, the effect of water temperature must be taken into account. Most immersion studies in men have been conducted at a thermoneutral temperature of 34–45°C because this has the least effect on the cardiovascular system. We selected a temperature of 37°C because it approaches 36°C, the most frequent mean temperature chosen by laboring women, and it is the most frequently used temperature in hydrotherapy studies in labor (Benfield, 2002; Schorn et al., 1993). The warmer temperature may affect endocrine levels as well as peripheral circulation, contributing to differences in study findings.

Study strengths include the within-subjects study design, which controls for the large individual variability in stress hormone baseline and response levels and provides considerably more power than would between-subjects designs. Strict adherence to study inclusion criteria during antepartal consent and at labor admission added rigor by controlling for extraneous variables known to affect labor pain and uterine contractility. It is particularly important that no participant received analgesia during study participation so that changes more clearly reflect the intervention effects rather than those of a narcotic. The 1 hr 30 min timeframe during early labor in which data were collected is most stable window of active labor in which to study the intervention effects.

Study results are limited by the short study period, which did not allow continued evaluation of the length of intervention effect. Additionally, we were unable to stratify findings based on demographics or parity because of the small sample size, although these factors affect both maternal perception of the experience and labor progress.

While our small sample size also limits interpretation of the findings, and the precise psychophysiological mechanisms of hydrotherapy remain unknown, our findings support the hypothesis that anxiety and pain decrease with bathing, particularly in women with high levels of pain. It may be that the central PVS% associated with immersion plays a role in reducing anxiety and pain. The effects of hydrotherapy are rapid, occurring within 15 min of a parturient entering the tub. This period of time may be adequate to allow for effects on the hypothalamic pituitary adrenal axis, as evidenced by decreases in cortisol levels. Although anxiety, vasopressin and oxytocin are temporally synchronized they are not statistically correlated in this sample. During hydrotherapy, uterine contraction frequency decreases, along with oxytocin and vasopressin, while contraction duration increases as plasma shifts into the vascular system.

As yet unknown modulatory relationships may exist between the complex psychophysiological processes in normal labor and the neuroendocrine responses to stress that are affected by hydrotherapy. Large, well-controlled randomized studies are needed, as along with additional smaller studies, to clarify the mechanisms of hydrotherapy action and its effects on labor progress. Knowledge gained from this and future studies can be directly applied in clinical practice by nurses, midwives and physicians caring for women in labor who desire noninvasive, low-risk yet efficacious anxiety and pain intervention.

Acknowledgments