Pilot Study Investigating the Effects of Ayurvedic Abhyanga Massage on Subjective Stress Experience

Abstract

Objectives:

Many providers of Ayurvedic massage treatments make unsubstantiated claims regarding their potential psychologic and physiologic benefits. While these claims are based on ancient traditions of the Indian medical system as well as personal experience, no attempt had yet been made to verify the effect of Ayurvedic oil massage within the Western-scientific context. This pilot study assessed the single-dose effect of Abhyanga, the classic Ayurvedic oil massage, in terms of its impact on subjective stress experience.

Design:

Ten (10) healthy women (mean [M] 36.4 years, standard deviation [SD] 4.62) and 10 healthy men (M 37.0 years, SD 3.23) underwent a 1-hour Abhyanga massage treatment. The study employed a repeated-measures design for the collection of stress data using a validated and published questionnaire, as well as for the collection of heart rate (HR) and blood pressure (BP) values, immediately pre- and postintervention.

Results:

Subjects showed high statistically and clinically significant reductions in subjective stress experience. A consequent analysis of physiologic data indicated likewise reductions in HR, while a BP reduction was not recorded for the total sample, but was only found in the prehypertensive subgroup.

Conclusions:

Findings indicate that Abhyanga massage is promising in reducing subjective stress experience. It may be beneficial in lowering HR in all, and BP in prehypertensive subjects. The results of this pilot study encourage further research with more stringent methodology.

Introduction

A bhyanga, the classic Ayurvedic oil massage, addresses the client's whole body with long, rhythmic, repetitive strokes, as well as circular movements around joints and important marma (pressure) points. A wide range of physiologic and psychologic benefits are traditionally claimed for this type of treatment.

In the only series of published studies assessing the impact of a form of Ayurvedic physical therapy, Uebaba et al. investigated the question of whether Shirodhara, the Ayurvedic oil-dripping treatment, had psychoneuroimmunologic effects.^1,2 The initial study concluded that it acted as an anxiolytic, and induced altered states of consciousness. This was followed by a second study where decreases in noradrenaline levels, as well as sympatholytic treatment effects, were found.

Several published studies assess the effects of “Western” massage modalities. Both physiologic and psychologic parameters were recorded; most studies found decreases in heart rate (HR)^{*,3,4,5,6,7,8} and in blood pressure (BP),^{*,3,4,5,6,7,8} and reductions in anxiety levels,^*,3,4,5,9 depression scores,⁹ salivary cortisol levels,^*,9 and stress levels, as well as improvements of mood.¹⁰

The article by Moyer et al. combined 37 primary studies, and concluded that massage treatments as such are effective in single-dose as well as multiple-dose applications.¹¹ Single-dose treatment analysis verified the ability of massage to reduce state anxiety and lower HR and BP. Contrary to some individual studies' results, possible effects on cortisol levels and mood could not be confirmed. Looking at multiple-dose applications, the authors additionally substantiated the effectiveness of massage in lowering trait anxiety and depression scores.

It can be concluded that while the effects of various “Western” styles of massage treatments are thus well documented, there have so far been no studies evaluating Ayurvedic massage at a level that satisfies the quality requirements of scientific investigation.

Textbooks as well as individual providers' advertising materials claim various far-reaching benefits of Ayurvedic massage treatments. While these claims rest upon the oral and written traditions of the Indian medical system, as well as on individual practitioners' observations over the years, the lack of scientific proof in the Western sense has so far been impeding Ayurveda's success in the West. Both health insurance providers and the majority of clients in these parts rely on proven facts when evaluating the potential benefits of treatment modalities. For this reason, it is of the greatest importance to anchor Ayurvedic knowledge, also regarding massage therapy, within the context of the scientific community.

This study aimed to fill the existing knowledge gap by investigating one of the recurring claims of Ayurvedic massage, namely, its ability to reduce stress. The following main hypothesis was established: A 1-hour classic Abhyanga treatment on healthy adults will result in a significant reduction in subjective stress experience. Should this hypothesis prove true, further subhypotheses should be investigated: Is this reduction in stress levels accompanied by reductions in HR and BP, as prior research on Western massage techniques suggests?

The purpose of this study was, for the first time within the Western-scientific context, to investigate some of the prominent physiologic and psychologic effects of Ayurvedic Abhyanga massage.

Materials and Methods

Sample

As this study was most interested in large effect sizes, a sample size of 20 test subjects was chosen in order to allow for the verification of large effects at 5% level of significance and power of 80% within a repeated-measures design. Subjects (Ss) were recruited by e-mail from the author's local circle of friends and acquaintances, where the only criterion for exclusion was an insufficient knowledge of the German language. The first 10 women (mean [M] 36.4 years, standard deviation [SD] 4.62, minimum 31, maximum 45) and 10 men (M 37.0 years, SD 3.23, minimum 32, maximum 41), all of them healthy by self-report, who responded to the e-mail were included in the study. They provided written informed consent. The ethical standards of the study are in accordance with the declaration of Helsinki developed by the World Medical Association (WMA) in 1964. Campus Naturalis Germany, the school at which the author was enrolled at the time of conducting this research, follows the ethical guidelines of the WMA.

Setting and implementation

All treatments were conducted over a period of 12 consecutive days in late March 2010, with a maximum number of three treatments per day. All conversations, massages, and the filling out of the questionnaires took place in the living room of the author's apartment, with all instructions and massages given by the author.

After a 5-minute instruction and familiarization session, subjects filled out an assessment instrument frequently used in Germany called Short Questionnaire Assessing Current Strain (KAB).¹² The KAB defines strain as an aspect of the current state of mind, and as a one-dimensional bipolar construct. It can be utilized to control short-term changes in subjective stress experience. After several stages of development, the questionnaire consists of six bipolar adjectives (tense–relaxed, mellow–uneasy, carefree–worried, calm–agitated, trusting–skeptical, comfortable–uncomfortable).^† The validity of this form has been verified over the course of several studies, and it is further indicated by the parallel run of physiologic parameters, and its high correlation with the related Spielberger State-Trait Anxiety Inventory (STAI) state questionnaire, as well as its prognostic value when it comes to the use of analgesics. Its internal consistency and the retest–coefficients in situations of comparable strain (from r = 0.72 to r = 0.91) support its reliability. Completion takes between 1 and 2 minutes.

Then Ss undressed, and could chose between wearing a disposable g-string (n = 6) and full nudity (n = 14), as is Ayurvedic tradition. They were then asked to lie down on the heated massage table in the prone position, and were covered with a blanket. After a 5-minute rest period, base readings for HR and BP were recorded, using an automatic device produced by Aponorm/Microlife, the blanket was removed, and the massage treatment began. The Abhyanga was performed as planned, with Ss adopting the seven classical positions (sitting, prone, side 1, prone, side 2, prone, sitting). A detailed description of this Kerala-style massage using long, medium-to-firm strokes away from the navel, predominantly in repeats of 7, can be found in Diehl et al.¹³ Even though traditionally medicated oils are used for this treatment, in order to reduce variance approximately 150 mL of warm, cured sesame oil, with a 10% addition of warm, pure coconut oil were used for all Ss.

After treatment, Ss rested in the prone position for about 5 minutes, at which point HR and BP values were again recorded. Then, and while still sitting on the massage table, Ss filled out the KAB questionnaire once more.

Statistical analysis

Statistical analysis relied on SPSS (Statistical Package for the Social Sciences) for Windows, Release 15.0. Statistic significance of changes between pre- and post-treatment was assessed via paired t tests. A simple analysis of variance was applied to evaluate differences between independent samples. Significance levels were set as p < 0.05. Effect sizes (ES) were calculated by relating pre–post differences to the SD post-treatment. ES was considered small at <0.2, medium at <0.5, and large at <0.8.

Results

Main hypothesis: Influence on subjective stress experience

The question was whether data collected post treatment would indicate a reduction in subjective stress experience. Average pretreatment values of the KAB are equivalent to those found in healthy blood donors after a period of rest or with employees of a company after participation in a health-promotion program.¹²

KAB values gathered prior to and post-treatment were compared, bearing in mind that an increase in values represented a decrease in stress experience. The averages of the total scores as well as individual item values are represented in Table 1. Note the high initial item values for items 3 and 5.

Table 1.

KAB Total Score and Individual Item Values

	M₁ (SD₁)	M₂ (SD₂)	n	Statistical significance	Clinical significance
KAB _{sum values}	22.95 (4.80)	32.10 (3.74)	20	<0.01	1.906
KAB _{item 1}	3.30 (1.08)	5.45 (0.76)	20	<0.01	2.074
KAB _{item 2}	3.40 (0.99)	5.00 (1.08)	20	<0.01	1.616
KAB _{item 3}	4.05 (1.43)	5.05 (0.76)	20	<0.01	0.699
KAB _{item 4}	3.30 (0.92)	5.65 (0.67)	20	<0.01	2.554
KAB _{item 5}	4.70 (1.49)	5.45 (0.83)	20	<0.01	0.503
KAB _{item 6}	4.20 (1.01)	5.50 (0.69)	20	<0.01	1.287

KAB, Short Questionnaire Assessing Current Strain; M₁, mean value pretreatment; SD₁, standard deviation pretreatment; M₂, mean value post-treatment; SD₂, standard deviation post-treatment, n, sample size.

It was found that the cumulative KAB values increased very significantly and with extremely high ES (t = 8.739; df = 19; p < 0.001; ES = 1.906), indicating a substantial decrease of strain.

Results for individual KAB items were also analyzed, and individual item results can be found in Table 2. It was found that all item values changed with the highest possible statistical significance while displaying medium to very high ES.

Table 2.

KAB Individual Item Results

	t-Value	df	p-Value	ES
KAB item 1 (tense–relaxed^a)	9.731	19	0.000	2.074
KAB item 2 (mellow–uneasy)	5.007	19	0.000	1.616
KAB item 3 (carefree–worried)	3.082	19	0.006	0.699
KAB item 4 (calm–agitated)	11.261	19	0.000	2.554
KAB item 5 (trusting–skeptical)	3.470	19	0.003	0.503
KAB item 6 (comfortable–uncomfortable)	4.772	19	0.000	1.287

All translations by the author.

KAB, Short Questionnaire Assessing Current Strain; df, degrees of freedom; ES, effect size.

An exploratory subdivision of the total sample into gender groups (n = 10/10) was performed. Gender-specific mean values are shown in Table 3.

Table 3.

KAB According to Gender

	M₁ (SD₁)	M₂ (SD₂)	n	Statistical significance	Clinical significance
Sample KAB_Women	22.40 (4.77)	33.80 (1.48)	10	<0.01	2.390
Sample KAB_Men	23.50 (5.02)	30.40 (4.58)	10	<0.01	1.375

A paired t test produced the following results for the group of women: (t = 7.386; df = 9; p < 0.001; ES = 2.390), while for the group of men the following results were found: (t = 6.517; df = 9; p < 0.001; ES = 1.375). The difference in ES was found to be remarkably high.

A simple analysis of variance showed no significant difference between groups prior to treatment (t = 0.502; df = 18; not significant [n.s.]), and values post-treatment displayed a significant difference at 5% level (t = 2.234; df = 18; p < 0.05).

As all results indicate that a classic Abhyanga massage treatment is very promising in reducing the subjective experience of stress, two subhypotheses dealing with physiologic indicators of stress were also investigated.

Subhypotheses: Influence on HR and BP

It was to be examined whether HR values would decrease as a result of the treatment. Mean HR values are shown in Table 4.

Table 4.

Heart Rate Total Sample

	M₁ (SD₁)	M₂ (SD₂)	n	Statistical significance	Clinical significance
Sample HR _total	63.40 (10.68)	59.65 (9.78)	20	<0.05	0.351

M₁, mean value pretreatment; SD₁, standard deviation pretreatment; M₂, mean value post-treatment; SD₂, standard deviation post-treatment, n, sample size; Sample HR _total, heart rate values for total sample.

Results obtained using a paired t test indicate that mean HR decreased significantly at the 5% level between the two lists of measurements, while showing a small-to-medium ES: (t = 2.437; df = 19; p < 0.05; ES = 0.351).

It was further to be investigated whether the intervention would significantly reduce systolic and diastolic BP readings. The data derived from the two lists of measurements are shown in Table 5.

Table 5.

Blood Pressure Total Sample

	M₁ (SD₁)	M₂ (SD₂)	n	Statistical significance	Clinical significance
Sample Sys _total	120.30 (14.19)	117.65 (11.78)	20	—	0.187
Sample Dia _total	70.15 (10.04)	69.25 (8.35)	20	—	0.090

M₁, mean value pretreatment; SD₁, standard deviation pretreatment; M₂, mean value post-treatment; SD₂, standard deviation post-treatment, n, sample size; Sample Sys _total, systolic blood pressure (BP) values for total sample; Sample Dia _total, diastolic BP values for total sample.

An assessment of the changes between lists of systolic values using a paired t test gave the following result: (t = 1.658; df = 19; n.s.; ES = 0.187), while the following results were obtained for diastolic BP: (t = 0.844; df = 19; n.s.; ES = 0.090). It is evident that for the total sample both systolic and diastolic BP values changed in a manner that was neither statistically nor clinically significant.

As no significant change in BP could be found for the total sample, an exploratory subdivision of the sample into normotensive and prehypertensive Ss according to the Joint National Committee's (JNC 07) definition was performed. The JNC 07 postulated that systolic BP values between 120 mm Hg and 139 mm Hg, as well as diastolic values between 80 mm Hg and 89 mm Hg, should be understood to be prehypertensive.¹⁴ This subdivision resulted in two subgroups: normotensive (n = 9) and prehypertensive (n = 11) Ss. The values for these groups are shown in Table 6.

Table 6.

Blood Pressure Normotensive/Prehypertensive Sample

	M₁ (SD₁)	M₂ (SD₂)	n	Statistical significance	Clinical significance
Sample Sys _norm.	107.00 (6.82)	108.11 (9.35)	9	—	0.163
Sample Sys _pre-hyp.	131.18 (7.49)	125.45 (6.71)	11	<0.01	0.770
Sample Dia _norm.	61.44 (4.77)	62.67 (2.45)	9	—	0.256
Sample Dia _pre-hyp.	77.27 (7.03)	75.00 (7.80)	11	—	0.323

M₁, mean value pretreatment; SD₁, standard deviation pretreatment; M₂, mean value post-treatment; SD₂, standard deviation post-treatment, n, sample size; Sample Sys _norm., systolic BP values for normotensive sample; Sample Sys _pre-hyp., systolic BP values for prehypertensive sample; Sample Dia _norm., diastolic BP values for normotensive sample; Sample Dia _pre-hyp., diastolic BP values for prehypertensive sample.

The normotensive subgroup continued to show no significant statistical change of the systolic reading (t = 0.425; df = 8; n.s.; ES = 0.163), but a small clinically significant increase of the diastolic values (t = 0.682; df = 8; n.s.; ES = 0.256). The systolic BP readings of the prehypertensive subgroup were significantly reduced at the 1% level, and displayed medium to high ES: (t = 3.838; df = 10; p < 0.01; ES = 0.77). For this group, the reduction in diastolic values was clinically, but not statistically significant (t = 1.972; df = 10; n.s.; ES = 0.323).

Discussion and Conclusions

Summary of results

The main hypothesis regarding a reduction in stress experience post-treatment was supported by a very significant increase in cumulative KAB values as well as significant value changes across all KAB items, while women's values increased more strongly than men's.

A reduction in HR as a result of treatment was found, so that subhypothesis 1 could be supported.

BP values for the entire sample did not produce a significant change as a result of treatment. So while subhypothesis 2 could not be supported, a very strong reduction in systolic BP values was recorded for the prehypertensive subgroup.

Interpretation: Main hypothesis

The reductions in subjective stress experience found in this study deserve attention, not so much because of their existence alone, as this had to have been expected from all previously published study results,^*,3,4,5 but because of the unexpectedly high statistical and clinical significance of these reductions.

Possible confounds

As all treatments and tests were conducted and evaluated by the author, Ss may have felt biased toward answering in a manner they felt supported the desired outcome. It also requires mention that some Ss were personally known to the author prior to the study (n = 14), while others were not (n = 6), and that this may have influenced individual Ss' feeling of trust as expressed in the KAB. Furthermore, Ss differed in their previous experience with Ayurveda massage. Two (2) of them reported substantial prior experience, 5 of them reported a very limited number of previous treatments (<2) of which 3 Ss had been treated by the same masseur as conducted the present study, and 13 of them reported no prior experience. The principal incentive was an expected increase of well-being.

While there is reason to believe that women profited more from treatment than men did, in that their subjective experience of stress appeared to have been reduced to a much greater degree, it must be considered that all treatments were administered by a female masseur. The greater stress-reducing effect on women may have had to do with women's greater general ease, also with regards to nudity, around a female therapist. Such gender-dependent situational ease may well have impacted on subjective stress experience results.

A further distorting aspect in this context may have been the audio-feedback caused by the so-called “filling of the ears,” a practice whereby cotton balls soaked in warm oil are inserted into Ss' ears within the first 10 minutes of treatment. This measure is an integral part of the Abhyanga treatment, and therefore cannot be considered a separate treatment measure. Nevertheless, several studies have shown audio-feedback itself to be inductive of a relaxation response.^15,16,17,18 Therefore, the comparability of this study with other massage studies that investigate manual manipulation alone is somewhat limited.

Finally, the heating of the massage table may have to be considered a confounder in the analysis of the stress-reducing impact of the manual treatment. Literature agrees that an increase of hand or foot temperature can be used as an indicator of parasympathetic activation.^{2,13,14,19,20,21} Therefore, it is possible that a causal connection exists also in reverse, and that an increase in body temperature may lead to a parasympathetic reaction, involving physical and mental relaxation.

Further exploration of results

Against the background of Ayurvedic medicine, an analysis of results according to Dosha (i.e., Ayurvedic constitution type) subgroups suggested itself. A not-yet-published objective test with high expert validity served to allocate the Ss to three subsamples. As subsamples were too small in this study to arrive at conclusive results, the following observations indicate trends only but are nevertheless very interesting in the context of Ayurvedic theory, and deserve further exploration in future studies.

Detailed analysis according to Dosha subgroups indicated that while all groups experienced a strong relaxation response, there were large differences in ES between groups. The group of Vata Ss showed the clinically most significant decrease in stress experience, while Pitta Ss recorded the lowest decrease by far, and marked a noteworthy low mean value in KAB item 2 (uneasy) post-treatment. Kapha Ss felt most relaxed overall, both before and after treatment, and marked notably higher mean values in KAB items 3 (carefree) and 5 (trusting) before treatment than other Dosha groups.

Interpretation: Subhypotheses

The reduction in HR found in this study was as expected. Along with all published study results, the present findings underpin the general consensus that a reduction in HR can be considered a statistically firm effect of any type of massage treatment.

The lack of change in BP values, on the other hand, came as a surprise, as it contradicts the findings of the majority of previous studies. Interestingly though, in the only other study that also assessed the effects of a 1-hour massage, Braziel was unable to find a change in BP values. It could be hypothesized that BP values decrease initially (durations of treatment in previous studies ranged between 3 and 30 minutes) during a treatment session but return to their initial values after a certain amount of treatment time. One other previous study, though, did not record a change in BP values from a single reading already after only 20 minutes of treatment time.¹⁰

Regarding the subdivision into normotensive and prehypertensive Ss, the contradictory nature of previous studies' results does not allow a conclusive contextual placement of the results found here. It has to be assumed that the parameters of massage methods, durations, and repeats have altogether been too widely spread across individual studies to allow for universal conclusions.

From the fact that the diastolic BP values of the normotensive group increased slightly, while the values of the prehypertensive group decreased slightly for diastolic, and even greatly for systolic BP in this study, it may be deduced that BP values for all massage Ss were changed toward individually healthy levels that—when averaged across the entire sample—manifested as an absence of change.

Future research

Due to the self-financed nature of this pilot study, some confounding factors were unable to be eliminated. The results and trends found in this study need to be re-assessed using a control group and a larger sample size that will allow for significant results to be conclusive also within subsamples. Readers are reminded that the sample size was chosen to statistically prove significant value changes between two lines of measures. Test power was not sufficient to statistically establish value changes within subgroups (BP/gender/Doshas), as for this purpose a larger sample size would have been required.

An extension of the study to include an investigation of the multiple-dose effect of the Abhyanga massage would be of interest. Traditional Ayurvedic literature recommends regular Abyangas as a preventative medicine measure for the general population. Therefore, a study design where massages are performed and assessed at regular intervals over the course of several months, including a treatment-free follow-up period, would be pertinent. Such a design might intend to find out whether the multiple-dose effect of Ayurvedic Abhyanga massages matches the effect Moyer et al. found for the administration of massages in general: Is it possible that regularly performed Ayurvedic Abhyanga massages also reduce trait anxiety and depression scores to a degree comparable with the effects of psychotherapy?¹¹ The impressive effect on stress experience found in this study appears to support such a hypothesis.

Footnotes

Acknowledgments

The author would like to thank Heinz-Dieter Basler, PhD, for his academic advice and support during the development and composition of this study.

Disclosure Statement

No competing financial interests exist.

*

Braziel A. The physiologic and psychological effectiveness of massage therapy in the management of stress, anxiety and depression. Dissertation Abstracts International: Section B: The Sciences and Engineering. 2002;63(6-B):3058.

†

All translations are by the author.

References

Uebaba

, Xu

, Tagawa

et al. Using a healing robot for the scientific study of shirodhara. IEEE Eng Med Biol Mag, 2005; 24:69–78.

Uebaba

, Xu

, Ogawa

et al. Psychoneuroimmunologic effects of Ayurvedic oil-dripping treatment. J Altern Complement Med, 2008; 14:1189–1198.

Fernandez-Perez

, Peralta-Ramirez

, Pilat

, Villaverde

. Effects of myofascial induction techniques on physiologic and psychologic parameters: A randomized controlled trial. J Altern Complement Med, 2008; 14:807–811.

Mok

, Woo

. The effects of slow-stroke back massage on anxiety and shoulder pain in elderly stroke patients. Complement Ther Nurs Midwifery, 2004; 10:209–216.

Ferrel-Torry

, Glick

. The use of therapeutic massage as a nursing intervention to modify anxiety and the perception of cancer pain. Cancer Nurs, 1993; 16:93–101.

Meek

. Effects of slow stroke back massage on relaxation in hospice patients. Image J Nurs Scholarsh, 1993; 25:17–21.

Olney

. The effect of therapeutic back massage on hypertensive persons: A preliminary study. Biol Res Nurs, 2005; 7:98–105.

Longworth

JCD

. Psychophysiological effects of slow stroke back massage in normotensive females. ANS Adv Nurs Sci, 1982; 4:44–61.

Field

, Ironson

, Scafidi

et al. Massage therapy reduces anxiety and enhances EEG pattern of alertness and math computations. Int J Neurosci, 1996; 86:197–205.

10.

Goodfellow

. The effects of therapeutic back massage on psychophysiologic variables and immune function in spouses of patients with cancer. Nurs Res, 2003; 52:318–328.

11.

Moyer

, Rounds

, Hannum

. A meta-analysis of massage therapy research. Psychol Bull, 2004; 130:3–18.

12.

Basler

, Müller

. Short Questionnaire Assessing Current Strain (KAB) [in German] Gõttingen: Germany Hogrefe, 1993.

13.

Diehl

, Kiel

. Classic Ayurveda Massage [in German] Neuhausen, Switzerland: Urania, 2005; 63–85.

14.

Chobanian

, Bakris

, Black

et al. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: The JNC 7 report. JAMA, 2003; 289:2560–2572.

15.

Bertilson

, Bartz

, Zimmerman

. Treatment program for borderline hypertension among college students: Relaxation, finger temperature biofeedback, and generalization. Psychol Rep, 1979; 44:107–114.

16.

Kappes

. Sequence effects of relaxation training, EMG, and temperature biofeedback on anxiety, symptom report, and self-concept. J Clin Psychol, 1983; 39:203–208.

17.

Pekala

, Wenger

, Levine

. Individual differences in phenomenological experience: States of consciousness as a function of absorption. J Pers Soc Psychol, 1985; 48:125–132.

18.

Vaitl

, Gruzelier

, Jamieson

et al. Psychobiology of altered states of consciousness. Psychol Bull, 2005; 131:98–127.

19.

McFarland

, Kadish

. Sex differences in finger temperature response to music. Int J Psychophysiol, 1991; 11:295–298.

20.

Kibler

. Effects of progressive muscle relaxation and music on stress as measured by finger temperature response. J Clin Psychol, 1983; 39:213–215.

21.

Vasilos

, Hughes

. Skin temperature control: A comparison of direct instruction, autogenic suggestion, relaxation, and biofeedback training in male prisoners. Correct Soc Psych Behav Tech Methods Ther, 1979; 25:119–122.