Yoga in Arterial Hypertension
A three-armed, randomized controlled trial
; ; ;
Background: Yoga seems to exert its effect against arterial hypertension mainly through the associated breathing and meditation techniques, and less so through yoga postures. The goal of this trial was to compare the blood pressure–lowering effect of yoga interventions with and without yoga postures in patients with arterial hypertension.
Methods: 75 patients taking medications for arterial hypertension (72% women, mean age 58.7 ± 9.5 years) were randomized into three groups: a yoga intervention group with yoga postures (25 patients, of whom 5 dropped out of the trial before its end), a yoga intervention group without yoga postures (25 patients, 3 dropouts), and a wait list control group (25 patients, one dropout). The interventions consisted of 90 minutes of yoga practice per week for twelve weeks. The data collectors, who were blinded to the intervention received, assessed the primary outcome measures “systolic 24-hour blood pressure” and “diastolic 24-hour blood pressure” before and after the intervention. In this report, we also present the findings on secondary outcome measures, including follow-up data.
Results: After the intervention, the systolic 24-hour blood pressure in the yoga intervention group without yoga postures was significantly lower than in the control group (group difference [Δ]= −3.8 mmHg; [95% confidence interval (CI): (−0.3; −7.4) p = 0.035]); it was also significantly lower than in the yoga intervention group with yoga postures (Δ = −3.2 mmHg; 95% CI: [−6.3; −0.8]; p = 0.045). Diastolic blood pressures did not differ significantly across groups. No serious adverse events were encountered in the course of the trial.
Conclusion: In accordance with the findings of earlier studies, we found that only yoga without yoga postures induced a short-term lowering of ambulatory systolic blood pressure. Yoga is safe and effective in patients taking medications for arterial hypertension and thus can be recommended as an additional treatment option for persons in this category.
The average blood pressure of the German population is decreasing, and a growing proportion of those with hypertension are receiving treatment. Nevertheless, arterial hypertension remains a serious health problem (1). Across the world, 9.4 million deaths each year are caused by excessively high blood pressure. The World Health Organization has pinpointed hypertension as one of the principal causes of premature morbidity and mortality in both developed and developing countries (2).
International medical guidelines recommend lifestyle changes for patients with arterial hypertension, e.g., physical activity and stress management (3). Yoga can comprise such a lifestyle-modifying intervention (4, 5). In western societies yoga is generally associated with specific body postures as well as breathing and/or meditation techniques (4, 6); various schools have developed with differing degrees of emphasis on the physical and mental aspects of yoga (6).
Three fourths of German residents who practice yoga state that they do so primarily for health reasons (7), and clinical studies have shown that the positive effects include a beneficial influence on physical and mental risk factors for hypertension (8, 9). However, no studies have yet been carried out in this country to ascertain the effect of yoga on hypertension, and the German guidelines do not take account of the existing evidence from other countries regarding yoga and related interventions (10). For example, a recent meta-analysis of studies on arterial hypertension and prehypertension performed in different countries (11) showed that yoga achieved significant reductions in systolic and diastolic blood pressure. In subgroup analyses, however, these effects were confirmed only for arterial hypertension. In studies that included both prehypertensive and hypertensive patients, the effects did not attain significance (mean difference: 0.9 mmHg systolic and 0.1 mmHg diastolic). Moreover, effects were demonstrated in studies that permitted individual antihypertensive co-medication, but not in those that excluded it (11). Finally, yoga was effective only in studies that included breathing and/or meditation techniques but no yoga postures. Studies that included yoga postures (alone or accompanied by breathing and meditation techniques) showed no significant effects (11). However, we are not aware of any studies in which the blood pressure–lowering effects of different yoga styles were compared directly. We therefore set out to investigate the efficacy of yoga interventions with and without yoga postures—as an accompaniment to antihypertensive medication—in lowering the 24-h blood pressure of patients with arterial hypertension. Our hypothesis was that the systolic and diastolic blood pressure would be lower following an intervention without yoga postures than in the absence of any specific intervention or after an intervention with yoga postures.
The study was carried out as a three-armed, single-center, randomized controlled trial between May 2016 and April 2017. Data acquisition ensued in blinded fashion. Study conduct adhered to the tenets of the Helsinki Declaration (12), and the report was written in conformance with the CONSORT 2010 Statement (13). Before the commencement of patient recruitment, approval was obtained from the ethics committee of the University of Duisburg–Essen (project number: 15–6726-BO) and the study was registered at ClinicalTrials.gov (registration number: NCT02727140).
Participants were recruited by means of appeals in the local press and via the e-mail distribution lists of local companies. Patients who came forward and fulfilled the inclusion criteria were recruited. They received detailed written and verbal information about the study before signing a consent form to indicate their agreement to participate.
To be included in the study, patients had to be at least 18 years old with primary arterial hypertension (>140 mmHg systolic and/or >90 mm Hg diastolic blood pressure). Moreover, they had to be receiving antihypertensive medication from their primary care physician or specialist at the time of investigation. Furthermore, they had to agree to the following conditions:
- No alteration of their antihypertensive medication in the next 7 months
- Participation in at least 10 of the 12 yoga sessions in the study program
- Additional unsupervised yoga sessions at home.
The reasons for exclusion included the following diseases and circumstances:
- Secondary hypertension
- Severe psychiatric comorbidities (major depression, dependency disorders, or psychosis)
- Coronary heart disease, myocardial infarction, pulmonary embolism, or stroke in the previous 3 months
- Heart failure of NYHA stage ≥ I
- Peripheral arterial occlusive disease of stage ≥ 1
- Renal failure of stage >2 with glomerular filtration rate (GFR) <60 mL/min/1.73 m²
- Participation in any other clinical studies at the time of commencement of our trial or planned participation in such studies in the next 28 weeks
- Pregnancy or breastfeeding.
The patients were stratified by sex and divided by block randomization with randomly varying block sizes into the three following groups (ratio 1 : 1 : 1):
- An intervention with yoga postures
- An intervention without yoga postures
- A wait list control group.
A biometrician who was not involved in patient recruitment or data acquisition used Random Allocation Software (14) to generate a password-protected randomization list to which only he had access. On the basis of this randomization list, opaque envelopes containing the study group allocation were numbered consecutively and sealed. Once a patient had signed the consent form and baseline data acquisition had been concluded, the study physician opened the envelope with the lowest remaining serial number and the patient was allocated to a group accordingly. The Cochrane Handbook for Systematic Reviews of Interventions describes this as a randomization method with a low risk of bias (15). At the end of the study, the biometrician checked the group allocations and found that all patients had been allocated correctly and according to the randomization list.
The two interventions were carried out by the same two yoga instructors. Each intervention comprised a 90-min session once each week for 12 weeks. In the intervention that included yoga postures, each session consisted of 45 min yoga postures and 45 min breathing, meditation, and relaxation techniques as well as short presentations and question-and-answer rounds. The yoga intervention without yoga postures consisted of only breathing, meditation, and relaxation techniques, short presentations, and question-and-answer rounds (eSupplement).
The study participants were encouraged to practice yoga and meditation every day at home, but no minimum duration was recommended. The exercises to be performed at home were explained in detail in the yoga sessions. Furthermore, the patients were given audio recordings of the yoga sessions and training manuals to support their unsupervised exercises. During the 12-week intervention period the participants noted in a diary how many minutes they spent on their exercises each day. At the end of the study (week 28), the patients were asked to state the average weekly duration of their exercises during the follow-up period.
The study participants allocated to the control group were put on a waiting list. They did not take part in any yoga interventions in the first 12 weeks of the study and were asked not to begin any yoga exercises or other new physical activities during this time. At the end of the 28th week of the study, the members of the control group were offered the opportunity to participate in an intervention corresponding to either of the other two study groups.
The person who acquired the data was not involved in patient recruitment, group allocation, or the interventions and was blinded to group allocation for the whole duration of the study. Systolic and diastolic 24-h blood pressure were defined as primary outcome measures and were determined using an internationally accepted digital blood pressure monitor (Mobil-O-Graph PWA, I.E.M., Stolberg, Germany) (16). The systolic and diastolic blood pressure at week 28 were documented as secondary outcome measures. The first measurement, at week 1, took place within the week immediately preceding the beginning of the intervention and the second measurement, at week 12, was performed in the week after the end of the intervention. The blood pressure measurement started at the same time of day at all three documentation points. Incorrect measurements were automatically deleted by means of standardized algorithms in the software of the blood pressure monitor. The secondary outcome measures are described in full in the eSupplement.
All adverse events that occurred during the study period were recorded and classified according to the criteria of the U.S. Food and Drug Administration (17). Any participant who experienced an adverse event was asked to consult the study physician so that he/she could assess the severity of the event and take any necessary steps. At weeks 12 and 28 the participants were asked open questions to document any previously unmentioned adverse events. The patients were requested to report all adverse events that occurred during the study period, whether or not they had any possible connection with the study intervention.
Sample size calculation and statistical analysis
The necessary sample size was calculated a priori based on a meta-analysis according to which the systolic blood pressure was 14.13 mmHg lower following yoga interventions, corresponding to an effect size of d = 1.35 (11). On the basis of this effect, 17 patients per study group would be needed for a t-test with a two-sided significance level of α = 0.05. This would enable detection of a corresponding group difference with 80% power. To rule out the loss of power that would result from a drop-out rate of up to 30%, 75 patients (25 per group) were included in the study.
The outcome criteria were evaluated on the basis of the intention-to-treat principle: all randomized patients were included in analysis, regardless of whether the data were complete or whether the study protocol was followed correctly. Missing values, i.e., variables for which there were no data for a given patient at a particular measurement time, were multiply imputed by means of the Markov chain Monte Carlo procedure (18, 19). This resulted in a total of 50 complete data sets that each contained full data for all 75 patients and were combined for purposes of analysis.
All linear outcome measures (blood pressure, heart rate, questionnaire data) were evaluated by means of univariate analyses of covariance (ANCOVA). The outcome measure (dependent variable) was modeled as function of the following parameters:
- Group allocation (independent variable)
- Expectation of effect (covariate)
- The respective baseline value (covariate).
This enabled determination of the overall effect estimates, the 95% confidence interval, and the p value. ANCOVA are recommended for evaluation of clinical studies because they control for possible baseline differences and ascertain group differences for all group comparisons and all measurement times. Moreover, they possess greater power than other statistical methods (20).
All analyses were carried out using Statistical Package for Social Sciences software (IBM SPSS Statistics for Windows, Version 22.0, Armonk, NY, USA: IBM Group). A p value <0.05 was interpreted as showing a significant difference.
Seventy-five patients fulfilled all the inclusion criteria and were included in the study after providing written informed consent. They were assigned randomly to the yoga group with yoga postures (n = 25), the yoga group without yoga postures (n = 25), or the control group (n = 25). Five members of the group with yoga postures (20.0%), three of the group without yoga postures (12.0%), and one patient in the control group (4.0%) did not complete the study program (Figure 1; p = 0.22). Therefore, missing values were multiply imputed for six participants at week 12 and nine patients at week 28. At each of these time points missing blood pressure values had to be imputed for three further patients who felt uncomfortable with the measuring device and broke off the measurement early. The patients’ characteristics are shown in Table 1.
Ambulatory blood pressure measurement
At 12 weeks the systolic 24-h blood pressure (a primary outcome measure) was significantly lower in the group without yoga postures than in the group with yoga postures (p = 0.045; Figure 2; eTable 1) or the control group (p = 0.035). At 28 weeks the reverse was found: the systolic 24-h blood pressure in the group with yoga postures was significantly lower than in the control group (p = 0.022) and in the group without yoga postures (p = 0.037; eTable 1). At no time were there any intergroup differences with regard to diastolic 24-h blood pressure (Figure 2, eTable 1). Further results of the ambulatory measurement of blood pressure are given in the eSupplement.
No severe adverse events occurred during the study. Nine patients (36.0%) in the intervention group with yoga postures experienced a total of 15 minor adverse events, while in the group without yoga postures four patients (16.0%) reported five minor adverse events. In the control group, two patients (8.0%) had four minor adverse events (eSupplement).
In this randomized controlled trial, a 12-week yoga intervention consisting purely of breathing and meditation techniques without any use of yoga postures lowered the joint primary outcome measure of 24-h systolic blood pressure by a greater amount than no treatment or an intervention with additional yoga postures. The other joint primary outcome measure, 24-h diastolic blood pressure, was not reduced. These results agree with the findings of a meta-analysis (11), although the effect sizes in our trial were smaller (−3.8 mmHg, 95% confidence interval [−7.5; −0.3] versus −14.1 mmHg, 95% confidence interval [−24.7; −3.6]). These differences from earlier studies are most likely attributable to differences in inclusion criteria, blood pressure recording methods (clinical versus ambulatory), and the country in which the study was carried out. In this last respect, yoga studies in India generally show larger effect sizes than studies performed in other countries (21).
Interestingly, though based only on secondary outcome measures, the pattern was reversed at week 28. In this final week of the study the group that had received the intervention with yoga postures exhibited greater reductions in ambulatory systolic blood pressure than the group without yoga postures.
In comparison with the untreated control group, 13 patients had to be treated with yoga to achieve substantial improvement in one patient.
A simultaneous increase in parasympathetic activity and decrease in sympathetic activity has been postulated as an important mechanism of action of therapeutic yoga (22). In this context, particularly yoga techniques of slow breathing and meditation may bring about a short-term reinforcement of parasympathetic activity (23–25). In experimental studies where yoga relaxation techniques were supplemented with yoga postures, no further potentiation of parasympathetic activity ensued (23). Together with the findings of the meta-analysis (11), this led to breathing and meditation techniques, rather than yoga postures, being viewed as the principal mechanism of action of yoga in patients with arterial hypertension (26). While this hypothesis was certainly confirmed immediately after the end of the yoga interventions, the results of our trial indicate that longer-term positive effects on arterial hypertension can be achieved only with the integration of yoga postures. The adoption of yoga postures can be seen as a form of isometric training, while dynamic yoga sequences (e.g., the well-known sun salutation) add an element from aerobic training (27, 28). Both isometric and aerobic training can lower blood pressure (29–31). These may then be the mechanisms responsible for the long-term effect of an intervention including yoga postures.
The main reason for the long-term group differences may, however, also lie in a discrepancy in treatment adherence following the end of the interventions: a higher proportion of patients in the group with yoga postures than in the group without yoga postures continued to practice the techniques they had been taught after study week 12. The group with yoga postures also practiced for a much longer time each week. It may be that yoga postures are easier to integrate into daily activities than yoga breathing and meditation techniques. When interpreting the findings in this way, however, it must be taken into account that the long-term group differences were greater in the intention-to-treat analysis than in the per-protocol analysis.
Our study featured some limitations. First, the wait list group did not control for the expectations of the patients, the time taken up, or the attentiveness of the instructors. Moreover, the study participants were not blinded to the intervention they received. The participants were predominantly female and well educated, and thus tended to be more typical of yoga practitioners (7, 32) than of patients with hypertension. This reduced the generalizability of the findings. No effects of the interventions on diastolic blood pressure were detected, the overall reduction in systolic blood pressure was only slight, and only 8% of hypertension patients will benefit substantially from a yoga intervention.
Only yoga without yoga postures had an immediate effect on the 24-h systolic (but not diastolic) blood pressure; yoga with yoga postures had no such action. However, the inclusion of yoga postures seems necessary if the effects are to be maintained beyond the end of the intervention. Since we found yoga in patients with arterial hypertension to be effective and safe, yoga can be recommended as a complement to medication for reduction of systolic blood pressure in this group.
This study was supported by the German Occupational Union for Yoga Teachers (Berufsverband der Yogalehrenden in Deutschland, BDY). The sponsor had no influence on the planning or conduct of the study, on the acquisition, management, analysis, or interpretation of the data, on the writing, revision, or approval of the manuscript, or on the decision to submit the manuscript for publication. We are particularly grateful to Peter Greve, Dr. Bettina Knothe, Nicole Verheyden, Frank Loddo, Eva Grandao, Janine Nagel, and Dr. Matthias Deparade for the development of the yoga interventions and to Eva Grandao and Janine Nagel for implementing the yoga interventions.
Conflict of interest statement
Dr. Cramer received support for this study from the BDY. The remaining authors declare that no conflict of interest exists.
After the conclusion of further analyses on the data used for this study, the authors will be willing to share the data with other researchers for scientific purposes. Interested parties should contact Dr. Holger Cramer.
Manuscript submitted on 12 April 2018, revised version accepted on 14 August 2018
Translated from the original German by David Roseveare
PD Dr. rer. medic. Dipl.-Psych. Holger Cramer
Klinik für Naturheilkunde und Integrative Medizin,
Am Deimelsberg 34a
45276 Essen, Germany
For eReferences please refer to:
subjects/ (last accessed on 11 April 2018).
|1.||Neuhauser H, Diederichs C, Boeing H, et al.: Hypertension in Germany—data from seven population-based epidemiological studies (1994–2012). Dtsch Arztebl Int. 2016; 113: 809–15 VOLLTEXT|
|2.||World Health Organization: Global Status Report on noncommunicable diseases www.apps.who.int/iris/bitstream/10665/148114/1/9789241564854_eng.pdf?ua=1 (last accessed on 11 April 2018).|
|3.||Woolf KJ, Bisognano JD: Nondrug interventions for treatment of hypertension. J Clin Hypertens (Greenwich) 2011; 13: 829–35 CrossRef MEDLINE|
|4.||Feuerstein G: The yoga tradition. Prescott: Hohm Press; 1998.|
|5.||Iyengar BKS: Light on yoga. New York: Schocken Books; 1966.|
|6.||De Michelis E: A history of modern yoga: Patanjali and western esotericism. London, Großbritannien: Continuum International Publishing Group; 2005.|
|7.||Cramer H: Yoga in Deutschland – Ergebnisse einer national repräsentativen Umfrage. Forsch Komplementmed 2015; 22: 304–10.|
|8.||Cramer H, Thoms MS, Anheyer D, Lauche R, Dobos G: Yoga in women with abdominal obesity—a randomized controlled trial. Dtsch Arztebl Int. 2016; 113: 645–52 VOLLTEXT|
|9.||Klatte R, Pabst S, Beelmann A, Rosendahl J: The efficacy of body-oriented yoga in mental disorders—a systematic review and meta-analysis. Dtsch Arztebl Int. 2016; 113: 195–202 VOLLTEXT|
|10.||Lund S, Angelow A, Baum E, et al.: Hausärztliche Risikoberatung zur kardiovaskulären Prävention – S3-Leitlinie. www.awmf.org/uploads/tx_szleitlinien/053–024l_Hausaerztliche_Risikoberat_kardiovask_Praevention_2017–11_1.pdf (last accessed on 11 April 2018).|
|11.||Cramer H, Haller H, Lauche R, Steckhan N, Michalsen A, Dobos G: A systematic review and meta-analysis of yoga for hypertension. Am J Hypertens 2014; 27: 1146–51 CrossRef MEDLINE|
|12.|| World Medical Association: WMA Declaration of Helsinki – ethical principles for medical research involving human subjects. www.wma.net/policies-post/wma-declaration-of-helsinki-ethical-principles-for-medical-research-involving-human-|
subjects/ (last accessed on 11 April 2018).
|13.||Schulz KF, Altman DG, Moher D: CONSORT 2010 statement: updated guidelines for reporting parallel group randomized trials. Ann Intern Med 2010; 152: 726–32 CrossRef MEDLINE|
|14.||Saghaei M: Random allocation software for parallel group randomized trials. BMC Med Res Methodol 2004; 4: 26 CrossRef MEDLINE PubMed Central|
|15.||Higgins JPT, Green S: Cochrane handbook for systematic reviews of interventions, John Wiley & Sons Ltd. West Sussex, UK, 2008 CrossRef|
|16.||Westhoff TH, Straub-Hohenbleicher H, Schmidt S, Tolle M, Zidek W, van der Giet M: Convenience of ambulatory blood pressure monitoring: comparison of different devices. Blood Press Monit 2005; 10: 239–42 CrossRef|
|17.||U.S. Food & Drug Association: Code of Federal Regulations,Title 21, Volume 5, Chapter I, Part 312. www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/cfrsearch.cfm?fr=312.32 (last accessed on 11 April 2018).|
|18.||Rubin DB: Multiple imputation for nonresponse in surveys, John Wiley & Sons. New York 1987 CrossRef|
|19.||Schafer JL: Analysis of incomplete multivariate data. Chapman and Hall. New York 1997 CrossRef|
|20.||Vickers AJ, Altman DG: Statistics notes: analysing controlled trials with baseline and follow up measurements. BMJ 2001; 323: 1123–4 CrossRef|
|21.||Cramer H, Lauche R, Langhorst J, Dobos G: Are Indian yoga trials more likely to be positive than those from other countries? A systematic review of randomized controlled trials. Contemp Clin Trials 2015; 41: 269–72 CrossRef MEDLINE|
|22.||Streeter CC, Gerbarg PL, Saper RB, Ciraulo DA, Brown RP: Effects of yoga on the autonomic nervous system, gamma-aminobutyric-acid, and allostasis in epilepsy, depression, and post-traumatic stress disorder. Med Hypotheses 2012; 78: 571–9 CrossRef MEDLINE|
|23.||Markil N, Whitehurst M, Jacobs PL, Zoeller RF: Yoga Nidra relaxation increases heart rate variability and is unaffected by a prior bout of Hatha yoga. J Altern Complement Med 2012; 18: 953–8 CrossRef MEDLINE|
|24.||Mourya M, Mahajan AS, Singh NP, Jain AK: Effect of slow- and fast-breathing exercises on autonomic functions in patients with essential hypertension. J Altern Complement Med 2009; 15: 711–7 CrossRef MEDLINE|
|25.||Telles S, Raghavendra BR, Naveen KV, Manjunath NK, Kumar S, Subramanya P: Changes in autonomic variables following two meditative states described in yoga texts. J Altern Complement Med 2013; 19: 35–42 CrossRef MEDLINE PubMed Central|
|26.||Cramer H: The efficacy and safety of yoga in managing hypertension. Exp Clin Endocrinol Diabetes 2016; 124: 65–70 MEDLINE|
|27.||Coulter D: Anatomy of hatha yoga: a manual for students, teachers and practitioners, Deep Books London UK, 2002.|
|28.||Mody BS: Acute effects of Surya Namaskar on the cardiovascular & metabolic system. J Bodyw Mov Ther 2011; 15: 343–7 CrossRef MEDLINE|
|29.||Millar PJ, Paashuis A, McCartney N: Isometric handgrip effects on hypertension. Current Hypertension Reviews 2009; 5: 54–60 CrossRef|
|30.||Brook RD, Appel LJ, Rubenfire M, et al.: Beyond medications and diet: Alternative approaches to lowering blood pressure. Hypertension 2013; 61: 1360–83 CrossRef MEDLINE|
|31.||Fagard RH, Cornelissen VA: Effect of exercise on blood pressure control in hypertensive patients. Eur J Cardiovasc Prev Rehabil 2007; 14: 12–7 CrossRef MEDLINE|
|32.||Cramer H, Ward L, Steel A, Lauche R, Dobos G, Zhang Y: Prevalence, patterns, and predictors of yoga use: results of a U.S. nationally representative survey. Am J Prev Med 2016; 50: 230–5 CrossRef MEDLINE|
|e1.||Streiner DL: The lesser of 2 evils: the ethics of placebo-controlled trials. Can J Psychiatry 2008; 53: 430–2 CrossRef MEDLINE|
|e2.||Cohen S, Kamarck T, Mermelstein R: A global measure of perceived stress. J Health Soc Behav 1983; 24: 385–96 CrossRef|
|e3.||Zigmond A, Snaith R: The hospital anxiety and depression scale. Acta Psychiatrica Scandinavica 1983; 67: 361–70 CrossRef|
|e4.||Ware JE, Sherbourne CD: The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992; 30: 473–83 CrossRef MEDLINE|
|e5.||Wagner P, Singer R: Ein Fragebogen zur Erfassung der habituellen körperlichen Aktivität verschiedener Bevölkerungsgruppen. Sportwissenschaft 2003; 33: 383–97.|
|e6.||Thompson FE, Kipnis V, Subar AF, et al.: Evaluation of 2 brief instruments and a food-frequency questionnaire to estimate daily number of servings of fruit and vegetables. Am J Clin Nutr 2000; 71: 1503–10 CrossRef MEDLINE|