Kundalini Yoga Meditation Versus the Relaxation Response Meditation for Treating Adults With Obsessive-Compulsive Disorder: A Randomized Clinical Trial

Abstract

Background: Obsessive-compulsive disorder (OCD) is often a life-long disorder with high psychosocial impairment. Serotonin reuptake inhibitors (SRIs) are the only FDA approved drugs, and approximately 50% of patients are non-responders when using a criterion of 25% to 35% improvement with the Yale-Brown Obsessive-Compulsive Scale (Y-BOCS). About 30% are non-responders to combined first-line therapies (SRIs and exposure and response prevention). Previous research (one open, one randomized clinical trial) has demonstrated that Kundalini Yoga (KY) meditation can lead to an improvement in symptoms of obsessive-compulsive severity. We expand here with a larger trial. Design: This trial compared two parallel run groups [KY vs. Relaxation Response meditation (RR)]. Patients were randomly allocated based on gender and Y-BOCS scores. They were told two different (unnamed) types of meditation would be compared, and informed if one showed greater benefits, the groups would merge for 12 months using the more effective intervention. Raters were blind in Phase One (0-4.5 months) to patient assignments, but not in Phase Two. Main Outcome Measures: Primary outcome variable, clinician-administered Y-BOCS. Secondary scales: Dimensional Yale-Brown Obsessive Compulsive Scale (clinician-administered), Profile of Mood Scales, Beck Anxiety Inventory, Beck Depression Inventory, Clinical Global Impression, Short Form 36 Health Survey. Results: Phase One: Baseline Y-BOCS scores: KY mean = 26.46 (SD 5.124; N = 24), RR mean = 26.79 (SD = 4.578; N = 24). An intent-to-treat analysis with the last observation carried forward for dropouts showed statistically greater improvement with KY compared to RR on the Y-BOCS, and statistically greater improvement on five of six secondary measures. For completers, the Y-BOCS showed 40.4% improvement for KY (N = 16), 17.9% for RR (N = 11); 31.3% in KY were judged to be in remission compared to 9.1% in RR. KY completers showed greater improvement on five of six secondary measures. At the end of Phase Two (12 months), patients, drawn from the initial groups, who elected to receive KY continued to show improvement in their Y-BOCS scores. Conclusion: KY shows promise as an add-on option for OCD patients unresponsive to first line therapies. Future studies will establish KY's relative efficacy compared to Exposure and Response Prevention and/or medications, and the most effective treatment schedule. Clinical Trial Registration: www.ClinicalTrials.gov, identifier NCT01833442.

Keywords: Dimensional Yale-Brown Obsessive-Compulsive Scale; Yale-Brown Obsessive Compulsive Scale; anxiety/anxiety disorders; depression; mental health; mindfulness.

Copyright © 2019 Shannahoff-Khalsa, Fernandes, Pereira, March, Leckman, Golshan, Vieira, Polanczyk, Miguel and Shavitt.

Figures

Figure 1

Consolidated standards of reporting trials diagram. Flow of patients through the study. OCD, obsessive compulsive disorder; CBT indicates cognitive-behavioral therapy.

Figure 2

Y-BOCS total scores Phase One: 0 month and 4.5 months % changes: intent-to-treat and completers. The Phase One clinician-administered Yale-Brown Obsessive Compulsive Scale (Y-BOCS) Total Scores Group % changes: 0 month vs. 4.5 months, are plotted for the intent-to-treat (ITT) analysis using the last observation carried forward (LOCF), and for the completers (Com) for the Kundalini Yoga meditation group and Relaxation Response control group. The Kundalini Yoga ITT mean group % improvement was 26.90% (SD = ± 27.628%; n = 24). The Relaxation Response ITT mean group % improvement was 8.214% (SD = ± 13.137; n = 24). The Univariate Analysis of Variance for the ITT LOCF indicates that Kundalini Yoga had a greater % improvement on the Y-BOCS, (f1.46 = 8.96, P = 0.004). The Com Kundalini Yoga baseline mean was 26.25 (SD ±4.74; n = 16) and the 4.5-month mean was 15.19, showing a 40.4% (SD ±24.32%) mean group improvement. The Relaxation Response baseline was 27.55 (SD ±4.204; n = 11) and 22.45 (SD ±4.34; n = 11) at 4.5 months. The Com Relaxation Response group mean % improvement was 17.92% (SD = ± 14.34%; n = 11). The univariate analysis of variance for the completers indicates that Kundalini Yoga had a greater improvement on the Y-BOCS (f1.25 = 7.50, P = 0.011).

Figure 3

DY-BOCS total scores Phase One: 0 Month and 4.5 Months % Changes: Intent-To-Treat and Completers. The Phase One clinician-administered Dimensional Yale-Brown Obsessive Compulsive Scale (DY-BOCS) Total Scores Group % changes: 0 month vs. 4.5 months, are plotted for the intent-to-treat (ITT) analysis with the last observation carried forward (LOCF), and for the completers (Com) for the Kundalini Yoga meditation group and the control group Relaxation Response meditation technique. The Kundalini Yoga ITT mean group % change improvement was 19.98% (SD = ± 27.629%; n = 24). The Relaxation Response ITT mean group % change improvement was 0.592% (SD = ± 12.282%; n = 24). The Univariate Analysis of Variance for the ITT LOCF indicates that Kundalini Yoga had a greater mean group % change improvement on the DY-BOCS, (f1.46 = 9.86, P = 0.003)]. The Com baseline means for Kundalini Yoga and Relaxation Response are 19.93 (SD ±4.46; n = 15) and 20.09 (SD ±3.36; n = 11), respectively. The 4.5-month means for Kundalini Yoga and Relaxation Response are 13.33 (SD ±5.81; n = 15) and for 19.55 (SD ±3.75; n = 11), respectively. The Com Kundalini Yoga group mean % improvement was 31.969% (SD ±29.04%; n = 15). The Com Relaxation Response group mean % improvement was 1.298% (SD ±18.599%; n = 11). The Univariate Analysis of Variance for the Com indicates that Kundalini Yoga had a greater improvement on the DY-BOCS, (f1.24 = 9.384, P = 0.005).

Figure 4

POMS total mood disorder scores, BAI scores, BDI scores, and SF-36 scores for Phase One: 0 month and 4.5 months % changes: intent-to-treat and completers. Figure 4shows the Profile of Moods States Total Mood Disorder (POMS TMD), the Beck Anxiety Index (BAI), the Beck Depression Inventory (BDI) scores, and the Short Form Health Survey (SF-36) scores for the Phase One % change improvement scores for the Kundalini Yoga and Relaxation Response groups when comparing the 0-month and 4.5-month scores. The POMS intent-to-treat (ITT) analysis with the last observation carried forward (LOCF) Kundalini Yoga group (0-month mean = 122.083; 4.5-month mean = 101.958) mean change score was −20.125 (SD ± 31.208; n = 24) and −0.709 (SD ± 15.398; n = 24) for the Relaxation Response group (0-month mean, 113.5; 4.5-month mean 112.791; n = 24). This difference of 19.417 was significant (P = 0.01). The ITT LOCF POMS mean % change improvement for 0 month to 4.5 months was 24.452% (SD ±37.72%; n = 24) for the Kundalini Yoga group and 1.773% (SD ± 15.38%, n = 24) for the Relaxation Response group. This difference was significant P = 0.009. The completer (Com) group baseline mean for Kundalini Yoga is 127.533 (SD ± 50.141; n 15) and 116 (SD ± 45.29; n = 11) for the Relaxation Response. The Com group 4.5-month mean for Kundalini Yoga is 95.33 (SD ± 36.976; n = 15) and 114.454 (SD ± 47.922; n = 11) for the Relaxation Response. The mean group % change score for Kundalini Yoga improvement was 39.123% (SD ± 41.505%; n = 15) and 3.869% (SD ± 23.138%; n = 11) for the Relaxation Response. The group differences are significant, (f1.25 = 6.42, P = 0.018). (Figure 4) Also shows the BAI % change improvement scores for the Kundalini Yoga and Relaxation Response groups. The ITT analysis with the LOCF for the BAI mean change score for 0 to 4.5 months was −4.00 (SD ± 5.741; n = 24) for the Kundalini Yoga group [0-month mean 17.708 (SD ± 8.094; n = 24); 4.5-month mean 13.708; (SD ± 8.379; n = 24)] and −0.708 (SD ± 4.982; n = 24) for the Relaxation Response group [0-month mean 15.375; (SD ± 11.336; n = 24); 4.5-month mean 14.667 (SD ± 10.98; n = 24)]. This change score difference of 3.292 was significant (2-tail) P = 0.039. The ITT LOCF BAI mean % change improvement for 0 to 4.5 months was 18.52% (SD ± 32.23%; n = 24) for the Kundalini Yoga group and –5.10% (SD ±42.86%; n = 24) for the Relaxation Response group. This difference was significant (2-tail) P = 0.036. The Com group baseline mean for Kundalini Yoga is 17.33 (SD ±7.724; n = 15) and 16.0 (SD 11.983; n = 11) for Relaxation Response. The completer group 4.5-month mean for Kundalini Yoga is 10.93 (SD ±6.766; n = 15) and 14.45 (SD ±11.228; n = 11) for Relaxation Response. The mean % change score improvement was 29.63% (SD ± 36.799%; n = 15) and –11.13% (SD ± 64.429%; n = 11) for Relaxation Response. The group differences are significant, (f1.24 = 4.2, P = 0.05, 2-tail). (Figure 4) Also shows the BDI % change scores for the Kundalini Yoga and Relaxation Response groups for the baseline to 4.5 months. The ITT analysis with the LOCF for the BDI mean change score for 0 to 4.5 months was −7.042 (SD ± 10.564; n = 24) for the Kundalini Yoga group [0-month mean 21.0 (SD ± 10.505); 4.5-month mean 13.958 (SD ± 8.518)] and –0.333 (SD ± 3.293; n = 24) for the Relaxation Response group [0-month mean, 17.208 (SD ±10823); 4.5-month mean, 16.875 (SD ± 10.617)]. This change score difference of 6.708 was significant P = 0.005. The BDI mean % change improvement for 0 to 4.5 months was 23.90% (SD ± 41.27%; n = 24) for the Kundalini Yoga group and −3.53% (SD ± 25.54%; n = 24) for the Relaxation Response group. This difference was significant (P = 0.008). The Com group baseline mean for Kundalini Yoga is 22.67 (SD ±11.81; n = 15) and 17.18 (SD ±10.84; n = 11) for the Relaxation Response. The Com group 4.5-month mean for Kundalini Yoga is 11.4 (SD ± 8.16; n = 15) and 16.45 (SD ± 10.35; n = 11) for the Relaxation Response. The group differences are significant, (f1.24 = 7.05, P = 0.014). The BDI Com Kundalini Yoga mean % change improvement was 38.242% (SD ± 47.012%; n = 15), and –7.705% (SD ± 38.278%; n = 11) for the Relaxation Response. Group differences are significant (f1.24 = 7.05, P = 0.014). Figure 4 also shows the SF-36 Com Scores Phase One 0 to 4.5 months % Changes. The Com group baseline mean for Kundalini Yoga is 89.73 (SD ± 22.53, n = 15) and 83.73 (SD ± 16.29, n = 11) for the Relaxation Response. The Com 4.5-month mean for Kundalini Yoga is 73 (SD ± 17.96, n = 15) and 76.91 (SD ± 20.11, n = 11) for the Relaxation Response. The Com mean Kundalini Yoga % change improvement was 16.4% (SD ± 3.87%; n = 15) and 8.1% (SD ± 5.98%; n = 11) for the Relaxation Response. The group differences are not significant (P = 0.18)

Figure 5

CGI Scores in Phases One and Two. Figure 5 shows the clinician-administered Clinical Global Impression Scale (GCI) scores for both Kundalini Yoga and the Relaxation Response for Phase One and Two for both the frequency and the relative frequency in %’s for the four possible scores of 1, 2, 3, or 4. At 4.5 months Kundalini Yoga had eight scores of 1, and 7 with a 2, 3, or 4, and the Relaxation Response had 0 scores of 1, and 11 with a 2, 3, or 4. Kundalini Yoga shows greater clinical improvement compared to the Relaxation Response with the distributions of 1’s versus 2, 3, or 4 (P = 0.007). The Kundalini Yoga group had only one 4, and 14 values of a 1, 2, or 3, and the Relaxation Response had four 4’s, and seven values of a 1, 2, or 3. This difference was also significant with P = 0.034. Phase Two did not show statistical differences when comparing the patients that were originally in the Kundalini Yoga group vs those originally in the Relaxation Response group (See Table S1 in Supplement).

Figure 6

Phase Two Y-BOCS Total % improvement. Figure 6 shows the Y-BOCS mean % change improvement for the patients remaining at 8.5 months (n = 17), 12.5 months (n = 9), and 16.5 months (n = 7) for those who were in their original groups (red = KY, black = RR), and also for the patients combined in the KY Phase Two group (blue = combined). Of the seven that completed the 16.5-month time point, three were from the original KY group and 4 from the RR group. The additional % mean change improvement for the seven subjects when comparing their 4.5-month and 16.5-month means was 18.76% (SD +36.91%), with a 31.07% (SD +36.65%) improvement for the three originally from KY and 9.53% (SD +7.53%) for the 4 originally from RR. When comparing the three patients originally from KY at 16.5 months to their Phase One 0-month baseline score, they had a 50.62% (SD +42.35%) mean % change improvement. The four RR patients had a 26.96% (SD +21.94%) overall improvement, and all seven combined showed an overall improvement of 37.1% (SD +37.91%). For reference, Figure 6 also shows the mean % improvement plotted for the Y-BOCS Total Phase One 4.5-month % improvement scores plotted on the y-axis for the completers for the KY meditation group (n = 16) with an open red circle and the RR control group (n = 11) with an open black circle. The KY mean showed a 40.4% improvement, and the RR mean was 17.9%.

Figure 7

A Phase Two Bayesian analysis for the Y-BOCS: The calculation of the Posterior Density π of the population showing the proportion of positive respondents, P. A Bayesian statistical analysis was performed for 17 patients comparing their 4.5-month value with their last Y-BOCS measure taken at dropout that included the seven that completed the 16.5-month trial end point. This analysis showed that a 50% or greater probability criteria of patient improvement in the Y-BOCS in Phase Two was P = 0.593, where 0 = no patients improving and 1 = all patients improving.