Yoga

What it is

Yoga is a mind-body practice combining postures, breathing exercises and meditation [1]. Most back-pain trials used Iyengar, hatha or viniyoga [2]. Hot yoga, done in rooms at 25 °C or warmer, is a separate form, most often Bikram-style [3].

How it might work

Proposed only: yoga may lower stress hormones and steady the sympathetic nervous system [4]. Breathing and meditation may matter most for blood pressure, since larger drops appeared in programmes that included them [5].

What the research says

Low back pain (strongest evidence). Reviews of randomised trials found less pain and better function than no exercise. In a review of 27 studies with 2,702 people, yoga reduced pain by 0.74 points on a numeric rating scale and disability by 2.28 points in the short term [6]. A Cochrane review of 21 trials with 2,223 people found a small pain improvement at three months, of about 4.5 points on a 0 to 100 scale, below the 15-point threshold usually taken as meaningful [2]. In one review, pain relief was no longer significant at 12 months, while disability gains persisted [7]. Compared with back-focused exercise, yoga made little or no difference to function at three months [2].

Blood pressure. Reviews show lower blood pressure with yoga, but the evidence is rated low to very low. In a 2025 review of 30 trials with 2,283 people, systolic pressure fell by 7.95 mmHg compared with waiting-list controls [8]. Against active controls the systolic difference was not significant [8]. A 2019 review of 49 trials with 3,517 adults found systolic pressure about 5 mmHg lower than controls [5]. In a menopause review of 24 trials, systolic and diastolic pressure were also lower with yoga [9].

Stress and anxiety. Results are mixed. Across eight small trials of 319 people with raised anxiety, yoga reduced anxiety compared with no treatment, but the effect was not seen in people with diagnosed anxiety disorders [10]. A review of hatha yoga found no effect on anxiety compared with active controls [11]. In pregnancy, a review of 29 studies with 2,217 women found lower anxiety and perceived stress [12]. A review of 17 studies in people recovering from heart bypass surgery reported less anxiety and stress biomarkers [13]. In seven trials of people with inflammatory bowel disease, yoga-based mind-body programmes reduced perceived stress [14]. Reviews of yoga asanas report lower evening and waking cortisol than active controls [4].

Sleep. Reviews point in a positive direction but disagree on size. In a review of 22 trials with 1,348 people with insomnia, yoga increased total sleep time by about 111 minutes compared with active controls, at moderate certainty [15]. Sleep efficiency rose by 15.6 percentage points, at low certainty [15]. In 19 studies of women with sleep problems, sleep quality improved, but not in breast-cancer or menopausal subgroups [16]. In a review of 35 trials in older adults, yoga was not among the exercise types that beat usual care [17]. A network review ranked yoga highly for sleep efficiency and daytime function, but aerobic exercise best for overall sleep quality score [18].

Balance and flexibility. Two meta-analyses of older adults found better balance and flexibility than controls. In 22 trials of older adults without a specific illness, yoga improved balance (effect size 0.7 against inactive controls) and lower-body flexibility (0.5) [19]. A review of 15 trials found balance improved (standardised mean difference 0.81, 95% CI 0.48 to 1.14) and flexibility improved slightly (0.38, 95% CI 0.07 to 0.68) [20].

Falls. One large trial found more falls with yoga. In 700 adults aged 60 and over, an Iyengar yoga programme led to 0.87 falls per person-year, against 0.64 with seated relaxation, a rate ratio of 1.33 (95% CI 1.01 to 1.75) over 12 months [21]. The same trial reported better balance confidence [21]. A small Hatha yoga trial found no difference in falls between yoga and relaxation groups [22].

Doses used in studies

In one review of 49 trials, yoga averaged about 4.8 sessions a week of about 59 minutes for 13 weeks [5]. A large falls trial used two one-hour classes a week for 12 months, with home practice on two more days [21]. A small falls trial used 16 Hatha sessions over eight weeks, plus 10 minutes of daily home practice [22]. In pregnancy, 12 or more sessions longer than 60 minutes had a significant effect on stress [12]. Frequency, intensity and duration varied widely between trials, so these are what was studied, not recommendations [12].

Safety and interactions

Side effects. Yoga caused more adverse events than no exercise, mainly back pain at 6 to 12 months (43 per 1,000 people, against 9 per 1,000) [2]. In the falls trial, six people in the yoga group had musculoskeletal side effects and none had serious ones [21]. A blood pressure review also found more adverse events with yoga than with usual care [23].

Upper limits and toxic doses. None of the reviews I read reported an upper limit or toxic dose. Only three of eight anxiety trials reported safety data [10].

Medicine interactions. No study I read examined interactions with medicines. In one blood pressure review, effects appeared only in trials that allowed people to keep taking blood pressure medicine [23].

Pregnancy and breastfeeding. Trials in pregnant women reported benefits for anxiety, stress and labour, but the review rated the evidence low to very low [12]. I found no studies on breastfeeding.

Relevant conditions. No kidney, liver or bleeding-disorder data appeared in the reviews I read [2]. In hot yoga, case studies report ill-health after practice [3].

Caveats and open questions

Most trials are small and short, and several reviews rate their risk of bias as high [2][8]. Blood pressure results vary widely between trials [23]. An older Cochrane review found a small diastolic drop of about 2.9 mmHg in small, short trials [24]. The largest recent stress review covered yoga nidra, a different practice, and its authors warned that effects may be inflated [25]. Long-term effects are uncertain [7]. Ongoing trials in atrial fibrillation and in pregnancy may add to these findings.