Is Yoga Good for You? Part 2

Today, we continue with Part 2 of our series on Yoga – and we start with the question: Which of the 50 yoga styles have been shown to be best?

Should doctors recommend their patients do yoga? The difficulties associated with recommending yoga stem from the low quality of the scientific evidence available regarding its effects. Oddly, this lack of evidence is partly due to a common failure among researchers to define what they actually studied. They just say yoga, without defining what they mean. And different types of yoga differ greatly in what they demand in terms of physical strength, depth of meditation, breathing control, and spiritual component. Yet, it’s very common for reports of studies not to define which type they used. This not only makes it hard to compare results between studies, it also makes it very hard to translate any findings to the bedside when counseling patients.

Yoga is broadly defined as a mind-body practice composed of physical postures, breathing techniques, and meditation. And so, if researchers say they put yoga to the test, you’d assume all three components would be present. However, some studies use meditation only and call it yoga, with no postures or breathing. Others only examine breathing exercises, and call it yoga. Others consider yoga only postures, and still others say they’re studying forms of yoga without any of the three components. And so, if a study shows no benefit, does that really mean yoga, as commonly practiced, didn’t help? And, if a study does show benefit, what exactly do you tell patients to do to achieve it?

We can take some comfort in the fact that a review entitled “Is one yoga style better than another?” attempted to compare the effects of the more than 50 different yoga styles used in hundreds of trials, and they concluded that there was no apparent winner in terms of whether the results turned out positive. So, to a certain extent they may be interchangeable.

I’ve talked about the need to have active controls to see if there’s some benefit to yoga beyond just the exercise component. But such controls may still not deal with confounders, such as socialization. Yoga usually involves going to classes. So, you can imagine if you were studying the effects of yoga on something like depression, does just breaking routine, getting out, and meeting other people have any effect independent of the yoga?

Sadly, for depression, yoga doesn’t even seem to have a benefit compared to doing nothing. Yoga does appear to be helpful with anxiety, but not for people anxious to the extent that they’re actually diagnosed with an anxiety disorder. And, all the anxiety benefit from yoga appears just to be from the physical activity, since there appears to be no benefit compared to other types of exercise. So, unfortunately, yoga does not appear to be effective for mood and anxiety disorders.

A review on the effects of yoga on depressive symptoms in people with mental disorders in general, including conditions like PTSD or schizophrenia, found no significant effect compared to things like social support that offer the same kind of attention, or to doing nothing, but did find an effect compared to so-called “waitlist” controls.

See, unlike in studies of drug effects, where a placebo sugar pill is relatively straightforward, selection of a control condition is much more complex for behavioral interventions like yoga, and one common solution is a passive control such as a waitlist. Basically, people are randomized to either yoga or onto a waitlist where they’ll still get yoga, but have to wait a while. Now, that’s better than having no control group at all, since it still controls for the passage of time––meaning whether or not your condition would have gotten better naturally on its own––but doesn’t really control for the placebo effect. If, instead, you were just told you were entering a study to see if some exercises may help, and then unbeknownst to you, randomized to do yoga versus some sort of sham yoga, then, you would have the same kind of expectation either way. But if you know going into it you’re going to get yoga now, which may help, or yoga in six months, and you end up in the waitlist control, this may spuriously amplify the difference in treatment effect, because people assigned to the waitlist may expect to not get better without active treatment.

For example, here’s a study for yoga on urinary incontinence. Women were randomized to a six weeks of yoga or a waitlist control group. After six weeks, the incontinence frequency decreased by 70 percent in the yoga therapy, versus only 13 percent fewer episodes in the control group. But it’s possible that the women in the yoga group reported some benefits due to what’s called an expectation bias, because they expected yoga to be helpful. So, based on the body of evidence, one might say there’s insufficient good-quality evidence to judge whether yoga is useful for women with urinary incontinence. What we’d like to see are studies comparing yoga to time-and-attention control interventions––some alternative to yoga that entails the same amount of time and attention––so people don’t even know they’re in a control group.

Now, the question arises: who cares if the benefits from a yoga intervention are due to the placebo effect or expectation bias? A benefit is a benefit. As long as people are getting better, who cares? But it would be nice to know if yoga really helps or not. And here we go. Women with urinary incontinence were randomly assigned to a yoga group or a rigorous time-and-attention control group involving nonspecific muscle stretching and strengthening exercises.

There are two different types of urinary incontinence: urgency incontinence, defined as an involuntary loss of urine associated with a sudden strong desire to urinate, or stress incontinence, where an activity such as sneezing triggers an involuntary leak. Three months of yoga didn’t seem to significantly help overall incontinence, but it did seem to help with stress incontinence, decreasing by an average of 61 percent in the yoga group vs. only 35 percent in controls with no significant changes in urgency incontinence.

In our next study, we look at how sham acupuncture underscores the necessity of controlling for expectancy effects.

Yoga as therapy: when is it helpful? The scientific literature on yoga is limited in scope and quality. I’ve talked about the long list of issues that plague so many of the yoga trials, but let’s continue down the list of conditions to get a sense of the best available balance of evidence. For example, yoga as a therapy for irritable bowel syndrome. Evidence from randomized, controlled trials found that yoga beat out drugs, and was equally effective as moderate intensity walking––so, whichever form of exercise you prefer.

What about yoga for inflammatory bowel disease? Since stress can be a trigger, might yoga help? In a trial of yoga as an add-on therapy for adolescents with inflammatory bowel disease, it didn’t seem to help at all. But in a trial of yoga vs written self-care advice for adults with ulcerative colitis, 12 weeks of yoga induced a stronger increase in quality of life, and also reduced disease activity––even three months after the study ended. Now, the so-called self-care group was really just handed a couple general self-care books. So, the two treatments—yoga vs. self-care—were not matched at all for time intensity, feeling part of a group, or for therapist contact—each of which are likely to have the “yoga” arm more likely to nonspecifically improve their quality of life. So, yes, yoga might be effective for improving quality of life in ulcerative colitis. However, the benefits may be no different to other group exercise, for example. Indeed, there have been 11 human studies on the influence of moderate exercise programs upon disease activity in chronic inflammatory bowel disease, and all 11 report benefit in terms of reduced disease activity.

Using passive controls doesn’t account for expectancy effects, attention given to the study subjects, and the time spent, the investment that the participants put in. Here’s an excellent example of why having more active control groups is so important. This details a series of studies on nonpharmacological interventions for menopausal hot flashes. For example, randomized to eight weeks of acupuncture, eight weeks of sham acupuncture, or nothing––just usual care. Sham acupuncture is when they still stick needles in you, but shallowly and not in traditional acupuncture points.

There was also a study of 10 weeks of yoga classes, or the attention control, health and wellness education classes, which lasted just as long as the yoga, and were done in the same kind of group social setting. Here’s what they found. The hot flashes in the usual care group, the passive control where they didn’t do anything special, stayed pretty stable. But the real acupuncture worked way better. But so did the sham acupuncture. So, yoga and acupuncture both worked just as well. But so did the nonyoga and fake acupuncture. There just appeared to be a general placebo effect such that women in all five intervention groups benefited simply by being in a study. That’s why it’s always better to have the control groups do something, so you can have some confidence that at least some of the effect of your intervention is real.

But hey, at least they had control groups. This study: “Twelve-Minute Daily Yoga Regimen Reverses Osteoporotic Bone Loss” caused quite a stir. We know that when it comes to bone health, it’s use it or lose it. Exercise early and exercise often. Physical activity is a widely-accessible, low-cost, and highly modifiable contributor to bone health. Exercise transmits forces through the skeleton, generating signals that are detected by your bone building cells. This is why the National Osteoporosis Foundation, International Osteoporosis Foundation, and other agencies recommend weight-bearing exercises for the prevention of osteoporosis. These include high-impact exercises, such as jumping, aerobics, and running, as well as lower-impact exercises, such as walking and weight training, to create those mechanical signals that spark bone growth. Lower impact activities, such as yoga, are generally not considered bone-building. That’s why the results of this study were so surprising.

It was a study of Internet-recruited volunteers comparing bone mineral density changes before and after yoga. The researchers devised a 12-minute DVD of 12 yoga poses that they believed would stimulate increased bone density in the spine and both parts of the hip. But in the end, out of the 741 patients recruited at the beginning, only a few dozen actually sent in their bone scans as instructed. So, that’s only like 5 percent. So, they lost 95 percent of the people that originally started. You can totally imagine how those who got positive results were more likely to send them in, and the others that failed just kind of slunk away. So, I don’t consider this convincing evidence.

The authors had the attitude of “Hey, look, so what do you got to lose? The side effects of yoga are all good; so, why not give it a try?”

What you have to lose is the opportunity to do higher impact exercise that has more decisive evidence of bone benefit.

And, all the side effects are not necessarily good. There have been vertebral compression fractures associated with yoga. In this series of nine cases from the Mayo Clinic, they describe spinal compression fractures occurring a month to years after initiating yoga-associated spinal flexion exercises. Both scientific and media reports continue to advertise yoga as a bone-protective activity. But there’s a need for selectivity in yoga poses in populations at increased fracture risk. Here’s the yoga poses they recommend, and these are the ones they encourage people at risk to avoid. And four of the nine patients developed a fracture in the setting of a normal or near-normal BMD. So, maybe everyone should be careful with those. Yoga-related injuries are not uncommon, resulting in thousands of emergency room visits a year, and approximately 5 percent of those are coming in with fractures. I’ll try to put that injury rate in context, comparing yoga to other physical activities, next.

Even though for certain health conditions we have not reached a point where we can say there is good scientific evidence that yoga is beneficial, many patients do seem to benefit. And isn’t that what matters? I mean, even in the studies that show that the benefit is mainly a placebo effect, does this really matter? After all, this editorial asserted, yoga therapy has never been shown to have adverse effects. But that’s simply not true. About 2,000 yoga-related injuries have been seen in hospital emergency rooms every year from 2001 to 2014.

It looks like one fatality has ever been reported in the medical literature. But if you look at the case report, a 16-year-old girl collapsing with a ruptured lung following voluntary mouth-to-mouth yoga breathing exercises with a teen-age boy in the hall, that can hardly be characterized as a typical yoga practice.

There are tragic rare cases tied to actual yoga, like this quad rupture in a young woman or, God, a femoral fracture. The femur is the bone in your thigh—longest, strongest, and heaviest bone in the body. Typically, you only see femur fractures with trauma, such as car crashes, but here it was just a guy practicing this yoga posture when he heard a large cracking sound and felt immense pain because yikes.

Now, two weeks earlier, he felt a pain in his thigh that he just thought was muscle strain, but bone biopsies showed evidence of previous microfractures. So, it sounds like he had been overdoing it, weakening it before the final crack. So, I guess the bottom line is that people should listen to their bodies and not to push themselves too far.

Now, I talked about the risk of spinal compression fractures, particularly those with weakened bones, osteopenia, or osteoporosis. Poses causing hyperflexion and hyperextension of the spine should probably be avoided. These were the poses tied to injury in a few dozen cases of musculoskeletal injuries described at the Mayo Clinic. But case reports, and series of case reports, are really just kind of like glorified anecdotes. They don’t give you an overall sense of the safety of yoga. If you look at randomized, controlled yoga trials, there’s actually no significant difference in injury rates between the yoga groups and the control groups that used some other sort of exercise intervention.

Now, this may not necessarily translate out into real-world experience, given the short duration of yoga in these clinical trials—as little as a single day—and the fact that they may be more closely supervised, with more highly-trained teachers. That’s why large-scale surveys like this can be helpful, looking at thousands of people taking yoga classes.

Looks like about 30 percent of yoga class attendees had experienced some type of adverse event, most commonly just something mild like muscle aches. But some individuals did experience more severe events, which caused them to stop going. Many causes were associated with “overexertion and overdoing” and starting out in “poor physical condition.”

This systematic review included nine observational studies with 9,000+ yoga practitioners and 9,000+ non-yoga practitioners from the U.S., Europe, Asia, and Australia. A considerable proportion of yoga practitioners experienced injuries or other adverse events. However, most were mild and transient, and risks were comparable to those of non-yoga practitioners. The risk of yoga-associated injuries was estimated as 1.45 per thousand hours of yoga practice––much lower than in higher-impact sports activities, such as soccer. Compared to 1.5 for yoga, runners have like 2.5 injuries per thousand hours, soccer 3.7, 5 for tennis, and an hour of yoga is more than five times safer than skiing.

Here’s where yoga falls on the spectrum. Yoga appears much safer, for example, than aerobic funk. On the other hand, yoga practitioners may suffer an increased risk of meniscus injury––the main stabilizing and cushioning cartilage in the knee. Yoga was found associated with significantly higher risk of meniscus injury compared with activities such as badminton, jogging, and climbing hills. Some fundamental yoga postures, like the lotus position, can be hard on the knees. Yes, yoga poses can undoubtedly improve the flexibility of the knee, but immoderate joint movement can pose a serious threat.

The bottom line is that like any other type of physical activity, yoga carries a risk of injury. Exercise is indeed medicine, but like any medicine, it must be prescribed appropriately. Many patients rely on their yoga practice for stress reduction. So, it is important to be mindful that being injured is the last thing you need. The higher-risk yoga poses appear to include headstand, shoulder stand, lotus and half-lotus, forward bend, backward bend, and handstand. And so, beginners should be beware, with particular attention paid to the spine, as this is where the highest number of injuries occur.

Hot yoga deserves special mention. The extreme heat and intensity of Bikram yoga may make it inappropriate for older adults and people with medical conditions. But there are case reports of sudden cardiac arrest even in a healthy 35-year-old. Pregnancy is an especially vulnerable time for heat exposure. With the increased risk of spinal defects and possibly of other birth defects among fetuses exposed to excessive heat, pregnant women should avoid practicing hot yoga during pregnancy. Maternal hyperthermia—whether from a sauna or electric blankets––had nearly a two-fold increased risk of spinal and brain malformations. For example, pregnant women who use hot tubs increase the risk of bearing babies with their intestines outside their bodies or being born without a brain.

As with any other physical or mental practice, yoga should be practiced carefully under the guidance of a qualified instructor. Beginners should avoid extreme practices, such as headstand, lotus position, and forceful breathing. Individuals with medical preconditions should work with their physician and yoga teacher to appropriately adapt postures. For example, patients with glaucoma should avoid upside-down positions, and patients with compromised bones should avoid forceful yoga practices, and practices like voluntary vomiting should perhaps be avoided completely. What? It’s evidently common practice in traditional yoga.