The research, which pooled 26 prior studies, found that spinal manipulation was linked to "modest improvements" in pain and function among people with short-term lower-back pain. Their pain improved an average of one point on a 10-point scale.
"For acute back pain, this is usually not considered a clinically meaningful improvement," said Dr. Wolf Mehling, who practices at the Osher Center for Integrative Medicine at the University of California, San Francisco. Mehling was not involved in the JAMA study.
The analysis did not include people with chronic low back pain, which lasts much longer. A 2011 Cochrane review failed to find good evidence that spinal manipulation was any better or worse than other treatments for chronic low back pain.
Mehling is one of two doctors he knows at UCSF who practice spinal manipulation and other chiropractic therapies. He said the new study lumps together many people who may have differing causes for their back pain.
For this reason, a lackluster average drop in pain may actually hide people who do see a noticeable benefit.
"It's not like everybody gets exactly a one-point benefit," said Dr. Paul Shekelle, one of the authors of the study and a physician at the VA Medical Center in West Los Angeles. "Some people get zero, but some people get a lot more."
Previous research has suggested that some people are more likely to benefit from spinal manipulation than others: for example, those whose symptoms last fewer than 16 days, who have no pain below the knee and who are willing to remain active.
Mehling, who trained in both Germany and the United States, said that most American doctors do not receive training in spinal manipulation, nor do they know how to test whether a patient might be a good candidate. Often, those who practice the technique may be chiropractors or physical therapists.
"Physicians in the US are not trained to put their hands on patients," he said. "In Germany, spinal manipulation ... is part of normal care."
Pills and motions
"We're trying to take joints in your spine that have become too stiff and give them back their normal motion again," said William Lauretti, a spokesman for the American Chiropractic Association and an associate professor at New York Chiropractic College.
Lauretti, who did not participate in the study, said spinal manipulation can take different forms, but the most common involves someone lying on their side on a low bench while a chiropractor twists and moves each spinal joint through its normal range of motion.
"There are people who respond very, very well to this," he said.
But some experts say that a sample of patients with acute low back pain is inherently skewed.
"We know that the vast majority of these patients will spontaneously get better whether or not we do anything," said Dr. Raj Rao, a spine specialist who has served on the American Academy of Orthopaedic Surgeons' board of directors. Rao was not involved in the new analysis.
Two of the studies in the analysis compared spinal manipulation to a fake, placebo version of the therapy, while additional studies compared it to other treatments. But Rao said that, because spinal manipulation is practiced in different ways by many types of practitioners -- physical therapists, chiropractors and doctors included -- it can be hard to standardize how the research is conducted.
Experts say there is a need for well-designed studies that can prove or disprove the benefits of spinal manipulation and whom it is most likely to help. In their analysis, Shekelle and his co-authors rated the quality of each study they included; over half were classified as "low quality."
"It makes me a little cautious about my interpretation of any results of the study," Rao said.
No magic bullet
"There's no magic bullet out there, on average," said Shekelle, the study author.
About four out of five people will have lower-back pain at some point, according to the National Institute of Neurological Disorders and Stroke, making it a leading cause of missed work and disability.
"Spinal manipulation is one of those therapies that on average has a small effect, and it's in your doctor's bag of things you can pull out and try," Shekelle said.
"All of this manipulation didn't change things that much but certainly adds to the cost of care," Rao said.
A first visit with a chiropractor, which typically includes an evaluation, may run between $75 and $200 depending on location and other factors, the ACA's Lauretti estimated. Followup visits may cost between $30 and $70. He said many insurance plans cover some sort of chiropractic treatment, but policies and co-payments can vary.
Over-the-counter pills, on the other hand, are cheaper and widely available, but come with potential side effects like stomach ulcers, Lauretti said. But side effects are more likely at high doses and over long periods, according to experts.
On the other hand, no "serious" harms were reported in the low back pain studies that were analyzed, but minor aches and pains were plenty. About half to two-thirds of patients treated with spinal manipulation had temporarily increased pain, muscle stiffness and headache.
Lauretti said, "This is similar to what you might get if you did a workout for the first time."
Back pain: Now what?
So what to do about short-term back pain? Shekelle said that if something is safe and has worked for patients in the past, he doesn't mess with a good thing.
He may ask his patients to start by applying heat and trying over-the-counter meds and said he generally follows the advice of the American College of Physicians.
In February, that group published its own recommendations for treating low back pain, suggesting starting with noninvasive remedies like superficial heat, massage, acupuncture or spinal manipulation. They also listed NSAIDs and muscle relaxants, if people prefer to take medicines.
"Physicians should avoid prescribing unnecessary tests and costly and potentially harmful drugs, especially narcotics, for these patients," Dr. Nitin Damle, the organization's president, said in a statement.
The group's review said that spinal manipulation was backed by "low-quality evidence," as well. Although the technique appeared to slightly improve function for short periods, there wasn't enough evidence to say that it had any impact on the back pain itself.
No matter which path people take, Shekelle said several things are key: reassurance that "the sky is not falling," that most patients get better no matter what and that it is important to remain active.
"When I first started out 30 years ago, we put everybody to bed for a week and told them to rest their back. It turns out we were probably causing more harm than good," Shekelle said, adding that the Department of Veterans Affairs, which requested the study, has integrated chiropractic care for years.
Rao said that anecdotally, some patients do well with chiropractic therapies, which may be appropriate for short-term pain.
"If you go to a chiropractor (one or two times) and it makes a difference, I'm all for it," he said. "What I think I'm a little bit more cautious about ... is prolonged chiropractic care.
"When any of us is in pain," he added, "we're willing to do almost anything to get relief."