August 8, 2018
Jennifer Abbasi
JAMA. Published online August 8, 2018. doi:10.1001/jama.2018.11068
A class of drugs called aromatase inhibitors (AIs) are a cornerstone of breast cancer treatment, but they come at a steep physical price. About half of women who take this form of hormone therapy develop arthritis-like joint stiffness and pain for still-unknown reasons.
“When we started using these medicines, the first thing that women would say is, ‘This medication makes me feel like an old lady,’” said oncologist Dawn Hershman, MD, who leads the breast cancer program at Columbia University Medical Center’s cancer center in New York City.
The drugs—anastrozole, exemestane, and letrozole—are prescribed for 5 to 10 years in postmenopausal women with hormone-sensitive breast cancer, but the life-limiting pain they cause compels many to abandon the treatments before that. And it’s feared that this could lead to cancer recurrence and death.
Eight years ago, in an effort to encourage women to keep taking their AIs, Hershman and her collaborators set out to find a way to alleviate the pain. A pill wouldn’t cut it. “We know that when a patient has side effects from the medicine, they don’t want to take another medicine that has the potential to cause more side effects,” she said. “Natural remedies or modalities are more acceptable to patients.”
So they chose to study acupuncture, a form of traditional Chinese medicine that some clinical evidence suggests can effectively treat chronic musculoskeletal pain. The results from their National Institutes of Health–funded randomized clinical trial recently appeared in JAMA.
The study involved 226 postmenopausal women at 11 academic centers or clinical sites that are part of the National Cancer Institute’s Clinical Trials Network. The participants had early-stage breast cancer, were taking AIs, and scored at least 3 on a 0- to 10-point worst pain scale. (Their baseline worst pain scores were higher than 6 points on average.)
The women were randomized to receive true or sham acupuncture or to be placed on a waiting list. The sham treatments used real acupuncture needles inserted at nonacupuncture points and at shallower depths. Patients receiving real or sham acupuncture had treatments twice a week for 6 weeks and then once a week for another 6 weeks, and all of the patients were followed up for an additional 12 weeks.
At the 6-week mark, true acupuncture resulted in a significant reduction of nearly 1 point in worst joint pain compared with sham or no treatment, the primary outcome.
Clinically Meaningful to Patients
However, the study was designed to detect a larger between-group difference than it did. The researchers powered their study to look for a 2-point average difference in worst pain between groups as an indicator of clinical meaningfulness, a bar that they didn’t meet.
But Hershman believes the findings are clinically important. She and her coauthors now say that a 0.7- to 1-point difference between groups is clinically meaningful in most pain studies. In her study, the average difference between the acupuncture and sham or no-intervention groups at 6 weeks was 0.92 point and 0.96 point, respectively, which she says meets the threshold for between-group differences used in other studies.
Additionally, patients consider a 2-point absolute decrease in their own pain to be clinically meaningful, according to recommendations from the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials. In a post hoc analysis of Hershman’s study, 58% of women who received true acupuncture had at least a 2-point reduction in their pain at 6 weeks compared with 33.3% in the sham acupuncture group and 31.4% in the wait-list control group.
There were also improvements in prespecified secondary outcomes including average pain, pain severity, and stiffness at 6 weeks in the true acupuncture group. And at the 6-month mark—12 weeks after the treatments stopped—average worst pain was still 0.59 point and 1.23 points lower in the acupuncture group compared with sham and wait list, respectively.
Despite the suggestion of a benefit for acupuncture, JAMA’s editors required the primary outcome measure—which did not achieve the prespecified 2-point difference between groups—to be reported as the main finding. This finding wasn’t acknowledged in a highly publicized abstract of the study presented at the San Antonio Breast Cancer Symposium in December.
“Therefore, we were required to say that the improvements were of uncertain clinical importance even though I think in our hearts we feel that this is a very significant study,” Hershman said.
Addressing Criticism
“In the formal sense, this study did not reach its primary end point,” said Deborah Schrag, MD, an oncologist at Dana-Farber Cancer Institute in Boston and the JAMA associate editor who handled the article. But, she added, “they did achieve a clear and consistent reduction in women’s symptoms.”
Ed Livingston, MD, a deputy editor at JAMA, said there is some debate regarding how to interpret the findings of a study when the primary end point is negative but many of the secondary end points trend in the positive direction, suggesting an effect of the intervention. “Some believe this is good enough to conclude that an intervention is effective,” he said.
In Schrag’s view, the study provided “important knowledge for women and their physicians to consider.”
The article caught her eye immediately for a number of reasons, starting with the common clinical problem it addressed. She was also impressed by the study’s multicenter design, including urban and rural sites; rigorous assessments of pain at multiple time points, which extended past the intervention period; clearly defined end points; and minimal missing data. Acupuncturists in the study also underwent extensive in-person training, video observation, and annual assessments. But what Schrag found most compelling was the study’s randomization with sham acupuncture. Many previous studies of acupuncture simply compare it with usual care or no treatment and don’t include an active control.
“One of the challenges that’s plagued many acupuncture studies is that they’re subject to placebo effects [and] it’s hard to know whether it’s the acupuncture or just the act of doing something,” Schrag said. “But these investigators conducted a blinded trial [with a sham control], and that’s a particularly strong design.”
Still, Hershman knows critics will take issue with the sham treatments, which weren’t double-blinded. Practitioners knew they were giving fake treatments and may have sent subconscious signals to their patients.
Using retractable needles designed to blind practitioners to the sham wouldn’t have helped, Hershman said, because studies suggest that both patients and acupuncturists can tell the difference.
“We looked at a whole slew of studies…of different sham modalities,” Hershman said. “They all had advantages and disadvantages. We felt in looking at the collective literature out there that this was probably the best approach.”
In the end, patients in the true acupuncture group were more likely to believe that they were receiving true acupuncture (68%) than those in the sham group (36%). But, Hershman said, their belief didn’t affect reductions in pain scores.
Acupuncture’s unknown mechanism of action is also likely to arouse criticism. Schrag understands the skepticism. “I agree that that’s perturbing about acupuncture,” she said, “but when there are rigorous studies that demonstrate a benefit, I think it’s still hard to ignore. If this were a very toxic or expensive intervention, there might be less enthusiasm, but acupuncture is widely available and well known to be quite safe.”
The experience of receiving acupuncture treatments—whether they’re real or fake—might be a factor in positive outcomes in acupuncture studies, Hershman acknowledged. “Whether or not patients were getting true acupuncture or sham acupuncture, they may have gotten other psychological benefits from just having the attention,” Hershman said of her study. Still, the magnitude of improvements in pain was larger in the true acupuncture group, which suggests additional mechanisms at work.
To investigate possible biological mechanisms, the researchers will measure effects on inflammatory and other biomarkers in serum collected from study participants at multiple time points. (This line of research might also shed light on why AIs cause musculoskeletal pain in the first place.)
But at the end of the day, Hershman’s belief is that it doesn’t really matter why acupuncture works, as long as it does: “Ultimately what matters is making sure that patients feel better and stay on their medicine. If 60% of patients had a significant reduction [in pain], then I think that we may have done them a real service.”