This article is more than 5 years old. Information may no longer be current.
Hyaluronic acid meta-analysis results, methodology questioned
Meta-analysis found that HA injections not much better than placebo; the experts disagree with conclusion.
Click Here to Manage Email Alerts
|
COURTESY OF GENZYME BIOSURGERY |
A meta-analysis of 22 trials found that hyaluronic acid intra-articular injections used to treat knee osteoarthritis produced results only slightly better than placebo injections.
“The presence of publication bias suggests even this effect may be overestimated,” Grace H. Lo, MD, clinical epidemiology research and training unit at Boston University Medical School, and her co-authors, wrote in their abstract. The study was published in the Journal of the American Medical Association.
Their conclusions, based on a meta-analysis of the trials using hyaluronic acid (HA) treatments, have been widely criticized by experts who have used or researched the compound. The study’s methodology was also questioned by some of those who spoke with Orthopedics Today.
HA is a local therapy that replaces lost lubrication in the knee joint in order to reduce or eliminate osteoarthritis (OA) pain. The Food and Drug Administration approved it in 1997. It is typically administered through three to five weekly injections. In their NIH-funded study, the researchers stated that high molecular weight (HMW) HA may be more effective than its lower-weight counterpart.
Orthopedics Today Chief Medical Editor Douglas W. Jackson, MD, said he was surprised that the studies selected for inclusion by the authors may not represent all the results that have shown HA injections to be more successful. “Studies involving placebo groups are always high for many of the treatments for OA,” he said.
Lo and colleagues searched several sources for published and unpublished single- or double-blinded randomized controlled trials that compared intra-articular HA with intra-articular placebo injection for treating knee OA. To be included, studies could be conducted by investigators in any country, but they had to use one of five pain outcome measurements included in guidelines established by the Osteoarthritis Research Society International (OARSI), including the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and the Lequesne index.
This yielded 22 studies for which effect sizes were calculated (small effect, 0.2 to 0.5; large effect, 1.0 to 1.8). Results from the studies were pooled, and their heterogeneity was assessed using the Cochrane Q tests. Publication bias was accounted for using a funnel plot and the Egger test.
The pooled effect size for HA was 0.32 (CI 95%, 0.17-0.47). Study heterogeneity was significant (P <.001). Two studies evaluating one HMW HA compound, hylan GF-20 (Synvisc, Genzyme Biosurgery), showed effect sizes that exceeded 1.5. But, when all three hylan GF-20 trials were eliminated from the data analysis, heterogeneity was not significant (P=.58) and the HA effect size changed to 0.19 (CI 95%, 0.10-0.27).
In the comment section of their study, the researchers wrote that the effect of HA in treating knee OA “is equivalent to the effect of nonsteroidal anti-inflammatory drugs over that of acetaminophen.” The placebo effect of the injection alone likely accounted for approximately 80% of the treatment effect seen with HA intra-articular injections. Acknowledging that meta-analyses cannot take into account design flaws in the original studies, they called for a re-evaluation of the use of HA compounds based on their results, perhaps including independent intent-to-treat analyses.
Criticisms aired
According to Jackson, who authored a study on viscosupplementation injection techniques published in the Journal of Bone and Joint Surgery, American Volume in 2002, the concept is appealing because it allows delivering a treatment locally in order to avoid effects from taking systemic medications. “This article raises questions that it may not be as successful [a treatment] as we hoped, but viscosupplementation is the beginning of our approach to local treatment for local disease within the knee joint.”
He speculated that the small effect detected might be due, in part, to poor technique used by investigators in the original studies. That is hard to measure, Jackson noted.
Effect demonstrated
David D. Waddell, MD, an orthopedic surgeon at Louisiana State
University in Shreveport, has studied the effect that hyaluronan preparations
have on synovial physiology. “We have been able to demonstrate that
hyaluronan can indeed alter the synovial site response to the proinflammatory
cytokines IL-1ß and TNF
, which
secondarily has a role in the control and release of the metalloproteinases,
which are integral to the osteoarthritic process. That research, in vitro, has
led us to believe there are definitely scientific reasons for the hyaluronans
to have a role other than a placebo role,” he said.
Clinically, Waddell reported a significant therapeutic positive effect from HA products, including “the ability for patients to delay their total knee replacement for extended periods of time, sometimes approaching five years.” He said using HA helps avoid the risks and complications associated with nonsteroidal medications. “If we can find an effective alternative and decrease our use of nonsteroidals, that’s a very positive thing.”
HMW vs. LMW hyaluronans
One criticism Waddell aired was the inclusion of studies that investigated HA products not available in the United States, like Orthovisc (Anika Therapeutics and OrthoBiotech) and Suplasyn (Bioniche). “The meta-analysis is somewhat skewed because that information does not apply to what we have available.”
|
In addition, HMW and low molecular weight (LMW) products, which are administered and work differently, were seemingly compared to one another. “Our basic science research seems to support that HMW products may have a more positive effect on the synovial physiology than the lower ones do,” said Waddell, who also questioned the validity of including studies conducted at different times, 1988 to 2003, and in different countries.
“What I took away from this is that probably more, better designed studies should be done ... We need to study all of these new treatments more extensively than we have,” he said.
There were major concerns with the JAMA study, according to rheumatologist Roland W. Moskowitz, MD, of Case Western Reserve University. Among them was the authors’ inclusion or exclusion of studies based on their own preferential ranking of five pain outcome measures. While outcome measures were recommended by an OARSI task force, as indicated in the study, not all of them answered measured pain outcomes. “I don’t think OARSI recommended one over the other,” he said. This preferential ranking was not validated.
As a result, this subjective selection process allowed studies to be included that used less stringently validated pain outcome scales. “The WOMAC and Lequesne are considered the best validated [ones], especially the WOMAC pain scale, which is considered the gold standard for pain,” Moskowitz said.
Key trials omitted
Calling the comparative treatment a placebo is a misnomer because the placebo group was actually a positive control group in most of the studies evaluated, Moskowitz said. For example, in the 1998 sodium hyaluronate (Hyalgan, Sanofi-Synthelabo Inc.) study included in the meta-analysis, in which he was involved, placebo patients received saline injections and could use accepted therapeutic OA approaches, such as up to 4 g/day acetaminophen with physical therapy.
Publication bias is practically unavoidable and it should be assumed that nearly every study has the potential for it. “At least 90% of major studies done today are with pharmaceutical support. We couldn’t do these studies without pharmaceutical support,” Moskowitz said.
He questioned the authors’ assumption of homogeneous disease and demographic characteristics. In addition, he wondered why some key studies, like the Cochrane review of hylan GF-20 for knee OA conducted by Nicholas Bellamy Jr., MD, et al, which was presented at the American College of Rheumatology (ACR) 67th Annual Meeting, were not included.
“The Bellamy study was a beautifully done, carefully done study.”
Also omitted was a recently published study by Maxime Dougados, MD, and colleagues that reviewed 24 Hyalgan trials. “They state that it’s an effective therapy in the treatment of OA,” Moskowitz said.
The meta-analysis conclusions are counter to ACR guidelines and those of the European League Against Rheumatism for this therapy, he said. “I’m concerned an article like this will dissuade people from using what we think is an effective therapy.”
Manufacturers speak up
Richard P. Polisson, MD, MSHc, rheumatologist and senior vice president of clinical research at Genzyme, which manufactures Synvisc, acknowledged that Lo and co-author David T. Felson, MD, MPH, are prominent, methodological researchers.
But, he agreed with some of the criticisms, namely the significance of the new information presented in Bellamy’s study, problems with pooling the data for HMW and LMW compounds, and that by virtue of meta-analyses methods, data from a key management trial were missing.
Because certain assumptions are made in meta-analyses, several Synvisc trials that would have amplified the overall study results were not included, such as ones by Wobig (1999) and Dickson (2001), Polisson said.
“They excluded from the analysis some trials we think should have been put in because they fulfilled all the right methodologic criteria of being placebo controlled, double blind, randomized, with a defined endpoint that can be measured.”
The meta-analysis “cites studies associated with products that have either not been submitted to FDA for approval in the United States, or have been submitted and rejected,” officials at Smith & Nephew, which manufactures the HA product Supartz, said in a prepared statement. They urged doctors to choose HA products based on high-quality clinical studies.
Dr. Moskowitz is a consultant to Sanofi-Synthelabo.
For more information:
- Lo GH, LaValley M, McAlindon T, Felson DT. Intra-articular hyaluronic acid in treatment of knee osteoarthritis. JAMA. 2003;290:3115-3121.
- Bellamy N, Campbell J, Gee T, et al. Hylan G-F 20 for knee osteoarthritis (OA): A Cochrane review. Presented at the American College of Rheumatology 67th Annual meeting. Oct. 23-28, 2003. Orlando, Fla.
- Raynauld JP, Torrance GW, Band PA, et al. A prospective, randomized, pragmatic, health outcomes trial evaluating the incorporation of hylan G-F 20 into the treatment paradigm for patients with knee osteoarthritis (part 1 or 2); clinical results. Osteoarthritis and Cartilage. 2002;10:506-517.