Could CMS (Centers for Medicare and Medicaid Services) be preparing to change its coverage policy for HA (hyaluronic acid) injections in the knee?
On December 15, 2014, CMS’ partner agency, the Agency for Healthcare Research and Quality (AHRQ) Technology Assessment (TA) program, issued an assessment report of HA (“Systematic Review for Effectiveness of Hyaluronic Acid in the Treatment of Severe Degenerative Joint Disease (DJD) of the Knee.”)and appeared to set the stage for second look at CMS’ national coverage policy for HA.
The AHRQ report said, essentially, that HA’s ability to improve knee function, patient quality of life (QoL) or the likelihood of a future knee replacement (KR) surgery was difficult to prove. In fact, said AHRQ, the evidence available was inconclusive.
(AHRQ’s Technology Assessment (TA) Program develops systematic reviews, health technology assessments, and other reports at the request of the Centers for Medicare & Medicaid Services Coverage and Analysis Group. These reports are used to inform Medicare’s national coverage policies.)
The report, which will be taken OFF the AHRQ website in a couple of weeks (January 12, 2015), will be available on Orthopedics This Week at https://ryortho.com/wp-content/uploads/2014/12/December-15-2014-AHRQ-Report-titled-djdkneepain-draft.pdf for as long as necessary.
Why This Initiative to Re-Visit HA Coverage?
Orthopedic and pain physicians have been using visco-supplementation for decades. They use it to treat pain and stiffness arising from osteoarthritis (OA) of the joints. CMS and every major payer have been reimbursing these physicians for using visco-supplementation since the 1990s.
By far, the most common form of visco-supplementation is hyaluronic acid (HA). All HA products, whether derived from rooster combs or bacterial fermentation, have been through the gauntlet of the FDA clinical study process and have met the FDA standards for safety and efficacy.
But the FDA is not CMS. CMS, to put it simply, writes the checks and has from time to time diverged from the FDA’s assessment of medical products. (Remember the Charité?)
Here is what CMS said as they were announcing this new initiative:
“If hyaluronic acid is effective, it is postulated that it might effectively prevent or delay the need for life-disrupting surgery and rehabilitation by relieving pain and improving function with minimal inconvenience or adverse effects; however, if the treatment delays arthroplasty but fails to halt progressive degeneration, patients could potentially experience worse outcomes, although thus far, evidence for such outcomes has been weak.”
Therefore:
“The scope of work for this task order includes an assessment of the evidence that hyaluronic acid injections prevent or delay the need for arthroplasty among individuals 65 and over.In addition to assessing the evidence for a role of HA in delaying or preventing the need for KR, this report aims to assess the evidence to date on the efficacy of intra-articular injections of HA with respect to the outcomes of function, ADLs/IADS [Activities of Daily Living/Instrumental Activities of Daily Living Scale], quality of life, and pain and on the safety of HA when used as indicated.”
Interesting.
So CMS wants to know if intra-articular injections of HA (which cost about $600 each, give or take) can reduce the rate of knee replacement (KR) surgery (which costs about $30, 000 each, give or take) and, furthermore, whether they affect knee function, quality of life or pain.
What’s In the Medicare Data Already?
Medicare, of course, has one of the largest existing databases of patient diagnosis, treatment and outcomes related to the use of HA injections.
PearlDiver Technologies, Inc., the medical database and software company which has approximately eight years of Medicare data and which just signed a huge deal with Humana to include their billions of patient records in their already massive database, checked the Medicare database for incidence rates of TKA (total knee arthroplasty) in HA patients.
PearlDiver’s analysts found 198, 391 patients who’d had a diagnosis of OA of the knee, had at least one round of HA knee injections and for whom two years of post-treatment data was available. Of that group, 1 out of 12 (8.1%) returned to have a total knee replacement within 24 months. The following table gives the results of the PearlDiver analysis.
AAOS and ACR HA Clinical Practice Guidelines
AHRQ’s inquiry comes right on the heels of AAOS’s (American Academy of Orthopaedic Surgeons) stunning 2013 announcement where it reversed its long standing support of HA and recommended against the use of hyaluronic acid to treat patients with symptomatic conditions. In 2012 the American College of Rheumatology (ACR) updated their 2000 guidelines for the treatment of osteoarthritis of the knee, hip, and hand and conditionally recommended hyaluronic acid injections for patients who had an inadequate response to initial (standard or more conservative) therapy.
AHRQ’s Review
AHRQ’s researchers started with 2, 214 peer-reviewed articles or abstracts which amounts to, essentially, the largest such literature review of HA injections to date.
Of those 2, 214 initial titles, AHRQ’s staff culled 415 out for full-text review. The remaining 1, 799 titles and abstracts were excluded because they were animal or in vitro studies (372), did not report on degenerative joint disease (DJD) of the knee (333), did not use intra-articular HA injections (357), did not report any outcomes of interest (53), had an inappropriate study design (e.g., obvious commentaries or non-systematic reviews) (390), did not enroll a population of interest (29), or were written in a non-English language (257). Eight articles were simply unobtainable.
Then the AHRQ researchers began reviewing those 415 studies and rejected 274 of those because the language was not in English (7); study design was flawed (78); participants were excluded (5); interventions were not of interest (4); outcomes were not of interest (75); the average age of the study participants was less than 65 (71); adverse event (AE) reports with sample size less than 500 (27); or duplicate data (7).
So…141studies formed the basis for AHRQ’s conclusions regarding the use of HA. Here are those conclusions.
HA’s ability to affect knee replacement surgery? Inconclusive.
AHRQ found only 3 randomized trials and 13 observational studies which in any way measured KR following HA treatment.
And…there were problems with those studies. Two of the trials didn’t explicitly test the hypothesis that HA injections could affect the future knee replacements. The average age of the patients in the studies were under 65—not good for a Medicare analysis.
The AHRQ researchers did find one large observational study which calculated the rates of KR by age group, pain at baseline, and various other factors. That study found that age was the only factor associated with the likelihood of undergoing KR.
OA patients between the ages 60-69 were significantly more likely to have KR than patients under 50. That same study also found that the median time to KR for HA recipients was 1.8 years. Seventy-five percent of the patients in this study had not had KR by 3.8 years. The authors stated that the average interval between entering care in their practice and undergoing KR for patients not treated with HA was 3 months.
AHRQ could draw no conclusions from these studies.
HA’s ability to improve knee function? Inconclusive.
AHRQ found 18 randomized trials reported that specifically measured the effects of HA to either sham-injected placebo control, another HA, or some other active treatment on function. In these trials, “function” was measured using WOMAC (Western Ontario and McMaster Universities Arthritis index), Lequesne index, KOOS (Knee injury and Osteoarthritis Outcome Score), or ADLs. Even better, the patients in these studies had an average age of 65 or older.
Those hardy AHRQ researchers then pooled the results of 10 of the placebo-controlled studies that reported outcomes using a WOMAC or Lequesne scale. That exercise uncovered a small increase in function in favor of HA (-0.23, 95%CI -0.34, -0.02).
The bad news, however, was that the difference translated to a change of 5.6 units on a 0-100mm VAS (Visual Analog Scale) scale which did NOT rise to a clinically significant amount. One of the studies actually said that the placebo was more effective than HA. Another study found no difference between control and HA.
Four of double-blind placebo-controlled trials reported on function and said that the proportion of patients who achieved a pre-specified level of overall improvement was higher for HA-treated patients than placebo-treated patients.
None of the studies AHRQ looked at for function stratified their findings by age, gender, or any other outcome of interest.
AHRQ’s conclusion: Strength of evidence for the effect of HA on function compared with placebo was low.
HA’s ability to reduce knee pain? It works, but not enough.
Bingo! The AHRQ folks found a large, comprehensive, relatively good quality 2012 systematic review that compared the effects of HA with sham or non-intervention controls and reported that HA injections significantly reduced pain, both statistically and clinically (that is, reaching the MCID–minimal clinically important difference) when measured at three months.
But…then the AHRQ researchers found that the ability of HA to reduce knee pain lessened to non-significance when only studies with blinded outcome assessment and at least 100 participants per study arm were included in the analysis.
When the AHRQ staff then conducted a stratified analysis that compared 54 studies with sham controls with 18 studies with non-blinded controls they found that HA DID have a statistically significant effect on pain. It, quoting directly from their study: “nearly met the MCID”—meaning it was almost clinically significant.
Almost.
So, again, AHRQ concluded that the strength of evidence that HA reduced pain was low. And when the evidence did show pain relief, it was minor and did not rise to minimum clinical importance.
HA’s ability to improve Quality of Life? Inconclusive.
AHRQ found three trials that tested and reported on the effects of HA treatment on quality of life. One trial, which compared Hyalgan to treatment with two different volumes of saline, found NO change in quality of life from baseline using the KOOS quality-of-life component.
The AHRQ folks then found two head-to-head trials that compared Synvisc with Hyalgan and where there was some attempt to assessed quality of life, one using the SF-36 mental component summary to assess QoL68 and one using the EuroQol-5D index to assess HRQoL.
The trial that used the SF-36 reported no increase in quality of life for either group.
The trial that used the EuroQol-5D reported a slight increase in QoL for Synvisc from 3 through 12 months but only an increase in HRQoL for patients on Hyalgan at 3 months.
Again, AHRQ concluded that the strength of evidence for any conclusions regarding an effect of HA on quality of life is insufficient.
AHRQ Questions HA’s Clinical Importance
After 200 pages of review and discussion, AHRQ concluded that it could not find clear, clinically important evidence that intra-articular HA injections delay or avoid KR nor improve function or quality of life for patients 65 years of age or older.
Next Steps
For the 27 million patients who have been told by their physicians that the pain they feel is due to osteoarthritis of the knee, this discussion is extraordinarily important. For decades physicians have relied on HA injections to treat patients with degenerative knee disease. And for decades CMS has reimbursed these injections.
Never mind the indisputable fact that the FDA has already reviewed the clinical safety and efficacy for HA, if AHRQ’s assessments become the basis for CMS’ national coverage policies regarding HA then Medicare coverage for HA could change.
Stay tuned, for sure.
