Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Avocado/soybean unsaponifiables for osteoarthritis

Background
Search
Trials
Results
Comment

Bandolier is occasionally taken to task over sins of omission. One reader commented that we paid insufficient attention to food additives for the treatment of osteoarthritis in a Cochrane review. The review [1] did indeed suggest that two trials of Avocado/soybean unsaponifiables (hereafter called ASU) suggested efficacy. A quick search found a third study, so a more detailed review seemed in order.

Background

ASU refers to extracts derived from one-third avocado oil and two-thirds soybean oil after hydrolysis. Chemical names and structures we do not have in detail, but this is presumably a fairly complex mix of chemicals. The main thing is that experiments show this mix to interfere with various cellular mediators, and prevent deterioration of synovial cells caused by interleukins, plus stimulating collagen synthesis in culture.

Search

The Cochrane review [1] had two trials [2,3], and a search of PubMed using free text terms avocado and arthritis produced two more [4,5].

Trials

All four trials (751 patients) were randomised and double blind, and compared ASU (predominantly 300 mg once a day) with placebo in people with osteoarthritis of knee or hip (Table 1). All used recommended standards for osteoarthritis trials, and all patients entered into the studies had osteoarthritis, pain of at least moderate intensity, and were using NSAIDs for pain control or had established pain of at least six months duration.

Table 1: Details of trials for ASU in OA


Reference Design Treatments Outcomes Results
Blotman et al. Revue du Rheumatisme 1997 64: 825-834
France
[2]
Randomised, double-blind, placebo-controlled, add-on to NSAID therapy
Patients aged 45-80 with OA hip or knee using ACR criteria, using NSAIDs (mean 80 mg diclofenac), and with pain 25/100 mm or more
Quality score 4/5
ASU 300 mg per day or placebo
Three months, with NSAID continued for first 45 days, with reduction according to need in second 45 days
164 patients recruited
Consumption of NSAID
Pain score
Global evaluation by patient
Adverse events
Discontinuations
Groups identical at baseline
Using NSAIDs at end: 33/77 ASU, 53/76 placebo
No difference in pain score
Global good and very good at end: 46/76 ASU, 25/73 placebo
Patients with AE second 45 days 1 ASU, 4 placebo
AE withdrawals 1 each group
Maheu et al. Arthritis & Rheumatism 1998 41: 81-91
France
[3]
Randomised, double-blind, placebo-controlled, add-on to NSAID therapy
Patients aged 45-75 with OA hip or knee using ACR criteria, using NSAIDs ,and with pain 30/100 mm or more
Quality score 4/5
ASU 300 mg per day or placebo
Eight months, with NSAID withdrawn for 15 days before start, and then allowed as necessary
164 patients recruited
Lequesne index
Pain score
Functional ability
NSAID use
Analgesic use
Global evaluation by patient
Adverse events
Discontinuations
Groups identical at baseline
Significant improvement in Lequesne index, pain score (by mean 10/100 mm), functional ability
Use of NSAID at end: 34/84 ASU, 44/78 placebo
Global good and very good 53/84 ASU, 30/78 placebo
Patients with AE 23/85 ASU, 20/79 placebo
AE withdrawals 1/85 ASU, 3/79 placebo
Appelboom et al. Scand J Rheumatology 2001 30: 242-247
Belgium
[4]
Randomised, double-blind, placebo-controlled, add-on to NSAID therapy
Patients aged 45-80 with knee OA using ACR criteria, using NSAIDs (mean 135 mg diclofenac), and with pain 30/100 mm or more
Quality score 3/5
ASU 300 mg per day, ASU 600 mg per day, or placebo
Three months, with symptomatic use of NSAID or analgesic
260 patients recruited
NSAID use over days 30-90
Pain score
Lequesne index
Global evaluation by patient
Adverse events
Discontinuations
Groups identical at baseline
NSAID use reduced significantly for ASU groups, with lower pain score and better Lequesne index
More than 50% decrease in NSAID use by days 60-90: 46/86 ASU 300, 49/86 ASU 600, 24/88 placebo
At least good score by patient at end: 58/86 ASU 300, 47/86 ASU 600, 25/88 placebo
Patients with AE: 28/86, 24/86, 23/88
Lequesne et al. Arthritis & Rheumatism 2002 47: 50-58
France
[5]
Randomised, double-blind, placebo-controlled, add-on to NSAID therapy
Patients aged 50-80 with hip OA using ACR criteria, using NSAIDs (mean 55 mg diclofenac), with pain for 6 months or more, and joint space of at least 1 mm
Quality score 4/5
ASU 300 mg per day or placebo
24 months, with symptomatic use of NSAID or analgesic
163 patients recruited
Joint space at 24 months
NSAID use
Pain score
Lequesne index
Global evaluation by patient
Adverse events
Discontinuations
Groups identical at baseline
Joint space significantly less narrow in those with initial joint space less than 2.5 mm, but not more than 2.5 mm or total sample
NSAID use declined from 55 mg diclofenac equivalent a day to 25 mg a day in both groups
No difference in pain score (reduced from 50 mm to 31 mm) or Lequesne index
Global slightly better or better: 31/85 ASU, 30/78 placebo
Patients with AE: 39/85 ASU, 39/78 placebo
AE withdrawals 2 each group

All four trials had similar outcomes, measuring pain, function, NSAID use and patient global outcome, and paid attention to adverse events and withdrawals. One [5] had a primary outcome of radiological joint space narrowing. Intention to treat and per-protocol analyses were both usually presented. Duration was three months in two, six to eight months in one, and 24 months in the fourth.

The design was an add-on to NSAID therapy, with an assumption that at least one month, and possible three, was needed for maximum effect of ASU. All the trials allowed for adjustment of NSAID dose according to requirement, either through withdrawal during the trial, or withdrawal at the start, followed by add-back as needed.

Results

Details of trials are given in Table 1. Three trials [2-4] had better pain and functional ability with ASU + NSAID compared with placebo + NSAID over three to six months. In all three, better pain relief was achieved with a lower dose of NSAID. Both these effects took some months to develop. For example, Figure 1 shows pain reduction over three months and Figure 2 shows the concomitant average daily equivalent dose of diclofenac in one trial [3].


Figure 1: Pain scores with ASU and placebo




Figure 2: NSAID consumption with ASU and placebo



A fourth trial [5] showed no difference in pain, functional ability, or daily NSAID use over 24 months. This trial examined only patients with osteoarthritis of the hip, with a relatively low (50 mg diclofenac equivalent) initial daily use of NSAID. Joint space narrowed from an initial mean of 1.9 mm to 1.7 mm over two years in both groups, but with a significant reduction in narrowing in those patients whose initial joint space was below the median.

The number needed to treat (NNT) for one more patient to achieve at least a global outcome of good or better at the end of the trial with ASU + NSAID compared with placebo + NSAID was 3.6 (2.8 to 5.0; Table 2) in three trials [2-4]. All three trials had similar levels of efficacy (Figure 3). For one patient to be on no NSAID or to have an NSAID dose reduction of more than 50% was 4.3 (3.2 to 6.5). The one trial looking at dose showed no difference between 300 mg or 600 mg of ASU a day.


Table 2: NNTs for ASU compared with placebo when added to NSAID therapy in OA


Number/total with outcome
Outcome
Trials
ASU + NSAID
Placebo + NSAID
Relative benefit
95% CI
NNT
95% CI
Not using NSAID or dose reduced by more than 50% at end of trial
3
189/333
(57%)
81/242
33%)
1.7 (1.4 to 2.1)
4.3 (3.2 to 6.5)
Patient global outcome at least good at end of trial
3
204/332
(61%)
80/239
(33%)
1.9 (1.5 to 2.3)
3.6 (2.8 to 5.0)
Patients reporting any adverse event during the trial
4
123/418
(29)
92/318
(29)
1.1 (0.8 to 1.5)
not calculated


Figure 3: Global outcomes with ASU and placebo



In all four trials there was no difference in the number of patients reporting at least one adverse event (Table 2). Adverse event withdrawals were low and the same in the three trials reporting it.

Comment

This is interesting. We may be a bit hazy about what ASU are chemically, and be without an overabundance of literature on what they do, but there is some evidence from three well-designed and conducted trials that ASU achieves the double whammy of both reducing pain and reducing NSAID consumption. The fly in the ointment is that a longer study of osteoarthritis in the hip showed no difference at all in any outcome. The evidence we have is of good quality, the trials are valid, and we have a reasonable number of patients in the trials.

What are we to make of this? All four trials used ASU from one source (a company called Pharmascience in France), which probably sponsored three of the trials. So differences in the results do not come about because of different preparations. In the two trials looking at both knee and hip, similar benefit was found. The one looking at only knee osteoarthritis was positive, the one looking at only hip osteoarthritis was negative. It is something of a mystery. The only major difference is duration, and we must wait to see if other longer duration trials produce negative results.

Our NNTs come only from the three positive trials, because the negative trial did not report outcomes in a way we could use for NNT calculations. Given that NSAIDs are implicated with renal, cardiovascular, and gastrointestinal harm, reducing their use must be a good thing.

So should people with osteoarthritis take ASU? On the evidence we have, it might make well make sense over the short term. The problems we have are with supply (where can we obtain it?), availability on prescription (depends on local rules), pharmaceutical consistency and quality (need for international quality and control), and perhaps above all a definition of what it is we might be taking. A definition of what is in ASU exists, with background science to support clinical observations [6], so there is some evidence that this is more than voodoo science.

References:

  1. CV Little et al. Herbal therapy for treating osteoarthritis (Cochrane Review). In: The Cochrane Library , Issue 1, 2004. Chichester, UK: John Wiley & Sons, Ltd.
  2. F Blotman et al. Efficacy and safety of avocado/soybean unsaponifiables in the treatment of symptomatic osteoarthritis of the knee and hip. Revue du Rheumatologie [English edition] 1997 64: 825-834.
  3. E Maheu et al. Symptomatic efficacy of avocado/soybean unsaponifiables in the treatment of osteoarthritis of the knee and hip. Arthritis & Rheumatism 1998 41: 81-91.
  4. T Appelboom et al. Symptoms modifying effect of avocado/soybean unsaponifiables (ASU) in knee osteoarthritis. Scandinavian Journal of Rheumatology 2001 30: 242-247.
  5. M Lequesne et al. Structural effect of avocado/soybean unsaponifiables on joint space loss in osteoarthritis of the hip. Arthritis & Rheumatism 2002 47: 50-58.
  6. YE Henrotin et al. Avocado/soybean unsaponifiables increase aggrecan synthesis and reduce catabolic and proinflammatory mediator production by human osteoarthritic chondrocytes. Journal of Rheumatology 2003 30: 1825-1834.

previous or next story