Osteoarthritis is one of the more common of the chronic degenerative diseases afflicting an increasingly older population. It affects primarily the weight-bearing joints (knee, hip), causing pain, stiffness and impotence, ultimately crippling patients and reducing quality of life. There is no curative treatment for this disease, and its final outcome is often joint replacement of the hip or knee. However, because of the limited life expectancy of these replacements, they are usually postponed as much as possible. In the meantime, symptomatic pain relief can be obtained with analgesics such as paracetamol, or the more effective non-steroidal anti-inflammatory drugs (NSAIDs) 123. NSAIDs, while generally safe when used at low doses and short term 456, can result in serious complications (gastrointestinal bleeding, renal failure, coronary heart disease) when used long-term or at higher doses in elderly patients, which is often the case in patients with osteoarthritis 789. Different products have been proposed for long-term symptomatic relief of osteoarthritis, some of them purporting to have some chondroprotective effects, such as chondroitin sulphate, diacerhein or glucosamine 10111213. Though meta-analyses are not very encouraging for direct effects of these adjuncts 14, we found that chondroitin sulphate might be associated with less use of analgesics or NSAIDs 15. Some food supplements might also have effects on osteoarthritis symptoms 161718192021, but few of these studies are methodologically valid. Authors have recommended testing fish-oils in patients with arthritis, and supplementation with minerals such as zinc or selenium, and vitamin E 22.

Patients using one such marketed multi-component food supplement reported reduced use of NSAIDS, at least for some time. Though it seemed highly unlikely that this supplement could suppress symptoms of OA, decreasing the use of NSAIDs might be a desirable objective in these high-risk long-term users. We therefore performed a randomized double-blind placebo-controlled clinical trial of this marketed food supplement, on the use of analgesics and NSAIDs on pain, stiffness and function scales of osteoarthritis to ensure that any decrease in analgesic and anti-inflammatory use was not at the expense of patient function.

Ninety-five patients were eligible for pre-inclusion. Of these 14 were excluded during the pre-inclusion period, for personal reasons (n = 9), for insufficient severity of OA (n = 3), for excess severity of OA resulting in hospitalization (n = 1), for severe depression resulting in psychiatric hospitalization during the pre-inclusion period (n = 1).

Eighty-one included patients were therefore randomized to active treatment (41 subjects) or placebo (40 subjects).

Five patients did not complete the study: one in the active treatment group because of an adverse event (diarrhoea with positive rechallenge); four in the placebo group: two for lack of treatment effect, one for adverse event (vomiting and epigastric pain), and the last became pregnant and was excluded from the study. All dropped out before the end of the first study month. None contributed any data to the study beyond inclusion. Patient disposition is shown in Figure 1.

In the active treatment group there were 14 men and 27 women, mean age was 56.8 years, (range 28 to 79). In the placebo group there were 12 men and 28 women, mean age was 57.5 (range 28 to 84). Overall 41 patients were fully assessable in the active treatment group, vs. 40 in the placebo group. The patients who dropped out of the study were not different from those who completed the study. Two patients, one in each study group, were wrongly included, their age (28) being below the inclusion limit. Both had post-traumatic knee OA, radiologically confirmed with osteophytes, combined with overweight. Removing them from the study did not change study results. One 84 year-old patient was above the age range.

There was no difference between the active and placebo groups for initial distribution of any of the studied parameters (Table 1).

This randomized double blind placebo-controlled clinical trial of Phytalgic^®, a composition marketed as a food supplement, showed that after three months patients randomized to Phytalgic^® used fewer concomitant analgesics and NSAIDs than patients randomized to placebo, and had better WOMAC scores. This effect increased regularly over the whole study period. There were no adverse reactions attributed to the drug other than digestive (flatulence and diarrhoea, fish-smelling eructations). The prespecified endpoints were reached and the study can be seen as a first demonstration of an effect of this product, confirming anecdotal reports of patient improvement and decreased use of NSAIDs, and the validity of some recommendations 22.

Even though this is the first known trial of this specific combination, so that we really did not know what to expect, the magnitude of the effect we found, which is homogeneous over the different metrics used, was greater than we expected. The study had been powered to detect a 20% difference in the use of NSAIDs and analgesics over the study period. What we found was a greater than 50% reduction in the use of analgesics and NSAIDs.

We attempted to avoid the more common biases: Treatment allocation was random, and conformed to the usual methodology of clinical trials; the study was double-blinded, placebo and active treatments were identical in all respects. Though two patients on active treatment did complain of fish-smelling burping that might compromise blinding in these patients, removing them from the statistical analysis did not change the results of the study. Of course, this being a single-centre study should not be considered as definitive proof. Further, large scale studies will be needed to confirm these results, and possibly clarify optimal dosing and duration of treatment, maybe identify responders and non-responders, and so on.

We included run-of-the-mill patients with diverse manifestations of OA: unilateral or bilateral osteoarthritis of the knee and/or hip, which is an advantage and a drawback: an advantage because it can be seen as a real-life population, increasing generalisability of the results, a drawback because the patients were heterogeneous. Even though there did not appear to be any difference in response between patients with knee of hip OA, the numbers in each group were not sufficient for subgroup analysis. Any differential response between hip and knee OA would need to be explored in further studies.

The composition tested here, Phytalgic^®, was devised empirically, from products used traditionally to treat osteoarthritis. It has been used commercially for several years, and the present study was initiated following patient reports of activity. The exact mechanism of action of the composition remains unknown, as are its possibly active components. Most of the components of Phytalgic^® are commonly used in practice 2126, but clinical trials are few and often of low methodological quality. This specific combination has never to our knowledge been tested before in a randomized double-blind clinical trial, though individual components may have been: Fish oils and oils rich in omega-3 and/or omega-6 fatty acids are thought to have a possible effect on rheumatoid arthritis 27 and osteoarthritis 18, with experimental support 28. A clinical trial of cod oil in osteoarthritis however found no clear effect 29, but the oils in the present composition come from cold-water fish rich in omega-3 and fatty acids (mackerel, herring, from the southern Pacific). Zinc is one of the essential elements for chondrocyte function, and has been associated with decreased inflammation 30 and interleukin production 31, but this would affect patients with rheumatoid arthritis more than those with OA. Nettles (Urtica dioica) are commonly used to treat osteoarthritis 2632. They may have an analgesic effect 33, and have also been found to reduce the use of NSAIDs in patients with osteoarthritis 34. Selenium (from Urtica dioica) alone or associated with vitamins A, C, E has been thought to modify the symptoms or the evolution of OA 35, but recent clinical trials and meta-analyses are rather negative in this regard 3637.

Despite these mixed results and the calls for more clinical trials such as this one 21, there seems to be an agreement that fish-oil and essential minerals are appropriate to use in patients with OA, and in this respect the preparation tested here appears to correspond with recommendations 22.

The effect of the present combination may be related to any one of the individual components, or more probably to the association of these compounds, since none of the compounds used alone seem to have demonstrated effects of the magnitude found here.

Osteoarthritis is a debilitating disease that is not usually by itself life-threatening, but is painful and crippling, and severely degrades quality of life. Its treatment is mainly symptomatic, using analgesics and non-steroidal anti-inflammatory drugs, whose adverse effects (mostly digestive or renal, but also possibly vascular) are such that they may in fact be one of the main life-threatening risks of the disease. Because of these risks inherent in the analgesic and anti-inflammatory drugs, decreasing drug use, especially of NSAIDs, could be a relevant treatment endpoint. However, reduced use of analgesics should not be associated with increased pain or reduced quality of life, but should indicate in fact less need of the drugs. This is clearly the case here: there is in our study an improvement of WOMAC scores, in all three areas covered by the scores. This symptomatic improvement was regular and increased over time, and in parallel the use of analgesics and NSAIDs decreased substantially.

WOMAC scores did not change much over the course of the study in the placebo group, though the use of analgesics and NSAIDs decreased by 20-30%. Though an improvement in OA scores with placebo is common, this is usually found in trials of single treatments, when a new treatment is given as a replacement of previous treatment or as new-onset treatment. In our patients, usual treatment was pursued. The fact that our patients were treated with analgesics and NSAIDs at baseline may also explain the relative modesty of the baseline WOMAC scores: These baseline scores are not raw untreated OA scores, but those obtained under usual, recognized effective treatment.

This is the first randomized clinical trial of this compound. Its results need to be confirmed in other, multicentre studies, and in more varied types of OA patients, to ensure these findings are not just the result of some undetected bias, or a statistical fluke, however homogeneous and consistent they may seem.

If these results are confirmed, longer-term studies will be needed to test for the persistence of this effect. Studies in wider populations will also indicate any real gain in terms of avoided adverse reactions because of less use of analgesics and NSAIDs. Considering the size of our population and the duration of our study, we could not demonstrate this endpoint, nor did we expect to.

In the meantime, this study demonstrates that use of three capsules a day over three months of this nutraceutical compound might decrease disease scores in patients with osteoarthritis of the knee and/or hip, and reduce their use of analgesics and NSAIDs.

BOCF: baseline observation carried forward; DDD: defined daily dose; ITT: intent to treat; NSAIDs: non-steroidal anti-inflammatory drugs; OA: osteoarthritis; PET: Paracetamol 500 mg-Equivalent Tablets; SD: standard deviation; WOMAC: Western Ontario-McMaster University Osteo-Arthritis Index.

The trial was funded by Laboratoires Phythea. Laurent Mallet is an employee of Laboratoire Phythea. Alain Jacquet is the main clinical investigator, and was compensated by Laboratoire Phythea for the time and activities related to the study. Nicholas Moore, Pierre Olivier Girodet and Antoine Pariente are employees of Bordeaux University and Hospitals, did not receive any compensation for the study, and have no financial or other conflict of interest with the study. Karelle Forest is a salaried employee of University of Bordeaux, and did not receive any compensation for this study.

AJ was the main investigator in the study. He contributed to the protocol, to the data acquisition, and to data analysis. LM initiated the study within the Phythea laboratories, discussed the protocol, provided the financing for the study, and contributed to the discussion concerning putative mode of action of the product tested. NM contributed to the protocol design, to the analysis and interpretation of the study, and wrote the initial manuscript. KF did the quality control and data management for this study. AP and POG contributed to study protocol, to patient selection and to data analysis, and to the evolution of the paper. All authors read the final version of this paper and were given the opportunity to comment. All approved the final version of this paper. NM is the guarantor of this paper.

AJ is an independent investigator working with the Department of Clinical Pharmacology, University of Bordeaux; AP is a university hospital practitioner in clinical pharmacology, Bordeaux University Hospitals and INSERM Clinical Investigation Centre at Bordeaux (CIC0005); POG is Master of Conferences and Hospital Practitioner in Clinical Pharmacology, University of Bordeaux, Bordeaux University Hospitals and INSERM Clinical Investigation Centre at Bordeaux (CIC0005); KF is clinical research assistant in the department of Pharmacology, University of Bordeaux, in charge of quality assurance and data monitoring; LM is Medical Director, Phythea Laboratories; NM is Professor of Clinical Pharmacology, Head of the Department of Pharmacology, and coordinator of the Clinical Investigation Centre, University of Bordeaux, Bordeaux University Hospitals, and INSERM CIC0005.