Nutraceutical supplements in management of pain and disability in osteoarthritis: a systematic review and meta-analysis of randomized clinical trials

This study designed to evaluate the effect of nutraceutical supplementation on pain intensity and physical function in patients with knee/hip OA. The MEDLINE, Web of Science, Cochrane Library, Scopus, EMBASE, Google Scholar, Science direct, and ProQuest in addition to SID, Magiran, and Iranmedex were searched up to March 2020. Records (n = 465) were screened via the PICOS criteria: participants were patients with hip or knee OA; intervention was different nutritional supplements; comparator was any comparator; the outcome was pain intensity (Visual analogue scale [VAS]) and physical function (Western Ontario and McMaster Universities Arthritis [WOMAC] index); study type was randomized controlled trials. The random effects model was used to pool the calculated effect sizes. The standardized mean difference (SMD) of the outcome changes was considered as the effect size. The random effects model was used to combine the effect sizes. Heterogeneity between studies was assessed by Cochran's (Q) and I2 statistics. A total of 42 RCTs were involved in the meta-analysis. Nutritional supplementation were found to improve total WOMAC index (SMD = − 0.23, 95% CI − 0.37 to − 0.08), WOMAC pain (SMD = − 0.36, 95% CI − 0.62 to − 0.10) and WOMAC stiffness (SMD = − 0.47, 95% CI − 0.71 to − 0.23) subscales and VAS (SMD = − 0.79, 95% CI − 1.05 to − 0.05). Results of subgroup analysis according to the supplementation duration showed that the pooled effect size in studies with < 10 months, 10–20 months and > 20 months supplementation duration were 0.05, 0.27, and 0.36, respectively for WOMAC total score, 0.14, 0.55 and 0.05, respectively for WOAMC pain subscale, 0.59, 0.47 and 0.41, respectively for WOMAC stiffness subscale, 0.05, 0.57 and 0.53, respectively for WOMAC physical function subscale and 0.65, 0.99 and 0.12, respectively for VAS pain. The result suggested that nutraceutical supplementation of patients with knee/hip OA may lead to an improvement in pain intensity and physical function.


Scientific Reports
| (2020) 10:20892 | https://doi.org/10.1038/s41598-020-78075-x www.nature.com/scientificreports/ Existing recommendations for the management of OA consist of three major classes: pharmacologic (i.e. opioids, non-steroidal anti-inflammatory drugs (NSAID), and COX-2 specific drugs), non-pharmacologic (i.e. rehabilitation to facilitate healthy body composition, lifestyle, and physical activity) and surgical treatment [4][5][6][7] . Present pharmacological treatments simply have a palliative effect on the relief of symptoms whereas not considering the essential problem of the cartilage disorder. Additionally, long-term consumption of these treatments has possible adverse events that might result drastic outcomes such as gastrointestinal problems, unwanted cardiovascular effects and adverse events on the cartilage 8 . Meanwhile, nutritional intervention demonstrates a continuing approach for management and inhibiting OA as an accompaniment to the traditional treatment of OA [9][10][11][12] . Nutraceutical supplements, such as chondroitin sulfate (CS), glucosamine sulfate (GS) and Methylsulfonylmethane (MSM), have been applied to manage OA and relieve symptoms in recent years 13 . Nutraceuticals are described as dietary supplements that comprise a condensed form of a considered bioactive ingredient, initially isolated from food, however existing in a nonfood matrix, and consumed to preserve or increase health situation in the amounts beyond those accessible from common foods 13 . Nevertheless, there is no agreement in regard to applying the term "nutraceutical" or "dietary supplement". The "active aging" is a principle objective of dietary supplements, as indicated by the developing sales of vitamins and minerals 14 . Dietary bioactive combinations have been revealed to be impressive in the improvement of clinical symptoms and in decreasing inflammatory indices in subjects with OA 15 . Presently 69% of subjects with OA receive various forms of dietary supplements for their problem 16 .
Even though there are several publications in the medical literature in regard to the use of nutraceuticals as a complementary treatment of OA, there have been variable findings concerning whether or not these nutrients have any beneficial consequence. The purpose of this study is to perform a systematic review and meta-analysis of relevant randomized controlled trials (RCTs) to assess the efficiency of different dietary supplements in the management of the symptoms of hip/knee OA.

Methods
The primary purpose of this systematic review and meta-analysis was to evaluate the efficacy and safety of dietary supplements in subjects with knee or hip OA. The current study has been planned based on the instructions in the Cochrane Collaboration handbook and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. The study question was framed according to the PICOS (participants, interventions, comparators, outcomes, study design) criteria (Table 1), is as follows: Do nutraceutical supplements influence pain and functional status in patients with hip/knee osteoarthritis? Literature search. Several search strategies were employed to recognize eligible studies. A medical librarian (FB) in an argument with the team (DA, ND and FB) performed a precise and comprehensive academic literature search of the titles, abstracts and keywords of all studies for competency independently through electronic databases (MEDLINE, Web of Science, Cochrane Library, Scopus, EMBASE, Google Scholar, Clininaltrial.gov, Science direct, and ProQuest in addition to SID, Magiran, Irandoc, and Iranmedex for Persian language literature) up to January 2020. Duplicate studies were excluded. At the same time, a hand search of the related references and cited articles of the included studies was conducted to recognize other appropriate studies that were lost by electronic search.
Search terms included a mix of Medical Subject Headings (MeSH) and a literature search was performed using the following MeSH terms for key concepts (with assistance from a librarian) targeting dietary supplements and hip or knee OA such as : ("supplement "(All Fields) OR "nutraceuticals"(All Fields) OR "vitamin"(All Fields) OR "mineral"(All Fields) OR "plant"(All Fields)) AND ("OA" OR "osteoarthritis"(All Fields) OR "knee osteoarthritis"(All Fields) OR "hip osteoarthritis"(All Fields) OR "knee OA"(All Fields) OR "hip OA"(All Fields)). After the primary search, titles and abstracts were sent out from EndNote X7 into Microsoft Excel to be screened. Three reviewers separately reviewed all titles and abstracts and full texts (DA, ND, and MH). A fourth reviewer was conferred if discrepancies happened.
Inclusion and exclusion criteria. Inclusion criteria to choose studies for this systematic review and metaanalysis were: (1) RCT (either parallel or crossover designs); (2) a nutraceutical as an intervention either as an adjunctive to standard medicine or as a monotherapy and (3) adults who have been diagnosed with hip or knee OA; (4) sufficient data reported about mean changes for Western Ontario and McMaster Universities Arthritis (WOMAC) index (total score and subscales) and/or Visual analogue scale (VAS) at baseline and at the end of the trial in both intervention and placebo/control groups. Then selected possible clinical trials were excluded based Data synthesis and analysis. The number of subjects in each intervention group with mean and SD of study outcomes before and after the intervention was extracted from the articles included in the study. Then, the mean difference of study outcome was calculated and the mean difference of study outcomes was compared between the two groups. Because of the different scales used in the articles included in the study for the WOMAC index and VAS, the standardized mean difference (SMD) of the outcome changes between the two groups was considered as the effect size in this study. The random effects model was used to combine the effect sizes calculated in the articles. Heterogeneity between studies was assessed by Cochran's (Q) and I 2 statistics, which expressed the percentage of variations between studies. In case of high heterogeneity between included studies, we performed subgroup analysis according to the treatment duration (< 10 months, 10-20 months and > 20 months) to evaluate the impression of these factors on the results. The Meta package in R software was used for data analysis. A p-value less than 0.05 was considered as significant level.
Publication bias. Egger's Regression Test and Funnel Plot were used to evaluate the presence or absence of publication bias. Publication bias was assessed for each study outcome. The Trim and Fill method was used to investigate the effect of publication bias on the results of the study.

Results
Study selection process. The systematic searching of the databases identified 1323 articles, of which 858 were excluded as duplicates, 372 were excluded by title and abstract and 52 were excluded after reviewing full texts (Fig. 1). Efficacy of the intervention. WOMAC (total). The total score of the WOMAC was evaluated in the 28 articles reviewed. There were 1404 cases in the intervention group and 1360 in the control group. The mean follow-up duration of patients (lowest to maximum) was 17.4 (6-144) weeks. There was a significant heterogeneity between studies (Q-value = 110.58, df = 37, p-value < 0.001, I 2 = 66.5%). Based on the meta-analysis results, it was observed that the Pooled Standardized Mean Difference between the intervention and control groups was 0.23 units (SMD = − 0.23, 95% CI − 0.37 to − 0.08, z-value = − 3.09, p-value = 0.002). Figure 4 shows the forest plot of the combination of results. Results of subgroup analysis according to the supplementation duration showed that the pooled effect size in studies with < 10 months as short term, 10-20 months as medium term and > 20 months as long term supplementation duration were 0.05, 0.27 and 0.36, respectively. Figure 5 shows the forest plot of the subgroups by the supplementation duration.  Fig. 6. The pooled effect size in studies with < 10 months as short term, 10-20 months as medium term and > 20 months as long term supplementation duration were 0.14, 0.55 and 0.05, respectively. The forest plot of the subgroups by the supplementation duration is presented in Fig. 7.

WOMAC (stiffness).
In the included articles, 29 articles assessed the WOMAC Stiffness subscale. There were 1539 subjects in the intervention group and 1513 subjects in the control group. The mean follow-up duration of patients (lowest to maximum) was 17.76  weeks. There was a significant heterogeneity between studies (Q-value = 353.55, df = 38, p-value < 0.001, I 2 = 88.8%). The Pooled Standardized Mean Difference between the intervention and control groups was 0.48 units (SMD = − 0.48, 95% CI − 0.72 to − 0.24, z-value = − 2.88, p-value < 0.001). The forest plot of the combination of results is presented in Fig. 8. The pooled effect size in studies with < 10 months as short term, 10-20 months as medium term and > 20 months as long term supplementation duration were 0.59, 0.47 and 0.41, respectively. The forest plot of the subgroups by the supplementation duration is presented in Fig. 9.

WOMAC (physical function).
In the included articles, 29 articles assessed the WOMAC Physical Function subscale. There were 1496 subjects in the intervention group and 1494 subjects in the control group. The mean follow-up duration of patients (lowest to maximum) was 7.21  weeks. There was a significant heterogeneity between studies (Q-value = 583.74, df = 37, p-value < 0.001, I 2 = 94.0%) The Pooled Standardized Mean Difference between the intervention and control groups was 0.25 units (SMD = − 0.25, 95% CI − 0.57 to − 0.07, z-value = − 1.55, p-value = 0.12). The forest plot of the combination of results is presented in Fig. 10. The pooled effect size in studies with < 10 months as short term, 10-20 months as medium term and > 20 months as long term supplementation duration were 0.05, 0.57 and 0.53, respectively. The forest plot of the subgroups by the supplementation duration is presented in Fig. 11.   Fig. 12. The pooled effect size in studies with < 10 months as short term, 10-20 months as medium term and > 20 months as long term supplementation duration were 0.65, 0.99 and 0.12, respectively. The forest plot of the subgroups by the supplementation duration is presented in Fig. 13.
Publication bias for WOMAC index total score. Figure 14 illustrates a Funnel Plot to investigate the publication bias for the WOMAC index total score. According to Eggers Regression Test, the publication bias was not significant (t-value = 1.51, df = 36, p-value = 0.13).   Figure 15 illustrates a Funnel Plot to investigate the publication bias for the WOMAC index pain subscale. According to Eggers Regression Test, the publication bias was not significant (t-value = − 0.42, df = 39, p-value = 0.67). Figure 16 illustrates a Funnel Plot to investigate the publication bias for the WOMAC index stiffness subscale. According to Eggers Regression Test, the publication bias was significant (t-value = − 2.13, df = 37, p-value = 0.03). Trim and Fill test was performed to modify the publication bias and 11 studies added to adjust for the missed study through this method. The results of the Trim and Fill test demonstrate that the pooled effect size was 0.08 (Adjusted SMD = 0.08, 95% CI − 0.33 to − 0.16).

Publication bias for WOMAC index physical function subscale. Figure 17 illustrates a Funnel
Plot to investigate the publication bias for the WOMAC index physical function subscale. According to Eggers Regression Test, the publication bias was not significant (t-value = − 0.41, df = 39, p-value = 0.68).
Publication bias for VAS. Figure 18 illustrates a Funnel Plot to investigate the publication bias for the VAS. According to Eggers Regression Test, the publication bias was significant (t-value = − 3.03, df = 29, p-value = 0.004). Trim and Fill test was performed to modify the publication bias and 9 studies added to adjust for the missed study through this method. The results of the Trim and Fill test demonstrate that the pooled effect size was 0.35 (Adjusted SMD = − 0.35, 95% CI − 0.64 to − 0.07).
Adverse events. The adverse events and dropout rates are summarized in Table 3. The dropout rate ranged from 0 to 41%.

Discussion
This meta-analysis demonstrated that nutraceutical supplementation may lead to an improvement in total and also pain and stiffness subscales of WOMAC and VAS but did not affect WOMAC physical function subscale. The existing modalities for managing OA are basically symptomatic and have not been confirmed to slow, arrest or inverse the joint subversion and cartilage degradation progression 8 . For this reason, over the past few years, attention has been focused on the impact of nutritional supplements in managing and preventing OA, considering its risk-benefit ratio and low cost and great acceptance by patients. Nutraceuticals provide a great variety of    Sierrasil is an indigenous mineral product isolated from the Sierra Mountains in the USA with a cultural history of usage in the treatment of joint pain and established cartilage degradation reducing properties 61 . SKI306X is a herbal mixture (Clematis mandshurica, Trichosanthes kirilowii and Prunella vulgaris) applied for the management of inflammatory diseases and is clinically accepted for the treatment of OA in Far East Asia 62 . In the systematic review of RCTs by Ameye and Chee 2 moderate evidence was established for SKI306X in improving the symptoms in OA patients. Hyaluronic acid or hyaluronan (sodium hyaluronate) is accountable for the viscoelasticity and lubricating impacts of synovial fluid of the joint and has been shown to have the biophysical and biochemical roles in synovial tissues 63 . However, in a recent systematic review and meta-analysis by Liu www.nature.com/scientificreports/ et al. 64 , collagen hydrolysate, extract of the skin of the passion fruit (PFP), Curcuma longa extract, Boswellia serrata extract, pycnogenol and L-carnitine exhibited clinically important effects for pain alleviation in short term and only two supplements (green-lipped mussel extract and undenatured type II collagen (UC-II) showed clinically important effects on pain reduction at medium term. However, we founded that long term UC-II supplementation had considerable effects on WOMAC total and also WOMAC pain and physical function scale scores in patients with mild Knee OA. UC-II is a natural component which comprises a glycosylated, Figure 11. Forest plot presenting the impact of nutraceutical supplementation on WOMAC physical function score (subgroup analysis based on duration of supplementation). Medium term nutraceutical supplementation in OA patients. In the subgroup analysis, the greatest efficacy of nutraceutical supplements on WOMAC index total score and its subscales and also VAS was related to medium term supplementation (10 to 20 months). Most of these studies involved patients with mild to moderate knee OA according to Kellgren and Lawrence scoring system for classification of knee OA 60 or American www.nature.com/scientificreports/ College of Rheumatology Classification Criteria for Knee Osteoarthritis 65 which supplements were administered as an adjunctive to symptomatic treatments (NSAIDs and/or analgesics) except nine of them (three 33,47,54 involved patients with knee and/or hip OA, four 24,32,44,52 involved patients with severe knee OA and two 53,55 involved patients for which supplements were administered as a monotherapy and no concomitant treatment were allowed). www.nature.com/scientificreports/ Among studies with medium term of supplementation, WOMAC total score was considerably improved through supplementation with CS in patients with mild to moderate knee OA, Deer bone extract (DBE) in patients with moderate to severe knee OA and PFP and collagen peptides isolated from pork skin (PCP) in patients with mild to severe knee OA.
OA is described by damage of type II collagen and GAGs, which are present in the joint. The lessening of GAGs is an essential factor leading to enhanced cartilage deprivation in the OA. CS, a central structural part of cartilage, is a sulfated GAG. Investigations in animal models have suggested that dietary supplements of CS prevent articular cartilage depreciation 66 . This protecting consequence is related to the anti-inflammatory activities of CS by increasing the synthesis of hyaluronic acid and proteoglycans, and decreasing the production of proteolytic enzymes and nitric oxide 57 . Deer horn extract has been considered as a noteworthy health restorative in traditional medicine amongst several Asian countries 67 . Oily DBE and CPC were recently revealed to have anti-inflammatory properties and reduce the morphological deviations related with osteoarthritic cartilage damage in animal models of OA 68,69 .
The WOMAC all subscale scores were improved through medium term supplementation with A. paniculata purified extract (ParActin) (in patients with mild knee OA), DBE (in patients with moderate to severe knee OA) and MSM (in knee OA patients with unknown severity). PFP improved only WOMAC pain and physical function subscales in patients with mild to severe knee OA, Boswellia serrata extract improved only WOMAC pain and stiffness subscales score and VAS in patients with mild to moderate knee OA and Artemisia annua extract (ART) improved considerably only WOMAC stiffness subscale in knee OA with unknown severity.
Long term nutraceutical supplementation in OA patients. Regarding long term supplementation, skimmed milk containing probiotic Lactobacillus casei Shirota (LcS) had considerably effects on WOMAC total and also WOMAC stiffness scale score and UC-II had considerably effects on WOMAC total and also WOMAC pain and physical function scale scores in patients with mild Knee OA according to Kellgren and Lawrence scoring system for classification of knee OA 60 . Boswellia serrata extract improved WOMAC stiffness scale score There is a growing field of interest and research indicating the protective benefits of dietary polyphenols in decreasing risk for chronic diseases 59 through accepting electrons from free radicals, distracting chain oxidation reactions, and improving cellular antioxidative capability 16 . The results of several studies suggested that supplementation with polyphenols and botanical extracts (e.g., Boswellia serrata extract, PFP, ParActin, ART and cat's claw extract) decrease the serum levels of TNF-α and MMP-3 in synovial fluid in patients with knee OA compared with the control groups 53,70,71 . Cellular and animal models have suggested also the benefits of such compounds and food ingredients (e.g., probiotics) in inhibiting inflammatory paths and reducing the production of iNOS, COX-2 and MMP enzymes to decrease the catabolic destruction of the cartilage 16,[72][73][74][75][76] .
A very important point in our findings which must be considered is that GS and vitamin D with the greatest interest in administration and consumption among OA patients, do not exhibit a clinically significant effect on knee or hip OA. GS is a water-soluble amino monosaccharide, considered as a desired substrate for the biosynthesis of GG chains and is in great amounts in cartilage matrix and synovial fluid. Glucosamine was thought to afford building substrates for the cartilage extracellular matrix biosynthesis. Later studies have established additional clarifications for its anti-inflammatory and anti-catabolic properties. A Cochrane review of RCTs of all GS formulations in OA patients, restricted to studies with satisfactory concealment, failed to display any advantage of GS for pain 77 . Hereafter, GS was firstly suggested by European League Against Rheumatism (EULAR) It has been theorized that vitamin D supplementation in patients with knee OA might be a practicable and cost-effective approach for managing clinical symptoms and making a structural advance. However most clinical trials showed that vitamin D supplementation does not improve cartilage volume or knee pain 41,46,78 . In line with our findings, the results of a systematic review of RCTs covering 1189 patients by Hussein 79 did not recommend vitamin D supplementation in patients with knee OA.
Our study opens new horizons for the managing of degenerative joint diseases. We collected clinical trials of nutraceuticals and dietary supplements and the findings were really hopeful and encouraging. However, there is a need for more well-designed randomized clinical trials which can confirm the safety and efficacy of such products. This could help clinicians in endorsing them for OA patients.
The present study has some limitations that need to be considered in explicating the results of this systematic review and meta-analysis. Firstly, in spite of an increasing body of nutraceutical researches in subjects with OA, the number of studies included in this specific review after a systematic review of the existing scientific literature www.nature.com/scientificreports/ was lower than what would have been predicted. We believe that our inclusion criteria had a significant role, because we considered variables (i.e. VAS and WOMAC) that are not measured in many studies. Secondly, there is probable publication bias. Some unpublished abstracts and articles were not included because of unavailability. Thirdly, the language may lead to bias as we selected only the English and Persian language due to limited resources. These may considerably reduce our sample size and accordingly our ability to delineate statistically significant findings. Fourthly, the heterogeneity between the results is an issue need to be considered. Although we did a subgroup analysis, we were not successful to completely minimize these heterogeneities. Finally, there may be some possible aspects not considered in the present systematic review and meta-analysis, such as the severity of OA, region, and race. In spite of the stated limitations, this systematic review and meta-analysis provides the first systematic work to consider clinical trials on nutraceutical supplementation in relation to pain and physical disability in patients with knee/hip OA. In addition, subgroup analysis was implemented according to the nutraceutical type and we applied more suitable consequence indicators to direct this meta-analysis.
In conclusion, nutraceutical supplementation mostly along with symptomatic treatments (NSAIDs/ COX-2 inhibitors and analgesics) may effectively improve pain and physical function in patients with knee/hip OA. In the subgroup analysis, the greatest efficacy of nutraceutical supplements was related to 10-20 month (medium www.nature.com/scientificreports/ term) supplementation especially in patients with mild to severe knee OA. Despite recognized supplements with no established significant efficacy in our study (such as glucosamine and vitamin D), some not well-known supplements (Boswellia serrata extract, DBE, PFP, PCP, ParActin, ART and Pycnogenol) seem to have largest benefits in decreasing pain and improving physical function with negligible adverse events. It is recommended to trying these supplements in a safe doses along with conventional symptomatic treatments and physical therapy for at least 10 weeks especially for those with mild to moderate knee OA except low dose Sierrasil in addition to cat's claw extract, fortified bioactive micronutrient beverage, SKI 306X, L-carnitine, MPC and hyaluronic acid which are expected to have beneficial effects in decreasing pain and/or disability in less than 10 weeks of supplementation and also probiotic LcS and UC-II which are not anticipated to have favorable effects in less than 20 weeks of supplementation even in patients with mild knee OA. Other more precise outcome measurement tools, such as inflammatory biomarkers or image study, should probably be introduced into future studies to make them more convincing evidence.