The efficacy and safety of Curcuma longa extract and curcumin supplements on osteoarthritis: a systematic review and meta-analysis

Abstract Objective: To assess the efficacy and safety of Curcuma longa extract and curcumin supplements on osteoarthritis (OA). Methods: The databases such as Pubmed and Cochrane Library were searched to collect the article about Curcuma longa extract and curcumin in the treatment of OA. Then, randomized controlled trials (RCTs) were selected and their data were extracted. Finally, the RevMan5.3 was utilized for risk of bias assessment and meta-analysis, the STATA15.0 were utilized for publication bias assessment, and GRADE tool were used for the evidence quality assessment of primary outcomes. Results: A total of 15 RCTs involving 1621 participants were included. (1) Compared with placebo, Curcuma longa extract and curcumin (C.) can decrease the visual analog scale (VAS) and The Western Ontario and McMaster Universities (WOMAC) score-pain, the WOMAC score-function and the WOMAC score-stiffness. In terms of adverse events, Curcuma longa extract and curcumin are comparable with those of placebo. (2) Compared with non-steroidal anti-inflammatory drugs (NSAIDs), Curcuma longa extract and curcumin have similar effects on joint pain, function and stiffness. The incidence of adverse events in Curcuma longa extract and curcumin was lower. (3) Compared with the NSAIDs group, C.+NSAIDs can also decrease the VAS and WOMAC score-pain, the WOMAC score-function and the WOMAC score-stiffness. In terms of adverse events, the addition of Curcuma longa extract and curcumin to NSAIDs did not increase adverse events. Conclusion: Curcuma longa extract and curcumin may be a safer and effective supplement for OA patients. It is recommended to use Curcuma longa extract and curcumin supplement for OA patients for more than 12 weeks.


Risk of Bias Assessment
Two reviewer independently used the Cochrane risk of bias assessment tool to evaluate the quality of RCTs, and if there were disagreements, they were discussed with all reviewer [21]. The tool includes the following six aspects: random sequence generation, allocation concealment, blinding, incomplete outcomes, selective reporting and other biases. Each item is recorded as: low risk of bias, high risk of bias, unclear risk of bias.

Statistical analysis
RevMan 5.3 was used for risk of bias assessment and meta-analysis. I2 is used to test the specificity between RCTs. If there is homogeneity between RCTs (I%<50%, P>0.1), the fixed effects model is used for Meta-analysis. If there is heterogeneity between RCTs (I%>50%, P<0.1), we would first discuss the source of heterogeneity and conduct subgroup analysis. If the heterogeneity is not reduced, the random effects model is used for Meta-analysis [22]. For continuous variables, if the measurement data units are different or the values differ greatly, the standard mean difference (SMD) is used as the effect size indicator, while in other cases, the mean difference (MD) is used as the effect size indicator. For dichotomous variables, the risk ratio (RR) is used as the effect size indicator. The size of the effect is expressed with a 95% confidence interval (95% CI). The publication bias was detected by STATA 15 with Egger method (continuous variable) and Harbord methods (dichotomous variable) for primary outcomes. P>0.1 is considered to have no publication bias.

Results of the Search
Of the 679 articles originally included, 22 articles were evaluated in detail, and finally 7 articles were excluded because they did not meet the inclusion and exclusion criteria. In the end, a total of 15 RCTs were included ( Figure 1). Among the exclusion articles, 4 of them were Curcuma longa Extracts and curcumin combined with other unconventional therapies [23][24][25][26], while 2 of them were not RCTs [27][28], 1 of them was not original article [29]. The basic characteristics of each RCTs are shown in Table 1.

Description of included trials
Most of the 15 RCTs are from different countries, among which Haroyan et al. 2018 [31] comes from Armenia, Wang et al. 2020 [16] comes from Australia, Henrotin et al. 2019 [18] comes from Belgium, Kertia et al. 2012 [32] comes from Indonesia, and Nakagawa et al. 2014 [36] comes from Japan. In addition, 3 RCTs come from India [20,[38][39], 3 RCTs come from Thailand [33,35,37], and 4 RCTs come from Iran [17,19,30,34]. These 15 RCTs involve a total of 1621 participants, and the scale of each RCT is 40-400 participants. There are 2 experimental groups in Henrotin et al. 2019 (high-dose group and low-dose group), so the control group is divided into 2 small subgroups accordingly, matching the high-dose group and the low-dose group respectively (Henrotin et Table 1.

Risk of bias of included studies
The summary and graph of risk of bias ware shown in figure 2 and 3.

Random Sequence generation
Three RCTs [32,[36][37] did not describe the random sequence generation methods, hence they were rated as unclear risk of bias. Other RCTs described the sequence generation methods, and were rated as low risk of bias. Among those RCTs, Panahi [18] utilized blocking randomization, Hashemzadeh et al. 2020 [19] utilized random number table, and Jamali et al. 2020 [17] utilized the online block randomization program. The other 7 RCTs used computer software to generate random sequences.

Allocation concealment
Two RCTs [30,32] did not describe whether allocation concealment was performed, and therefore were assessed as unclear risk of bias. The remaining RCTs use similar-looking drug packaging, or only allow pharmacists to see the random number, or package the random number in a similar-looking opaque box, or use computer-generated random sequences that cannot be guessed by researchers and participants, so they were considered to be a low risk of bias.

Blinding
Kuptniratsaikul et al. 2009 [33] only described the blinding for outcome assessment, but failed to describe the blinding for participants and its outcomes are subjective indicators (VAS), hence it was rated as having low risk of bias in blinding of outcome assessment and having high risk of bias in bling of participants and personnel. Madhu et al. 2013 [38] only used blinding to the participants, not blinding to the measurer, hence it was rated as having low risk of bias in bling of participants and personnel and having high risk of bias in blinding of outcome assessment. Although the four RCTs claimed to use blinding, they did not describe the process of blinding implementation in the paper, so they were assessed as unclear risk of bias. Panahi et al. 2016 [30] and Kertia et al. 2012 [32] did not state whether blinding was used, but because its outcomes are objective indicators (such as COX-2, SOD, MDA), which is less affected by blinding; hence, we assessed the risk of bias as low. Shep et al.
2019 [20] did not mention whether to use blinding, and its main outcome indicators are subjective evaluation indicators, hence it was rated as high risk of bias. Other RCTs describe the process of blind implementation and are therefore judged as low risk of bias.

Incomplete outcome data and Selective reporting
Although all RCTs exist and participants fall off, because the reasons for falling out and the number of people were balanced, they were considered to be low risk of bias. No selective reports were found in all RCTs, so they were considered low risk of bias.

Other potential bias
Other sources of bias were not observed in 15 RCTs; therefore, the risks of other bias of the RCTs were low.

Pain
The improvement of pain is represented by the results of VAS and WOMAC score-pain.

Function
The improvement of function is represented by the results of WOMAC score-function. 6 RCTs reported WOMAC score-function of patients [16-18, 20, 33-34, 38-39] not applicable], so the random effects model was used for meta-analysis. The meta-analysis results of each subgroup showed that: (1) Compared with placebo, Curcuma longa Extract and curcumin can reduce WOMAC score-function (SMD -0.79, 95%CI -1.27 to -0.31, P=0.001; random effect model).

Secondary outcomes
The results of KOOS score and MDA were shown in Table 2. Only Panahi et al. 2016 [30] reported the improvement of SOD and GSH. This RCT found that compared with placebo, the serum SOD activities in curcuminoids group was higher (P<0.001). However, the difference of GSH level between curcuminoids group and placebo group was of no statistical significance (P=0.064).
Only Kertia et al. 2012 [32] reported the improvement of COX-2. This RCT found that the difference in COX-2 between diclofenac sodium group and curcuminoid group was of no statistical

Adverse events
Ten (10)  showed the adverse events between in experimental group and control group was of no statistical significance (RR 0.77, 95%CI 0.56 to 1.05, P=0.10; random effect model). (Figure 8). The publication bias detection suggests that there may be no publication bias (P=0.372) ( Figure 9E). score-function; D: WOMAC score-stiffness; E: adverse events)

Impact of Time of treatment
In order to explore the influence of the duration of the intervention on the primary outcomes, we conducted a subgroup analysis of the main results according to the duration of the intervention (Table   3). (1)  week, but there was no difference in the results at other time points.  for NSAIDs [40][41]. A large number of pharmacological studies have also revealed that curcumin has the potential to become a clinical treatment for OA [41][42][43]. For example, curcumin inhibits inflammation by blocking inflammatory factor-mediated NF-κB, NLRP3 and other signaling pathways, and inhibits oxidation by removing free radicals and enhancing antioxidant enzyme activity, thereby protecting cartilage from damage [44][45][46]. Curcumin can also promote cartilage matrix repair by adjusting the levels of proteins such as synthin, inhibit chondrocyte apoptosis by promoting autophagy and increasing the activity of anti-apoptotic proteins, and affect chondrocyte proliferation by regulating the Wnt signaling pathway [41][42][43].
In this systematic review and meta-analysis, we found that: (1) compared with placebo, Curcuma longa Extract and curcumin can relieve pain (decrease the VAS and WOMAC score-pain), improve This may indicate that the 6-week intervention is the time point with the least adverse events, or it may be caused by differences in race, administration methods, and pharmaceutical preparations. In the future, it is still necessary to report more outcomes data at different time points of Curcuma longa Extract and curcumin's intervention to correct or confirm this result. Current research reports also show that curcumin can inhibit the inflammatory response upstream phospholipase A2 (phospholipase A2, In addition, we can pay more attention to the role of Curcuma longa in OA in the future. Curcuma longa contains more phenolic pigments (including curcumin, demethoxycurcumin, bisdemethoxycurcumin) and essential oils (including cineole, linalool, α-terpinene, caryophyllene, arcurcumene, zingiberen, curcumol, dl-turmerone, arturmerone, dehydrocurdione); it also contains campesterol, stigmasterol, β-sitosterol, cholesterol, fatty acids and metal elements potassium, sodium, magnesium, calcium, manganese, iron, copper, zinc and other multi-component botanicals [52][53][54][55].
Compared with curcumin monomer, because Curcuma longa extract contains more components, it may play a multi-target and multi-signal pathway transduction role in the treatment of OA pain and inflammation in the molecular pathology mechanism [56][57]. Meanwhile, Curcuma longa is a multicomponent botanical drug, and the synergy between its components may bring potential clinical effects in the treatment of OA [58][59]. These components may increase the concentration of each other in the blood of patients with OA through pharmacokinetics and increase the time of each other's stay in the body, thereby exerting a better clinical effect. Current research showed that the bioavailability of curcumin compound monomers is low [60][61]. However, through the combination with piperine and other substances, the blood concentration of curcumin increased, the elimination half-life was prolonged, the metabolic clearance rate was reduced, and the bioavailability was improved [62][63][64].
Curcuma longa is a multi-component botanical drug, and the synergy between its different components may also reduce potential side effects. Recent studies have shown that Curcuma longa is generally well tolerated even in large doses, although there are still some gastrointestinal side effects, such as nausea and diarrhea, and allergic reactions [65]. Recent studies have also shown the clinical efficacy of Curcuma longa extract in OA [16,18]. In the future, the synergistic relationship between the multiple components of Curcuma longa can be further explored.
To promote the conclusion, the GRADE tool was utilized to rate the quality of the evidence [66].
According to the GRADE handbook [53], the evidence was judged to be high to moderate ( Table 4).
The quality of WOMAC score-pain and WOMAC score-stiffness was high; the quality of VAS, WOMAC score-function, adverse events was moderate (Table 4).  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.

Very low quality:
We are very uncertain about the estimate. 1 Downgraded one level due to the probably substantial heterogeneity Some of our results agree with the meta-analysis of Bannuru et al. For example, we have found that Curcuma longa Extract and curcumin can improve pain, function and stiffness compared with placebo. We also found that there is no difference between Curcuma longa Extract and NSAIDs in improving pain, function and stiffness. In terms of adverse events, we all found that Curcuma longa Extract is as safe as placebo and safer than NSAIDs. However, our study included the RCT of Curcuma longa Extract and curcumin combined with NSAIDs, and showed that this combination is more effective than NSAIDs alone, and the addition of Curcumin does not increase the occurrence of adverse events. Our study also evaluated the effects of Curcuma longa Extract and curcumin in combination with Glucosamine, and found that the pain improvement and the incidence of adverse events in the Curcuma longa Extract and curcumin+Glucosamine group were similar to those in the Glucosamine group. However, because there are too few RCTs related to Curcuma longa Extract and curcumin+Glucosamine and Curcuma longa Extract and curcumin+NSAIDs, it is not enough to draw a very positive conclusion. In the future, more related RCTs are needed to verify or modify this results.
Our meta-analysis also showed that Curcuma longa Extract and curcumin can improve oxidative stress in patients with OA. Compared with previous meta-analysis, our risk of bias assessment results are different, but we list the reasons for the assessment in detail. And our GRADE score shows that the level of evidence is higher, possibly because our assessment of the risk of bias is lower, and the heterogeneity of RCTs is lower. Our meta-analysis also shows that Curcuma longa Extract and curcumin may need to be administered for at least 12 weeks to obtain the therapeutic effect. In addition, the RCTs we included are more novel, which increases the reliability of the conclusions. Our metaanalysis shows that the combination of Curcumin and NSAIDs does not increase the occurrence of adverse events and has better efficacy. This is a promising result, because adding Curcumin supplementation in the case of using NSAIDs may increase the efficacy and perhaps reduce the dosage of NSAIDs. This is a direction that can be studied in the future.
In view of the broad prospects of the current development and application of curcumin or Curcuma longa Extract in the treatment of OA, it is recommended that future RCT research can be indepth from the following aspects: (1) Explore the effects of different administration routes of Curcuma longa Extract and curcumin (such as oral, topical percutaneous application, joint cavity injection, etc.) on its curative effect, and find the best administration method, concentration and dosage of curcumin in the treatment of OA. (2) The role of Curcuma longa Extract and curcumin combined with other active ingredients (such as quercetin, etc.) in the treatment of OA. (3) Report outcomes at different intervention time points. In addition, due to the difference in the incidence of OA between male and female [68], we look forward to future RCTs to analyze the efficacy and safety of different genders, so as to provide more detailed guidance on the medication of patients of different genders.

Conclusion
This systematic review and meta-analysis show that Curcuma longa Extract and curcumin can relieve pain and joint stiffness in patients with OA, improve joint function, and would not increase the occurrence of adverse events. Based on current evidence, it is recommended to use Curcuma longa Extract and curcumin supplement for OA patients for more than 12 weeks. Future RCTs can focus on the different usage and dosage of Curcuma longa Extract and curcumin, and the curative effect of combination with other drugs.

Data Availability Statement
The data that support the findings of this study are openly available in supplementary materials.

Conflict of Interest
We declare no competing interests.

Author contributions
Liuting Zeng, Ganpeng Yu, Hua Chen are responsible for the study concept and design. Liuting Zeng, Kailin Yang, Wensa Hao, Ganpeng Yu, Hua Chen are responsible for the data collection, data analysis and interpretation; Liuting Zeng and Kailin Yang drafted the paper; Hua Chen and Ganpeng Yu supervised the study; all authors participated in the analysis and interpretation of data and approved the final paper.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Rationale
3 Describe the rationale for the review in the context of what is already known.

2-3
Objectives 4 Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, outcomes, and study design (PICOS).

METHODS
Protocol and registration 5 Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide registration information including registration number.

3-5
Eligibility criteria 6 Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale.

3-5
Information sources 7 Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched.

3-5
Search 8 Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated.

3-5
Study selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis).

3-5
Data collection process 10 Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators.

3-5
Data items 11 List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made.

3-5
Risk of bias in individual studies 12 Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis.

3-5
Synthesis of results 14 Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e.g., I 2 ) for each meta-analysis.