Quality of Evidence Supporting the Role of Curcuma Longa Extract/Curcumin for the Treatment of Osteoarthritis: An Overview of Systematic Reviews

Background Well known for its good anti-inflammatory effect, curcuma longa extract (CLE)/curcumin (C) has a potential effect on osteoarthritis (OA), and a large number of researchers have completed several systematic reviews/meta-analyses (SRs/MAs) in this research area. However, the methodological and evidentiary quality of these SRs/MAs need to be further evaluated, and whether these findings provide reliable evidence for clinicians remains controversial. Methods Two researchers collected data from seven databases for SRs/MAs that are about randomized controlled trials (RCTs) on CLE/C for OA. Assessment was made for the SRs/MAs included in this article by means of the Assessment System for Evaluating Methodological Quality 2 (AMSTAR-2), the Risk of Bias in Systematic (ROBIS) scale, the list of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), and the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system. Results Nine published SRs/MAs were included in our study. According to the results of the AMSTAR-2 assessment, only one SR/MA was assessed as high quality. According to the ROBIS evaluation results, only 2 SRs/MAs have a low risk of bias. According to the results of the PRISMA checklist assessment, only 2 SRs/MAs studies fully reported the checklist, while other studies had reporting flaws. According to GRADE, a total of 59 effect sizes extracted from the included SRs/MAs were evaluated, among which no effect size was rated as high. Conclusions CLE/C may be an effective and safe complementary treatment for OA. However, further standard SRs/MAs and RCTs are needed to provide an evidence-based medical rationale for this.


Introduction
Osteoarthritis (OA) is a global inflammatory joint disease. It is one of the main causes of joint disability [1]. With the combined effects of aging, rising obesity rate, and the increase in the number of joint injuries, OA has become an increasingly common disease on a worldwide scale, and a global estimate indicates that the affected population has reached 250 million. [2,3]. Pain, joint stiffness or deformity, and even terminal disability are commonly reported as typical symptoms of OA in studies around the world [4,5]. However, most current drug therapies focus only on pain relief and symptomatic treatment, including the use of nonsteroidal anti-inflammatory drugs (NSAIDs), intra-articular injection of glucocorticoids and opioids [6]. However, gastrointestinal discomfort and dose dependence are common problems with these drugs [7,8]. In addition, total knee arthroplasty (TKA) is mainly used to treat severe KOA, which has a high complication rate [9]. erefore, it is necessary to explore a safer treatment for OA. OA used to be considered a degenerative cartilage disease, but now, this concept has been transformed into a complex disease that affects the entire joint [10]. It is now recognized that OA involves mechanical, inflammatory, and metabolic factors, rather than a simple "wear and tear" disease. Inflammation plays a greater role in the pathogenesis of OA than previously recognized, and OA is now regarded as a low-grade inflammatory disease that affects all tissues of the joint, including cartilage degeneration, bone remodeling, osteophytes, and synovitis [11]. In view of the fact that inflammation may play a key role in the pathogenesis and progression of osteoarthritis, it may be a good idea to develop new treatments for OA from this perspective.
Recently, curcuma longa extract (CLE, an anti-inflammatory and antioxidant preparation) has been used in traditional Chinese medicine and Ayurveda to treat arthritis and has thus become an attractive treatment option for improving the joint condition of OA patients [12]. Often used as an alternative medicine or dietary supplement, turmeric is typically an extract that is standardized to 80-95% curcuminoids, which include curcumin (C), demethoxylated curcumin (DMC), and didemethoxylated curcumin (BDMC), among which C [13] is the most active ingredient in turmeric and is "generally regarded as safe" by the US FDA [14]. In addition, the CLE alone has anti-inflammatory properties similar to NSAIDs [15]. It has been shown that CLE affects the signal transduction of proinflammatory cytokines by influencing the activity of NF-κB, such as interleukin, phospholipase A2, 5-lipoxygenase, and COX-2 [16].
Many systematic reviews/meta-analyses (SRs/MAs) have been conducted to evaluate the potential therapeutic benefits of CLE/C for patients with OA. However, the conclusions are inconsistent due to the defects of the quality and methods of the preliminary research studies. e systematic review is a novel tool for solving specific and key issues related to policies and practices [17]. e purpose is to combine the evidence from multiple SRs/MAs to form a practical document that can be used to guide healthcare professionals and decision-makers [18]. e purpose of our research is to use a systematic overview to critically evaluate the scientific quality of related SRs/MAs in the CLE/C treatment of OA.

Protocol Registration.
is overview protocol has been registered with the INPLASY website (Registration number: INPLASY202220063).

Research Methods.
e SR/MA overview is based on the guidelines specified in Cochrane Handbook [19], the Preferred Reporting Project for System Reviews and Meta-Analyses (PRISMA) statement (Supplementary file 1) [20], and the overview of high-quality methods [21,22]. e search strategy adopts a combination of MeSH terms and free words. We searched the above databases with the following key terms: curcuma longa extract, curcumin, osteoarthritis, systematic reviews, and meta-analysis. We also manually searched the references of related articles. e specific search strategy was modified according to different databases. Supplementary file 2 provided the search strategy.

Eligibility Assessment and Data Extraction.
Two researchers (WQ-C and P-D) independently performed literature screening. After deleting duplicate content, researchers read the title and abstract to find potential SRs/ MAs based on the inclusion and exclusion criteria. en, full-text articles were obtained for further screening to determine their eligibility. Afterwards, two researchers (ZG-Y and WB-L) independently extracted data using a standardized data extraction form. e following specific characteristics are extracted from each SR/MA: first author, the year of publication, country, the number of included studies, sample size, treatment intervention, control intervention, quality assessment methods, results, and main conclusions. Evidence-Based Complementary and Alternative Medicine 2.6. SRs/MAs Quality Assessment. Quality assessment of included SRs/MAs was performed independently by two researchers (Q-L and CD-D).

Assessment of Methodological Quality.
e Assessment System for Evaluating Methodological Quality 2 (AMSTAR-2) [23] scale was used to assess the methodological quality of the included SRs/MAs. It consists of 16 items, 7 of which are critical areas (2, 4, 7, 9, 11, 13, and 15). Each item was assessed and rated as "yes," "partially yes," or "no."

Assessment of Risk of Bias.
e risk of bias of the included SRs/MAs was assessed by the risk of bias in systematic (ROBIS) scale [24]. e scale was completed in 3 stages to assess the overall risk of bias. e results are assessed and rated as "low," "unclear," or "high."

Assessment of Reporting Quality.
e list of PRISMA was used to assess the quality of each SR/MA report based on the following aspects: (a) title, (b) summary, (c) introduction, (d) method, (e) result, (f ) discussion, and (g) funding. It consists of 27 items, with a focus on reporting methods and results in a meta-analysis. Based on the completeness of the project information report, each project is assessed and rated as "yes" (full report), "partial yes" (partial report), or "no" (no report).

Assessment of Quality of Evidence.
e GRADE scale was used to assess the quality of the evidence of the included SRs/MAs from five aspects: research limitations, inconsistencies, indirectness, imprecision, and publication bias [25].

Data Synthesis and Presentation.
In this overview, a narrative synthesis was used. e characteristics and results of each SR/MA and the assessment results of AMSTAR 2, PRISMA, ROBIS, and GRADE were reported in the form of a list.

Results on Literature Search and Screening.
A total of 71 articles were retrieved from seven literature databases, and 21 duplicate articles were deleted. We filtered by the title and abstract of the literature and finally obtained 9 literature studies for full-text screening. After evaluation according to the inclusion and exclusion criteria, we finally obtained 9 literature studies [26,27] included for study. Figure 1 shows the screening flow chart. [26,27] published from 2016 to 2021 were included, and all the included papers were SR and MA. Of these published SRs/ MAs, five were from China [27][28][29][30][31], and the remaining four were from the United States [26], the United Kingdom [32], Australia [33], and South Korea, respectively [34]. Among them, 8 SRs/MAs [26,[28][29][30][31][32][33][34] were published in English and one [27] was published in Chinese. All SRs/MAs contained a total of 28 RCTs, and the number of RCTs included in each SR/MA ranged from 5 to 15, and the total number of individuals in the included RCTs for a single SR/MA ranged from 599 to 1,621. e intervention rendered to the treatment group was curcuma longa extract or curcumin, and the control group was treated with CT or placebo, and CT modalities include painkillers and NSAIDs. In terms of quality assessment scales, all the literature adopted the Cochrane risk of bias standard. e details of the SRs/MAs included are shown in Table 1.

Summary of the Results of the Included Studies.
e result indicators extracted from the included studies are listed in Table 2.
(3) Adverse events. Seven SRs/MAs [26][27][28][29][30][31]33] reported the occurrence of adverse events in CLE/C compared with placebo, none of which was statistically significant. Two SRs/ MAs [26,33] reported that, compared with placebo, there was no significant difference in the use of rescue drugs during the treatment with CLE/C. A SR/MA [26] reported that CLE/C was not statistically significant in the incidence of withdrawal from treatment due to adverse events compared with placebo.

Efficacy and Safety of CLE/C for OA (Compared with CT) (Table 2(b))
(1) Pain. Five SRs/MAs [26-28, 31, 33] reported the effects of CLE/C on pain in patients with OA compared with NSAIDs, and 2 SRs/MAs [31,33] indicated that CLE/C had similar effects on pain relief compared with NSAIDs. In addition, a SR/MA [27] showed that NSAIDs were superior to CLE/C in reducing pain in patients with OA.
(3) Adverse events. Six SRs/MAs [26][27][28][30][31][32][33] reported the occurrence of adverse events in CLE/C compared with NSAIDs. Among them, the results of 4 SRs/MAs [26,27,31,33] showed that patients using CLE/C had a lower incidence of adverse events. Two SRs/MAs [26,33] reported no difference between CLE/C and NSAIDs in the use of rescue drugs. A SR/MA [26] reported that CLE/C has a lower incidence of withdrawal due to adverse events than NSAIDs.

Methodological Quality Assessment.
e AMSTAR-2 assessment breakdown for each review is shown in Table 3. Only one SR/MA was of high quality [29]. Since more than one key item was missing in the remaining SRs/MAs, their quality was rated very low. e method restriction came from the following items: item 2 (only 2 SRs/MAs [26,29] had registered in the protocol), item 7 (only 2 SRs/MAs provided a research exclusion list), and item 15 (the 4 SRs/MAs [26,27,30,32] did not conduct publication bias studies or discuss their impact on SR/MA).   [26,29] in the first stage and Domain 1 of the ROBIS evaluation was assessed as low risk. In Domain 2, 6 SRs/MAs [26,27,29,30,32,33] were assessed as low risk. In Domain 3, 8 SRs/MAs [26,[28][29][30][31][32][33][34] were assessed as low risk of bias and only 2 SRs/MAs [28,29] were assessed as low risk of bias in Domain 4. In Phase 3, 7 SRs/MAs [26, 28-31, 33, 34] had a low risk of bias. e evaluation details of the included SRs/MAs on the ROBIS scale are shown in Table 4.

Report Quality.
e results of the PRISMA inventory evaluation are shown in Table 5. Among the 27 items, 24 items had a "yes" response rate of more than 70%, which showed that the report was relatively complete. However, there were some reporting deficiencies in other projects.
Items 5 (protocol and registration) was inadequately reported (the "yes" response rate is less than 50%). Table 6 shows the results of GRADE evaluation including SR/MA-related effect sizes. e 9 SRs/ MAs included 59 effect sizes related to the efficacy and safety of CLE/C for OA. In the evaluation results based on the effect sizes, 19 were rated as medium, 20 low, and 20 very low in terms of the quality of evidence. Publication bias (n � 42) was the most common downgrading factor, followed by imprecision (n � 38), inconsistency (n � 23), risk of bias (n � 22), and imprecision (n � 0).

Discussion
CLE/C may be a complementary treatment for OA, which is a common disease in the elderly. At the same time, more and

Summary of the Main Findings.
is overview incorporated 9 SR/MAs published between 2016 and 2021, 8(8/9, 88.9%) of which were published in the past 5 years, indicating that an increasing attention has been paid to the effectiveness and safety of CLE/C treatment of OA in recent years.
Based on the results of the AMSTAR-2 evaluation in this overview, only one methodological quality of SR/MA was evaluated as high, and the other methodological qualities were very low, and especially in items 2 (protocol registration, 2/9, 22.2%), 7 (exclusion list, 1/9, 11.1%), and 15 (publication bias, 55.6%). Only 2 SRs/MAs were registered for the protocol. Protocol registration is very important when conducting each SR/MA and should be carried out at the time of topic selection, which helps reduce the potential for selective reporting bias and ensures that each SR/MA is conducted in a more accurate manner [35]. Only 2 SRs/MAs provided a complete list of excluded literature, increasing publication bias. Providing a list of excluded literature can provide strong evidence of the accuracy and rigor of the literature screening process. In addition, publication of biased assessments may reduce the veracity of the final results. In this overview, we used the ROBIS scale to evaluate the risk of bias for the included SRs/MAs. Among them, incomplete literature retrieval and insufficient evaluation of publication bias are the main factors leading to a high risk of bias. Similar to the results of the AMSTAR-2 evaluation, the Note. * e 95% confidence interval does not cross the invalid line.   Author, year (country) Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14 Q15 Q16 Quality Raveendhara R. 2018 (USA) [26] Y

Evidence-Based Complementary and Alternative Medicine
Note. Y, yes; PY, partial yes; N, no; VL, very low; L, low; H, high.
8 Evidence-Based Complementary and Alternative Medicine   Note. ① e included studies have a large bias in methodology such as randomization, allocation concealment, and blinding. ② e confidence interval overlaps less, or the I2 value of the combined results was larger. ③ e sample size from the included studies does not meet the optimal sample size or the 95% confidence interval crosses the invalid line. ④ e funnel chart is asymmetry. ⑤Fewer studies were included, and their results were all positive, which may result in a large publication bias. evaluation using the PRISMA checklist found that only 2 SRs/MAs completed the program registration. e quality of the evidence is based on the GRADE system. Among the 59 effect sizes, no one was rated as high in terms of the quality of evidence. Publication bias is the most common downgrading factor, followed by imprecision, inconsistency, risk of bias, and imprecision. Further analysis revealed a lack of publication bias analysis or the presence of publication bias for the effect sizes included in the SRs/MAs included in this overview. In addition, the insufficient number of RCTs included in the publication bias assessment of the relevant effect size is a potential reason for the missing publication bias assessment. When assessing the quality of evidence for a relevant effect size, the insufficient number of study populations included in that effect size is also a significant contributor to the low quality of the final evidence. Descriptive analysis shows that CLE/C is an effective treatment for OA and may be safer than CT. Due to the low methodological and evidentiary quality of the included RCTs, conclusions from the inclusion of SRs/MAs may differ from real-world outcomes and caution should be exercised when recommending CLE/C as a complementary intervention for OA.

Implications for Future
Research. Various aspects of each SR/MA included were assessed using AMSTAR-2, PRISMA, ROBIS, and GRADE with the aim of facilitating future standardization of SRs/MAs. Researchers conducting SRs/MAs should register or publish study protocols in advance to minimize the risk of bias and ensure the credibility of SRs/MAs results and should provide a list of excluded literature with explanations to ensure transparency and minimize publication bias. In future RCTs, it is important to increase the sample size of the study in a reasonable way to increase the credibility of the evidence. In addition, a complete evaluation of publication bias will also increase the accuracy of the metaanalysis results. With the development of evidence-based complementary and alternative medicine, it is hoped that researchers will continue to promote the standardization of related single RCTs in the future. Well-designed and strictly implemented RCTs can minimize or avoid bias.
is is the gold standard for evaluating interventions [36].

Strengths and Limitations.
As far as we know, our study is the first overview of SRs/MAs on the use of CLE/C in the treatment of OA, which can provide a comprehensive evidence reference for clinical practice. In addition, the evaluation process of AMSTAR-2, PRISMA, ROBIS, and GRADE revealed the obvious limitations of SRs/MAs and RCT, which may help guide high-quality research in the future. However, we found that the methodological quality of most of the included SRs/MAs was poor, a limitation that also prevented our study from drawing firm conclusions about the use of CLE/C for OA.

Conclusion
According to the available published evidence, CLE/C may be effective and safe for the treatment of OA. However, due to the generally low quality of methodologies, reports, and evidence in the included SRs/MAs, clinicians should approach this finding with caution in their practice.
Data Availability e datasets analyzed during the current study are available from the corresponding author on reasonable request.

Disclosure
Wenqiang Chen and Hongshuo Shi are the co-first authors.

Conflicts of Interest
e authors declare no conflicts of interest.