1 Introduction

Knee osteoarthritis (OA) is a degenerative joint disease associated with progressive destruction of articular cartilage, inflammation, and diminished remodeling of adjacent bone.^1–3 Given the increasing prevalence of knee OA, novel therapies aimed at improving biomechanical function and patient symptomatology have been devised, including viscosupplementation.⁴ Recognized for its rheological properties, hyaluronic acid (HA) acts to absorb shock and lubricate during joint movement.^1,4,5 In osteoarthritis, HA is clears at higher rates than usual, which reduces synovial fluid viscoelasticity and leads to cartilage loss.^5,6 Intraarticular administration of HA aims to maintain joint lubrication, protect against cartilage erosion, lessen inflammation of synovium, and increase the synovial fluid elasticity.^1,5

The American Academy of Orthopaedic Surgeons (AAOS) 3rd edition evidence-based clinical practice guideline, “Management of Osteoarthritis of the Knee (Non-Arthroplasty),” reports a recommendation strength of moderate against the routine use of HA in the treatment of symptomatic OA of the knee.⁷ This is a downgrade in strength from the 2nd edition released in 2013, which made a strong recommendation against the use of HA for patients with symptomatic OA of the knee.⁸ The updated recommendation recognizes the potential for HA to benefit patients but is limited by inconsistency in the available evidence, highlighting the need for physicians to critically evaluate the published literature when trying to determine the true clinical efficacy of this treatment modality.⁷

An important consideration when critically reviewing published studies is the potential for bias, including author conflicts of interest (COI). COI refers to scenarios where a physician's secondary interest may compromise the integrity of the physician-patient relationship.^9,10 Although potential COI are sometimes unavoidable, there is a societal expectation that providers are free from of conflicted interests throughout medical decision making.¹⁰ Thus, transparency to the public is expected to maintain an impartial arrangement, and concerns may arise when industry-sponsored medical devices dictate clinical practice.^9,10 Both industry participation in scientific research and payments to surgery practices have been implicated with publication of positive outcomes and underreporting of negative findings.^9,10 Accordingly, the Physicians Payments Sunshine Act (PPSA) requires medical industries to disclose any transfers of value to physicians or hospitals to the Centers for Medicare and Medicaid Services (CMS).¹⁰ In addition, the CMS's Open Payments program encourages transparency by maintaining a public database of payments that reporting entities make to covered recipients.

Given the literature supporting the advantages of delaying total knee arthroplasty (TKA), conservative treatment modalities for OA, including viscosupplementation, have correspondingly been on the rise.^3,11,12 Evidence of inconclusive study results and publication bias in these studies has led to concern that financial COI have influenced viscosupplementation outcomes.¹¹ Previous investigations assessing the role of author COIs/industry funding as sources of bias in studies on viscosupplementation for knee OA have come to differing conclusions, with one study concluding that industry funding did not consistently affect the estimates of viscosupplementation efficacy, and another reporting an observed significant association between author COIs and qualitative study conclusions.^13,14 These investigations, however, were limited by incomplete COI/funding information and included studies only as recent as the year 2012, and thus the conclusions may not translate to the more recent body of published literature. The purpose of the present study was systematically review and assess bibliometric data, financial COIs, and overall outcomes of contemporary viscosupplementation studies. We hypothesize that manuscripts favorably detailing viscosupplementation with HA are more likely to have COI.

2 Materials and methods

2.2

2.2 Quality of studies

Articles were screened in 3 rounds using the web-based version of Rayyan Intelligent Systematic Review (Rayyan Systems Inc., Qatar Computing Research Institute, Doha, Qatar). Articles were first screened by title, by abstract, and then by full-text, respectively. A total of 3 independent reviewers completed the screening process, and a 2 out of 3 majority was required for study inclusion during each stage. Articles were included only if they were written in English and were available in their full-text form. All included manuscripts reported resulted that included functional outcomes utilizing a validated scoring tool. It was also required that the analysis included a comparison between a control and viscosupplementation. Also excluded were biomechanical and cadaveric studies. Any clinical studies containing treatment arms with fewer than 20 patients each, as well as commentaries, editorials, case reports, systematic reviews, and meta-analyses were excluded. Articles were excluded if disclosures or conflicts of interest were not reported within the manuscript. This review process was based upon and adapted from similar studies that have been successfully published regarding conflicts of interest.^9,15,16

A PRISMA flow diagram depicting the manuscript screening process is presented in [Fig. 1].

Close

Fig. 1 PRISMA flow diagram for viscosupplementation in primary knee OA.

Export to PPT

3 Results

After excluding studies lacking a COI statement or for which the level of evidence was <3, 67 studies were retained for analysis. A full list of the included studies, along with payment information is provided in [APPENDIX 1]. A large majority of studies presented Level I evidence (53, 79.1%). Over half were published in open access journals (42, 62.7%), with mean IF of 2.34 ± 1.65 and mean RCR of 3.77 ± 3.94. The large majority of studies reported favorable outcomes (48, 71.6%), with substantially fewer reporting equivocal (15, 22.4%) or negative (4, 6.0%) outcomes. Within this sample, 21 studies (31.3%) reported that at least one author had a COI, and all of those studies also disclosed industry funding. None of the remaining studies without a disclosed COI had any industry funding.

There were no relationships between reported COI and the outcomes of studies (Χ² = 0.31, P = 0.577), levels of evidence (Χ² = 3.48, P = 0.176), or RCR (S = 743, P = 0.591). However, studies reporting COIs/industry funding tended to be published in significantly higher impact journals (reporting COI: IF = 3.5 ± 2.0; no COI: IF = 1.8 ± 1.1; S = 950, P < 0.001). Study outcomes were not related to the probability of being published in an open access journal (Χ² = 0.01, P = 0.960), nor to level of evidence (Χ² = 2.67, P = 0.263), RCR (S = 618, P = 0.835), or IF (S = 563, P = 0.655). Level of evidence and the probability of being published in an open access journal were also unrelated (Χ² = 0.51, P = 0.776).

4 Discussion

The current study reviewed the contemporary literature regarding viscosupplementation for the treatment of knee OA and sought to determine whether author COIs and/or industry funding affect study conclusions, level of evidence, and scientific impact (as measured by RCR and journal IF). Due to an observed lack of consistency in the available evidence regarding the efficacy of viscosupplementation for knee OA, several authors have previously investigated the published literature for factors to explain the discrepant study results, including potential sources of bias.^4,11,18,19 One such source which has remained incompletely studied is potential investigator bias related to COIs and/or industry funding.

We hypothesized that “conflicted” studies (i.e., those in which authors disclosed COI and/or industry funding) would be more likely to report favorable conclusions than “nonconflicted” studies. This is due to the fact that author financial COIs and industry funding are more frequently associated with statistically significant results and favorable study outcomes.^20,21 We found, on the other hand, that neither author COIs nor industry funding significantly affected study outcomes and conclusions. Additionally, there was no relationship between author COIs/industry funding and the level of evidence or RCR of the included studies.

It was, however, found with statistical significance that “conflicted” studies tended to be published in journals with significantly higher IFs than “nonconflicted” studies (IF = 3.5 vs. IF = 1.8). The same finding was also observed in the earlier Printz et al. study.¹³ The exact reason(s) for these findings remains unclear. Such a finding could indicate that these conflicted studies have the potential for being read by a wider audience than studies published in lower impact journals. In turn, the conflicted articles could become more impactful on general practice through increased reader attention and future citations in other impactful publications. Despite the significance of IF, neither level of evidence nor RCR between conflicted and nonconflicted studies demonstrated statistical significance, suggesting that the scientific quality/impact of contemporary viscosupplementation studies is not substantially influenced by COI or funding status.

Intraarticular HA remains controversial in the treatment of knee OA, owing in part to conflicting reports in the literature regarding its efficacy; even systematic reviews have produced discordant results, and this has been attributed to a high degree of various forms of bias and/or methodological flaws among published studies.^{4,7,8,11,18,19,22,23} Investigator COIs/industry funding can be one potential source of bias and carries the risk of influencing study design along with conduction, analysis, and reporting of outcomes. However, little is known regarding the specific impact of COIs on the contemporary viscosupplementation literature.

Earlier studies in this area have come to different conclusions.^13,14 Wang et al. performed a meta-analysis of randomized controlled trials on the therapeutic effects of HA on OA of the knee, which included 20 trials published up to the year 2001.¹⁴ The authors concluded that industry funding did not consistently affect the estimates of HA efficacy based on 3 outcome end points (pain with activities, pain without activities, and function). More recently, Printz et al. performed an updated systematic review of prospective, randomized, placebo-controlled studies on the effects of HA injections for knee OA to evaluate whether industry sponsorship or author COIs were associated with study conclusions.¹³ Their investigation included 48 studies published from 1987 up to March/April 2012. The authors were unable to demonstrate any significant differences among qualitative study conclusions (favorable, neutral, unfavorable) between industry- and non-industry-sponsored studies, which was attributed to the small number of studies (only 3/48) which were non-industry-sponsored. However, they observed a significant difference in qualitative conclusions when studies were compared according to author affiliations. The studies in which any author had pharmaceutical company affiliations were significantly more likely to report favorable/neutral conclusions and significantly less likely to report unfavorable conclusions. The authors concluded that qualitative conclusions were associated with a financial COI with the sponsoring pharmaceutical company among viscosupplementation for knee OA studies.

The conflicting conclusions from these two previous investigations may be due to differences among the included studies, as the latter investigation considered studies published up to the year 2012, whereas the former only included studies up to the year 2001. Furthermore, the findings of those studies do not reflect the most recent decade of published literature and may have been limited by incomplete COI/funding information. For example, Printz et al. reported that study sponsorship information was unidentifiable in 31% of studies, which limited the strength of the findings and conclusions of that investigation.

For these reasons, in the current study we sought to review the contemporary literature and to include only those studies published during or after 2011, which coincides with the implementation of Physician Payments Sunshine Act. This ensured that the most accurate and detailed information pertaining to COIs/funding could be accessed. Having complete and accurate COI/funding information represents a strength of the current investigation, but it must be noted that the results of this study apply specifically to the contemporary viscosupplementation literature (i.e., those studies which have been published after the year 2010). The results may not translate to the older literature and it's unclear whether similar findings would have been observed when considering the entire body of viscosupplementation literature, had complete COI/funding been available for all studies.

Overall, only a small percentage of the studies included in this investigation, roughly one-third (31%), were “conflicted.” It should also be noted that all COIs were found to be related to industry funding of the studies. This percentage is lower than that reported in the previously mentioned systematic reviews, which both observed that the majority of studies were industry funded. Wang et al. reported that 60% of studies were industry funded, while Printz et al. reported that figure to be 63%.^13,14 This discrepancy may reflect a decreasing industry influence on the viscosupplementation literature in the most recent decade.

On the other hand, a large majority of studies, nearly three-fourths (72%), reported favorable outcomes, regardless of COI/funding status. Furthermore, no statistical association was found between the presence of COI or industry funding and how often studies described favorable outcomes, with 76% of “conflicted” studies and 70% of “nonconflicted” studies reporting favorable outcomes. These findings contrast with those of Printz et al., who observed a much lower percentage of all studies reporting favorable qualitative outcomes (only 39.5% favorable).¹³ In addition, they found that none (0/17) of the “industry-authored” studies (i.e., those in which any of the authors had pharmaceutical company affiliation), reported unfavorable conclusions, while 35% (11/31) of studies with “academic authorship,” i.e., those in which all authors had academic affiliations, reported unfavorable conclusions.

It was postulated that this difference may have been the result of publication bias, with industry-sponsored trials with unfavorable results less likely to be presented and/or published as a result of influence from sponsoring companies/authors. It is unclear why such a large difference in the percentage of studies reporting favorable outcomes has been observed between our investigation and that of Printz et al. (72% versus 39.5%, respectively). This may be due to differing methodological quality, patient populations, and outcome instruments among the published studies. It could also reflect continued publication bias in the contemporary viscosupplementation literature, with studies demonstrating favorable outcomes being more frequently published than those reporting negative outcomes, regardless of author COIs and/or funding sources.

There are several limitations to this study. As mentioned previously, the findings reflect only the most recent literature, as we relied in part on Open Payments Data to determine COI/funding status. In addition, the accuracy of our analysis depends on the correctness of the reported COI/funding data, and it is possible that inaccurate reporting exists. The authors also acknowledge that there may also be underlying relationships between COI/funding across time with the evolution of the molecules/drugs formulations of various viscosupplements and their respective frequencies of administration. Finally, our assessments regarding the qualitative conclusions of the included studies were based upon the reported outcome measures and statistical analyses from each individual study. We did not attempt to interpret each study's results in the context of clinical significance, and it's possible that some studies reported “favorable” conclusions despite having clinically insignificant results. If the frequency of such reporting was influenced by COI/funding status, this could potentially have had an impact on the results.

Despite these limitations, this study presents a comprehensive review of the contemporary literature regarding viscosupplementation for the treatment of knee OA, offering additional insights into the impact of an author's COIs and industry funding on a study's outcomes and scientific quality. Industry and commercial support is a valuable and often necessary aspect of modern research, and a better understanding of its influence on a particular field can help clinicians and researchers effectively interpret the body of published literature.

5 Conclusions

The results of this review suggest that investigator COIs, and more specifically commercial funding of studies, do not significantly influence the frequency of favorable outcomes or study level of evidence within the contemporary body of published literature on viscosupplementation in treating knee OA. The percentage of recent viscosupplementation studies that were sponsored by industry was lower than previously reported, but the frequency of favorable outcomes was higher than previously reported, regardless of COI/funding status. It is, however, worth noting that studies reporting COIs/industry funding tended to be published in journals with significantly higher impact factors. The authors recommend that physicians carefully assess the published literature for potential COIs and consider the effects of hidden biases when interpreting study results.

Funding/sponsorship

None.

Informed consent

Not applicable.

Institutional ethical committee approval

Institutional Review Board approval was not required as this manuscript did not involve the use of human subjects.

Authors contribution

GGV: Conceptualization, Methodology, Data curation, Writing – Original Draft, Writing – Review & Editing. DAB: Conceptualization, Methodology, Data curation, Writing – Review & Editing. JGL: Conceptualization, Methodology, Writing – Original Draft, Writing – Review & Editing. TCF: Data curation, Writing – Original Draft, Writing – Review & Editing. AWF: Conceptualization, Methodology, Formal analysis, Writing – Original Draft, Writing – Review & Editing. ABK: Conceptualization, Data curation, Supervision, Validation, Writing – Review & Editing.