A Three-Tier Framework for Orthobiologic Use in Arthrosis and Lumbar Degeneration: Separating What Is Proven Safe and Effective from What Is Merely Proven Safe

 

Márcio Hiroaki Kume¹*, Bianca Furlan², Camila Gobatto Boaventura², Mônica Andréa Probst², Edson Peracchi², Carmen Austrália Paredes Marcondes Ribas³

Citation: Kume MH, Furlan B, Boaventura CG, Probst MA, Peracchi E, et al. A Three-Tier Framework for Orthobiologic Use in Arthrosis and Lumbar Degeneration: Separating What Is Proven Safe and Effective from What Is Merely Proven Safe. J Stem Cell Res. 7(3):1-20.

Received: August 09,2026 | Published: August 27, 2026

Copyright^© 2026 by Kume MH, et al. All rights reserved. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

DOI: https://doi.org/10.52793/JSCR.2026.7(3)-95

Abstract

Background: Orthobiologic injections are offered to patients with arthrosis and degenerative lumbar disease on the strength of a large and reassuring safety literature. Safety and efficacy are, however, different claims resting on different evidence, and the two are routinely conflated in clinical conversation, in marketing and occasionally in review articles. We set out to separate them formally and to determine which product–indication pairs are currently supported by both.

Methods: Working from a three-stream evidence base of 82 randomised controlled trials in arthrosis (460 outcome rows), 24 clinical studies in lumbar degeneration (92 rows) and 63 systematic reviews and network meta-analyses (192 pooled estimates), we defined four pre-specified, operationally explicit criteria: (A) safety established in controlled comparative data on at least 300 exposed patients with no excess of serious adverse events; (B) a pooled effect against a non-inert (active) comparator whose 95% confidence interval excludes the null; (C) that effect surviving every leave-one-out iteration; and (D) at least five controlled comparisons. Tier 1 requires A, B, C and D together; Tier 2 requires A and B; everything else falls to Tier 3. Effects were expressed as Hedges g with negative values favouring the orthobiologic and pooled by DerSimonian–Laird random effects with one pain-preferred estimate per study.

Results: Eight product–indication pairs were assessable. Criterion A was met by 6 of 8, criterion B by 5, criterion D by 4 and criterion C by 0. No pair therefore reached Tier 1. Four pairs reached Tier 2 — PRP in knee osteoarthritis (g = -0.34 (-0.63, -0.05), k = 18), adipose-derived products in knee osteoarthritis (-0.44 (-0.86, -0.01), k = 7), orthobiologic injection in non-knee joints (-0.55 (-0.98, -0.13), k = 7) and the PRP family in lumbar degeneration (-1.77 (-3.31, -0.22), k = 4) — and four fell to Tier 3. Robustness was the universal failure point. The pooled class-level estimate against active comparators was stable (k = 33, g = -0.33 (-0.53, -0.14), p = 0.00093; leave-one-out range -0.38 to -0.28), but every product-specific estimate was fragile. PRP alone versus hyaluronic acid was not significant (k = 7, g = -0.29 (-0.63, +0.06), p = 0.1022), with 6 of 7 single-trial omissions abolishing any signal. Effects were strongly comparator-dependent: pooled g was -0.97 (-1.54, -0.41) against saline or sham (k = 10) versus -0.33 (-0.53, -0.14) against active comparators (k = 33), so 66% of the apparent benefit was not reproduced when the control arm received an injection — closely matching the 63% attributed to contextual effects by an independent analysis [1].

Conclusions: Under a strict product-specific standard the top tier of the orthobiologic evidence base is currently empty. This is not an argument for abandoning these treatments; it is an argument for describing them accurately. Safety is genuinely established for several products, class-level symptomatic benefit is genuinely established, and product-specific superiority over an active injection is not. The framework makes that distinction operational, states for every pair what evidence would promote it, and shows that for most pairs the required trial is small enough to be feasible.

Keywords

Orthobiologics; Platelet-rich plasma; Mesenchymal stromal cells; Evidence tiering; Osteoarthritis; Lumbar degeneration; Comparator dependence; Contextual effects; Informed consent.

Introduction

Two different claims, one sentence

A clinician who tells a patient that an orthobiologic injection is “safe and effective” is making two claims that rest on entirely different bodies of evidence, that were generated by different study designs, and that currently have very different levels of support. The safety claim is supported by large pooled series, by network meta-analyses of adverse events and by regulatory pharmacovigilance; it is, for several products, one of the better-evidenced statements in musculoskeletal medicine. The efficacy claim is supported by a body of randomised trials that is heterogeneous, dominated by inert comparators, rarely blinded to the degree required for a subjective outcome, and almost never powered to detect the modest effect that is actually present. Conflating the two is the single most consequential communication error in this field.

The companion synthesis to this article quantified the aggregate effect of orthobiologic injection across arthrosis and lumbar degeneration. It found a moderate, durable, statistically robust class-level benefit in arthrosis — g = -0.47 (-0.66 to -0.27) at 6 months across 43 controlled comparisons — that did not decay by 12 months (-0.53 (-0.78 to -0.27)), was largest against inert controls, was not demonstrably superior to corticosteroid, and was unaccompanied by any structural benefit on imaging [2,3,4]. That result is reassuring at the level of the class and unhelpful at the level of the consultation, because no patient is offered “a class”. They are offered a specific product, prepared in a specific way, for a specific joint or spinal level.

Why an explicit tier is necessary

Existing evidence appraisals in this field mostly grade the certainty of a body of evidence, in the manner of GRADE, or issue directional recommendations, in the manner of clinical practice guidelines. Both are valuable and neither answers the question a clinician actually faces, which is closer to: for this product, in this indication, do I have evidence that it works better than the injection I would otherwise give, and separately, do I have evidence that it will not harm the patient? Guideline positions diverge sharply on precisely this point. The American Academy of Orthopaedic Surgeons rates PRP in knee osteoarthritis as limited-strength evidence [5], several interventional-pain societies grade selected biologic applications as Level II to III with moderate recommendation strength [6,7], and a Cochrane review of cell therapy for knee osteoarthritis concluded that benefits were uncertain and that the certainty of the evidence was low to very low [8]. These are not contradictory readings of the same data so much as different questions asked of it.

We therefore constructed a framework with a deliberately narrow purpose: to sort product–indication pairs into three tiers using criteria that are pre-specified, arithmetically checkable, and phrased so that every pair carries an explicit statement of what evidence would move it up. A tier that cannot be promoted is a verdict; a tier that states its own promotion condition is a research agenda.

Objectives

• To define four operationally explicit criteria that separate established safety from established comparative efficacy.
• To apply them to every product–indication pair in the arthrosis and lumbar evidence base for which a pooled controlled estimate can be computed.
• To quantify how much of the apparent benefit of orthobiologic injection is contingent on the choice of control arm.
• To distinguish class-level conclusions, which are robust, from product-level conclusions, which are not.
• To state, for each pair, the specific trial that would change its tier, together with the sample size that trial would require.

 

Methods

Evidence base

Three streams were used. Stream 1 comprised arm-level continuous outcomes extracted from 82 randomised controlled trials of orthobiologic injection in arthrosis, yielding 460 outcome rows across the knee (69 trials), hip (6) and other joints. Stream 2 comprised 92 outcome rows from 24 clinical studies of orthobiologics in lumbar degeneration, of which 10 were randomised controlled trials. Stream 3 was an umbrella layer of 63 systematic reviews and network meta-analyses contributing 192 pooled estimates, used for safety accounting and for triangulation rather than for pooling. Every extracted value was bound to the URL of the document from which it was read.

Effect measure and pooling

Continuous pain and function outcomes were converted to Hedges g with the sign oriented so that negative values favour the orthobiologic arm. Where a study reported several eligible outcomes at the same timepoint, one estimate per study was retained, preferring pain over function and, within pain, the most frequently reported instrument. Pooling used DerSimonian–Laird random effects with Q, τ² and I² reported; 95% confidence intervals are Wald intervals on the pooled estimate. The 6-month timepoint was pre-specified as the primary anchor for tiering because it is the most densely reported and because it post-dates the period in which a procedural placebo effect is at its largest.

Comparator classification

Every controlled comparison was classified as inert or active. Inert comparators were saline injection, dry needling and sham procedure. Active comparators were hyaluronic acid, corticosteroid, structured exercise or conservative care delivered as the trial's control arm, and any other injectable given with therapeutic intent. This distinction is the analytical hinge of the framework: an effect measured against saline conflates the biological action of the product with the procedural and contextual effects of receiving an injection, whereas an effect measured against an active injection isolates the increment attributable to the product itself.

The four criteria

Criteria were specified before the tier assignments were computed and each was given a threshold that can be checked arithmetically from the data tables. They are set out in Table 2 and summarized here.

• Criterion A: safety established. Controlled comparative safety data covering at least 300 patients exposed to the product in the relevant indication, with no excess of serious adverse events relative to the comparator. Uncontrolled case series were not counted, because they cannot establish the absence of an excess.
• Criterion B: efficacy against a non-inert comparator. A pooled Hedges g at 6 months against active comparators whose 95% confidence interval excludes zero.
• Criterion C: robustness. That pooled estimate retains statistical significance in every leave-one-out iteration. A result that depends on the presence of any single trial is not a stable basis for a routine-care recommendation.
• Criterion D: volume. At least five controlled comparisons contribute to the pooled estimate, so that heterogeneity can be estimated with some stability and a single small trial cannot dominate the weight.

 

Tier 1 requires all four criteria. Tier 2 requires A and B, that is, established safety together with a pooled effect against an active comparator, whether or not that effect is robust or well replicated. Everything else, including products with excellent safety data but no demonstrated incremental efficacy, falls to Tier 3.

Sensitivity, fragility and promotion arithmetic

Three additional analyses support the tiering. First, a fragility analysis recorded, for each pooled estimate, the leave-one-out range and the identity of every trial whose removal abolished statistical significance. Second, a blinding-restricted sensitivity analysis repeated the active-comparator pools using double-blind trials only. Third, a promotion analysis computed, for each product, the total sample size that a single new two-arm trial would require to detect the currently observed active-comparator effect with 80% power at a two-sided α of 0.05, together with the sample size required to detect a minimum clinically important standardised effect of 0.30 irrespective of the observed point estimate.

Reporting and reproducibility

The analysis was performed in Python. All pooled estimates, subgroup pools, leave-one-out sequences and power calculations reported here were computed from the extracted data rather than transcribed from published reviews; where a published review is cited for a number, that number is attributed to the review and not presented as an independent result. Figures were generated directly from the same computed objects, so that no figure can disagree with the text.

Tier	Definition	What the clinician can honestly say	How the treatment should be offered
Tier 1	Proven safe and proven effective. Criteria A, B, C and D all satisfied.	This product has been shown to relieve symptoms better than the injection you would otherwise be offered, and its safety is established.	As routine care, within the indication studied, with standard consent.
Tier 2	Proven safe, efficacy partly established. Criteria A and B satisfied; C or D not satisfied.	This product is well tolerated and, taken across the published trials, performs somewhat better than an active comparator — but that finding is not yet stable.	After explicit consent naming the uncertainty; dose recorded; patient entered into a registry where one exists.
Tier 3	Proven safe only, or not characterised. Criterion A or B not satisfied.	This product appears to be well tolerated, but it has not been shown to work better than the alternatives — or it has not been tested adequately at all.	Inside a trial or a registry protocol. It should not be presented as an established treatment.

Table 8: Allocation of randomised evidence at 6 months. Counts are controlled comparisons and randomised patients contributing to the 6-month pooled analyses.

“Actively controlled” is the percentage of each family's randomised patients whose control arm received an active intervention rather than saline or sham.

Provenance of the pooled estimates

(Table 9) lists every controlled comparison contributing to the active-comparator pools that underpin criteria B, C and D, so that any reader can reconstruct a pool, repeat a leave-one-out iteration or identify the trial responsible for a fragile result without returning to the parent synthesis.

Product–indication pair	k	Pooled g (95% CI)	Contributing controlled comparisons
Platelet-rich plasma, knee osteoarthritis	18	-0.34 (-0.63, -0.05)	Abbadi 2022 [24], Elksnins-Finogejevs 2020 [25], Du 2020 [18], Patel 2024 [26], Rahimzadeh 2018 [31], Nouri 2022 [19], Kraeutler 2021 [20], Sun 2021 [32], Jubert 2017 [33], Raeissadat 2020 [21], Filardo 2012 [22], Zhou 2023 [34], Romandini 2024 [35], Baria 2022 [36], Pretorius 2022 [37], Arliani 2021 [23], Huang 2022 [38], Karaborklu Argut 2024 [39]
Adipose-derived products, knee osteoarthritis	7	-0.44 (-0.86, -0.01)	Lu 2025 [27], Jing 2026 [28], Lu 2019 [29], Wu 2023 [30], Baria 2024 [40], Molnar 2025 [41], Gobbi 2022 [42]
Bone marrow products, arthrosis	5	-0.07 (-0.49, +0.35)	Dulic 2021 [43], Dwyer 2021 [44], Lamo Espinosa 2020 [45], Lana 2026 [46], Anz 2020 [47]
Hypertonic dextrose, knee osteoarthritis	1	-0.57 (-1.16, +0.02)	Gul 2020 [48]
Umbilical cord or amniotic products, knee osteoarthritis	2	-0.77 (-2.81, +1.27)	Matas 2019 [49], Pill 2025 [50]
PRP family, lumbar degeneration	4	-1.77 (-3.31, -0.22)	Singh 2023, Wongjarupong 2023, Wang 2025, Won 2022

 No structural benefit for any product

None of the tier assignments incorporates a structural claim, because no product supports one. A meta-analysis of randomised trials with imaging endpoints reported a standardised mean difference of −0.01 (p = 0.91) for cartilage outcomes [2]; a second found g = 0.079 (p = 0.723) [3]; a third, examining WORMS scores across 16 studies and 875 patients, found no difference between orthobiologic and control [4]. Any consent conversation that describes an orthobiologic injection as regenerating, repairing or rebuilding cartilage is unsupported by the randomised imaging literature, regardless of tier.

Implications for consent and for practice

The framework translates into three concrete changes to the way these treatments should be discussed. For Tier 2 products, the phrase “safe and effective” should be replaced by something closer to: this is well tolerated; averaged across the trials it works somewhat better than the alternative injection; that finding is not yet stable enough for me to promise you it will hold. For Tier 3 products with good safety data and null efficacy data — bone marrow products in arthrosis being the clearest case — the honest statement is that the product is safe and has not been shown to work better than the alternatives. For Tier 3 products with no controlled data at all, including exosomes, the honest statement is that we do not know, and the appropriate setting is a trial.

Guideline bodies have converged on positions broadly compatible with this reading, though they express it differently: limited-strength evidence for PRP in knee osteoarthritis from the orthopaedic surgery side [5], low to very low certainty for cell therapy from Cochrane [8], and graded but qualified support for selected biologic applications from interventional pain medicine [6,7,51]. The tier framework does not overturn any of these; it makes the reason for their divergence visible, which is that they are weighing safety and comparative efficacy in different proportions.

A research agenda that follows from the tiers

Because each tier carries its promotion condition, the research agenda writes itself and is notable mostly for how modest it is.

• A double-blind trial of PRP against hyaluronic acid in knee osteoarthritis, 272 patients, standardised preparation, 12-month primary endpoint. This single trial would either promote PRP to Tier 1 or demote it to Tier 3.
• A double-blind trial of an adipose-derived product against an active comparator, 166 patients, with the preparation characterised well enough to be reproduced.
• A trial of hypertonic dextrose against an active comparator, 96 patients — the cheapest question in the field and the one closest to being answered.
• No further uncontrolled case series of any product, and no further saline-controlled trials of products already shown to beat saline; both add exposure without adding information relevant to any criterion.
• Controlled human safety data for exosome preparations before any efficacy trial, and mandatory registry entry for every culture-expanded cell product administered outside a trial.

Figure 8: Decision algorithm for applying the framework to a product–indication pair. The algorithm is deliberately sequential: safety is assessed first because a product that fails criterion A cannot be promoted by any amount of efficacy data.

Limitations 

The thresholds are defensible but arbitrary at the margin; 300 exposed patients, five comparisons and complete leave-one-out stability are conventions rather than derivations, and reasonable analysts would set them slightly differently. Heterogeneity is extreme throughout (I² 91–88% for the arthrosis pools), so every pooled point estimate should be read as a summary of a distribution rather than as an estimate of a common effect; the prediction interval for the primary arthrosis pool spans -1.64 to +0.71. Products were grouped into families, which conceals real variation in preparation — platelet concentration, leukocyte content, activation method, cell dose and passage number all differ across trials nominally testing the same thing. The lumbar evidence base is thin enough (8 controlled 6-month comparisons, one at 12 months) that its tier assignments should be regarded as provisional. Our contextual-effect computation assumes that inert-controlled and active-controlled trials are otherwise comparable, which is not guaranteed. Finally, this is a secondary analysis of a previously assembled evidence base: it inherits that base's search date, inclusion decisions and extraction errors, and it was not separately registered.

Conclusions

Applying four pre-specified criteria to eight product–indication pairs in arthrosis and lumbar degeneration leaves the top tier empty. Six pairs have established safety, five have a pooled effect against an active comparator, and none has an effect that survives every leave-one-out iteration. Four pairs — PRP and adipose-derived products in knee osteoarthritis, orthobiologic injection in non-knee joints, and the PRP family in lumbar degeneration — are proven safe with partly established efficacy. Four, including the entire bone marrow family and all exosome preparations, are proven safe only or not characterised at all.

Two quantitative findings should shape practice immediately. 66% of the benefit seen against saline is not reproduced against an active injection, a figure that matches an independently derived contextual-effect estimate of 63% [1]. And the class-level conclusion is robust while every product-level conclusion is fragile, which means the honest unit of claim is “injecting a biologic helps a little” rather than “this biologic helps”. The remedy is not more patients treated; it is three modest, double-blind, actively controlled trials totalling fewer than 550 participants. Until they are done, the accurate description of most orthobiologic practice is that it is proven safe and not yet proven better.

Declarations

Not applicable. This study is a secondary analysis of published aggregate data and involved no new human or animal participants.

Consent for publication

Not applicable.

Availability of data and materials

All extracted data tables, the analysis scripts and the figure-generation code are available from the corresponding author on reasonable request. Every extracted value is bound to the URL of its source document.

Competing interests

The authors declare that they have no competing interests. [To be completed by each author before submission.

Funding

No funding.

Authors' contributions

[MHK, CAPMR] conceived and designed the study. All authors read, revised and approved the final manuscript.

References

1. Yin F, Wu H, Tong D, Luo G, Deng Z, et al. (2025) Contextual effects of mesenchymal stem cell injections for knee osteoarthritis: systematic review and meta-analysis of randomized controlled trials. Front Med (Lausanne). 12:1636181.
2. Sax OC, Chen Z, Mont MA, Delanois RE. (2022) The Efficacy of Platelet-Rich Plasma for the Treatment of Knee Osteoarthritis Symptoms and Structural Changes: A Systematic Review and Meta-Analysis. J Arthroplasty. 37(11):2282-2290.e2.
3. Prodromidis AD, Charalambous CP, Moran E, Venkatesh R, Pandit H. (2022) The role of Platelet-Rich Plasma (PRP) intraarticular injections in restoring articular cartilage of osteoarthritic knees. A systematic review and meta-analysis. Osteoarthr Cartil Open. 4(4):100318.
4. Xie RH, Gong SG, Song J, Wu PP, Hu WL. (2024) Effect of mesenchymal stromal cells transplantation on the outcomes of patients with knee osteoarthritis: A systematic review and meta-analysis. J Orthop Res. 42(4):753-68.
5. Brophy RH, Fillingham YA. (2022) AAOS Clinical Practice Guideline Summary: Management of Osteoarthritis of the Knee (Nonarthroplasty), Third Edition. J Am Acad Orthop Surg. 30(9):e721-29.
6. Navani A, Manchikanti L, Albers SL, Latchaw RE, Sanapati J, et al. (2019) Responsible, Safe, and Effective Use of Biologics in the Management of Low Back Pain: American Society of Interventional Pain Physicians (ASIPP) Guidelines. Pain Physician. 22(1S):S1-S74.
7. Manchikanti L, Knezevic E, Knezevic NN, Kaye AD, Atluri S, et al. (2024) Effectiveness of Intradiscal Regenerative Medicine Therapies for Long-Term Relief of Chronic Low Back Pain: A Systematic Review and Meta-Analysis. Pain Physician. 27(10):E995-32.
8. Whittle SL, Johnston RV, McDonald S, Worthley D, Campbell TM, et al. (2025) Stem cell injections for osteoarthritis of the knee. Cochrane Database Syst Rev. 4:CD013342.
9. Pers YM, Soler-Rich R, Vadalà G, Ferreira R, Duflos C, et al. (2024) Allogenic bone marrow-derived mesenchymal stromal cell-based therapy for patients with chronic low back pain: a prospective, multicentre, randomised placebo controlled trial (RESPINE study). Ann Rheum Dis. 83(11):1572-83.
10. Du D, Liang Y. (2025) A meta-analysis and systematic review of the clinical efficacy and safety of platelet-rich plasma combined with hyaluronic acid (PRP + HA) versus PRP monotherapy for knee osteoarthritis (KOA). J Orthop Surg Res. 20(1):57.
11. Fucaloro S, Bragg JT, Feldman MW, Krivicich L, Salzler MJ. (2025) Complication rates of bone marrow aspirate concentrate injections versus other injectable therapies for knee osteoarthritis: A systematic review and meta-analysis. J Orthop. 62:36-42.
12. Sadeghirad B, Rehman Y, Khosravirad A, Sofi-Mahmudi A, Zandieh S, et al. (2024) Mesenchymal stem cells for chronic knee pain secondary to osteoarthritis: A systematic review and meta-analysis of randomized trials. Osteoarthritis Cartilage. 32(10):1207-19.
13. Tang X, Huang H, Hao L. (2024) Decadal analysis of efficacy and safety profiles of mesenchymal stem cells from varied sources in knee osteoarthritis patients: A systematic review and network meta-analysis. Exp Gerontol. 192:112460.
14. Wang J, Liang J, Yao J, Song HX, Yang XT, et al. (2022) Meta-analysis of clinical trials focusing on hypertonic dextrose prolotherapy (HDP) for knee osteoarthritis. Aging Clin Exp Res. 34(4):715-24.
15. Vadalà G, Russo F, Lavazza C, Petrucci G, Ambrosio L,  et al. (2025) Intradiscal Mesenchymal Stromal Cell Therapy for the Treatment of Low Back Pain Due to Moderate-to-Advanced Multilevel Disc Degeneration: A Preliminary Report of a Double-Blind, Phase IIB Randomized Clinical Trial (DREAM Study). JOR Spine. 8(2):e70086.
16. Kong Y, Wang Y, Yang Y, Hou Y, Yu J, et al. (2024) Intra-articular injection of exosomes derived from different stem cells in animal models of osteoarthritis: a systematic review and meta- analysis. J Orthop Surg Res. 19(1):834.
17. Wang Z, Hu Z, Niu L, Xu Y, Qi Y. (2025) Mesenchymal stem cell-derived exosomes for the treatment of knee osteoarthritis: a systematic review and meta-analysis based on rat model. Front Pharmacol. 16:1588841.
18. Du W, Cui HP, Fu BS, Li W, Liu Q, et al. (2020) [Platelet rich plasma intra-articular and extra-articular injection for the treatment of knee osteoarthritis]. Zhongguo Gu Shang. 33(3):209-13.
19. Nouri F, Babaee M, Peydayesh P, Esmaily H, Raeissadat SA. (2022) Comparison between the effects of ultrasound guided intra-articular injections of platelet-rich plasma (PRP), high molecular weight hyaluronic acid, and their combination in hip osteoarthritis: a randomized clinical trial. BMC Musculoskelet Disord.23(1):856.
20. Kraeutler MJ, Houck DA, Garabekyan T, Miller SL, Dragoo JL, et al. (2021) Comparing Intra-articular Injections of Leukocyte-Poor Platelet-Rich Plasma Versus Low-Molecular Weight Hyaluronic Acid for the Treatment of Symptomatic Osteoarthritis of the Hip: A Double-Blind, Randomized Pilot Study. Orthop J Sports Med. 9(1):2325967120969210.
21. Raeissadat SA, Gharooee Ahangar A, Rayegani SM, Minator Sajjadi M, Ebrahimpour A, et al. (2020) Platelet-Rich Plasma-Derived Growth Factor vs Hyaluronic Acid Injection in the Individuals with Knee Osteoarthritis: A One Year Randomized Clinical Trial. J Pain Res. 13:1699-711.
22. Filardo G, Kon E, Di Martino A, Di Matteo B, Merli ML, et al. (2012) Platelet-rich plasma vs hyaluronic acid to treat knee degenerative pathology: study design and preliminary results of a randomized controlled trial. BMC Musculoskelet Disord. 13:229.
23. Arliani GG, Durigon TS, Pedroso JP, Ferreira GF, Oksman D, et al. (2022) Intra-articular Infiltration of Platelet-Rich Plasma versus Hyaluronic Acid in Patients with Primary Knee Osteoarthritis: Preliminary Results from a Randomized Clinical Trial. Rev Bras Ortop (Sao Paulo). 57(3):402-408.
24. Abbadi W, Kara Beit Z, Al-Khanati NM. (2022) Arthrocentesis, Injectable Platelet-Rich Plasma and Combination of Both Protocols of Temporomandibular Joint Disorders Management: A Single-Blinded Randomized Clinical Trial. Cureus. 14(11):e31396.
25. Elksniņš-Finogejevs A, Vidal L, Peredistijs A. (2020) Intra-articular platelet-rich plasma vs corticosteroids in the treatment of moderate knee osteoarthritis: a single-center prospective randomized controlled study with a 1-year follow up. J Orthop Surg Res. 15(1):257.
26. Patel S, Gahlaut S, Thami T, Chouhan DK, Jain A, et al. (2024) Comparison of Conventional Dose Versus Superdose Platelet-Rich Plasma for Knee Osteoarthritis: A Prospective, Triple-Blind, Randomized Clinical Trial. Orthop J Sports Med. 12(2):23259671241227863.
27. Lu J, Wen T, Huang H, Xie C, Kang J, et al. (2025) Intraarticular injection of the stromal vascular fraction for the treatment of knee osteoarthritis a prospective randomized controlled clinical trial. Sci Rep. 15(1):28109.
28. Jing B, Zhang Z, Lv Y, Zhou S, Liu F, et al. (2026) Injection of Adipose-Derived Mesenchymal Stem Cell-Enriched Adipose Extract for Temporomandibular Joint Osteoarthritis: A Randomised Controlled Trial. Int Dent J. 76(4):109604.
29. Lu L, Dai C, Zhang Z, Du H, Li S, et al. (2019) Treatment of knee osteoarthritis with intra-articular injection of autologous adipose-derived mesenchymal progenitor cells: a prospective, randomized, double-blind, active-controlled, phase IIb clinical trial. Stem Cell Res Ther. 10(1):143.
30. Wu CZ, Shi ZY, Wu Z, Lin WJ, Chen WB, et al. (2023) Mid-term outcomes of microfragmented adipose tissue plus arthroscopic surgery for knee osteoarthritis: A randomized, active-control, multicenter clinical trial. World J Stem Cells. 15(12):1063-76.
31. Rahimzadeh P, Rahimzadeh P, Imani F, Faiz SHR, Entezary SR, et al. (2018) The effects of injecting intra-articular platelet-rich plasma or prolotherapy on pain score and function in knee osteoarthritis. Clin Interv Aging. 13:73-79.
32. Sun SF, Lin GC, Hsu CW, Lin HS, Liou IS, et al. (2021) Comparing efficacy of intraarticular single crosslinked Hyaluronan (HYAJOINT Plus) and platelet-rich plasma (PRP) versus PRP alone for treating knee osteoarthritis. Sci Rep. 11(1):140.
33. Joshi Jubert N, Rodríguez L, Reverté-Vinaixa MM, Navarro A. (2017) Platelet-Rich Plasma Injections for Advanced Knee Osteoarthritis: A Prospective, Randomized, Double-Blinded Clinical Trial. Orthop J Sports Med. 5(2):2325967116689386.
34. Zhou Y, Li H, Cao S, Han Y, Shao J, et al. (2023) Clinical Efficacy of Intra-Articular Injection with P-PRP Versus that of L-PRP in Treating Knee Cartilage Lesion: A Randomized Controlled Trial. Orthop Surg. 15(3):740-749.
35. Romandini I, Boffa A, Di Martino A, Andriolo L, Cenacchi A, et al. (2024) Leukocytes Do Not Influence the Safety and Efficacy of Platelet-Rich Plasma Injections for the Treatment of Knee Osteoarthritis: A Double-Blind Randomized Controlled Trial. Am J Sports Med. 52(13):3212-22.
36. Baria M, Pedroza A, Kaeding C, Durgam S, Duerr R, et al. (2022) Platelet-Rich Plasma Versus Microfragmented Adipose Tissue for Knee Osteoarthritis: A Randomized Controlled Trial. Orthop J Sports Med. 10(9):23259671221120678.
37. Pretorius J, Nemat N, Alsayed A, Mustafa A, Hammad Y, et al. (2022) Double-Blind Randomized Controlled Trial Comparing Platelet-Rich Plasma With Intra-Articular Corticosteroid Injections in Patients With Bilateral Knee Osteoarthritis. Cureus. 2022;14(9):e29744.
38. Huang HY, Hsu CW, Lin GC, Lin HS, Chou YJ, et al. (2022) Comparing efficacy of a single intraarticular injection of platelet-rich plasma (PRP) combined with different hyaluronans for knee osteoarthritis: a randomized-controlled clinical trial. BMC Musculoskelet Disord. 23(1):954.
39. Karaborklu Argut S, Celik D, Ergin ON, Kilicoglu OI. (2024) Does the Combination of Platelet-rich Plasma and Supervised Exercise Yield Better Pain Relief and Enhanced Function in Knee Osteoarthritis? A Randomized Controlled Trial. Clin Orthop Relat Res. 482(6):1051-61.
40. Baria M, Barker T, Durgam S, Pedroza A, Flanigan D, et al. (2024) Microfragmented Adipose Tissue Is Equivalent to Platelet-Rich Plasma for Knee Osteoarthritis at 12 Months Posttreatment: A Randomized Controlled Trial. Orthop J Sports Med. 12(3):23259671241233916.
41. Molnar V, Jeleč Ž, Rod E, Hudetz D, Brlek P, et al. (2025) Clinical and dGEMRIC Evaluation of Microfragmented Adipose Tissue Versus Hyaluronic Acid in Inflammatory Phenotype of Knee Osteoarthritis: A Randomized Controlled Trial. Biomedicines. 13(9):2301.
42. Gobbi A, Dallo I, D'Ambrosi R. (2023) Autologous microfragmented adipose tissue and leukocyte-poor platelet-rich plasma combined with hyaluronic acid show comparable clinical outcomes for symptomatic early knee osteoarthritis over a two-year follow-up period: a prospective randomized clinical trial. Eur J Orthop Surg Traumatol. 33(5):1895-1904.
43. Dulic O, Rasovic P, Lalic I, Kecojevic V, Gavrilovic G, et al. (2021) Bone Marrow Aspirate Concentrate versus Platelet Rich Plasma or Hyaluronic Acid for the Treatment of Knee Osteoarthritis. Medicina (Kaunas). 57(11):1193.
44. Dwyer T, Hoit G, Lee A, Watkins E, Henry P, et al. (2021) Injection of Bone Marrow Aspirate for Glenohumeral Joint Osteoarthritis: A Pilot Randomized Control Trial. Arthrosc Sports Med Rehabil. 3(5):e1431-40.
45. Lamo-Espinosa JM, Blanco JF, Sánchez M, Moreno V, Granero-Moltó F, et al. (2020) Phase II multicenter randomized controlled clinical trial on the efficacy of intra-articular injection of autologous bone marrow mesenchymal stem cells with platelet rich plasma for the treatment of knee osteoarthritis. J Transl Med. 18(1):356.
46. Lana JF, Pires L, Macedo A, Mainine S, Mosaner T, et al. (2026) Randomized clinical trial comparing intra-articular injection of bone marrow aspirate clot and bone marrow aspirate concentrate in grade 3 and 4 knee osteoarthritis. Stem Cell Res Ther. 17(1):168.
47. Anz AW, Hubbard R, Rendos NK, Everts PA, Andrews JR, et al. (2020) Bone Marrow Aspirate Concentrate Is Equivalent to Platelet-Rich Plasma for the Treatment of Knee Osteoarthritis at 1 Year: A Prospective, Randomized Trial. Orthop J Sports Med. (2):2325967119900958.
48. Gül D, Orsçelik A, Akpancar S. (2020) Treatment of Osteoarthritis Secondary to Developmental Dysplasia of the Hip with Prolotherapy Injection versus a Supervised Progressive Exercise Control. Med Sci Monit. 26:e919166.
49. Matas J, Orrego M, Amenabar D, Infante C, Tapia-Limonchi R, et al. (2019) Umbilical Cord-Derived Mesenchymal Stromal Cells (MSCs) for Knee Osteoarthritis: Repeated MSC Dosing Is Superior to a Single MSC Dose and to Hyaluronic Acid in a Controlled Randomized Phase I/II Trial. Stem Cells Transl Med. 8(3):215-224.
50. Pill SG, Ahearn B, Tokish JM, Cook C, Siffri PJ, et al. (2025) Amniotic Tissue Injections Are an Effective Alternative to Corticosteroid Injections for Pain Relief and Function in Patients With Severe Knee Osteoarthritis: A Double-Blind, Randomized, Prospective Study. J Am Acad Orthop Surg Glob Res Rev. 9(1).
51. Manchikanti L, Abd-Elsayed A, Jin MY, Kaye AD, Sanapati MR, et al. (2026) A Systematic Review and Meta-Analysis of Randomized Controlled Trials of Epidural Regenerative Medicine Therapies in Managing Spinal Pain. Curr Pain Headache Rep. 30(1):83.