Effectiveness of Platelet‐Rich Plasma in Knee Osteoarthritis: A Systematic Review and Meta‐Analysis

ABSTRACT

Introduction

Knee osteoarthritis (OA) causes chronic pain and disability, with limited benefit from conventional treatments. Platelet‐rich plasma (PRP) has been proposed as a biologic therapy with potential tissue‐repair properties; however, its clinical efficacy remains uncertain.

Objective

To assess the effectiveness of intra‐articular PRP injections in reducing pain and improving function in patients with knee OA.

Methodology

This systematic review and meta‐analysis included randomized controlled trials comparing intra‐articular PRP with placebo or no treatment. Pain outcomes were measured using the Visual Analog Scale (VAS), while functional outcomes were assessed using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) or Knee injury and Osteoarthritis Outcome Score (KOOS).

Results

PRP significantly reduced pain and improved function compared with placebo across short‐, mid‐, and long‐term follow‐up. The magnitude of benefit varied but remained clinically relevant.

Conclusion

PRP provides significant improvements in pain relief and functional outcomes compared with placebo or no treatment in patients with knee OA, supporting its potential as a therapeutic option. Standardized protocols and larger RCTs are warranted to further confirm its clinical effectiveness.

1. Introduction

Osteoarthritis (OA) is a joint disease characterized by the slow degradation of the cartilage that covers the joints. Various symptoms of this illness, such as pain, stiffness, and edema, have a major influence on functional ability and quality of life [1, 2]. The knee joint is the most impacted by osteoarthritis due to its regular use and increased strain throughout life. Knee osteoarthritis (OA) primarily affects people over the age of aged people, particularly those with a higher BMI. In response to population aging and the increasing prevalence of obesity, the risk and epidemiological trend of knee osteoarthritis are anticipated to rise significantly [3].

The fundamental characteristics of knee osteoarthritis, such as and functional impairment are routinely assessed with validated methods such as the Western Ontario and McMaster Universities Arthritis Index and the Visual Analog Scale [4, 5, 6, 7]. Treatment aims to relieve symptoms and improve functional ability safely and effectively. Physical therapy and oral medication are usually the first choices of treatment. Nevertheless, should the symptoms continue to manifest, the administration of intra‐articular injections may be considered as a potential intervention [7].

Corticosteroids, hyaluronic acid, and Platelet‐rich plasma (PRP) are among the options utilized in these injections, and continuing research is being conducted to assess their safety and effectiveness [8]. PRP is derived from the patient's blood and is processed to extract a high concentration of platelets. These platelets are rich in growth factors and bioactive molecules, which are believed to play a crucial role in promoting wound healing and modulating the repair process [9, 10]. However, the findings regarding PRP's effectiveness vary across different studies, indicating a need for further research to assess its role in treating knee osteoarthritis [11, 12, 13].

Given the rising prevalence of knee osteoarthritis, it is critical to assess the safety and efficacy of various treatments, especially PRP. The purpose of this systematic review and meta‐analysis is to determine the therapeutic efficacy of intra‐articular platelet‐rich plasma injections in knee osteoarthritis.

2. Methods

2.1. Search Strategy

This review followed the PRISMA guidelines and was registered with PROSPERO (CRD420251019846) [14]. A comprehensive literature search was performed on March 30, 2025, utilizing four electronic databases: PubMed, the Virtual Health Library (VHL), the Cochrane Library, and Open SIGLE. Only published articles were included, with no language restrictions applied. Eligible studies were randomized clinical trials (RCTs) involving participants aged 18 years and older.

In each database, advanced search options were utilized with specific keywords such as “clinical trial AND platelet‐rich plasma OR PRP AND knee osteoarthritis OR OA”.

2.2. Inclusion and Exclusion Criteria

Titles and abstracts were used to screen the search results from each database (Table 1). The inclusion criteria were as follows:

1. Studies published until 2025.

2. Clinical trials evaluating the safety and efficacy of intra‐articular PRP injections in patients with knee osteoarthritis.

3. Participants aged ≥ 18 years and diagnosed with knee OA (no further restrictions on patient demographics).

4. In order to assess the efficacy of PRP injections for the treatment of knee osteoarthritis, the analysis was restricted to studies that employed a true placebo‐controlled design.

TABLE 1.

Description of the five included randomized, placebo‐controlled clinical trials.

Author (Year)	Country	Average age (years)	Average BMI (kg/m2)	Injections received	Participants	Study assessment tool
Ghai et al. (2019) [15]	India	Average—49.8	Average weight—67 kg	PRP—1 P—1	PRP – 10 P—10	VAS & WOMAC
Murillo et al. (2021) [16]	Brazil	PRP—66.4 P—62.5	PRP – 28.3 P—27.6	PRP—2 P—2	PRP—20 P—21	VAS & WOMAC
Chu et al. (2022) [17]	China	PRP—53.91 P—54.51	PRP—27.5 P—27.9	PRP – 3 P—3	PRP—308 P—302	VAS & WOMAC
Mardani‐Kivi et al. (2023) [18]	Iran	PRP – 62.6 P—60.6	PRP—29.7 P—29.1	PRP – 1 P—1	PRP—30 P—30	VAS & WOMAC
Yoshioka et al. (2024) [19]	Japan	PRP – 65.9 P—67.9	PRP—28.7 P—28.1	PRP – 1 P—1	PRP—15 P—15	VAS & WOMAC

Abbreviations: P, Placebo; PRP, platelet‐rich plasma; VAS, Visual Analog Scale; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index.

The criteria for exclusion were as follows:

1. Studies with only abstracts, no full‐text articles, or case reports.

2. Studies using non‐human subjects.

3. Studies involving participants under 18 years of age were excluded.

Articles meeting the eligibility criteria were included, and duplicate records were removed (Table 1).

2.3. Data Extraction

Statistical analysis was performed using Review Manager 5.4. Two independent reviewers conducted the literature screening and data extraction for studies published up to 2025. Any disagreements between the reviewers were addressed through discussion with a third reviewer to ensure consistency and accuracy in the selection process.

2.4. Data Extraction Process

The systematic review was performed using Rayyan, an online platform specifically developed to facilitate the organization and screening of studies in systematic reviews [20]. To avoid observer bias, reviewers informed each other during the screening process whether they were including or excluding items. Rayyan also allowed reviewers to specify the reasons for their decisions to include or exclude studies. Once the initial screening was completed, the results were unblinded, and the reviewers discussed and resolved any discrepancies regarding article selection.

For data extraction, all eligible articles were independently reviewed using Google Forms, and reviewers were blinded to each other's inputs. The extracted data included the extraction date, journal, publication year, study location, design, objectives, outcomes, treatment details, demographics, effect sizes, and potential conflicts of interest. Statistical analyzes were then performed by the first reviewer based on this data.

2.5. Selective Studies and Sample Characteristics

A comprehensive search identified 6566 prospective studies from four databases: PubMed (n = 3642), Cochrane (n = 2666), VHL (n = 158), and Open Sigle (n = 100) (Figure 1). An automated screening procedure was implemented, utilizing publication date and study type as criteria for eligibility assessment. The process led to the exclusion of 3344 papers that didn't meet the criteria. Subsequent screening led to the removal of an additional 3200 papers due to issues of duplication or non‐compliance with qualification requirements. Ultimately, 22 studies remained for final screening and data extraction. During the screening process, the remaining eligible study data were collected in a blinded manner by 2 authors.

FIGURE 1.

PRISMA flow diagram.

This criterion, however, led to the exclusion of 17 studies, leaving 5 studies for the meta‐analysis.

2.6. Selected Studies and Sample Characteristics

Individual research data were presented utilizing several techniques in this study. Detailed tables were prepared to arrange each study's main data, including the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and Visual Analog Scale (VAS) scores. These tables provide the mean values, standard deviations (SD), and sample sizes (n) for both the PRP and placebo groups at baseline and 6 months post‐treatment. To assess the therapeutic efficacy of PRP, this study assessed the effect size (Cohen's d) and reported the corresponding standard errors and p‐values to quantify and compare treatment outcomes across studies.

2.7. Meta‐Analysis

Due to the diversity of our data, we employed meta‐analysis methods to aggregate the findings from numerous investigations. A random‐effect model was used to analyze each comparison, such as baseline PRP vs. 6 months, baseline placebo vs. 6 months, or 6‐month PRP vs. placebo for WOMAC and VAS, to calculate the combined effect size. The model takes into account both between and within‐study variances, as well as prospective variations. We quantified total heterogeneity in the study using the limited maximum likelihood (REML) approach and τ² estimator [21]. Cochran's Q statistic was employed to assess statistical heterogeneity. This statistic utilizes the chi‐square distribution to calculate variability between studies, rather than within individual study participants, and indicates high heterogeneity when the p‐value is < 0.0001 [22]. Furthermore, the I ² index was utilized to evaluate heterogeneity. This index determines heterogeneity by computing the ratio of the Cochran Q test result to the degrees of freedom, as well as the Q value [23]. The I ² index enhances the Cochran Q test by revealing greater heterogeneity [24].

2.8. Cochrane Risk of Bias Tool

The Cochrane Risk of Bias Tool was used in this systematic review and meta‐analysis to thoroughly analyze the methodological quality, credibility of the evidence, and potential sources of bias in the included studies. The analysis included only randomized controlled trials (RCTs), which are widely regarded as the gold standard for assessing the efficacy and safety of therapeutic interventions. To ensure objectivity and consistency, a blinded assessment was conducted based on the six domains specified in the Cochrane Risk of Bias Tool, which include: (1) random sequence creation, (2) allocation concealment, (3) participant and personnel blinding, (4) outcome assessment blinding, (5) missing outcome data, (6) selective reporting, and (7) other potential sources of bias.

The two authors found no domains with a “high” risk of bias among the five research studies included in the meta‐analysis. Additionally, one study raised “some concerns” about bias, as one author marked at least one area as “unclear.” However, we approached the evaluation of bias risk with great care, and these studies received a maximum of one or two “unclear”. No study was evaluated by both authors as exhibiting a “low” risk of bias in any domain. Although one author rated all five studies as “low” in each domain, the other author still provided an “unclear” rating in at least one domain [15, 16, 17, 18, 19]. Therefore, it was concluded that these studies raised “some concerns” regarding overall bias risk.

A post hoc sensitivity analysis was performed for each outcome by excluding studies with a high risk of bias and including only the five studies assessed as having a low risk of bias to determine the robustness of the pooled estimates. This method aimed to evaluate the robustness of the meta‐analysis findings by excluding the two studies that presented bias concerns. This action was taken before finalizing the meta‐analysis, and the authors were unaware of the outcomes of the bias risk evaluations (Figure 2A,B).

FIGURE 2.

(A) Bias and study summary. (B) Risk of bias graph.

This approach facilitated a comprehensive evaluation of the internal validity of the included trials [25, 26].

3. Results

All five studies in this systematic review were randomized, double‐blinded, placebo‐controlled clinical trials, indicating a high level of scientific rigor [15, 16, 17, 18, 19]. The frequency of interventions exhibited variability among the studies reviewed. In the five included studies, participants were administered three injections of either platelet‐rich plasma or saline, typically on a weekly basis. Five studies administered a single intra‐articular injection of either PRP or normal saline. In two of the included studies, ultrasound‐guided intra‐articular injections were administered into the knee joint to deliver either platelet‐rich plasma or normal saline as a placebo, with injection intervals set at 2 weeks. To evaluate treatment efficacy, five studies used the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and the Visual Analog Scale (VAS) as the primary outcome measures for assessing functional outcomes and pain intensity, respectively. Additionally, five studies incorporated both VAS and WOMAC scores at the six‐month follow‐up to evaluate sustained treatment effects (Table 1).

3.1. Visual Analog Scale (VAS)

The objective of this meta‐analysis was to evaluate the effectiveness of platelet‐rich plasma injections in comparison to placebo in alleviating pain related to knee osteoarthritis (OA), with a particular focus on alterations in the Visual Analog Scale. We analyzed data from five studies to determine the impact of PRP treatment over 6 months.

The results revealed a significant reduction in pain intensity following PRP treatment at 6 months, indicating a notable therapeutic benefit. At the six‐month follow‐up, the aggregated mean difference (MD) between platelet‐rich plasma and placebo was found to be 0.53, with a 95% confidence interval (CI) ranging from 0.39 to 0.68 and an I ² statistic of 2% (Figure 3). This statistically significant difference suggests PRP's potential efficacy in treating knee osteoarthritis‐related pain.

FIGURE 3.

The comparison of VAS scores between PRP and placebo.

3.2. Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC)

In addition to pain intensity, our meta‐analysis also utilized the WOMAC index as a primary evaluation tool to assess the functional outcomes of PRP treatment for knee osteoarthritis. This review included five trials that compared WOMAC scores at baseline and 6 months after treatment, assessing changes in pain, stiffness, and physical function in patients with knee osteoarthritis.

The aggregated data demonstrated a significant improvement in WOMAC ratings in the PRP group after 6 months, indicating improved joint function and fewer symptoms.

At 6 months post‐treatment, there was a significant difference in WOMAC ratings between the PRP and placebo groups (MD = 0.14, 95% CI: −0.00 to 0.28, I ² = 0%) (Figure 4). This study indicates that PRP injections had a consistent and beneficial effect on patients' reported functional outcomes.

FIGURE 4.

The comparison of WOMAC scores between PRP and placebo.

4. Discussion

This comprehensive review and meta‐analysis examined five randomized controlled trials with 761 patients in total. Platelet‐rich plasma therapy appears to considerably reduce pain in patients with knee osteoarthritis. Pain outcomes were predominantly assessed using the Visual Analogue Scale and the Western Ontario and McMaster Universities Osteoarthritis Index, with both methods consistently demonstrating greater pain alleviation following PRP injection. The results of this study suggest that PRP therapy has potential clinical benefits for controlling knee OA symptoms; however, more research with bigger sample numbers and longer follow‐up periods is required to establish its long‐term efficacy and stability.

The findings of this study suggest that PRP therapy may offer potential clinical benefits for improving knee osteoarthritis symptoms. However, further large‐scale studies with longer follow‐up periods are needed to confirm its long‐term efficacy and sustained clinical effects.

The pooled analysis of VAS scores revealed a significant effect of PRP treatment at 6 months, with an approximate mean reduction of 9 mm in pain intensity compared to placebo. Analysis of WOMAC scores also showed a significant overall effect with an estimated mean reduction of 6.15 points in the PRP group compared to placebo, indicating improved joint function and symptom relief. The heterogeneity observed across included studies suggests the need for further investigation into factors that may influence patient outcomes, such as injection protocols, treatment regimens, and population characteristics.

Intra‐articular PRP injections demonstrated superior efficacy over placebo in relieving pain and improving knee function within 6 months; however, the long‐term effectiveness remains uncertain. While this study focused on the therapeutic efficacy of PRP alone, a 2022 meta‐analysis reported that the incidence of adverse events was lower when PRP was combined with hyaluronic acid (HA) compared to PRP alone, suggesting that combination therapy may offer enhanced safety [27]. Though the combination of PRP and HA did not show superior efficacy over PRP alone, future research may explore the potential benefits of combining PRP with other agents, such as corticosteroids, to enhance clinical outcomes [27].

A meta‐analysis compared leukocyte‐rich and leukocyte‐poor PRP to determine which is more effective for treatment [27, 28, 29, 30, 31]; however, this issue was not further explored in the present study.

In our investigation, all control groups demonstrated Q‐statistics with p‐values less than 0.0001, signifying a substantial level of heterogeneity. These analyzes were carried out utilizing Review Manager version 5.4 software in conjunction with the “meta” package [32].

4.1. Limitations

This study has several limitations. First, there were no specific guidelines regarding the type of PRP or the injection protocols used. It is currently unknown whether a defined number of injections can provide the best therapeutic benefit [27, 28].

Furthermore, the data utilized in this systematic review and meta‐analysis were determined by the availability of information and the comparison methods employed, resulting in a restricted number of studies after screening. This limitation could impact the generalizability and strength of the findings.

4.2. Strengths

This analysis only examined randomized clinical studies (RCTs) that compared PRP to placebo or control groups, not other potential treatments. This design lowers variability when comparing findings. We also assessed the outcomes at 6 months, which is a key time point for evaluating the effect of PRP injections. Furthermore, we included the most recent articles (published from 2019 to 2025) with no restrictions on study location.

This study followed the PRISMA standards and used tools like the Cochrane Intervention Systematic Review Handbook to ensure that systematic measures were taken throughout the research process. Blinding was implemented at several phases, including data retrieval, screening, and extraction. All effect sizes were determined using the same approach (Cohen's d), allowing more exact comparisons between trials.

Several tests revealed high heterogeneity in the study, with no evidence of publication bias. Finally, we assessed two outcomes to gain a better understanding of the efficacy of PRP injections for treating knee osteoarthritis.

5. Conclusion

Overall, most studies discovered that PRP treatment reduced Visual Analogue Scale and Western Ontario and McMaster University Osteoarthritis Index scores at 6 months, indicating improved pain and function.

Author Contributions

Hsuan‐Wei Liu: writing – original draft, writing – review and editing.

Funding

The author has nothing to report.

Disclosure

The author confirm that no paper mills and artificial intelligence were used.

Consent

The author has nothing to report.

Conflicts of Interest

The author declares no conflicts of interest.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.