Platelet rich plasma versus corticosteroids for lateral epicondylitis: a meta-analysis of randomized clinical trials
Article information
Abstract
Background
Lateral epicondylitis, colloquially known as tennis elbow, is a common cause of elbow pain and daily task disability. Caused by repetitive movement, it is typically a degenerative rather than inflammatory event and affects mostly middle-aged patients. Despite its good prognostic nature, its economic burden on the healthcare system encourages research on the efficacity of non-operative injection treatments. This article aims to compare the clinical effectiveness of platelet-rich plasma (PRP) and corticosteroid (CS) injections in managing lateral epicondylitis.
Methods
PubMed, Cochrane, and Google Scholar (pages 1–20) were searched up to March 2024. Only randomized controlled trials were included. The clinical outcomes evaluated were the visual analog scale (VAS) and Disabilities of the Arm, Shoulder, and Hand (DASH) score.
Results
Twenty-six randomized controlled trials with 1.877 patients were included in this meta-analysis. In terms of VAS scores, short-term results (<2 months) favored CS over PRP (P=0.03; mean difference [MD], 0.67; 95% CI, 0.05 to 1.28), whereas long-term results (>6 months) favored PRP (P<0.001; MD, –1.60; 95% CI, –2.01 to –1.20]). Intermediate-term results (2–6 months) showed no significant difference between injection treatments. In terms of DASH scores, short- and intermediate-term results showed no significant difference, whereas long-term results favored PRP (P<0.001; MD, –4.87; 95% CI, –7.69 to –2.06).
Conclusions
CS provides significantly better short-term pain relief, while PRP provides better long-term functional improvement and clinical long-term pain relief. However, future studies should focus on other injection protocols or addition of other non-invasive modalities.
Level of evidence
I.
INTRODUCTION
Lateral epicondylitis (LE) is most commonly referred to as tennis elbow, although only 10% of the affected population are tennis players [1]. Generally, LE is considered the number one cause of symptoms related to elbow pain. This entity is frequently seen in orthopedic clinics with 1% to 3% of men and women being annually affected [1-3]. Initially, inflammation was thought to be the driving factor of tennis elbow, but histologically, the absence of the inflammatory cells [1,4] has classified tennis elbow as a degenerative process of the tendon. The most commonly affected tendon is the extensor carpi radialis brevis [3,5,6] which, due to its unique anatomical position, is exposed to stressful forces in every movement of the arm. In fact, the strongest risk factor of LE was shown to be repetitive or highly vigorous work, thus a depiction of occupational or athletic activities [1,6-10]. This debilitating disease causes lateral elbow pain with increasing weakness in grip strength, and any additional stress on an extended wrist is painful [7,11]. In addition to disturbing the patients’ daily life activities, tennis elbow adds a huge economic burden on the United States health care system. The mean reimbursement per patient per year has steadily risen [3] in the United States every year [2].
Different treatment modalities exist for LE, from nonoperative front-line treatments divided between non-invasive and invasive procedures such as injections [12-15], to operative treatment used as a last resort. Injection therapy, specifically corticosteroid (CS) and platelet-rich plasma (PRP), remain the most common management modality. Meta-analyses aimed to resolve this controversy by comparing the clinical efficacy of the injection types during the last 5 years [16-21]. However, many of these meta-analyses had flaws such as inclusion of a study by Varshney et al. [22] that analyzed 20 patients with medial epicondylitis (ME) or analyzing different scores together without considering the standardized mean difference (SMD). Therefore, our study was deemed important because it includes the most recent literature analyzed to date to compare the clinical effectiveness of CS and PRP injections in LE treatment.
METHODS
Search Strategy
This study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. PubMed, Cochrane, and Google Scholar (pages 1–20) were searched up to March of 2024, to compare the clinical effectiveness of PRP and steroids in the treatment of LE. Boolean operators were used with the following keywords: “PRP,” “plasma,” “predni,” “steroids,” “triamcinolone,” “epicondylitis,” and “elbow.” To enhance our literature search, reference lists from papers and online searches were also scanned for possible relevant studies. Initial title and abstract screening were performed for all articles, followed by a full-text review by two authors (RM, MD). Conflicts were discussed and resolved by consensus. Data extraction was performed by one author. The PRISMA flowchart (Fig. 1) provides a summary of the search strategy.
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart for article selection.
To be included, studies had to be randomized controlled trials comparing patients with LE managed with PRP to others managed with steroids and achieving the outcomes of interest. Studies with the following characteristics were excluded from this study: non-randomized controlled studies, missing outcomes, or (3) including both patients with medial or LE [22].
Data Extraction
The data extracted from the included studies consisted of clinical outcomes of visual analog scale (VAS) for pain score and Disabilities of the Arm, Shoulder, and Hand (DASH) score.
Risk of Bias Assessment
The risk of bias assessment was performed independently by two authors. To grade eligible studies, the Cochrane risk-of-bias tool was used. For each trial, six risk-of-bias categories were assessed and scored: random sequence generation, allocation concealment, blinding of participants and personnel to the study protocol, blinding to outcome assessment, incomplete outcome data, and selective reporting. Scores used were high, low, or unclear risk (Fig. 2A). Trials were labeled as high risk of bias if >1 key domain was scored as high. When all key domains were scored as low, the trial was labeled as low risk of bias. Other trials not meeting either of these criteria had an unclear risk of bias.
(A) Risk of bias items for each included trial. (B) Risk of bias items presented as percentages across all included trials.
Statistical Analysis
The statistical analysis was performed using Review Manager 5.4 (The Cochrane Collaboration, 2020). The studied outcomes were compared using mean differences (MD) and 95% CIs. Q tests and I2 statistics were used to evaluate heterogeneity. If P≤0.05 or I2 >50%, showing considerable heterogeneity, the random-effects model was adopted. Otherwise, the fixed-effect model was chosen. Statistical significance is shown by P<0.05. In addition, based on previous studies, short-term follow-up was defined as <2 months, intermediate follow-up as between 2 and 6 months, and long-term follow-up as >6 months [16,18].
RESULTS
Characteristics of the Included Studies
Twenty-six randomized controlled trials met the inclusion criteria and were included in this meta-analysis [23-48]. In total, 1,877 patients were included, with 943 in the PRP group and 934 in the steroid group. The main characteristics of the included studies are summarized in Table 1. The results of the bias assessment are summarized in Fig. 2B.
Characteristics of the included studies
Visual Analog Scale
Twenty randomized controlled trials on 1,540 subjects (774 PRP and 766 steroids) reported data on VAS at less than 2 months, 18 randomized controlled trials on 1,370 subjects (689 PRP and 681 steroids) reported data on VAS between 2 and 6 months, and 15 randomized controlled trials on 1,143 subjects (574 PRP and 569 steroids) reported data on VAS at more than 6 months. A random-effects model was used for all three of these analyses due to I2 >50%. At less than 2 months, PRP had significantly higher VAS scores (P=0.03; MD, 0.67; 95% CI, 0.05–1.28) (Fig. 3A). Between 2 and 6 months after the injection, there was no significant difference in VAS score between PRP and steroids (P=0.09; MD, –0.38; 95% CI, –0.81 to 0.05) (Fig. 3B). At more than 6 months, PRP had significantly lower VAS scores than steroids (P<0.001; MD, –1.60; 95% CI, –2.01 to –1.20) (Fig. 3C).
Forest plots showing visual analog scale ((VAS) score between platelet-rich plasma (PRP) and steroids (A) at <2 months, (B) at 2–6 months, and (C) at >6 months. SD: standard deviation, IV: inverse variance.
DASH Score
Twelve randomized controlled trials on 858 subjects (433 PRP and 425 steroids) reported data on DASH at less than 2 months, 10 randomized controlled trials on 738 subjects (373 PRP and 365 steroids) reported data on DASH between 2 and 6 months, and 9 randomized controlled trials on 611 subjects (308 PRP and 303 steroids) reported data on DASH at more than 6 months. A random-effects model was used for all three of these analyses due to the I2 >50%. At less than 2 months and between 2 and 6 months, there was no significant difference in DASH score between PRP and steroids (P=0.36; MD, 2.39; 95% CI, –2.69 to 7.47 [Fig. 4A] and P=0.10; MD, –3.24; 95% CI, –7.06 to 0.59 [Fig. 4B], respectively). At more than 6 months, PRP had significantly lower (better) DASH scores than steroids (P<0.001; MD, –4.87; 95% CI, –7.69 to –2.06) (Fig. 4C).
Forest plots showing the difference in Disabilities of the Arm, Shoulder, and Hand (DASH) score between platelet-rich plasma (PRP) and steroids (A) at <2 months, (B) at 2–6 months, and (C) at >6 months. SD: standard deviation, IV: inverse variance.
DISCUSSION
LE, in addition to hindering daily life activities, is also a significant healthcare financial burden [3]. Therefore, study of its most used treatments is of importance. The basis of this study was performed to include newly published randomized control trials in a meta-analysis. This study used a three-time-point system to minimize heterogeneity from different follow-up times in the original studies. Therefore, our study had the highest number of included studies and addressed shortcomings of previous published works. Our study answered previously posed questions regarding the superiority of one treatment over the other in terms of pain relief or functional improvement. Our findings avoided the limitation of Varshney et al [22], who included patients with ME, which affects different muscles from those in LE. This difference in ME can impact muscle functions and pain measures not observed in LE [49]. Other meta-analyses [17,50] included a study by Lebiedziński et al. [51] that compared autologous conditioned plasma (ACS] to CSs, which do not contain the same cytokines, causing a high risk of sampling bias. The most recent meta-analysis [16] mixed outcomes (DASH and QuickDASH), measured using multiple scales, and did not consider using SMD for the analysis. Furthermore, they analyzed pain outcomes measured through Patient-Rated Tennis Elbow Evaluation (PRTEE) mixed with those measured through VAS and did not adjust the DASH outcome of Peerbooms et al. [26], based on a 0–200 scale, to match the other results based on a 0–100 scale.
In terms of short-term (<2 months) pain relief, our analysis showed significantly better performance of CS injections. However, the change of pain scores did not exceed the published minimal clinically important difference (MCID) of 1.5 [52], and this result not significant clinically. Though intermediate-term (2–6 months] effects were not significantly different, the long term (>6 months) outcomes were statistically and clinically significantly better with PRP. A recent meta-analysis by Xu et al. [16] showed similar results. Another study by Hohmann et al. [17] reported similar results for short- and long-term pain relief but observed a significant difference in favor of PRP at the intermediate term. This discrepancy could be the result of a smaller number of analyzed trials and inclusion of the limited study by Varshney et al. [22].
The possibility of simultaneous use of PRP and steroid injections should be considered, with the latter being solely administered for short-term pain relief. However, these two injections may have counteracting effects and should be compared in further study. In terms of functional improvement, our study showed no significant difference in DASH scores in the short and intermediate term. Though there was significant difference in favor of PRP in the long term, it was not clinically significant as it did not exceed the published MCID of 15.8 [53]. Our short-term and long-term results seem to be in accordance with almost all published studies, but Xu et al. [16] showed a discrepancy in the short-term, favoring the use of CS. This can be explained by the lower number of included studies, in addition to their other mentioned analysis-related flaws. DASH and VAS results highlight the prolonged PRP effect that can be explained by the ability of PRP to reduce inflammatory receptors and increase anti-inflammatory ones, theoretically healing tissue over time [54,55]. The present study showed a long-term but not clinically significant functional improvement along with clinically significant long-term pain relief with PRP.
Meta-analyses by Xu et al. [21], Li et al. [19], and Huang et al. [18] showed similar results, although a precise comparison could not be performed due to heterogeneous data collection intervals. In addition, the Mayo performance index was analyzed in two of those meta-analyses, a measure that can produce possible bias and proximate conclusions stemming from very limited data. Additionally, our results are based on the latest recorded scores for each follow-up interval, which was not the case in Huang et al. [18], who used all recorded data within the analyzed interval. This study is not void of limitations and suffers those of the included studies, which may have resulted in heterogeneity. The analyzed trials had non-standardized PRP preparation protocols and different dosages of PRP or CS injections. In addition, inclusion and exclusion criteria varied by study, and the data used for analysis were pooled, limiting comprehensive analyses.
CONCLUSIONS
This meta-analysis and systematic review showed that CS provides significantly better short-term (<2 months) pain relief than PRP, although the difference was however this did not reach MCID for pain. PRP provides a clinically insignificant but a statistically better long-term (>6 months) functional improvement and both a clinically and statistically significant long-term pain relief.
Notes
Author contributions
Writing – original draft: RM, MD, PB, RL, AZK. Writing – review & editing: JAA. All authors read and agreed to the published version of the manuscript.
Conflict of interest
AZK would like to disclose research support from Stryker and DePuy and was a paid presenter or speaker at Enovis. JAA would like to disclose royalties from Osteocentric Technologies, Enovis, Zimmer-Biomet, Stryker, and Globus Medical, Inc. and ownership of stocks in Shoulder Jam, Aevumed, Oberd, OTS MEDICAL, ORTHOBULLETS, ATREON, and RESTORE 3D. Research support as a PI was received from ENOVIS and ARTHREX, and royalties or financial or material support was from Wolters Kluwer, Slack Orthopaedics, and Elsevier. Board member/committee appointments have been held for the American Shoulder and Elbow Society, Mid Atlantic Shoulder AND Elbow Society, Shoulder 360, Pacira.
Funding
None.
Data availability
Contact the corresponding author for data availability.
Acknowledgments
None.