Intra-Articular Platelet-Rich Plasma Injection After Anterior Cruciate Ligament Reconstruction

Key Points Question Does postoperative platelet-rich plasma (PRP) injection improve knee symptoms and function in patients undergoing anterior cruciate ligament reconstruction (ACLR)? Findings In this randomized clinical trial of 120 participants undergoing ACLR, the intervention group received 3 doses of postoperative intra-articular PRP injection at monthly intervals, and the control group received no injection. The between-group difference in the Knee Injury and Osteoarthritis Outcome Score (mean score of 4 subscales) at 12 months was not statistically significant. Meaning In this study among patients with ACLR, the addition of postoperative PRP injection did not significantly improve knee symptoms and function.


Introduction
The anterior cruciate ligament (ACL) is a primary stabilizer of the knee and vulnerable to sports injuries, which result in knee swelling, pain, and instability. 1,2ACL reconstruction (ACLR), one of the most common orthopedic surgeries, is performed in patients with a ruptured ACL to restore knee stability and facilitate a safe return to sports. 3,4During the first year postoperatively, knee symptoms and function gradually improve while the autograft or allograft undergoes a series of healing, remodeling, and maturation processes. 5,6atelet-rich plasma (PRP) is a class of product derived from autologous blood through centrifugation, with high concentrations of growth factors and cytokines and potential for augmenting tissue repair procedures. 7,8[11][12] Nevertheless, the clinical outcomes of PRP for ACLR are controversial [13][14][15][16][17] due to the highly variable levels of evidence, methods of preparation, timing and doses of injection, and sites of administration.
Recently, meta-analyses revealed that a single dose of intraoperative PRP injection (at graft harvest sites, in femoral and tibial tunnels, or into the joint cavity) was not associated with long-term clinical or radiologic improvements at more than 6 months postoperatively. 13,15 note, intra-articular PRP injection has been expected to promote graft healing and remodeling, 9,13 which typically occur during the months following ACLR. 5,180][21] However, there is a lack of high-level comparative studies on the clinical efficacy and safety of postoperative PRP injection in patients with ACLR.
The purpose of this study was to compare the subjective outcomes and graft maturity in patients undergoing ACLR with and without postoperative intra-articular PRP injection.We hypothesized that the addition of postoperative intra-articular PRP injection would improve knee symptoms and function as well as graft maturity after ACLR.

Study Design
This surgeon-and investigator-masked randomized clinical trial (RCT) was performed at Shanghai Sixth People's Hospital, a national medical center for orthopedics.The trial protocol and statistical analysis plan (Supplement 1) were approved by the institutional review board of the Shanghai Sixth People's Hospital prior to participant enrollment on March 12, 2021.Participants signed written informed consent.This report follows the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline. 22

Participants
Patients diagnosed with ACL injury by radiology or arthroscopy and scheduled to undergo ACLR were screened for eligibility.Eligible participants were aged 16 to 45 years with epiphyseal closure of the femur and tibia, and the affected knee had a nearly normal range of motion.Participant recruitment and baseline evaluation were completed during hospitalization within 3 days before surgery.Patients were excluded if they had multiple ligament injuries; bilateral ACL injuries; contralateral knee

Conventional Procedures
Participants underwent anatomic double-bundle ACLR by 1 of 3 fellowship-trained orthopedic surgeons (G.X., S.D., and J.Z.) specialized in arthroscopic knee surgery, following a standard operating procedure. 23,24The autologous semitendinosus tendon and gracilis tendon were harvested from the affected limb and braided to prepare two 4-stranded grafts.The tibial tunnel and femoral tunnel for the anteromedial bundle and posterolateral bundle were sequentially created using a transtibial method at 90°of knee flexion.The 4-stranded grafts were separately pulled into the femoral tunnels through the corresponding tibial tunnels, with the proximal ends suspended on cortical buttons.The fixation of the distal graft ends was implemented at full knee extension using a biocomposite interference screw and a cortical button with an adjustable loop setting through a transtibial ridge tunnel.
Participants received standardized rehabilitation education from a physical therapist through verbal instruction, printed brochures, and online videos before discharge (detailed training programs provided in the trial protocol in Supplement 1).The postoperative follow-up visits were scheduled at 4 weeks, 8 weeks, 3 months, 6 months, and 12 months.

Group Allocation
To avoid imbalanced exclusion between groups, the randomization and group allocation were performed after surgery, and the demographic, radiologic, and surgical details were rechecked against the inclusion and exclusion criteria.Participants were randomized 1:1 to the PRP group or control group according to a computer-generated random number sequence.An independent trial assistant who was blocked from contacting the participants and assessors prepared the random numbers and sealed them in sequentially numbered opaque envelopes.At the first follow-up visit, the envelope with group allocation was revealed to each participant by an assistant who was not involved in the outcome measures.
To avoid low motivation for participation and potential ethical concerns because of nontherapeutic blood withdrawal in the control group (only for masking), participants were aware of their group allocation from the first follow-up visit, while the surgeons, physical therapist, clinical and radiologic outcome assessors (H.C., Y.Q., and E.C.), and statistician (J.X.) were masked to the group allocation throughout the study.

Intervention
Participants in the PRP group received 3 doses of intra-articular PRP injection from a senior treatment nurse at 4 weeks (within 3 days), 8 weeks (within 5 days), and 3 months (within 7 days) postoperatively, referenced to graft remodeling processes and previous studies on knee osteoarthritis. 5,19,25At each follow-up visit, blood withdrawal and knee injection were performed after clinical assessments and radiologic scanning to avoid interference in outcome measures.
Participants in the control group did not receive postoperative knee injection.Participants in both groups were instructed to avoid cointerventions (eg, unscheduled knee injection, nonsteroidal antiinflammatory drugs) and to report additional medications or treatment to the clinician at followup visits.
The leukocyte-poor PRP was prepared using a 2-stage centrifugation method through a commercially used system (Platelet-Rich Plasma Preparation Kit; WEGO Ltd) as previously reported. 26,27In the first stage, 45 mL of venous blood was drawn and mixed with 5 mL of anticoagulant (acid-citrate-dextrose) preloaded in the syringe.The whole blood was centrifuged at 260g for 10 minutes and separated into 3 layers.In the second stage, the bottom layer (red blood cells) was discarded, and the upper and middle layers were aspirated and recentrifuged at 360g for 15 minutes.Then, the supernatant was discarded to obtain 5 mL of liquid-form leukocyte-poor PRP, of which the platelet concentration was 4.5 times that of the whole blood on average (range, 2.8-6.0 times) (detailed characteristics shown in eTable 1 in Supplement 2).
The intra-articular injection was performed immediately after PRP preparation with the participant in the sitting position at 90°of knee flexion, through an anterolateral approach, and without exogenous activation.To increase the accuracy of the knee injection, the puncture needle was located laterally to the patellar tendon and toward the intercondylar notch under ultrasound guidance. 28,29After injection, passive knee flexion and extension were performed for 10 cycles.

Outcomes
The primary outcome was the mean score at 12 months for 4 subscales of the Knee Injury and Osteoarthritis Outcome Score (KOOS 4 ). 30,31The KOOS has been validated for ACLR and includes 42 items covering 5 subscales: pain (9 items), symptoms (7 items), activities of daily living (17 items), sports and recreation (5 items), and quality of life (4 items). 32,33The responses are scored on a 5-point Likert scale (0-4), and scores for each subscale are standardized to range from 0 to 100, with higher scores indicating better knee-related results.The KOOS 4 score excludes the subscale of activities of daily living to avoid the ceiling effect. 34condary outcomes were patient-reported outcomes (PROs), graft maturity, and physical examinations.The overall change of condition resulting from the treatment was assessed at 12 months using the Global Rating of Change (GROC) scale (range, −7 to 7, with higher positive values indicating more perceived improvement). 35Other PROs, including the KOOS subscales, 4-Item Pain Intensity Measure score (range, 0-40, with higher values indicating worse pain), 36 Tegner score (range, 0 [sick leave or disability] to 10 [highest level of sports]), 37 Lysholm score (range, 0-100, with higher scores indicating fewer knee-related symptoms), 38 and subjective International Knee Documentation Committee (IKDC) score (range, 0-100, with higher scores indicating better performance in sports activities), 39 were also assessed at 3, 6, and 12 months.
Participants underwent knee magnetic resonance imaging (MRI) (MAGNETOM Prisma; Siemens Healthineers) at the 3-, 6-, and 12-month follow-up visits.Graft maturity was evaluated by 2 independent investigators (H.C. and Y.Q.) on sagittal slices of the proton density-weighted sequence (repetition time, 3010 ms; echo time, 44 ms).The signal intensities were measured by placing unified 0.1-cm 2 circular regions of interest on the background, quadriceps tendon, and ACL graft (separated into 6 segments). 40,41For each segment, the signal-to-noise quotient (SNQ) was calculated as (signal intensity of ACL graft − signal intensity of quadriceps tendon) / signal intensity of background.Lower SNQ values, representing reduced vascularity and water content, indicated better graft maturity. 42,43ysical examinations were performed by an experienced clinician (E.C.).The active assisted range of motion (knee flexion and extension) and knee circumference at the midpatellar level were measured at each follow-up visit. 44,45The knee laxity was evaluated using the anterior drawer, Lachman, and pivot-shift tests at 12 months (details provided in the trial protocol in Supplement 1). 46

Statistical Analysis
A priori sample size calculation was performed using G*Power, version 3.1 software (Heinrich-Heine-Universität Düsseldorf).To detect a 13-point, anchor-based minimal clinically important difference (MCID) of the KOOS 4 score between the 2 groups with a common SD of 20, 30,31,47,48 a total of 120 participants (60 in each group) was required for a 2-sided P < .05, a power of 90%, and an allowed dropout rate of 15%.The primary analysis was performed according to the intention-to-treat principle.To compare the primary outcome between the 2 groups, an analysis of covariance model was applied with adjustment for the baseline KOOS 4 . 30,31,49Other PROs during follow-up visits (except for GROC scale) were also compared using analysis of covariance models adjusting for the corresponding baseline scores.In addition, the Mann-Whitney-Wilcoxon test and Student t test were used to compare ordinal or nonnormally distributed outcomes (eg, GROC scale, graft SNQ, or knee laxity) and normally distributed outcomes, respectively.The Pearson χ 2 test or Fisher exact test was used to compare dichotomous outcomes.The interrater reliability of graft maturity was evaluated using the intraclass correlation coefficient and classified as good (Ն0.75),fair (0.50-0.74), and poor (<0.50).
Sensitivity analyses included missing data, multivariable, and per-protocol analyses.For outcomes with missing data of more than 5%, multiple imputation was planned using the fully conditional specifications method; variables in the imputation model included group allocation, age, sex, sports participation, graft diameter, meniscal treatment, and baseline KOOS 4 . 50These potential outcome variables and confounders were also included in the multivariable linear regression model for the primary outcome.The per-protocol (as-treated) analysis excluded participants receiving fewer than 3 doses of PRP injection in the PRP group and participants undergoing reoperation or unscheduled knee injection in both groups.
All statistical analyses were performed using SPSS, version 27.0 software (IBM Corporation), with a 2-sided P < .05.Because of the potential for type I error due to multiple comparisons, analyses of secondary outcomes were exploratory.) were thus randomized 1:1 and included in the primary analysis.The baseline characteristics were comparable between the PRP and control groups (Table 1).

Participants
At 12 months, 114 participants (95%) were available for analysis of the primary and secondary outcomes, and 6 participants (3 in the PRP group and 3 in the control group) were lost to follow-up (5% missing data).Three participants in the PRP group who received fewer than 3 doses of PRP injection and 5 participants (2 in the PRP group and 3 in the control group) who underwent reoperation within 12 months were excluded from the per-protocol analysis.

Secondary Outcomes
The GROC scale and other PROs assessed at 3 and 12 months were not statistically significantly different between groups (  3).
No statistically significant between-group differences were found for graft maturity on the 3-and

JAMA Network Open | Orthopedics
Intra-Articular Platelet-Rich Injection After ACL Reconstruction 12-month follow-up MRI.There were no statistically significant between-group differences in the results of follow-up physical examinations (eTable 4 in Supplement 2).These results were similar in the per-protocol analysis (eTables 5-7 in Supplement 2).

Adverse Events
The intervention-related adverse events in the PRP group were mild and temporary, with 4 participants (6.7%) reporting pain at the injection site and 3 (5.0%)reporting knee swelling after injection.These symptoms were relieved within 5 days, and no additional medications were required.
The rates of common adverse events and additional medications were comparable between groups (eTable 8 in Supplement 2).

Discussion
The most important finding of this study was that 3-dose postoperative intra-articular PRP injection compared with no injection did not significantly improve knee symptoms and function at 12 months after ACLR.The adjusted between-group difference in the primary outcome was not statistically significant, and the upper limit of the 95% CI of the mean difference did not achieve the MCID (ranging from 7.9 to 18.3), 47,48,51 excluding the clinical significance of this treatment.d The Lachman test represents knee anteroposterior laxity at 30°of flexion; grades include 0 (0-2 mm), 1 (3-5 mm), 2 (6-10 mm), and 3 (>10 mm) based on the side-to-side difference.The pivot-shift test is graded as 0 (negative), 1 (glide), 2 (clunk), and 3 (gross reduction), with higher grades indicating increased knee rotatory laxity.
e The amount of ACL remnant is graded as 0 (no remnant), 1 (less than one-third), 2 (one-to two-thirds), and 3 (greater than two-thirds) based on the length proportion covering the reconstructed ACL.
f The graft sizes of AMB and PLB are measured by passing the 4-stranded semitendinosus tendon and gracilis tendon, respectively, through a nonslotted diameter measurement tool.
g The KOOS includes 42 items covering 5 subscales (pain, symptoms, activities of daily living, sports and recreation, and quality of life).Scores for each subscale range from 0 to 100, with higher scores indicating better knee-related results.The KOOS 4 is calculated as the mean score for 4 of the 5 subscales (except for activities of daily living).
h The VAS for pain is assessed on a straight horizontal line, with scores ranging from 0 (no pain) to 10 (worst pain possible).The P 4 score is calculated as the sum of VAS scores at 4 time points (morning, afternoon, evening, and with activity) during the past 2 days.
i The Lysholm Knee Questionnaire includes 8 items on subjective perception (eg, instability, pain, or locking) related to knee ligament injury; scores range from 0 to 100, with higher scores indicating fewer symptoms and better function in daily living.j The IKDC Subjective Knee Form includes 18 items covering 3 domains (symptoms, sports activities, and function); scores range from 0 (worst condition) to 100 (best condition).
Most secondary outcomes were not statistically significantly different between groups except for sports and recreation level and graft maturity at 6 months.The between-group differences in the KOOS subscale of sports and recreation and Tegner score were statistically significant (favoring the PRP group); the mean differences did not achieve the MCIDs, while the respective upper limits of the 95% CIs slightly exceeded the MCIDs (12.5 for KOOS sports and recreation; 1 for Tegner score). 48vertheless, the results of the secondary outcomes were inconclusive, with variable levels of type I and type II errors because the significance level and a priori sample size calculations were based on the primary outcome.
The clinical significance of statistically improved graft maturity at 6 months in the PRP group remains ambiguous.1][42][43] Although graft maturity can be compared using the SNQ value, no studies have reported the MCID or grading system for this measure.3][54] Therefore, further investigations are warranted to better understand the clinical relevance of graft maturity (perhaps guiding an accelerated rehabilitation program for patients with superior graft maturity or a delayed return to sports for patients with inferior graft maturity). 554][15][16] Similarly, the 12-month clinical and radiologic outcomes of the present study showed no long-term efficacy of postoperative PRP injection in patients who underwent ACLR.A meta-analysis 13 stratified the PROs by follow-up timing and found that intraoperative PRP injection was favorable only in terms of visual analog scale IKDC scores at 3 months.Another meta-analysis 15 with pooled follow-up periods showed improved Lysholm and visual analog scale scores in the PRP group; the between-group differences were statistically but not clinically significant.Furthermore, the discrepant biological hypotheses and sites of PRP administration resulted in diverse and incomparable radiologic outcomes among studies, including graft remodeling, tunnel enlargement, and harvest site healing. 14,16In terms of graft remodeling, only 1 RCT 56 with intra-articular injection and 1 RCT 57 with intratunnel and intragraft injection reported improved short-term MRI results at 4 to 6 months and 4 to 6 weeks after ACLR, respectively, while the statistical significance diminished over time.
Preclinical studies have explored a wide range of biological augmentation (eg, growth factors, stem cells, and biomaterials) for tendon-to-bone healing after ACLR. 17With regard to PRP, the discrepancy between promising biological effects in animal studies 58,59 and limited clinical efficacy in RCTs may be attributed to the persistence and complexity of graft remodeling in humans, where The KOOS 4 is calculated as the mean score for the pain, symptoms, sports and recreation, and quality-oflife subscales of the KOOS.Scores range from 0 to 100, with higher scores indicating better knee-related results.The means and 95% CIs are unadjusted.The between-group differences at follow-up visits were adjusted for the baseline KOOS 4 using an analysis of covariance model, with positive values indicating higher scores in the platelet-rich plasma (PRP) group and negative values indicating higher scores in the control group (3 months: mean, −0. the MRI signal intensity peaks at 6 months while the complete maturation process lasts for more than 2 years (compared with several weeks to months for sheep). 5,6,41ven the consensus that single-dose intraoperative PRP injection was ineffective in long-term outcomes of ACLR, it has been necessary for subsequent studies to modify the conventional protocols, such as timing and dosage of injection.Based on the hypothesis of improving knee symptoms and function by continuously promoting graft healing and remodeling, we designed a 3-dose monthly interval protocol for postoperative intra-articular injection of liquid-form, delayed activated, and leukocyte-poor PRP. 19,60,61To our knowledge, this RCT is the first high-level study of the effects of postoperative PRP injection in patients undergoing ACLR, and standardized surgical procedures, rehabilitation programs, and intervention protocols were performed to lower the risk of confounding bias.The negative results of our study were also consistent with previous RCTs using 2 to 3 doses of postoperative PRP injection at weekly or monthly intervals for patients with rotator cuff tears or ankle instability. 21,62On the basis of current findings, postoperative PRP injection should not be advocated in patients undergoing ACLR unless new evidence emerges.Nevertheless, further studies may focus on identifying patients at a high risk of biological failure 55 and determining rigorous indications for PRP or other biological augmentation.

Limitations
This study has several limitations.First, participants were not masked to the group allocation, and the control group did not receive saline injection; the assessments of PROs may be affected by subjective tendency and the placebo effect.As a remedy, radiologic evaluation and physical examinations were performed with allocation concealment from investigators.Second, although our sample included participants treated at a national medical center using standardized protocols, generalizability of the results may be limited for clinicians applying different surgical techniques or treatment protocols.
Third, there were difficulties in quantifying the compliance and completion of rehabilitation programs and assessing the intergroup imbalance.Fourth, the analyses of whole blood and PRP were limited to blood cell counting, while the concentrations of growth factors and cytokines were not analyzed for participants in this trial.Fifth, the follow-up duration was relatively short.Although a 12-month follow-up is considered long enough for studies on PRP, 13,14 the progression of graft maturation and return to sports usually lasts for years, 5,6 and a longer follow-up period may provide more information.a The graft maturity is evaluated on the sagittal plane of fat-saturated proton density-weighted magnetic resonance imaging.The SNQ is calculated as (signal intensity of ACL graft -signal intensity of quadriceps tendon) / signal intensity of background, with lower SNQ values indicating better graft maturity.The quadriceps tendon and background are located at the patellar upper limit level and at 2 cm anterior to the patellar tendon, respectively.
b The SNQ values are separately calculated for 6 segments of the reconstructed ACL as follows: the femoral intratunnel segment, intra-articular segment, and tibial intratunnel segment of the anteromedial bundle and posterolateral bundle.In each segment, 3 regions of interest (0.1-cm 2 circles) are selected, and the mean signal intensity is calculated.

c
The range of motion is measured by placing the axis of a goniometer over the lateral femoral epicondyle and aligning the 2 arms along the femoral and fibular axes, with positive values indicating knee flexion and negative values indicating hyperextension.

Figure 2 .
Figure 2. Mean Scores of the 4 Subscales of the Knee Injury and Osteoarthritis Outcome Score (KOOS 4 ) at Baseline and During 12-Month Follow-Up by Group 100

Table 1 .
Baseline Characteristics of Study Participants

Table 2 .
Patient-Reported Outcomes During Follow-Up Visits g The IKDC Subjective Knee Form includes 18 items covering 3 domains (symptoms, sports activities, and function); scores range from 0 (worst condition) to 100 (best condition).

Table 3 .
Graft Maturity on Magnetic Resonance Imaging During Follow-Up Visits A, Zarei Kurdkandi H, Ebrahimpour A, Biglari F, Jafari Kafiabadi M. Efficacy of platelet-rich plasma for chronic lateral ankle instability after modified Broström-Gould surgery: a randomized, single-blinded, prospective controlled trial.Foot Ankle Orthop.Published online April 26, 2023.doi:10.1177/24730114231168633Processing, Characteristics, and Delivery of Platelet-Rich Plasma eTable 2. Primary and Sensitivity Analyses of KOOS 4 During Follow-Up Visits eTable 3. Multivariable Linear Regression Analysis of KOOS 4 at 12 Months eTable 4. Physical Examinations During Follow-Up Visits (Primary Analysis) eTable 5. Patient-Reported Outcomes During Follow-Up Visits (Per-Protocol Analysis) eTable 6. Graft Maturity on MRI During Follow-Up Visits (Per-Protocol Analysis) eTable 7. Physical Examinations During Follow-Up Visits (Per-Protocol Analysis) eTable 8. Medications and Adverse Events During Follow-Up Visits