Platelet-Rich Plasma Injection for Painful Trapeziometacarpal Osteoarthritis

Abstract

Background:

Intra-articular injection of platelet-rich plasma (PRP) for osteoarthritis (OA) has increased in popularity recently. Evidence of the effect in hand OA is weak. The aim of this study was to evaluate if PRP injection decreases pain and disability due to trapeziometacarpal OA compared with placebo.

Methods:

Double-blind trial of 90 patients with painful trapeziometacarpal OA that were randomized to a single intra-articular injection of either PRP (n = 45) or placebo (saline injection) (n = 45) in the trapeziometacarpal joint. The primary outcome was change in pain on load (Numerical Rating Scale (NRS) 0 to 100) from baseline to 6 months after injection. Secondary outcomes were change in pain on load after 3 months; change in pain at rest; Patient-Rated Wrist and Hand Evaluation; Disability of the Arm, Hand, and Shoulder score; range of motion; key pinch; and 3-finger pinch strength, after 3 and 6 months.

Results:

One PRP patient was lost to follow-up. The mean age was 63 years, and 58 participants were women. Before intervention, median pain on load was 65 (interquartile range [IQR] 45-80) in the PRP group and 60 (IQR 40-80) in the placebo group. At 6 months, median pain change was −10 (95% CI −20 to 0) in the PRP group and 0 (95% CI 0-10) in the placebo group (absolute difference −10 (−20 to 0), p = 0.22). No clinically relevant differences were found in secondary outcomes.

Conclusion:

A single PRP injection has no short-term effect on pain on load compared with placebo for trapeziometacarpal osteoarthritis.

Level of evidence:

Level 1. See Instructions for Authors for a complete description of levels of evidence.

Introduction

Radiological trapeziometacarpal osteoarthritis (OA) is a common condition in both men and women, particularly postmenopausal women, with prevalence increasing with age¹. Clinical symptoms can occur at various radiological stages of the disease and may cause considerable pain, weakness, and disability². Nonoperative treatment typically starts with hand therapy, orthosis treatment, and analgesics, such as nonsteroidal anti-inflammatory drugs and paracetamol. In addition, nonoperative treatment commonly includes intra-articular corticosteroid injections. The efficiency of corticosteroid injections in trapeziometacarpal OA has been questioned³. A recent meta-analysis of randomized controlled trials (RCTs) on intra-articular injections for trapeziometacarpal osteoarthritis demonstrated a short-term benefit of corticosteroid injections⁴. However, due to the low level of supporting evidence, the European Alliance of Associations for Rheumatology and the American College of Rheumatology do not recommend corticosteroid injections for hand OA⁵.

Intra-articular platelet-rich plasma (PRP) injections for OA have gained popularity in the last decade. PRP is plasma with a high platelet count produced by centrifuging autologous blood, to separate the plasma from other blood components. PRP has a high concentration of platelets, contains growth factors and inflammatory modulators, and may potentially reduce inflammation, protect cartilage, and even promote cartilage repair^6,7. In preclinical trials, PRP injections have shown to reduce synovial inflammation, induce chondrocyte proliferation, and protect chondrocytes against interleukins in rats^8,9. PRP also increase the proliferation and reduce apoptosis of human chondrocytes in vitro¹⁰.

A meta-analysis reports effective pain relief and functional improvement following PRP injections for knee and ankle OA¹¹. Conversely, a RCT, not included in the meta-analysis, found no beneficial effect of PRP compared with placebo injections for early-stage knee OA¹².

There are limited data on the effectiveness of PRP for hand osteoarthritis, and previous studies have been small and results contradictory. Malahias et al. compared steroid injection with PRP in trapeziometacarpal joint in 33 patients and found a statistically significant pain-relieving effect in the PRP group after 12 months unlike the steroid group¹³. Hasley et al.¹⁴ report good longstanding effect of PRP injection in the trapeziometacarpal joint in a retrospective study of 19 patients, and Loibl et al.¹⁵ showed similar findings in a retrospective study of 10 patients. Sabaah et al.¹⁶, on the other hand, found no effect of PRP on pain when comparing PRP with hyaluronic acid and corticosteroid injections in an RCT of 45 patients. Accordingly, Winter et al.¹⁷ compared 24 patients treated with PRP with 21 treated with placebo and report no differences in pain or disability. Swärd and Wilcke¹⁸ found uncertain effect on pain scores in the short term in a retrospective study of 29 patients with OA in trapeziometacarpal or scaphoid-trapezium-trapezoidal (STT) joint. A recent meta-analysis concludes that there are promising results for PRP in trapeziometacarpal OA, but the quality of present evidence is low and larger methodologically robust trials are needed¹⁹.

The primary aim of this study was to compare the 6-month effect of a single intra-articular PRP injection to placebo on pain on load in patients with painful trapeziometacarpal OA. The secondary aims were to assess the effect of PRP on other patient-reported outcome measures (PROMs), thumb range of motion, and strength up to 6 months after injection.

Methods

Study Design

This is a double-blind, randomized, controlled trial conducted at the Department of Hand Surgery, Södersjukhuset, Stockholm. Participants were recruited from outpatients referred to this tertiary care center. Recruitment started on February 24, 2020, and follow-up was completed on June 5, 2024. The study was approved by the Swedish Ethical Review Authority (2019-05678) and registered in clinicaltrials.gov (NCT04218591). Written informed consent was obtained from all participants before inclusion in the study.

Participants

Inclusion criteria were clinical signs of trapeziometacarpal osteoarthritis: pain at palpation and provocation of the thumb base, and radiologically verified OA in the trapeziometacarpal joint, radiographic Eaton class 1 to 3²⁰.

Exclusion criteria were rheumatoid arthritis, psoriasis arthritis, gout or pseudogout affecting the hand, ongoing infection in the hand, inability to co-operate with the follow-up protocol (due to language difficulties, severe psychiatric disorder, cognitive impairment, drug addiction), intra-articular injection of corticosteroids in the affected joint within 6 months, and STT joint osteoarthritis (radiographical Eaton class 4).

Randomization and Masking

The patients were randomly assigned (1:1) to a single intra-articular injection with PRP or a single intra-articular injection with saline (placebo) in the trapeziometacarpal joint base. The inclusion process and injections were conducted by an unblinded assessor (hand surgeon). Randomization was conducted by a dedicated research nurse drawing the allocation from opaque envelopes in blocks of 10. All patients were blinded during the injection. Assessments at 3 and 6 months were performed by another blinded assessor (hand surgeon). Five hand surgeons were engaged in the study.

Procedures

A 15-ml venous blood sample was taken from all the patients, including the placebo group. The Arthrex ACP double-syringe system (Arthrex) was used. The blood was centrifuged for 5 minutes at 1,500 rpm. The centrifugation separates the platelets from red blood cells, yielding a leukocyte poor plasma²¹. Under fluoroscopic guidance, 0.8 ml of either PRP or saline (placebo) was injected. The participants were not provided with an orthosis, hand therapy, or pain medication after the injection but were permitted to use such therapeutic measures.

Patients were informed about their allocation 2 weeks after the 6-month follow-up by letter. Patients in the placebo group were offered a PRP injection, and all patients were offered alternative treatments as clinically indicated.

Role of the Funding Source

The study was financed by grants from the Swedish state under the agreement between the Swedish government and the county councils (the ALF-agreement). The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Outcomes

Primary Outcome

The primary outcome was change in pain on load (self-administered 11-point Numerical Rating Scale (NRS) 0-100) from baseline to 6 months after injection. The participants were asked to rate their pain over the previous week when performing tasks that required the use of force.

Secondary Outcomes

Secondary outcomes were:

Change from baseline to 3 months after injection in pain on load (NRS 0-100).

Change from baseline to 3 and 6 months in the following parameters:

Pain at rest (NRS 0-100), Patient-Rated Wrist and Hand (PRWHE) score²². PRWHE is a 15-item hand-specific and wrist-specific PROM quantifying pain and disability, ranging from 0 (no pain or disability) to 100 (worst pain and disability).

Disability of the Arm, Shoulder, and Hand score (DASH)²³. DASH is a 30-item score assessing symptoms and disability in the upper limbs and range from 0 (no disability) to 100 (most severe disability).

Patients were also asked a yes/no question regarding perceived improvement.

Objective physical variables were assessed according to the manual of the Swedish quality registry for hand surgery (HAKIR.se)²⁴ including radial and palmar abduction of the first metacarpal measured with a goniometer, opposition according to Kapandji (0-10), key pinch, and three-finger pinch grip (Saehan medical mechanical pinch gauges SH5010).

The Pain Catastrophizing Score (PCS)²⁵ was assessed at baseline. PCS is a 13-item tool resulting in scores from 0 to 52 with higher score indicating greater levels of pain catastrophizing.

Any adverse events (e.g. severe pain, swelling, or infection) were recorded.

Statistical Analysis

We tested the null hypothesis that there is no difference in change in NRS pain in load after 6 months between the PRP group and the placebo group. The reported minimal important clinical difference (MCID) for pain NRS for base of trapeziometacarpal OA has been reported to be 20 on a 100-point scale²⁶. To detect this difference between the groups, presuming a SD of 30², 37 patients in each group were required (power 80%, p < 0.05). To account for nonparametric outcomes and loss to follow-up, we aimed to include 90 participants.

For description of baseline data, nonparametric data (Pain NRS, PRWHE, DASH, thumb opposition) are presented as median (interquartile range [IQR]). According to the Shapiro-Wilks test, none of the continuous objective physical parameters were normally distributed. Hence, they are also presented as median (IQR) and analyzed as nonparametric data.

For comparison of change in all outcome variables between baseline (before injection) and 3 or 6 months later between PRP and placebo, the Mann-Whitney U-test signed rank sum tests was used. Changes in outcomes and estimates are presented as median (95% CI). For the binary question of perceived improvement, χ² test was used for comparison between groups.

Generalized estimating equations (GEE) were used to analyze the repeated measurements. We used a robust estimator covariance matrix, an independent working correlation matrix (according to the lowest Quasi-likelihood under Independence Model Criterion), and a linear model for all variables. Sex, age, if the dominant hand was treated, Eaton class, PCS, and treatment*time were included as possible predictors in the model. The changes in outcome variables are presented as beta coefficients (β) with p-values. β is the expected population average change of the outcome variable between the baseline and the assessments at 3 and 6 months.

MCID for PRWHE has been approximated to minimum 10 points for trapeziometacarpal OA²⁶. For DASH, MCID has not been estimated specifically for trapeziometacarpal OA but for upper-limb musculoskeletal disorders in general, a MCID of 11 points has been suggested²⁷.

SPSS versions 28 and 29 and R were used.

Results

Between February 24, 2020, and December 5, 2023, we assessed 222 individuals for eligibility and recruited 90 participants (45 PRP and 45 placebo) (Fig. 1). One patient in the PRP group withdrew from the study 1 week after injection due to severe pain, leaving 44 participants in that group.

Fig. 1.

Trial profile.

Demographic data are summarized in Table I. The groups were similar in terms of sex, age, occupational status, radiographic Eaton classification, and PCS score. However, a higher proportion of participants in the placebo group were treated in their dominant hand. The outcome parameters at baseline are presented in Table II.

TABLE I.

Baseline Demographic and Clinical Characteristics

	PRP Group (n = 45)	Placebo Group (n = 45)	Total (n = 90)
Age, years mean (range)	63 (44-87)	62 (38-87)	63 (38-87)
Sex (female/male)	28/17	30/15	58/32
Dominant hand treated (yes/no)	17/28	27/18	45/45
Employment status
Manual labor	16	17	33
Nonmanual labor	10	9	19
Retired/not employed	19	19	38
Eaton class (1/2/3)	2/17/26	4/17/24	6/34/50
Pain catastrophizing score at inclusion (median, IQR)	14 (5-21)	15 (8-26)	15 (5-22)

TABLE II.

Baseline Values of Outcome Measures (Median [IQR])

	PRP Group (n = 44)	Placebo Group (n = 45)	Total (n = 89)
Pain on Load (NRS 0-100)	65 (45-80)	60 (40-80)	60 (30-80)
Pain at rest (NRS 0-100)	20 (10-55)	30 (19-50)	30 (10-50)
PRWHE (0-100)	51 (40-67)	53 (37-66)	51 (37-66)
DASH (0-100)	34 (21-47)	28 (15-40)	29 (17-45)
Radial abduction (°)	35 (35-40)	40 (35-45)	40 (35-45)
Palmar abduction (°)	40 (35-40)	40 (35-45)	40 (35-42)
Opposition (0-10)	10 (9-10)	10 (9-10)	10 (9-10)
Key pinch (kg)	6.1 (4.3-8.6)	5.8 (4.5-7.2)	6.0 (4.5-7.8)
Three-finger pinch (kg)	5.8 (4.0-7.9)	5.7 (4.0-7.5)	5.7 (4.0-7.5)

DASH = Disability of the Arm, Shoulder, and Hand Score, IQR = interquartile range, NRS = Numerical Rating Scale, and PRWHE = Patient-rated Wrist and Hand Score.

No difference was observed between the groups for the primary outcome (change of pain on load) at 6 months (Fig. 2, Table III). Similarly, no differences were found in change of pain on load at 3 months, pain at rest, or PRWHE at either 3 or 6 months. The PRP group showed significantly greater improvement in DASH scores after 3 and 6 months, but the difference did not reach MCID (Table III).

Fig. 2.

Box plot showing median pain on load (0-100) (IQR) before injection and after 3 and 6 months. Platelet-rich plasma injection for painful trapeziometacarpal osteoarthritis—a double-blind, randomized, placebo-controlled trial. IQR = interquartile range.

TABLE III.

Change in Patient-Rated Outcome Measures (Median 95% CI)

	PRP (n = 44)	Placebo (n = 45)	Absolute Difference	P
Primary outcome 6 months after injection
Change in pain on load (NRS 0-100)	−10 (−20 to 0)	0 (0 to 10)	−10 (−20 to 0)	0.217
Missing data	0	3
Secondary outcomes 3 months after injection
Pain on load (NRS 0-100)	0 (0 to 10)	0 (0 to 10)	0 (−10 to 10)	0.826
Pain at rest (NRS 0-100)	−10 (−10 to 0)	−10 (−10 to 0)	0 (−10 to 10)	0.610
PRWHE	−7 (−12 to −1)	−8 (−13 to 1)	−4 (−10 to 5)	0.425
DASH	−5 (−13 to −3)	−1 (−3 to 3)	−7 (−12 to −2)	0.005
Missing data	1	1
Secondary outcomes 6 months after injection
Pain at rest (NRS 0-100)	0 (0 to 10)	0 (0 to 10)	0 (−10 to 10)	0.912
PRWHE	−6 (−17 to 1)	−7 (−13 to −4)	0 (−7 to 8)	0.976
DASH	−8 (−13 to −4)	-1 (−6 to 3)	−7 (−12 to −1)	0.018
Missing data	0	3

DASH = Disability of the Arm, Shoulder, and Hand Score, NRS = Numerical Rating Scale, and PRWHE = Patient-rated Wrist and Hand score.

A minus sign (−) indicates a decrease in the score.

There were no differences in change of radial or palmar abduction of the first metacarpal, key pinch, three-finger pinch, or thumb opposition between the groups at any time (Table IV).

TABLE IV.

Change in Objective Physical Parameters (Median 95% CI)

	PRP (n = 44)	Placebo (n = 45)	Absolute Difference (95% CI)	p
3 months after injection
Radial abduction (°)	0.0 (0.0 to 5.0)	0.0 (0.0 to 5.0)	0.0 (−5.0 to 0.0)	0.311
Palmar abduction (°)	0.0 (0.0 to 5.0)	0.0 (0.0 to 5.0)	0.0 (0.0 to 5.0)	0.390
Opposition (0-10)	0.0 (0.0 to 1.0)	0.0 (0.0 to 1.0)	0.0 (0.0 to 0.0)	0.503
Key pinch (kg)	0.0 (−0.3 to 0.7)	0.0 (−0.1 to 0.4)	−0.1 (−0.6 to 0.4)	0.640
Three-finger pinch (kg)	0.2 (−0.8 to 0.7)	−0.2 (−0.7 to 0.1)	−0.3 (−1.0 to 0.3)	0.252
6 months after injection
Radial abduction (°)	1.0 (0.0 to 5.0)	0.0 (0.0 to 5.0)	−2.0 (−5.0 to 0.0)	0.071
Palmar abduction (°)	0.0 (0.0 to 5.0)	0.0 (0.0 to 5.0)	0.0 (0.0 to 5.0)	0.666
Opposition (0-10)	0.0 (0.0 to 1.0)	0.0 (0.0 to 1.0)	0.0 (0.0 to 0.0)	0.145
Key pinch (kg)	0.0 (-0.5 to 0.7)	−0.2 (−0.3 to 0.5)	−0.2 (−0.3 to 0.5)	0.542
Three-finger pinch (kg)	0.2 (−0.6 to 0.3)	−0.4 (−0.7 to 0.3)	−0.3 (−0.3 to 0.8)	0.329
Missing data (all parameters)	0	3

Regarding the binary question on perceived improvement, 18 patients in the PRP group and 16 in the placebo group answered “yes” at 3 months, while at 6 months, 17 patients in the PRP group and 15 in the control group answered “yes” (p = 0.40 and 0.57, respectively).

The GEE analysis indicated that PRP did not improve any outcome parameter that were not also improved in the placebo group during the follow-up, and no beta coefficient exceeded MCID. The beta coefficients for changes in PROMs are presented in Table V.

TABLE V.

The Effect of PRP and Placebo on Patient-Reported Outcome Measures

Beta Coefficients for Change in Outcomes (95% CI)
	PRP		Placebo
	Before Injection to 3 Months	Before Injection to 6 Months	Before Injection to 3 Months	Before Injection to 6 Months
Pain on load NRS	−3 (−13 to 7)	−11 (−21 to −1)	−6 (−12 to 1)	−6 (−13 to −0.5)
Pain at rest NRS	−8 (−16 to 1)	−8 (−17 to 1)	−6 (−10 to −1)	−6 (−11 to −0.3)
PRWHE	−7 (−15 to 1)	−8 (−16 to 0.3)	−6 (−11 to −2)	−8 (−11 to −4)
DASH	−2 (−7 to 4)	−3 (−9 to 3)	−1 (−4 to 2)	−2 (−6 to 2)

DASH = Disability of the Arm, Shoulder, and Hand score (0-100), NRS = Numerical Rating Scale (0-100), and PRWHE = Patient-rated Wrist and Hand score (0-100).

Beta coefficients: Expected population average change of the outcome variable between the assessments with placebo before injection as the reference. Adjusted for sex, age, if the dominant hand was treated, Eaton score, and Pain Catastrophizing Score.

Adverse events were the following: 1 severe pain after injection (PRP group), 1 transient numbness dorsally of the thumb base (PRP group), and 2 transient swelling around the thumb base (1 in each group).

Discussion

This double-blinded randomized study indicates that a single intra-articular PRP injection has no short-term effect on pain on load compared with placebo for trapeziometacarpal osteoarthritis. We found no difference in the primary outcome, pain on load, between PRP and placebo after 6 months.

PRP showed no beneficial effect on any secondary outcome except for the three-month and six-month DASH scores that indicated a statistically significant difference in favor for PRP. However, the absolute difference between the groups (-7 points at both assessments) did not reach the estimated MCID for DASH (11 points)²⁷. Furthermore, no effect on DASH was found in the GEE analysis which disagrees with a true effect of PRP. The PRP group had higher baseline DASH score than the placebo group, and the observed difference could possibly be explained by a larger regression to the mean in this group.

Previous studies reporting a beneficial effect of PRP in trapeziometacarpal OA all have methodological flaws that may lead to potentially misleading results. Malahias et al.¹³ compared PRP (n = 16) to corticosteroid injection (n = 17) in an open-label RCT. There were no significant differences in pain or DASH scores after 3 months, but after 12 months, the PRP group reported less pain and disability. Small retrospective case series^14,15 have reported improvements following PRP injections; however, measuring postintervention changes without a control group often suggests nonexisting benefits, due to factors such as regression to the mean and contextual effects²⁸.

In accordance with our results, a placebo-controlled study of a single PRP injection by Winter et al.¹⁷, a small comparison to hyaluronic acid and corticosteroid injections by Sabaah et al.¹⁶, and a case series by Swärd & Wilcke¹⁸ found no pain-relieving effect of PRP.

If the reported positive effects of PRP for painful osteoarthritis in knee and ankle¹¹ are true, the question arises of why PRP does not have a certain effect for trapeziometacarpal OA. A limitation of this study is that we did not measure platelet concentration leaving uncertainty if the platelet dose was sufficient. It has been suggested that higher platelet concentration might improve effectiveness of PRP in knee osteoarthritis²⁹. Furthermore, we evaluated the effect of a single injection which could be insufficient. In knee OA, a positive effect on pain has been reported after a single injection, while 2 or 3 injections resulted in a longer lasting effect³⁰. In guinea pigs, anti-inflammatory effects on the synovium in the short term have been reported after both single and multiple injections, while multiple injections also had a chondroprotective effect³¹. In our opinion, it is not likely that multiple PRP injections for trapeziometacarpal OA would have effect when a single injection had no effect on pain or disability after 3 and 6 months. However, considering reports on the beneficial effects on other joints, we cannot rule out that multiple injections or higher platelet concentrations could have a positive effect on trapeziometacarpal OA. Also, the possibility remains that PRP could exhibit effects beyond 6 months and the relatively short follow-up is a limitation of this study.

We included participants with a broad spectrum of radiographic osteoarthritis severity, which may be subject to critique. The distribution of radiographic Eaton classes was balanced between groups and has probably not biased the results. However, if PRP hypothetically is effective in the early stages of osteoarthritis, restricting inclusion to Eaton class 1 and 2 might have been preferable. On the other hand, radiological severity has not shown to correlate with perceived pain in trapeziometacarpal OA³² and previous studies reporting pain reduction after PRP have included patients with Eaton class 3^13-15. An implication for further research may be to evaluate PRP efficacy across specific stages of trapeziometacarpal osteoarthritis. Still, the challenge remains to stage trapeziometacarpal OA appropriately. By speculation, PRP may be effective in the early stage of OA even before radiographs signs occur or in osteoarthritic joints with more synovitis. Primary OA has generally been considered as noninflammatory, characterized by joint cartilage degeneration and changes in the bone. However, recent research has increasingly recognized the role of inflammation and synovitis in the pathogenesis and progression of OA³³. Ultrasound is reported to be effective in detecting and quantifying synovial abnormalities, which are present even before visible cartilage degeneration occurs³⁴. We did not use ultrasound to assess inflammation, as ultrasound is not routinely used in the clinical management of trapeziometacarpal OA.

The strengths of this study are the pragmatic randomized controlled trial format, the relatively large number of patients included, and the double-blind design. The study population is representative for patients with trapeziometacarpal OA in terms of sex and age¹, and we believe that the findings are generalizable to patients with primary trapeziometacarpal OA.

In addition to previous mentioned limitations, we did not assess if the blinding of participants and assessors was effective. However, we have no reason to believe that the blinding was ineffective and affected the results. Another weakness is that we did not record concomitant interventions such as use of orthoses or pain medication that could have influenced the outcomes.

Conclusion

In patients with painful trapeziometacarpal OA, a single injection PRP does not relieve pain on load in the short term and, hence, has no evident value in the treatment.