A Comparative Analysis of Surgical Treatment for First Carpometacarpal Osteoarthritis: A Systematic Review and Pooled Analysis of Randomized Controlled Trials

Abstract

Background:

First carpometacarpal osteoarthritis (CMC1-OA) is a common degenerative condition causing pain, functional loss, and reduced grip strength. While trapeziectomy remains the standard surgical treatment, reconstructive procedures and implant arthroplasties are increasingly used. However, differences in postoperative recovery between these approaches remain unclear. This study provides an indirect pooled comparison of reconstructive, non-reconstructive, and dual-mobility implant procedures for CMC1-OA.

Methods:

Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, randomized controlled trials (RCTs) evaluating surgical treatment of CMC1-OA were included. Procedures were categorized as reconstructive, non-reconstructive, or dual-mobility implants. Outcomes included visual analog scale (VAS), Disabilities of the Arm, Shoulder, and Hand, hand grip, tip pinch, lateral pinch, and complications at 3- and 12-month follow-up.

Results:

Seven RCTs comprising 409 hands were included. At 3 months, dual-mobility implants demonstrated the greatest reduction in VAS pain scores (Δ = 6.05), compared with reconstructive (Δ = 3.52) and non-reconstructive procedures (Δ = 2.86). Disabilities of the Arm, Shoulder, and Hand scores improved across all groups, most prominently in implant procedures (Δ = 29.48 at 3 months; Δ = 35.35 at 12 months). Strength outcomes varied between groups, though implants demonstrated the greatest improvement in lateral pinch strength at 12 months. Complication rates differed between reconstructive (20.61%), implant (24.49%), and non-reconstructive procedures (51.61%), though only 1 study reported complications in the non-reconstructive group.

Conclusion:

Although 12-month pain and functional outcomes were comparable, dual-mobility implants demonstrated greater early improvements in pain, function, and lateral pinch strength. These findings suggest that differences may exist between surgical approaches for CMC1-OA.

Introduction

First carpometacarpal osteoarthritis (CMC1-OA) is a prevalent and progressively debilitating condition that affects nearly one-third of adults over the age of 55, with a higher prevalence among postmenopausal women.^1,2 The unique saddle-shaped anatomy of the first carpometacarpal joint, combined with its repeated use in grip and pinch activities, makes it particularly susceptible to degenerative changes. Patients often report chronic pain, reduced manual dexterity, and loss of functional independence as a result of symptom progression.^3,4 These impairments can have a significant impact on activities of daily living and productivity, especially in patients who rely on fine motor function for work or caregiving. In addition, this condition leads to about 7% rate of work absenteeism, with a median loss of 42 work hours, which can result in indirect costs further compounding its economic burden and making effective treatment a matter of clinical and societal importance.^{5 -8}

First carpometacarpal osteoarthritis can be treated with several surgical approaches, broadly including non-reconstructive procedures, reconstructive procedures, and implant-based interventions. Non-reconstructive procedures, such as simple or partial trapeziectomy, primarily aim to relieve pain through excision of the arthritic joint surface without restoration of ligamentous support.^{9 -11} In contrast, reconstructive procedures, including ligament reconstruction with tendon interposition (LRTI) and suspensionplasty techniques, seek to restore thumb stability and function following trapeziectomy through reconstruction or augmentation of the volar beak ligament.^10,12,13 Implant-based procedures, often utilizing silicone, pyrocarbon, or metal devices, represent an evolving treatment category designed to preserve thumb column height and provide immediate mechanical stability, although concerns remain regarding implant longevity and revision risk.^6,14,15

Although all 3 procedural categories have demonstrated postoperative pain relief and functional improvement, comparative studies evaluating clinical outcomes among these approaches have reported mixed findings. ⁹ ,^{16 -18} Several randomized controlled trials (RCTs) have demonstrated improved early Disabilities of the Arm, Shoulder, and Hand (DASH) outcomes following reconstructive procedures compared with alternative surgical approaches.^7,19,20 However, the long-term clinical significance of these differences remains uncertain, particularly given concerns regarding complication and revision rates associated with implant-based interventions.^11,16,21

With limited studies, hand surgeons often rely on their training background or institutional setup when choosing between soft-tissue reconstruction, non-reconstructive techniques, and implants. ²² Therefore, we conducted a systematic review and pooled analysis using RCTs to compare postoperative outcomes for pain, disability, function, and complications across these 3 distinct surgical strategies for CMC1-OA. We hypothesize that reconstructive, non-reconstructive, and dual-mobility implant procedures will yield distinct short-term and mid-term clinical outcomes at 3- and 12-month follow-up.

Methods

Search Strategy and Data Extraction

This study is a systematic review using PubMed, Cochrane, Web of Science, Embase, and Scopus, in which articles were retrieved from inception until February 24, 2025. The search algorithm used was: (“CMC arthroplasty” OR “thumb arthroplasty” OR “CMC joint replacement” OR “CMC fusion” OR “CMC arthrodesis” OR “LRTI” OR “trapeziectomy” OR “carpometacarpal arthroplasty” OR “carpometacarpal fusion” OR “carpometacarpal arthrodesis”) AND (“long-term outcomes” OR “functional outcomes” OR “patient-reported outcomes” OR “PROM” OR “AROM” OR “DASH score” OR “VAS score” OR “Kapandji score” OR “grip strength” OR “pinch strength” OR “key pinch” OR “range of motion” OR “pain relief” OR “hand function”) AND (“CMC joint” OR “carpometacarpal joint”). This systematic review is in accordance with the most recent Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. ²³

Inclusion/Exclusion Criteria

The inclusion criteria for this study were RCTs of adult patients diagnosed with CMC1-OA, confirmed by radiograph and classified as Eaton criteria stage I, II, III, or IV OA, treated surgically by any procedure involving trapeziectomies, with outcomes including the visual analog scale (VAS) for pain, DASH to assess function, and hand/tip pinch/lateral pinch strength (all measured in kg). All studies must have measured outcomes at both 3 and 12 months. Exclusion criteria included cadaveric or animal studies, case reports, systematic reviews or meta-analyses, book chapters, editorials, non-English publications, and studies lacking extractable quantitative data. Of the 14 full-text articles assessed for eligibility, 7 were excluded because they did not report the prespecified postoperative follow-up intervals (n = 4), did not report one or more outcomes of interest, including pain or functional measures (n = 2), or lacked sufficient quantitative data for extraction despite meeting other eligibility criteria (n = 1). The study selection process is summarized in the PRISMA 2020 flow diagram (Figure 1).

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart 2020 of included studies.

*Consider, if feasible, reporting the number of records identified from each database or register searched rather than the total number across all databases/registers.

**If automation tools were used, indicate how many records were excluded by a human and how many were excluded by automation tools.

Data Extraction

Two independent authors (NM and JT) screened the completed initial title/abstract screening and then full-article selection using Rayyan.ai, an online, AI-assisted systematic review platform designed for article screening and selection. ²⁴ A third independent researcher resolved any conflicts (JL). Four independent researchers (JL, NM, JT, and ZC) performed the data extraction using a standardized data collection form. Data extracted included article information (first author and year of publication), patient demographics (age and sex), type of surgical intervention, outcomes, and complication rate at 3- and 12-month follow-up, including VAS, DASH, and hand/tip pinch/lateral pinch grip strength scores in kilograms. Summary of included studies are reflected in Table 1.

Table 1.

Summary of Included Studies.

Author	Country	Study design	Inclusion criteria	Interventions	Rehabilitation protocol	N hands	Notes
Guzzini et al 18	Italy	Prospective RCT	Age >18 with pain at the base of the thumb at rest or with activity, unresponsive to ≥ 3 months of conservative treatment (splinting, NSAIDs, or corticosteroid injections), and radiographic evidence of TMC osteoarthritis (Eaton-Littler stage II-early III).	1. Trapeziectomy with tendon interposition 2. Total joint arthroplasty using Touch TMC joint dual-mobility press-fit prosthesis	Not explicitly reported; standard postoperative hand therapy with serial follow-up at 1, 3, 6, 12, and 24 months.	n total = 139 Group A (n = 71) Group B (n = 68)	Group A = tendon interposition arthroplasty with trapeziectomy; Group B = total joint arthroplasty
Zarezadeh et al 25	Iran	Prospective single-blind RCT	Age >18, first CMC joint OA confirmed by clinical examination and wrist/thumb x-rays showing degenerative changes, unresponsive to medical treatment, and provided informed consent.	1. LRTI 2. HDA	Thumb spica splint with K-wire immobilization for 4 weeks, followed by ROM after K-wire removal; strengthening at 6 weeks and pinch/grip exercises by 8 weeks.	n total = 59 LRTI (n = 30) HDA (n = 29)	LRTI used autologous tendon; HDA involved hematoma formation and distraction without graft or implant.
De Jong et al 26	Netherlands	Double-blind RCT	Women ≥40 years old with symptomatic TMC osteoarthritis, Eaton-Glickel stage II or III, confirmed by clinical and radiological evaluation by a level 4 hand surgeon.	1. Simple dorsal trapeziectomy 2. Uncemented, semi-constrained total joint replacement using Maïa prosthesis	Forearm/thumb cast for 4 weeks (1 week bandage + 3 weeks cast), followed by supervised hand therapy weekly (weeks 5-8) and biweekly (weeks 9-12) using a standardized rehabilitation protocol.	n total = 62 Trapeziectomy (n = 31) Total joint arthroplasty (n = 31)	Trapeziectomy was uncemented; prosthesis was semi-constrained Maia implant placed without bone cement.
Vermeulen et al 27	Netherlands	Single-center, double-blind RCT	Women ≥40 years old with CMC I osteoarthritis (Eaton stage II-III), failed conservative treatment, and persistent symptoms prompting surgical intervention.	1. Trapeziectomy with ligament reconstruction and tendon interposition	Thumb spica cast for approximately 4 weeks, then supervised hand therapy with progressive ROM and strengthening as tolerated.	n total = 43 Arthrodesis (n = 22) LRTI (n = 21)	Trapeziometacarpal arthrodesis treatment was excluded from analysis
Marks et al 15	Switzerland	Single-center, parallel-group RCT	Age >18 with CMC I osteoarthritis (Eaton stage ≥II), failed conservative treatment (orthosis and/or ≤3 steroid injections), and persistent symptoms prompting surgical request.	1. FCR autograft LRTI: trapeziectomy with suspension-interposition using a modified Weilby technique 2. Allograft LRTI: trapeziectomy with suspension-interposition using a GraftJacket allograft sutured to the preserved FCR tendon	Removable orthosis for 6 weeks; active ROM began at 3 weeks; strengthening at 8 weeks; unrestricted daily activities thereafter.	n total = 60 FCR (n = 29) Allograft (n = 31)	Compared FCR autograft vs dermal allograft (GraftJacket); same suspension-interposition technique used.
Graesser et al 28	United States	Single-center, 2-arm, prospective RCT	Age >18 with thumb CMC osteoarthritis (Eaton stage ≥II), failed non-surgical management (orthosis and/or ≤3 steroid injections), and electing surgery due to persistent symptoms.	1. Trapeziectomy with ligament reconstruction and tendon interposition 2. Trapeziectomy with suture-based suspension arthroplasty (no tendon harvest)	Standardized postoperative therapy protocol for both groups; early removable splint and supervised hand therapy. Exact timeline not reported.	n total = 32 LRTI (n = 15) STS (n = 17)	LRTI used full FCR harvest; STS performed using suture-based suspension without tendon harvest.
Picchi et al 29	Italy	Prospective RCT	Age 40-75 years; Eaton-Littler stage III trapeziometacarpal OA; failure of conservative management (injections, therapy, orthosis); isolated TMJ OA with intact STT joint; no prior TMJ surgery; no inflammatory arthritis, infection, metabolic bone disease, or post-traumatic OA	1. Suspension tenoplasty (AST) using flexor carpi radialis tendon 2. Dual-mobility total joint arthroplasty using Maïa prosthesis	Cast immobilization for 3 weeks, followed by supervised rehabilitation with progressive ROM and strengthening.	n total = 36 AST (n = 18) Total joint arthroplasty (n = 18)	Arthroplasty preserved joint height and yielded greater grip strength; AST showed mild proximal migration.

Arthrodesis arms from Vermeulen et al were excluded from pooled analysis because they did not fit predefined intervention categories. Implant arthroplasty procedures included contemporary dual-mobility or semi-constrained prosthetic systems.

Abbreviations: AST, abductor suspension tenoplasty; CMC, carpometacarpal; FCR, flexor carpi radialis; HDA, hematoma distraction arthroplasty; LRTI, ligament reconstruction and tendon interposition; OA, osteoarthritis; RCT, randomized controlled trial; STS, suture-based suspension arthroplasty; TMC, trapeziometacarpal; NSAIDs, nonsteroidal anti-inflammatory drugs; ROM, range of motion; STT, scaphotrapeziotrapezoid joint; TMJ, trapeziometacarpal joint.

Categorization of Surgical Groups by Intervention

Procedures were categorized into soft-tissue reconstruction, non-reconstructive, and dual-mobility implant/prosthesis-based techniques, outlined in Table 2. Soft-tissue reconstruction included trapeziectomy combined with tendon interposition arthroplasty, LRTI, suspension-interposition methods (eg, modified Weilby), allograft-based LRTI, and suture-based suspension arthroplasty without tendon harvest. These procedures generally involved complete trapeziectomy followed by restoration of first metacarpal stability using autologous tendon, allograft, or suture constructs to suspend the metacarpal and occupy the trapezial space.

Table 2.

Categorization of Surgeries by Variation.

Types	Included variations
Non-reconstructive	Hematoma distraction arthroplasty; simple dorsal trapeziectomy.
Soft-tissue reconstruction	Trapeziectomy with tendon interposition arthroplasty; LRTI; trapeziectomy with ligament reconstruction and tendon interposition; FCR autograft LRTI; trapeziectomy with suspension-interposition (modified Weilby technique); allograft LRTI using allograft tissue sutured to the preserved FCR tendon; trapeziectomy with ligament reconstruction and tendon interposition; trapeziectomy with suture-based suspension arthroplasty (no tendon harvest).
Dual-mobility implants	Total joint arthroplasty includes devices such as the Touch TMC Dual-Mobility Press-Fit Prosthesis (such as the TOUCH Dual-Mobility CMC 1 Prosthesis [KeriMedical SA, Plan-les-Ouates, Switzerland]) and uncemented, semi-constrained total joint replacement systems such as the Maia prosthesis (Groupe Lépine, Genay, France).

Abbreviations: CMC, carpometacarpal; FCR, flexor carpi radialis; LRTI, ligament reconstruction with tendon interposition; TMC, trapeziometacarpal.

The non-reconstructive category consisted of hematoma distraction arthroplasty and simple trapeziectomy, which rely on trapezial excision alone or temporary K-wire distraction without ligament reconstruction or tendon interposition, allowing a hematoma to function as a biologic spacer during healing. Dual-mobility implant/prosthesis-based techniques comprised total joint arthroplasty using contemporary devices such as the Touch TMC (KeriMedical SA, Plan-les-Ouates, Switzerland) and Maïa prostheses (Groupe Lépine, Genay, France). These were defined as total joint replacements employing a dual-mobility/partially constrained ball-and-socket design that provides intrinsic stability without the need for tendon-based reconstruction.

Postoperative rehabilitation protocols were broadly comparable across studies. Immobilization with a thumb spica cast or removable orthosis was maintained for approximately 3 to 6 weeks (most commonly 4 weeks), followed by supervised hand therapy emphasizing progressive range-of-motion exercises. Strengthening exercises were generally initiated between 6 and 8 weeks postoperatively. Within individual randomized trials, rehabilitation protocols were typically standardized across treatment groups, although the duration of immobilization and specific therapy milestones varied modestly between studies.

Certainty of Evidence and Bias Assessment

The included articles were assessed for certainty of evidence using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria by 2 independent authors (JT and NM). ²⁰ For risk-of-bias assessment, RoB-2 was used by 2 independent authors (JL and ZC) to evaluate the included articles. ²¹ Any conflicts were resolved by NK.

Statistical Analysis

A pooled weighted mean analysis was performed using IBM SPSS Statistics (version 31.0.0.0) to summarize study-level outcomes across RCTs. A formal random-effects meta-analysis was omitted due to high heterogeneity and the limited number of studies. Weighted pooled means and standard deviations were calculated for each surgical category (reconstructive, non-reconstructive, and implants) at 3- and 12-month follow-up intervals. Change-from-baseline values (Δ) were calculated as weighted mean differences using study-level preoperative and postoperative means and standard deviations, allowing the pooled delta to reflect the proportional contribution of each study. Changes in dispersion (ΔSD) were also computed to evaluate variability across studies.

Continuous outcomes such as pain (VAS), disability (DASH), and grip strength measures were summarized using descriptive statistics. Complication data were extracted for all intervention groups at the longest common follow-up (12 months when available) and analyzed descriptively. Reported complications were categorized into soft-tissue complications (eg, tendon irritation or rupture, superficial wound infection, delayed wound healing), neurologic complications (eg, complex regional pain syndrome or transient sensory nerve symptoms), structural complications (eg, metacarpal subsidence/collapse), and implant-related complications (eg, prosthetic dislocation, loosening, or revision surgery). When a study reported no complications, this was recorded as zero events; when complication data were not reported, they were treated as missing and excluded from quantitative comparisons. Between-group differences in complication rates at 1 year were evaluated using χ² test, with statistical significance set at P < .05.

Results

Article Selection

Seven RCTs were included in this systematic review from the 685 articles initially retrieved from the database search. After the inclusion and exclusion criteria, 14 studies were assessed for eligibility, of which 7 were further excluded due to missing time points, outcomes, or data. In total, 7 RCTs (or 409 hands) were included in the review. Figure 1 depicts the search process for this systematic review.

All outcomes were assessed at 3- and 12-month postoperative follow-up. Subgroup pooled analyses were conducted for 3 intervention categories: (1) reconstructive procedures (LRTI or similar soft-tissue techniques); (2) non-reconstructive procedures (simple trapeziectomies alone or hematoma distraction arthroplasty); and (3) dual-mobility arthroplasties (Table 2).

Pain and Functional Outcomes at 3 and 12 months

At 3 months, the pooled weighted means demonstrated substantial reductions in pain scores across all surgical groups based on the change in VAS. Reconstructive procedures improved from 7.37 ± 1.54 preoperatively to 3.84 ± 1.27 postoperatively (Δ = 3.52), non-reconstructive procedures improved from 7.82 ± 1.67 to 4.96 ± 1.24 (Δ = 2.86), and dual-mobility implants demonstrated the greatest absolute early reduction from 7.00 ± 1.69 to 0.95 ± 0.93 (Δ = 6.05). Variability was lowest in the implant group at 3 months (SD = 0.93), compared with reconstructive (SD = 1.27) and non-reconstructive (SD = 1.24) groups. By 12 months, pain continued to improve across all groups from the preoperative baseline: reconstructive procedures decreased to 1.39 ± 1.14 (Δ = 5.98), non-reconstructive to 1.28 ± 1.18 (Δ = 6.54), and implant to 0.73 ± 0.87 (Δ = 6.27). Variability remained in reconstructive, non-reconstructive, and implant groups (SD = 1.14, 1.18, and 0.87, respectively) (see Table 3).

Table 3.

Pain and Function Pooled Weighted Means and Standard Deviations.

Category	Baseline mean (SD)	3-month mean (SD)	Δ 3-month mean	12-month mean (SD)	Δ 12-month mean
VAS
Implant	7.00 (1.69)	0.95 (0.93)	−6.05	0.73 (0.87)	−6.27
Non-reconstructive	7.82 (1.67)	4.96 (1.24)	−2.86	1.28 (1.18)	−6.54
Reconstructive	7.37 (1.54)	3.84 (1.27)	−3.52	1.39 (1.14)	−5.98
Total	7.30 (1.60)	3.03 (1.15)	−4.27	1.17 (1.06)	−6.13
DASH
Implant	51.76 (6.30)	22.28 (8.73)	−29.48	16.41 (5.53)	−35.35
Non-reconstructive	39.02 (4.91)	30.46 (3.53)	−8.56	17.96 (4.28)	−21.06
Reconstructive	47.98 (10.57)	32.47 (11.81)	−15.51	16.66 (9.89)	−31.32
Total	47.29 (8.18)	29.08 (9.25)	−18.21	16.77 (7.55)	−30.52

Values are reported as pooled weighted means with standard deviations. Δ values represent mean change from baseline to postoperative follow-up. Negative Δ values for VAS and DASH indicate improvement in pain and disability scores, respectively. Total values represent pooled weighted means across all intervention categories.

Abbreviations: DASH, Disabilities of the Arm, Shoulder, and Hand; VAS, visual analog scale.

For DASH scores, reconstructive procedures improved from 47.98 ± 10.57 to 32.47 ± 11.81 at 3 months (Δ = 15.51), non-reconstructive from 39.02 ± 4.91 to 30.46 ± 3.53 (Δ = 8.56), and implants from 51.76 ± 6.30 to 22.28 ± 8.73 (Δ = 29.48). At 12 months, reconstructive procedures improved to 16.66 ± 9.89 (Δ = 31.32), non-reconstructive to 17.96 ± 4.28 (Δ = 21.06), and implant to 16.41 ± 5.53 (Δ = 32.75). Variability was consistently lowest in the non-reconstructive group (SD = 4.91 preoperatively to 4.28 at 12 months) and remained wider in reconstructive (SD = 9.89) and implant (SD = 5.53) groups, reflecting more heterogeneous outcomes in these groups (see Table 3).

Hand, Tip Pinch, and Lateral Pinch Strength Outcomes at 3 and 12 Months

Grip strength was analyzed separately for hand, tip pinch, and lateral pinch grip. At baseline, weighted mean hand grip strength was 20.61 ± 7.76 kg for reconstructive procedures, 16.10 ± 6.70 kg for non-reconstructive procedures, and 21.32 ± 6.03 kg for implants. At 3 months, hand grip strength decreased to 18.90 ± 6.96 kg (Δ = −1.71 kg) in the reconstructive group and 12.20 ± 5.70 kg (Δ = −3.90 kg) in the non-reconstructive group, while remaining relatively unchanged in the implant group (21.48 ± 6.21 kg, Δ = +0.16 kg). By 12 months, all groups exceeded baseline strength levels. Reconstructive procedures reached 24.18 ± 7.94 kg (Δ = +3.57 kg), non-reconstructive procedures reached 20.10 ± 7.40 kg (Δ = +4.00 kg), and implants reached 27.24 ± 6.72 kg (Δ = +5.92 kg). The pooled total improved from 20.50 ± 7.03 kg at baseline to 24.97 ± 7.44 kg at 12 months (Δ = +4.47 kg).

Baseline tip pinch grip averaged 3.08 ± 1.17 kg in the reconstructive group, 2.72 ± 1.35 kg in the non-reconstructive group, and 3.22 ± 1.12 kg in the implant group. At 3 months, tip pinch strength increased across all groups: 3.47 ± 0.93 kg (Δ = +0.39 kg), 3.33 ± 1.04 kg (Δ = +0.61 kg), and 3.61 ± 1.03 kg (Δ = +0.39 kg), respectively. At 12 months, tip pinch grip continued to improve, reaching 4.98 ± 1.26 kg (Δ = +1.90 kg) for reconstructive procedures, 4.58 ± 0.95 kg (Δ = +1.86 kg) for non-reconstructive procedures, and 4.29 ± 0.86 kg (Δ = +1.07 kg) for implants. The pooled total improved from 3.00 ± 1.21 kg at baseline to 4.67 ± 1.06 kg at 12 months (Δ = +1.67 kg).

Lateral pinch strength at baseline was 3.56 ± 1.43 kg for reconstructive procedures, 4.31 ± 1.17 kg for non-reconstructive procedures, and 4.21 ± 1.26 kg for implants. At 3 months, values were 3.26 ± 1.05 kg (Δ = −0.30 kg), 3.66 ± 0.89 kg (Δ = −0.65 kg), and 5.45 ± 0.96 kg (Δ = +1.24 kg), respectively. At 12 months, lateral pinch strength improved across all groups, reaching 4.48 ± 1.41 kg (Δ = +0.92 kg) for reconstructive procedures, 4.60 ± 1.11 kg (Δ = +0.29 kg) for non-reconstructive procedures, and 6.36 ± 0.86 kg (Δ = +2.15 kg) for implants. The pooled total increased from 3.90 ± 1.33 kg at baseline to 5.11 ± 1.18 kg at 12 months (Δ = +1.21 kg). All grip strength outcomes are summarized in Table 4.

Table 4.

Grip Strength Pooled Weighted Means and Standard Deviations.

Category	Baseline mean (SD)	3-month mean (SD)	Δ 3-month mean	12-month mean (SD)	Δ 12-month mean
Hand grip (kg)
Implant	21.32 (6.03)	21.48 (6.21)	0.16	27.24 (6.72)	5.92
Non-reconstructive	16.10 (6.70)	12.20 (5.70)	−3.9	20.10 (7.40)	4.00
Reconstructive	20.61 (7.76)	18.90 (6.96)	−1.71	24.18 (7.94)	3.57
Total	20.50 (7.03)	19.29 (6.58)	−1.21	24.97 (7.44)	4.47
Tip pinch grip (kg)
Implant	3.22 (1.12)	3.61 (1.03)	0.39	4.29 (0.86)	1.07
Non-reconstructive	2.72 (1.35)	3.33 (1.04)	0.61	4.58 (0.95)	1.86
Reconstructive	3.08 (1.17)	3.47 (0.93)	0.39	4.98 (1.26)	1.9
Total	3.00 (1.21)	3.46 (0.99)	0.46	4.67 (1.06)	1.67
Lateral pinch (kg)
Implant	4.21 (1.26)	5.45 (0.96)	1.24	6.36 (0.86)	2.15
Non-reconstructive	4.31 (1.17)	3.66 (0.89)	−0.65	4.60 (1.11)	0.29
Reconstructive	3.56 (1.43)	3.26 (1.05)	−0.3	4.48 (1.41)	0.92
Total	3.90 (1.33)	4.04 (1.00)	0.14	5.11 (1.18)	1.21

Values are reported as pooled weighted means with standard deviations. Δ values represent mean change from baseline to postoperative follow-up. Positive Δ values indicate improvement in grip strength, whereas negative Δ values indicate postoperative decline relative to baseline. Total values represent pooled weighted means across all intervention categories.

Complication Rates at 12 Months

Overall, 5 total studies reported complication rates (as defined in the Methods), which varied significantly between groups at 12 months (χ² = 12.58; df = 2; P = .00185). The pooled complication rate for soft-tissue/reconstructive procedures (4 of the 5 studies) was 20.61%, compared with 24.49% for implants (2 of the 5 studies), and 51.61% for non-reconstructive techniques (1 of the 5 studies). Although the non-reconstructive category demonstrated the highest observed complication rate, this estimate is derived from a single study and therefore has limited generalizability.

Certainty of Evidence and Risk-of-Bias Assessment

Risk of bias was evaluated across 5 domains using the Cochrane Risk of Bias 2.0 (RoB-2) tool (Figure 2). ³⁰ All included studies were RCTs. Randomization methods were generally adequate and included computer-generated randomization or random number generation in most studies, whereas several studies provided insufficient detail regarding sequence generation or allocation procedures, resulting in judgments of “some concerns” for the randomization process. Allocation concealment was reported using sealed opaque envelopes or centralized randomization in studies with low risk of bias but was incompletely described in others. Blinding of participants and surgeons was inherently limited because of the nature of the surgical interventions; however, several trials minimized detection bias by using blinded outcome assessors and independent evaluators, particularly for patient-reported outcomes and functional assessments.

Figure 2.

Risk of Bias 2.0 traffic light plot.

Five studies (Guzzini et al, ¹⁵ Zarezadeh et al, ³¹ Vermeulen et al, ³² Marks et al, ¹² Picchi et al ³³ ) had “some concerns” overall, primarily due to insufficient reporting of the randomization process and allocation concealment. Two studies (de Jong et al, ³⁴ Graesser et al ³⁵ ) were rated as “low risk,” owing to clearly described randomization procedures, appropriate allocation concealment, and standardized outcome assessment. Across studies, bias due to deviations from intended interventions, outcome measurement, and selection of reported results was consistently judged as low risk, reflecting generally robust adherence to study protocols and prespecified outcome reporting. Overall, the primary source of bias was insufficient reporting of the randomization process and allocation concealment rather than clear evidence of inadequate randomization. All studies received “low-risk” judgments in bias due to deviations from intended intervention, bias in measurement of the outcome, and bias in selection of the reported result, indicating generally robust adherence to intended interventions, outcome measurement, and selection of reported results.

The overall certainty of evidence was evaluated using the GRADE framework (Table 5). ²⁵ High-certainty evidence supported 12-month VAS pain outcomes, indicating that further research is unlikely to substantially change confidence in the estimated treatment effects. Moderate-certainty evidence was identified for surgical complications, grip strength, and early (3-month) functional and strength outcomes, primarily because of clinical heterogeneity in surgical techniques and postoperative rehabilitation protocols. Lateral pinch strength was further downgraded for imprecision due to wide confidence intervals and the limited number of contributing studies. Overall, these certainty ratings suggest that conclusions regarding long-term pain relief are robust, whereas findings related to functional recovery, strength, and complications should be interpreted with greater caution.

Table 5.

Certainty Assessment Using GRADE Analysis for Selected Studies.

Participants (studies)	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Overall certainty of evidence
Pain measured using VAS at 3 months
n = 3 RCTs	Not serious (0)	Serious (−1)	Not serious (0)	Not serious (0)	Early pain relief variable; may reflect peri-op management more than true efficacy	Moderate ⨁⨁⨁◯
Pain measured using VAS at 12 months
n = 4 RCTs	Not serious (0)	Not serious (0)	Not serious (0)	Not serious (0)	Consistent long-term pain relief across techniques; minimal bias or attrition.	High ⨁⨁⨁⨁
Function measured using DASH at 3 months
n = 6 RCTs	Not serious (0)	Serious (−1)	Not serious (0)	Not serious (0)	Function varies by rehab and surgical method; early scores more sensitive to these.	Moderate ⨁⨁⨁◯
Function measured using DASH at 12 months
n = 6 RCTs	Not serious (0)	Serious (−1)	Not serious (0)	Not serious (0)	Functional improvement consistent, but rehab protocols differ across studies.	Moderate ⨁⨁⨁◯
Hand grip at 3 months
n = 6 RCTs	Not serious (0)	Serious (−1)	Not serious (0)	Not serious (0)	Measured objectively, heterogeneity is due to graft type and rehab onset timing.	Moderate ⨁⨁⨁◯
Hand Grip at 12 months
n = 6 RCTs	Not serious (0)	Serious (−1)	Not serious (0)	Not serious (0)	Directional consistency despite magnitude variation; adequate precision.	Moderate ⨁⨁⨁◯
Tip pinch strength at 3 months
n = 4 RCTs	Not serious (0)	Serious (−1)	Not serious (0)	Not serious (0)	Objective measurement, variation in technique, and rehab explain the inconsistency.	Moderate ⨁⨁⨁◯
Tip pinch strength at 12 months
n = 4 RCTs	Not serious (0)	Serious (−1)	Not serious (0)	Not serious (0)	Consistent trends favoring tendon-based techniques; good clinical relevance.	Moderate ⨁⨁⨁◯
Lateral pinch strength at 3 months
n = 4 RCTs	Not serious (0)	Serious (−1)	Not serious (0)	Serious (−1)	High variability in measurement/reporting; depends on first metacarpal stability.	Low ⨁⨁◯◯
Lateral grip strength at 12 months
n = 4 RCTs	Not serious (0)	Serious (−1)	Not serious (0)	Serious (−1)	Small sample sizes, CI overlap; affected by surgical positioning variability.	Low ⨁⨁◯◯
Surgical complications at 12 months
n = 5 RCTs	Not serious (0)	Not serious (0)	Not serious (0)	Not serious (0)	Objectively reported, consistent findings across techniques; well-powered.	High ⨁⨁⨁⨁

Certainty of evidence was assessed using the GRADE framework across domains of risk of bias, inconsistency, indirectness, imprecision, and other considerations. Overall certainty ratings were categorized as high-, moderate-, or low-quality evidence.

Abbreviations: DASH, Disabilities of the Arm, Shoulder, and Hand; GRADE, Grading of Recommendations Assessment, Development and Evaluation; RCTs, randomized controlled trials; VAS, visual analog scale.

Discussion

This systematic review synthesizes contemporary randomized evidence comparing reconstructive, non-reconstructive, and dual-mobility implants for CMC1-OA. Although all surgical strategies demonstrated substantial improvements in pain and function at 12 months, implants appeared to provide faster early postoperative recovery. Interestingly, despite greater early improvements in pain, function, and pinch strength, long-term patient-reported outcomes converged across groups by 1 year. In addition, the absolute postoperative strength values observed in the implant cohort were lower than those reported in several non-randomized dual-mobility series, which may reflect differences in study design, patient selection, rehabilitation, or reduced selection bias in randomized trials.

Previous literature has suggested broadly comparable results among surgical approaches to trapeziectomy. A 2015 Cochrane review ²⁷ concluded that no surgical technique had superior results over simple trapeziectomy and suggested that postoperative outcomes were broadly equivalent regardless of reconstructive method. However, that review did not include modern dual-mobility or semi-constrained prostheses. Our results differ from the opinion that all post-trapeziectomy strategies are equivalent. Early pain reduction, functional improvement, lateral pinch restoration, and complication patterns differed between surgical categories, suggesting that reconstruction of the metacarpo-trapezial complex, whether biologic or mechanical, has meaningful clinical implications. These differences can be explained by biomechanical and procedural factors. Implants offer early gains, such as mechanical stability through preservation of metacarpo-trapezial column height, which may facilitate early mobility and improved pinch mechanics.^14,15

Devices such as the Touch TMC (KeriMedical SA, Plan-les-Ouates, Switzerland) and Maïa prostheses (Groupe Lépine, Genay, France) were not included in older reviews,^7,27 yet are rapidly gaining popularity in Europe and increasingly recognized in emerging RCTs.^14,15,34 Reconstructive procedures depend on the graft’s ability to integrate with native tissue, which introduces variability in healing and recovery due to inherent differences between patient anatomy.^18,29 Non-reconstructive techniques lack structural stabilization, leading to the well-documented pattern of early metacarpal subsidence and temporary strength decline.^9,10 Surgeon experience and learning curve may also influence implant outcomes, as highlighted in a recent proficiency study. ²⁶ These factors may explain why early recovery differs in contrast to 12-month outcomes.

Overall, these findings challenge the dogma that removal of the trapezium is the only constant to achieve successful results. The method by which the trapezial void is managed appears to influence early pain, speed of functional recovery, strength restoration, and the nature of postoperative complications. As new implants gain further traction, we need clear and comparative benchmarks to guide patient selection, counseling, and expectations. This study may provide an early baseline for understanding how modern implants perform relative to established reconstructive strategies. Well-designed larger studies are required to further elucidate this complex relationship and the comparative effectiveness of implant arthroplasties compared to the traditional techniques.

Although the included randomized trials consistently demonstrated faster early recovery with contemporary implant arthroplasty, these findings should be interpreted alongside the long-term durability of each surgical strategy. In the present review, dual-mobility arthroplasty provided earlier improvements in pain, strength, and function than reconstructive or non-reconstructive procedures, although these differences largely converged by 12 months.^15,33,34 However, the included RCTs were limited to follow-up periods of 1 to 2 years and therefore cannot adequately assess implant longevity or revision risk. Long-term studies of trapeziectomy, with or without ligament reconstruction, have demonstrated durable pain relief and functional improvement extending beyond 15 years, with no clear long-term advantage of LRTI over simple trapeziectomy.^28,36 Conversely, recent studies of modern dual-mobility implants, including the Maïa and Touch prostheses evaluated in the present review, have reported encouraging mid- to long-term survivorship with relatively low revision rates, representing a substantial improvement over earlier prosthetic designs.^26,37,38 Nevertheless, implant-specific complications such as loosening, dislocation, polyethylene wear, and revision surgery remain unique considerations that are not encountered following simple trapeziectomy.^37,39 Therefore, although implant arthroplasty may offer clinically meaningful benefits through earlier restoration of strength and function, these advantages should be weighed against the potential need for future revision, particularly in younger or highly active patients with longer life expectancy. At present, the available evidence suggests that procedure selection should remain individualized, balancing the desire for faster recovery with the established long-term durability of trapeziectomy-based procedures until additional long-term randomized comparative studies become available.

This study has several important limitations. First, although data were pooled descriptively across randomized trials, a formal meta-analysis could not be performed due to heterogeneity in surgical technique, implant design, and patient selection, producing substantial variability across studies. Therefore, a true inter-group comparison could not be performed, as improvements reflect results within each group. Nevertheless, these descriptive results can still provide clinically useful insight into how each surgical category performs across each time point and highlight trends that may be helpful for future comparative research. Second, many RCTs were excluded after initial screening for not reporting either the desired outcomes (eg, VAS, DASH, strength) or the 3- and 12-month time points. However, this allowed a strong evaluation of specific outcomes across different surgical groups across by synthesizing only Level 1 evidence. Third, potential selection bias should be considered. The included trials enrolled patients with varying disease severity, implant eligibility, surgeon preferences, and institutional practices, which may have influenced treatment allocation and outcome reporting despite randomization within individual studies. Furthermore, studies with insufficient quantitative data or incomplete outcome reporting were excluded, introducing the possibility of reporting bias. Finally, several included trials were judged as having “some concerns” on the Cochrane RoB-2 assessment, primarily because of incomplete reporting of the randomization process and allocation concealment rather than clear methodological deficiencies. Although most studies demonstrated low risk of bias for deviations from intended interventions, outcome measurement, and selective reporting, these methodological limitations reduce the overall certainty of some outcome estimates. Despite these limitations, this study represents one of the few contemporary RCT-only pooled analyses incorporating modern implant arthroplasties to treat CMC1-OA. This information can guide surgeons as the use of modern implants becomes more prevalent.

Conclusions

Dual-mobility implants appear to provide earlier improvements in pain, function, and strength than traditional reconstructive or non-reconstructive procedures, although these differences largely converge by 12 months. Contemporary implant arthroplasty therefore represents a promising option for appropriately selected patients but should be balanced against long-term implant durability and revision risk.