Joint Distraction for Thumb Carpometacarpal Osteoarthritis: 2-Year Follow-up Results of 20 Patients

Abstract

Background  Joint distraction is a fairly new treatment for patients with symptomatic thumb carpometacarpal osteoarthritis (CMC1 OA). A previous pilot study of five patients showed that CMC1 joint distraction is technically feasible. The current study presents the results of CMC1 joint distraction in 20 patients with a 2-year follow-up period.

Purposes  The primary study aim was to assess if patients with CMC1 OA have better physical function and less pain 2 years after CMC1 joint distraction. Second, we assessed the number of patients who achieved a minimal clinically important difference (MCID) in patient-reported outcome measures at each follow-up time point. Furthermore, this study sought differences on magnetic resonance imaging (MRI) of the CMC1 joint before and after distraction. Adverse events were noted and reported.

Methods  Twenty patients (median age of 54 years) with symptomatic CMC1 OA and an established indication for a trapeziectomy were enrolled. An external distractor device was placed over the CMC1 joint and left in situ for 8 weeks. Disabilities of the Arm, Shoulder, and Hand (DASH) score, Michigan Hand Outcome Questionnaire (MHQ), visual analogue scale (VAS), and grip strength were recorded preoperatively and at 3, 6, 12, and 24 months postoperatively.

Results  Two years after joint distraction, physical function and pain scores had improved significantly compared with baseline: DASH from 48 to 17, MHQ from 56 to 83, and VAS for pain from 50 to 18 mm. Fourteen of 19 patients (74%) reached an MCID in DASH and MHQ scores. One patient was not satisfied with treatment outcome and chose to proceed with a trapeziectomy 14 months after initial distraction therapy.

Conclusions  This study demonstrates that CMC1 joint distraction can postpone more invasive surgical interventions (e.g., trapeziectomy) for at least 2 years. Larger comparative studies are needed to assess the value of CMC1 joint distraction in the treatment of CMC1 OA.

Level of Evidence  This is a Level IV, prospective case series study.

Carpometacarpal osteoarthritis of the thumb (CMC1 OA) affects approximately one-third of the population aged 55 years and older. ^{1 2} The radiographic prevalence increases to 90% in people over 80 years of age. ^{1 2} Initial treatment of patients with symptomatic CMC1 OA routinely involves nonoperative options first, including splints, oral analgesia, and hand therapy. ³ If nonoperative management does not offer sufficient relief, surgical treatment can be considered. ^{4 5} Surgical treatment options of CMC1 OA vary greatly. Also, there is no evidence for the superiority of each technique regarding pain and functional outcome. ^{4 5} Trapeziectomy with or without ligament reconstruction and tendon interposition (LRTI) or suspension arthroplasty is effective but carries the long-term risk of metacarpal subsidence, persistence, or recurrence of symptoms. ^{4 6 7} Prostheses are associated with loosening, subluxation, fracture, and synovitis, potentially requiring revision surgery. ^{4 6 7} With respect to patients requiring surgical intervention for CMC1 OA at a relatively young age, other less invasive techniques that preserve the joint may be considered more desirable.

Joint distraction is an innovative joint-sparing treatment for patients (<65 years) with ankle or knee OA. It aims to postpone or prevent an invasive surgical intervention. ^{8 9} Previous evidence reports that joint distraction can reduce pain, improve physical function, and increase joint cartilage thickness measured on magnetic resonance imaging (MRI). ^{8 10 11} van der Woude et al showed persisting pain reduction and greater physical function compared with baseline at 5-year follow-up in patients who underwent knee distraction for painful OA. ¹¹ A paper published in 2018 reported a 9-year knee joint survival rate of nearly 50% after joint distraction, with long-lasting clinical and structural improvements. ⁸ Our previous pilot study of five patients with a 1-year follow-up period demonstrated that CMC1 joint distraction is technically feasible and provides pain reduction and improved physical function. ¹²

We aimed to assess the effect of CMC1 joint distraction in a larger cohort including 20 patients with a 2-year follow-up period. The purposes of this study were: (1) to assess surgical outcomes of joint distraction—in terms of physical function and pain—for patients at a relatively young age (<65 years) with CMC1 OA; (2) to analyze the number of patients who achieved a minimal clinically important difference (MCID) in patient-reported outcome measures (PROMs) at each time point; (3) to assess MRI findings of the CMC1 joint before and after the surgery; and (4) to report the adverse events.

Materials and Methods

Study Design

This prospective cohort study was approved by our local institutional review board. We enrolled 20 patients from a single outpatient center and obtained written informed consent. The first five patients were enrolled between October and September 2014 to assess technical feasibility of the surgical technique. The following 15 patients were enrolled between September 2016 and November 2017. All patients were aged <65 years, had symptomatic CMC1 OA (Eaton–Glickel classification II or III on radiographs), and failed to obtain acceptable results with nonoperative treatment efforts (e.g., hand therapy or 3-month splint use). In all patients, the indication for invasive surgical treatment (e.g., trapeziectomy) was made. ¹³ Patients were excluded for the following reasons: severe radiographic CMC1 OA (Eaton–Glickel grade IV), previous surgical treatment of the affected CMC1 joint, a past medical history of inflammatory or rheumatoid arthritis, or use of immunosuppressive or chemotherapeutic drugs. Details of the inclusion and exclusion criteria are further described in the prior pilot study by Spaans et al. ¹²

Surgical Technique

All procedures took place in our day care surgery center and were performed by one of two hand surgeons (A. B. or A. B. M. v. d. M.). General anesthesia was performed in 5 of 20 patients who specifically preferred this over locoregional anesthesia. All patients received 2 g of intravenous cefazolin preoperatively. Under fluoroscopic guidance, the distractor device (Osteo-x, Osteotec; Dorset, UK) was placed over the CMC1 joint. The device is anchored percutaneously with two proximal K-wires into the trapezium bone and two distal K-wires in the first metacarpal by using the distractor as a drill guide ( Fig. 1 ). Subsequently, the device distracted 3 mm intraoperatively and the K-wires shortened to 1 cm above the distractor device. The average procedure time was 15 to 30 minutes, depending on surgeon experience and individual cases (e.g., height of trapezium bone for placement of K-wires). A custom-made thermoplastic splint was applied to cover and protect the distractor. The distractor was left in situ for the following 8 weeks, after which it was removed in our outpatient clinic. Hand therapy commenced thereafter.

Fig. 1.

( A ) The distractor device. ( B ) Drawing of the trapezium bone and metacarpal bone. ( C ) Distractor device in situ with two K-wires in the trapezium bone and two in the first metacarpal.

Patients Evaluation

Upper extremity–specific disability was measured with the Disabilities of Arm, Shoulder, and Hand (DASH) questionnaire (Dutch language version), which produces scores from 0 to 100, with higher scores indicating more disability. ¹⁴ Hand health status was assessed using the Michigan Hand Outcome Questionnaire (MHQ) (Dutch language version), which produces scores between 0 and 100, with higher scores indicating greater hand health. ¹⁵ The applied MCID, the smallest difference that patients perceive as beneficial, was based on those of previous studies: 10 points for DASH and 9 points for MHQ. ^{16 17} Thumb pain, thumb function, and overall patient satisfaction were scored on a 0- to 100-mm visual analogue scale (VAS). A score of 0 mm represented no pain, perfect thumb function, or perfect satisfaction. A score of 100 mm represented the worst imaginable pain, poorest satisfaction, or worst possible thumb function. ¹⁸ Patients were asked to complete these questionnaires preoperatively and at 3, 6, 12, and 24 months after the procedure. During the distraction period, only VAS scores were assessed every 2 weeks. The reported 2-year follow-up period is in accordance with the suggested guidelines of required minimal follow-up in hand surgery studies. ¹⁹

An independent certified hand therapist measured active range of thumb motion: (1) radial abduction (degrees), (2) extension (cm) with a Devore goniometer, (3) palmar abduction with the Pollexograph (degrees), (4) abduction by intermetacarpal distance (cm), and (5) opposition by Kapandji measurements (0–10). ²⁰ Grip strength was measured using a hand dynamometer with the shoulder adducted and in neutral rotation, elbow at 90 degree flexion, and the forearm and wrist in neutral position. Key, tip, and three-point pinch strength were measured by baseline pinch gauge (P100 Hand Kit, Biometrics Ltd; Gwent, UK). Strength measurements were recorded as the average of three attempts.

Diagnostic Imaging

Radiographs (anteroposterior, lateral, and Bett's view) were obtained at all visits to evaluate the joint space. MRI of the CMC1 joint was obtained in the first five patients preoperatively and 12 months postoperatively using a dedicated wrist coil (Siemens 3T Skyra; Erlangen, Germany). The Outcome Measures in Rheumatology Clinical Trials (OMERACT) Thumb base Osteoarthritis MRI Scoring system (TOMS) was used to assess subchondral bone defects, cartilage, osteophytes, synovitis, and bone marrow lesions on a scale from 0 to 3 (normal–severe) of the CMC1 and scaphotrapeziotrapezoidal joints. ²¹ Subluxation of the CMC1 joint was assessed and scored as absent (“0”) or present (“1”). ²¹ MR images were reviewed by two experienced independent researchers, blinded for patient treatment details, using the OMERACT TOMS tool. Previous reports define the average-measure intraclass correlation coefficients of the OMERACT TOMS tool for the CMC1 joint as good to excellent for most features (>0.70), except for cartilage assessment (0.39) and osteophytes (0.47). ²²

Adverse Events

Adverse events were scored according to the Clavien–Dindo classification of surgical complications, ranging from grade I (minor complication) to grade V (death) ( Appendix A ). ²³

Appendix A. The Clavien–Dindo classification of surgical complications ^a .

Grades	Definition	Adverse events in study cohort b
Grade I	Any deviation from the normal postoperative course without the need for pharmacological treatment or surgical, endoscopic, and radiological interventions	• Superficial pin tract infection ( N  = 5) • Pin infection and early device removal ( N  = 1) • Broken K-wire ( N  = 4)
	Allowed therapeutic regimens are: drugs as antiemetics, antipyretics, analgetics, diuretics, and electrolytes and physiotherapy. This grade also includes wound infections opened at the bedside
Grade II	Requiring pharmacological treatment with drugs other than such allowed for grade I complications
	Blood transfusions and total parenteral nutrition are also included
Grade III	Requiring surgical, endoscopic, or radiological intervention	De Quervain release ( N  = 1)
- IIIa	Intervention not under general anesthesia
- IIIb	Intervention under general anesthesia
Grade IV	Life-threatening complication (including central nervous system complications) requiring IC(U)-management
- IVa	Single-organ dysfunction (including dialysis)
- IVb	Multiorgan dysfunction
Grade V	Death of a patient

Abbreviations: IC, intermediate care; ICU, intensive care unit.

^aPresented with permission as the description from Dindo D, Demartines N, Clavien P-A. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240(2):205–213.

^bAdded column, not part of the Clavien–Dindo classification.

Sample Size Calculation

The aim of this proof-of-principle study was to give more insight in the effects of CMC1 distraction. An a priori power analysis was not performed. Results obtained in this study will be used for a priori power analysis of future studies.

Statistical Analysis

Continuous variables are reported as median with interquartile range (IQR) and discrete variables as absolute numbers with percentages. The Wilcoxon signed-rank test was used to compare the differences between DASH, MHQ, VAS scores, range of motion, and strength measurements. All p -values < 0.05 were considered statistically significant.

Results

Participants

The median age of 20 patients was 54 years (range, 41–64 years). Nineteen (95%) were women ( Table 1 ). Fourteen patients (70%) underwent distraction on their dominant hand ( Table 1 ). For one patient, the preoperative measurements of VAS, strength, and range of motion were not available for analysis. One patient proceeded with a trapeziectomy 14 months after initial distraction treatment because of ongoing pain. Results of this patient are not included in the 2-year follow-up results.

Table 1. Patient characteristics.

Variables	N  = 20
Age, median (range) (y)	54 (41–64)
Gender
Man	1 (5)
Woman	19 (95)
Dominant hand
Left	3 (15)
Right	17 (85)
Operated on dominated hand	14 (70)

Note: Discrete variables as number (percentage), unless otherwise specified.

Difference in Pain and Physical Function

At 1-year follow-up, all PROMs improved after CMC1 joint distraction compared with baseline. These improvements remained at 2-year follow-up ( Table 2 ). DASH scores improved by 28 points at 1 year and by 31 points (from 48 to 17; p  < 0.001) at 2 years; MHQ increased by 21 points at 1 year and by 27 points at 2 years (56 to 83; p  < 0.001).

Table 2. Baseline, 1-year, and 2-year follow-up results of all patients.

Variables	Baseline N  = 19 a	1 year N  = 20	p -Value	2 years N  = 19 b	p -Value (2 years vs. baseline)
DASH	48 (34–57)	20 (4–49)	<0.001	17 (3.3–28)	<0.001
MHQ	56 (41–62)	77 (62–95)	<0.001	83 (64–95)	<0.001
VAS pain	50 (37–72)	15 (0–37)	0.003	18 (7–28)	<0.001
VAS function	59 (40–75)	30 (5–56)	0.007	26 (13–32)	0.001
VAS satisfaction	75 (61–86)	13 (0–34)	0.002	24 (1–47)	0.002
Range of motion
Radial thumb abduction in degrees	60 (55–65)	57 (54–60)	0.020	58 (55–62)	0.601
Palmar thumb abduction, Pollexograph in degrees	40 (40–46)	44 (39–51)	0.115	44 (40–54)	0.091
Palmar thumb abduction, imd in cm	5.3 (5.0–5.8)	6.0 (5.5–6.5)	0.032	6.2 (5.5–6.7)	0.004
Extension in cm	4.0 (3.5–5.0)	3.0 (2.5–4.0)	0.033	3.5 (3.0–5.0)	0.554
Kapandji score	10 (9–10)	9 (9–10)	0.816	10 (9–10)	0.741
Strength measurements (kg/m 2 )
Grip	20 (13–26)	21 (17–28)	0.040	23 (21–26)	0.053
Key pinch	4.5 (3.5–5.6)	5.4 (4.0–6.2)	0.067	5.0 (4.5–6.7)	0.074
Three-point pinch	3.6 (2.8–4.7)	4.6 (3.3–5.0)	0.409	4.6 (3.3–5.0)	0.089
Tip pinch	3.2 (2.1–3.3)	2.9 (2.2–4.3)	0.334	3.3 (2.4–3.9)	0.223

Abbreviations: DASH, Disabilities of Arm, Shoulder, and Hand; imd, intermetacarpal distance; MHQ, Michigan Hand Outcome Questionnaire; VAS, visual analogue scale.

Note: All continuous variables as median (interquartile range). Bold indicates statistically significance, p  < 0.05.

^aOne missing value at baseline for all measures except DASH and MHQ.

^bOne patient excluded after trapeziectomy.

VAS scores for pain intensity improved from 50 to 18 mm ( p  < 0.001) at 2 years, thumb function from 59 to 26 mm ( p  = 0.001), and overall patient satisfaction from 75 mm at baseline to 24 mm 2 years postdistraction ( p  = 0.002) ( Table 2 ).

Some range of motion measures decreased 1 year postdistraction but improved back to baseline levels at 2-year follow-up. There was no difference in motion and strength measurements at 2 years compared with baseline, except for an increase of thumb palmar abduction ( Table 2 ).

Minimal Clinically Important Differences

Twelve patients ( N  = 12/20; 60%) reached the MCID for DASH score at 1 year and 14 patients (74%; N  = 14/19) did so at 2 years ( Appendix B ). For the MHQ, these numbers were 15 (75%; N  = 15/20) at 1 year and 14 (74%; N  = 14/19) at 2 years ( Appendix B ).

Appendix B. MCID of MHQ and DASH scores.

Patient	DASH			MHQ
	Baseline	1 year	2 years	Baseline	1 year	2 years
	N  = 20	N  = 20	N  = 19	N  = 20	N  = 20	N  = 19
1	36.7	1.7	10.0	54.7	96.1	95.1
2	59.2	24.2	27.5	33.4	64.1	64.0
3	56.7	50.0	24.0	61.3	71.8	67.7
4	62.5	42.5	13.8	48.5	65.3	85.9
5	50.8	18.3	19.2	41.0	81.7	89.8
6	35.0	17.5	10.8	56.3	77.5	83.4
7	70.0	65.8	71.7	36.9	31.6	31.5
8	55.0	50.8	45.0	53.5	54.4	46.1
9	44.8	60.0	X	56.5	43.0	X
10	51.8	0.83	2.5	41.0	100	98.3
11	27.5	6.7	3.3	62.6	79.8	86.3
12	33.3	0.83	0.0	65.4	100	89.3
13	39.8	0.83	0.0	60.6	93.9	98.0
14	62.5	18.3	17.5	32.7	76.9	58.4
15	33.3	24.2	26.7	60.2	81.6	79.8
16	25.0	4.2	4.2	66.2	98.8	98.9
17	44.0	4.2	0.0	52.2	96.9	95.0
18	24.2	22.5	16.7	68.3	69.9	81.8
19	57.5	48.3	59.2	37.4	47.0	30.2
20	51.7	65.8	55.8	64.8	60.4	64.7
MCID reached		12 (60%)	14 (74%)		15 (75%)	14 (74%)

Abbreviations: DASH, Disabilities of Arm, Shoulder, and Hand; MCID, minimal clinically important differences; MHQ, Michigan Hand Outcome Questionnaire.

Note: “X,” excluded. Bold indicates measures that surpassed the MCID of 10 points for DASH and 9 points for MHQ.

Diagnostic Imaging

Fig. 2 shows radiographs before, during, and 12 and 24 months after distraction therapy. Analysis of MRI scored by OMERACT TOMS showed no difference in osteophytes, cartilage loss, or subluxation ( Table 3 ). There was a small increase of synovitis, subchondral bone defects, and bone marrow lesions in the CMC1 joint 1 year after distraction ( Table 3 ; Fig. 3 ).

Fig. 2.

Radiographs ( A ) before and ( B ) during distraction therapy, and ( C ) at 1-year and ( D ) at 2-year follow-up. Image courtesy: Spaans et al. ¹²

Table 3. OMERACT TOMS scores on MRI.

Variables	Preoperative	1-year postoperative
Synovitis	0.25 (0.0–0.25)	0.50 (0.0–0.75)
Subchondral bone defects	0.30 (0.30–0.50)	0.75 (0.70–0.90)
Osteophytes	0.70 (0.60–0.90)	0.70 (0.60–0.90)
Cartilage loss	0.50 (0.25–0.75)	0.50 (0.25–0.75)
Bone marrow lesions	0.40 (0.10–0.50)	0.65 (0.60–0.75)
Subluxation, n (%)	0.0 (0.0)	0.0 (0.0)

Abbreviation: OMERACT TOMS, Outcome Measures in Rheumatology Clinical Trials Thumb base Osteoarthritis MRI Scoring system.

Note: All continuous variables as median (interquartile range), unless otherwise specified. Scores range from 0–3, except for the subluxation variable, which has scores of 0 or 1.

Fig. 3.

MRI ( A ) before and ( B ) 12 months after distraction therapy with slight increase of bone marrow lesions and subchondral bone defects.

Adverse Events

Adverse events were scored according to the Clavien–Dindo classification ( Appendix A ). ²³ In one patient, a single K-wire tip broke while putting on socks; in three other patients, a broken K-wire was noticed during removal of the distractor device and a small piece of K-wire was left in situ in the trapezium bone in these four patients. To avoid further K-wire breakage, we changed to using unthreaded K-wires and released tension of the distraction device before its removal. Thereafter, no further K-wire complications were noted. Six patients experienced a superficial pin tract infection, which was managed successfully with oral antibiotics in five patients. In one patient, the distractor device had to be removed at 6 weeks to control the infection. After removal, the infection resolved. One patient developed De Quervain's tenosynovitis on the operated side 6 months after distraction therapy and was treated successfully with surgical release of the affected compartment.

Discussion

Distraction is an innovative and joint-preserving technique. It has the potential to delay or prevent more invasive surgical procedures for patients with CMC1 OA seeking care. ^{8 11 12} This study shows a significant improvement in PROMs at 2 years after distraction therapy.

We acknowledge some limitations of the present study. First, this is a noncomparative trial. The contribution of a possible placebo effect and regression to the mean of the outcome measures could not be assessed. There is evidence that many people with CMC1 OA adjust over time and do not seek medical care. ^{5 24 25} Future studies comparing distraction with other operative and nonoperative treatments are required to address such effects. Second, this study describes a small number of patients, mainly woman, who were enrolled in a single urban center. Our results might not generalize to other populations, regions, or practice settings.

This proof-of-principle study demonstrates the ability of CMC1 joint distraction to improve physical function and reduce pain at 1 and 2 years after the procedure. The median improvements in DASH, MHQ, and VAS pain scores of 28, 21, and 35 at 1 year, respectively, and 31, 27, and 32 at 2 years are comparable to prior published outcomes after trapeziectomy (with or without interposition). ^{26 27 28 29} Wang et al retrospectively reviewed 20 patients who received trapeziectomy with LRTI and reported improvements of 30 points in DASH (from 52 to 22) and 54 in pain intensity (68 to 14) at 2-year follow-up. No MHQ scores were assessed. ²⁹ Although pain intensity improved more in their study, this can be related to the higher baseline intensity reported, since the results at 2 years are nearly the same (14 vs. 18). The improvement in MHQ of 22 points at 1 year is comparable to the results of a recent prospective cohort study among 233 Dutch patients who underwent trapeziectomy with interposition tendinoplasty: MHQ scores increased with 21 points to 69 points after 1 year. ²⁸ A randomized prospective trial by Davis and Pace comparing simple trapeziectomy versus trapeziectomy with LRTI and K-wire insertion among 128 patients showed 31 points improvement of DASH scores at 1 year after simple trapeziectomy and 28 points improvement after trapeziectomy with LRTI and K-wire insertion. These results are comparable to the 28 points improvement in DASH we noted at 1-year follow-up. ²⁷ Both studies did not report results at 2 years, which is the suggested minimum length of follow-up in hand surgery reports. ¹⁹ The finding that joint distraction results in improved patient-reported outcomes that are comparable to outcomes in other procedures for CMC1 OA supports the use of CMC1 joint distraction in selected patients.

Out of 19 patients, 14 (74%) reached an MCID in both DASH and MHQ scores 2 years after joint distraction. To our knowledge, there are no other studies describing the MCID of individual patients with CMC1 OA. Some studies mention the average MCID of all patients after treatment of CMC1 OA. ^{28 30} Our findings show that CMC1 joint distraction leads to improvements that most patients experience as beneficial.

The finding that no differences in cartilage quality were detectable on additional imaging deviates from previous distraction studies. ^{8 11} This is possibly related to the differences in shape and type of joints studied. The CMC1 joint has a smaller joint space compared with knee joints and is a non–weight-bearing joint with a unique saddle shape. ⁵ Additionally, the complexity of the shape and the oblique orientation of the trapezium make assessment of the CMC1 joint challenging on both radiographs and MRI. For these reasons, radiographic classification systems lack reliability. ³¹ Studies on knee distraction showed increased cartilage thickness and joint space width after treatment. ^{10 11} These findings, however, were not correlated with better patient outcomes. ^{10 11} Improved imaging techniques or invasive arthroscopic sampling of articular cartilage is needed to reliably assess CMC1 cartilage quality. Alternatively, biochemical analysis of synovial fluids can be a promising avenue for future research of the CMC1 joint.

One of the 20 patients experienced ongoing pain after joint distraction and proceeded with a trapeziectomy 14 months after the initial surgery. We defined this r-operation in one patient as treatment failure. A previous study on ankle OA reported that 17% (18/105) underwent arthrodesis or osteotomy within 2 years of distraction, defined as treatment failure. ³² In a cohort of knee OA patients, 17% (3/18) converted to knee arthroplasty between 3 and 5 years after distraction. ¹¹ The failure rate of 1/20 patients in this study illustrates that joint distraction can postpone invasive surgical interventions in the majority of patients. Future studies can assess factors associated with treatment failure, for example, patient characteristics, to better select patients who will likely benefit from joint distraction.

In 9 of 20 patients (45%), 11 adverse events occurred of which 10 events were classified as a grade I complication ( Appendix A ). This rate is comparable to a study by Davis et al after trapeziectomy and K-wire stabilization for 4 weeks: in 101 of 183 patients (55%), an adverse event occurred, mainly nerve- or tendon-related problems. ³³ Other studies on CMC1 OA seem to have fewer adverse events: in 14 of 65 patients (22%) after trapeziectomy with or without LRTI, as reported by Field and Buchanan. ³⁴ A comparative trial by Davis and Pace showed complication rates of 31% after single trapeziectomy and 46% after trapeziectomy with LRTI and K-wire stabilization. ²⁷ For arthrodesis of the CMC1 joint nonunion rates up to 26% have been reported in a previous review. Adverse event rates of 15% were reported for Swanson implants. ^{4 5} In knee distraction series, 17 of 20 patients (85%) suffered from pin tract infections and 2 patients (10%) from a pulmonary embolus. ¹⁰ The present small cohort distraction study had a relatively high number of adverse events compared with other CMC1 OA procedures but the severity of the events was classified as the lowest category of complications (grade I; Appendix A ). One-third of the adverse events in our study was related to the use of threaded K-wires. After changing to unthreaded wires, no more device-related complications were noted. A more reliable adverse event and treatment failure rate can be established by studying a larger cohort with longer follow-up.

In conclusion, this study demonstrates that 8 weeks of CMC1 distraction in patients younger than 65 years of age with symptomatic CMC1 OA can reduce pain and improve physical function after 1 year, which is sustained at 2 years. These findings suggest that joint distraction could postpone more invasive procedures, such as trapeziectomy, for most patients. To better determine the value of joint distraction in the treatment of CMC1 OA, future studies are required with a larger cohort of patients, a comparison group, and longer follow-up.

Funding Statement

Funding This study was financially supported by the Dutch Arthritis Foundation on 14-01-2013 to perform MRI in five patients.