Keywords trapeziectomy - carpometacarpal osteoarthritis - carpometacarpal joint - ligament
reconstruction and tendon interposition
The trapeziometacarpal articulation in the thumb is a joint that is second-most commonly
affected by osteoarthritis, and this can lead to considerable hand pain and disability.[1 ]
[2 ] Trapeziometacarpal arthritis typically occurs in those aged between 50 and 70 years,
and it is often associated with obesity and having a manual occupation.[3 ] Currently, there is a multiplicity of surgical options available to address this
problem, yet none has proven to be significantly superior to the others.[4 ] The most commonly performed procedure is a trapeziectomy augmented with ligament
reconstruction and tendon interposition (LRTI). The latter technical modification
was developed to help prevent postoperative loss of strength, thumb shortening, and
risk of scaphoid impingement.[5 ]
[6 ] Total joint replacement has evolved since its introduction in the 1970s,[7 ]
[8 ] and recent studies have suggested that it can provide a quicker rehabilitation period
and improved function in the short term when compared with the LRTI.[9 ]
[10 ]
[11 ]
[12 ]
In this article, the postoperative outcomes of trapeziectomy with LRTI and trapeziometacarpal
joint replacement have been compared. Up to 2019, there had been no randomized controlled
trials (RCTs), with most studies being nonrandomized and comparative in nature. Our
aim was to perform a systematic review of the literature and meta-analysis to examine
the patient related outcomes.
Materials and Methods
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines[13 ] formed the basis on which this systematic review was performed.
Eligibility Criteria
The Population, Intervention, Control, Outcome, and Time (PICOT) criteria[14 ] was used to formulate the eligibility criteria, therefore reducing ambiguity. All
RCTs and observational studies comparing the outcomes of trapeziectomy with LRTI and
trapeziometacarpal joint replacement for thumb carpometacarpal osteoarthritis that
met the criteria were included. Thumb carpometacarpal osteoarthritis was the pathology
of interest.
Outcome Measures
Five outcome measures were compared: The Quick Disabilities of the Arm, Shoulder,
and Hand (QDASH) score, Kapandji score, grip strength, pinch strength, and the Visual
Analog Scale (VAS) for pain score.
Literature Search Strategy
Two independent co-authors (M.K.Q. and U.H.) used the NICE Healthcare Databases Advanced
Search (HDAS) tool to search articles in the PubMed, Medline, EMBASE, and Cochrane
Central Register of Controlled Trials databases. The most recent search was performed
on August 16, 2020. Articles included in this review were critically evaluated by
using the Critical Appraisal Skills Program tool[15 ] and Cochrane Handbook for Systematic Reviews of Interventions.[16 ]
Study Selection
Two co-authors assessed the titles and abstracts of all the identified studies from
the literature search. Of the studies which were deemed applicable, the full texts
were evaluated and those which met the eligibility criteria were selected. A third
co-author was involved if there were any discrepancies identified within the study
selection.
Data Collection
The Cochrane's data collection criteria for intervention reviews were used to produce
a data extraction sheet. Study data extraction included authors, journal, study design,
sample size, follow-up period, and outcome measures recorded. Demographics such as
mean age and sex were collected. Two independent co-authors extracted the data, and
where there were any discrepancies a third co-author was consulted.
Data Synthesis and Statistical Analysis
For the key continuous outcome measures of interest (QDASH, VAS score, grip strength,
pinch strength, and Kapandji score), the mean differences between the trapeziectomy
with LRTI and thumb carpometacarpal joint replacement groups were calculated.
Review manager 5.3 software (Cochrane Community, Oxford, United Kingdom) was selected
for data analysis. Fixed and random-effect models were employed as necessary. Random-effects
modelling was only applied if substantial heterogeneity was present between the studies.
Forest plots using 95% confidence intervals (CI) were used to display the results.
The I2
was calculated to reveal any inconsistencies between the studies. However, 0 to 25%
suggested low heterogeneity, 25 to 75% suggested moderate heterogeneity, and 75 to
100% suggested high heterogeneity.
Methodological Quality and Risk of Bias Assessment
The risk of bias was assessed by using the Cochrane tool[17 ] and Newcastle-Ottawa Scale (NOS).[18 ] All articles were evaluated by two independent co-authors. The Cochrane tool analyses
bias specifically existing in RCTs in domains such as selection, reporting, and attrition.
The NOS employs a 1-star (lowest risk) to 9-star (highest risk) scale to judge the
quality of each study on three broad areas; the comparability of the groups, the selection
of each study group, and the ascertainment of outcome of interest. If any dispute
arose between assessments of both co-authors, a third independent co-author was consulted.
Sensitivity Analysis
To evaluate the strength of our data and to investigate possible sources of heterogeneity,
sensitivity analysis was performed. To investigate the effect each study had on the
overall effect heterogeneity, the statistical analyses were performed multiple times
once each study had been removed separately.
Results
Literature Search Results
A total of 654 articles were identified during the literature search, of which four
met the inclusion criteria for this review ([Fig. 1 ]). This consisted of one RCT,[9 ] one prospective cohort study,[11 ] and two retrospective cohort studies.[12 ]
[19 ]
Fig. 1 Study flow diagram.
Patient Demographics and Follow-Up
The total number of patients who underwent treatment in these articles was 583. Of
those, 358 underwent trapeziectomy with LRTI and 225 underwent an arthroplasty procedure.
The mean age of the patients was 68 years. The percentage of female patients was 84%.
The mean follow-up period was 54 months. Height and weight were not recorded in any
of the selected studies. [Table 1 ] presents the characteristics of each study included Methodological Quality and Risk
of Bias.
Table 1
Baseline characteristics of the included studies
Study (Year)
Journal title
Study design
Follow-up (mo)
Sample size
Trapeziectomy + LRTI
Arthroplasty
Mean age (y)
Female (%)
Thorkildsen et al[9 ] (2019)
Journal of Plastic Surgery and Hand Surgery
RCT
48
40
20
20
62.5
70
Cebrian-Gomez et al[11 ] (2019)
The Journal of Hand Surgery (European volume)
Prospective
45.5
146
62
84
60.65
92.15
Robles-Molina et al[12 ] (2017)
Orthopaedics
Retrospective
57.5
65
34
31
58.43
83
Vandenberghe et al[19 ] (2013)
The Journal of Hand Surgery (European volume)
Retrospective
64.8
332
242
90
59.5
91.57
Abbreviations: LRTI, ligament reconstruction and tendon interposition; RCT, randomized
controlled trial.
[Figure 2 ] represents the outcomes of the methodological quality assessment of all the studies
included.
Fig. 2 Risk of bias graph showing authors' judgements about each risk of bias item for observational
studies (A ) and randomized trials (B ).
Outcome: Quick Disabilities of the Arm, Shoulder, and Hand Score
The QDASH score was recorded in four studies[9 ]
[11 ]
[12 ]
[19 ] involving 583 patients ([Fig. 3 ]). There was a significant difference in the QDASH score between the trapeziectomy
with LRTI and joint replacement groups (Z = 3.37, p = 0.0008). The joint replacement group exhibited better QDASH scores than the trapeziectomy
group. Heterogeneity was deemed to be low among the studies (I
2 = 0%, p = 0.42). The mean difference in QDASH score was 4.32 (95% CI: 1.80–6.83) in favor
of joint replacement.
Fig. 3 Forest plots of the comparisons of the Quick Disabilities of the Arm, Shoulder, and
Hand score (QDASH), the Visual Analogue Scale (VAS) for pain score, Kapandji score,
grip strength and pinch strength.
Outcome: Visual Analog Scale Score
The VAS score was recorded in three studies[11 ]
[12 ]
[19 ] involving 543 patients ([Fig. 3 ]). There was no significant difference in the VAS score between the trapeziectomy
with LRTI and joint replacement groups (Z = 0.89, p = 0.38). Heterogeneity was deemed to be high between the studies (I
2 = 82%, p = 0.004). The mean difference in VAS score was 0.39 (95% CI: −0.47 to 1.25) in favor
of joint replacement.
Outcome: Kapandji Score
The Kapandji score was recorded in three studies[9 ]
[11 ]
[12 ] involving 251 patients ([Fig. 3 ]). There was a significant difference in the Kapandji score between the trapeziectomy
with LRTI and joint replacement groups (Z = 3.09, p = 0.002). The joint replacement group exhibited better Kapandji scores when compared
with the trapeziectomy with LRTI group. Heterogeneity was deemed to be moderate between
the studies (I
2 = 40%, p = 0.19). The mean difference in Kapandji score was 0.53 (95% CI: 0.87–0.19) in favor
of joint replacement.
Outcome: Grip Strength
The grip strength was recorded in two studies[9 ]
[11 ] involving 186 patients ([Fig. 3 ]). There was no significant difference in grip strength between the trapeziectomy
with LRTI and joint replacement groups (Z = 0.57, p = 0.57). Heterogeneity was deemed to be low among the studies (I
2 = 0%, p = 0.50). The mean difference in grip strength was −0.61 (95% CI: −2.71 to 1.48) in
favor of trapeziectomy with LRTI.
Outcome: Pinch Strength
The pinch strength was recorded in three studies[9 ]
[11 ]
[12 ] involving 251 patients ([Fig. 3 ]). There was no significant difference in grip strength between the trapeziectomy
with LRTI and joint replacement groups (Z = 0.94, p = 0.35). Heterogeneity was deemed to be high between the studies (I
2 = 88%, p = 0.0002). The mean difference in pinch strength was −0.87 (95% CI: −2.67 to 0.93)
in favor of trapeziectomy with LRTI.
Randomized Controlled Trials
One RCT involving 40 patients was included in this study. No significant differences
in the primary outcomes measured was noted between the groups at 2 years. No significant
difference in the objective measurements was noted between the groups at 2 years.
Radiological Assessment
Thorkildsen et al preoperatively assessed the patients with standardized anteroposterior
(AP) and lateral radiographs of both thumbs as well as computed tomography scans of
the involved hand. Repeat radiographs were taken 6 weeks postoperatively and at all
subsequent follow-up appointments. In the arthroplasty group, lesions were analyzed
by using the lateral radiograph in the zones described by Chakrabarti et al.[20 ] For the trapeziectomy group, the ratio of the trapezial space to the length of the
proximal phalanx was calculated pre- and postoperatively. At the final follow-up,
the trapezial space height was described according to Downing and Davis.[21 ]
Cebrian-Gomez et al assessed standard AP and lateral radiographs of the affected hand.
The Eaton classification[22 ] was used to assess the preoperative trapeziometacarpal status. In the joint replacement
group, the latest radiographs at follow-up were compared with 2-week postoperative
radiographs as described by Wacht et al.[23 ]
Robles-Molina et al and Vandenberghe et al did not describe their radiological assessment
within their studies.
Surgical Technique
Thorkildsen et al describe using a hydroxyapatite coated Elektra total joint replacement
(Small Bone Innovations International, ZA Les Bruyères, Pèronnas, France), which was
inserted by a single expert level 3 according to Tang and Giddins criteria[24 ] surgeon under general anesthetic via a dorsal approach. Trapeziectomy with LRTI
was also performed by using a dorsal approach. The trapezium was removed in two pieces
and flexor carpi radialis (FCR) was harvested and inserted into the trapezial void.
Capsular and skin closure were identical in both procedures. The postoperative rehabilitation
was identical in both groups.
Cebrian-Gomez et al describe using a cementless modular Ivory system (Stryker, Memometal,
Bruz, France) prosthesis, which was inserted via the dorsal approach by a single surgeon
with expertise level 3 according to Tang and Giddins criteria. In the trapeziectomy
group, the operative technique was similar to that described by Burton and Pellegrini.[25 ] Both procedures were undertaken under regional anesthesia.
Robles-Molina et al describe using the Arpe cementless prosthesis introduced by Biomet
(Valence, France) via the Wagner anterolateral approach.[26 ] A modified Burton-Pellegrini technique was performed in the trapeziectomy with LRTI
group, removing the trapezium via a lateral approach. Unlike the original technique,
the full thickness of the FCR tendon was used to suspend the first metacarpal through
a bone tunnel, and no intermetacarpal stabilization with Kirschner wire was performed.
In this study, there was neither mention of the number of surgeons nor their expertise
level.
Vandenberghe et al state that two different prosthesis were used: a cemented prosthesis
(De La Caffiniere [Stryker, Howmedical]) and the Roseland (DePuy International Ltd,
Leeds, England). A radiopalmar approach was used. In the trapeziectomy group, a Burton-Pellegrini
technique was employed via a dorsal approach.
Complications
Thorkildsen et al experienced complications in nine patients: six in the joint replacement
group and three in the trapeziectomy with LRTI group. There were two cup loosenings,
resulting in one being revised to trapeziectomy with LRTI and the other to a cemented
revision polyethylene cup. Three patients had dislocations. Of these, one patient
required a closed reduction only, the second was revised to trapeziectomy with LRTI
and the third patient underwent an arthrodesis. One patient was treated for suspected
periprosthetic joint infection and required a two-stage revision with the final procedure
being an arthrodesis. In the trapeziectomy group, one patient had transitory hematoma
in the forearm and two patients experienced long-term pain, but no cause was identified.[9 ] Cebrian-Gomez et al state that seven patients (8.3%) in the prosthesis group had
complications, including one superficial infection, two cases of dysesthesia in the
superficial radial nerve, and one algodystrophy. Two patients had dislocations, one
of which trapezial cup loosening and both required revision surgery to LRTI and removal
of metacarpal stems. One patient had his prosthesis revised to a new prosthesis. In
the trapeziectomy group, there were five patients (9.7%) with complications, but no
revisions. Two patients had collapse of the trapeziometacarpal space with continuous
pain, two had painful degenerative changes at the scaphotrapezial joint and one had
algodystrophy. Robles-Molina et al state that five patients in their prosthesis group
experienced complications. Two patients had dysesthesia of the superficial branch
of the radial nerve. Dislocations were observed in three patients who all underwent
revision surgery with prosthesis removal and LRTI. In the trapeziectomy group, four
patients experienced complications. Two patients had dysesthesia of the superficial
branch of the radial nerve, and two patients required further surgery to correct metacarpophalangeal
hyperextension.
Discussion
The results of this systematic review and meta-analysis demonstrate that joint replacement
confers a statistically significant benefit in physical function and symptoms compared
with trapeziectomy with LRTI, as evidenced by superior QDASH and Kapandji scores.
QDASH uses 11 items to measure physical function and symptoms in patients with musculoskeletal
disorders of the upper limb, so is regarded as a useful tool for assessing overall
upper limb performance. Joint replacement is shown to give significantly better thumb
opposition than trapeziectomy with LRTI, as demonstrated by a superior Kapandji score.
When other tools are used to study particular aspects of function and symptoms, however,
the differences between the two treatment modalities become less clear. No significant
differences in grip strength or pinch strength have been demonstrated by our meta-analysis.
Furthermore, we have found that neither treatment modality confers superior pain relief,
given their similar VAS scores.
Overall, the current body of evidence does not reveal convincing data to show that
joint replacement should supersede trapeziectomy with LRTI. Some may highlight the
better QDASH and Kapandji scores for joint replacement as evidence for its superiority.
However, certain caveats must be borne in mind. First, the mean difference in QDASH
score was 4.32 (95% CI: 1.80–6.83) in favor of joint replacement, but this is much
lower than the mean clinically important difference of 15 that was used in the only
RCT that has so far been conducted.[9 ] Second, any benefit in function and symptoms must be balanced against the higher
risk of complications which appear to come with joint replacement.
In a case series of 39 patients who received an Elektra joint replacement, there was
a 24% revision rate at 36 months.[27 ] Moreover, the dislocation rates have been estimated at 7%.[28 ] In comparison, the revision rate for trapeziectomy (including those with and without
LRTI) is said to be 4.6% at 10 years.[29 ] It may be hard to justify the small functional and symptom improvement conferred
by joint replacement when revision rates are much higher than that of trapeziectomy.
Moreover, trapeziectomy with LRTI has been shown to confer a greater degree of thumb
opposition than joint replacement, with equivalent VAS pain scores, pinch strength,
and grip strength.
There are several limitations to this systematic review and meta-analysis. First,
a total of only 583 patients have been studied in the literature. In addition, only
one RCT exists to date, and this was conducted in a single center, with all operations
performed by a single surgeon.[9 ] Furthermore, this study only had a 2-year follow-up and was subject to selection
and performance bias.[9 ] Moreover, the studies included in this meta-analysis used different implants and
implantation techniques in their joint replacement groups. Questions remain as to
which implant type and implantation method (cemented or uncemented) gives the best
outcome.
This systematic review and meta-analysis therefore highlights the need for a more
detailed and expansive comparisons of joint replacement with trapeziectomy and LRTI.
A priority is the need for large, adequately powered multicenter RCTs which are double
blinded. These must also have a sufficiently long follow-up to detect late complications,
including polyethylene wear and implant loosening. Such studies would also be useful
for investigating the properties of different components, such as cemented and uncemented
stems and cups, as well as metal and ceramic heads. Additionally, they would be required
to provide clear data on the frequency of the various complications of joint replacement.
Once the benefits of joint replacement can be more confidently ascertained it would
also be pertinent to perform cost-benefit analyses to see whether the additional cost
of joint replacement are justified.
Conclusion
The existing literature suggests that joint replacement gives superior upper limb
function and symptom improvement than trapeziectomy with LRTI, but the benefit does
not appear to be clinically significant. Moreover, trapeziectomy with LRTI appears
to confer a greater degree of thumb opposition than joint replacement. Our study reveals
that both treatment options are valid. While arthroplasty is less commonly done, we
have identified that it carries a higher risk of complications than trapeziectomy
with LRTI. As such, we advocate that this procedure is only performed by surgeons
who perform them regularly to reduce the risk of complications.
