EDITOR’S CHOICE

Advancements in de Quervain
Tenosynovitis Management: A
Comprehensive Network Meta-Analysis
Han Hong Chong, MBChB, MSc,* Akhilesh Pradhan, BSc (Hons), MBBS,* Mohit Dhingra, BMedSci, MBChB,*
William Liong, MS,† Melinda Y. T. Hau, MBChB, iBSc (Hons),* Rohi Shah, BMBS, MSc SEM‡

Purpose This study presents a network meta-analysis aimed at evaluating nonsurgical treat-
ment modalities for de Quervain tenosynovitis. The primary objective was to assess the
comparative effectiveness of nonsurgical treatment options.
Methods The systematic review was conducted following Preferred Reporting Items for
Systematic Reviews and Meta-Analysis (PRISMA) guidelines. Searches were performed in
multiple databases, and studies meeting predefined criteria were included. Data extraction,
risk of bias assessment, and statistical analysis were carried out to compare treatment mo-
dalities. The analysis was categorized into short-term (within six weeks), medium-term (six
weeks up to six months), and long-term (one year) follow-up.
Results The analysis included 14 randomized controlled trials encompassing various treatment
modalities for de Quervain tenosynovitis. In the short-term, extracorporeal shockwave therapy
demonstrated statistically significant improvement in visual analog scale pain scores compared with
placebo. Extracorporeal shockwave therapy also ranked highest in the treatment options based on its
treatment effects. Corticosteroid injections (CSIs) combined with casting and laser therapy with
orthosis showed favorable outcomes. Corticosteroid injection alone, platelet-rich plasma injections
alone, acupuncture, and orthosis alone did not significantly differ from placebo in visual analog
scale pain score. In the medium-term, extracorporeal shockwave therapy remained the top-ranking
option for visual analog scale pain score, followed by CSI with casting. In the long-term (one year),
CSI alone and platelet-rich plasma injections demonstrated sustained pain relief. Combining CSI
with orthosis also appeared promising when compared with CSI alone.
Conclusions Corticosteroid injection with a short duration of immobilization remains the pri-
mary and effective treatment for de Quervain tenosynovitis. Extracorporeal shockwave
therapy can be considered a secondary option. Alternative treatment modalities, such as
isolated therapeutic injection, should be approached with caution because they did not show
substantial benefits over placebo. (J Hand Surg Am. 2024;49(6):557e569. Copyright Ó 2024
by the American Society for Surgery of the Hand. All rights reserved.)
Type of study/level of evidence Therapeutic I.
Key words Comparative effectiveness, de Quervain’s, DQT, network meta-analysis, outcome.

From the *Department of Trauma & Orthopaedic, University Hospitals of Leicester National Corresponding author: Han Hong Chong, MBChB, MSc, Department of Trauma & Or-
Health Service (NHS) Trust, Leicester, United Kingdom; the †Department of Orthopaedic, thopaedic, University Hospitals of Leicester National Health Service (NHS) Trust, Infirmary
Hospital Shah Alam, Shah Alam, Selangor, Malaysia; and ‡Department of Trauma & Or- Square, Leicester LE1 5WW, United Kingdom; e-mail: [email protected].
thopaedic, Nottingham University Hospitals National Health Service (NHS) Trust, Nottingham, 0363-5023/24/4906-0005$36.00/0
United Kingdom. https://doi.org/10.1016/j.jhsa.2024.03.003
Received for publication October 13, 2023; accepted in revised form March 6, 2024.

Ó 2024 ASSH r Published by Elsevier, Inc. All rights reserved. r 557

558 DE QUERVAIN’S NETWORK META-ANALYSIS

D
E QUERVAINTENOSYNOVITIS (DQT) is a condition Search strategy
characterized by the thickening and myxoid The search strategy was conducted by a clinical
degeneration of the tendon sheath located librarian, as described in Appendix S1 (available
within the first dorsal compartment of the wrist.1,2 online on the Journal’s website at www.jhandsurg.
This leads to painful entrapment of the abductor org). We performed searches in MEDLINE,
pollicis longus and extensor pollicis brevis tendons. Embase, Emcare, Cochrane Central Register of
This condition is often attributed to repetitive overuse Controlled Trials, and Cumulative Index to Nursing
of the wrist in ulnar deviation while extending or and Allied Health Literature, covering the period
abducting the thumb, although it is also associated from their inception up to August 2023. Additionally,
with conditions such as pregnancy and rheumatoid we reviewed the reference lists of identified studies
arthritis.3,4 A study by Walker-Bone et al5 revealed and previous systematic reviews to identify further
that the prevalence of DQT is approximately 1.3% in relevant studies for potential inclusion.
women and 0.5% in men, with peak prevalence
among those in their 40s and 50s.
A range of treatment options are available for Selection of studies
DQT, encompassing both nonsurgical and surgical Four authors (A.P., M.D., W.L., and M.Y.T.H.)
approaches.6,7 Initial treatment typically involves conducted the screening and selection of studies. The
rest, ice, nonsteroidal anti-inflammatory drugs authors collaborated in pairs, with each pair allocated
(NSAIDs), physiotherapy, and splinting. Corticoste- an equal distribution of identified articles for review.
roid injections (CSIs) are considered the best practice This approach minimized the number of articles that
for patients who do not respond adequately to these each pair needed to review, with the aim to decrease
nonsurgical measures. Other reported treatments bias and improve inter-rater reliability. The articles
include acupuncture, platelet-rich plasma (PRP) in- retrieved through the searches underwent an initial
jections, NSAID injections, laser therapy, and extra- screening based on a review of their titles and ab-
corporeal shockwave therapy (ECSWT).8e13 Surgical stracts, followed by a more comprehensive review of
intervention is reserved for cases where nonsurgical the full texts. Any disagreements that arose during
treatments have proven ineffective and may involve this process were resolved by the senior authors
various techniques for releasing the first dorsal (H.H.C. and R.S.). The inter-rater reliability was
compartment.14 assessed using Cohen k score.
Although recent systematic reviews have focused
on the efficacy of CSI as a treatment for DQT, most Eligibility criteria
published randomized controlled trials (RCTs) have To ensure the comparability of the studies included in
primarily compared CSI against splinting or our analysis, we considered RCTs or quasi-RCTs that
casting.14e17 Limited evidence exists for comparing compared various treatments for DQT patients. These
other nonsurgical or surgical treatments. This has studies were deemed eligible if they met the
resulted in a gap in the evidence regarding the following criteria:
optimal treatment pathway for DQT. We therefore
conducted a systematic review and network meta-
Available in English or English-translated articles.
analysis (NMA) to comprehensively evaluate the
Sample: adult patients (aged 18 years) with
comparative effectiveness of various available DQT clinical diagnosis of DQT.
treatments. This approach allowed us to combine
Interventions: removable orthosis, casting (plaster
both direct and indirect evidence from a network of of Paris or fiberglass), physiotherapy, injection
published treatments, providing a more comprehen- therapy (eg, NSAID, CSI, and PRP), ECSWT,
sive assessment of their relative efficacy. laser therapy, ultrasound therapy, and acupuncture.

Comparison: control/placebo or other included in-
terventions mentioned.
METHODS
Primary outcome: visual analog scale (VAS) pain
We conducted and reported this systematic review in score.
adherence to the guidelines outlined in the Preferred
Secondary outcome: Disabilities of the Arm,
Reporting Items for Systematic Reviews and Meta- Shoulder, and Hand (DASH)/QuickDASH score.
Analysis (PRISMA) statement (Table S1, available Quasi-RCT is a study design resembling an RCT,
online on the Journal’s website at www.jhandsurg. but participant allocation to treatment and control
org).18,19 groups is not strictly randomized. Instead,

J Hand Surg Am. r Vol. 49, June 2024

 DE QUERVAIN’S NETWORK META-ANALYSIS 559

nonrandom methods such as alternation, birth dates, for each outcome during each of the three specified
hospital identification, or availability are employed. time periods.
Studies were required to provide sufficient data for
extraction and pooling, ideally reporting mean and RESULTS
SD.19 In cases where studies reported outcomes using The initial search produced a total of 2,754 results.
the median and range, we calculated the relevant Following the screening and comprehensive review
mean and SD using the statistical calculator devel- of full-text articles, 14 RCTs were identified that met
oped by Tong et al,20 Luo et al,21 Shi et al,22 and Wan the predefined inclusion criteria (Fig. 1); no quasi-
et al.23 RCTs were encountered for inclusion.8e12,27e35
During the screening process, there was agreement
Data extraction on 94% of screened studies between primary re-
Information from the included studies was retrieved viewers, resulting in the k score of 0.56, suggesting
and organized into a standardized data extraction moderate agreement.
template. The extracted data encompassed study These 14 studies encompassed a collective sample
characteristics, treatment modalities, comparative in- size of 823 patients, with individual study sample
terventions, the outcome of interest, and the duration sizes ranging from five to 67 participants. The anal-
of follow-up. ysis compared 12 distinct treatment modalities.
Table 1 summarizes the characteristics of the
Risk of bias assessment included studies.
Four authors (A.P., M.D., W.L., and M.H.) con- RoB analysis
ducted independent assessments of the risk of bias Figure S1 (available online on the Journal’s website
(RoB) for each study using the Cochrane Risk of Bias at www.jhandsurg.org) shows the RoB scores for the
tool 2.0 (ROB 2).24 In cases where disagreements included studies. Overall, three studies were classi-
arose among the authors, consensus was reached fied as serious RoB, three as moderate RoB, and eight
through discussion with senior authors (H.H.C. and as low RoB.
R.S.). The results were visually represented through
both traffic light plots and summary plots using the Short-term: follow-up within 6 weeks
robvis online tool.25 Nine studies were included to construct a compre-
hensive NMA of the VAS pain score within six
Statistical analysis weeks of follow-up, with a total of 499 patients
Age and follow-up duration were presented in the (range between five and 60) included and eight
form of either means (SD; range) or medians (range), different treatment modalities.9,11,12,27,30e33,35
consistent with the original articles. In our analysis, Figure S2 (available online on the Journal’s web-
we performed direct and indirect comparisons of in- site at www.jhandsurg.org) comprises both the
terventions through the frequentist NMA utilizing a network plot and the forest plot.
random effects model with Metainsight V1.1.26 To We generated a ranking matrix that relied on the
convey the results for continuous data, we utilized the comparative treatment effects, as presented in
mean difference along with a 95% CI. Network lea- Table 2. Among all the interventions, ECSWT
gue tables were produced to show details of the re- emerged as the top-ranking option, displaying supe-
sults of the comparisons between the interventions. rior treatment effects in terms of VAS pain score
Heterogeneity among studies was evaluated through when compared with the alternatives. Following
a thorough examination of the articles, visual in- closely were CSI in conjunction with casting and
spection of forest plots illustrating treatment effects laser therapy combined with removable orthosis, both
and their CIs, and an assessment of inconsistency. of which demonstrated favorable outcomes. In
The timeline for analysis follow-up was divided contrast, CSI alone, PRP injection alone, acupunc-
into three categories: studies that evaluated the ture, and orthosis alone were positioned below the
outcome within six weeks, at six weeks and up to six placebo in the rankings.
months, and at one year after intervention. In cases The study by Sharma et al8 was excluded from
where studies provided multiple outcome time points the NMA because their study did not establish a
within the same category (eg, both three and six network connection with the other studies (ultra-
months), the later follow-up time point was selected sound therapy vs laser therapy). Both interventions
for analysis. Separate analysis models were executed demonstrated significant improvements in VAS

J Hand Surg Am. r Vol. 49, June 2024

560 DE QUERVAIN’S NETWORK META-ANALYSIS

FIGURE 1: Preferred Reporting Items for Systematic Reviews and Meta-Analysis flowchart. CINAHL, Cumulative Index to Nursing
and Allied Health Literature.

pain score at the 2-week post intervention mark. The scarcity of data resulted in limited loop
However, there was no statistically significant dif- closure within the network, causing a lack of overlap
ference observed when comparing these two in- between direct and indirect evidence. The incapacity
terventions with each other. to conduct a consistency assessment restricts robust
Six studies were included to construct a compre- evaluation of the coherence regarding the compara-
hensive NMA of the DASH/QuickDASH score tive effectiveness of different interventions.
within six weeks of follow-up, with a total of 291
patients (range between nine and 32) included and six Medium-term: follow-up at 6 weeks and up to 6 months
different treatment modalities (Figure S3, available Eleven studies were included to construct a compre-
online on the Journal’s website at www.jhandsurg. hensive NMA of the VAS pain score at six weeks and
org).11,12,27,30,31,33 CSI with or without casting up to six months of follow-up.9,10,12,27e33,35 These
ranked above all other interventions, followed by studies encompassed a total of 654 patients, with
PRP injection, laser therapy with orthosis, acupunc- individual study sample sizes ranging from five to 67
ture, and orthosis alone (Table 3). participants, and investigated nine distinct treatment

J Hand Surg Am. r Vol. 49, June 2024

 TABLE 1. Characteristics of Studies*
Reported
Duration of Sample Outcome of Follow-
Study/Year Inclusion Criteria Exclusion Criteria Age (y) Onset Intervention Size (n) Interest Up

Peters- Age >18 y, radial border pain, Trauma, neoplastic, 52.3 (SD, 12.6) Not Sham (NaCl) 5 VAS 1y
Veluthamaningal crepitus over radial complications/reactions, 51.2 (SD, 20.2) specified CSI 7
et al,32 2009 styloid, þve Finkelstein test, unable
no previous CSI in last 6 mo to follow-up
Mehdinasab and Radial border pain, tenderness at Trauma, 32.6 (21e61) 5.59 wk CSI þ cast 37 VAS 6 mo
Alemohammad,28 the first dorsal compartment, history of fracture, (SD, 3.61)
2010 þve Finkelstein test/WHAT rheumatological disorder, 6.45 wk Cast 36
test, no previous CSI pregnancy (SD, 3.43)
Hadianfard Radial border Onset <4 wk, trauma, history 40.7 (22e76) >4 wk CSI 15 VAS, 6 wk
et al,12 2014 pain and/or swelling, þve of fracture/surgery, systemic Acupuncture 15 QuickDASH
Finkelstein test, disorder, abnormal blood

DE QUERVAIN’S NETWORK META-ANALYSIS

J Hand Surg Am.

no previous CSI/NSAID test/radiography, pregnant

Mardani-Kivi Age >18 y, radial border pain, Trauma, history of surgery, 44 (SD, 13) Not CSI 30 VAS, 6 mo
et al,27 2014 tenderness at the first dorsal rheumatological/neurological/ specified CSI þ cast 29 QuickDASH
compartment, þve Finkelstein dermatological disorder,
test, pain score >6, no pregnant, complications/
previous CSI in last 6 mo reaction, infection
r

þve Finkelstein test

Vol. 49, June 2024

Sharma History of fracture, cervical 36.6 (21e45) 2.7 mo USS therapy 15 VAS 2 wk
et al,8 2015 spondylosis, rheumatological/ (SD, 1.6)
systemic disorder 3.4 mo LILT 15
(SD, 1.6)
Kumar,35 2020 Radial border pain, tenderness Trauma, history of fracture/ 42.5 (SD, 16) Not specified CSI þ cast 60 VAS, 6 mo
at the first dorsal surgery, rheumatological/ 47 (SD, 17) CSI 60 QuickDASH
compartment, þve Finkelstein neurological disorder,
test, pain score >6, no previous pregnant
CSI in last 6 mo
Akhtar et Age 30e60 y, radial border Trauma, history of surgery, 40.73 (SD, 9.2) Not specified CSI þ cast 67 VAS, QuickDASH 6 wk
al,29 2020 pain, þve Finkelstein test, neoplasm, absolute 41.44 (SD, 8.5) Cast 67
failed 6 wk of conservative contraindication
therapy (oral or local NSAIDs),
no CSI
Ippolito et Age >18 y, radial border pain, Trauma, history of surgery, 46 Not specified CSI þ cast 11 VAS, DASH 6 mo
al,31 2020 tenderness at the first dorsal rheumatological/ CSI 9
compartment, þve Finkelstein dermatological/neurological
test, VAS pain score >4, no disorder, taking analgesia,
previous CSI in last 6 mo pregnant, complications/
reactions

(Continued)

561
 562
TABLE 1. Characteristics of Studies* (Continued)
Reported
Duration of Sample Outcome of Follow-
Study/Year Inclusion Criteria Exclusion Criteria Age (y) Onset Intervention Size (n) Interest Up

Das et al,30 2021 Age 30e50 y, no CSI in last 6 mo Trauma, history of fracture, Not specified <6 wk Orthosis 30 VAS, QuickDASH 6 mo
rheumatological/congenital/ CSI 30
systemic disorder, local
infection, pregnant
Haghighat Age >18 y, radial border Trauma, history of fracture/ 44.61 (SD, 11.36) Not specified ECSWT 13 VAS, DASH 6 wk
et al,9 2021 pain, þve Finkelstein test, no surgery, neurological/ 48.21 (SD, 14.45) Sham 13
previous physiotherapy/CSI in rheumatological/coagulation
last 1 mo disorder, unable to follow-up,
complications
Başar et al,34 2021 Tenderness at the first dorsal Trauma, history for fracture/ 50.6 (SD, 12.5) Not specified CSI þ orthosis 42 VAS, QuickDASH 1y
compartment, þve Finkelstein/ surgery, rheumatological/ 43.8 (SD, 11.6) CSI 34

DE QUERVAIN’S NETWORK META-ANALYSIS

J Hand Surg Am.

WHAT test, no previous CSI systemic disorder, pregnant

Suwannaphisit Age >18 y, radial border Trauma, history of surgery, 54.5 (SD, 10) 30 d (SD, 37) NSAID injection‡ 31 VAS, DASH 6 wk
et al,10 2022 pain, þve Finkelstein test/ neoplastic, inflammatory 54 (SD, 14) 30 d (SD, 47.1) CSI 29
WHAT test, no previous CSI in disorder, complications/
last 6 mo reaction, unable to follow-up
Kumar et al,33 þve Finkelstein test, no Trauma, presence or arthritis, 37.8 9 (SD, 6.44) Not specified CSI 30 VAS, DASH 1y
r

2022 previous CSI Dupuytren disease,

Vol. 49, June 2024

35.83 (SD, 8.48) PRP 30

rheumatological/inflammatory/
systemic disorder, pregnant
Dundar Ahi and Unilateral wrist pain with clinical Trauma, history of CSI/surgery, 40.4 (SD, 8.9) 6 mo (3e18) HILT þ orthosis 30 VAS, QuickDASH 5 wk
Sirzai,11 2023† diagnosis of DQT cervical radiculopathy/ 37.9 (SD, 8.4) 6 mo (3e24) Sham þ orthosis 32
myelopathy, neurological/
systemic/rheumatological
disorder

þve, positive; HILT, high-intensity laser therapy; LILT, low-intensity laser therapy; NaCl, sodium chloride; PRP, platelet-rich plasma; USS, ultrasound; WHAT, wrist hyperflexion and abduction of the thumb.
*Data are presented as n, mean (range) or mean (SD), unless otherwise indicated.
†Median and range were used to report VAS result. Data were converted to mean and SD using Tong20 calculator.
‡NSAID injection used by author was ketorolac.
 TABLE 2. Results (Mean Differences With 95% CIs) of the Pairwise and NMA for VAS Pain Score Within 6 Weeks of Follow-Up
ECSWT 3.25 (4.31 to
2.19)*
2.18 (4.62 to CSI þ TSC 1.20 (1.62 to
0.25) 0.77)*

DE QUERVAIN’S NETWORK META-ANALYSIS

J Hand Surg Am.

2.98 (5.92 to 0.80 (2.54 to HILT þ Orthosis 1.70 (3.16 to

0.04)* 0.95) 0.24)*
3.25 (4.31 to 1.07 (3.26 to 0.27 (3.01 to Placebo 0.13 (2.29 to 2.03)
2.19)* 1.13) 2.47)
3.38 (5.78 to 1.20 (1.62 to 0.40 (2.10 to 0.13 (2.29 to CSI 0.40 (1.75 to 1.37 (2.71 to 1.30 (2.17 to
r

0.98)* 0.77)* 0.03)* 0.43)*

Vol. 49, June 2024

1.30) 2.03) 0.95)

3.78 (6.53 to 1.60 (3.01 to 0.80 (2.97 to 0.53 (3.07 to 0.40 (1.75 to 0.95) PRP Injection
1.03)* 0.18)* 1.37) 2.01)
4.75 (7.50 to 2.57 (3.97 to 1.77 (3.93 to 1.50 (4.04 to 1.37 (2.71 to 0.97 (2.87 to Acupuncture
2.00)* 1.16)* 0.39) 1.04) 0.03)* 0.93)
4.68 (7.23 to 2.50 (3.46 to 1.70 (3.16 to 1.43 (3.76 to 1.30 (2.17 to 0.90 (2.50 to 0.07 (1.53 to Orthosis
2.13)* 1.53)* 0.24)* 0.90) 0.43)* 0.70) 1.67)

HILT, high-intensity laser therapy; PRP, protein-rich plasma; TSC, thumb spica cast.
*Statistically significant difference.

563
564 DE QUERVAIN’S NETWORK META-ANALYSIS

modalities (Figure S4, available online on the Jour-

15.08 (28.16 to 2.00)*

nal’s website at www.jhandsurg.org).

6.50 (18.93 to 5.93)

On the network league table (Table 4), ECSWT
remained the top-ranking option, followed by CSI

Orthosis
with casting when compared with placebo. Cortico-
steroid injection alone, PRP injection, acupuncture,
Results (Mean Differences With 95% CIs) of the Pairwise and NMA for DASH/QuickDASH Score Within 6 Weeks of Follow-Up

casting alone, and orthosis alone were all ranked

below the placebo. Notably, NSAID injection occu-
pied the lowest position in the table, demonstrating a
statistically significant worsening in VAS pain scores
10.60 (23.42 to 2.22)

4.48 (22.79 to 13.83)

compared with the other interventions during the

follow-up period of up to six months.
Acupuncture

We included data from eight studies to construct a

comprehensive NMA focusing on the DASH/Quick-
DASH score within a follow-up period of six weeks
up to six months.10,12,27,29e31,33,35 These studies
involved a total of 543 patients, with individual study
sample sizes ranging from nine to 67 patients, and
examined seven distinct treatment modalities
2.02 (24.15 to 20.11)
6.50 (18.93 to 5.93)

(Figure S5, available online on the Journal’s website

HILT þ Orthosis

at www.jhandsurg.org). The ranked league table

(Table 5) showed a similar trend to the VAS pain
score table, with CSI with casting ranking above all
other interventions as the top-performing option.
Haghighat et al9 was not considered in the NMA
because of the absence of a network connection with
the other studies, specifically in comparing ECSWT
0.53 (11.65 to 10.59)

8.05 (29.24 to 13.14)

10.07 (27.04 to 6.90)
14.55 (31.72 to 2.62)

versus placebo. It is noteworthy that at the 6-week

follow-up, ECSWT demonstrated a statistically sig-
PRP Injection

nificant superiority in improving the DASH score

when compared with placebo.
HILT, high-intensity laser therapy; PRP, protein-rich plasma; TSC, thumb spica cast.

The sparse data resulted in minimal loop closure

and an inability to perform a consistency assessment,
constraining a thorough evaluation of coherence in
terms of comparative effectiveness between
15.08 (28.16 to 2.00)*

interventions.
0.53 (11.65 to 10.59)

10.60 (23.42 to 2.22)

5.30 (15.28 to 4.67)

8.58 (26.62 to 9.46)

Long term: follow-up at 1 year

Peters-Veluthamaningal et al32 (2009) (CSI vs pla-
CSI

cebo), Başar et al34 (2021) (CSI vs CSI with

orthosis), and Kumar et al33 (2022) (PRP vs CSI)
reported their outcomes at one year after intervention.
In summary, the improvement in VAS pain scores
*Statistically significant difference.

was observed to be sustained for up to one year with

CSI alone and PRP injection. Additionally, there is a
20.38 (36.83 to 3.94)*

possibility of achieving a better outcome when

13.88 (34.50 to 6.73)
15.90 (32.15 to 0.34)
5.30 (15.28 to 4.67)
5.83 (20.77 to 9.10)

combining CSI with orthosis when compared with

CSI alone as reported by Başar et al.34

DISCUSSION
CSI þ TSC
TABLE 3.

Despite the plethora of RCTs and subsequent meta-

analyses conducted, the search for an optimal treat-
ment strategy for DQT remains inconclusive, with a

J Hand Surg Am. r Vol. 49, June 2024

 TABLE 4. Results (Mean Differences With 95% CIs) of the Pairwise and NMA for VAS Pain Score at 6 Weeks and up to 6 Months of Follow-Up
ECSWT 3.50 (5.89 to
1.11)*
3.67 (8.03 to CSI þ TSC 1.06 (2.42 to 3.32 (5.37 to
0.68) 0.31) 1.27)*
3.50 (5.89 to 0.17 (3.46 to Placebo 1.23 (4.60 to
1.11)* 3.81) 2.14)

DE QUERVAIN’S NETWORK META-ANALYSIS

J Hand Surg Am.

4.33 (9.07 to 0.66 (3.36 to 0.83 (4.93 to PRP Injection 0.40 (2.73 to
0.41) 2.04) 3.27) 1.93)
4.73 (8.86 to 1.06 (2.42 to 1.23 (4.60 to 0.40 (2.73 to CSI 0.10 (2.43 to 0.87 (3.46 to 4.60 (7.20 to
0.60)* 0.31) 2.14) 1.93) 2.23) 1.72) 2.00)*
4.83 (9.57 to 1.16 (3.86 to 1.33 (5.43 to 0.50 (3.79 to 0.10 (2.43 to Orthosis
r

0.09)* 1.54) 2.77) 2.79) 2.23)

Vol. 49, June 2024

5.60 (10.48 to 1.93 (4.85 to 2.10 (6.35 to 1.27 (4.75 to 0.87 (3.46 to 0.77 (4.25 to Acupuncture
0.72)* 1.00) 2.15) 2.21) 1.72) 2.71)
6.99 (11.81 to 3.32 (5.37 to 3.49 (7.67 to 2.66 (6.05 to 2.26 (4.73 to 2.16 (5.55 to 1.39 (4.97 to TSC
2.18)* 1.27)* 0.68) 0.72) 0.20) 1.22) 2.18)
9.33 (14.21 to 5.66 (8.59 to 5.83 (10.09 to 5.00 (8.49 to 4.60 (7.20 to 4.50 (7.99 to 3.73 (7.40 to 2.34 (5.92 to NSAID
4.45)* 2.72)* 1.57)* 1.51)* 2.00)* 1.01)* 0.06)* 1.25) Injection

PRP, protein-rich plasma; TSC, thumb spica cast.

*Statistically significant difference.

565
566 DE QUERVAIN’S NETWORK META-ANALYSIS

lack of robust evidence favoring one intervention

29.70 (38.54 to
Results (Mean Differences With 95% CIs) of the Pairwise and NMA for DASH/QuickDASH Score at 6 Weeks and up to 6 Months of Follow-Up
over the others. Extracorporeal shockwave therapy
leads to short and medium-term outcomes, but

20.86)*

Injection
limited trials against placebo necessitate further

NSAID
research. Corticosteroid injection with casting has
demonstrated moderate evidence, supporting its effi-
cacy in improving VAS pain scores and DASH/
QuickDASH outcomes, both in the short and medium
term when compared with placebo and other modal-
16.55 (23.42 to

20.73 (32.68 to
ities. Despite receiving a higher ranking, indicating
potential superior effectiveness, the presence of the
9.68)*

8.79)*
95% CI including zero suggests insufficient statistical
evidence for a definitive conclusion, particularly in
TSC

comparisons against placebo. Therefore, these results

should be interpreted cautiously, taking into account
other pertinent information and factors when deciding
26.00 (38.00 to

on interventions.
3.70 (11.82 to

5.27 (16.69 to

Orthosis alone, casting alone, CSI alone, laser

Acupuncture

therapy, and PRP injection rank lower than placebo

14.00)*

in VAS pain scores, signaling less preference and

4.42)

6.15)

uncertain benefits. Similarly, in DASH/QuickDASH

scores, orthosis alone, casting alone, laser therapy,
and PRP injection fare worse than CSI alone. Lower
rankings and the 95% CI including zero imply, on
27.66 (40.86 to
2.04 (11.84 to

1.66 (14.38 to

6.93 (19.60 to

average, less preference and uncertainty in their

effectiveness. Patients considering these treatments
14.46)*

for DQT should be informed that they may not pro-

11.06)
Orthosis
7.76)

5.74)

vide substantial benefits. NSAID injection showed a

statistical trend toward worsening the outcome mea-
sures. We would therefore recommend caution with
the use of these treatment modalities for DQT.
The Cochrane review conducted in 2009 provided
29.70 (38.54 to
7.58 (11.74 to

2.04 (11.84 to

3.70 (11.82 to

8.97 (17.00 to
0.64 (5.87 to

an initial evaluation of the available evidence, albeit

with limitations.17 It concluded that there was low-
20.86)*
3.42)*

0.93)*

level evidence supporting the superiority of CSI

4.59)

7.76)

4.42)

over the use of splints for pain relief in DQT. This

CSI

conclusion was primarily based on a single small-

scale RCT characterized by poor study quality.
Following this initial review, subsequent RCTs
studied various interventions for DQT, some of
30.34 (40.61 to
2.68 (13.78 to

4.34 (13.99 to

9.61 (19.19 to
0.64 (5.87 to

PRP, protein-rich plasma; TSC, thumb spica cast.

which are included in our NMA. Ashraf and Deva-

PRP Injection

doss15 conducted a study that directly compared CSI

20.07)*
0.02)*

with splinting and reached a similar conclusion to our

4.59)

8.42)

5.31)

findings, favoring CSI injection alone as an effective

*Statistically significant difference.

treatment option. Rowland et al16 investigated CSI in

comparison to both placebo and other active treat-
ments. Their study concluded in favor of the superi-
11.28 (20.40 to

16.55 (23.42 to

37.28 (47.05 to

ority of CSI, further supporting the evidence in this

6.94 (13.62 to

7.58 (11.74 to

9.62 (20.27 to

direction.16
CSI þ TSC

Cavaleri et al36 explored the effectiveness of CSI

27.51)*
TABLE 5.

0.26)*

3.42)*

2.16)*

9.68)*

when combined with hand therapy, concluding that

1.02)

CSI alongside orthosis provided better outcomes than

either intervention alone. However, it is noteworthy

J Hand Surg Am. r Vol. 49, June 2024

 DE QUERVAIN’S NETWORK META-ANALYSIS 567

that in their study, orthosis, casting, acupuncture, and selection. Additionally, we conducted a meticulous
dry needling were all classified under the umbrella of review of the methodology and characteristics of the
“hand therapy.” Although there is existing literature included studies to confirm their suitability for inte-
suggesting positive outcomes associated with phys- gration into the NMA. Our assessment indicated that
iotherapy for DQT, it is important to note that these these studies exhibited sufficient methodological
studies often feature small sample sizes and exhibit similarity to be included in the NMA.
methodological shortcomings.36e38 We also noted a Our NMA has several limitations that should be
gap in the literature because no RCTs were identified taken into consideration when interpreting the find-
that directly compared physiotherapy with any other ings. First, the studies included in our analysis
interventions. exhibited heterogeneity, which could potentially
In our analysis, we noted a lack of agreement affect the reliability and generalizability of our re-
among the included studies regarding the type, sults. One possible source of this heterogeneity is the
duration, and strictness of immobilization for DQT variation in symptom duration among the studies
treatment. It is important to highlight that the because the response to treatment may differ between
included studies that implemented immobilization acute and chronic cases of DQT. Despite categori-
primarily employed a thumb spica cast with either zation efforts, differences in treatment protocols,
Plaster of Paris or fiberglass with varying durations including injection dosage and orthosis types, were
(2e5 weeks). In clinical practice, rigid immobiliza- observed. Data limitations prevent accounting for
tion is often considered excessive owing to limited these variations, emphasizing the need to consider
patient acceptance. The Menendez et al39 RCT these sources of heterogeneity when interpreting our
comparing full-time splinting to patient-desired findings.
splinting found no significant outcome difference at Six of the 14 studies had a RoB categorized as
7.5 weeks, indicating that strict rest is not disease- “moderate” or “serious”, potentially affecting the
modifying for DQT. This indirectly questions the reliability and robustness of our review findings. The
necessity of rigid casting for immobilization. absence of a thorough evaluation of publication bias
In 2014, the European HANDGUIDE Study group may also affect the overall robustness, acknowl-
conducted a Delphi study involving 112 experts, edging potential selective reporting that could influ-
including 52 hand surgeons, 47 hand therapists, and ence synthesized evidence. This limitation should be
13 physical medicine and rehabilitation physicians.40 considered when interpreting and generalizing the
Their expert consensus outlined a treatment hierar- findings.
chy, suggesting initial instructions with NSAIDs, In the context of individual treatment arms, the
followed by splinting, CSI, and, if necessary, surgery. majority of RCTs compared interventions against
The experts often recommended a combination of CSI, with only one RCT comparing against PRP,
modalities, considering factors like pain severity, ECSWT, and laser therapy. Consequently, the wide
symptom duration, and prior interventions. Their CIs associated with these comparisons limit the
systematic review aligns with our NMA results, strength of any definitive conclusions that can be
emphasizing limitations in guiding DQT treatment drawn. The restricted loop closure in the network
decisions due to available evidence constraints. The resulted in insufficient overlap between direct and
experts’ consensus emphasizes the importance of a indirect evidence, thereby constraining the capacity to
tailored approach based on individual patient char- evaluate consistency across various interventions. We
acteristics and clinical circumstances while also anticipate that future research will provide additional
underscoring the need for further research to establish RCTs focusing on these treatments, which will
more robust treatment guidelines for DQT, address- contribute to more robust evidence.
ing existing uncertainties in this field. It is worth noting that there were no available
Network meta-analysis offers the advantage of RCTs comparing physiotherapy against other treat-
allowing comparisons between every treatment ment options in our analysis. Additionally, the studies
method for DQT against each other, in contrast to the included in our analysis did not provide sufficient
traditional pairwise meta-analysis. We categorized evidence to ascertain whether the patient cohort had
our investigation into different follow-up timings, previously undergone any form of physiotherapy
focusing on treatment effectiveness at these specific intervention. This gap in the literature highlights the
time points. To ensure comparability, we strictly need for further research in these areas to better un-
adhered to predefined criteria for study design, pop- derstand their comparative effectiveness in managing
ulation, intervention, and outcomes during study DQT.

J Hand Surg Am. r Vol. 49, June 2024

568 DE QUERVAIN’S NETWORK META-ANALYSIS

To address the ongoing controversies surrounding tenosynovitis: a double-blind randomized clinical trial. J Hand Surg
Am. 2017;42(9):S45eS46.
DQT treatment, there is a need for further high- 14. Larsen CG, Fitzgerald MJ, Nellans KW, Lane LB. Management of de
quality RCTs. One aspect to consider is the com- Quervain tenosynovitis: a critical analysis review. JBJS Rev.
parison of the efficacy of specific, standardized 2021;9(9):e21.00069.
physiotherapy, CSI with removable orthosis, and 15. Ashraf MO, Devadoss VG. Systematic review and meta-analysis on
steroid injection therapy for de Quervain’s tenosynovitis in adults.
ECSWT in treating DQT across various stages (acute, Eur J Orthop Surg Traumatol. 2014;24(2):149e157.
subacute, and chronic). Such research initiatives 16. Rowland P, Phelan N, Gardiner S, Linton KN, Galvin R. The
would contribute substantially to resolving the un- effectiveness of corticosteroid injection for De Quervain’s stenosing
tenosynovitis (DQST): a systematic review and meta-analysis. Open
certainties in this pathology and guide clinicians in Orthop J. 2015;9(1):437e444.
making informed treatment decisions. 17. Peters-Veluthamaningal C, Van Der Windt DAWM, Winters JC,
Meyboom-de Jong B. Corticosteroid injection for de Quervain’s
tenosynovitis. Cochrane Database Syst Rev. 2009;(3):CD005616.
CONFLICTS OF INTEREST 18. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020
No benefits in any form have been received or will be statement: an updated guideline for reporting systematic reviews.
BMJ. 2021;372:n71.
received related directly to this article. 19. Higgins JPT, Thomas J, Chandler J, et al. Cochrane Handbook for
Systematic Reviews of Interventions version 6.3 (Updated February
2022). Published 2022. Accessed November 4, 2022. https://training.
ACKNOWLEDGMENTS cochrane.org/handbook
The authors would like to thank Pip Divall, Clinical 20. Tong T. (Online Calculator) Estimating the sample mean and stan-
Librarian of University Hospital of Leicester National dard deviation (SD) from the five-number summary and its applica-
tion in meta-analysis. Accessed November 28, 2022. https://www.
Heatlh Service (NHS) Trust, United Kingdom, for her math.hkbu.edu.hk/wtongt/papers/median2mean.html
assistance in conducting literature search for this 21. Luo D, Wan X, Liu J, Tong T. Optimally estimating the sample mean
meta-analysis. from the sample size, median, mid-range, and/or mid-quartile range.
Stat Methods Med Res. 2018;27(6):1785e1805.
22. Shi J, Luo D, Wan X, et al. Detecting the skewness of data from the
REFERENCES five-number summary and its application in meta-analysis. Stat
Methods Med Res. 2023;32(7):1338e1360.
1. Ahuja NK, Chung KC. Fritz de Quervain, MD (1868-1940): sten- 23. Wan X, Wang W, Liu J, Tong T. Estimating the sample mean and
osing tendovaginitis at the radial styloid process. J Hand Surg Am. standard deviation from the sample size, median, range and/or
2004;29(6):1164e1170. interquartile range. BMC Med Res Methodol. 2014;14(1):135.
2. Clarke MT, Lyall HA, Grant JW, Matthewson MH. The histopa- 24. Sterne JAC, Savovic J, Page MJ, et al. RoB 2: a revised tool for
thology of de Quervain’s disease. J Hand Surg Br. 1998;23(6): assessing risk of bias in randomised trials. BMJ. 2019;366:l4898.
732e734. 25. McGuinness LA, Higgins JPT. Risk-of-bias VISualization (robvis):
3. Weiss APC, Akelman E, Tabatabai M. Treatment of de Quervain’s an R package and Shiny web app for visualizing risk-of-bias as-
disease. J Hand Surg Am. 1994;19(4):595e598. sessments. Res Synth Methods. 2021;12(1):55e61.
4. Schned ES. DeQuervain tenosynovitis in pregnant and postpartum 26. Owen RK, Bradbury N, Xin Y, Cooper N, Sutton A. MetaInsight: an
women. Obstet Gynecol. 1986;68(3):411e414. interactive web-based tool for analyzing, interrogating, and visual-
5. Walker-Bone K, Palmer KT, Reading I, Coggon D, Cooper C. izing network meta-analyses using R-shiny and netmeta. Res Synth
Prevalence and impact of musculoskeletal disorders of the upper limb Methods. 2019;10(4):569e581.
in the general population. Arthritis Rheum. 2004;51(4):642e651. 27. Mardani-Kivi M, Karimi Mobarakeh M, Bahrami F, Hashemi-
6. Ilyas AM. Nonsurgical treatment for de Quervain’s tenosynovitis. Motlagh K, Saheb-Ekhtiari K, Akhoondzadeh N. Corticosteroid in-
J Hand Surg Am. 2009;34(5):928e929. jection with or without thumb spica cast for de Quervain tenosyno-
7. Pensak MJ, Bayron J, Wolf JM. Current treatment of de Quervain vitis. J Hand Surg Am. 2014;39(1):37e41.
tendinopathy. J Hand Surg Am. 2013;38(11):2247e2249; quiz 2250. 28. Mehdinasab SA, Alemohammad SA. Methylprednisolone acetate
8. Sharma R, Aggarwal AN, Bhatt S, Kumar S, Bhargava SK. Outcome injection plus casting versus casting alone for the treatment of de
of low level lasers versus ultrasonic therapy in de Quervain’s teno- Quervain’s tenosynovitis. Arch Iran Med. 2010;13(4):270e274.
synovitis. Indian J Orthop. 2015;49(5):542e548. 29. Akhtar M, Faraz Ul Hassan Shah Gillani S, Nadeem RD,
9. Haghighat S, Vahdatpour B, Ataei E. The effect of extracorporeal Tasneem M. Methylprednisolone acetate injection with casting
shockwave therapy on de quervain tenosynovitis; a clinical trial. versus casting alone for the treatment of De-Quervain’s tenosyno-
Shiraz E-Med J. 2021;22(8):106559. vitis: a randomized controlled trial. J Pak Med Assoc. 2020;70(8):
10. Suwannaphisit S, Suwanno P, Fongsri W, Chuaychoosakoon C. 1314e1318.
Comparison of the effect of ketorolac versus triamcinolone acetonide 30. Das R, Bimol N, Deb D, Meethal SA, Singh YN. Efficacy of thumb
injections for the treatment of de Quervain’s tenosynovitis: a double- abduction orthosis versus local methylprednisolone acetate injection
blind randomized controlled trial. BMC Musculoskelet Disord. in de quervain’s tenosynovitis: a randomized controlled trial. J Med
2022;23(1):831. Soc. 2021;35(1):35e39.
11. Dundar Ahi E, Sirzai H. Short-term effectiveness of high-intensity 31. Ippolito JA, Hauser S, Patel J, Vosbikian M, Ahmed I. Nonsurgical
laser therapy in De Quervain tenosynovitis: a prospective, random- treatment of De Quervain tenosynovitis: a prospective randomized
ized, controlled study. Medeni Med J. 2023;38(1):24e31. trial. Hand (N Y). 2020;15(2):215e219.
12. Hadianfard M, Ashraf A, Fakheri M, Nasiri A. Efficacy of 32. Peters-Veluthamaningal C, Winters JC, Groenier KH, Meyboom-
acupuncture versus local methylprednisolone acetate injection in De Dejong B. Randomised controlled trial of local corticosteroid in-
Quervain’s tenosynovitis: a randomized controlled trial. J Acupunct jections for de Quervain’s tenosynovitis in general practice. BMC
Meridian Stud. 2014;7(3):115e121. Musculoskelet Disord. 2009;10(1):131.
13. Chadderdon C, Gaston RG, Loeffler BJ, Lewis D. Betamethasone 33. Kumar V, Talwar J, Rustagi A, Krishna LG, Sharma VK. Com-
versus ketorolac injection for the treatment of De Quervain’s parison of clinical and functional outcomes after platelet-rich

J Hand Surg Am. r Vol. 49, June 2024

 DE QUERVAIN’S NETWORK META-ANALYSIS 569

plasma injection and corticosteroid injection for the treatment of 37. Katana B, Jaganjac A, Bojicic S, et al. Effectiveness of physical
de Quervain’s tenosynovitis. J Wrist Surg. 2022;12(2):135e142. treatment at De Quervain᾽s disease. J Health Sci. 2012;2(1):80e84.
34. Başar B, Aybar A, Basar G, Başar H. The effectiveness of cortico- 38. Goel R, Abzug JM. de Quervain’s tenosynovitis: a review of the
steroid injection and splint in diabetic de Quervain’s tenosynovitis rehabilitative options. Hand (N Y). 2015;10(1):1e5.
patients: a single-blind, randomized clinical consort study. Medicine. 39. Menendez ME, Thornton E, Kent S, Kalajian T, Ring D.
2021;100(35):e27067. A prospective randomized clinical trial of prescription of full-time
35. Kumar R. Management of de Quervain tendinitis using corticosteroid versus as-desired splint wear for de Quervain tendinopathy. Int
injection (CSI) with or without thumb spica cast (TSC). Eur J Mol Orthop. 2015;39(8):1563e1569.
Clin Med. 2020;7(8):5635e5640. 40. Huisstede BMA, Coert JH, Fridén J, Hoogvliet P, European
36. Cavaleri R, Schabrun SM, Te M, Chipchase LS. Hand therapy versus HANDGUIDE Group. Consensus on a multidisciplinary treatment
corticosteroid injections in the treatment of de Quervain’s disease: a guideline for de Quervain disease: results from the European
systematic review and meta-analysis. J Hand Ther. 2016;29(1):3e11. HANDGUIDE study. Phys Ther. 2014;94(8):1095e1110.

J Hand Surg Am. r Vol. 49, June 2024