Meta-analysis

Comparative Effectiveness of Different Nonsurgical

Treatments for Patellar Tendinopathy: A Systematic
Review and Network Meta-analysis
Po-Cheng Chen, M.D., Kuan-Ting Wu, M.D., Wen-Yi Chou, M.D., Yu-Chi Huang, M.D.,
Lin-Yi Wang, M.D., Tsung-Hsun Yang, M.D., Ka-Kit Siu, M.D., and
Yu-Kang Tu, D.D.S., Ph.D.

Purpose: To investigate the functional improvement and pain reduction of different nonsurgical treatments for patellar
tendinopathy (PT), a systematic review with network meta-analysis was performed. Methods: Studies were compre-
hensively searched for without language restrictions in the CENTRAL, MEDLINE, EMBASE, Web of Science, Physio-
therapy Evidence Database, and SPORTDiscus databases from inception to May 2018. Randomized controlled trials about
nonsurgical treatments for PT were included. The outcome measurements were the Victorian Institute of Sports
Assessment (VISA) scale and pain scores (such as the visual analog scale or Numerical Rating Scale). Study quality was
evaluated using the Physiotherapy Evidence Database score. Direct comparisons were performed using pairwise meta-
analysis, whereas network meta-analysis was performed using a frequentist method in a multivariate random-effects
model. Results: Eleven studies with 430 affected patellar tendons were included in the systematic review. The sum-
mary mean difference of improvement in the VISA scale versus the control group for corticosteroid injection was e23.00
(95% confidence interval [CI] e36.73 to -9.27), for leukocyte-rich platelet-rich plasma (LR-PRP) was 13.22 (95% CI 2.37-
24.07), for focused extracorporeal shockwave therapy (ESWT) was e1.28 (95% CI e6.25 to 3.68), for radial ESWT was
e6.68 (95% CI e20.20 to 6.84), for ultrasound was e0.70 (95% CI e11.23 to 9.83), for autologous blood injection was
e0.60 (95% CI e9.30 to 8.10), for dry needling was 17.51 (95% CI e2.57 to 37.60), for topical glyceryl trinitrate was
e0.90 (95% CI e13.07 to 11.27), and for skin-derived tendon-like cells was 10.40 (95% CI e1.59 to 22.39). LR-PRP
(Surface Under the Cumulative Ranking curve [SUCRA] ¼ 87.5%) or dry needling (SUCRA ¼ 90.5%) was most likely
to be ranked the best in terms of improvement on the VISA scale. Compared with the control group, the summary mean
difference of the change in pain score for corticosteroid injection was 0.80 (95% CI e3.48 to 5.08), for LR-PRP was e1.87
(95% CI e3.28 to e0.46), for focused ESWT was 0.13 (95% CI e0.68 to 0.93), for radial ESWT was 0.03 (95% CI e1.92 to
1.98), for ultrasound was e0.20 (95% CI e1.49 to 1.09), for autologous blood injection was 0.60 (95% CI e0.73 to 1.93),
for dry needling was e0.37 (95% CI e2.71 to 1.97), and for topical glyceryl trinitrate was e0.50 (95% CI e2.55 to 1.55).
The treatment most likely to be ranked the best in terms of change in pain score was LR-PRP (SUCRA ¼ 94.9%).
Conclusions: The network meta-analysis demonstrated that LR-PRP has the greatest functional improvement and pain
reduction for PT compared with other treatment options. However, the treatment effect estimates can be biased by the
possible intransitivity and should not be overestimated. Level of Evidence: Level I, meta-analysis of Level I studies.

See commentary on page 3132

From the Department of Physical Medicine and Rehabilitation, Kaohsiung Po-Cheng Chen and Kuan-Ting Wu contributed equally to this work and
Chang Gung Memorial Hospital and Chang Gung University College of should be considered co-first authors.
Medicine, Kaohsiung (P-C.C., Y-C.H., L-Y.W., T-H.Y.); Department of Or- Received December 11, 2018; accepted June 13, 2019.
thopedic Surgery, Kaohsiung Chang Gung Memorial Hospital and Chang Address correspondence to Yu-Chi Huang, M.D., Department of Physical
Gung University College of Medicine, Kaohsiung (K-T.W., W-Y.C., K-K.S.); Medicine and Rehabilitation, Kaohsiung Chang Gung Memorial Hospital,
Medical Mechatronic Engineering Program, Cheng Shiu University, Kaoh- Chang Gung University College of Medicine, 123 Ta-Pei Road, Niao-Sung,
siung (W-Y.C.); Institute of Epidemiology and Preventive Medicine, College of Kaohsiung, Taiwan. E-mail: [email protected]
Public Health, National Taiwan University, Taipei (Y-K.T.); and Department Ó 2019 by the Arthroscopy Association of North America
of Dentistry, National Taiwan University Hospital, Taipei (Y-K.T.), Taiwan. 0749-8063/181503/$36.00
The authors report that they have no conflicts of interest in the authorship https://doi.org/10.1016/j.arthro.2019.06.017
and publication of this article. Full ICMJE author disclosure forms are
available for this article online, as supplementary material.

Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 35, No 11 (November), 2019: pp 3117-3131 3117
3118 P-C. CHEN ET AL.

P atellar tendinopathy (PT) is commonly seen among

elite and recreational athletes owing to repetitive
activities.1,2 The overall prevalence of PT was reported
reduction in patients with PT has not been compared at
the same time in a meta-analysis. The aim of this
network meta-analysis was to investigate the functional
to be 8.5%, with greater rates in volleyball and soccer improvement and pain reduction of different nonsur-
players, according to one cross-sectional study.3 The gical treatments for PT. The hypothesis of this study was
chronic and refractory nature of PT can affect athletes’ that PRP would be the most effective treatment in
training frequencies and competitive performance. functional improvement and pain reduction for patients
There is no consensus with regards to the gold-standard with PT.
treatment for PT, and conservative treatment is usually
the first-line therapy, including adequate rest, nonste-
roidal anti-inflammatory drugs, local cryotherapy, Methods
stretching and eccentric exercise training, patellar Search Methods for Identification of Studies
tendon straps, or therapeutic ultrasound.4,5 In the This systematic review was conducted following the
literature, eccentric exercise training has been identified rules of the Preferred Reporting Items for Systematic
as a stimulation treatment with the capability to in- Reviews and Meta-Analyses23 and was registered pro-
crease remodeling of collagen fibers in the tendon in spectively in the PROSPERO International Prospective
response to stress. Functional improvement and pain- Register of Systematic Reviews (registration number:
reducing effect of this training in the early stages of CRD 42018086696). We comprehensively searched the
PT have been demonstrated in several studies.6,7 Cochrane Central Register of Controlled Trials (CEN-
Patellar tendon straps are effective in reducing pain TRAL), MEDLINE, EMBASE, Web of Science, Physio-
and increasing proprioception of knee joints in patients therapy Evidence Database (PEDro), and SPORTDiscus
with PT, but the study designs are caseecontrol, cohort, from inception to May 31, 2018. We also searched us-
or single-arm with pretestepost-test research de- ing the World Health Organization International Clin-
signs.8-10 Therapeutic ultrasound was used to treat ical Trials Registry Platform search portal, which
chronic PT in 1 randomized controlled trial11 but did includes various trial registers such as ISRCTN and
not provide additional benefit over placebo treatment. ClinicalTrials.gov, to identify further studies. The search
For patients with refractory symptoms, further strategies are presented in Appendix 1. Search filters
nonsurgical interventions, including dry needling, developed by the Hedges Project (http://hiru.mcmaster.
corticosteroid injection, extracorporeal shockwave ca/hiru/HIRU_Hedges_home.aspx) were employed to
therapy (ESWT), and platelet-rich plasma (PRP) have achieve the best balance of sensitivity and specificity.
been administered for reactivation of the repair process; References included in identified studies also were
however, the reported results are inconsistent.12-15 searched to avoid non-inclusion of related clinical trials.
Regarding ESWT, mechanical load is considered a ma- All of the literature search was performed in June 2018.
jor factor in the reactivation of tissue regeneration.
ESWT is considered a good alternative to surgery in Inclusion Criteria
patients who decline surgery or are poor surgical can- Types of Study
didates.16 It also has shown good analgesic effect for The study was level of evidence 1 for all included
chronic tendinitis around the shoulder and elbow,17,18 studies and included randomized controlled trials.
but the outcome of ESWT in PT remains inconsis- Studies comparing different types of exercise program
tent.12,19 In the category of PRP, in which whole blood were excluded. Studies were excluded if the follow-up
drawn from the patient is spun in a centrifuge and cells period was shorter than 3 months. Studies also were
are separated out of the blood, the condensed platelets excluded if outcome measures could not be extracted.
in PRP can initiate the healing process in the tendon
when they reach the site of injury.20 In a meta-analysis, Types of Participant
PRP showed promising results in cases of refractory Participants with clinically- or image-confirmed PT
PT.21 However, despite PRP exhibiting positive results were included. Participants who had received surgery,
for the treatment of PT, Liddle and Rodriguez-Mer- corticosteroid injection, ESWT, or other interventions
chan22 reported no significant superiority of PRP over for affected knees previously were excluded. To avoid
other alternative treatments. In addition, the standard interference from other chronic diseases, patients with
protocol of PRP therapy has not yet been established rheumatoid arthritis, septic arthritis, or other inflam-
with regards to the number and interval of injections matory arthritis were excluded.
for PT.
After thorough searching of the existed literature, we Types of Intervention
found that the effectiveness of different nonsurgical There were no restrictions in terms of dose or dura-
treatment options for functional improvement and pain tion of nonsurgical treatments. As eccentric exercise
 TREATMENTS FOR PATELLAR TENDINOPATHY 3119

training usually is prescribed along with other conser- Data Collection and Analysis
vative management techniques, it could be considered a Selection of Studies
control group in these comparisons. There was no re- Two authors (one physiatrist, P-C. C., and one ortho-
striction regarding the type, frequency, intensity, or pedic surgeon, K-T. W.) independently screened the titles
duration of eccentric exercise training. and abstracts of studies and identified those that could be
included. After initial screening, the full text of studies
Outcome Measures
with the potential to be included was reviewed by at least
The outcome measures assessed in our systematic
one of the authors for further selection. Any ongoing trials
review were improvement in functional and pain
identified primarily through the World Health Organiza-
scores for each treatment strategy for PT. The severity
tion International Clinical Trials Registry Platform also
of PT can be evaluated by a simple and practical
were followed-up until data analysis began in June 2018.
questionnaire-based index, the Victorian Institute of Discrepancies in the selection of studies were resolved
Sport Assessment (VISA) questionnaire.24 This brief
through discussion.
questionnaire assesses symptoms, function, and ability
to play sport. The VISA score ranges from 0 (greatest Data Extraction and Management
severity) to 100 points (an asymptomatic, fully- Two authors independently extracted the following
performing individual). We also focused on patellar data:
pain symptoms during activities of daily living and 1. Author and publication year.
used the visual analog scale (VAS) or Numeric Rating 2. Journal of publication.
Scale (NRS) for evaluation. The pain severity on the 3. Inclusion and exclusion criteria.
VAS or NRS is rated from 0 (no pain) to 10 points 4. Participant characteristics, such as age, sex (male/
(worst pain). female), and duration of symptoms of PT.

Fig 1. Flow diagram of the

literature search and identi-
fication of articles for review.
(VISA, Victorian Institute of
Sport Assessment.)
Table 1. Characteristics of the Included Studies

3120
Diagnosis
Confirmed Sex (Male/ Duration of Level of
Study by Imaging N Age, y Female) Symptoms Treatment Follow-up Outcome Measures Evidence
Warden et al., 200811 Yes 17 27  7 12/5 3.4  3.1 y Therapeutic ultrasound 12 wk (1) VAS-U (usual pain) and VAS-W I
20 27  7 18/2 4.1  3.8 y Control (worst pain)
(2) VISA
Kongsgaard et al., 200915 Yes 12 34.3  10.0 12/0 18.3  14.1 mo Corticosteroid injection 12, 24 wk (1) VISA I
12 31.3  8.3 12/0 18.8  13.0 mo Control (2) VAS
(3) Ultrasonographic findings
(4) Biomechanical outcomes
Clarke et al., 201139 Yes 33 Unknown Unknown Unknown SDTLC þ ABI 6 wk, 3 mo, (1) VISA I
27 Unknown Unknown Unknown ABI 6 mo (2) Ultrasonographic findings
Zwerver et al., 2011 19
No 31 24.2  5.2 20/11 7.3  3.6 mo fESWT 1, 12, 22 wk (1) VISA I
31 25.7  4.5 21/10 8.1  3.8 mo Control (2) VAS during ADL, sports, during 1
and 10 single-legged decline squats,
after 3 maximum single-legged
jumps, and after the triple-hop test
(3) Improvement ratio
Steunebrink et al., 201342 No 16 31.9  9.6 11/5 47  39 wk TGT 6, 12, 24 wk (1) VISA I
17 33.8  10.5 14/3 49  36 wk Control (2) VAS
(3) Satisfaction rate
van der Worp et al., 201444 Yes 21 28.8  10.3 16/5 32.3  28.7 mo fESWT 7, 14 wk (1) VISA I
33.4  10.7 38.6  56.9 mo

P-C. CHEN ET AL.

22 16/6 rESWT (2) VAS during ADL, sport, 1 single-
leg decline squat, 10 single-leg
decline squats
Vetrano et al., 201345 Yes 23 26.9  9.1 20/3 18.9  19.1 mo LR-PRP 2, 6, 12 mo (1) VISA I
23 26.8  8.5 17/6 17.6  20.2 mo fESWT (2) VAS
(3) Modified Blazina scale
Dragoo et al., 201440 Yes 9 28  8 8/1 Unknown LR-PRP 12, 26 wk (1) VISA I
12 40  14 12/0 Unknown Dry needling (2) Tegner
(3) Lysholm
(4) VAS
(5) SF-12
Resteghini et al., 201641 Yes 11 38.9  8.9 10/1 16.7  9.7 mo ABI 1, 3, 12 mo (1) VAS I
11 42.5  12.9 8/3 19.2  12.2 mo Control (2) VISA
(3) SF-MPQ
Lee et al., 20184 Yes 16 21.1  2.2 16/0 35.6  22.4 mo fESWT 12 wk (1) VISA I
14 24.1  4.6 14/0 31.5  30.0 mo Control (2) VAS during activity
(3) Tendon force and mechanical
properties
Thijs et al., 201743 No 22 30.5  8.0 14/8 65.1  72.7 wk fESWT 6 wk, 12 wk, (1) VISA I
30 27.3  5.2 24/6 99.4  126.3 wk Control 24 wk (2) NRS during 10 decline squats, 3
single-leg jumps, and 3 maximal
vertical jumps
(3) Satisfaction ratio
ABI, autologous blood injection; ADL, activities of daily living; fESWT, focused extracorporeal shockwave therapy; LR-PRP, leukocyte-rich platelet-rich plasma; NRS, Numerical Rating Scale;
rESWT, radial extracorporeal shockwave therapy; SDTLC, skin-derived tenocyte-like cell; SF-12, Short Forme12; SF-MPQ, Short-FormdMcGill Pain Questionnaire; TGT, topical glyceryl
trinitrate; VAS, visual analog scale; VISA, Victorian Institute of Sports Assessment score.
 TREATMENTS FOR PATELLAR TENDINOPATHY 3121

5. Details of interventions and treatment strategies. Dealing with Missing Data

6. Outcome measures. We used data from intention-to-treat analysis25
7. Risk of bias. whenever possible; otherwise, we used the data avail-
If any unclear or missing information were encoun- able from the included studies. For numerical variables,
tered, we contacted the authors of the individual pa- the median value was substituted for the mean value if
pers for assistance. Any differences in opinion were the latter was unavailable. Standard deviations (SDs)
resolved through discussion with the corresponding were input from P values according to guidance given
authors. in the Cochrane Handbook for Systematic Reviews of
Intervention.26 If only means and SDs for baseline and
Assessment of Risk of Bias in Included Studies follow-up measurements for each group were reported,
We used the PEDro scale to evaluate the quality of we calculated the changes in the means and SDs for
the included studies. The PEDro scale contains 11 each group,27,28 as per the following mathematical
items that result in a score of 10. It also includes formula:
additional questions regarding the follow-up of study
subjects and between-group statistical comparisons. Change in mean ¼ Meanpost  Meanpre
The risk of bias of the included studies is summarized qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
in Appendix 2. Change in SD ¼ SD2pre þ SD2post  2 r SDpre SDpost

Measures of Treatment Effect Meanpre and SDpre are the mean and SD for the baseline
Because VISA and pain scores (VAS or NRS) are measurement, and Meanpost and SDpost are the mean
numerical variables, we displayed these outcome and SD for the follow-up measurement; r is the corre-
measures as the mean difference (MD) with a 95% lation between the matched pairs of baseline and
confidence interval (CI). follow-up measurements, and was set at r ¼ 0.5 for
each group.
Unit of Analysis Issues
The unit of analysis was knees affected by PT ac- Data Synthesis
cording to the intervention group to which they were Data were extracted into a spreadsheet software
randomly assigned. (Excel version 2013, Microsoft, Redmond, WA) for

Fig 2. Forest plot of the

traditional pairwise meta-
analysis for improvement on
the VISA scale for patellar
tendinopathy. (ABI, autolo-
gous blood injection; CI,
confidence interval; CSI,
corticosteroid injection; DN,
dry needling; fESWT, focused
extracorporeal shockwave
therapy; LR-PRP, leukocyte-
rich platelet-rich plasma;
PRP, platelet-rich plasma;
rESWT, radial extracorporeal
shockwave therapy; SDTLC,
skin-derived tendon-like
cells; TGT, topical glyceryl
trinitrate; WMD, weighted
mean difference.)
3122 P-C. CHEN ET AL.

Fig 3. Forest plot of the

traditional pairwise meta-
analysis for change in pain
score for patellar tendin-
opathy. (ABI, autologous
blood injection; CI, confi-
dence interval; CSI, cortico-
steroid injection; DN, dry
needling; fESWT, focused
extracorporeal shockwave
therapy; LR-PRP, leukocyte-
rich platelet-rich plasma;
PRP, platelet-rich plasma;
rESWT, radial extracorporeal
shockwave therapy; SDTLC,
skin-derived tendon-like
cells; TGT, topical glyceryl
trinitrate; WMD, weighted
mean difference.)

preliminary management and were later imported into Q test were calculated for evaluation of heterogeneity.
the statistical software STATA (StataCorp. 2013. Stata Subgroup analysis, meta-regression, or sensitivity
Statistical Software: Release 13; StataCorp LP College analysis was performed when substantial heterogeneity
Station, TX). Traditional pairwise meta-analysis for was recognized. Network meta-analysis compares
direct comparisons and network meta-analysis to multiple treatments by combining both direct and in-
combine direct and indirect evidence were undertaken. direct evidence,29,30 yielding more precise and robust
We used random-effects models for meta-analysis estimates under certain assumptions.31,32 The assump-
throughout this systematic review due to various tion of transitivity can be thought of as any participant
participant groups and treatment regimens. In the in the network could have been given any of the
traditional pairwise meta-analysis, I2 and the Cochrane treatments in the network. We used the frequentist

Fig 4. Network plot of the treatments for (A) improvement on the VISA scale and (B) change in pain score for patellar ten-
dinopathy. (ABI, autologous blood injection; CSI, corticosteroid injection; DN, dry needling; fESWT, focused extracorporeal
shockwave therapy; LR-PRP, leukocyte-rich platelet-rich plasma; rESWT, radial extracorporeal shockwave therapy; SDTLC, skin-
derived tendon-like cells; TGT, topical glyceryl trinitrate; VISA, Victorian Institute of Sport Assessment.)
Table 2. Summary of Pairwise and Network Meta-Analyses of Treatment Effectiveness in Terms of Improvement on the VISA Scale
Effect Expressed as Improvement on VISA Scale With 95% CI
Pairwise Meta-Analysis (B vs A)
B Control CSI LR-PRP fESWT rESWT Ultrasound ABI DN TGT SDTLC

TREATMENTS FOR PATELLAR TENDINOPATHY

Network A Control e23.00 (e36.73 e e1.27 (e6.24 e e0.70 (e11.05 e0.60 (e9.30 e e0.90 (e13.01 e
meta-analysis to e9.27) to 3.69) to 9.65) to 8.10) to 11.21)
(A vs B) CSI e23.00 (e36.73 e e e e e e e e
to e9.27)
LR-PRP 13.22 (2.37 36.22 (18.72 e14.50 (e24.14 e e e 4.30 (e13.97 e e
to 24.07) to 53.72) to e4.86) to 22.57)
fESWT e1.28 (e6.25 21.72 (7.12 e14.50 (e24.15 e5.40 (e18.00 e e e e e
to 3.68) to 36.32) to e4.85) to 7.20)
rESWT e6.68 (e20.20 16.32 (e2.95 e19.90 (e35.75 e5.40 (e17.98 e e e e e
to 6.84) to 35.59) to e4.05) to 7.18)
Ultrasound e0.70 (e11.23 22.30 (4.99 e13.92 (e29.04 0.58 (e11.06 5.98 (e11.16 e e e e
to 9.83) to 39.61) to 1.21) to 12.23) to 23.12)
ABI e0.60 (e9.30 22.40 (6.14 e13.82 (e27.72 0.68 (e9.33 6.08 (e10.00 0.10 (e13.56 e e 11.00 (2.74
to 8.10) to 38.65) to 0.09) to 10.70) to 22.16) to 13.76) to 19.26)
DN 17.51 (e2.57 40.51 (16.19 4.30 (e12.60 18.80 (e0.66 24.20 (1.03 18.21 (e4.46 18.11 (e3.77 e e
to 37.60) to 64.84) to 21.20) to 38.26) to 47.37) to 40.89) to 40.00)
TGT e0.90 (e13.07 22.10 (3.75 e14.12 (e30.42 0.38 (e12.76 5.78 (e12.41 e0.20 (e16.29 e0.30 (e15.26 e18.41 (e41.90 e
to 11.27) to 40.45) to 2.19) to 13.52) to 23.97) to 15.89) to 14.66) to 5.07)
SDTLC 10.40 (e1.59 33.40 (15.17 e2.82 (e18.99 11.68 (e1.29 17.08 (e0.99 11.10 (e4.86 11.00 (2.75 e7.12 (e30.50 11.30 (e5.78
to 22.39) to 51.63) to 13.35) to 24.66) to 35.15) to 27.06) to 19.25) to 16.27) to 28.38)
NOTE. Effect expressed as MD with 95% CI for network meta-analysis or pairwise meta-analysis.
ABI, autologous blood injection; CI, confidence interval; CSI, corticosteroid injection; DN, dry needling; fESWT, focused extracorporeal shock wave therapy; LR-PRP, leukocyte-rich platelet-
rich plasma; MD, mean difference; rESWT, radial extracorporeal shock wave therapy; SDTLC, skin-derived tendon-like cells; TGT, topical glyceryl trinitrate; VISA, Victorian Institute of Sport
Assessment.

3123
 3124
Table 3. Summary of Pairwise and Network Meta-Analyses of Treatment Effectiveness in Terms of the Change in Pain Score
Effect Expressed as Change of Pain Scores With 95% CI
Pairwise Meta-Analysis (B vs A)
B Control CSI LR-PRP fESWT rESWT Ultrasound ABI DN TGT
Network A Control 0.80 (e3.48 0.13 (e0.68 e0.20 (e1.49 0.60 (e0.73 e0.50 (e2.55
meta-analysis (A vs B) to 5.08) to 0.93) to 1.09) to 1.93) to 1.55)
CSI 0.80 (e3.48
to 5.08)
LR-PRP e1.87 (e3.28 e2.67 (e7.18 2.00 (0.84 1.50 (e0.35

P-C. CHEN ET AL.

to e0.46) to 1.84) to 3.16) to 3.35)
fESWT 0.13 (e0.68 e0.67 (e5.03 2.00 (0.84 e0.10 (e1.89
to 0.93) to 3.69) to 3.16) to 1.69)
rESWT 0.03 (e1.92 e0.77 (e5.48 1.90 (e0.22 e0.10 (e1.88
to 1.98) to 3.93) to 4.02) to 1.68)
Ultrasound e0.20 (e1.49 e1.00 (e5.47 1.67 (e0.24 e0.33 (e1.85 e0.23
to 1.09) to 3.47) to 3.58) to 1.19) (e2.57, 2.11)
ABI 0.60 (e0.73 e0.20 (e4.68 2.47 (0.53 0.47 (e1.08 0.57 (e1.79 0.80 (e1.05
to 1.93) to 4.28) to 4.41) to 2.02) to 2.93) to 2.65)
DN e0.37 (e2.71 e1.17 (e6.05 1.50 (e0.37 e0.50 (e2.70 e0.40 (e3.22 e0.17 (e2.84 e0.97 (e3.66
to 1.97) to 3.71) to 3.37) to 1.70) to 2.42) to 2.50) to 1.72)
TGT e0.50 (e2.55 e1.30 (e6.05, 1.37 (e1.12 e0.63 (e2.83 e0.53 (e3.36 e0.30 (e2.72 e1.10 (e3.54 e0.13 (e3.24
to 1.55) 3.45) to 3.86) to 1.57) to 2.30) to 2.12) to 1.34) to 2.98)
NOTE. Effect expressed as MD with 95% CI for network meta-analysis or pairwise meta-analysis.
ABI, autologous blood injection; CI, confidence interval; CSI, corticosteroid injection; DN, dry needling; fESWT, focused extracorporeal shock wave therapy; LR-PRP, leukocyte-rich platelet-
rich plasma; MD, mean difference; rESWT, radial extracorporeal shock wave therapy; SDTLC, skin-derived tendon-like cells; TGT, topical glyceryl trinitrate; VISA, Victorian Institute of Sport
Assessment.
 TREATMENTS FOR PATELLAR TENDINOPATHY 3125

Fig 5. Ranking of treatment

strategies based on (A) rank
probabilities and (B) cumu-
lative probabilities with
regards to their effects on
improvement on the VISA
scale. (ABI, autologous blood
injection; CSI, corticosteroid
injection; DN, dry needling;
fESWT, focused extra-
corporeal shockwave ther-
apy; LR-PRP, leukocyte-rich
platelet-rich plasma; rESWT,
radial extracorporeal shock-
wave therapy; SDTLC, skin-
derived tendon-like cells;
TGT, topical glyceryl trini-
trate; VISA, Victorian Insti-
tute of Sport Assessment.)

approach to network meta-analysis by specifying the Surface Under the Cumulative Ranking curve
consistency and inconsistency models as multivariate (SUCRA), as described by Salanti et al.,36 in which
random-effects regression models.33 We checked the SUCRA is 1 when a treatment is always the best and
inconsistency between direct and indirect comparisons 0 when a treatment is always the worst. All statistical
using the design-by-treatment interaction models, loop analyses were performed using the mvmeta package for
inconsistency models, and node-splitting models.34,35 the statistical software STATA. The statistical signifi-
For the ranking of treatments, we calculated the cance level was set at 5%.
3126 P-C. CHEN ET AL.

Assessment of Heterogeneity between-group statistical comparisons were reported

We carefully examined the characteristics and designs for at least 1 key outcome in all but one of the included
of the included studies for clinical and methodologic studies.39 Both point measures and measures of vari-
heterogeneity. Differences in patients’ ages, sex, and ability for at least 1 key outcome were provided in all
the duration of symptoms of PT could result in clinical included studies.
heterogeneity, and the risk of bias and methodologic
characteristics of included studies could contribute to Effects of Interventions
methodologic heterogeneity. If substantial heterogene- Pairwise Meta-Analysis (Direct Comparisons)
ity was recognized, we explored the causes of hetero- Figures 2 and 3 show the pooled estimates for each
geneity by subgroup analysis, sensitivity analysis, or outcome measure.
meta-regression.
Improvement on the VISA Scale. Eleven studies were
Assessment of Reporting Bias included in the analysis of the improvement on the
Funnel plot asymmetry is the most common tool used VISA scale for PT. The weighted MD of improvement
for evaluating reporting bias28,37; however, there was on the VISA scale was e1.27 (95% CI e6.24 to 3.69,
no single reference line against which symmetry could I2 ¼ 0%) for focused ESWT versus control studies.4,19,43
be judged owing to the estimated effects for different Other comparisons consisted of only one clinical trial,
comparisons, and a comparison-adjusted funnel plot and the MD of improvement on the VISA scale was
was used instead.38 e23.00 (95% CI e36.73 to e9.27) for corticosteroid
injection versus control study,15 14.50 (95% CI 4.86-
24.14) for leukocyte-rich (LR)-PRP versus focused
Results ESWT study,45 e0.70 (95% CI e11.05 to 9.65) for
ultrasound versus control study,11 e0.60 (95% CI
Study Selection and Description
e9.30 to 8.10) for autologous blood injection (ABI)
The details of the study selection process are presented
versus control study,41 e4.30 (95% CI e22.57 to
in Figure 1. We identified 865 references through elec-
13.97) for LR-PRP versus dry needling study,40 e0.90
tronic searches and other sources. Following screening of
(95% CI e13.01 to 11.21) for topical glyceryl
titles and abstracts, 822 inappropriate references were
trinitrate (TGT) versus control study,42 11.00 (95% CI
excluded. After reading the full text, we finally retrieved
2.74 to 19.26) for skin-derived tendon-like cells
11 references for further evaluation.4,11,15,19,39-45 These
versus ABI study,39 and 5.40 (95% CI e7.20 to
studies included 430 affected patellar tendons. Although
18.00) for focused ESWT versus radial ESWT study.44
1 study39 did not report the number of male and female
participants, participants in other included studies were Change in Pain Score. Ten studies were enrolled to
mostly male (295 males and 75 females). The average analyze the change in pain score for PT. The weighted
age of the participants was 21 to 42 years, and the MD of the change in pain score was 0.13 (95% CI e0.68
duration of symptoms varied from 7 months to 4 years. to 0.93, I2 ¼ 0%) for focused ESWT versus control
Most studies assessed outcome measures from 3 to studies.4,19,43 Other comparisons were composed of only
12 months. All included studies compared the one clinical trial, and the MD of change in pain score was
improvement on the VISA scale, whereas only 10 e2.00 (95% CI e3.16 to e0.84) for LR-PRP versus
studies4,11,15,19,40-45 compared the change in pain scores. focused ESWT study,45 0.60 (95% CI e0.73 to 1.93) for
The characteristics of the included studies are summa- ABI versus control study,41 e1.50 (95% CI e3.35 to
rized in Table 1. 0.35) for LR-PRP versus dry needling study,40 0.10
(95% CI e1.69 to 1.89) for focused ESWT versus radial
Quality of Studies ESWT study,44 0.80 (95% CI e3.48 to 5.08) for
Appendix 2 shows the methodologic quality of the corticosteroid injection versus control study,15 e0.50
included studies. The eligibility criteria were specified (95% CI e2.55 to 1.55) for TGT versus control study,42
in all studies. The allocation concealment methods and e0.20 (95% CI e1.49 to 1.09) for ultrasound
were described in 6 studies.11,19,40,42-44 The groups versus control study.11
were similar at baseline with regards to the most
important prognostic indicators in all studies except Network Meta-Analysis (Combination of Direct and In-
2.39,41 In terms of blind assessment, 8 studies11,19,39-44 direct Comparisons)
had blinding of all subjects, and 9 studies4,11,19,39-44 Figure 4 presents a network plot of treatments in
had blinding of all assessors, whereas 4 terms of improvement on the VISA scale and change in
studies11,39,41,42 mentioned blinding of all therapists pain score for PT. The area of each circle is proportional
who administered treatment. Two studies11,43 had a to the number of studies involving the specific treat-
dropout rate greater than 15%. Intention-to-treat ment. Any 2 circles connected by lines represent direct
analysis was used in 9 studies.11,19,39-45 The results of comparisons in the studies, and the thickness of each
 TREATMENTS FOR PATELLAR TENDINOPATHY 3127

Fig 6. Ranking of treatment

strategies based on (A) rank
probabilities and (B) cumula-
tive probabilities with regards
to their effects on the change
in pain score. (ABI, autolo-
gous blood injection; CSI,
corticosteroid injection; DN,
dry needling; fESWT, focused
extracorporeal shockwave
therapy; LR-PRP, leukocyte-
rich platelet-rich plasma;
rESWT, radial extracorporeal
shockwave therapy; SDTLC,
skin-derived tendon-like
cells; TGT, topical glyceryl tri-
nitrate; VISA, Victorian Insti-
tute of Sport Assessment.)

line is proportional to the number of comparisons Improvement on the VISA Scale. Compared with the
included in the network meta-analysis. control group, the summary MD of improvement on the
The results of the network meta-analysis, including VISA scale for corticosteroid injection was e23.00 (95%
the summary MDs with 95% CIs, are reported in CI e36.73 to e9.27), for LR-PRP was 13.22 (95% CI 2.37-
Table 2 and Table 3. The rank probabilities and cumu- 24.07), for focused ESWT was e1.28 (95% CI e6.25 to
lative probabilities are plotted in Figure 5 and Figure 6. 3.68), for radial ESWT was e6.68 (95% CI e20.20 to
3128 P-C. CHEN ET AL.

6.84), for ultrasound was e0.70 (95% CI e11.23 to 9.83), eccentric-specific strength of the quadriceps plays a
for ABI was e0.60 (95% CI e9.30 to 8.10), for dry significant role in protection of the patellar tendons
needling was 17.51 (95% CI e2.57 to 37.60), for TGT during sports activities.52
was e0.90 (95% CI e13.07 to 11.27), and for skin- ESWT has gained growing popularity for the treat-
derived tendon-like cells was 10.40 (95% CI e1.59 to ment of orthopedic conditions, especially tendino-
22.39). LR-PRP (SUCRA ¼ 87.5%) or dry needling pathies.53 Published papers have reported various
(SUCRA ¼ 90.5%) was most likely to be ranked the energy levels, numbers of treatment sessions, and
best in terms of improvement on the VISA scale. numbers of impulses for ESWT in the treatment of
tendinopathies, and there exist no guidelines with
Change in Pain Score. Compared with the control
regards to the recommended dose of ESWT. It is being
group, the summary MD of the change in pain score for
used more and more frequently in lower-limb tendi-
corticosteroid injection was 0.80 (95% CI e3.48 to
nopathies, although the mechanism of pain reduction
5.08), for LR-PRP was e1.87 (95% CI e3.28 to e0.46),
and functional improvement remains unclear. Possible
for focused ESWT was 0.13 (95% CI e0.68 to 0.93), for
mechanisms of ESWT include stimulation of healing,
radial ESWT was 0.03 (95% CI e1.92 to 1.98), for
neovascularization, suppressive effects on nociceptors,
ultrasound was e0.20 (95% CI e1.49 to 1.09), for
and a hyperstimulation mechanism blocking the gate-
ABI was 0.60 (95% CI e0.73 to 1.93), for dry
control mechanism.54-57 Recent evidence showed that
needling was e0.37 (95% CI e2.71 to 1.97), and for
there exists associations between tendon repair and
TGT was e0.50 (95% CI e2.55 to 1.55). The
increases in growth factors in patients receiving ESWT,
treatment most likely to be ranked the best in terms
and these growth factors are assumed to be responsible
of change in pain score was LR-PRP (SUCRA ¼ 94.9%).
for the success of treatment.58-60
Inconsistency in these outcomes was evaluated using
ABI and PRP represent one kind of regenerative
the design-by-treatment, loop inconsistency, and node-
medicine that has become popular in recent years. They
splitting models. As there was no loop in the network,
are used in the treatment of many orthopedic condi-
there was no need to evaluate the inconsistency.
tions, including muscle or ligament injuries, tendino-
Reporting Bias pathies and enthesopathies, osteoarthritis, and as an
Appendix Figure 1 presents the comparison-adjusted adjunct to surgical treatment.61 Several randomized
funnel plots of improvement on the VISA scale and controlled trials compared ABI with PRP for chronic
change in pain score. There was no significant funnel lateral epicondylitis,62-65 and PRP reduced pain and
plot asymmetry. The Egger test found no evidence of improved function better than ABI did. However, there
small study bias. was no direct comparison of ABI versus PRP for PT in
this systematic review. The primary effect of ABI or PRP
in the treatment of PT is enhancement of tissue-
Discussion healing. What differentiates ABI or PRP from ESWT
The network meta-analysis of improvement on the for the treatment of PT is that high concentrations of
VISA scale showed that LR-PRP or dry needling was growth factors are found within ABI or PRP,45 whereas
most likely to be ranked the best treatment, and the ESWT only triggers specific responses in the injured
network meta-analysis of the change in pain score tissue to increase the concentrations of growth factors.
demonstrated that LR-PRP had the greatest probability Dry needling involves repeated lancing of the area of
to be ranked the best. abnormal tendons. It has been evaluated as a treatment
PT, also called jumper’s knee, commonly affects the option for chronic tendinosis, such as Achilles tendi-
patellar tendon inserted on the inferior pole of the pa- nosis, lateral epicondylitis, or patellar tendinosis.66 It is
tella bone,46,47 and its symptoms are not usually performed to stimulate an inflammatory response
induced by an acute inflammatory process.48,49 PT is within the tendon, resulting in focal disruption of the
instead a degenerative change in the patellar tendon collagen fibers within the area of tendinosis and inciting
resulting in knee pain and dysfunction, and the favored internal hemorrhage. It is hypothesized that the in-
pathogenesis is chronic repetitive tendon overload that flammatory response induces the formation of granu-
contributes to increased fibroblast production of pros- lation tissue, strengthening the tendon. However, as
taglandin E2 and leukotriene B4 or neo- with the clinical response to ESWT, dry needling lacks
vascularization.50 In addition to activity modification or high concentrations of growth factors to enhance the
analgesics, eccentric exercise training is associated with healing process of injured tissue.
improvement in knee pain or function.7 Eccentric ex- Liddle and Rodriguez-Merchan22 compared the
ercise training refers to muscle contractions as the effectiveness of PRP with alternative treatment options
muscle elongates, and is important for athletes to avoid (eccentric exercise training, ESWT, and dry needling) in
sports injury.51 Because the quadriceps muscles are a pairwise meta-analysis. The results of the comparative
embryologically related to the patellar tendons, the studies were inconsistent, and superiority of PRP over
 TREATMENTS FOR PATELLAR TENDINOPATHY 3129

control treatments could not be conclusively demon- each treatment strategy. Finally, we did not find any
strated. This meta-analysis viewed eccentric exercise related literature with follow-up of greater than 2 years.
training, ESWT, and dry needling as the same The follow-up time of the included studies was quite
comparator, but these 3 treatment options were actu- short, which limited the results to the short-term
ally different in clinical practice. This would lead to high effectiveness.
heterogeneity in a meta-analysis. Dupley and Char-
alambous21 performed a similar pairwise meta-analysis Conclusions
and compared PRP with controls (ESWT and dry The network meta-analysis demonstrated that LR-
needling). The results showed that PRP was statistically PRP has the greatest functional improvement and
better than control with regards to VISA scale at long- pain reduction for PT compared with other treatment
term follow-up. Likewise, the results were also of options. However, the treatment effect estimates can be
high heterogeneity because of the same reasons biased by the possible intransitivity and should not be
encountered in Liddle and Rodriguez-Merchan.22 In overestimated.
our network meta-analysis, we regarded nonsurgical
treatments other than PRP as different treatment op-
tions. Besides, we also included studies without PRP to
References
1. Blazina ME, Kerlan RK, Jobe FW, Carter VS, Carlson GJ.
make the multiple comparisons of nonsurgical treat-
Jumper’s knee. Orthop Clin North Am 1973;4:665-678.
ment options for PT more complete.
2. Eifert-Mangine M, Brewster C, Wong M, Shields C Jr,
Our study depicted a practical and complete picture of Noyes FR. Patellar tendinitis in the recreational athlete.
the success of various treatments for PT in terms of Orthopedics 1992;15:1359-1367.
major outcomes. The statistical model of network meta- 3. Lian OB, Engebretsen L, Bahr R. Prevalence of jumper’s
analysis provided results including both direct and in- knee among elite athletes from different sports: A cross-
direct comparisons. We also presented the probabilities sectional study. Am J Sports Med 2005;33:561-567.
of ranking for the treatment strategies and calculated 4. Lee WC, Ng GY, Zhang ZJ, Malliaras P, Masci L, Fu SN.
the SUCRA for ranking purposes: this could help phy- Changes on tendon stiffness and clinical outcomes in
sicians to make better clinical decisions. athletes are associated with patellar tendinopathy after
eccentric exercise [published online December 17, 2018].
Limitations Clin J Sport Med. doi: 10.1097/JSM.0000000000000562.
Our study nevertheless had some limitations. First, 5. Visnes H, Bahr R. The evolution of eccentric training as
treatment for patellar tendinopathy (jumper’s knee): A
the dose of each treatment was not standardized in the
critical review of exercise programmes. Br J Sports Med
related studies. The doseeresponse effect should not be
2007;41:217-223.
ignored when performing meta-analyses, and meta- 6. Rudavsky A, Cook J. Physiotherapy management of
regression could be considered with sufficient avail- patellar tendinopathy (jumper’s knee). J Physiother
able data. Second, all participants received eccentric 2014;60:122-129.
exercise training in the control groups. Although the 7. Larsson ME, Kall I, Nilsson-Helander K. Treatment of
exercise protocols (such as frequency, intensity, types of patellar tendinopathyda systematic review of random-
exercise, time) are similar in most studies, some dif- ized controlled trials. Knee Surg Sports Traumatol Arthrosc
ferences may exist in these exercise programs, calling 2012;20:1632-1646.
into question the assumption of transitivity. This is a 8. Dar G, Mei-Dan E. Immediate effect of infrapatellar strap
common problem when studies involving exercise on pain and jump height in patellar tendinopathy among
young athletes. Prosthet Orthot Int 2018:
training were compared, and the differences in treat-
309364618791619.
ment effect should be interpreted with caution. Third,
9. Rosen AB, Ko J, N Brown C. Single-limb landing
the included participants were mostly male, and the biomechanics are altered and patellar tendinopathy
results might be biased by the sex difference. Future related pain is reduced with acute infrapatellar strap
studies could focus on the effects of treatment for PT in application. Knee 2017;24:761-767.
female athletes relative to male athletes. Fourth, 10. de Vries AJ, van den Akker-Scheek I, Diercks RL,
number of knees rather than patients were used as a Zwerver J, van der Worp H. The effect of a patellar strap
unit of analysis in our systematic review, and this kind on knee joint proprioception in healthy participants and
of non-independence of the data might bias the anal- athletes with patellar tendinopathy. J Sci Med Sport
ysis. Fortunately, the number of knees did not far 2016;19:278-282.
exceed patients in our systematic review, so the analysis 11. Warden SJ, Metcalf BR, Kiss ZS, et al. Low-intensity
pulsed ultrasound for chronic patellar tendinopathy: A
would not be overly biased. Fifth, adverse events after
randomized, double-blind, placebo-controlled trial. Rheu-
each treatment were not analyzed due to the paucity of
matology (Oxford) 2008;47:467-471.
the data retrieved from these studies. Patient safety is 12. Wang CJ, Ko JY, Chan YS, Weng LH, Hsu SL. Extracor-
an extremely important issue in clinical decisions, and it poreal shockwave for chronic patellar tendinopathy. Am J
is essential to compare the rates of adverse events for Sports Med 2007;35:972-978.
3130 P-C. CHEN ET AL.

13. Fredberg U, Bolvig L, Pfeiffer-Jensen M, Clemmensen D, 29. Mills EJ, Ioannidis JP, Thorlund K, Schunemann HJ,
Jakobsen BW, Stengaard-Pedersen K. Ultrasonography as Puhan MA, Guyatt GH. How to use an article reporting a
a tool for diagnosis, guidance of local steroid injection and, multiple treatment comparison meta-analysis. JAMA
together with pressure algometry, monitoring of the 2012;308:1246-1253.
treatment of athletes with chronic jumper’s knee and 30. Cipriani A, Higgins JP, Geddes JR, Salanti G. Conceptual
Achilles tendinitis: a randomized, double-blind, placebo- and technical challenges in network meta-analysis. Ann
controlled study. Scand J Rheumatol 2004;33:94-101. Intern Med 2013;159:130-137.
14. James SL, Ali K, Pocock C, et al. Ultrasound guided dry 31. Lu G, Ades AE. Combination of direct and indirect evi-
needling and autologous blood injection for patellar ten- dence in mixed treatment comparisons. Stat Med 2004;23:
dinosis. Br J Sports Med 2007;41:518-521. discussion 522. 3105-3124.
15. Kongsgaard M, Kovanen V, Aagaard P, et al. Corticoste- 32. Glenny AM, Altman DG, Song F, et al. Indirect compari-
roid injections, eccentric decline squat training and heavy sons of competing interventions. Health Technol Assess
slow resistance training in patellar tendinopathy. Scand J 2005;9:1-134. iii-iv.
Med Sci Sports 2009;19:790-802. 33. White IR, Barrett JK, Jackson D, Higgins JP. Consistency
16. Everhart JS, Cole D, Sojka JH, et al. Treatment options for and inconsistency in network meta-analysis: Model esti-
patellar tendinopathy: A systematic review. Arthroscopy mation using multivariate meta-regression. Res Synth
2017;33:861-872. Methods 2012;3:111-125.
17. Rompe JD, Zoellner J, Nafe B. Shock wave therapy versus 34. Higgins JP, Jackson D, Barrett JK, Lu G, Ades AE,
conventional surgery in the treatment of calcifying White IR. Consistency and inconsistency in network
tendinitis of the shoulder. Clin Orthop Rel Res 2001:72-82. meta-analysis: Concepts and models for multi-arm
18. Pettrone FA, McCall BR. Extracorporeal shock wave studies. Res Synth Methods 2012;3:98-110.
therapy without local anesthesia for chronic lateral epi- 35. White IR. Multivariate random-effects meta-regression:
condylitis. J Bone Joint Surg Am 2005;87:1297-1304. Updates to mvmeta. Stata J 2011;11:255-270.
19. Zwerver J, Hartgens F, Verhagen E, van der Worp H, van 36. Salanti G, Ades AE, Ioannidis JP. Graphical methods and
den Akker-Scheek I, Diercks RL. No effect of extracor- numerical summaries for presenting results from
poreal shockwave therapy on patellar tendinopathy in multiple-treatment meta-analysis: an overview and tuto-
jumping athletes during the competitive season: A ran- rial. J Clin Epidemiol 2011;64:163-171.
domized clinical trial. Am J Sports Med 2011;39:1191-1199. 37. Begg CB. A comparison of methods to detect publication
20. Creaney L, Hamilton B. Growth factor delivery methods bias in meta-analysis. by P. Macaskill, S. D. Walter and L.
in the management of sports injuries: the state of play. Br Irwig, Statistics in Medicine, 2001; 20:641-654. Stat Med
J Sports Med 2008;42:314-320. 2002;21:1803.
21. Dupley L, Charalambous CP. Platelet-rich plasma in- 38. Chaimani A, Higgins JP, Mavridis D, Spyridonos P,
jections as a treatment for refractory patellar tendinosis: A Salanti G. Graphical tools for network meta-analysis in
meta-analysis of randomised trials. Knee Surg Rel Res STATA. PLoS One 2013;8:e76654.
2017;29:165-171. 39. Clarke AW, Alyas F, Morris T, Robertson CJ, Bell J,
22. Liddle AD, Rodriguez-Merchan EC. Platelet-rich plasma Connell DA. Skin-derived tenocyte-like cells for the
in the treatment of patellar tendinopathy: A systematic treatment of patellar tendinopathy. Am J Sports Med
review. Am J Sports Med 2015;43:2583-2590. 2011;39:614-623.
23. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. 40. Dragoo JL, Wasterlain AS, Braun HJ, Nead KT. Platelet-
Preferred reporting items for systematic reviews and rich plasma as a treatment for patellar tendinopathy: A
meta-analyses: The PRISMA statement. J Clin Epidemiol double-blind, randomized controlled trial. Am J Sports Med
2009;62:1006-1012. 2014;42:610-618.
24. Visentini PJ, Khan KM, Cook JL, Kiss ZS, Harcourt PR, 41. Resteghini P, Khanbhai TA, Mughal S, Sivardeen Z.
Wark JD. The VISA score: An index of severity of symp- Double-blind randomized controlled trial: Injection of
toms in patients with jumper’s knee (patellar tendinosis). autologous blood in the treatment of chronic patella
Victorian Institute of Sport Tendon Study Group. J Sci Med tendinopathy-a pilot study. Clin J Sport Med 2016;26:
Sport 1998;1:22-28. 17-23.
25. Newell DJ. Intention-to-treat analysis: Implications for 42. Steunebrink M, Zwerver J, Brandsema R,
quantitative and qualitative research. Int J Epidemiol Groenenboom P, van den Akker-Scheek I, Weir A.
1992;21:837-841. Topical glyceryl trinitrate treatment of chronic patellar
26. Higgins JPT, Green S, (editors) Cochrane Handbook for tendinopathy: A randomised, double-blind, placebo-
Systematic Reviews of Interventions Version 5.1.0 [upda- controlled clinical trial. Br J Sports Med 2013;47:34-39.
ted March 2011]. The Cochrane Collaboration, 2011. 43. Thijs KM, Zwerver J, Backx FJ, et al. Effectiveness of
Available at: www.cochrane-handbook.org. Accessed shockwave treatment combined with eccentric training
February 26, 2018. for patellar tendinopathy: A double-blinded randomized
27. Tu YK, Baelum V, Gilthorpe MS. The problem of ana- study. Clin J Sport Med 2017;27:89-96.
lysing the relationship between change and initial value in 44. van der Worp H, Zwerver J, Hamstra M, van den Akker-
oral health research. Eur J Oral Sci 2005;113:271-278. Scheek I, Diercks RL. No difference in effectiveness be-
28. Egger M, Davey Smith G, Schneider M, Minder C. Bias in tween focused and radial shockwave therapy for treating
meta-analysis detected by a simple, graphical test. BMJ patellar tendinopathy: A randomized controlled trial. Knee
1997;315:629-634. Surg Sports Traumatol Arthrosc 2014;22:2026-2032.
 TREATMENTS FOR PATELLAR TENDINOPATHY 3131

45. Vetrano M, Castorina A, Vulpiani MC, Baldini R, Pavan A, junction. A study in rabbits. J Orthop Res 2003;21:
Ferretti A. Platelet-rich plasma versus focused shock 984-989.
waves in the treatment of jumper’s knee in athletes. Am J 58. Abrahamsson SO. Similar effects of recombinant human
Sports Med 2013;41:795-803. insulin-like growth factor-I and II on cellular activities in
46. Fredberg U, Bolvig L, Jumper’s knee. Review of the flexor tendons of young rabbits: Experimental studies
literature. Scand J Med Sci Sports 1999;9:66-73. in vitro. J Orthop Res 1997;15:256-262.
47. Khan KM, Maffulli N, Coleman BD, Cook JL, Taunton JE. 59. Chen YJ, Wang CJ, Yang KD, et al. Extracorporeal shock
Patellar tendinopathy: Some aspects of basic science and waves promote healing of collagenase-induced Achilles
clinical management. Br J Sports Med 1998;32:346-355. tendinitis and increase TGF-beta1 and IGF-I expression.
48. Cook JL, Feller JA, Bonar SF, Khan KM. Abnormal J Orthop Res 2004;22:854-861.
tenocyte morphology is more prevalent than collagen 60. Banes AJ, Horesovsky G, Larson C, et al. Mechanical load
disruption in asymptomatic athletes’ patellar tendons. stimulates expression of novel genes in vivo and in vitro
J Orthop Res 2004;22:334-338. in avian flexor tendon cells. Osteoarthritis Cartilage 1999;7:
49. Khan KM, Cook JL, Kannus P, Maffulli N, Bonar SF. Time 141-153.
to abandon the "tendinitis" myth. BMJ 2002;324:626-627. 61. Foster TE, Puskas BL, Mandelbaum BR, Gerhardt MB,
50. Schwartz A, Watson JN, Hutchinson MR. Patellar ten- Rodeo SA. Platelet-rich plasma: From basic science
dinopathy. Sports Health 2015;7:415-420. to clinical applications. Am J Sports Med 2009;37:2259-2272.
51. Lorenz D, Reiman M. The role and implementation of 62. Creaney L, Wallace A, Curtis M, Connell D. Growth
eccentric training in athletic rehabilitation: Tendinopathy, factor-based therapies provide additional benefit beyond
hamstring strains, and acl reconstruction. Int J Sports Phys physical therapy in resistant elbow tendinopathy: A pro-
Ther 2011;6:27-44. spective, single-blind, randomised trial of autologous
52. Reinking MF. Current concepts in the treatment of patellar blood injections versus platelet-rich plasma injections. Br J
tendinopathy. Int J Sports Phys Ther 2016;11:854-866. Sports Med 2011;45:966-971.
53. Huisstede BM, Gebremariam L, van der Sande R, Hay EM, 63. Raeissadat SA, Rayegani SM, Hassanabadi H, Rahimi R,
Koes BW. Evidence for effectiveness of extracorporal Sedighipour L, Rostami K. Is Platelet-rich plasma superior
shock-wave therapy (ESWT) to treat calcific and non- to whole blood in the management of chronic tennis
calcific rotator cuff tendinosisda systematic review. Man elbow: One year randomized clinical trial. BMC Sports Sci
Ther 2011;16:419-433. Med Rehabil 2014;6:12.
54. Orhan Z, Cam K, Alper M, Ozturan K. The effects of 64. Raeissadat SA, Sedighipour L, Rayegani SM,
extracorporeal shock waves on the rat Achilles tendon: Is Bahrami MH, Bayat M, Rahimi R. Effect of platelet-rich
there a critical dose for tissue injury? Arch Orthop Trauma plasma (PRP) versus autologous whole blood on pain
Surg 2004;124:631-635. and function improvement in tennis elbow: A randomized
55. Maier D, Bornebusch L, Salzmann GM, Sudkamp NP, clinical trial. Pain Res Treat 2014;2014:191525.
Ogon P. Mid- and long-term efficacy of the arthroscopic 65. Thanasas C, Papadimitriou G, Charalambidis C,
patellar release for treatment of patellar tendinopathy Paraskevopoulos I, Papanikolaou A. Platelet-rich plasma
unresponsive to nonoperative management. Arthroscopy versus autologous whole blood for the treatment of chronic
2013;29:1338-1345. lateral elbow epicondylitis: A randomized controlled clin-
56. Wang CJ, Huang HY, Pai CH. Shock wave-enhanced ical trial. Am J Sports Med 2011;39:2130-2134.
neovascularization at the tendon-bone junction: An 66. Housner JA, Jacobson JA, Misko R. Sonographically
experiment in dogs. J Foot Ankle Surg 2002;41:16-22. guided percutaneous needle tenotomy for the treatment
57. Wang CJ, Wang FS, Yang KD, et al. Shock wave of chronic tendinosis. J Ultrasound Med 2009;28:
therapy induces neovascularization at the tendon-bone 1187-1192.
3131.e1 P-C. CHEN ET AL.

Appendix 1 EMBASE
SEARCH STRATEGY #1. random*:ab,ti
#2. placebo*
The Central Register of Controlled Trials (CENTRAL) #3. (double next/1 blind*):ab,ti
(Wiley) #4. (single next/1 blind*):ab,ti
#1. random*:ab,ti #5. #1 or #2 or #3 or #4
#2. placebo* #6. ’patellar tendinopathy’/exp
#3. (double next/1 blind*):ab,ti #7. (patella$ next/3 tend$):ab,ti
#4. (single next/1 blind*):ab,ti #8. pt:ab,ti
#5. #1 or #2 or #3 or #4 #9. #6 or #7 or #8
#6. mh ’patellar tendinopathy’ #10. #5 and #9
#7. (patella$ next/3 tend$):ab,ti
#8. pt:ab,ti Web of Science
#9. #6 or #7 or #8 #1 TS ¼ (random* OR rct* OR crossover OR masked
#10. #5 and #9 OR blind* OR placebo*)
#2 TS ¼ (patellar tendinopathy OR patellar tendinosis
OR patellar tendinitis OR patellar tendon pain)
MEDLINE #3 #2 AND #1
1. randomized controlled trial.pt.
PEDro
2. controlled clinical trial.pt
‘patellar tendinopathy’
3. randomized.ab.
4. placebo.ab. SportDiscus
5. therapy.fs. (SU (patellar tendinopathy) OR TI (patellar tendin-
6. randomly.ab. opathy) OR AB (patellar tendinopathy) OR SU (patellar
7. trial.ab. tendinitis) OR TI (patellar tendinitis) OR AB (patellar
8. group.ab. tendinitis) OR SU (patellar tendinosis) OR TI (patellar
9. or/1e8 tendinosis) OR AB (patellar tendinosis) OR SU (PT) OR
10. humans.sh. TI (PT) OR AB (PT))
11. 9 and 10 World Health Organization International Clinical
12. patella*.mp Trials Registry Platform Search Portal (apps.who.int/
13. tendinopathy/ or tendinosis/ or tendinitis/ trialsearch/Default.aspx)
14. 12 and 13 #1 Patellar tendinopathy
15. 11 and 14
 TREATMENTS FOR PATELLAR TENDINOPATHY 3131.e2

Appendix 2. Methodologic Quality Assessment of the Included Studies (PEDro Scale)

Study 1* 2 3 4 5 6 7 8 9 10 11 Total
Zwerver et al., 201119 Y Y Y Y Y N Y Y Y Y Y 9
Thijs et al., 201743 Y Y Y Y Y N Y N Y Y Y 8
Lee et al., 20184 Y Y N Y N N Y Y N Y Y 5
Dragoo et al., 201440 Y Y Y Y Y N Y Y Y Y Y 9
Resteghini et al., 201641 Y Y N N Y Y Y Y Y Y Y 8
Vetrano et al., 201345 Y Y N Y N N N Y Y Y Y 7
Kongsgaard et al., 200915 Y Y N Y N N N Y N Y Y 5
Warden et al., 200811 Y Y Y Y Y Y Y N Y Y Y 9
Steunebrink et al., 201342 Y Y Y Y Y Y Y Y Y Y Y 10
Clarke et al., 201139 Y Y N N Y Y Y Y Y N Y 7
van der Worp et al., 201444 Y Y Y Y Y N Y Y Y Y Y 9
NOTE. The following is the website address for the PEDro Scale: https://www.pedro.org.au/english/downloads/pedro-scale/.
N, no; PEDro, Physiotherapy Evidence Database; Y, yes.
*Item 1 influences external validity; hence, it is not used to calculate the PEDro score. The 11-item scale gives a score out of 10. The 11 items of
the PEDro scale are as follows:
1. Eligibility criteria were specified.
2. Subjects were randomly allocated to groups (in a crossover study, subjects were randomly allocated an order in which treatments were
received).
3. Allocation was concealed.
4. The groups were similar at baseline regarding the most important prognostic indicators.
5. There was blinding of all subjects.
6. There was blinding of all therapists who administered the therapy.
7. There was blinding of all assessors who measured at least 1 key outcome.
8. Measures of at least 1 key outcome were obtained from more than 85% of the subjects initially allocated to groups.
9. All subjects for whom outcome measures were available received the treatment or control condition as allocated or, where this was not the
case, data for at least 1 key outcome was analyzed by “intention to treat.”
10. The results of between-group statistical comparisons are reported for at least 1 key outcome.
11. The study provides both point measures and measures of variability for at least 1 key outcome.

Appendix Figure 1. Comparison-adjusted funnel plots to assess reporting bias for (A) improvement on the VISA scale, and (B)
change in pain score. (ABI, autologous blood injection; CSI, corticosteroid injection; DN, dry needling; fESWT, focused extra-
corporeal shock wave therapy; LR-PRP, leukocyte-rich platelet-rich plasma; rESWT, radial extracorporeal shock wave therapy;
SDTLC, skinederived tendonelike cells; TGT, topical glyceryl trinitrate; VISA, Victorian Institute of Sport Assessment.)