Effectiveness and safety of prolotherapy injections for management of lower limb tendinopathy and fasciopathy: a systematic review

A total of 203 studies were identified through electronic search. Following the review
of titles and abstracts, 26 studies were extracted for full review and finally, 8
studies were considered appropriate for inclusion (Additional file 1). The reasons for the exclusion of studies are available in Additional file 2. A flow diagram, as described by Moher et al. 43], is presented in Fig. 1 and highlights the study selection process. The included studies were assigned to
one of three categories, based on the following comparisons: prolotherapy injection
for Achilles tendinopathy, prolotherapy injection for plantar fasciopathy, and prolotherapy
injection for Osgood-Schlatter Disease. Five studies were included in the Achilles
tendinopathy section, two studies in the plantar fasciopathy section and one study
in the Osgood-Schlatter Disease section of the review. All studies that met the inclusion
criteria reported dextrose as the primary prolotherapy agent. See Table 1 for characteristics of included studies.

Fig. 1. Quorum flowchart of the reviewing process

Table 1. Characteristics of included studies

Quality assessment of included studies

The methodological quality scores for the eight included studies are shown in Table 2. The maximum possible score per study is 18 and each study was also graded for allocation
concealment (A – concealed assignment, B – possible disclosure/unclear, C – disclosure
likely).

Table 2. Quality of assessment for included trials

The methodological quality of the studies in this review was generally poor. Allocation
concealment (the procedure for protecting the randomisation process so that the treatment
to be allocated is not known before the participants enter into the study) was poorly
conducted in the majority of studies analysed in this review. The trials by Topol
et al. 37] and Kim et al. 35] were two exceptions, however concealment was implied but not defined in both studies.
All studies scored poorly for participant blinding, which is particularly difficult
in injection intervention trials, with Topol et al. 37] the only study providing any reference of participant blinding. Only Yelland et al.
34] reported an adequate method of allocation concealment. In general, the selection
of outcome measures was clinically appropriate and the duration of surveillance was
adequate. The blinding of assessors was attempted in only two trials, although it
is recognised that it may be difficult to blind an assessor when there may be evidence
of the procedure being undertaken from post-injection signs. Many studies also failed
to conduct, or provide data for, intention to treat analysis.

Yelland et al. 34] and Topol et al. 37] were the only studies with a ‘good’ overall quality assessment (see Table 2), with Yelland et al. 34] the only study in which adequate evidence of concealed allocation was provided, stating
the randomisation schedule was generated and administered by telephone independently
by the NHMRC Clinical Trials Centre in Sydney, Australia. Topol et al. 37] reported use of a random numbers table for assignment to each intervention group
and stated that this was blinded to the participants, guardian and treating/evaluating
physician; however, again, no specific mention was made of how this allocation was
concealed. Kim et al. 35] used alternation of participants as a method of randomisation but did not define
the method of concealment.

All other studies used a prospective single intervention case-series methodology,
thus no randomisation or allocation concealment was carried out, with resulting ‘poor’
to ‘moderate’ quality assessment overall scores. Although all studies had reasonable
scores regarding the definition and usefulness of outcome measurements (item I), they
were generally very poor in quality regarding intention to treat analysis (item B)
and blinding procedures (items C, D), although this may not have been possible in
several of the studies.

Study characteristics

There were three randomised controlled trials and five prospective case-series studies
identified, the characteristics of which are presented in Table 1. Lyftogt 30] conducted a prospective case-series of 16 participants with non-insertional Achilles
tendinopathy, to evaluate the clinical effectiveness of prolotherapy for relieving
painful symptoms. Subcutaneous prolotherapy injections of 1 mL were performed weekly
where possible with final follow-up at 3 months post-final treatment. Lyftogt 31] carried out a significantly larger study with similar methods 2 years later and used
three different prolotherapy regimens: (i) dextrose 20 %/lignocaine 0.1 %, (ii) dextrose
30 %/lignocaine 0.1 %/ropivacaine 0.1 % and (iii) dextrose 40 %/ropivacaine 0.1 %
for 87, 31 and 26 participants respectively. In this study, Lyftogt’s 31] final follow-up was undertaken independently after a mean period of 20 months. Both
of Lyftogt’s studies involved mainly male participants and treated only midportion
Achilles tendinopathy 30], 31].

Maxwell et al. 32] conducted a prospective case-series assessing pain, satisfaction and sonographic
changes following sonographically-guided hyperosmolar dextrose injections in 36 participants
(25 men and 11 women) with chronic Achilles tendinopathy. VAS pain during rest, VAS
pain during normal daily activity and VAS pain during or after sporting activity scores
and sonographic assessments were recorded 6 weeks after final treatment. In addition,
a telephone interview with each study participant was performed a mean of 12 months
(range 4.5–28 months) after the last treatment to assess the medium- to long-term
efficacy of dextrose injection therapy. Ryan et al. 33] built on evidence presented in the study by Maxwell et al. 32] in a prospective case-series published 3 years later. Ninety-nine participants (58
men and 48 women) with painful midportion (86) and insertional (22) Achilles tendinopathy
pain received the same prolotherapy regimen, outcome measures and treatment schedule
as described by Maxwell et al. 32], however improvements were made on the length of follow-up with a mean final follow-up
time of 28.6 months.

Yelland et al. 34] compared weekly prolotherapy injections with eccentric loading exercises and a combination
of eccentric loading exercises plus prolotherapy injections in 43 participants with
painful Achilles tendinopathy. The participants were randomly allocated using a computer-generated
randomisation schedule to receive prolotherapy injections alone (15 participants),
eccentric loading exercisesalone (14 participants) or a combination therapy of prolotherapy
and eccentric loading exercises (14 participants). Between 0.5 and 1 mL of solution
was injected at each tender point, most commonly the anterolateral and anteromedial
margins of the tendon and on the most posterior aspect of the tendon 2–7 cm from the
calcaneal attachment, as illustrated in Fig. 2. A maximum total of 5 mL of solution was used, with the number of treatments determined
by the time it took to reach a pain-free activity or until the participant requested
to cease treatment. Outcome assessments of VISA-A, VAS pain, stiffness and limitation
of activity scores, satisfaction rating and treatment costs were made at 6 weeks,
3, 6 and 12 months from initial treatment.

Fig. 2. Posterior photograph of right lower leg showing injection points most commonly used
by Yelland et al. 34] for management of Achilles tendinopathy. The ‘X’ markings represent the anteromedial,
posterior midline and anterolateral margins of the tendon, with orange lines demarking the Achilles tendon

Kim et al. 35] compared the efficacy of autologous platelet-rich plasma with prolotherapy injections
in 21 participants with chronic recalcitrant plantar fasciopathy aged from 19 to 57 years.
Participants with odd participant sequence numbers were allocated to the prolotherapy
group (11 participants), while participants with even sequence numbers were placed
in the autologous platelet-rich plasmagroup (10 participants). The injection procedure
was performed under aseptic conditions using a 22-gauge needle. Figure 3 illustrates the location of needle insertion through the medial heel, which was performed
under ultrasound-guidance toward the target area. Approximately 2 mL of platelet-rich
plasmaor prolotherapy solution was injected using a peppering technique, which involved
a single skin portal followed by five penetrations of the fascia. Both interventions
consisted of two injections at 2-weekly intervals. Final follow-up occurred between
10 and 28 weeks from initial treatment date, with all participants completing the
follow-up with the exception of one in the platelet-rich plasmagroup. Ryan et al.
36] conducted a prospective case-series to evaluate the effectiveness of sonographically-guided
prolotherapy injections in reducing pain in 20 participants (17 women) with chronic
plantar fasciopathy. Outcome measures including the Foot Function Index (FFI) and
VAS pain were recorded at baseline and the final treatment consultation, with a follow-up
telephone interview conducted a mean of 11.8 months (range 6–20 months) following
the participant’s final session.

Fig. 3. Plantar photograph of left foot illustrating the injection site used by Kim et al.
35] for management of plantar faciopathy. The ‘X’ marking represents the medial heel
site used for the ultrasound-guided platelet-rich-plasma and prolotherapy injections,
with the orange lines demarking the medial band of the plantar fascia

In the only study included in this analysis that assessed prolotherapy injections
for Osgood-Schlatter Disease, Topol et al. 37] performed a double-blinded randomised controlled trial involving 65 knees in 54 participants.
Participants (mean age 13.3 years) were randomly allocated to receive prolotherapy
injections (21 knees), lignocaine injections (22 knees) or supervised usual care (22
knees). Locations and technique for the lignocaine and prolotherapy injections can
be seen in Fig. 4. Topol et al. 37] used the Nirschl Pain Phase Scale (NPPS), which is a 7-level measure of sports inhibition
and sports-related symptoms. The threshold goals in the study were a NPPS score 4,
where symptoms may be present however sport is uninhibited, or a NPPS score below
4 where asymptomatic sport occurs. This study assessed sport alteration and sport-related
symptoms in adolescent athletes with Osgood-Schlatter Disease. All participants that
still reported sports-related symptoms after 3 months, the point at which the actual
injectant was revealed to the treating physician and participant, could then choose
to receive prolotherapy injections. This was offered monthly until 12 months after
either elimination of symptoms or plateau of improvement. Athletes were not required
to receive dextrose injection if they were satisfied with their status at 3 months.

Fig. 4. Anteroposterior photograph of knee illustrating injection points marked ‘X’ starting
over the most distal area of pain on the tibial tuberosity and moving proximally in
1-cm increments to the most proximal painful point with pressure as described by Topol
et al. 37]. The orange lines represent the attachment of the patellar tendon from the patella to the tuberosity
or its fragments

Effects of interventions

Pooled standardised mean differences for short-, intermediate- and long-term data

Meta-analysis was only possible for three of the included studies. SMDs were calculated
for Yelland et al. 34], Kim et al. 35] and Topol et al. 37] and are presented in the forest plots in Fig. 5. Substantial heterogeneity was present in the comparison between prolotherapy and
other interventions for short- (P??0.001, I
²
?=?79 %), intermediate- (P?=?0.029, I
²
?=?72 %) and long-term data (P?=?0.001, I
²
?=?81 %). A statistically significant total effect was found between prolotherapy
and other intervention for long-term data (P?=?0.024), with the total (random effects) SMD found to be 0.48 (0.15 to 2.03). Short-term
data approached statistical significance (P?=?0.052) with a total SMD of 0.38 (?0.01 to 1.48) while the intermediate-term total
SMD of 0.47 (?0.55 to 1.33) was not found to be statistically significant (P?=?0.410). Data from the remaining five studies could not be pooled due lack of a
comparator group, and qualitative and quantitative analysis of all individual studies
are discussed in the following pathology-specific sections.

Fig. 5. Standardised mean differences (SMD) for improved pain after prolotherapy vs comparator
intervention for plantar fasciopathy, Osgood Schlatter disease and Achilles tendinopathy

Prolotherapy injections for treatment of Achilles tendinopathy

Five studies of moderate to good methodological quality involving 338 participants
reported on management of painful Achilles tendinopathy with prolotherapy injections.
The studies included: prolotherapy as the only intervention with no control group
(Lyftogt 30], Lyftogt 31], Maxwell et al. 32], Ryan et al. 33]); and prolotherapy versus eccentric loading exercises versus combination of prolotherapy
plus eccentric loading exercises (Yelland et al. 34]).

Pain

All five studies reported a significant reduction in pain following prolotherapy treatment.
Lyftogt 30] reported a mean reduction in VAS pain score of 62.0 mm by final 16-week follow-up.
In another study, Lyftogt 31] reported a mean reduction in VAS scores for three different prolotherapy regimens
with results as follows: 60.2 mm for dextrose 20 %/lignocaine 0.1 %, 56.0 mm for dextrose
30 %/lignocaine 0.1 %/ropivacaine 0.1 % and 56.0 mm for dextrose 40 %/ropivacaine
0.1 % with a mean length of treatment 8.0, 7.6 and 8.7 weeks respectively. Not enough
data was supplied to calculate the 95 % CIs for these studies. Maxwell et al. 32] reported a mean percentage reduction at final follow-up for VAS1 (pain during rest)
of 88.2 % (p??0.001), for VAS2 (pain during normal daily activity) of 84.0 % (p??0.001), and for VAS3 (pain during or after sporting activity) of 78.1 % (p??0.001) following sonographically-guided intratendinous prolotherapy injections.
The same injection regimen was repeated 3 years later in a study by Ryan et al. 33] who reported significant improvement in pain scores for both midportion Achilles
tendinopathy [VAS1 90.3 % (p??0.001), VAS2 81.07 (p??0.001), VAS3 76.3 % (p??0.001)] and insertional Achilles tendinopathy [VAS1 91.5 % (p??0.001), VAS2 80.5 % (p??0.001), VAS3 74.6 % (p??0.001)] participants from baseline to final follow-up.

Yelland et al. 34] were contacted for unpublished VAS pain data used in the study from which the SMD
between intervention groups was calculated. As can be seen in the forest plot in Fig. 5, a large effect was seen in favour of prolotherapy vs eccentric loading exercises
for short- (SMD 0.82, 95 % CI ?2.18 to 2.25), intermediate- (1.26, 95 % CI ?0.77 to
1.54) and long-term pain (0.84, 95 % CI ?1.54 to 3.23). For prolotherapy compared
to combination (prolotherapy plus eccentric loading exercises) therapy, a small and
moderate effect was found in favour of prolotherapy for short- (0.04, ?2.18 to 2.25)
and intermediate-term (0.56, 95 % CI ?1.60 to 2.72) pain, however a small effect in
favour of combination therapy was seen long-term (?0.04, 95 % CI ?2.04 to 1.97). It
should be noted that although the SMDs for VAS pain favoured prolotherapy and combination
therapy over eccentric loading exercises, none of these differences were found to
be statistically significant. Similarly, Yelland et al. 34] reported decreases in pain scores from baseline for eccentric loading exercises,
which was significantly less than for prolotherapy at 6 months and for combined treatment
at 12 months. However, more importantly, the difference between treatment groups over
time was not significant overall.

Pathology or location-specific outcome measures

At 12 months, Yelland et al. 34] reported that the proportions achieving the minimum clinically important change for
VISA-A were 73 % for eccentric loading exercises, 79 % for prolotherapy and 86 % for
combined treatment. Mean improvements in VISA-A scores at 12 months were 23.7 points
for eccentric loading exercises, 27.5 points for prolotherapy and 41.1 points for
combined treatment. At 6 weeks and 12 months, these increases were significantly less
for eccentric loading exercises than for combined treatment. Compared with eccentric
loading exercises, reductions in stiffness and limitation of activity occurred earlier
with combined treatment.

Satisfaction

Lyftogt 30] reported 88 % satisfaction for the single cohort at 16-week follow-up, while the
satisfaction for each treatment group in Lyftogt 31] was: A) 96 % at mean 21-month follow-up, B) 88 % at mean 12-month follow-up, and
C) no data supplied. Twenty of 32 participants were asymptomatic at 12 months after
treatment in the study by Maxwell et al. 32], with 19 participants giving a satisfaction level of 95–100 %. Only one participant
had moderate symptoms and described only 50 % satisfaction level. Yelland et al. 34] found the percentage of participants reporting “satisfaction” or “extreme satisfaction”
with treatment at 12 months was 50 % for eccentric loading exercises, 69 % for prolotherapy
and 71 % for combined treatment, however the differences between groups was not statistically
significant.

Sonographic measurements of tendon thickness, size of hypoechoic region and intratendinous
tear size

Maxwell et al. 32] reported the mean Achilles tendon thickness decreased from 11.7 to 11.1 mm (p??0.007) following prolotherapy treatment. The number of tendons with anechoic clefts
or foci was reduced by 78 %. Echogenicity improved in 6 tendons (18 %), but was unchanged
in 27 tendons (82 %). Neovascularity was unchanged in 11 tendons (33 %) but decreased
in 18 tendons (55 %); no neovascularity was present before or after treatment in the
remaining tendons. Similarly, Ryan et al. 33] also noted a reduction in painful symptoms corresponded with improvements in some
aspects of the sonographic appearance of the Achilles tendon. There were no differences
in the recorded thicknesses of the Achilles tendon at either injury site. There was
a greater reduction in the number of participants with both grade 3 or 2 hypoechogenicity
and neovascularization in the midportion group at post-test compared with the insertional
group.

Adverse events

Lyftogt 30], Lyftogt 31] and Ryan et al. 33] did not provide information on post-intervention complications. Maxwell et al. 32] reported no adverse effects or complications following prolotherapy injections, and
concluded hyperosmolar dextrose has “an excellent safety profile”. Yelland et al.
34] was the only study to report an adverse effect; in which one participant in the eccentric
loading exercises group had a partial calf tear while playing tennis. An independent
sports physician did not attribute this to the eccentric loading exercises program.

Prolotherapy injections for treatment of plantar fasciopathy

Two studies of moderate to good methodological quality involving 41 participants reported
on management of painful plantar fasciopathy with prolotherapy injections. The studies
included: prolotherapy versus platelet-rich plasma (Kim et al. 35]); and prolotherapy as the only intervention with no control group (Ryan et al. 36]).

Pain

Both Kim et al. 35] and Ryan et al. 36] reported a significant reduction in painful symptoms following prolotherapy injections
in participants with painful plantar fasciopathy. Figure 5 shows the SMD for improvement in pain after prolotherapy vs platelet-rich plasma
injections, with Kim et al. 35] reporting a small short-term (SMD ?0.26, 95 % CI ?1.88 to 1.37) and moderate intermediate-term
(?0.52, 95 % CI ?2.56 to 1.52) effect in favour of platelet-rich plasma. However,
both prolotherapy and platelet-rich plasma interventions had significant improvements
on pain levels, with a 17.1 % and 29.7 % improvement at 28 weeks respectively. Ryan
et al. 36] found a significant decrease (p??0.001) in all mean VAS items from pre-test to post-test: VAS1 (36.8 to 10.3), VAS2
(74.7 to 25.0) and VAS3 (91.6 to 38.7) with no significant changes at the final follow-up
interview.

Pathology or location-specific outcome measures

Kim et al. 35] found the mean Foot Function Index total and subcategory score improvements were
greater in the platelet-rich plasma group (30.4 %) compared with the prolotherapy
group (15.1 %), however no statistically significant differences were noted at any
follow up. In the pain and disability subcategories, both groups showed significant
improvements at the last re-evaluation.

Adverse events

Ryan et al. 36] reported no complications from prolotherapy injections into the plantar fascia. Kim
et al. 35] found most participants in both intervention groups reported local pain or discomfort
that started on the day of injection and subsided gradually, however no other complications
of either injection therapy were reported.

Prolotherapy injections for treatment of Osgood-Schlatter disease

One trial of moderate to good methodological quality involving 54 participants reported
on management of painful Osgood-Schlatter disease with prolotherapy injections. The
study included prolotherapy versus usual care versus lignocaine injections (Topol
et al. 37]).

Pain

SMD calculations of the results reported by Topol et al. 37] found a large effect size in favour of prolotherapy versus usual care or lignocaine
injections for short-term (prolotherapy versus usual care SMD 2.39, 95 % CI 1.69 to
3.08; prolotherapy versus lignocaine 1.16, 95 % CI 0.85 to 1.48) and long-term pain
levels (prolotherapy versus usual care 2.43, 95 % CI 1.67 to 3.18; prolotherapy versus
lignocaine 1.33, 95 % CI 0.57 to 2.09) – illustrated in Fig. 5. Figure 5 also shows that the prolotherapy intervention had a large positive effect when compared
to usual care and lignocaine, with the lower confidence interval not crossing zero
(the line of no effect) for both short- and long-term outcomes.

Return to activity

Topol et al. 37] reported unaltered sport was more common in both prolotherapy-treated (21 of 21 versus
13 of 22; p?=?0.001) and lignocaine-treated (20 of 22 versus 13 of 22; p?=?0.034) knees when compared with usual care at 3 months In addition, asymptomatic
sport was more frequent in the prolotherapy-treated knee group than either lignocaine-treated
(14 of 21 versus 5 of 22; p?=?0.006) or usual care-treated (14 of 21 versus 3 of 22; p??0.001) knee groups. At 1 year, asymptomatic sport was more common in the prolotherapy-treated
knee group than in groups that had knees treated with only lignocaine (32 of 38 versus
6 of 13; p?=?0.024) or only usual care (32 of 38 versus 2 of 14; p??0.001).

Adverse events

Topol et al. 37] did not provide information on post-intervention complications for any of the participant
groups, however the study reported 100 % of the prolotherapy intervention group participants
achieved unaltered sport by the final blinded period follow-up without missing a monthly
injection, indicating no complications occurred following prolotherapy injections
in the trial.