INTRODUCTION
Chronic musculoskeletal pain is defined as pain that lasts for three to six months or beyond the time of normal healing [1]. Musculoskeletal disorders are the most common source of chronic musculoskeletal pain, and their increasing prevalence has led to a need for effective non-surgical solutions, such as physical therapy, pharmacologic treatment, and injection-based treatment [2]. Injection therapies can be introduced when pain or functional limitations are significant despite oral medication or exercise [3]. Corticosteroid injections are the most common regimen for musculoskeletal disorders; they provide short-term symptomatic improvement, but aggravate cartilage damage, thus increasing the risk of tissue atrophy [4]. Therefore, physicians have become interested in alternative injectants, such as prolotherapy or platelet-rich plasma (PRP) [5].
Prolotherapy is a nonsurgical regenerative injection technique that administers small amounts of an irritant solution to the degenerated tendon insertions (entheses), joints, ligaments, and adjacent joint spaces over a series of several treatment sessions [6-8].The mechanism of action behind prolotherapy is not completely understood, but the current theory is that the injected proliferate causes a healing process that is similar to the body’s natural healing process, whereby a local inflammatory cascade is initiated, which triggers the release of growth factors and collagen deposition [2]. To date, many studies which support the benefits of the use of prolotherapy in patients with chronic musculoskeletal pain have been reported [9,10]. However, few meta-analyses have analyzed the effect of prolotherapy in patients with chronic musculoskeletal pain. Therefore, we designed a meta-analysis to evaluate the effect of prolotherapy in the treatment of chronic musculoskeletal pain and compare the effect of prolotherapy with other treatments.
MATERIALS AND METHODS
Study design
This meta-analysis was performed according to the recommendations of the PRISMA and Cochrane Collaboration. The protocol was registered with PROSPERO (no. CRD42019130609).
Information sources and search strategy
Two reviewers (WL, YL) systematically searched electronic databases such as Medline, Embase, and the Cochrane Library (CENTRAL) with no limitations on the year of publication. Additionally, KoreaMed (https://koreamed.org) and KMbase (http://kmbase.medric.or.kr) were used to search for manually relevant domestic articles. Broad search terms such as “prolotherapy”, “chronic osteoarthritis”, and “randomized controlled trials”, were included to achieve higher sensitivity, and Medical Subject Heading (MeSH) terms were used. The languages of the articles were limited to Korean and English. The last search was conducted on March 10, 2019.
We did not search grey literature, despite its important contribution to a systematic review, because we wanted to present an effective basis for treatment to clinicians with as little bias as possible, based on the results of RCTs.
Study selection and eligibility criteria
All relevant studies were independently screened by two reviewers (WL and YL). Selection of relevant articles was done primarily at the title and abstract level, then after at the full-text level. Studies for the final assessment were selected based on the agreement of the two reviewers. Any disagreement was resolved by discussion with a third reviewer (SL).
Studies were included in the meta-analysis if they satisfied the following criteria: (1) patients with chronic musculoskeletal pain lasting for more than 3 months; (2) prolotherapy using dextrose for any joints, tendon, and/or ligaments; (3) results of the non-prolotherapy group were reported; and (4) the post-injection pain score was reported as the primary outcome.
Studies were excluded for the following reasons: (1) use of prolotherapy solutions containing anything other than glucose (polidocanol, manganese, zinc, human growth hormone, phenol-glucose-glycerine, pumice, ozone, glycerin, phenol, PRP, bone marrow, lipoaspirate, stem cells, or sodium morrhuate); (2) injection into the epidural space; (3) did not report appropriate outcomes or outcome measurements as mentioned; (4) non-randomized controlled trials; (5) non-human studies; (6) articles not in English or Korean.
Risk of bias in individual studies
Two independent authors (WL and YL) reviewed the articles to assess the risk of bias (ROB) using the ROB tool provided in the Review Manager software version 5.3 (The Cochrane Collaboration, UK) based on Cochrane’s assessment of the risk of bias [11]. If necessary, a third reviewer (SL) was included in the discussion to sort out the disagreements. The following eight domains were used to assess the risk of bias in each trial: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other bias. The methodology for each trial was graded as “high”, “low”, or “unclear” to reflect a high risk of bias, low risk of bias, or uncertainty of bias, respectively. The agreement between the two independent reviewers for the level of risk of bias regarding the eight domains was assessed using Cohen’s kappa. Kappa values were interpreted as follows: 1) less than 0: less than chance agreement, 2) 0.01 to 0.20: slight agreement, 3) 0.21 to 0.40: fair agreement, 4) 0.41 to 0.60: moderate agreement, 5) 0.61 to 0.80: substantial agreement, and 6) 0.8 to 0.99: almost perfect agreement.
Data collection process and extracted items
Two authors (WL and YL) extracted data from the original articles, and another author (SL) independently confirmed all of the extracted data. The general characteristics (i.e., the study design, publication year, and name of the first author), intervention types and methods, and outcomes were extracted for each study based on the inclusion criteria. Each method of the intervention, such as the prolotherapy regimen, interval, and duration, was extracted. The measured outcomes included the number of patients analyzed in each group, tools for pain assessment, and pain scores.
The main outcome was determined by the severity of the pain, derived from the results of the pain scale. The first priority of pain measurement extraction was the pain score for 6 months to 1 year. To assess the effectiveness of dextrose prolotherapy, we used the standardized mean difference of pain scores between the prolotherapy group and other comparator groups using exercise, saline, PRP, and steroid injection.
Subgroup analysis
We grouped the analyses of VAS for pain into less than three months, three to six months, and more than six months while registering our review in PROSPERO. However, we were unable to classify the subgroups as originally planned because not all the individual studies followed the patients and reported the resulting variables on these criteria. Using the common denominator of the results of the individual studies, we were able to synthesize results that could be divided into three subgroups: baseline to 1 month, 1 month to 3 months, and 6 months to 1 year.
Statistical analysis
Continuous data (e.g., post-injection pain scores) were pooled as standardized mean differences (SMDs) because different outcome measurement scaling was expected across trials. We also calculated the 95% confidence intervals (CIs) for all estimates. A random-effect model was used to pool the study results, taking into account possible variations in effect sizes across trials. The heterogeneity statistic Cochrane Q and its corresponding degrees of freedom (df) and P value, as well as Higgins’ I2 as a measure of heterogeneity were calculated. P values < 0.05 were considered to be representative of statistically significant heterogeneity, and I2 values > 50% were considered to represent significant heterogeneity. Post-hoc subgroup analyses were performed where possible for each outcome to explore heterogeneity based on the different sites of injection. Chi-squared tests for heterogeneity were performed to identify differences between subgroups. Publication bias was not evaluated because only a few (< 10) studies were included in this meta-analysis. We conducted a sensitivity analysis to evaluate the influence of each study on the long-term (six months to one year) therapeutic effect of prolotherapy compared with saline by excluding one trial at a time from the pooled effects. All analyses were performed using R 3.51 (R Foundation for Statistical Computing, Austria) and Review Manager (RevMan, version 5.3, The Cochrane Collaboration).
RESULTS
Study selection and characteristics
We retrieved 680 articles after the initial database search: Medline (n = 250), EMBASE (n = 64), CENTRAL (n = 168), and Korean databases (n = 198).
After excluding 567 duplicate articles, primary selection was performed on 131 articles. First, we excluded 66 unrelated articles based on titles and abstracts. Second, we excluded 27 articles that only included abstracts. Thereafter, full-text reviews were conducted for 38 articles. Of these 38 full-text articles, 28 were excluded for the following reasons: not controlled with placebo or other treatment (n = 14), patients’ pain period not clearly described or less than three months (n = 9), duplication (n = 4), and articles not in English or Korean (n = 1). The reasons for exclusion of these papers are given in detail in Table 1. Finally, ROB evaluation and data extraction were performed on 10 articles (Fig. 1).
All studies were randomized controlled trials. Various injection sites, including large joints such as the knee and small joints such as finger joints and carpometacarpal joints were investigated. The comparator groups were saline injection, exercise, steroid injection, PRP injection, and extracorporeal shock wave therapy. The severity of pain as the primary outcome was measured using the Visual Analog Scale (VAS), the Western Ontario and McMaster Universities Osteoarthritis Index, the Karnofsky Performance Score, and Foot Function Index. The concentration, volume of dextrose solution, and interval between injection sessions were different between studies. The dextrose concentration ranged from 5% to 25%, and the injection interval ranged from weeks to months (Table 2).
Quality assessment of the included studies (risk of bias within studies)
ROB evaluation revealed an overall low risk for selection and reporting bias, while almost half of the studies showed a high risk of performance bias because they could not be blinded to the differences in procedures (Figs. 2, 3). All studies reported detailed information regarding the randomization techniques that were used, such as manual random number selection or a computer-generated random number table. Allocation concealment was unclear in five studies which did not mention the specific allocation concealment method.
For performance bias, four studies [12-15] were unable to blind the participants and five studies [12-14,16,17] could not blind the physicians because the injection site was different or because exercise was included in a control group. These studies were considered “high” bias. Two studies had high detection bias [15,17].
The risk of incomplete outcome data was “high” in four studies that did not mention a minimal sample size [15,16,18,19]. Four other studies did not meet a minimal sample size, resulting in an unclear risk of bias [14,17,20,21].
Reporting bias was low because there was no selective reporting in any of the studies. Regarding the potential bias, four studies were rated as unclear because there was no detailed description of the sample size calculation, and one study was high risk because the number of samples was very small [17].
The kappa value between the two reviewers for the 10 selected articles was 0.81.
Effectiveness of prolotherapy compared with other therapies
Prolotherapy with dextrose compared to saline
The effectiveness of prolotherapy compared to saline was reported in five studies [15,16,18-20] (n = 246; prolotherapy group = 126, normal saline group = 120), which suggested that prolotherapy with dextrose significantly reduced the pain score from 6 months to 1 year (SMD, −0.44; 95% CI [-0.76 to -0.11]; P = 0.008; I² = 36%; Fig. 4A). However, there was no difference between the effects of both therapies during the other periods analyzed. (SMD, 0.42; 95% CI [0.51 to 1.35]; P = 0.003; I² = 88% at baseline to 1 month; SMD, -0.07; 95% CI [-0.37 to 0.23]; P = 0.66; I² = 0% at 1 month to 3 months) Sensitivity analysis using a single study removal method did not significantly change the pooled results. The therapeutic effect of prolotherapy was 33% lower (SMD, -0.29; 95% CI [-0.57 to -0.01]; P = 0.040) than the pooled estimate effect size (SMD, -0.44; 95% CI [-0.91 to -0.13]; P = 0.009) after omitting one trial [16].
Prolotherapy with dextrose compared to exercise
Two studies [15,18] (n = 128; prolotherapy group = 63, exercise group = 65) provided data on pain scores comparing prolotherapy and exercise. Compared to exercise, dextrose therapy significantly reduced the pain score from 1 month to 3 months (SMD, -0.44; 95% CI [-0.84 to -0.04]; P = 0.11; I² = 55%) and 6 months to 1 year (SMD, -0.42; 95% CI [-0.77 to -0.07]; P = 0.02; I² = 0%; Fig. 4B). However, there was no difference in the effects of both therapies during the baseline to 1-month-period (SMD, -0.42; 95% CI [-1.14 to 0.30]; P = 0.02; I² = 83%).
Prolotherapy with dextrose compared to PRP
Two studies [12,17] (n = 99; prolotherapy group = 51, PRP group = 48) reported data on pain scores comparing prolotherapy and PRP. Prolotherapy with dextrose had a therapeutic effect corresponding to that of PRP, and there was no significant difference from 1 month to 3 months (SMD, 0.05; 95% CI [-0.34 to 0.45]; P = 0.96; I² = 0%) and 6 months to 1 year. (SMD 0.19; 95% CI [-0.20 to 0.59]; P = 0.34; I²= 0%; Fig. 4C).
Prolotherapy with dextrose compared to a steroid
Two studies [12,21] (n = 135; prolotherapy group = 68, steroid group = 67) suggested that prolotherapy with dextrose had a therapeutic effect comparable to that of steroids from 1 month to 3 months (SMD, 0.22; 95% CI [-1.27 to 1.70]; P < 0.001; I² = 94%) and 6 months to 1 year (SMD, 0.45; 95% CI [0.57 to 1.47]; P = 0.39; I²= 88%; Fig. 4D).
DISCUSSION
Previous studies have reported that prolotherapy is effective for treating musculoskeletal pain. However, their analyses included a small number of studies, which was not thought to be enough to compare prolotherapy with common regimens such as corticosteroids or PRP [2,22].
Our principal findings revealed that prolotherapy with dextrose has a clear and positive effect on chronic musculoskeletal pain ranging from 6 months to 1 year. In comparison with saline injection or exercise, treatment with prolotherapy showed a moderately superior therapeutic effect. In particular, prolotherapy was found to be more effective than exercise from one month after treatment. It was also found to have a similar effect to steroids or PRP one month after treatment.
Corticosteroid injection has been widely used as it is known to be effective in the treatment of musculoskeletal disorders. In vitro studies have shown that corticosteroids have therapeutic effects on the tendon and the surrounding connective tissues by inhibiting collagen, extracellular matrix molecules, and granulation tissue production, in addition to inflammatory suppression [23]. However, such positive therapeutic effects of corticosteroids may exist only in the short term [24]. Uğurlar et al. [12] reported that corticosteroid injection was an effective treatment in the first 6 months, but lost its effectiveness after the first 6 months. The effect of pain relief in prolotherapy was seen within 3 to 12 months. In another study, Jahangiri et al. [21] compared the effects of corticosteroid injection and prolotherapy in patients with first carpometacarpal osteoarthritis and reported that the corticosteroid injection group had better results of pain score at 1 month. However, after 2 months, prolotherapy had a more favorable outcome than corticosteroid injection. Although not shown in our study, another concern of corticosteroids is adverse effects, such as focal inflammation, necrosis, fragmentation of collagen bundles in the subacromial space, tendon/ligament weakening or rupture, and worsening osteoarthritic changes [25-27]. In contrast, prolotherapy has no serious side effects and is effective, safe, and sustainable [10]. In this study, three RCTs reported only minor transient complications such as mild to moderate pain and self-limiting bruising after prolotherapy.
We found that PRP and dextrose prolotherapy were shown to be effective for treating degenerative conditions and injuries. Both PRP therapy and prolotherapy commonly have regenerative therapeutic properties, but the central mechanisms of prolotherapy and PRP are different. In prolotherapy, hyperosmolar dextrose triggers an inflammatory response, increases platelet-derived growth factor expression, and upregulates several mitogenic factors that may act as signaling mechanisms in tendon repair [28-30]. In PRP therapy, it aims to augment the natural healing process of tendon repair and regeneration by delivering high concentrations of growth factors directly to a lesion [31]. For preparation, following the extraction of autologous venous blood with a large-gauge needle to prevent premature platelet activation [32], platelets are separated from other blood components and further concentrated [33]. This occurs through a centrifuge process, in which platelets can be isolated from the other cell components of blood based on their physiological size [33]. Further concentration of platelets occurs with subsequent centrifuge cycles [34]. As such, several steps are needed to prepare PRP, whereas the preparation of the prolotherapy is simple. And PRP involves an invasive procedure (i.e., blood drawing) and lacks an optimized standardized protocol. In this regard, prolotherapy can provide more convenience to both patients and treatment providers.
Of the ten papers included in the study, nine papers showed generally positive results of achieving pain relief and patient satisfaction regardless of the injection site. Yelland et al. [18] reported that prolotherapy was not more effective than injections of normal saline for low back pain. Nevertheless, participants exhibited marked and sustained improvements in their pain and disability, even with saline injections. They assumed that these therapeutic effects could be achieved by other factors such as patients were enrolled in a trial during severe pain and then spontaneously recovered naturally, or by the therapeutic effect by direct needling of entheses, or the placebo effect by clinical visits.
In the case of using physiotherapy as a control group [13,14], the positive result from the comparison with prolotherapy was within expectations because injection carries a strong placebo effect, which usually leads to a superior response to the noninvasive treatment.
The present study mainly analyzed the pain measurement outcomes, and functional improvement measurements were not considered. Among the RCTs, investigations of functional improvements were conducted in eight studies. Six studies reported that the prolotherapy group had a significant improvement in function compared to the control group [13,15,17,19-21]. One study showed functional improvement at 90 days after treatment, but after 360 days, both the prolotherapy and control groups showed similar results [14]. In one study, no significant improvement was noted in any of the groups at the end of the follow-up period [12]. However, unlike other studies which used a dextrose concentration of 10% or higher, this study only used a 5% concentration. When used clinically, dextrose concentrations higher than 10% are partly affected by inflammatory mechanisms, while concentrations less than 10% are considered noninflammatory [35,36]. Considering this, it is possible that a low concentration of dextrose could have affected the therapeutic effect. Although the degree of pain reduction and functional improvement is not completely consistent, there seems to be a correlation between the two in the studies that were included in this meta-analysis.
Although there were several positive aspects of our study, there are some limitations. First, despite recent studies being added, the number of trials eligible for inclusion in the meta-analysis was limited. Since the results regarding prolotherapy corresponding to the effects of corticosteroids and PRP were derived by analyzing only two studies, additional studies are needed. Second, there is heterogeneity in the pooled analyses; this is likely attributable to multiple factors, including differences in patient characteristics, control treatment, study design, injection protocol methods, dextrose concentrations, follow-up duration, and outcome assessment methods. A limited number of studies and heterogeneity have inhibited more detailed meta-analyses of subgroups. Third, due to a lack of a uniform longer-term follow-up duration across the studies, pooling of results could only be done with data collected between 6 months and one year of follow-up. Considering that prolotherapy is hypothesized to work by healing and regeneration over several months, reported results of effects may underestimate long-term benefits. Therefore, further studies (including cohort studies) are needed to evaluate the long-term effects. Fourth, since prolotherapy has been shown to have comparable effects to steroid injection and PRP, further studies should be conducted regarding cost effectiveness. Jahangiri et al. [21] compared prolotherapy and corticosteroids and mentioned that there was no significant difference in cost. In previous study, prolotherapy was more effective [14], and has a better cost advantage compared to PRP [37].
In the future, subgroup analysis should be performed to identify patients who respond most favorably to prolotherapy. There are several ways in which treatment strategies can vary; for example, dextrose concentrations/volumes may differ, the interval and total duration of treatment may differ, and the site of injection (intra- or extra-articular areas) may differ. Since there are no clear criteria or standard treatment, this should be discussed in the future. Reducing pain, improving functionality, and increasing patient satisfaction provide a solid foundation for further research in attempt of treatment standardization.
In conclusion, dextrose-based prolotherapy has been shown to have a positive and significantly beneficial effect for patients with chronic musculoskeletal pain, ranging from 6 months to 1 year. There is evidence that dextrose-based prolotherapy has a better therapeutic effect than exercise, and that it has a similar effect compared to PRP and steroid injection. Adequately powered, longer-term trials with uniform endpoints are needed to better elucidate the efficacy of prolotherapy.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
