A systematic review on conservative treatment options for OSGOOD-Schlatter disease Physical Therapy in Sport

Objectives: Osgood-Schlatter disease (OSD) is a sport- and growth-associated knee pathology with locally painful alterations around the tibial tuberosity apophysis. Up to 10% of adolescents are affected by OSD. Treatment is predominantly conservative. The aims of this systematic review are to comprehen- sively identify conservative treatment options for OSD, compare their effectiveness in selected outcomes, and describe potential research gaps. Methods: A systematic literature search was conducted using CENTRAL, CINAHL, EMBASE, MEDLINE, and PEDro databases. In addition, ongoing and unpublished clinical studies, dissertations, and other grey literature on OSD were searched. We also systematically retrieved review articles for extraction of treatment recommendations. Results: Of 767 identi ﬁ ed studies, thirteen were included, comprising only two randomised controlled trials (RCTs). The included studies were published from 1948 to 2019 and included 747 patients with 937 affected knees. Study quality was poor to moderate. In addition to the studies, 15 review articles were included, among which the most prevalent treatment recommendations were compiled. Conclusion: Certain therapeutic approaches, such as stretching, have apparent ef ﬁ cacy, but no RCT comparing speci ﬁ c exercises with sham or usual-care treatment exists. Carefully controlled studies on well-described treatment approaches are needed to establish which conservative treatment options are most effective for patients with OSD. © 2021 The Authors. Published by Elsevier This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).


Introduction
OsgoodeSchlatter disease (OSD) is a common osteochondrosis pathology during adolescent growth (Lohrer, Nauck, Scholl, Zwerver, & Malliaropoulos, 2012). In most cases, OSD is a clinical diagnosis with locally painful alterations around the tibial tuberosity apophysis. It results in knee pain, often severe enough to cause limping, sometimes accompanied by swelling or deformity, and frequently resulting in long-term symptoms with functional impairment. Patients experience pain on descending stairs, after prolonged periods of sitting with the knee immobile, while kneeling, and during sports activities (Gholve, Scher, Khakharia, Widmann, & Green, 2007). The knee pain can often be severe enough to cause limping. OSD as a growth-related condition is a relevant problem in young athletes. Suzue et al. investigated the prevalence of osteochondrosis in 494 child and adolescent soccer playersd198 players (40%) had positive knee findings; thirteen of those (6.5%) an OSD diagnosis (Suzue et al., 2014). Another study reported that up to 30% of OSD patients had bilateral involvement . Although OSD is more common in boys, with more girls becoming involved in sports the gender gap is narrowing (Domingues, 2013). OSD most frequently occurs between the ages of 8 and 13 years in girls, and between 10 and 15 years in boys (de Lucena, dos Santos Gomes, & Oliveira Guerra, 2011). A Brazilian study showed an OSD prevalence of 9.8% (11.0% in boys and 8.3% in girls) in a sample of 956 adolescent students (de Lucena et al., 2011). Hall et al. analysed data from 357 multi-sport and 189 single-sport female athletes and found that single-sport athletes have a four-times-higher risk of developing patellar tendinopathy and OSD than multi-sport athletes (Hall, Barber Foss, Hewett, & Myer, 2015).
The exact cause of OSD is unknown. It could be secondary to repetitive microtrauma of the tibial tuberosity or due to a tight quadriceps (de Lucena et al., 2011). A common hypothesis on the aetiology of OSD suggests an asynchronous development of bone and soft tissues, in particular the rectus femoris muscle, during the maturation stage (Tzalach, Lifshitz, Yaniv, Kurz, & Kalichman, 2016). This force results in irritation and, in severe cases, a partial avulsion of the tibial tubercle apophysis. The force is increased with higher levels of activity and especially after the periods of rapid growth typically seen in adolescence (Smith & Varacallo, 2019).
There is a remarkable lack of information on growth-related injuries in young athletes and their prevention and rehabilitation (Faude, R€ ossler, & Junge, 2013). OSD is frequently considered a selflimiting condition, but this perspective should be challenged. The healing period in the bradytrophic tissue of a growth plate under traction can last one to two years (Hefti, 2015)da significant period of time in the context of adolescent timespans. An effective treatment approach and the implementation of prevention programs before OSD onset are strongly needed (Guldhammer, Rathleff, Jensen, & Holden, 2019).
Although a wide range of treatment philosophies exist for OSD, it is predominantly treated conservatively. For the purposes of this review, conservative treatment is defined as icing, bracing, casting, or splinting (Hefti, 2015;Smith & Varacallo, 2019), and physiotherapy (e.g., stretching, strengthening, rest, and activity modification) (Smith & Varacallo, 2019). Non-operative treatment is defined as pharmacological treatment for pain relief, either orally with non-steroidal anti-inflammatory drugs (NSAIDs) (Smith & Varacallo, 2019) or by local anaesthetic injected into the painful area of the tibial tuberosity .
A recently published systematic review by Cairns et al. focused on the therapeutic interventions in children and adolescents with patellar-tendon-related pain . Despite the unlikeliness of OSD-related pain complaints spontaneously improving when treated with a "wait-and-see" approach, systematic evaluations of treatment strategies are currently lacking. Hamstring and quadriceps stretching and strengthening exercises are frequently recommended. However, the exact muscle-stretching techniques as well as the overall bundle of prescribed exercises are rarely welldescribed. An evidence-based, multi-management program should be available to facilitate affected adolescents' return to sport.
The literature agrees that surgery is not the most effective treatment for OSD . With regard to conservative and non-operative treatment options, some anecdotal and practical experience from conference presentations, and some empirical data on the potential benefits are available (Core Advantage, 2019;Strickland Protocol, 2016). However, an integrative view of conservative or non-operative treatment options and their effectiveness to improve recovery from OSD is currently missing.
Thus, the aims of this systematic review were: 1. To comprehensively identify conservative and non-operative treatment options for OSD and compare their effectiveness in selected outcomes. 2. To provide recommendations for evidence-based treatment options and for future research.

Methods
The review was conducted according to the Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) statement (Moher, Liberati, Tetzlaff, Altman, & Group, 2009). The systematic review was registered in PROSPERO (CRD42018106215).

Data sources and search strategy
Based on scoping searches, we expected a low number of eligible high-quality studies for inclusion. Therefore, we chose a highly sensitive search strategy in order to detect the largest possible set of relevant articles. We searched the following bibliographic databases: MEDLINE and Embase via OVID, the Cochrane Central Register of Controlled Trials (CENTRAL), the Cumulative Index to Nursing and Allied Health Literature (CINAHL), and the Physiotherapy Evidence Database (PEDro). The original search was completed on 13th April 2018 and the updated search on 6th January 2020. The search terms and strategy were developed by an information specialist (CAH). Text words (synonyms and word variations) and database-specific subject headings for OSD were used (Appendix 1). In order to identify possible additional studies, the bibliographic references of all included articles and key reviews were screened (reference chasing). The key reviews were retrieved during title abstract screening and also used for the extraction of therapy recommendations (see below). Furthermore, ongoing and unpublished clinical trials, dissertations and theses, congress abstracts, and other grey literature were retrieved from the following URLs using the search terms "Osgood AND Schlatter" and reviewed for inclusion/exclusion according to the eligibility criteria (ongoingtrials: www.science.gov, greylit.org, projectreporter. nih.gov/reporter.cfm, apps. who.int/trialsearch; dissertations/theses: www.opengrey.eu, search. proquest.com/pqdt/advanced/dissertations, www.dart-europe.eu, www.ndltd.org, oatd.org, www. openthesis.org; other grey literature: search.datacite.org). We also hand-searched the following journals: "Sportphysio" (from Volume November 1, 2013 to Volume February 6, 2018; update to Volume December 7, 2019), "Physiopraxis" (from Volume January 9, 2011 to Volume March 16, 2018; update to Volume January 18, 2020), "Pediatric Physical Therapy" (from Volume Spring 25, 2013 to Volume April 30, 2018; update to Volume January 32, 2020), "Physiotherapy" (from Volume March 96, 2010 to Volume 104 March 2018; update to Volume 105 December 2019), "Monatsschrift Kinderheilkunde" (from Volume 163 2015 to Volume 166 2018; update to Volume 168 January 2020), "Journal of Children's Orthopaedics" (from Volume March 1, 2007 to Volume December 10, 2016; update to Volume December 13, 2019), and "Physical Therapy in Sport" (from Volume February 11, 2010 to Volume March 30, 2018; update to Volume January 41, 2020) using the terms "Osgood AND Schlatter".

Eligibility criteria and study selection
The PICOS framework (Population, Intervention, Comparison/ Control, Outcome, and Study design) was used to define the following inclusion criteria: patients with OSD of six to 28 years of age (P); analysis of at least one conservative or non-operative intervention in either single-arm or controlled study design (I); if applicable, compared with an additional intervention or no treatment (C); assessment of at least one of the following outcomes: pain, symptoms duration, function (e.g., capacity to kneel), range of motion, muscle length or sport participation (O); the study design was either a prospective-or retrospective-observational study, a case control study, a case series, a randomised or non-randomised trial, or an abstract-only publication, (S). Only records written in English, German, or French were considered. Animal studies, case reports, cross-sectional studies, letters, editorials, and diagnostic or other assessment studies were excluded. We applied no publication date restrictions. Inclusion criteria for the key reviews were English, German, or French language and a content specific to OSD (e.g., reviews with general overuse sports injuries were excluded). Studies and key reviews were independently selected by two investigators (CN, OF). A final decision on eligibility was achieved by consensus.

Data extraction
Data extraction was carried out by one author (CN). The following data were extracted: authors, year, study design, country, participants (e.g., sex, age, sample size for intervention, and control groups), type of intervention, duration and time of intervention, outcome measures (pain, function, sport participation, and any additional outcome reported), and main conclusions.

Quality assessment
The Physiotherapy Evidence Database (PEDro) scale, a valid measure of the methodological quality of a clinical trial, was used (Maher, Sherrington, Herbert, Moseley, & Elkins, 2003). This scale comprises 11 dichotomous items with a maximal score of 10. Studies were rated by one researcher (CN), who was not blinded to study authors, place of publication, and results. A PEDro score of seven or greater was considered as "high quality", studies with a score of five or six were considered to be of "moderate quality", and those with a score of four or less "poor quality".

Data analysis
Due to the heterogeneous nature of the included studies (e.g., different study designs, interventions, outcome measures, and quality of data), a quantitative analysis was not applicable. We narratively synthesised the results based on the domains of interest. In addition, all retrieved reviews were evaluated with regard to recommendations and referenced sources for the treatment of OSD.
The included studies were published between 1948 and 2019 and comprised 747 patients with 937 affected knees (one study (Strickland et al., 2008) did not report the number of affected knees). Of these patients, 563 were male and 119 female. The sex of the remaining 65 patients from two studies was not reported (Trail, 1988;Yatsuka et al., 1992). Age range was 9e28 years. There was substantial heterogeneity among the studies in terms of population size, patient age, and sex (Table 1). Follow-up periods varied from one month to nine years. Treatment methods were analgesics (NSAIDs), avoiding sports activity, ice, injections, thigh-muscle stretching, immobilisation of the knee joint with a resin cast, plaster or an infrapatellar strap, tape or knee bandaging, physiotherapy, massage, and extracorporeal shock wave therapy. No detailed descriptions of exercises or physiotherapy programs were provided.
The quality assessment using the PEDro scale indicated that the overall study quality was very low (Table 2): only one study was of high quality (Nakase et al., 2019), another study was considered to be of moderate quality , ten studies had a PEDro score of four or less, indicating poor quality (Duperron et al., 2016;Ehrenborg, 1962;Hussain & Hagroo, 1996;Krause et al., 1990;Kridelbaugh & Wyman, 1948;Levine & Kashyap, 1981;Lohrer et al., 2012;Reichmister, 1969;Trail, 1988;Yatsuka et al., 1992), and one study could not be assessed because only the abstract was available (Strickland et al., 2008).

Detailed description of studies
The studies are described in chronological order starting from the oldest. Kridelbaugh (Kridelbaugh & Wyman, 1948) showed a subjective improvement after anterior thigh taping in 46% of patients. No further details were described. In 1962, Ehrenborg (Ehrenborg, 1962) retrospectively analysed a series of 170 patients. Of the 218 affected knees, 144 were treated with a cast for four to six weeks, and 74 knees were not immobilised. The average duration of symptoms was 14.6 months in the plaster-treated group versus 27.8 months in the non-immobilised group, suggesting effectiveness of immobilisation. Reichmister (Reichmister, 1969) applied combined injections of corticosteroid and anaesthetics (Decadron® and Xylocaine®) into the infrapatellar bursa. All ten treated cases were completely cured by the time of the final injection (on average 1.9 injections). Levine (Levine & Kashyap, 1981) used an infrapatellar strap, which showed an improvement in 19 out of 24 knees after a period of 6e8 weeks. In 1988, Trail et al. (Trail, 1988) compared surgery (tibial sequestrectomy) with conservative treatment in a retrospective study involving 51 patients . Conservative treatment and surgery showed no relevant differences. Conservative treatment was, however, not appropriately described. In the retrospective study by Krause et al. (Krause et al., 1990), 50 OSD patients (69 knees) were instructed to do what they could do during the acute phase of the disorder and no treatment or activity restrictions were documented. At the last follow-up, 36 (76%) had no limitations, but for 60%, kneeling continued to be uncomfortable. An additional 12 OSD patients had spent some time in plaster. Only ten patients are mentioned in the article: three had chronic symptoms and seven were unable to kneel. Yatsuka et al. (Yatsuka et al., 1992) examined 15 knees with OSD, which were treated with hamstring stretching exercises without any further therapy. Hamstring stretching resulted in pain relief for 11 out of the 15 knees. Hussain and Hagroo (Hussain & Hagroo, 1996) followed 261 patients (365 knees) for 1e2 years and reported that 237 patients (91%) responded well to heterogeneous conservative measures including activity modification, rest with NSAID medication, and knee bandaging. Strickland et al. (Strickland et al., 2008) conducted a pilot study with 25 patients with OSD suffering from symptoms for 8 months on average (range of 1 weeke36 months). Physiotherapy treatment consisted of myofascial release massage and stretching of the quadriceps group. When patients achieved a wall-squat with full range of motion of the knees (on average after 20 days), they were discharged and able to return to their sporting activities as normal, with no reported further problems. Patients returned to their sport in a shorter time than the authors anticipated. At various follow-up dates (1e5 years), only two patients reported recurrence, though they likely had not followed the recommended advice on stretching. Topol et al.  randomly assigned 54 patients to usual care (hamstring stretching, quadriceps strengthening exercises), local anaesthetic (lidocaine injection), or local anaesthetic plus dextrose injection. Average Nirschl Pain Phase Scale scores improved more in the dextrose-treated knees (from 4.6 to 0.7) than the lidocaineonly-treated (from 4.2 to 1.8) or usual-care-treated knees (from 4.3 to 3.1). Moreover, the duration of both sports limitation and sports-related symptoms was reduced in the dextrose injection group compared to the other groups, and all patients who gave up sports or were unable to perform exercises were part of the usual care group. Lohrer et al. (Lohrer et al., 2012) treated 14 patients (16 knees) with radial extracorporeal shock waves. After 5.6 years, 12 knees (75%) reached the best score on a patellar tendinopathy questionnaire. Duperron et al. (Duperron et al., 2016) immobilised 30 OSD patients' knees with a plaster for 4 weeks. Time until resuming sports was on average 14.4 ± 14.2 weeks, but 10 out of 30 patients still suffered from pain after plaster removal. The latest study by Nakase et al. (Nakase et al., 2019) readdressed the effectiveness of the dextrose injection. Thus, 38 patients who received non-invasive therapy for more than 1 month and had no improvement were randomly assigned into 2 groups to receive an injection with dextrose (plus anaesthetic) or saline (plus anaesthetic) in a double-blind procedure. Although both groups displayed marked improvements, negligible differences were found between the two groups at any follow-up time, which challenged the previous results from Topol et al. .
The most frequently recommended treatments were activity modification (   parent education, modification of sports activities, NSAIDs, ice, hamstring stretching, and shock-absorbing insoles were recommended. Patients classified with grade three (pain does not disappear between sports activities) were advised to rest, be immobilised in a cast, and undergo specific rehabilitation programs ). Circi et al. and Ladenhauf et al. recommended to reduce sports activity and perform non-impact exercise such as swimming or cycling Ladenhauf et al., 2019). Nührenborger et al.  recommended ice, however application techniques and characteristics were not described in detail.

Discussion
The main result of this review is the absence of high-quality studies evaluating the effectiveness of interventions for the treatment of OSD. The number of included studies was low and the studies were heterogeneous. Hence, conducting a quantitative analysis was impossible. The number of available review articles covering OSD treatment options is even larger than the number of available original studies. The problem of OSD in athletically active children and adolescents is being recognised, but evidence-based guidelines do not exist, implying that treatment recommendations are based on clinicians' experience and anecdotal evidence.
OSD is a long-term pain condition that occurs during adolescent growth with a potential to develop into chronic knee pain. As for any health condition, clinicians aim to avoid chronic problems and offer patients evidence-based treatment options. Currently, a lack of evidence or consensus causes uncertainty on what can be recommended for OSD  and treatment is based merely on clinical experience and expert opinion (Ladenhauf, Seitlinger, & Green, 2020). For example, therapists individually adapt physiotherapy exercises. In the available review articles, the discussed treatment options are quite comparable. Remarkably, the same publications were repeatedly referenced and only one review article based its recommendationsdin partdon an RCT. Only one review article advised core stability and balance training . In the included articles, the most Table 2 PEDro scale scores of the included studies.
PEDro Scale Item Kridelbaugh & Wyman, 1948Ehrenborg, 1962Reichmister, 1969Levine & Kashyap, 1981Trail, 1988Krause et al., 1990Yatsuka et al., 1992Hussain & Hagroo, 1996Strickland et al., 2008  frequently mentioned therapy was injections (Nakase et al., 2019;Reichmister, 1969;Topol et al., 2011), followed by splinting methods using a patellar strap, tape, or bandage (Hussain & Hagroo, 1996;Kridelbaugh & Wyman, 1948;Levine & Kashyap, 1981). Immobilisation with a cast was studied twice (Duperron et al., 2016;Krause et al., 1990). Lohrer et al. tested shock wave therapy in a pilot study, which is another type of passive intervention (Lohrer et al., 2012). One study investigated the use of two different hamstring stretching techniques (Yatsuka et al., 1992). Other researchers used a combination of surgery (tibial sequestrectomy), casting, injections, and physiotherapy (Ehrenborg, 1962;Trail, 1988). Of the two identified RCTs, both examined injection therapy with or without a hypertonic dextrose solution, which is also known as prolotherapy. Apart from OSD, hypertonic dextrose is also used in other tendinopathies and fasciopathies with unclear effectiveness (Sanderson & Bryant, 2015). Whereas Nakase et al. (Nakase et al., 2019) ran a double-blind comparison of two injection groups, Topol et al. (Topol et al., 2011) conducted a three-armed RCT that also analysed a usual care group. Concerning the controversial conclusions of the two RCTs on the effectiveness of dextrose, there is no obvious resolution. Different factorsdincluding the use of inappropriate comparator (saline), quality of blinding, divergent outcome scales, and statistical methodsdcould have played a role Nakase et al., 2019;Rabago, Topol, Podesta, Cheng, & Fullerton, 2020). More research is needed to conclusively learn about the potential benefits of hypertonic dextrose injections for OSD. When considering all of the studies and their results, a "consensus" treatment recommendation may be the modification of physical activity. Taping or patellar strapping is frequently recommended, which is similar in motivation to the approach of isometric strengthening of the quadriceps in knee extension and stretching the hamstrings. However, there is a lack of RCTs that provide high-level evidence in favour of stretching or strengthening exercises for the lower extremity in children or adolescents with OSD. Existing evidence on exercise therapies is contradictory and controversial. Considering the lack of evidence with regard to the efficacy of treatment options in general, it is obvious that recommendations on specific intervention characteristics (e.g., duration, frequency, intensity) are missing all the more . Thus, unfortunately, the formulation of evidencebased treatment recommendations, as defined in our aims for this systematic review, is currently not feasible.
Review of the grey literature offers valuable information about possible treatment programs, which are frequently commercially advertised. A prominent example is the so-called Strickland protocol, which was presented at the European College of Sports Science Conference in Portugal in 2008 (Strickland et al., 2008). The protocol is mainly a combination of myofascial release massage (2 min daily) and active stretching of the quadriceps femoris muscle. Comparable therapeutic approaches may appear effective and successful.
Although it is important to seek evidence-based therapy for existing OSD conditions, it is also important to focus on prevention strategies that reduce overuse injuries. The long-term impact of a person with OSD being unable to participate in typical physical activity and sports-team peer groups should not be underestimated . The underlying reasons why one adolescent develops OSD yet an equally active peer does not, is not well-understood. With regard to injuries in general, scientifically evaluated, exercise-based prevention programmes already exist for young athletes (R€ ossler et al., 2014). For instance, a multi-national cluster-RCT found that an injury prevention warm-up programme is effective in reducing overuse injury rate to the lower extremities in young football players (R€ ossler et al., 2018). However, though included in this injury category, no specific data on OSD were reported. The prevention of pediatric overuse injuries requires a comprehensive, multidimensional approach that may include improved injury surveillance, identification of risk factors for injury, thorough physical examination prior to participation, supervision and education, improved training and conditioning programs, and delayed specialisation (Valovich McLeod et al., 2011).

Methodological considerations
Strengths of this systematic review are the inclusion of German, French, and English articles and the searching of grey literature sources. To the best of our knowledge, this level of detail has not been achieved in previous reviews. Limitations of this systematic review are the heterogeneity of the included studies, the inclusion of non-peer-reviewed studies, the widespread lack of control groups and the missing data in patient characteristics (e.g., number of affected knees). Furthermore, the PEDro scale quality assessment confirmed that the overall study quality is a major limitation.

Conclusion and future directions
No evidence exists on the effectiveness of specific exercise programs for patients with OSD; only poor evidence exists for the use of injections with local anaesthetic. In such absence of highquality evidence, the first step should be to rely on expert consensus for best-practice recommendations (Table 3) (Minas & Jorm, 2010). It is desirable that, in a subsequent step, high-quality clinical RCTs be conducted. Future investigations should focus both on well-described and approved treatment approaches and on specific exercise programs. Children are generally regarded as the future of our society and, therefore, their health should be of particular importance (Faude, R€ ossler, & Junge, 2013).

Funding statement
No funding.

Author contributions
CN and OF designed and organised this systematic review. CAH organised the literature search process. CN and OF analysed the data and wrote the manuscript draft. CAH critically revised the article.

Ethical statement
As this is a Systematic Review, Ethical Approval was not required.

Declaration of competing interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.