*Review* **Extraosseous Ewing Sarcoma in Children: A Systematic Review and Meta-Analysis of Clinicodemographic Characteristics**

**Maher Ghandour <sup>1</sup> , Burkhard Lehner <sup>1</sup> , Matthias Klotz <sup>2</sup> , Andreas Geisbüsch <sup>1</sup> , Jakob Bollmann <sup>1</sup> , Tobias Renkawitz <sup>1</sup> and Axel Horsch 1,\***


**\*** Correspondence: axel.horsch@med.uni-heidelberg.de

**Abstract: Background:** We conducted this systematic review to provide comprehensive evidence on the prevalence, clinical features and outcomes of young extraosseous Ewing sarcoma (EES) cases. **Methods:** PubMed, Scopus, Web of Science, and Google Scholar were searched for articles reporting the occurrence of EES among children and adolescents (<21 years). The primary outcome included the rate of occurrence of EES among children and adolescents, while the secondary outcomes included the descriptive analyses of the demographic characteristics, tumor characteristics, and clinical outcomes of the affected cases. The data are reported as the effect size (ES) and its corresponding 95% confidence interval (CI). **Results:** A total of 29 studies were included. Twenty-four reported instances of childhood disease among all the EES cases [ES = 30%; 95%CI: 29–31%], while five studies reported extraosseous cases among the pediatric EES cases [ES = 22%; 95%CI: 13–31%]. The thorax is the most common location of childhood EES [33%; 95%CI: 20–46%] followed by the extremities [31%; 95%CI: 22–40%]. Concurrent chemotherapy and radiotherapy [57%; 95%CI: 25–84%] was the most commonly implemented management protocol in the pediatric EES cases. The rate of no evidence of disease and 5-year overall survival was 69% for both outcomes. Mortality occurred in 29% of cases, while recurrence and secondary metastasis occurred in 35% and 16% of cases, respectively. **Conclusions:** Our findings provide insight into the clinical features and outcomes of EES among children and adolescents.

**Keywords:** Ewing sarcoma; extraosseous; children

#### **1. Introduction**

The Ewing sarcoma family of tumors (ESFT) is a collection of small, rounded tumor cells that have similar neural histological and genetic characteristics [1–4]. ESFT is categorized into four types based on the origin of the tumor: Ewing sarcoma of the bone, peripheral primitive neuroectodermal tumor (pPNET), Askin tumor, which originates from the chest wall, and, finally, the extraosseous or extraskeletal Ewing sarcoma (EES). EES, which occurs in around 20% of ES cases, typically originates from the soft tissues of the trunk and extremities [5], and the majority of these cases are reported among patients who are 10–30 years of age [6].

Based on a previous report, the incidence of EES is 0.4 per million individuals, which is lower than that of ES of the bone by 10-fold [7]. Although uncommon, the occurrence of EES seems to have a bimodal distribution, where there is a peak in the occurrence rate among children (<5 years) and adults (>35 years) [8], with an increased likelihood of presenting among older populations compared to ES of the bone. Unlike Ewing earcoma of the bone, no evidence supports a link between the tumor and race or biological sex [8–10].

The management of EES includes surgery [11] and chemotherapy [10,12,13] in resectable tumors. Under unresectable conditions, radiotherapy is usually considered [14].

**Citation:** Ghandour, M.; Lehner, B.; Klotz, M.; Geisbüsch, A.; Bollmann, J.; Renkawitz, T.; Horsch, A. Extraosseous Ewing Sarcoma in Children: A Systematic Review and Meta-Analysis of Clinicodemographic Characteristics. *Children* **2022**, *9*, 1859. https:// doi.org/10.3390/children9121859

Academic Editor: Vito Pavone

Received: 17 October 2022 Accepted: 26 November 2022 Published: 29 November 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

According to the National Comprehensive Cancer Network (NCCN), the optimum management of EES remains not clearly defined [15,16], although some studies have highlighted an added value of surgery among EES cases compared to Ewing sarcoma of the bone in terms of better survival rates [17,18]. In general, the prognosis of EES is more favorable than that of the bone [9,10].

To date, there is no clear picture regarding the occurrence rate of EES among children and adolescents (<21 years), as well as their demographic characteristics, tumor characteristics (i.e., location), treatment modalities, and clinical outcomes (i.e., survival, mortality, recurrence). Therefore, we conducted this systematic review and meta-analysis to provide collective evidence regarding the clinical characteristics and outcomes in this patient population.

#### **2. Materials and Methods**

### *2.1. Study Design and Search Strategy*

This systematic review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [19]. A protocol was not registered, since it is not mandatory, as per several recommendations [20,21]. On 26 July 2022, PubMed, Scopus, Web of Science, and Google Scholar were searched for articles that report the presentation of EES in the pediatric population (children and young adolescents <21 years of age). Of note, only the first 200 records from Google Scholar were retrieved and screened according to recently published guidelines [22]. We updated the database search on 24 August 2022 to ensure that no additional relevant reports had been published prior to the qualitative and quantitative analyses [23].

We used a combination of keywords and terms in our search, which included the following: ("Ewing Sarcoma" OR "Ewing's Sarcoma") AND (adolescen\* OR Child\* OR Pediatric\* OR "young adult") AND ("soft tissue" OR extraskeletal OR extraosseous) AND (clinicopathologic\* OR "clinical feautre" OR "clinical characteristic\*" OR "clinical outcome\*"). The terms of the Medical Subject Headings (MeSH) were also added (particularly in PubMed) to retrieve all the possibly relevant articles. The detailed search criteria used for each database are described in Supplementary Table S1.

Moreover, we conducted a manual search to find any relevant articles that may have potentially been excluded during the screening phase or were not found during the database search [24,25]. This strategy was conducted through three different approaches: (1) screening the titles of the reference list of the final included papers, (2) reading the titles and abstracts of articles similar to final included studies through the "similar articles" function on PubMed, and (3) conducting a random search on Google using keywords similar to those of the original database search, such as: "Ewing sarcoma" + "child". It is noteworthy that no filters were used during the database search regarding the language of the research paper, year in which the paper was published, or the country of the first author.

#### *2.2. Eligibility Criteria*

The methodology and design of this review were conducted as per the PICO framework [26,27], including the population (pediatric cases of EES), intervention (none), comparison (none), and outcome (primary outcome: prevalence rate of EES in children and adolescents; secondary outcomes: clinicodemographic characteristics, tumor characteristics, and clinical outcomes in pediatric cases of EES).

For articles to be included, a study had to: (1) report original data, (2) include cases of EES, (3) report cases aged <21 years. On the other hand, studies were excluded if they were compliant with one of the following criteria: (1) non-original research (i.e., review articles, editorials without human data, commentaries, theses, conference abstracts/posters, and books), (2) animal, in vivo, and in vitro studies, (3) case reports and case series of <5 cases, (4) studies reporting EES cases of mixed ages (children, adolescents, adults, and elderly) without stratifying the cases according to their age, (5) studies reporting Ewing sarcomas of

mixed origin (extraosseous and skeletal) in children without stratifying the cases according to their origin, and (6) duplicated records.

#### *2.3. Screening and Study Selection*

Following the retrieval of records through the database search [28], the references were imported to EndNote (Version 8) for duplicate removal and to organize the screening sheet [29]. The screening sheet included the following: article ID, list of authors' names, year of publication (YOP), research paper's title, DOI, journal name, and abstract. The screening was carried out in three separate stages: title, abstract, and full-text screening. All of these steps were performed by two sets of two reviewers each. Any differences between the reviewers were reviewed and resolved by the senior author [30].

Significantly, upon reviewing the literature, two categories of articles were found to be consistent with our eligibility criteria. The first group of articles included patients with EES, among whom pediatric cases were counted, and the second group of articles included pediatric cases, of whom the origin of Ewing sarcoma was determined (extraosseous or skeletal). Both of these categories were included, extracted, and presented separately in our review.

#### *2.4. Extraction and Quality Assessment*

The data extraction process was carried out in a similar manner as the screening stage [31]. The senior author designed a pilot data extraction sheet through the Excel software (version 2021) that was consistent with the study objectives. The sheet included 5 domains. The first domain highlighted the baseline characteristics of the included studies (authors' names, year of publication, country, study design, sample size, and follow-up duration). The second domain included the demographic characteristics of the included participants, such as age and biological sex. The third domain included the location of the EES among the pediatric cases (i.e., cranium, female genital tract, orbit, head and neck, pelvis, extremities, thorax, abdomen). The fourth domain included the tumor's characteristics (i.e., management modalities (i.e., surgery alone, surgery combined with radiotherapy, surgery combined with chemotherapy, etc.). The final domain included the patients' clinical outcomes in terms of the overall survival (OS), progression-free survival (PFS), disease-specific survival (DSS), secondary metastasis, no evidence of disease (NED), mortality, and recurrence. Two reviewers extracted the data from the included studies for further qualitative and quantitative synthesis, as per the recommended guidelines [32,33].

#### *2.5. Data Synthesis*

All quantitative analyses were conducted using the STATA software (version 17) with the metaprop command [34]. The exact cimethod [34] was used to pool the effect size (ES)—occurrence rate of EES in the pediatric cases—along with its 95% confidence interval (CI). Importantly, for the purposes of discussing the findings of our review, the term ES will refer to the effect size and not Ewing sarcoma (which will not be abbreviated in this manuscript). The random-effects and fixed-effects models were used according to the presence or absence of heterogeneity, respectively [35,36]. Heterogeneity was measured using the *I* 2 statistic, where a value of >50% or a *p*-value of <0.05 indicates significant heterogeneity.

#### **3. Results**

#### *3.1. Search Results*

A summary of the results of the electronic database search, as well as the screening stage, is provided in Figure 1. The initial database search resulted in 2611 references, out of which 179 duplicated records were found and removed using the EndNote software (version 8). The titles and abstracts of 2432 articles were screened, resulting in 276 articles eligible for full-text screening. The full texts of six studies were not found and, therefore, these were excluded. A total of 241 studies were excluded as follows: adult cases (n = 22), skeletal involvement (n = 23), case reports (n = 73), duplicated records (n = 2),

elderly cases (n = 2), in vitro studies (n = 3), mixed-age populations (n = 82), no data on sarcoma origin (n = 5), non-Ewing sarcoma (n = 23), and reviews (n = 6). The updated and manual search yielded no more studies, so that the final number of included studies was equal to 29 reports. Twenty-four articles reported the rate of childhood cases among those with EES (of all ages), while five studies reported the rate of extraosseous involvement in pediatric Ewing sarcoma (of mixed origin—skeletal and extraosseous) cases. (n = 2), in vitro studies (n = 3), mixed-age populations (n = 82), no data on sarcoma origin (n = 5), non-Ewing sarcoma (n = 23), and reviews (n = 6). The updated and manual search yielded no more studies, so that the final number of included studies was equal to 29 reports. Twenty-four articles reported the rate of childhood cases among those with EES (of all ages), while five studies reported the rate of extraosseous involvement in pediatric Ewing sarcoma (of mixed origin—skeletal and extraosseous) cases.

A summary of the results of the electronic database search, as well as the screening stage, is provided in Figure 1. The initial database search resulted in 2611 references, out of which 179 duplicated records were found and removed using the EndNote software (version 8). The titles and abstracts of 2432 articles were screened, resulting in 276 articles eligible for full-text screening. The full texts of six studies were not found and, therefore, these were excluded. A total of 241 studies were excluded as follows: adult cases (n = 22), skeletal involvement (n = 23), case reports (n = 73), duplicated records (n = 2), elderly cases

*Children* **2022**, *9*, 1859 4 of 16

**3. Results** 

*3.1. Search Results* 

**Figure 1.** A PRISMA diagram showing the database search and screening results of the review. **Figure 1.** A PRISMA diagram showing the database search and screening results of the review.

#### *3.2. Baseline Characteristics of the Included Studies*

*3.2. Baseline Characteristics of the Included Studies*  The baseline characteristics of the included studies are presented in Table 1. Among the 24 studies that reported the rate of childhood cases among all the EES cases (adults and children), two were conducted in the United Kingdom (UK), two in China, one in Germany, three in India, three in Italy, one in Japan, one in Korea, one in Turkey, and ten in the United States (US). Five studies were case series, fourteen were retrospective chart reviews, one was a registry-based study, three were SEER-based studies, and one was a The baseline characteristics of the included studies are presented in Table 1. Among the 24 studies that reported the rate of childhood cases among all the EES cases (adults and children), two were conducted in the United Kingdom (UK), two in China, one in Germany, three in India, three in Italy, one in Japan, one in Korea, one in Turkey, and ten in the United States (US). Five studies were case series, fourteen were retrospective chart reviews, one was a registry-based study, three were SEER-based studies, and one was a secondary analysis of three prospective studies. The number of included patients with EES in the individual studies ranged from 8 [37–39] to as high as 3178 [40] patients, with a total sample size of 5752 patients with EES. The follow-up was reported in 16 studies, ranging from as low as 0.9 [41] months to as high as 349 months [42].

Among the five studies that reported the rate of EES among children with ES, one study was conducted in the Netherlands, three in the US, and one in China. Three studies were retrospective chart reviews, one was a multicenter cohort study, and one was a secondary analysis of two clinical trials. The sample size of included pediatric patients with Ewing sarcoma ranged from 18 [43] to 1039 cases [9], with an overall sample size of 1190. The follow-up duration was reported in only two studies ranging from 56.4 [44] to 120 months [45].


**Table 1.** Baseline characteristics of the included studies.

\* Data are reported as the median and not the mean. YOP: year of publication; USA: United States of America; NR: not reported; UK: United Kingdom; FU: follow-up; EES: extraosseous Ewing sarcoma.

#### *3.3. Demographic Characteristics of the Included Participants*

The demographic characteristics of the included patients are illustrated in Table 2. Among the studies that included patients with EES regardless of their age, the rate of affected children ranged from as low as 5.63% (31 out of 550 cases) [60] to as high as 100% [55,56]. The pooled rate of childhood EES among the patients with EES was 30% [3001 patients, 95%CI: 29–31%; I<sup>2</sup> = 99.01%]. Out of the pediatric cases diagnosed with EES, 52.45% were males (224 out of 427 patients).


**Table 2.** The demographic characteristics of the included participants in each study with an overall estimation of the rate of the presentation with EES in childhood.

YOP: year of publication, N: number, T: total sample size, EES: extraosseous Ewing sarcoma.

Among the studies that included pediatric cases of Ewing sarcoma regardless of its origin, the rate of presentation with a disease of extraosseous origin ranged from 10% (6 out of 60 cases) [45] to 55.55% (10 out of 18 cases) [42], with an overall pooled rate of 22% [1190 patients, 95%CI: 13–31%; I<sup>2</sup> = 19.28%]. Out of the pediatric cases diagnosed with EES, 54.70% (122 out of 223) were males.

#### *3.4. The Location of EES in the Pediatric Cases*

Among the included studies, only 13 reported data regarding the location of EES among the pediatric cases (Table 3). The pooled meta-analysis revealed that the thorax was the predominant site where EES occurred [33%; 95%CI: 20–46%] followed by the extremities [31%; 95%CI: 22–40%], the head and neck [14%; 95%CI: 7–21%], the pelvis [13%; 95%CI: 9–16%], the abdomen [10%; 95%CI: 4–16%], the spine [8%; 95%CI: 6–11%], the intracranial space [8%; 95%CI: 1–33%], and finally the orbit [2%; 95%CI: 0–4%]. Of note, among the pediatric cases, the occurrence of EES in the skin, the kidney, and the female genital tract was scarcely reported, and the performance of a meta-analysis was not feasible due to the lack of sufficient data.

**Table 3.** The location of extraosseous Ewing sarcoma in the pediatric cases.


YOP: year of publication; N: number of cases; T: total number of pediatric cases of EES; EES: extraosseous Ewing sarcoma; H&N: head and neck; ES: effect size; CI: confidence interval.

#### *3.5. The Characteristics of EES among the Pediatric Cases*

Among the included studies, the management modalities in childhood EES were described and reported in ten studies (Table 4), among which concurrent chemotherapy and radiotherapy [13 patients, 57%; 95%CI: 25–84%] was the most frequently employed treatment protocol, followed by surgery combined with radiotherapy [236 patients, 55%; 95%CI: 28–82%], surgery alone [223 patients, 53%; 95%CI: 37–68%], surgery combined with chemotherapy [36 patients, 29%; 95%CI: 5–52%], and finally radiotherapy alone [219 patients, 16%; 95%CI: 11–21%].

**Table 4.** Trends in the management modalities of pediatric cases of extraosseous Ewing sarcoma reported in the literature.


YOP: year of publication; CT: chemotherapy; RT: radiotherapy; ES: effect size; CI: confidence interval; N: number of cases; T: total sample of pediatric cases of EES; EES: extraosseous Ewing sarcoma.

#### *3.6. The Clinical Outcomes of EES among the Pediatric Cases*

The clinical outcomes associated with childhood EES are presented in Table 5. The 5-year OS was reported in 11 studies, out of which 664/1066 pediatric EES cases survived. The pooled 5-year OS rate was 69% [95%CI: 56–81%]. The 5-year PFS, DSS, and DMFS were reported in only a single study, which was not enough to derive conclusions or be for

the data to be included in a meta-analysis. Seven studies reported no evidence of disease among 257 out of the 288 cases, with a pooled rate of 69% [95%CI: 51–87%]. Morality was reported in ten studies, where 120 deaths occurred among 404 pediatric cases of EES, with a pooled mortality rate of 29% [95%CI: 25–33%]. Meanwhile, recurrence was reported in five studies (19 cases out of 60 pediatric EES cases), with a pooled recurrence rate of 35% [95%CI: 16–54%]. Finally, secondary metastasis was reported in three studies, occurring in 38 cases out of 236 pediatric EES patients, with a pooled rate of 16% [95%CI: 11–21%].


**Table 5.** The clinical outcomes of pediatric patients with extraosseous Ewing sarcoma.

YOP: year of publication; N: number of cases of the outcome; T: total number of pediatric cases of EES; EES: extraosseous Ewing sarcoma; ES: effect size; CI: confidence interval; OS: overall survival; PFS: progressionfree survival; DSS: disease-specific survival; NED: no evidence of disease; DMFS: distant-metastatic free survival; N/A: not applicable for meta-analysis.

#### **4. Discussion**

There is limited evidence regarding the occurrence rate and clinical characteristics of EES in children. Our systematic review is the first to comprehensively discuss the prevalence, clinical features, and outcomes of EES patients of pediatric age (less than 21 years). A summary of our key findings can be found in Table 6. Overall, a total of

29 studies reporting on 5752 patients were analyzed. In our study, we found that the rate of affected children and adolescents with EES in a population with EES (mixed age) varied substantially between the studies, ranging from 5.63% [60] to as high as 100% [55]. This discrepancy could be related to the design and methodology of the included studies, since some studies included patients with EES regardless of the age group at baseline, while a few studies included pediatric cases of EES at baseline [55,56]. Overall, our meta-analysis revealed that 30% of the EES cases occurred among children and adolescents. Consistent with previous observations [8,9], no link was noted between EES presentation in children and biological sex. The pooled rate of male pediatric patients with EES was 52.45%, which is relatively similar to that of female cases (47.55%).


**Table 6.** Summary of the key findings on the pediatric EES cases in our review.

In addition, five studies included children affected with ES at baseline, and then the origin of the tumor was analyzed in these cases. The rate of EES out of all the ES types ranged from 10% to 55.55% among the individual studies. Again, the difference in reported rates could be related to the design and methodology implemented in each study. That being said, in our meta-analysis, the rate of EES occurrence among the pediatric ES cases was 22%, of whom 54.70% were males.

Data on the location of EES among pediatric cases is scarce, since the majority of the available studies in the literature include patients with mixed ages and tend to stratify the outcomes (i.e., survival) based on age (children vs. adults or the elderly), without stratifying the clinical characteristics or tumor characteristics based on the age of the examined patients. Therefore, the data reported in our review regarding the EES location in the pediatric cases rely mainly on case series with a case-by-case description of the tumor characteristics. Thirteen studies reported relevant data on the location of EES, and our pooled meta-analysis revealed that the thorax is the most predominant origin for EES in children and adolescents, followed by the extremities, the head and neck, the pelvis, the abdomen, the spine, and the intracranial space, respectively. In certain cases, the EES originated in the orbit among the pediatric cases; however, the occurrence rate did not surpass the rare event assumption (>5%). Additionally, other sites, such as the great toe [63], the mesocolon [64], the frontal sinus [65], and the penis [66], have been described as rare cases. Moreover, the kidneys [41,53], the skin [37], and the female genital tract [49] have been reported as sites of origin of EES in pediatric cases in several case series; however, not enough data were present to perform a meta-analysis of the prevalence in this case.

There is a debate on the best management approach for EES cases occurring in children, and this uncertainty is related to the rarity of EES, the discrepancy in its clinical presentation, and the differences in the patients' characteristics [67]. In addition, this patient population is underrepresented in clinical trials directed towards the investigation of the efficacy and safety of various treatment modalities among pediatric cases of EES. In our review, only ten studies reported the treatment modalities according to different age groups, and the majority of the data were pooled from case series. Overall, the majority of cases were treated with concurrent chemotherapy and radiotherapy (57% of cases), followed by surgery and radiotherapy (55%), surgery alone (53%), surgery and chemotherapy (29%), and radiotherapy alone in cases of unresectable tumors (16%). It is important to mention that the confidence interval of these reported rates is wide, reflecting the imprecision of the reported effect estimates. Therefore, these data should be interpreted with caution and should not be perceived as representative of the EES pediatric population. More data from properly designed research studies are still needed to confirm this observation. Additionally, the available data did not present survival outcomes stratified by these treatment modalities in the pediatric cases separately. Therefore, future studies should carefully consider stratifying data (clinical characteristics and outcomes) based on the origin of the tumor (skeletal vs. extraskeletal) and age of the included patients (children vs. adults vs. the elderly).

In our study, we found that a great proportion of pediatric EES patients have a preferable prognosis in terms of their 5-year overall survival (with an overall rate of 69%), which is consistent with that of cases with no evidence of disease following treatment (an overall rate of 69%). However, mortality was documented in almost one-third of the pediatric EES population (120 deaths out of 404 cases, an overall rate of 29%). Additionally, recurrence was reported in 35% of cases, while secondary metastasis was reported in 16%. That being said, these rates should be based on the available data of 11 studies out of the 29 studies included in our review. Therefore, the presented data are not generalizable to the whole EES pediatric population.

Meanwhile, our review has several limitations. The most important is the fact that our estimates regarding the prevalence of childhood EES among EES cases (of all ages) or the prevalence of cases of extraosseous origin among the pediatric Ewing sarcoma cases could be overestimated, since the majority of the included studies investigated EES cases and not the Ewing sarcoma population as a whole. In addition, most of these studies are based on retrospective analyses and not cross-sectional in design, which further limits the generalizability of our findings.

#### **5. Conclusions**

Although it is difficult to draw solid conclusions, our results highlight the proportion of children affected by extraosseous Ewing sarcoma, with a special focus on the demographic characteristics, tumor characteristics, and clinical outcomes of the affected patients.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/children9121859/s1, Table S1: The detailed search strategy employed in each electronic database.

**Author Contributions:** Conceptualization: M.G. and A.H.; methodology: M.G. and A.H.; software: M.G., A.H. and M.K.; validation: M.G., A.H. and M.K.; investigation: M.G., A.H. and M.K.; writing—original draft preparation: A.H. and M.G.; writing—review and editing: A.H., M.G., B.L., M.K., J.B., A.G. and T.R.; visualization: M.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this manuscript can be provided by the corresponding author upon reasonable request.

**Conflicts of Interest:** The authors declare no conflict of interest. Regardless, Tobias Renkawitz has received research support and personal fees from Arbeitsgemeinschaft Endoprothetik (AE), DGOU, DGOOC, BVOU, DePuy International, the Otto Bock Foundation, Deutsche Arthrose Hilfe, Aesculap, Zimmer, Stiftung Oskar Helene Heim Berlin, Vielberth Foundation Regensburg, the German Ministry of Education and Research, as well as the German Federal Ministry of Economic Cooperation and Development. Axel Horsch received research support from Arthrose Hilfe and Ipsen.

### **References**


### *Article* **Defining Equinus Foot in Cerebral Palsy**

**Axel Horsch , Lara Petzinger, Maher Ghandour , Cornelia Putz, Tobias Renkawitz and Marco Götze \***

Department of Orthopedics, Heidelberg University Hospital, 69118 Heidelberg, Germany; axel.horsch@med.uni-heidelberg.de (A.H.); lara.petzinger@med.uni-heidelberg.de (L.P.); mghandourmd@gmail.com (M.G.); cornelia.putz@med.uni-heidelberg.de (C.P.); tobias.renkawitz@med.uni-heidelberg.de (T.R.)

**\*** Correspondence: marco.goetze@med.uni-heidelberg.de

**Abstract:** Background: Equinus foot is the deformity most frequently observed in patients with cerebral palsy (CP). While there is widespread agreement on the treatment of equinus foot, a clear clinical definition has been lacking. Therefore, we conducted this study to evaluate functional changes in gait analysis in relation to maximum possible dorsiflexion (0◦ , 5◦ , 10◦ and 15◦ ) and in two subgroups of CP patients (unilateral and bilateral). Methods: In this retrospective study, CP patients with different degrees of clinically measured maximum dorsiflexion were included. We further subdivided patients into unilaterally and bilaterally affected individuals and also included a healthy control group. All participants underwent a 3D gait analysis. Our goal was to determine the degree of maximum clinical dorsiflexion where the functional changes in range of motion (ROM) and ankle moment and power during gait were most evident. Then, a subgroup analysis was performed according to the affected side. Results: In all, 71 and 84 limbs were analyzed in unilaterally and bilaterally affected subgroups. The clinically 0◦ dorsiflexion group barely reached a plantigrade position in the 3D gait analysis. Differences in ROM were observed between subgroups. Ankle moment was quite similar between different subgroups but to a lower extent in the unilateral group. All CP patients had reduced ankle power compared to controls. Conclusions: A cutoff value of clinical ≤5 ◦ dorsiflexion is the recommended value for defining a functionally relevant equinus foot in CP patients.

**Keywords:** cerebral palsy; equinus; dorsiflexion; plantarflexion; definition

#### **1. Introduction**

Equinus foot is the deformity most frequently reported in patients with cerebral palsy (CP) [1]. There are two main types of equinus deformities, and the management approach differs for each type. The first type is known as dynamic equinus, which occurs when the calf muscles become spastic and, in most cases, there is no actual shortening of the gastrosoleus muscle structure yet. Therefore, it is not managed surgically but rather by physical therapy, foot orthosis, casting or botulinum toxin injection i.a. [2]. The second type is referred to as fixed or static equinus with evident contract shortening of the gastrosoleus muscle [3]. In this case, the condition is usually treated surgically by lengthening the affected muscle or its tendon. A wide variety of surgical interventions have been reported in the literature for this type of equinus, and international consensus for treatment has been reached [4,5].

Importantly, proper identification and diagnosis of equinus foot are considered the first step in implementing an appropriate intervention and in reaching a successful outcome [3]. However, there is a clear lack of consensus in the literature on how to properly identify and diagnose equinus foot in CP. A recent survey among 223 orthopedic surgeons highlighted that most of them regularly perform a clinical gait assessment in CP patients they encounter, not specifying if visual or instrumented [6]. However, when it comes to diagnosing equinus foot, their practices differed: 14% rely solely on foot dorsiflexion above plantigrade with an

**Citation:** Horsch, A.; Petzinger, L.; Ghandour, M.; Putz, C.; Renkawitz, T.; Götze, M. Defining Equinus Foot in Cerebral Palsy. *Children* **2022**, *9*, 956. https://doi.org/10.3390/ children9070956

Academic Editor: Reinald Brunner

Received: 30 May 2022 Accepted: 22 June 2022 Published: 25 June 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

extended knee and neutral hindfoot, and 86% use the "Silfverskjöld test" in addition to the previous approach [6].

In the same context, orthopedic surgeons depend mainly on ankle dorsiflexion to assess equinus; however, there is no clear cutoff value of ankle dorsiflexion to define equinus foot. It is of course difficult to determine such a value, as in CP, equinus must never be considered in isolation; this applies especially for bilateral but also for unilateral CP. The sagittal gait pattern should be identified and described, as the ankle and knee levels are linked by coupling (e.g., the plantar flexion knee extension couple), and treatment should therefore be indicated with caution. Rang's aphorism that "a little equinus is better than calcaneus" can usually be agreed on, and individual decisions have to made with each patient [7].

Surprisingly, only a minority of studies report a cutoff value used to define equinus. For example, Horsch et al. [8] defined equinus foot as ≤5 ◦ of clinical ankle dorsiflexion in the extended knee. However, most reports assessing the efficacy of different surgical interventions for equinus foot did not report a clear diagnostic criterion [9–13]. This lack of a standard diagnostic criterion further demonstrates why the actual prevalence rate of equinus foot in CP patients has been called into question [14]. In addition, most surgeons rely mainly on the change in the degree of ankle dorsiflexion (pre- vs. postintervention) in defining successful surgery or improved clinical outcomes in CP patients with equinus foot [15–19].

Based on the aforementioned observations, a standard clinical criterion needs to be established to define the basis for managing equinus foot in CP patients surgically or conservatively. Such clinical criteria would be of great importance in determining how patients are selected for treatment based on their functional impairment, and this might change the way we look at this condition entirely - from the real prevalence of equinus, to the choice of the proper intervention to the reliable definition of a successful outcome. Therefore, we conducted this current research, using a gait analysis, in an attempt to reach a clinically relevant criterion to define CP patients with equinus foot who are eligible for surgery from those who can be managed otherwise. We also aimed to investigate whether this criterion would be applicable to different subtypes of patients (those unilaterally and bilaterally affected).

#### **2. Materials and Methods**

#### *2.1. Study Design and Eligibility Criteria*

For this study, data from 173 patients diagnosed with CP were retrospectively reviewed. The data in this work were collected in our motion lab in the time period from 2002 to 2021. Only patients with gross motor function system (GMFCS) I-II were included [20]. Other inclusion criteria were fully available clinical examinations and an instrumented 3D gait analysis, including measurements of the kinetic and kinematic parameters. We searched for patients with a limited range of motion (ROM) in the left leg measured in the clinical examination. More specifically, we included patients with a 0◦ , 5◦ , 10◦ or 15◦ maximum clinical dorsiflexion in the affected ankle joint. It is common to use 5◦ steps in stating a joint position, as smaller steps are usually not feasible to determine. The ROM was measured in the knee extension. Therefore, we grouped the patients according to their structural shortening of the calf muscles and did not take the additionally underlying amount of spasticity into account. We excluded patients with a GMFCS level III-IV and ankle dorsiflexion >15◦ ; patients with prior soft tissue surgeries (calf, ischios, psoas, etc.) or botox within the past 6 months before gait analysis; patients that had botulinum toxin injections within 6 months prior to gait analysis; and patients that underwent soft tissue lengthening procedures of the lower limbs in general. We also excluded patients with a maximum clinical dorsiflexion of less than 0◦ , as this is unanimously seen as a functionally impairing equinus foot, and we tried to establish a value of dorsiflexion that marks the functional shift from good to impaired gait.

#### *2.2. Study Participants and Measurements*

We divided the subjects in two groups depending on the characteristics of CP. One group contained patients diagnosed with unilateral, and another group with bilateral, CP. The instrumented 3D gait analysis (IGA) and the clinical examination were performed by a specialized study nurse and a physiotherapist with years of neuro-orthopedic experience, and angles were assessed by a single examiner using a standard goniometer and the neutral zero method. Markers were applied on the skin according to a standard protocol (Plug in Gait Model) [21]. Three markers were used on each foot to measure the ankle movements. One of them was placed on the lateral malleolus, another one on the calcaneus and a third one being the toe marker, which is attached to the second metatarsal bone. Gait patterns were captured by 16 cameras from a Vicon System (Oxford Metrics, Oxford, UK). The data were captured with a 120 Hz frequency. A Woltring filter was used. Additionally, three force plates (Kistler Instruments, Winterthur, Switzerland) measured the kinetic parameters of the patients. All participants were asked to walk a distance of 7 m several times at their own walking speed. This examination was performed barefoot and without any walking aids. All data captured in the gait analysis were visualized as different diagrams for every joint in the sagittal, frontal and transverse plane.

#### *2.3. Statistical Analysis*

Study subjects were recruited from a patient cohort of the neuro-orthopedic department of our University Hospital and our established CP patient register. Data were documented in Microsoft Excel. Additionally, we used MatlabR for a graphic representation of the gait analysis data. Normal values were obtained from a group of 50 age-matched healthy patients without any form of gait pathology. The baseline characteristics of participants were presented as numbers and percentages for categorical/dichotomous variables, and as means and standard deviations (SDs) for continuous variables.

#### **3. Results**

#### *3.1. Baseline Characteristics of Included Participants*

Overall, a total of 155 limbs were analyzed: 71 in the unilateral and 84 in the bilateral group. Meanwhile, 5 patients were excluded from the unilateral groups [GMFC score > 2 (*n* = 2), data gaps in clinical examinations (*n* = 2) and ankle dorsiflexion > 15◦ (*n* = 1)] and 13 patients were excluded from the bilateral group [GMFCS > 2 (*n* = 6), data gaps in clinical examinations (*n* = 4) and dorsiflexion > 15◦ (*n* = 3)]. In the unilateral group, 54.9% of the patients were male, 77.5% were GMFCS level I, mean age at the time of IGA was 16.97 (12.13) years, mean weight was 47.16 (21.6) kg and mean height was 151.03 (22.5) cm. Most patients in the bilateral group were also males (56.0%), with GMFCS level II of 72.6% in this group, a mean age of 16 (10.11) years, mean weight of 43.98 (17.12) kg and a mean height of 147.73 (20.64) cm.

#### *3.2. Measured Outcomes in the Bilateral CP Group*

The results are demonstrated in three different graphs for every group, divided into ROM, ankle moment and ankle power. Each graph shows a different gait pattern according to the clinically measured dorsiflexion. The norm data are visualized with a gray band in all diagrams for better comparison.

In the bilateral group, 27 patients had a limited maximum dorsiflexion with 0◦ in the clinical exam and 27 patients could reach 5◦ dorsiflexion. In 22 patients, 10◦ dorsiflexion was measured and in 8 patients 15◦ (Figure 1). The graph shows different curves describing the gait pattern during the gait cycle. The curve representing patients with 0◦ dorsiflexion in the clinical examination did not reach a plantigrade position of the ankle joint in the gait analysis either. These patients walked in plantarflexion without any heel contact during the whole gait cycle. The gaits corresponding to the patients with 5◦ and 10◦ dorsiflexion were similar. In comparison to the 0◦ patients, they could reach a dorsiflexion in the first 15–50% of the gait cycle. Contemplating the plantarflexion in patients with 5◦ , 10◦ and 15◦ after

toe-off (60%), the ROM was approximately comparable to a normal gait. Additionally, the curve for the 15◦ patients followed a nearly normal course with a slight vaulting pattern in the early stance phase. 10° and 15° after toe-off (60%), the ROM was approximately comparable to a normal gait. Additionally, the curve for the 15° patients followed a nearly normal course with a slight vaulting pattern in the early stance phase. Additionally, the curve for the 15° patients followed a nearly normal course with a slight vaulting pattern in the early stance phase.

in the first 15–50% of the gait cycle. Contemplating the plantarflexion in patients with 5°,

in the first 15–50% of the gait cycle. Contemplating the plantarflexion in patients with 5°, 10° and 15° after toe-off (60%), the ROM was approximately comparable to a normal gait.

*Children* **2022**, *9*, x FOR PEER REVIEW 4 of 11

*Children* **2022**, *9*, x FOR PEER REVIEW 4 of 11

**Figure 1.** Range of motion measurement in the bilateral group (*n* = 84). **Figure 1.** Range of motion measurement in the bilateral group (*n* = 84). **Figure 1.** Range of motion measurement in the bilateral group (*n* = 84).

The moments of dorsiflexion and plantarflexion show a typical M-Shape with two maxima in every curve (Figure 2). All the curves, regardless of the dorsiflexion measured, were comparable. The first maxima were between 0 and 20% of the gait cycle, which represented the loading response and early midstance [22]. After that, the plantarflexion moment was pathologically reduced, followed by the physiological maxima in 40–60% of the gait cycle. In comparison with the norm data, the second peaks were lower than the peak seen in the gait of the norm group. The plantarflexion moment of the patients with 0° dorsiflexion in the clinical exam was the lowest. The moments of dorsiflexion and plantarflexion show a typical M-Shape with two maxima in every curve (Figure 2). All the curves, regardless of the dorsiflexion measured, were comparable. The first maxima were between 0 and 20% of the gait cycle, which represented the loading response and early midstance [22]. After that, the plantarflexion moment was pathologically reduced, followed by the physiological maxima in 40–60% of the gait cycle. In comparison with the norm data, the second peaks were lower than the peak seen in the gait of the norm group. The plantarflexion moment of the patients with 0◦ dorsiflexion in the clinical exam was the lowest. The moments of dorsiflexion and plantarflexion show a typical M-Shape with two maxima in every curve (Figure 2). All the curves, regardless of the dorsiflexion measured, were comparable. The first maxima were between 0 and 20% of the gait cycle, which represented the loading response and early midstance [22]. After that, the plantarflexion moment was pathologically reduced, followed by the physiological maxima in 40–60% of the gait cycle. In comparison with the norm data, the second peaks were lower than the peak seen in the gait of the norm group. The plantarflexion moment of the patients with 0° dorsiflexion in the clinical exam was the lowest.

**Figure 2.** Ankle moment measurement in the bilateral group (*n* = 84). **Figure 2. Figure 2.** Ankle moment measurement in the bilateral group ( Ankle moment measurement in the bilateral group (*nn* = 84). = 84).

The generation of power was reduced in all patients in comparison to the norm data (Figure 3). The power generated in the terminal stance to toe-off was lower than usual in every curve, especially in the group consisting of patients with 0◦ dorsiflexion. In the beginning of the gait cycle, all graphs show an increased absorption of power, followed by increased power during midstance (10–30% of gait cycle) and a reduced power generation before toe-off (40–60% of gait cycle). The generation of power was reduced in all patients in comparison to the norm data (Figure 3). The power generated in the terminal stance to toe-off was lower than usual in every curve, especially in the group consisting of patients with 0° dorsiflexion. In the beginning of the gait cycle, all graphs show an increased absorption of power, followed by increased power during midstance (10–30% of gait cycle) and a reduced power generation before toe-off (40–60% of gait cycle).

**Figure 3.** Ankle power measurement in the bilateral group (*n* = 84). **Figure 3.** Ankle power measurement in the bilateral group (*n* = 84).

#### *3.3. Measured Outcomes in the Unilateral Group 3.3. Measured Outcomes in the Unilateral Group*

In the unilateral group, the subjects were separated into 35 patients with a 0° dorsiflexion at the clinical examination and 15 patients with 5°. Ten subjects could reach a 10° dorsiflexion and 11 patients showed a 15° dorsiflexion. In patients with unilateral CP, the ROM was quite similar to that in the bilaterally affected group (Figure 4). The limbs from the 0° group barely attained dorsiflexion in the gait analysis; instead, the ankle remained in plantarflexion during most of the gait cycle. The other curves (orange, blue and black) stay in the area of the norm data until the swing phase, where all the legs showed an In the unilateral group, the subjects were separated into 35 patients with a 0◦ dorsiflexion at the clinical examination and 15 patients with 5◦ . Ten subjects could reach a 10◦ dorsiflexion and 11 patients showed a 15◦ dorsiflexion. In patients with unilateral CP, the ROM was quite similar to that in the bilaterally affected group (Figure 4). The limbs from the 0◦ group barely attained dorsiflexion in the gait analysis; instead, the ankle remained in plantarflexion during most of the gait cycle. The other curves (orange, blue and black) stay in the area of the norm data until the swing phase, where all the legs showed an increased plantarflexion.

increased plantarflexion. Referring to the ankle moment, the plantarflexion moment in the loading response was also higher in every curve, but in comparison to the bilaterally affected patients to a lower extent (Figure 5). Additionally, the maximum of the plantarflexion moment during 40–60% of the gait cycle was decreased in every group. The highest moment of plantarflexion appeared in patients with 5◦ dorsiflexion at the clinical examination.

In patients with unilateral CP, the ankle power was also decreased in all groups during the terminal stance phase to toe-off (Figure 6). Other than in the bilateral group, the absorption-generation pattern of power in the early and midstance was less prominent. Only the 0◦ patients showed slightly increased power generation in midstance.

*Children* **2022**, *9*, x FOR PEER REVIEW 6 of 11

**Figure 4.** Range of motion measurement in the unilateral group (*n* = 71). **Figure 4.** Range of motion measurement in the unilateral group (*n* = 71). flexion appeared in patients with 5° dorsiflexion at the clinical examination.

**Figure 5.** Ankle moment measurement in the unilateral group (*n* = 71). **Figure 5.** Ankle moment measurement in the unilateral group (*n* = 71).

**Figure 5.** Ankle moment measurement in the unilateral group (*n* = 71).

Only the 0° patients showed slightly increased power generation in midstance.

In patients with unilateral CP, the ankle power was also decreased in all groups during the terminal stance phase to toe-off (Figure 6). Other than in the bilateral group, the

In patients with unilateral CP, the ankle power was also decreased in all groups during the terminal stance phase to toe-off (Figure 6). Other than in the bilateral group, the absorption-generation pattern of power in the early and midstance was less prominent.

**Figure 6.** Ankle power measurement in the unilateral group (*n* = 71). **Figure 6.** Ankle power measurement in the unilateral group (*n* = 71).

#### **4. Discussion**

**4. Discussion**  Normally, during the stance phase (0–60% of the gait cycle), the greatest degree of dorsiflexion is needed just before lifting the heel in a fully extended knee, in which case the ankle has to be in a dorsiflexed, but perpendicular position for smooth ambulation [23,24]. Yet, there is still controversy related to the proper degree of dorsiflexion that is truly required for this to happen. Consequently, it would seem reasonable to use a normal range of values for defining the physiological gait, rather than to apply a definitive value. Normally, during the stance phase (0–60% of the gait cycle), the greatest degree of dorsiflexion is needed just before lifting the heel in a fully extended knee, in which case the ankle has to be in a dorsiflexed, but perpendicular position for smooth ambulation [23,24]. Yet, there is still controversy related to the proper degree of dorsiflexion that is truly required for this to happen. Consequently, it would seem reasonable to use a normal range of values for defining the physiological gait, rather than to apply a definitive value. Based on the literature, this accepted range of normal ankle dorsiflexion lies between 3◦ and 15◦ beyond the perpendicular plane when the knee is fully extended [23–25].

Based on the literature, this accepted range of normal ankle dorsiflexion lies between 3° and 15° beyond the perpendicular plane when the knee is fully extended [23–25]. Equinus foot is described as decreased ankle joint dorsiflexion; although, there is a lack of consensus on the exact definition and diagnostic criteria. While studies widely agree that static ankle joint equinus represents a reduced range of dorsiflexion at the ankle joint, there is no agreement concerning the degree of dorsiflexion reduction needed for this condition to manifest. Of course, different amounts of spasticity can lead to functionally different patterns. In the present work, we examined the differences in ROM, ankle moment and ankle power in different groups of patients based on the degree of maximum clinical ankle dorsiflexion. Here, among patients in the 0° dorsiflexion group, the sub-Equinus foot is described as decreased ankle joint dorsiflexion; although, there is a lack of consensus on the exact definition and diagnostic criteria. While studies widely agree that static ankle joint equinus represents a reduced range of dorsiflexion at the ankle joint, there is no agreement concerning the degree of dorsiflexion reduction needed for this condition to manifest. Of course, different amounts of spasticity can lead to functionally different patterns. In the present work, we examined the differences in ROM, ankle moment and ankle power in different groups of patients based on the degree of maximum clinical ankle dorsiflexion. Here, among patients in the 0◦ dorsiflexion group, the subgroup of patients with bilateral CP could not reach plantigrade, and the unilateral subgroup barely reached plantigrade in midstance (40%). Thus, we hypothesize that this value would not be the best cutoff value to define a functionally relevant equinus foot in children with CP.

group of patients with bilateral CP could not reach plantigrade, and the unilateral subgroup barely reached plantigrade in midstance (40%). Thus, we hypothesize that this value would not be the best cutoff value to define a functionally relevant equinus foot in children with CP. Owing to this lack of consensus, physicians have used a wide variety of restrictions on dorsiflexion for diagnosis [26]. Sobel et al. [27] proposed that patients have less than 0° of dorsiflexion in order to be diagnosed with equinus (i.e., no step beyond plantigrade), while Orendurff et al. [28] recommended a cutoff value of 5°. On the other hand, DiGiovanni et al. [29] proposed a minimum value of less than 10° of dorsiflexion, which is consistent with the need for at least 10° of dorsiflexion to maintain a normal gait and avoid Owing to this lack of consensus, physicians have used a wide variety of restrictions on dorsiflexion for diagnosis [26]. Sobel et al. [27] proposed that patients have less than 0◦ of dorsiflexion in order to be diagnosed with equinus (i.e., no step beyond plantigrade), while Orendurff et al. [28] recommended a cutoff value of 5◦ . On the other hand, Di-Giovanni et al. [29] proposed a minimum value of less than 10◦ of dorsiflexion, which is consistent with the need for at least 10◦ of dorsiflexion to maintain a normal gait and avoid possibly increased forefoot loading throughout locomotion [28,30]. The aforementioned recommendations are consistent with Meyer's more recent proposal that, instead of basing an equinus diagnosis on a specific range of dorsiflexion motion, a diagnosis should be verified when the reduction in dorsiflexion reaches a magnitude that increases stress on the Achilles tendon and loading on the forefoot [31]. While it would be expected that forefoot

possibly increased forefoot loading throughout locomotion [28,30]. The aforementioned recommendations are consistent with Meyer's more recent proposal that, instead of bas-

verified when the reduction in dorsiflexion reaches a magnitude that increases stress on

pressure rises during locomotion in order to base an equinus diagnosis on a limit of 10◦ of dorsiflexion, there is no evidence that this would contribute to the development of defects in the foot or lower leg.

Our results indicate that ankle ROM in children with CP in the unilateral subgroup with 5◦ , 10◦ and 15◦ dorsiflexion was functionally quite similar to that in healthy controls from the beginning of the gait cycle and up to the mid-swing phase (0–75%), typically with more plantarflexion during the end of the cycle as compared to controls. On the other hand, ROM in the bilateral subgroup with 15◦ dorsiflexion was also similar to that in controls; however, both the 5◦ and 10◦ dorsiflexion groups showed reduced dorsiflexion compared to controls during the period from midstance to toe-off (40–60%).

Notably, the ankle joint moment was quite similar in both the uni- and bilateral CP subgroups during the gait cycle; however, the extent of plantarflexion in the bilateral group was higher. In all dorsiflexion groups (0◦ , 5◦ , 10◦ and 15◦ ), the ankle moment throughout the gait cycle was quite comparable. When compared to healthy controls, all of the dorsiflexion groups revealed a higher plantarflexion moment during the loading response that lowered during midstance and never reached a physiological peak in terminal stance. This was consistent in both uni- and bilateral CP groups. This clearly shows that during gait, function as seen in kinetics is often worse than clinical examination and kinematics would suggest. This might be due to differences in the amount of spasticity of the calf muscles and weakness of the foot levers.

So far, equinus deformity itself has not been considered separately for uni- or bilateral involvement in most of the studies published; although, studies often present one group or the other. In our experience, the two groups present differently, with the amount of foot lever weakness being greater in unilaterally affected patients. Recent studies suggest that a functionally separate approach is beneficial, which is also in line with our experience and this assessment [32]. In terms of ankle power, both groups (uni- and bilateral) in our study had reduced power when compared to healthy controls. Similarly, the gait analysis of both groups revealed that reduced ankle power was evident in the period from terminal stance to toe-off (40–60%) as compared to controls. However, both groups had increased power during the loading phase in early stance as compared to healthy controls. The only difference between the two groups was found in the midstance (10–30%): the bilateral group revealed higher power than controls; while in the unilateral group, power was similar to that of the healthy or typically developing (TD) controls.

In addition, although a cutoff point of 10◦ may increase forefoot loading during locomotion, Orendurff et al. [28] proposed that a dorsiflexion ≤5 ◦ should be used for the diagnosis of equinus foot as they found that forefoot pressure was higher in patients with ≤5 ◦ dorsiflexion than in patients with more than 5◦ dorsiflexion (*p* < 0.05). This is consistent with our findings, as we noted that the lowest extent of dorsiflexion was observed in the ≤5 ◦ dorsiflexion group as compared to all other groups (0◦ , 10◦ , 15◦ and healthy controls), especially at the loading and midstance phases of the gait cycle where the ankle is at a perpendicular plane and the knee is normally fully extended. In the same context, ankle moment and power in the ≤5 ◦ dorsiflexion group were similar to that in other dorsiflexion groups (0◦ , 10◦ and 15◦ ) during the whole gait cycle.

Therefore, we recommend taking ≤5 ◦ dorsiflexion value as a cutoff point for defining equinus foot deformity at the early phases of gait (loading and midstance) when the knee is normally fully extended and the ankle is in a perpendicular position, especially in bilaterally affected CP patients.

Our study provides a clinically relevant criterion for distinguishing CP patients with impaired function due to equinus from those in whom function is not impaired. This criterion is applicable and easy to use in a clinical setting to identify patients who are in need of treatment and those who possibly are not. That being said, this is a simplistic approach to define a functional relevant equinus, and sagittal gait patterns should always be considered before treatment suggestions are made. Surgeons have to be careful not to contribute to, e.g., crouch gait and impairment of global gait function by trying to improve ankle kinematics solely based on the presence of functionally relevant equinus foot.

The main aim of this study was to estimate a clinically relevant criterion to define CP patients with equinus foot, by reaching a cutoff point of dorsiflexion that can help discriminate CP patients with functional impairment who might be in need of surgical management. In the case of instability in the talonavicular joint, the joint was held during the clinical examination, and then the extent of movement was examined. Since it is not possible to avoid the occurrence of a midfoot break during a gait analysis, this should be considered a clear limitation. However, the purpose of this study was to find a clinical definition for equinus foot with the help of a gait analysis. Our patients were first clinically examined, which ensured that no midfoot break was present. It would also be worth discussing the integration of a foot model in future studies in order to investigate this question more closely.

Our study has several limitations. The most important limitation is the small number of participants within each subgroup (0◦ , 5◦ , 10◦ and 15◦ ) included in our analysis, which further limits the implications drawn from our study and renders it difficult to apply our findings more generally. Additionally, the retrospective nature of our study limits the interpretation of our results. Moreover, we describe the differences observed between the groups studied clinically; thus, more work is required to determine whether these minimal clinical differences also reach statistical significance. Therefore, and based on the aforementioned limitations, future research in this area is still warranted by larger studies with more diverse populations and consideration of sagittal gait patterns to be able to determine whether such definition criteria are applicable for all CP patients, regardless of their clinical and demographic characteristics.

#### **5. Conclusions**

In conclusion, a cutoff value ≤5 ◦ maximum clinical dorsiflexion should be used to distinguish between patients who might need treatment from those who do not among CP patients with equinus foot. However, the applicability of this criterion among patients who are unilaterally and bilaterally affected is slightly different; therefore, we recommend considering the laterality of CP when diagnosing patients with equinus foot and when choosing the appropriate management approach for these individuals. This simplistic approach should not leave the sagittal gait pattern unconsidered, as this might lead to impairment of the global gait function.

**Author Contributions:** Conceptualization: A.H.; methodology: A.H.; software: A.H. and L.P.; validation: A.H. and L.P.; formal analysis: A.H., L.P. and M.G. (Marco Götze); investigation: A.H. and M.G. (Maher Ghandour), L.P. and M.G (Marco Götze); resources: M.G. (Marco Götze) and A.H.; data curation: L.P. and A.H.; writing—original draft preparation: A.H., L.P. and M.G. (Maher Ghandour); writing—review and editing: A.H., C.P., M.G. (Marco Götze) and T.R.; visualization: L.P.; supervision: A.H.; project administration: A.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. For the publication fee we acknowledge financial support by Deutsche Forschungsgemeinschaft within the funding programme "Open Access Publikationskosten" as well as by Heidelberg University.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki, and approved on 25 September 2018 by the Ethics Committee of Heidelberg University (protocol registration number: S-576/2018) for studies involving humans.

**Informed Consent Statement:** Patient consent was waived due to minimal risk.

**Data Availability Statement:** All the Data analyzed in this manuscript can be provided upon request by contacting the corresponding author.

**Conflicts of Interest:** None declared. There was no external funding for the study. Regardless, Tobias Renkawitz has received research support and personal fees from Arbeitsgemeinschaft Endoprothetik (AE), DGOU, DGOOC; BVOU, DePuy International, Otto Bock Foundation, Deutsche Arthrose Hilfe, Aesculap, Zimmer, Stiftung Oskar Helene Heim Berlin, Vielberth Foundation Regensburg, the German Ministry of Education and Research as well as the German Federal Ministry of Economic Cooperation and Development. Axel Horsch received research support from Arthrose Hilfe and Ipsen.

### **References**

