Ultrasound findings in lateral elbow tendinopathy: A retrospective analysis of radiological tendon features [version 1; peer review: awaiting peer review]

Background: Lateral elbow tendinopathy (LET) is prevalent in the upper extremity, with various therapeutic options. Understanding the types and the relations between the radiological tendon features would help to develop more specific treatments. This study reviewed ultrasound exams of LET to investigate the types of degenerative findings and the relationships between them in one of the most prominent sports medicine clinics in Latin America. Methods: A retrospective study was performed. We evaluated 4335 ultrasonographic exams with LET from 2017 and 2018. Five principal degenerative ultrasound criteria with subtypes were selected: hypoechogenicity, neovascularity, calcification, enthesopathy, and intrasubstance tear. A multiple linear regression model was conducted to explore the association between the findings, sex, and age. Results: Overall, 4324 ultrasound exams were analyzed; 2607 (60.29%) were males. Multiple degenerative tendon findings were found in adults (≥18 years) with LET. Hypoechogenicity (67.77%) and neovascularity (37.8%) were the most frequent. The mean length of a tendon tear in both sexes was 4.44 (± 2.81) millimeters. Mild hypoechogenicity (P < .001), and depth intrasubstance tear (P < .01) were statistically significant between them. Severe hypoechogenicity was associated with an increase in all tendon tear dimensions for length 1.37 ([95% Confidence interval (CI), 0.57, 2.17]; P < .001), for Open Peer Review Reviewer Status AWAITING PEER REVIEW Any reports and responses or comments on the article can be found at the end of the article. Page 1 of 12 F1000Research 2022, 11:44 Last updated: 21 JAN 2022


Introduction
Tendinopathies are a crucial global burden of musculoskeletal disorders. 1 Specifically, lateral elbow tendinopathy (LET), or tennis elbow, 2 is one of the most prevalent degenerative conditions in the upper extremity. 3 It has been shown that LET affects both sexes, between 35 to 54 years 4 with severe pain, 5 and functional disability in wrist extension and handgrip activities. 6,7 The LET symptoms can persist for more than 12 months, 8 directly impacting a patient's quality of life, 9 sports activities, 10 occupational settings 11 and the health system. 12 The LET pathogeneses are similar to other tendinopathies. These includes collagen disorganization, neurovascular ingrowth, 13 tissue necrosis with myxoid and hyaline degeneration and fibrosis. 14 Most of the degenerative process is concentrated on the extensor carpi radialis brevis (ECRB) tendon. 15,16 These histological changes were first described in 1979. 17 In 2009, two researchers proposed an alternative model to describe continuous tendon changes based on three states of its structure: reactive, disrepair and degenerative. 18 Additionally, in 2016, another investigation included a final phase with gross structural disruption and tendon tear. 19 These structural tendon changes can be identified with imaging methods, 20 ultrasound (US) findings in the degenerative stage has been well documented 21 and can be a useful tool to detect signs such as hypoechoic areas, neovascularization, calcifications, enthesopathy, and intrasubstance tears 22 with better sensitivity and accuracy than magnetic resonance imaging (MRI). 23,24 In this context, there is no consensus in the literature about the main ultrasound findings in LET. For example, a study evaluated 240 patients with this condition with the highlighted presence of calcifications, 25 but another article identified hypoechoic areas, partial and total tears principally. 26 Currently, articles have been published with standardized graduations of US findings, allowing to subclassify the magnitude of the tendon's structural changes. 21,27 For instance, hypoechogenicity and hyperemia have been divided in severity values of 0% to 30%, between 30 to 50% and more than 50%. 28 Intrasubstance tears of the tendon have also been added. 29 However, the dimensions and magnitudes have not yet been specified. 22 Still, most of the articles have reported individual degenerative tendon characteristics without studying the relationship and occurrence of more than one degenerative characteristic simultaneously. And this is vital because abnormalities in the tendon structure are considered a risk factor for developing symptoms of LET. 30 For example, an author suggested vascular in-growth into tendons may be the cause of tendon weakness and tears. 31 Another article confirmed that 97% of tendon tears have degenerative changes. 32 More recently, a meta-analysis indicated that there was a higher relative risk for developing clinical tendinopathy when multiple US findings were considered. 33 We know that the tendon tear is the final phase of the degenerative process in this tissue and the number of US findings could determine the severity of the tendinopathy. However, the frequency of degenerative findings, the size of tendon tears, and the association between them in LET are poorly known. Indeed, to date, no studies have investigated the types of degenerative findings and the relationships between them. Hence the main aim of this study was to identify the frequency and evaluate associations between standardized degenerative tendon findings with tendon tears in three sizes dimensions (length, width, and depth). This is a unique study with one of the most extensive collections of ultrasound exams analyzed to the best of the authors' knowledge.

Ethics statement
This study has been performed in keeping with the latest version of the Declaration of Helsinki, in accordance with Chilean legislation. The study was approved by the Comité de Ética Científico Adulto del Servicio Metropolitano Oriente de la ciudad de Santiago de Chile (SSMO). The ethical committee required no informed consent from patients given the retrospective nature of the study. The project was approved on August 17, 2018. No approval number was recorded.

Study design
This research was designed as a descriptive, retrospective, multicentric study and was written in accordance to the Strengthening the Reporting of Observation studies in Epidemiology (STROBE) guideline. 34 This study started on January 14, 2019. All patients with documentation of elbow ultrasound at MEDS Clinic in Santiago, Región Metropolitana, Chile, from January 1, 2017, to December 31, 2018, were selected for the study. The included patients had LET suspicion with any unilateral or bilateral signs or symptoms in this period. If the patient had multiple episodes of LET at different times, each episode was included. Our Medical Imaging Software (Carestream RIS, v.11) allows a word search of the whole electronic ultrasound exams record database between 2017 to 2018. Authors GD, FF, and MT had access to the database. Any structural tendon change of all the patients with exams of any search containing the word LET were included. The exams were diagnosed by five musculoskeletal radiologists with more than ten years of experience. Of these, only JR was an author of this article. The professionals used an Aplio 500 US system (Toshiba America Medical Systems, Inc, Tustin, CA) equipped with a multifrequency linear transducer with a frequency of 18 MHz to inform the LET condition. Demographics such as sex, age, and size of the injury were recorded. The exams with a history of musculoskeletal ultrasound guided treatment, such as injections, or any surgery, were excluded from the analysis. Physical activity, pain score, and time of symptoms were not reported.

Degenerative ultrasound findings and classification
We proposed a standardized categorization to assess the presence of five principal US criteria with subdivisions based on previous reports in the literature. 16,21,26,35,36 A normal tendon was defined as no changes in the tendon structure. Hypoechogenicity was defined as being rounded and not associated with tendon disruption and was evaluated depending on the tendon extension structure, with an ordinal score from 0 to 3, with 0 representing a normal tendon echogenicity, mild level (HE1) is 1% to 30%, moderate level (HE2) is 31% to 50% and severe level (HE3) is over 50%. Neovascularity was assessed with the maximum number of color pixels occupying the tendon origin, an ordinal score from 0 to 2, where 0 represents no neovascularity presence, 1 represents focal neovascularity (NV1) with 30% blood vessels, and 2 refers to multiple neovascularity (NV2) consisting of more than 31% blood vessels. A linear intrasubstance tear was defined as a linear hypoechoic focus associated with discontinuity of tendon fibers. The number of tendon tears was recorded as a focal intrasubstance tear (IST1) or multiple intrasubstance tear (IST2). The size of the abnormality was measured in millimeters. The site of the abnormality was examined in two planes and assigned as predominantly affecting the superficial, mid, or deep fibers and the anterior, mid, or posterior fibers. Calcification (C) was defined as calcium deposits in the substance of the tendon and enthesopathy (E) was defined as the pathologies that affect the entheses. Both findings were categorized as a binary score (absent and present).

Statistical analysis
Data was reported with frequencies, means, and standard deviations (SD), when corresponding. The Chi-square test was used to compare categorical data and T-test for continuous variables. Univariate comparisons between the tear's size in each dimension and degenerative findings were performed using the Kruskal Wallis test. A separate multiple linear regression model was fit to analyze the association between sex, age, and degenerative findings with tendon tear in each dimension (length, width, and depth).
Neither final exposure variables nor outcome data presented missing values. Missing data analysis showed that missingness were at random; multivariate regression imputation was employed as required. P value < .05 was considered a statistically significant, two-tailed statistics were used. All graphs and analyses were performed using R (The R Foundation for Statistical Computing, v. 3.6.2) and (RStudio, v.4.1.0).

Results
A total of 4335 US exams from 2926 patients with suspicion of LET were included in our analysis. Of these, 11 exams were excluded as they did not present any final report. Summary data from ultrasonographic exams are available in Underlying data. 45 The age of patients with LET presented a minimum value of 18 years and a maximum of 91 years. The data shows that women have a mean age 47.18 AE11.00 (P < .001), and are on average, one year older than males. In 70% of the patients, the affected side was the right elbow.
The HE1 was the most frequent structural change detected on US (53.07%), followed by enthesopathy (35.01%), focal neovascularity (33.55%), and focal intrasubstance tear (31.91%). However, when the results considered the subtypes classification, the frequency distribution changed. Frequency (number of cases diagnosed) for most of the degenerative findings were very similar between females and males. More than a third of exams with LET showed an intrasubstance tear in the tendon, see Table 1.
Additionally, men with focal and multiple intrasubstance tears, in all three dimensions (length, width, depth), presented a larger tear size than women. Overall, the variables age (P < .001), HE1 (P < .001), and depth IST1 (P < .01) were statistically significant between females and males. A synthesis is provided in Table 1.

Relationship between tendon tear and degenerative findings
The category of hypoechogenicity depends on the size of the intrasubstance tear. The length compared with width and depth dimensions presented the most significant variability among hypoechogenicity levels. Also, an increase in the median length of the tear occurs when the tendon presented higher levels of hypoechogenicity (HE3).
Particularly, Figure 1 shows that the length of the tear for HE3 cases, when compared with HE2, is significantly different (P = .035). Nevertheless, when the width of the tendon tear and different levels of hypoechogenicity were considered, the biggest difference (P < .001) occurred between HE1 to HE2 (not shown in the figure).
Similarly, in Figure 2, we show the relationship for the width dimension of tendon tear with the presence of two neovascularity levels (NV1 and NV2). Again, higher values of width tear dimensions were associated with the presence of the greater neovascularity level (P < .005). However, only NV0 to NV1 transitions demonstrated a significant difference (P < .004). Something similar happened with the depth of tear. This case did not show statistically significant differences among neovascularity levels. This may be due to the low frequency of cases in the deep dimension (not shown in figure).
Finally, Figure 3 shows that the depth dimension presented the highest statistical difference (P = .007) between HE0 and HE1, with no essential differences among hypoechogenicity HE1 and HE2 levels. Also, HE2 compared to HE3 shows statistical significance (P = .025). However, the number of cases with depth tear and HE3 is deficient, hence it might not be representative.
Subsequently, we examined if the three dimensions of tendon tear presented statistical differences with the presence of other degenerative findings. We performed linear regression models for each tear dimension as shown in Table 2. Hypoechogenicity had an effect in all tendon tear dimensions with an average increase from H0 to H3 of 1.37 Finally, it was evaluated whether interactions between the exposure variables could improve the models in any of the three proposed dimensions of tendon tear. However, none of the interactions among the different types of degenerative findings presented statistical significance.

Discussion
The presence of tendinopathies is a frequent problem for the general population and athletes. Notably, LET is the most commonly diagnosed elbow musculoskeletal disorder and still represents an unresolved problem for public health. 37 Multiple studies have discussed the structural changes that can be found exclusively in LET such as tissue thickening, hypoechoic areas, neovascularization areas, fibrillary disruption, and bone abnormalities. 21,26,34 As demonstrated in this research, it is infrequent that the degenerative findings are present in isolation and unrelated. Our study allows us to determine the presence of four large groups of degenerative findings and the presence of gradual lesions, which can determine the presence of tear within the tendon.
In a large sample of ultrasounds exams obtained over two years, our analysis indicates that the age and tendon degenerative findings of patients were the most critical factors associated with increased tendon tear in any dimension. Particularly, we found a positive relationship between the presence of hypoechogenicity levels and the length of tear and neovascularity levels with the tear width. This result suggests that if the goal is to reduce the magnitude and the variability of tendon tear in patients with LET, hypoechogenicity, and neovascularity should influence the importance of degenerative findings, so preventive ultrasound exams could be useful to avoid potential tendon tears. 38 Furthermore, this information can be used to assess lateral elbow tendinopathy severity. For example, investigators concluded that the size of intrasubstance tear on an ultrasound could determine which patients respond to nonoperative treatment and the correlation with functional disability. 39 However, the tendon tear description did not include the magnitude of the tear dimension, HE, and NV findings. Therefore, it is essential to have this information in future studies.
On the other hand, there are several possible explanations for the positive association of hypoechogenicity and neovascularity with the dimension of the tendon tear. Firstly, the ECRB is the central tendon involved in a degenerative process of LET, and in this study, we only considered the exhaustive descriptions and analysis for degenerative ultrasound findings in the interior of this tendon. However, most of the existing publications describe the relation between LET and different anatomical aspects 40 or other adjacent tissues such as nerve, ligaments, muscles, bone structures or more tendons 41 and even other health conditions that are not unique to the alteration of the origin of the wrist extensors. Secondly, researchers have proposed that the tendon pathogenesis model is continuous, arguing that there is a 3-state transition from an initial stage characterized by reactive tendinopathy, a second disruptive stage, and finally, a third degenerative stage. 42 We believe that this last stage also presents graduation, particularly with some degenerative findings because they can be subdivided according to their degree of commitment. Furthermore, the existence of the demonstrated associations could support future predictive models that determine whether the existence of one finding anticipates the appearance of another.
Recently, a study showed that LET primarily affects middle-aged people 43 and that the probability of having LET decreases with older age. 12 Our study shows that age is strongly associated with the tendon tear's size, so the older the patient, the higher the magnitude of the tear independent of the dimension evaluated. The results confirm that men have LET almost a year earlier than women. However, the clinical importance lies in identifying the tendons' structural characteristics using ultrasound.
Some authors even propose that US examinations be performed even in subjects who do not present symptoms but could present degenerative findings. 27 As we could observe in our study, the degenerative findings could present graduations before the tendon tear, so earlier identification of these findings would allow monitoring future tendon tears and could improve the therapeutic resolution capacity. It is important to have future publications that can assign a time factor to degenerative findings to understand the natural evolution of such tendon tear.

Limitations
The study's principal limitation was the retrospective design, which is often considered inferior to prospective, randomized, and controlled clinical trials. 44 Additionally, we couldn't manage the imaging procedure and settings, and inter and intra rater reliability was not considered, losing external validity. Further limitations include assessment only of ECRB tendon, hindering results generalization. However, this is the study with the highest number of exams  Statistical significance between the groups, P < .05.
analyzed to the best of the authors' knowledge. An additional limitation was that the images correspond to patients with different conservative treatments, pain scores, evolution times, physical activity levels, comorbidities, and clinical manifestations. This study focused solely on the characteristics of ultrasound exams, not incorporating these elements. Future studies should consider a cohort design to assess these considerations.

Conclusions
In summary, this study provides evidence of a positive association between the presence of hypoechogenicity and neovascularity findings with the increased size of a tendon tear in patients with LET. Additionally, we reaffirm the higher frequency of tendon tears between 40 to 60 years of age but a greater magnitude of the injury from 60 years of age. This article demonstrates the importance of degenerative findings in the magnitude of tendon injury.

Data availability
Underlying data The ultrasonographic exams are not available publicly because consent for publication of raw data was not obtained, and the dataset could pose a threat to confidentiality. Researchers and reviewers interested in accessing the data will need to submit an official request letter to the Academic Unit, MEDS Clinic. They must also confirm that they will not violate the ethical standards of the ethical committee and protect the anonymity of the participants. Interested parties can contact the corresponding author, Guillermo Droppelmann (guillermo.droppelmann@meds.cl) who can facilitate this process.