Factors associated with long head of the biceps tendon tear severity and predictive insights for grade II tears in rotator cuff surgery

Article information

J Korean Shoulder Elbow Soc. 2024;.cise.2023.01053

Publication date (electronic) : 2024 April 22

doi : https://doi.org/10.5397/cise.2023.01053

Dong-Hyun Lee¹, Gyu-Min Lee¹, Hyung Bin Park^,1,2,3

¹Department of Orthopedic Surgery, Gyeongsang National University Changwon Hospital, Changwon, Korea

²Department of Orthopedic Surgery, College of Medicine, Gyeongsang National University, Jinju, Korea

³Gyeongsang Institute of Health Sciences, Gyeongsang National University, Jinju, Korea

Correspondence: Hyung Bin Park Department of Orthopedic Surgery, Gyeongsang National University Changwon Hospital, 11 Samjeongja-ro, Seongsan-gu, Changwon 51472, Korea Tel: +82-55-214-3744 Fax: +82-55-214-3259 E-mail: hbinpark@gnu.ac.kr

Received 2023 April 22; Revised 2024 February 13; Accepted 2024 March 5.

Abstract

INTRODUCTION

The long head of the biceps tendon (LHB) has been reported as a mechanical stabilizer of the glenohumeral joint, though that is still a debatable issue [1,2]. LHB tears are a common source of shoulder pain [3,4] and have been considered part of a degenerative aging process, similar to a rotator cuff tear (RCT) [5]. On the other hand, the LHB is used in various shoulder repair and reconstruction surgeries. For large to massive RCT management, LHB rerouting [6], anterior cable reconstruction [7], biceps augmentation using the tenotomized biceps [6,8], and the snake technique (similar to superior capsular reconstruction) [9] have been reported. For recurrent shoulder instability, transfer of the LHB has been reported as a viable option in patients with subcritical glenoid bone loss [10]. Several factors have been proposed as risk factors for LHB tears: age, a posterosuperior rotator cuff tear (PSRCT) and its size [11,12], a superior labrum anterior-posterior lesion [13], a subscapularis tendon (SSC) tear [14,15], glucocorticoid use [16], diabetes [17], and serum high-sensitivity C-reactive protein (hs-CRP) >1 mg/L [18].

Serum hs-CRP is a marker of inflammation and is widely accepted as a potential risk predictor for several chronic inflammatory diseases, including atherosclerosis and cardiovascular disease [19]. Several studies have reported that an elevated hs-CRP level is associated with chronic orthopedic disorders, including hip and knee osteoarthritis [20], frozen shoulder [21], chronic lower back pain [22], and sciatic pain [22]. A previous study reported that serum hs-CRP >1 mg/L is independently associated with LHB tears [18]. Lafosse et al. [23] classified LHB tears according to the tear severity in arthroscopic findings: grade 0 (without gross injury), grade I (minor lesion with fraying or erosion involving <50% of the tendon diameter), and grade II (major lesion with fraying or erosion involving ≥50% of the tendon diameter, including complete tears). To use the LHB for augmentation or as a graft material during shoulder surgery, the status of an LHB tear is recommended to be ≤grade I [24]. However, no readily accessible relevant studies have evaluated factors associated with LHB tear severity or predictive of grade II LHB tears. This study tests the hypothesis that hs-CRP is significantly associated with LHB tear severity and predictive of grade II LHB tears. In addition, this study aimed to determine any factors associated with LHB tear severity and any factors that might predict the presence of a grade II LHB tear before rotator cuff surgery.

METHODS

This study, which was approved by the Institutional Review Board of Gyeongsang National University Changwon Hospital (No. 2022-01-030), retrospectively reviewed the data of 795 patients who underwent arthroscopic shoulder surgery performed by one shoulder and elbow surgeon (HBP) from January 2010 to February 2021. Requirement for informed consent was waived by the Institutional Review Board because of the study retrospective design. Some of those patients were then excluded from the analyses for the following reasons: a history of infection (n=8), greater tuberosity fractures (n=7), revision rotator cuff surgery (n=9), arthroscopy after arthroplasty or fracture fixation (n=5), or an acute traumatic RCT (n=16). Therefore, 750 patients were enrolled in this study (Fig. 1). The mean age of the patients at the time of arthroscopic surgery was 60.7±12.3 years; 361 patients (48.1%) were male, and 389 patients (51.9%) were female. All patients underwent arthroscopic surgery in the lateral decubitus position under general anesthesia.

Fig. 1.

Flowchart showing the inclusion and exclusion criteria for this study. A total of 750 patients were enrolled in this study in keeping with those criteria.

The demographic, physical, social, and serologic parameters; medical comorbidities; intrinsic shoulder lesion; factors related to RCT; and pain severity were evaluated. The demographic variables were age and sex. Patients were grouped by age: <40 years, 40–49 years, 50–59 years, 60–69 years, and ≥70 years. The physical variables were body mass index (BMI) and involvement of the dominant side. BMI was categorized as underweight (<18.5 kg/m²), normal weight (18.5–22.9 kg/m²), overweight (23.0–24.9 kg/m²), and obese (≥25.0 kg/m²). The social variable was smoking. The serologic parameter was the serum hs-CRP level, and it was categorized as ≤1 mg/L (low level) and >1 mg/L (high level). The evaluated medical comorbidities were diabetes, hypertension, hyperthyroidism, hypothyroidism, and dyslipidemia. The studied intrinsic shoulder lesion was RCT, which included both PSRCTs and SSC tears regardless of tear thickness. PSRCTs were further classified as partial- and full-thickness tears. The RCT-related variables were the size, Patte retraction degree [25], and tangent sign [26]. The size was classified as intact and full-thickness tear, with a full-thickness tear further classified as small (<1 cm), medium (1–3 cm), large (3–5 cm), or massive (>5 cm) based on arthroscopic findings [27]. The Patte retraction degree of an RCT was categorized as stage 1 (proximal tendon stump near bony insertion), stage 2 (proximal stump at humeral head level), or stage 3 (proximal stump at glenoid level) [25]. The tangent sign was assessed on T1-weighted oblique-sagittal images. It was graded negative if the supraspinatus muscle crossed a tangent line drawn between the superior margins of the scapular spine and coracoid process and positive if it did not cross the line [26]. Pain severity was evaluated one day before surgery using a visual analog scale (VAS). Pain reported on the VAS was categorized as ≤3, 4 to 7, and ≥8. Table 1 summarizes the prevalence and medians of all studied variables.

Table 1.

The studied variables and their medians or prevalences

LHB tear severity was determined as grade 0, grade I, and grade II using the Lafosse classification described above (Fig. 2) [23]. The distributions of continuous variables were determined using the Kolmogorov-Smirnov test. When variables were not distributed normally, the median and interquartile range are used to present them.

Fig. 2.

Arthroscopic findings about the integrity of the long head of the biceps tendon. (A) Grade 0, without gross injury. (B) Grade I, minor lesion with fraying or erosion involving <50% of the tendon diameter. (C) Grade II, major lesion with fraying or erosion involving ≥50% of the tendon diameter, including a complete tear.

Ordinal logistic regression analyses, reporting odds ratios (ORs) and 95% confidence intervals (CIs), were performed to determine the strengths of associations between the studied variables and LHB tear severity in both the overall cohort and the large-to-massive RCT cohort. Univariable ordinal logistic regression analyses were performed for all variables, and then multivariable ordinal logistic regression analyses were performed using the variables that were significant in the univariate analyses. The goodness of fit for the multivariable ordinary logistic regression model was assessed using the parallel lines test and McFadden R². A significance level of P>0.05 is the standard for the parallel lines test, and McFadden R² values from 0.2 to 0.4 indicate that the model has a very good fit [28].

Binary logistic regression analyses were conducted to assess the associations between the variables and grade II LHB tears. ORs with 95% CIs were calculated for both the overall cohort and the large-to-massive RCT cohort. Univariate analyses were initially conducted for all variables, followed by multivariable analyses using variables significant in the univariate analyses while considering multicollinearity. Multicollinearity was assessed using Spearman correlations, the variance inflation factor, and condition index. A Spearman correlation coefficient of ≥0.7 between variables indicates multicollinearity [29]. The multivariable binary logistic analysis was conducted with variables included separately. Multicollinearity was considered absent when the variance inflation factor and condition index were <10 [30]. The Hosmer-Lemeshow test was used to assess the goodness of fit for the multivariable binary logistic regression model, with significance set at a P-value >0.05. ORs were calculated to determine the effect size according to Cohen’s d scale: small (0.20–0.49), medium (0.50–0.79), and large (≥0.80) [31]. The number of variables in the final model was limited to 15 events per variable to prevent overfitting of the binary logistic regression analysis [32]. The predictive accuracy of variables significant in the multivariable binary logistic regression analyses for grade II LHB tears was determined using the area under the receiver operating characteristic curve (AUC) for both the overall studied cohort and the large-to-massive RCT cohort. AUC predictive accuracy was assessed as unsatisfactory (0.5≤ AUC <0.6), satisfactory (0.6≤ AUC <0.7), good (0.7≤ AUC <0.8), very good (0.8≤ AUC <0.9), and excellent (0.9≤ AUC <1.0) [33]. The AUCs of the significant variables were compared using the DeLong test [34]. All statistical analyses were performed using SPSS software version 26.0 (IBM Corp.) except the DeLong test, which was performed using MedCalc statistical software (MedCalc Software, version 19.2.6; https://www.medcalc.org; 2020). Statistical significance was set at a P-value of <0.05, with the exception of the parallel line test and Hosmer-Lemeshow test, for which significance was set at a P-value of >0.05.

RESULTS

The 750 patients enrolled had the following LHB tear: grade 0 for 406 patients, grade I for 218 patients, and grade II for 126 patients. Among the 126 grade II patients, 22 had complete tears of the LHB. In the overall cohort, both univariate and multivariable ordinal logistic regression analyses, hs-CRP >1 mg/L, SSC tear, hypothyroidism, and tangent sign were significantly associated with LHB tear severity (P<0.05) (Table 2). The univariate binary logistic regression analyses showed that being male, hs-CRP >1 mg/L, a full-thickness RCT, PSRCT, SSC tear, tangent sign, Patte retraction degree ≥stage I, hypertension, hypothyroidism, hyperthyroidism, and age ≥70 years were significantly associated with grade II LHB tears (P<0.05) (Table 3). Because multicollinearity existed between a full-thickness RCT (as the tear size) and the Patte retraction degree, those two variables were included in separate multivariable analyses. The first multivariable binary logistic regression analysis showed that hs-CRP>1 mg/L and an SSC tear were significantly associated with grade II LHB tears (P<0.001) (Table 4). Both hs-CRP >1 mg/L and SSC tears had satisfactory predictive accuracy for grade II LHB tears, with AUCs of 0.67 and 0.68, respectively. The second multivariable binary logistic regression analysis showed that hs-CRP >1 mg/L and the Patte retraction degree were significantly associated with grade II LHB tears (P<0.001), with AUCs of 0.67 and 0.72, respectively (Table 4). The combination of hs-CRP >1 mg/L and SSC tear had a significantly larger AUC than either hs-CRP >1 mg/L or SSC tear alone (P<0.01). That combination had good predictive accuracy for grade II LHB tears; its AUC, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were 0.72, 0.82, 0.63, 0.31, and 0.94, respectively. The combination of hs-CRP >1 mg/L and Patte retraction degree did not have a significantly larger AUC than that of hs-CRP >1 mg/L alone (P=0.069) or Patte retraction degree alone (P=0.630), but it did have good predictive accuracy for grade II LHB tears; its AUC, sensitivity, specificity, PPV, and NPV were 0.72, 0.93, 0.50, 0.27, and 0.97, respectively. The AUCs of the combination of hs-CRP >1 mg/L and SSC tear and the combination of hs-CRP >1 mg/L and Patte retraction degree did not differ significantly from each other (P=0.783).

Table 2.

Strengths of the associations between the studied factors and LHB tear severity

Table 3.

Strengths of the associations between each studied factor and a grade II LHB tear in the overall studied cohort

Table 4.

Strengths of the associations between each studied factor and a grade II LHB tear in the overall studied cohort

In the large-to-massive RCT cohort, hs-CRP>1 mg/L, hypertension, SSC tear, and age ≥50 years were significantly associated with LHB tear severity in the univariate ordinal logistic regression analyses (P<0.05). In the multivariable ordinal logistic regression analysis, hs-CRP >1 mg/L, hypertension, and age ≥50 years were significantly associated with LHB tear severity (P<0.05) (Table 5). Both the univariate and multivariable binary logistic regression analyses showed that hs-CRP>1 mg/L and hypertension were significantly associated with grade II LHB tears (P<0.05) (Table 6). Serum hs-CRP >1 mg/L had good predictive accuracy for grade II LHB tears; its AUC, sensitivity, specificity, PPV, and NPV were 0.70, 0.58, 0.81, 0.62, and 0.80, respectively (Fig. 3). Hypertension had satisfactory predictive accuracy for grade II LHB tears; its AUC, sensitivity, specificity, PPV, and NPV were 0.60, 0.44, 0.76, 0.49, and 0.73, respectively (Fig. 3). The combination of hs-CRP >1 mg/L and hypertension had only satisfactory predictive accuracy for grade II LHB tears with an AUC of 0.69, which was no increase over the AUC of hs-CRP>1 mg/L alone (P=0.710).

Table 5.

Strengths of the associations between the studied factors and LHB tear severity in the large-to-massive rotator cuff tear cohort

Table 6.

Strengths of the associations between the studied factors and a grade II LHB tear in the large-to-massive rotator cuff tear cohort

Fig. 3.

In the large-to-massive rotator cuff tear cohort, the areas under the receiver operating characteristic (ROC) curves (AUCs) of high-sensitivity C-reactive protein (hs-CRP) >1 mg/L and hypertension were 0.70 and 0.60, respectively. Serum hs-CRP >1 mg/L had good predictive accuracy for a grade II biceps long head tendon tear (P<0.001). The combination of hs-CRP >1 mg/L and hypertension had satisfactory predictive accuracy for a grade II LHB tear. The AUC of the combination of hs-CRP >1 mg/L and hypertension was 0.69, which was not an increase over the AUC of hs-CRP >1 mg/L alone (P=0.710).

DISCUSSION

The results of this study show that only hs-CRP >1 mg/L was significantly associated with LHB tear severity and grade II LHB tears in both the overall and large-to-massive RCT cohorts. Therefore, this study has confirmed the hypothesis that hs-CRP >1 mg/L is a factor significantly associated with LHB tear severity and a good predictor of grade II LHB tears.

Serum hs-CRP is a well-known inflammatory marker. It has been used in the diagnosis of atherosclerosis, cardiovascular disease, and stroke [35]. Several studies have reported that musculoskeletal disorders, including knee osteoarthritis, frozen shoulder, and sciatica, are also associated with subclinical systemic inflammatory responses that can be detected using hs-CRP [21,22,36,37]. Carp et al. [38] reported that elevated CRP can be initiated by a local response and is proportionally amplified in the presence of greater tissue injury and inflammation. Biomolecular studies have reported that in biceps tendinopathy, the gene expressions for proinflammatory or inflammatory cytokines are upregulated, and the gene expression for an anti-inflammatory cytokine is downregulated [3,39]. Histological studies have reported that inflammatory processes are involved in LHB tendinopathy [40,41]. One study reported that serum hs-CRP >1 mg/L is an independent risk factor for LHB tears without any making distinction in tear severity [18]. The results of this study show that serum hs-CRP >1.0 mg/L is associated with LHB tear severity, which suggests that inflammatory processes are involved in LHB tears and tear severity. Serum hs-CRP >1.0 mg/L was also found to be associated with grade II LHB tears and a good predictor of grade II LHB tears in the large-to-massive RCT cohort, which is a clinically helpful finding for determining preoperatively whether the LHB can be used for rotator cuff repair surgery.

The population of this study contained 22 patients with complete LHB tears. Although complete tears of the LHB can be readily diagnosed using magnetic resonance imaging (MRI), these patients were included in this study to determine factors associated with LHB tear severity. An additional study that excluded patients with complete LHB tears from the cohort was conducted to ascertain whether those tears affected the study results. The results of that analysis closely mirror those of the original study, as follows. In the overall cohort, age ≥50 years, hs-CRP >1 mg/L, and an SSC tear were significantly associated with LHB tear severity (P≤0.001). In the large-to-massive RCT cohort, only serum hs-CRP>1 mg/L was significantly associated with grade II LHB tears (P<0.001), and it had satisfactory predictive accuracy (AUC, 0.69). Therefore, in a clinical situation, a surgeon planning to repair a large to massive cuff tear, which frequently requires biceps tendon augmentation, could consider hs-CRP test results to predict a biceps tendon injury before surgery.

In this study, in addition to hs-CRP, an SSC tear, hypothyroidism, tangent sign, Patte retraction degree, age ≥50 years, and hypertension were significantly associated with LHB tear severity and/or grade II LHB tears, depending on the studied cohort. SSC tears have been reported to be associated with LHB tears in several studies [18,23]. Medial instability of the LHB has been reported to be associated with SSC tears [23]. Sahu et al. [15] reported that an abrasion or partial tear of the anterior portion of the LHB, the sentinel sign, indicates a coexisting SSC tear. Biomechanical studies have demonstrated that the increase in the load on the LHB is more significant after an SSC tear than after an infraspinatus tendon tear [42,43]. Those previous studies’ findings suggest that a meaningful association is present between SSC tears and LHB tears, supporting this study’s finding that an SSC tear was significantly associated with LHB tear severity.

Hypothyroidism has been reported to cause deterioration of the extracellular matrix and then induce tendon injury through an accumulation of glycosaminoglycans in the extracellular matrix [44]. Oliva et al. [45,46] demonstrated the presence of thyroid hormone receptors in healthy and pathologic rotator cuff tendons and showed in vitro that thyroid hormone enhances tenocyte growth and counteracts apoptosis in healthy tenocytes in a dose- and time-dependent manner. In a clinical observational study, those authors also reported a relationship between thyroid pathology and nontraumatic RCT [45]. Those previous studies' findings that hypothyroidism is associated with tendon injury support this study’s finding that hypothyroidism is significantly associated with LHB tear severity.

The chronicity of RCT has been defined as the duration of a tear, which is represented as fatty infiltration and muscle atrophy [47]. The tangent sign is an index of supraspinatus muscle atrophy that was reported to be highly correlated with the retraction grade of the tendon and tear size in RCT [26,48]. Melis et al. [47] reported that a positive tangent sign appeared an average of 4.5 years after the onset of symptoms. Chronic RCT with symptom duration longer than three months or a massive tear has been reported to be associated with LHB lesions [49]. RCT size is known to increase over time [50]. Wu et al. [39] indicated that the size of a coexisting RCT plays a role in the severity of LHB tendinopathy. Those previous findings suggest that the chronicity of RCT, as presented by the tangent sign and Patte retraction degree, is a risk factor for LHB tear severity or grade II LHB tears.

Age has been reported as a risk factor for LHB tears. Takeshima et al. [51] reported that the prevalence of LHB lesions increased with age. LHB tears mostly occur in persons older than 50 years [52]. A histologic study also reported that age is significantly related to LHB degeneration [5]. This study’s finding that age ≥50 years was significantly associated with LHB tear severity in the large-to-massive RCT cohort supports the previous studies' findings of a significant association between LHB tears and increasing age.

Hypertension has been reported to be associated with symptomatic RCT [53]. Some studies reported that hypertension is associated with the severity and prevalence of RCT [54,55]. Hypertension has been reported to be prevalent in posterior tibial tendon ruptures and Achilles tendinopathy [56,57]. Previous studies’ findings suggest that hypertension deteriorates tendon integrity and increases RCT severity, and then that increased RCT severity affects LHB integrity [58].

This study has some limitations. Although biceps tendon status was evaluated after intra-articular retraction of the biceps tendon, hidden biceps lesions, which would have been far distal to the transverse humeral ligament, were not assessed. Hypertrophied biceps tendons and lipstick sign have been reported as biceps pathologies, but they were not evaluated here [59]. Another limitation is that this study did not include a physical examination to diagnose LHB injury. Several physical tests, including the Speed, Yergason, and uppercut tests, are known to diagnose LHB tears, though without considering severity [60-62]. However, it is difficult to find studies showing whether any physical tests can predict LHB tear severity [63]. Also, physical examinations have issues with intraobserver and interobserver reproducibility [60]. On the other hand, hs-CRP is an objective indicator for diagnosing grade II LHB tears because it has no reliability issues with intraobserver and interobserver reproducibility. However, to increase the predictive accuracy for grade II LHB tears, further research that includes those physical tests should be performed. Nevertheless, given the difficulty of diagnosing LHB tears using conventional 1.5T MRI [64], this study’s findings are helpful for predicting biceps tendon status before arthroscopic surgery. Among the studied variables, an SSC tear, hypothyroidism, tangent sign, Patte retraction degree, age ≥50 years, and hypertension were found to be significantly associated with LHB tear severity or grade II LHB tears, depending on the studied cohorts. Conversely, hs-CRP >1 mg/L consistently predicted both LHB tear severity and grade II LHB tears across both studied cohorts. Therefore, hs-CRP >1 mg/L can serve as a reliable predictor of LHB tear severity and grade II LHB tears before surgery.

CONCLUSIONS

Serum hs-CRP >1 mg/L is associated with LHB tear severity and serves as a reliable predictor of grade II LHB tears, facilitating the preoperative assessment of the LHB as a potential graft material in arthroscopic rotator cuff repair.

Article information Continued