The role of subscapularis muscle denervation in the pathogenesis of shoulder internal rotation contracture after neonatal brachial plexus palsy: A study in a rat model


  • This study was approved by the Ethic commission of VHIR center, act number 24, extraordinary session November 20, 2008.


We assessed the role of subscapularis muscle denervation in the development of shoulder internal rotation contracture in neonatal brachial plexus injury. Seventeen newborn rats underwent selective denervation of the subscapular muscle. The rats were evaluated at weekly intervals to measure passive shoulder external rotation. After 4 weeks, the animals were euthanized. The subscapularis thickness was measured using 7.2T MRI axial images. The subscapularis muscle was then studied grossly, and its mass was registered. The fiber area and the area of fibrosis were measured using collagen-I inmunostained muscle sections. Significant progressive decrease in passive shoulder external rotation was noted with a mean loss of 58° at four weeks. A significant decrease in thickness and mass of the subscapularis muscles in the involved shoulders was also found with a mean loss of 69%. Subscapularis muscle fiber size decreased significantly, while the area of fibrosis remained unchanged. Our study shows that subscapularis denervation, per se, could explain shoulder contracture after neonatal brachial plexus injury, though its relevance compared to other pathogenic factors needs further investigation. © 2014 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 32:1675–1679, 2014.

Shoulder impairment is the most common long-term complication and the major cause of morbidity in children with upper trunk neonatal brachial plexus palsy (NBPP).[1]

The subscapularis muscle has been implicated as a primary deforming force leading to the development of shoulder internal rotation contracture (SIRC). This constant internal rotation posture leads to early glenohumeral joint deformity (i.e., glenohumeral dysplasia).[2-4] While the natural history of the changes in bone and articular alignment has been extensively studied, the pathogenesis of muscular changes preceding them remains unclear and is currently receiving more attention.[5-10] The mechanism leading to subscapularis muscle growth impairment with subsequent volume loss and contracture has been explained by two differing hypotheses: denervation[8, 9, 11, 12] and muscle imbalance.[5-7, 10, 13, 14] The imbalance theory infers that the subscapularis changes are secondary to the lack of passive stretching in the absence of active shoulder external rotation, which leads to progressive shortening and stiffness of subscapularis muscle due to an abnormal muscle growth or muscle hypoplasia.[5]

The denervation hypothesis states that denervation of the subscapularis muscle causes progressive muscle fibrosis, shortening, and contracture.[9] Recent experiments in a mouse model of upper trunk neurectomy supported this second hypothesis, showing that muscle reinnervation prevents contracture.[8, 12] However, these studies showed that muscle shortening and contracture occurred without fibrosis. We assessed the contribution of selective subscapular denervation through surgical neurectomy in the neonatal period in a rat model.


This study was conducted following NIH Guidelines for the use of laboratory animals and with the approval of the local ethics committee for experimental animal use. Seventeen rat newborns from one pregnant Sprague–Dawley Oncins France Strain A (OFA) rat were used. A sham group, undergoing a surgical approach without neural injury, was not considered necessary as previous studies proved this group had no alterations in passive shoulder external rotation (PSER) or subscapularis growth.[13]

PSER measurement, magnetic resonance imaging (MRI), and subscapularis muscle weighting were done 4 weeks after surgery as described below and following the methodology of previous studies.[13] Sectional muscle fiber size and muscle area of fibrosis of subscapularis muscles were determined using images from collagen-I inmunostaining and also followed the methods of a previous study.[13]

Rat Newborn Surgery

Five-day-old rat pups underwent right brachial plexus surgery under general anesthesia with isofluorane. A surgical microscope was needed for dissection. A transverse incision below the clavicle was made, splitting the pectoralis major and minor muscles to expose the brachial plexus. The superior subscapular nerve was identified, as it branched from the suprascapular nerve. The inferior subscapular nerve was also identified as it branched from the posterior cord.[13] A neurectomy with complete excision, from the origin to the entrance to the muscle, of both nerves was performed using microscissors to prevent potential spontaneous nerve regeneration and muscle reinnervation. The skin incision was closed using a running 7/0 polypropylene suture (Prolene, Ethicon, Johnson & Johnson).

Passive Shoulder External Rotation Measurement (PSER)

PSER was determined at weekly intervals for 4 weeks. Measurements from 3 weeks were performed with the rats anesthetized with an isofluorane-oxygen mixture while the measurement at the 4th week was made after the animal was euthanized using intraperitoneal sodium pentobarbital after proper sedation. PSER was measured using a previous method.[13]

The measurements were performed in full adduction with neutral flexion and extension of the shoulder. The neutral position was defined as the shoulder in 0° of abduction and the elbow in 90° of flexion with the front limbs up ventrally perpendicular to the examination table. After scapular stabilization with the thumb, each shoulder was placed in maximal external glenohumeral joint rotation and photographed using a 12-megapixel digital reflex camera (Canon EOS 1100D, Tokyo) stabilized on a tripod perpendicular to the rat.

The angle formed by the forearm and the animal's midline was defined as glenohumeral external rotation. Measurements were performed digitally using Osirix (Apple Inc, Cupertino, CA). The non-operative side was measured and used as a control. A previous pilot study of 10 rats tested the reliability of this measurement method. Both limbs were positioned and photographed three times each, and each photograph was measured twice. The intraclass correlation coefficient was >0.93, suggesting high reliability.[13]

MRI Evaluation

Each animal was then placed supine with the front limbs onto its abdomen into a 7.2T MRI Biospect scanner (Bruker, Germany). Both shoulders were independently evaluated using RARE 1 mm axial TR 4000 T 30 sequences in sagittal oblique and oblique axial planes parallel to the scapula. To standardize the cross-sectional slice chosen for measurements, the axial image selected to optimally visualize the subscapularis and infraspinatus muscles was just inferior to the scapular spine. The point of maximum thickness was measured on the MRI of the involved and uninvolved shoulders that could include different slices for each side. Measurements were made using Osirix (Apple Inc) by a pediatric musculoskeletal radiologist. Previous reliability analysis in which each measurement was repeated three times over three days and calculated by the mean of the three measures showed an intraclass correlation coeficient >0.96, proving high reliability.[13]

Subscapularis Muscle Mass

After MRI, the right and left subscapularis muscles were excised and weighed. Through an anterior midline incision, both the pectoralis major and the clavicle were sectioned to gain access to the subscapularis fossae. The entire subscapularis was cleaned and dried of any surface moisture, then placed on a high precision analytic scale (AV114, Adventurer Pro, OHAUS, Switzerland).

Analysis of Subscapularis Muscle Fiber Cross-Sectional Area and Area of Fibrosis

Measurements were performed using the images obtained from inmunofluorescence of collagen-I stained subscapularis muscle sections. For homogeneity purposes, all images were obtained from cuts of the same region (middle part of the muscle) and at same area of the muscle (the inferior area adjacent to the upper subscapularis nerve). Fresh, whole subscapularis muscles were embedded and suspended in OCT compound (Tissue-Tek) and immediately frozen in liquid nitrogen. The tissue blocks were stored at −80°C until used. Ten µm serial frozen muscle transverse sections were cut at below −20°C using an HM505E microtome (Microm, Walldorf, Germany) and mounted on 0.1% (w/v) poly-L-lysine pre-treated slides. Muscle sections were fixed in cold (−20°C) acetone for 5 min, rinsed twice with PBS, and blocked with 10% goat serum in PBS for 20 min. Primary antibody against collagen-I (ab292, Abcam) were diluted in PBS with 1% BSA and incubated in a humid chamber for 16 h at 4°C. Next, muscle sections were washed three times with PBS and incubated with Alexa Fluor® 488 anti-rabbit (A-11008, Invitrogen) secondary antibody diluted in PBS with 3% BSA in a dark humid chamber for 45 min at room temperature. Finally, slides were washed three times with PBS and mounted using a coverslip and fluoromount-g (Southern Biotech) mounting medium. Immunostained sections were imaged with a microscope (DM6000, Leica) at 20× magnification. Skeletal muscle fibers from cross-sectional areas of each muscle were then measured with ImageJ software (version 1.45 s; NIH, Bethesda, MD). Quantification of fibrosis in the subscapularis muscles was made using the software on the same images used for measurement of fiber size, avoiding intramuscular tendons and epimysium. The collagen I stain was separated by color deconvolution, and the resulting images were converted to binary format for pixel counts to determine the ratio of collagen to muscle area.

Statistical Analysis

Non-parametric analysis was performed for comparisons. The Wilcoxon test was used to check differences in paired comparisons (control vs. affected limbs).

The level of significance was set at p ≤ 0.05.


The results are reproduced in Table 1.

Table 1. Shoulder Rotation, Subscapularis Thickness, and Subcapularis Weight: (a) Wilcoxon Test Between Involved and Uninvolved Shoulder in the Suprascapular Group
 Shoulder Passive External Rotation (degrees)Subscapularis Thickness (cm)Subcapularis Mass (gr)
Involved shoulder58.750.140.15
Suprascapular group m (SD)(12.18)(0.03)(0.04)
Uninvolved shoulder79.380.210.26
Suprascapular group m (SD)(3.59)(0.03)(0.05)

Passive Shoulder Joint Motion

A significant decrease in PSER occurred in the involved shoulder starting the 2nd week. Comparing between weeks, a significant progressive decrease in PSER was found in the involved shoulder (p < 0.01) with a mean loss of 4.12° (95%, CI: 0.1–8.2) in the 1st week after surgery, 38.81° (95%, CI: 34.9–42.8) in the 2nd week, 43.54° (95%, CI: 39.4–47.6) in the 3rd week with a mean loss of 58.38° (95%, CI: 53.5–63.1) in the 4th week (Fig. 1).

Figure 1.

Weekly shoulder passive external rotation showed a progressive shoulder internal rotation contracture after subscapular nerve injury in the right side after the first week of denervation.

MRI Evaluation and Muscle Changes

The involved shoulder showed a significant decrease of subscapularis muscle thickness with a mean loss of 68,4% (95% CI: 62,79%–74,08%) (p < 0.01) (Fig. 2).

Figure 2.

Shoulder rat MRI four weeks after subscapular neurectomy showing asymmetrical subscapular thickness with severe atrophy in the right side, but preservation of infraspinatus thickness.

Muscle Mass

A significant decrease in involved subscapularis weight occurred compared to the nonoperated shoulder with a mean loss of 68,77% (95% CI: 63,46%–74,07%) (p < 0.01).

After dissection, all involved subscapularis muscles showed a severe muscle wasting and pale color compared to the contralateral subscapularis while supraspinatus, teres major, latissimus dorsi, and anterior serratus muscles were macroscopically similar to the contralateral ones.

Cross-Sectional Subscapularis Muscle Fiber Area and Fibrosis

Significant differences were found in the cross-sectional subscapularis muscle fiber area, measured in relative area units, between left (control), and right (subscapular neurectomy) subscapularis: 0.273 (95%, CI: 0.268–0.278) and 0.038 (95%, CI: 0.028–0.048), respectively (p < 0.000) (Fig. 3). No differences were found in collagen-I deposition area, measured as the percentage of the total muscle section area, between the left and right: 30.2%(95% CI: 39.2–21.2) and 35.27 (95% CI: 25–45.54), respectively (p < 0.03) (Fig. 3).

Figure 3.

Subscapularis collagen-I immunofluorescence microphotographs (×20). Denervated subscapularis muscle (B) shows a marked fiber atrophy and increase of collagen compared to uninvolved subscapularis (A).


Selective subscapularis muscle denervation in the neonatal age of a rat led to marked muscle volume loss and contracture without fibrosis, which resulted in a progressive loss of shoulder external rotation. The precise sequence of events leading to shoulder contracture in children with NBPP remains speculative. Subscapularis thickness loss is a consistent MRI finding in patients with an internal rotation contracture and glenohumeral joint dysplasia.[6, 9, 10, 14] The pathogenesis of subscapularis changes seems to be multifactorial and derived from an alteration in muscle growth both in length and thickness, by either shoulder rotation imbalance, subscapularis denervation, or both.[5, 8, 11-13]

The rodent neonatal model was described as a valid preclinical model for muscle anomalies following NBPP as upper trunk neurectomy showed subscapularis changes similar to those occurring in children with upper NBPP.[8, 15] However, in these models, both subscapularis and shoulder external rotators are denervated. This forbids us from attributing subscapularis contracture solely to the subscapularis denervation or to the shoulder imbalance resulting from external rotation weakness. Experimental isolation of these hypothetical pathogenic factors by selective intervention on nerve or muscle while creating models of denervation or imbalance might help to clarify the pathogenesis.

In Nikolau's study, a first group of neonatal mice underwent an excision of external rotator muscles (infraspinatus and teres minor) to achieve a muscular model of shoulder imbalance.[5] This shoulder imbalance group did not develop an SRIC. On the contrary, a second group of neonatal mice that underwent an upper trunk neurectomy developed a shoulder contracture. Based on these findings, the authors indirectly inferred that the primary cause of glenohumeral joint contracture may be a denervation phenomenon of the subscapularis muscle, leading to altered growth of the muscle.

We isolated the denervation factor by developing a model of selective subscapularis denervation in the neonatal period and, although this is not a reproducible situation clinically, the experiment analyzes directly the contribution of muscle denervation in the development of shoulder contracture. By preserving shoulder external rotator muscles' innervations, we could reduce their confounding contribution through the hypothesized imbalance model. We also minimized the possible confounding effect of the reinnervation by excising a relevant segment of the subscapular nerves and thus, avoiding a potential spontaneous regeneration.[12]

We isolated shoulder imbalance factors by performing a selective neonatal neurectomy of the suprascapularis nerve and keeping the subscapularis innnervation.[13] Neonatal rats developed a fast and severe shoulder contracture within the first week after the neural injury. Around 25% decrease in the thickness and weight of the non-denervated subscapularis occurred without muscle fiber atrophy or fibrosis. These findings were similar to those found in subscapularis histology samples taken from children with GHD.[7] The authors concluded that a muscle hypoplasia occurred as a result of rotator shoulder imbalance with lack of passive stretching in the absence of active shoulder external rotation.[5]

In our study, a selective subscapularis muscle denervation produced a slower shoulder contracture as compared to the neural imbalance model[4] (Fig. 1). In fact, shoulder contracture appeared only after the first week of the neural injury. Subscapularis thickness and weight decrease were also substantially higher than those in the imbalance model with a severe subscapularis muscle fiber atrophy.[11] In our model of complete neurotmesis of both superior and inferior subscapularis nerves, reinnervation was impossible. This could explain the differences with the subscapularis muscle histology findings in children with NBPP and glenohumeral dysplasia. The absence of muscle fiber atrophy in children could be rationalized by muscle reinnervation from C5-C6 and preservation of innervation arising from C7 level.[4, 7, 11]

No significant subscapularis muscle fibrosis occurred supporting other neonatal mice studies showing that muscle fibrosis appears at a later stage, after muscle shortening and stiffness in models of NBPP.[6]

Our findings showed that denervation of the subscapularis is a prime factor for the internal rotation contracture seen after NBPP, but its pathogenic significance needs further research. Further refinements in the rat model are required to determine the relative contribution of denervation/reinnervation of the subscapularis and muscle imbalance. Muscle mechanical studies should also be performed and would be of paramount importance in this sense. The contribution of changes in anterior capsular contracture will also need to be evaluated to fully understand the pathogenesis of shoulder dysplasia following NBPP.

Our study supports the theory that subscapularis denervation is a cause of shoulder contracture in patients with neonatal brachial plexus palsy.


This work was funded by Instituto de Salud Carlos III, grant FIS PI10/01357 co-financed by the European Regional Development Fund (ERDF), Fundació' Privada A. Bosch and Fundação Santa-Maria-Silva. The grant was fully destined to fungible. No author, immediate families, or any research foundation with which they are affiliated, received any financial payments or other benefits from any commercial entity related to the subject of this article.