Partial­Thickness Rotator Cuff Tears Andrew B. Wolff, MD

Abstract Paul Sethi, MD

Partial­thickness rotator cuff tears are not a single entity; rather,

Karen M. Sutton, MD

they represent a spectrum of disease states. Although often

Aaron S. Covey, MD

asymptomatic, they can be significantly disabling. Overhead

David P. Magit, MD

Michael Medvecky, MD

throwing athletes with partial­thickness rotator cuff tears differ with respect to etiology, goals, and treatment from older, nonathlete patients with degenerative tears. Pathogenesis of degenerative partial­thickness tears is multifactorial, with evidence

Dr. Wolff is Resident, Department of

of intrinsic and extrinsic factors playing key roles. Diagnosis of

Orthopaedics and Rehabilitation, Yale

partial­thickness rotator cuff tears should be based on the patient’s University School of Medicine, New Haven, CT. Dr. Sethi is Orthopaedic Surgeon, The Greenwich Sports and

symptoms together with magnetic resonance imaging studies. Conservative treatment is successful in most patients. Surgery

Shoulder Service, Orthopaedic and

generally is considered for patients with symptoms of sufficient Neurosurgical Specialists, Greenwich, CT. Dr. Sutton is Resident, Department of Orthopaedics and Rehabilitation, Yale

duration and intensity. The role of acromioplasty has not been clearly delineated, but it should be considered when there is

University School of Medicine. Dr. Covey

evidence of extrinsic causation for the partial­thickness rotator is Resident, Department of

cuff tear. Orthopaedics and Rehabilitation, Yale University School of Medicine. Dr. Magit is Fellow, Kerlan­Jobe Orthopaedic Clinic, Los Angeles, CA. Dr. Medvecky is Assistant Professor, Department of Orthopaedics and Rehabilitation, Yale University School of Medicine.

A dvances py, basic in science shoulder research, arthrosco­ and imaging modalities continue to in­ crease our understanding of and abil­ ity to treat rotator cuff disease. Com­ pared to studies on full­thickness rotator cuff tears, limited published data are available

regarding the opti­ mum treatment of partial­thickness rotator cuff tears. In part, this is be­ cause of the emerging concept that

partial­thickness rotator cuff tears are not a single entity with a single etiology. Similarly, treatment of symptomatic partial­thickness rota­ tor cuff tears should be based on the patient’s goals, injury site, and cause of the tear.

Anatomy and Classification

The rotator cuff represents the coa­ lescence of the subscapularis, su­

praspinatus, infraspinatus, and teres minor tendons. Clark and Harry­ man1 showed a significant amount of interdigitation of these tendons, as well as between the tendons, the None of the following authors or the

shoulder capsule, and the coraco­ departments with which they are

humeral ligament (Figure 1). Most affiliated has received anything of value

partial­thickness rotator cuff tears in from or owns stock in a commercial

older patients occur on the articular company or institution related directly or

side of the supraspinatus tendon; indirectly to the subject of this article:

isolated involvement of the bursal Dr. Wolff, Dr. Sethi, Dr. Sutton, Dr.

side of the supraspinatus tendon or Covey, Dr. Magit, and Dr. Medvecky.

of the infraspinatus and subscapu­

Reprint requests: Dr. Wolff, Department

laris tendons is less common. In con­

of Orthopaedics and Rehabilitation, Yale

trast, tears in younger overhead

University School of Medicine, Yale

throwing athletes are articular­sided, Physicians Building, First Floor, 800

partial­thickness tears of the domi­ Howard Avenue, New Haven CT 06520.

nant arm at the supraspinatus­

J Am Acad Orthop Surg 2006;14:715­ 725

infraspinatus interval.

Partial­thickness tears of the rota­ tor cuff can be articular­sided, Copyright 2006 by the American

bursal­sided, intratendinous, or a Academy of Orthopaedic Surgeons.

combination thereof. Ellman2 de­ scribed a classification system that

Volume 14, Number 13, December 2006 715

Partial­Thickness Rotator Cuff Tears

716 Journal of the American Academy of Orthopaedic Surgeons

Figure 1

Transverse section of the rotator cuff with the orientations of the fascicles indicated by the lines on their upper surfaces. Layer 1 is composed of superficial fibers that overlie the cuff tendons and form an extension of the coracohumeral ligament (chl). Layers 2 and 3 contain the fibers of the supraspinatus (SP) and infraspinatus (IS) tendons. The fibers in layer 2 are oriented parallel to the axes of the supraspinatus and infraspinatus tendons; the fibers of layer 3 are smaller and obliquely oriented compared with the fibers of layer 2. The fibers of layer 4 make up the deep extension of the coracohumeral ligament. Layer 5 is the joint capsule. (Adapted with permission from Clark JM, Harryman DT II: Tendons, ligaments, and capsule of the rotator cuff: Gross and microscopic anatomy. J Bone Joint Surg Am 1992;74:713­ 725.)

A, Arthroscopic view through the midlateral port of a small bursal­sided partial­ thickness tear (Ellman grade B2) after débridement, showing intact anterior and posterior attachments of the rotator cuff. B, Arthroscopic view through the midlateral portal of a nearly full­thickness tear of the bursal side of the rotator cuff (Ellman grade B3).

Table 1

Ellman’s Classification of Partial­Thickness Rotator Cuff Tears

Location Grade

A: Articular surface

differentiates between partial­ and full­thickness tears of the rotator cuff based on findings at the time of arthroscopy. Partial­thickness tears are subdivided based on location and thickness of tear (Table 1) (Figures 2

1: <3 mm deep

and 3). Ruotolo et al3 showed the mean thickness of the supraspinatus tendon to be 11.6 mm anteriorly, B: Bursal surface 2: 3­6 mm deep

12.1 mm at midtendon, and 12 mm

C: Interstitial 3: >6 mm deep

posteriorly.

(Adapted with permission from

Incidence Ellman H: Diagnosis and treatment of incomplete rotator cuff tears. Clin Orthop Rel Res 1990;254:64­74.)

Cadaveric and imaging studies have been performed in an attempt to de­ fine the incidence of rotator cuff

Figure 3

Figure 2

A, Arthroscopic view through the posterior portal of a small partial­thickness tear of the articular side of the rotator cuff (Ellman grade A1). B, Arthroscopic view through the posterior portal of a deep partial­thickness tear of the articular side of the rotator cuff (Ellman grade A3).

tears. In 1934, Codman4 noted a 32% prevalence of supraspinatus rupture. More recently, Fukuda et al5 reported a prevalence of 7% full­ thickness and 13% partial­thick­ ness tears (of which 18% were bur­ sal­sided, 27% articular­sided, and 55% intratendinous) in 249 cadav­ ers. However, Fukuda later cau­ tioned that “the statistics from ca­ daveric studies are skewed, because most of the specimens are from peo­ ple older than the patients in clini­ cal practice.”6

Data from imaging studies show an increasing incidence with age. In

Andrew B. Wolff, MD, et al

a 1995 magnetic resonance imaging

sible explanation for the increased (MRI) study of 96 asymptomatic

incidence of articular­sided tears fol­

Figure 4

shoulders, Sher et al7 reported no

lowing a traumatic event. full­thickness and 4% partial­

The etiology of partial­thickness thickness tears in subjects 19 to 39

rotator cuff tears can be broadly dif­ years of age. In contrast, among indi­

ferentiated into intrinsic, extrinsic, viduals older than 60 years of age,

or traumatic causes. Evidence sug­ there was a 28% incidence of full­

gests that the pathogenesis of bursal­ thickness and a 26% incidence of

sided and articular­sided partial­ partial­thickness tears. Similarly, ul­

thickness tears may differ. In 1934, trasound studies of asymptomatic

Codman4 postulated that the origin volunteers has demonstrated a 5%

of rotator cuff tears was intrinsic and to 11% incidence of full­ or partial­

degenerative. In 1972, Neer13 pro­ thickness tears in subjects in their

posed that the etiology of rotator fourth and fifth decades, increasing

cuff tears was primarily extrinsic to 80% in the eighth decade.8

and secondary to subacromial im­ Throughout the literature, clini­

pingement. Although the pathogen­ cally noted articular­sided tears are

esis of rotator cuff tears is still the approximately two to three times

subject of debate, evidence suggests more common than bursal­sided

that both intrinsic and extrinsic fac­

Typical stress­strain curve for the supraspinatus tendon. Bursal and articular sides of the rotator cuff tendon show nonlinear deformation initially, followed by elastic elongation. The bursal side is able to undergo a greater deformation

and has a greater tensile strength. (Adapted with permission from Nakajima T, Rokuuma N, Hamada tears.

tors play a role. The intrinsic tendi­

K, Tomatsu T, Fukuda H: Histologic nopathy from degenerative changes

and biomechanical characteristics of

Pathogenesis

is most often thought to result in articular­sided tears in older pa­ Bursal and articular surfaces of the

tients. Similarly, most partial­

the supraspinatus tendon: Reference to rotator cuff tearing. J Shoulder Elbow Surg 1994;3:79­87.) rotator cuff differ with respect to

thickness tears in young overhead their vascularity, biomechanical

throwing athletes are on the articu­ properties, and histologic composi­

lar side. Bursal­sided tears may have

Thus the differences in blood

sup­ tion. In 1965, Rothman and Parke9

a greater association with extrinsic

ply, biomechanical and histologic described the “critical zone” of hy­

factors such as coracoacromial arch

properties, associated changes of the povascularity near the insertion of

narrowing or impingement from the

acromion, and association with trau­ the supraspinatus on the humerus.

distal clavicle.

ma suggest that the pathogenesis of Rathbun and Macnab10 histological­

In a cadaveric study, Ozaki et al14

articular­ and bursal­sided partial­ ly correlated areas of hypovasculari­

found that all bursal­sided tears were

thickness tears of the rotator cuff ty and degeneration of the rotator

associated with attritional lesions on

may differ. Evidence points to intrin­ cuff in this region. Lohr and

the coracoacromial ligament as well

sic factors such as hypovascularity Uhthoff11 found that the critical

as the anterior third of the undersur­

and decreased tensile strength put­ zone hypovascularity predominated

face of the acromion; this was not

ting the articular side of the rotator on the articular side and extended

true of articular­sided tears, in

cuff in greater jeopardy, while both from the musculotendinous junc­

which the undersurface of the acro­

intrinsic and extrinsic factors sub­ tion to within 5 mm of the inser­

mion was almost always intact.

ject the bursal side of the rotator cuff tion.

Similarly, Burkhead et al15 reported

to greater wear. Nakajima et al12 reported differ­

that iatrogenic impingement of the ences in biomechanical and histo­

rotator cuff in rats caused exclusive­ logic properties in intact rotator

ly superior surface lesions on the

Clinical Presentation

cuffs taken from individuals at the

cuff. The impingement observed in

The most common reports by indi­ time of autopsy. They found that the

patients with bursal­sided cuff inju­

viduals with partial­thickness rota­ bursal side was composed mostly of

ries still may be secondary to a pri­

tor cuff tears are pain and stiffness. tendon bundles, whereas the articu­

marily intrinsic cause that produces

Pain is the predominant symptom, lar side was a complex of tendon, lig­

weakness of the rotator cuff. This

often most troubling at night and ament, and capsule. Furthermore,

weakness in turn may lead to supe­

with overhead activities. Many pa­ they found that the bursal side was

rior migration of the humerus that

tients have a painful arc of motion able to undergo a greater deforma­

causes increased impingement, initi­

with impingement signs, with or tion and had a greater tensile

ating what Ozaki et al14 described as

without apparent or real muscle strength (Figure 4). This offers a pos­

a “vicious cycle.”

weakness.

Volume 14, Number 13, December 2006 717

Partial­Thickness Rotator Cuff Tears

These partial tendon lesions are

full­thickness tears; the higher levels

imens have shown macrophages and often much more painful than full­

of substance P correlated with the

multinucleated giant cells, which thickness tears.6 In contrast to full­

significantly (P < 0.001) higher pain

exhibit strong immunoreactivity for thickness tears, partial­thickness de­

levels in the group of patients with

IL­1β, cathepsin D, and matrix me­ fects of the cuff have been theorized

partial­thickness tears. Gotoh et

talloprotease­1; none was found in to give rise to stiffness and nonphys­

al19 also found significantly higher

intact tendons.23 This led the au­ iologic tension on the remaining fi­

levels of interleukin (IL)­1β and IL­1

thors to conclude that granulation bers. Halder et al16 found that de­

receptor agonists (IL­1ra), which are

tissue around the insertion of a torn tachment of one or two thirds of the

inflammatory cytokine mediators,

supraspinatus tendon contributes to supraspinatus tendon from the cres­

in the subacromial bursae of patients

the progression of rotator cuff tears cent area had a negligible effect on

with rotator cuff pathology; the

by weakening the insertion of the re­ the force transmission of the rotator

amount of these mediators correlat­

maining tendon. cuff. These authors disputed the idea

ed with patient pain levels. Further­

Although relatively few direct that nonphysiologic tension plays a

more, they found that increased IL­

data are available on the natural his­ role in symptomatology. They con­

1β and IL­1ra in the glenohumeral

tory of partial­thickness rotator cuff cluded that neighboring fibers com­

joint did not correlate with pain in

tears, there is a substantial body of posing the cables of the crescent­

patients with rotator cuff patholo­

circumstantial evidence to suggest and­cable theory of rotator cuff

gy.20

that most partial tears do not heal on function were sufficient for force

their own. Patients may have wax­ transmission. However, Bey et al17 reported significantly increased

Natural History

ing and waning symptoms, but the clinical, biomechanical, epidemio­ strain on the remaining supraspina­

Although the natural history of

logic, and biologic data suggest that tus tendon at arm abduction angles

partial­thickness tears of the rotator

most of these tears progress to be­ of 30°, 45°, and 60° in a cadaveric

cuff is incompletely understood, ev­

come larger rather than smaller with model after creation of a one­third

idence suggests that progression

time. thickness articular­sided partial­

may occur over time.7,14 Yamanaka thickness tear of a supraspinatus

and Matsumoto21 reported on 40 pa­ tendon. They concluded that

tients with articular­sided tears diag­

Diagnostic Imaging

articular­sided tears predispose the

nosed by arthrography repeated at a

Various imaging modalities have remaining rotator cuff to further

mean of 412 days. They found an en­

been used to assist in the diagnosis damage. Thus in vitro studies have

largement of tear size in 21 patients

of partial­thickness rotator cuff shown that although mechanical

(53%), progression to full­thickness

tears. Arthrography and bursography strength may be preserved, strain on

tears in 11 patients (28%), reduction

have been described, but these are of neighboring fibers is increased. This

of tear size in 4 patients (10%), and

uncertain value given the wide rang­ increased strain likely plays a role in

disappearance of the tear in 4 pa­

es of reported accuracy in the litera­ the symptoms experienced by those

tients (10%). This early study indi­

ture—15% to 83% for arthrogra­ with partial­thickness tears of the

cates that many articular­sided tears

phy24,25 and 25% to 67% for rotator cuff.

can progress over relatively short pe­

bursography.5,25 Sensitivity and spec­ Fukuda6 reported that 73.3% of

riods. It also suggests that although

ificity of preoperative ultrasound has patients with either subacromial

some patients become asymptomat­

been reported to be as high as 94% bursitis or a partial­thickness rotator

ic over time, few tears heal anatom­

and 93% and as low as 41% and cuff tear reported “more than mod­

ically.

91%, respectively, for arthroscopi­ erate” nocturnal pain, whereas only

The limited healing potential of

cally confirmed partial­thickness 50% of patients with full­thickness

partial­thickness cuff tears is sup­

tears.26,27 tears reported this much pain. Fuku­

ported by histologic studies that

Other authors have reported a da also found that bursal­sided tears

have observed no active repair; prox­

similarly variable accuracy level for produced notably more pain than did

imal stumps of the cuff were round­

MRI in the evaluation of partial­ intratendinous or articular­sided

ed, retracted, and avascular. At the

thickness tears, with sensitivity as partial­thickness tears.6 Gotoh et

molecular level, cells at the tendi­

low as 56%28 and false­negative rates al18 reported significantly (P < 0.01)

nous margin of partial­thickness

as high as 83%.24 However, with im­ higher levels of substance P, an affer­

tears have been shown to contain α1

proved technology and the increas­ ent nerve pain mediator, in the sub­

type­I procollagen mRNA, a precur­

ing usage of MRI arthrography, the acromial bursae of patients with

sor of type I collagen, thereby dem­

accuracy of MRI has improved. A re­ partial­thickness rotator cuff tears

onstrating the potential for repair.22

cent study using magnetic resonance than in the bursae of patients with

However, other studies of tear spec­

arthrograms with gadopentetate

718 Journal of the American Academy of Orthopaedic Surgeons

Andrew B. Wolff, MD, et al

dimeglumine contrast agent and

Initially, these tears should be man­

In Neer’s original article describ­ coronal oblique T1­weighted fat­

aged with rest, activity modification,

ing open anterior acromioplasty,13 he suppression images yielded a sensi­

and nonsteroidal anti­inflammatory

noted that 15 of 16 patients with tivity of 84% and a specificity of

drugs. Physical therapy for range of

fraying of the bursal side of the su­ 96%.29

motion should then begin, with the

praspinatus who underwent anterior In comparing preoperative ultra­ sound to MRI in arthroscopically confirmed partial­thickness rotator cuff tears, Teefey et al30 found that 13 of 19 tears were correctly identi­ fied by ultrasound, whereas 12 of 19 tears were correctly identified by MRI. Iannotti et al31 found the pre­ operative diagnostic accuracy of ul­ trasound and MRI to be 70% and 73%, respectively.

Although ultrasound and MRI can be of similar utility in the diag­ nosis of partial­thickness tears, each has its limitations. Ultrasound can provide a less expensive and nonin­ vasive alternative for evaluation of these tears, but it is highly operator­ dependent and does not provide in­ formation regarding concomitant pathologies. In many patients with vague presenting signs and symp­ toms, MRI will offer a more com­ plete evaluation of the shoulder. MRI evaluation of partial­thickness rotator cuff tears demonstrates alter­ ation of the morphology, without discontinuity of the rotator cuff, on T1­weighted images in areas corre­ sponding to increased signal on T2­ weighted images.

For most patients with suspected partial­thickness cuff tears—espe­ cially young, overhead throwing ath­ letes—magnetic resonance arthrog­ raphy is the imaging modality of choice to best visualize these tears and assess for concomitant patholo­ gy. However, given the variable rates of accuracy and the high proportion of the population with asympto­ matic partial­thickness tears, MRI should be considered in conjunction with clinical evaluation.

goal of regaining any motion lost be­

acromioplasty had satisfactory re­ cause of capsular contractures. Al­ though recent reports have ques­ tioned the efficacy of corticosteroid injections,32 we have found them to be a useful adjunct. Strengthening of the rotator cuff and periscapular musculature should be initiated af­ ter range of motion has improved and inflammation has subsided. Al­ though there is a paucity of reliable reports on the clinical outcome of conservative treatment of partial tears, most patients will improve with conservative measures over 6 months; some continue to improve for up to 18 months.

Surgical intervention generally is considered for patients with symp­ toms of sufficient duration and in­ tensity. According to the literature, the timing of surgery has ranged from a few months to 1.5 years, but it should be based on the patient’s symptoms, improvement, and rate of improvement with nonsurgical therapy as well as on the goals of the patient. Many surgical procedures have been described to address the pathology and pathogenesis of partial­thickness tears; these include acromioplasty alone, débridement of tears with or without acromioplasty, and both mini­open and arthroscop­ ic repair of tears, with or without acromioplasty.

Ogilvie­Harris and Wiley33 report­ ed that approximately half of 57 pa­ tients with partial­thickness tears had successful results with arthro­ scopic débridement without acro­ mioplasty at a minimum of 1 year after surgery. Budoff et al34 reported 89% good and excellent results at 2­ to 5­year follow­up and 81% good and excellent results at >5­year fol­ low­up; other authors

have reported success rates of 50% to 89% after ar­ throscopic débridement without acromioplasty.33­35

Volume 14, Number 13, December 2006 719 sults at 9­month to 5­year follow­up. The articular side of the rotator cuff was not evaluated in these patients. Snyder et al35 reported 84% satisfac­ tory results with arthroscopic dé­ bridement of partial­thickness rota­ tor cuff tears at 10 to 43 months. Bursoscopy was performed in 71% of these patients, with bursal­sided partial­thickness tears found in 82%; these patients underwent ar­ throscopic acromioplasty as well as débridement, resulting in an average postoperative University of Califor­ nia, Los Angeles (UCLA) score of 33 (range, 0­35). For patients found to have a normal bursal side of the ro­ tator cuff, acromioplasty was not performed, resulting in an average UCLA score of 30.

Conclusions based on these stud­ ies are difficult to draw because most are small retrospective studies with short follow­ups and differing surgical techniques. Additionally, many studies combine data from pa­ tients with articular­sided and bursal­sided tears, traumatic and atraumatic tears, superficial fraying of the cuff and nearly full­thickness tears, and young overhead throwing athletes and older patients. From this group of reports it is difficult to define (1) the indications for surgery, (2) which aspects of the patients’ pa­ thologies were responsible for their symptoms, (3) why up to 50% of pa­ tients failed to achieve a satisfactory result, and (4) which aspect of the surgery (acromioplasty or débride­ ment) was responsible for improve­ ment after surgery.

In a 1999 study, Weber36 com­ pared the results at 2­ to 7­year

Treatment

follow­up of 65 patients with grade 3A or 3B partial­thickness rotator Treatment of partial­thickness rota­

cuff tears treated either with acro­ tor cuff tears varies according to the

mioplasty with mini­open repair or cause and location of the pathology.

with acromioplasty and arthroscop­

Partial­Thickness Rotator Cuff Tears

ic débridement. He reported a mean

Thus the lack of concordance with

Torpey et al39 demonstrated in a ca­ UCLA score of 31.6 for patients un­

the study of Park et al37 cannot be

daveric model that a 4­mm and a dergoing mini­open repair versus

explained by the longer follow­up in

6­mm acromioplasty, respectively, 22.7 for those undergoing arthro­

the work of Weber36 and Cordasco et

would be expected to release 43% scopic débridement. For the bursal­

al.38 Based on these data, Cordasco et

and 72% of the anterior and lateral sided tear subgroup, he reported a

al38 stated that they have begun to

surfaces of the deltoid tendon origin mean UCLA score of 33.0 for those

repair grade 2B partial­thickness

from the acromion; however, the treated with mini­open repair versus

tears.

clinical implications of this are un­ a UCLA score of 13.6 for those treat­

The literature thus suggests that

known. Nevertheless, acromioplas­ ed with arthroscopic débridement.

repair of partial­thickness tears of

ty, when done, should be performed Thus Weber concluded that débride­

the rotator cuff should be considered

judiciously, with removal of no more ment and acromioplasty was not ad­

in articular­sided tears with a depth

bone than is necessary to achieve a equate treatment of most grade 3

>6 mm and in bursal­sided tears with

flat acromion. tears, with bursal­sided tears faring

a depth >3 mm. Completion of the

Snyder40 originally described an especially poorly.36

tear, followed by repair in the pre­

arthroscopic transtendinous repair More recently, Park et al37 report­

ferred manner (ie, all­arthroscopic,

technique that does not involve ed 86% satisfactory results in 37 pa­

mini­open, or open), should be con­

completion of the tear. This tech­ tients with partial­thickness cuff

sidered when the tear is near full­

nique has the theoretic advantage of tears (24 articular­sided, 13 bursal­

thickness and the remaining tissue is

retaining the lateral portion of the sided) 2 years after arthroscopic acro­

thin and tenuous, or when there is a

original footprint of the rotator cuff mioplasty and débridement of

concomitant tear on the opposite

insertion and minimizing the articular­ and bursal­sided tears of

side of the rotator cuff. Alternatively,

length­tension mismatch of the re­ <50% thickness. They noted that

arthroscopic repair of the tear with­

paired rotator cuff. Excellent results bursal­sided tears fared significantly

out creation of a full­thickness tear

were recently reported by Ide et al,41 better with respect to pain score and

can be performed when the opposite

who used this technique to repair function (P < 0.05 for both) at 6

side of the cuff is intact and tissue is

grade 3A partial­thickness cuff tears. months, but that the groups were

of good quality.

Of 17 patients, 14 were rated as ex­

not significantly different at 1­year

The role of acromioplasty has not

cellent, 2 as good, and 1 as fair at an and 2­year follow­ups.

been clearly delineated. Multiple

average 39­month follow­up (range, Cordasco et al38 reported results

studies demonstrate good results

25 to 57 months). Mean UCLA and of arthroscopic subacromial decom­

with and without acromioplasty in

Japanese Orthopaedic Association pression and débridement of partial­

the treatment of partial­thickness

scores improved significantly (17.3 thickness tears that were in concor­

rotator cuff tears.34­38 We feel that, in

and 68.4 to 32.9 and 94.8, respective­ dance with those of Weber36 but

all patients undergoing shoulder ar­

ly; P < 0.01). differed from those of Park et al.37

throscopy for a suspected partial­

For articular­sided tears >1.5 cm They demonstrated that patients

thickness tear, a comprehensive bur­

in the anterior­to­posterior direc­ with bursal­sided tears have a much

sectomy should be performed to

tion, two bioabsorbable suture an­ higher rate of unsatisfactory out­

visualize completely the bursal side

chors double­loaded with no. 2 non­ comes than do patients with

of the rotator cuff, as well as to re­

absorbable polyester are used. For articular­sided tears at 2­ to 10­year

move any potential inflammatory

tears <1.5 cm, only one suture an­ follow­up (mean, 4.5 years). The au­

mediators. Consideration should be

chor is used. For tears >1.5 cm, the thors reported a 5% failure rate (2/

given to performing an acromioplas­

suture anchors are placed through 44) for grade 2A tears versus a 38%

ty when an extrinsic etiology of a

the rotator cuff at the lateral articu­ failure rate (3/8) for patients with

partial­thickness tear is suspected,

lar margin at the medial extent of grade 2B tears (P = 0.02). When grade

as evidenced by damage to the bursal

the footprint after localization of the 1A and 1B tears were included, the

side of the rotator cuff and fraying of

angle with a spinal needle. The spi­ overall treatment failure rate for

the coracoacromial ligament and/or

nal needle and suture anchors are bursal­sided tears was 29% (4/14),

anterior acromial osteophyte, or

passed just off the lateral aspect of whereas the failure rate in the

when there is evidence of impinge­

the acromion to obtain an angle articular­sided tears was 3% (2/63)

ment of the cuff on the undersurface

≤45°. From the subacromial space, (P = 0.008). The authors stated that

of the acromion under arthroscopic

one limb from each anchor is all of the patients in their study in

visualization.

grasped and pulled through the later­ whom the failure occurred garnered

The goal of acromioplasty is the

al portal, where it is tied over an in­ no benefits from surgery, and the

creation of a flat acromion that does

strument. Using the opposite ends of

failure was apparent immediately.38

not impinge on the rotator cuff.

these sutures, the knot is pulled into

720 Journal of the American Academy of Orthopaedic Surgeons

Andrew B. Wolff, MD, et al

Figure 5

1.5 cm anterior to posterior) A3B0 (Ellman grade A3) articular­sided partial­thickness rotator cuff tear, as seen through a standard posterior portal after arthroscopic transtendinous fixation using two suture anchors.

the subacromial space, cinching down the rotator cuff on top of the suture anchors, reestablishing the footprint. The opposite ends are then tied with static knots through a lat­ eral subacromial portal. This process is repeated with the remaining su­ ture limbs (Figure 5).

For articular­sided tears <1.5 cm in the anterior­to­posterior direc­ tion, the single double­loaded suture

Volume 14, Number 13, December 2006 721

Arthroscopic repair of a partial­thickness articular­sided tear. A, A shuttle relay is passed through the spinal needle and retrieved through the anterior portal. B, One end of the suture is also retrieved through the anterior portal. C, One end of the suture is engaged in the eyelet of the shuttle relay and then pulled back out of the healthy portion of the partially torn tendon. D, A complete repair of an articular­sided partial­thickness rotator cuff tear. (Adapted with permission from Ide J, Maeda S, Takagi K: Arthroscopic transtendon repair of partial­thickness articular­side tears of the rotator cuff: Anatomical and clinical study. Am J Sports Med 2005;33: 1672­ 1679.)

anchor is placed in the same manner as described above. However, be­ cause the entire suture is passing through the same puncture of the ro­ tator cuff, a tissue bridge must be created. This is done either by a shuttle suture passed through a spi­ nal needle or with a suture passer, such as the Suture Lasso (Arthrex, Naples, FL) or BirdBeak (Arthrex). In either manner, one limb of each su­

ture is passed posteromedially through different percutaneous por­ tals. Each suture is then retrieved and tied sequentially through the lateral subacromial portal (Figures 5 and 6).

For partial­thickness bursal­sided tears, bursectomy and acromioplas­ ty are performed. The frayed edges of the bursal flap are then gently débrided, and the tuberosity is exco­ riated to bleeding bone using a round burr. One or two bioabsorbable double­loaded suture anchors are placed through the full thickness of the rotator cuff (one anterior, one posterior) using a percutaneous su­ ture lasso (Figures 7 and 8).

Based on the available literature, we propose the following treatment algorithm for patients with degener­ ative partial­thickness rotator cuff tears who are not overhead throwing athletes (Figure 9). At presentation, nonsurgical management should be the mainstay of treatment in most cases. When progress is lacking with nonsurgical therapy, surgical inter­ vention can be considered as early as 3 months (but generally at 6 to 12 months) after initiation of nonsurgi­ cal management, depending on the patient’s desires and goals.

Figure 6

A large (

>

Partial­Thickness Rotator Cuff Tears

Surgical treatment is based on lo­ cation and depth of tear. Depth of tear can be determined either by ex­ amining the tear with a probe of a known length or, for articular­sided tears, by measuring the amount of bone lateral to the articular margin. A distance >7 mm represents a tear >50% (grade 3).3 Ellman grade 1 tears (<3 mm deep) of either the articular or bursal side, and grade 1 and 2 tears (3 to 6 mm deep) of the articular side, should be treated with débridement of the tear, with or without acro­ mioplasty. Although there is little ev­ idence to suggest that débridement of partial­thickness tears stimulates a healing response or improves the bio­ mechanics of the rotator cuff, it is vi­ tal in the assessment of the depth of the tear. Débridement also may re­ lieve mechanical irritation in the subacromial space and the glenohu­ meral joint. In addition, débridement likely removes inflammatory cells and inflammatory mediators present in the torn rotator cuff tissue.23

Grade 2 tears of the bursal side (>3 mm deep) and grade 3 tears of the ar­

722 Journal of the American Academy of Orthopaedic Surgeons

Figure 7

Placement of bioabsorbable double­loaded suture anchors. A, The lasso is passed through the full thickness of the cuff through a percutaneous portal, which optimizes the angle required for repair. This is often accomplished through the Neviaser portal, the superior medial portal bordered by the clavicle, the acromioclavicular joint, and the spine of the scapula. The nitinol wire is shuttled out of a cannula, along with the more medial of the suture limbs, and then is pulled back out of the percutaneous portal along with the suture limb, passing the suture through the full thickness of the cuff. The procedure is repeated for the posterior limb of the suture after again passing a lasso through the full thickness of the rotator cuff in a more posterior position. B, After passing both the anterior and posterior sutures through the bursal­sided tear, both sutures are tied securely over the bursal side of the rotator cuff, thereby restoring the bursal side of the cuff to its anatomic location.

ticular or bursal side (>6 mm deep) should be treated with repair, with or without acromioplasty. We prefer to complete articular­sided tears and perform double­arrow arthroscopic repairs because we have found long­ er convalescence and stiffness to oc­ cur with intratendinous repairs. In contrast, we arthroscopically repair bursal­sided tears without comple­ tion of the tear, using the articular footprint as an internal splint.

Partial­Thickness Rotator Cuff Tears in Overhead Throwing Athletes

Overhead throwing athletes repre­ sent a distinct population of individ­ uals in whom partial­thickness tears of the rotator cuff develop. The etiol­ ogy and pathogenesis of these injuries often differ from those in the non– overhead throwing population. Ac­ cordingly, the treatment of this sub­ group can be very different as well, particularly when return to throwing is the goal. Partial­thickness tears in

athletes are seen mostly in overhead throwing sports. Onset is usually in­ sidious, with symptoms of pain at rest, loss of velocity, and “popping” or “catching” while throwing.

Most partial­thickness tears are articular­sided tears of the dominant arm that begin at the supraspinatus­ infraspinatus interval, posterior to the location of tears commonly seen in the older population. Partial­ thickness tears are often associ­ ated with other shoulder pathology. Superior labral injury, including pos­ terosuperior labral injury and de­ tachment, posteroinferior capsular contracture, anterior capsular atten­ uation, and internal rotation deficit, are frequently associated with articular­sided tears.42

The cause of partial­thickness ro­ tator cuff tears in the overhead throwing athlete is a subject of con­ siderable debate. Andrews et al43 originally theorized that these articular­sided tears resulted from repetitive trauma of the massive ec­ centric traction force to the su­ praspinatus and infraspinatus ten­ dons, which keep the humeral head in the correct position from the de­ celeration phase through the release phase of throwing. Davidson et al44 espoused a theory of internal im­ pingement from repetitive mi­ crotrauma of the rotator cuff result­

Figure 8

1.5 A small (

<

cm anterior to posterior) bursal­sided partial­thickness rotator cuff tear after repair with a single suture

anchor.

Andrew B. Wolff, MD, et al

ing from anterior subluxation secondary to minor instability and

Figure 9

fatigue of the dynamic stabilizers, which results in a secondary im­

Partial­thickness rotator cuff tear pingement of the rotator cuff on the posterior edge of the glenoid. Citing the association with posterior type 2

Nonsurgical management superior labrum anterior­posterior

(minimum of 3 months) (SLAP) tears via a peel­back mecha­ nism, Burkhart et al42 have theorized that a combination of repetitive ten­ sile loading of the cuff, and the sub­

Symptom improvement No improvement

sequent partial­thickness tear, may be caused by posterosuperior sublux­ ation of the humerus from posterior

Diagnostic arthroscopy

capsule contracture and by repeti­ tive torsional and shear overload generated by hyperexternal rotation

Bursal­sided tear Articular­sided tear

in throwing athletes.

Treatment of partial­thickness cuff tears in the overhead throwing

<3 mm deep >3 mm deep athlete should begin with nonsurgi­ cal measures. All of the following may be used to effect resolution of

Débridement ± acromioplasty symptoms: stretching of the posteri­ or capsular tissues to address loss of internal rotation, rotator cuff strengthening programs (including eccentric and plyometric exercises), trunk and lower extremity strength­ ening exercises,

restoration of prop­ er throwing mechanics, nonsteroidal anti­inflammatory drugs, and corti­ costeroid injections.

Surgical treatment is reserved for patients who fail to respond to non­ surgical management. Diagnostic ar­ throscopy often reveals concomitant pathology that may be the primary cause of the patient’s symptoms. Débridement of the articular­sided tear and treatment of concomitant capsular or labral pathology should be the mainstay of surgical interven­ tion; repair of the partial­thickness tear or acromioplasty is rarely indi­ cated.

Summary

Partial­thickness rotator cuff tears are not a single entity but rather rep­ resent a spectrum of disease states affecting the rotator cuff. From asymptomatic to significantly dis­

Volume 14, Number 13, December 2006 723

<6 mm deep >6 mm deep

Repair ±

Débridement;

Repair; acromioplasty

rarely acromioplasty

rarely acromioplasty

Proposed treatment algorithm for partial­thickness rotator cuff tears.

Additional Resources

DVD/video: Surgical Techniques in Orthopaedics: “Arthroscopic Rota­ tor Cuff Repair”: http://www4.aaos.org/product/productpage.cfm?code =02640

Related clinical topics articles available on Orthopaedic Knowledge Online: “Glenohumeral Arthritis and the Rotator Cuff

Deficient Shoul­ der,” by Gregory P. Nicholson, MD, and Guido Marra, MD: http:// www5.aaos.org/oko/shoulder_elbow/ga_rcdeficient/pathophysiology/ pathophysiology.cfm

“Shoulder Arthroscopy,” by Stephen J. Snyder, MD, and Petra Waldherr, MD: http://www5.aaos.org/oko/shoulder_elbow/arthroscopy/anatomy/ anatomy.cfm

“Rotator Cuff Tears,” by Evan Flatow, MD, and Leesa Galatz, MD: http:// www5.aaos.org/oko/shoulder_elbow/rotator_cuff/pathophysiology/ pathophysiology.cfm

Patient information: Free information for your patients about rotator cuff tears and treatment is available on Your Orthopaedic Connection at http://www.orthoinfo.com

Partial­Thickness Rotator Cuff Tears

abling, partial­thickness tears of the

1. Clark JM, Harryman DT II: Tendons,

Contemp Orthop 1995;31:262­271. rotator cuff can affect patients in dif­ ferent ways. Young overhead athletes presenting with partial­thickness tears differ with respect to etiology,

ligaments, and capsule of the rotator

16. Halder AM, O’Driscoll SW, Heers G, cuff: Gross and microscopic anatomy.

et al: Biomechanical comparison of ef­ J Bone Joint Surg Am 1992;74:713­

fects of supraspinatus tendon detach­ 725.

ments, tendon defects, and muscle re­ 2. Ellman H: Diagnosis and treatment of

tractions. J Bone Joint Surg Am 2002; goals, and treatment from older pa­

incomplete rotator cuff tears. Clin

84:780­785. tients with degenerative tears.

Pathogenesis of degenerative tears is multifactorial. Articular­ sided tears are more likely caused by

Orthop Relat Res 1990;254:64­74.

17. Bey MJ, Ramsey ML, Soslowsky LJ: 3. Ruotolo C, Fow JE, Nottage WM: The

Intratendinous strain fields of the su­ supraspinatus footprint: An anatomic

praspinatus tendon: Effect of a surgi­ study of the supraspinatus insertion.

cally created articular­surface rotator Arthroscopy 2004;20:246­249.

cuff tear. J Shoulder Elbow Surg primarily intrinsic factors, while

4. Codman EA: The Shoulder: Rupture

2002;11:562­569. both intrinsic and extrinsic factors

of the Supraspinatus Tendon and

18. Gotoh M, Hamada K, Yamakawa H, may play roles in the development of bursal­sided tears. Diagnosis of clin­ ically significant partial­thickness

Other Lesions in or About the Sub­ acromial Bursa. Boston, MA: Thomas Todd, 1934. 5. Fukuda H, Mikasa M, Yamanaka K:

Inoue A, Fukuda H: Increased sub­ stance P in subacromial bursa and shoulder pain in rotator cuff diseases. J Orthop Res 1998;16:618­621. tears should be based on the pa­

Incomplete thickness rotator cuff

19. Gotoh M, Hamada K, Yamakawa H, et tient’s symptoms and clinical find­ ings in conjunction with magnetic resonance arthrography.

tears diagnosed by subacromial bur­ sography. Clin Orthop Relat Res 1987;223:51­58. 6. Fukuda H: Partial­thickness rotator

al: Interleukin­1­induced subacromi­ al synovitis and shoulder pain in rota­ tor cuff diseases. Rheumatology (Oxford) 2001;40:995­1001. Treatment should be based on the

cuff tears: A modern view on Cod­

20. Gotoh M, Hamada K,

Yamakawa H, et patient’s goals, etiology, and depth of tear. Considerable thought should be given to differentiating between articular­ and bursal­sided tears.

man’s classic. J Shoulder Elbow Surg

al: Interleukin­1­induced glenohu­ 2000;9:163­168.

meral synovitis and shoulder pain in 7. Sher JS, Uribe JW, Posada A, Murphy

rotator cuff diseases. J Orthop Res BJ, Zlatkin MB: Abnormal findings on

2002;20:1365­1371. magnetic resonance images of asymp­

21. Yamanaka K, Matsumoto T: The joint Nonsurgical treatment is successful

tomatic shoulders. J Bone Joint Surg

side tear of the rotator cuff: A fol­ in most patients. Surgical treatment, when necessary, should consist of tear débridement in articular­sided tears <6 mm in depth and bursal­

Am 1995;77:10­15.

lowup study by arthrography. Clin 8. Milgrom C, Schaffler M, Gilbert S,

Orthop Relat Res 1994;304:68­73. van Holsbeeck M: Rotator­cuff chang­

22. Hamada K, Tomonaga A, Gotoh M, es in asymptomatic adults: The effect

Yamakawa H, Fukuda H: Intrinsic of age, hand dominance and gender.

healing capacity and tearing process sided tears <3 mm in depth. Consid­

J Bone Joint Surg Br 1995;77:296­298.

of torn supraspinatus tendons: In situ eration should be given to repairing

9. Rothman RH, Parke WW: The vascu­

hybridization study of α1(I) procol­ tears >6 mm in depth on the articu­ lar side and >3 mm of depth on the bursal side. The role of acromioplas­

lar anatomy of the rotator cuff. Clin

lagen mRNA. J Orthop Res 1997;15: Orthop Relat Res 1965;41:176­186.

24­32. 10. Rathbun JB, Macnab I: The microvas­

23. Gotoh M, Hamada K, Yamakawa H, cular pattern of the rotator cuff.

Tomonaga A, Inoue A, Fukuda H: Sig­ ty has not been clearly delineated,

J Bone Joint Surg Br 1970;52:540­553.

nificance of granulation tissue in torn but it should be considered when there is evidence of extrinsic causa­ tion of the tear. To optimize future treatment of patients with partial­

11. Lohr JF, Uhthoff HK: The microvascu­ lar pattern of the supraspinatus ten­ don. Clin Orthop Relat Res 1990; 254:35­38. 12. Nakajima T, Rokuuma N, Hamada K,

supraspinatus insertions: An immu­ nohistochemical study with antibod­ ies against interleukin­1β, cathepsin D, and matrix metalloprotease­1. J Orthop Res 1997;15:33­39. thickness tears of the rotator cuff,

Tomatsu T, Fukuda H: Histologic and

24. Gartsman GM, Milne JC: Articular further research at the basic science

biomechanical characteristics of the

surface partial­thickness rotator cuff and clinical levels is needed.

supraspinatus tendon: Reference to

tears. J Shoulder Elbow Surg 1995;4: rotator cuff tearing. J Shoulder Elbow

409­415.

References

Surg 1994;3:79­87.

25. Itoi E, Tabata S: Incomplete rotator 13. Neer CS II: Anterior acromioplasty for

cuff tears: Results of operative treat­

Evidence­based Medicine: Level I/II prospective studies on partial­

the chronic impingement syndrome

ment. Clin Orthop Relat Res 1992; in the shoulder: A preliminary report.

284:128­135. J Bone Joint Surg Am 1972;54:41­50.

26. Wiener SN, Seitz WH Jr: Sonography thickness rotator cuff tears: referenc­

14. Ozaki J, Fujimoto S, Nakagawa Y,

of the shoulder in patients with tears es 32 and 36. The remaining refer­ ences are case­control case series or expert opinion studies.

Masuhara K, Tamai S: Tears of the ro­

of the rotator cuff: Accuracy and value tator cuff of the shoulder associated

for selecting surgical options. AJR with pathological changes in the acro­

Am J Roentgenol 1993;160:103­107. mion: A study in cadavera. J Bone

27. Brenneke SL, Morgan CJ: Evaluation

Citation numbers printed in bold type indicate references published

Joint Surg Am 1988;70:1224­1230. 15. Burkhead WZ Jr, Burkhart SS, Gerber C, et al: Symposium: The rotator cuff.

of ultrasonography as a diagnostic technique in the assessment of rotator cuff tendon tears. Am J Sports Med within the past 5 years.

Debridement versus repair: Part I.

1992;20:287­289.

724 Journal of the American Academy of Orthopaedic Surgeons