O

F

Aa

b

c

d

AA

KSAT

1

aoapia

T

1

Orthopaedics & Traumatology: Surgery & Research 100S (2014) S365–S369

Available online at

ScienceDirectwww.sciencedirect.com

riginal article

ailed subacromial decompression. Risk factors

. Bouchardb,∗, J. Garreta, L. Favardc, H. Charlesd, D. Ollatb

Clinique du Parc, 69000 Lyon, FranceHôpital d’Instruction des Armées Bégin Saint-Mandé, 69, avenue de Paris, 94160 Saint-Mandé, FranceHôpital Trousseau, CHU de Tours, 37000 Tours, France13, place Philippe Lebon, 59000 Lille, France

a r t i c l e i n f o

rticle history:ccepted 17 September 2014

eywords:ubacromial impingement syndromerthroscopic subacromial decompressionreatment failure

a b s t r a c t

Background: Arthroscopic subacromial decompression (acromioplasty) is widely held to be effective,although pain may persist after the procedure. The objective of this study was to evaluate the proportionof patients with residual pain (i.e., the failure rate) after isolated subacromial decompression and to lookfor predictors of failure.Material and method: We conducted a retrospective multicentre study of 108 patients managed withisolated arthroscopic subacromial decompression between 2007 and 2011, for any reason. We excludedpatients in whom surgical procedures on the rotator cuff tendons were performed concomitantly. Datawere collected from the medical records, a telephone questionnaire, and radiographs obtained beforesurgery and at last follow-up. Failure was defined as persistent pain (visual analogue scale score > 3)more than 6 months after surgery and at last follow-up.Results: The failure rate was 29% (31/108). Two factors significantly predicted failure, namely, receiv-ing workers’ compensation benefits for the shoulder condition and co-planing. Heterogeneous calcifictendinopathy and deep partial-thickness rotator cuff tears were also associated with poorer outcomes,but the effect was not statistically significant.Discussion: Co-planing may predict failure of subacromial decompression, although whether this effectis due to an insufficient degree of co-planing or to the technique itself is unclear. Nevertheless, in patientswith symptoms from the acromio-clavicular joint, acromio-clavicular resection is probably the bestoption. Receiving workers’ compensation benefits was also associated with treatment failure, as a resultof well-known parameters related to the social welfare system.

Conclusion: Isolated arthroscopic subacromial decompression is effective in 70% of cases. We recommendthe utmost caution if co-planing is considered and/or the patient receives workers’ compensation benefitsfor the shoulder condition, as these two factors are associated with a significant increase in the failurerate.Level of evidence: IV (retrospective study).

© 2014 Published by Elsevier Masson SAS.

. Introduction

Subacromial impingement syndrome is a common cause ofnterior shoulder pain. In 1972, Neer was the first to describepen acromioplasty and its outcomes [1]. In 1983, Ellman reportedn arthroscopic method for performing acromioplasty [2]. The

rinciple was to achieve subacromial decompression by remov-

ng the bursa, resecting the undersurface of the anterior acromion,nd severing the coraco-acromial ligament. Today, acromioplasty

∗ Corresponding author. 69, avenue de Paris, 94160 Saint-Mandé, France.el.: +33 6 99 07 23 99.

E-mail address: antoine.bouch@gmail.com (A. Bouchard).

http://dx.doi.org/10.1016/j.otsr.2014.09.006877-0568/© 2014 Published by Elsevier Masson SAS.

is almost consistently performed as an arthroscopic procedureand remains widely used, although perhaps less often in isola-tion. Published studies of outcomes after isolated acromioplastyhave produced fairly consistent results, with success rates of 77%to 90% [3,4]. Nevertheless, the finding in some studies that over25% of patients experiencing residual pain may call into ques-tion the effectiveness of isolated acromioplasty in ensuring painrelief. Studies have shown that risk factors for failed acromio-plasty include inappropriate patient selection and technical errors[5–7].

We performed a retrospective study with the dual objective ofdetermining whether isolated arthroscopic acromioplasty is effec-tive in ensuring pain relief and of identifying factors that predictfailure.

dx.doi.org/10.1016/j.otsr.2014.09.006http://www.sciencedirect.com/science/journal/18770568http://crossmark.crossref.org/dialog/?doi=10.1016/j.otsr.2014.09.006&domain=pdfmailto:antoine.bouch@gmail.comdx.doi.org/10.1016/j.otsr.2014.09.006

S366 A. Bouchard et al. / Orthopaedics & Traumatology: Surgery & Research 100S (2014) S365–S369

efore

2

fftaPccodt

Fig. 1. A. Antero-posterior radiograph of the right shoulder b

. Material and methods

We conducted a retrospective multicentre (four centres) studyor a symposium held by the French Society for Arthroscopy (Sociétéranç aise d’arthroscopie) in 2013. Investigators in the four cen-res reviewed the data from patients who underwent isolatedrthroscopic acromioplasty between 2007 and 2011 for any reason.atients in whom acromio-clavicular co-planing was performedoncomitantly were included but those who underwent acromio-

lavicular resection, a procedure on the long head of biceps tendon,r a procedure on the rotator cuff tendons were excluded. Theata sources were the medical records, surgical report, responseso a questionnaire administered during a telephone interview, and

Fig. 2. A. Antero-posterior radiograph of the right shoulder after s

surgery. B. Lateral view of the same shoulder before surgery.

radiographs obtained before surgery and at last follow-up. Pain wasassessed using a Visual Analogue Scale (VAS). The radiographs wereused to determine the acromio-humeral interval, thickness of theacromion, and shape of the acromion according to Bigliani and toPark (Figs. 1–4).

Failure of the acromioplasty procedure was defined as a VASpain score greater than 3/10 6 months after the procedure and atlast follow-up.

Pain intensity was the primary outcome measure for the statisti-

cal analysis. Fisher’s exact test was used to compare groups. Valuesof P ≤ 0.05 were considered statistically significant. We identifiedfactors predicting failure by computing the odds ratios (ORs) withtheir 95% confidence intervals (95% CIs).

urgery. B. Lateral view of the same shoulder before surgery.

A. Bouchard et al. / Orthopaedics & Traumatology: Surgery & Research 100S (2014) S365–S369 S367

Fig. 3. Bigliani classification of acromial shape. 1, flat; 2, curved; 3, hooked.

3

m(2aot7dr

Table 1Effect of age, gender, and workers’ compensation status on the acromioplasty failurerate.

% Failures Odds ratio P value

Age 0.11< 50 years 29 150–65 years 35 0.76> 65 years 11 3.4

Gender 0.58Male 25 1Female 41 1.3

Workers’ compensation < 0.001No 16 1Yes 69 11.9

Table 2Outcomes according to the reason for acromioplasty.

Reason for acromioplasty No of patients % Failures P value

Full-thickness distal or intermediate tear 6 0 NSIrreparable tear 6 17Superficial partial-thickness (< 50%) tear 18 17Tendinopathy with no tear

Isolated subacromial bursitis62 29

Calcific tendinopathy type C 8 37

by 72% of patients after 6 months versus only 17% at last follow-up;

TO

Fig. 4. Park classification of acromial shape.

. Results

During the study period, 108 patients (one shoulder per patient)et the study inclusion criteria. Among them, 60% were women

male/female ratio, 0.4). Mean age at surgery was 52 years (range,7–76 years). The dominant shoulder was involved in 63% of casesnd the shoulder injury was recognised as a work-related injuryr occupational disorder in 24% of patients. All patients had failedo respond to medical treatment given for longer than 6 months,0% had received a preoperative subacromial test injection of a

elayed-action corticosteroid, and 75% had received preoperativeehabilitation therapy.

able 3utcomes according to the surgical report data.

% of patients

Complete division of the coraco-acromial ligamentNo 18 Yes 82

Acromial release to the lateral deltoid muscle fibresNo 4 Yes 96

Co-planingNo 78 Yes 22

Deep partial-thickness (< 50%) tear 8 75

NS: non-significant.

The acromioplasty procedure was performed on a day-case basisin 1 patient. The remaining 107 patients spent a mean of 2.5 nightsin the hospital.

Mean time from surgery to last follow-up was 45 months (range,14–82 months). The failure rate was 29% (31/108 patients). The31 patients with failed acromioplasty had a mean VAS pain scoreof 6 and a mean Subjective Shoulder Value (SSV) of 57, althoughonly 25% of them used analgesics daily. Of these 31 patients, 52%reported pain at the same site as before surgery. Anterior sub-acromial pain was reported by 68% of patients and posterior andcervical pain non-specific for subacromial impingement by half thepatients.

Neither age nor gender was significantly associated with failureby univariate analysis. In contrast, receiving workers’ compensa-tion benefits because of the shoulder disorder was very significantlyassociated with failure (OR, 11.9; P < 0.001) (Table 1).

No statistically significant associations were demonstratedbetween the reason for acromioplasty and failure of the proce-dure (Table 2). The most common reason for acromioplasty wasisolated subacromial bursitis with rotator cuff tendinopathy butno tears (69% of patients). This diagnosis was associated with a29% failure rate. A superficial partial-thickness (< 50%) tear in thesupraspinatus was associated with good outcomes, although theimprovement was slow to develop (persistent pain was reported

thus, the failure rate was 17%). In the group with partial-thickness(< 50%) tears of the deep aspect of the rotator cuff, the failure rate

% Failures Odds ratio P value

10 133 4.42

25 129 1.22

18 1 < 0.00167 9.2

S368 A. Bouchard et al. / Orthopaedics & Traumatology:

Table 4Outcomes according to the shape of the acromion.

% of patients % Failures Odds ratio P value

Bigliani1 13 14 1 NS2 48 35 0.323 39 26 0.48

Park1 8 0 NS2 74 31

N

wr7c

pr

ltva

ap

4

tfOaftttrspKpm

cesiTes[gasiibps[

p

3 18 33

S: non-significant.

as 75%, and 3 patients in this group required revision surgicalepair. A preoperative subacromial test injection was performed in0% of patients. Failure to respond to the injection was not signifi-antly associated with failure of the acromioplasty procedure.

Co-planing performed concomitantly with the acromioplastyrocedure was a highly significant predictor of failure (67% failureate; OR, 9.2; i < 0.001) (Table 3).

The comparison of radiographs obtained preoperatively and atast follow-up showed a mean 20% decrease in anterior acromialhickness and a mean 20% increase in the acromio-humeral inter-al. The degree of anterior acromial resection was not significantlyssociated with failure of the acromioplasty procedure (Table 4).

The shape of the acromion assessed according to Bigliani (Fig. 3)nd/or Park (Fig. 4) was not significantly associated with acromio-lasty failure.

. Discussion

The 29% acromioplasty failure rate in our population is higherhan in earlier studies, which demonstrated success rates rangingrom 77% to 90% after isolated arthroscopic acromioplasty [3,4].ur study has several limitations. The design was retrospective,nd the numbers of patients varied widely according to the reasonor acromioplasty (tendinopathy with no rotator cuff tear, partial-hickness tears of the deep or superficial cuff, irreparable tears, andype C calcifications). In addition, we used a more stringent defini-ion of failure compared to other studies. For example, in a similaretrospective study of 105 patients, Klintberg et al. obtained an 88%uccess rate but specified that only half the patients were com-letely free of pain [4]. Using pain as the primary outcome measure,etola et al. found no significant difference after 2 years betweenatients treated with arthroscopic acromioplasty and those treatededically [8], thus challenging the effectiveness of acromioplasty.Many factors acting at various steps of the acromioplasty pro-

edure can cause treatment failure. At the diagnostic stage, strongvidence must be obtained that subacromial impingement is theource of the pain. However, accurately diagnosing subacromialmpingement syndrome is difficult even for experienced clinicians.he diagnosis rests on a converging set of findings from the clinicalxamination and diagnostic investigations, and appropriate patientelection exerts a major influence on the outcomes. Magaji et al.9] reported that the outcomes of acromioplasty were best in theroup that exhibited the following four criteria: pain in the mid-rc of abduction, consistently positive Hawkins test, response to aubacromial corticosteroid injection, and radiological evidence ofmpingement. Outcomes were less favourable in the groups meet-ng only two or three of these criteria [9]. We found no associationetween the result of the corticosteroid injection test and acromio-lasty failure. However, several prospective studies consistently

howed that this test is an excellent diagnostic and prognostic tool10,11].

The second cause of failure lies in the indication for the acromio-lasty procedure. Dopirak and Ryu reported poorer outcomes in

Surgery & Research 100S (2014) S365–S369

patients with shoulder stiffness or osteoarthritis and in those withuntreated symptomatic acromio-clavicular arthropathy or unfusedacromion [5]. They also reported that one reason for failure was thepresence of an undiagnosed concomitant regional disorder as thesource of pain, such as tendinopathy of the long head of biceps ten-don or a tear in the upper third of the subscapularis. Lashgari et al.confirmed that poor patient selection was among the causes of fail-ure and added neck pain and peripheral neurological disorders asalternative sources of pain [6]. We identified several other causesof failure, most notably receiving workers’ compensation benefitsfor the shoulder disorder, which was associated with a 70% failurerate (18 of 26 patients). Arcand et al. reported similar results [12].Furthermore, we found that acromioplasty in patients with super-ficial partial-thickness tears produced good results but that thesewere slow to develop. In contrast, deep partial-thickness tears wereassociated with a 75% failure rate and a need in some patients forrevision repair surgery. This finding has not been clearly reported inearlier studies, which failed to separate deep and superficial partial-thickness tears. Another condition associated with a high failurerate was type C heterogeneous calcific tendinopathy, a finding thathas not been reported previously.

The third cause of failure is related to the operative technique.Co-planing performed concomitantly with acromioplasty was asso-ciated with a 67% failure rate in our study and with an odds ratio forfailure of 9.2. Previously reported data on the effects of co-planingare conflicting. Cadaveric and biomechanical studies show that theinferior incision performed to open the acromio-clavicular joint forco-planing destabilises the joint [13,14]. Based on a clinical study,Barber [15] advocated co-planing, as no differences in the clavic-ular symptoms were apparent between the groups managed withremoval of inferior clavicle osteophytes, partial resection of the dis-tal clavicle to the level of the acromion, or complete resection of theacromio-clavicular joint. In contrast, Fischer et al. reported a 39%rate of secondary acromio-clavicular symptoms when co-planingwas performed concomitantly with acromioplasty and advocatedan all-or-none approach consisting in either no procedure on theacromio-clavicular joint and distal clavicle or complete resection[16]. Kharrazi et al. suggested a more subtle strategy based ona retrospective study of 1482 patients divided into two groupsbased on whether surgery consisted in isolated acromioplasty orin acromioplasty with co-planing or acromio-clavicular resection.The rate of re-operation for acromio-clavicular symptoms was 1.5%in both groups. The authors concluded that co-planing or acromio-clavicular joint resection should not be performed routinely inpatients without acromio-clavicular joint symptoms [17].

Finally, excessive or inadequate acromial resection may beamong the causes of failed acromioplasty. Inadequate resectionresults in persistent impingement and excessive resection in acro-mial fractures [5]. These two extremes are rare in clinical practice.No clear recommendations are available regarding the optimalthickness of bone to be removed or the posterior extent of theresection. Our results are in agreement with a study by Soyer et al.,showing no association between the amount of acromial resectionand the outcomes [7].

5. Conclusion

Isolated acromioplasty is effective in 75% of cases. Publisheddata on this point are fairly consistent. However, acromioplastyshould not be viewed as a default procedure: instead, the patients

must be selected based on valid and well-defined criteria. Althoughwe found no significant differences across reasons for acromio-plasty, the outcomes seem poorer in patients with heterogeneouscalcific tendinopathy or deep partial-thickness rotator cuff tears.

logy:

frc

D

c

R

[

[

[

[

[

[

[

A. Bouchard et al. / Orthopaedics & Traumato

We recommend the utmost caution when co-planing is per-ormed concomitantly with acromioplasty and when the patienteceives workers’ compensation, since these two factors are asso-iated with significantly higher failure rates.

isclosure of interest

The authors declare that they have no conflicts of interest con-erning this article.

eferences

[1] Neer CS. Anterior acromioplasty for the chronic impingement syndrome in theshoulder: a preliminary report. J Bone Joint Surg Am 1972;54:41–50.

[2] Ellman H. Arthroscopic subacromial decompression: analysis of 1-to 3-yearresults. Arthroscopy 1987;3:173–81.

[3] Odenbring S, Wagner P, Atroshi I. Long-term outcomes of arthroscopic acromio-plasty for chronic shoulder impingement syndrome: a prospective cohort studywith a minimum of 12 years’ follow-up. Arthroscopy 2008;24(10):1092–8.

[4] Klintberg IH, Karlsson J, Svantesson U. Health-related quality of life, patientsatisfaction, and physical activity 8–11 years after arthroscopic subacromialdecompression. J Shoulder Elbow Surg 2011;20(4):598–608.

[5] Dopirak R, Ryu RK. Management of the failed arthroscopic subacromialdecompression: causation and treatment. Sports Med Arthrosc 2010;18(3):207–12.

[6] Lashgari CJ, Galatz L, Purcell DB. The failed subacromial decompression: furtherevaluation and treatment. Oper Tech Orthop 2003;13(4):252–9.

[

Surgery & Research 100S (2014) S365–S369 S369

[7] Soyer J, Vaz S, Pries P, Clarac JP. The relationship between clinical outcomes andthe amount of arthroscopic acromial resection. Arthroscopy 2003;19(1):34–9.

[8] Ketola S, Lehtinen J, Arnala I, Nissinen M, Westenius H, Sintonen H, et al. Doesarthroscopic acromioplasty provide any additional value in the treatment ofshoulder impingement syndrome? A two-year randomised controlled trial. JBone Joint Surg Br 2009;91(10):1326–34.

[9] Magaji SA, Singh HP, Pandey RK. Arthroscopic subacromial decompression iseffective in selected patients with shoulder impingement syndrome. J BoneJoint Surg Br 2012;94(8):1086–9.

10] Lim JT, Acornley A, Dodenhoff RM. Recovery after arthroscopic subacromialdecompression: prognostic value of the subacromial injection test. Arthroscopy2005;21(6):680–3.

11] Mair SD, Viola RW, Gill TJ, Briggs KK, Hawkins RJ. Can the impingement testpredict outcome after arthroscopic subacromial decompression? J ShoulderElbow Surg 2004;13(2):150–3.

12] Arcand MA, O’Rourke P, Zeman CA, Burkhead Jr WZ. Revision surgery after failedsubacromial decompression. Int Orthop 2000;24(2):61–4.

13] Deshmukh AV, Perlmutter GS, Zilberfarb JL, Wilson DR. Effect of subacromialdecompression on laxity of the acromioclavicular joint: biomechanical testingin a cadaveric model. J Shoulder Elbow Surg 2004;13(3):338–43.

14] Edwards SG. Acromioclavicular stability: a biomechanical comparison ofacromioplasty to acromioplasty with coplaning of the distal clavicle.Arthroscopy 2003;19(10):1079–84.

15] Barber FA. Long-term results of acromioclavicular joint coplaning. Arthroscopy2006;22(2):125–9.

16] Fischer BW, Gross RM, McCarthy JA, Arroyo JS. Incidence of acromioclav-

icular joint complications after arthroscopic subacromial decompression.Arthroscopy 1999;15(3):241–8.

17] Kharrazi FD, Busfield BT, Khorshad DS. Acromioclavicular joint reoperationafter arthroscopic subacromial decompression with and without concomitantacromioclavicular surgery. Arthroscopy 2007;23(8):804–8.

http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0005http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0010http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0015http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0020http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0025http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0030http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0035http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0040http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0045http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0050http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0055http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0060http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0065http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0070http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0075http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0080http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085http://refhub.elsevier.com/S1877-0568(14)00242-4/sbref0085

Failed subacromial decompression. Risk factors1 Introduction2 Material and methods3 Results4 Discussion5 ConclusionDisclosure of interestReferences