(PDF) Humeral Shaft Fractures in Adults – Effectiveness of Surgery versus Functional Bracing

Helsinki University Hospital

Department of Orthopaedics and Traumatology

FICEBO – Finnish Centre for Evidence-Based Orthopaedics

Faculty of Medicine

Doctoral Programme in Clinical Research

University of Helsinki

HUMERAL SHAFT FRACTURES IN ADULTS –

EFFECTIVENESS OF SURGERY VERSUS

FUNCTIONAL BRACING

Lasse Rämö

DOCTORAL DISSERTATION

To be presented for public discussion

with the permission of the Faculty of Medicine of the University of Helsinki,

in Auditorium 1 of Töölö Hospital,

on 13 August 2021 at 12 noon.

HELSINKI 2021

Supervised by

Docent Mika Paavola, MD, PhD

Division of Musculoskeletal Surgery

Helsinki University Hospital, Helsinki, Finland

University of Helsinki, Finland

Docent Simo Taimela, MD, PhD

FICEBO – Finnish Centre for Evidence-Based Orthopaedics

Division of Musculoskeletal Surgery

Helsinki University Hospital, Helsinki, Finland

University of Helsinki, Finland

Reviewed by

Docent Antti Eskelinen, MD, PhD

Coxa Hospital for Joint Replacement, Tampere, Finland

University of Tampere, Finland

Docent Inari Laaksonen, MD, PhD

Turku University Hospital, Turku, Finland

University of Turku, Finland

Opponent

Docent Petri Virolainen, MD, PhD

Turku University Hospital, Turku, Finland

University of Turku, Finland

The Faculty of Medicine uses the Urkund system (plagiarism recognition) to

examine all doctoral dissertations.

ISBN 978-951-51-7374-4 (print) Unigrafia

ISBN 978-951-51-7375-1 (online) Helsinki 2021

ISSN 2342-3161 (print)

ISSN 2342-317X (online)

http://ethesis.helsinki.fi

To my family

CONTENTS

ORIGINAL PUBLICATIONS ............................................................................. 3

ABBREVIATIONS ............................................................................................. 4

ABSTRACT ......................................................................................................... 5

TIIVISTELMÄ .................................................................................................... 7

1 INTRODUCTION ....................................................................................... 9

2 REVIEW OF THE LITERATURE ............................................................. 11

2.1 Surgical anatomy and exposures of the humeral shaft ..................... 11

2.2 Epidemiology of humeral shaft fractures and associated injuries .. 15

2.3 Humeral shaft fracture classification ............................................... 16

2.4 Nonsurgical treatment of humeral shaft fractures .......................... 18

2.4.1 Hanging cast .............................................................................. 18

2.4.2 Coaptation splint ....................................................................... 18

2.4.3 Others ........................................................................................ 18

2.4.4 Functional bracing .................................................................... 19

2.5 Surgical treatment of humeral shaft fractures ................................. 20

2.5.1 Open reduction and plate osteosynthesis ................................. 20

2.5.2 Intramedullary nailing .............................................................. 22

2.5.3 Minimally invasive plate osteosynthesis .................................. 22

2.5.4 External fixation ........................................................................ 23

2.6 Outcome measures in clinical studies of humeral shaft fractures ... 24

2.7 Adverse events in humeral shaft fracture treatment ....................... 27

2.7.1 Nonsurgical treatment .............................................................. 27

2.7.2 Surgical treatment ..................................................................... 29

2.8 Surgery versus nonsurgical treatment in humeral shaft fracture

studies ............................................................................................... 30

3 RESEARCH QUESTIONS ........................................................................ 34

4 PATIENTS AND METHODS ................................................................... 35

4.1 Patients ............................................................................................. 35

4.1.1 RCT comparing the effectiveness of surgical and nonsurgical

treatments of humeral shaft fractures in adults (I, II) ............. 35

4.1.2 Two-year follow-up of the FISH trial comparing patients with

secondary surgery with patients with successful fracture

healing (III) ............................................................................... 38

4.2 Interventions ..................................................................................... 39

4.2.1 Surgical treatment (II, III) ........................................................ 39

4.2.2 Nonsurgical treatment, functional bracing (II, III) .................. 39

4.2.3 Rehabilitation (II, III) .............................................................. 40

4.3 Follow-up ......................................................................................... 40

4.4 Outcome measures (II, III) ............................................................... 43

4.4.1 Disabilities of the Arm, Shoulder, and Hand score .................. 43

4.4.2 Pain at rest and on activity ........................................................ 43

4.4.3 Constant-Murley score .............................................................. 43

4.4.4 15D quality-of-life tool ............................................................... 43

4.4.5 Elbow range of motion .............................................................. 44

4.4.6 DASH work and sports or performing arts modules ................ 44

4.4.7 General satisfaction ................................................................... 44

4.4.8 Patient acceptable symptom state (PASS) ................................ 44

4.4.9 Clinical recovery ........................................................................ 45

4.5 Ethics ................................................................................................. 45

4.6 Statistical methods ............................................................................ 45

5 RESULTS ................................................................................................. 48

5.1 One-year follow-up (II) ..................................................................... 48

5.2 Two-year follow-up (III) ................................................................... 54

6 DISCUSSION ............................................................................................ 61

6.1 Main findings .................................................................................... 61

6.2 FISH trial versus other trials comparing surgery and nonsurgical

treatments in humeral shaft fractures .............................................. 62

6.3 Clinical implications ......................................................................... 63

6.4 Strengths and limitations of this study ............................................ 64

6.5 Future aspects ................................................................................... 65

7 CONCLUSIONS ....................................................................................... 66

ACKNOWLEDGMENTS .................................................................................. 67

REFERENCES ................................................................................................ 69

APPENDIX ...................................................................................................... 84

ORIGINAL PUBLICATIONS ........................................................................... 85

3

ORIGINAL PUBLICATIONS

This thesis is based on the following original publications. Roman numerals I

to III are used in the text to refer to these publications, which have been

reprinted with the kind permission of their copyright holders.

I Rämö L, Taimela S, Lepola V, Malmivaara A, Lähdeoja T, Paavola M.

Open reduction and internal fixation of humeral shaft fractures versus

conservative treatment with a functional brace: a study protocol of a

randomised controlled trial embedded in a cohort. BMJ Open 2017 Jul

9;7(7):e014076.

II Rämö L, Sumrein BO, Lepola V, Lähdeoja T, Ranstam J, Paavola M,

Järvinen TLN, Taimela S. Effect of Surgery vs Functional Bracing on

Functional Outcome Among Patients With Closed Displaced Humeral

Shaft Fractures: The FISH Randomized Clinical Trial. JAMA

2020;323(18):1792–1801.

III Rämö L, Paavola M, Sumrein BO, Lepola V, Lähdeoja T, Ranstam J,

Järvinen TLN, Taimela S. Surgery or Functional Bracing for Humeral

Shaft Fractures: Effect of Healing Problems Requiring Secondary

Surgery After Initial Nonoperative Treatment: A Pre-Specified

Secondary Analysis of the FISH Randomized Clinical Trial. JAMA Surg

2021;156(6):526–34.

4

ABBREVIATIONS

AO Arbeitsgemeinshaft für Osteosynthesegefragen

CI Confidence interval

DASH Disabilities of the Arm, Shoulder, and Hand

FISH Finnish Shaft of the Humerus

HRQoL Health-related quality of life

IMN Intramedullary nailing

LCP Locking compression plate

MIPO Minimally invasive plate osteosynthesis

NRS Numerical rating scale

ORIF Open reduction and internal fixation

OTA Orthopaedic Trauma Association

PASS Patient acceptable symptom state

PRNP Primary radial nerve palsy

PROM Patient-reported outcome measure

RCT Randomized controlled trial

RUSHU Radiographic union score for humeral fractures

SRNP Secondary radial nerve palsy

VAS Visual analog scale

5

ABSTRACT

Introduction

Humeral shaft fractures account for 1–3% of all adult fractures. They are

usually caused by simple falls, traffic accidents, and sports injuries.

Historically, the treatment of these injuries has been mainly nonsurgical.

However, there has been a marked increase in the rate of surgical treatment

for humeral shaft fractures in recent years without high-quality evidence

supporting this trend.

Aim

The Finnish Shaft of the Humerus (FISH) randomized clinical trial was

planned to compare the effectiveness of surgery versus nonsurgical care in the

treatment of humeral shaft fractures in patients traditionally deemed suitable

for nonsurgical care with functional bracing (Study I).

Patients and methods

Patient recruitment was conducted at the Helsinki and Tampere University

hospitals between November 2012 and January 2018. Consenting adult

patients with displaced, closed, unilateral humeral shaft fractures were

randomized to either surgical care using open reduction and plate fixation or

nonsurgical care using functional bracing. Patients with a history or condition

affecting the function of the injured upper limb, pathological fracture, other

concomitant injury affecting the same upper limb, other trauma requiring

surgery (e.g., fracture, internal organ, brachial plexus, or vascular injury),

polytrauma, multimorbidity with high anesthesia risk, or inadequate

cooperation for any reason were excluded. Altogether, 321 patients were

assessed for eligibility and of these 140 were eligible for randomization. After

informed consent, 82 were willing to undergo randomization. The primary

outcome was the Disabilities of the Arm, Shoulder, and Hand (DASH) score

(range, 0 to 100 points; 0 = no disability, 100 = extreme disability). In Study

II, the patients were analyzed according to the initially allocated treatment

method (surgery group and bracing group). In Study III, the patients were

analyzed in three groups according to their final treatment method: 1) initial

surgery group, 2) bracing group with successful healing, and 3) secondary

surgery group including patients randomized to functional bracing but who

underwent late surgery due to fracture healing problems.

Results

Study II. The mean DASH score was 8.9 (95% confidence interval [CI], 4.2 to

13.6) in the surgery group and 12.0 (95% CI, 7.7 to 16.4) in the bracing group

at 12 months. The between-group difference was -3.1 (95% CI, -9.6 to 3.3). This

difference was not statistically significant, and it was below the predefined

minimal clinically important difference of 10 points. Of the patients allocated

6

to functional bracing, 13/44 (30%) underwent late surgery due to healing

problem during the 12-month follow-up. In the post hoc analysis, the results

of those with initial surgery were superior to those with late surgery due to

healing problem (between-group difference, -11.1; 95% CI, -20.1 to -2.1) at 12

months.

Study III. The mean DASH score was 6.8 (95% CI, 2.3 to 11.4) in the initial

surgery group (n=38), 6.0 (95% CI, 1.0 to 11.0) in the bracing group (n=30),

and 17.5 (95% CI, 10.5 to 24.5) in the secondary surgery group (n=14) at the

2-year follow-up. The between-group difference was -10.7 (95% CI, -19.1 to -

2.3) between the initial and secondary surgery groups and -11.5 (95% CI, -

20.1 to -2.9) between the bracing and secondary surgery groups.

Conclusions

Study II. Surgery with plate fixation, compared with functional bracing in the

treatment of adult patients with closed humeral shaft fractures, did not

significantly improve functional outcomes at 12 months. However, 30% of the

patients allocated to functional bracing underwent late surgery due to healing

problem.

Study III. Shared decision-making between the clinicians and patients with

closed humeral shaft fractures should weigh the prospect that 2/3 of patients

undergo successful healing and have good functional outcomes using

functional bracing against the 1/3 risk of fracture healing problems leading to

secondary surgery and inferior outcomes even at 2 years after the injury.

7

TIIVISTELMÄ

Johdanto

Olkavarren murtumat muodostavat 1–3 % aikuisväestön murtumista. Tämän

vamman syynä on yleensä kaatuminen, liikenneonnettomuus tai

urheiluvamma. Aiemmin näitä murtumia on hoidettu pääsääntöisesti ilman

leikkausta käyttäen esimerkiksi ulkoista tukilastaa, ns. toiminnallista ortoosia.

Kuitenkin parin viime vuosikymmenen aikana näiden vammojen

leikkausmäärät ovat selvästi lisääntyneet ilman leikkaushoitoa tukevaa

laadukasta tieteellistä näyttöä.

Tavoite

Tässä satunnaistetussa vertailevassa tutkimuksessa verrattiin kirurgisen

hoidon ja ilman leikkausta toteutetun hoidon (ortoosihoito) vaikuttavuutta

sulkeisen olkavarren murtuman saaneilla aikuispotilailla. Näiden vammojen

on perinteisesti katsottu soveltuvan hyvin ortoosihoitoon. Tutkimuksen

menetelmät julkaistiin protokollajulkaisussa (tutkimus I).

Aineisto ja menetelmät

Tutkimusaineisto kerättiin Helsingin ja Tampereen yliopistollisissa

sairaaloissa marraskuun 2012 ja tammikuun 2018 välisenä aikana. Tänä

aikana arvioitiin 321 olkavarren murtuman saaneen aikuispotilaan

soveltuvuus tutkimukseen. Tutkimuksesta suljettiin pois potilaat, joilla oli

aiempi toiminnallista haittaa aiheuttava yläraajan vamma tai sairaus,

pahanlaatuisesta sairaudesta johtuva murtuma, muu kirurgista hoitoa

edellyttävä vamma, monivamma, merkittävä anestesiariskiä nostava sairaus

tai ilmeinen hoitomyöntyvyyteen vaikuttava tila. Yhteensä 140 potilasta

soveltui mukaan tutkimukseen ja heistä 82 suostui satunnaistettuun hoitoon,

missä potilas hoidettiin joko kirurgisesti käyttäen avointa murtuman

paikalleen asetusta ja levykiinnitystä tai ilman leikkausta ortoosihoidolla.

Tutkimuksen päätulosmuuttuja oli potilaan itse raportoima yläraajan

toimintakykyä ja vammaan liittyvää haittaa kartoittava Disabilities of the Arm,

Shoulder, and Hand (DASH) mittari; asteikko 0-100; 0 = ei toiminnan

rajoitetta, 100 = äärimmäinen toiminnan rajoite). Tutkimuksessa II tulokset

arvioitiin sen mukaisesti, mihin ryhmään potilas oli tutkimuksen alussa

satunnaistettu (kirurginen ryhmä vs. ortoosiryhmä). Tutkimuksessa III

tulokset arvioitiin kolmessa ryhmässä lopullisen hoitomuodon mukaisesti: 1)

välittömän kirurgian ryhmä, 2) ortoosiryhmä, joilla murtuma parani ilman

ongelmia ja 3) myöhäisen kirurgian ryhmä sisältäen ne potilaat, joilla

ortoosihoito epäonnistui ja heidät jouduttiin leikkaamaan myöhäisvaiheessa

murtuman paranemisongelman vuoksi.

8

Tulokset

Tutkimus II. DASH-pisteiden keskiarvo oli 8.9 (95 %:n luottamusväli [CI],

4.2–13.6) kirurgisessa ryhmässä ja 12.0 (95 % CI, 7.7–16.4) ortoosiryhmässä

12 kuukauden kuluttua vammasta. Ryhmien välinen ero oli -3.1 pistettä (95 %

CI, -9.6–3.3) kirurgisen ryhmän eduksi, mutta ero ei ollut tilastollisesti

merkitsevä ja se oli selvästi alle ennalta määritetyn pienimmän kliinisesti

merkitsevän 10 pisteen erotuksen. 30 % (13/44) ortoosiryhmän potilasta

jouduttiin kuitenkin leikkaamaan vuoden kuluessa vammasta

paranemisongelman vuoksi. Jälkikäteen tehdyssä analyysissa heti hoidon

alussa leikattujen potilaiden DASH-pisteet olivat merkittävästi paremmat

kuin myöhään leikattujen potilaiden pisteet. Ryhmien välisen keskiarvojen

erotus oli -11.1 pistettä (95 % CI, -20.1 – -2.1) heti alussa leikattujen hyväksi 12

kuukauden kuluttua vammasta.

Tutkimus III. DASH-pisteiden keskiarvo oli 2 vuoden kuluttua vammasta 6.8

(95 % CI, 2.3–11.4) välittömän kirurgian ryhmässä (n=38), 6.0 (95 % CI, 1.0–

11.0) ortoosiryhmässä (n=30) ja 17.5 (95 % CI, 10.5–24.5) myöhäisen

kirurgian ryhmässä (n=14). Ryhmien keskiarvojen väliset erot olivat -10.7

pistettä (95 % CI, -19.1 – -2.3) välittömän ja myöhäisen kirurgian ryhmien

välillä ja -11.5 pistettä (95 % CI, -20.1 – -2.9) ortoosiryhmän ja myöhäisen

kirurgian ryhmän välillä.

Yhteenveto

Tutkimus II. Kirurgisella hoidolla ei saavutettu parempia tuloksia

ortoosihoitoon verrattuna sulkeisen olkavarren murtuman saaneilla

aikuispotilailla vuoden kuluttua vammasta. Toisaalta 30 % potilaista, joilla

hoito aloitettiin ortoosilla, jouduttiin leikkaamaan paranemisongelman

vuoksi.

Tutkimus III. Olkavarren murtuman saaneen potilaan ja häntä hoitavan

lääkärin tulee huomioida parasta hoitomuotoa valitessaan, että 2/3

ortoosihoidolla hoidetuista potilaista paranee hyvin tuloksin ilman kirurgisen

hoidon haittoja, mutta 1/3:lla on ortoosihoitoa käytettäessä

paranemisongelmia, minkä vuoksi heidät joudutaan leikkaamaan

viivästetysti. Tällöin heidän toiminnalliset tuloksensa ovat vielä 2 vuoden

kuluttua vammasta huonompia, kuin heti leikatuilla potilailla sekä niillä

potilailla, joilla murtuma paranee ortoosihoidolla ilman leikkausta.

9

1 INTRODUCTION

Fractures of the humerus account for 7.5% of all fractures in adults (Court-

Brown et al. 2006). Fractures of the humerus are generally divided according

to anatomical location to proximal, shaft, and distal humeral fractures.

Humeral shaft fractures comprise around 15% of all humeral fractures

(Knowelden et al. 1964, Rose et al. 1982, Court-Brown et al. 2006, Bergdahl et

al. 2016). In Finland, the incidence for inpatient care due to humeral shaft

fracture is around 7/100 000 person-years (Somersalo et al. 2014).

Before the 20th century and the advent of modern surgery, the

treatment of humeral shaft fractures was mainly nonsurgical. Surgery was

used in very rare cases, with treatment being nothing short of exotic compared

with modern surgical standards—for instance, rubbing of caustic potash

against the end of non-united humeral shaft ends (Earle 1823). In reports from

the 1930s, a surgical approach was employed in approximately 15% of cases,

using mainly steel plates, screws, and wires (Smyth 1934). The pioneer of

intramedullary nailing, Gerhard Küntscher, introduced his technique in the

1930s, first in femoral and later in humeral shaft fractures (Küntscher 1940,

1958). Another alternative, intramedullary pinning, was introduced nearly at

the same time (Rush et al. 1949, 1950). Plate osteosynthesis with compression

plating was introduced by a Belgian surgeon, Robert Danis, in 1949 (Danis

1949). The technique was popularized by Maurice Müller, a founding member

of the Swiss research group Arbeitsgemeinshaft für Osteosynthesegefragen

(AO) (Müller 1956, 1963). Regardless of advances in surgical care, there were

strong advocates for nonsurgical management of humeral shaft fractures

(Böhler 1965) and even the AO members leading the development of modern

surgical fracture management suggested that good results are achieved with

nonsurgical methods, and this should be considered as the first line of

treatment (Rüedi et al. 1974). The methods for nonsurgical care included

bandages, hanging casts, splints, and shoulder spicas (Klenerman 1966).

In 1977, Augusto Sarmiento published his method of nonsurgical

management of humeral shaft fractures using functional bracing (Sarmiento

et al. 1977). This was later followed by a large retrospective cohort study of

nearly 1000 patients treated with this method, and the publication became a

landmark study for functional bracing of humeral shaft fractures (Sarmiento

et al. 2000). Of the patients, 97% healed with functional bracing without the

need for additional interventions and with good reported outcomes. The study

can, however, be criticized for its large attrition, as one-third of patients were

lost to follow-up. Several studies have subsequently tried to repeat the

excellent outcomes of Sarmiento’s publication without achieving their results

(Koch et al. 2002, Toivanen et al. 2005, Ali et al. 2015, Harkin et al. 2017).

10

We have witnessed a marked increase in the surgery of humeral shaft

fractures during the last few decades, with up to two-thirds of fractures

operated (Mahabier et al. 2015, Schoch et al. 2017). In Finland, there was a

two-fold increase in the number of surgeries performed for these fractures

from 1987 to 2009 (Huttunen et al. 2012). At our unit—Töölö Hospital (Level

1 Trauma Center, Helsinki University Hospital)—approximately 50% of

patients underwent surgery as the first line of treatment between 2005 and

2009. Of those who were initially treated with functional bracing, one-third

had late surgery to promote the healing of their fracture (Penttilä et al. 2012).

However, there is scant evidence to support the trend of treating these

fractures primarily with surgery. Despite humeral shaft fractures being a

common injury, the first randomized controlled trial (RCT) comparing surgery

with nonsurgical care was published only in 2017 (Matsunaga et al. 2017). No

clinically meaningful difference emerged in any of the outcomes. However,

15% of the patients treated initially with functional bracing required surgery

to promote healing of their fracture.

The aim of this thesis was to compare the effectiveness of surgery with

nonsurgical treatment of closed humeral shaft fractures in adults.

11

2 REVIEW OF THE LITERATURE

2.1 SURGICAL ANATOMY AND EXPOSURES OF THE

HUMERAL SHAFT

Surgical anatomy of the humerus is generally divided into proximally located

head, shaft, and distal end. The head articulates with the glenoid fossa of the

scapula, and the distal part forms the elbow joint together with the ulna and

radius. This section focuses on surgical anatomy related to humeral shaft

fractures. Important structures around the humerus are the brachial plexus

and the brachial artery running medially along the humeral shaft, and the

radial nerve running with the deep brachial artery around the radial groove of

the humerus through the intermuscular septum.

Surgical exposures of the humeral shaft are generally divided into

anterior, lateral, posterior, and (rarely used) medial approaches. The approach

used depends on the location of the fracture, condition of the soft tissues, and

the surgeon’s preference. Here, the three most common approaches for open

reduction and internal fixation (ORIF) are described.

Henry in 1924 (Henry 1924). The incision starts from the coracoid process,

following the course of the cephalic vein to the anterior aspect of the cubital

fossa. The biceps brachii muscle is moved medially with the accompanying

musculocutaneous nerve in the posterior aspect of the muscle belly. The

brachialis muscle is divided longitudinally at the outer fourth of the muscle,

exposing the humeral shaft. The radial nerve stays protected on the lateral side

of the brachialis muscle fibers, but it is easily found if necessary (Fig. 1 A). The

term ‘anterolateral approach’ is commonly used in conjunction with the

anterior approach in the literature. The distinction can be made according to

how the humeral shaft is exposed at the distal end of the exposure. In the

anterolateral approach (Fig. 1 B), the plane between the brachialis and

brachioradialis is used, instead of splitting the brachialis muscle (true anterior

approach). Recently, a modification of the approach was introduced to prevent

unnecessary transection of diagonally oriented superficial head fibers of the

brachialis muscle at the distal end of the humeral shaft (Chang et al. 2019).

The anterior approach is a useful option in the treatment of proximal and

middle third fractures of the humeral shaft.

12

Lateral

The lateral approach (Fig. 2) was described in the context of humeral shaft

fractures rather recently (Mills et al. 1996). The incision starts from the lateral

epicondyle of the humerus proximally towards the deltoid insertion. The plane

between the brachioradialis and triceps muscles is divided, and the humeral

shaft is exposed. It is paramount to locate and protect the radial nerve as it

pierces the lateral intermuscular septum within 5 mm from a junction of the

middle and distal thirds of the line running from the lateral border of the

acromion to the lateral epicondyle (Fleming et al. 2004). The lateral approach

is mainly used in distal third shaft fractures.

A.

B.

Fig. 2. Lateral approach to the humeral shaft. Copyright by AO Foundation,

Switzerland. Source: AO Surgery Reference, www.aosurgery.org.

Fig. 1.

A. Anterior approach to the humerus, where the brachialis muscle is split.

B. Anterolateral approach, where the brachialis muscle is retracted medially.

Copyright by AO Foundation, Switzerland. Source: AO Surgery Reference, www.aosurgery.org.

13

Posterior

The posterior approach gives good exposure of middle and distal third shaft

fractures. The bone can be exposed either around the lateral and medial border

of the triceps muscle (paratricipital approach, Fig. 3) or by splitting the muscle

longitudinally (triceps splitting approach, Fig. 4) (Gerwin et al. 1996). This

approach gives good visibility to the radial nerve, which should always be

visualized and protected before exposing the humeral shaft. This approach

allows placing of the implant to either the medial, posterior, or lateral border

of the humerus. This approach is useful especially in the cases having both

distal intra-articular humeral fracture and ipsilateral shaft fracture.

Fig. 4. Posterior triceps splitting approach to the humeral shaft. Copyright by AO Foundation,

Switzerland. Source: AO Surgery Reference, www.aosurgery.org.

Fig. 3. Posterior paratricipital approach to the humeral shaft with radial (left) and ulnar

(right) windows around the triceps muscle. Copyright by AO Foundation, Switzerland. Source: AO

Surgery Reference, www.aosurgery.org.

14

Approach for minimal invasive plate osteosynthesis

Two separate incisions are made to enable plate fixation above and below the

fracture. The proximal part can be made either through the deltoid muscle

(transdeltoid approach) or using the upper part of the anterior approach

proximally (deltopectoral interval) and the anterolateral approach distally

(Fig. 5). Care must be taken at the distal part to protect the radial nerve on the

lateral side of the incision.

Approaches for intramedullary nailing

An intramedullary nail is introduced to the medullary canal from either the

proximal (antegrade, Fig. 6) or distal (retrograde, Fig. 7) direction. For

antegrade nailing, an approximately 4 cm incision is made from the

anterolateral border of the acromion downwards. The muscle fibers of the

deltoid muscle are split, and the rotator cuff interval is opened. Care must be

taken to avoid excessive opening of the deltoid muscle, as the axillary nerve

runs approximately 7 cm below the edge of the acromion. For retrograde

nailing, a longitudinal midline incision is made over the tendinous part of the

triceps right above the olecranon. Sharp dissection is carried out through the

tendon by splitting the tendon fibers going towards the upper part of the

olecranon fossa.

Fig. 5. Anterior approach for minimal invasive plate osteosynthesis. Copyright by AO Foundation,

Switzerland. Source: AO Surgery Reference, www.aosurgery.org.

Fig. 6. Approach for antegrade nail

insertion. Copyright by AO Foundation, Switzerland.

Source: AO Surgery Reference, www.aosurgery.org.

15

2.2 EPIDEMIOLOGY OF HUMERAL SHAFT FRACTURES

AND ASSOCIATED INJURIES

Humeral shaft fractures account for 1–3% of all fractures and the incidence is

around 10–30/100 000 person-years (Rose et al. 1982, Court-Brown et al.

2006, R Ekholm et al. 2006, Bergdahl et al. 2016, Oliver et al. 2020). The

incidence starts to rise in people over 50 years of age, reaching 100/100 000

person-years in those aged 80+ years (Tytherleigh-Strong et al. 1998). In

Finland, the incidence of inpatient care due to humeral shaft fractures is

7/100 000 person-years (Somersalo et al. 2014). In the United States, humeral

shaft fractures accounted for 60 000 visits to emergency departments in 2008

(Kim et al. 2012).

Humeral shaft fractures are caused mainly by simple falls, especially in

elderly patients, while sports and traffic accidents predominate as the

mechanism of injury in younger patients (Tytherleigh-Strong et al. 1998).

The traumatic event fracturing the humerus can cause associated injuries

to surrounding tissues, i.e., neural, vascular, or soft tissue injuries. The most

common associated injury with humeral shaft fractures is radial nerve palsy,

found in approximately 12% of cases (Hegeman et al. 2020, Ilyas et al. 2020).

Spiral distal shaft fractures (Holstein-Lewis fracture, AO/OTA 12A1c, Fig. 9)

have been reported to have a rate of radial nerve palsy of up to 22%

(Ekholm et al. 2008a).

Open humeral shaft fractures are rare, with a reported incidence of

0.4/100 000 person-years constituting 3% of all open long bone fractures

(Court-Brown et al. 1998) and 2–5% of all humeral shaft fractures

(Tytherleigh-Strong et al. 1998, R Ekholm et al. 2006, Bergdahl et al. 2016).

At our institute (Helsinki University Hospital), there were 26 open humeral

shaft fractures in 938 cases (2.8%) treated between 2006 and 2016

(unpublished data). Associated injuries to the brachial plexus (Brien et al.

1990) or brachial artery (Gainor et al. 1986) are very rare, and the incidence is

not reported in the literature.

Fig. 7. Approach for retrograde nail

insertion. The medullary canal is opened 2.5

cm proximally from the proximal border of

olecranon fossa (green triangle). Copyright by

AO Foundation, Switzerland. Source: AO Surgery

Reference, www.aosurgery.org.

16

2.3 HUMERAL SHAFT FRACTURE CLASSIFICATION

The most common classification system for humeral shaft fractures is

AO/OTA (Arbeitsgemeinshaft für Osteosynthesegefragen and Orthopaedic

Trauma Association) classification (Kellam et al. 2018). Fractures are divided

into three main types according to their morphology (Fig. 8): type A (simple

fracture line), type B (separate wedge fragment), and type C (segmental

fracture). All main types are further divided into groups: type A in spiral,

oblique, and transverse fractures; type B in intact and fragmented wedge

fractures; and type C in intact and fragmented segmental fractures. Moreover,

fractures have qualifications according to the location of the center of the

fracture (Fig. 9). Even though the severity of bony injury increases from type

A to type C, the AO/OTA classification has not been validated to guide

treatment decisions of humeral shaft fractures.

Fractures can also be classified according to the severity of soft tissue

injury. The most commonly used classifications for soft tissue injuries are the

Gustilo-Anderson (open fractures) and Tscherne (closed and open fractures)

classifications (Gustilo et al. 1976, Tscherne et al. 1982). As open fractures

were excluded from the studies of this thesis and the closed fractures were not

classified according to the Tscherne classification, the classifications are not

introduced here.

Fig. 8. AO/OTA classification of humeral shaft fracture types.

Copyright by AO Foundation,

Switzerland. Source: AO Surgery Reference, www.aosurgery.org.

17

Fig. 9. Groups and qualifications of humeral shaft fractures according to the AO/OTA

classification. Copyright by AO Foundation, Switzerland. Source: AO Surgery Reference,

www.aosurgery.org.

18

2.4 NONSURGICAL TREATMENT OF HUMERAL SHAFT

FRACTURES

Treatment of humeral shaft fractures has historically been mainly nonsurgical

(Smyth 1934, Mitchell et al. 1942), with some authors advocating strongly

against surgical care (Böhler 1964, 1965). The rationale behind nonsurgical

care is to splint the broken bone with external support with several suggested

methods. There is no comparative evidence of the superiority of any

nonsurgical treatment method (Mukerjee et al. 2008).

2.4.1 HANGING CAST

The technique was published in 1933 (Caldwell

1933). The elbow is bent at 90° of flexion and the

cast is molded around the arm, elbow, and

proximal forearm to support the fractured

humerus. A sling is placed around the collar and

the cast to support the hanging cast (Fig. 10). The

weight of the cast pulls the fractured humerus.

Care must be taken to avoid excessive pull of the

fractured humerus with the cast.

2.4.2 COAPTATION SPLINT

Coaptation splint or U-splint was introduced to

prevent an excessive diastasis of the fracture gap

with a heavy hanging cast. A U-shaped plaster is placed starting from the top

of the shoulder, going laterally around the elbow, and ending medially to the

axilla. The splint allows some movement of the elbow, but full extension is not

possible. The hand is supported with a sling around the collar.

2.4.3 OTHERS

The use of a simple collar and hand sling has been suggested to avoid stiffness

of surrounding joints (Spak 1978). Other methods of nonsurgical care used

earlier were open Velpeau-type thoracobrachial or shoulder spica casts (Holm

1970), but these types of weighty casts surrounding the whole thoracic region

are rarely seen nowadays.

Fig. 10. Hanging cast.

Photo credit: Ca ldwell JA. Treatmen t of

Fractures in the Cincinnati General

Hospital. Ann Surg. 1933;97(2):161–76.

19

2.4.4 FUNCTIONAL BRACING

The current mainstay of nonsurgical

care is functional bracing. The method

was popularized by Augusto Sarmiento

in the 1970s (Sarmiento et al. 1977). The

technique was first developed for tibial

shaft fractures (Sarmiento 1967). The

technique was soon adapted to humeral

shaft fractures. A custom-made or

prefabricated brace is placed around the

fractured arm, leaving the motion of the

elbow free (Fig. 11). The biomechanical

rationale for functional bracing is that

some motion is needed in the fracture

site to enable the best possible

environment for fracture healing

(Sarmiento et al. 1995). A soft tissue

envelope around the fractured humerus

is supported by the brace (Fig. 12). Early motion of the elbow is advocated to

prevent stiff elbow. The muscle contraction around the fracture is beneficial

for the fracture healing (Sarmiento et al. 1974). The brace should be tightened

regularly to function properly, while avoiding overtightening, as this might

result in pressure ulcers. The brace is kept in place at all times, except during

showering, until radiological and clinical healing is observed (Zagorski et al.

1988).

A

B

C

D

Fig. 11. Custom-made functional brace.

Fig. 12.

A. A 36-year-old male sustained a

humeral shaft fracture (AO/OTA

12A3b) due to fall from a standing

height. The patient was randomized

to functional bracing. The brace was

applied at the emergency department

and the X-ray was taken with the

brace.

B. There was some callus formation

at 6 weeks.

C. The arm was pain-free and stable

with a strong callus formation at 3

months.

D. A strong fracture union was

observed at 6 months.

20

2.5 SURGICAL TREATMENT OF HUMERAL SHAFT

FRACTURES

Surgical treatment of humeral shaft fractures has been suggested in several

clinical situations (Table 1). The indications are derived from retrospective

case series. There are no studies validating the superiority of surgical care over

nonsurgical treatment in these situations, but a common sense justifies an

operative approach. According to current literature, the most common

associated injury, primary radial nerve palsy (PRNP) with no other indication

for surgery, is generally not considered an indication for early nerve

exploration (Böstman et al. 1986, Shao et al. 2005, Ekholm et al. 2008b,

Heckler et al. 2008, Liu et al. 2012, Korompilias et al. 2013). However, the

most recent systematic review combining 58 observational studies with

890/7262 PRNPs found that patients with early exploration had a higher

chance of nerve recovery than those with late exploration (89.8% vs. 68.1%)

(Ilyas et al. 2020). This conclusion can be criticized for overestimating the

effect of early exploration. Some of the patients with early exploration could

have recovered without intervention, but the patients with late exploration do

not include those with spontaneous nerve recovery. There are no RCTs

comparing surgery with nonsurgical care in humeral shaft fractures

complicated with PRNP.

Table 1. Generally accepted indications for surgical treatment of humeral shaft fractures.

Clinical situation

References

Bilateral fracture

Brug et al. 1994

Pathological fractures

Flinkkilä et al. 1998, Laitinen et al. 2011

Floating elbow (arm and

forearm fracture)

Rogers et al. 1984, Sarmiento et al. 2001

Multiple trauma patient

Bell et al. 1985, Brumback et al. 1986

Brachial plexus or artery injury

Gainor et al. 1986, Brien et al. 1990

Open fractures (grade II or III)

Brug et al. 1994, Sarmiento et al. 2001

Symptomatic nonunion

Campbell 1937, Barquet et al. 1989, Jupiter et al. 1998

2.5.1 OPEN REDUCTION AND PLATE OSTEOSYNTHESIS

Originally, the aim of ORIF was to anatomically reduce the fracture by

exposing the fracture site, compressing the fracture either with an

interfragmentary lag screw or by using dynamic compression with the plate

(Müller 1963). Compression is appropriate especially in simple fractures and

has been shown to improve bony union (King 1957). However, in comminuted

fractures, compression of the fracture is not possible. This led to the concept

of bridge plating, where the alignment of the bone is corrected without

interfering with the fracture zone and fixing the plate from both sides of the

fracture (Heitemeyer et al. 1987).

21

The plate is fixed with locking or nonlocking screws according to the

surgeon’s preference and quality of the bone (Fig. 13). Often use of locking

screws is recommended in osteoporotic bone with three bicortical screws on

both sides of the fracture (Gautier et al. 2003). However, in a cadaveric study

of osteoporotic bone, addition of a third locking screw did not strengthen the

construction (Hak et al. 2010). In a biomechanical study with a bone model,

the construction with two locking screws on both sides of the fracture gap

showed similar biomechanical properties in an osteoporotic model compared

with three nonlocking screws on both sides of the 1 cm fracture gap (Grawe et

al. 2012). In the same study with a good-quality bone model, the construction

with three nonlocking screws was slightly superior to that of two locking

screws. These findings have not been validated in vivo.

The length of the plate is often debated. There is no good evidence for

an optimal plate length. Empirically, in comminuted fractures, a plate length

of 2 to 3 times and in simple fractures 8 to 10 times the length of the fracture

zone has been suggested (Gautier et al. 2003).

A potential benefit of plate fixation of the humeral shaft fracture is that

it seems safe to use the extremity for weight-bearing, as this did not increase

the number of hardware failures, malunions, or nonunions compared with

patients not bearing weight with upper extremity (Tingstad et al. 2000). This

has clinical implications in patients with multiple injuries who have an injured

lower extremity in conjunction with a humeral shaft fracture. It enables

ambulation by using crutches even when weight-bearing of the lower limb is

not permitted. Also, patients depending on the use of canes or other walking

aids can remain ambulatory after fractured humerus.

Fig. 13.

A. A 62-year-old woman fractured her

left humeral shaft (AO/OTA 12A3b)

due to a fall from a standing height.

B. The patient was randomized to

surgery. The fracture was reduced

through an anterolateral approach and

fixed with a locking compression plate

using dynamic compression.

C. The fracture healed uneventfully,

and a strong fracture union was

observed at 6 months after surgery.

B

C

A

22

2.5.2 INTRAMEDULLARY NAILING

Intramedullary fixation was proposed in the 1930s by the Rush brothers (Rush

et al. 1950) with a flexible pin. Gerhard Küntscher popularized the technique

by developing a more rigid intramedullary implant (Küntscher 1940, 1958).

Intramedullary nailing (IMN) became popular, especially in Northern

Finland, where Küntscher served in the German troops during World War II.

Later, an intramedullary bundle nailing technique with several intramedullary

pins was introduced (Hackethal 1961), and this technique is today used in, for

instance, Morocco, Brazil, Cote d’Ivore, and the Czech Republic (Rodríguez-

Merchán 1995, Obruba et al. 2012, Sié et al. 2014, Mohamed et al. 2018). It has

not gained popularity in Finland.

The most common method of IMN in Finland is locked nailing, where

the nail is introduced to the intramedullary canal of the fractured humerus

either from the proximal (antegrade nailing) or distal (retrograde nailing) end

of the bone, as described in the exposures for intramedullary nailing above.

The nail is locked from both sides of the fracture with screws. The fracture site

is normally left untouched but can be opened in the cases with radial nerve

palsy for identification and possible repair of the nerve. The natural benefit of

IMN compared to ORIF is less prominent surgical scars as it can be introduced

with only small incisions.

The use of IMN in humeral shaft fractures has been studied extensively.

There are several RCTs (Chiu et al. 1997, Chapman et al. 2000, McCormack et

al. 2000, Changulani et al. 2007, Putti et al. 2009, Li et al. 2011, C. Wang et al.

2013, Singh et al. 2014, Fan et al. 2015, Akalın et al. 2020) and meta-analyses

(Bhandari et al. 2006, Heineman et al. 2010, Kurup et al. 2011, Liu et al. 2013,

Ma et al. 2013, Ouyang et al. 2013, X. Wang et al. 2013, Dai et al. 2014,

Zarkadis et al. 2018) comparing IMN with ORIF. According to these

publications, the outcomes are comparable, except for the higher risk of

shoulder pain and reoperation with IMN than with ORIF. The popularity of

IMN varies among countries, with less interest than with ORIF in treating

traumatic shaft fractures in Finland and the USA (Huttunen et al. 2012,

Gottschalk et al. 2016, Schoch et al. 2017). For pathological humeral shaft

fractures, IMN has proven to be a valuable tool and is commonly used for this

indication in Finland (Flinkkilä et al. 1998, Flinkkilä 2004, Laitinen et al.

2011).

2.5.3 MINIMALLY INVASIVE PLATE OSTEOSYNTHESIS

Minimally invasive plate osteosynthesis (MIPO) has recently gained in

popularity (Tetsworth et al. 2018). The principle of using MIPO in humeral

shaft fractures was published in 2002 (Fernández Dell’Oca 2002), with reports

of case series following thereafter (Livani et al. 2004, Zhiquan et al. 2007,

Apivatthakakul et al. 2009). The rationale behind MIPO is to combine the

benefits of ORIF and the minimal surgical soft tissue trauma of IMN. The plate

23

is inserted and attached using two separate small incision without disturbing

the fracture site. The major concern has been the possible secondary radial

nerve palsy (SRNP), as the plate is introduced without visibility to the nerve

(Livani et al. 2009). However, in comparative studies the risk has not proven

to be an issue; in fact, the reported incidences of SRNPs are higher with IMN

or ORIF in studies comparing MIPO with IMN or ORIF (Beeres et al. 2020,

van de Wall et al. 2021). It is noteworthy that in some studies the rate of SRNP

after ORIF is unacceptably high (>30%) (An et al. 2010). The RCTs from

China, Egypt, Brazil, Korea, and the Czech Republic, with low numbers of

participants comparing MIPO with either ORIF or IMN, showed similar

outcomes with regard to complications and function (Lian et al. 2013, Benegas

et al. 2014, Smejkal et al. 2014, Esmailiejah et al. 2015, Hadhoud et al. 2015,

Kim et al. 2015). In Finland, use of MIPO in humeral shaft fractures has not

yet gained popularity.

2.5.4 EXTERNAL FIXATION

The use of external fixation (Fig. 14)

in treatment of humeral shaft

fractures was first published in 1907

(Lambotte 1907). The first case

series, including 8 cases, was

published in 1978 (Kamhin et al.

1978). This was followed by a larger

series of 164 patients, where the

method was used for complex

proximal and distal metaphyseal

shaft fractures (Hinsenkamp et al.

1984). More recently, it has been

used mainly as a temporary fixation

before definitive treatment with either plate or nail in patients with severe soft

tissue injuries or polytrauma (Sarmiento et al. 2001, Suzuki et al. 2010). It has

proven to be a valid definitive treatment method in infected nonunions

(Bassiony et al. 2009, Xiao et al. 2016), with some authors, many from Italy,

recommending it in acute noncomplicated fractures as well (Tartaglia et al.

2016, Basso et al. 2017, Alhammoud et al. 2019, Costa et al. 2019). To date, no

RCTs have compared external fixation with other treatment methods in

humeral shaft fractures.

In Finland, external fixation is seldom used even in complicated

fractures. At Helsinki University Hospital, only 4 of 938 fractures were treated

temporarily with this method between 2006 and 2016 (unpublished data).

Fig. 14. Albin Lambotte (1866–1955)

performed the first external fixation in 1902.

24

2.6 OUTCOME MEASURES IN CLINICAL STUDIES OF

HUMERAL SHAFT FRACTURES

Fracture union

One of the most important goals of any long bone fracture treatment is fracture

union. In clinical studies on humeral shaft fractures, the rate of fracture union

or often nonunion is practically always reported. The problem in interpreting

the results from different studies is the lack of a universally accepted definition

for fracture union or nonunion (Corrales et al. 2008, Morshed et al. 2008,

Morshed 2014, Cunningham et al. 2017).

Fracture malunion (i.e., fracture healing in a nonanatomic position)

rate is often reported. However, the definition of malunion and especially its

effect on other outcomes are not well defined. Generally, 20° of antecurvatum

and 30° of varus are considered limits for acceptable alignment, having no

clinical consequences, but these limits are based on a cohort of 32 patients

treated in the 1950s and 1960s (Klenerman 1966). Fracture unions within

these limits were later shown to have no correlation with patient-reported

outcomes and patient satisfaction in a cohort of also 32 patients (Shields et al.

2016).

In addition, time to fracture union is often reported. The caveat in

interpreting the union times is the fact that radiographs are taken at intervals

of several weeks. This causes an imprecise estimate of union times, as patients

having a fracture union at any follow-up point are defined as having achieved

fracture union at that time point (Fig. 15). In reality, the fracture union has

occurred between the previous and the current time point. A more precise

method for assessing fracture union time would be to report the proportion of

patients reaching fracture union at different time points (Table 2).

Fig. 15. Time to fracture union from the

Finnish Shaft of the Humerus (FISH)

trial data. This example reflects the

problem of calculating the fracture union

time as the radiographs are taken at

predefined time points and this does not

reflect the true time to fracture union.

The blue and red lines represent the

mean (solid line) and the median (dotted

line) times to fracture union and

individual dots represent patients with

fracture union.

25

Table 2. Proportion of patients reaching fracture union at different time points.

An example from the FISH trial data.

Pain

Pain is generally measured on a visual analog scale (VAS) or on an 11-point

numerical rating scale (NRS) (Revill et al. 1976, Ferreira-Valente et al. 2011).

Some earlier studies on humeral shaft fractures report residual pain on a

dichotomous ‘no pain’ or ‘pain’ scale (Wallny et al. 1997a, 1997b, Koch et al.

2002) or on a verbal scale (Klestil et al. 1997). Unfortunately, only a few studies

on humeral shaft fractures report either VAS or NRS for pain (van Middendorp

et al. 2011, Matsunaga et al. 2017), making a comparison of different treatment

options difficult with regard to this outcome.

Function

Functional outcomes of humeral shaft fracture treatment have been reported

using several outcome measures. In the early publications of humeral shaft

fracture treatment, the functional outcomes consisted of a description of the

range of motion of shoulder and elbow joints (Sarmiento et al. 1977, 2000,

Spak 1978, Balfour et al. 1982, Zagorski et al. 1988). Hunter (1982) used his

own grading for functional outcome, but this simple 5-point scale has been

used only once in later publications (Naver et al. 1986).

One of the first widely accepted functional outcome measures still in

use was developed by Constant and Murley (Constant et al. 1987). The

Constant-Murley Score contains subjective measures of pain and activities of

daily living, and objective measures of range of motion and strength of

shoulder. The scale ranges from 0 to 100, with a higher score indicating better

outcome. The normative values for different age groups have been published

elsewhere (Constant et al. 2008).

Patient-reported outcome measures (PROMs)

Upper extremity-specific PROMs

The Disabilities of the Arm, Shoulder, and Hand (DASH) score is a widely used

PROM to assess the upper limb function and symptoms in daily living (Hudak

et al. 1996). The score ranges from 0 to 100, with lower score indicating better

outcome. DASH score has proven to be a valid tool in assessing outcomes of

humeral shaft fractures (Mahabier et al. 2017, Van Lieshout et al. 2020) and

has the best clinimetric properties out of several shoulder scores (Bot et al.

2004). Currently, it is one of the most popular outcomes measures in humeral

Follow-up

time point

Surgery group

n (%)

Bracing group

n (%)

6 weeks

13 (34)

10 (23)

3 months

26 (68)

28 (64)

6 months

35 (92)

33 (75)

12 months

37 (97)

40 (91)

26

shaft fracture studies, as all the published and upcoming RCTs comparing

surgical and nonsurgical care use DASH score as the primary outcome (Kumar

et al. 2017, Matsunaga et al. 2017, Hosseini Khameneh et al. 2019, W. M. Oliver

et al. 2019, Berry 2020, Karimi 2020). There is a culturally adapted Finnish

translation of the DASH score without a proper validation study (Hacklin et

al. 2009).

Other functional outcome measures used in reporting results of

humeral shaft fracture treatment are American Shoulder and Elbow Surgeons

Shoulder Score (ASES), the University of California at Los Angeles Shoulder

Score (UCLA Shoulder Score), Mayo Elbow Performance Index (MEPI), and

Oxford Shoulder Score. There is no universal consensus on the best outcome

measure in upper limb conditions.

Generic quality-of-life PROMs

There are only 10 studies reporting the health-related quality of life (HRQoL)

after humeral shaft fracture treatment, including two RCTs (Matsunaga et al.

2017, Akalın et al. 2020). The instrument used in these 10 publications was

the 36-item Short Form Health Survey (SF-36). The two ongoing RCTs

comparing nonsurgical and surgical care in humeral shaft fractures will use

EQ-5D in reporting HRQoL (W. M. Oliver et al. 2019, Karimi 2020). A 15-

dimensional (15D) tool was developed in Finland (Sintonen 2001) and with

the reported normative values in a Finnish population (Koskinen et al. 2012)

it was chosen as the HRQoL instrument in the FISH trial. None of the HRQoL

instruments are validated and their responsiveness in patients with humeral

shaft fractures is unknown.

Important concepts of PROMs

The minimal clinically important difference (MCID)

The MCID is the smallest difference in the outcome measure that the patients

perceive as important (Jaeschke et al. 1989). This concept may help clinicians

and patients when contemplating different treatment options and the

magnitude of their effect on outcomes. The MCID for DASH has been

estimated to be 10 points in patients with different upper extremity conditions

(Gummesson et al. 2003), 1.5 points for pain on 11-point NRS and 8.3 points

for Constant Score in patients with shoulder conditions (Hao et al. 2019), and

0.03 points for 15D (Alanne et al. 2015). Recently, the MCID of 6.7 (95% CI,

5.0 to 15.8) points in DASH score for patients with humeral shaft fractures has

been reported, but the area under the curve was only 0.66 (95% CI, 0.58 to

0.73; sensitivity 45%, specificity 81%), suggesting a poor discrimination

(Mahabier et al. 2017).

Patient acceptable symptom state (PASS)

The PASS is defined as the level of symptoms below which patients consider

themselves well (Tubach et al. 2005). The PASS is often determined with a

question like: “Considering the activities of your daily life and current

27

symptoms, including pain and functional impairment, is your current state

satisfactory?”. It has been suggested that achieving PASS is more important to

patients than having a treatment effect above MCID (Tubach et al. 2006). The

PASS has been estimated to be 43 points for DASH score among patients with

rheumatoid diseases undergoing shoulder surgery (Christie et al. 2011) and 1.5

points for pain among patients after total shoulder arthroplasty (Chamberlain

et al. 2017).

Both the MCID and PASS estimates are dependent on medical

condition and patient population characteristics.

2.7 ADVERSE EVENTS IN HUMERAL SHAFT FRACTURE

TREATMENT

As in any treatment, there is a possibility of adverse events or complications

with all of the treatment options for humeral shaft fractures, and these should

be taken into consideration in the shared decision-making process.

2.7.1 NONSURGICAL TREATMENT

Secondary radial nerve palsy

The risk of SRNP in nonsurgical care is 0.4% (Hendrickx et al. 2020). The

suggested mechanisms for SRNP are manipulation of the fracture during

splinting (Shaw et al. 1967, Bleeker et al. 1991) or capture inside the fracture

callus (Soustelle et al. 1970, Vural et al. 2008, Ravinsky et al. 2020). Surgical

treatment for SRNP after initial nonsurgical care is generally suggested

(Holstein et al. 1963, Shaw et al. 1967, Abdelgawad et al. 2010), even though a

recent review (Vaishya et al. 2019) found only 8 cases with SRNP after

conservative management in the recent literature. Thus, no firm conclusions

can be made regarding the optimal management for these patients.

Nonunion

The most common complication of nonsurgical treatment of humeral shaft

fractures with functional bracing is nonunion. The reported nonunion rates

vary from 0 to 33%, with large variation in the lost to follow-up rate (Table 3).

Three recently published systematic reviews report overall nonunion rate of

15–18% with nonsurgical care (B. van de Wall et al. 2020, Lode et al. 2020,

Sargeant et al. 2020). Nonunion (Fig. 16) often leads to prolonged impairment

and is generally treated with surgery, with over 90% healing rate (Peters et al.

2015, Wiss et al. 2020).

28

Table 3. Nonunion rates of humeral shaft fractures treated with functional bracing.

Publication

Country

Patients

analyzed

Nonunion

rate (%)

Lost to

follow-up

(%)

Sarmiento et al. 1977

USA

51

2

N/A

Balfour et al. 1982

USA

42

2

44

Ricciardi-Pollini et al. 1985

Italy

14

0

N/A

Naver et al. 1986

Denmark

20

10

0

Zagorski et al. 1988

USA

170

2

27

Wallny et al. 1997

Germany

79

6

0

Sarmiento et al. 2000

USA

620

3

33

Fjalestad et al. 2000

Norway

67

9

7

Pehlivan 2002

Turkey

21

0

16

Koch et al. 2002

Switzerland

67

13

9

Toivanen et al. 2005

Finland

93

23

0

Radford Ekholm et al. 2006

Sweden

78

10

0

Rutgers et al. 2006

USA

49

10

6

Broadbent et al. 2010

UK

96

17

13

Denard et al. 2010

USA

63

21

N/A

Ali et al. 2015

UK

138

17

11

Pal et al. 2015

India

66

2

0

Singhal et al. 2015

UK

20

25

0

Harkin et al. 2017

Australia

80

33

17

Westrick et al. 2017

USA

69

23

0

Basa et al. 2020

Turkey

46

13

0

Olson et al. 2020

USA

70

17

0

Serrano et al. 2020

USA

1182

18

20

Fig. 16.

A. A 27-year-old male sustained a humeral shaft

fracture (AO/OTA 12A1c) due to arm wrestling.

B. Atrophic nonunion at 6 months after injury.

C. Healed fracture at 12 months after surgery.

A

B

C

29

Recently, a scoring system for humeral shaft fracture union, the

Radiographic Union Score for Humeral fractures (RUSHU), was introduced to

help in predicting fracture nonunion at 6 weeks after injury (W. Oliver et al.

2019). Each of the four cortices receive a score from 1 to 3 according to

radiographic appearance (1=no callus, 2=nonbridging callus, 3=bridging

callus). Those having a sum score of 7 points or less will end up having a

fracture nonunion with 65% probability (area under the curve = 0.84, 95% CI

0.74 to 0.94). This scoring system has been validated with one external patient

cohort, and it seems to be a promising tool for selecting the patients most likely

to end up having fracture nonunion. Surgery could be offered to individuals

with 7 points or less already 6 weeks after injury (Dekker et al. 2021). Also, the

fracture site mobility at 6 weeks is a fairly accurate predictor for fracture

nonunion (Driesman et al. 2017, Dekker et al. 2021).

Others

Functional bracing has been reported to cause cutaneous problems in 1–5% of

cases (Zagorski et al. 1988, Koch et al. 2002, Pehlivan 2002, Jawa et al. 2006,

Rutgers et al. 2006). These are best avoided by careful skincare with creams

and proper hygiene. Rarely, the fracture can threaten skin integrity, especially

in proximal oblique or spiral fractures, where the deltoid pulls the proximal

humerus in abduction (Woon 2010).

2.7.2 SURGICAL TREATMENT

Secondary radial nerve palsy

The risk of SRNP after surgical care is around 4–9% (Claessen et al. 2015,

Schwab et al. 2018, Hendrickx et al. 2020). At our unit (Helsinki University

Hospital), the rate of SRNP was 7% (23/323) after initial surgical care of the

humeral shaft fracture between 2006 and 2016 (unpublished data).

Interestingly, MIPO and IMN seem to carry a lower risk for SRNP than ORIF

(Beeres et al. 2020, Hendrickx et al. 2020, van de Wall et al. 2021). However,

comparisons between surgical methods are subject to uncertainty due to the

small numbers of participants and nonrandomized settings in most of the

published data. Recommendations for the treatment of SRNP after surgery

vary from watchful waiting (Vaishya et al. 2019) to early exploration (Schwab

et al. 2018), even though up to 96% have been reported to recover from SRNP

without revision surgery (Hendrickx et al. 2020). Careful surgical technique

and handling of the injured arm during surgery cannot be overemphasized, as

SRNP clearly causes limitations in daily activities often for 3–12 months.

Nonunion

The nonunion rate after surgical treatment varies between surgical methods.

In recent meta-analyses combining RCTs and observational studies, the risk

for nonunion was 8.5% with ORIF, 9.0% with IMN, and 1.2–2.0% with MIPO

(Beeres et al. 2020, van de Wall et al. 2021). However, when looking at the

RCTs only, the rate of nonunion was 2/47 (4.2%) with ORIF versus 0/51 (0%)

30

with MIPO (Beeres et al. 2020), and 3/42 (7.1%) with IMN versus 1/45 (2.2%)

with MIPO (van de Wall et al. 2021). The small sample sizes of these trials

cause uncertainty regarding the observed nonunion rate. Furthermore, direct

comparison of the nonunion risk with surgical and nonsurgical care should not

be done based on these results since the patients treated in surgical trials

usually have more severe injuries than those in observational studies with

functional bracing.

Infection

The reported infection rate with surgical care varies substantially depending

on the injury characteristics of the studies. In a meta-analysis of 14 RCTs and

comparative studies between ORIF and IMN, the rate of infection was 17/357

(4.8%) with ORIF and 6/370 (1.6%) with IMN (Dai et al. 2014). In a relatively

large single-center cohort with 102 patients treated with ORIF (16% with open

fracture), only one patient with open fracture had a deep surgical site infection

and no superficial surgical site infections were observed (B. J. M. van de Wall

et al. 2020).

Others

The incidence of shoulder impairment is higher in patients treated with

antegrade IMN (13%) than in those treated with ORIF (1%) (Dai et al. 2014).

Restriction in elbow range of movement has been reported to be higher after

ORIF (9%) than after antegrade IMN (0%) (Chapman et al. 2000). Injury to

the musculocutaneous nerve has been described during nailing of the humeral

shaft fracture (Blyth et al. 2003). Brachial artery injury is a very rare

complication during humeral shaft fracture surgery (Kumar et al. 2013,

Kyurkchiev et al. 2020). This rare injury has been noted also in conjunction

with nonsurgical care of humeral shaft fracture (Kemp et al. 2014).

2.8 SURGERY VERSUS NONSURGICAL TREATMENT IN

HUMERAL SHAFT FRACTURE STUDIES

Recent meta-analyses summarized the current comparative studies between

nonsurgical and surgical care of humeral shaft fractures (B. van de Wall et al.

2020, Lode et al. 2020, Sargeant et al. 2020). Three RCTs have been published

on this subject prior to the publication of the FISH trial. One is from Brazil

(Matsunaga et al. 2017), one from India (Kumar et al. 2017), and one from Iran

(Hosseini Khameneh et al. 2019). The trial published by Matsunaga et al. was

properly registered in a trial registry and had a protocol publication with

predefined outcome measures and a sample size calculation (Matsunaga et al.

2013). The results of these 3 RCTs and 14 observational studies are

summarized in Table 4.

In summary, there is one RCT comparing MIPO and functional bracing

with proper predefined outcome measures (Matsunaga et al. 2017) that

showed no clinically meaningful difference in DASH at 6 months. Of the

31

patients randomized to functional bracing, 15% subsequently underwent

surgical care due to fracture nonunion. The other two RCTs have inherent

methodological flaws, and the results of these two trials should be interpreted

with caution.

The observational studies have abundant variation in outcome

measures, follow-up time points, and reporting of adverse events. None of the

observational studies refer to a predefined statistical analysis plan that was

published before the actual study. The nonsurgical and surgical groups are

often not comparable due to the retrospective nature of the studies and the

indication for surgery has usually been at the discretion of the treating

surgeon. In many of the studies, the patients in surgical groups had more high-

energy injuries and open fractures, making comparison between treatment

methods unreliable. Generally, the nonunion rates are higher with nonsurgical

care and there is a trend towards a better functional result at least in the early

phase of recovery. The patients with surgical care have a higher rate for

additional procedures for reasons other than nonunion (B. van de Wall et al.

2020).

At the time of planning of the FISH trial (2011–2012), there was no

properly executed trial comparing ORIF and functional bracing, the most

common treatment methods in, for instance, Finland and the USA, in the

treatment of closed humeral shaft fractures. The FISH trial (Study II) was the

first of its kind at the time of its publication (Obremskey 2021).

32

(Stewart et al. 1955, Neer 1970, Kwasny et al. 1990, Wallny et al. 1997,a, Klestil

et al. 1997, Osman et al. 1998, Jawa et al. 2006, Ekholm et al. 2008,a,

Broadbent et al. 2010, Denard et al. 2010, Ristić et al. 2011, van Middendorp

et al. 2011, Mahabier et al. 2013, Matsunaga et al. 2017, Westrick et al. 2017,

Dielwart et al. 2017, Harkin et al. 2017, Kumar et al. 2017, Hosseini Khameneh

et al. 2019)

Table 4. Summary of studies comparing surgery and nonsurgical care in humeral shaft fractures.

33

Table 4 continued. Summary of studies comparing surgery and nonsurgical care in humeral shaft fractures.

34

3 RESEARCH QUESTIONS

The aims, research questions, and hypotheses were as follows:

I The first aim was to plan and describe the methods used in the RCT

comparing surgical and nonsurgical care in the treatment of closed

humeral shaft fractures in adults.

II Study question 1: What is the effectiveness of surgical versus

nonsurgical treatment of closed humeral shaft fractures in adult

patients?

Hypothesis: There is no clinically meaningful difference in the

outcomes of closed humeral shaft fractures in adults between surgical

and nonsurgical care as measured with the DASH score at 12 months

after injury.

III Study question 2: Is there a difference in the outcomes of patients who

underwent secondary surgery to promote healing of their humeral

shaft fracture compared with patients who had successful fracture

healing regardless of initial treatment method at 2 years after injury?

Hypothesis: There is no clinically meaningful difference in the

outcomes at 2 years after injury.

35

4 PATIENTS AND METHODS

4.1 PATIENTS

4.1.1 RCT COMPARING THE EFFECTIVENESS OF SURGICAL AND

NONSURGICAL TREATMENTS OF HUMERAL SHAFT

FRACTURES IN ADULTS (I, II)

Adult patients with humeral shaft fractures treated at the Helsinki and

Tampere University Hospitals between November 2012 and January 2018

were screened for eligibility. Inclusion criteria were acute (10 days or less),

unilateral, displaced, closed humeral shaft fracture. Patients with history or

condition affecting the function of the injured upper limb, pathological

fracture, other concomitant injury affecting the same upper limb, other

fracture, internal organ, brachial plexus or vascular injury requiring surgery,

inadequate cooperation, multimorbidity, or polytrauma were excluded. The

fracture had to lie between the upper border of pectoralis major insertion and

the line 5 cm above the upper border of the olecranon fossa (Fig. 17A). Fracture

type in which the deltoid muscle pulls the proximal part in abduction and the

pectoralis major pulls the distal part medially was excluded (Fig. 17B).

The suspected diagnosis was confirmed with X-ray of the arm. A

surgeon member of the FISH study group (Appendix 1) screened the patient

for eligibility and after receiving written consent the patient was randomized

to either surgical or nonsurgical treatment.

A

B

Fig. 17. A. The fracture had to be

between the level of the upper

border of pectoralis major tendon

insertion and the line 5 cm

proximally to the upper border of

the olecranon fossa (blue area).

B. The excluded fracture type

where the deltoid muscle abducts

the proximal part and pectoralis

major pulls the distal part

36

A total of 321 patients were assessed for eligibility. After excluding 181

noneligible patients (Table 5) and 58 patients not willing to undergo

randomization, 82 patients remained to be randomized. Forty-two of the 58

patients not willing to participate in the RCT consented to be in the concurrent

prospective cohort (‘declined cohort’). The cohort was introduced to further

support the results of the RCT. Table 6 shows the baseline characteristics of

the randomized and the declined cohort. Originally, we planned to evaluate

the results of the noneligible patients (I), but due to poor adherence to the 12-

month follow-up and large variation in reasons for exclusion, comparisons

between the treatment methods proved unfeasible.

Table 5. Reasons for exclusion of the 181 patients in the FISH trial. Some of the patients

had more than one exclusion reason.

Reason for exclusion

Number of patients

Too proximal fracture

56

Too distal fracture

35

Compliance problem

30

Significant health problem

30

Other trauma affecting the same upper limb

14

History of older trauma or disease affecting the same upper limb

13

Polytrauma

10

Language problem

8

Pathological fracture

7

Open fracture

6

Fracture between deltoid and pectoralis major attachment

5

Other fracture warranting operation

4

Periprosthetic fracture

3

Advancing radial nerve palsy

3

Needs walking aid

3

Foreign patient

2

Fracture not dislocated enough

2

Plexus injury

1

Fracture older than 10 days

1

Bilateral fracture

0

Floating shoulder

0

Floating elbow

0

Vascular injury

0

Total

233

37

Table 6. Baseline variables of the RCT and the cohort not willing to undergo randomization

(declined cohort).

Characteristic

Surgery group

(n=38)

Bracing group

(n=44)

Declined cohort

(n=42)

Mean age, years (SD1)

[range]

49.6 (18.2)

[19 to 81]

48.4 (16.2)

[19 to 80]

44.6 (17.3)

[20 to 83]

Female/male

18/20

20/24

16/26

Mean weight, kg (SD)

83.5 (21.2)

85.0 (15.6)

84.4 (17.2)

Mean height, cm (SD)

173 (9)

174 (9)

175 (10)

Mean body mass index,

kg/m2 (SD)

27.7 (5.9)

28.1 (4.1)

27.4 (4.4)

Smoker, n (%)

12 (31.6)

9 (20.4)

10 (23.8)

Radial nerve palsy, n (%)

3 (7.9)

2 (4.5)

3 (7.1)

AO/OTA classification, type

A (simple), n (%)

34 (89.5)

36 (81.8)

30 (68.2)

B (wedge fragment), n (%)

4 (10.5)

7 (15.9)

11 (25.0)

C (segmental), n (%)

0

1 (2.3)

1 (2.2)

Fracture location

Proximal shaft, n (%)

2 (5.3)

5 (11.4)

2 (4.8)

Mid-shaft, n (%)

35 (92.1)

37 (84.1)

32 (76.2)

Distal shaft, n (%)

1 (2.6)

2 (4.5)

8 (19.0)

Injury mechanism

Low energy, n (%)

34 (89.5)

38 (86.4)

38 (90.5)

High energy, n (%)

4 (10.5)

6 (13.6)

4 (9.5)

Dominant limb injured, n

(%)

20 (52.6)

18 (40.9)

22 (52.4)

Median pre-injury DASH

score, (IQR2)

0.0 (0.0 to 2.5)

0.4 (0.0 to 2.5)

0.4 (0.0 to 1.9)

Mean pre-injury DASH

work module score, (SD) [n]

0 (0.0) [26]

0.2 (1.2) [27]

0.2 (1.2) [28]

Mean pre-injury DASH

sports / performing arts

module score, (SD) [n]

0 (0.0) [23]

0.3 (1.4) [19]

1.0 (3.4) [26]

Mean pre-injury 15D score,

(SD)

0.95 (0.05)

0.94 (0.05)

0.94 (0.09)

1 standard deviation

2 interquartile range

38

4.1.2 TWO-YEAR FOLLOW-UP OF THE FISH TRIAL COMPARING

PATIENTS WITH SECONDARY SURGERY WITH PATIENTS WITH

SUCCESSFUL FRACTURE HEALING (III)

The 2-year follow-up study consisted of 82 randomized patients of the FISH

trial. The patients were divided into three groups according to the final

treatment method: 1) initial surgery without further surgery (Initial surgery

group), 2) functional bracing with successful healing (Bracing group), and 3)

patients randomized to functional bracing but who had secondary surgery to

promote the healing of their fracture (Secondary surgery group). The number

of patients in each group was 38, 30, and 14, respectively. Table 7 shows the

baseline characteristics of these groups.

Table 7. Baseline variables in the 2-year follow-up study. The patients are divided according

to the final treatment modality.

Initial randomization

in the FISH trial

Randomized to

surgery (N=38)

Randomized to bracing (N=44)

Characteristic

Initial surgery

group

(N=38)

Bracing group

(N=30)

Secondary

surgery group

(N=14)

Mean age at allocation, years (SD1)

[range]

49.6 (18.2)

[19 to 81]

44.8 (16.7)

[19 to 80]

56.0 (12.4)

[34 to 74]

Female/male

18/20

13/17

6/8

Mean weight, kg (SD)

83.5 (21.2)

87.4 (16.4)

80.0 (13.0)

Mean height, cm (SD)

173 (9)

174 (9)

172 (11)

Mean body mass index, kg/m2 (SD)

27.7 (5.9)

28.6 (4.2)

26.9 (3.7)

Smoker, n (%)

12 (31.6)

5 (16.7)

4 (28.6)

Radial nerve palsy, n (%)

3 (7.9)

1 (3.3)

1 (7.1)

AO/OTA classification, type

A (simple), n (%)

34 (89.5)

26 (86.7)

10 (71.4)

B (wedge fragment), n (%)

4 (10.5)

3 (10.0)

4 (28.6)

C (segmental), n (%)

0

1 (3.3)

0

Fracture location

Proximal shaft, n (%)

2 (5.3)

1 (3.3)

4 (28.6)

Mid-shaft, n (%)

35 (92.1)

27 (90.0)

10 (71.4)

Distal shaft, n (%)

1 (2.6)

2 (6.7)

0 (0.0)

Injury mechanism

Low energy, n (%)

34 (89.5)

26 (86.7)

12 (85.7)

High energy, n (%)

4 (10.5)

4 (13.3)

2 (14.3)

Dominant limb injured, n (%)

20 (52.6)

13 (43.3)

5 (35.7)

Median pre-injury DASH score, (IQR2)

0.0 (0.0 to 2.5)

0.8 (0.0 to 3.5)

Mean pre-injury DASH work module

score, (SD) [n]

0 (0.0) [26]

0.3 (1.5) [18]

0 (0.0) [9]

Mean pre-injury DASH sports /

performing arts module score, (SD) [n]

0 (0.0) [23]

0.4 (1.6) [15]

0 (0.0) [4]

Mean pre-injury 15D score, (SD)

0.95 (0.05)

0.96 (0.04)

0.91 (0.06)

1 standard deviation

2 interquartile range

39

4.2 INTERVENTIONS

The interventions in Studies II and III were prespecified in the study protocol

(I).

4.2.1 SURGICAL TREATMENT (II, III)

The patients in the surgery group were operated by or under the supervision

of an experienced orthopedic trauma surgeon within 14 days of injury using

ORIF with 4.5 mm locking compression plate (LCP). The surgical approach,

the mode of the plating (bridging, dynamic compression, or neutralizing), and

the use of either locking or nonlocking screws was left to the discretion of the

treating surgeon. A minimum 10-hole LCP was used with at least 3 bicortical

screws on each side of the fracture segment. Patients were instructed to move

their upper extremity without extra weight after the surgery. Gradual weight-

bearing was introduced after 6 weeks.

In case of secondary surgery, the surgical approach, use of bone

grafting, and fixation method was left to the discretion of the treating surgeon.

All 14 patients with secondary surgery had plate fixation using ORIF.

4.2.2 NONSURGICAL TREATMENT, FUNCTIONAL BRACING (II, III)

Patients in the bracing group had a functional bracing applied by a trained

plaster technician. The brace covered the arm from the shoulder to elbow,

leaving the motion of both joints free (Fig. 18). Patients received written and

verbal instructions on how to tighten the brace as the swelling subsided. The

patients were told to wear the brace until the fracture had healed. The

prespecified rehabilitation protocol followed the protocol by Sarmiento et al.

(2000). Some minor modifications were introduced; the patients were

instructed to start immediate active non-weight-bearing exercises of the elbow

and wrist and pendulum exercises of the shoulder. Assisted exercises of the

shoulder were started at 3 weeks, and gradual weight-bearing at 6 weeks.

Fig. 18. The functional brace can be

tightened as the swelling of the arm

resolves.

40

4.2.3 REHABILITATION (II, III)

A detailed rehabilitation protocol (Table 8) was used in each study group.

Patients met with a physical therapist at 3 and 9 weeks.

Table 8. Rehabilitation protocol used in the FISH trial.

4.3 FOLLOW-UP

The patients had follow-up visits at 6 weeks, 3 and 6 months, and 1 year in

Study II, and at 2 years in Study III. A clinical examination with X-ray was

carried out at all follow-up visits. A physical therapist unaware of the patient

treatment method performed the objective measurements. The flow charts of

the studies are illustrated below (Fig. 19 and Fig. 20).

Weeks

Surgery group

Bracing group

0–3

Active non-weight-bearing exercises of the

upper extremity

Active non-weight-bearing exercises of the

elbow and hand, pendulum exercises of the

shoulder. Patient instructed to tighten the brace

as the swelling subsided.

3–6

Visit to physical therapist at 3 weeks

Previous exercises continued

Visit to physical therapist at 3 weeks

Passive ROM1 exercises of the shoulder started

6–9

Active exercises of the upper extremity

Gradual weight-bearing started

9–12

Visit to physical therapist at 9 weeks

Scapulohumeral rhythm exercises

12–

Free mobilization if no problems with consolidation

1 range of motion

41

181 Excluded (See Table 5)

91 Fracture outside the predefined fracture zone

30 Significant health problem

30 Compliance problem

14 Other trauma affecting the same upper limb

13 History of older trauma or disease affecting the limb

10 Polytrauma

8 Language problem

7 Pathological fracture

6 Open fracture

24 Other

321 Adults with humeral shaft fracture assessed for eligibility

38 Randomized to surgery

38 Received intervention as

randomized

82 Randomized

36 Completed trial

2 lost to follow-up prior to

12-month follow-up

44 Randomized to bracing

44 Received intervention as

randomized

42 Completed trial

2 lost to follow-up prior to

12-month follow-up

38 Included in primary analysis

44 Included in primary analysis

58 Declined randomization

42 Consented to

’declined cohort’

16 Declined participation

33 Preferred bracing

9 Preferred surgery

35 Completed follow-up

7 lost to follow-up prior

to 12-month follow-up,

2 of the surgery group,

5 of the bracing group

42 Included in analysis

Fig. 19. Flow chart of Study II.

42

Fig. 20. Flow chart of Study III.

181 Excluded

91 Fracture outside the predefined fracture zone

30 Significant health problem

30 Cognitive barriers

14 Other trauma affecting the same upper limb

13 History of older trauma or disease af fecting the limb

10 Polytrauma

8 Language problem

7 Pathological fracture

6 Open fracture

24 Other

321 adults with humeral shaft fracture assessed

for eligibility

38 Randomized to surgery

38 Received intervention

as randomized

82 Randomized

44 Randomized to bracing

44 Received intervention as

randomized

30 Healed with bracing

38 Included in the analysis as

the initial surgery group

30 Included in the analysis

as the bracing group

58 Refused randomization

16 Declined participation

9 Preferred surgery

9 Included in the analysis

as the initial surgery

group

5 lost to follow-up (did not

come to 2-y follow-up)

3 lost to follow-up

(did not come to 2-y

follow-up)

5 lost to follow-up

(did not come to

2-y follow-up)

14 Had secondary surgery

0 lost to follow-up

14 Included in the analysis

as the secondary

surgery group

33 Preferred bracing

4 lost to follow-up

(did not come to 2-y

follow-up)

7 Had secondary surgery

26 Healed with

bracing

1 lost to follow-up (did

not come to 2-y follow-

up)

26 Included in the

analysis as the

bracing group

7 Included in the analysis

as the secondary

surgery group

Randomized cohort

Declined cohort

42 Consented to follow-up

140 Eligible for randomization

43

4.4 OUTCOME MEASURES (II, III)

4.4.1 DISABILITIES OF THE ARM, SHOULDER, AND HAND SCORE

The primary outcome was DASH score at 12 months. DASH is a validated,

responsive, widely used PROM for upper limb-related physical function and

symptoms (Hudak et al. 1996, Beaton et al. 2001). The scale ranges from 0 to

100; 0 denotes no disability and 100 extreme disability. The MCID of DASH

has been evaluated to be 10 points (Gummesson et al. 2003). The patients

filled in the questionnaire at baseline, at 6 weeks, at 3, 6, and 12 months, and

at 2 years post-randomization, with time points other than 12 months being

secondary outcomes. As the patients who were recruited already had a broken

humerus at the time of filling in the baseline questionnaires, they were asked

to recall the situation before the fracture. The DASH score can be calculated if

3 or less values out of 30 are missing.

4.4.2 PAIN AT REST AND ON ACTIVITY

The patients reported the pain at rest and on activity using a 11-point scale (0

to 10 NRS); 0 on the far left of the scale denotes ‘no pain’ and 10 on the far

right the ‘worst imaginable pain’; MCID 1.5 (Hao et al. 2019). Pain-NRS is an

easy-to-use, responsive, and validated tool for assessing pain (Ferreira-

Valente et al. 2011). Patients reported the pain they experienced at 6 weeks, at

3, 6, and 12 months, and at 2 years post-randomization.

4.4.3 CONSTANT-MURLEY SCORE

The Constant-Murley score is a widely used instrument for shoulder function.

The instrument contains clinically assessed functions of the shoulder and

patient-reported parts for pain and function in activities of daily living

(Constant et al. 1987). The scale ranges from 0 to 100; 0 denotes extreme

disability and 100 no disability; MCID 8.3 (Hao et al. 2019). Shoulder flexion,

abduction, and external rotation angle was measured using a goniometer;

shoulder abduction strength was measured using a calibrated spring balance.

The measurements were done by a physical therapist unaware of the treatment

allocation. The patients wore a long-sleeved shirt to hide a possible surgical

scar and were instructed not to reveal the treatment method.

4.4.4 15D QUALITY-OF-LIFE TOOL

15D is a generic health-related quality-of-life tool (Sintonen 2001). The tool

comprises 15 dimensions of various health-related issues such as physical

activity, activities of daily living, and anxiety. Scores range from 1 (no

problems in any dimension) to 0 (patient is dead); MCID 0.03 (Alanne et al.

2015).

44

4.4.5 ELBOW RANGE OF MOTION

The range of motion of elbow was measured by a physical therapist using a

goniometer. The results were reported as a single value (the difference of the

angle between full extension and full flexion). The MCID of elbow range of

motion has been reported to be 14.1° using a distribution-based method and

25.0° using an anchor-based method in patients undergoing elbow arthrolysis

(Sun et al. 2021).

4.4.6 DASH WORK AND SPORTS OR PERFORMING ARTS MODULES

DASH work and sports or performing arts modules are optional modules for

DASH. Modules comprise four questions assessing the effect of medical

condition on work, sports, or performing arts. The scale ranges from 0 to 100;

0 denotes no disability and 100 extreme disability. The score can be calculated

if no values are missing.

4.4.7 GENERAL SATISFACTION

Patients’ satisfaction was measured at all follow-up time points using the

following scores:

-Satisfaction with shoulder condition, 0–10 NRS (0 worst, 10 best)

-Satisfaction with elbow condition, 0–10 NRS (0 worst, 10 best)

-Satisfaction with the entire upper limb, 0–10 NRS (0 worst, 10 best)

The satisfaction was also assessed at 12 months and 2 years by asking whether

the patient would choose the same treatment again if sustaining a similar

humeral shaft fracture (yes or no).

4.4.8 PATIENT ACCEPTABLE SYMPTOM STATE (PASS)

The proportion of patients with acceptable symptom state was determined

using a 7-point Likert scale. The patients were asked “How satisfied are you

with the overall condition of your injured upper limb and its effect on your

daily life?”. The answering options were:

-“Very satisfied”

-“Satisfied”

-“Somewhat satisfied”

-“Neither satisfied nor dissatisfied”

-“Somewhat dissatisfied”

-“Dissatisfied”

-“Very dissatisfied”

We categorized the patients answering “very satisfied” and “satisfied” as

having PASS. All other responses were categorized as not having PASS.

45

4.4.9 CLINICAL RECOVERY

There is no definition in the literature for good clinical recovery after humeral

shaft fracture. We decided to categorize those with a DASH score of pre-injury

DASH + 10 points (MCID) or less as having an adequate clinical recovery (e.g.,

pre-injury DASH 6 points; 16 points or less at follow-ups categorized the

patient as having a good clinical recovery). The rationale behind our novel

definition is that if the primary outcome measure between the current and pre-

injury status is within the limits of MCID, the difference should not be

clinically important.

4.5 ETHICS

The institutional review board of the Helsinki and Uusimaa hospital district

approved the study protocol (I) of the FISH trial (II, III). All patients gave

written informed consent.

4.6 STATISTICAL METHODS

Stata version 15.1 (StataCorp LLC) was used for the statistical analyses. An

independent statistician performed the analyses for Studies II and III based

on a prespecified statistical analysis plan (I).

Study II

Primary statistical analyses were done according to an intention-to-treat

principle. The primary comparison between the treatment groups (surgery

and bracing groups) was performed using a mixed-model repeated-measures

analysis of variance (MMRM ANOVA). Treatment group and time of

assessment (baseline, 6 weeks, 3, 6, and 12 months) were included as fixed

factors and patients as random factors. The model included interactions

between study group and time of assessment. Change from baseline at the

different follow-up time points was estimated with the baseline as a

covariate. The model was used to quantify the treatment effect as the absolute

difference between the group means in DASH score with associated 95%

confidence intervals (CIs) and P-value at 12 months.

A similar model was used to analyze continuous secondary outcomes

(pain at rest and on activity, 15D, Constant-Murley score). For categorical

response variables, effects were analyzed using marginal logistic regression

analysis. Because of the potential for type I error due to multiple comparisons,

findings for analyses of secondary end points were interpreted as exploratory.

The MMRM ANOVA model allows missing data, and no data were thus

imputed. All patients with at least some data were included in the analysis.

In addition, preplanned per-protocol and as-treated analyses were

performed using the same models described above. In the per-protocol

analysis, patients were analyzed in three groups, 1) surgery group, 2) bracing

46

group with no surgery, and 3) bracing group with secondary surgery due to

healing problem. In the as-treated analysis, patients were analyzed according

to their latest treatment modality during the follow-up visits (the number of

patients increased in the surgery group and decreased in the bracing group as

the patients randomized to bracing had secondary surgery during the follow-

up). The data of the declined cohort were analyzed according to the same

principles as the data of the actual RCT. The threshold for statistical

significance was set at the level of 0.05 with two-sided testing.

Study III

The results of the 12-month follow-up revealed that nearly 1/3 of the patients

randomized to functional bracing had undergone late surgery due to healing

problems, and the results of these patients were inferior to those with

successful healing, irrespective of the treatment method (surgery or functional

bracing). This prompted us to make a pre-specified analysis plan for the 2-year

follow-up data, and the plan was published as a supplement to Study III. We

decided to analyze the patients in three groups (i.e., per-protocol analysis):

initial surgery group, bracing group with successful healing, and a secondary

surgery group with patients randomized to functional bracing but who had

secondary surgery due to a fracture healing problem. Similar MMRM ANOVA

was used as in Study II: Study group, time of assessment (baseline, 6 weeks, 3,

6, and 12 months, and 2 years), and study site were included as fixed factors,

patients as random factors. For categorical response variables, we used

Fisher’s exact test due to convergence problems with logistic regression

analysis.

In addition, intention-to-treat and as-treated analyses were performed

according to same principles as in Study II. Also, the results of the declined

cohort were analyzed according to the per-protocol principle described above.

Randomization

Consenting patients were randomized using a block size of 4 by a closed

envelope method. Both study centers had separate randomization lists

generated by an independent biostatistician. Stratification was used according

to radial nerve status (intact, paresthesia, or mild motor deficit; or either

subtotal or total motor palsy) and fracture type (AO/OTA type A [simple

fracture] or either type B [separate wedge fragment] or type C [segmental

fracture], Fig. 8 and Fig. 9). The rationale behind stratification according to

these factors was the assumption of PRNP affecting the early functional results

and the previous finding of type A fractures having a higher risk for a fracture

nonunion (Ekholm 2007, Papasoulis et al. 2010).

47

Sample size

The study was powered to detect a 10-point difference (Gummesson et al.

2003) in the primary outcome DASH with a standard deviation of 14.7 points

(Hunsaker et al. 2002). With a two-sided error rate of 5% and 80% power, 35

participants per group were needed. With an anticipated 12.5% attrition, 80

patients were planned to be recruited.

48

5 RESULTS

5.1 ONE-YEAR FOLLOW-UP (II)

A total of 82 patients were randomized, with 78 completing the 12-month

follow-up: 36/38 patients of the surgery group and 42/44 of the bracing group

(Fig. 19).

Primary outcome

The mean DASH score was 8.9 points (95% CI, 4.2 to 13.6) in the surgery group

and 12.0 points (95% CI, 7.7 to 16.4) in the bracing group at 12 months. The

between-group mean difference was -3.1 points (surgery minus bracing; 95%

CI, -9.6 to 3.3) in favor of surgery, but the difference was neither statistically

significant nor clinically meaningful (Fig. 21).

Fig. 21. The DASH scores in the intention-to-treat analysis. The error bars indicate 95% CIs

of the mean estimates.

Secondary outcomes

The results of the secondary outcomes at all time points are given in Table 9

and in Fig. 22. There was a statistically significant difference in the DASH

score in favor of surgery at 6 weeks (between-group mean difference; -9.9

points; 95% CI, -16.3 to -3.5) and at 3 months (between-group mean

difference; -10.1 points; 95% CI, -16.6 to -3.6), the latter reaching the pre-

specified level of MCID. At other time points, the difference was neither

statistically nor clinically significant.

Before

Fracture

612 26 52

0

20

40

60

80

100

Follow-up, weeks

post-randomization

DASH score (0 = no disability)

No. of patients

reporting DASH

Surgery

Bracing

37

44

37

43

34

42

34

43

36

42

Surgery

Bracing

49

The pain-NRS on activities was 4.4 points (95% CI, 3.6 to 5.2) in the

surgery group and 5.6 points (95% CI, 4.8 to 6.3) in the bracing group, with a

mean difference of -1.2 points (95% CI, -2.3 to -0.1) between the groups at 6

weeks. The statistically significant difference was below the level of MCID (1.5

points) and was not clinically important.

The Constant-Murley score showed a statistically significant and

clinically meaningful difference between the groups at 6 weeks (30.7 points;

95% CI, 22.8 to 38.7), at 3 months (14.9 points; 95% CI, 6.9 to 22.9), and at 6

months (8.8 points; 95% CI, 0.8 to 16.9) in favor of surgery.

The proportion of patients having an adequate clinical recovery was

86% (95% CI, 74% to 98%) in the surgery group and 73% (95% CI, 59% to 87%)

in the bracing group at 12 months. The difference was not statistically

significant (difference 13%; 95% CI, -5% to 31%). The proportion of patients

willing to have same the treatment again was 97% (95%, 91% to 100%) in the

surgery group and 71% (95% CI, 58% to 85%) in the bracing group, with a

significant difference of 26% (95% CI, 11% to 40%) between the groups.

Table 9. Outcomes of the 1-year follow-up at all time points.

Outcome

Surgery group

(N=38)

mean (95% CI)

Bracing group

(N=44)

mean (95% CI)

Between-group

mean difference

(95% CI)

6 weeks

DASH score

39.8 (35.1 to 44.5)

49.7 (45.4 to 54.0)

-9.9 (-16.3 to -3.5)

Pain at rest

2.1 (1.5 to 2.7)

1.9 (1.4 to 2.5)

0.2 (-0.6 to 0.9)

Pain on activities

4.4 (3.6 to 5.2)

5.6 (4.8 to 6.3)

-1.2 (-2.3 to -0.1)

Constant-Murley score

53.3 (47.5 to 59.2)

22.6 (17.2 to 28.0)

30.7 (22.8 to 38.7)

Elbow ROM – degrees

125 (119 to 131)

96 (91 to 101)

29 (21 to 37)

15D score

0.88 (0.86 to 0.90)

0.85 (0.83 to 0.87)

0.03 (-0.01 to 0.07)

DASH work module score

61.9 (51.3 to 72.4)

83.6 (71.7 to 95.5)

-21.7 (-37.6 to -5.8)

DASH sports/performing arts

module score

77.6 (65.4 to 89.9)

99.6 (85.7 to 100)

-21.9 (-40.5 to -3.4)

Patients with acceptable

symptomatic state – %

24 (10 to 38)

11 (1 to 21)

12 (-4 to 28)

Adequate clinical recovery – %

6 (0 to 14)

2 (0 to 6)

3 (-5 to 11)

Satisfaction with shoulder

function

7.1 (6.3 to 7.8)

5.8 (5.1 to 6.5)

1.2 (0.2 to 2.3)

Satisfaction with elbow

function

7.3 (6.6 to 7.9)

6.5 (5.8 to 7.1)

0.8 (-0.1 to 1.7)

Satisfaction with upper limb

function

6.5 (5.8 to 7.3)

4.6 (3.9 to 5.4)

1.9 (0.8 to 3.0)

Patients able to return to

previous daily activities – %

68 (52 to 84)

66 (52 to 80)

3 (-17 to 23)

Patients able to return to

previous hobbies – %

16 (4 to 28)

11 (1 to 21)

5 (-11 to 21)

50

Table 9 continued. Outcomes of the 1-year follow-up at all time points.

3 months

Surgery group

Bracing group

Mean difference

DASH score

23.8 (18.9 to 28.6)

33.8 (29.5 to 38.1)

-10.1 (-16.6 to -3.6)

Pain at rest

1.5 (1.0 to 2.1)

1.3 (0.7 to 1.8)

0.3 (-0.5 to 1.1)

Pain on activities

3.5 (2.7 to 4.4)

4.3 (3.5 to 5.1)

-0.8 (-1.9 to 0.4)

Constant-Murley score

61.9 (56.0 to 67.8)

46.9 (41.5 to 52.3)

14.9 (6.9 to 22.9)

Elbow ROM – degrees

134 (129 to 140)

121 (115 to 126)

14 (6 to 21)

15D score

0.91 (0.89 to 0.93)

0.88 (0.86 to 0.90)

0.03 (-0.01 to 0.07)

DASH work module score

33.0 (22.5 to 43.6)

45.2 (33.3 to 57.2)

-12.2 (-28.2 to 3.8)

DASH sports/performing arts

module score

55.5 (42.1 to 68.9)

83.0 (69.5 to 96.6)

-27.5 (-46.6 to -8.4)

Patients with acceptable

symptomatic state – %

46 (30 to 62)

20 (8 to 32)

26 (6 to 46)

Adequate clinical recovery – %

23 (9 to 37)

11 (1 to 21)

12 (-6 to 30)

Satisfaction with shoulder

function

6.8 (6.0 to 7.6)

6.1 (5.4 to 6.8)

0.6 (-0.4 to 1.7)

Satisfaction with elbow

function

7.6 (6.9 to 8.3)

7.4 (6.8 to 8.1)

0.2 (-0.8 to 1.1)

Satisfaction with upper limb

function

7.1 (6.3 to 7.9)

5.4 (4.7 to 6.2)

1.6 (0.6 to 2.7)

Patients able to return to

previous daily activities – %

81 (67 to 95)

78 (66 to 90)

3 (-15 to 21)

Patients able to return to

previous hobbies – %

40 (24 to 56)

27 (13 to 41)

12 (-10 to 34)

6 months

DASH score

13.5 (8.7 to 18.3)

18.4 (14.1 to 22.7)

-4.9 (-11.3 to 1.6)

Pain at rest

1.0 (0.4 to 1.6)

0.7 (0.1 to 1.2)

0.3 (-0.5 to 1.1)

Pain on activities

2.4 (1.6 to 3.2)

2.5 (1.7 to 3.3)

-0.1 (-1.2 to 1.0)

Constant-Murley score

73.1 (67.1 to 79.0)

64.3 (58.9 to 69.7)

8.8 (0.8 to 16.9)

Elbow ROM – degrees

139 (133 to 145)

133 (127 to 138)

6 (-2 to 14)

15D score

0.93 (0.91 to 0.95)

0.91 (0.89 to 0.93)

0.02 (-0.02 to 0.06)

DASH work module score

12.3 (1.8 to 22.9)

25.4 (14.9 to 36.0)

-13.1 (-28.1 to 1.8)

DASH sports/performing arts

module score

19.4 (6.8 to 32.1)

37.4 (24.6 to 50.2)

-18.0 (-36.0 to 0.0)

Patients with acceptable

symptomatic state – %

71 (55 to 87)

51 (35 to 67)

21 (-1 to 43)

Adequate clinical recovery – %

66 (50 to 82)

48 (32 to 64)

17 (-5 to 39)

Satisfaction with shoulder

function

8.3 (7.5 to 9.1)

7.0 (6.3 to 7.7)

1.3 (0.3 to 2.4)

Satisfaction with elbow

function

8.9 (8.2 to 9.6)

8.0 (7.3 to 8.6)

0.9 (0 to 1.9)

Satisfaction with upper limb

function

8.4 (7.6 to 9.2)

6.8 (6.1 to 7.6)

1.5 (0.5 to 2.6)

Patients able to return to

previous daily activities – %

95 (87 to 100)

93 (85 to 100)

2 (-8 to 12)

Patients able to return to

previous hobbies – %

73 (57 to 89)

60 (46 to 74)

13 (-9 to 35)

51

Table 9 continued. Outcomes of the 1-year follow-up at all time points.

12 months

Surgery group

Bracing group

Mean difference

DASH score

8.9 (4.2 to 13.6)

12.0 (7.7 to 16.4)

-3.1 (-9.6 to 3.3)

Pain at rest

0.9 (0.4 to 1.5)

0.7 (0.1 to 1.2)

0.3 (-0.5 to 1.1)

Pain on activities

2.2 (1.4 to 3.0)

1.7 (1.0 to 2.5)

0.5 (-0.7 to 1.6)

Constant-Murley score

78.1 (72.1 to 84.0)

76.4 (70.9 to 81.8)

1.7 (-6.4 to 9.8)

Elbow ROM – degrees

143 (138 to 149)

137 (131 to 142)

7 (-1 to 15)

15D score

0.95 (0.93 to 0.97)

0.92 (0.90 to 0.94)

0.03 (-0.01 to 0.07)

DASH work module score

5.2 (0 to 15.3)

8.0 (0 to 18.4)

-2.9 (-17.4 to 11.6)

DASH sports/performing arts

module score

6.7 (0 to 19.3)

27.9 (15.4 to 40.3)

-21.2 (-38.9 to -3.4)

Patients with acceptable

symptomatic state – %

82 (70 to 94)

68 (54 to 82)

14 (-6 to 34)

Adequate clinical recovery – %

86 (74 to 98)

73 (59 to 87)

13 (-5 to 31)

Satisfaction with shoulder

function

8.5 (7.7 to 9.3)

8.0 (7.3 to 8.7)

0.5 (-0.6 to 1.6)

Satisfaction with elbow

function

9.0 (8.3 to 9.7)

8.8 (8.2 to 9.4)

0.2 (-0.7 to 1.2)

Satisfaction with upper limb

function

8.6 (7.8 to 9.4)

7.6 (6.9 to 8.4)

1.0 (-0.1 to 2.1)

Patients able to return to

previous daily activities – %

96 (90 to 100)

91 (83 to 99)

5 (-5 to 15)

Patients able to return to

previous hobbies – %

85 (73 to 97)

80 (68 to 92)

5 (-13 to 23)

Patients willing to repeat the

same treatment – %

97 (91 to 100)

71 (58 to 85)

26 (11 to 40)

52

Fig. 22. Trajectories of secondary outcomes in Study II.

612 26 52

0

20

40

60

80

100

Weeks

Constant-Murley score

612 26 52

0

20

40

60

80

100

Weeks

% in PASS

612 26 52

0

2

4

6

8

10

Weeks

Satisfaction with upper limb function

(NRS 0–10)

612 26 52

0

2

4

6

8

10

Weeks

Pain on activities (NRS 0–10)

Before

Fracture

612 26 52

0.5

0.6

0.7

0.8

0.9

1.0

Weeks

15D score

612 26 52

0

20

40

60

80

100

Weeks

Adequate clinical recovery - %

612 26 52

0

2

4

6

8

10

Weeks

Pain at rest (NRS 0–10)

Surgery

Bracing

53

Adverse events and reasons for late surgery

Two serious adverse events occurred. One patient in the surgery group had a

cardiac arrhythmia warranting cardioversion directly after the surgery and

one patient in the bracing group had pulmonary embolism at 4 weeks. Eleven

patients in the bracing group had late surgery, with fracture nonunion being

the most common reason (8 patients). All adverse events and reasons for late

surgery are shown in Table 10. The mean DASH score of the 13 patients with

delayed surgery due to healing problem was 20.0 (95% CI, 12.3 to 27.7) at 12

months (4.5 to 12 months after surgery), exceeding the MCID of 10 points,

compared with patients randomized to surgery (8.9 points; 95% CI, 4.3 to

13.6) or those with successful healing with bracing (8.5 points; 95% CI, 3.4 to

13.6).

Table 10. Adverse events and reasons for late surgery.

Description

Surgery group

(N=38)

Bracing group

(N=44)

Serious adverse event

Cardiovascular event

1

Minor adverse event

Fracture nonunion

0

11

Refracture

0

1

Secondary temporary radial nerve palsy

3

1

Superficial wound infection

2

1

Wound seroma

1

0

Shoulder adhesive capsulitis

1

Loss of reduction

0

1

Sensory disturbance in forearm

0

1

Reason for late surgery

Nonunion (time range 3–7.5 months)

8

Loss of reduction (at 6 weeks)

1

Refracture (at 8 months)

1

Intolerable pain in fracture site (at 1 week)

1

Not tolerating the bracing (at 1 and 6 weeks)

2

Declined cohort

The DASH scores of the declined cohort are given in Table 11. There was a

clinically meaningful but statistically insignificant between-group mean

difference of -13.6 points (95% CI, -27.3 to 0.2) in favor of the surgery group

at 6 weeks only. The nonunion rate was 2/9 (22%) in the surgery group and

6/33 (18%) in the bracing group, with all healing after the secondary surgery.

54

Table 11. DASH scores of the declined cohort.

DASH

score

Surgery group (n=9)

mean

(95% CI)

Bracing group (n=33)

mean

(95% CI)

Between-group

mean difference

(95% CI)

Baseline

4.8 (0 to 15.5)

3.0 (0 to 8.5)

1.8 (-10.3 to 13.8)

6 weeks

36.7 (24.2 to 49.1)

50.2 (44.5 to 56.0)

-13.6 (-27.3 to 0.2)

3 months

28.1 (15.6 to 40.5)

28.0 (22.2 to 33.9)

0.04 (-13.7 to 13.8)

6 months

18.3 (5.9 to 30.8)

13.5 (7.6 to 19.5)

4.8 (-9.0 to 18.6)

12 months

13.8 (2.1 to 25.6)

10.3 (4.4 to 16.1)

3.6 (-9.5 to 16.7)

5.2 TWO-YEAR FOLLOW-UP (III)

In the primary analysis of the 2-year follow-up, the patients were analyzed in

three groups as specified in the Methods section above (page 46). Seventy-four

of the 82 randomized patients completed the follow-up, with 33 in the initial

surgery group, 27 in the bracing group with successful healing, and 14 in the

secondary surgery group (Fig. 20).

Primary outcome

The DASH score was 6.8 points (95% CI, 2.3 to 11.4) in the initial surgery

group, 6.0 points (95% CI, 1.0 to 11.0) in the bracing group, and 17.5 points

(95% CI, 10.5 to 24.5) in the secondary surgery group at 2 years (Fig. 23). There

was a statistically significant and clinically meaningful difference between the

initial surgery and the secondary surgery groups (initial surgery minus

secondary surgery, -10.7 points; 95% CI, -19.1 to -2.3) and between the bracing

and the secondary surgery groups (bracing minus secondary surgery, -11.5

points; 95% CI, -20.1 to -2.9). The difference was not statistically significant

between the initial surgery and bracing groups (initial surgery minus bracing,

0.8 points; 95% CI, -6.0 to 7.6).

Fig. 23. The DASH scores of

the 2-year follow-up (Study III).

The error bars indicate 95% CIs

of the mean estimates.

Before

Fracture

612 26 52 104

0

20

40

60

80

100

Weeks

DASH score (0 = no disability)

No. of patients

reportin g DASH

Surgery

Bracing

Crossovers

37

30

14

37

29

14

34

29

13

34

30

13

36

28

14

33

27

14

Initial surgery

Bracing

Secondary surgery

55

Secondary outcomes

The secondary outcomes are shown in Fig. 24 and Table 12. There was a

statistically significant and clinically meaningful difference between the

secondary surgery group and either the initial surgery or the bracing group in

many of the outcomes, in favor of the latter two groups.

612 26 52 104

0

20

40

60

80

100

Week s

Constant-Murley score

612 26 52 104

0

20

40

60

80

100

Week s

% in PASS

612 26 52 104

0

2

4

6

8

10

Week s

Pain at rest (NRS 0–10)

Before

Fracture

612 26 52 104

0.5

0.6

0.7

0.8

0.9

1.0

Weeks

15D score

612 26 52 104

0

20

40

60

80

100

Week s

Adequate clinical recovery - %

612 26 52 104

0

2

4

6

8

10

Week s

Pain on activities (NRS 0–10)

Fig. 24. Trajectories of secondary outcomes of the 2-year follow-up.

Initial surgery

Bracing

Secondary surgery

56

612 26 52 104

0

20

40

60

80

100

Week s

Return to activities of daily living - %

612 26 52 104

0

2

4

6

8

10

Week s

Satisfaction with shoulder function

(NRS 0–10)

612 26 52 104

60

80

100

120

140

160

Week s

Elbow ROM (degrees)

612 26 52 104

0

20

40

60

80

100

Week s

Return to hobbies - %

612 26 52 104

0

2

4

6

8

10

Week s

Satisfaction with elbow function

(NRS 0–10)

612 26 52 104

0

2

4

6

8

10

Week s

Satisfaction with upper limb function

(NRS 0–10)

Fig. 24 continued. Trajectories of secondary outcomes of the 2-year follow-up.

Initial surgery

Bracing

Secondary surgery

57

Table 12. Outcomes of the 2-year follow-up at all time points.

Randomized to

surgery (N=38)

Randomized to functional bracing

(N=44)

Outcome

Initial surgery

group

(N=38)

mean (95% CI)

Bracing group

(N=30)

mean (95% CI)

Secondary

surgery group

(N=14)

mean (95% CI)

Between-group

mean difference

Initial surgery – Bracing

(95% CI)

Between-group

mean difference

Initial surgery – Secondary

surgery (95% CI)

Between-group

mean difference

Bracing – Secondary

surgery (95% CI)

6 weeks

DASH score

39.8 (35.4 to 44.1)

47.9 (43.0 to 52.7)

53.3 (46.2 to 60.3)

-8.1 (-14.7 to -1.5)

-13.5 (-21.8 to -5.3)

-5.4 (-14.0 to 3.2)

Pain at rest

2.1 (1.6 to 2.6)

1.8 (1.2 to 2.4)

2.2 (1.3 to 3.1)

0.3 (-0.5 to 1.1)

-0.1 (-1.1 to 1.0)

-0.4 (-1.4 to 0.7)

Pain on activities

4.4 (3.7 to 5.2)

5.1 (4.3 to 5.9)

6.6 (5.3 to 7.9)

-0.7 (-1.8 to 0.5)

-2.2 (-3.6 to -0.7)

-1.5 (-3.0 to 0.0)

Constant-Murley score

53.3 (47.9 to 58.7)

23.7 (17.7 to 29.7)

20.3 (11.5 to 29.1)

29.6 (21.6 to 37.7)

33.0 (22.7 to 43.3)

3.4 (-7.3 to 14.0)

Elbow ROM – degrees

125 (120 to 130)

100 (94 to 106)

88 (79 to 96)

25 (17 to 33)

37 (27 to 47)

12 (2 to 22)

15D score

0.85 (0.81 to 0.89)

0.87 (0.85 to 0.89)

0.82 (0.78 to 0.86)

-0.02 (-0.08 to 0.04)

0.02 (-0.04 to 0.08)

0.04 (-0.02 to 0.10)

DASH work module score

62.5 (53.0 to 72.0)

77.2 (64.0 to 90.3)

95.3 (76.7 to 100)

-14.7 (-30.9 to 1.6)

-32.8 (-53.7 to -11.9)

-18.1 (-40.9 to 4.6)

DASH sports/performing arts module score

78.6 (67.8 to 89.4)

95.4 (82.1 to 100)

100 (78.4 to 100)

-16.8 (-34.1 to 0.5)

-30.6 (-63.0 to -1.8)

-13.8 (-47.4 to 19.9)

Patients with acceptable symptomatic state – %

Group comparisons: relative risk ratio, RRR (95% CI)

24 (11 to 40)

13 (4 to 31)

7 (0 to 34)

RRR 1.78

(0.61 to 5.21)

RRR 3.32

(0.46 to 23.85)

RRR 1.87

(0.23 to 15.21)

Adequate clinical recovery – %

Group comparisons: RRR (95% CI)

6 (1 to 19)

3 (0 to 18)

0 (0 to 23)

RRR 1.61 (0.15 to 16.90)

∞

Satisfaction with shoulder function

7.1 (6.4 to 7.7)

6.1 (5.3 to 6.9)

5.2 (4.1 to 6.3)

0.9 (-0.1 to 2.0)

1.9 (0.5 to 3.2)

0.9 (-0.5 to 2.3)

Satisfaction with elbow function

7.3 (6.6 to 7.9)

6.4 (5.7 to 7.1)

6.6 (5.6 to 7.6)

0.9 (-0.1 to 1.8)

0.6 (-0.5 to 1.8)

-0.2 (-1.5 to 1.0)

Satisfaction with upper limb function

6.5 (5.8 to 7.2)

4.9 (4.1 to 5.7)

4.0 (2.8 to 5.2)

1.6 (0.5 to 2.7)

2.5 (1.2 to 3.9)

1.0 (-0.5 to 2.4)

Patients able to return to activities of daily living – %

Group comparisons: RRR (95% CI)

68 (51 to 82)

67 (47 to 83)

64 (35 to 87)

RRR 1.03 (0.74 to 1.43)

RRR 1.06 (0.68 to 1.66)

RRR 1.04 (0.65 to 1.65)

Patients able to return to previous hobbies – %

Group comparisons: RRR (95% CI)

16 (6 to 32)

17 (6 to 35)

0 (0 to 23)

RRR 0.97 (0.33 to 2.88)

∞

3 months

DASH score

23.7 (19.2 to 28.2)

29.4 (24.5 to 34.3)

42.8 (35.6 to 50.0)

-5.7 (-12.3 to 1.0)

-19.1 (-27.6 to -10.6)

-13.4 (-22.1 to -4.7)

Pain at rest

1.5 (1.0 to 2.1)

1.1 (0.5 to 1.7)

1.5 (0.6 to 2.4)

0.4 (-0.4 to 1.2)

0.0 (-1.0 to 1.0)

-0.4 (-1.5 to 0.7)

Pain on activities

3.5 (2.8 to 4.3)

3.6 (2.8 to 4.5)

5.7 (4.5 to 7.0)

-0.1 (-1.3 to 1.1)

-2.2 (-3.7 to -0.7)

-2.1 (-3.7 to -0.6)

Constant-Murley score

61.9 (56.4 to 67.3)

52.3 (46.3 to 58.3)

35.5 (26.7 to 44.3)

9.6 (1.5 to 17.7)

26.4 (16.1 to 36.8)

16.9 (6.2 to 27.5)

Elbow ROM – degrees

134 (129 to 140)

129 (123 to 135)

103 (94 to 111)

5 (-3 to 13)

31 (21 to 41)

26 (16 to 37)

15D score

0.87 (0.83 to 0.91)

0.91 (0.89 to 0.93)

0.81 (0.77 to 0.85)

-0.04 (-0.10 to 0.02)

0.06 (0.00 to 0.12)

0.10 (0.04 to 0.16)

DASH work module score

33.4 (23.9 to 42.9)

35.2 (22.0 to 48.5)

64.4 (45.8 to 83.0)

-1.9 (-18.2 to 14.5)

-31.0 (-51.9 to -10.1)

-29.2 (-52.0 to -6.3)

DASH sports/performing arts module score

56.2 (44.5 to 67.9)

81.3 (68.4 to 94.1)

74.3 (43.8 to 100)

-25.1 (-42.7 to -7.5)

-18.2 (-50.8 to 14.5)

6.9 (-26.3 to 40.2)

Patients with acceptable symptomatic state – %

Group comparisons: RRR (95% CI)

47 (30 to 65)

27 (12 to 46)

8 (0 to 36)

RRR 1.76 (0.88 to 3.53)

RRR 6.12 (0.90 to 41.58)

RRR 3.47 (0.48 to 24.97)

Adequate clinical recovery – %

Group comparisons: RRR (95% CI)

24 (11 to 42)

14 (4 to 32)

8 (0 to 36)

RRR 1.76 (0.59 to 5.24)

RRR 3.15 (0.44 to 22.76)

RRR 1.79 (0.22 to 14.52)

Satisfaction with shoulder function

6.8 (6.1 to 7.5)

7.1 (6.3 to 7.8)

4.0 (2.9 to 5.2)

-0.3 (-1.4 to 0.8)

2.7 (1.4 to 4.1)

3.0 (1.7 to 4.4)

Satisfaction with elbow function

7.6 (6.9 to 8.2)

7.9 (7.2 to 8.5)

6.5 (5.5 to 7.6)

-0.3 (-1.2 to 0.7)

1.1 (-0.2 to 2.3)

1.3 (0.1 to 2.6)

Satisfaction with upper limb function

7.0 (6.3 to 7.8)

6.3 (5.5 to 7.1)

3.6 (2.4 to 4.8)

0.7 (-0.4 to 1.8)

3.5 (2.1 to 4.9)

2.7 (1.3 to 4.2)

Patients able to return to activities of daily living – %

Group comparisons: RRR (95% CI)

79 (62 to 91)

87 (69 to 96)

62 (32 to 86)

RRR 0.92 (0.87 to 1.08)

RRR 1.29 (0.81 to 2.05)

RRR 1.41 (0.90 to 2.21)

Patients able to return to previous hobbies – %

Group comparisons: RRR (95% CI)

39 (23 to 58)

Humeral Shaft Fractures in Adults – Effectiveness of Surgery versus Functional Bracing

Abstract and Figures