The N-brace trial: Does arm position during immobilisation of proximal humerus fractures influence outcome - A preliminary study

Abstract

Background

Traditional initial management of proximal humerus fractures (PHF) involves arm immobilisation in a simple sling (SS) in an internally rotated position. We believe this risks fracture displacement and imbalance of soft tissues, encouraging malunion and stiffness. A neutral-rotation brace (NRB) maintains an arm position which may prevent this, leading to quicker and superior recovery.

Methods

We randomised patients with two- to four-part PHF into 4 weeks of immobilisation with either a SS or NRB, independent of surgery. Range of motion (ROM), subjective shoulder value (SSV), DASH, Constant–Murley (CMS) and Oxford Shoulder (OSS) scores were assessed at 6-weeks, 3-months and 1-year post-injury.

Results

The SS group included 11 patients vs 9 in the NRB group. At final follow-up, the SS and NRB groups had mean DASH scores of 42 vs 35, OSS 42 vs 46, CMS 71 vs 86, SSV 84% vs 92%, respectively. ROM was superior with the NRB (elevation 159°, ER 47° and IR score 8 vs 140°, 37° and 7 with SS).

Conclusions

Despite being a small series, our results demonstrate a trend towards NRB providing better outcomes. This feasibility study supports the need for a larger multi-centre randomised controlled trial comparing these immobilisation methods for PHF.

Introduction

Proximal humerus fractures (PHF) are common injuries (70 per 100,000) with the majority of them still being treated non-operatively.^1–3 In recent years, studies such as PROFHER ⁴ along with the financial and social-distancing pressures associated with the coronavirus pandemic have steered us in the United Kingdom further towards considering non-operative treatment modalities in a growing number of patients for proximal humerus fractures (PHFs). Irrespective of whether a PHF is treated operatively or not, conventionally these injuries are associated with a period of immobilisation with the arm in a simple sling (SS). Traditional slings hold the elbow in flexion while placing the forearm across the trunk of the body which effectively places the humerus in internal rotation.

In a simple two-part surgical neck of humerus fracture, the functioning rotator cuff ensures the humeral head remains well-balanced and in a neutral position. However, the humeral diaphysis will tend to be internally rotated by the strong deforming forces of pectoralis major, teres major and latissimus dorsi.^5,6 A SS will compound this malrotation by further internally rotating the humeral shaft. In a three- or four-part fracture, the greater tuberosity (GT) is often displaced posteriorly and superiorly by the pull of the active rotator cuff. Again, application of a SS has the potential to internally rotate the head, increasing the tension in the posterior rotator cuff and exacerbating the degree of GT displacement. It is well recognised that a proximal humeral fracture that heals with a displaced GT typically has a poor clinical outcome.^7–10 Further to this, internal rotation of the head may lead to shortening of subscapularis and the anterior capsule, with subsequent contractures and tightening of these structures, similar to that seen in post-traumatic adhesive capsulitis, becoming a hindrance to recovery of glenohumeral movement.^11–13

There is a lack of work in the available literature assessing the effect of alternate bracing or positions of immobilisation for PHFs. No published studies to date have made comparisons of immobilisation in neutral rotation vs the internal rotation sling for non-operatively managed PHFs. In theory, improved conditions for anatomical healing and restoration of normal function would result from placing the arm in an anatomical position using a neutral rotation brace (NRB) (Figures 1A & 1B). This should restore alignment of the humeral shaft and head in surgical neck fractures and with displaced GT fractures it would effectively externally rotate the humeral head which may possibly reduce the displacement, discouraging further migration by de-tensioning the posterior rotator cuff and capsule.

Figure 1.

A. Neutral-rotation brace viewed from face-on. B. Neutral-rotation brace viewed from the affected side.

Figure 2.

Graph demonstrating CMS results in the SS and NB groups.

We hypothesise that following sustaining a PHF, the proposed anatomical fracture reduction and restoration of glenohumeral joint (GHJ) soft tissue tension achieved from immobilisation in mid-range with an NRB results in a shorter recovery, better functional outcome and greater range of motion (ROM), compared to that achieved in internal rotation provided by a SS. Given the lack of evidence assessing the position of arm immobilisation after these injuries, we propose a feasibility randomised clinical study comparing the NRB and SS.

Method

We carried out a single-centre prospective randomised controlled clinical trial in a National Health Service district general hospital trauma unit following attaining ethical approval from the local research and ethics committee. Patients with acute PHFs were recruited according to our inclusion and exclusion criteria (Table 1) and consented accordingly. They were randomised, using sealed envelopes, to be immobilised in either a SS or NRB, irrespective of whether the fracture was treated non-operatively or operatively. The decision to operate and method of fixation was made by the treating consultant according to their experience and usual practice. A CT scan was performed at initial presentation for assessment of fracture configuration and management planning. Due to the time required for acute adult fractures to start to unite with primary or secondary healing and for the injured GHJ soft tissues to settle, both groups had a 4-week period of immobilisation before physiotherapy was delivered according to a standardised local protocol. For those who underwent operative fixation, the start of the 4-week period was reset at the point of surgery.

Table 1.

Inclusion and exclusion criteria.

Inclusion	Exclusion
Age 18–75	Patient refusal to be recruited
Proximal humerus anatomical and surgical neck fractures	ASA grade 4–5
2, 3 or 4 part fractures (according to the Neer classification)	Impaired cognitive function
ASA grade 1–3	Isolated GT fractures
Isolated injury	Shoulder dislocations
Within 7 days of injury	All other humerus fractures
	Previous injury or operations to shoulder or humerus
	Polytrauma
	Greater than 7 days from injury
	Metabolic bone disease
	Clinically significant upper limb neuromuscular disability
	Non-English speakers

Patients underwent routine follow-up with clinical and radiographic assessment, as deemed appropriate by the responsible consultant surgeon. At 6 weeks, 3 months and 12 months, primary outcome measures were assessed using validated functional outcome scores; the Oxford shoulder score (OSS),^14,15 subjective shoulder value (SSV), ¹⁶ Constant–Murley score (CMS)^17,18 and disability of arm, shoulder and hand (DASH) scoring system.^19,20 ROM was determined as a secondary outcome measure at each of the same time intervals, including shoulder elevation, external rotation (ER) and internal rotation (IR). IR was measured according to the CMS.

For all of the data analyses we used the intention-to-treat principle. Primary and secondary outcome measures, at 6 weeks, 3 months and 12 months following injury (or surgery in those patients operated on), were compared between the SS and NRB groups using unpaired t-tests. Outcomes within the same groups at the different time points were assessed using paired t-test P < 0.01 was considered significant.

Results

We recruited 20 patients for this trial who met the inclusion criteria, with a mean age of 53 years (range 27–69) and 45% being female. Eighteen patients were followed up for the full 12-month study period with two lost to follow-up after 3 months. According to Neer's classification, there were ten 2-part, eight 3-part and two 4-part fractures. All patients had definitive management initiated within 7 days (1–6) of their injury, 7 of which included operative management (35%) using dedicated proximal humerus locking plates (2), intramedullary nails (2) or all-suture fixation (3). Eleven patients were randomised to the SS group (55%) and nine to NRB group, both groups being relatively well matched for age, sex and fracture type (Table 2). In the SS group, 27% were operated on vs 44% in the NRB group. All patients were fully compliant with their selected method of bracing.

Table 2.

Patient randomisation and demographics.

	SS	NB
Number of patients	11	9
Female	5	4
Average age	54	51
Two-part fracture	6	4
Three-part fracture	4	4
Four-part fracture	1	1
ORIF	3	4
12-month Follow-up	10	8

When considering the primary outcomes, (Table 3), all functional scores for both bracing positions demonstrated significant progress throughout the 12-month follow-up (Figures 2–5). Apart from a lesser OSS at 6 weeks, the NRB had superior scores compared with the SS throughout the follow-up period in all four scoring modalities. At 12 months, the NRB had a greater CMS of 86, OSS of 46, DASH of 35 and SSV of 92 compared with the SS who scored 71, 42, 42 and 84, respectively. However, the disparity in scores was not statistically significant. A similar trend was seen assessing movement (Table 4), including a significant overall increase in ROM in both bracing groups and a greater 12-month ROM with the NRB. The NRB had superior elevation and ER (Figures 6 & 7) throughout the follow-up period but mean IR (Figure 8) was only greater at the final assessment. Final ROM included elevation of 158.8°, ER of 46.9° and an IR score of 8.0 for the NRB vs 139.9, 37.0 and 7.0, respectively, for the SS but none of the differences was of statistical significance.

Table 3.

Primary outcome measures – functional outcome scores.

Functional scores		SS	NB	Difference (P Value)
CMS	6 weeks	26 (4−42)	35 (18−59)	0.161
	3 months	43 (11−71)	55 (31−85)	0.183
	12 months	71 (32−95)	86 (63−96)	0.123
	Progression (P value)	<0.0001	<0.0001
OSS	6 weeks	21 (3−44)	16 (3−34)	0.340
	3 months	28 (8−48)	36 (23−43)	0.118
	12 months	42 (34−48)	46 (42−48)	0.127
	Progression (P value)	<0.0001	<0.0001
DASH	6 weeks	92 (62−144)	77 (45−105)	0.226
	3 months	74 (30−125)	58 (34−97)	0.209
	12 months	42 (30−48)	35 (30−38)	0.058
	Progression (P value)	<0.0001	0.0015
SSV	6 weeks	41 (20–75)	47 (25–60)	0.445
	3 months	60 (20–90)	63 (40–80)	0.574
	12 months	84 (80–100)	92 (80–100)	0.045
	Progression (P value)	<0.0001	0.0003

Figure 3.

Graph demonstrating SSV results in the SS and NB groups

Figure 6.

Graph demonstrating elevation in degrees in the SS and NB groups

Table 4.

Secondary outcome measures – range of motion.

ROM		SS	NB	Difference (P value)
Elevation (degrees)	6 weeks	45 (10–85)	68 (40–110)	0.062
	3 months	78 (−20–180)	112 (80–170)	0.112
	12 months	140 (80–180)	159 (130–180)	0.185
	Progression (P value)	<0.0001	<0.0001
ER (degrees)	6 weeks	4 (−20–30)	15 (−10–40)	0.149
	3 months	14 (−10–45)	18 (−10–40)	0.530
	12 months	37 (0–80)	47 (20–70)	0.698
	Progression (P value)	0.0008	<0.0001
IR (See CM tool)	6 weeks	3.3 (0–6)	2.2 (0–8)	0.051
	3 months	3.8 (2–8)	3.1 (0–8)	0.236
	12 months	7 (2–10)	8 (4–10)	0.886
	Progression (P value)	0.0011	0.0008

Figure 7.

Graph demonstrating ER results in degrees in the SS and NB groups

Figure 8.

Graph demonstrating IR in points (as per the CM scoring tool) in the SS and NB groups

Figure 4.

Graph demonstrating DASH results in the SS and NB groups

Figure 5.

Graph demonstrating OSS results in the SS and NB groups

Complications

There were no cases of avascular necrosis, nor clinical or radiological non- or delayed union, identified. Neither neurological injury nor infection was seen in patients undergoing operative fixation. Further surgery was not required in any patient.

Discussion

This is the first study assessing the effect of arm position on outcomes of acute PHFs which includes those managed non-operatively. Similarly, there are no other published articles comparing the use of a traditional internal-rotation sling with a brace that holds the arm in neutral rotation for operatively and non-operatively treated PHFs. While the numbers in this feasibility study are small and statical analysis is not possible, the results seen do support the hypothesis of greater function, ROM and speed of recovery with a NRB compared to a SS for management of PHF.

While the use of either bracing modality permitted improved function and ROM in our study, all outcomes (barring the 6-week OSS) demonstrated superiority of the NRB group at all-time points. The outcome differences between positions of immobilisation did not demonstrate statistical significance due to the small sample size but probably reflects the more anatomical healing position seen with the NRB which is associated with improved balancing of the soft tissues. Although immobilisation in mid-range does not prevent glenohumeral stiffness, it should decrease the amount of tightening of the anterior capsule and subscapularis that might be expected with a SS. Even with near anatomical fracture healing, such anterior soft tissue contractures, in the short term, will likely restrict ER and elevation and may well have a detrimental effect on the final outcome. Interestingly, although IR was improved in the SS group initially, by the end of study period the NRB group was able to internally rotate further. This may be a reflection of the overall benefit of anatomical immobilisation. From our study, use of the NRB saw a trend towards a quicker recovery of movement and function, with excellent compliance and an absence of additional complications.

In a prospective RCT assessing 36 patients, Baumgarten compared a traditional SS with an NRB following anatomic total shoulder arthroplasty used to treat glenohumeral joint osteoarthritis. Pain relief, forward flexion and overall function were superior with the NRB but without statistical significance. The NRB did provide statistically significant greater passive and active ER in neutral abduction after a year compared to the internal rotation sling. ²¹ Pain relief, movement and function may be improved in the NRB because a more anatomical position should lead to decrease in the tension in the posterior structures.

Boileau and Walch have long since advocated use of an NRB following hemiarthroplasty for management of three- or four-part PHFs. Following appropriately seating and securing both the implant and tuberosities, post-operative use of an NRB is accompanied by passive elevation exercises but with an initial 6-week exclusion of internal or external rotation to avoid excessive tensions in the infraspinatus and subscapularis respectively so as to protect tuberosity position and healing.^7,9 If the NRB reduces the GT more accurately in three- or four-part fractures, this may contribute to reduced pain and greater function.^7–10

Chen et al. more recently compared the NRB with a SS in surgically fixed PHFs. ²² They found that the NRB group produced superior function and reduced pain which they were able to demonstrate was statistically significant. Unlike our study, they only looked at fracture fixation. It may be there is more benefit with the NRB from those treated non-operatively due to the balancing effect on soft tissues with less chance of secondary displacement occurring (especially of the GT).^7,9,21

There are limitations of this study which are mainly attributed to its small sample size but given the lack of previous work and evidence on position of arm immobilisation following a PHF we were obliged to run this feasibility study first. This study has confirmed that research in this area is both possible and safe. Although differences in outcomes between the SS and NRB do not reach statistical significance, they do indicate that there may be benefit from the use of NRB for managing these injuries. This is a safe and inexpensive intervention that is highly likely to be cost-effective.

The results from this study along with other recent published evidence suggest there would be value in a more extensive fully powered multi-centred RCT with a similar design. While the advantages of using an NRB in operatively treated PHF have been discussed, the value of utilising this mode of immobilisation for solely non-surgically managed patients may provide us with a more accurate assessment of its benefits, as surgical intervention with rigid internal fixation probably traverses the effect of balancing the soft tissues through position of immobilisation alone. Our proposed protocol for further study includes additional evaluation of fractures with repeat CT, 3 months after initiation of immobilisation, to determine the extent of secondary displacement with each bracing modality and subsequent correlation with existing primary and secondary outcome measures.

Conclusion

This small feasibility study assessing the optimal position for arm immobilisation after a PHF has demonstrated a trend towards a superior outcome from an NRB with both ranges of movement and function both during recovery and at a year after the injury. A larger multi-centre randomised controlled trial would help confirm the benefit of this safe and inexpensive intervention.