PROFHER Trial Review

By Arman Memarzadeh

Professor Rangan et al; published in JAMA (Journal of American Medical Association). This was a multi-centre RCT in the UK.

Backgroung: Increasing trend of operative intervention for proximal humerus fractures and half are displaced and majority involve the surgical neck.

Aim: whether non-operative management resulted in better outcomes compared with operative intervention.

Methods: Pragmatic, open, multicentre, parallel-group, superiority, randomized clinical trial between September 2008 and April 2011 from orthopaedic departments (fracture clinics or wards) at 32 acute UK National Health Service hospitals.

Inclusion criteria:

  • 16 years or older
  • Patients presenting within 3 weeks of injury;
  • DISPLACED proximal humerus fracture involving the surgical neck
    • The degree of displacement had to be significant enough for the surgeon to consider operative intervention but did not have to meet Neer’s criteria

Exclusions:

  • Dislocations
  • Open fracture
  • Insufficient mental capacity
  • Comorbidities precluding surgery or anaesthesia
  • CLEAR indication for surgery e.g. severe soft tissue compromise, multiple injuries, pathological fractures
  • terminal illness
  • non-resident in the catchment area

Approved by York Trials Unit, ethical approval obtained, written consent from participants. Randomisation performed remotely on a 1:1 basis. Blinding was not done.

Outcome measures: Oxford Shoulder Score (primary), SF-12 (secondary), EuroQol 5D

Complications recorded. 2 year follow up, measures taken at 6, 12 and 24 months.

Process:

2 independent and blinded specialists classified the fractures based on the Neer classification.

Interventions were plate fixation, hemiarthroplasty or IM nailing.

Non-op = sling for as long as required (3+ weeks suggested) followed by active ROM physiotherapy

Stats:

MCID of 5 points on the OSS

Power calculations: 80% power = 100 patients per arm (recruited 125 in each to account for a 20% loss to follow up)

Results:

1250 patients screened, 43% eligible. Of the 563 eligible participants, 250 (44%) consented to take part in the study. Mean age 66 (24-92); 77% female, 99.6% white

In the 125 participants allocated to surgery, 109 received this, 16 were treated non-operatively (8 changed their mind, 6 were unfit, 2 difference of opinion with treating surgeon). Of the 125 in the non-operative group, 2 received surgery (1 changed mind, 1 surgeon changed mind). Overall >100 participants in each analysis arm (over the value required by the power calculations). Similar baseline demographics, more smokers in the non-operative group. Well matched in the spread of fracture patterns.

94% of operations were performed by or under supervision of a consultant surgeon (specialty not disclosed). Non-operative management broad arm sling (65.6%); collar and cuff (28%); hanging cast (2.4%); missing data (4%).

OSS 3840 from 624 months; no statistically significant difference at any point although a trend towards better early (6 months) results surgically

Similarly, trend towards better SF-12 physical scores at 6 months in surgical group (not statistically significant) and trend towards better OSS at 6 months with increasing age, and regardless of tuberosity involvement in the surgical group.

Slightly more complications in the surgical group (not statistically significant).

Discussion:

The statistical results showed no difference between primary or secondary outcome measures over 24 months. Baseline demographics, slight differences and subgroup analyses were discussed. PROMs were used, which are dependent on patient reporting.

Six points discussed but no mention of the large percentage of excluded patients at all.

Critique

POSITIVES

  • Valid question
  • Solid statistical basis
  • Multi-centre trial, well-coordinated and classified fractures independently
  • Good follow up rate
  • Similar demographics and fracture pattern spread between the two groups.

NEGATIVES

  • 250 patients included from a potential 1250 (20%).
  • Did not include the types of fractures in the body of the article
  • Mostly 2 part +/- GT fractures (>80%)
  • No consideration of other types of arthroplasty such as reverse shoulder replacement, which may be considered in an elderly, cuff deficient patient with a displaced proximal humerus fracture

Relevance to current practice

This was a well-designed, multi-centres RCT which showed no difference in outcomes of proximal humerus fractures involving the surgical neck when treated operatively or sling immobilisation and physiotherapy.

It should be noted that due to the low number of 4-part fractures, the results could be extrapolated to 2-part fractures (involving the surgical neck) or 3-part fractures involving the greater tuberosity. However, care should be taken not to extend this to isolated greater tuberosity, lesser tuberosity, four part, head-splitting fractures, or fracture dislocations.