Surgical interventions for treating distal humeral fractures in adults

  • Review
  • Intervention

Authors


Abstract

Background

Distal humeral fractures in adults are relatively uncommon injuries that require surgical intervention in most cases. There is a lack of consensus regarding the best management of distal humeral fractures in adults, including the role of conservative treatment, appropriate surgical approach, fixation strategies, the role of total elbow arthroplasty and handling of nerves such as the ulnar nerve.

Objectives

To assess the effects (benefits and harms) of surgical interventions for distal humeral fractures in adults.

Search methods

We searched the Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (May 2012), the Cochrane Central Register of Controlled Trials (The Cochrane Library, 2012 Issue 4), MEDLINE (1946 to April Week 4 2012), EMBASE (1980 to 2012 Week 17), Current Controlled Trials (1st May 2012), the WHO International Clinical Trials Registry Platform (1st May 2012) and the bibliographies of trial reports and relevant articles.

Selection criteria

All randomised and quasi-randomised controlled trials pertinent to the management of distal humeral fractures in adults were included.

Data collection and analysis

Two review authors independently performed study selection, assessed of risk of bias and extracted data. Pooling of data was impossible due to study heterogeneity.

Main results

Three small randomised controlled trials, with a total of 109 participants with Orthopaedic Trauma Association/Arbeitsgemeinschaft für Osteosynthesefragen (OTA/AO) type C distal humeral fractures, were included. Overall, the quality of the available evidence is limited. As well as the small sample sizes and detection bias from the lack of blinding of subjective outcomes, the methods and results of all three trials were incompletely reported.

One trial, involving 42 participants, compared open reduction-internal fixation (ORIF) with total elbow arthroplasty (TEA) in patients aged over 65 years. Of the 40 participants followed up for two years, five allocated ORIF underwent intraoperative conversion to TEA. These participants were crossed-over to the TEA group in the analyses. The reported Mayo Elbow Performance Score (MEPS) results were consistently better in the TEA group at follow-up after 6, 12, and 24 months, whereas the Disability of the Arm, Shoulder, Hand (DASH) scores showed short term (after 6 months), but not longer term (after 12 and 24 months), superiority in the TEA group. The reoperation rate, complication rate and elbow range of motion results showed no statistically significant differences between the two groups. While an intention-to-treat analysis of treatment failure, where the five cross-over participants are placed in their original allocated group, is in favour of TEA, the result did not reach statistical significance (9/21 versus 3/21; RR 3.00, 95% CI 0.94 to 9.55).

The second trial, involving 38 patients but reporting results for 35, compared perpendicular versus parallel double plate fixation strategies. There was a consistent finding of a lack of significant differences between the two treatment groups in terms of MEPS, re-operation for complications (3/17 versus 3/18; RR 1.06, 95% CI 0.25 to 4.54), complications and elbow joint range of motion.

The third trial, which included 29 patients with preoperative ulnar nerve compression symptoms, compared anterior subfascial transposition with in situ decompression of the ulnar nerve. Although results for complete recovery of ulnar nerve function (12/15 versus 8/14; RR 1.4, 95% CI 0.83 to 2.35) and grades based on the Bishop rating system tended to favour the transposition group, none of the differences were statistically significant.

Authors' conclusions

Overall, this review found there is either no or insufficient evidence from randomised or quasi-randomised controlled trials to determine whether surgery is, and which surgical interventions are, the most appropriate for the management of different types of distal humerus fractures. Well designed and reported large and multi-centre randomised controlled trials testing current interventions, such as pre-contoured and locking plating systems, are needed.

Résumé scientifique

Interventions chirurgicales pour le traitement des fractures distales de l'humérus chez l'adulte

Contexte

Les fractures distales de l'humérus chez l'adulte sont des blessures relativement rares qui nécessitent une intervention chirurgicale dans la plupart des cas. Il n'y a pas de consensus concernant la meilleure prise en charge des fractures distales de l'humérus chez l'adulte, notamment sur le rôle du traitement conservateur, l'approche chirurgicale appropriée, les stratégies de fixation, le rôle de l'arthroplastie totale du coude et la manipulation des nerfs tels que le nerf cubital.

Objectifs

Évaluer les effets (bénéfices et préjudices) des interventions chirurgicales pour les fractures distales de l'humérus chez l'adulte.

Stratégie de recherche documentaire

Nous avons effectué une recherche dans le registre spécialisé du groupe Cochrane sur les traumatismes ostéo-articulaires et musculaires (mai 2012), le registre Cochrane des essais contrôlés (CENTRAL) (The Cochrane Library, 2012, numéro 4), MEDLINE (de 1946 jusqu'à la 4ème semaine d'avril 2012), EMBASE (de 1980 jusqu'à la 17ème semaine de 2012), Current Controlled Trials (1er mai 2012), le système d'enregistrement international des essais cliniques de l'OMS (ICTRP pour « International Clinical Trials Registry Platform ») (1er mai 2012) et les bibliographies des rapports sur les essais et des articles pertinents.

Critères de sélection

Tous les essais contrôlés randomisés et quasi-randomisés relatifs à la prise en charge des fractures distales de l'humérus chez l'adulte ont été inclus.

Recueil et analyse des données

Deux auteurs de la revue ont sélectionné les études, évalué leurs risques de biais et extrait les données de manière indépendante. Le regroupement des données n'a pas été possible en raison de l’hétérogénéité des études.

Résultats principaux

Trois essais contrôlés randomisés de petite taille, totalisant 109 participants présentant des fractures distales de l'humérus de type C selon l'Association d'orthopédie et de traumatologie/Arbeitsgemeinschaft für Osteosynthesefragen (OTA/AO), ont été inclus. Globalement, la qualité des preuves disponibles est limitée. Il en est de même pour les petits effectifs et les biais de détection dus à l'absence de mise en aveugle des critères de jugement subjectifs, et les méthodes et les résultats des trois essais qui ont été partiellement rapportés.

Un essai, portant sur 42 participants, a comparé la réduction à ciel ouvert avec fixation interne (ORIF) à l'arthroplastie totale du coude (ATC) chez des patients âgés de plus de 65 ans. Sur les 40 participants ayant fait l'objet d'un suivi pendant deux années, cinq ORIF assignées ont été converties en peropératoire en une arthroplastie totale du coude (ATC). Ces participants ont été croisés dans le groupe d'arthroplastie totale du coude (ATC) dans les analyses. Les résultats rapportés du score MEPS (Mayo Elbow Performance Score) ont été systématiquement meilleurs dans le groupe d'arthroplastie totale du coude (ATC) au bout d'un suivi de 6, 12 et 24 mois, tandis que les résultats du score DASH (Disability of the Arm, Shoulder, Hand) ont montré une supériorité à court terme (au bout de 6 mois), mais pas à plus long terme (au bout de 12 et 24 mois), dans le groupe d'arthroplastie totale du coude (ATC). Les résultats du taux de réinterventions, du taux de complications et de l'amplitude de mouvement du coude n'ont pas révélé de différences statistiquement significatives entre les deux groupes. Tandis qu'une analyse en intention de traiter de l'échec du traitement, lorsque les cinq participants croisés sont inclus dans le groupe auquel ils ont été assignés initialement, est en faveur de l'arthroplastie totale du coude (ATC), le résultat n'a pas atteint de signification statistique (9/21 contre 3/21 ; RR 3,00, IC à 95 % 0,94 à 9,55).

Le deuxième essai, portant sur 38 patients mais rapportant des résultats pour 35, a comparé les stratégies de fixation par doubles plaques perpendiculaires aux doubles plaques parallèles. Un manque de différences significatives a systématiquement été observé entre les deux groupes de traitement en termes de score MEPS, de taux de réinterventions pour les complications (3/17 contre 3/18 ; RR 1,06, IC à 95 % 0,25 à 4,54), de taux de complications et d'amplitude de mouvement de l'articulation du coude.

Le troisième essai, qui avait inclus 29 patients présentant des symptômes préopératoires de compression du nerf cubital, a comparé la transposition subfasciale antérieure à la décompression in situ du nerf cubital. Bien que les résultats concernant la récupération complète de la fonction du nerf cubital (12/15 contre 8/14 ; RR 1,4, IC à 95 % 0,83 à 2,35) et les grades basés sur le système d'évaluation Bishop aient semblé favoriser le groupe de transposition, aucune des différences n'était statistiquement significative.

Conclusions des auteurs

Globalement, cette revue a constaté que les preuves issus des essais contrôlés randomisés ou quasi-randomisés sont soit inexistantes soit insuffisantes pour déterminer si la chirurgie est la plus appropriée, et quelles interventions chirurgicales sont les plus appropriées, pour la prise en charge des différents types de fractures distales de l'humérus. Des essais contrôlés randomisés à grande échelle et multicentriques, bien conçus et convenablement rapportés, examinant les interventions actuelles, telles que les systèmes de plaques à verrouillage et prédécoupées, sont nécessaires.

Plain language summary

Surgical interventions for treating distal humeral fractures in adults

The distal humerus is the end of the upper arm bone (the humerus) and forms the upper part of the elbow joint. Its structure is highly complex as it connects with both forearm bones (the radius and ulna) to allow a wide range of motion: in bending and straightening out of the elbow, and rotating of the forearm. Fractures to the distal humerus most often occur in young men from high-energy trauma; or older women, aged 60 years and over, who typically have osteoporosis and whose fracture results from a low-energy fall. Most distal humeral fractures need surgical intervention because elbow motion is either very difficult or impossible. Open reduction and internal fixation, to hold the bone fragments in place until the bone is healed, with various plates and fixation techniques is the standard surgical treatment, especially in younger patients. Total elbow replacement or arthroplasty is where the distal humerus and ulna bone ends forming the elbow are replaced by an artificial joint. Imposed lifting restrictions, irrespective of age, are necessary for successful total elbow arthoplasty.

Despite a comprehensive search, we identified only three small randomised controlled trials with a total of 109 participants. Each trial tested different interventions for treating a complete intra-articular fracture in which the joint surface is separated from the shaft of the humerus. One trial compared open reduction and internal fixation versus total elbow replacement in people aged 65 years and older. This found some limited evidence that internal fixation is sometimes not practical for some more complex and difficult fractures and that people treated total elbow replacement may have a better outcome early on (around six months). The second trial compared two ways of placing two plates used for internal fixation. These were either perpendicular (where the plates were at right angles to each other along the bone) or parallel (where the plates were on either side of the bone). The trial did not find any major differences in outcome between the two methods. The third trial did not provide clear evidence of a difference in two surgical approaches of managing pre-operative ulnar nerve dysfunction.

In all, none of these three under-sized trials provided adequate evidence to determine which of the surgical interventions under test was the most appropriate. The review found no evidence to inform on the use of more current methods of surgical fixation, specifically the use of locking plates.

Résumé simplifié

Interventions chirurgicales pour le traitement des fractures distales de l'humérus chez l'adulte

L'humérus distal est l'extrémité de l'os du bras (l'humérus) et forme la partie supérieure de l'articulation du coude. Sa structure est très complexe car il est relié aux deux os de l'avant-bras (le radius et l'ulna) afin de permettre une grande amplitude de mouvement : la flexion et l'extension du coude, et la rotation de l'avant-bras. Les fractures de l'humérus distal se produisent le plus souvent chez les jeunes hommes suite à un traumatisme à haute énergie ; ou chez les femmes plus âgées, de 60 ans et plus, qui souffrent généralement d'ostéoporose et dont la fracture est provoquée par une chute à faible énergie. La plupart des fractures distales de l'humérus nécessitent une intervention chirurgicale parce que le mouvement du coude est soit très difficile soit impossible. La réduction à ciel ouvert avec fixation interne, afin de maintenir les fragments d'os en place jusqu'à ce que l'os soit guéri, avec différentes plaques et techniques de fixation est le traitement chirurgical standard, en particulier chez les patients très jeunes. Le remplacement total du coude ou arthroplastie consiste à remplacer l'articulation du coude formée par les extrémités osseuses de l'humérus distal et de l'ulna par une articulation artificielle. Les restrictions imposées concernant le soulèvement, indépendamment de l'âge, sont nécessaires pour réussir une arthroplastie totale du coude.

Malgré des recherches exhaustives, nous n'avons identifié que trois essais contrôlés randomisés de petite taille totalisant 109 participants. Chaque essai a examiné des interventions différentes pour le traitement d'une fracture intra-articulaire complète dans laquelle la surface de l'articulation est séparée de la diaphyse de l'humérus. Un essai a comparé la réduction à ciel ouvert avec fixation interne au remplacement total du coude chez des personnes âgées de 65 ans et plus. Celui-ci a trouvé certaines preuves limitées que la fixation interne parfois n'est pas pratique pour certaines fractures plus complexes et difficiles et que les personnes traitées par le remplacement total du coude peuvent obtenir un meilleur résultat très tôt (autour de six mois). Le deuxième essai a comparé deux techniques de positionnement de deux plaques utilisées pour la fixation interne. Celles-ci étaient soit perpendiculaires (lorsque les plaques étaient à angle droit les unes par rapport aux autres le long de l'os) soit parallèles (lorsque les plaques étaient d'un côté ou de l'autre de l'os). L'essai n'a pas permis de mettre en évidence des différences majeures au niveau du résultat entre les deux méthodes. Le troisième essai n'a pas fourni de preuves probantes d'une différence entre les deux approches chirurgicales pour la prise en charge préopératoire du dysfonctionnement du nerf cubital.

Dans l'ensemble, aucun de ces trois essais de taille insuffisante n'a fourni de preuves probantes permettant de déterminer laquelle des interventions chirurgicales testées était la plus appropriée. La revue n'a trouvé aucune preuve pour orienter l'utilisation des méthodes plus récentes de fixation chirurgicale, spécifiquement l'utilisation de plaques à verrouillage.

Notes de traduction

Traduit par: French Cochrane Centre 5th February, 2013
Traduction financée par: Minist�re du Travail, de l'Emploi et de la Sant� Fran�ais

Background

Description of the condition

The distal humerus is the end of the upper arm bone (the humerus) that forms the upper part of the elbow joint (AAOS 2011). Approximately 7% of all adult fractures are elbow fractures, with fractures of the distal humerus accounting for 30% of these (Anglen 2005). An annual incidence of 5.7 distal humeral fractures per 100,000 people was reported for Edinburgh over the period 1988 to 1997 (Robinson 2003). Although this may seem low, the number of cases is growing rapidly as the population ages. In their study of low-trauma distal humeral fractures among women 60 years of age or older, Palvanen 2010 reported a greater than five-fold increase in the annual total number of these fractures (there were 42 fractures in 1970 and 224 in 1998) and a dramatic increase in the age-adjusted incidence over the same period (1970 to 1998) from 12 to 34 per 100,000 people per year. There is a well-documented bimodal distribution of these fractures, with an early peak in young people, mainly males, aged between 12 and 19 years, from high-energy trauma, and a second larger peak in older people, mainly women, aged 60 years and over, who typically have osteoporosis and whose fracture results from a low-energy fall (Robinson 2003).

The anatomy of the distal humerus is highly complex, reflecting its articulation with both forearm bones (the radius and ulna). It allows for wide range of motion in multiple planes (the ulnohumeral joint for flexion (bending) and extension (straightening out) of the elbow, and the radiocapitellar joint for forearm rotation). Elbow motion is generally difficult or impossible when the distal humerus is fractured. In addition, the structure of the elbow leaves the ulnar and radial nerves vulnerable to injury: for example, 3.8% (12/320) of patients in Robinson 2003 had nerve palsies before undergoing surgery.

There are several classification systems for distal humeral fractures. The traditional descriptive classification was based on the concept of the distal humerus comprising two triangular 'columns' of bone (lateral (outer) side of the distal humerus aligned with the radius and medial (inner) side of the distal humerus aligned with the ulna) and the relationship of fracture location to the olecranon fossa (a deep triangular depression on the posterior (back) side of the humerus). The fractures can thus be broadly classified into supracondylar (above the olecranon fossa), single column (condyle) fractures and bi-column fractures (these are the most common). Jupiter and Mehne (Jupiter 1992) advocated a classification scheme based on intraoperative observations of fracture patterns, and described as 'high T', 'low T', 'Y', 'H', 'medial lambda' and 'lateral lambda' patterns.

A more comprehensive and the most commonly used system is the Orthopaedic Trauma Association/Arbeitsgemeinschaft für Osteosynthesefragen (OTA/AO) classification system. In this system, similar to that for other periarticular fractures, type 'A' describes an extra-articular fracture, type 'B' fractures extend into the articular surface but are partial articular often involving one 'column', and 'C' is a complete intra-articular fracture in which articular surface is separated from the shaft of the humerus. With the further use of numerical attachments from 1 to 3, fractures of distal humerus are subdivided into subtypes based on the extent of comminution and specific fracture pattern (Rüedi 2007).

Description of the intervention

Surgical management is widely considered to be indicated for displaced fractures of the distal humerus (Riseborough 1969). Surgical interventions for distal humeral fractures include: 1) open reduction (to reposition the bone fragments) and internal fixation (to hold the bone fragments in place until the bone is healed) with various devices such plates and screws; 2) total elbow replacement/arthroplasty (TEA) where the distal humerus and ulna bone ends forming the elbow are replaced by an artificial joint; 3) closed reduction and external fixation (the device holding the pins placed in the bones is external to the skin); and 4) elbow hemiarthroplasty (currently this uncommon intervention where only the distal humerus is replaced is not approved in many countries).

Numerous surgical approaches have been described, almost all of which employ a universal posterior midline skin incision (Wong 2008). The ulnar nerve is almost always identified and protected at the elbow, though whether the ulnar nerve is left in situ or transposed is left to the preference of the surgeon (Nauth 2011).

It is notable that for open reduction and internal fixation (ORIF) there has been a variety of implant types, including precontoured metal plates and locking plates, and fixation techniques that have been described and recently developed.

Non-operative treatment is generally viewed to have a limited role to play in treating distal humeral fractures (Baratz 2006). It is generally reserved for completely undisplaced fractures, and for patients who cannot tolerate anaesthesia or who have medical co-morbidities that render them unsuitable for surgery, such as advanced dementia, and those with limited arm function such as stroke-related paralysis. Non-operative (conservative) treatment typically involves the so-called 'bag of bones' treatment that consists of immobilisation of the elbow in 60° of flexion for two to three weeks, followed by gentle motion exercises (Baratz 2006). Pehlivan 2002 reported good outcomes from functional bracing of extra-articular supracondylar distal one-third humeral shaft fractures.

How the intervention might work

Surgical intervention is widely considered to have superior clinical outcomes for displaced distal humeral fractures. Currently, there are a variety of internal fixation techniques The goal of ORIF of a fracture of distal humerus is to accomplish: 1) anatomic reduction and reconstruction of the elbow joint and the articular surface, 2) rigid stabilisation and fixation of the articular surface to the humeral shaft (diaphysis), and 3) early motion exercises to minimize joint stiffness. However, this goal becomes difficult to achieve in a group of patients who are greater than 65 years, have sustained a highly comminuted type C fracture, and have low bone quality due to osteoporosis. In such patients, total elbow arthroplasty is often considered to be a better choice compared with ORIF (Thomas 2011).

Since the introduction of bi-columnar plating for the fixation of fractures of distal humerus by the AO, a number of implants and fixation methods are available for ORIF. Typically, the decisions regarding the choice of both the implants and fixation techniques will be dictated by the fracture pattern and degree of comminution. Precontoured periarticular distal humerus locking plates are currently the most popular choice of treatment for distal humeral fractures, though their use is still controversial (Galano 2010). Regarding dual plate fixation, there is considerable amount of debate pertaining to its position or orientation as orthogonal plates (90-90 on separate columns), or parallel plates (medial and lateral supracondylar ridges). Other currently less popular modalities for fixation include simple Kirschner wires or screws, 3.5-mm pelvic reconstruction plates, and limited-contact dynamic compression plates.

Total elbow arthroplasty has emerged as a viable option for some patients (Thomas 2011). Common indications include patients with a severe articular comminution of the distal humerus; and those with pre-existing inflammatory or degenerative symptomatic elbow arthritis. Irrespective of age, functional low-demand and compliancy is necessary for successful total elbow arthoplasty. The main disadvantages of TEA include a lifting restriction of 5 pounds or 2.3 kg for repetitive lifting and 10 pounds or 4.5 kg for a single lift with the affected arm, the risk of implant loosening, periprosthetic fractures, infection and prosthetic failure. Although elbow hemiarthroplasty could be a better option in appropriately selected patients, it is not approved in most countries.

Non-operative treatment involves immobilisation of the injured elbow followed by physiotherapy and motion exercises. Immobilisation (e.g. in a long-arm cast) is helpful in terms of maintaining fracture stabilisation and providing pain relief and reducing swelling during the healing process. However, a major disadvantage is the high risk of the elbow becoming stiff, with a diminished range of motion subsequently. For displaced distal humeral fractures, closed reduction with immobilisation is difficult to achieve and maintain, and joint articular incongruity can result. Rehabilitation generally comprises motion exercises and physiotherapy.

Why it is important to do this review

Although fractures of the distal humerus account for only around two per cent of all adult fractures, they are increasingly common and can be a debilitating injury leading to long-term impairment and disability, including that from post-traumatic elbow arthritis. They are also difficult to treat because of the complexity of bony anatomy of the distal humerus and increasing number of osteoporotic fractures. Most distal humeral fractures need surgical intervention with open reduction and internal fixation being the standard surgical treatment, especially in younger patients. Moreover, a variety of new techniques for exposure and pre-contoured locking plates are considered to give a better outcome. However, there are scenarios where osteoporosis in older patients renders ORIF impossible. While total elbow arthroplasty and distal humeral hemiarthroplasty may be viable options, they are not without their own limitations and complications.

Given the challenges and controversies regarding treating these fractures, there is a need to systematically review the current evidence of the effectiveness of surgical interventions for treating distal humeral fractures in adults.

Objectives

To assess the effects (benefits and harms) of surgical interventions for distal humeral fractures.

The main comparisons are between:

  1. surgical and conservative treatment

  2. different methods of surgical treatment

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials (RCTs) and quasi-randomised controlled trials (method of allocating participants to a treatment which is not strictly random: e.g. by date of birth, hospital record number, alternation) evaluating surgical interventions for treating distal humeral fractures.

Types of participants

Skeletally mature people of either gender with a fracture of the distal humerus. Trials containing participants with other elbow injuries were included provided the primary focus is on the treatment of the distal humeral fracture. Studies testing secondary surgery after failed non-surgical treatment were excluded.

Types of interventions

All types of surgical interventions including open reduction and internal fixation (ORIF: e.g. locking plates, reconstruction plate, lag screws and K-wires) and joint arthroplasty/hemiarthroplasty were included.

The two main comparisons were:

  • Any surgery compared with conservative treatment (e.g. arm sling, splint, closed reduction/traction if displaced and cast immobilisation)

  • ORIF as a whole compared with joint arthroplasty

Different types of surgical interventions (e.g. different types of plating) were also included.

Studies testing secondary surgery after failed non-surgical treatment were excluded.

Types of outcome measures

Primary outcomes
  1. Functional outcomes: patient-reported measures of joint functions (e.g. the Mayo Elbow Performance Score (MEPS), the Disability of the Arm, Shoulder, Hand (DASH) score and other validated clinical rating scales)

  2. Treatment failure (e.g. further or secondary surgery indicated)

Secondary outcomes
  1. Pain (long-term)

  2. Adverse events (e.g. malunion/nonunion, ulnar neuropathy, post-traumatic arthritis, heterotopic ossification, infection)

  3. Elbow range of motion (elbow extension/flexion, forearm supination/pronation)

  4. Anatomical outcomes (e.g. distal humeral tilt in radiographic observations)

Where possible, adverse events were pooled.

Timing of outcome assessment

Depending on data availability, we presented outcome data for the following time periods: short term follow-up (up to six months following treatment); intermediate follow-up (more than six months and up to 12 months after the end of treatment) and long-term (greater than 12 months after the end of treatment). Longer term outcome data (two years and above) were extracted and assessed if available. Different time periods were analysed separately.

Search methods for identification of studies

Electronic searches

We searched the Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (May 2012), the Cochrane Central Register of Controlled Trials (The Cochrane Library, 2012 Issue 4), MEDLINE (1946 to April Week 4 2012), and EMBASE (1980 to 2012 Week 17). No language or publication restrictions were applied.

In MEDLINE (Ovid Online), the subject-specific strategy was combined with the sensitivity-maximising version of the Cochrane Highly Sensitive Search Strategy for identifying randomised trials (Lefebvre 2011). Search strategies for The Cochrane Library, MEDLINE and EMBASE are shown in Appendix 1.

In order to identify recently completed, unpublished and ongoing trials, we searched Current Controlled Trials and the WHO International Clinical Trials Registry Platform (1st May 2012).

Searching other resources

We checked the reference lists of published studies to identify additional trials. Where necessary, authors of identified studies and companies who produce relevant products were contacted.

Data collection and analysis

Selection of studies

Two review authors (ZQ, WY) assessed the results of the searches. The authors independently scanned the titles, abstracts and keywords of every record retrieved. Where necessary, full articles were retrieved for further assessment. Discrepancies were resolved by group discussion.

Data extraction and management

Each trial was assessed independently by two review authors (ZQ, YW). Information on trial design, study population, interventions and outcomes were extracted independently by two authors (ZQ, YW) using a data extraction form specifically designed for this review. Discrepancies were resolved by scrutinizing the original article until a consensus was reached. We tried to contact the original authors of identified studies for missing information.

Assessment of risk of bias in included studies

Risk of bias was assessed independently by two authors (ZQ, YW) using the criteria described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). The following sources of bias were assessed:

  • Randomisation process: assessment for selection bias

  • Allocation concealment process: assessment of selection bias

  • Blinding of participants and care personnel: assessment for performance bias

  • Blinding of outcome assessors: assessment for detection bias. Subjective and objective outcomes were assessed separately

  • Incomplete data bias

  • Selective reporting bias

  • Other potential source of bias including major imbalances in key baseline characteristics (e.g. gender, age, fracture type) and in care programmes and surgical experience

Where necessary, disagreements were resolved by discussion.

Measures of treatment effect

The data extracted comprised both event (dichotomous) data (e.g. adverse events as in secondary outcomes) and continuous data (e.g. primary outcomes). Risk ratios (RR) with 95% confidence intervals (CI) were used for reporting dichotomous data. Continuous data were presented as mean differences (MD) with 95% confidence intervals (CI).

Unit of analysis issues

The unit of analysis was individual patients but we remained alert to other potential unit of analyses issues, such as the inclusion of patients undergoing more than one intervention, repeated observation from more than one time-point, and multiple observations for the same outcome (e.g. total adverse events).

Dealing with missing data

When necessary, missing data, particularly denominators and standard deviations, were sought from the authors of the primary studies. Intention-to-treat analyses were performed wherever possible. Missing standard deviations were not imputed.

Assessment of heterogeneity

Besides the visual inspection of forest plot analyses, we intended to examine heterogeneity using the Chi² statistic with significance set at P < 0.1. The I² statistic (Higgins 2003) was to be used to estimate total variation across studies due to heterogeneity rather than chance or sampling error where 0% to 40% might not be important, 30% to 60% may represent moderate levels, 50% to 75% may represent substantial heterogeneity and 75% to 100% considerable heterogeneity (Higgins 2011).

Assessment of reporting biases

Should sufficient data (10 or more trials) become available in future updates of the review, publication bias will be explored by producing funnel plots (Egger 1997).

Data synthesis

Should sufficient homogeneous studies become available, we will decide whether to pool data in a meta-analysis based on statistical and clinical heterogeneity between included studies. A fixed-effect model will be used unless there is significant or unexplained heterogeneity, in which case a random-effects model will be used.

Subgroup analysis and investigation of heterogeneity

Should sufficient data become available in future updates of the review, we plan to explore possible sources of heterogeneity using subgroup analysis based on:

  1. Fracture types (the AO classification system categories: A, B and C)

  2. Age (under versus over 65 years old)

  3. Different interventions (main categories)

  4. Open versus closed fractures

Sensitivity analysis

Should data be available in the future, we plan to perform sensitivity analyses in order to explore the influence of the following factors on effect size, by repeating the analysis and:

  1. excluding unpublished studies;

  2. excluding studies at unclear or high risk of bias relating to non concealment of allocation, non blinding of outcome assessors, incomplete outcome data;

  3. using different measures of calculating treatment effects (e.g. risk difference, odds ratio) and different statistical models (fixed-effect and random-effects models).

Results

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of studies awaiting classification.

See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of studies awaiting classification.

Results of the search

For this search (completed May 2012), we screened a total of 926 records from the following databases: Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (8 records); Cochrane Central Register of Controlled Trials (73), MEDLINE (266), EMBASE (404), Current Controlled Trials (48) and the WHO International Clinical Trials Registry Platform (127). 

The search resulted in the identification of 12 potentially eligible studies. Of these, three were included (McKee 2009; Ruan 2009; Shin 2010), six were excluded (Aktekin 2008; Jain 2009; Mulpuri 2009; Mulpuri 2011; Prasad 2008; Schmale 2009) and three await classification (Frankle 1996; IIyas 2001; Kruger 1985). In addition, a further report (conference proceeding) was identified for one included study (McKee 2009). No ongoing studies met our inclusion criteria.

Included studies

We included three randomised controlled trials (McKee 2009; Ruan 2009; Shin 2010). Detailed descriptions of these are provided in the Characteristics of included studies.

Design

The three included trials were randomised controlled clinical trials. However, Ruan 2009 provided no details about the method of randomisation and thus their use of quasi-randomisation for sequence generation cannot be ruled out. Both McKee 2009 and Shin 2010 employed a computer random number generator as the method of randomisation, and assessors were blinded in both trials, or at best partially in McKee 2009. McKee 2009 reported the use of sealed envelopes for allocation concealment.

Sample size

The sample sizes of the three trials were 42 participants in McKee 2009, 29 in Ruan 2009 and 38 in Shin 2010.

Setting

All three trials took place in university-affiliated hospitals in Canada (McKee 2009), China (Ruan 2009) and South Korea (Shin 2010).

Participants

All three studies exclusively recruited patients with complete intra-articular fractures (OTA/AO type C fractures). One study (McKee 2009) set a lower age limit of 65 years. In McKee 2009 and Shin 2010, the majority of participants were women (87.5% and 65.7% respectively).

Interventions

The three studies compared different surgical interventions: open reduction and internal fixation (ORIF) versus total elbow arthroplasty (TEA) in McKee 2009, perpendicular versus parallel plate fixation strategy in Shin 2010 and anterior subfascial transposition versus in situ decompression of the ulnar nerve in Ruan 2009.

Outcomes

Focusing on primary outcomes only. McKee 2009 reported reoperation rate and patient reported measures of elbow and upper-limb function, namely the Mayo Elbow Performance Score (MEPS) and the Disability of the Arm, Shoulder, Hand (DASH) score, and adverse effects. Shin 2010 reported MEPS as well as the incidence of adverse effects. Ruan 2009 reported on ulnar nerve recovery.

Excluded studies

Brief reasons for the exclusion of six studies are given in the Characteristics of excluded studies. Five were excluded because they focused on children, and another one was excluded because it tested surgery after failed primary treatment.

Risk of bias in included studies

Judgments of the risk of bias, based on eight items for each individual study, are described in the risk of bias tables in the Characteristics of included studies. Figure 1 and Figure 2 also provide summaries.

Figure 1.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Figure 2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Allocation

This was mainly assessed on whether true randomisation and allocation concealment were carried out. McKee 2009 was judged as a low risk of selection bias due to its adequate sequence generation and allocation concealment. Shin 2010 employed a true randomisation method, but lacked information on the concealment of allocation. As Ruan 2009 did not describe the methods of randomisation or allocation concealment, we judged the trial to be at unclear risk of selection bias.

Blinding

Given that the blinding of patients and surgeons is impractical due to the physical nature of the surgical interventions and related ethical issues, all three trials were judged at high risk of performance bias relating to lack of blinding.

The blinding of both participants and outcome assessors are necessary when judging a low risk of detection bias for primarily subjective outcomes such as pain and functional. Only McKee 2009 reported assessor blinding, in this case of the research assistants undertaking outcome measurement. McKee 2009 was rated at unclear risk of bias for such outcomes, whereas the other two trials were rated at high risk of bias. We judged that the risk of bias was low for the assessment of outcomes such as re-operation and treatment failure.

Incomplete outcome data

Ruan 2009 reported no losses of follow-up, and was thus considered at a low risk of attrition bias. McKee 2009, which excluded one patient who had died from each treatment group, was judged at high risk of bias from its failure to present intention-to-treat analyses. In this trial, five patients allocated to ORIF but who received TEA were analysed in the TEA group. Shin 2010 reported a loss of follow-up of three patients, and as in McKee 2009, no description was made regarding at what stage these losses occurred.

Selective reporting

The lack of information regarding trial registration and absence of trial protocols for the three included studies resulted in a judgement of unclear bias for selective reporting. However, there was mention of "the study protocol" in the Discussion section of McKee 2009.

Other potential sources of bias

Baseline characteristics

No trial was judged at high risk of bias due to confounding factors resulting from major imbalances in baseline characteristics. However, Ruan 2009 was considered at unclear risk of bias for this item because separate data for the two treatment groups were not available for severity of preoperative ulnar nerve dysfunction.

Care programmes and surgical experience

Risk of performance bias from significant disparities in care programmes and surgical experience was judged to be unclear for all three trials. In McKee 2009, there were six surgeons performing the operations, but although their surgical seniority or experience was not reported, they all followed standardised protocols. In addition, different types of plates and plating techniques were used in the ORIF group, which was also the case in Shin 2010. In Ruan 2009, the number of surgeons involved was not mentioned. All three trials used the same postoperative rehabilitation protocol between comparison groups within the trial.

Another potential source of other bias pertaining to Ruan 2009 and Shin 2010 is that the follow-up was performed at particular time points rather than having set follow-up times. As a consequence, the length of follow-up in the two trials ranged from 25 to 38 months, and 24 to 93 months respectively.

Effects of interventions

The outcomes reported in the three included studies are listed in the Characteristics of included studies. Meta-analysis was not performed because all three included trials tested different comparisons.

Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA)

McKee 2009 compared ORIF with TEA in 42 elderly patients (> 65 years) with displaced intra-articular distal humeral fractures (OTA/AO type C), 40 of whom who were followed up for two years. On-treatment (treatment received), rather than intention-to-treat, results were presented in the trial report. Thus, the results for five participants originally allocated to ORIF but who received TEA instead were reported for the TEA group. The authors reported statistically significantly poorer MEPS (Mayo Elbow Performance Score) results in the ORIF group than in the TEA group at six months, 12 months and 2 years (see Analysis 1.1). Categorisation of the MEPS scores, showed a non-statistically significant trend to fewer ORIF participants having an excellent or good result (8/15 versus 21/25; RR 0.63, 95% CI 0.38 to 1.05; see Analysis 1.2). The DASH (Disability of the Arm, Shoulder, Hand) scores also favoured the TEA group but the difference between the two groups was statistically significant at six months only (see Analysis 1.3).

As mentioned above, five out of 21 patients allocated ORIF in McKee 2009 underwent intraoperative conversion to TEA. This was because of "extensive comminution and an inability to achieve stable fixation". These patients were then reported as part of the TEA group in the analyses. Furthermore there were two patients, one in each group, whose data were excluded as they had died from "unrelated cause" before the end of follow-up. Based on a treatment received analysis, there was no statistically significant difference between the two groups in the re-operation rate difference between the two groups (4/15 versus 3/25; RR 2.22, 95% CI 0.57 to 8.60; see Analysis 1.4). The original group allocation of the three TEA group participants having a re-operation was not given. Assuming this applied to none of these, an intention-to-treat analysis of treatment failure, where the cross-over participants are placed in their original treatment group, is in favour of TEA although the result did not reach statistical significance (9/21 versus 3/21; RR 3.00, 95% CI 0.94 to 9.55; see Analysis 1.5).

Similar numbers of participants in the two groups had one or more complications (8/15 versus 10/25; RR 1.33, 95% CI 0.68 to 2.62; see Analysis 1.6). Analysis 1.6 also shows the incidence of individual complications of which ulnar nerve dysfunction (sensory and motor deficit included) was the most common.

Based on the on-treatment data, McKee 2009 reported no statistically significant difference between the two groups regarding mean flexion-extension arc and mean pronation-supination arc (see Analysis 1.7).

The duration of surgery was significantly less in the TEA group (MD 32.00 min, 95% CI 11.08 to 52.92 min; see Analysis 1.8). There was no statistically significant difference in length of hospital stay (1.60 days, 95% CI -3.89 to 7.09 days; see Analysis 1.9).

Perpendicular versus parallel configuration of double plate fixation in ORIF

Shin 2010 compared perpendicular versus parallel configuration of double plate fixation in 38 adults with closed intraarticular distal humeral fractures (OTA/AO type C). After exclusion of three participants who were lost to follow-up, 35 participants were followed up for a minimum of two years. Intention-to-treat (ITT) analysis was not conducted.

The authors reported finding no significant differences in outcome between the two groups. This included the MEPS results (see Analysis 2.1). Similarly, there was no significant difference between the two groups in the numbers with good or excellent results from MEPS (14/17 versus 16/18; RR 0.93, 95% CI 0.70 to 1.22; see Analysis 2.2). Three participants in each group had a re-operation for a complication, including non-union (both in the perpendicular group) and heterotopic ossification (RR 1.06, 95% CI 0.25 to 4.54; see Analysis 2.3). Eleven participants in all requested hardware removal (5/17 versus 6/18; RR 0.88, 95% CI 0.33 to 2.36; see Analysis 2.3). Similar numbers of participants in the two groups had one or more complications (5/17 versus 7/18; RR 0.76, 95% CI 0.30 to 1.93; see Analysis 2.4). Most of the individual complications shown in Analysis 2.4 resulted in a re-operation.

There was no significant difference found in the mean flexion-extension arc (see Analysis 2.5: MD -6.00°, 95 CI -20.02° to 8.02°). A similar lack of difference was reported for time to bony union (see Analysis 2.6).

Ulnar nerve transposition versus in situ decompression (ISD)

Ruan 2009 randomised 29 patients with displaced, comminuted distal humeral fractures (OTA/AO type C) and concomitant preoperative ulnar nerve compression symptoms to either anterior subfascial transposition or in situ decompression (ISD). Medial and lateral plates were used for internal fixation of the fractures. There were no losses to follow-up.

Although favouring the transposition group, there was no statistically significant difference between the two groups in those with complete recovery of ulnar nerve function (see Analysis 3.1: 12/15 versus 8/14; RR 1.40, 95% CI 0.83 to 2.35). The remaining participants (3 versus 6) recovered only 'partially'.

Based on the Bishop rating system, used by the study as the major outcome measure, the results again favoured the transposition group in terms of those with excellent results (11/15 versus 5/14; RR 2.05, 95% CI 0.95 to 4.42) or those with good or excellent results (13/15 versus 8/14; RR 1.52, 95% CI 0.92 to 2.49). Neither result was statistically significant (see Analysis 3.2).

Discussion

Summary of main results

This review aimed to cover all surgical treatment methods, including when compared with conservative treatment, for distal humeral fractures in the adult population. Despite a comprehensive search, only three small randomised controlled trials met our inclusion criteria. Each tested different surgical interventions for treating OTA/AO type C distal humeral fractures.

Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA)

McKee 2009, a multicenter trial, compared ORIF and TEA in 42 patients aged over 65 years with OTA/AO type C distal humeral fractures. Forty participants were followed up for two years. Five patients allocated ORIF underwent intraoperative conversion to TEA because of "extensive comminution and an inability to achieve stable fixation". These participants were crossed-over to the TEA group in the analyses. The on-treatment analysis results showed consistently higher MEPS results for the TEA group compared with the ORIF group at 6, 12 and 24 months. For DASH scores, TEA showed only a short-term advantage at six months, which no longer applied at 12 and 24 months. The reoperation rate, complication rate and elbow range of motion results showed no statistically significant differences between the two groups. While an intention-to-treat analysis of treatment failure, where the five cross-over participants are placed in their original allocated group, is in favour of TEA, the result did not reach statistical significance (9/21 versus 3/21; RR 3.00, 95% CI 0.94 to 9.55). McKee 2009 reported that the results of the intention-to-treat analysis were not significantly different to those of the on-treatment analyses but did not present data to confirm this.

Perpendicular versus parallel configuration of double plate fixation

Shin 2010 compared these two different configurations of dual plate fixation and reported results for 35 participants. There was a consistent finding of no significant differences between perpendicular and parallel plate positioning, in terms of MEPS, re-operation, complications and elbow joint range of motion. However, it is worth noting that this study may be underpowered to detect any clinically meaningful differences between the two groups.

Ulnar nerve transposition versus in situ decompression (ISD)

Ruan 2009 randomised 29 patients with preoperative ulnar nerve compression symptoms to either anterior subfascial transposition or in situ decompression of the ulnar nerve. There were no losses to follow-up. Although results for complete recovery of ulnar nerve function and grades based on the Bishop rating system tended to favour the transposition group, none of the differences were statistically significant.

Overall completeness and applicability of evidence

Since all eligible studies exclusively examined patients with OTA/AO type C distal humeral fractures, and one (McKee 2009) only with patients over the age of 65 years, the applicability of evidence may be limited to the population and fracture pattern of the included studies. For instance, it is generally considered that every effort should be made to treat young adults with these fractures with ORIF rather than TEA (Nauth 2011).

To assess the external validity of the evidence from individual trials, we considered following items to be of vital importance: a clearly defined research question in terms of population, intervention and well-accepted outcome measures, and proper timing of follow-up (see Table 1).

Table 1. Items relating to applicability of evidence
 Clearly defined study population?Clearly defined intervention?Well-accepted and comprehensive outcome system?Adequate timing of follow-up?
McKee 2009YesYesYesPartial: the length of follow-up in ORIF is adequate (2 years); but 2 years is a relatively short time for joint arthroplasty assessment.
Shin 2010YesYesPartial: due to the lack of the more detailed validated patient-based functional outcome such as DASH scoring system.Partial: although an average time of follow-up being 40.3 months, the author did not clarify when each outcomes measure in the study were assessed.
Ruan 2009Partial: no clearly defined exclusion criteria reportedYesPartial: we considered Bishop rating system as a partially validated scoring system.Yes (29.5 months on average, ranging from 25 to 38 months)

The study population was relatively clearly defined in all three included studies, except for Ruan 2009 in which the exclusion criteria were not explicitly delineated. Ruan 2009 is also exceptional in that it reported a relatively high incidence of pre-operative ulnar neuropathy (25% = 29 of the consecutive series of 117 patients) among patients with a distal humeral fractures, despite the preponderance of type C fractures in their series. Just eight (2.5%) of the consecutive series of 320 patients in Robinson 2003 were reported to have pre-operative ulnar nerve palsy. All three included studies in the review were conducted by researchers and surgeons in university affiliated hospitals.

All three studies provided sufficient information on their interventions. However, it is worth noting that different types of osteosynthetic plates were used between treatment groups in Shin 2010 as well as with the ORIF procedures in McKee 2009, which may suggest that the final outcome is likely to be influenced by such variation, and potential confounding. Furthermore, the recent development and increasing popularity of locking plating system, which may allow stronger constructs and fixation with even less soft tissue and blood supply compromise in more complex and osteoporotic fractures, is challenging the use of the traditional plates. Unfortunately, no locking plate was employed in all three studies, thus limiting the applicability of the current review, especially in the hospitals where locking plates are available and popular.

McKee 2009 employed well-accepted and comprehensive scoring systems for functional outcome, including physician-based MEPS and patient-based DASH score. Shin 2010, by contrast, was weaker due to lack of evaluations based on patients' subjective outcomes such as DASH score. Ruan 2009 used the Bishop rating system as its major outcome measure; however, we regard this scoring system only as partially validated.

In terms of the timing and length of follow-up, McKee 2009 was labelled as partially inadequate, because, although sufficient for the ORIF group, two years represents a relatively short time in terms of joint arthroplasty surveillance. Shin 2010 and Ruan 2009, with mean length of follow-ups of 40.3 months and 29.5 months respectively, were both deemed adequate for their outcomes.

Quality of the evidence

Overall, the quality of the evidence base of this review is limited. There were a number of methodological drawbacks leading to risks for bias affecting the validity of the findings of all three trials. These include a potential risk of selection bias associated with a lack of allocation concealment (Ruan 2009; Shin 2010), and high risks of performance bias relating to the unavoidable lack of blinding of surgeons in all three trials, and of detection bias relating to lack of full blinding of outcome assessors for subjective outcomes in two trials (Ruan 2009; Shin 2010). There was also high risk of attrition bias, reflecting losses to follow-up and lack of intention-to-treat analysis, in two trials (McKee 2009; Shin 2010).

The major weakness for all three included studies is that the sample size was small (42 for McKee 2009, 29 for Ruan 2009, and 38 for Shin 2010). This makes them susceptible to being underpowered to detect clinically significant differences. Only McKee 2009 reported a sample size calculation, which was based on the reoperation rate.

None of the studies strictly abided by the reporting criteria laid down in the CONSORT statement (CONSORT 2001; CONSORT 2010).

Potential biases in the review process

A comprehensive and systematic search was conducted of key electronic databases and the reference lists of relevant articles. However, there is always a possibility that we have missed unreported trials or those that only appeared in unpublished conference abstracts. Examples of the lattermost are two trials in the Characteristics of studies awaiting classification (Frankle 1996; Kruger 1985). Both abstracts report small trials (10 for two groups and 45 for four groups respectively), which were conducted some time ago (26 years and 16 years). While both abstracts mention randomisation, there is very little other information on methods and it is likely that, as well as being underpowered, both trials are at high risk of bias.

Unavoidably, on treatment, rather than intention-to-treat analyses were presented for McKee 2009 due to the availability of data. Attempts made to contact the authors to obtain relevant data were unsuccessful.

Agreements and disagreements with other studies or reviews

A recently published systematic review (Mauffrey 2012) comparing ORIF with TEA for displaced intraarticular fractures of the distal humerus included one randomised trial (McKee 2009) and one non-randomised retrospective comparative series (Frankle 2003). The authors concluded that "there is insufficient evidence to advocate the routine use of primary TEA in favour of ORIF for displaced intraarticular fractures of the distal humerus". Another review (Hanso 2010) making the same comparison included the same two studies. Its conclusion, however, was a thoughtfully considered opinion of treatment options that exceeded the reviewed evidence. A traditional (not systematic) review (Nauth 2011), which covered all interventions for these fractures, summarised the findings of randomised trials along with several non-randomised comparative studies and case series and provided overall graded recommendations.

All three reviews noted the limitations in the available evidence and emphasized the importance and necessity of well-designed prospective, multicentre, large-scale trials to inform clinical decision-making in the treatment of distal humeral fractures. We further argue that such trials should involve the random allocation of patients.

Authors' conclusions

Implications for practice

Overall, this review found there is either no or insufficient evidence from randomised or quasi-randomised controlled trials to determine whether surgery is, and which surgical interventions are, the most appropriate for the management of different types of distal humerus fractures.

Implications for research

There are numerous issues relating to the management of distal humeral fractures that have not been addressed. These include when surgery is required and the evaluation of current interventions, such as pre-contoured and locking plating systems. In light of this, well-designed multicentre randomised controlled trials with larger sample sizes are required to inform clinical practice in the treatment of distal humeral fractures. When planning such trials, trialists need to consider whether the trial should include both or just one of the two distinct patient populations (the early and the later peak populations) that incur these injuries. Better still, specialists in this area should come together to identify the priority questions for future multicentre trials.

Fully reporting such trials, and abiding by the CONSORT statement (CONSORT 2010), is highly recommended for future studies. Specifically, it is impractical to blind the surgeons and the patients due to the physical nature of the surgical intervention and related ethical issues. However, blinding of outcome assessors was perfectly justified and extremely important for the functional outcomes such as MEPS, DASH and pain. To further avoid performance bias, it is important that all surgeons performing the trial interventions have comparable and sufficient experience with the interventions under test. In particular for TEA surveillance, a longer term follow-up is strongly recommended. We also recommend that standardised and validated functional outcomes such as DASH and MEPS be used in all future studies.

Also crucial is the maintenance and update of this Cochrane review when further evidence in the form of randomised or quasi-randomised controlled trials becomes available.

Acknowledgements

Our gratitude goes to Helen Handoll, David Ring, Cathie Sherrington and Haris Vasiliadis for valuable comments about the protocol and review. We also acknowledge the help of Catherine Deering for her assistance with the development of the search strategies and Lindsey Elstub for kind help and helpful comments along the way. We thank Joanne Elliott for kindly and patiently providing several search results.

Data and analyses

Download statistical data

Comparison 1. Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA)
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Mayo elbow performance score (MEPS): 0 to 100 (best outcome)  Other dataNo numeric data
1.2 6 months (on treatment data)  Other dataNo numeric data
1.3 12 months (on treatment data)  Other dataNo numeric data
1.4 2 years (on treatment data)  Other dataNo numeric data
2 Excellent or good MEPS at 2 years1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
3 Disability of the Arm, Shoulder, Hand score (DASH): 0 to 100 (worst outcome)  Other dataNo numeric data
3.2 6 months (on treatment data)  Other dataNo numeric data
3.3 12 months (on treatment data)  Other dataNo numeric data
3.4 2 years (on treatment data)  Other dataNo numeric data
4 Reoperation rate (treatment received analysis))1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
5 Treatment failure (conversion and reoperation) (intention-to-treat analysis)1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
6 Adverse events1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
6.1 One or more complications1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
6.2 Ulnar nerve dysfunction1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
6.3 Nonunion1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
6.4 Post-traumatic stiffness1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
6.5 Heterotopic ossification (type III)1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
6.6 Infection1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
6.7 Would complication1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
6.8 Radial nerve palsy (temporary)1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
6.9 Malalignment > 5°1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
6.10 Incongruent reduction >2mm1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
7 Range of motion at 2 years (degrees)  Other dataNo numeric data
7.1 Extension (on treatment data)  Other dataNo numeric data
7.2 Flexion (on treatment data)  Other dataNo numeric data
7.3 Arc of motion of flexion-extension (on treatment data)  Other dataNo numeric data
7.4 Arc of motion of pronation-supination (on treatment data)  Other dataNo numeric data
8 Duration of operation (minutes)1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
9 Length of hospital stay (days)1 Mean Difference (IV, Fixed, 95% CI)Totals not selected

Analysis 1.1.

Comparison 1 Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA), Outcome 1 Mayo elbow performance score (MEPS): 0 to 100 (best outcome).

Mayo elbow performance score (MEPS): 0 to 100 (best outcome)
StudyOutcome measureORIF (n = 15)TEA (n = 25)Reported significance
6 months (on treatment data)
McKee 2009MEPSmean = 68mean = 86P = 0.003
12 months (on treatment data)
McKee 2009MEPSmean = 72mean = 88P = 0.007
2 years (on treatment data)
McKee 2009MEPSmean = 73mean = 86P=0.015
Analysis 1.2.

Comparison 1 Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA), Outcome 2 Excellent or good MEPS at 2 years.

Analysis 1.3.

Comparison 1 Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA), Outcome 3 Disability of the Arm, Shoulder, Hand score (DASH): 0 to 100 (worst outcome).

Disability of the Arm, Shoulder, Hand score (DASH): 0 to 100 (worst outcome)
StudyOutcome measureORIF (n = 15)TEA (n = 25)Reported significance
6 months (on treatment data)
McKee 2009DASHmean = 47mean = 31P = 0.04
12 months (on treatment data)
McKee 2009DASHmean = 47mean = 31P = 0.07
2 years (on treatment data)
McKee 2009DASHmean = 43mean = 32P = 0.18
Analysis 1.4.

Comparison 1 Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA), Outcome 4 Reoperation rate (treatment received analysis)).

Analysis 1.5.

Comparison 1 Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA), Outcome 5 Treatment failure (conversion and reoperation) (intention-to-treat analysis).

Analysis 1.6.

Comparison 1 Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA), Outcome 6 Adverse events.

Analysis 1.7.

Comparison 1 Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA), Outcome 7 Range of motion at 2 years (degrees).

Range of motion at 2 years (degrees)
StudyOutcome measureORIF
(n = 15)
TEA
(n = 25)
Reported significance
Extension (on treatment data)
McKee 2009Extension

mean = 28°

range = 5° to 60°

mean = 26°

range = 0° to 72°

Statistical significance was not reported. Quote: " Although there was a trend toward improved motion in the TEA group (by 12°), the mean extension, flexion, and arc of motion of flexion-extension were not significantly different between the ORIF group and TEA group at 2 years."
Flexion (on treatment data)
McKee 2009Flexion

mean = 123°

range = 90° to 150°

mean = 133°

range = 90° to 155°

Statistical significance was not reported. Quote: " Although there was a trend toward improved motion in the TEA group (by 12°), the mean extension, flexion, and arc of motion of flexion-extension were not significantly different between the ORIF group and TEA group at 2 years."
Arc of motion of flexion-extension (on treatment data)
McKee 2009Arc of motion of flexion-extension

mean = 95°

range = 30° to 140°

mean = 107°

range = 42° to 145°

Statistical significance was not reported. Quote: "Although there was a trend toward improved motion in the TEA group (by 12°), the mean extension, flexion, and arc of motion of flexion-extension were not significantly different between the ORIF group and TEA group at 2 years."
Arc of motion of pronation-supination (on treatment data)
McKee 2009Arc of motion of pronation-supination

mean = 160°

range = 90° to 180°

mean = 172°

range = 110° to 180°

P = 0.13
Analysis 1.8.

Comparison 1 Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA), Outcome 8 Duration of operation (minutes).

Analysis 1.9.

Comparison 1 Open reduction - internal fixation (ORIF) versus total elbow arthroplasty (TEA), Outcome 9 Length of hospital stay (days).

Comparison 2. Perpendicular versus parallel double plate fixation
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Mayo elbow performance score (MEPS) at 2+ years: 0 to 100 (best outcome)  Other dataNo numeric data
2 Excellent or good results from MEPS1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
3 Re-operation1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
3.1 Re-operation for complication1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
3.2 Requested hardware removal1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
4 Adverse events1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
4.1 Patients with complications1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
4.2 Nonunion1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
4.3 Arthrofibrosis1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
4.4 Heterotopic ossification1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
4.5 Ulnar never dysfunction persisting after 3 months postoperatively1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
5 Range of motion (degrees)1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
5.1 Arc of flexion1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
5.2 Flexion1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
5.3 Extension1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
6 Bony union time (months)  Other dataNo numeric data

Analysis 2.1.

Comparison 2 Perpendicular versus parallel double plate fixation, Outcome 1 Mayo elbow performance score (MEPS) at 2+ years: 0 to 100 (best outcome).

Mayo elbow performance score (MEPS) at 2+ years: 0 to 100 (best outcome)
StudyOutcome measurePerpendicular group (n=17)Parallel group (n=18)Reported significance
Shin 2010MEPS

mean = 91.5

range = 70 to 100

mean = 94.3

range = 70 to 100

P = 0.928
Analysis 2.2.

Comparison 2 Perpendicular versus parallel double plate fixation, Outcome 2 Excellent or good results from MEPS.

Analysis 2.3.

Comparison 2 Perpendicular versus parallel double plate fixation, Outcome 3 Re-operation.

Analysis 2.4.

Comparison 2 Perpendicular versus parallel double plate fixation, Outcome 4 Adverse events.

Analysis 2.5.

Comparison 2 Perpendicular versus parallel double plate fixation, Outcome 5 Range of motion (degrees).

Analysis 2.6.

Comparison 2 Perpendicular versus parallel double plate fixation, Outcome 6 Bony union time (months).

Bony union time (months)
StudyOutcome measurePerpendicular group (n = 17)Parallel group (n = 18)Reported significance
Shin 2010Bony union timemean = 6.3mean = 5.4P = 0.436
Comparison 3. Ulnar nerve transposition group versus in situ decompression (ISD) group
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Complete recovery of ulnar nerve function1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
2 Bishop rating system of ulnar nerve function grades1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
2.1 Good or excellent results1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
2.2 Excellent results1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
Analysis 3.1.

Comparison 3 Ulnar nerve transposition group versus in situ decompression (ISD) group, Outcome 1 Complete recovery of ulnar nerve function.

Analysis 3.2.

Comparison 3 Ulnar nerve transposition group versus in situ decompression (ISD) group, Outcome 2 Bishop rating system of ulnar nerve function grades.

Appendices

Appendix 1. Search strategies

Completed May 2012

The Cochrane Central Register of Controlled Trials (Wiley Online Library)

#1 MeSH descriptor Humerus, this term only (50)
#2  MeSH descriptor Elbow, this term only (120)
#3  (#1 OR #2) (168)
#4  MeSH descriptor Fractures, Bone, this term only (994)
#5  MeSH descriptor Fracture Fixation explode all trees (973)
#6  MeSH descriptor Fracture Healing, this term only (318)
#7  (#4 OR #5 OR #6) (1931)
#8   (#3 AND #7) (13)
#9   MeSH descriptor Humeral Fractures, this term only with qualifier: SU (40)
#10 (humor* or elbow*) NEAR/3 (fractur* or fixat*):ti,ab,kw (173)
#11 (#8 OR #9 OR #10) (175)
#12 (distal* or elbow* or condyl* or supracondyl* or epicondyl* or transcondyl* or intercondyl* or bicondyl* or transchondral*):ti,ab,kw (6266)
#13 (#11 AND #12) [Trials] (73)

MEDLINE (Ovid Online)

1  Humerus/ or Elbow/ (12117)
2   Fractures, Bone/ (43616)
3   exp Fracture Fixation/ (43154)
4   Fracture Healing/ (8077)
5   2 or 3 or 4 (80933)
6   1 and 5 (1153)
7   exp Humeral Fractures/su [Surgery] (3173)
8   ((humor$ or elbow$) adj3 (fractur$ or fixat$)).tw. (4533)
9   6 or 7 or 8 (6614)
10 (distal$ or elbow$ or condyl$ or supracondyl$ or epicondyl$ or transcondyl$ or intercondyl$ or bicondyl$ or transchondral$).tw. (188754)
11 9 and 10 (3200)
12 Randomized controlled trial.pt. (325981)
13 Controlled clinical trial.pt. (84018)
14 randomized.ab. (230161)
15 placebo.ab. (130821)
16 Drug therapy.fs. (1527419)
17 randomly.ab. (166144)
18 trial.ab. (238081)
19 groups.ab. (1093016)
20 or/12-19 (2835150)
21 exp Animals/ not Humans/ (3710538)
22 20 not 21 (2407260)
23 11 and 22 (266)

EMBASE (Ovid Online)

1   Humerus/ or Elbow/ (17874)
2   Fracture/ (51878)
3   Fracture Healing/ (14122)
4   exp Fracture Treatment/ (71656)
5   2 or 3 or 4 (121552)
6   1 and 5 (2252)
7   Humerus Fracture/ or Humerus Supracondylar Fracture/ (7645)
8   ((humor$ or elbow$) adj3 (fractur$ or fixat$)).tw. (5606)
9   6 or 7 or 8 (10487)
10 (distal$ or elbow$ or condyl$ or supracondyl$ or epicondyl$ or transcondyl$ or intercondyl$ or bicondyl$ or transchondral$).tw. (224287)
11 9 and 10 (4459)
12 Randomized Controlled Trial/ (320525)
13 Clinical Trial/ (864496)
14 Controlled Clinical Trial/ (388165)
15 Randomization/ (57934)
16 Single Blind Procedure/ (15770)
17 Double Blind Procedure/ (108422)
18 Crossover Procedure/ (33662)
19 Placebo/ (197016)
20 prospective Study/ (201807)
21 ((clinical or controlled or comparative or placebo or prospective$ or randomi#ed) adj3 (trial or study)).tw. (618456)
22 (random$ adj7 (allocat$ or allot$ or assign$ or basis$ or divid$ or order$)).tw. (150311)
23 ((singl$ or doubl$ or trebl$ or tripl$) adj7 (blind$ or mask$)).tw. (142490)
24 (cross?over$ or (cross adj1 over$)).tw. (60730)
25 ((allocat$ or allot$ or assign$ or divid$) adj3 (condition$ or experiment$ or intervention$ or treatment$ or therap$ or control$ or group$)).tw. (187331)
26 RCT.tw. (9039)
27 or/12-26 (1659860)
28 Case Study/ or Abstract Report/ or Letter/ (844248)
29 27 not 28 (1624732)
30 11 and 29 (404)

Contributions of authors

Yan Wang, Qi Zhuo, Wei Yang, and Peifu Tang contributed equally to the review. Qi Zhuo is the guarantor of the review.

Declarations of interest

None known.

Sources of support

Internal sources

  • Chinese PLA General Hospital, China.

External sources

  • No sources of support supplied

Differences between protocol and review

For the risk of bias assessment tool in the review, we have split the blinding of outcome assessors into two aspects, with one focusing on the functional and clinical outcomes, where blinding is extremely important to avoid risk of detection bias, and the other on objective outcomes, where blinding is unlikely to affect their assessment. We have also specified Other biases in the tool; these were bias associated major differences in baseline characteristics and performance bias associated with differences in care programmes and surgical experience.

We added a table showing our evaluation of items relating to the applicability of evidence and review findings.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

McKee 2009

Methods

Double-blind RCT

Method of randomisation: a random number generator

Assessor blinded

Number lost to follow-up: 2

Length of follow-up: 24 months

Participants

42 patients over the age of 65 years with displaced, comminuted, intraarticular fractures of the distal humerus (OTA/AO type 13C fracture)

4 university-affiliated Canadian medical centres

Date of study: not fully stated (only "beginning at year 2001" was mentioned)

5 male, 35 female; mean age 77 years in ORIF group, 78 years in TEA group

Exclusion criteria: extra-articular or partial articular fractures of the distal humerus, Gustilo grade I open fractures that had not had irrigation and debridement within 12 hours of occurrence, Gustilo grade II, IIIa, IIIb, and IIIc open fractures, associated vascular injury, previous ipsilateral distal humeral fracture, pathologic fractures, fractures with diaphyseal extension of 8 cm or greater, definitive surgery more than 21 days after injury, pre-existing severe joint disease, and limited life expectancy because of significant medical comorbidity.

Interventions

All fractures were treated within 12 hours of injury. General anaesthesia used in all cases. Preoperative prophylactic antibiotics given to all cases. Six surgeons performed the procedures. The ulnar nerve was transposed in all cases. Both the two groups had the same rehabilitation protocol.

ORIF group: dual plates placed on medial and lateral columns in all cases, with 10 placed at 90°, and 5 placed parallel to each other. Different kinds of plates were used, including 15 precontoured plates, nine 3.5-mm pelvic reconstruction plates, and six limited-contact 3.5-mm compression plate. No 1/3 tubular or locking plates were used in the ORIF group. All ORIF was performed via a posterior approach with anatomic reduction of the fracture and provisional Kirschner wire fixation. [25]

TEA group: Semiconstrained (Coonrad - Morrey elbow prosthesis, Zimmer, Warsaw, IN) TEA was implanted with antibiotic-impregnated bone cement (40 grams of Simplex Bone Cement powder plus 1 gram of tobramycin; Stryker Canada, Hamilton, Ontario) in all cases. A midline triceps split or triceps-sparing approach was used for TEA. [15]

Outcomes

Primary outcome: reoperation rate.

Secondary outcome: MEPS and DASH scores.

NotesInitially 42 patients were randomised into two groups with each one comprising 21, but one died in each group and for 5 patients of the ORIF group a decision was made intra-operatively for conversion to TEA. Patients converted from ORIF to TEA were considered in the TEA group in the trial analyses.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "A random number generator was used to create a 1:1 allocation scheme"
Allocation concealment (selection bias)Low riskQuote: "A sealed envelope containing the randomly assigned treatment group allocation was opened by the study coordinator, not the surgeon."
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of the surgeon and the patients in this trial was not practical due to the physical nature of the interventions.
Blinding of outcome assessment (detection bias) (functional outcomes, pain, clinical outcomes, complications)Unclear riskQuote: "A research assistant, blinded to the treatment group, and the attending physician, who was not blinded, performed a standardized clinical assessment of all patients preoperatively and postoperatively." Insufficient information to judge whether the attending physician's awareness of the grouping status would influence the judgment of the research assistant, and thus bias the final conclusion.
Blinding of outcome assessment (detection bias) (reoperation, treatment failure, range of motion)Low riskLack of blinding unlikely to affect assessment of these outcomes.
Incomplete outcome data (attrition bias)
All outcomes
High riskParticipant flow provided, but two (4.8%) died of unrelated causes before complete follow-up and were excluded from the analysis of the data. In addition, there was a 5 out of 20 patients (25%) intraoperative conversion to TEA in the ORIF group because of "extensive comminution and an inability to achieve stable fixation". This means no ITT analysis was followed, despite the author's claim that no significant difference existed (data not presented).
Selective reporting (reporting bias)Unclear riskInsufficient information to judge this.
Balance in baseline characteristicsLow riskQuote: "There were no significant differences between the groups in terms of age, gender, weight, baseline pre-injury DASH, injury severity score, mechanism of injury, smoking status, baseline activity, and fracture type."
Extra performance bias (care programmes and surgical experience, etc.)Unclear riskThere were 6 surgeons performing the operations, and their seniority or experience was not mentioned. Different types of plates and plating techniques were used in the ORIF group. All this increased the heterogeneity, but not necessarily the bias concerned. Conversely, a standardized protocol was followed in the trial. Therefore, McKee 2009 was labelled as unclear risk.

Ruan 2009

Methods

RCT

Method of randomisation not described

Assessor blinding not described

Number lost to follow-up: 0

Length of follow-up: 29.5 months on average, ranging from 25 to 38 months

Participants

29, out of 117 consecutive patients with AO type C fractures of distal humerus, who exhibited preoperative ulnar nerve compression symptoms.

One university affiliated hospital in Shanghai, China

Date of study: from June 1998 to October 2005

9 male and 6 female, mean age 45.1 years in the group with anterior subfascial (meaning subcutaneous) transposition of the ulnar nerve.
10 male and 4 female, mean age 40.7 years in the in situ decompression (ISD) group.

Interventions

All fractures were treated using ORIF within 2 weeks of injury. General anaesthesia was used in all cases. The number of surgeons performing the procedures was not reported. Both groups followed the similar post-operative rehabilitation protocol.

Transposition Group: the ulnar nerve underwent anterior subfascial  transposition along with its arteries.[15]

ISD group: the ulnar nerve underwent decompression in the original position.[14]

OutcomesUlnar nerve recovery and Bishop rating system (No details given)
NotesTwo 3.5-mm reconstruction plates (DePuy ACE, Johnson & Johnson Company, Switzerland) were used in all ORIF procedures, but neither details regarding the position of the dual plates, nor preoperative severity of ulnar nerve dysfunction were provided.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo details provided.
Allocation concealment (selection bias)Unclear riskNo details provided.
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of the surgeon in this trial was not practical, due to the physical nature of the interventions. Blinding of the patients was not reported.
Blinding of outcome assessment (detection bias) (functional outcomes, pain, clinical outcomes, complications)High riskNo details provided, but this was taken to suggest that there was no blinding, and rated as "high risk" for functional outcomes.
Blinding of outcome assessment (detection bias) (reoperation, treatment failure, range of motion)Low riskN/A
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo losses to follow up.
Selective reporting (reporting bias)Unclear riskInsufficient information to judge this.
Balance in baseline characteristicsUnclear riskThe age, time to operation, and fracture subtypes between the two groups was not significantly different. But the severity of ulnar nerve dysfunction was not differentiated between the two groups.
Extra performance bias (care programmes and surgical experience, etc.)Unclear riskThe number of surgeons performing the procedures were not mentioned, but the author suggested that the surgical techniques, plates used, and the postoperative rehabilitation protocols were the same for the two groups.

Shin 2010

  1. a

    DASH = Disability of the Arm, Shoulder, Hand score
    MEPS = Mayo Elbow Performance Score
    ORIF = open reduction and internal fixation
    TEA = total elbow arthroplasty

Methods

RCT

Method of randomisation: a computer random number generator

Assessor not blinded

Number lost to follow-up: 3

Length of follow-up: 40.3 months on average, ranging from 24 to 93 months

Participants

38 consecutive patients (3 lost to follow-up, but no details were reported) with closed intraarticular fractures of the distal humerus (OTA/AO type C fracture)

Study setting: Seoul, South Korea

Date of study: from March 2002 to November 2006

11 male, 23 female; mean age 52.5 years (range 18 to 94) in group using orthogonal plating, 56.1 years (range 18 to 83) in the group using parallel plating

Exclusion criteria: extraarticular distal humerus fractures, or with a previous operative treatment, or of a pathologic fracture around elbow.

Interventions

All fractures were treated within five days of injury, except for one whose surgery was delayed due to internal organ injuries. Method of anaesthesia was not reported. A single surgeon performed the procedures. The ulnar nerve was transposed in all cases. Both groups followed the same post-operative rehabilitation and heterotopic ossification prevention protocol.

Group I: using 2 orthogonal/perpendicular plates, with 1 placed along the medial supracondylar ridge and another placed posterolaterally. A 3.5-mm reconstruction plate was used for the lateral column, while in the medial column, 2 different plates were used, including ten 3.5-mm reconstruction plates, seven pre-contoured anatomical plates (Acumed, Hillsboro, OR). [17]

Group II: dual plates, parallel with each other, placed along each supracondylar ridge. Pre-contoured anatomical plates were employed in all patients of the group. [18]

Outcomes

Incidence of complications, including nonunion, metal failure, and heterotopic ossification.

Range of motion

Union time

MEPS

Notes

Postoperative rehabilitation protocol: "passive and active assisted range of elbow motion exercises commenced at 3 days postoperatively."

Heterotopic ossification prevention: "patients underwent prophylactic radiotherapy 3 times after the operative day, and on each occasion received 200 Cgy. Indomethacin (oral dosage of 75 mg) was administered to all patients for 2 weeks postoperatively."

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "Randomization was performed based on a computer random number generator determining the order of 1 of 2 fixation methods."
Allocation concealment (selection bias)Unclear riskNo details provided.
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of the surgeon in this trial was not practical, due to the physical nature of the interventions. Blinding of the patients was not reported.
Blinding of outcome assessment (detection bias) (functional outcomes, pain, clinical outcomes, complications)High riskNo details provided, but this was taken to suggest that there was no blinding, and rated as "high risk" for functional outcomes.
Blinding of outcome assessment (detection bias) (reoperation, treatment failure, range of motion)Low riskNo details, but lack of blinding unlikely to affect assessment of these outcomes.
Incomplete outcome data (attrition bias)
All outcomes
High risk3 patients (7.9%) were lost to follow-up, but no details were provided. No intention-to-treat (ITT) analysis has been conducted.
Selective reporting (reporting bias)Unclear riskInsufficient information to judge this.
Balance in baseline characteristicsLow riskThe age, male/female ratio, time to operation, fracture subtypes, and common cause of injuries between the 2 groups was not significantly different.
Extra performance bias (care programmes and surgical experience, etc.)Unclear riskPlates used in the ORIF procedures were not always the same between the 2 groups. Otherwise, a single surgeon performed all operations, and the postoperative rehabilitation (etc) were the same for both groups.

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Aktekin 2008Participants were children.
Jain 2009Ongoing study focusing on type I supracondylar humerus fractures in children.
Mulpuri 2009Ongoing study focusing on supracondylar humerus fractures in children.
Mulpuri 2011Ongoing study focusing on type III supracondylar humerus fractures in children.
Prasad 2008Study testing secondary surgery after failed non-surgical or ORIF treatment.
Schmale 2009Ongoing study focusing on supracondylar humerus fractures in children.

Characteristics of studies awaiting assessment [ordered by study ID]

Frankle 1996

MethodsPatients were "randomised by the surgeon" for either total elbow arthroplasty or ORIF. "The patients were evaluated by a research nurse and not under the auspices of either surgeon."
Length of follow-up: "at least eighteen months postoperatively".
Participants10 elderly female with elbow fractures
Interventions

1. Total elbow arthroplasty. [6]

2. ORIF. [4]

OutcomesPatients were evaluated postoperatively "according to the scale of Broberg and Morey".
NotesOf note, the authors described their study as "retrospective" in the study title. There were patients with elbow dislocation and coronoid fractures involved. All operations were performed by the same surgeon. So far, we have no access to the report of this study other than the abstract provided in the AAOS scientific program.

IIyas 2001

MethodsNo randomisation was mentioned in the abstract of the study. Study period: 1993 to 1997. 6 losses to follow-up.
Participants

50 patients with intra-articular fractures of distal humerus.

Study setting: Teaching hospital in Quetta.

Interventions

All patients underwent ORIF, with one of the two following surgical approaches:

1. Triceps Tongue approach

2. Trans-olecranon approach

No information was given regarding the number of participants in each group.

OutcomesNo details were given regarding outcome measures.
NotesWe were only able to find the abstract of the study. No single methodological item was mentioned, and thus it was regarded as highly biased.

Kruger 1985

Methods"In a randomised series", 33 patients out of 45 were followed up ranging from 6 weeks to 2 years.
Participants45 patients with distal humeral fractures
Interventions

1. Overhead skeletal traction

2. Closed reduction

3. "Bag of bone" methods

4. ORIF

No information was given regarding the number of participants in each of the four groups.

Outcomes

Elbow range of motion

Grip strength

Pain

NotesWe were only able to find the abstract of the study, which was conducted 27 years ago.

Ancillary