Clavos intramedulares para la fractura extracapsular de cadera en adultos
Resumen
Antecedentes
Los clavos intramedulares pueden utilizarse para la fijación quirúrgica de fracturas extracapsulares de cadera en adultos. Ésta es una actualización de una revisión Cochrane publicada por primera vez en 2005 y actualizada por última vez en 2008.
Objetivos
Evaluar los efectos (beneficiosos y perjudiciales) de diferentes diseños de los clavos intramedulares para el tratamiento de las fracturas extracapsulares de cadera en adultos.
Métodos de búsqueda
Se hicieron búsquedas en el registro especializado del Grupo Cochrane de Lesiones Óseas, Articulares y Musculares (Cochrane Bone, Joint and Muscle Trauma Group) (6 enero 2014), Registro Cochrane Central de Ensayos Controlados (Cochrane Central Register of Controlled Trials) (The Cochrane Library número 12, 2013), MEDLINE (1966 hasta noviembre, semana 3, 2013), MEDLINE In‐Process & Other Non‐Indexed Citations (3 enero 2014), EMBASE (1988 hasta 2014, semana 1) y en la World Health Organization (WHO) International Clinical Trials Registry Platform (acceso enero 2014).
Criterios de selección
Todos los ensayos aleatorios o cuasialeatorios que comparan diferentes tipos o modificaciones en el diseño de los clavos intramedulares para el tratamiento de la fractura extracapsular de cadera en adultos.
Obtención y análisis de los datos
Al menos dos autores de la revisión seleccionaron los estudios de forma independiente, evaluaron el riesgo de sesgo y extrajeron los datos. Se realizó el metanálisis limitado mediante el modelo de efectos fijos.
Resultados principales
Se incluyeron ocho nuevos ensayos, probando siete nuevas comparaciones en esta actualización. En términos generales, se incluyeron 17 ensayos, que probaban 12 comparaciones de diferentes diseños de clavos cefalocondilares. Los ensayos incluyeron a un total de 2130 adultos (predominantemente mujeres y personas mayores) principalmente con fracturas trocantéricas inestables.
Todos los ensayos estuvieron en riesgo incierto de sesgo para la mayoría de los dominios y la mayoría en riesgo alto de sesgo de detección para los resultados subjetivos. Los tres ensayos cuasialeatorios tenían un alto riesgo de sesgo de selección.
Cuatro estudios (910 participantes) compararon el clavo femoral proximal (CFP) con el clavo Gamma. No hubo diferencias significativas entre los dos implantes en el resultado funcional (las pruebas de muy baja calidad estuvieron limitadas a resultados de ensayos individuales), la mortalidad (pruebas de baja calidad: 86/415 versus 80/415; cociente de riesgos [CR] 1,08; intervalo de confianza [IC] del 95%: 0,82 a 1,41), las complicaciones graves de la fijación (fractura quirúrgica del fémur, expulsión, consolidación viciosa y fractura posterior del fémur) ni en las nuevas cirugías (pruebas de baja calidad: 45/455 versus 36/455; CR 1,25; IC del 95%: 0,83 a 1,90).
Dos ensayos (185 participantes) aportaron pruebas de calidad muy baja de una falta de diferencias clínicamente significativas en el resultado (puntuación funcional, mortalidad, complicaciones con la fijación de la fractura y nueva cirugía) entre el clavo trocantérico ACE y el clavo Gamma.
Dos ensayos (200 participantes) aportaron pruebas de calidad muy baja de una falta de diferencias significativas en el resultado (puntuación de movilidad, dolor, complicaciones con la fijación de la fractura o nuevas cirugías) entre el clavo femoral proximal antirrotación (CFPA) y el clavo Gamma 3.
Siete de los nueve ensayos que evaluaban diferentes comparaciones aportaron pruebas de calidad muy baja de una falta de diferencias significativas entre los grupos en todos los resultados principales informados para las siguientes comparaciones: clavo trocantérico ACE versus clavo Gamma 3 (112 participantes); clavo deslizante versus clavo Gamma (80 participantes); clavo Russell‐Taylor Recon versus clavo Gamma largo (34 participantes, todos menores de 50 años de edad); clavo femoral proximal antirrotación (CFPA) versus clavo Targon PF (80 participantes); clavo con tornillo intramedular de cadera (TIMC) colocado de forma dinámica versus estática (81 participantes); clavo Gamma 3 deslizante versus no deslizante (80 participantes, todos menores de 60 años de edad); y CFPA largos versus estándar (40 participantes con fracturas oblicuas inversas).
Los otros dos ensayos de comparaciones únicas también aportaron pruebas de calidad muy baja de una falta de diferencias significativas entre los grupos en todos los resultados principales con excepciones únicas. El ensayo (215 participantes) que comparó el clavo ENDOVIS versus clavo TIMC encontró pruebas de baja calidad de una movilidad más deficiente en el grupo de clavo ENDOVIS, en el cual hubo más participantes encamados después de la cirugía (29/105 versus 18/110; CR 1,69; IC del 95%: 1,00 a 2,85; P = 0,05). El ensayo (113 participantes) que comparó el clavo InterTan versus clavo CFPA II encontró pruebas de muy baja calidad de que más participantes del grupo de CFPA II experimentaron dolor en el muslo (3/47 versus 12/46; CR: 0,24; IC del 95%: 0,07 a 0,81).
Conclusiones de los autores
Las pruebas limitadas de los ensayos aleatorios realizados hasta la fecha son insuficientes para determinar si hay diferencias importantes en el resultado entre diferentes diseños de clavos intramedulares usados para el tratamiento de las fracturas extracapsulares de cadera. Dadas las pruebas de la superioridad del tornillo deslizante para cadera comparado con los clavos intramedulares para la fractura extracapsular de cadera, los estudios adicionales que comparen diferentes diseños de clavos intramedulares no constituyen una prioridad. Cualquier nuevo diseño debe evaluarse en una comparación aleatoria con el tornillo deslizante para cadera.
PICO
Resumen en términos sencillos
Clavos intramedulares para la fractura extracapsular de cadera en adultos
¿Cuál es el problema médico?
Las fracturas de la parte superior del hueso del muslo (fémur) se denominan fracturas proximales de fémur o fracturas de cadera. Estas fracturas son más frecuentes en las mujeres mayores de 65 años. Cerca de dos de cada cinco fracturas de cadera son "extracapsulares", ya que se producen fuera de la cápsula de la articulación coxofemoral.
¿Qué tratamientos están disponibles?
La mayoría de estas fracturas se fija quirúrgicamente con implantes metálicos. Un tipo de implante de uso cada vez mayor es el “clavo intramedular”. Consta de una varilla metálica, que habitualmente puede insertarse desde la epífisis superior en la cavidad interna (médula) del hueso femoral y sujetarse con tornillos. Hoy en día se utilizan diferentes tipos de clavos, generalmente producidos por diferentes fabricantes.
¿Algunas clavos intramedulares son mejores que otros para estas fracturas?
Esta revisión se propuso examinar las pruebas de los ensayos que compararon diferentes diseños de clavos en la práctica clínica.
Se efectuaron búsquedas en las bases de datos médicas y los registros de estudios nuevos (hasta enero de 2014) y se encontraron 17 ensayos que comparaban clavos de diferentes diseños. Comprendieron un total de 2130 participantes. La mayoría de los participantes eran mujeres mayores.
La calidad de las pruebas de estos ensayos es baja o muy baja, en parte debido a que la mayoría de los ensayos utilizó métodos deficientes que provocan que los resultados no sean fiables. Además, varios ensayos no informaron la función ni proporcionaron datos que pudieran utilizarse. De las 12 comparaciones diferentes que se probaron, nueve fueron evaluadas en un ensayo solamente.
Cuatro ensayos compararon el clavo femoral proximal (CFP) con el clavo Gamma en 910 adultos mayores. Dos ensayos compararon el clavo intramedular ACE con el clavo Gamma en 185 pacientes de edad avanzada. Dos ensayos compararon el clavo femoral proximal antirrotación (CFPA) con el clavo Gamma 3 en 200 adultos de edad avanzada. Los otros nueve ensayos incluyeron comparaciones únicas de diferentes tipos de diseños de clavos.
En términos generales, las pruebas débiles disponibles sobre las 12 comparaciones no mostraron ninguna diferencia importante en el resultado (función, movilidad, dolor, muerte, complicaciones de la fijación de la fractura e intervención quirúrgica de revisión) entre los dos clavos o los dos diseños de clavos que se probaron. Hubo una posible excepción. Hubo pruebas débiles provenientes de un ensayo con 215 pacientes de edad avanzada de que el clavo ENDOVIS dio lugar a una movilidad más deficiente (más pacientes no pudieron caminar después de la cirugía) en comparación con el tornillo intramedular de cadera (TIMC). Sin embargo, se necesitan más pruebas para tener seguridad en cuanto a este resultado.
En conclusión, las pruebas disponibles son insuficientes para determinar si hay diferencias importantes en el resultado entre diferentes diseños de clavos intramedulares utilizados para fijar las fracturas extracapsulares de cadera. En cuanto a la investigación futura, se propone que se asigne prioridad a las comparaciones de los clavos intramedulares con otro tipo de dispositivo que se utiliza comúnmente, el tornillo deslizante de cadera.
Authors' conclusions
Background
Description of the condition
Hip fracture is the general term for fracture of the proximal (upper) femur. These fractures can be subdivided into intracapsular fractures (those occurring within or proximal to the attachment of the hip joint capsule to the femur) and extracapsular (those occurring outside or distal to the hip joint capsule). Extracapsular hip fractures are defined as those fractures that occur within the area of bone bounded by the attachment of the hip joint capsule and extending down to a level which is five centimetres below the distal (lower) border of the lesser trochanter. Other terms used to describe these fractures include trochanteric, subtrochanteric, pertrochanteric and intertrochanteric fractures. These terms reflect the proximity of these fractures to the greater and lesser trochanters, which are two bony protuberances (bulges) at the upper end of the femur outside the joint capsule.
Hip fractures occur predominantly in older people (aged over 65 years), especially women. The incidence of hip fracture varies considerably between different populations (Bjorgul 2007). An incidence of 1024 per 100,000 for women over 50, and 452 per 100,000 for men over 50 was reported for Norway between 1998 and 2003 (Bjorgul 2007). The relative proportion of extracapsular fractures also varies: 39% of hip fractures were extracapsular fractures in Bjorgul 2007 and 48% in Karagas 1996. A summary of the casemix for 61,508 hip fractures occurring between 1 April 2012 and 31 March 2013 in 180 hospitals in England, Wales and Northern Ireland is presented by an annual report of the National Hip Fracture Database (NHFD 2013). This shows that around three‐quarters of hip fractures (73.2%) occurred in women, and over 90% of cases were aged over 70 years. In each of four years from 2009, around 40% of fractures were extracapsular.
Numerous subdivisions and classification methods exist for these fractures (e.g. the AO classification (Muller 1991)). The most practical classification, and that used for this review, is the basic division into four types: stable trochanteric fractures (AO classification type A1); unstable trochanteric fractures (AO classification type A2); fractures at the level of the lesser trochanter (transtrochanteric or AO classification type A3); and subtrochanteric fractures. Stable trochanteric fractures are two part fractures in which the fracture line runs obliquely (at an angle) between the lesser and greater trochanter of the femur. Unstable trochanteric fractures again have an oblique fracture line running between the trochanters, but in addition there is comminution (fragmentation) of the fracture site. The comminution fragments may be the lesser trochanter, greater trochanter or both trochanters. Transtrochanteric fractures, sited at the level of the lesser trochanter, have a slightly more distally located (lower) fracture line that either runs transversely (across the bone) at the level of the lesser trochanter or in an oblique direction that is opposite (or 'reverse') to that of the stable and unstable trochanteric fractures. Transtrochanteric fractures may be two part or comminuted. This fracture pattern is unstable as the femur is displaced medially (inwards) due to the pull of the adductor muscles. Subtrochanteric fractures are those fractures in which the fracture crossing the femur is predominately found within the five centimetres of bone immediately below the lesser trochanter. These fractures may be two part or comminuted, and in some instances the fracture may extend proximally into the trochanteric region or distally into the shaft of the femur.
Description of the intervention
Operative treatment of hip fractures was introduced in the 1950s using a variety of different implants. Implants may be either extramedullary or intramedullary in nature. The most commonly used extramedullary implant is the sliding hip screw (SHS), which is synonymous with the term compression hip screw and equivalent models such as the Dynamic, Richards or Ambi hip screws. Intramedullary nails used for the internal fixation of extracapsular fractures can either be inserted from proximal to distal (cephalocondylic nails) or from distal to proximal (condylocephalic nails).
Cephalocondylic nails are inserted through the greater trochanter of the femur and secured by a cross pin or screw, which is passed up the femoral neck into the femoral head. A number of different designs have been developed and marketed by different manufacturers. Examples include the Gamma nail (Stryker‐Howmedica), the intramedullary hip screw (Smith and Nephew Richards), the proximal femoral nail (Synthes) and the ACE trochanteric nail (DePuy Orthopaedics). Table 1 presents further information of the nails to date examined by the included trials in this review.
| Name | Description |
| Endovis nail | The Endovis nail (Citieffe Ltd) is available in 3 sizes (195 to 400 mm) and has a neck shaft angle of 130°. It has two cephalic screws for the femoral head to facilitate fracture compression. The distal section is slotted to produce a graduated variation of stiffness |
| Gamma nail | The Gamma nail (Stryker Ltd) was introduced in the late 1980s for the treatment of extracapsular hip fractures. The implant consists of a sliding lag screw which passes through a short intramedullary nail placed via the trochanteric entry point. One or two screws may be passed through the nail tip to secure it to the femoral shaft (distal locking). Theoretical advantages of this implant are due to a percutaneous insertion technique and include reduced blood loss, minimal soft tissue trauma and short operating time. Modifications to the design of the Gamma nail and its instrumentation have occurred since its introduction. The long Gamma nail has a range of different lengths from 280 to 460 mm with two distal locking screw options. An Asian‐Pacific version of the nail is available for use in the Asian population and has reduced length, diameter and mediolateral angle to accommodate small femurs typically seen in this group |
| Gamma 3 nail | The Gamma 3 nail (Stryker Ltd) is the third generation of the gamma nail fixation system for proximal femoral fractures. It is a trochanteric entry nail with a reduced proximal nail diameter (15.5 mm versus 17 mm) to facilitate a shorter incision. Its length options range from 280 mm to 460 mm. Its neck shaft angle options include 120°, 125° and 130°. The lag screw shape has also been modified to provide superior cutting behaviour and greater resistance to cut‐out. One trial in this review compared a sliding and a non‐sliding lag screw mechanism in the Gamma 3 nail |
| Gliding nail | The gliding nail (Smith‐Nephew) is a trochanteric entry nail designed to avoid the complications of implants such as the Gamma nail. It utilises a T‐shaped femoral neck blade that has an extensive surface area relative to other intramedullary devices which provides good rotational stability and a high resistance to fatigue fracture. The T‐shaped blade can slide through the nail, facilitating fracture compression and healing. The standard length is 220 mm with long nail options ranging from 340 mm to 440 mm. Blade shaft angles include 125° and 135° |
| Intramedullary hip screw (IMHS) | The IMHS (Smith and Nephew), length 210 mm, was introduced in 1991 for the treatment of extracapsular femoral fractures. Like the Gamma nail, it consists of a nail inserted via the greater trochanter into the medullary cavity. It utilises a single screw in the femoral head that can slide through a barrel in the nail allowing fracture compression. Three different neck angles are available, 125°, 130° and 135°. Nail lengths are available from 195 mm to 440 mm |
| Proximal femoral nail (PFN) | The PFN (Synthes Ltd), length 240 mm, was introduced in 1998 for the treatment of extracapsular fractures. Like the Gamma and IMHS, it consists of a nail inserted via the greater trochanter into the medullary cavity. Two lengths are available, 200 mm and 240 mm. Two proximal lag screws are passed up the femoral neck to the head. Distal locking can be performed in static or dynamic mode via two distal locking screws |
| Proximal femoral nail antirotation (PFNA) | The PFNA (Synthes Ltd), length 170 mm, 200 mm or 240 mm, is a modification of the PFN nail. It is similar to the PFN nail apart from not having two proximal lag screws, but instead a single helically‐shaped blade which is designed to provide increased angular and rotational stability. The helical blade is designed to avoid bone loss that occurs during drilling and insertion of a standard hip screw. It has 2 distal locking screw options for either dynamic or static locking. Blade shaft angle options include 125°, 130° and 135° |
| Proximal femoral nail antirotation II (PFNA II) | The PFNA II (Synthes Ltd) is a modification of the PFNA nail to address the different proximal femoral anatomy of Asian patients. The PFNA has a large proximal diameter (17 mm) which was thought to account for the increase in femoral shaft fracture, lateral cortex splitting and thigh pain reported in Asian patients. The PFNA II has a smaller proximal diameter (16.5 mm versus 17 mm) and a flatter lateral shape (5° versus 6°) |
| Targon PF (proximal femoral) nail | The Targon PF nail (B Braun Ltd), length 220 mm, is inserted into the intramedullary cavity via a trochanteric entry point. Proximally, this nail has a sliding lag screw and an antirotation pin. The Targon PF nail facilitates fracture dynamisation via a gliding screw that glides through a sleeve that is attached to the nail, thereby avoiding protrusion of the screw into peritrochanteric tissues |
| ACE trochanteric nail | The ACE nail (Depuy) has a 10.5 mm lag screw and an optional antirotation lag screw. It has 2 distal holes for static or dynamic locking. Its proximal diameter is 16 mm and length is 180 mm or 200 mm |
| Russell‐Taylor Recon nail | The Russel‐Taylor Recon nail (Smith‐Nephew) is an intramedullary nail that utilises a piriformis entry point. 2 screws are available for fixation in the femoral head. It is a full length femoral nail with no short versions available for proximal femoral fixation only |
| InterTan nail | The InterTan nail (Smith‐Nephew) uses 2 cephalocervical screws in an integrated mechanism allowing intra‐operative compression and rotational stability of the head‐neck fragments. It has a cannulated set screw mechanism that allows for the device to be used in fixed angle mode or in sliding/compression mode. Its length ranges from 180 mm to 460 mm (long nail option) |
Condylocephalic nails are inserted into the distal femur and passed up the intramedullary cavity across the fracture site and up into the femoral head. The best known and tested type of such nails is the Ender nail.
How the intervention might work
Successive updates of our Cochrane review (Parker 2010) comparing the Gamma and other cephalocondylic intramedullary nails with extramedullary implants for extracapsular hip fractures have consistently found that cephalocondylic nails incur the complications of intra‐operative fracture, and later fracture around the implant. Based primarily on the higher rate of complications and re‐operations of these nails for trochanteric fractures, we suggested that the SHS (an extramedullary implant) appears to be the better device for these fractures. We also suggested that "Further studies are required to determine if different types of intramedullary nail produce similar results, or if intramedullary nails have advantages for selected fracture types (for example, subtrochanteric fractures)."
Our Cochrane review of randomised trials comparing condylocephalic nails with extramedullary fixation (Parker 1998) concluded that the use of condylocephalic nails could not be recommended because of the markedly increased risk of fracture‐healing complications and other problems associated with condylocephalic nails (in particular Ender nails).
Why it is important to do this review
Despite the evidence of poor performance of intramedullary nails in comparison with the SHS (an extramedullary implant), developments and modifications to intramedullary nails, especially cephalocondylic nails, continue. Additionally, the use of these nails is increasing (Anglen 2008). This systematic review of randomised trials examines studies that have compared different types, or modifications to the design, of intramedullary nails for extracapsular proximal femoral fractures. This is an update of a Cochrane review first published in Parker 2005 and last updated in Parker 2008.
Objectives
To assess the effects (benefits and harms) of different designs of intramedullary nails for treating extracapsular hip fractures in adults.
Methods
Criteria for considering studies for this review
Types of studies
All randomised or quasi‐randomised (for example, alternation) controlled trials comparing different types of intramedullary nails.
Types of participants
Skeletally mature patients with an extracapsular proximal femoral fracture. Given that one of the authors (MJP) has become aware of the growing use of intramedullary nails in intracapsular fractures, we note here that in a future update we will consider including trials with a mixed population of intracapsular and extracapsular proximal femoral fractures. We will, however, request separate data for the two fracture types.
Types of interventions
Surgical fixation of the fracture with either a cephalocondylic intramedullary nail (for example, the Gamma nail, the intramedullary hip screw (IMHS) and the proximal femoral nail (PFN)) or a condylocephalic nail (for example, the Ender nail). In setting out our comparisons we generally selected the older, more conventional, or static, or both, implant design as our control group. Before we undertake our next update, we will consider setting up comparisons addressing more general design concepts, such as long versus short nails.
Types of outcome measures
The primary focus is on long‐term functional outcome, preferably measured at one year or more.
Primary outcomes
-
Functional outcomes: preferably, validated patient‐reported measures of lower limb or hip function (e.g. Oxford hip score; Western Ontario and McMaster Universities Arthritis Index (WOMAC)) and activities of daily living and health related quality of life scores (SF‐36). Composite scores of subjectively and objectively rated function and overall outcome (e.g. Harris hip score, Merle D'Aubigne hip score).
-
“Poor outcome”, defined as death or deterioration of functional status leading to markedly increased dependency in the community or admission to institutional care.
-
Serious adverse events and technical complications of fixation (e.g. deep infection, avascular necrosis, later fracture of the femur, non‐union, cut‐out, implant breakage) for which substantive treatment, such as revision surgery, is indicated or performed.
Secondary outcomes
-
Mobility, use of walking aids, presence of a limp.
-
Hip, lower limb pain (chronic).
-
Medical complications: pneumonia; thromboembolism (symptomatic deep vein thrombosis or pulmonary embolism); pressure sore; urinary tract infection; delirium.
-
Less serious local complications: intra‐operative periprosthetic fracture; surgical site infection (superficial); wound haematoma; minor operation for removal of hardware.
Other outcomes
Data for the following outcomes were collected for completeness, but not presented as main results for this review.
-
Operative details: length of surgery, operative blood loss, number of patients transfused, radiographic screening time.
-
Functional impairment: range of motion, muscle strength.
-
Anatomical restoration: leg shortening (preferably > 2 cm), varus deformity of the femoral neck, external rotation deformity (preferably > 20 degrees).
Economic outcomes
Each trial report was reviewed for costs and resource data, such as length of hospital stay and number of outpatient attendances, that would enable economic evaluation.
Search methods for identification of studies
Electronic searches
We searched the Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (6 January 2014), the Cochrane Central Register of Controlled Trials (The Cochrane Library Issue 12, 2013), MEDLINE (1966 to November Week 3, 2013), MEDLINE In‐Process & Other Non‐Indexed Citations (3 January 2014) and EMBASE (1988 to 2014, Week 1). We searched the WHO International Clinical Trials Registry Platform (accessed January 2014) for ongoing and recently completed trials. We did not apply any language restrictions.
Previously, we searched the UK National Research Register (accessed June 2007, now archived) and Current Controlled Trials (accessed June 2007) for ongoing and recently completed trials.
Search strategies developed for The Cochrane Library (2007 onwards), MEDLINE (2007 onwards) and EMBASE (2007 onwards) are shown in Appendix 1. The subject specific MEDLINE search was combined with the sensitivity‐maximizing version of the Cochrane Highly Sensitive Search Strategy for identifying randomised trials (Lefebvre 2011). Previous search strategies can be found in Appendix 2.
Searching other resources
We searched reference lists of articles and our own reference databases.
In our previous update, we included the findings from handsearches of the British Volume of the Journal of Bone and Joint Surgery supplements (1996 onwards) and abstracts of the OTA (Orthopaedic Trauma Association) annual meetings (1996 to 2006) and AAOS (American Academy of Orthopaedic Surgeons) annual meetings (2004 to 2007). We also included handsearch results from the final programmes of SICOT (Société Internationale de Chirurgie Orthopédique et de Traumatologie) (1996 and 1999) and SICOT/SIROT (Société Internationale de Recherche en Orthopédie et Traumatologie) (2003), EFORT (European Federation of National Associations of Orthopaedics and Traumatology) (2007) and the BOA (British Orthopaedic Association) Congress (2000, 2001, 2002, 2003, 2005 and 2006). We scrutinised weekly downloads of "Fracture" articles in new issues of 15 journals (Acta Orthopaedica Scandinavica; American Journal of Orthopedics; Archives of Orthopedic and Trauma Surgery; Clinical Journal of Sport Medicine; Clinical Orthopaedics; Foot and Ankle International; Injury; Journal of the American Academy of Orthopaedic Surgeons; Journal of Arthroplasty; Journal of Bone and Joint Surgery American volume; Journal of Bone and Joint Surgery British volume; Journal of Foot and Ankle Surgery; Journal of Orthopaedic Trauma; Journal of Trauma; Orthopedics) from AMEDEO.
Data collection and analysis
Selection of studies
Three review authors (JQ, EH and HH) independently screened search results and, after obtaining full reports, assessed potentially eligible trials for inclusion. The other review author (MJP) provided feedback on selection and notification of results from his ongoing scrutiny of the hip fracture literature. We did not mask the titles of journals, names of authors or supporting institutions at any stage.
Data extraction and management
Using a data extraction form, two review authors (JQ and EH) independently extracted data for the outcomes listed above and resolved any differences by discussion. Data entry by these two review authors was checked by a third author (HH). We contacted all trialists for additional data and clarification when necessary.
Assessment of risk of bias in included studies
At least two review authors (always JQ and EH) independently assessed risk of bias for all trials without masking of the source and authorship of the trial reports, including those that had been assessed in previous versions of the review. We piloted the assessment form on one trial. JQ checked between rater consistency in assessment at data entry, and this was subsequently checked by HH. We resolved all differences by discussion. We used the tool outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). This tool incorporates assessment of randomisation (sequence generation and allocation concealment), blinding (of participants and treatment providers, and outcome assessment), completeness of outcome data, selection of outcomes reported, and other sources of bias. We considered 'subjective' outcomes (e.g. functional outcome scores, pain) and 'objective' outcomes (mortality, complications) separately in our assessment of blinding (outcome assessment) and short‐term (in hospital; up to four months) and longer‐term (four months and above; post‐hospital discharge) outcomes for completeness of outcome data. We assessed three additional sources of bias: bias resulting from major imbalances in key baseline characteristics (e.g. age, gender, type of fracture, prior mobility); performance bias, particularly 'differential expertise' bias resulting from lack of comparability in surgeon's expertise with the devices under test; and bias relating to a commercial conflict of interest.
Measures of treatment effect
For each study, we calculated risk ratios (RRs) with 95% confidence intervals (CIs) for dichotomous outcomes, and mean differences (MDs) with 95% CIs for continuous outcomes.
Unit of analysis issues
The unit of analysis was individual patients in these trials but we remained alert to other potential unit of analyses issues, such as the repeated observation from more than one time‐point, and multiple observations for the same outcome (e.g. total adverse events).
Dealing with missing data
We contacted trial authors for missing data and information. Where reported, we used the numbers of participants reported as providing data for any particular outcome. In studies for which a number of events were reported, but the denominator was unclear, we used numbers randomised or alive at follow‐up. We did not impute missing standard deviations (SDs) but derived these from standard errors, 95% CIs or exact P values, if these were presented instead.
Assessment of heterogeneity
We used both the Chi² statistic and I² test (Higgins 2003), as well as visual inspection, to determine whether heterogeneity was present and whether data pooling was appropriate.
Assessment of reporting biases
Should data for meta‐analyses be available for 10 or more trials in a future update, we will consider the generation of funnel plots to explore the potential for publication bias.
Data synthesis
Where appropriate, we pooled results of comparable groups of trials using the fixed‐effect model. We would have used the random‐effects model to compare the results where there was substantial and unexplained heterogeneity.
Subgroup analysis and investigation of heterogeneity
No subgroup analyses were specified a priori in the protocol. For this update, we prespecified two subgroups (type of fracture ‐ initially, intertrochanteric versus subtrochanteric ‐ and gender) and indicated that we would test whether the subgroups were statistically significantly different from one another by inspecting the overlap of CIs and performing the test for subgroup differences available in Review Manager (RevMan 2014). However, we found there were insufficient data to conduct either subgroup analysis.
Sensitivity analysis
There were insufficient data to conduct our planned exploratory sensitivity analyses based on allocation concealment and on the reporting of surgical experience. Sensitivity analyses using numbers randomised were done for any outcome for which denominators other than number randomised had been used, in order to assess any impact of missing data on results.
Quality assessment
We used the GRADE approach to assess the quality of evidence relating to the primary outcomes for the individual comparisons (Schünemann 2011).
Results
Description of studies
Results of the search
For this update (search completed January 2014), we screened a total of 852 records from the following databases: Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (20 records), Cochrane Central Register of Controlled Trials (276), MEDLINE (255) and EMBASE (301). We also screened further records for 76 trials from a search of the WHO International Clinical Trials Registry Platform conducted in January 2014, and obtained references for two abstracts in Journal of Bone and Joint Surgery (British Volume) supplements from one review author (MJP). The results from the previous searches (up to June 2007) are shown in Appendix 3. Upon assessment, we excluded Gahr 2003, which had been in 'Studies awaiting classification', because a full trial report has not been forthcoming of the currently inadequately reported quasi‐randomised trial.
The search update resulted in the identification of 22 potentially eligible studies, which consisted of 17 full reports, three trial reports and two conference abstracts. Of the 17 fully reported trials, we included eight new trials (De Grave 2012; Makridis 2010; Okcu 2013; Vaquero 2012; Wild 2010; Xu 2010a; Zhang 2013; Zhu 2012), excluded seven others (Cao 2009; Dall'Oca 2010; Huang 2012; Ouyang 2010; Pan 2009; Xu 2010b; Yang 2011) and left two in 'Studies awaiting classification' pending further information (Park 2010; Stern 2011). Of the three potentially eligible studies identified from our search of the WHO International Clinical Trials Registry Platform, one trial (NCT00736684) was identified as having been published (Vaquero 2012), one is excluded (NTR1133) and one is ongoing (NCT01437176). One conference abstract also awaits classification (Mora 2011). (The other conference abstract was another report of Makridis 2010).
Figure 1 presents the study flow diagram for this review update. Overall, there are now 17 included trials (De Grave 2012; Efstathopoulos 2007; Fritz 1999; Hardy 2003; Herrera 2002; Makridis 2010; Marques 2005; Okcu 2013; Papasimos 2005; Schipper 2004; Starr 2006; Vaquero 2012; Vidyadhara 2007; Wild 2010; Xu 2010a; Zhang 2013; Zhu 2012), 12 excluded studies (Cao 2009; Dall'Oca 2010; Gahr 2003; Huang 2012; Merenyi 1995; NTR1133; Ouyang 2010; Pan 2009; Suckel 2006; Wagner 1998; Xu 2010b; Yang 2011), one ongoing trial (NCT01437176) and three studies awaiting classification (Mora 2011; Park 2010; Stern 2011).
Study flow diagram
Included studies
Sixteen included trials were reported in full in English language journals; a full translation from Spanish being obtained for Marques 2005. Thirteen were single centre trials based in seven different countries (Belgium: De Grave 2012; Hardy 2003; China: Xu 2010a; Zhu 2012; Germany: Fritz 1999; Wild 2010; Greece: Efstathopoulos 2007; Makridis 2010; Papasimos 2005; India: Vidyadhara 2007; Spain: Herrera 2002; Marques 2005; USA: Starr 2006). Four were multi‐centre trials: Okcu 2013 and Zhang 2013 were two‐centre trials based respectively in Turkey and China; Schipper 2004 was a multi‐centre trial based in The Netherlands and Vaquero 2012 was a multi‐centre trial carried out in Spain. Papasimos 2005 tested three implants: those patients allocated to the sliding hip screw (SHS) are included in a separate Cochrane review (Parker 2010). Fifteen trials had predominantly older populations, with mean ages ranging between 69 and 85 years. The two exceptions were Starr 2006, which only included adults under 50 years of age with high‐energy fractures and Zhu 2012 which included adults under 60 years of age. Ten trials (Fritz 1999; Hardy 2003; Marques 2005; Schipper 2004; Okcu 2013; Papasimos 2005; Vaquero 2012; Vidyadhara 2007; Xu 2010a; Zhang 2013) included only patients with unstable trochanteric proximal femoral fractures, whereas a minority of patients in the other seven trials (De Grave 2012; Efstathopoulos 2007; Herrera 2002; Makridis 2010; Starr 2006; Wild 2010; Zhu 2012) had stable fractures. Further details of the 17 included studies are given in the Characteristics of included studies. The trials tested 12 different comparisons between various cephalocondylic nail designs (the nails are described in Table 1). There were no trials evaluating condylocephalic nails.
Proximal femoral nail (PFN) versus Gamma nail
The PFN was compared with the standard Gamma nail in 250 participants in Herrera 2002, in 156 participants in Marques 2005, in 80 participants in Papasimos 2005, and in 424 participants in Schipper 2004.
ACE trochanteric nail versus Gamma nail
The ACE trochanteric nail was compared with the trochanteric Gamma nail in 112 participants in Efstathopoulos 2007, and with the Gamma AP nail in 73 participants in Vidyadhara 2007.
ACE trochanteric nail versus Gamma 3 nail
The ACE trochanteric nail was compared with the Gamma 3 nail in 112 participants in De Grave 2012.
Gliding nail versus Gamma nail
One trial (Fritz 1999) involving 80 participants compared the gliding nail (where the lag screw of a Gamma nail is replaced with a double T‐shaped blade) with the Gamma nail.
ENDOVIS nail versus intramedullary hip screw (IMHS)
One trial (Makridis 2010) involving 215 participants compared the ENDOVIS nail (contains two holes for cephalic screw insertion) and the IMHS nail.
Russell‐Taylor recon nail versus long Gamma nail
One trial (Starr 2006) involving 34 participants compared the Russell‐Taylor Recon nail with the long Gamma nail.
Proximal femoral nail antirotation (PFNA) versus Targon PF nail
One trial (Wild 2010) compared the PFNA (intramedullary device with a helical blade instead of a screw) with the Targon PF nail (a device that has an extra antirotation pin in the femoral neck) in 80 participants.
PFNA versus Gamma 3 nail
The PFNA nail was compared with the Gamma 3 nail in 136 participants in Xu 2010a and 64 participants in Vaquero 2012.
Dynamic versus static locked intramedullary hip screw (IMHS)
One trial (Hardy 2003) involving 81 participants compared a modified intramedullary hip screw (IMHS) featuring a single slotted hole that allowed dynamic locking of the nail versus the standard IMHS, which is locked distally with two screws.
Sliding versus non‐sliding lag screw Gamma 3 nail
One trial (Zhu 2012) compared a sliding versus a non‐sliding lag screw in 80 participants, all of whom were under 60 years of age, who were treated with a Gamma 3 nail.
InterTan nail versus PFNA II
One trial (Zhang 2013) involving 113 participants compared the InterTan nail with the PFNA II nail
Long versus standard PFNA nail
One trial (Okcu 2013) compared long versus standard (short) PFNA nails in 40 participants with reverse oblique type proximal femoral fractures.
Excluded studies
We excluded 12 studies for the reasons given in the Characteristics of excluded studies. Eight trials were either not randomised or very unlikely to be randomised trials. We excluded Xu 2010b because the population overlapped to an unknown extent with that of Xu 2010a. We excluded Yang 2011 because of concerns raised from similarities of this report with that of an earlier trial (Makridis 2010). We excluded NTR1133 because it seems very unlikely that this trial, if started, will be published. We excluded Dall'Oca 2010 as it did not compare two different nails but the same nail (Gamma nail), used with and without cement augmentation.
Risk of bias in included studies
The risk of bias assessments for the individual trials are shown in Figure 2 and as a composite for all trials in Figure 3.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Allocation
All three quasi‐randomised trials were deemed at high risk of selection bias in terms of sequence generation (allocation was based on odd and even record numbers in Herrera 2002 and Marques 2005; and on admission sequence in Wild 2010) and lack of allocation concealment.
Eight trials (De Grave 2012; Efstathopoulos 2007; Fritz 1999; Hardy 2003; Makridis 2010; Papasimos 2005; Starr 2006; Zhu 2012) did not specify their method of random sequence generation and were judged at unclear risk of bias for this domain. The remaining five trials (Okcu 2013; Vaquero 2012; Vidyadhara 2007; Xu 2010a; Zhang 2013) used a computer‐generated random numbers table or list and were judged at low risk of bias.
Efstathopoulos 2007, Makridis 2010, Starr 2006, Vaquero 2012, Xu 2010a, Zhang 2013 and Zhu 2012 used sealed envelopes. These were reported to be numbered in Starr 2006, Xu 2010a and Zhang 2013. We judged these seven studies and the six studies not providing details of their method of randomisation (De Grave 2012; Fritz 1999; Hardy 2003; Okcu 2013; Papasimos 2005; Vidyadhara 2007) to be at unclear risk of bias for allocation concealment.
We only considered Schipper 2004, which used computer‐generated randomisation (stratified by participating centre and balanced in blocks of four and six patients), and numbered and blinded envelopes, to be at low risk of bias for both domains.
Blinding
Blinding of the participants did not appear to have occurred in any of the trials. The surgeons could not be blinded. We judged all trials to be at unclear risk of performance bias relating to blinding.
Complete assessor blinding does not appear to have occurred in any of the trials. Okcu 2013 had a blinded assessor for the mobility and Harris hip scores and so we judged this as unclear risk of bias for 'subjective' outcomes. We also judged Efstathopoulos 2007 (which did not report subjective outcomes) and Hardy 2003 (which included some cross‐checking of results) at unclear risk of bias for subjective outcomes, whereas we judged the rest to be at high risk of detection bias for these outcomes. The final classification of complications was done using anonymised date in Vaquero 2012. We judged all studies to be at unclear risk of detection bias for objective outcomes.
Incomplete outcome data
We judged 10 trials to be at low risk of attrition bias for short‐term outcome assessment; six were deemed to be at 'unclear risk', either because of lack of information on early losses (De Grave 2012; Herrera 2002; Marques 2005; Papasimos 2005) or because data were provided for survivors only (Okcu 2013) or people without complications (Zhu 2012). We judged Vaquero 2012 to be at high risk of short‐term bias because of a 33% loss at three months, with only a quarter of these attributable to mortality.
We judged four trials (Fritz 1999; Starr 2006; Zhang 2013; Zhu 2012) to be at low risk of attrition bias for long‐term outcome assessment; and four trials (Papasimos 2005; Schipper 2004; Vaquero 2012; Xu 2010a), at high risk. Of special note is that in Schipper 2004, follow‐up was discontinued at four months for participants with complete radiological consolidation; this considerably reduced the number of participants available at one year follow‐up. For Vaquero 2012, there was no explanation provided for the high attrition rate of 61% at long‐term follow‐up. We deemed the remaining nine trials to be at unclear risk of bias.
Selective reporting
Protocols were not available for any of these trials, which were judged to be at unclear risk of bias with the exception of Herrera 2002 and Marques 2005. We judged these two trials to be at high risk because several outcome measures mentioned in the methods sections were not presented in their results. Zhu 2012 excluded three cases with complications from the final analysis for unspecified reasons.
Other potential sources of bias
Baseline characteristics
The baseline characteristics of the intervention groups were well matched in the majority of studies. We judged three trials to be at unclear risk of bias relating to this item: data for age and gender were not presented in Herrera 2002; group allocation and baseline data were missing for 21 participants not available at follow‐up in Papasimos 2005 and baseline mobility status was not documented in Wild 2010.
Performance bias relating to surgeon expertise and care programmes
Eight trials were judged at low risk of performance bias (De Grave 2012; Efstathopoulos 2007; Okcu 2013; Starr 2006; Vidyadhara 2007; Xu 2010a; Zhang 2013; Zhu 2012) as the surgeons were reported as being experienced in the interventions under test and care programmes were comparable in the two groups. The remaining trials were all judged as being at unclear risk of bias. Of particular note is that both Marques 2005 and Hardy 2003 had a higher number of junior surgeons performing the surgery in one group and Papasimos 2005 reported that the four surgeons involved had less experience with the PFN. There was a lack of information to judge whether post‐operative care was the same for both groups in Fritz 1999 and Wild 2010.
Funding source or conflict of interest
We judged the nine trials that explicitly indicated there was no conflict of interest to be at low risk of bias and that six provided no information at unclear risk. Schipper 2004 and Vaquero 2012 were judged to be at high risk reflecting the financial support received from the manufacturer of one of the implants in their respective trials.
Effects of interventions
In the following we have presented the outcomes in five categories, starting with 'Final outcome measures'. Appendix 4 shows the relationship between the outcomes listed in Types of outcome measures and these categories.
Proximal femoral nail (PFN) versus Gamma nail
Four trials (Herrera 2002; Marques 2005; Papasimos 2005; Schipper 2004) compared the PFN with the standard Gamma nail in a total of 910 participants. All participants had unstable fractures other than 32 participants with stable fractures in Herrera 2002. Aside from Herrera 2002, which included 13 patients with "neoplasia", pathological fractures were excluded.
Final outcome measures
There was no statistically significant difference between the two groups in mortality at 12 months for the three trials that provided data (86/415 versus 80/415; risk ratio (RR) 1.08; 95% confidence interval (CI) 0.82 to 1.41; seeAnalysis 1.1). Papasimos 2005 excluded from their analyses the data from the 10 people who had died by one year follow up.
The few functional outcome data that could be presented are shown in Analysis 1.2 and Analysis 1.3. Data from Herrera 2002 showed no statistically significant difference between the two groups in the failure to recover pre‐fracture walking ability (RR 1.03; 95% CI 0.80 to 1.33). Pain in the thigh at follow‐up was reported as being statistically significantly less in the PFN group in Marques 2005 (4.7% versus 27.3%; reported P = 0.004) but this difference was not apparent when the actual numbers of participants with pain were obtained from the trialist (seeAnalysis 1.2). Marques 2005 reported there was no statistically significant difference between the two groups in the final independent mobility scores. Papasimos 2005 reported there was no significant difference between the two groups in the return to pre‐fracture level of ambulation and independence. The Harris hip scores at four weeks, six months and one year reported in Schipper 2004 showed no statistically significant difference between groups (seeAnalysis 1.3); however, this was for a subgroup of patients at each time point and thus may not be representative of the outcome for the population of survivors.
Fracture fixation complications
Operative details as presented by each study are summarised in Analysis 1.4. None of the differences between the two groups in the various aspects and intra‐operative complications of fracture fixation was statistically significant other than an increased risk of greater trochanteric fractures, or intra‐operative comminution of the fracture around the trochanteric region, for those treated with the Gamma nail (seeAnalysis 1.4.6, 6/165 versus 20/165; RR 0.30, 95% CI 0.12 to 0.73). Herrera 2002 did not reveal the surgical consequences of these fractures; both cases in Papasimos 2005 were treated conservatively. The difference between the two groups in the more important outcome of operative fracture of the femur was not statistically significant (seeAnalysis 1.4.7, 1/455 versus 5/455; RR 0.33, 95% CI 0.07 to 1.63). Marques 2005 attributed all three intra‐operative femoral fractures to bad surgical technique. The operative fracture in the Gamma nail group of Papasimos 2005 was managed conservatively. In Schipper 2004, both operative fractures of the femur, featuring a subtrochanteric extension, occurred in the Gamma nail group.
Analysis 1.5 presents the fracture healing complications as reported by each study. None of the differences in outcomes between the two implants in the pooled data from three or four studies was statistically significant. The tendency to a higher rate of secondary varus, reflecting a loss of reduction, in Herrera 2002 was stated as not being linked with subsequent clinical problems. The most common fracture healing complication was cut‐out of the implant (17/455 versus 24/455; RR 0.71, 95% CI 0.39 to 1.30). It should be noted that in Schipper 2004, participants whose fractures were judged to be healed at four months had no further radiological follow‐up. There was no significant difference in the incidence of local complications, which included cut‐out, infection, haematoma, migration of hip screws, malrotation, shaft fracture and nail fatigue, at four months in Schipper 2004 (45/211 versus 47/213); and similarly at 12 months (51/211 versus 50/213).
Pooled data from all four trials for re‐operation showed no statistically significant difference between the two groups (seeAnalysis 1.6, 45/455 versus 36/455; RR 1.25, 95% CI 0.83 to 1.90). There were no significant differences between the two implants in any of the reported wound complications (seeAnalysis 1.7).
Post‐operative complications and hospital stay
None of the differences between the two implant groups in specific post‐operative complications were statistically significant in Herrera 2002, Marques 2005 or Papasimos 2005 (seeAnalysis 1.8). Schipper 2004 reported no difference between groups in medical complications that had occurred by one year follow up.
Herrera 2002 reported that trial participants remained in hospital for an average of 14.1 days. The mean length of stay in hospital was nearly a week longer in Schipper 2004; there being no significant difference between the two groups (seeAnalysis 1.9). Similarly, the difference between the two groups in the mean hospital stays were reported to be not statistically significant for both Marques 2005 (11.1 days for the PFN group versus 12.2 days for the Gamma nail group) and Papasimos 2005 (8.6 days versus 8.8 days).
Anatomical restoration
These outcomes were not reported in any of the trials. However, two participants of the PFN group and one of the Gamma nail group had re‐operations for "rotational defect of the leg" in Herrera 2002.
Operative details
The mean length of surgery of the PFN group reported as being significantly shorter in Herrera 2002 (49 versus 68 minutes), whereas it was reported to be significantly longer in Papasimos 2005 (71 versus 51 minutes). The mean length of surgery was 60 minutes in both groups of Schipper 2004. Marques 2005 reported the difference between the two groups in the median length of surgery (45 versus 40 minutes) was not statistically significant.
Schipper 2004 found intra‐operative blood loss was statistically significantly lower in the PFN group (mean difference (MD) ‐67.00 mL, 95% CI ‐111.40 to ‐22.60 mL: seeAnalysis 1.10). Papasimos 2005 found the difference between the two groups in mean operative blood loss (265 mL versus 250 mL) was not statistically significant. Though significantly more participants in the PFN group of Herrera 2002 received blood transfusion, the converse was true for Marques 2005 (seeAnalysis 1.11). These results were not pooled since visual inspection of the transfusion results from the two trials shows substantial heterogeneity (I² = 88.9% when pooled).
Neither trial reporting radiographic screening time found a statistically significant difference in this outcome between the two groups: for Marques 2005, the median times were 100 versus 120 seconds; for Papasimos 2005, the mean times were 0.26 minutes in both groups.
ACE trochanteric nail versus Gamma nail
Two trials (Efstathopoulos 2007; Vidyadhara 2007) made this comparison although with some variations in the intervention. Efstathopoulos 2007 compared the ACE nail used with one proximal screw versus the trochanteric Gamma nail in 112 people, 82% of whom had unstable fractures. In Vidyadhara 2007, the ACE nail had two proximal screws and was compared with the AP (Asian/pacific) Gamma nail in 73 people with unstable fractures. Neither study included subtrochanteric fractures.
Final outcome measures
There was no difference between the two groups in mortality (seeAnalysis 2.1). Vidyadhara 2007 found no statistically significant differences between the two groups in hip pain at one month after injury, or the presence of a limp or difficulty in squatting at two years (seeAnalysis 2.2). Efstathopoulos 2007 found no difference between the two groups in mobility scores at follow‐up (seeAnalysis 2.3). Although the Harris hip scores at four months, one year and two years were significantly different in the two groups of Vidyadhara 2007, the very small differences were clinically insignificant (seeAnalysis 2.4).
Fracture fixation complications
There were no fracture healing complications reported in Efstathopoulos 2007 and only one in Vidyadhara 2007 (seeAnalysis 2.5). This was a cut‐out in the Gamma nail group that was treated by removal of the implant followed by bed rest for three months. Vidyadhara 2007 reported no wound infection whilst Efstathopoulos 2007 reported four cases of superficial wound infection in the ACE nail group versus three in the Gamma nail group (seeAnalysis 2.6).
Post‐operative complications and hospital stay
There were no significant differences between the two groups in the limited data provided for post‐operative complications (seeAnalysis 2.7). Efstathopoulos 2007 reported similar mean lengths of hospital stay for the two groups (7.2 versus 7.0 days; reported as not significant).
Anatomical restoration
Three people had limb shortening in Vidyadhara 2007, with no significant differences between the two groups (seeAnalysis 2.8).
Operative details
Efstathopoulos 2007 reported no difference in the mean length of surgery between groups (seeAnalysis 2.9). Vidyadhara 2007 reported a higher median length of surgery for the ACE nail (43 versus 32 minutes). The statistically significantly greater blood loss found in Vidyadhara 2007 for the ACE nail is clinically minor (seeAnalysis 2.9: mean difference 13 mL; 95% CI 6.78 to 19.22). Efstathopoulos 2007 found no statistically significant differences in the units of blood transfused, number of patients transfused or the radiographic screening time (seeAnalysis 2.9 and Analysis 2.10).
ACE trochanteric nail versus Gamma 3 nail
One trial (De Grave 2012), with 112 participants with stable or unstable trochanteric fractures, compared the ACE trochanteric nail versus the Gamma 3 nail. The numbers of participants available at each follow‐up was not provided.
Final outcome measures
De Grave 2012 reported no statistically significant differences between the two groups at fracture consolidation (between three and 12 months post‐operatively) in the mean Merle d'Aubigne scores for pain, walking function, mobility or overall (0 to 18: best outcome) between the two groups (14.12 (SD 2.95) versus 14.19 (SD 2.86); reported P = 0.92). Walking ability was restored in 83% of participants in the ACE group and in 80% of the Gamma 3 group. There was no significant difference between the two groups in mortality at one year (12/51 versus 14/61; RR 1.03, 95% CI 0.52 to 2.01; seeAnalysis 3.1).
Fracture fixation complications
There were no statistically significant differences between the two groups for these outcomes which included fixation failure, non‐union, wound infection and re‐operation (seeAnalysis 3.2). There were no intra‐operative complications. Two people in each group underwent another operative procedure, each receiving a hip arthroplasty, because of either cut‐out or secondary displacement of their fracture. There was no wound infection or non‐union.
Post‐operative complications and hospital stay
No trial participant had a deep vein thrombosis. The only recorded complication was a peripheral nerve injury that resulted in foot drop in one participant of the Gamma 3 nail group (seeAnalysis 3.3).
Anatomical restoration
These outcomes were not reported.
Operative details
De Grave 2012 reported that the mean operative time was 51 minutes in the ACE group and 41 minutes in the Gamma 3 nail group.
Gliding nail versus Gamma nail
Fritz 1999 compared the gliding nail (a modification of the Gamma nail) with a standard Gamma nail. There were 40 participants, all with an unstable trochanteric fracture, in each group.
Final outcome measures
No statistically significant difference was found between the two groups for mortality (seeAnalysis 4.1), for residence of survivors in a geriatric institution (seeAnalysis 4.2) or overall unfavourable outcome, defined as residence in a geriatric institution or dead, at six months. Fritz 1999 reported there were no statistically significant differences between the two groups in the Merle d'Aubigne scores for pain, walking function, mobility or overall.
Fracture fixation complications
There were no statistically significant differences between the two groups for these outcomes (seeAnalysis 4.3). There was one intra‐operative complication (a minor shaft fracture) in the gliding nail group and seven intra‐operative complications (six were due to failed placement of the second locking screw) in the Gamma nail group. One woman in the gliding nail group fell during mobilisation, fracturing her femur shaft. Cut‐out of the implant occurred in three cases in the standard nail group. Re‐operations (three versus four) resulted from these two complications, as well as from wound infection and a haematoma.
Post‐operative complications and hospital stay
There were no statistically significant differences between the two groups in those with any post‐operative medical complication or for specific complications as presented in Analysis 4.4. Fritz 1999 reported there was no statistically significant difference between the two groups in the mean hospital stay (9.2 versus 10.4 days).
Anatomical restoration
There were no statistically significant differences between the two groups in those with leg shortening or rotational deformity (seeAnalysis 4.5).
Operative details
Fritz 1999 reported there were no statistically significant differences between the two groups for length of surgery (mean duration: 63 versus 62 minutes) or operative blood loss (mean loss: 338 mL versus 296 mL).
ENDOVIS nail versus intramedullary hip screw (IMHS)
Makridis 2010 compared the Endovis nail versus the IMHS nail in 215 participants with either stable or unstable pertrochanteric fractures.
Final outcome measures
More participants in the ENDOVIS nail group were bedridden, and thus unable to walk, after their operation (29/105 versus 18/110; RR 1.69, 95% CI 1.00 to 2.85; seeAnalysis 5.1). The post‐operative Parker‐Palmer mobility scores also reflected poorer mobility after ENDOVIS nails (mean scores 4.7 versus 6.4; 9 = fully mobile); the difference between the two groups was reported to be statistically significant (P < 0.05). There was no difference between the two groups in mortality, either in hospital (3/105 versus 2/110; RR 1.57, 95% CI 0.27 to 9.22) or at one year (16/105 versus 15/110; RR 0.99; 95% CI 0.53 to 1.85); seeAnalysis 5.2.
Fracture fixation complications
The majority of these complications occurred in the ENDOVIS group. None of the differences between the two groups for the more serious complications were statistically significant (seeAnalysis 5.3); these include cut‐out (3/105 versus 1/110), later femoral shaft fracture (0/105 versus 1/110), intra‐operative femoral shaft fracture (0/105 versus 1/110); Z‐phenomena (2/105 versus 0/110) and nail breakage (1/105 versus 1/110). The operative fracture was successfully treated with circular wires. Overall, there were five re‐operations in the ENDOVIS group (three for cut‐out, one for a Z‐phenomenon and one for a reverse Z‐phenomenon) versus two re‐operations in the IMHS group (one for cut‐out, one for periprosthetic femoral shaft fracture): RR 2.62, 95% CI 0.52 to 13.21. Similar rates (two in each group) of infection were seen in both groups; all were superficial wound infections successfully treated by intravenous antibiotics.
Post‐operative complications and hospital stay
Post‐operative medical complications were not assessed.
Anatomical restoration
Anatomical restoration outcomes were not assessed. However, shortening was reported for five patients in the ENDOVIS nail group who had medial displacement of the femur shaft.
Operative details
The mean operative times were similar in both groups (24.8 minutes versus 25.4 minutes). There were no significant differences reported between the groups regarding blood loss, haemoglobin levels or numbers of patients receiving transfusion (28/105 versus 29/105; (RR 1.01, 95% CI 0.65 to 1.58; seeAnalysis 5.4).
Russell‐Taylor Recon nail versus long Gamma nail
Starr 2006 compared the Russell‐Taylor Recon nail with the long Gamma nail in 34 people, aged between 19 and 50 years, with high energy extracapsular hip fracture. Five trial participants had stable and 21 unstable trochanteric fractures, and the other eight had subtrochanteric fractures. Three patients had open fractures and 17 had concurrent surgery for other injuries.
Final outcome measures
No deaths occurred within the one year follow‐up period. Starr 2006 found no statistically significant differences between the two groups in the numbers of participants who were unable to walk independently or unable to return to the same work (seeAnalysis 6.1). The person who was unable to walk had sustained a spinal cord injury at the time of her initial trauma. Similarly, there were no statistically significant differences between the two groups in the range of hip or knee movements. The Harris hip scores were similar for the two groups (mean scores: 86 versus 84; reported P = 0.60).
Fracture fixation complications
There were no fracture fixation complications reported (seeAnalysis 6.2). One patient in the long Gamma nail group had wound debridement for sepsis and a further 12 patients had elective removal of their implants for persistent pain (seeAnalysis 6.2).
Post‐operative complications and hospital stay
These outcomes were not reported in Starr 2006.
Anatomical restoration
These outcomes were not reported in Starr 2006.
Operative details
Starr 2006 reported there were no statistically significant differences between the two groups for length of surgery (mean duration: 106 versus 88 minutes; reported P = 0.26) or operative blood loss (mean loss: 328 versus 282 mL; reported P = 0.15).
PFNA versus Targon PF nail
Wild 2010 compared the PFNA with the Targon PF nail (a device that has an extra antirotation pin) in 80 patients with a pertrochanteric femoral fracture.
Final outcome measures
There were no perioperative deaths. Overall, 18 patients died in the first post‐operative year but their group allocation was not reported. Wild 2010 found no difference between the two groups in the modified Harris hip score (0 to 100: best outcome) at one year: mean 78.5 versus 78.1; reported P = 0.83). There were also no significant differences in the range of motion measures between the two groups.
Fracture fixation complications
There were no statistically significant differences between the two groups for these outcomes, which included cut‐out (three versus two), periprosthetic fracture (one versus zero); fracture non‐union (zero versus one), implant breakage (zero versus zero) and superficial wound infection (four versus two): seeAnalysis 7.1. Two cut‐outs in the PFNA group required a re‐operation as there was irritation of the iliotibial band/tract. Revision surgery was implied for all six cases of wound infection but we suspect this was a typographical error in the report given the infections were all superficial. The sequelae of the periprosthetic fracture were not recorded.
Post‐operative complications and hospital stay
These outcomes were not reported.
Anatomical restoration
These outcomes were not reported.
Operative details
Duration of operation (66.2 minutes versus 84.7 minutes; reported P < 0.01) and fluoroscopy (103.6 sec versus 164.5 sec; reported P < 0.01) were significantly shorter in the PFNA group.
PFNA versus Gamma 3 nail
Two trials (Vaquero 2012; Xu 2010a) compared the PFNA nail with the Gamma 3 nail in 200 participants with unstable proximal femoral fractures.
Final outcome measures
Separate group data were not reported for the 43/136 participants in Xu 2010a who were not available at final follow‐up (17.68 months, range 12 to 27 months). Of these, three participants died in the immediate post‐operative period and a further 12 participants died before the final follow‐up. Twenty‐one participants were too ill to attend and seven were lost to follow‐up in terms of functional outcomes. Thirty‐five participants (55%) were lost to follow‐up at 12 months in Vaquero 2012. All four recorded deaths that had occurred by 12 months in Vaquero 2012 were in the PFNA group (4/31 versus 0/30; RR 8.72, 95% CI 0.49 to 155.27; seeAnalysis 8.1). There were no significant differences found in a variety of functional scores and return to mobility assessment. Vaquero 2012 found no significant difference between the groups in the Harris Hip Score at six and 12 months (65.1 versus 72.6; MD 7.50, 95% CI ‐12.19 to 27.19; seeAnalysis 8.2). Similarly, Xu 2010a found no statistically significant differences between the two groups in mobility scores (0 to 9: best score; 6.30 versus 6.10; MD 0.20, 95% CI ‐0.51 to 0.91, seeAnalysis 8.3) and similar proportions of participants in each group recovered their pre‐operative weight bearing ability (29/46 versus 32/47, RR 0.93, 95% CI 0.69 to 1.24 seeAnalysis 8.4). Vaquero 2012 found no significant difference between the two groups at six or 12 months in the SF‐36 Physical and Mental Health Scores (seeAnalysis 8.5 and Analysis 8.6), or the Katz ADL scores (0 to 6, higher score meaning best function) (4 versus 3.6, seeAnalysis 8.7) at 12 months. Xu 2010a found no difference between the two groups in mean hip flexion (98.3 versus 94.9 degrees, seeAnalysis 8.8). Slightly more participants in the PFNA had hip and thigh pain in Xu 2010a but this difference was not statistically significant (19/46 versus 11/47; RR 1.76, 95% CI 0.95 to 3.29, seeAnalysis 8.10). Vaquero 2012 used a numeric pain score (0 to 10, higher score meaning worse pain) to assess thigh pain and found no significant difference between the groups at 12 months follow‐up (mean score 1 versus 1.5, MD ‐0.50, 95% CI ‐1.80, 0.80, seeAnalysis 8.9).
Fracture fixation complications
None of the differences between the two groups in specific fracture fixation complications were statistically significant; seeAnalysis 8.11. Xu 2010a reported no significant difference in intra‐operative femoral shaft fractures (2/66 versus 1/70), all of which were treated with delayed weight bearing for six to eight weeks. One (1/97) femoral shaft fracture occurred in the PFNA group one month post‐operatively and was treated with plate osteosynthesis; this was the only reported re‐operation in Xu 2010a. One post‐operative femoral (1/100) shaft fracture occurred in the Gamma 3 group and was treated with a secondary procedure, the exact details of which were not given. There were no significant differences between the groups in blade/screw cut‐out (2/97 versus 0/100, RR 4.84, 95% CI 0.24 to 96.89, seeAnalysis 8.11) or non‐union rates (2/30 versus 3/31). The two cases of cut‐out reported occurred in the PFNA group and were treated with a secondary procedure. There were no significant differences between the groups in superficial wound infection (2/66 versus 1/70), wound haematoma (5/66 versus 6/70) or deep wound infection (2/31 versus 0/30). There was no significant difference in proximal screw migration in Xu 2010a who reported nine cases of proximal screw migration (6/66 versus 3/70), all of which were treated conservatively. The mean time to fracture healing was similar in the two groups of this trial (9.65 weeks versus 10.21 weeks; reported P = 0.183).
Post‐operative complications and hospital stay
Vaquero 2012 reported that 16 participants in each group had a "general" complication, which mainly related to the need for transfusion. There were similar numbers in the two groups of patients with post‐operative medical complications of chest infection, decubitus ulcers and urinary tract infection in Xu 2010a (seeAnalysis 8.12). In Vaquero 2012, the mean Sangha scores (questionnaire based co‐morbidity score; 1 to 6, higher score equals more comorbidity) were not significantly different at 12 months follow‐up (mean 4.6 versus 4.5, seeAnalysis 8.13). There was no significant difference between the two groups in length of hospital stay (seeAnalysis 8.14).
Anatomical restoration
No significant difference in the amount of femoral shortening was observed between the two groups at final follow‐up in Xu 2010a (5.30 mm versus 5.49 mm; MD ‐0.19 mm, 95% CI ‐1.23 to 0.85; seeAnalysis 8.15).
Operative details
There was no significant difference between the groups in operating time (two trials, MD ‐3.03 minutes, 95% CI ‐6.88 to 0.82 minutes; seeAnalysis 8.16); notably surgery took between 27 to 33 minutes longer in Xu 2010a. There was no significant difference in fluoroscopy time (2.7 versus 3.2 minutes; MD ‐0.50 minutes, 95% CI ‐0.88 to ‐0.12; seeAnalysis 8.16). Intra‐operative blood loss was significantly lower in the PFNA group (217.4 mL versus 272.7 mL; MD ‐55.30, 95% CI ‐94.70 to ‐15.90; seeAnalysis 8.16) in Xu 2010a. However, there was no significant difference between the two groups in the number of participants transfused (24/66 versus 31/70; RR 0.82, 95% CI 0.54 to 1.24; seeAnalysis 8.17) nor in the mean number of units transfused (1.95 versus 2.03 units) in this trial.
Dynamic versus static locked intramedullary hip screw
Hardy 2003 compared a dynamically locked intramedullary hip screw (IMHS), which was allocated to 42 patients, with the usual statically locked IMHS allocated to 39 patients.
Final outcome measures
No statistically significant difference was found between the two groups for mortality (seeAnalysis 9.1). Pain in the mid‐thigh region was reported at follow‐up for two participants of the dynamic group and six in the static group (seeAnalysis 9.2). The pain impaired walking in four of the latter group. All six participants reporting mid‐thigh pain in the static group had cortical hypertrophy. The other instance of cortical hypertrophy occurred in a participant of the dynamic group who did not report mid‐thigh pain. Hardy 2003 reported similar results in the two groups for accommodation, mobility scores and independence rating of survivors at one year.
Fracture fixation complications
There were no statistically significant differences between the two groups for these outcomes (seeAnalysis 9.3). Cut‐out of the implant occurred in one case in the dynamic group and a fracture below the tip in one case in the static group. Re‐operations (one versus three) resulted from these two complications as well as from two operations for hardware removal in the static group.
Post‐operative complications and hospital stay
Medical complications and length of hospital stay were not reported in Hardy 2003. Though there were some discrepancies between text and tables in the trial report for discharge destination and in‐hospital deaths, there was clearly no difference between the two groups in these outcomes.
Anatomical restoration
Incomplete data for leg shortening (seeAnalysis 9.4) showed no statistically significant difference between the two groups (subsidence of the nail in the femoral shaft was seen in nine participants of the dynamic group compared with none in the static group). No information on deformity was presented in Hardy 2003.
Operative details
There were no statistically significant differences between the two groups for length of surgery, operative blood loss, haemoglobin levels or transfusion requirements (seeAnalysis 9.5).
Sliding versus non‐sliding Gamma 3 nail
Zhu 2012 compared a sliding versus a non‐sliding lag screw in the Gamma 3 nail in 80 participants with intertrochanteric fractures. Separate data for some outcomes were presented by subgroups based on the AO fracture classification with group A being participants with the most stable fractures (AO 31A1.1, 1.2 and 1.3 fractures), group B being participants with less stable fractures (AO 31A2.1) and group C being participants with the least stable fractures (AO 31A2.2 and 2.3).
Final outcome measures
There was no statistical or clinically significant difference between the sliding and non‐sliding groups in their Harris hip scores (MD ‐1.27, 95% CI ‐4.98 to 2.43; seeAnalysis 10.1).
Fracture fixation complications
Although all four reported complications occurred in the non‐sliding group, the differences between the two groups for these outcomes, which were non‐union (0/40 versus 1/40), cut‐out (0/40 versus 1/40) and femoral shaft fracture (0/40 versus 2/40), were not statistically significant (seeAnalysis 10.2). There was also no significant difference between the two intervention groups in time to fracture healing (MD ‐0.06 months, 95% CI ‐0.55 to 0.43 months; seeAnalysis 10.3).
Post‐operative complications and hospital stay
There was no between‐group difference in the length of hospital stay (seeAnalysis 10.5), which was four days in each group. Post‐operative complications other than fracture fixation complications were not reported.
Anatomical restoration
Although in the 41 participants with unstable comminuted intertrochanteric fractures (Group C, AO 31A2.2 and A2.3 fractures) a significant difference in leg length was reported (seeAnalysis 10.6, 0.573 mm versus 0.955 mm; MD 0.38 mm, 95% CI 0.37 to 40 mm), this is not a clinically significant leg length discrepancy. No other anatomical parameters were reported.
Operative details
There were no significant differences between the two groups in the operation time (46.73 min versus 48.35 min) or intra‐operative blood loss (141.1 mL versus 138.5 mL); seeAnalysis 10.4). No participants received a blood transfusion.
InterTan nail versus PFNA II nail
Zhang 2013 compared the InterTan nail with the PFNA II nail in 113 participants with unstable intertrochanteric fractures.
Final outcome measures
Of the 113 participants in the study, 15 died within 12 months and five were lost to follow‐up due to illness.There was no difference in mortality at 12 months after the procedure (8/57 versus 7/56; RR 1.12, 95% CI 0.44 to 2.89, seeAnalysis 11.1). For the 93 participants followed‐up for one year or longer, there were no significant differences between the groups in their Harris hip scores (mean 80.2 versus 82.6, MD ‐2.40, 95% CI ‐7.50 to 2.70) or walking ability scores (mean 5.8 versus 6.1; MD ‐0.30, 95% CI ‐1.03 to 0.43; seeAnalysis 11.2). There was no difference between the groups in the number of participants complaining of hip pain at final follow‐up (2/47 versus 2/46). However, significantly fewer participants in the InterTan nail group had thigh pain (3/47 versus 12/46; RR: 0.24, 95% CI 0.07 to 0.81; seeAnalysis 11.3).
Fracture fixation complications
There were no significant differences between the two groups for outcomes such as cut‐out (2/47 versus 0/46), intra‐operative femoral shaft fracture (1/57 versus 2/56) or later femoral shaft fracture (0/47 versus 1/46), blade migration (0/47 versus 4/46), problems with distal locking or prominence of the nail proximally in the greater trochanter (seeAnalysis 11.4). A similar finding applied to re‐operation, the reasons for the five re‐operations were not detailed in the report (2/47 versus 3/46). There was a significant difference in time to fracture healing, with fractures healing quicker in the InterTan group (14 versus 17 weeks; MD ‐3.0, 95% CI ‐4.88 to ‐1.12 weeks; seeAnalysis 11.5).
Post‐operative complications and hospital stay
There was no difference between the groups in the occurrence of superficial (3/47 versus 2/46) or deep (2/47 versus 1/46) wound infection, deep vein thrombosis, pulmonary embolism, pressure sores or urinary tract infection (seeAnalysis 11.6). There was no significant difference in the length of hospital stay (mean 8.33 versus 8.03 days; seeAnalysis 11.7).
Anatomical restoration
These outcomes were not reported. (While results for femoral neck shortening were reported, the clinical significance of these is unclear given that the results in both groups were less than the 5 mm of shortening required to affect abductor function (Zlowodzki 2008)).
Operative details
Zhang 2013 found that the length of surgery was longer in the InterTan nail group (mean 66.5 versus 53.7 minutes; MD 12.80 min, 95% CI 7.87 to 17.73 minutes) as was the fluoroscopy time, which took 1.5 minutes longer on average (seeAnalysis 11.8). While the intra‐operative blood losses were also greater in this group, the between‐group difference was not significant (mean 235.5 mL versus 197.5 mL; MD 37.80 mL, 95% CI ‐4.12 to 79.72 mL).
Long versus standard PFNA nail
Okcu 2013 compared long versus standard PFNA nails in 40 participants with reverse oblique fractures. Aside from mortality, the results were presented for the 33 survivors at a minimum of one year follow‐up.
Final outcome measures
There was no significant difference in mortality at one year between the two groups (34/22 versus 4/22; RR 1.09, 95% CI 0.28 to 4.26; seeAnalysis 12.1). Similarly there was no significant difference at one year between the two groups in the reported functional outcomes, namely the Harris hip score (0 to 100; top score equals best outcome): 79 versus 74; MD 5.00, 95% CI ‐1.14 to 11.14; and the Parker and Palmer mobility score (0 to 9; top score equals best outcome): 5.5 versus 5.2; MD 0.30, 95% CI ‐0.94 to 1.54); seeAnalysis 12.2.
Fracture fixation complications
There were no statistically significant differences between the two groups for re‐operation or other outcomes in this category (seeAnalysis 12.3). Both re‐operations were in the long nail group, separately these were for blade cut‐out and deep infection. There were no cases of non‐union in either group. There was one case of superficial wound infection in the standard nail group.
Post‐operative complications and hospital stay
There was no significant difference in the length of hospital stay (5.4 days versus 4.9 days, reported P = 0.51). There was no report of post‐operative complications.
Anatomical restoration
There was no significant difference in the number of malunions, defined as angulation or rotation deformity more than 10 degrees or limb shortening more than one centimetre (6/18 versus 3/15; RR 1.67, 95% CI 0.50 to 5.56; seeAnalysis 12.3).
Operative details
As would be expected given that reaming of the femoral canal is required when placing a long nail, both duration of operation (71.8 versus 52.6 minutes; reported P < 0.001) and fluoroscopy (75.3 versus 58.6 seconds; reported P < 0.001) were significantly longer in the long PFNA group.
Discussion
Summary of main results
The 17 trials included in this review, involving a total of predominantly female and older participants with predominantly unstable trochanteric fractures, tested 12 comparisons of different cephalocondylic nail designs. There were no trials evaluating condylocephalic nails. A summary of the risk of bias assessment and findings for each of the comparisons are provided below.
Proximal femoral nail (PFN) versus Gamma nail
All four trials contributing to this comparison were at high risk of bias. Of note, is that two trials (Herrera 2002; Marques 2005) were quasi‐randomised and thus at high risk of selection bias; and the other two trials (Papasimos 2005; Schipper 2004) were at high risk of attrition bias.
There were no significant differences between groups in functional outcome, these data being limited to results from single trials. Pooled results from three trials showed no difference in mortality between the two groups (86/415 versus 80/415; risk ratio (RR) 1.08; 95% confidence interval (CI) 0.82 to 1.41). There were no statistically significant differences between the two implants in serious fixation complications (operative fracture of the femur, cut‐out, non‐union and later fracture of the femur) nor re‐operations (45/455 versus 36/455; RR 1.25, 95% CI 0.83 to 1.90). These revision rates (9.8% versus 7.9%) are high, particularly in the context of those found for short femoral nails (5.4%) or, indeed, the sliding hip screw (3.4%) (Parker 2010). Schipper 2004 acknowledged the high revision rates, for both implants in their study but stressed their inclusion of exclusively unstable fractures.
ACE trochanteric nail versus Gamma nail
The two small trials (Efstathopoulos 2007; Vidyadhara 2007) addressing this comparison used different implants or techniques and had different populations. They were judged as being at unclear risk of bias for most domains; Vidyadhara 2007 being at high risk of detection bias for subjective outcomes.
The outcome of Vidyadhara 2007, which had no deaths or loss to follow‐up, was very favourable for both groups as shown by the usually high Harris hip scores with very little variation within each group. Supposing that the correct statistics were presented, the clinical significance of the statistically significant differences in the Harris hip scores at one year follow‐up (MD 1.00, 95% CI 0.28 to 1.72) is questionable. Two participants died in each group of Efstathopoulos 2007. The only fracture fixation complication reported was a cut‐out which resulted in a re‐operation in Vidyadhara 2007.
ACE trochanteric nail versus Gamma 3 nail
The single trial (De Grave 2012) in this comparison was at unclear risk of bias for most domains but at high risk of detection bias for subjective outcomes. De Grave 2012 found no differences between the two implants in any of the outcomes assessed (functional score, mortality, fracture fixation complications and re‐operation). Two patients in each group underwent a revision procedure for either cut‐out or fracture displacement.
Gliding nail versus Gamma nail
The single trial (Fritz 1999) in this comparison was at unclear risk of bias for most domains, but at high risk of detection bias for subjective outcomes. Fritz 1999 found no differences between the two implants in any of the outcomes assessed (mortality, poor outcome, fracture fixation complications and re‐operation).
ENDOVIS nail versus intramedullary hip screw (IMHS)
The single trial (Makridis 2010) in this comparison was at unclear risk of bias for most domains but at high risk of detection bias for subjective outcomes.
Makridis 2010 reported poorer mobility scores in the ENDOVIS nail group: this reflected that more participants in the ENDOVIS nail group were bedridden, and thus unable to walk, after their operation (29/105 versus 18/110; RR 1.69, 95% CI 1.00 to 2.85). There were no significant differences between the two groups in other outcomes (mortality, fracture fixation complications, re‐operation).
Russell‐Taylor Recon nail versus long Gamma nail
The single trial (Starr 2006) included people with high energy fractures aged between 18 and 50 years; 17 of whom had concurrent surgery for other injuries. Thus, this was a very different population to the other trials in this review. Starr 2006 was at unclear risk of bias for most domains but at high risk of detection bias for subjective outcomes.
Starr 2006 found no notable differences in outcome between the two groups. However, there was a very high rate of elective removal of implants for pain (8/17 (47%) versus 5/17 (29%)) compared with the other trials. Overall, Starr 2006 was too small to conclude that the lack of differences between the two groups is a true finding.
Proximal femoral nail antirotation (PFNA) versus Targon PF nail
The quasi‐randomised trial (Wild 2010) for this comparison was at high risk of selection bias and at high risk of detection bias for subjective outcomes. This was a poorly reported trial, including the failure to provide separate group data for mortality or complete data for re‐operations.
Wild 2010 found no difference between the two groups in functional outcome; nor were there statistically significant between‐group differences in fracture fixation complications.
PFNA versus Gamma 3 nail
Both trials (Vaquero 2012; Xu 2010a) for this comparison were at high risk of detection bias for subjective outcomes and high risk of attrition bias.
No statistically significant differences between implants were found in a variety of functional scores including the Harris Hip Score, SF‐36 mental and physical health scores, a mobility score or recovery of pre‐operative weight bearing ability, hip range of movement, in hip or thigh pain, in fracture fixation complications or re‐operations (four re‐operations occurred in the PFNA group, two operations occurred in the Gamma 3 group). No significant difference in mortality was observed between the groups in Vaquero 2012. Separate group data for mortality were not provided in the other study (Xu 2010a). No significant difference was seen in post‐operative complications or length of stay. Intra‐operative blood loss was lower in the PFNA group though this did not translate to a significant difference in the amount of participants transfused or the mean number of units transfused.
Dynamic versus static locked intramedullary nail
The single trial (Hardy 2003) making this comparison was at unclear risk of bias for most domains. Though none of the differences between the two groups reached statistical significance, Hardy 2003 suggested that lower incidence of cortical hypertrophy of the bone at the level of the distal locking screws in the dynamic group was linked with the lower number of participants with mid‐thigh pain in the dynamic group.
Sliding versus non‐sliding Gamma 3 nail
The single trial (Zhu 2012) was at unclear risk of bias for most domains. This trial, which aimed to avoid osteoporotic fractures, included participants aged less than 60 years with intertrochanteric fractures. This is a different population to all the other trials in this review (except Starr 2006).
Zhu 2012 reported no significant differences in terms of the Harris hip scores, fracture fixation complications, length of stay and operative details such as operative time and blood loss. A significant leg length discrepancy was reported in Group C (most unstable fractures, AO A2.2, 2.3) with the mean difference being 0.38 mm. This is not a clinically significant leg length discrepancy.
InterTan nail versus PFNA II nail
The single trial (Zhang 2013) was at low risk of bias for most domains, but at high risk of bias for reporting of subjective outcomes. No significant differences were reported for the majority of domains including mortality, functional scores (Harris hip score, walking ability score), fracture fixation complications, post‐operative complications and length of hospital stay. Although participants in the PFNA II group were four times more likely to experience thigh pain, the implications of this were not clear and not apparent from the Harris hip score. Similarly, the implications of the fracture healing occurring on average three weeks earlier in the InterTan nail group were unclear as was the actual assessment of this outcome. InterTan nailing surgery took longer, with greater fluoroscopy exposure.
Long versus standard (short) PFNA nail
The single trial (Okcu 2013) was at low or unclear risk of bias for individual domains. This small trial, which reported results for 33 participants with reverse oblique fractures, was described as a "pilot" and no power analysis being conducted beforehand. No significant differences were found for the majority of outcomes including mortality, functional scores (Harris hip score, Parker and Palmer mobility score), fracture fixation complications and re‐operation. There was no significant difference in the length of hospital stay. As expected, overall operating and fluoroscopy times were shorter in the standard PFNA group; reaming of the femoral canal is rarely required with this implant.
Overall completeness and applicability of evidence
Nine of the 12 comparisons in the review were made in single trials, whose populations ranged from 34 participants to 215 participants. The maximum number of participants available in any pooled analysis for the two remaining comparisons was 910 and 197 respectively. There was no pooling of functional outcomes, such as Harris hip scores, which were under‐reported and recorded. For each comparison, there were insufficient patient numbers to rule out important differences, particularly in final outcomes, between the implants under test.
The trial populations were generally representative of the populations with these fractures. As indicated above, Starr 2006 and Zhu 2012 were two exceptions to this with an upper age limit of 50 years being applied in Starr 2006 and 60 years being applied in Zhu 2012. Starr 2006 was also exceptional in its inclusion of a few subtrochanteric fractures. The other exception is Okcu 2013, which included reverse oblique fractures only. Particular emphasis was made in Zhang 2013 on the use of nails that reflect a difference in geometry between Asian and Caucasian femur geometry.
It is noteworthy that this review is predominantly a set of comparisons of intramedullary nails from different manufacturers. Newer nails have different features aimed at enhancing stability and reducing known complications (such as operative or later femoral shaft fractures). But as well as testing for improved performance of these different features in the clinical setting, it could be conjectured that it is also the market place that has set the research agenda and the associated aims of these generally underpowered trials.
As stated above, functional outcomes were under‐recorded and reported. Several trials failed to report separate statistics for mortality. We have already alerted the reader to the unusually high re‐operation rates in the PFN versus Gamma nail comparison and also the very high rate of elective removal of implants for pain in Starr 2006.
Quality of the evidence
As summarised above for the individual comparisons, all trials were at unclear risk of bias for several domains and most trials were at high risk of detection bias for subjective outcomes. Poor reporting of methods of randomisation and participant flow were commonplace and resulted in concerns regarding both selection and attrition biases in several trials.
Proximal femoral nail (PFN) versus the Gamma nail
The quality of the evidence for this comparison was downgraded three levels for function (one for limitations in design and implementation that related to potential risk of bias; one for imprecision; and one because data were only available from one trial); two levels for mortality (one for limitations in design and implementation that related to potential risk of bias; and one because data were absent from one trial and a substantial number of participants for another trial) and two levels for fracture fixation complications and re‐operation (one for limitations in design and implementation that related to potential risk of bias; and one for either imprecision or that substantial amounts of data were absent).
ACE trochanteric nail versus the Gamma nail
The quality of the evidence for this comparison was downgraded three levels for all primary outcomes (one for limitations in design and implementation that related to potential risk of bias; one for imprecision; and one because data were only available from one trial or were highly unrepresentative of the general population).
PFNA versus the Gamma 3 nail
The quality of evidence for this comparison was downgraded three levels for mortality and function (one for limitation in design and implementation that related to potential risk of bias in multiple areas including allocation, blinding, attritional and reporting bias; one for imprecision of results due to small numbers of trial participants; and one because data were only available from one trial for the majority of these outcomes). The quality of evidence for fracture fixation complications was downgraded two levels (one for limitation in design that related to potential risk of bias; and one for imprecision related to the small number of participants in each trial or the substantial loss to follow‐up).
Remaining comparisons
With one exception, the quality of the evidence for each comparison involving single trials was downgraded three levels for all outcomes. This generally included the downgrading by one for limitations in design and implementation that related to potential risk of bias; one for imprecision; and one because data were from one underpowered trial. The consistent finding of poorer mobility for the ENDOVIS nail when compared with the IMHS was considered to merit an upgrading by one point.
This overall, reflected a triple downgrading; we judged the evidence to be of very low quality, which indicates that we are very uncertain about the estimates for all outcomes. For the first comparison, however, there was a double downgrading for mortality. For the first (PFN versus the Gamma nail) and third (PFNA versus the Gamma 3 nail) comparisons there was a double downgrading for fracture fixation complications. For these, we judged the evidence to be low quality and thus we consider that further research is very likely to change the estimated effect and affect our confidence in this result.
Potential biases in the review process
While our search was comprehensive, it is possible that we have failed to identify some trials, especially those reported in conference proceedings only. Changing methodology and authorship between updates can be challenging and while we have taken a systematic approach, extra vigilance and checks have been required to ensure a satisfactory transition and consistency. Inevitably, the risk of bias judgements of similar aspects of trial quality do not neatly correspond to the previous ratings for previous trials; this difference is probably greater given that risk of bias assessment was done by a different pair of reviewers. The restructuring of the Types of outcome measures and reporting of the results presented the greatest challenge. Rather than completely rewrite the previous review, we took a pragmatic decision to reorder the categories of outcomes and highlight the primary outcomes.
Study flow diagram
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Comparison 1 Proximal femoral nail (PFN) versus Gamma nail, Outcome 1 Mortality.
Comparison 1 Proximal femoral nail (PFN) versus Gamma nail, Outcome 2 Final functional outcomes.
Comparison 1 Proximal femoral nail (PFN) versus Gamma nail, Outcome 3 Harris hip scores (0 to 100: high values = best function).
Comparison 1 Proximal femoral nail (PFN) versus Gamma nail, Outcome 4 Intra‐operative complications.
Comparison 1 Proximal femoral nail (PFN) versus Gamma nail, Outcome 5 Fracture healing complications.
Comparison 1 Proximal femoral nail (PFN) versus Gamma nail, Outcome 6 Re‐operation.
Comparison 1 Proximal femoral nail (PFN) versus Gamma nail, Outcome 7 Wound complications.
Comparison 1 Proximal femoral nail (PFN) versus Gamma nail, Outcome 8 Post‐operative complications.
Comparison 1 Proximal femoral nail (PFN) versus Gamma nail, Outcome 9 Length of hospital stay (days).
Comparison 1 Proximal femoral nail (PFN) versus Gamma nail, Outcome 10 Operative details: length of surgery and blood loss.
Comparison 1 Proximal femoral nail (PFN) versus Gamma nail, Outcome 11 Number of patients transfused.
Comparison 2 ACE trochanteric nail versus Gamma nail, Outcome 1 Mortality.
Comparison 2 ACE trochanteric nail versus Gamma nail, Outcome 2 Final functional outcomes.
Comparison 2 ACE trochanteric nail versus Gamma nail, Outcome 3 Mobility score (0: no difficulties to 9: most difficulties).
Comparison 2 ACE trochanteric nail versus Gamma nail, Outcome 4 Harris hip score (1 to 100: high values = best function).
Comparison 2 ACE trochanteric nail versus Gamma nail, Outcome 5 Fracture healing complications.
Comparison 2 ACE trochanteric nail versus Gamma nail, Outcome 6 Wound complications.
Comparison 2 ACE trochanteric nail versus Gamma nail, Outcome 7 Post‐operative complications.
Comparison 2 ACE trochanteric nail versus Gamma nail, Outcome 8 Anatomical restoration.
Comparison 2 ACE trochanteric nail versus Gamma nail, Outcome 9 Operative details: length of surgery, blood loss and radiographic screening time.
Comparison 2 ACE trochanteric nail versus Gamma nail, Outcome 10 Number of patients transfused.
Comparison 3 ACE trochanteric nail versus the Gamma 3 nail, Outcome 1 Mortality at 1 year.
Comparison 3 ACE trochanteric nail versus the Gamma 3 nail, Outcome 2 Fracture healing complications.
Comparison 3 ACE trochanteric nail versus the Gamma 3 nail, Outcome 3 Post‐operative complications.
Comparison 4 Gliding nail versus Gamma nail, Outcome 1 Mortality at 6 months.
Comparison 4 Gliding nail versus Gamma nail, Outcome 2 Residence and unfavourable outcome (geriatric institution or death) at 6 months.
Comparison 4 Gliding nail versus Gamma nail, Outcome 3 Fracture fixation complications.
Comparison 4 Gliding nail versus Gamma nail, Outcome 4 Post‐operative complications.
Comparison 4 Gliding nail versus Gamma nail, Outcome 5 Anatomical deformity.
Comparison 5 ENDOVIS nail versus intramedullary hip screw (IMHS), Outcome 1 Unable to walk (bedridden) post‐operatively.
Comparison 5 ENDOVIS nail versus intramedullary hip screw (IMHS), Outcome 2 Mortality.
Comparison 5 ENDOVIS nail versus intramedullary hip screw (IMHS), Outcome 3 Fracture fixation complications.
Comparison 5 ENDOVIS nail versus intramedullary hip screw (IMHS), Outcome 4 Number of patients transfused.
Comparison 6 Russell‐Taylor Recon nail versus long Gamma nail, Outcome 1 Final outcome measures.
Comparison 6 Russell‐Taylor Recon nail versus long Gamma nail, Outcome 2 Fracture healing and wound healing complications.
Comparison 7 PFNA versus Targon PF nail, Outcome 1 Fracture fixation complications.
Comparison 8 PFNA versus Gamma 3 nail, Outcome 1 Mortality (12 months).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 2 Harris hip score (1 to 100; higher values = best function).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 3 Mobility at 12+ months (Parker and Palmer mobility score: 0 to 9: best).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 4 Recovery of pre‐operative mobility (12+ months).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 5 SF‐36 Physical Health (0 to 100; higher scores = best function).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 6 SF‐36 Mental Health (0 to 100; higher scores = best function).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 7 Katz ADL score at 12 months (0 to 6; higher score = best function).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 8 Range of hip flexion (degrees).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 9 Thigh pain at 12 months (Numeric pain scale, 1 to 10, higher scores = most pain).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 10 Hip or thigh pain (12+ months).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 11 Fracture fixation complications.
Comparison 8 PFNA versus Gamma 3 nail, Outcome 12 Post‐operative complications.
Comparison 8 PFNA versus Gamma 3 nail, Outcome 13 Sangha Score at 1 year (1 to 6; higher score = more comorbidity).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 14 Length of stay (days).
Comparison 8 PFNA versus Gamma 3 nail, Outcome 15 Femoral shortening.
Comparison 8 PFNA versus Gamma 3 nail, Outcome 16 Operative details.
Comparison 8 PFNA versus Gamma 3 nail, Outcome 17 Number of patients transfused.
Comparison 9 Dynamic versus static locked intramedullary nail, Outcome 1 Mortality at 1 year.
Comparison 9 Dynamic versus static locked intramedullary nail, Outcome 2 Pain and cortical hypertrophy.
Comparison 9 Dynamic versus static locked intramedullary nail, Outcome 3 Fracture fixation complications.
Comparison 9 Dynamic versus static locked intramedullary nail, Outcome 4 Leg shortening (mm) in those able to undergo a radiographic assessment.
Comparison 9 Dynamic versus static locked intramedullary nail, Outcome 5 Operative details.
Comparison 10 Sliding versus non‐sliding lag screw for Gamma 3 nail, Outcome 1 Harris hip score (0 to 100: high values = best function).
Comparison 10 Sliding versus non‐sliding lag screw for Gamma 3 nail, Outcome 2 Fracture fixation complications.
Comparison 10 Sliding versus non‐sliding lag screw for Gamma 3 nail, Outcome 3 Average healing time (months).
Comparison 10 Sliding versus non‐sliding lag screw for Gamma 3 nail, Outcome 4 Operative details.
Comparison 10 Sliding versus non‐sliding lag screw for Gamma 3 nail, Outcome 5 Length of stay (days).
Comparison 10 Sliding versus non‐sliding lag screw for Gamma 3 nail, Outcome 6 Leg length discrepancy (mm) ('Group C' ‐ unstable fractures ‐ only).
Comparison 11 InterTan nail versus the PFNA II nail, Outcome 1 Mortality (1 year).
Comparison 11 InterTan nail versus the PFNA II nail, Outcome 2 Final functional outcomes.
Comparison 11 InterTan nail versus the PFNA II nail, Outcome 3 Hip and thigh pain.
Comparison 11 InterTan nail versus the PFNA II nail, Outcome 4 Fracture fixation complications.
Comparison 11 InterTan nail versus the PFNA II nail, Outcome 5 Time to fracture healing (weeks).
Comparison 11 InterTan nail versus the PFNA II nail, Outcome 6 Post‐operative complications.
Comparison 11 InterTan nail versus the PFNA II nail, Outcome 7 Length of stay (days).
Comparison 11 InterTan nail versus the PFNA II nail, Outcome 8 Operative details.
Comparison 12 Long versus standard proximal femoral nail antirotation (PFNA nail), Outcome 1 Mortality (1 year).
Comparison 12 Long versus standard proximal femoral nail antirotation (PFNA nail), Outcome 2 Final functional outcomes.
Comparison 12 Long versus standard proximal femoral nail antirotation (PFNA nail), Outcome 3 Fracture fixation complications.
| Name | Description |
| Endovis nail | The Endovis nail (Citieffe Ltd) is available in 3 sizes (195 to 400 mm) and has a neck shaft angle of 130°. It has two cephalic screws for the femoral head to facilitate fracture compression. The distal section is slotted to produce a graduated variation of stiffness |
| Gamma nail | The Gamma nail (Stryker Ltd) was introduced in the late 1980s for the treatment of extracapsular hip fractures. The implant consists of a sliding lag screw which passes through a short intramedullary nail placed via the trochanteric entry point. One or two screws may be passed through the nail tip to secure it to the femoral shaft (distal locking). Theoretical advantages of this implant are due to a percutaneous insertion technique and include reduced blood loss, minimal soft tissue trauma and short operating time. Modifications to the design of the Gamma nail and its instrumentation have occurred since its introduction. The long Gamma nail has a range of different lengths from 280 to 460 mm with two distal locking screw options. An Asian‐Pacific version of the nail is available for use in the Asian population and has reduced length, diameter and mediolateral angle to accommodate small femurs typically seen in this group |
| Gamma 3 nail | The Gamma 3 nail (Stryker Ltd) is the third generation of the gamma nail fixation system for proximal femoral fractures. It is a trochanteric entry nail with a reduced proximal nail diameter (15.5 mm versus 17 mm) to facilitate a shorter incision. Its length options range from 280 mm to 460 mm. Its neck shaft angle options include 120°, 125° and 130°. The lag screw shape has also been modified to provide superior cutting behaviour and greater resistance to cut‐out. One trial in this review compared a sliding and a non‐sliding lag screw mechanism in the Gamma 3 nail |
| Gliding nail | The gliding nail (Smith‐Nephew) is a trochanteric entry nail designed to avoid the complications of implants such as the Gamma nail. It utilises a T‐shaped femoral neck blade that has an extensive surface area relative to other intramedullary devices which provides good rotational stability and a high resistance to fatigue fracture. The T‐shaped blade can slide through the nail, facilitating fracture compression and healing. The standard length is 220 mm with long nail options ranging from 340 mm to 440 mm. Blade shaft angles include 125° and 135° |
| Intramedullary hip screw (IMHS) | The IMHS (Smith and Nephew), length 210 mm, was introduced in 1991 for the treatment of extracapsular femoral fractures. Like the Gamma nail, it consists of a nail inserted via the greater trochanter into the medullary cavity. It utilises a single screw in the femoral head that can slide through a barrel in the nail allowing fracture compression. Three different neck angles are available, 125°, 130° and 135°. Nail lengths are available from 195 mm to 440 mm |
| Proximal femoral nail (PFN) | The PFN (Synthes Ltd), length 240 mm, was introduced in 1998 for the treatment of extracapsular fractures. Like the Gamma and IMHS, it consists of a nail inserted via the greater trochanter into the medullary cavity. Two lengths are available, 200 mm and 240 mm. Two proximal lag screws are passed up the femoral neck to the head. Distal locking can be performed in static or dynamic mode via two distal locking screws |
| Proximal femoral nail antirotation (PFNA) | The PFNA (Synthes Ltd), length 170 mm, 200 mm or 240 mm, is a modification of the PFN nail. It is similar to the PFN nail apart from not having two proximal lag screws, but instead a single helically‐shaped blade which is designed to provide increased angular and rotational stability. The helical blade is designed to avoid bone loss that occurs during drilling and insertion of a standard hip screw. It has 2 distal locking screw options for either dynamic or static locking. Blade shaft angle options include 125°, 130° and 135° |
| Proximal femoral nail antirotation II (PFNA II) | The PFNA II (Synthes Ltd) is a modification of the PFNA nail to address the different proximal femoral anatomy of Asian patients. The PFNA has a large proximal diameter (17 mm) which was thought to account for the increase in femoral shaft fracture, lateral cortex splitting and thigh pain reported in Asian patients. The PFNA II has a smaller proximal diameter (16.5 mm versus 17 mm) and a flatter lateral shape (5° versus 6°) |
| Targon PF (proximal femoral) nail | The Targon PF nail (B Braun Ltd), length 220 mm, is inserted into the intramedullary cavity via a trochanteric entry point. Proximally, this nail has a sliding lag screw and an antirotation pin. The Targon PF nail facilitates fracture dynamisation via a gliding screw that glides through a sleeve that is attached to the nail, thereby avoiding protrusion of the screw into peritrochanteric tissues |
| ACE trochanteric nail | The ACE nail (Depuy) has a 10.5 mm lag screw and an optional antirotation lag screw. It has 2 distal holes for static or dynamic locking. Its proximal diameter is 16 mm and length is 180 mm or 200 mm |
| Russell‐Taylor Recon nail | The Russel‐Taylor Recon nail (Smith‐Nephew) is an intramedullary nail that utilises a piriformis entry point. 2 screws are available for fixation in the femoral head. It is a full length femoral nail with no short versions available for proximal femoral fixation only |
| InterTan nail | The InterTan nail (Smith‐Nephew) uses 2 cephalocervical screws in an integrated mechanism allowing intra‐operative compression and rotational stability of the head‐neck fragments. It has a cannulated set screw mechanism that allows for the device to be used in fixed angle mode or in sliding/compression mode. Its length ranges from 180 mm to 460 mm (long nail option) |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mortality Show forest plot | 3 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 1.1 At 4 months | 1 | 424 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.27 [0.82, 1.96] |
| 1.2 At 12 months | 3 | 830 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.08 [0.82, 1.41] |
| 2 Final functional outcomes Show forest plot | 3 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2.1 Pain at follow up | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Symptoms or restriction from the hip | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.3 Incomplete recovery of walking ability (including death) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Harris hip scores (0 to 100: high values = best function) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 3.1 At 4 weeks | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 At 4 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.3 At 1 year | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4 Intra‐operative complications Show forest plot | 4 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 4.1 Changed method of fixation | 1 | 424 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.76 [0.17, 3.34] |
| 4.2 Open reduction | 1 | 424 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.15 [0.95, 4.86] |
| 4.3 Poor reduction | 2 | 504 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.01 [0.62, 6.57] |
| 4.4 Difficult surgery | 2 | 504 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.46 [0.98, 2.19] |
| 4.5 Difficult proximal or distal screw insertion | 1 | 424 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.01 [0.36, 2.83] |
| 4.6 Intra‐operative comminution of the fracture around the trochanteric region | 2 | 330 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.3 [0.12, 0.73] |
| 4.7 Operative fracture of the femur | 4 | 910 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.33 [0.07, 1.63] |
| 4.8 Suboptimal position of fixation devices | 1 | 424 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.23 [0.77, 1.95] |
| 5 Fracture healing complications Show forest plot | 4 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 5.1 Cut‐out | 4 | 910 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.71 [0.39, 1.30] |
| 5.2 Later fracture of femur | 4 | 910 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.82 [0.24, 2.84] |
| 5.3 Implant breakage | 3 | 754 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.34 [0.01, 8.21] |
| 5.4 Non‐union/pseudoarthrosis | 3 | 754 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.60 [0.14, 2.50] |
| 5.5 Secondary varus (> 10%) | 1 | 250 | Risk Ratio (M‐H, Fixed, 95% CI) | 4.5 [0.99, 20.41] |
| 5.6 Fracture site collapse due to screw migration | 1 | 250 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.5 [0.81, 7.76] |
| 5.7 Medial or lateral hip screw migration | 1 | 424 | Risk Ratio (M‐H, Fixed, 95% CI) | 6.06 [1.37, 26.74] |
| 5.8 Muscle pain due to 'point effect' | 1 | 250 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.57 [0.17, 1.90] |
| 6 Re‐operation Show forest plot | 4 | 910 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.25 [0.83, 1.90] |
| 7 Wound complications Show forest plot | 4 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 7.1 Seroma | 1 | 250 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.90 [0.51, 1.60] |
| 7.2 Haematoma | 4 | 910 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.97 [0.62, 1.51] |
| 7.3 Superficial infection | 3 | 754 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.65 [0.32, 1.29] |
| 7.4 Deep infection | 3 | 830 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.01 [0.34, 2.95] |
| 8 Post‐operative complications Show forest plot | 3 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 8.1 Pneumonia | 2 | 236 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.93 [0.12, 70.72] |
| 8.2 Pressure sores | 2 | 406 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.08 [0.51, 2.30] |
| 8.3 Deep vein thrombosis | 2 | 236 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.65 [0.22, 12.29] |
| 8.4 Pulmonary embolism | 3 | 486 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.5 [0.25, 8.85] |
| 8.5 Acute post‐operative mental confusion | 2 | 330 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.78 [0.44, 1.39] |
| 8.6 Urinary infection | 2 | 330 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.13 [0.44, 2.84] |
| 8.7 Digestive haemorrhage | 1 | 250 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.33 [0.01, 8.10] |
| 8.8 Acute kidney failure | 1 | 250 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.5 [0.05, 5.44] |
| 9 Length of hospital stay (days) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 10 Operative details: length of surgery and blood loss Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 10.1 Length of surgery (minutes) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 10.2 Blood loss (mL) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 11 Number of patients transfused Show forest plot | 2 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mortality Show forest plot | 2 | 185 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.0 [0.15, 6.85] |
| 2 Final functional outcomes Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2.1 Hip pain at 1 month | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Limp | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.3 Difficulty in squatting | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Mobility score (0: no difficulties to 9: most difficulties) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 4 Harris hip score (1 to 100: high values = best function) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 4.1 At 4 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.2 At 1 year | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.3 At 2 years | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5 Fracture healing complications Show forest plot | 2 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 5.1 Operative fracture of femur | 2 | 185 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 5.2 Later fracture of femur | 2 | 185 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 5.3 Cut‐out | 2 | 185 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.34 [0.01, 8.14] |
| 5.4 Non‐union | 2 | 185 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 5.5 All fracture healing complications | 2 | 185 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.34 [0.01, 8.14] |
| 5.6 Re‐operation | 2 | 185 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.34 [0.01, 8.14] |
| 6 Wound complications Show forest plot | 2 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 6.1 All wound infection | 2 | 185 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.33 [0.31, 5.69] |
| 6.2 Deep wound infection | 2 | 185 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 7 Post‐operative complications Show forest plot | 2 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 7.1 Deep vein thrombosis | 2 | 185 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.33 [0.01, 8.01] |
| 7.2 All medical complications | 1 | 88 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.24 [0.67, 2.27] |
| 8 Anatomical restoration Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 8.1 Shortening (1 cm or more) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9 Operative details: length of surgery, blood loss and radiographic screening time Show forest plot | 2 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 9.1 Length of surgery (minutes) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9.2 Operative blood loss (mls) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9.3 Units of blood transfused | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9.4 Radiographic screening time (minutes) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 10 Number of patients transfused Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mortality at 1 year Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2 Fracture healing complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2.1 Fixation failure (cut‐out or redisplacement) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Non‐union | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.3 Wound infection | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.4 Re‐operation | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Post‐operative complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 3.1 Deep vein thrombosis | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 Peripheral nerve injury | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mortality at 6 months Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2 Residence and unfavourable outcome (geriatric institution or death) at 6 months Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2.1 Living in a geriatric institution | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Unfavourable outcome (institutionalised or dead) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Fracture fixation complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 3.1 Operative fracture of femur | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 Later fracture of femur | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.3 Cut‐out | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.4 Technical complications of fixation | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.5 Re‐operation | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4 Post‐operative complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 4.1 Participants with a complication | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.2 Pressure sores (decubitus) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.3 Pneumonia | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.4 Cerebrovascular accident | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.5 Apoplexy | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.6 Forearm fracture | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5 Anatomical deformity Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 5.1 Leg shortening > 2 cm | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.2 External rotation > 20 degrees | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.3 Internal rotation > 20 degrees | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Unable to walk (bedridden) post‐operatively Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2 Mortality Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2.1 During hospital stay | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 At 1 year | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Fracture fixation complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 3.1 Joint penetration | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 Periprosthetic fracture | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.3 Nail breakage | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.4 Cut‐out | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.5 MIssed proximal hole | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.6 Misplaced proximal screws | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.7 Failure of distal locking | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.8 Z‐phenomenon | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.9 Reverse Z‐phenomenon | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.10 Proximal screw back‐out | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.11 Femoral shaft medialization | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.12 Femoral shaft fracture | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.13 Infection | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.14 Re‐operation | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4 Number of patients transfused Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Final outcome measures Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 1.1 Mortality | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 1.2 Non‐independent ambulator | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 1.3 Unable to do the same work | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2 Fracture healing and wound healing complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2.1 Operative fracture of femur | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Later fracture of femur | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.3 Cut‐out | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.4 Non‐union | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.5 All fracture healing complications | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.6 Wound infection (any type) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.7 Deep wound infection | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.8 Re‐operation | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Fracture fixation complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 1.1 Femoral neck cut‐out | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 1.2 Re‐operation to change femoral neck components | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 1.3 Periprosthetic fracture | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 1.4 Implant breakage | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 1.5 Fracture nonunion | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 1.6 Infection (superficial) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mortality (12 months) Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2 Harris hip score (1 to 100; higher values = best function) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2.1 6 months post‐operative score | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 12 months post‐operative score | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Mobility at 12+ months (Parker and Palmer mobility score: 0 to 9: best) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 4 Recovery of pre‐operative mobility (12+ months) Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 5 SF‐36 Physical Health (0 to 100; higher scores = best function) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5.1 6 months post‐operative score | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.2 12 months post‐operative score | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6 SF‐36 Mental Health (0 to 100; higher scores = best function) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 6.1 6 months post‐operative score | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6.2 12 months post‐operative score | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 7 Katz ADL score at 12 months (0 to 6; higher score = best function) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 8 Range of hip flexion (degrees) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 9 Thigh pain at 12 months (Numeric pain scale, 1 to 10, higher scores = most pain) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 10 Hip or thigh pain (12+ months) Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 11 Fracture fixation complications Show forest plot | 2 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 11.1 Intra‐operative femoral shaft fracture | 2 | 197 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.12 [0.20, 22.85] |
| 11.2 Cut‐out | 2 | 197 | Risk Ratio (M‐H, Fixed, 95% CI) | 4.84 [0.24, 96.89] |
| 11.3 Later femoral shaft fracture | 2 | 197 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.01 [0.15, 6.92] |
| 11.4 Deep wound infection | 1 | 61 | Risk Ratio (M‐H, Fixed, 95% CI) | 4.84 [0.24, 96.89] |
| 11.5 Superficial wound infection | 2 | 197 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.39 [0.36, 15.92] |
| 11.6 Wound haematoma | 2 | 197 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.77 [0.27, 2.21] |
| 11.7 Proximal screw/blade migration | 1 | 136 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.12 [0.55, 8.14] |
| 11.8 Delayed healing | 1 | 61 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.97 [0.31, 3.01] |
| 11.9 Non‐union | 2 | 197 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.45 [0.26, 8.09] |
| 11.10 Implant breakage | 1 | 61 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.32 [0.01, 7.63] |
| 11.11 Failure of fixation | 1 | 61 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.26 [0.64, 7.93] |
| 12 Post‐operative complications Show forest plot | 2 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 12.1 Number with "general" complications (mainly need for transfusion) | 1 | 61 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.97 [0.60, 1.56] |
| 12.2 Chest infection | 1 | 136 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.33 [0.37, 4.73] |
| 12.3 Decubitus ulcer | 1 | 136 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.06 [0.15, 7.31] |
| 12.4 Urinary tract infection | 1 | 136 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.06 [0.15, 7.31] |
| 13 Sangha Score at 1 year (1 to 6; higher score = more comorbidity) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 14 Length of stay (days) Show forest plot | 2 | 197 | Mean Difference (IV, Fixed, 95% CI) | ‐0.17 [‐0.63, 0.29] |
| 15 Femoral shortening Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 16 Operative details Show forest plot | 2 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 16.1 Operating time (minutes) | 2 | 197 | Mean Difference (IV, Fixed, 95% CI) | ‐3.03 [‐6.88, 0.82] |
| 16.2 Fluoroscopy time (minutes) | 1 | 136 | Mean Difference (IV, Fixed, 95% CI) | ‐0.5 [‐0.88, ‐0.12] |
| 16.3 Intra‐operative blood loss (mL) | 1 | 136 | Mean Difference (IV, Fixed, 95% CI) | ‐55.30 [‐94.70, ‐15.90] |
| 17 Number of patients transfused Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mortality at 1 year Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2 Pain and cortical hypertrophy Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2.1 Mid‐thigh pain | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Cortical hypertrophy | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Fracture fixation complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 3.1 Later fracture of the femur | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 Cut‐out | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.3 Technical complications of fixation | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.4 Re‐operation | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4 Leg shortening (mm) in those able to undergo a radiographic assessment Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5 Operative details Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5.1 Length of surgery (minutes) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.2 Intra‐operative blood loss (minutes) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.3 Haemoglobin level: 48 hours post‐op (g/dL) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.4 Transfused packed blood cells | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Harris hip score (0 to 100: high values = best function) Show forest plot | 1 | 80 | Mean Difference (IV, Fixed, 95% CI) | ‐1.27 [‐4.98, 2.43] |
| 1.1 Group A (AO A1.1, 1.2, 1.3) | 1 | 11 | Mean Difference (IV, Fixed, 95% CI) | ‐2.25 [‐10.07, 5.57] |
| 1.2 Group B (AO A2.1) | 1 | 28 | Mean Difference (IV, Fixed, 95% CI) | ‐0.80 [‐8.24, 6.64] |
| 1.3 Group C (AO A2.2, 2.3) | 1 | 41 | Mean Difference (IV, Fixed, 95% CI) | ‐1.08 [‐6.19, 4.03] |
| 2 Fracture fixation complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2.1 Femoral shaft fracture (timing not known) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Cut‐out | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.3 Non‐union at 6 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.4 Non‐union at 12 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Average healing time (months) Show forest plot | 1 | 80 | Mean Difference (IV, Fixed, 95% CI) | ‐0.06 [‐0.55, 0.43] |
| 3.1 Group A (AO A1.1,1.2,1.3) | 1 | 11 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [‐0.84, 0.84] |
| 3.2 Group B (AO, A2.1) | 1 | 28 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [‐0.65, 0.65] |
| 3.3 Group C (AO, A2.2,2.3) | 1 | 41 | Mean Difference (IV, Fixed, 95% CI) | ‐0.62 [‐2.25, 1.01] |
| 4 Operative details Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 4.1 Operation time (mins) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.2 Intra‐operative blood loss (mL) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5 Length of stay (days) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 6 Leg length discrepancy (mm) ('Group C' ‐ unstable fractures ‐ only) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mortality (1 year) Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2 Final functional outcomes Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2.1 Harris hip score (1 to 100; high values = best function) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Walking ability score (0 to 9; high value = best function) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Hip and thigh pain Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 3.1 Hip pain | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 Thigh pain | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4 Fracture fixation complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 4.1 Cutout | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.2 Femoral shaft fracture (post‐operative) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.3 Femoral shaft fracture (intra‐operative) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.4 Lateral greater trochanter fracture | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.5 Proximal end of nail penetrating trochanter | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.6 Distal interlocking problem | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.7 Blade migration | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.8 Re‐operation | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5 Time to fracture healing (weeks) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 6 Post‐operative complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 6.1 Superficial wound infection | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6.2 Deep infection | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6.3 Deep venous thrombosis | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6.4 Pulmonary embolism | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6.5 Pressure sore | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6.6 Urinary tract infection | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 7 Length of stay (days) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 8 Operative details Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 8.1 Operative time (min) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 8.2 Blood loss (mL) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 8.3 Fluoroscopy time (seconds) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mortality (1 year) Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2 Final functional outcomes Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2.1 Harris hip score (1 to 100, top score = best function) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Parker and Palmer mobility score (0 to 9, top score = best function) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Fracture fixation complications Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 3.1 Re‐operation | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 Blade cut‐out | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.3 Wound infection (deep) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.4 Wound infection (superficial) | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.5 Non‐union | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.6 Malunion | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
