Intervention Review

You have free access to this content

Surgical versus conservative interventions for treating ankle fractures in adults

  1. Christian CMA Donken1,*,
  2. Hesham Al-Khateeb2,
  3. Michael HJ Verhofstad3,
  4. Cornelis JHM van Laarhoven4

Editorial Group: Cochrane Bone, Joint and Muscle Trauma Group

Published Online: 15 AUG 2012

Assessed as up-to-date: 6 MAY 2012

DOI: 10.1002/14651858.CD008470.pub2


How to Cite

Donken CCMA, Al-Khateeb H, Verhofstad MHJ, van Laarhoven CJHM. Surgical versus conservative interventions for treating ankle fractures in adults. Cochrane Database of Systematic Reviews 2012, Issue 8. Art. No.: CD008470. DOI: 10.1002/14651858.CD008470.pub2.

Author Information

  1. 1

    Radboud University Nijmegen Medical Center, Department of General and Trauma Surgery, Nijmegen, Netherlands

  2. 2

    Princess Alexandra Hospital, Department of Trauma & Orthopaedics, Harlow, UK

  3. 3

    Saint Elisabeth Hospital, Department of General and Trauma Surgery, Tilburg, Noord Brabant, Netherlands

  4. 4

    Radboud University Nijmegen Medical Center, Department of Surgery, Nijmegen, Netherlands

*Christian CMA Donken, Department of General and Trauma Surgery, Radboud University Nijmegen Medical Center, PO Box 9101, Nijmegen, 6500 HB, Netherlands. christiandonken@hotmail.com.

Publication History

  1. Publication Status: New
  2. Published Online: 15 AUG 2012

SEARCH

 
Characteristics of included studies [ordered by study ID]
Bauer 1985

MethodsMethod of randomisation: random number tables.

Blinding: unclear.

Loss of follow-up: 11 (7 died, 4 lost to follow-up) out of 111. However, type A fractures (n = 8) are excluded in the analysis. The study reports the follow-up results of only the type B fractures (n = 92).

Patients' consent to randomisation: unclear.

Number of surgeons: 10 surgeons with at least one year of experience in fracture treatment participated.


ParticipantsStudy locations: departments of orthopaedic surgery of the Malmö General Hospital in Malmö and the Östersund Hospital in Östersund, both in Sweden.

Study period: between 1 February 1968 and 21 June 1970.

Study participants: 111 randomised, baseline data for 108. Follow-up in 92 patients.

Age: 44 years, range 16 to 77 years.

Gender: 44 men and 64 women (of which 8 patients with type A fractures are excluded).

Inclusion criteria: patients with a diagnosis of a malleolar fracture.

Exclusion criteria: pilon fractures or tibial plafond fractures, type C fractures, patients with open fractures, other concomitant injuries to the lower limb, nondisplaced fractures, patients with open growth zones and those with diseases which entail increased risk of complications.


InterventionsSurgery: according to the AO principles (Müller 1979). Ligamentous injuries were sutured. Suction drainage as applied for 24 hours. Active movements started on the first postoperative day. The patients were kept in bed with the operated leg elevated in a splint until the fifth postoperative day. After 6 weeks, 20% weight bearing was allowed. Full weight bearing was allowed after 9 weeks.

Conservative: closed reduction was performed and a plaster cast was applied. The patients were discharged as soon as they were able to walk on crutches. Weight bearing was not allowed during the first 6 weeks in either group to prevent load-induced displacement. After 6 weeks, the cast was removed and 20% weight bearing was allowed. Full weight bearing was allowed after 12 weeks.


OutcomesLength of follow-up: mean 7 years, range 6 to 8 years.

A questionnaire (not validated and only introduced in the present study) measuring ankle girth, range of motion, presence of foot deformity, ability of walking, radiological outcome (bony union, anatomical result and osteoarthritis (according to Cedell (Cedell 1967) and Magnusson (Magnusson 1944)), hospital stay, mean time of treatment, sick-leave, complications (infection, DVT, pulmonary embolism, tenderness).


NotesLong-term results in this study only reported for displaced type B fractures. Early results were provided for 108 patients.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskA list of random numbers was used.

Allocation concealment (selection bias)Unclear riskThe authors only state: "random allocation was performed with the aid of a list of random numbers".

Blinding (performance bias and detection bias)
All outcomes
High riskThe authors state: "all films were evaluated by the same radiologist" and "at follow-up all patients were seen by two of the authors".

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskThe results at follow-up of 8 patients with type A fractures were not incorporated in the study. However baseline data were shown. After exclusion of these 8 patients, 'protocol deviants' were kept in their allocated group.

Selective reporting (reporting bias)Unclear riskThere is no study protocol available.

Other biasUnclear riskNo information to judge this.

Makwana 2001

MethodsMethod of randomisation: computer-generated random numbers.

Blinding: unclear.

Loss of follow-up: 7 (6 unwilling to attend, 1 died) out of 43 (16%). Eight (of which only 5 were available at follow-up) conservatively treated patients with secondary dislocation were excluded and not analysed (separately).

Patients' consent to randomisation: yes.

Number of surgeons: unknown number of registrars, senior registrars and consultants.


ParticipantsStudy location: level 1 trauma centre; Leicester Royal Infirmary, United Kingdom.

Study period: between January 1995 and February 1997.

Study participants: 43 randomised, baseline data for 43. Follow-up in 31 patients.

Age: 66 years, range 55 to 81 years.

Gender: 12 men and 31 women at baseline. Gender distribution at follow-up is unknown.

Inclusion criteria: all patients with a displaced ankle fracture over 55 years of age requiring operative treatment.

Exclusion criteria: pre-existing ipsilateral or contralateral ankle pathology, a pathological fracture (e.g. a stress fracture), an intra-articular fracture (e.g. a pilon fracture), inability to attend follow-up or inability to follow the postoperative regimen, unsatisfactory closed reduction (according to criteria by Phillips (Phillips 1985)), blisters precluding operative treatment, pilon fractures or tibial plafond fractures.


InterventionsSurgery: standard AO implants and techniques according to the AO/ASIF manual (Müller 1979). Additional information from correspondence (Ramasamy 2001): Nearly all the fractures of the ankle (96%) were fixed using a 1/3 semitubular plate on the lateral side, with or without lag screws. In one patient the lateral malleolus was treated by a Rush pin. AO cancellous screws were used in 82% of cases on the medial side, and this was supplemented in one by a Kirschner wire. No fixation was necessary in three medial malleoli. Postoperative regimen: a below-knee cast was retained for six weeks. The leg was elevated for 48 hours after which protected weight-bearing was allowed.

Conservative: a moulded below-knee plaster cast, with 48 hours limb elevation, after which protected weight-bearing was allowed for 6 weeks.


OutcomesLength of follow-up: mean 27 months, range 15 to 42 months.

Olerud and Molander ankle score (Olerud 1984), visual analogue score for pain, ankle girth, range of motion, inpatient stay, radiological outcome (bony union, malunion, nonunion, talar tilt and degenerative changes using the criteria of Magnusson (Magnusson 1944)), complications (wound problems, tenderness of the scar, malunion, nonunion, Complex Regional Pain Syndrome type 1 (Stanton-Hicks 1995), secondary dislocation and further surgery).


NotesIt is unclear from the report whether there were 47 randomised patients (Abstract states: "Forty-seven patients ... were entered into a prospective, randomised study") or 43 (according to the flow chart and 'Patients and methods section', where randomisation occurred after assessment of a satisfactory closed reduction). We have opted for the latter.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskA computer random number generator was used.

Allocation concealment (selection bias)Unclear riskThe authors only state: "patients had been randomised by computer-generated random numbers in theatre".

Blinding (performance bias and detection bias)
All outcomes
High riskThe authors only state: "all patients were assessed clinically and radiographically by the same observer (first author) at follow-up in order to minimise interobserver error".

Incomplete outcome data (attrition bias)
All outcomes
High riskOf the 22 patients in the ORIF group only 19 had successful manipulative reduction; the 3 patients with unsuccessful reduction were excluded, because they required surgery. Eight (of which only 5 were available at follow-up) conservatively treated patients with secondary dislocation were excluded and not analysed (separately). No intention-to-treat-analysis was performed.

Selective reporting (reporting bias)Unclear riskThere is no study protocol available.

Other biasHigh riskThe conservative group had significantly more smokers (6 versus 0).

Phillips 1985

MethodsMethod of randomisation: random number tables.

Blinding: unclear.

Loss of follow-up: 47 (could not be located or refused to be evaluated) of 96 (49%).

Patients' consent to randomisation: yes.

Number of surgeons: two members of the attending orthopaedic staff in combination with a resident.


ParticipantsStudy location: level 1 trauma centre; University of Chicago Hospitals, Chicago, United States of America.

Study period: between June 1975 and June 1979.

Study participants: 96 randomised, baseline characteristics for 93 (3 lost charts). Follow-up in 71 patients.

Age: 41.6 years, range 15 to 78 years.

Gender: 42 men and 54 women.

Inclusion criteria: patients with a diagnosis of a closed supination-external rotation grade-4 or a pronation-external rotation grade-4 ankle fracture, classified according to the modification of the Lauge Hansen system (Lauge Hansen 1942).

Exclusion criteria: open ankle fractures, pilon or tibial plafond fractures, open epiphyses, a previous fracture, concomitant injuries, pregnant women, a medical contraindication to a specific form of treatment that would hamper randomisation.


InterventionsThere were two treatment categories relevant to this review:

Surgery: satisfactory closed reduction and then open reduction and internal fixation by the Association for the Study of Internal Fixation (ASIF) technique. The leg was immobilised in a below-the-knee plaster cast. Walking on crutches without weight bearing started a few days after surgery until the tenth week. The cast was removed after 1 week to allow early active motion exercises. The trans-syndesmotic screw (if present) was removed after 10 weeks.

Conservative: satisfactory closed reduction followed by continued closed treatment. A toe-to-thigh cast with the knee flexed in 30 degrees was applied for 6 weeks, without weight-bearing. At the end of 6 weeks a below-the-knee walking cast was applied for an additional 4 weeks.


OutcomesLength of follow-up: 3.5 years, range 1.7 to 6.0 years.

The authors devised 3 scoring systems for this study that provided clinical (maximum 100 points for pain, range of motion and function), anatomical (maximum 35 points for talocrural angle, medial clear space, integrity of tibiofibular syndesmosis, medial malleolar displacement, size of posterior malleolar fragment, lateral malleolar displacement and shortening, talar tilt, talar subluxation, aspect of anteromedial corner) and arthritis scores (maximum15 points for non-union, synostosis, osteoporosis, irregular articular surface, osteophytes, periarticular cysts and joint-space narrowing). These three scores were combined to give a total score, for which a perfect result was 150 points.
Complications (secondary operation; pressure sores; superficial skin sloughs; non-union; deep infection)
Death (post-operative)


NotesThe trial also compared two methods of surgery where there was unsatisfactory closed reduction in 42 patients (open reduction and internal fixation by the ASIF technique versus open reduction and internal fixation of the medial malleolus). These groups are not included in this review.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskA random numbers table was used.

Allocation concealment (selection bias)Low riskSealed envelopes were used.

Blinding (performance bias and detection bias)
All outcomes
High riskNo information about blinding is provided in the study.

Incomplete outcome data (attrition bias)
All outcomes
High riskIn 47 patients follow-up could not be performed because patients could not be located or refused to be evaluated. In three of these patients baseline characteristics could not be provided, because of the loss of their charts.

Selective reporting (reporting bias)Unclear riskThere is no study protocol available.

Other biasUnclear riskNo information to judge this.

Rowley 1986

MethodsA quasi-randomised study using Accident and Emergency number.

Blinding: unclear.

Loss of follow-up: 2 (excluded because of open reduction and internal fixation after failed manipulation) of 42 (5%).

Patients' consent to randomisation: unclear.

Number of surgeons: unclear.


ParticipantsStudy location: The Royal Hallamshire Hospital, University of Sheffield, United Kingdom.

Study period: unclear.

Study participants: 42 randomised, no baseline characteristics provided (only type of fracture). Follow-up in 40 patients.

Age: mean not provided, range 16 to 70 years.

Gender: not provided.

Inclusion criteria: patients with a diagnosis of a displaced type ankle fracture.

Exclusion criteria: patients with pilon or tibial plafond fracture.


InterventionsSurgery: Open reduction and internal fixation, using the standard AO technique (Müller 1979), fixing the fibula with compression screws and a buttress plate, and the medial malleolus with screws or tension bands according to the fragment size. No attempt was made to institute any ligamentous repair. After fixation, the limb was placed in a below-knee backslab; early active movement was started on the first day after operation and continued for five days until a reasonable range of movement was established. The ankle was then placed in a below-knee plaster for 6 weeks and early weight-bearing was encouraged. At 6 weeks, patients were mobilised out of plaster and weight bearing was encouraged as soon as possible. No formal physiotherapy was offered.

Conservative: After manipulation, the position of the fracture was subsequently maintained in a long-leg plaster for 6 weeks and early weight-bearing was encouraged. The criteria of a good reduction were defined as correction of talar shift on the anterolateral radiograph and of posterior talar subluxation on the lateral film. On both views, the fibula had to be corrected in terms of rotation and length. The medial malleolus had to have a smooth articular profile, although a small step on its medial aspect was acceptable. Further radiographs were taken at one and two weeks from injury; if there was significant loss of position the patient was re-admitted and open reduction and internal fixation was carried out. At 6 weeks, patients were mobilised out of plaster and weight bearing was encouraged as soon as possible. No formal physiotherapy was offered.


OutcomesLength of follow-up: 20 weeks, no range provided.

Time to normal weight-bearing. Time for dorsiflexion to return to normal. Time for foot angle to return to normal (the position of the foot during walking by means of a footprint analysis).
Complications (none)
Treatment failure


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskThe study was quasi-randomised according to the last digit of the Accident and Emergency number.

Allocation concealment (selection bias)High riskThe authors provide no information about allocation concealment and was quasi-randomised.

Blinding (performance bias and detection bias)
All outcomes
High riskNo information about blinding is provided in the study.

Incomplete outcome data (attrition bias)
All outcomes
Low riskTwo patients were excluded in the analysis, because they required secondary surgery after failed manipulation. The authors' intention was to analyse these patients separately, but this was not done. The effects of these two missing patients were explored in analysis 1.2 and revealed no significant impact on the results.

Selective reporting (reporting bias)Unclear riskThere is no study protocol available.

Other biasUnclear riskNo information to judge this.

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Dietrich 2002Not a randomised controlled trial.

Doleżal 1995Not a randomised controlled trial.

Hughes 1979Not a randomised controlled trial. Depending on the hospital of presentation patients received only operative treatment (Kantonsspittal-St. Gallen, Switzerland) or conservative and operative treatment (University Surgery Clinic, Freiburg, Germany and Kantonsspital-Liestal, Switzerland).

Richter 1994Not a randomised controlled trial.

Salai 2000The absence of clarification of study methods from the trial author has failed to convince us that this was a randomised trial. Even if it was randomised, the reported results from the study could not be used given that the data from 19 surgically-treated patients who "met the above criteria for randomisation" were combined with 30 other surgically-treated patients, and compared with those of "an additional 16 [who] completed treatment by closed reduction".

Strömsöe 1995This randomised controlled trial including patients with an unstable Weber B or C ankle fracture without a medial malleolar fracture compared different surgical approaches. No patients were treated conservatively.

Velkovski 1995Not a randomised controlled trial.

Yde 1980aNot a randomised controlled trial. Depending on the hospital of presentation, patients with supination-eversion stage II fractures received conservative treatment (Fredrikhavn Hospital, Denmark) or operative treatment (Hjørring Hospital, Denmark).

Yde 1980bNot a randomised controlled trial. Depending on the hospital of presentation, patients with supination-eversion stage IV fractures received conservative treatment (Fredrikhavn Hospital, Denmark) or operative treatment (Hjørring Hospital, Denmark).

 
Characteristics of ongoing studies [ordered by study ID]
Gray

Trial name or titleInfluence of ankle skin perfusion on fracture treatment outcomes in the elderly: a randomised controlled trial comparing total contact cast with open reduction and internal fixation, the effect on cutaneous blood supply and relationship to skin complications

MethodsRandomised interventional trial

ParticipantsPatients aged over 60 years old with an unstable ankle fracture.

Estimated enrolment: 120.

InterventionsPatients will be anaesthetised and undergo either closed manipulation of the fracture and then application of a close contact cast or have a surgical open reduction and internal fixation by a contemporary method using metal plates and screws

OutcomesOlerud and Molander score, skin complications, radiological assessment, quality of life at 6 weeks and 6 months

Starting dateMay 2004

Contact informationGray B, Kadoorie Centre for Critical Care Research, Level 3, Headley Way, Headington, Oxford, United Kingdom, OX3 9DU

NotesExpected end date: 2010

Harris

Trial name or titleCombined randomised and observational study of type B ankle fracture treatment (CROSSBAT)

MethodsA prospective randomised multi-centre study

ParticipantsPatients aged between 18 and 65 years with a diagnosis of undisplaced (AO type 44-B1) distal fibula fracture
Estimated enrolment: 160

InterventionsOperative versus non operative treatment

OutcomesAmerican Academy of Orthopaedic Surgeons (AAOS) Foot and Ankle Questionnaire and SF-12 version 2 Health Survey

Starting dateAugust 2010

Contact informationMittal R, Bsc (Med), MBB, The University of New South Wales, Australia, +61 2 9828 3947, rajatmittal.syd@gmail.com

NotesDecember 2016

Pakarinen

Trial name or titleSyndesmotic injury and fixation in supination-external (SE) ankle fractures

MethodsRandomised controlled trial

ParticipantsAll skeletally mature patients (≥16 years old) with a unilateral Lauge-Hansen supination-external rotation type 4 ankle fractures treated within one week after injury at their hospital.

Estimated enrolment: 140

InterventionsComparing syndesmotic transfixation to no fixation in AO/OTA Weber B-type ankle fracture

OutcomesOlerud-Molander scoring system, RAND 36-Item Health Survey, and visual analogue scale (VAS) to measure pain and function after a minimum 1-year of follow-up

Starting dateJune 2007

Contact informationPakarinen HJ, Oulu University Hospital, Finland

NotesExpected end date: 2010

Sanders

Trial name or titleOperative versus non operative treatment for unstable ankle fractures

MethodsA prospective randomised multi-centre study

ParticipantsPatients aged between 18 and 65 years with a diagnosis of unstable Weber B unilateral fibular fracture
Estimated enrolment: 80

InterventionsOperative versus non operative treatment

OutcomesComparison of physical functioning score on SF36

Starting dateNovember 2001

Contact informationSanders D, M.D., FRCSC, University of Western Ontario, Canada, 519-685-8055

NotesExpected end date: 2010

Willett

Trial name or titleComparison of close contact cast (CCC) technique to open surgical reduction and internal fixation (ORIF) in the treatment of unstable ankle fractures in patients over 60 years

MethodsMulti-centre pragmatic individually randomised controlled equivalence study

ParticipantsPatients aged over 60 years with a diagnosis of an isolated displaced unstable ankle fracture
Estimated enrolment: 620

InterventionsClose contact cast versus open reduction and internal fixation

OutcomesOlerud and Molander ankle score, Iowa ankle score, radiological measurements of fracture and ankle joint congruence, Euroqol EQ-5D and SF-12 health survey

Starting dateOctober 2009

Contact informationProf. K. Willett, Kadoorie Centre John Radcliffe Hospital, OX3 9DU, University of Oxford, United Kingdom

NotesExpected end date: 2014

 
Comparison 1. Surgical versus conservative treatment

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Complaints of symptoms (pain, restricted range of ankle motion, unsteadiness, swelling) and walking difficulties at 7 years1Risk Ratio (M-H, Fixed, 95% CI)Totals not selected

    1.1 Complaints of significant symptoms
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    1.2 Pain
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    1.3 Restricted range of motion
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    1.4 Unsteadiness
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    1.5 Swelling
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    1.6 Passing stiffness
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    1.7 Walking difficulties on rough ground
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    1.8 Walking difficulties on even ground
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

 2 Functional or combined scores2Mean Difference (IV, Fixed, 95% CI)Totals not selected

    2.1 Olerud scores at mean 27 months (0 to 100: best score)
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    2.2 Combined (clinical, anatomical + arthritis: 0 to 150: best) scores at mean 3.5 years
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

 3 Pain scores (VAS: assume maximum of 100: worst) at mean of 27 months1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 4 Adverse events4Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    4.1 Early treatment failure (usually secondary operation)
4242Risk Ratio (M-H, Fixed, 95% CI)0.18 [0.06, 0.54]

    4.2 Deep / more serious infection
3172Risk Ratio (M-H, Fixed, 95% CI)0.23 [0.01, 4.61]

    4.3 Superficial infection
3172Risk Ratio (M-H, Fixed, 95% CI)0.97 [0.21, 4.37]

    4.4 Skin ulcer(s)
280Risk Ratio (M-H, Fixed, 95% CI)0.22 [0.03, 1.93]

    4.5 Deep vein thrombosis
192Risk Ratio (M-H, Fixed, 95% CI)0.76 [0.13, 4.34]

    4.6 Complex regional pain syndrome
131Risk Ratio (M-H, Fixed, 95% CI)3.25 [0.17, 62.40]

    4.7 Wound closure problem
131Risk Ratio (M-H, Fixed, 95% CI)1.95 [0.09, 44.32]

    4.8 Removal of internal fixation
131Risk Ratio (M-H, Fixed, 95% CI)4.55 [0.26, 81.03]

    4.9 Surgical scar tenderness
131Risk Ratio (M-H, Fixed, 95% CI)3.25 [0.17, 62.40]

 5 Functional impairment at 20 weeks1Risk Ratio (M-H, Fixed, 95% CI)Totals not selected

    5.1 Restricted dorsiflexion
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    5.2 Abnormal foot angle
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

 6 Differences in range of motion (injured ankle - other ankle)1Mean Difference (IV, Fixed, 95% CI)Totals not selected

    6.1 Dorso-plantar (degrees)
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    6.2 Dorsiflexion (degrees)
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    6.3 Pro-supination (degrees)
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

 7 Radiological results: non-union and malunion4Risk Ratio (M-H, Fixed, 95% CI)Totals not selected

    7.1 Non union
4Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    7.2 Malunion
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

 8 Radiological signs of osteoarthritis2141Risk Ratio (M-H, Fixed, 95% CI)1.05 [0.83, 1.31]

    8.1 Findings at mean of 7 years
192Risk Ratio (M-H, Fixed, 95% CI)1.10 [0.84, 1.45]

    8.2 Findings at mean of 3.5 years
149Risk Ratio (M-H, Fixed, 95% CI)0.94 [0.64, 1.40]

 9 Length of hospital stay (days)1Mean Difference (IV, Fixed, 95% CI)Totals not selected