- Email: [email protected]
Rest easy? Is bed rest really necessary after surgical repair of an ankle fracture?
Injury, Int. J. Care Injured 43 (2012) 766–771
Contents lists available at SciVerse ScienceDirect
Injury journal homepage: www.elsevier.com/locate/injury
Rest easy? Is bed rest really necessary after surgical repair of an ankle fracture? Lara A. Kimmel a,b,*, Elton R. Edwards b,c, Susan M. Liew c,d, Leonie B. Oldmeadow a, Melissa J. Webb a, Anne E. Holland a,e a
Department of Physiotherapy, The Alfred Hospital, Melbourne, Australia Department of Epidemiology and Preventative Medicine, Monash University, Melbourne, Australia c Associate Professor, Department of Orthopaedics, The Alfred Hospital, Melbourne, Australia d Department of Surgery, Monash University, Melbourne, Australia e Associate Professor of Physiotherapy, La Trobe University, Melbourne, Australia b
A R T I C L E I N F O
A B S T R A C T
Article history: Accepted 22 August 2011
Introduction: Bed rest with elevation of the affected limb is commonly prescribed postoperatively following ankle fracture fixation although there is no evidence that this is necessary. Aim: The aim of this prospective, randomised study was to investigate the effects of early mobilisation following surgical fixation of an ankle fracture on wound healing and length of stay (LOS). Method: A total of 104 patients underwent primary internal fixation of an ankle fracture at The Alfred hospital, Melbourne between July 2008 and January 2010. Intervention: The strategy included either early mobilisation group (first day post surgery) or control group (bed rest with elevation until day 2 post surgery). Outcome measures: Data collected included demographic, injury type and surgical procedure. Outcome data included inpatient LOS, wound condition at 10–14 days, opioid use and re-admission rate. Results: Groups were comparable at baseline. Wound breakdown rate was 2.9% (3 patients in the control group). Median LOS of the early mobilisation group was 55 h compared with 71 h in the control group (p < 0.0001). Opioid use for the control group was an average of 90 mg morphine equivalent in the first 24 h post surgery compared with 67 mg morphine equivalent for the early mobilisation group (p = 0.32). Conclusion: This study indicates that early mobilisation following surgical fixation of an ankle fracture results in a shorter hospital stay without evidence of an increased risk of re-admission or wound complication. ß 2011 Elsevier Ltd. All rights reserved.
Keywords: Early mobilisation Ankle Fracture Wound healing Length of stay
Introduction A common presentation to any emergency trauma service is the patient with an ankle fracture requiring surgical fixation.1 Subsequent inpatient management may include bed rest with the operated limb elevated for a time after surgery although this practice may be considered conservative in a health-care climate of shorter lengths of hospital stay (LOS). Tull et al. have described elevation as a well-established modality in postoperative patient care2 presumably based on the belief that wound complication rates, reported to be between 3% and 8%3,4 will be minimised. Studies in the area of hand surgery however suggest that elevation should be used judiciously as it may impair wound healing.5,6 Examining the presumed benefits of bed rest with elevation after
ankle surgery would be useful as unnecessary bed rest may prolong hospital discharge, increase health-care costs and possibly be associated with deleterious physical effects such as venous thrombo-embolism (VTE) and pneumonia.7 The aim of this study was to compare the effects of early mobilisation (day 1 postoperatively) versus day 2 mobilisation post ankle fracture fixation with the main outcomes investigated being LOS and surgical wound condition. Our hypothesis was that early mobilisation would be associated with a shorter length of hospital stay, without any increase in wound complication rates compared with mobilisation delayed 24 h or more. Patients and methods Design
* Corresponding author at: C/O Physiotherapy Department, The Alfred, Commercial Rd., Prahran, Melbourne 3181, Australia. Tel.: +61 3 9076 3450; fax: +61 3 9076 5430. E-mail address: [email protected] (L.A. Kimmel). 0020–1383/$ – see front matter ß 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.injury.2011.08.031
This is a prospective randomised controlled single-blinded study conducted at The Alfred, a large Level 1 Trauma Centre in Melbourne, Australia. The study was approved by The Alfred Research and Ethics committee.
L.A. Kimmel et al. / Injury, Int. J. Care Injured 43 (2012) 766–771
Subjects Between July 2008 and January 2010, consecutive adult participants (aged 18–80 years) who underwent internal fixation of their ankle fracture were recruited for the study. All fracture types (uni-, bi- and trimalleolar) were included. Participants were excluded if there were any conditions or associated injuries limiting their ability to mobilise early after their surgery, if they had a Pilon fracture or a Gustilo grade II or greater open fracture.8 Participants who were admitted from a nursing home or were previously non-ambulant or whose ankle fracture was the result of a pathological process were also excluded. Consenting participants were then randomised to one of two ‘time to first ambulation’ intervals: early (within 24 h/day 1 post operation) or usual practice (day 2 postoperatively). Those assigned to the ‘usual practice’ group acted as the control group. Randomisation was via a computer-generated program and allocation was concealed using opaque envelopes. All patients were admitted from the emergency department and reviewed by the treating surgeon to determine suitability for immediate (within 24 h) surgery. The timing of the surgical procedure (determined by the treating surgeon) was variable due to delays in patient presentation, the clinical condition of the patient on arrival at the Trauma Centre and operating room availability. Any preoperative period of bed rest and elevation for soft-tissue management was recorded. Fracture stabilisation Standardised stabilisation protocols were used. A fracture of the medial malleolus was stabilised using screws or a tension band wire technique, and a fracture of the lateral malleolus was stabilised with a plate and screws. A posterior malleolus fracture was stabilised with a screw if more than one-third of the articular surface was involved and a syndesmotic screw was used when a diastasis was present. The wound was closed in layers with vicryl in the subcutaneous tissues and either nylon or staples in the skin. All operations were performed by a consultant surgeon or a senior registrar under the supervision of a consultant surgeon. The ankle was then immobilised in a plaster backslab until wound review for 10–14 days, with no weight bearing allowed for 6 weeks. Intervention The gait retraining programme was conducted by an independent physiotherapist from whom group allocation was withheld by both staff and patient. Participants in the early mobilisation group ambulated with an appropriate gait aid and supervision on the first morning after their surgery (within 24 h). Participants in the control group rested in bed with their affected leg elevated on two or more pillows until the second postoperative day, when they commenced the same ambulation retraining. All participants were instructed in a bed-based, home exercise strength-maintenance programme. Discharge criteria Participants were discharged (directly home whenever possible) when medically stable, physically safe and independent with transfers and ambulation using an appropriate gait aid on flat ground and step (if required for access to their home). Participants were educated regarding leg elevation at all times other than when performing activities of daily living such as going to the toilet or shower. Physical readiness for discharge was determined by the independent physiotherapist who had no knowledge of group allocation. This physiotherapist was provided with the patient
767
demographics and operative notes with the date of operation removed. Data collection A research assistant prospectively collected the following demographic data: gender, age, cause of trauma, pre-morbid mobility (walking distance and gait aid), medical co-morbidities (including presence of diabetes and previous cardiovascular or neurological insults), alcohol intake, smoking history, medications, body mass index, fracture type, time to surgery and method of fracture fixation. At discharge, the following parameters were collected: LOS as measured from admission to discharge (in h) and from time of surgery to time of discharge (in h) and discharge destination (home or off-site rehabilitation facility). To remove the effects of other non-physical influences to discharge, we also looked at ‘physical readiness to discharge’, which was the time from operation to when the blinded physiotherapist deemed the participant physically ready to go home. Opioid requirement during 0–24 h postoperatively (described as morphine equivalence score in mg) was also collected as were any postoperative complications. Wound condition was determined by three measures that were collected at 10–14 days at the outpatient review by 2 independent assessors (1 nurse and 1 doctor) who were blinded to group allocation: (1) ASEPSIS wound score9: The ASEPSIS wound score was chosen as it best describes wound characteristics (erythema and separation of wound edges) with the greater percentage of the wound affected, the higher the score. One point is awarded to each 20% segment of wound that is deemed to have erythema, purulent exudate or separation of wound edges, with further points awarded if there is separation of deep tissues or a requirement for antibiotics. (2) Documented presence of infection: If an infection was suspected, a wound swab was obtained for confirmation. (3) Documented presence or absence of wound breakdown requiring re-operation. All patients were followed until the wound was deemed healed and no further wound review was required. All medical records were then reviewed at day 30 to ensure that no re-admissions were missed. Sample size For an 80% probability of detecting a difference between the two groups in LOS, 104 participants were required. This was based on the assumption that the true difference between groups would be 2 days with a standard deviation of 3.6 days. This was determined by a pilot study performed at The Alfred in 2006 looking at 20 patients with ankle fractures. Statistical analysis Differences between groups for LOS, time to safe ambulation, pain scores and ASEPSIS wound scores were unlikely to be normally distributed and were analysed using the Mann–Whitney U test. Categorical variables (presence of infection and admission to rehabilitation) were assessed using the chi-squared statistic. The effect of group allocation on time to hospital discharge, including both acute hospital and rehabilitation days, was evaluated using Kaplan–Meier survival curves.10 This method estimates the probability of discharge from hospital at each time point for the early ambulation and usual care (control) groups. The significance of any difference between the survival curves for each group is
L.A. Kimmel et al. / Injury, Int. J. Care Injured 43 (2012) 766–771
768
examined using the log rank test. Statistical significance was determined by a p value of <0.05. The impact of demographic characteristics on LOS was evaluated using multiple regression analysis. Based on a review of the literature,11–13 8 factors were identified a priori as being likely to impact on length of hospital stay: age, American Society of Anaesthesiologists (ASA) score,14 sex, insurance status, support at home, current smoking, average daily alcohol use at time of fracture and presence of a bimalleolar fracture. The ASA score provides a six-category physical status classification system for assessing a patient before surgery with grading from normal healthy patient (1) to brain-dead (6). Potential predictor variables with a relationship to LOS of p 0.1 on univariate analysis were included in the model. The LOS variables were not normally distributed and were transformed to natural logarithms prior to analysis. Results Between July 2008 and January 2010, 104 eligible participants with ankle fractures (60 males) were recruited for the study. All
attended for follow-up in 10–14 days. The flow of participants through the study is shown in Fig. 1. The mean age of the total group was 41.7 years (range 18–75 years) and male participants were younger than females (mean 36 (SD 14) vs. 50 (SD 16) years, p < 0.001). Most were healthy with few co-morbidities (54% had ASA score = 1). Baseline demographics (Table 1) and injury characteristics (Table 2) for the two groups were comparable. The timing of the surgical procedure was variable due to delays in patient presentation, condition of the soft tissues, and availability of the operating room and surgeon. Most participants had a preoperative period of bed rest and elevation with an average in hospital waiting time to operation of 21 h. Nine patients, with significant swelling, waited longer than 48 h at the surgeon’s request, as it was felt that there were significant risks of wound problems if immediate surgery was performed. In the first 24 h, those who ambulated early required an average 67 mg of morphine equivalent whilst those who rested in bed until day 2 postoperatively required an average of 90 mg of morphine equivalent (p = 0.32). Participants in the early mobilisation group were discharged from hospital earlier than those in the control group (total acute
Consort diagram Assessed for eligibility (n= 118)
Excluded (n=14)
Randomised (n=104)
Allocated to control (n= 53) Received allocated intervention (n= 48 ) Participants mobilised prior to 2 days post operatively (n=5)
Not meeting inclusion criteria (n= 11) Refused to participate (n= 2) unable to attend follow up (n= 1)
Allocated to experimental (n=51 ) Received allocated intervention (n= 51 )
Lost to follow-up (n=0 )
Lost to follow-up (n=0 )
Discontinued intervention (n= 0)
Discontinued intervention (n= 0)
Analysed (n=53 )
Analysed (n=51 )
Excluded from analysis (n= 0)
Excluded from analysis (n= 0)
Fig. 1.
L.A. Kimmel et al. / Injury, Int. J. Care Injured 43 (2012) 766–771
769
Table 1 Comparison of demographic characteristics between groups. Patient Characteristic
Control n = 53
Experimental n = 51
p Value
Gender, male Age, years (mean, SD) BMI (kg m 2) (mean, SD) Public patient ASA 3 History of diabetes (non insulin dependent) Employed Heavy/daily alcohol consumption Current smoker Carer at home
33 (62.1) 42.7 (16.6) 26.4 (3.9) 38 (71.7) 8 (15.1) 7 (13.2) 32 (60.4) 11 (20.8) 23 (43.4) 38 (71.7)
27 (52.9) 40.76 (15.8) 26.2 (4.3) 41 (80.1) 5 (9.8) 2 (3.9) 33 (64.7) 10 (19.6) 20 (39.2) 36 (70.6)
0.34 0.54 0.54 0.3 0.42 0.16 0.65 0.34 0.67 0.9
BMI: body mass index, and ASA: American Society of Anaesthesiologists. All data are n (%) unless indicated.
Table 2 Injury event and surgical management factors. Factor Mechanism of injury Low fall Sport High fall Motor vehicle Hit by object Pedestrian Assault Involvement of step Involvement of alcohol Weber B fracture Number of malleoli 1 2 3 Mean delay to theatre from time of injury in hours (SD) Closure with staples >1 incision
Control n = 53
Experimental n = 51
p Value
29 12 2 7 1 1 1 43 21 41
(54.7) (22.6) (3.8) (13.2) (1.9) (1.9) (1.9) (81.1) (39.6) (77.4)
35 (68.6) 11 (21.6) 1 (2.0) 1 (2.0) 1 (2.0) 0 2 (3.9) 36 (70.6) 14 (27.5) 40 (78.4)
0.35
22 13 18 54 13 23
(41.5) (24.5) (34.0) (79.1) (24.5) (43.4)
22 (43.1) 13 (25.5) 16 (31.4) 50.9 (67.3) 11 (21.6) 28 (54.9)
0.96
0.21 0.19 0.96
0.73 0.72 0.24
All data in n (%) unless indicated.
hospital median LOS 55 vs. 71 h, p < 0.0001). Only two participants in the early mobilisation group required rehabilitation compared with 8 in the control group (Table 3). The probability of discharge from hospital in the early mobilisation group was greater than in the control group at all time points (log rank test p < 0.001, Fig. 2). Factors other than time to ambulation that were associated with a longer postoperative LOS (acute and inpatient rehabilitation) included female gender (p = 0.008), age (p < 0.001), presence of a bimalleolar fracture (p = 0.08) and higher ASA score (p < 0.001). There were no significant differences in postoperative LOS between those who had a high impact injury (median 48 h) compared to those with other types of injuries (median 46 h, p = 0.66). Multiple regression shows that only time to ambulation and ASA score were independent predictors of postoperative LOS (Table 4), with time to ambulation making the greatest contribution. This model explained 50% of the variability in postoperative LOS. The overall wound scores for 64 participants (62.5%) were excellent, with both blinded assessors observing no noteworthy
redness, exudate or wound issues. The postoperative wound outcomes are recorded in Table 5. Post hoc analysis of ASEPSIS wound scores according to fracture severity (number of malleoli) found no difference between groups. There were 3 wound breakdowns requiring admission for re-operation; all were in the control group. A further review 1 week later was requested for 6 other participants who at that time were assessed as healed. Three other participants were re-admitted to hospital for reasons other than concerns regarding wound healing; one deep vein thrombosis (control group), one alcohol-related inability to cope at home (control group) and one asystolic arrest following drug and alcohol consumption (early mobilisation group). Discussion The results of this study suggest that bed rest and elevation after surgical fixation of an ankle fracture, commonplace in our institution and others,15 may be unnecessary. Mobilising from the bed within 24 h after surgery is associated with earlier hospital
Table 3 Post operative hospital related outcomes between groups. Outcome
Control group n = 53
Experimental group n = 51
p Value
Morphine equivalence score 0–24 h, mean (SD) (mg) Time to ambulation (h) Time to physical readiness for discharge (h) Total acute hospital LOS (h) Postoperative LOS (acute hospital) (h) Inpatient rehabilitation, n (%)
89.6 (82.3) 44.5 (41–47) 45 (42–47.8) 71 (62–114) 50 (46–65) 8 (15.1)
67 (42.4) 22 (18.5–24) 24 (20.5–30) 55 (35.3–70.5) 30 (24.5–48) 2 (3.9)
0.32 <0.0001 <0.0001 <0.0001 <0.0001 0.09
SOOB: sit out of bed, LOS: length of stay, and h: hours. All data are medians (interquartile range) unless indicated.
L.A. Kimmel et al. / Injury, Int. J. Care Injured 43 (2012) 766–771
770
Table 4 Multiple regression model for natural logarithm of post-operative length of stay. Unstandardised coefficients SE
B Constant Treatment group Age Sex ASA score Bimalleolar fracture R2
3.40 0.452 0.005 0.140 0.206 0.061
0.11 0.070 0.003 0.078 0.067 0.071
Standardised coefficients
p Value
Beta <0.001 <0.001 0.13 0.08 0.003 0.393
0.472 0.158 0.144 0.302 0.063
0.50
Larger standardised coefficients indicate variables that make a greater contribution to the model. The R2 value indicates that 50% of the variability in length of stay can be explained by the variables included in the model.
Fig. 2. Kaplan Meier curve for time to discharge.
discharge without an increase in wound complications compared to longer bed rest after surgery. Our acute hospital LOS (average 3.2 days) compares favourably with other published studies (LOS of 4 and 6 days).16 Our study used ‘physical readiness for discharge’ as a primary criterion for discharge to remove any confounding influence of non-biophysical factors such as social circumstances or inpatient bed availability.17 Those who mobilised earlier achieved our discharge criteria earlier (mean difference in acute hospital LOS 33 h, p < 0.001). Regression analysis demonstrated that early mobilisation and the general health of the patient (as measured by ASA score) were the most predictive of a short hospital LOS. The demonstrated effect of early mobilisation on shorter LOS may appear obvious at first but we had expected that early mobilisation would be more painful and thereby delay the achievement of our discharge criteria. This was not the case and our results demonstrate that early mobilisation
after ankle fracture fixation contributes to desirable health-care efficiency.3 We did not observe any wound complications in those who mobilised early. Three participants had wound complications requiring re-operation and all were in the control group. The overall wound complication rate of 2.9% is comparable to others (Carragee et al. (5%),4 Thangarajah et al. (8%)18 and Hoiness et al. (3.2%)3) although slightly higher than the 1.44% reported in a study by SooHoo et al.19 Time to first mobilisation was not reported in these studies and may have influenced the results. Our early mobilisation group was out of bed within the first postoperative day (range 7–24 h) and it would seem clinically significant that no wound complications occurred in this group. Other researchers have suggested that elevation may lead to a decrease in skin and tissue perfusion, resulting in worse outcomes after hand surgery.5,6 For the three participants in the control group who experienced a serious wound complication, mean bed rest and elevation time was 44.5 h. However, we cannot say whether the elevation or the bed rest contributed to the wound
Table 5 Follow up outcomes between groups. Outcome
Control group n = 53
Experimental group n = 51
p Value
Opioids at follow up Erythema > 20% wound Modified ASEPSIS wound score (medical), (mean, SD) Modified ASEPSIS wound score (nursing), (mean, SD) Return to theatre within 30 days due to wound Readmission within 30 days not related to wound
16 (30.2) 4 (7.5) 0.92 (4.2) 1.1 (3.9) 3 (5.7) 2 (3.8)
10 (19.6) 6 (11.8) 0.65 (1.6) 0.31 (0.7) 0 (0) 1 (2.0)
0.19 0.53 0.98 0.03 0.24 1.0
All data in n (%) unless indicated.
Table 6 Characteristics of 3 wound breakdown patients (all in control group). Characteristic
Patient 1
Patient 2
Patient 3
Age Sex ASA Comorbidites
65 Male 3 Peripheral vascular disease/ischaemic heart disease/ acute myocardial infarction Low fall Ex smoker Heavy Yes Unimalleolar, Weber B 38 Nylon
26 Male 1 Nil
52 Female 2 Depression
Low fall Ex smoker Social Yes Trimalleolar, Weber C 23 Nylon
Low fall Smoker Heavy Yes Trimalleolar, Weber B 10 Nylon
Cause of injury Smoking status Alcohol consumption Involvement of alcohol in fall Type of fracture Wait time to theatre post injury (h) Closure of wound
L.A. Kimmel et al. / Injury, Int. J. Care Injured 43 (2012) 766–771
breakdown and further research regarding this is needed. It is of interest that the three participants who suffered wound breakdowns had all consumed alcohol prior to their injury and had been involved in a low fall, (Table 6) although neither factor was statistically associated with wound breakdown. None of the three participants with wound breakdown had diabetes nor did we find other factors reported in the literature (medical conditions,19–21 high-energy trauma, alcohol intake,3 smoking, bimalleolar fracture type,18 age19 and ASA score14) to be associated with wound breakdown in our study population, although it is possible that these were not detected due to the low rate of wound complications and relatively young age of our sample. We were surprised to find that those who mobilised early did not require more analgesic support than those who rested for longer. We had previously observed that dependency of the surgical limb appeared to be a painful experience, presumably due to gravitational effects on circulation. However, this observation was not supported in this study. One possible explanation for this clinically significant finding may be that improving limb blood flow may decrease the pain associated with any increase in intracompartmental pressure.5 The absence of a validated method by which to quantify wound status may have weakened our study. To improve our judgement, we used 3 separate measures22 and 2 independent assessors (1 medical and 1 nurse). Although the validated ASEPSIS wound score9 is the most frequently used quantitative grading tool,23 we observed interrater differences between the nursing and medical assessor’s findings. Despite this, we are confident that wound outcomes were not compromised by the early mobilisation protocol. We do recognise that our study was powered to determine a reduction in LOS rather than wound complications, which may have introduced error. We also did not measure ankle swelling as an outcome of elevation, primarily as water displacement (the gold standard measurement tool) was thought to be unsuitable in the postoperative environment.24 Ankle circumference could have been considered but we did not perform this measurement as it was felt that the wound healing was the outcome measure of relevance. The strengths of this study were the randomisation, blinded assessors and high compliance with follow-up at 10–14 days. Conclusion This study is the first to suggest that mobilisation within 24 h of surgery for ankle fracture fixation does not result in an increased rate of wound complication or re-admission. Early mobilisation was associated with a decreased LOS and we now include this practise in our standardised treatment protocol. Conflict of interest The principal author received an Allied Health grant through the National Trauma Research Institute (sponsored by AAMI) to assist with completion of this study. The NTRI did not influence the study or results in any way.
771
References 1. Urquhart DM, Edwards ER, Graves SE, Williamson OD, McNeil JJ, Kossmann T. Characterisation of orthopaedic trauma admitted to adult Level 1 trauma centres. Injury 2006;37:120–7. 2. Tull F, Borrelli J. Soft tissue injury associated with closed fractures: evaluation and management. Journal of American Academy Orthopaedic Surgery 2003;11(6):431–8. 3. Hoiness P, Engebretsen L, Stromsoe K. Soft tissue problems in ankle fractures treated surgically: a prospective study of 154 consecutive closed ankle fractures. Injury 2002;34(12):928–31. 4. Carragee EJ, Csongradi JJ, Bleck EE. Early complications in the operative treatment of ankle fractures: influence of delay before operation. Journal of Bone and Joint Surgery (British) 1991;73:79–82. 5. Mars M. Hands Up? A preliminary study on the effect of post-operative hand elevation. The Journal of Hand Surgery 1988;B13(4):430–4. 6. Khan IU, Southern SJ, Nishikawa H. The effect of elevation on digital blood pressure. British Journal of Plastic Surgery 2001;54:137–9. 7. Allen C, Glasziou P, Del Mar C. Bed rest: a potentially harmful treatment needing more careful evaluation. The Lancet 1999;354(9186):1229–33. 8. Gustilo RB, Mendoza RM, Williams DN. Problems in the management of type III (severe) open fractures: a new classification of type III open fractures. Journal of Trauma 1984;24:742–6. 9. Wilson AP, Treasure T, Sturridge MF, Gruneberg RN. A scoring method (ASEPSIS) for postoperative wound infections for use in clinical trials of antibiotic prophylaxis. The Lancet 1986:311–2. 10. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. Journal of American Statistical Association 1958;53:457–81. 11. Leyes M, Torres R, Guillen P. Complications of open reduction and internal fixation of ankle fractures. Foot and Ankle Clinics 2003;8:133–47. 12. Nirula R, Nirula G, Gentilello LM. Inequity of rehabilitation services after traumatic injury. Journal of Trauma 2009;66(1):255–9. 13. Tønnesen H, Pedersen A, Jensen MR, Møller A, Madsen JC. Ankle fractures and alcoholism. The influence of alcoholism on morbidity after malleolar fractures. Journal of Bone and Joint Surgery (British) 1991;73(3):511–3. 14. Woodfield JC, Beshay NM, Pettigrew RA, Plank LD, van Rij AM. American society of anesthesiologists classification of physical status as a predictor of wound infection. ANZ Journal of Surgery 2007;77(9):738–41. 15. Konrad G, Markmiller M, Lenich A, Mayr E, Ruter A. Tourniquets may increase postoperative swelling and pain after internal fixation of ankle fractures. Clinical Orthopaedics and Related Research 2005;443:189–94. 16. Lindenhovious AL, Helmerhorts GT, Schnellen AC. Differences in prescription of narcotic pain medication after operative treatment of hip and ankle fractures in the United States and The Netherlands. Journal of Trauma Injury Infection and Critical Care 2009;67:160–4. 17. Jacobs DG, Sarafin JL, Norton J, Christmas AB, Huynh T. Wasted hospital days impair the value of length of stay variables in the quality assessment of trauma care. The American Surgeon 2009;75(9):794–803. 18. Thangarajah T, Prasad PS, Narayan B. Surgical site infections following open reduction and internal fixation of ankle fractures. The Open Orthopaedics Journal 2009;3:56–60. 19. SooHoo NF, Krenek L, Eagen MJ, Gurbani B, Ko CY, Zingmond DS. Complication rates following open reduction and internal fixation of ankle fractures. Journal of Bone and Joint Surgery (American) 2009;A91(5):1042–9. 20. McCormack RG, Leith JM. Ankle fractures in diabetics. Journal of Bone and Joint Surgery (British) 1998;B80:392–689. 21. Blotter RH, Connolly E, Wasan A, Chapman MW. Acute complications in the operative treatment of isolated ankle fractures in patients with diabetes mellitus. Foot and Ankle International 1999;20(11):687–94. 22. Wilson AP, Gibbons C, Reeves BC, Hodgson B, Liu M, Plummer D. Surgical wound infection as a performance indicator: agreement of common definitions of wound infection in 4773 patients. British Medical Journal 2004. 10.1136/ bmj.38232.646227. 23. Bruce J, Russell EM, Mollison J, Krukowski Z. The quality of measurement of surgical wound infection as the basis for monitoring: a systematic review. Journal of Hospital Infection 2001;49:99–108. 24. Peterson EJ, Irish SM, Lyons CL. Reliability of water volumetry and the figure of eight method on subjects with ankle joint swelling. Journal of Orthopaedics Sports and Physical Therapy 1999;29:609–15.
Contents lists available at SciVerse ScienceDirect
Injury journal homepage: www.elsevier.com/locate/injury
Rest easy? Is bed rest really necessary after surgical repair of an ankle fracture? Lara A. Kimmel a,b,*, Elton R. Edwards b,c, Susan M. Liew c,d, Leonie B. Oldmeadow a, Melissa J. Webb a, Anne E. Holland a,e a
Department of Physiotherapy, The Alfred Hospital, Melbourne, Australia Department of Epidemiology and Preventative Medicine, Monash University, Melbourne, Australia c Associate Professor, Department of Orthopaedics, The Alfred Hospital, Melbourne, Australia d Department of Surgery, Monash University, Melbourne, Australia e Associate Professor of Physiotherapy, La Trobe University, Melbourne, Australia b
A R T I C L E I N F O
A B S T R A C T
Article history: Accepted 22 August 2011
Introduction: Bed rest with elevation of the affected limb is commonly prescribed postoperatively following ankle fracture fixation although there is no evidence that this is necessary. Aim: The aim of this prospective, randomised study was to investigate the effects of early mobilisation following surgical fixation of an ankle fracture on wound healing and length of stay (LOS). Method: A total of 104 patients underwent primary internal fixation of an ankle fracture at The Alfred hospital, Melbourne between July 2008 and January 2010. Intervention: The strategy included either early mobilisation group (first day post surgery) or control group (bed rest with elevation until day 2 post surgery). Outcome measures: Data collected included demographic, injury type and surgical procedure. Outcome data included inpatient LOS, wound condition at 10–14 days, opioid use and re-admission rate. Results: Groups were comparable at baseline. Wound breakdown rate was 2.9% (3 patients in the control group). Median LOS of the early mobilisation group was 55 h compared with 71 h in the control group (p < 0.0001). Opioid use for the control group was an average of 90 mg morphine equivalent in the first 24 h post surgery compared with 67 mg morphine equivalent for the early mobilisation group (p = 0.32). Conclusion: This study indicates that early mobilisation following surgical fixation of an ankle fracture results in a shorter hospital stay without evidence of an increased risk of re-admission or wound complication. ß 2011 Elsevier Ltd. All rights reserved.
Keywords: Early mobilisation Ankle Fracture Wound healing Length of stay
Introduction A common presentation to any emergency trauma service is the patient with an ankle fracture requiring surgical fixation.1 Subsequent inpatient management may include bed rest with the operated limb elevated for a time after surgery although this practice may be considered conservative in a health-care climate of shorter lengths of hospital stay (LOS). Tull et al. have described elevation as a well-established modality in postoperative patient care2 presumably based on the belief that wound complication rates, reported to be between 3% and 8%3,4 will be minimised. Studies in the area of hand surgery however suggest that elevation should be used judiciously as it may impair wound healing.5,6 Examining the presumed benefits of bed rest with elevation after
ankle surgery would be useful as unnecessary bed rest may prolong hospital discharge, increase health-care costs and possibly be associated with deleterious physical effects such as venous thrombo-embolism (VTE) and pneumonia.7 The aim of this study was to compare the effects of early mobilisation (day 1 postoperatively) versus day 2 mobilisation post ankle fracture fixation with the main outcomes investigated being LOS and surgical wound condition. Our hypothesis was that early mobilisation would be associated with a shorter length of hospital stay, without any increase in wound complication rates compared with mobilisation delayed 24 h or more. Patients and methods Design
* Corresponding author at: C/O Physiotherapy Department, The Alfred, Commercial Rd., Prahran, Melbourne 3181, Australia. Tel.: +61 3 9076 3450; fax: +61 3 9076 5430. E-mail address: [email protected] (L.A. Kimmel). 0020–1383/$ – see front matter ß 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.injury.2011.08.031
This is a prospective randomised controlled single-blinded study conducted at The Alfred, a large Level 1 Trauma Centre in Melbourne, Australia. The study was approved by The Alfred Research and Ethics committee.
L.A. Kimmel et al. / Injury, Int. J. Care Injured 43 (2012) 766–771
Subjects Between July 2008 and January 2010, consecutive adult participants (aged 18–80 years) who underwent internal fixation of their ankle fracture were recruited for the study. All fracture types (uni-, bi- and trimalleolar) were included. Participants were excluded if there were any conditions or associated injuries limiting their ability to mobilise early after their surgery, if they had a Pilon fracture or a Gustilo grade II or greater open fracture.8 Participants who were admitted from a nursing home or were previously non-ambulant or whose ankle fracture was the result of a pathological process were also excluded. Consenting participants were then randomised to one of two ‘time to first ambulation’ intervals: early (within 24 h/day 1 post operation) or usual practice (day 2 postoperatively). Those assigned to the ‘usual practice’ group acted as the control group. Randomisation was via a computer-generated program and allocation was concealed using opaque envelopes. All patients were admitted from the emergency department and reviewed by the treating surgeon to determine suitability for immediate (within 24 h) surgery. The timing of the surgical procedure (determined by the treating surgeon) was variable due to delays in patient presentation, the clinical condition of the patient on arrival at the Trauma Centre and operating room availability. Any preoperative period of bed rest and elevation for soft-tissue management was recorded. Fracture stabilisation Standardised stabilisation protocols were used. A fracture of the medial malleolus was stabilised using screws or a tension band wire technique, and a fracture of the lateral malleolus was stabilised with a plate and screws. A posterior malleolus fracture was stabilised with a screw if more than one-third of the articular surface was involved and a syndesmotic screw was used when a diastasis was present. The wound was closed in layers with vicryl in the subcutaneous tissues and either nylon or staples in the skin. All operations were performed by a consultant surgeon or a senior registrar under the supervision of a consultant surgeon. The ankle was then immobilised in a plaster backslab until wound review for 10–14 days, with no weight bearing allowed for 6 weeks. Intervention The gait retraining programme was conducted by an independent physiotherapist from whom group allocation was withheld by both staff and patient. Participants in the early mobilisation group ambulated with an appropriate gait aid and supervision on the first morning after their surgery (within 24 h). Participants in the control group rested in bed with their affected leg elevated on two or more pillows until the second postoperative day, when they commenced the same ambulation retraining. All participants were instructed in a bed-based, home exercise strength-maintenance programme. Discharge criteria Participants were discharged (directly home whenever possible) when medically stable, physically safe and independent with transfers and ambulation using an appropriate gait aid on flat ground and step (if required for access to their home). Participants were educated regarding leg elevation at all times other than when performing activities of daily living such as going to the toilet or shower. Physical readiness for discharge was determined by the independent physiotherapist who had no knowledge of group allocation. This physiotherapist was provided with the patient
767
demographics and operative notes with the date of operation removed. Data collection A research assistant prospectively collected the following demographic data: gender, age, cause of trauma, pre-morbid mobility (walking distance and gait aid), medical co-morbidities (including presence of diabetes and previous cardiovascular or neurological insults), alcohol intake, smoking history, medications, body mass index, fracture type, time to surgery and method of fracture fixation. At discharge, the following parameters were collected: LOS as measured from admission to discharge (in h) and from time of surgery to time of discharge (in h) and discharge destination (home or off-site rehabilitation facility). To remove the effects of other non-physical influences to discharge, we also looked at ‘physical readiness to discharge’, which was the time from operation to when the blinded physiotherapist deemed the participant physically ready to go home. Opioid requirement during 0–24 h postoperatively (described as morphine equivalence score in mg) was also collected as were any postoperative complications. Wound condition was determined by three measures that were collected at 10–14 days at the outpatient review by 2 independent assessors (1 nurse and 1 doctor) who were blinded to group allocation: (1) ASEPSIS wound score9: The ASEPSIS wound score was chosen as it best describes wound characteristics (erythema and separation of wound edges) with the greater percentage of the wound affected, the higher the score. One point is awarded to each 20% segment of wound that is deemed to have erythema, purulent exudate or separation of wound edges, with further points awarded if there is separation of deep tissues or a requirement for antibiotics. (2) Documented presence of infection: If an infection was suspected, a wound swab was obtained for confirmation. (3) Documented presence or absence of wound breakdown requiring re-operation. All patients were followed until the wound was deemed healed and no further wound review was required. All medical records were then reviewed at day 30 to ensure that no re-admissions were missed. Sample size For an 80% probability of detecting a difference between the two groups in LOS, 104 participants were required. This was based on the assumption that the true difference between groups would be 2 days with a standard deviation of 3.6 days. This was determined by a pilot study performed at The Alfred in 2006 looking at 20 patients with ankle fractures. Statistical analysis Differences between groups for LOS, time to safe ambulation, pain scores and ASEPSIS wound scores were unlikely to be normally distributed and were analysed using the Mann–Whitney U test. Categorical variables (presence of infection and admission to rehabilitation) were assessed using the chi-squared statistic. The effect of group allocation on time to hospital discharge, including both acute hospital and rehabilitation days, was evaluated using Kaplan–Meier survival curves.10 This method estimates the probability of discharge from hospital at each time point for the early ambulation and usual care (control) groups. The significance of any difference between the survival curves for each group is
L.A. Kimmel et al. / Injury, Int. J. Care Injured 43 (2012) 766–771
768
examined using the log rank test. Statistical significance was determined by a p value of <0.05. The impact of demographic characteristics on LOS was evaluated using multiple regression analysis. Based on a review of the literature,11–13 8 factors were identified a priori as being likely to impact on length of hospital stay: age, American Society of Anaesthesiologists (ASA) score,14 sex, insurance status, support at home, current smoking, average daily alcohol use at time of fracture and presence of a bimalleolar fracture. The ASA score provides a six-category physical status classification system for assessing a patient before surgery with grading from normal healthy patient (1) to brain-dead (6). Potential predictor variables with a relationship to LOS of p 0.1 on univariate analysis were included in the model. The LOS variables were not normally distributed and were transformed to natural logarithms prior to analysis. Results Between July 2008 and January 2010, 104 eligible participants with ankle fractures (60 males) were recruited for the study. All
attended for follow-up in 10–14 days. The flow of participants through the study is shown in Fig. 1. The mean age of the total group was 41.7 years (range 18–75 years) and male participants were younger than females (mean 36 (SD 14) vs. 50 (SD 16) years, p < 0.001). Most were healthy with few co-morbidities (54% had ASA score = 1). Baseline demographics (Table 1) and injury characteristics (Table 2) for the two groups were comparable. The timing of the surgical procedure was variable due to delays in patient presentation, condition of the soft tissues, and availability of the operating room and surgeon. Most participants had a preoperative period of bed rest and elevation with an average in hospital waiting time to operation of 21 h. Nine patients, with significant swelling, waited longer than 48 h at the surgeon’s request, as it was felt that there were significant risks of wound problems if immediate surgery was performed. In the first 24 h, those who ambulated early required an average 67 mg of morphine equivalent whilst those who rested in bed until day 2 postoperatively required an average of 90 mg of morphine equivalent (p = 0.32). Participants in the early mobilisation group were discharged from hospital earlier than those in the control group (total acute
Consort diagram Assessed for eligibility (n= 118)
Excluded (n=14)
Randomised (n=104)
Allocated to control (n= 53) Received allocated intervention (n= 48 ) Participants mobilised prior to 2 days post operatively (n=5)
Not meeting inclusion criteria (n= 11) Refused to participate (n= 2) unable to attend follow up (n= 1)
Allocated to experimental (n=51 ) Received allocated intervention (n= 51 )
Lost to follow-up (n=0 )
Lost to follow-up (n=0 )
Discontinued intervention (n= 0)
Discontinued intervention (n= 0)
Analysed (n=53 )
Analysed (n=51 )
Excluded from analysis (n= 0)
Excluded from analysis (n= 0)
Fig. 1.
L.A. Kimmel et al. / Injury, Int. J. Care Injured 43 (2012) 766–771
769
Table 1 Comparison of demographic characteristics between groups. Patient Characteristic
Control n = 53
Experimental n = 51
p Value
Gender, male Age, years (mean, SD) BMI (kg m 2) (mean, SD) Public patient ASA 3 History of diabetes (non insulin dependent) Employed Heavy/daily alcohol consumption Current smoker Carer at home
33 (62.1) 42.7 (16.6) 26.4 (3.9) 38 (71.7) 8 (15.1) 7 (13.2) 32 (60.4) 11 (20.8) 23 (43.4) 38 (71.7)
27 (52.9) 40.76 (15.8) 26.2 (4.3) 41 (80.1) 5 (9.8) 2 (3.9) 33 (64.7) 10 (19.6) 20 (39.2) 36 (70.6)
0.34 0.54 0.54 0.3 0.42 0.16 0.65 0.34 0.67 0.9
BMI: body mass index, and ASA: American Society of Anaesthesiologists. All data are n (%) unless indicated.
Table 2 Injury event and surgical management factors. Factor Mechanism of injury Low fall Sport High fall Motor vehicle Hit by object Pedestrian Assault Involvement of step Involvement of alcohol Weber B fracture Number of malleoli 1 2 3 Mean delay to theatre from time of injury in hours (SD) Closure with staples >1 incision
Control n = 53
Experimental n = 51
p Value
29 12 2 7 1 1 1 43 21 41
(54.7) (22.6) (3.8) (13.2) (1.9) (1.9) (1.9) (81.1) (39.6) (77.4)
35 (68.6) 11 (21.6) 1 (2.0) 1 (2.0) 1 (2.0) 0 2 (3.9) 36 (70.6) 14 (27.5) 40 (78.4)
0.35
22 13 18 54 13 23
(41.5) (24.5) (34.0) (79.1) (24.5) (43.4)
22 (43.1) 13 (25.5) 16 (31.4) 50.9 (67.3) 11 (21.6) 28 (54.9)
0.96
0.21 0.19 0.96
0.73 0.72 0.24
All data in n (%) unless indicated.
hospital median LOS 55 vs. 71 h, p < 0.0001). Only two participants in the early mobilisation group required rehabilitation compared with 8 in the control group (Table 3). The probability of discharge from hospital in the early mobilisation group was greater than in the control group at all time points (log rank test p < 0.001, Fig. 2). Factors other than time to ambulation that were associated with a longer postoperative LOS (acute and inpatient rehabilitation) included female gender (p = 0.008), age (p < 0.001), presence of a bimalleolar fracture (p = 0.08) and higher ASA score (p < 0.001). There were no significant differences in postoperative LOS between those who had a high impact injury (median 48 h) compared to those with other types of injuries (median 46 h, p = 0.66). Multiple regression shows that only time to ambulation and ASA score were independent predictors of postoperative LOS (Table 4), with time to ambulation making the greatest contribution. This model explained 50% of the variability in postoperative LOS. The overall wound scores for 64 participants (62.5%) were excellent, with both blinded assessors observing no noteworthy
redness, exudate or wound issues. The postoperative wound outcomes are recorded in Table 5. Post hoc analysis of ASEPSIS wound scores according to fracture severity (number of malleoli) found no difference between groups. There were 3 wound breakdowns requiring admission for re-operation; all were in the control group. A further review 1 week later was requested for 6 other participants who at that time were assessed as healed. Three other participants were re-admitted to hospital for reasons other than concerns regarding wound healing; one deep vein thrombosis (control group), one alcohol-related inability to cope at home (control group) and one asystolic arrest following drug and alcohol consumption (early mobilisation group). Discussion The results of this study suggest that bed rest and elevation after surgical fixation of an ankle fracture, commonplace in our institution and others,15 may be unnecessary. Mobilising from the bed within 24 h after surgery is associated with earlier hospital
Table 3 Post operative hospital related outcomes between groups. Outcome
Control group n = 53
Experimental group n = 51
p Value
Morphine equivalence score 0–24 h, mean (SD) (mg) Time to ambulation (h) Time to physical readiness for discharge (h) Total acute hospital LOS (h) Postoperative LOS (acute hospital) (h) Inpatient rehabilitation, n (%)
89.6 (82.3) 44.5 (41–47) 45 (42–47.8) 71 (62–114) 50 (46–65) 8 (15.1)
67 (42.4) 22 (18.5–24) 24 (20.5–30) 55 (35.3–70.5) 30 (24.5–48) 2 (3.9)
0.32 <0.0001 <0.0001 <0.0001 <0.0001 0.09
SOOB: sit out of bed, LOS: length of stay, and h: hours. All data are medians (interquartile range) unless indicated.
L.A. Kimmel et al. / Injury, Int. J. Care Injured 43 (2012) 766–771
770
Table 4 Multiple regression model for natural logarithm of post-operative length of stay. Unstandardised coefficients SE
B Constant Treatment group Age Sex ASA score Bimalleolar fracture R2
3.40 0.452 0.005 0.140 0.206 0.061
0.11 0.070 0.003 0.078 0.067 0.071
Standardised coefficients
p Value
Beta <0.001 <0.001 0.13 0.08 0.003 0.393
0.472 0.158 0.144 0.302 0.063
0.50
Larger standardised coefficients indicate variables that make a greater contribution to the model. The R2 value indicates that 50% of the variability in length of stay can be explained by the variables included in the model.
Fig. 2. Kaplan Meier curve for time to discharge.
discharge without an increase in wound complications compared to longer bed rest after surgery. Our acute hospital LOS (average 3.2 days) compares favourably with other published studies (LOS of 4 and 6 days).16 Our study used ‘physical readiness for discharge’ as a primary criterion for discharge to remove any confounding influence of non-biophysical factors such as social circumstances or inpatient bed availability.17 Those who mobilised earlier achieved our discharge criteria earlier (mean difference in acute hospital LOS 33 h, p < 0.001). Regression analysis demonstrated that early mobilisation and the general health of the patient (as measured by ASA score) were the most predictive of a short hospital LOS. The demonstrated effect of early mobilisation on shorter LOS may appear obvious at first but we had expected that early mobilisation would be more painful and thereby delay the achievement of our discharge criteria. This was not the case and our results demonstrate that early mobilisation
after ankle fracture fixation contributes to desirable health-care efficiency.3 We did not observe any wound complications in those who mobilised early. Three participants had wound complications requiring re-operation and all were in the control group. The overall wound complication rate of 2.9% is comparable to others (Carragee et al. (5%),4 Thangarajah et al. (8%)18 and Hoiness et al. (3.2%)3) although slightly higher than the 1.44% reported in a study by SooHoo et al.19 Time to first mobilisation was not reported in these studies and may have influenced the results. Our early mobilisation group was out of bed within the first postoperative day (range 7–24 h) and it would seem clinically significant that no wound complications occurred in this group. Other researchers have suggested that elevation may lead to a decrease in skin and tissue perfusion, resulting in worse outcomes after hand surgery.5,6 For the three participants in the control group who experienced a serious wound complication, mean bed rest and elevation time was 44.5 h. However, we cannot say whether the elevation or the bed rest contributed to the wound
Table 5 Follow up outcomes between groups. Outcome
Control group n = 53
Experimental group n = 51
p Value
Opioids at follow up Erythema > 20% wound Modified ASEPSIS wound score (medical), (mean, SD) Modified ASEPSIS wound score (nursing), (mean, SD) Return to theatre within 30 days due to wound Readmission within 30 days not related to wound
16 (30.2) 4 (7.5) 0.92 (4.2) 1.1 (3.9) 3 (5.7) 2 (3.8)
10 (19.6) 6 (11.8) 0.65 (1.6) 0.31 (0.7) 0 (0) 1 (2.0)
0.19 0.53 0.98 0.03 0.24 1.0
All data in n (%) unless indicated.
Table 6 Characteristics of 3 wound breakdown patients (all in control group). Characteristic
Patient 1
Patient 2
Patient 3
Age Sex ASA Comorbidites
65 Male 3 Peripheral vascular disease/ischaemic heart disease/ acute myocardial infarction Low fall Ex smoker Heavy Yes Unimalleolar, Weber B 38 Nylon
26 Male 1 Nil
52 Female 2 Depression
Low fall Ex smoker Social Yes Trimalleolar, Weber C 23 Nylon
Low fall Smoker Heavy Yes Trimalleolar, Weber B 10 Nylon
Cause of injury Smoking status Alcohol consumption Involvement of alcohol in fall Type of fracture Wait time to theatre post injury (h) Closure of wound
L.A. Kimmel et al. / Injury, Int. J. Care Injured 43 (2012) 766–771
breakdown and further research regarding this is needed. It is of interest that the three participants who suffered wound breakdowns had all consumed alcohol prior to their injury and had been involved in a low fall, (Table 6) although neither factor was statistically associated with wound breakdown. None of the three participants with wound breakdown had diabetes nor did we find other factors reported in the literature (medical conditions,19–21 high-energy trauma, alcohol intake,3 smoking, bimalleolar fracture type,18 age19 and ASA score14) to be associated with wound breakdown in our study population, although it is possible that these were not detected due to the low rate of wound complications and relatively young age of our sample. We were surprised to find that those who mobilised early did not require more analgesic support than those who rested for longer. We had previously observed that dependency of the surgical limb appeared to be a painful experience, presumably due to gravitational effects on circulation. However, this observation was not supported in this study. One possible explanation for this clinically significant finding may be that improving limb blood flow may decrease the pain associated with any increase in intracompartmental pressure.5 The absence of a validated method by which to quantify wound status may have weakened our study. To improve our judgement, we used 3 separate measures22 and 2 independent assessors (1 medical and 1 nurse). Although the validated ASEPSIS wound score9 is the most frequently used quantitative grading tool,23 we observed interrater differences between the nursing and medical assessor’s findings. Despite this, we are confident that wound outcomes were not compromised by the early mobilisation protocol. We do recognise that our study was powered to determine a reduction in LOS rather than wound complications, which may have introduced error. We also did not measure ankle swelling as an outcome of elevation, primarily as water displacement (the gold standard measurement tool) was thought to be unsuitable in the postoperative environment.24 Ankle circumference could have been considered but we did not perform this measurement as it was felt that the wound healing was the outcome measure of relevance. The strengths of this study were the randomisation, blinded assessors and high compliance with follow-up at 10–14 days. Conclusion This study is the first to suggest that mobilisation within 24 h of surgery for ankle fracture fixation does not result in an increased rate of wound complication or re-admission. Early mobilisation was associated with a decreased LOS and we now include this practise in our standardised treatment protocol. Conflict of interest The principal author received an Allied Health grant through the National Trauma Research Institute (sponsored by AAMI) to assist with completion of this study. The NTRI did not influence the study or results in any way.
771
References 1. Urquhart DM, Edwards ER, Graves SE, Williamson OD, McNeil JJ, Kossmann T. Characterisation of orthopaedic trauma admitted to adult Level 1 trauma centres. Injury 2006;37:120–7. 2. Tull F, Borrelli J. Soft tissue injury associated with closed fractures: evaluation and management. Journal of American Academy Orthopaedic Surgery 2003;11(6):431–8. 3. Hoiness P, Engebretsen L, Stromsoe K. Soft tissue problems in ankle fractures treated surgically: a prospective study of 154 consecutive closed ankle fractures. Injury 2002;34(12):928–31. 4. Carragee EJ, Csongradi JJ, Bleck EE. Early complications in the operative treatment of ankle fractures: influence of delay before operation. Journal of Bone and Joint Surgery (British) 1991;73:79–82. 5. Mars M. Hands Up? A preliminary study on the effect of post-operative hand elevation. The Journal of Hand Surgery 1988;B13(4):430–4. 6. Khan IU, Southern SJ, Nishikawa H. The effect of elevation on digital blood pressure. British Journal of Plastic Surgery 2001;54:137–9. 7. Allen C, Glasziou P, Del Mar C. Bed rest: a potentially harmful treatment needing more careful evaluation. The Lancet 1999;354(9186):1229–33. 8. Gustilo RB, Mendoza RM, Williams DN. Problems in the management of type III (severe) open fractures: a new classification of type III open fractures. Journal of Trauma 1984;24:742–6. 9. Wilson AP, Treasure T, Sturridge MF, Gruneberg RN. A scoring method (ASEPSIS) for postoperative wound infections for use in clinical trials of antibiotic prophylaxis. The Lancet 1986:311–2. 10. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. Journal of American Statistical Association 1958;53:457–81. 11. Leyes M, Torres R, Guillen P. Complications of open reduction and internal fixation of ankle fractures. Foot and Ankle Clinics 2003;8:133–47. 12. Nirula R, Nirula G, Gentilello LM. Inequity of rehabilitation services after traumatic injury. Journal of Trauma 2009;66(1):255–9. 13. Tønnesen H, Pedersen A, Jensen MR, Møller A, Madsen JC. Ankle fractures and alcoholism. The influence of alcoholism on morbidity after malleolar fractures. Journal of Bone and Joint Surgery (British) 1991;73(3):511–3. 14. Woodfield JC, Beshay NM, Pettigrew RA, Plank LD, van Rij AM. American society of anesthesiologists classification of physical status as a predictor of wound infection. ANZ Journal of Surgery 2007;77(9):738–41. 15. Konrad G, Markmiller M, Lenich A, Mayr E, Ruter A. Tourniquets may increase postoperative swelling and pain after internal fixation of ankle fractures. Clinical Orthopaedics and Related Research 2005;443:189–94. 16. Lindenhovious AL, Helmerhorts GT, Schnellen AC. Differences in prescription of narcotic pain medication after operative treatment of hip and ankle fractures in the United States and The Netherlands. Journal of Trauma Injury Infection and Critical Care 2009;67:160–4. 17. Jacobs DG, Sarafin JL, Norton J, Christmas AB, Huynh T. Wasted hospital days impair the value of length of stay variables in the quality assessment of trauma care. The American Surgeon 2009;75(9):794–803. 18. Thangarajah T, Prasad PS, Narayan B. Surgical site infections following open reduction and internal fixation of ankle fractures. The Open Orthopaedics Journal 2009;3:56–60. 19. SooHoo NF, Krenek L, Eagen MJ, Gurbani B, Ko CY, Zingmond DS. Complication rates following open reduction and internal fixation of ankle fractures. Journal of Bone and Joint Surgery (American) 2009;A91(5):1042–9. 20. McCormack RG, Leith JM. Ankle fractures in diabetics. Journal of Bone and Joint Surgery (British) 1998;B80:392–689. 21. Blotter RH, Connolly E, Wasan A, Chapman MW. Acute complications in the operative treatment of isolated ankle fractures in patients with diabetes mellitus. Foot and Ankle International 1999;20(11):687–94. 22. Wilson AP, Gibbons C, Reeves BC, Hodgson B, Liu M, Plummer D. Surgical wound infection as a performance indicator: agreement of common definitions of wound infection in 4773 patients. British Medical Journal 2004. 10.1136/ bmj.38232.646227. 23. Bruce J, Russell EM, Mollison J, Krukowski Z. The quality of measurement of surgical wound infection as the basis for monitoring: a systematic review. Journal of Hospital Infection 2001;49:99–108. 24. Peterson EJ, Irish SM, Lyons CL. Reliability of water volumetry and the figure of eight method on subjects with ankle joint swelling. Journal of Orthopaedics Sports and Physical Therapy 1999;29:609–15.