Early mobilisation after conventional knee
replacement may reduce the risk of
postoperative venous thromboembolism
- E. O. Pearse, MA, FRCS(Orth), Specialist Registrar in Orthopaedics, Clinical Knee Fellow1;
- B. F. Caldwell, FRACS(Orth), Orthopaedic Surgeon1;
- R. J. Lockwood, BHlth Sc(Nursing), RN Surgical Nursing Unit Manager2; and
- J. Hollard, BMed(Hons), FANZCA, Consultant Anaesthetist3
Abstract
We carried out an audit on the result of achieving early walking in total knee replacement after instituting a new rehabilitation protocol, and assessed its influence on the development of deep-vein thrombosis as determined by Doppler ultrasound scanning on the fifth post-operative day. Early mobilisation was defined as beginning to walk less than 24 hours after knee replacement.
Between April 1997 and July 2002, 98 patients underwent a total of 125 total knee replacements. They began walking on the second post-operative day unless there was a medical contraindication. They formed a retrospective control group. A protocol which allowed patients to start walking at less than 24 hours after surgery was instituted in August 2002. Between August 2002 and November 2004, 97 patients underwent a total of 122 total knee replacements. They formed the early mobilisation group, in which data were prospectively gathered. The two groups were of similar age, gender and had similar medical comorbidities. The surgical technique and tourniquet times were similar and the same instrumentation was used in nearly all cases. All the patients received low-molecular-weight heparin thromboprophylaxis and wore compression stockings post-operatively.
In the early mobilisation group 90 patients (92.8%) began walking successfully within 24 hours of their operation. The incidence of deep-vein thrombosis fell from 27.6% in the control group to 1.0% in the early mobilisation group (chi-squared test, p < 0.001). There was a difference in the incidence of risk factors for deep-vein thrombosis between the two groups. However, multiple logistic regression analysis showed that the institution of an early mobilisation protocol resulted in a 30-fold reduction in the risk of post-operative deep-vein thrombosis when we adjusted for other risk factors.
The reported incidence of venous thromboembolism after total knee replacement (TKR) ranges from 30% to 60%.1 Chemical and mechanical prophylaxis regimes are often used, either in isolation or in combination, to reduce the risk of post-operative venous thromboembolism. Several studies have shown that mechanical devices such as foot pumps,2 continuous-passive movement devices3 and intermittent pneumatic compression4 increase venous blood flow and reduce the incidence of post-operative deep-vein thrombosis (DVT).5 These devices are used to enhance the natural mechanism for venous return in the lower limb during the immediate post-operative period when patients have a reduced mobility mainly because of pain. Some studies have suggested that patients who undergo minimally-invasive TKR are able to walk sooner than those who undergo conventional knee replacement.6,7 This has been attributed to reduced soft-tissue trauma and post-operative pain. However, Berend, Lombardi and Mallory8 suggested that rapid recovery after knee replacement may be less related to the size of the incision and more to the pre-operative education of the patient, peri-operative nutrition, pre-emptive analgesia and postoperative rehabilitation.
To our knowledge, there is no accepted definition of early mobilisation. We defined it as beginning to walk less than 24 hours after knee replacement. Our hypothesis was that early mobilisation could be achieved after conventional knee replacement if an adequate protocol was instituted, and that this could reduce the incidence of post-operative venous thromboembolism. We tested the hypothesis in a prospective outcome study using a historical control group of patients.
Patients and Methods
All patients who underwent TKR performed by a single surgeon (BFC) between 1997 and 2004 at a single institution (Toronto Private Hospital, Toronto, New South Wales) were included in the study.
Control group.
Between April 1997 and July 2002, 98 patients, 48 men and 50 women, with a mean age of 68.8 years (49 to 87) routinely began to walk on the second postoperative day unless there was a medical contraindication. Synchronous bilateral TKR had been undertaken in 27, giving a total of 125 knee replacements performed. These patients were used as a historical control group. Clinical data for this group were collected retrospectively from inpatient medical, nursing and physiotherapy records.
Early mobilisation group.
The senior author (BFC), his anaesthetist (JH) and the senior nurse on the surgical ward (RJL) developed a protocol (Table I⇓), the aim of which was to allow patients to begin walking independently less than 24 hours after surgery. They walked to the bathroom for a shower and then walked approximately 30 yards from the high-dependency bay to their individual rooms. Gradually they increased the distances walked throughout the day and were encouraged to be increasingly independent. They increased the distances walked on subsequent days. This protocol was instituted at the start of August 2002 and patients operated on between August 2002 and November 2004 were prospectively studied as the early mobilisation group.
Table I.
Details of the protocol instituted to achieve early mobilisation
There were 97 patients (54 men and 43 women) with a mean age of 68.7 years (42 to 85). Synchronous bilateral TKR had been undertaken in 25, giving a total of 122 knee replacements performed.
Operative treatment.
Under tourniquet control at an inflation pressure of 300 mmHg, a standardised surgical approach was followed, using a midline incision approximately 16 cm long and a medial parapatellar approach with eversion of the patella. Intramedullary alignment was used for the femoral cutting jig and rotational orientation was determined with reference to the inter-epicondylar axis. External alignment was used for the tibial cutting jig.
Most operations were undertaken using the NexGen prosthesis (Zimmer, Warsaw, Indiana) which was used either as a cemented or uncemented implant (Table II⇓). A suction drain was used, and was removed the morning after surgery. In all patients who had undergone bilateral synchronous replacement, blood loss was managed by routine transfusion with two units of heterologous blood. Haemoglobin levels were checked 24 hours after the operation and the threshold for transfusion was < 10g/dl in a symptomatic patient. All the patients received subcutaneous low-molecular-weight heparin (LMWH) (Clexane; Aventis Pharma Pty Ltd, New South Wales, Australia) and compression stockings for thromboprophylaxis. The first dose of heparin (40 mg subcutaneously) was administered before surgery, the second on the evening of surgery and subsequent doses were administered daily. An experienced consultant radiologist examined all patients for DVT using Doppler ultrasound on the fifth post-operative day. He was not aware of our study. Any suspected pulmonary emboli were confirmed by spiral CT.
Table II.
Clinical details of the early mobilisation (122 knees) and control (125 knees) groups
Throughout the period of study all patients received instructions on physiotherapy using active and passive movements on the first post-operative day.
Statistical analysis.
Data were analysed using SPSS version 10 (SPSS Inc., Chicago, Illinois). Continuous data were plotted as a histogram to determine whether they were normally distributed. Normally distributed data were analysed using the independent samples Student’s t-test, and data which were not normally distributed by the Mann-Whitney U test. Nominal data were analysed using the chi-squared test.
We included the following established risk factors for DVT in our analysis: increasing age (> 60 years), history of previous venous thromboembolism, obesity (body mass index of ≥ 33 kg/m2 for females and ≤ 32.5 kg/m2 for males) evidence of venous stasis and hormone replacement therapy.9Smoking was not included as a risk factor due to lack of clarity on this subject.
Potential influences on the risk of DVT were: spinal anaesthesia (potentially protective because of an increase in blood flow), synchronous bilateral case (potential risk because of increased operating time), female gender (presence of oestrogens), walking within 24 hours, and cemented fixation.
All identified potential risk factors for DVT were included simultaneously as explanatory variables in a logistic regression analysis using the presence or absence of DVT as the outcome. The exponential values of the logistic regression coefficients provided the odds ratios, giving an indication of the quantitative effect of each factor (adjusted for the remaining factors) on the risk of developing post-operative DVT. Differences were considered to be significant if p < 0.05.
Results
There were no significant differences between the two groups in regard to gender, age, the number of synchronous bilateral cases and the number of significant medical conditions (cardiovascular, respiratory or endocrine disease). There were more cemented knee replacements in the early mobilisation group than in the control group (Table I⇑) because of a change in surgical preference. The tourniquet time was similar in both groups.
In the early mobilisation group, 90 patients (92.8%) began walking less than 24 hours after their operation, while in the control group, 20 (20.4%) did so (Table III⇓). This difference was statistically significant (chi-squared test, p < 0.001). In the early mobilisation group, 89 of 90 patients who successfully began walking less than 24 hours after their operation progressively increased their amount of walking through the first post-operative day and on subsequent days. In the control group, 11 of the 20 patients who began walking less than 24 hours after their operations had a reduction in their walking ability because of nausea, syncope or poorly-controlled pain.
Table III.
Details of potential risks for deep-vein thrombosis (DVT) in patients in both groups
A similar proportion of patients in both groups used patient-controlled analgesia pumps in the first 48 postoperative hours, 59 (60.8%) in the early mobilisation group and 52 (53%) in the control group (chi-squared test, p = 0.274). Of those who used patient-controlled analgesia pumps, there was a significant difference in the amount of morphine required between the two groups. A median of 19 mg (1 to 590) was used by those in the early mobilisation group compared with median of 61.5 mg (4 to 1135) in the control group (Mann-Whitney U test, p < 0.001). Intramuscular injections of morphine were used for supplementary pain control when a patient-controlled analgesia device was not used or after it had been removed. When compared with the early mobilisation group, a greater proportion of patients in the control group required at least one intramuscular injection of morphine for pain, 62 (63.2%) patients in the control group compared with 22 (22.6%) in the early mobilisation group (chi-squared test, p < 0.001).
There were fewer episodes of syncope because of postural hypotension in the early mobilisation group (3 (3.1%) compared with 11 (11.2%)). In the control group, 97 units of blood were transfused compared with only 67 units in the early mobilisation group. There were more minor wound problems consisting of serosanguinous ooze and erythema in the early mobilisation group (23 (23.7%) compared with 11 (11.2%)), all of which settled within 72 hours without antibiotic treatment. There were no wound dehiscences or deep infections in either group.
Overall, 28 DVTs were identified after 247 knee replacements in the 195 patients, of which 25 (89%) were distal calf and three (11%) were proximal. They were identified in one patient (1.0%) in the early mobilisation group and in 27 patients (27.6%) in the control group (chi-squared test, p < 0.001). The three proximal DVTs, as well as two non-fatal pulmonary emboli occurred in the control group.
There were several significant differences between the two groups which included some risk factors for post-operative venous thromboembolism: history of previous venous thromboembolism, obesity, venous stasis, hormone replacement therapy, spinal anaesthesia, cemented fixation (Table III⇑). The quantitative effect of each factor on the risk of developing a post-operative DVT was determined using multiple logistic regression analysis (Table IV⇓). All identified potential risk factors for DVT were included simultaneously as explanatory variables and the presence or absence of DVT was the outcome. When we adjusted for other risk factors the risk of developing DVT if patients were in the early mobilisation group was approximately 30 times lower than if they were in the control group (odds ratio 0.036, p = 0.002).
Table IV.
Logistic regression analysis to determine the quantitative effect of the early mobilisation protocol (adjusted for the remaining factors) on the risk of developing a post-operative deep-vein thrombosis (DVT)
We performed a secondary logistic regression analysis to determine the effect of early mobilisation itself (Table V⇓). When other risk factors were adjusted for, patients who began walking within 24 hours were over three times less likely to develop a DVT than those who began walking after 24 hours (odds ratio 0.295, p = 0.031).
Table V.
Logistic regression analysis to determine the effect of early mobilisation per se on the risk of developing a post-operative deep-vein thrombosis (DVT)
Post-operative DVT occurred in six patients who began walking less than 24 hours after their operations. Of these, five were in the control group and experienced a decline in their walking ability because of syncope. The patient who developed a post-operative DVT in the early mobilisation group was the only patient to experience a decline in his walking ability caused by pain. This patient had polycythaemia rubra vera.
Discussion
Our results show that early mobilisation can be achieved after conventional TKR when a multidisciplinary protocol involving the surgeon, anaesthetist, nursing staff and physiotherapist is instituted. In the early mobilisation group 92.8% of patients were able to walk less than 24 hours after their knee replacement. Patients in the early mobilisation group used their patient-controlled analgesia less and required less morphine than those in the control group, which suggested that they did not have more pain. As well as being associated with the use of adequate oral analgesia and anti-inflammatory medication, this may also be related to psychological preparation. There is good evidence that patients behave differently as a result of their interpretation and understanding of their bodily symptoms.10 A number of studies11,12 have shown the efficacy of cognitive behavioural therapies in improving pain and physical disability in patients with various pain conditions.
Patients in the early mobilisation group had fewer episodes of syncope suggesting that this programme was safe. A number of features of the protocol (Table I⇑) contributed to this outcome. First, the administration of two units of blood to patients who had synchronous bilateral TKRs on the day of operation was important, and secondly, changes in the nursing care were instituted with emphasis on the measured postural blood pressure before walking and encouragment of adequate hydration and a good urinary output. The patient’s standard antihypertensive medication was delayed or omitted in response to the blood pressure readings at an early stage. Patients who underwent synchronous bilateral TKRs had sufficient muscular control to walk using either a frame or two crutches without further assistance.
The incidence of DVT detected by venous Doppler ultrasound on the fifth post-operative day was 27.6% in the control group. Westrich et al13 reported a similar prevalence of 29% in patients who underwent knee replacement and received thromboprophylaxis with low-molecular-weight heparin. In our study, the incidence of post-operative DVT fell from 27.6% to 1.0% after the institution of an early mobilisation protocol, a difference which was highly significant.
The drawbacks of the use of radiological rather than clinical indicators to determine the incidence of postoperative DVT and the efficacy of thromboprophylaxis have been highlighted.14 Most DVTs diagnosed by Doppler ultrasound are confined to the calf, are clinically silent and remain so without any adverse consequences.9 Kakkar et al15 found no cases of symptomatic pulmonary embolism in 31 patients in whom DVT had resolved spontaneously or remained confined to the calf. The post-thrombotic syndrome does not occur any more frequently in patients with a history of asymptomatic DVT. Five years after hip or knee arthroplasty, its prevalence is the same (approximately 5%) whether patients have isolated calf DVT, proximal DVT or no DVT.16 However, it has been reported that 10% to 20% of calf thrombi extend to the proximal veins.1 Most symptomatic DVTs start in the calf veins: 99% of patients with proximal DVT have associated calf-vein thrombosis and these have been shown to be continuous in over 90%.17Consequently, a reduction in the prevalence of detected DVT by Doppler scanning can be assumed to be a reasonable goal of thromboprophylaxis, but efficacy, cost, patient acceptance and adverse effects must be considered.
We have shown that early mobilisation can be achieved after conventional knee replacement, is effective in reducing the incidence of radiologically detected DVT, requires no additional costly equipment, has a high level of patient acceptance and a low incidence of associated morbidity.
Our protocol was effective in initiating walking in patients less than 24 hours after their operation and ensuring that there was a progression in their walking ability beyond this initial period. Although 20 patients walked early in the control group, more than half of these had a decline in their walking ability because of nausea, syncope or poorly controlled pain. The five patients who developed DVTs despite starting to walk less than 24 hours after their operations all subsequently had a decline in their walking ability.
The major limitation of our study was its retrospective nature and the fact that it was not a randomised trial with stratification for risk factors. Both groups of patients were similar in many respects including gender, age and co-existing medical conditions. A single surgeon (BFC) performed all the operations using the same surgical technique and implants in most replacements. However, there was a greater incidence of risk factors for DVT in the control group (Table III⇑), and this difference was statistically significant. This may have biased our results in favour of the early mobilisation group. We performed multiple logistic regression analysis to determine which factors had a significant independent bearing on the chances of developing a post-operative DVT detected by venous Doppler ultrasound examination. The only factor which had a statistically significant effect on the risk of DVT, reducing the risk by a factor of 30, was whether patients were in the early mobilisation group (Table IV⇑). Walking within 24 hours of surgery was associated with a threefold reduction in risk of DVT (Table V⇑).
The result of logistic regression analysis must be interpreted with caution. First, our analysis cannot adjust for unidentified risk factors which could have occurred with different frequencies in the two groups thus having a confounding action. Secondly, it would not make sense to discount established risk factors for DVT simply because their effect on the risk of developing post-operative DVT was not statistically significant in our analysis. Logistic regression is sensitive to the number of explanatory variables and the number of patients in the study. It may simply be that we did not have enough patients in our study for the influence of several risk factors to achieve statistical significance.
Footnotes
- We wish to thank the Toronto Private Hospital Surgical Team, Toronto Physiotherapy and Toronto Imaging, Toronto and The Australia Institute of Musculoskeletal Studies, Sydney, New South Wales, Australia.
- No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.
- Received May 31, 2006.
- Accepted December 1, 2006.
- © 2007 British Editorial Society of Bone and Joint Surgery
Nenhum comentário:
Postar um comentário