Challenging the evidence for graduated compression stockingsBMJ 2013; 346 doi: https://doi.org/10.1136/bmj.f3653 (Published 11 June 2013) Cite this as: BMJ 2013;346:f3653
All rapid responses
As authors of the 2010 Cochrane review (1) on the use of elastic compression stockings for prevention on deep vein thrombosis (DVT), we would like to respond to comments by Whittaker et al (2) regarding this review. It should be noted that NICE guidance (3) on the prevention of venous thromboembolism (CG92, published in 2009) cited a previous version of this review (4).
Whittaker et al highlighted the exclusion of two large trials (5, 6) from the updated review, noting that one including only patients with acute stroke (5) was “too specific” and another trial including orthopaedic patients was “too pragmatic”(6).
Stockings reduce the cross-sectional area of the deep veins, making the calf muscle pump more effectively and thereby improve blood flow. Authors of the CLOTS trial have suggested that severe leg weakness in patients with acute stroke may therefore account for the ineffectiveness of stockings in this patient group (5). Trials on patients with acute stroke were thus excluded from this review but included in a separate Cochrane review focussing on patients with acute stroke (7).
The trial on orthopaedic patients (6) was excluded as it was unclear whether a standardised protocol was used throughout this study. We noted a number of discrepancies in this trial (such as variation in the use of imaging modalities and how an “objective” diagnosis of symptomatic DVTs was made), and felt it was unclear whether all patients or only selected patients were routinely scanned for DVT. We had attempted to contact the authors of this trial to seek further information, but received no response.
2. Unit of randomisation
Whittaker et al also reported potential bias by the application of stockings to one leg in 6 of 18 included trials, while the other leg acted as a control. The authors suggested the possibility of impeding blood flow in the both legs, citing the Spiro et al 1970 study (8), as also noted in our review. However, Spiro et al used a method of intermittent compression, rather than graduated compression, to assess the haemodynamic effects on the contralateral leg (8). Therefore, it is unclear whether graduated compression stockings (GCS) have a similar effect on the contralateral leg (9). Additionally, potential biases can occur in studies where the randomisation unit is “the leg” rather than the patient (10).
Nevertheless, we have now addressed this potential source of bias by performing a sensitivity analysis to assess the influence of the unit of randomisation (i.e. an individual leg vs. an individual patient). The use of stockings as the sole method of thromboprophylaxis (Figure 1) is favoured both in trials randomising individual legs (OR 0.22, 95% CI 0.11-0.43, p < 0.001), and also in trials randomising individual patients (OR 0.39, 95% CI 0.27-0.56, p < 0.001). A similar effect was noted when stockings were used over a background method of thromboprophylaxis (Figure 2), irrespective of whether individual legs (OR 0.15, 95% CI 0.06-0.36, p< 0.001) or individual patients were randomised (OR 0.24, 95% CI 0.14-0.43, p< 0.001). Bias associated with the unit of randomisation in this review is, therefore, unlikely.
3. Symptomatic/asymptomatic DVTs
The authors also questioned the use of radiologically diagnosed asymptomatic DVTs, citing their potential limited clinical significance, as the incidence of symptomatic DVTs was not reported in our review. The predefined primary outcome of our review was the “diagnosis of DVT identified by ultrasound, venogram or isotope studies”. Therefore, the use of an objective diagnostic test to confirm the presence of DVT was an inclusion criterion for this review. Only two of 18 included trials reported the presence of clinical signs of DVT, and none reported the incidence of symptomatic DVTs.
Furthermore, asymptomatic proximal DVTs have been shown to be clinically significant, being associated with higher mortality compared to patients without DVT (11). We, therefore, assessed included trials for incidence of radiologically diagnosed asymptomatic proximal DVT. Eight of 18 trials have reported the incidence of proximal DVTs – 2 of 8 trials using GCS as sole thromboprophylaxis and 6 of 10 trials using GCS over background thromboprophylaxis. Fewer proximal DVTs were diagnosed amongst patients fitted with GCS in both groups of trials (Figures 3 & 4), though the difference was only statistically significant in trials using GCS over a background method of thromboprophylaxis (OR 0.30, 95% CI 0.16-0.58, p <0.001).
Furthermore, we would like to highlight that significantly fewer PEs were diagnosed when GCS was used over background thromboprophylaxis (Analysis 3.2 in the review, p 0.048), though only 2 of the 3 studies included in this analysis routinely assessed all participants for PE, and scintigraphy was only used for confirmation for patients with clinical manifestations of PE in one additional study.
4. Funding from stocking manufacturers
Whittaker et al highlight support from manufacturers of stockings for 10 of 14 trials, as reported in our review. Five included trials were fully or partially sponsored by the manufacturers, whereas manufacturers supplied stockings for an additional 5 trials. However, none of the trials reported whether the manufacturers participated in trial design, data collection or analysis, or preparation of the manuscript.
5. Funnel plots
We have sought statistical advice from the Cochrane Peripheral Vascular Disease group, and can confirm that our use of two separate funnel plots is appropriate, as we have analysed the data using two separate comparisons (‘GCS as the sole method of thromboprophylaxis’ and ‘GCS over a background method of thromboprophylaxis’). Nevertheless, we acknowledge the author’s concern that there may be a slight possibility of publication bias on visual inspection of the funnel plots. We are currently in the process of updating the review, and this issue will be addressed in the update.
Nevertheless, as noted in our review, we emphasise that all but one included trial (12) was based on surgical patients. Therefore, there remains paucity of evidence to support the use of stockings in medical patients, though we have good evidence favouring the use of stockings in surgical patients.
1. Sachdeva, A., Dalton, M., Amaragiri, S. V., & Lees, T. (2010). Elastic compression stockings for prevention of deep vein thrombosis. Cochrane database of systematic reviews, (7), CD001484.
2. Whittaker, L., Baglin, T., & Vuylsteke, A. (2013). Challenging the evidence for graduated compression stockings. BMJ (Clinical research ed), 346(jun11 1), f3653–f3653.
3. National Institute for Health and Care Excellence. Venous thromboembolism: reducing the risk of venous thromboembolism (deep vein thrombosis and pulmonary embolism) in patients admitted to hospital. CG92. London: National Institute for Health and Clinical Excellence; 2010
4. Amaragiri, S. V., & Lees, T. A. (2000). Elastic compression stockings for prevention of deep vein thrombosis. Cochrane database of systematic reviews, (3), CD001484.
5. CLOTS Trials Collaboration, Dennis, M., Sandercock, P. A. G., Reid, J., Graham, C., Murray, G., et al. (2009). Effectiveness of thigh-length graduated compression stockings to reduce the risk of deep vein thrombosis after stroke (CLOTS trial 1): a multicentre, randomised controlled trial. Lancet, 373(9679), 1958–1965.
6. Cohen, A. T., Skinner, J. A., Warwick, D., & Brenkel, I. (2007). The use of graduated compression stockings in association with fondaparinux in surgery of the hip. A multicentre, multinational, randomised, open-label, parallel-group comparative study. The Journal of bone and joint surgery British volume, 89(7), 887–892.
7. Naccarato, M., Chiodo Grandi, F., Dennis, M., & Sandercock, P. A. (2010). Physical methods for preventing deep vein thrombosis in stroke. Cochrane database of systematic reviews, (8), CD001922.
8. Spiro, M., Roberts, V. C., & Richards, J. B. (1970). Effect of externally applied pressure on femoral vein blood flow. BMJ, 1(5698), 719–723.
9. Bergqvist, D., & Lindblad, B. (1984). The thromboprophylactic effect of graded elastic compression stockings in combination with dextran 70. Archives of surgery (Chicago, Ill: 1960), 119(11), 1329–1331.
10. Nesbitt, C. I., & Stansby, G. (2011). Regarding “Prospective randomized trial comparing endovenous laser ablation and surgery for treatment of primary great saphenous varicose veins with a 2-year follow-up.” Journal of Vascular Surgery, 53(5), 1456.
11. Vaitkus, P. T., Leizorovicz, A., Cohen, A. T., Turpie, A. G. G., Olsson, C.-G., Goldhaber, S. Z., & PREVENT Medical Thromboprophylaxis Study Group. (2005). Mortality rates and risk factors for asymptomatic deep vein thrombosis in medical patients. Thrombosis and haemostasis, 93(1), 76–79.
12. Kierkegaard, A., & Norgren, L. (1993). Graduated compression stockings in the prevention of deep vein thrombosis in patients with acute myocardial infarction. European Heart Journal, 14(10), 1365–1368.
Sensitivity analysis for unit of randomisation
Figure 1: GCS as sole thromboprophylaxis
Figure 2: GCS over background thromboprophylaxis
Incidence of proximal DVT
Figure 3: GCS as sole thromboprophylaxis
Figure 4: GCS over background thromboprophylaxis
Competing interests: AS, MJD, SVA & TAL are authors of the 2010 Cochrane review.
Is extended anticoagulation after knee replacement necessary in 2013?
Whittaker et al challenge the evidence for compression stockings in preventing
symptomatic venous thromboembolism (VTE). There may also be grounds to
question the benefits of prophylactic anticoagulation.1
Current National Institute for Health and Care Excellence (NICE)
recommendations advocate chemical prophylaxis for 10‐14 days following
elective total knee replacement (TKR), a procedure considered by many as high
risk for VTE. 2
The most significant consequence of VTE following TKR is fatal pulmonary
embolism (PE). Prior to the introduction of anticoagulants, the fatal PE rate was
0.1‐0.43%.3 The introduction of chemical thromboprophylaxis has not clearly
decreased the prevalence of PE with little advantage in using aspirin or
Early papers documented a 40‐60% prevalence of deep vein thrombosis (DVT)
following TKR. Concern existed about the connection between DVT and PE. But
their association is not strong. And surveys focused on asymptomatic DVTs, most
of which resolve spontaneously. Also, post‐thrombotic limb ulceration has not
featured in the literature.6 Current literature suggests a symptomatic DVT rate of
0.6% following TKR, using either aspirin or heparin.
Seeing is believing. Orthopaedic surgeons review patients for two or more years
after TKR .From a personal series of more than 800 knee replacements,
rehabilitated with next‐day walking and injected anticoagulants only whilst in
hospital, I am aware of 6 DVTs and 3 PEs, none fatal. Perhaps the answer lies in
rapid mobilisation. One study suggests that delaying walking by only one day
increases risks of VTE 30‐fold. 7
Post‐operative bleeding is more likely with anticoagulation, undermines wound
healing and risks deep infection, particularly in superficial joints such as the
knee.5 And shorter hospital stays will increase potential for evolving wound
problems to be missed.
The current default to provide extended thromboprophylaxis has a shaky
evidence base. Knee replacement is no longer strongly associated with DVT nor
As with compression stockings, perhaps NICE should review its guidelines.
1. Whittaker L, Baglin T, Vuylsteke A. Challenging the evidence for
graduated compression stockings. BMJ 2013;346:f3653.
2. National Institute for Health and Care Excellence. Venous
thromboembolism: reducing the risks of venous thromboembolism (deep
vein thrombosis and pulmonary embolism) in patients admitted to
hospital. CG92. 2010. http://guidance.nice.org.uk/CG92.
3. Warwick DJ, Whitehouse S. Symptomatic venous thromboembolism after
total knee replacement. J Bone Joint Surg Br. Vol 79‐B, No. 5 September
4. Cusick LA, Beverland DE. The incidence of fatal pulmonary embolism after
primary hip and knee replacement in a consecutive series of 4253
patients. J Bone Joint Surg Br. Vol 91‐B, No 5, May 2009. 645‐648.
5. Jameson SS, Baker PN, Charman SC, Deehan DJ, Reed MR et al. The effect
of aspirin and low molecular weight heparin on venous
thromboembolism after knee replacement. A non‐randomised
comparison using National Joint Registry Data. J Bone Joint Surg Br. Vol
94‐B, No 7, July 2012. 914‐918.
6. Wang C.‐J, Wang J.W, Weng L.‐H, Hsu C.‐C, Lo C.‐F. Outcome of calf deepvein
thrombosis after total knee arthroplasty. J Bone Joint Surg Br. Vol 85‐
B, No 6, August 2003. 841‐844.
7. Pearse EO, Caldwell BF, Lockwood RJ, Hollard J. Early mobilisation after
conventional knee replacement may reduce the risk of post‐operative
venous thromboembolism. J Bone Joint Surg Br. Vol 89‐B, No 3, March
Mr Jai Chitnavis
Consultant Orthopaedic Surgeon,
The Cambridge Knee Clinic.
Cambridge University Hospitals NHS Trust
Competing interests: No competing interests