Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

A question of stockings

Systematic review

How long should your stockings be? Not a frivolous question this, but a serious issue about whether graduated compression stockings should be thigh length, or whether we can get away with stockings just up to the knee to prevent deep venous thrombosis (DVT). A somewhat difficult (either confused, or confusing, depending on point of view) systematic review [1] provides some evidence, and a lesson in when not to combine dissimilar data.

Systematic review

It looked for randomised trials on the use of knee versus thigh length graduated compression stockings for thromboprophylaxis in surgical patients, and the use of knee length stockings versus no stockings in long haul flight passengers. The trials had to use objective tests for deep venous thrombosis interpreted blind to allocation and with predefined criteria for an abnormal test, and investigate DVT above or below the knee.


In surgical patients, five trials with 592 patients showed no difference between knee and thigh length stockings (Figure 1, Table 1). As Figure 1 shows, the individual trials had widely varying event rates. In different trial arms the percentage of DVTs ranged from 2% to 65%. We don't know very much about these patients, except that in three trials they underwent general surgery, and in two they were orthopaedic patients (two small trials with higher event rates).

Figure 1: Individual trials in surgical patients

Table 1: Results of a comparison of knee length graduated compression stockings versus thigh length stockings in surgical patients using, versus no stockings in long haul flight passengers

Number of
DVT (%) with
Patient group
Relative risk
(95% CI)
Surgical patients
1.1 (0.7 to 1.6)
Long haul flight passengers
0.1 (0.05 to 0.26)
In surgical patients, control was thigh length compression stockings. In long haul passengers, the control was no stockings

In long haul flight individuals the effect of using knee length graduated compression stockings was considerable (Figure 2, Table 1). Only 0.2% of people using stockings had a DVT, compared with 3.7% in those without stockings. This was a highly significant reduction in risk, corresponding to a number needed to treat with knee length graduated compression stockings of 28 (95% CI 21 to 42) to prevent a DVT in one of them. Again, we are told nothing about these participants without reading the original trials.

Figure 2: Individual trials in long haul flight passengers


What can we make of all this? Bandolier would make no hard judgements without a more detailed look at the evidence because of obvious inconsistencies and mistakes in the paper. For knee versus thigh length graduated compression stockings the lack of a statistically significant result in this limited trial set probably rules out any large difference, but does not exclude a small one. Whether a small difference would be clinically significant or overcome any practical or cost differences depends on circumstances.

The reduction in DVT incidence compared with no stockings comes from a larger data set, and with a big effect. For people at higher risk on longer flights, it certainly makes the use of these stockings worth considering. It would be better if we knew the exact definition of higher risk, so we could judge for ourselves.


1 MS Sajid et al. Knee versus thigh length graduated compression stockings for prevention of deep venous thrombosis: a systematic review. European Journal of Endovascular Surgery 2006: doi:10.1016/j.ejvs.2006.06.012