Journal Club : The PROTECT study

Neil Orford on 27-03-2011

The PROTECT trial is another incredible ANZICS CTG effort, in conjunction with the Canadian CTG. This 3764 participant, double blinded, multicentre, international, randomised controlled trial, compared dalteparin to unfractionated heparin in adult ICU patients (exclusions included orthopaedic, neurosurgery, major trauma) with proximal DVT detected by twice weekly compression ultrasound the primary outcome. The interesting points are; 1. No difference in primary outcome, with overall incidence of proximal DVT 5.6% 2. Most VTE occurs in the ICU, not the ward 3. A reduction in PE in the dalteparin group (HR 0.51, 95% CI 0.30-.88) NNT=100. PE was sought on clinical suspicion, not screening, so it is possible that some bias exists, although the mechanism by which 1 group would be tested more for PE is unclear 4. No difference in bleeding 5. No difference in HITs, although the per-protocol analysis reported a decrease in the dalteparin group (HR 0.27, 95% CI 0.08-.98). This should be interpreted with caution Questions 1. Is the reduction in PE and possibly thrombocytopenia, combined with safety profile, enough to justify switch to LMWH? 2. Does it simply come down to a cost-benefit analysis?

2 Comments

Got something to say?

Log in to your account

Forgotten your password?

Todd Fraser wrote 03-30-2011 06:42:17 pm
I'm having a little trouble rationalising the difference in PE rates with the lack of difference in proximal DVT, or CVC-related DVT. The mechanism by which this could occur is a bit obscure.

I'm also curious why the study chose DVT as the end-point rather than DVT. Presumably its because the incidence of DVT is thought higher than PE, and thus would be easier to prove a difference, when in fact, the opposite occurred.

The finding that only about 5% of patients in ICU have DVTs when anticoagulated is certainly reassuring, and far lower than expected. The power calculation was based on an expected incidence of 8% - perhaps this underpowers the trial somewhat. Major bleeding, while not different between the groups, is quite high, in fact, as high as the rate of DVT.

The suggestion that other drugs in this class may have different effects is frustrating, particularly given many recommendations use enoxaparin as the LMWH of choice.

Alex McKenzie wrote 07-02-2011 10:50:00 am
The difference in PE rates is interesting. There was almost a 50% reduction in PEs, which is nothing to sneeze at. I guess its the mechanism that is puzzling - the rate of PE fell though there was no fall in leg DVT rates. Is this because DVTs, when they do form, are less friable and likely to embolise? Or are the significant PEs coming from other sites not screened for by the protocol (upper limb, CVC etc)? If the latter is true, why would the drug affect these and not peripheral DVTs?

Or is this just statistical anomoly?