Introduction
The influence of unfractionated heparin (UFH) and other anticoagulants on the spread of cancer has been reported since the early 1960s.1 However, clinical studies investigating the use of heparins in cancer patients have not produced consistent results.2 Intravenous, adjusted-dose UFH for 5 to 10 days has been the standard initial treatment for venous thrombosis. More recently, subcutaneous, fixed-dose, low molecular weight heparins (LMWHs), which are fractions of the parent compound, have been shown to be safe and effective alternatives to UFH in the initial treatment for venous thromboembolism.3-5
In one of our randomized clinical trials comparing LMWH and UFH in the initial treatment of deep vein thrombosis (DVT), we observed an unexpected difference in 6-month mortality among cancer patients in favor of LMWH, which could not be attributed to a difference in the incidence of thrombotic or bleeding complications.6 A similar observation in favor of LMWH was reported in a subsequent study and in a meta-analysis of trials.7,8 The number of cancer patients included in these studies was small, and adjustment of the observed effect for the baseline characteristics of the cancer patients was not possible. However, these findings suggested an inhibitory effect of LMWH on tumor growth or metastasis, which is less apparent or absent for UFH, resulting in a beneficial effect on the survival of cancer patients. This hypothesis is supported by the observations, in experimental studies, that LMWH and low molecular weight heparan sulfate, in comparison to UFH, effectively suppressed angiogenesis, a process necessary for tumor growth and metastasis.9,10
On the other hand, animal studies that investigated the effect of chemically-modified heparins on the spread of cancer did not detect a superior anti-tumor effect of LMWH compared to UFH; both were found to inhibit metastasis.11,12 To date, the effect of LMWH on cancer survival in humans has not been investigated as a primary objective. If a consistent and beneficial effect of LMWH on mortality is indeed present, such a study would be warranted.
We performed a meta-analysis of all available randomized clinical trials where LMWH was compared with UFH in the treatment of venous thromboembolism (VTE) to estimate the crude treatment effect of LMWH on mortality in cancer patients compared to UFH. Subsequently, we adjusted this treatment effect for age, gender, and primary malignancy site by reanalyzing data from three of those trials.3-5 This effect was further adjusted for other prognostic factors, including cancer histology, tumor stage, presence of metastases, duration of cancer, and concomitant use of cancer treatment, by analyzing individual patient data from the largest randomized trial.5
