We read with interest the report by De Stefano et al (2006) regarding the risk of recurrent venous thromboembolism in pregnancy and puerperium. This comprehensive study showed that prevalence of thrombophilia did not differ significantly in the cohort of pregnant women with or without deep-vein thrombosis (DVT) and the presence of a provoking factor for previous DVT did not increase the risk for DVT in subsequent pregnancies. Between 2002 and 2006, we studied all the pregnancy-related DVT cases that presented at two of the busiest hospitals in Mumbai. Out of 25 132 multigravida women examined, 32 had DVT during their current pregnancy, based on the clinical features that were confirmed by Doppler studies. Clinically inapparent cases (absence of oedema, redness, pain and calf tenderness) were not investigated for subclinical DVT. Nine out of 462 women with previous history of fetal loss had DVT whereas 23 out of 24 670 with no history of previous fetal loss and no other precipitating risk factor, such as toxaemia, dehydration or smoking, had a pregnancy-related DVT in the current pregnancy. Hence, the risk of clinically apparent DVT during pregnancy was 22-fold higher if there was a previous history of one or more pregnancy loss (chi-squared test P < 0·001). Laboratory investigations for thrombophilia, i.e. lupus antibodies, IgG/IgM antibodies for anticardiolipin antibodies, β2glycoprotein, annexin V, protein C, protein S, antithrombin III, factor V Leiden, F2 (PT gene), MTHFR C677T, FBG (fibrinogen gene Arg/Lys), PROCR (also known as EPCR) 23 bp insertion and SERPINE1 (also known as PAI) 4G/5G polymorphisms, did not show any significant difference in their prevalence in these two groups P > 0·05), as was reported by De Stefano et al (2006). In conclusion, we believe that previous pregnancy loss should be considered as a significant risk factor for DVT in subsequent pregnancies. The risk might have been significantly underestimated in our study as we only identified clinically significant DVT and did not investigate the prevalence of DVT in the puerperium period, which is reported to be several times higher than that of the antenatal period (Andersen et al, 1998; Simpson et al, 2001).