Incidence of delayed venous thromboembolic events in patients undergoing abdominal and pelvic surgery for cancer: a systematic review and meta-analysis

Background Incidence of venous thromboembolism (VTE) following discharge for abdominal cancer surgery is uncertain. Methods We searched MEDLINE and Embase for studies evaluating the incidence of VTE at 3 months from surgery. Studies indicating use of post-hospital VTE prophylaxis were excluded. Two independent reviewers performed study selection, data abstraction and risk of bias. Random-effects model was used to estimate pooled incidence, and weights were estimated using inverse variance method. Statistical heterogeneity was explored via subgroup analysis. Results Of 4215 abstracts retrieved, 11 reported the incidence of VTE at 3 months. There were three randomized trials (n = 520), one prospective cohort study (n = 284) and seven retrospective cohort studies (n = 65 308). VTE incidence among prospective studies was 9.6% (95% confidence interval (CI) 2.9-16.4), while for retrospective studies was 2.2% (95% CI 1.4-3.0). Heterogeneity was high (I-2 = 92% and 81%, respectively). The incidence of symptomatic VTE was 1.3% (95% CI 0.4-2.3) for prospective studies. VTE was diagnosed by screening venography in most of the prospective studies, whereas retrospective studies did not use a screening method. Subgroup analysis based on the type of organ surgery performed explained the heterogeneity. Conclusions VTE incidence following abdominal cancer surgery varies greatly depending on the study type, with differences largely explained by the method of assessment of VTE. The fact that VTE incidence among retrospective studies was closer to the incidence of symptomatic events (non-screen detected) in the prospective studies, suggests that the screened events were mostly asymptomatic and their clinical significance is unclear.