The risk of venous thrombosis after air travel: contribution of clinical risk factors

Long distance air travel is associated with an increased risk of venous thrombosis (VT). A meta-analysis (Chandra et al, 2009) of 14 studies concerning 4055 VT cases demonstrated that the overall pooled relative risk for air travel is 2·8 (95% confidence interval [CI] 2·2–3·7) after exclusion of studies with referred control participants. The absolute risk of developing symptomatic VT within 8 weeks of a long-haul flight is 1 in 4500 flights (Kuipers et al, 2007). This absolute risk is not high enough to promote the widespread use of potentially harmful preventive measures, such as anticoagulant therapy. However, for some subgroups of travellers at highly increased risk, the risk-benefit ratio may favour the use of preventive measures. The objective of the current study among frequently flying employees of international companies and organizations was to assess the effect of transient classical risk factors for VT on the occurrence of air travel-related VT. We performed a cohort study of frequently travelling employees of large international companies and organizations. The occurrence of VT was related to exposure to air travel. The design of the study has previously been described in detail (Kuipers et al, 2007). All participating employees received a web-based questionnaire to determine occurrence of VT and the prevalence of risk factors, such as surgery, plaster-cast, pregnancy, delivery, hormone use (hormone replacement therapy or oral contraceptives) and active cancer. To assess the effect of these risk factors for VT, we performed a nested case–control analysis within the cohort. We compared exposure to these risk factors as well as to at least one long haul flight, using flight data, for a 3 month period before the event occurred for the VT cases, and a random period of 3 months for the subjects without VT. All odds ratios (ORs) were adjusted for age and sex using logistic regression. Only symptomatic first episodes of VT, diagnosed by compression ultrasonography or venography, were considered. Pulmonary embolism had to have been diagnosed by spiral-CT scanning, high probability ventilation-perfusion scanning or angiography. A total of 8755 employees of eight different companies or organizations participated. More than half of the responders (n = 4915, 56%) were men and the mean age was 40 years. Of these 8755 employees, 1163 participated in a pilot study and were excluded from the current analysis, because the questions regarding transient risk factors for VT were added after this pilot study. The remaining 7592 employees were followed for a total of 33 279 person years (py), with a mean follow-up of 4·4 years. In this period, 44 employees were diagnosed with VT, of which 18 occurred within 3 months of a long haul flight. All ORs are shown in Table 1. The risk of VT was increased by almost 20-fold in travellers who had undergone recent surgery, compared with subjects who had been exposed to neither (OR 19·8 [95% CI 5·6–70·1]). There was only one VT patient who suffered from cancer and had travelled by air in the 3 months before the diagnosis of VT, which resulted in an 18-fold increased risk (OR 18·0, 95% CI 2·2–148·7). The combination with hormone use led to a seven to eightfold increased risk. Three women with thrombosis were pregnant in the 3 months before their diagnosis of VT. The OR for both pregnancy and air travel was 14·3 (95% CI 1·7–121·0). The risk of air travel in subjects who also have transient risk factors for VT has not been studied extensively before. Two previous studies only found a synergistically increased risk for the combination of oral contraceptive use and long distance travel (Martinelli et al, 2003; Cannegieter et al, 2006; Kuipers et al, 2009). A limitation of our study is the relatively small number of cases with a combination of air travel and the risk factor of interest. Therefore, the confidence intervals were wide. Nevertheless, we found consistently increased risks for almost all classical risk factors for VT, which appeared to double in combination with air travel. Another possible limitation of our study is that, due to the nested case–control design, it was only possible to calculate ORs. However, one could estimate absolute risks using the numbers in our study. Assuming that the percentage of employees having surgery during the 3-month period in question is representative for the complete follow-up period, it can be calculated that the estimated absolute risk is 1/164 flights for travellers with recent surgery, 1/141 flights for those with a malignant disease and 1/140 flights for those with a plaster cast. The estimated absolute risk for women travelling by air who used oral contraceptives is 1/259 flights, 1/405 flights for women using HRT and 1/109 flights for those who were pregnant. From this nested-case control study, we conclude that the risk of travel-related VT is most increased in travellers who have recently undergone surgery, those with a malignant disease and pregnant women. In these populations, preventive measures might be considered. Large randomized controlled studies in travellers at increased risk are needed to assess which preventive measure is most effective and least harmful. We thank all participating companies and organizations for their help in conducting this study. This study was funded by grant number 2002B53 from the Netherlands Heart Foundation and sponsored by the UK government and the European Commission. Study concept and design: SC, HB, and FR. Acquisition of data: SK, SC. Analysis and interpretation of data: SK, SM, SC. Drafting of the manuscript: SC, AV. Final approval of the version to be published: SK, AV, SM, HB, SC, FR.