Risk of venous thromboembolism in patients undergoing gastric cancer surgery: a systematic review and meta-analysis

Despite the varying incidence of VTE among cancers, current guidelines on preventive methods are not well stratified by cancer type. This study found that the overall 1-month incidence rate of VTE, DVT, and PE after GC surgery was low. These results are consistent with those of a previous meta-analysis of postoperative symptomatic VTE in abdominal and pelvic tumours [37]. However, we only focused on GC and did not limit VTE presentation (symptomatic or asymptomatic). Most current guidelines do not describe VTE prevention after GC surgery, and recommendations for VTE prophylaxis are not specified according to the types of cancer and surgery [11, 12, 15, 38, 39]. Considering that the reported incidence of postoperative VTE varies widely among pelvic and abdominal tumours (e.g., 1.9% for colorectal cancer and as high as 8% for ovarian cancer) [40, 41], it is necessary to understand the accurate incidence of this complication in patients undergoing GC surgery.

Although the 1-month incidence of VTE (1.8%) and DVT (1.2%) after GC surgery is low, the 7-day incidence of postoperative VTE (8.8%) and DVT (8.8%) is high. Such a high rate of VTE may be due to the routine postoperative imaging screening in both studies with a 7-day follow-up; all patients diagnosed with VTE were asymptomatic [31, 34]. Although routine imaging screening can reveal asymptomatic VTE, it is not recommended because of the cost and limited accuracy of imaging examination in the diagnosis of asymptomatic VTE [42]. Asymptomatic VTE events are very common in cancer patients, and the associated mortality is not lower than that of symptomatic VTE events [43‐45]. In the current study, analysis stratified according to VTE diagnostic methods showed that the 1-month incidence of postoperative DVT was significantly higher in the routine DUS group than in the non-routine DUS group, which indicated that the incidence of asymptomatic DVT was higher than that of symptomatic DVT. Considering that PE is mainly a complication of DVT, and PE accompanying DVT accounts for 80.6% of the total incidence of PE [46], clinicians have to pay more attention to PE due to its potential lethality. At present, routine imaging screening is rarely performed in the clinic and is only used to confirm the diagnosis when patients have relevant clinical manifestations, which may mean that many asymptomatic VTEs are missed. Therefore, whether postoperative imaging screening should be routinely performed remains controversial.

Many previous studies have shown that the incidence of VTE in Asian populations is significantly lower than that in Western populations [47‐52]. Surprisingly, in our study, the pooled 1-month incidence of VTE after GC surgery was similar between Asian and Western countries. In addition, although the rate of postoperative DVT was higher in Asia than in the West, the difference was not statistically significant. Meanwhile, the incidence of PE after GC surgery was significantly lower in Asian countries than in Western countries. Sakon et al. reported that the PE rate after general surgery was lower in Japan than in the West (0.33% vs. 1.6%) [53], and consistent findings of a difference between Asian and Western populations were found in the current study. Currently guidelines, Western guidelines put more emphasis on pharmacological thromboprophylaxis, while Asian guidelines do not. Our study found that the 1-month incidence of VTE in patients undergoing GC surgery in both Asian and Western populations was not high at 1.8%, which suggested that there may be futile treatment in the Western guidelines. However, among the included Western studies, none reported specific methods for VTE prevention. Considering that drug prophylaxis in patients undergoing cancer surgery is used routinely in Western countries but rarely in Asian countries, this consistency between the VTE rates in the East and West may only be a facade. Therefore, this result should be interpreted cautiously, and more complete data need to be obtained for further analyses.

The appropriate strategy for pharmacological thromboprophylaxis in patients undergoing cancer surgery also remains unclear to date. Anticoagulant use for patients undergoing cancer surgery is routine practice in Western countries, and the debate is mostly focused on drug selection and the duration of prophylaxis. The latest ITAC and ESMO guidelines recommended the highest prophylactic dose of LMWH once per day for patients undergoing major cancer surgery, starting 2–12 h preoperatively and extending to 4 weeks after surgery [54, 55]. In the guidelines from the Japanese Circulation Society, patients over the age of 40 undergoing major cancer surgery were considered to be at high risk of VTE, and anticoagulation therapy was recommended as a preventive treatment for patients undergoing abdominal surgery [56]. The choice of anticoagulants includes enoxaparin, fondaparinux, or low-dose unfractionated heparin. The medication regimen was as follows: starting 24 h postoperatively, enoxaparin was administered subcutaneously twice daily at a dose of 2000U with a treatment duration of ≤ 2 weeks, or fondaparinux subcutaneously once daily at a dose of 2.5 mg with a treatment duration of ≤ 8 days. However, in Korean guidelines for patients undergoing GC surgery, patients aged < 60 years were defined to have very low risk, and only early ambulation was recommended; meanwhile patients aged ≥ 60 years were defined to have low risk, and mechanical prevention was recommended [57]. The latest Asian venous thromboembolism guidelines recommend LMWH for VTE prevention in patients undergoing cancer surgery, but lack details such as drug dosage [15]. But in fact, most doctors in Japan and Korea still prefer mechanical prophylaxis over pharmacological prophylaxis. Currently, guidelines from the Asian region are still fewer and slower to update compared with Western countries. In the final analysis, the crux of the controversy is postoperative bleeding, the most common side effect of pharmacological prophylaxis in cancer surgery. A previous RCT showed that mechanical prophylaxis combined with fondaparinux was associated with significantly lower incidence of postoperative VTE than was mechanical prophylaxis alone in patients undergoing abdominal surgery, but there was more severe postoperative bleeding [58]. Similarly, another study on abdominal tumour surgery also reported that although the 1-month incidence of postoperative VTE in patients who received in-hospital drug prevention before and after surgery was as low as 0.35%, up to 42% of the patients had major bleeding [59]. In patients undergoing GC surgery, Joe et al. found that although the use of LMWH brought significant benefits, it also significantly increased the risk of bleeding [60]. In our study, mechanical prophylaxis combined with anticoagulants was associated with a significantly lower risk of VTE within 1 month after GC surgery than was mechanical thromboprophylaxis alone. However, postoperative bleeding events were also significantly higher in the study by Jung 2018 [28]. Interestingly, all the included studies comparing between mechanical prophylaxis alone and mechanical plus pharmacological prophylaxis were from Japan and Korea. This may reflect the scepticism of Asian surgeons about the benefits of chemoprophylaxis. In addition, only one of the included studies reported postoperative bleeding events with different preventive measures, so more studies are needed to provide reliable data.

Only few studies have focused on the long-term VTE risk after GC surgery, and most related studies have focused on VTE events within 1 month. An earlier prospective cohort study of national populations found that although the VTE risk peaked at approximately 3 weeks postoperative, it is still high at 12 weeks postoperative [61]. In addition, in two analyses of the national readmission database, the readmission rate related to VTE after GC surgery was remained high over a long period [33, 62]. In the current analysis, only 3 of the 13 studies reported the long-term risk of VTE after GC surgery. Despite the limited number of included studies, it still showed an increasing trend in the rate of VTE from the time after GC surgery: 1 m, 1.8%; 90d, 2.4%; 24 m, 2.0%; 36 m, 3.7% for VTE; 1 m, 1.2%, 90d, 1.4%; and 36 m, 3.7% for DVT; and 1 m, 0.4%; 90d, 1.0%, and 36 m, 0% for PE. However, the association between GC surgery and long-term VTE risk is still unconvincing, because the long-term risk may be due to other causes, such as tumour recurrence, rather than the cancer surgery itself.

Laparoscopic surgery has become widely used in cancer surgery; however, although this surgical approach has the advantages of minimal trauma and rapid postoperative recovery, its impact on VTE is still controversial. One study suggested that the relatively longer duration of laparoscopic surgery and the compression caused by the pneumoperitoneum might increase postoperative VTE events [63]. However, some studies found that the risk of VTE after laparoscopic surgery was comparable to that after open surgery [34, 59]. The results of two studies using large clinical databases supported that laparoscopic surgery has a significantly lower incidence of postoperative VTE than has open surgery [64, 65]. In the current analysis, the incidence of postoperative VTE was higher after open surgery than after laparoscopic surgery, but the difference was not statistically significant. Few studies have compared VTE risk after different surgical procedures for GC, and thus, more studies are needed to further evaluate this risk in the future.

Among the 13 studies included in the current analysis, 6 studies involved 5 national or provincial databases. The incidence of postoperative VTE after GC surgery was 0.89–2.98% from Western databases [24, 25, 27, 33, 35]. Meanwhile, the incidence of VTE reported from only one Asian database was 0.47% [36]. Although some of the data reported from these studies, such as those on thromboprophylaxis methods, are incomplete, they are still of great reference for real-world data. Eastern and Western countries showed significant differences in both the main sites of GC and the treatment methods. Proximal GC is common in Western populations, whereas distal GC is more prevalent in Asia [66]. In addition, gastrectomy with D1 lymphadenectomy is mostly used in the West, and attention is paid to perioperative chemoradiation; meanwhile, D2 lymphadenectomy is most commonly performed for GC in Asian countries [67, 68]. Chemoradiation is an important risk factor for postoperative VTE in cancer [69, 70], and this may be one of the reasons for the clinical use of drug prophylaxis in the West. In contrast, the more aggressive surgical procedures for GC in Asia than in Western countries may explain why Asian surgeons are more worried about postoperative bleeding caused by anticoagulants. Therefore, these differences may have contributed to the disagreement between the East and West on pharmacological thromboprophylaxis for patients undergoing GC surgery.

Our study determined the postoperative incidence of VTE in patients undergoing GC surgery, providing a reference for clinicians to formulate preventive measures. This meta-analysis has several strengths. First, the analyses only considered studies whose primary outcomes included VTE events to ensure reliability of the results. Second, the studies were required to include a clear follow-up period, and the VTE data were classified and analysed according to similar lengths of follow-up. Third, the risk of postoperative VTE and bleeding complications with pharmacological thromboprophylaxis was evaluated, although this was observed only in Asian populations.

Despite its strengths, this study also has some limitations. There was substantial heterogeneity among the included studies, which might have resulted from the different methods used to identify VTE, lack of information about prevention methods, and unknown dosage of LMWH, etc. In addition, the included studies were only from six countries, and most of the studies were from the USA, Japan, and Korea, which may affect the representativeness of the estimates. Therefore, further research is needed in other Eastern and Western countries. Moreover, as none of the included studies from Western countries described the use of anticoagulation and postoperative bleeding complications, the benefit of chemoprophylaxis in Western populations could not be assessed. More prospective studies are needed to optimise the risk stratification of patients and develop more personalised treatment plans to achieve the optimal balance between the risk of VTE and bleeding prevention. Finally, because the demographic information provided by most of the included studies only reported pooled data for multiple tumours (e.g. esophagogastric, gastrointestinal tract, and pelvic/abdominal cancers) and did not provide biological data for GC alone, there were insufficient data to further investigate the biological factors underlying the differences in VTE incidence between Asian and Western populations. Therefore, more detailed data are needed to elucidate the biological factors that may cause the difference in VTE rates.

In conclusion, the 1-month incidence of VTE in patients undergoing GC surgery is not very high. The incidence rate of VTE after GC surgery is comparable between Eastern and Western patients; however, evidence, such as for thromboprophylaxis measures, in different regions is still lacking. Additionally, pharmacological prophylaxis significantly increases the risk of postoperative bleeding complications after GC surgery in Asian populations, although VTE events are also notably reduced. Compared with open GC surgery, laparoscopic GC surgery does not increase or reduce the risk of postoperative VTE.