br Characteristics br Female br Surgical Site br Upper
Variables with a p value less than 0.05 (see Table 2) in addition to sex and age were included. OR: odds ratio.
a Upper gastrointestinal (GI): esophagus, stomach, and duodenum.
b Lower GI: intestine, colon, rectum, cecum, and anus.
c Hepato-pancreato-biliary (HPB) tumor: liver, pancreas, and gallbladder.
d Others included 93 breast cancer, 15 gynecologic cancer, 5 genitourinary cancer, and 9 other cancer types.
transfusion (p Z 0.054). The AOR of transfusion in patients with streptokinase/streptodornase administration before surgery was 6.0 compared to those not receiving this medication. The area under receiver operating character-istic curve (AUROC) value for intraoperative transfusion was 0.779 (95% CI 0.687e0.872, p < 0.001) (Fig. 1).
CGA was originally developed by geriatricians as a multi-domain evaluation for older patients. CGA comprises functional status, comorbidity, polypharmacy, cognition, psychological status, social support, and nutritional status. It has been reported that CGA is a fundamental aid for evaluation and treatment planning of older cancer pa-tients. Ample evidence supports the value of CGA for pre-dicting overall morbidity and mortality in geriatric cancer patients.26e28 It has become a predictor for intra- and postoperative complications, medical treatment, cogni-tion, geriatric syndromes, and extrinsic frailty in elderly cancer patients.29e33
Unlike long-acting anticoagulants, which showed signif-icant effects on intraoperative transfusion, antiplatelet drugs did not affect blood transfusion, regardless of their half-lives or duration of action. The CGA was performed within 7 days prior to surgery; therefore, we assumed that patients included in the medication-intake group received the antithrombotics during the 7 days preceding surgery. According to a previous study, patients who stopped taking
aspirin 3e7 days preoperatively had little or no increased requirement for blood transfusion.34,35 Although the surgi-
cal bleeding risks associated with clopidogrel, another an-tiplatelet drug, have not been completely characterized, some experts suggest that this medication has a safety profile similar to that of aspirin in this setting.36 In contrast, bleeding complications have been shown to be increased in patients taking warfarin. Currently, it AZD-5991 is recommended to discontinue warfarin 5 days before surgery, as it takes 5 days for the INR to normalize after stopping warfarin.37
We found that blood transfusion was clearly associated with low BMI. Similar results of other studies suggest that
low BMI is a common risk factor for intraoperative trans-fusion.38,39 Low BMI has been associated with an increased risk of excessive blood loss. Especially in heart surgery, low BMI has been shown to be a risk for reoperation because of excessive bleeding.40 However, for hip and knee arthro-plasty, BMI was not found to be associated with blood loss in either hip- or knee-replacement patients.41,42 The
Please cite this article as: Jeong YM et al., Association between preoperative use of antithrombotic medications and intraoperative transfusion in older patients undergoing cancer surgery, Asian Journal of Surgery, https://doi.org/10.1016/j.asjsur.2019.06.005
Antithrombotics and intraoperative transfusion 7
discrepancy is attributable to the different surgery types. It is not surprising that a low BMI was associated with intra-operative blood transfusion in our study, because many patients with cancer experience significant weight loss before surgery due to poor oral intake and malnourishment.
HPB cancer was found to be associated with transfusion in our study. Other studies have reported that HPB surgery
is among the most common factors associated with trans-fusion.43,44 The liver is a highly vascular organ, and sub-
stantial blood loss is common during liver surgery.45e47 Therefore, patients undergoing liver resection may be at increased risk for excessive blood loss and a subsequent need for blood transfusion.
The study is limited by its retrospective, single-center design. Nevertheless, this study demonstrates the value of a CGA team-directed comprehensive medication assess-ment using the most current evidence-based screening tools to detect medication-related risk factors of intra-operative transfusion. Especially, it was recommended that long-acting anticoagulants be withdrawn at least seven days before surgery.
The results of this study show that medication use in geri-atric cancer patients has an impact on transfusion during surgery; in particular, the preoperative use of long-acting anticoagulants is associated with an increased risk of intraoperative transfusion. This finding also demonstrates the value of medication screening, as a component of a multidisciplinary comprehensive geriatric assessment, for reducing the risks of medication-related complications in the perioperative period.