Global practice variation in pharmacologicthromboprophylaxis for general and gynaecologicalsurgery: systematic review

risk of bleeding 2 . Expert recommendations regarding VTE prophylaxis in surgery vary 3 , but the extent of practice remains


Dear Editor
Venous thromboembolism (VTE) and bleeding are serious complications of surgery. Pharmacological prophylaxis decreases VTE but increases bleeding 1 . The decision to use thromboprophylaxis requires balancing decreasing VTE versus increased risk of bleeding 2 . Expert recommendations regarding VTE prophylaxis in surgery vary 3 , but the extent of practice variation remains uncertain.
We performed comprehensive literature searches in Embase, MEDLINE, Web of Science, and Google Scholar for observational studies with procedure-specific information on VTE and/or bleeding for 16 general abdominal and 22 gynaecological surgery procedures until November 2020 3 (PROSPERO CRD42021234119) (Supplementary material). Two reviewers independently assessed eligibility and extracted data using standardized, piloted data forms, guided by written instructions. We included surgical procedures that had been investigated in at least five studies, with the majority of participants enrolled from 2000 onwards (Supplementary material). For each study, we extracted data on the proportion of patients receiving pharmacological prophylaxis, and the duration of prophylaxis. For each procedure, we calculated the proportion of discretionary use of pharmacological prophylaxis, and the mean or median duration of prophylaxis.
All 50 studies reported whether prophylaxis was used; 29 of 46 (63 per cent) studies in which prophylaxis was used, also reported the duration of use.
The proportion receiving prophylaxis varied widely in laparoscopic cholecystectomy, open groin hernia repair, and open liver resection. The duration of pharmacological prophylaxis varied between 2 and 3 days after laparoscopic cholecystectomy, between 6 and 27 days after open liver resection, and only one study reported the duration for hernia repair (Fig. 1).
Studies of laparoscopic cholecystectomy, open hernia repair, and open liver resection reported large variation in the use of pharmacological prophylaxis (Fig. 1). Studies in cancer and obesity surgery consistently reported high rates of use of pharmacological prophylaxis, but with substantial variation in the duration of prophylaxis.
Earlier studies have addressed pharmacological prophylaxis in one or more centres 4,5 , but our study examines variation in practice across studies. The other strengths of this review include a thorough search of contemporary studies, application of explicit eligibility criteria, and standardized piloted data forms for data collection.
This article has limitations. Although we screened many potential studies, only a small proportion proved eligible (1.4 per cent of full texts screened). These 50 studies (of which 15 had low risk of bias, 19 had moderate risk, and 16 had high risk of bias; Tables S3 and S4) published between 2000 and 2020 represent seven procedures each with 5-10 studies.
We identified substantial practice variation, within and between countries, in the use of pharmacological prophylaxis in most types of benign surgery, and in the duration of prophylaxis after cancer surgery. Rationalization of practice will require evidence that provides a better understanding of procedure-specific risks of VTE and bleeding as well as creation of procedure-specific, evidence-based guidelines for thromboprophylaxis. Rationalization of current practice would decrease both under-and overuse of thromboprophylaxis improving patient outcomes.

Funding
This study was supported by the Academy of Finland (309387 and 340957), Sigrid Jusélius Foundation and Competitive Research Funding of the Helsinki University Hospital (TYH2019321, TYH2020248, and TYH2022330). The sponsors had no role in the analysis and interpretation of the data or the manuscript preparation, review, or approval.

Fig. 1 Proportion (in percentages) of patients with reported use of pharmacological prophylaxis
Depending on the study, the mean or median duration of pharmacological prophylaxis (in days) for those who received any is noted on each bar (if reported in the article).