Hospital Episode Statistics data analysis of postoperative venous thromboembolus in patients undergoing urological surgery: a review of 126,891 cases

Introduction Current guidelines on venous thromboembolism (VTE) prevention do not reflect the potential varying risk for patients undergoing different urological procedures. Our study aimed to establish the procedure specific rate of postoperative VTE in patients undergoing urological surgery. Methods Hospital Episode Statistics were obtained for all patients undergoing common urological procedures between April 2009 and April 2010. This cohort was followed up to identify all patients reattending with either deep vein thrombosis (DVT) or pulmonary embolism (PE) within 12 months. Results A total of 126,891 individuals underwent urological surgery during the study period. This included 89,628 men (70.6%) and 37,236 women (29.3%) with a mean age of 65.2 years. At the 12-month follow-up, 839 patients (0.66%) were readmitted with VTE. Of these, 373 (0.29%) were admitted with DVT and 466 (0.37%) with PE. The procedure-specific rate of VTE varied significantly between 2.86% following cystectomy and 0.23% following urethral dilatation. Procedures performed in the lithotomy position carried a significantly lower risk of VTE than those performed in the supine position (0.60% vs 1.28%, p<0.0001). Furthermore, of all procedures performed in the lithotomy position, those performed on benign conditions carried a significantly lower risk than those performed on malignant disease (0.52% vs 0.79%, p<0.0001). Conclusions Procedure specific rates of postoperative VTE vary widely among patients undergoing urological procedures. These findings suggest the potential benefit of prolonging the use of thromboprophylaxis in high-risk patients but also exploring the apparent lack of need for routine thromboprophylaxis in patients undergoing low-risk procedures.

The House of Commons Health Committee 2005 report on venous thromboembolism (VTE) documents the significant problems associated with hospital-acquired VTE. 1 It estimates the annual incidence of fatal hospital-acquired pulmonary embolism (PE) at 25,000 annually and highlights the substantial morbidity associated with deep vein thrombosis (DVT). Consequently, in 2010 the National Institute for Health and Clinical Excellence (NICE) published guidance for the care of hospitalised patients at risk of developing VTE. 2 These guidelines detail the suggested mechanical, medical and surgical measures that should be employed in hospitalised patients suffering with medical illness and undergoing a range of surgical procedures.
Clearly, both the risk of VTE and the consequence of postoperative bleeding differ between different surgical procedures. This is reflected in the NICE guidance with different recommendations for surgery in different specialties (orthopaedic, ophthalmology, gastrointestinal etc). 2 There is, however, no procedure-specific variation in guidance for patients undergoing urological surgery. This is particularly important in urological surgery as there is wide variation in surgical techniques (open, endoscopic and laparoscopic), in patient positioning (supine, prone and lithotomy) and in diseases (benign and malignant). Furthermore, much of the data on which conclusions for current guidelines are based date from the 1970s and 1980s. [3][4][5][6][7] By interrogating the National Health Service (NHS) Hospital Episode Statistics (HES) database, our study aimed to elucidate the rate of procedure-specific postoperative VTE in patients undergoing a range of urological surgery.

DYER WYKE LYNCH
HOspitAL episOde stAtistics dAtA AnALYsis Of pOstOpeRAtive venOUs tHROmBOemBOLUs in pAtients UndeRGOinG UROLOGicAL sURGeRY: A Review Of 126,891 cAses methods HES data were obtained for all patients undergoing common urological procedures in NHS trusts throughout England between April 2009 and April 2010. The codes and groupings are given in Table 1. This dataset was employed as it was the most recently available complete dataset that permitted a minimum follow-up period of 12 months. The cohort was followed to identify those patients reattending with either DVT or PE within 12 months of their surgery. Patients were stratified according to sex, procedure and operative position. Statistical significance was determined using a chi-squared test. In order to preserve patient confidentiality, data were not included in subsequent analysis if the output from HES was fewer than six in any given category.   There was no significant difference between the incidence of VTE between men and women (0.67% vs 0.63%, p=0.413).
Procedure-specific VTE rates varied considerably (Fig 1,  Table 2) with major open and laparoscopic surgery carrying the highest risk of VTE. Of this group, cystectomy carried the highest risk with an incidence of 2.86%. Benign cystoscopic procedures and urethral dilatation carried the lowest risk of postoperative VTE with 0.48% and 0.23% respectively. Procedures performed in the lithotomy position were associated with approximately half the risk of developing VTE than those performed in the supine position (0.60% vs 1.28%, p<0.0001) ( Table 3). Furthermore, there was a reduction in VTE risk when comparing cystoscopic surgery on benign and malignant conditions (0.52% vs 0.79%, p<0.0001) ( Table 4).

discussion
As expected, procedure-specific rates of symptomatic VTE varied greatly among patients undergoing urological surgery. Cystectomy appears to carry the highest risk of VTE, which is greater than the published incidence for known high-risk procedures such as total hip arthroplasty (2.4%), partial hip arthroplasty (2.0%) and embolectomy of lower limb artery (2.8%). 8 Current guidelines for major urological, gastrointestinal and gynaecological pelvic cancer surgery suggest continuing medical thromboprophylaxis following discharge for 28 days. 2 Although the assumption from our findings would be to support a prolonged duration of postoperative thromboprophylaxis, it is necessary to reiterate that our study indicates annual incidence of VTE rather than the incidence in a different specific time period. There is a need for further clarification on the VTE risk for such patients, particularly in terms of the time to development of symptomatic VTE. Such knowledge will aid the development of recommendations for patients undergoing a cystectomy with regard to method, duration and dosage of thromboprophylaxis. Conversely, our study showed that urethral dilatation and cystoscopic procedures performed in the lithotomy position for A systematic review of community-based epidemiological studies estimated the annual incidence of VTE to be around 0.1%. 9 However, beyond the age of 40, there seems to be an exponential increase in VTE incidence, rising to 0.3-0.5% annually in patients aged between 70 and 79 years.
Unfortunately, there are inherent problems in using HES for the purposes of this study. It is not possible to control for of the use of thromboprophylaxis, pre-existing VTE risk factors (age, previous VTE, immobility, etc) or procedures performed under general or regional anaesthesia. Despite laparoscopic surgery being the established gold standard for a variety of procedures, HES do not account for different surgical techniques (open, laparoscopic or robotic). Furthermore, there is reliance on accurate clinical coding and also inevitable underestimation of VTE incidence given that some patients will be diagnosed with VTE out of the hospital setting.
Nevertheless, HES have been shown to be increasingly accurate and an effective tool in comparative audit. 10-12 HES will no doubt continue to be used as a research tool in the future. The introduction of a greater number of codes for urological procedures performed with differing techniques would increase the usefulness of analogous analysis and is a recommendation of our study.
Despite its limitations, this study is the first large multicentre study of postoperative VTE across a range of urological procedures. Comparing VTE incidence in our study with community-based control studies reveals a group of procedures that carry a VTE risk similar to and often less than the risk of developing a symptomatic VTE in the absence of any surgery. This calls into question the role for routine thromboprophylaxis in patients undergoing low-risk urological procedures.
The current guidance draws its conclusions from a number of single-centre randomised controlled trials conducted in the 1970s and 80s. The incidence of symptomatic VTE in those patients not receiving thromboprophylaxis was as high as 10% for DVT and 9% for PE following both transvesical and transurethral resection of prostate. [3][4][5][6][7] We suggest that in the last 30 years there has been shift towards shorter operating times as well as earlier mobilisation and discharge from hospital. Despite increasing use of mechanical and medical thromboprophylaxis, these factors are highly likely to contribute to our study findings. The combination of dated evidence and apparently low VTE incidence in our study suggest value in conducting further randomised controlled trials to delineate the health benefit and economical implications of routine thromboprophylaxis, particularly for patients undergoing low-risk surgery.
conclusions It would appear that there is scope for improvement and streamlining of care with regard to VTE prevention for patients undergoing both high-risk and low-risk urological surgery. Given the limitations of our study, alterations in local practice should await further investigation.