AIM: To compare venous thromboembolism (VTE) in hospitalized ulcerative colitis (UC) patients who respond to medical management to patients requiring colectomy.

METHODS: Population-based surveillance from 1997 to 2009 was used to identify all adults admitted to hospital for a flare of UC and those patients who underwent colectomy. All medical charts were reviewed to confirm the diagnosis and extract clinically relevant information. UC patients were stratified by: (1) responsive to inpatient medical therapy (n = 382); (2) medically refractory requiring emergent colectomy (n = 309); and (3) elective colectomy (n = 329). The primary outcome was the development of VTE during hospitalization or within 6 mo of discharge. Heparin prophylaxis to prevent VTE was assessed. Logistic regression analysis determined the effect of disease course (i.e., responsive to medical therapy, medically refractory, and elective colectomy) on VTE after adjusting for confounders including age, sex, smoking, disease activity, comorbidities, extent of disease, and IBD medications (i.e., corticosteroids, mesalamine, azathioprine, and infliximab). Point estimates were presented as odds ratios (OR) with 95%CI.

RESULTS: The prevalence of VTE among patients with UC who responded to medical therapy was 1.3% and only 16% of these patients received heparin prophylaxis. In contrast, VTE was higher among patients who underwent an emergent (8.7%) and elective (4.9%) colectomy, despite greater than 90% of patients receiving postoperative heparin prophylaxis. The most common site of VTE was intra-abdominal (45.8%) followed by lower extremity (19.6%). VTE was diagnosed after discharge from hospital in 16.7% of cases. Elective (adjusted OR = 3.69; 95%CI: 1.30-10.44) and emergent colectomy (adjusted OR = 5.28; 95%CI: 1.93-14.45) were significant risk factors for VTE as compared to medically responsive UC patients. Furthermore, the odds of a VTE significantly increased across time (adjusted OR = 1.10; 95%CI: 1.01-1.20). Age, sex, comorbidities, disease extent, disease activity, smoking, corticosteroids, mesalamine, azathioprine, and infliximab were not independently associated with the development of VTE.

CONCLUSION: VTE was associated with colectomy, particularly, among UC patients who failed medical management. VTE prophylaxis may not be sufficient to prevent VTE in patients undergoing colectomy.

In the western countries the prevalence of ulcerative colitis (UC) has been reported as high as 500 cases per 100000 persons[1]. Approximately, 16% of UC patients will require colectomy within the first 10 years of diagnosis; though, colectomy rates have been decreasing overtime[2,3]. Colectomy for UC has been associated with postoperative morbidity and mortality[4-6]. An important contributor to postoperative complications is the occurrence of venous thromboembolism (VTE)[7,8], VTE carries risk of significant morbidity and mortality, particularly due to pulmonary embolism[9-11].

The inflammatory bowel diseases (IBD) have been established as an independent risk factor for developing VTE as well as recurrent VTE[12,13]. In recent years, several population-based studies showed that IBD patients are at increased risk of developing VTE[7,14]. Bernstein et al[7] demonstrated that IBD patients have a 3-4 fold increased risk of developing a VTE requiring hospitalization. Similar rates have been described in both ambulatory and hospitalized IBD patients[12,15], while specific studies focusing on hospitalized IBD patients showed an increased risk of 1.5-2 fold compared to hospitalized non-IBD patients[16]. In a meta-analysis the risk of VTE among IBD patients was 2-fold higher than individuals without IBD[17].

Studies have also highlighted some of the potential risk factors behind the development of VTE in IBD. One study demonstrated that IBD patients experiencing an acute flare that required steroid use were associated with an over 8-fold increased risk of VTE when compared to non-IBD patients[15]. The risk of VTE in IBD was driven by disease severity with the risk steadily increasing across IBD patients in remission, in a flare managed as an outpatient, and a flare managed in-hospital[18]. Among hospitalized IBD patients risk factors for VTE included UC, advanced age, IBD-related surgery, malnutrition, and medical co-morbidities[16,19]. Further, the prevalence of VTE among asymptomatic flaring IBD patients was low, which suggests that in-hospital factors drive the development of VTE[20]. Surgery has been recognized as an important risk factor for VTE among IBD patients[21,22]. However, studies are needed to explore the impact of VTE in UC patients who respond to in-hospital medical management as compared to those who are refractory to medical management and require colectomy.

Guidelines recommend prophylaxis for VTE using anti-coagulants (e.g., subcutaneous heparin) in IBD patients who have been hospitalized for a flare of disease and following surgery[23,24]. While the use of VTE prophylactic doses of heparin in UC is recognized as safe[25,26], utilization of VTE prophylaxis for hospitalized UC patients in clinical practice may be suboptimal[27,28]. In a tertiary care referral center, nearly half of UC patients admitted to medical service were not provided VTE prophylaxis as compared to over 90% who were prescribed in a surgical service[27]. However, population-based studies that include VTE prescribing practices in community and academic hospitals have not been explored.

We designed a large population-based study to determine the rate of VTE among UC patients who underwent an emergent or elective colectomy as compared to those who responded to in-hospital medical management. Furthermore, we explored heparin utilization in clinical practice and the efficacy of VTE prophylaxis in this population.

Between January 1^st 1997 and December 31^st 2009, 638 individuals with UC who underwent a colectomy (309 emergent and 329 elective) were identified. A further 382 UC patients were admitted emergently with a flare, but were discharged after responding to in-hospital medical management. The characteristics of patients with UC are presented in Table 1. The overall VTE rate was 4.7%, and 18.8% of these individuals experienced a pulmonary embolism. The most common site of VTE was intra-abdominal (45.8%) followed by lower extremity (19.6%). Table 2 describes the location of VTE.

VTE occurred in 1.3% of medically responsive, 8.7% of emergent colectomy, and 4.9% of elective colectomy UC patients (Table 1). Among the emergent colectomy group 88.9% were diagnosed with a VTE postoperatively. Among the 48 UC patients who were diagnosed with VTE, 16.7% were diagnosed after discharge from hospital. No patients died as a complication of the VTE. Postoperative VTE prophylaxis with anticoagulants was prescribed in over 90% of patients. However, VTE prophylaxis was prescribed preoperatively or prior to discharge in less than 20% of patients with UC who underwent emergent colectomy or were emergently admitted to hospital (i.e., medically responsive), respectively (Table 1).

Among all UC patients (n = 1020), UC patients requiring elective colectomy (adjusted OR = 3.69; 95%CI: 1.30-10.44) and emergent colectomy (adjusted OR = 5.28; 95%CI: 1.93-14.45) were significantly more likely to be diagnosed with a VTE as compared to medically responsive UC patients (Table 3). Furthermore, the odds of a VTE significantly increased across time (adjusted OR = 1.10; 95%CI: 1.01-1.20). Patients who underwent emergent colectomy (i.e., elective colectomy excluded) were significantly more likely to develop a VTE (adjusted OR = 4.62; 95%CI: 1.66-12.88) as compared to UC patients who were medically responsive. Furthermore, an in-hospital complication was associated with an increased odds of VTE (adjusted OR = 2.68; 95%CI: 1.42-5.05) (Table 3).

This study identified a population-based cohort of over 1000 hospitalised UC patients. All charts were identified to confirm the diagnosis of VTE both in-hospital and following discharge. The study population was stratified according to the indication for admission (i.e., medically responsive UC flare, emergent colectomy, or elective colectomy) in order to explore the effect of surgery on the risk of VTE. We demonstrated that colectomy was significantly associated with the development of VTE. VTE occurred in about 1% of the medically responsive UC patients, whereas UC patients who failed to respond to medical management and underwent a colectomy had an over 5-fold increased odds of VTE. Similarly, UC patients undergoing elective colectomy had an over 3-fold increased odds of VTE when compared to medically responsive UC inpatients. In contrast, the risk of VTE was not associated with disease location, or drug utilization including the use of corticosteroids, immunomodulators, and biologics.

VTE occurred in 4.7% of UC patients. These occurred in a number of locations including extremity, intra-abdominal (i.e., portal and hepatic vein thrombosis), and pulmonary embolism. VTE were confirmed using multiple different testing modalities including chest X-ray, D-Dimer, Doppler ultrasound of the extremities, V/Q scan, CT PE protocol, and angiography. While most VTE were identified following the manifestation of clinical symptoms, most of the intra-abdominal clots were detected incidentally in asymptomatic patients on CT scans performed to assess postoperative symptoms or complications. While incidental cases were primarily identified in postoperative patients following colectomy (11 cases among elective and 10 cases among emergent colectomy), one incidental intra-abdominal VTE was identified in the medically responsive group. Thus, some cases of VTE that were asymptomatic may have been missed. However, the proportion of clinically relevant VTE that we missed was likely low. For example, a recent study of inpatient and outpatient UC patients who were actively flaring, but were asymptomatic for a DVT, were not found to have any incidentally diagnosed DVT on screening ultrasound[20]. Also, we evaluated every case for readmission to hospital for a VTE within 6 mo of the discharge from hospital admission. Thus, we were able to capture clinically meaningful cases of VTE that were missed in hospital or developed post-discharge. Sixteen percent of UC patients were readmitted to hospital for VTE following discharge from hospital including one patient who was readmitted 6 d after discharge for mesenteric vein thrombosis that led emergent resection of ischemic bowel. Consequently, physicians should be aware of this complication, recognize the signs and symptoms, and have a low threshold to screen for VTE in the extremities and in the abdomen in UC patients.

Current American College of Gastroenterology guidelines recommend VTE prophylaxis for all patients admitted to hospital with an IBD flare as well as all postoperative patients undergoing colectomy[23]. The American College of Chest Physicians (ACCP) guidelines recommend VTE prophylaxis in IBD patients confined to a bed[30]. Though, ambulation alone would not prevent intra-abdominal VTE. In our study, greater than 90% of emergent and elective colectomy patients received VTE prophylaxis postoperatively. Despite the high utilization of postoperative heparin, the occurrence of VTE remained high. Similarly, a smaller study found that despite adequate postoperative heparin use, rates of VTE in post-colectomy patients were over 7-fold higher in UC patients as compared to patients undergoing surgery for colorectal cancer[31]. A randomized controlled trial that included nearly 600 patients with inflammatory bowel disease demonstrated that low dose heparin had significantly lower rates of VTE (2.9%) as compared to patients randomized to enoxaparin (9.0%)[32].

The current ACCP guidelines recommend that patients undergoing abdominal or pelvic surgery for cancer, with a high risk of postoperative VTE (i.e., > 6%) and who do not have increased risk of bleeding, should be considered for both mechanical and pharmacological VTE prophylaxis and extended duration of the latter[30]. Our study was unique from prior work because we explored VTE occurrence up to 6 mo following discharge from hospital. We demonstrated that 16% of our detected VTE were diagnosed after discharge from hospital. Our findings raise the issue of whether the same consideration for extended postoperative VTE prophylaxis should be given for IBD surgeries. Prior studies have reported lower DVT levels (< 3%) postoperatively for IBD surgeries[31,32], as compared to our study. However, nearly half of our VTE population experienced an intra-abdominal VTE. The clinical significance of intra-abdominal VTE is controversial and depends on the acuity, size, and location of the thrombosis[33,34]. Given the uncertainty, these findings warrant a randomized controlled trial to evaluate the efficacy of VTE prophylaxis for 4 wk postoperatively.

Only 19% of patients admitted with an UC flare who underwent emergent colectomy received preoperative VTE prophylaxis. Similarly, only 16% of patients who were admitted for an UC flare and discharged without colectomy received VTE prophylaxis in-hospital. A recent multicenter chart audit of 29 Canadian hospitals also demonstrated similarly low prophylaxis rates for hospitalized medical patients requiring VTE prophylaxis[35]. Moreover, patients who received VTE prophylaxis were more likely to be diagnosed with a VTE. However, this finding likely reflected disease severity (i.e., confounding by indication) because a meta-analysis indicated that heparin is safe to use in an acute flare of UC[25,36] and randomized controlled trials support the use of VTE prophylaxis in IBD[32]. Thus, in the absence of hemodynamically significant haemorrhage VTE prophylaxis should be prescribed. However, the adequate agent, dose, and duration of VTE prophylaxis in medically managed UC patients need further evaluation.

Our study could not differentiate whether the increased risk of VTE among UC patients undergoing colectomy was due specifically to the colectomy or was caused by disease severity necessitating colectomy. Though, most likely it was due to the combination of disease severity and surgery. Thus, the optimal approach of preventing VTE is to induce and maintain remission in UC. The majority of UC patients were admitted to hospital prior to publications demonstrating efficacy of infliximab in the hospital[37] and outpatient setting[38]. Additionally, recent studies have demonstrated that infliximab significantly reduced the risk of colectomy for up to 3 years following initiation of treatment[39-41]. Because less than 4% of the patients in our study were prescribed infliximab, greater utilization of anti-TNF therapies may lead to reduced colectomy and hence, burden of VTE among UC patients.

Several limitations to this study should be considered. While VTE events that resulted in a readmission to hospital within 6 mo of hospital discharge were captured, UC patients diagnosed and managed for VTE as an outpatient following discharge from hospital would not have been accounted for in this study. Thus, the occurrence of VTE may have been underestimated. While chart review improves the accuracy of the data and minimizes misclassification bias that occur with studies using administrative databases[29,42], some data were missed due to incomplete recording in the medical chart. For example, we were not able to account for the use of oral contraceptive pills prior to hospitalization due to unreliable recording of this data. This study was not able to confirm if other forms of VTE prophylaxis such as graduated compression stockings or intermittent pneumatic compression devices for hospitalized UC patients were utilised. However, these modalities are only recommended if heparin is contraindicated because they are less effective at preventing proximal extremity and intra-abdominal VTE. Also, due to the retrospective nature of chart reviews we were not able to assess the number of flares from diagnosis to admission to hospital or to define disease severity using a validated disease activity index (e.g., Mayo score). Finally, despite the advantages of designing a population-based study, regional practice pattern differences may limit generalizability. Consequently, we recommend hospitals assess their own utilization of heparin prophylaxis and correlate these with VTE outcomes.

VTE occurred commonly in UC patients who underwent emergent (8.7%) and elective (4.9%) colectomy despite greater than 90% rate of postoperative heparin prophylaxis. In contrast, about 1% of medically responsive UC inpatients experienced a VTE, despite a 16% prophylaxis rate. Thus, VTE prophylaxis should be prescribed at time of admission in all IBD patients, and VTE prophylaxis for flaring UC patients should be considered a quality indicator of best clinical practice. This population-based study confirmed that the necessity of colectomy was significantly associated with VTE. The high rates of VTE in UC patients who underwent colectomy despite postoperative VTE prophylaxis highlights potentially serious outcomes associated with surgical management of UC. Consequently, these findings emphasize the importance of optimizing medical therapy for UC patients.

We acknowledge the DIMR department for providing data from the Calgary Health Zone. Dr. Kaplan is supported through a New Investigator Award from the Canadian Institute of Health Research and a Population Health Investigator Award from the Alberta Heritage Foundation for Medical Research.