See “Thromboembolism in Pediatric Inflammatory Bowel Disease: The Blood Stops Here” by Zacur and Saeed on page 265.
Patients with inflammatory bowel disease (IBD) have been increasingly recognized to be at high risk for venous thromboembolism (VTE) and associated complications (1–7). Risk of VTE in IBD versus non-IBD patients is 3-fold higher overall (1), and rises as high as 15-fold during disease flares (2). In adults with both Crohn disease (CD) and ulcerative colitis (UC), colonic disease location has been associated with higher risk of VTE (3–5). Although the incidence of thrombosis in children appears lower than that of adults, pediatric patients with IBD hospitalized with disease flares have been reported to have a 6-fold higher risk of both deep vein thrombosis and pulmonary embolism when compared with non-IBD pediatric inpatients (6). In patients of all ages, IBD-associated VTE has a high risk for reoccurrence and in-hospital mortality (5,7).
There are many risk factors, which may contribute to thrombotic events in patients with IBD (8). These include malnutrition, malabsorption, dehydration, surgery, and medications, such as thalidomide. Patient risk factors for thrombosis that may also compound baseline IBD risk include obesity, oral contraceptive use, and genetic predisposition. Hospitalized patients with IBD are often at particular risk for thrombosis because of infection and venous stasis induced by immobilization. Indwelling central line and other catheters necessary for administering drugs or intravenous nutrition may be a specific risk factor. These lines cause endothelial damage, loss of mucosal integrity, and venous stasis, as well as provide a nidus for both infection and thrombosis. Finally, the use of total parenteral nutrition (TPN) can contribute to line-associated thrombosis (9).
Present guidelines for hospitalized adult patients with IBD include prophylactic anticoagulation (10,11). In contrast, evidence-based strategies are lacking for pediatric patients; 1 recent publication on risk stratification for hospitalized children noted IBD as a risk factor (12). The difference in practice recommendations may be because of the overall lower incidence of VTE in pediatrics (6). In addition, there may be general misconceptions about the risks and feasibility of administering prophylactic anticoagulation to children, as well as the effects of anticoagulation on patients with active gastrointestinal (GI) bleeding.
To date, several trials of low-molecular-weight heparin (LMWH) as a possible treatment for adult UC have demonstrated its safety in patients with active colitis (12–14). One double-blinded, randomized placebo-controlled trial used subcutaneous tinzaparin for 6 weeks and found that none of the 48 patients with active colitis enrolled developed clinically significant rectal hemorrhage (13). A second double-blinded, randomized, placebo-controlled trial of reviparin for mild-to-moderate UC for 8 weeks also reported no increase in rectal hemorrhage in the 15 subject treatment group (14). The study with the longest treatment period used enoxaparin daily for 12 weeks and again found no increased rectal hemorrhage in the 16 subjects treated or in the placebo group (15).
The aim of our study was to evaluate the incidence and severity of thrombotic events in an inpatient IBD population with colonic inflammation at a large pediatric tertiary care center. Specifically, we reviewed all patients with IBD who developed VTE at our center during the study period to determine morbidity, length of stay, cost, and associated complications, as well as any medical and surgical management used in their treatment. We also were interested in characterizing clinical and laboratory risk factors present in our incident cases, as well as the safety of using therapeutic anticoagulation.
We performed a retrospective review of pediatric inpatients with IBD at Boston Children's Hospital (BCH) to determine the prevalence of thrombotic complications in the inpatient setting. Patients were identified from 2 databases: thrombotic complications were identified from an institutional review board–approved database of all patients on anticoagulation while inpatient at BCH, and IBD admissions from 2006 to 2011 were compiled based on billing codes beginning with 555 or 556 from a BCH inpatient billing database. Diagnosis code 555 (ie, regional enteritis, chronic inflammation of the intestines) was used to capture all subjects with CD. Diagnosis code 556 (ie, UC) captured all subjects with UC, defined as chronic, recurrent inflammation of the colonic mucosa and submucosa. Given that previous reports have suggested that increased VTE risk is associated with CD with colonic involvement (3,4,16), we then excluded subjects with isolated small bowel inflammation, a primary or secondary diagnosis code of 555.0.
To avoid repetitively counting risk factors in readmitted patients, incidence data were calculated per patient rather than per admission. Incident cases were reviewed to identify documented prothrombotic risk factors including obesity (body mass index ≥35), smoking, indwelling central venous catheters (including peripherally inserted central catheter, Broviac or Port-a-Cath), prothrombotic medications (thalidomide, oral contraceptives), recent surgery, and first-degree family history of VTE. When documented in the medical record, inflammatory markers (elevated sedimentation rate and C-reactive protein) and disease activity indices (Pediatric Ulcerative Colitis Activity Index [PUCAI]) at the time of thrombosis were collected. The length and cost of admission of incident cases of IBD and thrombosis were compared with average costs of admission and length of stay for IBD flares for a 1-year period at BCH. Variable hereditary or acquired thrombophilia evaluations were performed at the discretion of involved clinicians. Length of anticoagulation, type, and reason for discontinuation were also reported for incident cases of IBD with VTE.
Incidence, Demographics, and Relation to Inflammation
Of 532 patients admitted during the study period (2006–2011) with IBD involving the colon, 10 patients (1.9%; 4 male patients, 6 female patients) were identified to have had thromboembolic events (9 venous, 1 arterial). Patients ranged in age from 8 to 23 years (mean 15.8 years). Of the patients with thrombotic complications, 8 had panulcerative colitis and 2 had CD with colonic involvement. Of 104 pediatric IBD inpatients with an indwelling line, 4 (3.8%) developed line-associated thrombosis, including 1 patient with pulmonary embolism (Fig. 1). Length of follow-up from VTE ranged from 1.1 to 7.5 years (mean 4.3 years). Of the 10 patients with thrombotic complications, 2 developed recurrent thrombosis (Tables 1 and 2).
Subjects whose inflammatory markers and disease activity indices were assessed at the time of thrombosis were found to be severely inflamed with fulminant disease activity: median elevated sedimentation rate 62.0 mm/hour (interquartile range [IQR] 42.5–104.5, normal range 0–20 mm/hour), median C-reactive protein 3.8 mg/dL (IQR 3.1–13.5, normal range <0.5 mg/dL), and PUCAI median 70 (IQR 65–70), consistent with severe disease (PUCAI >65) (17). Six of 8 patients with UC underwent colectomy shortly after their thrombotic event, suggesting severe disease refractory to medical management (Table 1).
Thromboembolic Events Are Associated With Increased Morbidity and Cost
Four of 10 patients had cerebrovascular thrombosis, 1 of whom has resulting permanent hemiparesis and cognitive defects as a result of cerebral sinus thrombosis. Subject 5 required intracranial vascular surgery. Subject 4 required placement of a permanent inferior vena cava filter and surgical closure of an atrial septal defect to prevent stroke, and had permanent occlusion of both proximal femoral veins with resulting postthrombotic syndrome (chronic pain and edema). Both subjects 1 and 4 had persistent thrombi despite therapeutic anticoagulation.
Thrombotic complications resulted in increased length of hospitalization, with a median of 18 days (IQR 7.5–26), compared with a median of 5 (IQR 3–8) days in those without VTE. Subjects with thrombosis had multiple imaging studies to assess thrombi, increasing both radiation exposure and cost of hospitalization. Indeed, thrombotic events more than doubled the median cost of hospitalization: $37,263 for patients with VTE versus $15,665 for patients with IBD without VTE.
We identified the following additional risk factors in our patient cohort: indwelling catheters (4/10), first-degree family history of VTE (2/10), hereditary thrombophilia (3/10), smoking (1/10), oral contraceptive use (1/5 female patients of appropriate age), and thalidomide (1/10). Hereditary thrombophilia was identified 4 times in 3 patients: prothrombin gene mutation G20210A (2/10), protein S deficiency (1/10), and factor V Leiden (1/10). Two patients were found to have elevated lipoprotein-a, which is inherited, but has a milder effect on thrombotic risk. Most subjects (8/10) had concomitant acquired thrombophilias (Table 2). Elevated factor VIII, D-dimer, anti-cardiolipin antibodies, and von Willebrand factor antigen were noted in ≥2 of the subjects at the time of thrombosis. Although protein C deficiency is associated with thrombotic risk, we identified elevated protein C function in 4 of 8 patients tested. Acquired thrombophilias may persist after VTE (Table 3), most prominently factor VIII, but also D-dimer, and anti-cardiolipin antibodies. Persistent thrombophilia contributes to the risk of thrombus recurrence; however, acquired thrombophilias were not uniformly tested at the time of VTE or at follow-up in all subjects (Tables 2 and 3).
Central Venous Line Clots Are Less Severe
Of 532 inpatients admitted with IBD-related colonic inflammation, 104 (19.5%) had a central venous line (CVL) placed during their hospitalization. Among patients with VTE, 4 had line-associated VTE (3.8%), as opposed to 6 whose VTE was not assessed to be associated with their CVL (1.4%) (Fig. 1). All patients (4/4) with CVL-associated VTE had ≥1 of the previously identified acquired risk factors for thrombosis in addition to their CVL. Fifty percent (3/6) of subjects with non-CVL–related VTE events had additional risk factors (Table 2). Subjects with CVL-associated thrombi had a median length of therapeutic anticoagulation of 3 months (IQR 2.6–9.3) compared with a median of 7.5 months (IQR 6.3–11.0) in those with non-CVL–associated thromboembolism (TE). Furthermore, all (4/4) subjects with CVL-associated VTE had complete resolution of their thrombi compared with 4 of 6 with non-CVL–associated VTE.
Safety and Efficacy of Anticoagulation
Nine of the 10 cases with IBD colitis with thrombosis were treated with therapeutic enoxaparin (1 mg/kg injected subcutaneously twice per day). The tenth patient developed thrombosis of the inferior mesenteric vein 8 days after colectomy and was managed without anticoagulation. Despite active GI bleeding, no subjects required transfusion or experienced a decrease in hemoglobin >2 g/dL (data not shown). One patient (subject 3) had increased GI bleeding after 4.5 months of therapeutic anticoagulation (higher dose than prophylactic); imaging showed the thrombus had completely resolved and anticoagulation was stopped. Length of anticoagulation ranged from 1.5 to 28 months (ongoing) (Table 4). Thrombus resolution was documented in 7 cases, persistence in 2 cases, and recurrence in 2 cases.
The incidence of pediatric inpatient VTE is rising, and was recently reported at 0.58% using data from the Pediatric Health Information System (18). We identified an incidence of 1.9% (10/532 subjects) in our inpatient cohort with IBD colitis, 4 of which were CVL associated and 6 non-CVL associated. These events were often severe and resulted in significant morbidity, most notably permanent neurologic deficits, persistent or recurrent thrombosis, embolization, and postthrombotic syndrome. The severity of these events influenced discussions regarding the need to undergo colectomy for refractory colitis. VTE events also increased length of stay and cost of treatment.
The absolute incidence of VTE is lower in younger patients, and multiple risk factors are often present when thrombosis does occur (6). Our results are consistent with a recent meta-analysis of pediatric and young adult subjects with IBD, which showed an increased incidence of VTE with active colonic disease and the presence of at least 1 additional risk factor for VTE (19). We found known risk factors to be present in a majority of patients who developed thrombosis, most commonly central venous lines. The primary risk factor for thrombosis in IBD appears to be the inflammatory state. Most patients with VTE in our cohort had high laboratory markers of inflammation and measurements of disease activity. In our cohort, 7 of 10 incident cases of TE in IBD inpatients had ≥1 additional risk factors for TE. These data suggest that assessment for additional risk factors may identify patients at higher risk for VTE, allowing for risk stratification preventive strategies.
Whether or not prophylactic anticoagulation is warranted in pediatric IBD inpatients remains an unsettled question. Inflammatory bowel disease has been recognized as a risk factor for TE in adult patients and the patients older than 17 years in our study would have received thromboprophylaxis had they been admitted to an adult facility (10,11); however, no standard approach to thromboprophylaxis in pediatric IBD has been proposed, caused in part by lower absolute incidence and concerns about anticoagulation. Children with active colitis who are admitted to the hospital may have significant rectal bleeding, so there is often resistance to initiating prophylactic anticoagulation. In our cohort, all patients were anticoagulated safely and effectively on therapeutic (higher than prophylactic) dosages of low-molecular-weight heparin. Encouraging a standard approach toward prophylactic anticoagulation in pediatric patients with IBD may represent a prime target for reducing morbidity in this already-complicated population.
The severity of thromboembolic events generated interest in risk stratification and thromboprophylaxis strategies at our institution. Numerous multidisciplinary discussions led to the acceptance of a thromboprophylaxis protocol. We have begun to use a risk-stratification algorithm for pediatric inpatients with IBD and colonic inflammation to identify patients at high risk of developing TE. A similar strategy for thromboprophylaxis for pediatric inpatients at a major pediatric center was published in which IBD was identified as a risk factor (12). At our institution, patients with IBD who are admitted with a flare or for surgery are evaluated for risk factors using a standardized form in our electronic medical record. Upon admission, evaluation of thrombotic risk factors is performed. Patients with IBD are defined as being at high risk for thrombosis if they have 1 major criterion and 1 minor criterion. Major criteria include admission for IBD with colonic involvement or major surgery. Minor criteria include personal history of thrombosis, first-degree family history of VTE, known thrombophilia, persistent anti-phospholipid antibody >12 weeks, oral contraceptive use, smoking, obesity, thalidomide, or central venous catheter. Patients screened as being at high risk for thrombosis and low risk for bleeding (some GI bleeding acceptable) are offered prophylactic enoxaparin, 0.5 mg/kg twice daily if <60 kg or prepubertal and 40 mg daily if >60 kg and postpubertal. Prophylaxis continues until discharge, return to baseline mobility, or resolution of colitis. We continue to monitor the safety of anticoagulation in these patients and the rate of thromboembolic events to evaluate the utility of our algorithm. A multicenter initiative to determine biomarkers of thromboembolic risk and best preventive strategies is strongly encouraged for this high-risk population.
The authors thank Juliann McSweeney and the anticoagulation service for maintenance of the database of all patients anticoagulated at our institution.
1. Bernstein CN, Blanchard JF, Houston DS, et al. The incidence of deep venous thrombosis and pulmonary embolism among patients with inflammatory bowel disease: a population-based cohort study. Thromb Haemost
2. Grainge MJ, West J, Card TR. Venous thromboembolism during active disease and remission in inflammatory bowel disease: a cohort study. Lancet
3. Solem CA, Loftus EV, Tremaine WJ, et al. Venous thromboembolism in inflammatory bowel disease. Am J Gastroenterol
4. Novotny DA, Rubin RJ, Slezak FA, et al. Arterial thromboembolic complications of inflammatory bowel disease. Report of three cases. Dis Colon Rectum
5. Nguyen GC, Sam J. Rising prevalence of venous thromboembolism and its impact on mortality among hospitalized inflammatory bowel disease patients. Am J Gastroenterol
6. Kappelman MD, Horvath-Puho E, Sandler RS, et al. Thromboembolic risk among Danish children and adults with inflammatory bowel diseases: a population-based nationwide study. Gut
7. Novacek G, Weltermann A, Sobala A, et al. Inflammatory bowel disease is a risk factor for recurrent venous thromboembolism. Gastroenterology
2010; 139:779–787.87 e1.
8. Zitomersky NL, Verhave M, Trenor CC 3rd. Thrombosis and inflammatory bowel disease: a call for improved awareness and prevention. Inflamm Bowel Dis
9. Kakzanov V, Monagle P, Chan AK. Thromboembolism in infants and children with gastrointestinal failure receiving long-term parenteral nutrition. JPEN J Parenter Enteral Nutr
10. Geerts WH, Bergqvist D, Pineo GF, et al. Prevention of venous thromboembolism: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest
2008; 133 (suppl 6):381S–453S.
11. Kornbluth A, Sachar DB. Ulcerative Colitis
Practice Guidelines in adults: American College of Gastroenterology Practice Parameters Committee. Am J Gastroenterol
12. Raffini L, Trimarchi T, Beliveau J, et al. Thromboprophylaxis in a pediatric hospital: a patient-safety and quality-improvement initiative. Pediatrics
13. Bloom S, Kiilerich S, Lassen MR, et al. Low molecular weight heparin (tinzaparin) vs. placebo in the treatment of mild to moderately active ulcerative colitis
. Aliment Pharmacol Ther
14. de Bievre MA, Vrij AA, Schoon EJ, et al. Randomized, placebo-controlled trial of low molecular weight heparin in active ulcerative colitis
. Inflamm Bowel Dis
15. Zezos P, Papaioannou G, Nikolaidis N, et al. Low-molecular-weight heparin (enoxaparin) as adjuvant therapy in the treatment of active ulcerative colitis
: a randomized, controlled, comparative study. Aliment Pharmacol Ther
16. Oldenburg B, Van Tuyl BA, van der Griend R, et al. Risk factors for thromboembolic complications in inflammatory bowel disease: the role of hyperhomocysteinaemia. Dig Dis Sci
17. Raffini L, Huang YS, Witmer C, et al. Dramatic increase in venous thromboembolism in children's hospitals in the United States from 2001 to 2007. Pediatrics
18. Lazzerini M, Bramuzzo M, Maschio M, et al. Thromboembolism in pediatric inflammatory bowel disease: systematic review. Inflamm Bowel Dis