Inflammatory Bowel Disease and Risk of Major Bleeding During Anticoagulation for Venous Thromboembolism

Abstract

INTRODUCTION

Venous thromboembolism (VTE), which includes deep vein thrombosis and pulmonary embolism, is a frequent disease with an annual incidence of 1 to 2 per 1000 persons.¹ It is one of the most common causes of cardiovascular death and is thus a major disease burden.^{2, 3} In half of patients, VTE occurs in association with a transient or permanent provoking risk factor including surgery, trauma, hospitalization, immobilization, pregnancy, female hormone intake, or cancer.^{4, 5} There is compelling evidence that patients with inflammatory bowel disease (IBD) are at increased risk not only for this type of VTE but also for recurrent VTE.^6-19 In addition to the aforementioned thrombotic risk factors, active and/or more extensive intestinal disease and the use of corticosteroids contribute to the increased risk of VTE, specifically in patients with IBD.^12-19 Furthermore, VTE is associated with increased morbidity and mortality among hospitalized patients with IBD.⁸

Patients with an acute VTE require immediate and intense anticoagulant therapy (AC) for a minimum of 3 months to prevent embolization, progression, and recurrence.²⁰ In patients at high risk of recurrent VTE including patients with IBD, long-term anticoagulation should be considered. However, AC does not come without risks in that it can cause severe or even fatal bleeding. Thus, the decision on the optimal duration of secondary thromboprophylaxis is based on weighing the benefits and harms of the AC. Long-term anticoagulation is only acceptable if the recurrence risk outweighs the risk of bleeding.²¹ Patients with IBD may suffer from gastrointestinal blood loss, less often in Crohn disease (CD) than in ulcerative colitis (UC), where bloody diarrhea is the hallmark of active disease.^22-25 Furthermore, severe intestinal hemorrhage, although a rare complication in IBD, may require emergency surgery.²⁶ Thus, it is conceivable that any anticoagulant treatment will contribute to the bleeding risk of patients with IBD, which may consequently influence decisions on the management of these patients.

However, limited and in part conflicting data exist about the bleeding risk during AC among patients with IBD. In an analysis of a multicenter observational registry of patients with VTE, AC at therapeutic dose was found to be not only effective but also safe in patients with IBD vs patients without IBD.²⁷ In a retrospective analysis of 63 patients with IBD-related portal vein thrombosis, major bleeding (MB) was described in 4 patients.²⁸ Furthermore, AC at a prophylactic dosage for preventing primary VTE^{29, 30} has been reported to be safe without an increased rate of bleeding events in some studies, whereas others have reported an increased red cell transfusion requirement.^31-33

We therefore sought to assess the rate of MB events in a well-defined cohort of patients with IBD during AC after VTE. In addition, we evaluated the clinical characteristics and risk factors of the bleeding events and the outcome and management of bleeding.

PATIENTS AND METHODS

Study Population

The present study is a follow-up investigation of an Austrian multicenter cohort study with the aim to evaluate the risk of a first and recurrent VTE in patients with IBD. Details of the multicentre study have been previously reported.^{9, 34} In that study, patients older than age 18 years with an established diagnosis of IBD^{23, 24} were evaluated for a history of VTE that had occurred after the diagnosis of IBD. One hundred fifty-seven patients with a first VTE objectively verified by compression ultrasound, venography, spiral computed tomography, or ventilation/perfusion lung scanning were identified.

The objective of the present study was to evaluate the impact of exposure to anticoagulation after VTE on the risk of MB in patients with IBD. The study is a project of the Austrian IBD Study Group.

Data Collection and Evaluation of MB Events

Data on MB events and IBD-related parameters of interest such as type of IBD, disease extent and behavior, surgery, medication, and disease activity were collected either via telephone interview by the study coordinator (coauthor ST) or by the treating physician, respectively, using a standardized questionnaire during 2011. The information given by the patients was double-checked, and missing data were added through the medical records (coauthors ST and GN, and the treating physicians of the participating IBD centers, respectively). We defined MB according to the criteria of the International Society on Thrombosis and Haemostasis³⁵ as follows: (1) fatal bleeding; (2) symptomatic bleeding in a critical area or organ, such as intracranial, intraspinal, intraocular, retroperitoneal, intraarticular or pericardial, or intramuscular with compartment syndrome; and/or (3) bleeding causing a fall in hemoglobin level of ≥2 g/dL (1.24 mmol/L) or leading to a transfusion of ≥2 units of whole blood or red cells.

The extent of IBD and disease behavior were classified according to the Montreal classification.³⁶ Active disease at the time of MB was defined if IBD-related intestinal symptoms were present, by means of a physician’s global assessment. Markers for disease severity such as mucosal lesions, inflammatory markers, ongoing use of immunosuppressive treatment such as azathioprine or biologics, and recent use of steroids were added.³⁷ Any IBD-related surgery was defined as bowel resection only. Bleeding-related parameters of interest included start and end, type, dose of anticoagulant therapy, occurrence of MB, and outcome. If a patient had died since the last control, then the cause of death was captured by use of a nationwide registry (Statistik-Austria death registry; https://www.statistik.at/web_en/statistics/PeopleSociety/population/deaths/index.html).

Ethical Considerations

The study was approved by the Ethics Committee of the Medical University of Vienna, and informed consent was obtained from all participants.

Primary Endpoint

The primary endpoint was the occurrence of MB.

Statistical Analysis

The rate of MB events in patients with IBD was evaluated as a sequential therapy comparison in the same patients with and without anticoagulation. The observation period started with the diagnosis of IBD and ended with the last follow-up. This overall observation time was divided into periods with exposure to anticoagulation and periods without it.

Data are presented as numbers and percentages or median and interquartile range (IQR), if not otherwise specified, respectively. Patients contributed time periods with variable length, each of which were characterized by anticoagulation or no anticoagulation. To enable the longitudinal structure, a random-effects Poisson regression model was used, allowing for the length of each separate episode and multiple episodes per patient. The incidence rate ratios (IRRs) were calculated with 95% confidence intervals (CI) and tested the null hypothesis of IRR = 1. Anticoagulation (yes vs no) was the main covariate. To adjust for other clinically predetermined influencing variables, we used multivariable regression models, entering the other variables as covariates. These consisted of sex, age (years), and a diagnosis of UC (yes vs no). For calculation, we used MS Excel for Mac 2011 and Stata 11 (Stata Corp, College Station, TX). Generally, a 2-sided P < 0.05 was considered statistically significant.

RESULTS

Patient Characteristics

We identified 157 patients with a first VTE after the diagnosis of IBD. Of those, 107 patients with ≥1 period of anticoagulation because of ≥1 VTE were included in the present analysis (Table 1). The remaining 50 patients had to be excluded because of missing consent (n = 4), loss to follow-up (n = 36), and death (n = 10). Reasons for death were myocardial infarction (n = 2), cardiomyopathy (n = 2), liver cirrhosis (n = 2), renal failure, abscess of the lung, and emphysema. One patient died because of CD-related complications other than bleeding. Furthermore, none of the patients lost to follow-up died on account of an MB event.

Anticoagulation After VTE

The overall observation time was 1833 patient-years, divided into an observed time with anticoagulation (388 patient-years) and without anticoagulation (1445 patient-years). This framework corresponded to 144 periods with and 195 periods without anticoagulation. Upon VTE diagnosis, patients were treated with low molecular weight heparin (LMWH) in therapeutic dosage in 41 episodes and with vitamin K-antagonist (VKA) in 103 episodes (Table 2).

MB

We recorded 23 MB events: 13 events in 195 periods without anticoagulation and 10 MB events in 144 periods with anticoagulation. Characteristics of these patients are presented in detail in Tables 3 and 4, respectively.

Those 13 MB events during periods without anticoagulation happened in 12 patients (7 women, 5 men). The median age at the time of MB was 38 years (IQR, 32-48 years). Eight patients (66.7%) had extensive (n = 6) or left-sided (n = 2) UC, and 4 patients had ileocolonic (n = 3) or colonic (n = 1) CD. Active disease was described in 9 out of 13 patients (69.2%) at the time of MB, inactive disease in 3 patients, and enough information was not available for 1 patient. The majority of patients with active disease and endoscopic procedures performed around the time of MB had the presence of mucosal lesions. Furthermore, elevated levels of C-reactive protein (CRP), ongoing immunosuppressive treatment with azathioprine or biologics, and recent steroid use were common in these patients, surrogates for more severe disease. All MB events were associated with hematochezia and led to ≥2 red cell transfusions. Surgery was required in 5 patients, including right-sided hemicolectomy (n = 1) and colectomy (n = 2). One patient had severe bleeding from a jejunal ulcer twice, requiring surgery both times. Despite intervention, the patient died from bleeding because of an erosion of the splenic artery resulting from a perforated ulcer. In all other patients, start or escalation of the medical IBD treatment was reported.

Ten MB events occurred during AC (5 women, 5 men) 5 days to 4 years after the start of anticoagulation (median, 15 months; IQR, 3.5-41 months). Six events were observed during VKA treatment and 4 during LMWH treatment. One patient had another MB event when diagnosed with active UC, without receiving AC therapy at that time. The median age at the time of MB in these patients was 47 years (IQR, 40-50 years). Six patients had extensive (n = 5) or left-sided (n = 1) UC (60.0%), and 4 patients had ileocolonic (n = 3) or colonic (n = 1) CD. At the time of the MB events, 7 out of these 10 patients (70.0%) had active disease, 2 patients had inactive disease, and enough information was not available for 1 patient.

In patients with active disease and in whom endoscopy was performed concurrently with the MB event, mucosal lesions were present. Similarly, as in the group of patients with MB without AC, elevated CRP levels, ongoing immunosuppressive treatment with azathioprine or biologics, and recent steroid use indicated disease severity. One patient had hematemesis because of esophageal and fundus varices resulting from portal vein thrombosis. Bleeding was stopped by band ligation and histoacryl injection, respectively. In all other patients, MB was associated with hematochezia. In 6 patients, red cell transfusions were required, and in 4 patients, a fall in the hemoglobin level of ≥2 g/dL occurred. No surgery was required, and no patient died of bleeding. In 3 patients, AC was continued; in another 3 patients, anticoagulation using VKAs was switched to LMWH; and in 4 patients, AC was stopped prematurely.

Incidence Rate for MB and Random-Effects Poisson Regression Model

We found that MB occurred more often during anticoagulation. The incidence rate for major bleeding was 2.6 per 100 patient-years during periods when patients were exposed to AC and 0.9 per 100 patient-years during the unexposed time. The unadjusted IRR was 2.9 (95% CI, 1.1-7.1; P = 0.018). In a random-effects Poisson regression model, only exposure to anticoagulation (adjusted IRR, 3.7; 95% CI, 1.5-9.0; P = 0.003) and UC (adjusted IRR, 3.5; 95% CI, 1.5-8.1; P = 0.003) were independent risk factors for MB (Table 5). We also found that MB was not associated with either sex or age. When we replaced UC in the random-effects Poisson regression model with colonic IBD (UC + CD with colonic involvement), colonic IBD was a significant risk factor for MB (adjusted IRR, 8.1; 95% CI, 1.1-60; P = 0.041), and the overall estimate for the effect of anticoagulation remained virtually unchanged (adjusted IRR, 3.6; 95% CI, 1.5-8.9; P = 0.005).

DISCUSSION

Our analysis shows a higher incidence rate for MB in patients with IBD during periods with AC vs periods without AC (2.6 per 100 patient-years vs 0.9 per 100 patient-years, respectively). We did not find a numerical difference in fatal bleeding, nor was surgery required because of MB under AC.

Patients with IBD are prone to developing VTE, which is associated with increased morbidity and mortality.^{8, 18, 19} Long-term anticoagulation therapy in patients with VTE aims to prevent VTE recurrence, at the expense of the risk of MB,²¹ especially in patients who already experience gastrointestinal bleeding because of luminal disease activity. However, limited and in part conflicting data exist in the literature on bleeding risk during AC among patients with IBD.

What is already known about bleeding risk during therapeutic AC? Analysis from a large, multicenter, observational registry of consecutive patients, including 265 patients with IBD with symptomatic objectively confirmed acute VTE, revealed that therapeutic AC for patients with IBD and VTE is effective and safe.²⁷ Patients with and without IBD had a similar rate of MB during the course of AC. In the group of patients in that study with IBD, depending on disease activity, the rate of MB ranged from 3 per 100 patient-years during remission to 7.3 per 100 patient-years during acute flares. In that study, further IBD-related data such as the type of IBD were not provided. Furthermore, in a noncontrolled analysis of 63 patients with IBD-related portal vein thrombosis, MB was described in 4 patients.²⁸ In a systematic review and a small case series, gastrointestinal hemorrhagic complications were reported in, respectively, 1 of 35 and in none of 9 patients with IBD receiving AC after developing a VTE.^{38, 39} A meta-analysis of the utility and safety of heparin in the treatment of active UC included 8 randomized controlled trials, investigating a total of 457 patients.⁴⁰ Six patients in the heparin group developed an increase in rectal bleeding. Three of these patients had to be withdrawn from the study, and 1 of these 3 patients required urgent surgery.

Because of the limited number of controlled trials pertaining to therapeutic anticoagulation in patients with IBD, we assessed the possible existence of more bleeding episodes in patients treated with prophylactic doses of anticoagulation. Some studies, which have included up to 974 patients, described similar rates of minor bleeding and MB events between patients who did and did not receive prophylactic AC, even if the patients had hematochezia associated with their IBD flare.^{30, 31, 33} On the contrary, in a study with 717 patients presenting with IBD exacerbation, more bleeding events with an increased red cell transfusion requirement were observed.^{30, 32} However, these results can be distorted by the underutilization of VTE prophylaxis in patients with IBD, most often associated with patients with active IBD disease with gastrointestinal hemorrhage and coexisting anemia.²⁹

We further evaluated the clinical characteristics and the outcome and management of bleeding. All MB events, which occurred during episodes without AC, required the transfusion of red cells during patients’ hospital stay, confirming that IBD can cause chronic anemia and gastrointestinal bleeding associated with IBD itself. In addition, MB during AC required the transfusion of red cells in 6 out of 10 patients. No surgery was required, and no fatal bleeding occurred. In only 4 of these 10 patients, AC was suspended, and the other patients continued with their treatment or were switched from VKA to LMWH.

Are there any further risk factors for MB in patients with IBD? Our results revealed that aside from anticoagulation, UC was an independent risk factor for MB. This finding could be expected because hematochezia and bloody diarrhea are hallmarks of active disease, independent of AC. However, UC has not previously been described as an independent risk factor for MB under AC. Along with this finding, in most patients the source of bleeding was the colon. When we replaced UC with colonic IBD, including UC and CD with colonic involvement, colonic IBD was a significant risk factor for major bleeding; be that as it may, this result seemed to be mainly influenced by the effect of UC.

Some patients with IBD presented additional risk factors for gastrointestinal bleeding, eg, gastrointestinal ulcers or esophageal and fundus varices. These bleeding events were included in the total number of MB events. Underlying risk factors or conditions added to the bleeding risk in this vulnerable patient cohort and should be considered in clinical practice.

Another clinically important observation is that 70% of the patients in our study had active disease at the time of MB occurring during AC. Furthermore, mucosal lesions discovered through endoscopic procedures performed around the time of the bleeding event, elevated levels of CRP, and recent steroid use were common in these patients, confirming disease activity. These findings are in accordance with Lobo et al,²⁷ who reported higher incidence rates of MB under AC during acute flares compared to clinical remission. We included sex and age in our regression model, but these variables were not associated with MB. No further risk factors have been described in the literature.

The choice of treatment for VTE in patients with IBD has been given much attention and remains to be fully clarified. A systematic review and meta-analysis of data from randomized controlled trials and real-world studies compared treatment using direct-acting oral anticoagulants vs conventional treatment, finding no significant difference when it came to the risk of major gastrointestinal bleeding events.⁴¹ However, in that study no information about patients with IBD was available. Solely focusing on treatment using direct-acting oral anticoagulants in patients with IBD with VTE, little information has been published concerning the risk of bleeding.^{27, 28}

Our study has several strengths and limitations that warrant further comment. Our patient cohort included both the time with and without AC—thus the entire time from diagnosis to end of follow-up. Every patient was his or her own control patient, which is important because IBD itself may be associated with bleeding events. Furthermore, our study consisted of a well-defined cohort of patients with IBD with an established diagnosis of IBD and VTE for which patients received AC. On the other hand, one important limitation of our study is the retrospective design. Moreover, a potential selection bias may have occurred because patients with IBD who were bleeding may have received AC at a reduced dose or for a limited time. Thus, our findings may not be representative of all patients with IBD. Another limitation is the lack of a control group; therefore, we could not compare the results to a non-IBD group. In addition, because our study focused on the use of conventional AC, only including patients receiving LMWH and VKA, the results cannot be extrapolated to patients receiving direct-acting oral anticoagulants. Finally, although we were able to describe clinical activity at the time of MB, we could not include this variable in the regression model because this information was not available for the overall observation time.

CONCLUSIONS

Our data suggest that MB episodes in patients with IBD exposed to AC are more common than during the unexposed time. However, we did not find a difference in mortality between time periods with and without AC. Most of the patients with MB had active disease at the time of the bleeding event with mucosal lesions at endoscopy. Patients with IBD are at an increased risk of recurrent VTE. Thus, after weighing the benefits and harms, we conclude that AC for VTE in patients with IBD should be performed according to the guidelines.

Presented at: This work has been presented in part as a poster at the ECCO Congress, February 2018.

Abbreviations

Abbreviations
 
• AC

  anticoagulation

 
• CD

  Crohn disease

 
• CI

  confidence interval

 
• CRP

  C-reactive protein

 
• IBD

  inflammatory bowel disease

 
• IQR

  interquartile range

 
• IRR

  incidence rate ratio

 
• LMWH

  low molecular weight heparin

 
• MB

  major bleeding

 
• UC

  ulcerative colitis

 
• VKA

  vitamin K-antagonist

 
• VTE

  venous thromboembolism

Conflicts of interest: SS has nothing to disclose. CP has served as a speaker/consultant/advisory board member for AbbVie, MSD, Takeda, Janssen, Merck, Ferring, and Astro Pharma. SE has served as a speaker/consultant/advisory board member of Bayer, BMS, Boehringer-Ingelheim, CSL Behring, Daiichi-Sankyo, and Pfizer. ST has nothing to disclose. MK has nothing to disclose. RK has served as a speaker/consultant/advisory board member for AbbVie, MSD, Takeda, Janssen, Merck, Pfiser, Sandoz, and Astro Pharma. AB has served as a speaker for AbbVie, Janssen, Takeda, Genericon, Pfizer, MSD, and Vifor. WR has served as a speaker for Abbott Laboratories, AbbVie, Aesca, Aptalis, Astellas, Centocor, Celltrion, Danone Austria, Elan, Falk Pharma GmbH, Ferring, Immundiagnostik, Mitsubishi Tanabe Pharma Corporation, MSD, Otsuka, PDL, Pharmacosmos, PLS Education, Schering-Plough, Shire, Takeda, Therakos, Vifor, and Yakult; as a consultant for Abbott Laboratories, AbbVie, Aesca, Algernon, Amgen, AM Pharma, AMT, AOP Orphan, Arena Pharmaceuticals, Astellas, Astra Zeneca, Avaxia, Roland Berger GmBH, Bioclinica, Biogen IDEC, Boehringer-Ingelheim, Bristol-Myers Squibb, Cellerix, Chemocentryx, Celgene, Centocor, Celltrion, Covance, Danone Austria, DSM, Elan, Eli Lilly, Ernst & Young, Falk Pharma GmbH, Ferring, Galapagos, Genentech, Gilead, Grunenthal, ICON, Index Pharma, Inova, Intrinsic Imaging, Janssen, Johnson & Johnson, Kyowa Hakko Kirin Pharma, Lipid Therapeutics, LivaNova, Mallinckrodt, Medahead, MedImmune, Millenium, Mitsubishi Tanabe Pharma Corporation, MSD, Nash Pharmaceuticals, Nestle, Nippon Kayaku, Novartis, Ocera, OMass, Otsuka, Parexel, PDL, Periconsulting, Pharmacosmos, Philip Morris Institute, Pfizer, Procter & Gamble, Prometheus, Protagonist, Provention, Robarts Clinical Trial, Sandoz, Schering-Plough, Second Genome, Seres Therapeutics, Setpointmedical, Sigmoid, Sublimity, Takeda, Therakos, Theravance, Tigenix, UCB, Vifor, Zealand, Zyngenia, and 4SC; and as an advisory board member for Abbott Laboratories, AbbVie, Aesca, Amgen, AM Pharma, Astellas, Astra Zeneca, Avaxia, Biogen IDEC, Boehringer-Ingelheim, Bristol-Myers Squibb, Cellerix, Chemocentryx, Celgene, Centocor, Celltrion, Danone Austria, DSM, Elan, Ferring, Galapagos, Genentech, Grunenthal, Inova, Janssen, Johnson & Johnson, Kyowa Hakko Kirin Pharma, Lipid Therapeutics, MedImmune, Millenium, Mitsubishi Tanabe Pharma Corporation, MSD, Nestle, Novartis, Ocera, Otsuka, PDL, Pharmacosmos, Pfizer, Proctor & Gamble, Prometheus, Sandoz, Schering-Plough, Second Genome, Setpointmedical, Takeda, Therakos, Tigenix, UCB, Zealand, Zyngenia, and 4SC and has received research funding from Abbott Laboratories, AbbVie, Aesca, Centocor, Falk Pharma GmbH, Immundiagnostik, and MSD. AM has nothing to disclose. TH has served as a speaker/consultant/advisory board member for AbbVie, MSD, Takeda, Janssen, Sandoz, Pfizer, Astro Pharma, Falk Pharma, Ferring, and Vifor. TF has served as a speaker/consultant/advisory board member for AbbVie, MSD, Takeda, Pfizer, and Janssen. HF has served as a speaker/consultant for AbbVie, MSD, Astro Pharma, Janssen, and Takeda. PS has served as a speaker for Abbvie, MSD, Janssen, Pfizer, and Falk. OL has nothing to disclose. RP has served as a speaker/consultant/advisory board member for AbbVie, MSD, Takeda, Janssen, and Astro Pharma. WM has served as a speaker/consultant/advisory board member for AbbVie, MSD, Takeda, Janssen, Sandoz, Falk Pharma GmbH, and Ferring. WT has served as a speaker for Takeda, MSD, Janssen, and AbbVie. SA has served as a speaker/consultant/advisory board member for AbbVie, MSD, Janssen, Takeda, and Shire. AS has nothing to disclose. KS has served as a speaker for AbbVie. HV has served as a speaker/consultant/advisory board member for AbbVie, Amgen, Gilead, MSD, Takeda, Janssen, Merck, Ferring, Astro Pharma, Pfizer, Roche, Sandoz, Panaceo, Covidien, Falk Pharma GmbH, and Montavit. CD has served as a speaker/consultant/advisory board member for AbbVie, Astro Pharma, Bayer, Falk Pharma GmbH, Janssen, MSD, Pfizer, Takeda, and Vifor. HH has nothing to disclose. GN has served as a speaker/consultant/advisory board member for AbbVie, MSD, Takeda, Janssen, Sandoz, Pfizer, Astro Pharma, Falk Pharma GmbH, Ferring, and Vifor.

REFERENCES

Author notes