SEARCH

SEARCH BY CITATION

Keywords:

  • antagonists;
  • anticoagulants;
  • bleeding;
  • hospitalization;
  • mortality;
  • vitamin-K

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Funding
  10. Disclosure of Conflict of Interest
  11. References

Background

Bleeding complications are common side effects of vitamin-K antagonist (VKA) therapy. Data on the in-hospital management and outcomes of these bleeding events are scarce and information is mostly derived from trial cohorts.

Objectives

The objective was to collect data on the management and clinical outcome of hospitalizations owing to VKA-related bleeding in real-world practice.

Patients and methods

We performed a multicenter observational cohort study involving 21 secondary and tertiary care hospitals in the administrative district Dresden, Saxony, Germany throughout the year 2005. All consenting patients presenting with VKA-related bleeding complications were included. No exclusion criteria applied. Data were collected at admission, at discharge and at 90 days to evaluate resource consumption, length of hospital stay and risk factors for in-hospital- and 3-month mortality.

Results

Two hundred and ninety patients were included (median age 74 years; 50.7% male). The main indications for VKA therapy were atrial fibrillation (63.4%), prior thromboembolism (18.6%) and mechanical heart valves (11.4%), and most common bleeding localizations were large hematoma (23.1%), upper gastrointestinal (GI) tract (17.9%) and intracranial bleeding (14.1%). On hospital admission, the median International Normalized Ratio (INR) was 3.0 (range 0.9–12.5, interquartile range [IQR] 2.1–3.9). In-hospital mortality was 7.6% with impaired renal function as the most relevant risk factor. At 90 days mortality was 14.1% and 15.3% of survivors were help-dependent.

Conclusions

VKA-related bleeding leading to hospitalization is associated with long hospitalization, relevant resource utilization, high mortality or persistent sequlae. Patient-related factors such as impaired renal function, chronic cardiac or pulmonary disease and dementia are predictive of in-hospital and 3-month mortality.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Funding
  10. Disclosure of Conflict of Interest
  11. References

Oral anticoagulants are potent drugs for preventing thromboembolic complications but have the downside of increased bleeding risks. While new oral anticoagulants are currently being introduced into clinical practice, vitamin K antagonists (VKA) including warfarin and phenprocoumon are still the mainstay of long-term anticoagulation [1-3] based on data from randomized trials in venous thromboembolism (VTE) and atrial fibrillation (AF) [4-7]. Randomized controlled trials have demonstrated an annual risk for major bleeding of around 1.1–3.6% [8-10].

Bleeding-related hospitalizations are common in VKA-treated patients [11] and the clinical impact of VKA-related bleeding is considerable: in a meta-analysis covering 33 studies and 4374 patient-years, Linkins et al.[12] found a case-fatality rate for major bleeding complications of 13.4%. They further demonstrated that bleeding risks decreased after the initial 3-month period (case-fatality-rate of 9.1%). However, this meta-analysis only included VTE treatment studies, whereas the majority of long-term anticoagulated patients are treated for AF [13]. These patients usually are older than VTE patients and have different risk profiles [14]. Furthermore, only randomized controlled trials or prospective cohort studies were included in this meta-analysis and the results may not necessarily reflect the clinical impact of VKA-related bleeding in a real-world setting.

Bleeding complications are likely to influence patients' morbidity and mortality and potentially require significant medical and socioeconomic resources [15, 16]. Data on the management of these bleeding events in the hospital in terms of resource utilization, therapeutic strategies and outcome are scarce. Therefore, we established a multicenter registry in which episodes of hospitalization for VKA-related bleeding complications were documented. Patient characteristics on admission, management of bleeding, resource utilization as well as in-hospital and 3-months outcome were prospectively collected.

Patients and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Funding
  10. Disclosure of Conflict of Interest
  11. References

The study was designed as a multicenter observational cohort study conducted in the former administrative district Dresden, Saxony, Germany throughout the year 2005. At the time the district had 1.7 million residents. Twenty-one secondary and tertiary care hospitals enrolled patients during 2005. Patients were eligible if admitted as a result of a bleeding episode while receiving long-term therapy with oral anticoagulants and no exclusion criteria were applied.

Site and patient selection

Physicians at the participating sites were instructed to fax to our department all reports of hospital admission of patients with bleeding events related to or occurring during VKA therapy. After notification, our study personnel went to the site to inform the patient about the registry and to obtain written informed consent and baseline data. Written informed consent was obtained from the patient or relatives; otherwise, the patient was not included and no data were documented. Episodes of asymptomatic International Normalized Ratio (INR) elevations (i.e. no bleeding complication) leading to hospitalization were initially included in the registry but excluded from the statistical analysis unless bleeding occurred during the hospital stay.

The study was approved by the ethics committee of the Technical University Dresden, Germany.

Data collection

During hospital stay and at discharge, data on patient history, co-morbidity and current hospitalization were collected. Patient interviews, a patient questionnaire, treatment charts, laboratory reports, discharge letters, autopsy reports and death certificates were used as sources. To evaluate bleeding risk factors, all available documents were specifically checked for pre-existing conditions with potential impact on anticoagulation quality or bleeding risk such as diabetes, hypertension, history of stroke, renal insufficiency, dementia, chronic obstructive lung disease, coronary artery disease, malignancy, prior bleeding complications, liver cirrhosis or a prior systemic thromboembolic event. This information was confirmed by patient interview (whenever possible) or review of the discharge documentation and collected using a standardized case report form with check boxes for every item. Of note, risk factors clearly developed during hospitalization (such as renal insufficiency as a direct consequence of acute bleeding) were excluded from risk analysis.

To establish the outcome, a telephone follow-up was performed 90 ± 15 days after hospital discharge.

At hospital admission, the following data were collected: demographic data, indication for VKA therapy, target INR, co-medication and -morbidity, INR values (test method used and last measurement), and bleeding site. Impaired renal function was defined as an impaired renal function with creatinine clearance below 50 mL min−1 (either tested or calculated). Concomitant medication including prescription and over the counter medication was documented for every patient. Every drug was evaluated for potential interactions with VKA according to the Summary of Product Characteristics provided by the manufacturer. Drugs were classified into ‘enhancing VKA action’, ‘reducing VKA action’ or ‘no interaction’.

During hospitalization and at discharge, the following data were collected: management of bleeding (packed red blood cells, oral vitamin K, fresh frozen plasma [FFP], prothrombin complex concentrate [PCC]), surgical/endoscopic or conservative treatment as measures of bleeding-related resource consumption. Furthermore, duration of hospitalization, mortality, extent of disability and presence of residual sequelae were evaluated at discharge.

At day 90 ± 15 follow-up (FU) the following data were collected either by direct contact or telephone: demographic data, morbidity, mortality, extent of disability and residual sequelae.

Residual sequelae were defined as any persisting physical or mental deficit in relationship with the bleeding episode. The extent of disability was evaluated by the need for daily professional health care support either at home or in a dedicated health care institution, or legal acceptance of a degree of disability of > 30%.

Indication for VKA and target range of INR

The indication for VKA use was documented for each patient. If more than one indication was present, patients were allocated according to the following rule: mechanical heart valves (MHV) irrespective of concomitant disease as ‘MHV’; arterial fibrillation (AF) and no MHV present irrespective of concomitant disease as ‘AF'; and venous thromboembolism (VTE) and either MHV or AF present as ‘VTE’. All other indications were allocated to ‘various’.

In general, an INR value between 2.0 and 3.0 was considered to be adequate for patients with VTE, AF, St. Jude aortic MHV, bio-prosthesis, stroke, myocardial infarction and peripheral bypass grafts. For patients with MHV including tilting disk valves and bicuspid valves (mitral or aortic or combined) with additional risk factors (AF, myocardial infarction, low left ventricular ejection fraction) or caged-ball valves and systemic embolism in spite of VKA use an INR between 2.5 and 3.5 was deemed to be adequate.

Statistical analysis

Data were collected using Microsoft Access and transferred to spss (Version 19, IBM, Armonk, NY, USA) for statistical analysis. The following analyzes were pre-specified:

  1. Descriptive statistics for the entire cohort and subgroups of different bleeding localizations with regard to baseline characteristics and resource consumption during in-hospital treatment of the bleeding episode.
  2. In-hospital mortality, total mortality and length of hospital stay for the entire cohort and subgroups of different bleeding localizations.
  3. Risk assessment for potential predictors of in-hospital- and total mortality with regard to baseline characteristics, co-morbidities, concomitant medications, bleeding treatment and different bleeding localizations. Owing to the number of different bleeding locations and small numbers in various groups, risk assessment was performed for groups of bleeding localizations, classified as ‘immediately life-threatening’ (intracranial, gastrointestinal, pulmonal, pericardial, abdominal, ‘retroperitoneal’ and ‘others’).

The following statistical tests were applied: continuous variables were compared using the Student's t-test, univariate comparisons between groups were performed using the Fisher's exact test or Chi-square test where appropriate. INR baseline values across VKA indications and bleeding sites were performed using both the independent-samples median test and the Kruskal–Wallis test. Further tests used were anova, Tamhane II and Tukey-HSD. Factors influencing morbidity and in-hospital and 3-month mortality were identified by univariate regression analysis (significance level of 0.05). Potentially relevant variables then were included in a Cox's proportional hazard analysis (in-hospital-mortality, FU mortality and total mortality with time to event analysis).

Data are shown as absolute, percentage and standard deviation and 95% confidence intervals (95% CI), where appropriate. A P-value of below 0.05 was regarded to be significant.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Funding
  10. Disclosure of Conflict of Interest
  11. References

During the enrolment period, we received 311 notifications of potential subjects with a median of 13 cases per participating hospital (range 3–39). Of these, 11 patients had already been registered as a result of a previous bleeding complication. Another 10 patients presented with asymptomatic supratherapeutic INR without bleeding complications. These patients were also enrolled in our registry but not included in our statistical analysis. Consequently, a cohort of 290 consecutive patients was analyzed.

Patient characteristics at hospital admission

Patients had a median age of 74 ± 9.7 years (range 27–92) and 50.7% were male. Hypertension (55.9%), diabetes (29.3%) and coronary heart disease (29.3%) were the most frequent co-morbidities (Table 1).

Table 1. Baseline characteristics
  n %/Mean ± SD
  1. SD, standard deviation; PAD, peripheral arterial disease; LVEF, left ventricular ejection fraction; VKA, vitamin-K antagonists.

Age29074 years [± 9.7]
Males14750.7%
Primary indication for VKA prescription
Non-valvular atrial fibrillation18463.4%
Venous thromboembolic event5418.6%
Mechanical heart valves3311.4%
Other (e.g. PAD, low LVEF)196.5%
Co-morbidities
Hypertension16255.9%
Diabetes mellitus8529.3%
Coronary artery disease8529.3%
Prior ischemic stroke3712.8%
Impaired renal function3411.7%
Chronic obstructive lung disease248.3%
Dementia124.1%
Malignancy134.5%
Prior gastrointestinal bleeding103.4%
Liver cirrhosis51.7%
Prior systemic thromboembolic event41.4%
Concomitant medication
Enhancing VKA effect8027.6%
Reducing VKA effect124.1%
Platelet inhibitors103.4%
Non-steroidal anti-inflammatory drugs72.4%
Heparin62.1%

The main indications for VKA therapy were AF (184 patients; 63.4%), followed by VTE (54; 18.6%), aortic MHV (26; 9.0%), mitral MHV (3; 1.0%) or aortic and mitral MHV (4; 1.4%). Furthermore, 19 patients (6.5%) received VKA for various other reasons.

Concomitant medication interfering with VKA effect was found in 31.0% of all patients; some patients had more than one. In 80 patients (27.6%), drugs enhancing VKA effect were present: thyroid hormones (n = 19; 6.6%), antibiotics (n = 16; 5.6%); amiodarone (n = 7; 2.4%); miscellaneous (n = 42; 14.6%). Furthermore, 32 patients (11.1%) received medications lowering the VKA effect: spironolactone (n = 18; 6.2%), theophylline (n = 10; 3.5%) and miscellaneous (n = 4; 1.4%). The use of interacting comedication was equally distributed between the different VKA indication groups, P = 0.84. The use of platelet inhibitors (3.8%), non-steroidal anti-inflammatory drugs (2.4%) and heparin (2.1%) was not frequent.

Bleeding localizations are shown in Table 2. On hospital admission, the median INR was 3.0 (range 0.9–12.5, interquartile range [IQR] 2.1–3.9) with therapeutic levels present in only 30.7%, whereas 37.2% had supratherapeutic and 19.0% subtherapeutic INR values. In 13.1% of patients, determination of INR was not performed immediately on admission. The proportion of patients with INR outside or inside the therapeutic range was not different between the different VKA indication groups (independent-samples median test P = 0.717; Kruskal–Wallis test P = 0.989). Also, the mean INR was not statistically significant between the different bleeding localizations (P = 0.447 and 0.496, respectively).

Table 2. Patient characteristics, indication for VKA and management of bleeding according to site of bleeding
Bleeding sitePatient characteristicsVKA-Indication (%)Management of bleeding
N %/290Age ± SDMale %MHV%AFTEOtherHospital stay median (25th/75th percentile)Oral VK n (%)FFP n (%)PCC n (%)RBC n (%)Conservative n (%)Endoscopic n (%)Surgical n (%)
  1. VKA, vitamin-K antagonists; MHV, mechanical heart valves; AF, atrial fibrillation; TE, thromboembolic event; SD, standard deviation; FFP, fresh frozen plasma; PCC, prothrombin complex concentrate, RBC, red blood cells; GI, gastrointestinal.

Upper GI bleeding52 (17.9)73 ± 965.421.257.721.20.011 (7;16)24 (46.2)4 (7.7)10 (19.2)25 (48.1)13 (25)32 (61.5)7 (13.5)
Intracranial bleeding41 (14.1)74 ± 1053.67.365.914.612.215 (7;21)12 (29.3)0 (0)9 (22.0)3 (7.3)25 (61.0)0 (0)16 (39.0)
Haematoma67 (23.1)75 ± 835.84.573.113.49.013 (7;18)27 (40.3)5 (7.5)1 (1.5)13 (19.4)36 (53.7)2 (3.0)29 (43.3)
Epistaxis30 (10.3)70 ± 960.016.746.730.06.76 (4;10)15 (50.0)1 (3.3)2 (6.7)5 (16.7)27 (90.0)2 (6.7)1 (3.3)
Hematuria28 (9.7)74 ± 746.40.071.417.910.79 (5;14)8 (28.6)1 (3.6)4 (14.3)7 (25.0)19 (67.9)3 (10.7)6 (21.4)
Lower GI bleeding19 (6.6)75 ± 947.40.0.78.915.85.310 (8;15)5 (26.3)2 (10.5)1 (5.3)7 (36.8)3 (15.8)13 (68.4)3 (15.8)
Intra-/retroperitoneal20 (6.9)75 ± 1135.015.045.035.05.018 (12;26)12 (60.0)6 (30.0)4 (20.0)10 (50.0)11 (55.0)4 (20.0)5 (25.0)
Hemoptysis6 (2.1)67 ± 19100.033.333.333.30.07 (5;21)3 (50.0)0 (0)1 (16.7)2 (33.3)3 (50.0)3 (50.0)0 (0)
Hemarthrosis3 (1.9)67 ± 1233.366.70.00.033.317 (16;22)0 (0)0 (0)0 (0)1 (33.3)0 (0.0)1 (33.3)2 (66.7)
Other24 (8.3)76 ± 1154.216.775.08.30.09 (7;17)11 (45.8)0 (0)1 (4.2)4 (16.7)16 (66.7)1 (4.2)7 (29.1)
Total290 (100)74 ± 1050.711.463.418.66.611 (7;17)117 (40.3)19 (6.6)33 (11.4)77 (26.6)153 (52.8)61 (21.0)76 (26.2)

In-hospital management

Oral vitamin-K application was the most widely used anticoagulation reversal strategy (n = 117; 40.3%), followed by prothrombin complex concentrate (n = 33; 11.4%; mean dosage 1.570 ± 764 IU) and fresh frozen plasma (n = 19; 6.6%; mean dosage 3.0 ± 1.7 U). Further treatment options included surgical intervention (n = 76; 26.2%), endoscopy (n = 61; 21.0%) or conservative treatment, which mainly consisted of monitoring the patient with or without packed red blood cell (RBC) transfusions and fluid supplementation (n = 153; 52.8%).

Seventy-seven patients (26.6%) received a mean of 3.1 ± 1.8 U of RBC. The necessity for a RBC transfusion was highest with upper GI bleeding (mean 4.3 ± 2.5 U). Rates and numbers of RBC transfusions were not different in surgical vs. endoscopic or conservative bleeding management (one-way anova testing, P = 0.12).

The mean duration of hospitalization was 13 ± 10 days (median 11 days; IQR 7–17, range 1–60; Table 2). Patients without any intervention had a significantly shorter hospitalization than those with either endoscopic or surgical intervention (10 vs. 13 vs. 19 days; P < 0.001).

Outcome of patients

During hospitalization, 22 of the 290 patients (7.6%) died. During the 3-month follow-up, a further 19 patients (6.6%) died. Therefore, total 3-month mortality after a VKA-related bleeding complication was found to be 14.1% (Fig. 1). Intracranial bleeding (ICB) at hospitalization was associated with the highest total mortality (n = 15; 36.6% of all ICB cases), followed by retro-/intraperitoneal bleeding (n = 4; 20.0% of all retro-/intraperitoneal bleeding) and hematuria (n = 5; 17.9% of all hematurias) (Table 3).

Table 3. Outcomes during hospitalization and follow-up
Site of initial bleedingMortality3-month outcome of survivors
In-hospital n (%)During follow-up n (%)Total n (%)Total number of 3-month survivors/All patientsPatients without any sequelae n (% of survivors)Independent with sequelae n (% of survivors)Help-dependent n (% of survivors)
  1. GI, gastrointestinal tract.

Upper GI4/52 (7.7)1/52 (1.9)5/52 (9.6)47/5241/47 (87.2)0/476/47 (12.8)
Intracranial8/41 (19.57/41 (17.1)15/41 (36.6)26/4113/26 (50.0)3/26 (11.5)10/26 (38.5)
Hematoma2/67 (3.0)4/67 (6.0)6/67 (9.0)61/6750/61 (82.0)10/61 (16.4)1/61 (1.6)
Epistaxis0 (0.0)1/30 (3.3)1/30 (3.3)29/3026/29 (89.7)1/29 (3.4)2/29 (6.9)
Hematuria2 (7.1)3 (10.7)5 (17.9)23/2822/23 (95.7)0 (0.0)1/23 (4.3)
Lower GI0 (0.0)1 (5.3)1 (5.3)18/1916/18 (88.9)0 (0.0)2/18 (11.1)
Intra-/retroperitoneal3/20 (15.0)1/20 (5.0)4/20 (20.0)16/2013/16 (81.3)1/16 (6.3)2/16 (12.6)
Hemoptysis0 (0.0)0 (0.0)0 (0.0)6/65/6 (83.3)0 (0.0)1 (16.7)
Hemarthrosis0 (0.0)0 (0.0)0 (0.0)3/32/3 (66.7)0 (0.0)1/3 (33.3)
Other3/24 (12.5)1/24 (4.2)4/24 (16.7)20/2415/20 (75.0)1/20 (5.0)4/20 (20.0)
Total22 (7.6)19 (6.6)41 (14.1)249/290 (85.9%)203 (81.5)8 (3.2)38 (15.3)
image

Figure 1. Kaplan–Meier analysis of cumulative survival of patients after hospitalization for VKA-related bleeding during 90 (± 15)-days of follow-up. Survival is different for different bleeding types (A), for different bleeding managements (B) and for patients with dementia, renal impairment or chronic obstructive lung disease (C).

Download figure to PowerPoint

At day 90 ± 15, 38 of the 249 surviving patients (15.3%) were help-dependent or disabled and another 8 (3.2%) were independent with sequelae (Table 3). Patients surviving intracranial bleeding had the highest rates of disability and sequelae (38.5 and and 11.5%, respectively), followed by hemarthrosis patients with 33.3% remaining disabled.

Patients who died during hospitalization were older than survivors (77.3 ± 6.6 vs. 73.3 ± 9.9; P = 0.014). Likewise, patients still living at day 90 were younger than patients dying during follow-up (73.0 ± 10.05 vs. 77.20 ± 6.61 years; P = 0.01). In-hospital and 3-month mortality was not different between VKA indication groups (Fisher's exact test; P = 0.541 and P = 0.166, respectively).

To evaluate potential risk factors for in-hospital mortality, a uni- and multivariate analysis using Cox's proportional hazard model was performed which identified impaired renal function (hazard ratios [HR] 3.1; 95% CI 1.2–8.3; P = 0.021) and potentially fatal bleeding localizations (HR 2.7; 95% CI 1.1–6.9; P = 0.031) as independent risk factors while surgical or interventional therapy compared with conservative treatment was found to reduce the risk (HR 0.2; 95% CI 0.1–0.6; P = 0.002) (Table 4).

Table 4. Uni- and multivariate analysis using Cox's proportional hazard model to evaluate hazard ratios (HR) of potential risk factors for in-hospital mortality in patients after hospitalization for VKA-related bleeding (= 290). Bleeding types were grouped as potentially fatal bleeding (intracranial, gastrointestinal, intra-abdominal, retroperitoneal and hemoptysis) vs. others
Variablen (%)UnivariateP-valueMultivariateP-value
HR95% CIHR95% CI
  1. VKA, vitamin K antagonists; HR, hazard ratio.

Surgical/interventional vs. conservative treatment137 (47.2)0.2180.083–0.573 0.2220.086–0.5730.002
Impaired renal function34 (11.7)2.4920.960–6.467 3.1421.188–8.3050.021
Potentially fatal bleeding132 (45.5)2.4651.002–6.066 2.7511.095–6.9160.031

Finally, a uni- and multivariate analysis using Cox's proportional hazard model was performed to evaluate HR of potential risk factors for 90 (± 15)-days mortality. In this, dementia (HR 7.3; 95% CI 3.1–16.8; P < 0.001), potentially fatal bleeding (HR 2.7; 95% CI 1.4–5.2; P = 0.003) and chronic obstructive lung disease (COLD; HR 2.9; 95% CI 1.2–7.2; P = 0.017) were found to be independent risk factors whereas surgical or interventional therapy compared with conservative treatment (HR 0.4; 95%-CI0.2–0.9; P = 0.015) was found to reduce the risk of 3-month mortality (Table 5).

Table 5. Uni- and multivariate analysis using Cox's proportional hazard model to evaluate hazard ratios (HR) of potential risk factors for 90(± 15)-days mortality in patients after hospitalization for VKA-related bleeding (= 290). Bleeding types were grouped as potentially fatal bleeding (intracranial, gastrointestinal, intra-abdominal, retroperitoneal and hemoptysis) vs. others
Variablen (%)UnivariateP-valueMultivariateP-value
HR95% CIHR95% CI
  1. VKA, vitamin K antagonists; HR, hazard ratio; COLD, chronic obstructive lung disease.

Dementia12 (4.1)6.1502.716–13.926< 0.0017.2633.137–16.817< 0.001
Potentially fatal bleeding132 (45.5)2.6471.079–6.4920.0332.7081.414–5.1880.003
Surgical/interventional vs. conservative treatment137 (47.2)0.5440.285–1.0380.0650.4370.224–0.8530.015
COLD24 (8.3)2.0460.860–4.8660.1052.9551.212–7.2060.017

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Funding
  10. Disclosure of Conflict of Interest
  11. References

In VKA-treated patients hospitalizations due to bleeding complications are common [11], but little is known about the distribution of bleeding localizations, bleeding management and outcome in real-world practice. In fact, in the meta-analysis by Linkins et al., only 13 major bleeding events were derived from observational studies [12]. Therefore, our data on 290 VKA patients with bleeding-related hospitalizations add relevant information to the field. In our cohort, around 40% of patients suffered from GI bleeding or intracranial haemorrhage, both of which have been shown to be associated with high treatment costs [16]. We found a considerable length of hospital stay (median 11; IQR 7–17 days) in patients hospitalized for VKA-related bleedings and a high necessity of blood product transfusions (fresh frozen plasma, PCC and RBC) to correct both blood loss and anticoagulant effect. Furthermore, surgical (23.6%) and endoscopic procedures (22.6%) were often necessary to treat bleeding events. In spite of all treatment efforts, only half of the patients had no sequelae of the bleeding event after 3 months and 7.6% of all patients died during hospital stay. Furthermore, at 3 months the total mortality rate was found to be 14.2%, which is in agreement with findings from a recently reported large cohort study, which found a short-term mortality of 18% after VKA-related hospitalization [17]. Consequently, the mortality after hospitalization for VKA-related bleeding is higher than the reported mortality of patients of an equal age hospitalized for other (non-malignant) reasons such as decompensated heart failure or severe community-acquired pneumonia [18, 19].

Our data confirmed previous studies which found high rates of concomitant medications enhancing VKA effects in this population [13, 20, 21], which has been shown to increase bleeding-related treatment costs [13, 20, 21]. As novel oral anticoagulants (NOAC) have less of an interaction with concomitant medication and have been found to reduce the risk of major and especially intracranial bleedings [22-26], future studies will need to evaluate rates, distribution pattern, treatment costs and outcome of bleedings during NOAC treatment in real-world practice.

Total mortality

Mortality in our cohort was particularly high among patients with ICB (36.6%) and substantially lower in patients with upper GI bleeding, hematoma or hematuria. The rate of mortality among patients with ICB is in line with published data, where ICB was reported to have a mortality of 30–46% [12, 27, 28]. Mortality was especially high in the first 20 days (n = 21; 92.8% cumulative survival), probably caused by very severe bleeding episodes or a critically impaired physical condition of the patient with reduced compensation reserve. This first phase is followed by a constant decline of the survival curve until day 105, which probably reflects mortality of recurrent bleeding [29], of bleeding sequelae or hospitalization-acquired complications.

Factors related to in-hospital and total mortality

Impaired renal function and chronic obstructive lung disease were identified as an independent predictor of mortality with HR of 3.1 and 2.9, which is consistent with established data [30, 31] and directs attention to the severe comorbidities in our cohort. As a result of these underlying diseases, compensation reserve of these patients is small and mortality remains high for weeks following discharge after VKA-bleeding hospitalizations.

Naturally, bleeding complications at sites of immediate risk of fatality such as intracerebral, gastrointestinal or intra-abdominal bleedings were found to increase the risk of in-hospital mortality (HR 2.7). On the other hand, active control of bleeding using surgical or interventional approaches was found to significantly reduce the risk of in-hospital mortality (HR 0.2) compared with conservative treatment, which indicates that, whenever possible, aggressive treatment with surgical and interventional techniques will reduce the bleeding risk immediately.

In our cohort, 61% of patients with intracranial bleeding, 55% of patients with intra- or retroperitoneal bleeding and 68% of patients with hematuria were treated without surgical or interventional therapies and had a high mortality. In contrast, gastrointestinal bleeding was the most common bleeding site (24.4%) and 77% of these GI bleedings were treated by endoscopy or surgery. Patients with gastrointestinal bleeding had a mortality rate of 8.5%, which still is high but much lower than the average mortality of our entire cohort. It has to be concluded that active treatment of VKA-related bleedings was more often considered in gastrointestinal bleeding than in other bleeding localizations, which might have been because of the availability of emergency GI endoscopy in most settings, whereas neurosurgical, urological or other interventional strategies are not generally available for emergency treatments. Consequently, early referral strategies to specialized centers should be considered in patients with potentially fatal, non-GI bleedings to reduce mortality.

In our evaluation, dementia (HR 7.3) was found to be the most important risk factor for total mortality, probably owing to lower patient compliance, the potential of unintentional overdosing of VKA and food–drug as well as drug–drug interactions [32, 33]. A recent large meta-analysis of long-term VKA use demonstrated that patients with dementia are less likely to receive VKA therapy in spite of an indication, possibly indicating that physicians expect adverse reactions such as major bleeding complications in this population [13]. In addition, dementia is more common in older patients and anticoagulant effect is increased in old patients with enhanced responses to small changes in dose of VKA [34-36]. Finally, a retrospective study by van Deelen et al. [37] identified dementia with a Mini-Mental State Examination of < 23 to be independently associated with an inadequate INR control, mainly because of an increased number of supra-therapeutic INR values.

Limitations

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Funding
  10. Disclosure of Conflict of Interest
  11. References

We present a large case series with patients hospitalized for bleeding complications under long-term VKA use. However, as a result of the design of our study a number of potential limitations need to be addressed.

First of all, the enrolment of patients relied entirely on the attending physicians at the participating hospital, introducing the potential for selection bias. Furthermore, while every effort was made to collect all relevant data on patient history from the hospital charts and the patient's family doctor, there were missing data in the history. However, all data regarding the bleeding event, management and resource utilization were collected prospectively at the time patients consented to participate. Therefore, the data sets on management and outcome are complete.

Participating hospitals contributed unequally to recruitment, so we cannot exclude a bias in reporting bleeding episodes. Furthermore, the total number of bleeding complications (being treated on an outpatient basis or leading to immediate death outside of hospital) and the total number of patients with long-term VKA treatment in the community is unknown. Therefore, the study does not allow an estimate of the incidence of bleeding.

It proved difficult to gain exact information about the duration of anticoagulation before the bleeding event. This is of importance because bleeding complications are more frequent within the first months of treatment [12, 38]. However, the spectrum of indications for VKA treatment (predominantly AF and MHV) makes long-term VKA treatment necessary and many patients develop additional comorbidities and bleeding risk factors such as concomitant therapies over the years. Therefore, also late bleeding complications are common in VKA patients [13, 39, 40].

The use of over-the-counter medication containing aspirin is frequent in the general population. If available, this information was included in our assessment of concomitant medications. However, we cannot exclude that the use of aspirin was more frequent than reported in our analysis.

Information on residual sequelae and extent of disability were derived from patient's hospital charts or via patient contact and not objectively confirmed and, therefore, may be subject to reporting bias. An autopsy in case of a fatal event was not routinely performed. Consequently, the exact cause of death could not be established in every case.

Finally, information on resumption or cessation of anticoagulant treatment was not available for all patients. This is a relevant limitation, as permanent cessation of VKA after GI tract bleeding has recently been shown to be associated with increased 90-days mortality increased risk for thrombosis [41]. Consequently, the high mortality in our entire cohort as well as in the subgroup of patients without potentially fatal bleeding (such as epistaxis, hematoma and hematuria) could be at least in part attributable to thromboembolic events. Furthermore, it is possible that dementia's association with overall mortality is at least in part associated with permanent discontinuation of VKA treatment in this difficult-to-anticoagulate subgroup of patients.

However, the large set of data derived from this multicentric registry provides new insights into the severity and the medical, social and economic burden of VKA-related bleeding complications in real-world practice. The prospective design of our study, the completeness and the consistency of our data are a significant strength and, from our point of view, outweigh the above addressed limitations.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Funding
  10. Disclosure of Conflict of Interest
  11. References

VKA-related bleeding episodes leading to hospitalization impact heavily on patient's health in terms of morbidity and mortality. Furthermore, duration of hospitalization and resource utilization to treat VKA-related bleeding constitute a relevant economic burden. Patient-related factors such as uremia, COLD, age beyond 70 years, chronic ischemic heart disease and dementia are predictive of poor in-hospital and 3-month outcome.

Funding

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Funding
  10. Disclosure of Conflict of Interest
  11. References

The conduct of the study was financed by the research funds of our department and partly by a research grant by ROCHE Diagnostics Ltd., who provided funding for a documentation assistant.

Disclosure of Conflict of Interest

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Funding
  10. Disclosure of Conflict of Interest
  11. References

No competing interests with respect to the current study have to be declared. Kai Halbritter, Jan Beyer-Westendorf and Sebastian M. Schellong received honoraria and research funding from a number of pharmaceutical companies producing low molecular heparins and other anticoagulants but declared no conflict of interest with regard to the presented data.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Funding
  10. Disclosure of Conflict of Interest
  11. References
  • 1
    Ageno W, Gallus AS, Wittkowsky A, Crowther M, Hylek EM, Palareti G. Oral anticoagulant therapy: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012; 141: e44S88S.
  • 2
    Kearon C, Akl EA, Comerota AJ, Prandoni P, Bounameaux H, Goldhaber SZ, Nelson ME, Wells PS, Gould MK, Dentali F, Crowther M, Kahn SR. Antithrombotic therapy for VTE disease: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012; 141: e419S94S.
  • 3
    You JJ, Singer DE, Howard PA, Lane DA, Eckman MH, Fang MC, Hylek EM, Schulman S, Go AS, Hughes M, Spencer FA, Manning WJ, Halperin JL, Lip GY. Antithrombotic therapy for atrial fibrillation: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012; 141: e531S75S.
  • 4
    Camm AJ, Kirchhof P, Lip GY, Schotten U, Savelieva I, Ernst S, van Gelder IC, Al-Attar N, Hindricks G, Prendergast B, Heidbuchel H, Alfieri O, Angelini A, Atar D, Colonna P, De Caterina R, De Sutter J, Goette A, Gorenek B, Heldal M, et al. Guidelines for the management of atrial fibrillation: the Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC). Europace 2010; 12: 1360420.
  • 5
    Hart RG, Pearce LA, Aguilar MI. Adjusted-dose warfarin versus aspirin for preventing stroke in patients with atrial fibrillation. Ann Intern Med 2007; 147: 5902.
  • 6
    Aguilar MI, Hart R, Pearce LA. Oral anticoagulants versus antiplatelet therapy for preventing stroke in patients with non-valvular atrial fibrillation and no history of stroke or transient ischemic attacks. Cochrane Database Syst Rev 2007; 3: CD006186.
  • 7
    Hart RG, Pearce LA, Aguilar MI. Meta-analysis: antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation. Ann Intern Med 2007; 146: 85767.
  • 8
    van der Meer FJ, Rosendaal FR, Vandenbroucke JP, Briet E. Assessment of a bleeding risk index in two cohorts of patients treated with oral anticoagulants. Thromb Haemost 1996; 76: 126.
  • 9
    Palareti G, Leali N, Coccheri S, Poggi M, Manotti C, D'Angelo A, Pengo V, Erba N, Moia M, Ciavarella N, Devoto G, Berrettini M, Musolesi S. Bleeding complications of oral anticoagulant treatment: an inception-cohort, prospective collaborative study (ISCOAT). Italian study on complications of oral anticoagulant therapy. Lancet 1996; 348: 4238.
  • 10
    Landefeld CS, Beyth RJ. Anticoagulant-related bleeding: clinical epidemiology, prediction, and prevention. Am J Med 1993; 95: 31528.
  • 11
    Vitry AI, Roughead EE, Ramsay EN, Preiss AK, Ryan P, Gilbert AL, Caughey GE, Shakib S, Esterman A, Zhang Y, McDermott RA. Major bleeding risk associated with warfarin and co-medications in the elderly population. Pharmacoepidemiol Drug Saf 2011; 20: 105763.
  • 12
    Linkins LA, Choi PT, Douketis JD. Clinical impact of bleeding in patients taking oral anticoagulant therapy for venous thromboembolism: a meta-analysis. Ann Intern Med 2003; 139: 893900.
  • 13
    Neidecker M, Patel AA, Nelson WW, Reardon G. Use of warfarin in long-term care: a systematic review. BMC Geriatr 2012; 12: 14.
  • 14
    Friberg L, Rosenqvist M, Lip GY. Net clinical benefit of warfarin in patients with atrial fibrillation: a report from the Swedish atrial fibrillation cohort study. Circulation 2012; 125: 2298307.
  • 15
    Hamby L, Weeks WB, Malikowski C. Complications of warfarin therapy: causes, costs, and the role of the anticoagulation clinic. Eff Clin Pract 2000; 3: 17984.
  • 16
    Ghate SR, Biskupiak J, Ye X, Kwong WJ, Brixner DI. All-cause and bleeding-related health care costs in warfarin-treated patients with atrial fibrillation. J Manag Care Pharm 2011; 17: 67284.
  • 17
    Gomes T, Mamdani MM, Holbrook AM, Paterson JM, Hellings C, Juurlink DN. Rates of hemorrhage during warfarin therapy for atrial fibrillation. CMAJ 2013; 185: E1217. doi:10.1503/cmaj.121218
  • 18
    Sanchez CE, Richards DR. Contemporary in-hospital management strategies for acute decompensated heart failure. Cardiol Rev 2011; 19: 1229.
  • 19
    Venditti M, Falcone M, Corrao S, Licata G, Serra P. Outcomes of patients hospitalized with community-acquired, health care-associated, and hospital-acquired pneumonia. Ann Intern Med 2009; 150: 1926.
  • 20
    Suh DC, Nelson WW, Choi JC, Choi I. Risk of hemorrhage and treatment costs associated with warfarin drug interactions in patients with atrial fibrillation. Clin Ther 2012; 34: 156982.
  • 21
    Baillargeon J, Holmes HM, Lin YL, Raji MA, Sharma G, Kuo YF. Concurrent use of warfarin and antibiotics and the risk of bleeding in older adults. Am J Med 2012; 125: 1839.
  • 22
    Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, Breithardt G, Halperin JL, Hankey GJ, Piccini JP, Becker RC, Nessel CC, Paolini JF, Berkowitz SD, Fox KA, Califf RM. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011; 365: 88391.
  • 23
    Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, Pogue J, Reilly PA, Themeles E, Varrone J, Wang S, Alings M, Xavier D, Zhu J, Diaz R, Lewis BS, Darius H, Diener HC, Joyner CD, Wallentin L. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009; 361: 113951.
  • 24
    Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, Hanna M, Al-Khalidi HR, Ansell J, Atar D, Avezum A, Bahit MC, Diaz R, Easton JD, Ezekowitz JA, Flaker G, Garcia D, Geraldes M, Gersh BJ, Golitsyn S, Goto S, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2011; 365: 98192.
  • 25
    Bauersachs R, Berkowitz SD, Brenner B, Buller HR, Decousus H, Gallus AS, Lensing AW, Misselwitz F, Prins MH, Raskob GE, Segers A, Verhamme P, Wells P, Agnelli G, Bounameaux H, Cohen A, Davidson BL, Piovella F, Schellong S. Oral rivaroxaban for symptomatic venous thromboembolism. N Engl J Med 2010; 363: 2499510.
  • 26
    Buller HR, Prins MH, Lensin AW, Decousus H, Jacobson BF, Minar E, Chlumsky J, Verhamme P, Wells P, Agnelli G, Cohen A, Berkowitz SD, Bounameaux H, Davidson BL, Misselwitz F, Gallus AS, Raskob GE, Schellong S, Segers A. Oral rivaroxaban for the treatment of symptomatic pulmonary embolism. N Engl J Med 2012; 366: 128797.
  • 27
    Hylek EM, Singer DE. Risk factors for intracranial hemorrhage in outpatients taking warfarin. Ann Intern Med 1994; 120: 897902.
  • 28
    Dowlatshahi D, Butcher KS, Asdaghi N, Nahirniak S, Bernbaum ML, Giulivi A, Wasserman JK, Poon MC, Coutts SB. Poor prognosis in warfarin-associated intracranial hemorrhage despite anticoagulation reversal. Stroke 2012; 43: 18127.
  • 29
    Lee JK, Kang HW, Kim SG, Kim JS, Jung HC. Risks related with withholding and resuming anticoagulation in patients with non-variceal upper gastrointestinal bleeding while on warfarin therapy. Int J Clin Pract 2012; 66: 648.
  • 30
    Lip GY, Frison L, Halperin JL, Lane DA. Comparative validation of a novel risk score for predicting bleeding risk in anticoagulated patients with atrial fibrillation: the HAS-BLED (Hypertension, Abnormal Renal/Liver Function, Stroke, Bleeding History or Predisposition, Labile INR, Elderly, Drugs/Alcohol Concomitantly) score. J Am Coll Cardiol 2011; 57: 17380.
  • 31
    Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJ, Lip GY. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest 2010; 138: 1093100.
  • 32
    Levy RH, Collins C. Risk and predictability of drug interactions in the elderly. Int Rev Neurobiol 2007; 81: 23551.
  • 33
    Monastero R, Camarda R, Camarda C. Potential drug-drug interaction between duloxetine and acenocoumarol in a patient with Alzheimer's disease. Clin Ther 2007; 29: 27069.
  • 34
    Schwartz JB, Kane L, Moore K, Wu AH. Failure of pharmacogenetic-based dosing algorithms to identify older patients requiring low daily doses of warfarin. J Am Med Dir Assoc 2011; 12: 6338.
  • 35
    Henderson MC, White RH. Anticoagulation in the elderly. Curr Opin Pulm Med 2001; 7: 36570.
  • 36
    Hylek EM. Oral anticoagulants. Pharmacologic issues for use in the elderly. Clin Geriatr Med 2001; 17: 113.
  • 37
    van Deelen BA, van den Bemt PM, Egberts TC, van‘t Hoff A, Maas HA. Cognitive impairment as determinant for sub-optimal control of oral anticoagulation treatment in elderly patients with atrial fibrillation. Drugs Aging 2005; 22: 35360.
  • 38
    Schulman S, Beyth RJ, Kearon C, Levine MN. Hemorrhagic complications of anticoagulant and thrombolytic treatment: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest 2008; 133: 257S98S.
  • 39
    Reynolds MW, Fahrbach K, Hauch O, Wygant G, Estok R, Cella C, Nalysnyk L. Warfarin anticoagulation and outcomes in patients with atrial fibrillation: a systematic review and metaanalysis. Chest 2004; 126: 193845.
  • 40
    Merli G, Weitz HH. The decision to anticoagulate: assessing whether benefits outweigh the risks for patients with atrial fibrillation. Clin Cardiol 2004; 27: 31320.
  • 41
    Witt DM, Delate T, Garcia DA, Clark NP, Hylek EM, Ageno W, Dentali F, Crowther MA. Risk of thromboembolism, recurrent hemorrhage, and death after warfarin therapy interruption for gastrointestinal tract bleeding. Arch Intern Med 2012; 172: 18.