J Thromb Thrombolysis DOI 10.1007/s11239-013-1011-z
Bleeding complications related to warfarin treatment: a descriptive register study from the anticoagulation clinic at Helsingborg Hospital Monica Navgren • Johan Forsblad Mattias Wieloch
•
Ó Springer Science+Business Media New York 2013
Abstract The most common indication for treatment with warfarin is the prevention of ischemic stroke in patients with atrial fibrillation. Time in therapeutic range (TTR) is an important tool to evaluate the quality of anticoagulation treatment. The aim of this study was to investigate the quality of treatment and the incidence of bleeding complications in patients on warfarin treatment treated by the anticoagulation clinic in Helsingborg. This is the first study that has specifically focused on the spontaneous reporting of bleeding complications in a real-world population. A total of 4,400 patients with a total of 8,394 patient years were registered, in the database Journalia AVK, during the time period November 1, 2007 to November 1, 2010. The mean age was 72 years. TTR was 73.3 % for the whole population. 421 patients suffered from haemorrhagic events. The frequency of major and fatal bleedings and intracranial haemorrhage (ICH) were 1.6, 0.2 and 0.5 % per patient-year, respectively. A correlation between age and severe bleeding (major, fatal and ICH) (p = 0.003) was seen, but no correlation between gender and severe bleeding (p = 0.27). In 60 out of 455 bleeding events the complication had been reported to the anticoagulation clinic. At the anticoagulation clinic in
M. Navgren (&) J. Forsblad Department of Internal Medicine, Helsingborg Hospital, Helsingborg, Sweden e-mail: [email protected] M. Wieloch Department of Emergency Medicine, Ska˚ne University Hospital Malmo¨, Malmo¨, Sweden M. Wieloch Department of Clinical Sciences, Lund University, Lund, Sweden
Helsingborg the quality of warfarin treatment is good compared to previous results described in the literature, with regards to bleeding complications and efficacy. However, in our study, we confirm that the spontaneous reporting of bleeding complications related to warfarin is inadequate, and that review of patient records is needed to assure proper follow-up. Keywords Warfarin TTR Atrial fibrillation Spontaneous reporting Bleeding complications
Background The most common indication for long-term treatment with warfarin is the prevention of ischemic stroke in patients with atrial fibrillation [1–3]. Other accepted indications include the treatment and prevention of venous thromboembolism (VTE), acquired heart disease with special risk of embolism in addition to atrial fibrillation, such as cardiomyopathy, large transmural myocardial infarction, and valvular- and vascular prostheses [2–4]. About 1 % of the population in Europe are treated with warfarin and in recent years, this proportion has increased [4–6]. In Sweden the proportion of warfarin treated population is about 1.5 % [2, 3]. Established risk factors for bleeding are advanced age, hypertension, cerebrovascular disease, stroke, alcoholism, and malignancy [7]. Reports of the incidence of bleeding complications from warfarin vary. This is partially due to the absence of a standard definition of major bleedings [4, 5] making assessments of bleeding complications difficult to compare. Another reason is a lack of reporting of haemorrhagic complications [4, 7, 8]. According to Peter Svensson, MD, PhD, and associate professor at the
123
M. Navgren et al.
Department for coagulation disorders, University of Lund, Sweden, only 0.2–0.4 % of major bleeding events are traditionally reported among warfarin-treated patients in Sweden. The most serious bleeding complication is intracranial haemorrhage (ICH) and mortality in spontaneous ICH in warfarin-treated patients is reported to be as high as 67 % [9]. Several factors increase the risk of bleeding. Studies have demonstrated a correlation between the intensity of warfarin therapy and the risk of bleeding [10–13]. Time in therapeutic range (TTR) has become an important tool to evaluate the quality of the anticoagulation treatment. A TTR of 70–80 % is desirable [13] and according to European guidelines, TTR should be above 70 % [14]. High levels of TTR, indicating tight control of anticoagulation treatment, reduces the risk of both vascular events (thrombosis) and major haemorrhage [10, 11, 15, 16]. A study from Sweden has demonstrated good therapeutic control with a high TTR of 76.2 % [11]. The multicenter RELY trial [17] demonstrated a TTR of 64 %, similar to previous published studies [18], and a subgroup analysis of the RELY trial demonstrated that a high TTR ([72.4 %) correlated to a reduced risk of both thrombo-embolic complications and major bleeding [19]. Recently a study demonstrated an increasing benefit from anticoagulation treatment with increasing TTR, and at TTR’s above 58 % it is reasonable to expect that a patient benefits significantly from anticoagulation treatment [10]. The purpose of this study was to investigate the frequencies and types of bleeding complications related to warfarin treatment at an anticoagulation clinic in southern Sweden. Since there are differences in healthcare contact patterns between different age groups there could be differences in reporting of bleeding events with regard to age. Also, physicians could be negligible of elderly patients and more prone to report bleedings in younger patients. Hence, we also wanted to investigate and to compare the proportion of patients with bleeding complications spontaneously reported to the anticoagulation clinic in different age groups with regards to both age and gender.
Materials and methods Between January 1 1995 and October 31 2010 all patients treated with warfarin in the catchment area of Helsingborg, were enrolled in a database named Journalia AVK. On November 1 2010, AuriculA, a national prospective quality register of anticoagulation, was started at Helsingborg Hospital. In connection to the change of registry system the aim of this study was to investigate the incidence of bleeding complications in patients on warfarin treatment being assessed and treated by the anticoagulation
123
clinic in Helsingborg. In the Helsingborg catchment area all patients treated with warfarin are registered in the database and prescriptions and follow-up are managed by the anticoagulation clinic. The study is a descriptive, retrospective registry study and includes all 4,400 patients registered in Journalia AVK during the time period November 1, 2007 to November 1, 2010. A list of clearly defined diagnoses of bleeding was made to select patients affected by a bleeding complication during the time period (Appendix). Patient data from Journalia AVK was then cross-matched with patient hospital records from a regional database and finally matched with the previously mentioned bleeding diagnoses. A systematic review of all medical records was done by the physician listing age, gender, indication of warfarin treatment, bleeding diagnosis and extent of bleeding according to ISTH definitions of bleeding [20]. Furthermore, bleeding events were divided into fatal, major, intracranial and minor bleeding. INR at the time of bleeding and duration of warfarin treatment was registered. Bleeding complications were investigated if they had been reported to the anticoagulation clinic and if the patient had died independent of the bleeding. During the systematic review of all patient records 215 patients were excluded from the study. These patients were excluded due to bleeding episodes that occurred, when the patient was no longer using warfarin or before warfarin treatment had commenced, during the observed time frame. Four cases of bleeding had been reported to the clinic but not originally found during the search among the medical records. These cases were the result of an incorrect diagnosis of bleeding recorded in the medical records. Those four patients were included in the study before statistical analysis. An additional five patients were included in our study. These patients had been reported to the anticoagulation clinic but missed in the search among the hospital records since the bleeding episodes were registered as other diagnoses than those used in our study. Two of these patients suffered from major gastrointestinal bleeding and three from intracranial bleeding associated with trauma. Time in therapeutic range was calculated according to F.R. Roosendaal’s algorithm with linear interpolation [21]. Only patients with target INR 2.0–3.0 were included in the analysis of TTR. As a result 440 patients with other target INR levels were excluded from the analysis of TTR. The age of any patient in the Journalia AVK population was defined as the age of that patient on May 1, 2009. This date is the mid-point of the study period. If a patient had more than one bleeding episode only the first was used for TTR calculation but all events were used to calculate rates of bleeding complications. Bleeding complications were converted to rates (per cent per patient-
Bleeding complications related to warfarin treatment
year) and tested for differences in patient characteristics. In all patients with bleeding complications only INR values at the time of the bleeding episode was used for statistical analysis. Mean INR was calculated based on 316 samples (138 were missing). Among the major and fatal bleedings the proportion of INR values within the therapeutic range 2.0–3.0, at the time of bleeding was calculated based on 161 samples (18 were missing). Differences in age and gender were tested using t test. Student‘s t test was used for testing age and bleeding complications. Differences in gender were tested using Chi Square test. The p value threshold for significance was \ 0.05. For statistical analysis SPSS v 20.0 was used.
Table 1 Incidence of bleeding events per patient-year
Results
ICH intracranial hemorrhage
A total number of 4,400 patients (1,803 female and 2,597 male) with a total of 8,394 patient-years were registered in Journalia AVK during the observation period. A total of 123,117 INR samples were registered. 79,793 of the samples were within the target range 2–3 (64.8 %). 37.1 % of the patients had an episode with INR \ 1.5, 31.5 % with INR [ 4.0 and 4.9 % with INR [ 6. TTR was 73.3 % for the whole population during the study period. The mean age in the whole population was 72 years. The mean age was higher in women (74) than in men (71). The most common indication of treatment with warfarin was atrial fibrillation (61.3 %) and VTE (23.8 %). Bleeding complications A total of 421 (9.6 %) patients treated with warfarin suffered a haemorrhagic event during the study (Table 1). The majority of bleeding events (n = 272) were minor bleedings. A total of 138 major bleedings were registered i.e. 1.6 % per patient-year (Table 1). The most frequent major bleeding was gastrointestinal (n = 64). Other major bleedings were haematuria, intraocular, hematoma, intramuscular with and without compartment, intra-articular, epistaxis, anaemia requiring transfusion, intra-abdominal, retroperitoneal and undefined bleeding. 18 patients suffered a fatal bleeding (4 %) i.e. 0.2 % per patient-year. ICH was the main cause of fatal bleeding (12/18). The other fatal bleedings were gastrointestinal, hematuria, deep bleeding in the hip region, hemoptysis and retroperitoneal bleeding. An additional 27 patients suffered from ICH without a fatal outcome with a total incidence of 0.5 % per patient-year. Thirty-four of a total of 421 patients suffered a second bleeding event of which 30 were minor bleedings and 4 were major bleedings. During the observation period 93 patients died of causes other than warfarin-related bleeding.
Bleeding diagnosis
Bleeding events (n)
Risk of bleeding ( % per patient-year)
Total bleeding episodes
455
5.4
18
0.2
Fatal ICH
12
Gastrointestinal bleeding
2
Other bleeding
4
Major bleeding
138
1.6
ICH (including fatal ICH) Minor
39 272
0.5 3.2
The total number of patient-years during the study was 8,394. Bleeding events could be recognised as both fatal, and intracranial (ICH)
Table 2 Distribution of bleeding diagnoses between genders Bleeding diagnosis
Female (n)
Male (n)
p-value
Total bleeding episodes
158
309
0.017*
6
12
ICH
5
7
Gastrointestinal bleeding
1
1
Other bleeding
0
4
Major bleeding
44
94
ICH (including fatal ICH)
22
17
Minor bleeding
86
186
Fatal
0.002**
Bleeding events could be recognised as both fatal and intracranial (ICH). The total number of females and males that suffered one or more bleeding events was 145 and 276 respectively. p values indicate differences between genders. There was a significant difference in mean age at bleeding between the sexes (female 76.9 years vs. male 74.2 years; p = 0.02), and female patients suffered more ICH than men (p = 0.002) ICH intracranial hemorrhage * **
Level of significance p \ 0.05 Level of significance p \ 0.01
Age and gender Among the 421 patients who suffered a bleeding, 145 were female and 276 were male (Table 2). In total the females suffered 158 bleeding events and the men 309. The mean age in the population with bleeding complications was 75 years. The women were significantly older than the men (p = 0.024) (Table 2). The distribution of bleeding complications between the genders is seen in Table 2. When comparing ICH, fatal and major bleedings with minor bleedings there was no significant difference between genders (p = 0.27). When comparing ICH against
123
M. Navgren et al.
Bleedingevents and age 0,8 fatal 0,7
Events per patient-year
Fig. 1 Distribution of bleeding events in different age groups. A significant correlation between age and severe bleeding events (fatal, major and ICH); p = 0.03, is seen. Bleeding events in the figure includes all events, also those not reported to the anticoagulation clinic
major ICH
0,6 0,5 0,4 0,3 0,2 0,1 0
Bleeding complications related to warfarin treatment: a descriptive register study from the anticoagulation clinic at Helsingborg Hospital Monica Navgren • Johan Forsblad Mattias Wieloch
•
Ó Springer Science+Business Media New York 2013
Abstract The most common indication for treatment with warfarin is the prevention of ischemic stroke in patients with atrial fibrillation. Time in therapeutic range (TTR) is an important tool to evaluate the quality of anticoagulation treatment. The aim of this study was to investigate the quality of treatment and the incidence of bleeding complications in patients on warfarin treatment treated by the anticoagulation clinic in Helsingborg. This is the first study that has specifically focused on the spontaneous reporting of bleeding complications in a real-world population. A total of 4,400 patients with a total of 8,394 patient years were registered, in the database Journalia AVK, during the time period November 1, 2007 to November 1, 2010. The mean age was 72 years. TTR was 73.3 % for the whole population. 421 patients suffered from haemorrhagic events. The frequency of major and fatal bleedings and intracranial haemorrhage (ICH) were 1.6, 0.2 and 0.5 % per patient-year, respectively. A correlation between age and severe bleeding (major, fatal and ICH) (p = 0.003) was seen, but no correlation between gender and severe bleeding (p = 0.27). In 60 out of 455 bleeding events the complication had been reported to the anticoagulation clinic. At the anticoagulation clinic in
M. Navgren (&) J. Forsblad Department of Internal Medicine, Helsingborg Hospital, Helsingborg, Sweden e-mail: [email protected] M. Wieloch Department of Emergency Medicine, Ska˚ne University Hospital Malmo¨, Malmo¨, Sweden M. Wieloch Department of Clinical Sciences, Lund University, Lund, Sweden
Helsingborg the quality of warfarin treatment is good compared to previous results described in the literature, with regards to bleeding complications and efficacy. However, in our study, we confirm that the spontaneous reporting of bleeding complications related to warfarin is inadequate, and that review of patient records is needed to assure proper follow-up. Keywords Warfarin TTR Atrial fibrillation Spontaneous reporting Bleeding complications
Background The most common indication for long-term treatment with warfarin is the prevention of ischemic stroke in patients with atrial fibrillation [1–3]. Other accepted indications include the treatment and prevention of venous thromboembolism (VTE), acquired heart disease with special risk of embolism in addition to atrial fibrillation, such as cardiomyopathy, large transmural myocardial infarction, and valvular- and vascular prostheses [2–4]. About 1 % of the population in Europe are treated with warfarin and in recent years, this proportion has increased [4–6]. In Sweden the proportion of warfarin treated population is about 1.5 % [2, 3]. Established risk factors for bleeding are advanced age, hypertension, cerebrovascular disease, stroke, alcoholism, and malignancy [7]. Reports of the incidence of bleeding complications from warfarin vary. This is partially due to the absence of a standard definition of major bleedings [4, 5] making assessments of bleeding complications difficult to compare. Another reason is a lack of reporting of haemorrhagic complications [4, 7, 8]. According to Peter Svensson, MD, PhD, and associate professor at the
123
M. Navgren et al.
Department for coagulation disorders, University of Lund, Sweden, only 0.2–0.4 % of major bleeding events are traditionally reported among warfarin-treated patients in Sweden. The most serious bleeding complication is intracranial haemorrhage (ICH) and mortality in spontaneous ICH in warfarin-treated patients is reported to be as high as 67 % [9]. Several factors increase the risk of bleeding. Studies have demonstrated a correlation between the intensity of warfarin therapy and the risk of bleeding [10–13]. Time in therapeutic range (TTR) has become an important tool to evaluate the quality of the anticoagulation treatment. A TTR of 70–80 % is desirable [13] and according to European guidelines, TTR should be above 70 % [14]. High levels of TTR, indicating tight control of anticoagulation treatment, reduces the risk of both vascular events (thrombosis) and major haemorrhage [10, 11, 15, 16]. A study from Sweden has demonstrated good therapeutic control with a high TTR of 76.2 % [11]. The multicenter RELY trial [17] demonstrated a TTR of 64 %, similar to previous published studies [18], and a subgroup analysis of the RELY trial demonstrated that a high TTR ([72.4 %) correlated to a reduced risk of both thrombo-embolic complications and major bleeding [19]. Recently a study demonstrated an increasing benefit from anticoagulation treatment with increasing TTR, and at TTR’s above 58 % it is reasonable to expect that a patient benefits significantly from anticoagulation treatment [10]. The purpose of this study was to investigate the frequencies and types of bleeding complications related to warfarin treatment at an anticoagulation clinic in southern Sweden. Since there are differences in healthcare contact patterns between different age groups there could be differences in reporting of bleeding events with regard to age. Also, physicians could be negligible of elderly patients and more prone to report bleedings in younger patients. Hence, we also wanted to investigate and to compare the proportion of patients with bleeding complications spontaneously reported to the anticoagulation clinic in different age groups with regards to both age and gender.
Materials and methods Between January 1 1995 and October 31 2010 all patients treated with warfarin in the catchment area of Helsingborg, were enrolled in a database named Journalia AVK. On November 1 2010, AuriculA, a national prospective quality register of anticoagulation, was started at Helsingborg Hospital. In connection to the change of registry system the aim of this study was to investigate the incidence of bleeding complications in patients on warfarin treatment being assessed and treated by the anticoagulation
123
clinic in Helsingborg. In the Helsingborg catchment area all patients treated with warfarin are registered in the database and prescriptions and follow-up are managed by the anticoagulation clinic. The study is a descriptive, retrospective registry study and includes all 4,400 patients registered in Journalia AVK during the time period November 1, 2007 to November 1, 2010. A list of clearly defined diagnoses of bleeding was made to select patients affected by a bleeding complication during the time period (Appendix). Patient data from Journalia AVK was then cross-matched with patient hospital records from a regional database and finally matched with the previously mentioned bleeding diagnoses. A systematic review of all medical records was done by the physician listing age, gender, indication of warfarin treatment, bleeding diagnosis and extent of bleeding according to ISTH definitions of bleeding [20]. Furthermore, bleeding events were divided into fatal, major, intracranial and minor bleeding. INR at the time of bleeding and duration of warfarin treatment was registered. Bleeding complications were investigated if they had been reported to the anticoagulation clinic and if the patient had died independent of the bleeding. During the systematic review of all patient records 215 patients were excluded from the study. These patients were excluded due to bleeding episodes that occurred, when the patient was no longer using warfarin or before warfarin treatment had commenced, during the observed time frame. Four cases of bleeding had been reported to the clinic but not originally found during the search among the medical records. These cases were the result of an incorrect diagnosis of bleeding recorded in the medical records. Those four patients were included in the study before statistical analysis. An additional five patients were included in our study. These patients had been reported to the anticoagulation clinic but missed in the search among the hospital records since the bleeding episodes were registered as other diagnoses than those used in our study. Two of these patients suffered from major gastrointestinal bleeding and three from intracranial bleeding associated with trauma. Time in therapeutic range was calculated according to F.R. Roosendaal’s algorithm with linear interpolation [21]. Only patients with target INR 2.0–3.0 were included in the analysis of TTR. As a result 440 patients with other target INR levels were excluded from the analysis of TTR. The age of any patient in the Journalia AVK population was defined as the age of that patient on May 1, 2009. This date is the mid-point of the study period. If a patient had more than one bleeding episode only the first was used for TTR calculation but all events were used to calculate rates of bleeding complications. Bleeding complications were converted to rates (per cent per patient-
Bleeding complications related to warfarin treatment
year) and tested for differences in patient characteristics. In all patients with bleeding complications only INR values at the time of the bleeding episode was used for statistical analysis. Mean INR was calculated based on 316 samples (138 were missing). Among the major and fatal bleedings the proportion of INR values within the therapeutic range 2.0–3.0, at the time of bleeding was calculated based on 161 samples (18 were missing). Differences in age and gender were tested using t test. Student‘s t test was used for testing age and bleeding complications. Differences in gender were tested using Chi Square test. The p value threshold for significance was \ 0.05. For statistical analysis SPSS v 20.0 was used.
Table 1 Incidence of bleeding events per patient-year
Results
ICH intracranial hemorrhage
A total number of 4,400 patients (1,803 female and 2,597 male) with a total of 8,394 patient-years were registered in Journalia AVK during the observation period. A total of 123,117 INR samples were registered. 79,793 of the samples were within the target range 2–3 (64.8 %). 37.1 % of the patients had an episode with INR \ 1.5, 31.5 % with INR [ 4.0 and 4.9 % with INR [ 6. TTR was 73.3 % for the whole population during the study period. The mean age in the whole population was 72 years. The mean age was higher in women (74) than in men (71). The most common indication of treatment with warfarin was atrial fibrillation (61.3 %) and VTE (23.8 %). Bleeding complications A total of 421 (9.6 %) patients treated with warfarin suffered a haemorrhagic event during the study (Table 1). The majority of bleeding events (n = 272) were minor bleedings. A total of 138 major bleedings were registered i.e. 1.6 % per patient-year (Table 1). The most frequent major bleeding was gastrointestinal (n = 64). Other major bleedings were haematuria, intraocular, hematoma, intramuscular with and without compartment, intra-articular, epistaxis, anaemia requiring transfusion, intra-abdominal, retroperitoneal and undefined bleeding. 18 patients suffered a fatal bleeding (4 %) i.e. 0.2 % per patient-year. ICH was the main cause of fatal bleeding (12/18). The other fatal bleedings were gastrointestinal, hematuria, deep bleeding in the hip region, hemoptysis and retroperitoneal bleeding. An additional 27 patients suffered from ICH without a fatal outcome with a total incidence of 0.5 % per patient-year. Thirty-four of a total of 421 patients suffered a second bleeding event of which 30 were minor bleedings and 4 were major bleedings. During the observation period 93 patients died of causes other than warfarin-related bleeding.
Bleeding diagnosis
Bleeding events (n)
Risk of bleeding ( % per patient-year)
Total bleeding episodes
455
5.4
18
0.2
Fatal ICH
12
Gastrointestinal bleeding
2
Other bleeding
4
Major bleeding
138
1.6
ICH (including fatal ICH) Minor
39 272
0.5 3.2
The total number of patient-years during the study was 8,394. Bleeding events could be recognised as both fatal, and intracranial (ICH)
Table 2 Distribution of bleeding diagnoses between genders Bleeding diagnosis
Female (n)
Male (n)
p-value
Total bleeding episodes
158
309
0.017*
6
12
ICH
5
7
Gastrointestinal bleeding
1
1
Other bleeding
0
4
Major bleeding
44
94
ICH (including fatal ICH)
22
17
Minor bleeding
86
186
Fatal
0.002**
Bleeding events could be recognised as both fatal and intracranial (ICH). The total number of females and males that suffered one or more bleeding events was 145 and 276 respectively. p values indicate differences between genders. There was a significant difference in mean age at bleeding between the sexes (female 76.9 years vs. male 74.2 years; p = 0.02), and female patients suffered more ICH than men (p = 0.002) ICH intracranial hemorrhage * **
Level of significance p \ 0.05 Level of significance p \ 0.01
Age and gender Among the 421 patients who suffered a bleeding, 145 were female and 276 were male (Table 2). In total the females suffered 158 bleeding events and the men 309. The mean age in the population with bleeding complications was 75 years. The women were significantly older than the men (p = 0.024) (Table 2). The distribution of bleeding complications between the genders is seen in Table 2. When comparing ICH, fatal and major bleedings with minor bleedings there was no significant difference between genders (p = 0.27). When comparing ICH against
123
M. Navgren et al.
Bleedingevents and age 0,8 fatal 0,7
Events per patient-year
Fig. 1 Distribution of bleeding events in different age groups. A significant correlation between age and severe bleeding events (fatal, major and ICH); p = 0.03, is seen. Bleeding events in the figure includes all events, also those not reported to the anticoagulation clinic
major ICH
0,6 0,5 0,4 0,3 0,2 0,1 0
