Incidence of Oral Anticoagulant–Associated Intracerebral Hemorrhage in the Netherlands
Background and Purpose—The aim of this study was to estimate the annual adult incidence and risk of intracerebral hemorrhage (ICH) and oral anticoagulant–associated ICH (OAC-ICH) in the Netherlands.
Methods—We retrospectively selected all consecutive adult patients with a nontraumatic ICH seen in 1 of 3 hospitals in the region South-Limburg, the Netherlands, from 2007 to 2009. Crude incidences were age-adjusted to Dutch and European population.
Results—We identified 652 ICH cases, of which 168 (25.8%) were OAC associated. The adult Dutch age-adjusted annual incidence of ICH and OAC-ICH was 34.8 (95% confidence interval, 32.0–37.8) and 8.7 (95% confidence interval, 7.3–10.3) per 100 000 person-years, respectively. The absolute risk of OAC-ICH was estimated at 0.46% per patient-year of OAC treatment.
Conclusions—The annual incidences of ICH and OAC-ICH are relatively high in the Netherlands when compared with international literature.
A serious complication of oral anticoagulants (OAC) is the occurrence of an OAC-associated intracerebral hemorrhage (OAC-ICH). In the Netherlands, acenocoumarol and phenprocoumon are mainly used, and strict dose-regulation of OAC is performed by a nation-wide network of specialized anticoagulation clinics. Although it is thought that OAC-ICH is low because of strict dose-regulation, recent Dutch data on the adult incidence of OAC-ICH are scarce. The aim of this study is to estimate the annual adult incidence of ICH and OAC-ICH, as well as the risk of OAC-ICH in the Netherlands.
We retrospectively analyzed all consecutive adult (≥18 years) patients with a brain scan-confirmed nontraumatic ICH, seen in the emergency department, inpatient or outpatient clinic, in 3 hospitals of South-Limburg, the Netherlands, from 2007 to 2009 (Figure). Patients were selected using diagnosis-treatment codes retrieved from hospital Medical Registration Archives complemented with hospital stroke registries. OAC-ICH was defined as an ICH while on treatment with oral vitamin-K antagonists at admission. Recurrent ICH cases were included. The study was approved by the medical ethical committees of all the 3 hospitals.
The crude annual adult incidence rates of ICH and OAC-ICH were calculated by dividing the number of ICH or OAC-ICH cases by the total adult person-years at risk for ICH. Person-years at risk were defined as the number of adult inhabitants in South-Limburg during 2007 to 2009, resulting in 1 595 800 person-years.1 We age-adjusted the incidence rates to the Dutch and European population of 2008.2 For all rates, we calculated the 95% or 1-sided 97.5% Poisson confidence interval (95% CI or 97.5% CI), if applicable.
The absolute risk of OAC-ICH per year of OAC treatment was calculated by dividing the number of OAC-ICH cases by the total patient-years of OAC treatment, which was estimated with data provided in the annual reports of the anticoagulation clinics. See Methods in the online-only Data Supplement for details.
We included 652 ICH cases in 617 patients; 588 (90.2%) patients had a first-ever ICH. Of all ICH cases, 168 (25.8%) were OAC associated and 153 of these (91.1%) were first-ever OAC-ICH. Acenocoumarol was used by 134 (79.8%) patients, phenprocoumon was used by 20 (11.9%), and a combination of OAC with another antithrombotic drug was used by 14 (8.3%). Indications for OAC were atrial fibrillation (77.4%), cardiac valve prosthesis (7.1%), venous thrombosis (6.0%), peripheral artery disease (4.2%), and other indications (5.4%). Table 1 shows clinical characteristics of (OAC-)ICH cases.
Incidence of (OAC-)ICH
The adult Dutch age-adjusted annual incidence of ICH and OAC-ICH was 34.8 (95% CI, 32.0–37.8) and 8.7 (95% CI, 7.3–10.3) per 100 000 person-years, respectively. The crude incidence rates and the age-adjusted rates of the European population of 2008 are presented in Table 2. For results on first-ever (OAC-)ICH incidence rates, see Table I in the online-only Data Supplement.
Absolute Risk of OAC-ICH
Total patient-years of OAC treatment in the study period was estimated at 36 494. Hence, the absolute risk of OAC-ICH was estimated at 0.46% per patient-year of OAC treatment.
We found an adult Dutch age-adjusted annual ICH incidence of 34.8 (95% CI, 32.0–37.8) per 100 000 person-years. Recent data on ICH incidence in the Netherlands are scarce. The Dutch Rotterdam population cohort study showed a crude annual incidence of 80 ICH cases per 100 000 person-years in subjects aged >55 years.3 This is comparable with our study when selecting subjects aged >55 years (data not shown). The ICH incidence is relatively high in our study when compared with international studies.4–6 A meta-analysis summarizing all population-based data up to November 2008 showed a crude annual incidence of 24.6 ranging from 1.8 to 129.6 per 100 000 person-years.6 The heterogeneity between studies is high (eg, regarding hospital versus population-based studies, inclusion criteria, and organization of healthcare). We determined the adult incidence rate, whereas many studies reported life-time incidence. Besides methodological issues, lifestyle and genetic factors could also explain differences in the incidence rates.
In our study, a quarter of ICH were OAC related, which is relatively high when compared with previous studies in which the proportion of OAC-ICH ranges from 5% to 27%.4,5,7–10 Highest proportions are seen in more recent studies, which may reflect the rise in OAC use.5 Other factors could also be involved in explaining a high proportion of OAC-ICH in our study, such as poor INR monitoring or vulnerability to bleeding.
We estimated the absolute risk of OAC-ICH at 0.46% per patient-year OAC treatment. Although this is an estimation and a prospective study would be needed to determine an exact risk, it is within the range found in clinical trials using warfarin, in which the risk of OAC-ICH ranged from 0.3% to 0.6% per patient-year.10 However, it must be noted that ICH rates in trials may be lower than in common practice because of patient selection and controlled care.
We have conducted a hospital-based study in a well- demarcated study region with easily accessible hospital care and where patients with stroke are always referred on short notice. Hence, only few patients hospitalized outside the region and patients not seen in a hospital may have been missed. Study limitations include patient selection using diagnosis-treatment codes, which may result in missing incorrectly coded patients. We used stroke registries to limit this possibility. Second, patient-years of OAC treatment were estimated, which may have led to a slight over- or underestimation. Third, acenocoumarol or phenprocoumon were used, whereas warfarin is more common in international literature. However, bleeding risks between OAC types are similar.11 Finally, South-Limburg has slightly poorer socioeconomic and lifestyle factors, which may negatively affect the representation of our data for the overall Dutch population.1
Notwithstanding the earlier raised methodological issues, we conclude that the adult annual incidence of ICH and the proportion of OAC-ICH in the Netherlands are relatively high when compared with previous studies, despite a nation-wide network of specialized anticoagulation clinics.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.113.003003/-/DC1.
- Received August 15, 2013.
- Accepted October 1, 2013.
- © 2013 American Heart Association, Inc.
- 1.↵Statistics Netherlands, Den Haag/Heerlen. http://www.cbs.nl. Accessed February 2013.
- 2.↵Eurostat. Statistics database, population and social conditions. http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=demo_pjangroup&lang=en. Accessed March 2013.
- Hollander M,
- Koudstaal PJ,
- Bots ML,
- Grobbee DE,
- Hofman A,
- Breteler MM
- Huhtakangas J,
- Tetri S,
- Juvela S,
- Saloheimo P,
- Bode MK,
- Hillbom M
- Kelly PJ,
- Crispino G,
- Sheehan O,
- Kelly L,
- Marnane M,
- Merwick A,
- et al
- Dequatre-Ponchelle N,
- Hénon H,
- Pasquini M,
- Rutgers MP,
- Bordet R,
- Leys D,
- et al
- Hart RG,
- Tonarelli SB,
- Pearce LA