SEARCH

SEARCH BY CITATION

Keywords:

  • haemorrhagic stroke;
  • oral anticoagulation;
  • warfarin

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

Abstract.  Själander A, Engström G, Berntorp E, Svensson P (Malmö University Hospital, University of Lund, Sweden). Risk of haemorrhagic stroke in patients with oral anticoagulation compared with the general population. J Intern Med 2003; 254: 434–438.

Objectives.  To compare the incidence of haemorrhagic stroke (HS), and the risk of fatal outcome after HS in patients with oral anticoagulation (OA) treatment and in the general population.

Design.  Five-year cohort study.

Setting.  The Anticoagulation Clinic, Malmö University Hospital, Lund, Sweden.

Subjects.  A total of 4434 patients treated with OA (6693 treatment years) from 1 Oct 1993 to 30 Sept 1998. The population-based Stroke Register of Malmö, Lund, Sweden (STROMA).

Results.  Forty-eight patients had HS according to ICD  9 code 430 and 431. HS occurred at a higher age in women compared with men (mean age 79.5 years vs. 74.7 years, P = 0.009). The age-adjusted relative risk of HS during OA treatment was 10.9 (CI 6.7–17.6) for men and 9.3 (CI 5.7–15.0) for women, as compared with the untreated general population. Number needed to harm (NNH) (person-years) was 103 for men and 188 for women. Adjusted for age and sex, OA treatment was significantly associated with fatal outcome in patients with HS (OR = 2.6, CI 1.4–4.8).

Conclusions.  Patients with OA treatment had approximately 10 times higher risk of HS as compared with the general population, and the risk increased markedly with age. OA treatment is associated with an increased case fatality in patients with HS.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

Almost 1% of the population in Sweden is under treatment with oral anticoagulants (OA), mostly warfarin [1]. Bleeding during OA treatment is a known and hazardous side-effect, amongst which haemorrhagic stroke (HS) is most feared [2]. It is likely that the frequency of OA-related complications will increase because of the ageing population in many countries, and because of the evaluation of new indications for OA treatment [3–5]. However, to what extent OA is associated with an increased risk of HS or confers higher case fatality ratio in HS has not been extensively studied in unselected samples of patients and controls.

Material and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

All patients on OA treatment at the Anticoagulation Clinic, Malmö University Hospital, Lund, Sweden were followed from 1 Oct 1993 to 30 Sept 1998. The cohort consists of 4434 patients, 2218 women and 2216 men, contributing to 6693 treatment years and comprising over 90% of the patients treated with OA in Malmö. The main indications for OA were atrial fibrillation (45%), venous thromboembolism (25%), mechanical valve prosthesis (20%) and others (10%). The distribution of the patients with HS was approximately the same: atrial fibrillation (47%), venous thromboembolism (21%), mechanical valve prosthesis (15%) and others (17%). The mean age at the start of OA treatment was 68.3 and 70.4 years for men and women, respectively. The incidence of HS in patients with OA was compared with the incidence of first-ever HS in the remaining (i.e. non-treated) population of the city of Malmö (approx. 240 000 citizens) during the same period of time (1 Oct 1993 to 30 Sept 1998). The stroke register of Malmö (STROMA) [6], which continuously and actively searched for and validated patients with stroke, was used for case-retrieval. A total of 1958 men and 2194 women were registered with stroke during this period. A total of 217 men and 222 women of the non-treated population had a first HS. The same experienced research nurse has, in cooperation with a senior neurologist, validated all cases of HS amongst patients with OA and in the general population. Medical records were reviewed in all patients with intracranial bleeding on OA treatment, and the outcome was verified 1 year after the event.

The sex-specific incidence rates were standardized for age (5-year bands) with direct standardization and relative risks (95% confidence intervals) was computed. The incidence was weighted for the age distribution of patients with OA. Logistic regression was used to compare case-fatality rates with adjustment for age and sex.

A total of 217 men and 222 women suffered HS in the non-treated general population during the 5-year period. The crude incidence rate of HS in the Malmö population was 0.04/100 person-years for both women and men. The standardized rates were 0.1/100 and 0.06/100 for men and women, respectively.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

In patients with OA, 66 intracranial bleedings were observed in 64 individuals during the period, 26 women and 38 men. Of these, 48 patients (19 women, 29 men) had HS according to ICD  9 code 430 (subarachnoidal haemorrhage, n = 1) and 431 (intracerebral haemorrhage, n = 47), and therefore fulfilled the criteria of HS [6]. Apart from these, another 11 patients had subdural haematomas, three patients had epidural haematomas, one had a bleeding from an intracranial tumour metastasis and one had a traumatic HS in the OA cohort. However, these 16 patients were excluded from the risk calculations since only the 48 patients with HS according to ICD  9 code 430 and 431 could be compared with the incidence in the general population. The crude rate of HS in OA patients was 0.76 for men and 0.55 for women per 100 treatment years. The age-standardized rates of HS were 1.07 and 0.59 per 100 treatment years, respectively. The age-standardized relative risk of HS during OA treatment was 10.9 (CI 6.7–17.6) for men and 9.3 (CI 5.7–15.0) for women, as compared with the untreated general population, corresponding to numbers needed to harm (NNH) per year of 103 and 188, respectively.

HS occurred at a higher age in women compared with men (mean age 79.5 years vs. 74.7 years, P = 0.009). The distribution of HS in relation to age and time on OA treatment is shown in Fig. 1. In the age group 60–69 years, 13 men and no women bled. Only one patient under the age of 60 bled.

image

Figure 1. Haemorrhagic stroke per 100 treatment years on OA in different age- and sex-groups.

Download figure to PowerPoint

Most of the bleedings occurred under therapeutic (2.0–4.0) or subtherapeutic (< 2.0) international normalized ratio (INR), with stable INR values in even weeks prior to the bleeding. Of the 48 patients with HS in the OA cohort, seven patients (15%) had INR > 4 at the time of bleeding (Fig. 2). Twenty-two (46%) had a known diagnosis of hypertension prior to the HS. Twenty-one (44%) of the bleedings were fatal within 4 weeks, an additional seven patients died within 1 year. In the HS patients without OA, 21% died within 4 weeks. Adjusted for age and sex in a logistic regression, OA treatment was significantly associated with fatal outcome amongst patients with HS (OR = 2.6, CI 1.4–4.8).

image

Figure 2. Distribution of INR values at time of haemorrhagic stroke.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

Bleeding during OA treatment is a known and hazardous side-effect, but the most feared side-effect of all is HS. This study shows that patients with OA had an approximately 10-fold increased risk for HS compared with the general, untreated population. Furthermore, OA treatment was associated with a significantly higher case-fatality rate in patients with HS. The study cohort reflects the clinical setting in the Malmö region, the results are based on large numbers of individuals and events, and these results should therefore be a good approximation of the true risk associated with OA treatment.

Previous studies on OA and its complications have mostly been performed in highly selected populations. Some studies addressing the risk of HS conferred by OA are confined to treatment of a specific diagnosis, i.e. myocardial infarction or atrial fibrillation [3–5], whilst others have inclusion criteria that exclude patients over or under a certain age [5, 7, 8]. Most studies have reported approximately a seven- to 10-fold increased risk of HS by OA treatment. One clinical trial reported a NNH for OA of 200, but in this study, patients over the age of 70 years were not included, and the control group was treated with aspirin [5]. As all patients on OA treatment, regardless of age, were included in this study, the patients were considerably older resulting in a higher bleeding risk. This can explain why a lower NNH was found (103 in men, 188 in women). A HS frequency of 0.3–1.7% per treatment year has been previously reported [3, 6, 9–11]. As this figure is strongly age-dependent (Fig. 1), the age distribution could partly explain differences between studies. In a large study from Finland [12], the age-adjusted odds ratio of being on anticoagulant treatment at the time of primary HS was 6.7. However, the Finnish study compared the frequency of OA treatment amongst the patients with HS and in the general population and not the other way round as in this study, i.e. the frequency of HS amongst OA patients and in the general population. HS during OA treatment is often greater and carries a higher mortality rate (44–68% in 1–6 months) [7, 8, 12–15]. This has previously been shown in relatively small studies and was proven significant by aggregating 185 patients from six different studies or in a subset of patients in one study [15]. The increased case fatality rate in HS during OA treatment is hereby verified.

The risk data must however be viewed in the context of the potential benefit by OA in the same patient group, i.e. preventing ischaemic strokes. That data is unfortunately not known in general, but it has been estimated that the risk of stroke and the recurrence of venous thrombosis without OA in similar patient groups may be as large as 10% and even larger during certain circumstances [16–18]. Thus OA should still be considered to be very beneficial for patients with the correct treatment indications, although with special caution when treating the elderly.

The STROMA register has actively and conscientiously searched for and validated patients with stroke in the city. All cases have been validated by the same experienced research nurse, and the procedures for case-retrieval have been unchanged during the study period. A validation study has shown that the STROMA register identifies more cases than routine registers, like the National Cause of Death register and patient administrative registers [19]. Computed tomography (CT) was required for the diagnosis, and more than 90% of all patients with stroke in the city underwent CT during this period. There is no reason to assume that the incidence rates have been confounded by biased case-retrieval or validation of cases.

A limitation of this study is that no information on other cardiovascular risk factors was available. It is likely that OA treatment is instituted more often in patients with an increased cardiovascular risk. The risk of HS attributable to OA could therefore be overestimated.

Differences between the sexes have not been addressed by most previous studies on OA-associated HS. There is a pronounced difference between the sexes regarding the prevalence of cardiovascular diseases in different age groups, with women developing atherosclerosis, hypertension and myocardial infarctions at a higher age [20–22]. This might be one reason why women on OA treatment were bleeding at a higher age than men. HS during OA treatment was uncommon under the age of 60 years, and OA therefore seems to be less hazardous amongst younger patients. As OA treatment increases the risk of HS bleeding approximately 10 times compared with the age-adjusted local general population, 10% of the bleedings are not due to the OA treatment but can be attributed to the risk in the general population.

In the near future, new and potent drugs, i.e. oral direct thrombin inhibitors are expected to replace warfarin in a number of indications [23]. These drugs will not need the close monitoring system with frequent INR samplings that are being performed today, making it more convenient for the patient but with a risk of worse compliance to the treatment. Therefore, it is of great importance to evaluate the current and coming therapeutic settings in order to compare the possible benefits and risks with new treatments.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

The Swedish Council for Work Life and Social Research. Anna and Edwin Berger's Foundation. Miss Eva Lindén at the Department of Coagulation Disorders, Malmö University Hospital, is gratefully acknowledged for her assistance.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References
  • 1
    Johnsson H. Anticoagulant therapy can be safer: an active follow-up and written care programs reduce the number of complications. Läkartidningen 1999; 96: 338890 [in Swedish].
  • 2
    Palareti G, Leali N, Coccheri S et al. Bleeding complications of oral anticoagulant treatment: and inception-cohort, prospective collaborative study (ISCOAT). Lancet 1996; 348: 4238.
  • 3
    Stroke Prevention in Atrial Fibrillation Investigators. Warfarin versus aspirin for prevention of thromboembolism in atrial fibrillation: stroke prevention in Atrial Fibrillation II Study. Lancet 1994; 343: 68791.
  • 4
    Anticoagulants in the Secondary Prevention of Events in Coronary Thrombosis (ASPECT) Research Group. Effect of long-term anticoagulant treatment on mortality and cardiovascular morbidity after myocardial infarction. Lancet 1994; 343: 68791.
  • 5
    Hurlen M, Abdelnoor M, Smith P, Erikssen J, Arnesen H. Warfarin, aspirin or both after myocardial infarction. N Engl J Med 2002; 347: 96974.
  • 6
    Engström G, Jerntorp I, Pessah-Rasmussen H, Hedblad B, Berglund G, Janzon L. Geographical distribution of stroke incidence within an urban population. Relationships to socioeconomic circumstances and prevalence of cardiovascular risk factors. Stroke 2001; 32: 1098103.
  • 7
    Franke CL, de Jonge J, van Swieten JC, Op de Coul AAW, van Gijn J. Intracerebral hematomas during anticoagulant treatment. Stroke 1990; 21: 72630.
  • 8
    Wintzen AR, de Jonge H, Loeliger EA, Bots GTAM. The risk of intracerebral hemorrhage during oral anticoagulant treatment: a population study. Ann Neurol 1984; 16: 5538.
  • 9
    Ericsson SE, Link H. Evaluation of anticoagulants in patients with cerebral infarction with slight to moderate neurological deficit. Acta Neurol Scand 1983; 68: 96106.
  • 10
    Odén A, Fahlén M. Oral anticoagulation and risk of death: a medical record linkage study BMJ 2002; 325: 10735.
  • 11
    van Walraven C, Hart RG, Singer DE et al. Oral anticoagulation vs aspirin in nonvalvular atrial fibrillation: an individual patient meta-analysis. JAMA 2002; 288: 24418.
  • 12
    Fogelholm R, Eskola K, Kiminkinen T, Kunnamo I. Anticoagulant treatment as a risk factor for primary intracerebral hemorrhage. J Neur Neurosurg 1992; 55: 11214.
  • 13
    Hart RG, Boop BS, Anderson DC. Oral anticoagulants and intracranial hemorrhage: facts and hypotheses. Stroke 1995; 26: 14717.
  • 14
    Neau JP, Couderq C, Ingrand P, Blanchon P, Gil R. Intracranial hemorrhage and oral anticoagulant treatment. Cerebrovasc Dis 2001; 11: 195200.
  • 15
    Sjöblom L, Hårdemark HG, Lindgren A et al. Management and prognostic features of intracerebral hemorrhage during anticoagulant therapy. A Swedish multicenter study. Stroke 2001; 32: 256774.
  • 16
    Hutten BA, Prins MH. Duration of treatment with vitamin K antagonists in symptomatic venous thromboembolism (Cochrane Review). In: The Cochrane Library, Issue 2, 2003. Oxford: Update Software.
  • 17
    Segal JB, McNamara RL, Miller MR et al. Anticoagulants or antiplatelet therapy for non-rheumatic atrial fibrillation and flutter (Cochrane Review). In: The Cochrane Library, Issue 2, 2003. Oxford: Update Software.
  • 18
    Cannegieter SC, Rosendaal FR, Briët E. Thromboembolic and bleeding complications in patients with mechanical heart valve protheses. Circulation 1994; 89: 63541.
  • 19
    Merlo J, Lindblad U, Pessah-Rasmussen H et al. Comparison of different procedures to identify probable cases of myocardial infarction and stroke in two Swedish prospective cohort studies using local and national routine registers. Eur J Epidemiol 2000; 16: 23543.
  • 20
    August P, Oparil S. Hypertension in women. J Clin Endocrinol Metab 1999; 84: 18626.
  • 21
    Giardina EG. Heart disease in women. Int J Fertil 2000; 45: 3507.
  • 22
    Hendricks AS, Goodman B, Stein JH, Carnes M. Gender differences in acute myocardial infarction: the University of Wisconsin experience. WMJ 1999; 98: 303.
  • 23
    Eriksson H, Wåhlander K, Lundström T, Billing Clason S, Schulman S. Extended secondary prevention with the oral direct thrombin inhibitor Ximegalatran for 18 months after 6 months of anticoagulation in patients with venous thromboembolism: a randomised, placebo-controlled trial. Blood 2002; 100: 81a (abstract no. 297).