Mortality After Ischemic Stroke in Patients With Acute Myocardial Infarction
Predictors and Trends Over Time in Sweden
Background and Purpose—Acute myocardial infarction (AMI) increases the risk of ischemic stroke, and mortality among these patients is high. Here, we aimed to estimate the 1-year mortality reliably after AMI complicated by ischemic stroke. We also aimed to identify trends over time for mortality during 1998–2008, as well as factors that predicted increased or decreased mortality.
Methods—Data for 173 233 unselected patients with AMI were collected from the Swedish Register of Information and Knowledge about Swedish Heart Intensive Care Admissions registry for 1998–2008. Specifically, we analyzed 1-year follow-up and mortality data for patients with AMI with and without ischemic stroke. Kaplan–Meyer analysis was used to analyze mortality trends over time, and Cox regression analysis was used to identify uni- and multivariate predictors of mortality.
Results—The 1-year mortality was 36.5% for AMI complicated by ischemic stroke and 18.3% for AMI without stroke. Mortality decreased over time in patients with and without ischemic stroke. The absolute decreases in mortality were 9.4% and 7.5%, respectively. Reperfusion and secondary preventive therapies were associated with a decreased mortality rate.
Conclusions—Mortality after AMI complicated by an ischemic stroke is very high but decreased from 1998 to 2008. The increased use of evidence-based therapies explains the improved prognosis.
Patients surviving an acute myocardial infarction (AMI) have an increased risk of stroke.1–3 The risk of stroke after AMI is ≈0.9% to 1.4% during the hospital stay, 2.1% within 1 year,1,2,4,5 and 5.7% within 4 years.6 The most important risk factors for the occurrence of stroke are older age, diabetes mellitus, hypertension, prior stroke, heart failure, atrial fibrillation, and coronary artery by-pass grafting.1,4,7
Stroke after AMI is associated with an increased risk of death. One study found that patients with AMI that have a stroke have an increased risk of dying during hospitalization with an odds ratio of 4.3.8 The short-term mortality rate (death in-hospital or during the 30-day follow-up) has been estimated to be 17.0% to 45.9%, whereas the long-term mortality rate (1.5–3.4 years after discharge) is estimated to be 16.9% to 27.9%.2,4,7–10 There are several limitations to these estimates. Most previous studies did not analyze ischemic and hemorrhagic stroke separately, probably because of small sample sizes; however, hemorrhagic stroke is associated with higher mortality compared with ischemic stroke.11 One study found a 1-year mortality rate after hemorrhagic stroke of 71.8% in a selected population of patients with AMI treated with thrombolysis.12 The risk of hemorrhagic stroke after AMI is closely related to the use of thrombolytic therapy, whereas ischemic stroke has other, partly unknown, mechanisms. It is, therefore, of interest to analyze ischemic stroke separately.
Most studies describe short-term mortality, often using the duration of the hospital stay as the follow-up duration. However, the hospitalization period has decreased over time making comparisons between studies rather difficult. Moreover, the treatment of AMI has changed markedly during the past decade resulting in decreased mortality. It is unknown whether mortality after an AMI-associated stroke has changed as well. Only 1 previous study has presented a time trend of mortality after AMI-associated stroke. The results of that study suggested that the mortality increased from 1986 to 2005. However, the calculations were based on a few stroke cases making the conclusion unreliable.2 Factors associated with increased or decreased mortality after an AMI-associated ischemic stroke remain unclear.
In the present study, our primary aim was to determine the mortality rate associated with ischemic stroke during the first year after an AMI. Second, we wished to identify trends in mortality during the 1998–2008 period and to identify factors associated with increased or decreased risk of death after AMI-associated ischemic stroke.
Materials and Methods
Patient data were obtained from the Swedish Web-system for Enhancement and Development of Evidence-based Care in Heart Disease Evaluated According to Recommended Therapies (SWEDEHEART) that includes Register of Information and Knowledge about Swedish Heart Intensive Care Admissions (RIKS-HIA), a national quality register of all patients admitted to a coronary care unit in Swedish hospitals. Patient data are reported on case record forms that include >100 variables; information is recorded on hospital admission, while the patient is in-hospital and at discharge. These records include patient demographics, risk factors, medical history, biochemical markers, discharge diagnosis, and discharge medications. The validity of the entered data is examined annually and shows 93% to 97% conformity between the Register of Information and Knowledge about Swedish Heart Intensive Care Admissions data and patient records. In 2008, 74 of 74 hospitals with coronary care units participated, covering >95% of Swedish coronary care units admissions, when compared with the Swedish National Patient Register.13 We collected data from patients with the index AMI during 1998–2008 with follow-up until 2009; this included a total of 173 233 patients.
We combined data from Register of Information and Knowledge about Swedish Heart Intensive Care Admissions with data from the National Patient Register to identify patients with ischemic stroke and to identify all causes of death. The National Patient Register includes dates for admission and discharge as well as diagnoses at discharge for all hospital stays in Sweden. The International Classification of Diseases-Ninth Revision (433 and 434) and International Classification of Diseases-Tenth Revision (I63 and I64) codes for cerebral infarction were used. The National Patient Register has been validated and a diagnosis of stroke or transient ischemic attack has a positive predictive value of 98.6%.14 Both registers are based on the entire population in Sweden. There were no criteria for exclusion.
Patients were registered from 1998 to 2008 at the time of their index AMI. The follow-up time for the major outcomes ischemic stroke and death was 1 year from AMI admission.
All patients for whom data were entered into the Register of Information and Knowledge about Swedish Heart Intensive Care Admissions were informed of their participation in the registry (patients could request to be excluded from the registry) and the long-term follow-up. The registry was approved by the National Board of Health and Welfare and the Swedish Data Inspection Board. The merging of registries was approved by the local ethics committee.
The variable “heart failure during hospitalization” was defined as the occurrence of pulmonary rales or treatment with intravenous diuretics during hospitalization. The variable “smoking” was defined as smoking within the past month. “Atrial fibrillation” includes patients with atrial fibrillation (all types) before or during hospitalization.
The study period was divided into 5 time periods (1998–2000, 2001–2002, 2003–2004, 2005–2006, and 2007–2008) to study changes over time. The results are presented as mean values for continuous variables and as percentages for categorical variables. Comparisons between patient groups were performed using the Mann–Whitney U test for continuous variables and the χ2 test for categorical variables.
Kaplan–Meyer curves were calculated for the occurrence of death during the 5 different time periods and compared between groups using the log-rank test. Uni- and multivariable predictors of risks were assessed using Cox proportional hazards regression analysis. The end point was defined as death within 1 year from admission to hospital for AMI. For our multivariable model, we used previously established predictors of risk for death and added the factors from our univariable analysis. Patients who died during hospitalization were excluded from the analysis when variables on treatment at discharge were included. We excluded variables from the multivariable analyses that had <90% valid cases (body mass index, habitual smoking, heart rate, systolic blood pressure, diastolic blood pressure, plasma cholesterol, plasma low-density lipoprotein, plasma glucose, and the use of AT2-inhibitors at discharge). A P value <0.05 was considered significant. All statistical analyses were performed using SPSS version 20.0 software.
This study included 173 233 patients with AMI from 1998 to 2008. The patients had a mean age of 71.2 years, and 36.8% were women. The patient characteristics are shown in Table 1. Of the 173 233 patients with AMI, 7185 (4.1%) had an ischemic stroke within 1 year of AMI and 2624 (36.5%) of these patients died during that year. Of the patients that did not have ischemic stroke in the year after AMI, 30 461 (18.3%) died. Overall, patients with ischemic stroke were older and had more comorbidities; this was also true for the patients who died during the first year after AMI.
Table 2 and Table I in the online-only Data Supplement describe the variation in patient characteristics and comorbidities over time for patients with and without ischemic stroke stratified for survival. Prior AMI decreased in all patient groups, probably because of the selection criterion (only the first event of AMI in the registry was included in the analysis). Hypertension increased in all groups, probably because of improved diagnosis. The proportion of ST-segment–elevation myocardial infarction decreased in all groups. The use of thrombolysis decreased, whereas the use of percutaneous coronary intervention increased.
The changes in medication at discharge over time are shown in Figure 1 and Figure I in the online-only Data Supplement. The use of β-blockers, acetylsalicylic acid, P2Y12-inhibitors, and statins increased in all patient groups. Overall, the use of evidence-based medication was lower in patients with stroke, as well as in nonsurvivors.
The overall mortality decreased during 1998–2008. In patients with and without stroke, the absolute decreases were 9.4% and 7.5%, respectively, whereas the relative decreases were 23.8% and 34.6%, respectively. The Kaplan–Meyer survival curves for patients with and without ischemic stroke stratified by time period are shown in Figure 2. The decrease in mortality was similar for men and women. In men, the 1-year mortality was 36.2%, 37.4%, 34.8%, 30.3%, and 27.1% and in women 44.2%, 41.3%, 40.4%, 38.8%, and 33.8% during the time periods 1998–2000, 2001–2002, 2003–2004, 2005–2006, and 2007–2008, respectively.
Predictors of Mortality After Ischemic Stroke
Table 3 shows the results of the Cox regression analyses in patients who had an ischemic stroke after AMI. Heart failure, renal disease, peripheral arterial disease, and diabetes mellitus had the strongest association with death. Evidence-based treatments were associated with a decreased risk of death. Diuretic use was associated with an increased risk of death. Table II in the online-only Data Supplement shows the corresponding analyses for patients without ischemic stroke.
This is the first study with an adequate size and duration of follow-up to assess trends over time and predictors of mortality after AMI-associated ischemic stroke reliably. The main findings were as follows: first, the mortality after AMI-associated ischemic stroke decreased during the 1998–2008 period. Second, patients with an AMI-associated ischemic stroke have an absolute mortality rate during the first year after the AMI that is 16% to 19% higher than the mortality rate of those without ischemic stroke. Third, evidence-based treatments are associated with decreased mortality in an adjusted analysis.
The large number of patients in this study allowed us to make a reliable estimate of the mortality in this fairly unselected population. The 1-year mortality of 36.5% was only slightly higher than in most previous studies of short-term mortality (in-hospital and 30 days).2,4,7,8 Most studies have not separated ischemic and hemorrhagic stroke, which causes a higher mortality estimate. The mortality, as well as the risk of stroke, rapidly declines after an AMI, which can explain the rather low difference between 30-day and 1-year mortality.3,5 We did not subclassify stroke events because ischemic strokes occurring after AMI are generally classified as cardioembolic. This classification can be questioned because other mechanisms may be more important than embolism.15
We found that mortality decreased during the 1998–2008 period, both in patients with and without an ischemic stroke after AMI. The relative risk reduction was larger in the population without ischemic stroke, whereas the absolute decrease was slightly greater for patients with stroke. There was a similar decrease in mortality in men and in women with stroke, but the women had a higher mortality during the study period. There is only 1 previous study, by Saczynski et al,2 that reported results for a time trend of mortality after stroke following an AMI. On the basis of 132 stroke cases and 43 deaths during 1986–2005, the authors suggested that the mortality after stroke has increased. However, a reliable time trend cannot be based on such a small patient population.
The present results are in agreement with earlier studies of mortality after stroke without a previous AMI, which show a decreasing mortality even though time periods and populations differ.16–19 There are probably several explanations for the decrease in mortality. Some studies have shown that specific stroke units have improved the prognosis of patients with stroke,19,20 and the 30-day mortality is lower for patients admitted to a stroke unit.21 Other explanations include improved secondary preventive treatment and improvement of cardiovascular risk factors.22
During the study period, the use of recommended treatments according to guidelines increased, particularly the use of statins and P2Y12-inhibitors. Overall, the use of recommended therapies was lower among patients with ischemic stroke and even lower among patients who died. The patients in these groups are older, have more comorbidities, and may have a short expected survival time and contraindications for medications. However, they may also have received suboptimal treatment.
Predictors of Mortality
In the multivariate Cox regression analysis, older age, heart failure, prior ischemic stroke, prior AMI, prior diabetes mellitus, prior renal failure, prior peripheral arterial disease, and atrial fibrillation were predictors of death. All secondary preventive medications lowered the risk, except for diuretics, which increased the risk for mortality. Although multivariate adjustments were made, a possible selection bias make it difficult to estimate the true size of the risk reduction achieved by different treatments.
These results agree with those of previous studies of risk factors for mortality after AMI not complicated by ischemic stroke. In those studies, older age, history of heart failure or stroke, and heart failure during the hospital stay were shown to predict the 2-year mortality rate.23
Strengths and Limitations
The strength of this study is the large number of included patients, which enabled precise estimates even in subsets of patients. In combination with the long study period, the large study population made it possible to analyze trends reliably over time. As for limitations, there were too many missing data for several variables, which could thus not be included in the multivariate Cox regression analysis. Furthermore, we did not have information about kidney function, which has prognostic importance. As always, there is the risk of confounders that are not taken into account because they were not recorded as variables in the database.
The 1-year mortality after AMI complicated by ischemic stroke is very high. Notably, the 1-year mortality for AMI plus ischemic stroke is ≈50% higher than for patients with AMI without stroke. Mortality decreased during the 1998–2008 period in parallel with increased use of evidence-based secondary preventive medications.
Sources of Funding
This study was supported by grants from the Research and Development Unit at Jamtland County Council.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.113.001434/-/DC1.
- Received March 12, 2013.
- Revision received July 13, 2013.
- Accepted July 15, 2013.
- © 2013 American Heart Association, Inc.
- Budaj A,
- Flasinska K,
- Gore JM,
- Anderson FA Jr.,
- Dabbous OH,
- Spencer FA,
- et al
- Mooe T,
- Olofsson BO,
- Stegmayr B,
- Eriksson P
- Mahaffey KW,
- Harrington RA,
- Simoons ML,
- Granger CB,
- Graffagnino C,
- Alberts MJ,
- et al
- Wienbergen H,
- Schiele R,
- Gitt AK,
- Schneider S,
- Heer T,
- Gottwik M,
- et al
- Xian Y,
- Holloway RG,
- Pan W,
- Peterson ED
- Giugliano RP,
- Giraldez RR,
- Morrow DA,
- Antman EM,
- Gibson CM,
- Mohanavelu S,
- et al
- Jernberg T,
- Attebring MF,
- Hambraeus K,
- Ivert T,
- James S,
- Jeppsson A,
- et al
- Rautio A,
- Eliasson M,
- Stegmayr B