Author + information
- Received October 20, 2009
- Revision received January 20, 2010
- Accepted January 25, 2010
- Published online March 30, 2010.
- Klas Gränsbo, MD⁎,
- Olle Melander, MD⁎,
- Lars Wallentin, MD†,
- Johan Lindbäck, PhD†,
- Ulf Stenestrand, MD‡,
- Jörg Carlsson, MD§ and
- Jan Nilsson, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Jan Nilsson, CRC Lund University, Building 91:12, Malmö University Hospital, Entrance 72, SE-205 02 Malmö, Sweden
Objectives The purpose of this study was to determine whether statin treatment is effective and safe in very elderly (80 years and older) acute myocardial infarction (AMI) patients.
Background Elderly individuals constitute an increasing percentage of patients admitted to hospitals for AMI. Despite that these patients have a higher mortality risk, the application of evidence-based medicine remains much lower than for younger patients.
Methods We included all patients 80 years and older who were admitted with the diagnosis of AMI in the Register of Information and Knowledge About Swedish Heart Intensive Care Admissions between 1999 and 2003 (n = 21,410). Of these, complete covariate and follow-up data were available for 14,907 patients (study population A). To limit the bias related comorbidity on statin therapy, we also performed analyses excluding patients who died within 14 days of the acute event (study population B) and all patients who died within 365 days (study population C). A propensity score was used to adjust for initial differences between treatment groups.
Results All-cause mortality was significantly lower in patients receiving statin treatment at discharge in study population A (relative risk: 0.55, 95% confidence interval: 0.51 to 0.59), in study population B (relative risk: 0.65; 95% confidence interval: 0.60 to 0.71), and in study population C (relative risk: 0.66; 95% confidence interval: 0.59 to 0.76). Similar observations were made for cardiovascular mortality as well as for AMI mortality. There was no increase in cancer mortality in statin-treated patients.
Conclusions Statin treatment is associated with lower cardiovascular mortality in very elderly post-infarction patients without increasing the risk of the development of cancer.
With increased life expectancy, the population of older patients is growing and cardiovascular disease remains the major cause of mortality in this age group. More than 80% of all coronary deaths occur in patients older than the age of 65 years (1). Despite that elderly patients with acute coronary syndromes have a higher short- and long-term mortality risk, the application of evidence-based medicine remains much lower than for younger patients (2–6). A large number of clinical trials have established that treatment with lipid-lowering statins significantly reduces cardiovascular mortality in post-myocardial infarction patients (7). However, data from observational studies such as GRACE (Global Registry of Acute Coronary Events) (8) and the Euro Heart Survey on ACS (9) suggest that <40% of myocardial infarction patients older than 75 years are prescribed statins at discharge. Several circumstances may to contribute to a lower use of statins in elderly post-myocardial infarction patients. The association between plasma cholesterol and cardiovascular risk diminishes with increasing age (10,11), and most lipid trials have excluded older patients. There may also be the fear of more side effects when treating older patients. The PROSPER (PROspective Study of Pravastatin in the Elderly at Risk) is the only randomized, controlled trial that specifically studied the effect of statin treatment in older (70 to 82 years) patients (12). In this trial, treatment with 40 mg pravastatin daily was found to reduce fatal/nonfatal cardiovascular events by 15% and fatal/nonfatal acute myocardial infarction (AMI) by 19%, but pravastatin treatment was also associated with a 25% increase in cancer incidence. Although meta-analyses of all major statin trials have shown no increase in cancer incidence (12), it cannot be excluded that older patients are at higher risk in this respect. In the present study, we used the Register of Information and Knowledge About Swedish Heart Intensive Care Admissions (RIKS-HIA) to analyze the association of statin treatment with all-cause mortality, cardiovascular mortality, and cancer mortality in a cohort of 14,907 very elderly (80 years and older) myocardial infarction patients.
RIKS-HIA includes all consecutive patients admitted to the coronary care units of all participating Swedish hospitals. Data on approximately 100 different variables regarding baseline characteristics, examinations, interventions, and complications in hospital and discharge medication and diagnosis were reported in case records, as described elsewhere (13). The variables in RIKS-HIA comply with the international Cardiology Audit and Registration Data Standards). To ensure the validity of the information entered into the database, a single specially trained monitor visited participating hospitals and compared information in the patient records, including electrocardiography, with the information entered into the RIKS-HIA database in 30 to 40 randomly chosen patients for each hospital. Data quality was monitored in 5,446 random records from all participating hospitals, comprising 299,530 measurements. There was a 94% overall agreement between the registered information and patient records. Between 1999 and 2001, the number of participating hospitals increased from 65 to 72, out of all 74 Swedish hospitals, where it remained through 2003.
All patients for whom data were entered into RIKS-HIA were informed of their participation in the registry (patients could request to be excluded) and the long-term follow-up. The registry and the merging with other registries were approved by the National Board of Health and Welfare and the Swedish Data Inspection Board. The Ethics Committee of Uppsala University Hospital approved the study.
We included all patients 80 years of age and older who were admitted with the diagnosis of AMI in the RIKS-HIA between January 1, 1999, and December 31, 2003 (n = 21,410). To be included in the end point analyses, we required complete data on all covariates that were adjusted for and specific cause of death in those who died during follow-up, leaving 14,907 patients for survival analyses (study population A) (Table 1).
Furthermore, to limit the bias related to effects of short life expectancy and comorbidity on physicians' choice of treatment, we excluded patients who died within 14 days from baseline (study population B) (Online Table 1) and all patients who died within 365 days (study population C) (Table 1). The study design is summarized in Figure 1. Cardiovascular drug therapies were entered in a structured formula on admission and at discharge. We used data from the Swedish National Patient Register to record a diagnosis of stroke, kidney failure, chronic obstructive pulmonary disease, dementia, congestive heart failure, myocardial infarction, peripheral artery disease, and cancer before the registration in RIKS-HIA.
Follow-up and end points
Patients were followed for end points with a median follow-up time of 296 days (interquartile range: 44 to 738 days, maximum of 5 years) by linking the Swedish 10-digit personal number with the Swedish National Cause of Death Register and the National Patient Register from baseline until the time of first event, death, or until December 31, 2003. End points were defined according to the International Classification of Diseases 10th Revision. Mortality end points were retrieved from the Swedish National Cause of Death Register with codes I21-I22 defining AMI mortality, codes I00-I99 defining cardiovascular mortality, and codes C00-D48 defining cancer mortality. In analyses of fatal and nonfatal AMI, end points were defined as codes I21-I22 in the National Patient Register or Swedish National Cause of Death Register. The date of hospital discharge was defined as the baseline.
Apart from exclusion of patients with short survival time (i.e., restricting the study population to study populations B and C), we attempted to further decrease bias related to comorbidities and the physicians' probability to prescribe statins at discharge by creating and adjusting for a propensity score.
The propensity score is defined as the conditional probability to receive treatment given the known baseline characteristics. At best, the propensity score captures all the initial differences among the treatment groups in 1 score that can be used for adjustments in subsequent analyses. The propensity score was estimated using a logistic regression model including the baseline variables, including cardiovascular medications at admission, as presented in Table 1.
We used Cox regression models to establish the relationship between statin treatment at the time of discharge and time to event. The models included other cardiovascular medications at discharge (beta-blockers, acetylsalicylic acid, other platelet inhibitors, and angiotensin-converting enzyme inhibitors), statin treatment on admission, the propensity score, and year of admission. The results are presented as relative risk (RR) and 95% confidence interval (CI). All statistical analyses were done using R version 2.8.1 (R Foundation for Statistical Computing, Vienna, Austria).
Of the total number of patients in study population A (n = 14,907), 8,817 (59.1%) patients died during follow-up. Of those who died, 6,929 (78.6%) patients died of cardiovascular causes (myocardial infarction, other ischemic heart diseases, congestive heart failure, stroke, cardiac arrhythmias, and other cardiac causes), 4,423 (50.2%) patients died of myocardial infarction, and 477 (5.4%) patients died of cancer. Because a significant proportion of the original study population was excluded due to missing data, we compared demographic and risk factor variables between those with complete data (n = 14,907) and those excluded due to missing data (n = 6,503). In general, the patients with missing data (Online Table 2) had a higher burden of prevalent disease compared with patients with complete data (Table 1).
All-cause mortality was markedly lower in patients receiving statin treatment at discharge in study populations A, B, and C (Table 2,Fig. 2). However, the RR for mortality associated with statin treatment was clearly dependent on whether patients who died early after discharge were excluded (Fig. 3). The RR reduction for mortality in statin-treated compared with non–statin-treated patients seemed to be less pronounced in study population B than in study population A, and it decreased even further as we excluded patients who died during the first 180 days from baseline in a stepwise manner. This suggested that part of the statistical relationship between statin treatment and mortality was attributed to bias related to comorbidities and the physicians' inclination to prescribe statins at discharge. Exclusion of all patients who died during the first year from baseline did not seem to further influence RR for mortality in statin-treated versus nonstatin-treated patients, suggesting that such bias was of less importance in study population C compared with study populations A and B (Fig. 3).
We subsequently performed stratified analyses in patients belonging to different quartiles of the propensity score, in those with and without myocardial infarction or congestive heart failure before admission, and by sex in study population C. The lower risk of all-cause mortality in patients treated with statins compared with those not treated with statins was significant in all these subgroups except in the lowest quartile of the propensity score (Table 3).
In study population C, the RR of cardiovascular mortality as well as AMI mortality was markedly lower in patients treated with statins compared with patients not treated with statins at discharge (Figs. 4 and 5).⇓⇓ Results were similar in study populations A and B (Table 2). The RR for the combination of fatal and nonfatal AMI during follow-up was reduced to a somewhat lesser degree compared with the RR for AMI mortality in study population C (RR: 0.69; 95% CI: 0.56 to 0.84), study population B (RR: 0.84; 95% CI: 0.76 to 0.92), and study population A (RR: 0.70; 95% CI: 0.65 to 0.76).
There was no increase in cancer mortality in statin-treated compared with non–statin-treated patients regardless of whether patients who died at different times during the first year from baseline were excluded, and it was even lower in statin-treated patients in study populations B and A (Table 2). The RR for cancer mortality was similar in statin-treated and nonstatin-treated patients in study population C (Table 2, Fig. 6).
This large observational study with complete long-term follow-up for up to 5 years (median 296 days) provides strong evidence of an association between statin treatment in very elderly (80 years of age and older) post-myocardial infarction patients and reduced cardiovascular mortality. Statin treatment at hospital discharge after AMI was associated with a reduction of all-cause mortality by 42% in the entire study cohort and by 34% if the analysis was restricted to patients who survived at least 1 year after the event. A total of 9,576 patients died during follow-up and 76.3% of these died of cardiovascular disease. Statin treatment was associated with a reduction of cardiovascular mortality by 41% and AMI mortality by 44% in the entire cohort and by 37% and 37%, respectively, in the cohort of patients surviving at least 1 year. Collectively, these observations suggest that the protective effect of statin treatment in very elderly patients post-myocardial infarction is of a similar relative magnitude as that demonstrated in randomized clinical trial for middle-age subjects (7) and that it may in absolute terms be even greater. Although the present study did not include patients younger than 80 years of age, it is of interest to note that a previous study of RIKS-HIA patients younger than 80 years of age revealed a 25% reduction of mortality in subjects prescribed a statin at discharge (13).
Patients discharged on statin treatment were more likely to have been taking statins at hospital admission (Table 1). It has been reported that pre-treatment with statins is associated with smaller myocardial infarction size (14,15), and there is also evidence from the follow-up of randomized trials of a long-term protective carryover effect of statins (16). Accordingly, there is a possibility that a more frequent statin pretreatment may have contributed to the increased survival observed in patients discharged on statin treatment. However, we did not observe an association between statin treatment at hospital admission and increased follow-up survival in the present study (data not shown).
Despite a higher risk (3,8,9), older AMI patients are less likely to receive evidence-based medication (4). In accordance, only 1 of 4 patients in the present study population received statins at discharge. One factor that may contribute to the lower use of evidence-based medicine in this age group is that controlled intervention trials mostly have excluded older patients. Only 1 randomized clinical statin trial, PROSPER (12), was restricted to older patients (70 to 82 years). This study, which included both primary and secondary prevention cohorts, showed a 15% reduction in cardiovascular events in the statin group but no effect on all-cause mortality. The Heart Protection Study was not designed to specifically address the effect of statin treatment in elderly patients, but a subgroup analysis of subjects between 70 and 80 years of age at baseline demonstrated an RR reduction equal to that for younger patients in the trial (17). Another factor that may have contributed to the lower prescription of statins to elderly post-AMI patients is the concern for an increased risk of cancer. The potential of an increased risk of cancer by cholesterol-lowering treatment was widely debated in the pre-statin era. Although meta-analysis of long-term statin trials have revealed no support for an increased cancer risk (18–20), the observation of a higher incidence of cancer in the pravastatin group of the PROSPER trial raised concerns that elderly patients could be at particular risk. Reports of inverse associations between plasma cholesterol and cancer rates in older persons (10) have also argued for precautions in treating elderly patients with statins. In the present studies that included statin-treated post-AMI patients older than 80 years of age, we observed no increase in cancer mortality. In contrast, in an analysis including the entire population of post-AMI patients for whom complete data were available (study population A; n =14,907), we observed a decreased incidence of cancer mortality among subjects receiving statin treatment (RR: 0.65; 95% CI: 0.49 to 0.86). However, no reduction of cancer mortality rates was observed in statin-treated patients if the study population was restricted to subjects surviving at least 1 year after the acute event.
There are several limitations of the present study that need to be considered. First, the inherent limitations of a nonrandomized registry study should be acknowledged. Despite appropriate statistical adjustments, unknown confounders may have affected the results. Although our analyses included controlling for the prevalence of cancer at the original admission, it is likely that the lower incidence of cancer mortality among patients given statins in study population A in some way reflects a bias in not prescribing preventive treatment to patients with decreased life expectancy. It is also necessary to critically consider whether the reduced cardiovascular mortality observed among elderly statin-treated post-AMI patients in the present study can be explained by a similar bias. However, the fact that statin treatment remained significantly associated with lower cardiovascular mortality risk also when all subjects who died during the first year were excluded argues against this possibility. Moreover, a propensity score was used to statistically control for the possible influence of baseline factors associated with the increased probability of being prescribed statin treatment. Another limiting factor of this study is that data regarding drug treatment are based solely on hospital discharge records. Accordingly, it can be assumed that some of the patients who were prescribed statins stopped taking the medication during the follow-up period and also that some patients discharged without statins began taking the medication at a later stage. However, assuming that this is correct, such a bias is likely to reduce the difference in cardiovascular mortality between the groups. It has also been reported that elderly patients have a higher compliance with statin medication (21,22). Finally, it should be kept in mind that the short median follow-up time of the present study may not be sufficient to accurately assess a possible association between statins and cancer mortality in elderly patients.
The present observational study strongly supports the concept that statin treatment provides cardiovascular protection in very elderly post-infarction patients without increasing the risk of the development of cancer.
For supplemental tables, please see the online version of this article.
Funding from the Swedish Medical Research Council, the Swedish Heart-Lung Foundation, Söderberg Foundation, Knut and Alice Wallenberg Foundation, Lars Hierta Memorial Foundation, Malmö University Hospital Foundation, Albert Påhlsson Foundation, and the Lundström Foundation.
- Abbreviations and Acronyms
- acute myocardial infarction
- Register of Information and Knowledge About Swedish Heart Intensive Care Admissions
- relative risk
- Received October 20, 2009.
- Revision received January 20, 2010.
- Accepted January 25, 2010.
- American College of Cardiology Foundation
- American Heart Association
- Halon D.A.,
- Adawi S.,
- Dobrecky-Mery I.,
- Lewis B.S.
- Wong C.K.,
- Newby L.K.,
- Bhapker M.V.,
- et al.
- Rosengren A.,
- Wallentin L.,
- Simoons M.,
- et al.
- Alsheikh-Ali A.A.,
- Trikalinos T.A.,
- Kent D.M.,
- Karas R.H.