The New England Journal of Medicine
e-mail icon  FREE NEJM E-TOC    HOME   |   SUBSCRIBE   |   CURRENT ISSUE   |   PAST ISSUES   |   COLLECTIONS   |    Advanced Search
Sign in | Get NEJM's E-Mail Table of Contents — Free | Subscribe
 
Original Article
PreviousPrevious
Volume 329:1442-1448 November 11, 1993 Number 20
NextNext

Age-Related Increase in Mortality among Patients with First Myocardial Infarctions Treated with Thrombolysis
Aldo A. Maggioni, Attilio Maseri, Claudio Fresco, Maria G. Franzosi, Francesco Mauri, Eugenio Santoro, Gianni Tognoni, for The Investigators of the Gruppo Italiano por lo Studio della Sopravvivenza nell'Infarto Miocardico (GISSI-2)

 

This Article
-Abstract

Tools and Services
-Add to Personal Archive
-Add to Citation Manager
-Notify a Friend
-E-mail When Cited

More Information
-PubMed Citation
ABSTRACT

Background The overall rate of mortality due to ischemic heart disease is known to increase progressively with age. We evaluated the relation between the mortality rate and age in patients with first myocardial infarctions treated with thrombolytic therapy.

Methods We studied 9720 patients with first infarctions who had been enrolled in the GISSI-2 trial. (This trial compared the efficacy of tissue plasminogen activator with that of streptokinase in patients with myocardial infarction.) Of these, only 35 percent had a history of angina. The relation between age and mortality during hospitalization and during the six months after discharge was determined by unadjusted and adjusted analyses.

Results The in-hospital mortality rate was 1.9 percent among patients 40 years old or younger, but it increased to 31.9 percent among those more than 80 years old; however, values for indicators of infarct size did not increase with age. Autopsies were performed in 20 percent of the 772 patients who died in the hospital; the findings showed that the frequency of cardiac rupture increased from 19 percent among patients 60 years old or younger to 86 percent among those more than 70 years old. The mortality rate for the first six months after hospital discharge also increased significantly with age. After adjustment for confounding variables, older age continued to be significantly associated with a higher risk of in-hospital and post-discharge death. When age was introduced into a multivariate model as a continuous variable, the risk of death was estimated to increase by about 6 percent per year for both in-hospital and six-month mortality rates.

Conclusions In patients with first myocardial infarctions who received thrombolytic therapy, age was a powerful independent predictor of both in-hospital and post-discharge mortality rates. The exponential, age-related increase in the mortality rate did not appear to be explained by larger infarcts.


The rate of mortality due to ischemic heart disease increases progressively with age. In the United States, about 80 percent of deaths from ischemic heart disease occur among patients over the age of 651.

The findings of recent large multicenter trials of thrombolytic therapy for acute myocardial infarction have shown that among patients over the age of 70 years, mortality during hospitalization is considerably higher than among younger patients2,3,4,5. However, it is unclear whether the increased incidence of fatal outcomes in the elderly is the result of more severe coronary atherosclerosis (which could lead to larger infarcts), greater susceptibility of the aging heart to myocardial necrosis, or other variables, such as coexisting illnesses.

The aim of this study was to assess the effect of clinical and epidemiologic variables on in-hospital and six-month mortality as a function of age in patients with first confirmed myocardial infarctions who were enrolled in the trial conducted by the Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto Miocardico (GISSI-2).

Methods

Trial Design, Selection of Patients, and Treatment

The details of the design and main results of the GISSI-2 trial have been presented elsewhere4. In brief, patients of any age were eligible for randomization if they presented to the participating centers within six hours after the onset of symptoms, had no contraindications to thrombolytic therapy, and had electrocardiographic evidence of ST-segment elevation of at least 1 mm in two limb leads or of at least 2 mm in one precordial lead. A total of 12,381 patients with acute myocardial infarction who met these criteria were randomly assigned to treatment with streptokinase (1.5 million units given over a period of 30 to 60 minutes) or alteplase (tissue plasminogen activator, 100 mg given over a period of 3 hours), with or without heparin (12,500 units given subcutaneously twice daily). Of the total studied, 9720 patients (78.5 percent) had first confirmed myocardial infarctions.

Clinical Data

Information about smoking, hypercholesterolemia, hypertension, and diabetes was abstracted from a standardized questionnaire by cardiologists specifically trained for this purpose. When the patients visited the clinic, the cardiologists who completed the study forms collected information about any earlier event suggesting angina pectoris, as indicated by recurring chest pain caused by effort and relieved by rest and also pain occurring during emotional stress or spontaneously and lasting less than 15 minutes, with a typical location or radiation or with the same features as the pain experienced during the infarction. Whenever possible, the investigators also reviewed the patients' previous clinical records documenting the clinical history. The Killip scale6 was used to stratify the severity of impairment of left ventricular function on admission. Whenever possible, left ventricular function was also assessed in survivors by two-dimensional echocardiography before discharge. Standard 12-lead electrocardiography was also performed on discharge to calculate the QRS score as a measure of the size of the infarct7.

An ad hoc committee blinded to the patients' thrombolytic treatment reviewed the clinical records of those who had died, to assess the causes of death. Cardiac rupture and cardiogenic shock resulting from rupture of the interventricular septum or papillary muscle were diagnosed only if they were found at autopsy.

Postmortem findings were available for 158 (20 percent) of the 772 patients who died in the hospital. The degree of coronary arteriosclerosis and the percentage of the left ventricle that was infarcted were estimated qualitatively during the postmortem examinations.

For the 1073 patients (11 percent of the total population) who could not be reached by the local monitor of the study, vital status six months after discharge was determined through the census offices of their cities of residence. The deaths of 30 patients were confirmed in this manner.

Estimation of Infarct Size and Left Ventricular Dysfunction

Three indicators of infarct size were used: the number of leads with ST-segment elevation on the admission electrocardiogram, the ratio of the peak creatine kinase level to the upper limit of normal at each study-center laboratory, and the QRS score calculated at discharge6.

Left ventricular dysfunction was defined as the presence of late (after the fourth hospital day) clinical congestive heart failure (indicated by the presence of an S3 gallop, rales, dyspnea, or radiologic evidence of pulmonary congestion), or as extensive left ventricular injury in the absence of clinical heart failure (indicated by a left ventricular ejection fraction of 35 percent or less on echocardiography, or by injury to 45 percent or more of myocardial segments [akinetic-dyskinetic scores] when a measurement of the left ventricular ejection fraction was not available8). The ejection fraction was calculated from left ventricular two-dimensional tomograms by either the area-length method or the modified Simpson rule9. Ejection fractions and akinetic-dyskinetic scores at discharge were available for 3228 and 8044 patients, respectively (33.2 percent and 82.8 percent of the study population).

Quality Control

Local monitors received a two-day training course on methods of data collection. The importance of validating the patients' answers was given special emphasis, and investigators were asked to review the patients' clinical records whenever possible. A random sample of about 10 percent of all echocardiographic assessments was also reviewed centrally. Data on the overall frequency of agreement were adjusted for the agreement expected by chance alone (kappa values). The kappa statistic showed highly satisfactory agreement for echocardiographic markers of left ventricular dysfunction (0.75 for the ejection fraction and 0.77 for the akinetic-dyskinetic score).

Statistical Analysis

The patients were divided into three age groups: patients 60 years old or younger, those 61 to 70 years old, and those more than 70 years old. All relations between variables were first determined by an unadjusted analysis10. Statistical significance was analyzed by chi-square tests. An adjusted analysis was performed with models constructed by multiple logistic regression11. Factors considered in the multivariable model for in-hospital mortality were age (<= 60, 61 to 70, and >70 years); sex; number of hours between the onset of symptoms and admission; Killip class on admission; location of the infarct (anterior or elsewhere); history of smoking, hypertension, diabetes, or angina; body-mass index; peak creatine kinase ratio; and the number of leads with ST-segment elevation. A second adjusted analysis (Cox model) of six-month mortality included the same variables and the indicators of the degree of left ventricular dysfunction at discharge. A third model incorporated the same variables and age as a continuous variable in an analysis of both in-hospital and six-month mortality. The results are presented in terms of Mantel-Haenszel odds ratios for in-hospital mortality and in terms of relative risks for six-month mortality, both with 95 percent confidence intervals.

Results

The base-line clinical characteristics of the patients are presented according to age group in Table 1.

View this table:
[in this window]
[in a new window]
 
Table 1. Distribution of Clinical Variables of Patients with First Myocardial Infarctions, According to Age Group.

 
A history of angina was obtained from about 35 percent of the study population, and the incidence was similar in the age groups. In this trial, in which only patients with ST-segment elevation could be enrolled, the number of leads showing ST-segment elevation at admission was similar among the age groups. The number of patients in Killip classes 3 and 4 (classes indicating more severe hemodynamic impairment) and the number with anterior infarctions increased with age. A history of smoking was significantly more common among younger patients, whereas a history of diabetes, hypercholesterolemia, and treated hypertension was obtained significantly more frequently among older patients.

In-Hospital Mortality

Of the 9720 patients enrolled in the study, 1035 (10.6 percent) had died within six months after discharge -- 772 (7.9 percent) during hospitalization and 263 (2.9 percent of the survivors) during the six months after discharge.

In the univariate analysis (Table 2), age was a powerful predictor of in-hospital mortality, with an odds ratio of 2.9 (95 percent confidence interval, 2.3 to 3.6) for patients 61 to 70 years old as compared with younger patients, and an odds ratio of 8.9 (7.4 to 10.8) for patients more than 70 years old. A history of smoking was found to be a "protective" factor in relation to the outcome of a first infarction. This paradoxical finding may be due to the base-line characteristics of the patients who smoked and, in particular, to the fact that smokers were younger than nonsmokers. The apparent protective effect of a history of smoking must be interpreted as a confounded result, because the statistical significance of this variable disappeared when the data were adjusted for age. All the other variables tested in the univariate analysis were associated with higher mortality, but the magnitude of these associations differed greatly among the variables. Weak associations (odds ratio, <2) were found between mortality and a longer interval between the onset of symptoms and arrival at the hospital, a history of hypertension, the presence of non-insulin-dependent diabetes, and chronic angina (angina present for more than one month). Stronger associations (odds ratio, >2) were found for advanced age, female sex, high Killip class at admission, insulin-dependent diabetes, anterior-wall infarction, and a high number of leads ( >= 6) with ST-segment elevation on the admission electrocardiogram.

View this table:
[in this window]
[in a new window]
 
Table 2. In-Hospital Mortality, According to Analysis of Study Variables.

 
To assess the independent contribution of age, the same variables were introduced into a multivariate model (Table 2). Older age was confirmed to be an independent predictor of mortality, with an odds ratio of 2.2 (95 percent confidence interval, 1.6 to 2.9) for patients 61 to 70 years old and 3.9 (95 percent confidence interval, 2.9 to 5.3) for patients more than 70 years old. The other variables that were confirmed as significantly associated with in-hospital mortality after adjustment were high Killip class on admission, anterior location of the infarct, and a high number of leads with ST-segment elevation on the admission electrocardiogram.

When age was introduced into the multivariate model as a continuous variable, its predictive value for in-hospital mortality was confirmed; the risk was estimated to increase exponentially by about 6 percent per year (odds ratio, 1.06; 95 percent confidence interval, 1.05 to 1.07). The overall in-hospital mortality was 1.9 percent among patients 40 years old or younger, which increased to 31.9 percent among patients more than 80 years old; values for the indicators of infarct size did not increase significantly with age (Figure 1).


View larger version (66K):
[in this window]
[in a new window]
 
Figure 1. Mortality among Patients with First Myocardial Infarctions, According to Age and Indicators of Infarct Size.

Age was considered a continuous variable, with the age of 40 years as the reference value. Values for the indicators of the size of the myocardial infarction did not show any increase with age.

Mortality during hospitalization increased exponentially with age; it is shown as an odds ratio (OR, calculated with the formula x = 0.1002588 x e0.0575x) and 95 percent confidence interval (CI). The curves for the indicators of infarct size show the percentage of patients with peak creatine kinase ratios more than six times the upper normal value of the study-center laboratory (CK >6 x normal) and the percentage with involvement of more than six electrocardiographic leads at entry.

Mortality during the interval from discharge to six months after discharge also increased exponentially with age; it is shown as a relative risk (RR, calculated with the formula x = 0.1094814 x e0.0553x). The curves for the indicators of infarct size show the percentage of patients with peak creatine kinase ratios more than six times normal and the percentage with QRS scores above 10.

 
Causes of In-Hospital Deaths

About half the deaths in each age group occurred on the day of the infarction or the day after -- i.e., in 70 (56 percent) of the 125 patients 60 years old or younger who died, 112 (46 percent) of the 244 patients 61 to 70 years old who died, and 210 (52 percent) of the 403 patients more than 70 years old who died.

Electromechanical dissociation occurred in 13 percent of the patients 60 years old or younger, 20 percent of those 61 to 70 years old, and 25 percent of those more than 70 years old. Conversely, the rate of death due to ventricular fibrillation decreased, from 15 percent among patients 60 years old or younger to 10 percent among those 61 to 70 years old and to 6 percent among those more than 70 years old. Autopsies were performed in 158 (20 percent) of the 772 patients who died during hospitalization (25 percent, 18 percent, and 21 percent of the three age groups, respectively). Cardiac rupture was found post mortem in 103 of these patients, or 65 percent, and its incidence increased progressively, from 19 percent among those 60 years old or younger (6 of 31 patients) to 58 percent among those 61 to 70 years old (25 of 43 patients) and to 86 percent among those more than 70 years old (72 of 84 patients). The variable incidence of cardiac rupture in previous reports may well be due to the inclusion of patients from different age groups,12 but a possible role of thrombolytic therapy cannot be excluded. At least two coronary arteries with severe stenoses were found in 40 percent of patients 60 years old or younger, 35 percent of those 61 to 70 years old, and 32 percent of those more than 70 years old. Infarcts involving less than 10 percent of the left ventricle were found in 20 percent of patients 60 years old or younger, 23 percent of those 61 to 70 years old, and 32 percent of those more than 70 years old.

Evaluation of Survivors before Discharge

The QRS score decreased progressively with age, but the differences among the three age groups were not significant. The percentage of patients with injury to more than 45 percent of myocardial segments (expressed by the akinetic-dyskinetic score) was 3 percent in the group 60 years old or younger, 4 percent in the group 61 to 70 years old, and 6 percent in the group more than 70 years old; the mean left ventricular ejection fraction was 50.8 ±11.5, 49.1 ±11.6, and 47.4 ±12.9, respectively.

Mortality from Discharge to Six Months

The mortality rate during the six months after discharge increased significantly, from 1.3 percent among patients 60 years old or younger to 3.0 percent among those 61 to 70 years old and to 7.4 percent among those more than 70 years old (Table 3). After adjustment, age continued to be associated with a higher risk of death after discharge, with a relative risk of 2.0 (95 percent confidence interval, 1.4 to 2.8) for patients 61 to 70 years old and 3.9 (95 percent confidence interval, 2.7 to 5.6) for those more than 70 years old. When age was introduced into the adjusted model as a continuous variable, the risk of death during the first six months after discharge was estimated to increase exponentially by 5.7 percent per year (relative risk, 1.06; 95 percent confidence interval, 1.04 to 1.07). Overall mortality after hospital discharge was 0.8 percent among patients 40 years old or younger and 11.6 percent among those more than 80 years old (Figure 1); the QRS scores at discharge were similar in these two age groups.

View this table:
[in this window]
[in a new window]
 
Table 3. Post-discharge Mortality, According to Analysis of Study Variables.

 
Discussion

In this study of a large group of patients presenting with first myocardial infarctions who were eligible to receive thrombolytic therapy, we observed that in-hospital mortality increased from 2.8 percent among patients 60 years old or younger to 19.0 percent among those more than 70 years old, confirming the findings of previous studies2,3,4,5. The mortality rate during the first six months after discharge was 1.3 percent among patients 60 years old or younger, but 7.4 percent among those more than 70 years old. The age-related increase in mortality did not appear to be explained by larger infarcts, as assessed by peak creatine kinase ratios and QRS scores on the electrocardiogram. However, both those measurements have limited precision in determining infarct size. As the age of the patients increased, we observed an increase in the frequency of Killip classes 3 and 4 (reflecting more severe hemodynamic dysfunction), echocardiographic evidence of left ventricular dysfunction, clinical signs of left ventricular failure, electromechanical dissociation, and cardiac rupture. Thus, older patients had more severe hemodynamic compromise than younger patients, despite the fact that the measurements of infarct size were similar across all age groups. Although the patients from all age groups who survived until discharge had a similar degree of estimated myocardial injury, the mortality rate during the first six months after discharge increased exponentially with age and was about six times higher among patients more than 70 years old than among those 60 years old or younger.

The explanation for our findings is unclear. Although the extent of coronary atherosclerosis increases with age, the number of coronary obstructive lesions was shown to plateau after the age of 60 in a series of 600 men who underwent autopsy13 and in a similar series of women14. Also, myocardial infarction was the very first manifestation of ischemic heart disease in about 60 percent of our patients, irrespective of age. Such patients often do not have severe stenoses in the infarct-related artery15 and have less severe and less extensive coronary atherosclerosis detectable by coronary angiography than patients presenting with a long history of uncomplicated stable angina16. Furthermore, in our selected group of patients with myocardial infarction who presented with ST-segment elevation, and who died and subsequently underwent autopsy (20 percent of patients who died during hospitalization), the number and degree of critical coronary stenoses did not differ according to age group. Consequently, the higher mortality rate among older patients may not be related to more extensive coronary artery disease.

Echocardiographic measurements of left ventricular dimensions remain unchanged up to the ninth decade of life, whereas left-ventricular-wall thickness increases by approximately 30 percent between the ages of 25 and 80 years17. The rate of myocardial contraction and relaxation decreases slightly with age, and the response to beta-adrenergic stimulation is considerably reduced18. Furthermore, postmortem studies indicate a sustained loss of myocardial cells in old age, with considerable expansion of the interstitial space and hypertrophy of myocardial cells19 -- findings similar to those in the hearts of old rats20 -- which could result in a greater impairment of left ventricular function and a greater risk of cardiac rupture in older patients even if the estimated size of their infarct is similar to or smaller than that of younger patients. Reductions in lung compliance, renal function, and catecholamine response may be additional factors responsible for the worse outcome of myocardial infarction in older patients. The tendency of coexisting illnesses (such as pneumonia or pulmonary emboli) to develop in elderly patients may also have a role.

Our findings may be applicable only to patients presenting with ST-segment elevation who are eligible to receive thrombolytic therapy, but our data on this large group of patients suggest that besides infarct size, older age is a major, independent risk factor for death among patients with acute myocardial infarction. Furthermore, research should be directed toward the prevention of cardiac failure, rupture, and electromechanical dissociation in the elderly. For example, beta-blockers may reduce myocardial rupture due to infarction21. Ongoing multicenter trials22,23 may eventually show whether nitrates or angiotensin-converting-enzyme inhibitors lower mortality after acute myocardial infarction in older patients with cardiac failure by reducing their vulnerability to myocardial necrosis.


Source Information

A complete list of the investigators and centers participating in GISSI-2 has been published (Lancet 1990;336:65-71). The steering committee consisted of G.A. Feruglio, A. Lotto, F. Rovelli, P. Solinas, L. Tavazzi, and G. Tognoni. These studies were endorsed by the Associazione Nazionale Medici Cardiologi Ospedalieri and the Istituto di Ricerche Farmacologiche Mario Negri.

Address reprint requests to Dr. Maggioni at the GISSI-2 Coordinating Center, Via Eritrea 62, 20157 Milan, Italy.

References

  1. Gurwitz JH, Goldberg RJ, Gore JM. Coronary thrombolysis for the elderly? JAMA 1991;265:1720-1723. [Free Full Text]
  2. Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto Miocardico (GISSI). Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Lancet 1986;1:397-402. [CrossRef][Medline]
  3. ISIS-2 (Second International Study of Infarct Survival) Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17 187 cases of suspected acute myocardial infarction: ISIS-2. Lancet 1988;2:349-360. [Medline]
  4. Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto Miocardico. GISSI-2: a factorial randomised trial of alteplase versus streptokinase and heparin versus no heparin among 12 490 patients with acute myocardial infarction. Lancet 1990;336:65-71. [Medline]
  5. ISIS-3 (Third International Study of Infarct Survival) Collaborative Group. ISIS-3: a randomised comparison of streptokinase vs tissue plasminogen activator vs anistreplase and of aspirin plus heparin vs aspirin alone among 41 299 cases of suspected acute myocardial infarction. Lancet 1992;339:753-770. [Medline]
  6. Killip T III, Kimball JT. Treatment of myocardial infarction in a coronary care unit: a two year experience with 250 patients. Am J Cardiol 1967;20:457-464. [CrossRef][Medline]
  7. Wagner GS, Freye CJ, Palmeri ST, et al. Evaluation of a QRS scoring system for estimating myocardial infarct size. I. Specificity and observer agreement. Circulation 1982;65:342-347. [Free Full Text]
  8. De Vita C, Franzosi MG, Geraci E, et al. GISSI-2: mortality plus extensive left-ventricular damage as "end-point." Lancet 1990;335:289-289. [CrossRef][Medline]
  9. Folland ED, Parisi AF, Moynihan PF, Jones DR, Feldman CL, Tow DE. Assessment of left ventricular ejection fraction and volumes by real-time, two-dimensional echocardiography: a comparison of cineangiographic and radionuclide techniques. Circulation 1979;60:760-766. [Free Full Text]
  10. Basic methods of analysis. In: Schlesselman JJ, Stolley PD. Case-control studies: design, conduct, analysis. New York: Oxford University Press, 1982:171-226.
  11. Multivariate analysis. In: Schlesselman JJ, Stolley PD. Case-control studies: design, conduct, analysis. New York: Oxford University Press, 1982:227-90.
  12. Bates RJ, Beutler S, Resnekov L, Anagnostopoulos CE. Cardiac rupture -- challenge in diagnosis and management. Am J Cardiol 1977;40:429-437. [CrossRef][Medline]
  13. White NK, Edwards JE, Dry TJ. The relationship of the degree of coronary atherosclerosis with age, in men. Circulation 1950;1:645-654. 
  14. Ackerman RF, Dry TJ, Edwards JE. Relationship of various factors to the degree of coronary atherosclerosis in women. Circulation 1950;1:1345-1354. [Medline]
  15. Hackett D, Davies G, Maseri A. Pre-existing coronary stenoses in patients with first myocardial infarction are not necessarily severe. Eur Heart J 1988;9:1317-1323. [Free Full Text]
  16. Bogaty P, Brecker SJ, White SE, et al. Comparison of coronary angiographic findings in acute and chronic first presentation of ischemic heart disease. Circulation 1993;87:1938-1946. [Free Full Text]
  17. Gerstenblith G, Frederiksen J, Yin FCP, Fortuin NJ, Lakatta EG, Weisfeldt ML. Echocardiographic assessment of a normal adult aging population. Circulation 1977;56:273-278. [Free Full Text]
  18. Vestal RE, Wood AJJ, Shand DG. Reduced beta-adrenoceptor sensitivity in the elderly. Clin Pharmacol Ther 1979;26:181-186. [Medline]
  19. Olivetti G, Melissari M, Capasso JM, Anversa P. Cardiomyopathy of the aging human heart: myocyte loss and reactive cellular hypertrophy. Circ Res 1991;68:1560-1568. [Free Full Text]
  20. Capasso JM, Palackal T, Olivetti G, Anversa P. Severe myocardial dysfunction induced by ventricular remodeling in aging rat hearts. Am J Physiol 1990;259:H1086-H1096. [Free Full Text]
  21. ISIS-1 (First International Study of Infarct Survival) Collaborative Group. Mechanisms for the early mortality reduction produced by beta-blockade started early in acute myocardial infarction: ISIS-1. Lancet 1988;1:921-923. [Erratum, Lancet 1988;2:292.] [CrossRef][Medline]
  22. ISIS-4 Collaborative Group. Fourth International Study of Infarct Survival: protocol for a large simple study of the effects of oral mononitrate, of oral captopril, and of intravenous magnesium. Am J Cardiol 1991;68:Suppl:87D-100D. [CrossRef][Medline]
  23. GISSI-3-Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto Miocardico. GISSI-3 study protocol on the effects of lisinopril, of nitrates, and of their association in patients with acute myocardial infarction. Am J Cardiol 1992;70:Suppl:62C-69C. [Medline]

 

This Article
-Abstract

Tools and Services
-Add to Personal Archive
-Add to Citation Manager
-Notify a Friend
-E-mail When Cited

More Information
-PubMed Citation

This article has been cited by other articles:



HOME  |  SUBSCRIBE  |  SEARCH  |  CURRENT ISSUE  |  PAST ISSUES  |  COLLECTIONS  |  PRIVACY  |  TERMS OF USE  |  HELP  |  beta.nejm.org

Comments and questions? Please contact us.

The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved.