Thirty-Year Trends (1975 to 2005) in the Magnitude of, Management of, and Hospital Death Rates Associated With Cardiogenic Shock in Patients With Acute Myocardial Infarction
A Population-Based Perspective
Background— Limited information is available about potentially changing and contemporary trends in the incidence and hospital death rates of cardiogenic shock complicating acute myocardial infarction. The objectives of our study were to examine 3-decade-long trends (1975 to 2005) in the incidence rates of cardiogenic shock complicating acute myocardial infarction, patient characteristics and treatment practices associated with this clinical complication, and hospital death rates in residents of a large central New England community hospitalized with acute myocardial infarction at all area medical centers.
Methods and Results— The study population consisted of 13 663 residents of the Worcester (Mass) metropolitan area hospitalized with acute myocardial infarction at all greater Worcester medical centers during 15 annual periods between 1975 and 2005. Overall, 6.6% of patients developed cardiogenic shock during their index hospitalization. The incidence rates of cardiogenic shock remained stable between 1975 and the late 1990s but declined in an inconsistent manner thereafter. Patients in whom cardiogenic shock developed had a significantly greater risk of dying during hospitalization (65.4%) than those who did not develop cardiogenic shock (10.6%) (P<0.001). Encouraging increases in hospital survival in patients with cardiogenic shock, however, were observed from the mid-1990s to our most recent study years. Several patient demographic and clinical characteristics were associated with an increased risk for developing cardiogenic shock.
Conclusions— Our findings indicate improving trends in the hospital prognosis associated with cardiogenic shock. Given the high death rates associated with this clinical complication, monitoring future trends in the incidence and death rates and the factors associated with an increased risk for developing cardiogenic shock remains warranted.
Received August 14, 2008; accepted December 31, 2008.
Although numerous clinical complications are associated with the development of acute myocardial infarction (AMI), none are more potentially devastating or carry a worse prognosis than cardiogenic shock.1–7
Clinical Perspective p 1219
Despite marked advances in medical treatment, revascularization techniques, and mechanical support during the past 2 decades, cardiogenic shock is still the most common cause of hospital mortality associated with AMI.7,8 On the other hand, data obtained from a limited number of recent studies suggest possible declines in the hospital mortality associated with cardiogenic shock,5,7,9–13 partially associated with the implementation of early revascularization therapy based on recent practice guidelines.14 Few studies, however, have examined contemporary and changing trends in the magnitude of or mortality associated with cardiogenic shock, particularly from the more generalizable perspective of a population-based investigation.
In 2 prior publications from the Worcester (Mass) Heart Attack Study, we described changing trends in the magnitude of and hospital outcomes associated with cardiogenic shock in residents of this large central New England metropolitan area hospitalized with AMI at all area medical centers.6,7 In the present study, we provide an extended 3-decade-long (1975 to 2005) perspective into changing trends in the incidence rates of cardiogenic shock, factors associated with the occurrence of cardiogenic shock complicating AMI, hospital treatment practices, and short-term death rates.15–17
The study population consisted of greater Worcester residents hospitalized with a discharge diagnosis of AMI at all teaching and community hospitals in the Worcester metropolitan area during the 15 individual study years of 1975 (n=781), 1978 (n=845), 1981 (n=998), 1984 (n=714), 1986 (n=765), 1988 (n=659), 1990 (n=766),1991 (n=848), 1993 (n=953), 1995 (n=949), 1997 (n=1059), 1999 (n=1027), 2001 (n=1239), 2003 (n=1157), and 2005 (n=903). Originally, 16 hospitals were included in this population-based investigation, but currently only 11 are included as a result of hospital closures or conversion to chronic care or rehabilitation facilities. Potentially eligible patients were identified through a review of computerized hospital databases of patients with International Classification of Disease discharge diagnoses consistent with the possible presence of AMI (eg, AMI, unstable angina). The medical records of all potentially eligible patients, who had to be residents of the Worcester metropolitan area because this study is population based, were reviewed in a standardized manner, and the diagnosis of AMI was confirmed according to preestablished criteria that have been previously described.15–17
Cardiogenic shock was defined as systolic blood pressure <80 mm Hg in the absence of hypovolemia and associated with cyanosis, cold extremities, changes in mental status, persistent oliguria, or congestive heart failure.6,7 The definition of cardiogenic shock remained the same during all periods studied. This disorder was defined so that patients with classic signs and symptoms of this clinical syndrome would be included.
Information was abstracted from the hospital medical records of greater Worcester residents with independently validated AMI by trained nurse and physician-reviewers. Information was collected on the patient’s demographic characteristics, medical history, clinical findings, and use of cardiac medications and specialized cardiac procedures as these therapies became available in clinical practice.
Differences in the demographic and clinical characteristics and treatment practices of patients with compared with those without cardiogenic shock, as well as short-term death rates, were examined using χ2 tests of statistical significance; differences in selected continuous variables between patients with and without cardiogenic shock were examined with t tests (Tables 1 and 2⇓). Differences in the characteristics of patients with cardiogenic shock who survived the acute hospitalization compared with those who did not were examined through the use of similar analytic tests.
Changes over time in the incidence rates of cardiogenic shock were assessed with the Mantel-Haenszel χ2 test for trends (Figure 1). Trends in hospital case fatality rates (CFRs), stratified according to the presence of cardiogenic shock, were analyzed with Mantel-Haenszel methods (Figure 2). Multivariable logistic regression models were used to assess differences from baseline in the incidence rates of cardiogenic shock over the period under study while controlling for potentially confounding demographic and clinical factors (Table 3). The variables controlled for in our regression models were included either because they differed between our respective comparison groups or because they had been previously shown to be possible risk factors for cardiogenic shock.
Because of the nonrandomized nature of the present study and because our methods of data collection did not allow us to determine whether a medical therapy or surgical intervention preceded or followed the occurrence of cardiogenic shock, we did not control for the use of various coronary reperfusion/revascularization procedures or medical therapies in our regression analyses. Our approach to model building focused on the hypothesis that changes in the incidence rates of cardiogenic shock over time were the result of changes in the characteristics of the hospitalized study sample.
We examined the impact of cardiogenic shock on hospital mortality by calculating in-hospital CFRs. Multivariable logistic regression analyses were used to assess the overall effect of cardiogenic shock on hospital mortality and to study changes over time in the hospital CFRs associated with cardiogenic shock while controlling for several potentially confounding factors that have previously been shown to be of prognostic importance (Table 4). This approach to model development was similar to that described for the development of cardiogenic shock. Human subjects’ approval for the review of hospital medical records in this cold pursuit disease surveillance project was obtained from the Committee for the Protection of Human Subjects in Research at the University of Massachusetts Medical School.
The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
The total study sample consisted of 13 663 greater Worcester residents hospitalized with AMI, of whom 6.6% (n=905) developed cardiogenic shock. The mean age of the study sample was 68.9 years; 58.5% were men; and 95.2% were white.
Baseline Demographic and Clinical Characteristics
Patients who developed cardiogenic shock during hospitalization for AMI were significantly older and were more likely to be women; to have a do-not-resuscitate order; to have a history of diabetes mellitus, heart failure, or MI; to present to greater Worcester hospitals with dyspnea; and to develop a Q-wave MI during hospitalization compared with patients who did not develop cardiogenic shock (Table 1). Patients with cardiogenic shock had significantly lower blood pressure, total serum cholesterol, and estimated glomerular filtration rate findings at the time of hospital admission but higher heart rates and serum glucose levels than patients who did not develop shock.
To provide more contemporary insights into the characteristics of patients with AMI who were at risk for developing cardiogenic shock, we examined differences in these characteristics in patients hospitalized with AMI during our 3 most recent study years (Table 1). Differences in the demographic characteristics of patients who did compared with those who did not develop cardiogenic shock were no longer apparent, whereas differences in previously observed clinical and physiological factors remained.
Hospital Treatment Practices
Patients who developed cardiogenic shock during hospitalization for AMI were significantly less likely to be treated with aspirin, β-blockers, calcium channel blockers, and lipid-lowering agents during hospitalization than patients who did not develop shock. Patients who developed cardiogenic shock were more likely to have been prescribed thrombolytic therapy and to have undergone cardiac catheterization, coronary artery bypass surgery, and a percutaneous coronary intervention (PCI) than patients who did not experience cardiogenic shock (Table 2). Patients who developed cardiogenic shock were significantly more likely to have received mechanical support during hospitalization through intraaortic balloon counterpulsation.
With regard to the elderly (≥65 years of age) specifically, those who developed cardiogenic shock were significantly less likely to have undergone cardiac catheterization (39.9%, 28.8%, 21.7%), a PCI (23.0%, 15.8%, 15.8%), or coronary artery bypass surgery (7.0%, 6.0%, 0%) and to have received intraaortic balloon counterpulsation (34.1%, 23.5%, 13.3%) with advancing age (65 to 74, 75 to 84, ≥85 years, respectively; P<0.001). Similar age-related differences in the use of cardiac diagnostic and interventional procedures were observed in patients who did not develop cardiogenic shock.
In examinations of the differences in the use of these treatment practices in patients who were hospitalized in our 3 most recent study years (Table 2), differences in the use of various treatment approaches either no longer remained apparent or became attenuated. The use of beneficial cardiac medications increased markedly over time in patients with AMI, regardless of the development of cardiogenic shock. The use of thrombolytic therapy declined considerably in both patient groups during recent years, whereas the use of PCI and intraaortic balloon counterpulsation increased markedly.
Changing Trends in the Incidence Rates of Cardiogenic Shock
Between 1975 and the late 1980s, the incidence of cardiogenic shock remained relatively stable, averaging ≈7.5% over this period (Figure 1). Although the proportion of patients with AMI who developed cardiogenic shock during hospitalization was somewhat inconsistent thereafter, declines in the incidence rates of cardiogenic shock were observed beginning in 1990, reaching a nadir in incidence rates of 4.1% in 2003. Overall, significant changes were found in the frequency of cardiogenic shock complicating AMI during the years under study (P<0.01). Of the patients who developed cardiogenic shock, ≈78% either presented with or developed this hemodynamic disturbance during the first day of hospitalization during our first 3 study years (1975, 1978, 1981) compared with ≈77% in our 3 most recent years under investigation.
Among the elderly, the overall incidence rates of cardiogenic shock did not vary appreciably (65 to 74 years of age, 7.3%; 75 to 84 years of age, 8.2%; ≥85 years of age, 7.0%). In 2003 and 2005, the incidence rates of cardiogenic shock declined with advancing age (65 to 74 years of age, 6.8%; 75 to 84 years of age, 4.9%; ≥85 years of age, 4.1%).
We carried out a series of regression analyses to examine differences from the baseline study year of 1975 in the rates of cardiogenic shock while controlling for several factors that might affect the likelihood of developing shock in patients with AMI (Table 3). The results of these analyses were similar to those of our unadjusted analyses of declines in the incidence rates of cardiogenic shock during the most recent years under investigation. Statistically significant declines in the proportion of patients with AMI who developed cardiogenic shock were noted during 2001, 2003, and 2005 (compared with the referent year of 1975). Similar, albeit attenuated, changes were observed in the risk of developing shock between 1997 and 2005 when information about whether the MI was an ST-segment or non-ST-segment elevation MI was available, as well as when data were collected about additional physiological findings.
Overall, 65.4% of patients with AMI who developed cardiogenic shock died during hospitalization compared with 10.6% of patients who did not develop cardiogenic shock (P<0.001). In elderly patients with AMI, the overall hospital CFRs associated with cardiogenic shock increased with advancing age (65 to 74 years of age, 65.1%; 75 to 84 years of age, 75.8%; ≥85 years of age, 75.2%).
A logistic regression analysis was carried out to examine the association between occurrence of cardiogenic shock and hospital CFRs while controlling for several demographic and clinical factors of prognostic importance. The results of this analysis confirmed the markedly higher risk of dying during hospitalization among patients who developed cardiogenic shock compared with those who did not (adjusted odds ratio [OR], 17.8; 95% confidence interval [CI], 14.5 to 20.9). When we restricted our analysis to patients hospitalized during 2003 and 2005, patients with cardiogenic shock remained at markedly increased risk for dying during hospitalization compared with patients who did not develop cardiogenic shock, although the absolute risk of dying for patients with cardiogenic shock was considerably lower than during earlier study years (adjusted OR, 12.5; 95% CI, 7.81 to 19.83).
Examination of the changing short-term death rates associated with cardiogenic shock (Figure 2) showed that in 1975 and 1978, 76.1% of patients who developed cardiogenic shock died in the hospital compared with 16.5% of patients who did not develop this complication. However, in 2003 and 2005, 45.4% of patients with shock died during hospitalization compared with 7.3% of patients who did not develop shock (P<0.001) (Table 4).
Patients who developed cardiogenic shock during the most recent years under study were significantly less likely to have died compared with patients hospitalized with AMI in the 1970s and 1980s (Table 4). These trends were apparent regardless of the demographic or clinical characteristics controlled for.
In 2003 and 2005, the hospital death rates associated with cardiogenic shock increased with advancing age in the elderly from 35.7% in patients 65 to 74 years of age to 57.1% and 64.7% in patients 75 to 84 and ≥85 years of age, respectively.
Because the length of hospital stay has declined markedly in our study population over the past 30 years (mean, 18 days in 1975; mean, 5 days in 2005), we examined changing trends in the 30-day CFRs after hospital presentation in patients who developed cardiogenic shock. The results of this analysis confirmed the declining short-term death rates in patients who developed cardiogenic shock over time. In 1975 and 1978, the 30-day death rate after hospital admission for patients with cardiogenic shock was 20.7%; this death rate was 17.0% in 1990 and 1991 and 12.8% in 2003 and 2005.
Characteristics of Hospital Survivors With Cardiogenic Shock
Patients who survived an episode of cardiogenic shock were significantly younger, were less likely to have had a history of coronary disease or heart failure, were more likely to have higher diastolic blood pressure findings and higher levels of estimated glomerular filtration rate, and were more likely to have been treated with effective cardiac medications and interventional procedures than patients who died of cardiogenic shock (Table 5). Similar findings were observed when we compared patients who died of cardiogenic shock with those who survived this clinical syndrome in our 3 most recent study years.
The results of our population-based observational study provide insights into changing trends in the magnitude, management, and hospital outcomes of patients with cardiogenic shock complicating AMI and the characteristics of patients likely to develop and survive cardiogenic shock.
Incidence Rates of Cardiogenic Shock
The incidence rates of cardiogenic shock after AMI have ranged from 5% to 15% in previously published studies. This relatively wide range reflects the various definitions of AMI and cardiogenic shock used, the use of representative as opposed to more highly selected patient samples, the time periods under study, and the use of therapeutic options that may reduce the risk of cardiogenic shock. The overall incidence rates of cardiogenic shock observed in the present study fall within this range.
The relatively few studies that have examined changing trends in the incidence rates of cardiogenic shock after AMI have yielded conflicting results. In the National Hospital Discharge Survey, declines in the frequency of cardiogenic shock were observed between 1979 and the early 2000s, in concert with increasing rates of PCI use.12 In contrast, findings from the National Registry of Myocardial Infarction showed either stable or slightly increasing incidence rates of cardiogenic shock over the 10-year period beginning in 1995 in patients hospitalized with ST-segment elevation MI.5
The results of our study suggest that patients hospitalized with AMI in the 2000s were less likely to develop cardiogenic shock than greater Worcester residents hospitalized with AMI during earlier study years. This finding is all the more impressive given the considerable aging of our patient population and the increasing prevalence of serious comorbidities, including diabetes mellitus, heart failure, and prior stroke.17
Although we cannot ascertain the reasons behind these encouraging trends, a number of likely contributory factors can be suggested. We have previously documented progressive declines since the mid-1970s in the hospital death rates of greater Worcester residents hospitalized with AMI at all area medical centers.15–17 Patients presenting with AMI in recent study years were more likely to have had prior MIs or to present with a non-Q-wave MI than during early study years. This may explain some of the changes in the incidence rates of cardiogenic shock found in the present study and in our earlier investigations.
Early revascularization, particularly via percutaneous intervention in patients with ST-segment elevation AMI, has been shown to decrease infarct size and hemodynamic compromise and to improve patient survival. Because of our data collection methods, we were unable to examine the association between the increased use of coronary reperfusion and revascularization strategies and the risk of developing or dying of cardiogenic shock. Nevertheless, our data and the findings from other recent investigations show a significant increase in the use of primary PCI over the past decade for the treatment of patients hospitalized with AMI. Significant resources in our community (as in others) have been spent attempting to improve the timeliness with which revascularization is performed in these patients. Although gains in time from hospital presentation to balloon initiation appear promising, efforts to decrease the extent of delay from symptom onset to hospital presentation remain frustrated. Indeed, the average duration of prehospital delay in patients presenting with cardiogenic shock during our most recent study years was nearly 4 hours.
Other possible reasons for the observed decline in the incidence rates of cardiogenic shock may include increased monitoring efforts, particularly of patient populations at increased risk for developing shock, use of increasingly effective cardiac medications for the secondary prevention of recurrent events, and careful attention to the maintenance of desirable hemodynamic parameters in patients with AMI.
Factors Associated With Cardiogenic Shock
Efforts to decrease the risk of developing cardiogenic shock in patients with AMI should focus on identifying patients who are at high risk for this serious complication and instructing them to seek care immediately after the onset of acute coronary symptoms so that appropriate monitoring, risk stratification, and intervention can be undertaken. Aggressive intervention may result in improved survival rates among patients in whom cardiogenic shock has developed.9,11,18 In the present study, patients who developed cardiogenic shock were older and were more likely to be women, to present with prior cardiovascular disease, or to experience various physiological findings than patients without shock.
Hospital CFRs After Cardiogenic Shock
Patients in whom cardiogenic shock developed continue to have a markedly higher risk of dying in the hospital than patients who did not develop cardiogenic shock. However, the hospital prognosis for greater Worcester residents hospitalized with cardiogenic shock has improved considerably during the most recent years under investigation. Between the mid-1970s and late 1980s, slightly more than three quarters of patients who developed cardiogenic shock died during hospitalization; in contrast, ≈2 of every 5 patients who developed shock between 2001 and 2005 died. Much of this decline occurred coincidentally with the increasingly effective management of patients with AMI with cardiac medications and coronary intervention approaches. This improvement in hospital survival may be due to the increasingly aggressive intervention strategies used or to changes in the natural history of shock, with fewer cases that subsequently result in death identified early in the course of the illness.
A number of nonrandomized studies suggest that PCI improves short-term survival in patients with cardiogenic shock, with survival contingent on the successful establishment of coronary reperfusion.19,20 Uncontrolled studies of coronary artery bypass grafting show that this revascularization approach improves short-term survival among patients with cardiogenic shock when they are treated soon after shock has developed.21 We were unable, however, to assess the role of these interventional procedures because we could not determine whether shock preceded or followed the use of these treatment strategies or to determine other reasons why certain patients received these therapeutic regimens and others did not.
Analyses from the Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries (GUSTO-I) trial suggest that the use of a more aggressive revascularization strategy in patients with cardiogenic shock, after initial treatment with a thrombolytic regimen, is associated with a reduction in short-term mortality after AMI.11 Findings from the Should We Emergently Revascularize Occluded Coronaries for Cardiogenic Shock (SHOCK) trial suggest benefits on 1-year survival with emergency revascularization and use of intraaortic counterpulsation.18,22 In the National Registry of Myocardial Infarction, a marked increase was found in the use of PCI in patients hospitalized for AMI at >650 medical centers between 1995 and 2004, and the increased use of this treatment strategy was associated with a marked decline in short-term mortality associated with cardiogenic shock.5
In current guidelines, AMI complicated by cardiogenic shock is listed as a class IA indication for PCI and a class IA indication for coronary artery bypass graft surgery if the patient has suitable coronary anatomy.14 In the National Hospital Discharge Survey12 and the National Registry of Myocardial Infarction,5 mortality rates from cardiogenic shock declined to <50% during the most recent years under investigation of these studies. In the Global Registry of Acute Coronary Events (GRACE) project, improvements in the management of patients with an acute coronary syndrome were accompanied by reductions in the magnitude and mortality associated with cardiogenic shock between 1999 and 2006.23
Cardiogenic shock continues to develop at a relatively high, although apparently declining, rate after AMI. The hospital death rate among patients with this complication remains high but continues to decline over time. Although the results of the SHOCK trial are encouraging, it is unclear to what extent the broader pool of patients who develop cardiogenic shock are currently managed with an aggressive coronary revascularization or reperfusion approach. It remains to be seen whether current efforts aimed at reducing the extent of prehospital delay and door-to-balloon times may lead to further declines in the incidence and CFRs of cardiogenic shock after AMI. It remains important to examine contemporary trends in the magnitude and short-term outcomes associated with cardiogenic shock and to develop a risk prognostication index for identifying patients most likely to develop this compromised hemodynamic state.
We wish to express our appreciation to all persons involved in the review of data for this project during the years under study.
Source of Funding
Grant support for this project was provided by the National Heart, Lung, and Blood Institute (RO1 HL35434).
Hochman JS, Boland J, Sleeper LA, Porway M, Brinker J, Col J, Jacobs A, Slater J, Miller D, Wasserman H. Current spectrum of cardiogenic shock and effect of early revascularization on mortality: results of an international registry. Circulation. 1995; 91: 873–881.
Holmes DR Jr, Califf RM, Van de Werf F, Berger PB, Bates ER, Simoons ML, White HD, Thompson TD, Topol EJ. Difference in countries’ use of resources and clinical outcome for patients with cardiogenic shock after myocardial infarction: results from the GUSTO trial. Lancet. 1997; 349: 75–78.
Berger PB, Holmes DR Jr, Stebbins AL, Bates ER, Califf RM, Topol EJ. Impact of an aggressive invasive catheterization and revascularization strategy on mortality in patients with cardiogenic shock in the Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries (GUSTO-I) trial: an observational study. Circulation. 1997; 96: 122–127.
Urban P, Stauffer JC, Bleed D, Khatchatrian N, Amann W, Bertel O, van den Brand M, Danchin N, Kaufmann U, Meier B, Machecourt J, Pfisterer M. A randomized evaluation of early revascularization to treat shock complicating acute myocardial infarction: the (Swiss) Multicenter Trial of Angioplasty for Shock-(S)MASH. Eur Heart J. 1999; 20: 1030–1038.
Antman EM, Anbe DT, Armstrong PW, Bates ER, Green LA, Hand M, Hochman JS, Krumholz HM, Kushner FG, Lamas GA, Mullany CJ, Ornato JP, Pearle DL, Sloan MA, Smith SC Jr, Alpert JS, Anderson JL, Faxon DP, Fuster V, Gibbons RJ, Gregoratos G, Halperin JL, Hiratzka LF, Hunt SA, Jacobs AK, for the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 1999 Guidelines for the Management of Patients With Acute Myocardial Infarction. ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2004; 110: 588–636.
Hochman JS, Sleeper LA, Webb JG, Sanborn TA, White HD, Talley JD, Buller CE, Jacobs AK, Slater JN, Col J, McKinlay SM, LeJemtel TH, Picard MH, Menegus MA, Boland J, Dzavik V, Thompson CR, Wong SC, Steingart R, Forman R, Aylward PE, Godfrey E, Desvigne-Nickens P, for the SHOCK Investigators. Early revascularization in acute myocardial infarction complicated by cardiogenic shock. N Engl J Med. 1999; 341: 625–634.
Bolooki H. Emergency cardiac procedures in patients in cardiogenic shock due to complications of coronary artery disease. Circulation. 1989; 79 (suppl): I-137–I-148.
The results of this population-based epidemiological study demonstrate that cardiogenic shock remains a relatively frequent complication of acute myocardial infarction, affecting ≈1 in every 15 patients hospitalized with acute myocardial infarction. Although the incidence rates of cardiogenic shock have remained relatively stable over the past 30 years (1975 to 2005) in our investigation of residents of a large New England metropolitan area hospitalized with acute myocardial infarction at all area medical centers, encouraging improvements in the hospital survival of these high-risk patients have occurred coincidentally with the increasingly aggressive management of patients who developed cardiogenic shock. Our findings also provide insights into the characteristics of patients who died after an episode of cardiogenic shock to whom targeted surveillance and therapeutic efforts might be directed.
Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.
Go to http://cme.ahajournals.org to take the CME quiz for this article.