Author + information
- Received December 30, 1999
- Revision received March 31, 2000
- Accepted June 2, 2000
- Published online October 1, 2000.
- Edward L Hannan, PhD∗,*,
- Michael J Racz, MA∗,
- Djavad T Arani, MD, FACC†,
- Thomas J Ryan, MD, FACC‡,
- Gary Walford, MD, FACC§ and
- Ben D McCallister, MD, FACC‖
- ↵*Reprint requests and correspondence: Dr. Edward L. Hannan, Department of Health Policy, Management and Behavior, School of Public Health, State University of New York, University at Albany, One University Place, Rensselaer, New York
The goal of this study was to learn more about the risk factors and short- and long-term outcomes for primary angioplasty.
Primary angioplasty (direct angioplasty without antecedent thrombolytic therapy) has been an effective alternative to thrombolytic therapy for patients with acute myocardial infarction (AMI). However, most reported studies have been compromised by small sample sizes and short observation times.
New York’s coronary angioplasty registry was used to identify New York patients undergoing angioplasty within 6 h of AMI between January 1, 1993 and December 31, 1996. Statistical models were used to identify significant risk factors for in-patient and long-term survival and to estimate long-term survival for all patients as well as various subsets of patients undergoing primary angioplasty.
The in-hospital mortality rate for all primary angioplasty patients was 5.81%. When patients in preprocedural shock (who had a mortality rate of 45%) were excluded, the in-hospital mortality rate dropped to 2.60%. Mortality rates for all primary angioplasty patients at one year, two years and three years were 9.3%, 11.3% and 12.6%, respectively. Patients treated with stent placement did not have significantly lower risk-adjusted in-patient or two-year mortality rates.
Primary angioplasty is a highly effective option for AMI.
Numerous studies have been conducted to determine the outcomes and benefits of primary percutaneous transluminal coronary angioplasty (PTCA) in the treatment of acute myocardial infarction (AMI), as well as its relative success in comparison to thrombolytic therapy (1–12).
A meta-analysis of various randomized clinical trials that was published in 1995 found a 44% reduction in six-week mortality with primary PTCA relative to thrombolytic therapy (odds ratio [OR] = 0.56; 95% confidence interval, 0.35, 0.94) (13). However, most of the studies were based on relatively small sample sizes. The authors of the meta-analysis concluded that, although the combined results suggest that PTCA may have superior outcomes to thrombolytic therapy, larger studies are needed to confirm the results.
Although this study did not have access to patients undergoing thrombolytic therapy without subsequent PTCA, it did include a large number (n = 2,291) of patients who underwent PTCA within 6 h of AMI without antecedent thrombolytic therapy (primary PTCA patients). The purposes of this study were to identify significant predictors of short-term (in-hospital) and long-term mortality for primary PTCA, to determine three-year survival for primary PTCA patients and to determine long-term survival for subsets of primary PTCA patients (patients with preprocedural shock and diabetes and patients undergoing stent placement).
In 1991, the New York State Department of Health (the Department) and its Cardiac Advisory Committee (CAC) established a registry for coronary angioplasty performed in the state, the Coronary Angioplasty Reporting System (CARS). The CAC is a group of cardiac surgeons, cardiologists, internists and patient advocates that advises the Department on issues related to the quality of coronary care, access to cardiac procedures and the prevention of coronary heart disease.
The cardiac catheterization laboratories in the 33 hospitals in which angioplasty is performed in New York are responsible for coding the CARS forms to capture the relevant information. When data fields are found to be missing, hospitals are contacted by the Department and asked to complete the missing information. Also, comprehensive audits of approximately half of the hospitals in each registry are conducted on behalf of the Department each year, and several hospitals have been asked to recode all or part of their data as a result of the audits.
For each patient undergoing PTCA in New York State, CARS contains demographic information; clinical risk factors; patient, operator and hospital identifiers; complications and discharge status from the hospital. An important data element related to this study is whether the patient suffered a recent AMI and the time period since the onset (coded as 0 to 6 h, 6 h to 24 h, number of days).
Another data base needed for the study was New York’s vital statistics death file, which identifies all residents of the state who have died each year. Since CARS and the death file contain patient social security numbers, the two databases were matched on social security numbers to obtain deaths subsequent to discharge after the index hospitalization.
Patients in the study consisted of all New York State residents in all of the 33 hospitals who underwent PTCA from 1993 to 1996 within 6 h after experiencing an AMI and did not undergo thrombolytic therapy within seven days before the PTCA. The study was limited to residents of the state because the vital statistics death file does not contain deaths for out-of-state residents, so those patients could not be tracked for long-term outcomes. In the interest of having the short-term (in-hospital) and long-term samples identical, out-of-state patients were also excluded from the short-term analyses. Although 12 h from the time of onset is the period most commonly used in the randomized clinical trials for PTCA versus thrombolysis, the 6-h time period for prior AMI was used because the only times available in the Registry were 6 h, 24 h and number of days (up to 21 days). Six hours was chosen rather than 24 h to ensure that the patient population consisted of patients who underwent primary angioplasty for acute evolving infarction.
As yet, there is no registry for patients treated with thrombolytic therapy in New York. Percutaneous transluminal coronary angioplasty patients who have undergone thrombolytic therapy within seven days before the procedure are identified in CARS, but no other time interval between thrombolytic therapy and PTCA is available. Thus, because we could not distinguish between patients who had undergone thrombolytic therapy subsequent to the recent AMI and patients who had undergone the therapy before the AMI, all patients who underwent thrombolytic therapy were not included in the study.
Study end points
Patients in the study were tracked to determine if they died at any time during the study period, and if so, how long they lived after undergoing the procedure. Deaths during the same admission as the procedure were identified using CARS, and deaths after discharge following the procedure were identified using the death file. For deaths after discharge, the time between the date of the procedure and the date of death was noted.
The prevalence and in-hospital mortality rate associated with each available determinant of mortality were calculated. These risk factors included the number of vessels diseased (with at least 70% stenosis) and lesion type attempted; patient age, gender, race and ethnicity; a variety of comorbidities and measures of the patient’s hemodynamic state and ventricular function. The use of coronary stents was also examined. All variables, including age and ejection fraction, were treated as categorical variables, and for some variables (ejection fraction, previous myocardial infarction, vessels diseased, lesion type attempted, race), more than two categories were defined. Chi-square tests were used to identify significant differences in mortality rates between categories.
A stepwise logistic regression model was developed using the LOGISTIC procedure in SAS, Version 6.12 (SAS Institute, Cary, North Carolina) in order to identify significant independent predictors of in-hospital mortality. The binary dependent variable was discharge status from the hospital after PTCA, with in-hospital mortality coded as a “1.” Candidates for the independent variables included all the demographic and clinical variables available in CARS, including the use of stents. Variables were chosen for inclusion in the model using a 0.05 level of significance. The C statistic was used to measure the discrimination of the model, and the Hosmer-Lemeshow statistic was used to measure the model’s calibration. Another stepwise logistic regression model was developed after excluding patients with preprocedural shock. The same rules and candidate covariates (except shock) were used in the second model as in the original one.
Significant risk factors for long-term survival (up to three years) were identified using stepwise Cox proportional hazards models and the SAS procedure PHREG (Version 6.12). All demographic and clinical variables available in CARS were used as candidate covariates, including the use of stents.
Table 1presents the prevalences and in-hospital mortality rates for various risk factors associated with primary PTCA. The total number of patients undergoing primary PTCA in New York between 1993 and 1996 was 2,291. A total of 133 (5.81%) patients died in the hospital before discharge. The 1993 to 1996 risk-adjusted in-patient mortality rates were 6.57, 6.07, 5.03 and 5.97. None of these rates were statistically different from the four-year rate.
Twenty-seven percent of the patients were at least 70 years old, and 5% were at least 80 years old. Thirty percent were women, and 9% were noncaucasian. Nearly 8% were in shock (systolic blood pressure less than 80 mm Hg or cardiac index less than 2.0 liters/min/m2), and another 15% were hemodynamically unstable (required pharmacologic or mechanical support to maintain blood pressure or output). Four percent underwent previous open heart surgery, and 4% underwent a previous PTCA. Eleven percent of the patients had congestive heart failure (CHF) during the same admission, and 14% had diabetes. Most patients (63%) had no more than one diseased vessel (stenosis ≥70%), and 1.5% had left main disease. Twenty-one percent of the patients underwent stent placement during the primary angioplasty. However, the percentages in the respective years 1993 to 1996 were 0.3%, 3.3%, 19.8% and 42.1%.
The in-hospital mortality rate rose with increasing age from 3% for patients under age 50 to 15.65% for patients between 80 and 89, and the trend was significant (p < 0.001, Table 1). Women had a significantly higher mortality rate than men (9.04% vs. 4.42%, p < 0.001) although subsequent multivariate analyses demonstrated that the difference is accounted for by their older ages. Hispanics had higher than average mortality rates, but the differences were not significant. Nearly all of the clinical risk factors available in the database were significantly (bivariately) related to in-hospital mortality. Exceptions were two intervention-related variables: lesion type attempted and stent placement. Risk factors with the highest mortality rates were shock (45% mortality), left main disease (40%), renal failure (creatinine >2.5 mg/dl or on dialysis, 31%), femoral/popliteal disease (23%) and CHF during the same admission (27%). All were significantly related to mortality at the 0.001 level. There was also a progressive increase in mortality rate with the number of vessels diseased (p < 0.001) and with decreasing left ventricular ejection fraction.
Table 2presents the significant risk factors for in-hospital mortality after primary PTCA along with their logistic regression coefficients, p values, OR and 95% confidence intervals for the ORs. The risk factor with by far the highest OR was shock (OR = 23.62), followed by left main disease (5.03) and hemodynamic instability (3.62). Other significant risk factors were age, CHF during the same admission, malignant ventricular arrhythmia, diabetes and stroke. Both the C statistic (0.913) and the Hosmer-Lemeshow statistic (p = 0.73) indicated that the model fit was highly acceptable.
Primary angioplasty patients in preprocedural shock had a 45% in-hospital mortality rate compared with a 2.60% mortality rate for primary angioplasty patients not in preprocedural shock. Thus, the mortality rate for shock patients was 17 times as high as it was for other primary angioplasty patients. In addition, 59% of all deaths among primary angioplasty patients were among patients with preprocedural shock, despite the fact that these patients comprised only 7.6% of all primary angioplasty patients.
Since preprocedural shock was such a dominant factor in mortality of primary angioplasty patients and since many previously reported randomized trials have excluded shock patients, Table 3presents the significant risk factors for primary angioplasty patients who were not in shock. Risk factors with the highest OR for in-hospital mortality were CHF during the same admission (OR= 4.42), previous PTCA (3.59) and malignant ventricular arrhythmia (3.52). Age, carotid disease, hemodynamic instability and diabetes were also significant. Again, the C statistic (0.847) and the Hosmer-Lemeshow statistic (p = 0.57) indicated excellent model fit.
Table 4presents the risk factors for long-term mortality among patients who underwent primary angioplasty after AMI. A comparison with Table 2 demonstrates that there are several long-term risk factors that were not short-term risk factors (female gender, ejection fraction, aortoiliac disease, creatinine above 2.5, renal failure with dialysis and two- or three-vessel disease). Risk factors with the highest risk ratios were shock (risk ratio = 4.66), dialysis (3.45) and left main disease (2.88). It is also notable that stent placement did not prove to yield significantly higher long-term survival than other types of angioplasty, as evidenced by its absence in the stepwise results.
Table 5presents the rates for three adverse outcomes (mortality, subsequent coronary artery bypass graft [CABG] surgery, subsequent PTCA) for five time periods after primary angioplasty (three months, six months, one year, two years, three years). As indicated, the mortality rate rose from 5.8% at three months to 12.6% at three years. A total of 13.1% of primary angioplasty patients underwent CABG surgery within three months of the index procedure, and this percentage rose to 16.7% at one year, 17.9% at two years and 19.4% at three years. The percentage of primary angioplasty patients who underwent a subsequent PTCA was initially much lower than the percentage undergoing CABG surgery (5.9% at three months and 12.1% at one year) but eventually approached the CABG surgery rate (15.6% at two years and 17.6% at three years).
Table 6presents the mortality rates for six time periods (in-hospital, three months, six months, one year, two years and three years) for all patients and for three subgroups of patients having the following characteristics: shock, diabetes and stent placement. Figure 1 provides graphs of the survival curves for all primary angioplasty patients and for the shock and diabetes patients.
As indicated in Table 6, diabetic patients undergoing primary angioplasty had a 12.4% in-hospital mortality rate, compared with the 5.8% mortality rate for all primary angioplasty patients. This differential gradually increased as the time after discharge from the hospital increased, to 8.3% at one year after discharge and 8.4% two years after discharge. The maximum differential of 10.2% (22.8% to 12.6%) occurred at three years after discharge.
Shock patients had an extremely high in-hospital mortality rate of 45.1%. However, once they were discharged alive from the hospital, their prognosis was similar to that of the average patient who underwent primary angioplasty. Between discharge from the hospital and three years later, 6.8% of all primary angioplasty patients died. For patients in shock during the same time span, 4.0% died.
Patients undergoing stent placement could be followed for only two years because most stent placements were performed during or after 1994. The in-hospital mortality for stent placement was 4.5% compared with 5.8% for all patients. This 1.3% differential increased gradually as time from discharge increased, to 2.5% at three months, 2.8% at one year and a maximum of 3.1% (11.3% to 8.2%) at two years.
Numerous studies have investigated the effectiveness of primary angioplasty in the treatment of AMI. Most of these studies have been small, randomized clinical trials that compared the relative effectiveness of primary angioplasty and thrombolytic therapy on the basis of short-term outcomes. Also, many of these randomized clinical trials excluded high-risk patients and do not reflect outcomes to be expected in patients treated in daily practice.
With regard to primary angioplasty mortality rates, Grines et al. (1), using a 12-h period as the maximum time between the onset of AMI and the performance of primary angioplasty, reported a 2.6% in-hospital mortality rate for 195 patients not in shock who underwent primary angioplasty. The Global Use of Strategies To Open occluded arteries (GUSTO) IIb Investigators, who used the same 12-h maximum period between AMI and angioplasty, reported a 5.7% 30-day mortality rate among 565 primary angioplasty patients, but no patients were in preprocedural shock (2).
In a report from the Second National Registry of Myocardial Infarction based on 4,939 patients who underwent primary angioplasty, Tiefenbrunn et al. (3) reported a 5.2% in-hospital mortality rate for nonshock patients and a 32% in-hospital mortality rate for the 4.2% of patients who had shock. Again, a 12-h maximum period between onset of AMI and primary angioplasty was used. Garcia et al. (4) reported a 2.8% in-hospital mortality rate for 109 patients undergoing primary angioplasty and a 4.6% mortality rate at six months. It does not appear that any of the patients were in preprocedural shock.
Ribeiro et al. (8), who used a maximum 6-h period between onset of the AMI and primary angioplasty, reported a 6% in-hospital mortality rate for 50 patients undergoing primary angioplasty. It is not clear whether any of those patients were in preprocedural shock. Every et al. (10), using a 6-h time frame between onset and performance of the procedure, reported a 5.5% in-hospital mortality rate for 1,272 primary angioplasty patients and a 12.1% mortality rate at three years. The incidence of preprocedural shock in the study was 11.7%. In a study of 1,000 consecutive primary angioplasty patients, O’Keefe et al. (11) reported a 7.8% in-hospital mortality rate. Patients in shock comprised 7.9% of the population and had a mortality rate of 44%; patients not in shock had a mortality rate of 5%.
Summary of results
In this study, which used a maximum 6-h period between onset of the AMI and performance of primary angioplasty, the in-hospital mortality rate was 5.81%. A total of 7.6% of the patients were in preprocedural shock, and the in-hospital mortality rate for those patients was 45%. The in-hospital mortality rate for patients not in preprocedural shock was 2.60%. For all primary angioplasty patients in the study, the mortality rate was 9.3% at one year, 11.3% at two years and 12.6% at three years.
In summary, in the previous studies, the range of in-hospital mortality rates for primary angioplasty patients not in shock varied from 2.6% to 5.2%. The in-hospital rate in our study was 2.6%. The two in-hospital mortality rates that were reported in other studies for shock patients undergoing primary angioplasty were 32% and 44%, compared with a 45% mortality rate in our study. The overall in-hospital mortality rates for primary angioplasty in the previously reported studies ranged from 5.7% to 7.8%; in our study we found a 5.81% mortality rate for all patients. The only study in the literature that reported a three-year mortality rate found a rate slightly lower than the rate found in our study (12.1% vs. 12.6%) (10).
Thus, the rates reported in our study are, for the most part, similar to those reported by other studies in the literature. Furthermore, the mortality rates reported in this study for primary angioplasty are lower than the rates reported for thrombolytic therapy in recent studies. For example, the National Registry of Myocardial Infarction-2 trial reported a 52% in-hospital mortality rate for patients in shock and a 5.4% mortality rate for patients not in shock (our corresponding rates were 45% and 2.6%, with a 5.81% overall rate) (3). The Primary Angioplasty in Myocardial Infarction trial reported a 6.5% in-hospital mortality rate (1). An obvious caveat is that our study did not randomize patients to primary angioplasty, so a selection bias may have existed.
There are differences between this study and most other studies reported in the literature. First, this was not a randomized clinical trial, and, in fact, no patients undergoing thrombolytic therapy as an alternative intervention for AMI were in the study. No such registry exists for patients undergoing thrombolytic therapy in New York. Although patients who underwent thrombolytic therapy before primary angioplasty are included in the Registry, this group was excluded from the study because it was impossible to determine from the Registry whether they had undergone thrombolytic therapy before or after the onset of the recent AMI. An advantage of our study is that, to the best of our knowledge, this is the first statewide population-based study of primary angioplasty. Except for patients undergoing prior thrombolytic therapy, the study includes all patients who underwent the procedure in New York State between 1993 and 1996.
The maximum allowable time frame used in this study for the period between onset of AMI and the performance of primary angioplasty was 6 h. Most studies in the literature use a 12-h maximum onset period and some use the 6 h time frame. There was not an option to use a 12-h period in this study because the only time periods less than 24-h that are recorded in the Registry are 0 to 6 h and 6 to 24 h. The 6-h time period was chosen rather than the 24 h time period to ensure that the patient population consisted of patients who underwent primary angioplasty for acute evolving infarction.
Another limitation of the study was that there was no information in the Registry that indicated whether patients undergoing primary angioplasty were ineligible for thrombolytic therapy. It should be noted that a contraindication for thrombolytic therapy is now a data element in the Registry.
An important ancillary finding of the study is that shock patients who were discharged alive from the hospital after primary angioplasty had lower three-year mortality rates than other primary angioplasty patients (4% vs. 6.8%, respectively). A caveat is that there were only 95 shock patients discharged alive from the hospital (out of 173 admissions). Another important finding was that patients treated with coronary stent placements did not have superior short-term or long-term outcomes compared with patients not receiving stent placements for primary angioplasty.
The authors would like to thank Kenneth Shine, MD, the Chair of New York State’s CAC, and the remainder of the CAC for their encouragement and support of this study. We would also like to thank Donna Doran, Rhonda O’Brien, MA, and the cardiac catheterization laboratories of the 33 participating hospitals for their tireless efforts to ensure the timeliness, completeness and accuracy of the registry data.
☆ Supported, in part, by the New York State Department of Health.
- acute myocardial infarction
- coronary artery bypass graft
- cardiac advisory committee
- coronary angioplasty reporting system
- congestive heart failure
- odds ratio
- percutaneous transluminal coronary angioplasty
- Received December 30, 1999.
- Revision received March 31, 2000.
- Accepted June 2, 2000.
- American College of Cardiology
- Grines C.L,
- Browne K.F,
- Marco J,
- et al.
- Tiefenbrunn A.J,
- Chandra N.C,
- French W.J,
- Gore J.M,
- Rogers W.J
- Garcia E,
- Elizaga J,
- Perez-Castellano N,
- et al.
- Mayo Coronary Care Unit and Catheterization Laboratory Groups,
- Gibbons R.J,
- Holmes D.R,
- Reeder G.S,
- Bailey K.R,
- Hopfenspirger M.R,
- Gersh B.J
- Ribeiro E.E,
- Silva L.A,
- Carneiro R,
- et al.
- Zilstra F,
- Beukema W.P,
- van’t Hof A.W.J,
- et al.
- Myocardial Infarction Triage and Intervention Investigators,
- Every N.R,
- Parsons L.S,
- Hlatky M,
- Martin J.S,
- Weaver W.D
- Stone G.W,
- Brodie B.R,
- Griffin J.J,
- et al.
- Michels K.B,
- Yusuf S