Author + information
- Received February 11, 1999
- Revision received June 18, 1999
- Accepted August 18, 1999
- Published online November 15, 1999.
- Mauricio G Cohen, MDa,∗,†,‡,§,* (, )
- Christopher B Granger, MDa,∗,†,‡,§,
- E.Magnus Ohman, MDa,∗,†,‡,§,
- Amanda L Stebbins, MSa,∗,†,‡,§,
- Liliana R Grinfeld, MD∗,
- Arturo M Cagide, MD∗,
- Marcelo V Elizari, MD†,
- Amadeo Betriu, MD‡,
- David F Kong, MDa,∗,†,‡,§,
- Eric J Topol, MD§ and
- Robert M Califf, MDa,∗,†,‡,§
- ↵*Reprint requests and correspondence: Dr. Mauricio G. Cohen, Box 3375, Duke University Medical Center, Durham, North Carolina 27710
We sought to describe the differences in the process of care and clinical outcomes between Hispanics and non-Hispanics receiving thrombolytic therapy for myocardial infarction (MI).
Hispanics are the fastest growing and second largest minority in the U.S. but most cardiovascular disease data on Hispanics has been derived from retrospective studies and vital statistics. Despite their higher cardiovascular risk-factor profile, better outcomes after MI have been reported in Hispanics.
We studied the baseline characteristics, resource use and outcomes of 734 Hispanics and 27,054 non-Hispanics treated for MI in the GUSTO-I and -III trials. The primary end point of both trials was 30-day mortality.
Hispanics were younger, shorter, lighter and more often diabetic and began thrombolysis 9 min later, compared with non-Hispanics. Measures of socioeconomic status (educational level, employment and health insurance) were lower among Hispanics. Fewer Hispanics than non-Hispanics underwent in-hospital angiography (70% vs. 74%, p = 0.013) or bypass surgery (11% vs. 13.5%, p = 0.04). Hispanics received more angiotensin-converting enzyme (ACE) inhibitors and less calcium-channel blockers, prophylactic lidocaine and inotropic agents. Mortality at 30 days and at one year did not differ significantly between Hispanics and non-Hispanics (6.4% vs. 6.7% and 9.0% vs. 9.7%, respectively). We noted no interactions between thrombolytic strategy and Hispanic status on major outcomes (30-day death, stroke and major bleeding).
The care of Hispanics with MI differed slightly from that of non-Hispanics. Nevertheless, these differences in care did not affect long-term outcomes.
The Hispanic population is the fastest-growing minority in the U.S. Hispanics are projected to become the largest minority by the year 2005, with an estimated population of over 31 million (about 12% of the total population) (1). However, the term Hispanic, used exclusively in the U.S., was coined without a scientific basis by the Office of Management and Budget in 1978, to standardize data collection and publication among federal agencies (2). Hispanics are a conglomerate of different cultures and nationalities; their origins are diverse, and they may be of any race.
Most data on cardiovascular disease prevalence, morbidity and mortality in Hispanics come from vital statistics (3–8), retrospective analysis (9), surveillance programs in areas with large Hispanic population (10), household sampling (11)and case-control studies (12). Investigations have shown a higher prevalence of cardiovascular risk factors in Hispanics compared with non-Hispanics, including diabetes (13–15), central obesity (13,16)and a less favorable lipid profile (6,9,17). In addition, socioeconomic factors such as poverty and limited access to health care may unfavorably influence Hispanics’ morbidity and mortality (18,19).
Two contradictory studies (20,21)have prospectively evaluated Hispanics with myocardial infarction (MI). Taylor et al. (20)reported a trend toward better outcomes for Hispanics compared with blacks and non-Hispanic whites, even though the known coronary risk factors were more prevalent in Hispanics. On the other hand, Canto et al. (21)found no significant in-hospital mortality differences between Hispanics and non-Hispanic whites.
The large databases of the Global Utilization of Streptokinase and TPA (alteplase) for Occluded Coronary Arteries (GUSTO-I) and Global Use of Strategies to Open Occluded Coronary Arteries (GUSTO-III) trials allowed us to maximize the power to detect differences between Hispanics and non-Hispanics and to observe outcomes over a seven-year period with all currently-used thrombolytic agents (streptokinase, alteplase and reteplase). We also determined the resource use and process of care for these two groups.
The GUSTO-I and GUSTO-III trials have been described in detail (22,23). These studies enrolled a total of 27,788 patients in the U.S. Briefly, the entry criteria included patients with <6 h of typical symptoms of MI and ST-segment elevation. Patients were randomly allocated to different thrombolytic strategies according to the trial protocols. GUSTO-I randomized patients to one of four thrombolytic regimens: streptokinase with subcutaneous heparin, streptokinase with intravenous heparin, accelerated alteplase with intravenous heparin, or a combination of streptokinase plus alteplase with intravenous heparin. GUSTO-III compared reteplase with alteplase, with adjunctive aspirin and intravenous heparin in both arms. Informed consent was obtained from all patients, and the protocols were approved by the institutional review board at each hospital. The primary end point of both trials was all-cause, 30-day mortality. One-year mortality was a secondary end point.
Information on the study patients was prospectively collected with case-report forms that captured demographic and clinical characteristics at enrollment and during hospitalization up to discharge. Data on patients who were transferred to another hospital were also collected until discharge from that hospital. Deaths within 30 days and one year after enrollment were ascertained by postcards returned by the patients’ families, through telephone contact or via registered mail.
Data on the availability of cardiac catheterization facilities were obtained from the site-descriptor database for the 366 hospitals that participated in both trials.
The U.S. Census Bureau considers a Hispanic to be any person who classifies himself or herself in one of these origin categories: Mexican, Puerto Rican, Cuban or other Spanish/Hispanic origin (24). Hispanics may be of any race. In this study, Hispanic ethnic origin was determined from the initial patient history and exam by study investigators, as recorded in source documents. Only patients enrolled in the U.S. were considered for this analysis, because interpretations of Hispanic ethnicity may not be uniform worldwide, and nations may vary in their use of resources to treat MI (25,26).
Continuous variables are presented as mean ± SD, and medians are given with 25th and 75th percentiles. Discrete variables are expressed as frequencies with percentages. Differences in continuous variables between groups were tested using the Wilcoxon rank-sum test. Differences in discrete variables were examined with chi-square tests. All significance tests were two-tailed. Results were interpreted as statistically significant when p < 0.05. Multivariable logistic regression analysis was used to assess any independent association between Hispanic ethnicity and survival, based on a multivariable survival model from the GUSTO-I population (27). Logistic modeling also was used to determine any interaction between thrombolytic strategy and Hispanic ethnicity. Mortality differences at one year were assessed using the log-rank test.
As these analyses are multiple, retrospective comparisons, they should be considered hypothesis-generating rather than definitive.
Of the 27,788 U.S. patients enrolled in GUSTO-I and -III, 734 (2.64%) were Hispanic. The geographic distribution of the Hispanic cohort paralleled the national Hispanic distribution reported by the U.S. Census Bureau (24). High proportions of Hispanics (86.1%) were enrolled in four regions: 32.8% in the Southwest, 21.5% in the Northeast, 16.8% in the West and 15% in the South Atlantic (Table 1). Compared with non-Hispanics, Hispanics were more often treated at hospitals without cardiac catheterization laboratories (9.3% vs. 13.5%, p = 0.09).
Baseline characteristics of the Hispanic populations were similar between trials. For hypercholesterolemia, systolic blood pressure, history of smoking and time from symptom onset to treatment, clinically small differences were statistically significant (Table 2). Overall, Hispanics tended to be younger, shorter and weigh less than non-Hispanics. Hispanics were more likely to be diabetic and less likely to have known elevated cholesterol. More Hispanics presented with an anterior MI than non-Hispanics (43.3% vs. 38.4%, p = 0.007). Other baseline characteristics did not differ substantially.
Socioeconomic information was collected in the GUSTO-III trial (Table 3). Fewer Hispanics finished high school (46.9% vs. 72.7%, p = 0.001) or college (15.5% vs. 28.6%, p = 0.005). Hispanics were less likely to have private insurance (43.4% vs. 62.6%, p = 0.001) and were more likely to be uninsured (18% vs. 6.5%, p = 0.001) than non-Hispanics. Hispanics were more often employed as laborers and less often in management positions. Finally, Hispanics were less likely to live alone (9.8% vs. 19.3%, p = 0.008).
Time to treatment
Both time from symptom onset to presentation and time from presentation to treatment were longer for Hispanics versus non-Hispanics (Table 4). Hispanics arrived at the hospital a median 5 min later than non-Hispanics after symptom onset. Once in the hospital, the time to treatment was 4 min longer for Hispanics than for non-Hispanics (p = 0.019). Of note was that 25% of Hispanics had a time to treatment that was 18 min longer than that of non-Hispanics.
Use of cardiac medications
The use of cardiac medications differed slightly between Hispanics and non-Hispanics (Table 5). During hospitalization, Hispanics patients were more likely to receive angiotensin-converting enzyme (ACE) inhibitors and less likely to receive prophylactic lidocaine, positive inotropic agents and calcium-channel blockers than non-Hispanics. Discharge prescriptions followed a similar pattern, with Hispanics receiving more ACE inhibitors and nitrates and less digitalis and calcium-channel blockers than non-Hispanics.
Hispanic patients received less invasive treatment (Table 6). Significantly fewer Hispanics underwent angiography during hospitalization (69% vs. 73%, p = 0.013), even after excluding patients enrolled in the GUSTO-I Angiographic Substudy (68.7% vs. 72.9%, p = 0.013) (28). The extent of coronary disease differed significantly between Hispanics and non-Hispanics, however. Hispanics less often had two-vessel disease compared with non-Hispanics, but the presence of one- or three-vessel disease was similar. There were no significant differences in left ventricular ejection fraction between groups. Despite the difference in the use of angiography, Hispanics and non-Hispanics were equally likely to undergo subsequent angioplasty (30.3% vs. 29.3%, p = 0.69). Fewer Hispanics, however, underwent bypass surgery (11% vs. 13.5%, p = 0.04).
The duration of critical care unit stay (median 3.5 days) and hospitalization (median 8 days) was identical for both groups. The use of other cardiac procedures in the acute coronary unit differed between Hispanics and non-Hispanics. Fewer Hispanics had a temporary transvenous pacemaker inserted (6% vs. 9.2%, p = 0.003), a pulmonary-artery catheter placed (14.1% vs. 17.7%, p = 0.01) or mechanical ventilation (11.6% vs. 15.4%, p = 0.005), probably reflecting their lower rate of bypass surgery. The use of intraaortic balloon pumps was similar in both groups (4.9% vs. 6.2%, p = 0.2).
Major clinical outcomes
Unadjusted mortality at 30 days did not differ significantly between groups (Table 7). Adding Hispanic ethnicity to the GUSTO-I mortality model did not increase the model’s ability to predict death at 30 days. Therefore, even after statistical adjustments for age, gender, pulse, blood pressure and Killip class, the difference in 30-day mortality was not significant between groups (odds ratio 1.24; 95% confidence interval, 0.89 to 1.73, p = 0.21). Unadjusted 1-yr mortality did not differ significantly between Hispanics and non-Hispanics (9.03% vs. 9.74%, p = 0.51).
The rates of stroke (hemorrhagic and ischemic) were similar for both groups, as were the rates of moderate or severe bleeding (Table 7). The median activated partial thromboplastin time (aPTT) at 12 h was significantly higher in Hispanics (73.8 vs. 66.7 s, p ≤ 0.01) but still within the 50- to 75-s target therapeutic range specified by the protocols in both groups. Other clinical outcomes, including recurrent ischemia, congestive heart failure and cardiogenic shock were similar among Hispanics and non-Hispanics (Table 7).
No interactions were found between thrombolytic strategy and Hispanic ethnicity on major adverse outcomes including 30-day death, stroke, cardiogenic shock and severe bleeding. However, the rates of these outcomes tended to be higher for patients treated with streptokinase compared with those treated with alteplase or reteplase (Table 8).
This study explored the outcomes of 734 Hispanic patients treated with thrombolytic therapy for acute MI. The findings suggest differences in both the process of care and resource use for the fastest-growing minority in the U.S. Despite these differences, however, 30-day and one-year outcomes were similar to those of other patients treated in the two GUSTO trials.
Hispanics showed slightly lower in-hospital, 30-day and one-year unadjusted mortality rates, probably reflecting their younger age, the strongest predictor of survival. The mortality differences between Hispanic and non-Hispanics were not significant. The differences in the process of care of Hispanics during hospitalization did not affect long-term mortality. After adjustment for baseline predictors of 30-day mortality in the GUSTO population (27), there was a trend towards increased mortality among Hispanics. Of note, the effect of the thrombolytic strategies on outcomes followed the same patterns seen in the parent GUSTO-I and GUSTO-III studies.
Results of previous studies are contradictory regarding mortality after MI in Hispanics. A prospective study of outcomes after MI in 147 Hispanics, 2,564 whites and 174 blacks reported a trend towards decreased mortality among Hispanics (20). An analysis of the second National Registry of Myocardial Infarction found no significant differences in in-hospital mortality between non-Hispanic whites and the largest cohort of Hispanics with MI (21). In an age-specific analysis, Hispanics less than 40 years old had a mortality risk four times higher than non-Hispanic whites. These results must be interpreted cautiously, because patients under 40 years old represented only 4% of that study’s population. Goff et al. (10)reported increased mortality after MI among Mexican-Americans compared with non-Hispanic whites. This study was confounded by a retrospective design and the use of ICD-9 codes to classify MI discharge diagnoses as possible or definite.
Other reports based on vital statistics have shown lower heart-disease mortality among Hispanics (3–7,9). The limitations of vital statistics include misclassification of Hispanic ethnicity and underestimation of Hispanic deaths in national death registries (28). Definitions of Hispanic ethnicity are not uniform across studies; they can be based on patient preference, death certificates, medical records, Hispanic surnames, algorithms or responses to national census questions. These methods are less reliable than those of randomized trials, where all patients are enrolled based on prospectively defined inclusion criteria.
Previous investigations have compared the outcomes of Hispanics versus non-Hispanic whites and blacks, categorizing Hispanics as a racial group. In this study, we differentiated by ethnicity rather than race; as a result, we compared the outcomes of Hispanics treated for MI with the rest of the study patients (non-Hispanics). For this reason, we purposely did not identify other racial minorities in the comparisons.
The high prevalence of diabetes among Hispanics compared with non-Hispanics in our study agrees with previous observational studies (13–15). However, other risk factors, including known hypercholesterolemia, hypertension and cigarette smoking did not follow observed patterns (13,17,29). This may reflect a lower ascertainment of hypertension and hypercholesterolemia in Hispanics, who are less likely to regularly visit a physician (30). Interestingly, Hispanics were more likely to present with anterior infarction than non-Hispanics. This has been observed in the Thrombolysis In Myocardial Infarction (TIMI)-II study (20)and the second National Registry of Myocardial Infarction (21).
Time to treatment
Time from symptom onset to thrombolytic therapy was longer in Hispanics than in non-Hispanics. This difference was mainly at the expense of a longer presentation delay, which reflects patient-related delays. Several factors, including cultural differences in the perceptions of disease, less comprehensive health-insurance coverage, access to care and educational level (18,31)may have influenced the decision to seek medical care among Hispanics. Clark et al. (32)studied inner-city patients with suspected MI and showed longer prehospital delays for Hispanics than non-Hispanic whites. Most of these delays were attributed to the time needed for patients to decide to seek medical care after symptom onset (32). One study has shown that Hispanics tend to underuse emergency medical systems (21). The GUSTO studies did not ascertain use of ambulance services. The influence of household members in the delay in seeking medical attention remains unclear. Alonzo et al. (33)reported increased prehospital delays when the decision to call was given to a family member. We found that Hispanics were less likely than non-Hispanics to live alone.
The in-hospital component of treatment delay, the “door-to-needle” time, was also longer among Hispanics. Language barriers and interpretation of the informed consent for enrollment in a clinical trial may have contributed to these delays. The differences in treatment delays in GUSTO-III compared with GUSTO-I do show that changes in the health-care system over the last five years to reduce delays in patients with MI also have included the Hispanic minority.
Angiography and revascularization
Although a high proportion of Hispanics (70%) underwent angiography, this rate was 4% lower than for non-Hispanics (74%). Younger age and recurrent ischemia are strong predictors of the use of angiography (34). We expected that Hispanics would undergo angiography at least as often as non-Hispanics because of their younger age and similar incidence of post-MI ischemic complications. On the other hand, Hispanics tended to seek medical care in hospitals without cardiac catheterization facilities, which are a negative predictor for the use of angiography in post-MI patients (34). Studies have reported similar findings on the use of invasive cardiac procedures (35). In a report from the Corpus Christi Heart Project, Mexican-Americans admitted for MI were catheterized 7% less often than non-Hispanic whites. However, the 52% rate of angiography in that investigation was substantially lower than it was in this study (36). Our findings contrast with those of Canto et al. (21), who showed that Hispanics more often went to hospitals with catheterization facilities and were as likely as non-Hispanic whites to have invasive procedures.
The extent of coronary disease, an independent predictor for bypass surgery, was greater in non-Hispanics. Non-Hispanics had a higher prevalence of two-vessel disease than Hispanics, whereas the rate of three-vessel disease was similar in both groups. This difference may have contributed to the greater use of surgical revascularization in non-Hispanics, but it is three-vessel (and not two-vessel) disease that has been correlated with the use of bypass surgery (34). The rates of single-vessel disease were similar and may explain the similar use of angioplasty among the two groups. Other factors that may have influenced the lower use of angiography and bypass surgery in Hispanics include less comprehensive insurance coverage (37), cultural perceptions of invasive procedures or major surgery and educational level. Schecter et al. (38)reported that coronary care unit patients with lower educational levels were significantly more likely to disagree with a physician’s recommendation to undergo angiography.
The use of cardiac medications during admission and at discharge also differed between Hispanics and non-Hispanics. Hispanics were more likely to receive ACE inhibitors, which are known to reduce mortality after MI. A higher use of ACE inhibitors among Hispanics also was observed in a Texas surveillance study (39). Perhaps the greater prevalence of diabetes among Hispanics led physicians to use ACE inhibitors more often. Non-Hispanics were more commonly prescribed drugs shown to have no benefit or even a detrimental effect, such as prophylactic lidocaine, calcium-channel blockers and inotropic agents. These results contrast with other investigations that have shown lower beta-adrenergic blocking agent use among Hispanics (21,39).
We measured socioeconomic status by educational level and occupation in this investigation. The association between these two variables and all-cause mortality has been validated in the literature. In general, mortality decreases as educational level increases. The same pattern also has been shown for different occupational categories, with laborers having the highest mortality and professionals, the lowest. The same association with socioeconomic status has been seen for cardiovascular mortality (40). Our results indicated a lower socioeconomic status for Hispanics than for non-Hispanics, but we did not detect a significant association with mortality. Other factors that favorably influence outcomes (such as diet, greater social support and lifestyle) may interact with indicators of socioeconomic status. Farmer et al. (41)reported an association between the greater social support observed among Hispanics and greater survival after MI.
This analysis is a multiple, retrospective comparison and is subject to the limitations of this type of study. In addition, subcategories of Hispanics based on national origin, such as Mexican, Puerto Rican, Central and South American or Cuban were, unfortunately, not collected by the case report form. The GUSTO trials considered patients presenting with a diagnosis of MI who were candidates for thrombolytic therapy. This may account for the lower proportion of Hispanics (2.64%) enrolled in the trial compared with the general population (10.65%). This lower proportion of Hispanics may reflect: 1) underreported Hispanic ethnicity, 2) nonuniform enrollment across geographic regions and urban centers, with higher enrollment in areas with few Hispanic residents, 3) selective nonenrollment of Hispanics at individual centers, 4) reduced access to health care, or 5) a lower incidence of MI due to the lower median age of Hispanics, which is unlikely (42). Finally, the categorization of Hispanic ethnicity did not follow a specific algorithm, leaving open the possibility some truly Hispanic patients were not categorized as such. Out-of-hospital deaths were not captured by the GUSTO trials.
This study provides detailed information on short- and long-term mortality and process of care of Hispanics hospitalized for MI. After adjusting for differences in baseline characteristics, we noted a trend towards increased mortality in Hispanics compared with non-Hispanics. At one-year follow-up, the difference in mortality was not statistically significant. Different aspects of the process of care, such as time to thrombolysis and resource use, show slight but statistically significant differences between Hispanics and non-Hispanics. Hispanics had a higher prevalence of diabetes and a lower socioeconomic status. We observed longer delays in time to hospital arrival and time to treatment in Hispanics compared with non-Hispanics. Fewer Hispanics underwent angiography and bypass surgery, which could reflect lesser insurance coverage and less severe coronary disease, respectively. Nevertheless, these differences in care did not affect long-term outcomes.
Public education campaigns directed to Hispanics should be considered, to reduce patient-related delays in treatment. To identify possible biases, studies should further explore access to and processes of health care for Hispanic subgroups.
☆ This study was supported by grants from Bayer, New York, New York, CIBA-Corning, Medfield, Massachusetts, Genentech, South San Francisco, California, ICI Pharmaceuticals, Wilmington, Delaware, Sanofi Pharmaceuticals, Paris, France and Boehringer-Mannheim Therapeutics, Mannheim, Germany and Gaithersburg, Maryland.
- angiotensin-converting enzyme
- activated partial thromboplastin time
- Global Utilization of Streptokinase and TPA (alteplase) for Occluded Coronary Arteries
- Global Use of Strategies To Open Occluded Coronary Arteries
- myocardial infarction
- Received February 11, 1999.
- Revision received June 18, 1999.
- Accepted August 18, 1999.
- American College of Cardiology
- ↵Day J. Population projections of the United States by age, sex, race and Hispanic origin: 1995 to 2050. Washington, DC: US Government Printing Office, US Bureau of the Census, Current Population Reports, 1996:P25–1130.
- Liao Y,
- Cooper R.S,
- Cao G,
- Kaufman J.S,
- Long A.E,
- McGee D.L
- Becker T.M,
- Wiggins C,
- Key C.R,
- Samet J.M
- Stern M.P,
- Bradshaw B.S,
- Eifler C.W,
- Fong D.S,
- Hazuda H.P,
- Rosenthal M
- Goff D.C Jr.,
- Ramsey D.J,
- Labarthe D.R,
- Nichaman M.Z
- Mitchell B.D,
- Hazuda H.P,
- Haffner S.M,
- Patterson J.K,
- Stern M.P
- Rewers M,
- Shetterly S.M,
- Baxter J,
- Marshall J.A,
- Hamman R.F
- Burchfiel C.M,
- Hamman R.F,
- Marshall J.A,
- Baxter J,
- Kahn L.B,
- Amirani J.J
- Orlander P.R,
- Goff D.C,
- Morrissey M,
- et al.
- Wei M,
- Mitchell B.D,
- Haffner S.M,
- Stern M.P
- Mitchell B.D,
- Stern M.P,
- Haffner S.M,
- Hazuda H.P,
- Patterson J.K
- Taylor H.A,
- Chaitman B.R,
- Rogers W.J,
- et al.
- GUSTO Investigators
- Global Use of Strategies to Open Occluded Coronary Arteries (GUSTO-III) Investigators
- ↵US Bureau of the Census. Statistical Abstract of the United States: 1997 (117th edition). Washington (DC) 1997.
- Lee K,
- Woodlief L,
- Topol E,
- et al.
- Kaplan G.A,
- Keil J.E
- Goff D.C,
- Nichaman M.Z,
- Chan W,
- Ramsey D.J,
- Labarthe D.R,
- Ortiz C