INTRODUCTION

Primary angioplasty has shown better clinical outcomes than fibrinolysis in patients with ST-segment elevation acute myocardial infarction (STEAMI) when specific conditions are met.1,2 Numerous clinical trials show its benefits, both when performed in the hospital where the patient presents with STEAMI1 and when undertaken after transfer to another center.3-7 Despite these results and current clinical practice guideline recommendations,8 fibrinolysis continues to be the most widely-used reperfusion strategy. In Spain, with >40 000 infarctions per year,9 only 5101 primary angioplasties were performed in 2005.10

The most important limitation of primary angioplasty is the need for an adequate infrastructure. Few Spanish hospitals have a catheterization laboratory and 43% of those that do lack a 24-hour emergency team.10 Clinical trials have demonstrated the benefits of primary angioplasty following interhospital transfer but results in clinical practice differ from those of trials. The unavoidable delay of transfer and concomitant risks make it a controversial strategy when compared to on-site, out-of-hospital fibrinolysis or pharmacologic facilitation and deferred angioplasty.1-14 Some authors suggest the clinical benefit of primary angioplasty over fibrinolysis is only a function of the delay.15

Conducting adequately structured and controlled programs with a follow-up of delays and rationalization of resources could reproduce, in clinical practice, the clinical trial results for primary angioplasty and extend its benefits to a greater number of patients than those currently receiving them.16,17 In Spain, data on the results of primary angioplasty, whether performed at the admitting hospital or following interhospital transfer, are currently limited.18

In the present study, we describe and compare the short- and long-term results of a systematic, primary angioplasty program in patients with STEAMI in 2 hospitals, one with a catheterization laboratory and the other needing interhospital transfer for patients to undergo catheterization.

METHODS

Design

A prospective, observational study of 2 consecutive cohorts of patients admitted to 2 hospitals, one with and one without a catheterization laboratory.

Patients

We enrolled patients from 2 hospitals (hospital 1, with a catheterization laboratory, 881 beds; and hospital 2, 10 km from hospital 1, without a catheterization laboratory, 444 beds), consecutively indicated for primary angioplasty from January 1, 2000 thru April 30, 2001.

Since 2000, primary angioplasty has been the treatment of choice for STEAMI in these 2 centers. In both hospitals, primary angioplasty is considered to be indicated when 2 criteria are simultaneously fulfilled: a) symptoms compatible with myocardial ischemia lasting >30 min despite administration of antianginal treatment, and b) electrocardiographic ST-segment elevation ≥0.1 mV in at least 2 contiguous leads, lasting >30 min (or presence of new, complete left bundle-branch block), within 12 hours after onset of symptoms or later if symptoms of myocardial ischemia persist.

During the period included in this study, exceptionally, primary angioplasty was not performed in some of the following situations: patient refusal to undergo catheterization; prior knowledge of highly unfavorable coronary anatomy; absence of percutaneous arterial access; involvement of the interventional cardiology team in another procedure that could not be postponed; explicit decision of the physician responsible; lack of an available mobile intensive care unit (ICU) to transfer the patient to hospital 2.

Inclusion and Exclusion Criteria in This Study

Inclusion Criteria

We followed the principle of intention to treat and excluded all patients for whom the interventional cardiologist responsible for performing primary angioplasty had been contacted.

Criteria of Exclusion

1. Prior administration of fibrinolytic agents to treat current infarction.

2. Patients previously included in the study.

3. Patients indicated for primary angioplasty proceeding from a hospital other than the 2 previously-mentioned centers.

Strategy of Applying for Urgent Catheterization

In hospital 1, patients with suspected STEAMI were immediately attended by the duty cardiologist. After clinical examination and initial electrocardiogram, the duty cardiologist telephoned the interventional cardiologists, present in the catheterization laboratory during working hours (08.00-15.00), or the emergency team on call outside of working hours. The team had 6 interventional cardiologists available for this procedure.

In hospital 2, the duty intensive care specialist indicating the procedure established contact with the duty cardiologist of hospital 1 and the latter contacted the catheterization team. When the transfer took place outside of working hours, patient transfer to the catheterization laboratory in a mobile ICU was simultaneous with the duty team's journey to the hospital.

To avoid delays, the program recommended patient transfer from the place of diagnosis (generally, the emergency room entrance) to the catheterization laboratory.

The procedural technique and later treatment remained at the discretion of the interventional cardiologist responsible in each case, in line with clinical practice guidelines current at the time.

Variables

At admission, clinical variables were taken from the clinical record and the catheterization laboratory register by an assistant cardiologist and a trainee cardiologist. We conducted clinical follow-up of patients during hospitalization and later at 1, 6, and 12 months post-AMI, through medical visits or telephone contact. We defined 3 time intervals:

1. "Door-to-angiography" time: from the arrival of the patient at the first hospital to the start of the angiography.

2. "Door-to-artery open" time: from the arrival of the patient at the hospital to the opening of the artery.

3. "Onset of symptoms-to-angiography" time: from the onset of symptoms to the start of the angiography.

Cardiogenic shock was defined as maintained arterial hypotension (invasive systolic arterial pressure <90 mm Hg during at least 30 min despite administration of fluids), accompanied by signs of tissue hypoperfusion secondary to cardiac dysfunction. We considered the procedure successful when it ended with residual stenosis ≤20% and TIMI (thrombolysis in myocardial infarction) flow 3 in the artery causing the infarction, without more severe complications in the catheterization laboratory (death, worsening of initial Killip class, or stroke).

Statistical Analysis

Qualiative variables are expressed as number of cases and percentage and compared using χ2 or the Fisher exact test when criteria for applying the former were not met.

Quantitative variables are expressed as mean (SD) or median and 25% and 75% percentiles if distributions were asymmetric. Comparison of these variables was by Student t test for independent means or the nonparametric Mann-Whitney U test for variables with non-normal distribution.

Analysis of factors determining in-hospital mortality present at admission or after performing primary angioplasty was by logistic regression. Follow-up was with Kaplan-Meier survival analysis, and survival curves were compared with the log rank test. We used the Cox proportional risk logistic regression method to determine the factors present at admission or after performing primary angioplasty that associated independently with long-term mortality. In the 2 multivariable analyses, the selection of variables in the final model was conducted after exploring the different models obtained with the forward stepwise method of sequential inclusion and the backward stepwise exclusion of those variables that had shown a different distribution between the hospitals or that, in the literature, associate with a different prognosis after primary angioplasty. In both final models we confirmed their assumed applicability (residual analysis and fulfillment of criteria of proportionality, and log-linear, respectively). In all cases we considered P<.05 significant. Statistical analysis was with SPSS® 12.02 and STATA® 8.2.

RESULTS

From January 1, 2000 thru April 30, 2001, in hospital 1 or hospital 2, 245 patients who were diagnosed with STEAMI met reperfusion criteria. Hospital 1 admitted 172 patients, indicating primary angioplasty in 158 (92% of patients admitted had STEAMI and met criteria to undergo reperfusion treatment). The remaining 14 patients were not indicated for primary angioplasty because: a) the interventional cardiology team were involved with another infarction patient (2 patients); b) the patient refused treatment (1); c) highly unfavorable coronary anatomy was known (2); d) the physician responsible decided against the procedure (2); and e) the patient presented a very poor baseline situation prior to infarction (7). Six of these 14 patients received fibrinolysis (3.5%).

Hospital 2 admitted 73 patients with STEAMI who met reperfusion criteria. Primary angioplasty was indicated in 64 (88%). Non-indication for primary angioplasty in the remaining 9 patients was because: a) a mobile ICU was not available for the transfer (2 patients); b) the interventional cardiology team were involved with another infarction patient with (1); and c) the physician responsible decided against the procedure (6). Four of these 9 patients received fibrinolysis.

The final sample consisted of 222 patients which represents 90.6% of admissions with STEAMI and indication for reperfusion treatment in one or the other hospital.

The baseline characteristics of patients appear in Table 1. Patients from hospital 2 showed a more favorable cardiovascular profile, with less frequency of dyslipidemia, prior AMI and prior coronary revascularization. They tended to be younger, with lower incidence of diabetes and other known, modifiable cardiovascular risk factors.

We found no differences between hospitals in characteristics of presentation of the infarction (Table 2).

In 79.9% of patients, onset of symptoms-to-angiography time was <6 h (Figure 1). We only found significant differences between the hospitals in the intervals that included interhospital transfer time (Table 3).

Figure 1. Distribution of patients as a function of onset of symptoms-to-angiography time.

Two patients at hospital 2 died prior to starting the diagnostic procedure leaving 220 who underwent diagnostic coronary angiography. Of these, 192 finally underwent coronary intervention. The reasons why primary angioplasty was not performed in the remaining 28 patients appear in Figure 2. Table 4 summarizes baseline angiographic characteristics and interventional procedures.

Figure 2. Patients undergoing coronary angiography and primary angioplasty. Reasons for not performing the procedure.

In-Hospital Evolution

Table 5 shows in-hospital results. Estimated incidence of more severe complications was 24.3%, fundamentally due to the presence of cardiogenic shock. Ventricular function at discharge, quantified through echocardiography in the first days after the acute episode, was evaluated in 192 patients and proved to be >40% in >75% of patients. In the remaining 14.1% ventricular function is unknown, fundamentally due to early death before echocardiography. Overall in-hospital mortality was 12.2%; 5.4% of patients died in the catheterization laboratory and the remaining 6.8% in-hospital. In 85.2% of deaths, the cause was cardiac, principally cardiogenic shock (70.4%). In the 14.8% of non-cardiac deaths, the cause was multiorgan failure once the shock situation had been overcome, digestive hemorrhage and ischemic stroke (2 patients).

In the final logistic regression model, factors independently associated with greater in-hospital mortality were advanced age, female gender, presence of cardiogenic shock at admission, and procedure failure (Table 6). The hospital of origin did not associate significantly with greater mortality (odds ratio [OR],1.42; 95% confidence interval [CI], 0.3-7.8).

Follow-Up Phase

A total of 195 patients survived the in-hospital phase. We obtained a complete follow-up at 1 month and 1 year of 97.4% and 96.4%, respectively, with a mean follow-up time of 362.2 days (95% CI, 356.8-369.5 days). During follow-up, 7 patients died (5 of cardiac cause), and accumulated mortality was 3.1% (3.7% of patients discharged alive). Total mortality at 1-year follow-up, including the in-hospital phase, was 15.3%.

We found no significant differences in mortality or in frequency of other adverse cardiovascular events during follow-up (Table 7). The log rank test was nonsignificant when comparing the 2 hospitals for mortality, both on including (P=.49) and excluding (P=.39) the hospitalization phase (Figure 3).

Figure 3. Follow-up survival curves (days) for total mortality, including (upper) and excluding (lower) the in-hospital phase. HR indicates hazard ratio; CI, confidence interval.

In the Cox model study, of the variables present at admission or following performance of primary angioplasty and influencing long-term survival, only advanced age, presence of diabetes, shock at admission and procedure failure associated independently and significantly with greater mortality at 1-year follow-up (Table 8). The hospital of origin did not associate significantly with greater mortality in 1-year follow-up (relative risk [RR],2.04; 95% CI, 0.74-5.61).

DISCUSSION

The present study indicates results of systematic primary angioplasty in STEAMI in a hospital without a catheterization laboratory can be similar to those in a hospital with a laboratory despite the need to transfer patients, if the transfer is conducted so as to minimize delays and guarantee safety.

Five clinical trials3-7 and 1 metaanalysis19 have shown that transfer of patients with STEAMI from hospitals without a catheterization laboratory for them to undergo primary angioplasty is feasible, safe, and offers greater clinical benefits than on-site fibrinolysis. Despite these results, primary angioplasty remains an infrequently used treatment for STEAMI, probably because it is difficult to reproduce these results in clinical practice. The motivation to participate in a clinical trial, geographic, and economic characteristics of countries, infrastructure, stimulation, and resources for the transfer and treatment of patients favored results in these studies-results that may be considered ideal. Despite distances of >100 km in the 2 principal studies (PRAGUE II5 and DANAMI 27), mean delays were optimal and complications during transfers, practically non-existent. In contrast, studies evaluating results of real-world primary angioplasty in centers without a catheterization laboratory have reported very different results and conclude that the recommendation to guarantee the potential benefits of primary angioplasty (arrival at the first hospital to start of revascularization time <90 min)8 was seldom achieved.20

In our series, with non-selected patients in optimal condition for interhospital transfer, the median delay times observed in patients from hospital 2 were <65 min for admission-to-angiography, 85 min for admission-to-artery open, and <190 min from onset of symptoms-to-angiography. These are all lower than the times reported in clinical trials3-7 and far below those in registers.20-22

When comparing the 2 hospitals in our study, we found differences in door-to-angiography, door-to-artery open, and symptoms-to-angiography times, of only 13, 15, and 25 min, respectively. Such short delays, together with the few events associated with transport probably justify the absence of differences in the success of the procedure, infarction size, or the rate of clinical events.

Two factors may have conditioned these reduced delays: a) the design of an infrastructure to favor interhospital transfer, including a mobile ICU almost always immediately available; an experienced interventional cardiology team, on call 24 hours a day, mobilized immediately on establishing the indication in hospital 2; and a system of direct transfer of patients to the catheterization laboratory from the service where the diagnosis was made avoiding the delay produced by the intermediate stages in the diagnosis and decision-making chain23; and b) the distance of just 10 km separating hospital 2 from the reference hospital making interhospital transfer time similar to that needed for the interventional cardiology team to arrive at the hospital of reference, meaning delays are practically identical in both hospitals.

The risk profile of patients included in this study is typical of STEAMI in daily clinical practice. In DANAMI 2,7 patients with diabetes represented 7.5% versus 22% in our series; 8.1% had a history of myocardial infarction versus 14%; and, especially, patients with cardiogenic shock represented 0% versus 12.5%. The selection bias of clinical trials may explain how, despite shorter delays and greater availability of technical advances in the treatment of STEAMI, in-hospital mortality in our series is twice that of these studies (6%-8% vs 12.2%) and highlights the need to know the results of applying scientific tests to the specific context in which clinical activity is undertaken. Despite the fact that 12.5% of patients transferred were in cardiogenic shock, interhospital transport was safe, with only 1 case of severe complications (1.5%), consisting of primary ventricular fibrillation, electrically cardioverted efficiently. On concluding the procedure, 80% of patients from hospital 2 who underwent a coronary intervention returned to the ICU of their hospital without incidents during the transfer.

After adjustment for the remaining variables, hospital 2 did not associate significantly with either in-hospital or mid-term mortality. In hospital 1, we even found a higher (but non-significant) rate of mortality, both absolute and of risk indicators that, given the sample size, may be due to chance or to the fact we are dealing with a reference hospital where the population attending may present a worse clinical profile than that attending hospital 2.

In Spain, relatively few hospitals have a catheterization laboratory and even the centers that do are not equipped with the resources needed to offer the procedure to patients with STEAMI on a 24-hour basis.10 Studies such as ours encourage both transfer from centers without central catheterization laboratories and the development of plans for various centers with this facility but unable to offer it continuously, to implement a transfer system to rationalize efforts and resources, with shift-work systems in laboratories or other possible approaches, in order to offer each patient the best treatment possible.

Limitations

The inclusion of a much greater number of patients may have produced significant differences in some of the variables analyzed, although the minimal differences found minimize the clinical importance of these possible differences. The tendency towards greater mortality found in hospital 1 indicates a larger sample size would be unlikely to invert the direction of the prognostic association, thus maintaining the idea that primary angioplasty in centers without catheterization laboratories is safe when certain conditions can be met. The conclusions of the present study are only applicable to centers close to a catheterization laboratory with strictly organized programs and sufficient resources similar to the ideal situation of the study. These data do not enable us to generalize that results obtainable in daily clinical practice in more distant centers might be superimposed on those of hospitals with a catheterization laboratory. However, the conclusions could be extended to many Spanish urban hospitals located in areas near centers able to perform primary angioplasty and which could do so if they made the necessary structural modifications.

CONCLUSIONS

In order to perform primary angioplasty, interhospital transfer from a center without a catheterization laboratory, to a nearby center, with adequate transport available and within a highly structured program, is feasible and meets the time requirements for the intervention established by the scientific societies. Under these conditions, in-hospital mortality and 1-year follow-up of patients undergoing transfer presented no differences by comparison with patients admitted to a hospital with a catheterization laboratory. The transfer of patients from a nearby hospital to the catheterization laboratory did not associate with greater mortality in any phase of follow-up.

ABBREVIATIONS
ICU: intensive care unit
STEAMI: ST-segment elevation acute myocardial infarction

See editorial on pages 791-3

Correspondence: Dra. P. Carrillo Sáez.
Antonio Machado, 2, 3 C. 30009 Murcia. España.
E-mail: pcarrillo@ya.com

Received October 16, 2006.
Accepted for publication April 11, 2007.