Primary percutaneous coronary intervention (PCI) is the optimal method of reperfusion when performed expeditiously. Factors contributing to delays in PCI for ST-segment elevation myocardial infarction (STEMI) have not been thoroughly characterized or quantified. We sought to identify the factors associated with the delays to reperfusion in patients with STEMI undergoing primary PCI. Primary PCI was performed in 3,340 patients with STEMI in the international, multicenter Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction trial. Multivariate analysis was used to identify independent predictors of delay in achieving reperfusion from 38 baseline and procedural variables. A total of 905 patients (27.1%) presented to non-PCI hospitals and were subsequently transferred; the remainder presented to PCI hospitals. The most powerful independent predictor of the interval from symptom onset to arrival at the PCI hospital and the first door-to-balloon time was an initial presentation at a non-PCI hospital (median incremental 58- and 54-minute delay, respectively, both p <0.001). Other independent predictors of prolonged door-to-balloon times included presentation with respiratory failure (42-minute incremental delay, p = 0.003), presentation during off-work hours (11-minute incremental delay, p <0.001), and co-morbid conditions such as diabetes and heart failure. In conclusion, among patients undergoing primary PCI, presentation to a non-PCI hospital was the variable associated with the greatest delay to reperfusion. Systems of care that encourage ambulance diagnosis and direct delivery of patients with STEMI to a PCI hospital might shorten the overall door-to-balloon times and improve the clinical outcomes.
Compared to fibrinolytic therapy, percutaneous coronary intervention (PCI) results in superior clinical outcomes in patients with ST-segment elevation myocardial infarction (STEMI) when both reperfusion modalities are available at the presenting hospital. When patients with STEMI present to a non-PCI hospital, prompt transfer to a PCI hospital results in superior outcomes compared to on-site fibrinolytic therapy or fibrinolysis followed by transfer for PCI. However, delays in transfer for primary PCI might negate its benefits. Recent studies have identified strategies to allow individual PCI hospitals and larger healthcare systems to decrease the delays to reperfusion by PCI but have involved patient populations limited in size or geography. We therefore attempted to identify the factors associated with delays to reperfusion by primary PCI in a large international trial, the Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction (HORIZONS-AMI) study.
Methods
The HORIZONS-AMI was a prospective, open-label, dual-arm factorial randomized, multicenter trial in which bivalirudin alone was compared to heparin plus a glycoprotein IIb/IIIa inhibitor and paclitaxel-eluting stents were compared to bare metal stents in patients with STEMI undergoing primary PCI. The study enrollment of 3,602 patients occurred from March 25, 2005 to May 7, 2007 at 123 hospitals in 11 countries (United States, Europe, South America, and Israel). Consecutive patients ≥18 years who presented within 12 hours after the onset of symptoms and who had ST-segment elevation of ≥1 mm in ≥2 contiguous leads, new left bundle branch block, or true posterior myocardial infarction were considered for enrollment. The exclusion criteria have been previously reported and included contraindications to the study medications, bleeding diathesis, planned surgery within 6 months that would require interruption of thienopyridine therapy, previous (within 30 days) stenting, or life expectancy <1 year. The patients were not excluded for cardiogenic shock, previous cardiac arrest, or other high-risk conditions. The study was approved by the institutional review board or ethics committee at each participating hospital, and all patients provided written informed consent.
The patients enrolled in the present study either presented directly to a tertiary PCI hospital or to a non-PCI hospital with subsequent transferral to a PCI hospital (where randomization occurred). No study-mandated protocols were in place for transfer or pre-enrollment treatment. After assessment at the study PCI hospital and screening for exclusion criteria, the patients provided informed consent and were randomly assigned in a 1:1 ratio to treatment with unfractionated heparin plus a glycoprotein IIb/IIIa inhibitor or to bivalirudin alone. Telephone randomization was performed using a computerized, interactive voice-response system. Aspirin (324 mg) and clopidogrel (300 or 600 mg) were given before angiography. Emergency coronary angiography with left ventriculography was performed after randomization followed by immediate primary PCI (n = 3,340, 92.7%), delayed PCI (n = 8, 0.2%), coronary artery bypass grafting (n = 63, 1.7%), or medical management (n = 191, 5.3%). After restoration of patency to the infarct vessel, the eligible patients undergoing primary PCI were randomized again in a 3:1 ratio to either a paclitaxel-eluting stent (Taxus Express, Boston Scientific) or bare metal stent (Express, Boston Scientific). The 3,340 patients who underwent PCI constituted the study cohort for the present retrospective substudy.
Of the 3,340 PCI patients, 905 (27.1%) had presented at non-PCI hospitals and were subsequently transferred to undergo immediate angiography and primary PCI, and the remaining 2,435 patients had presented and were enrolled at the PCI hospital. Missing time data included symptom onset times (n = 29), arrival times at the non-PCI hospital (n = 80), PCI hospital arrival times (n = 946 patients, including 386 transferred and 580 nontransferred patients), catheterization laboratory arrival times (n = 5), and time of first balloon inflation (n = 113). Thus, symptom onset to first hospital door time data were available for 2,716 patients, first hospital door-to-balloon time data were available for 2,639, non-PCI hospital to PCI hospital door time data were available for 448, PCI hospital door to catheterization laboratory arrival time data were available for 2,372, and catheterization laboratory arrival to balloon time data were available for 3,222 patients.
The raw time data were used to calculate the intervals between symptom onset and balloon inflation. Categorical outcomes were compared using the chi-square test or Fisher’s exact test. Continuous data are reported as the median (interquartile ranges) and were compared using the Wilcoxon rank-sum test. The univariate associations between 29 baseline demographic variables ( Table 1 ), 10 preprocedural adverse events ( Table 1 ), and 3 procedural variables (infarct-related artery, ejection fraction, and number of stents implanted) and each of the intervals were examined. Linear regression analysis was then performed to identify the independent correlates of each interval. Statistical significance was assumed at p = 0.05.
| Variable | All Patients (n = 3,340) | Patients With Door-to-Balloon Time Data (n = 2,639) | Patients Without Door-to-Balloon Time Data (n = 701) | p Value |
|---|---|---|---|---|
| Age (years) | 60.1 (52.5–69.7) | 59.7 (52.3–69.7) | 61.2 (53.7–69.8) | 0.06 |
| Men | 77.2% | 76.7% | 79.0% | 0.18 |
| Diabetes mellitus | 16.3% | 16.1% | 17.1% | 0.41 |
| Diabetes mellitus requiring insulin | 4.4% | 4.3% | 5.0% | 0.41 |
| Hypertension | 52.7% | 51.9% | 55.9% | 0.06 |
| Hyperlipidemia | 42.8% | 42.6% | 43.7% | 0.59 |
| Smoker | 64.4% | 64.1% | 65.5% | 0.50 |
| Previous coronary intervention | 10.5% | 10.6% | 10.4% | 0.91 |
| Previous bypass surgery | 2.6% | 2.5% | 3.0% | 0.50 |
| Previous angina pectoris | 21.9% | 21.2% | 24.4% | 0.07 |
| Previous myocardial infarction | 10.5% | 10.4% | 11.1% | 0.56 |
| Previous heart failure | 2.6% | 2.7% | 2.6% | 0.91 |
| Peripheral vascular disease | 4.4% | 3.9% | 6.3% | 0.007 |
| Previous major cardiac rhythm/rate disturbance | 2.9% | 2.8% | 3.1% | 0.63 |
| Enrolled in United States | 21.8% | 25.4% | 8.4% | <0.0001 |
| Presentation to noncoronary angioplasty hospital | 27.1% | 29.6% | 7.6% | <0.0001 |
| Presentation during daytime | 53.0% | 53.7% | 41.1% | 0.004 |
| Weight (kg) | 80.0 (71.0–90.0) | 80.0 (71.0–90.7) | 80.0 (70.0–90.0) | 0.23 |
| Body mass index (kg/m 2 ) | 27.1 (24.5–30.2) | 27.1 (24.6–30.3) | 27.0 (24.5–30.1) | 0.29 |
| Anemia (hematocrit <39% [men] or <36% [women]) | 10.4% | 9.9% | 12.4% | 0.06 |
| Platelet count (cells/mm 3 ) | 248 (208–292) | 248 (209–292) | 245 (204–291) | 0.28 |
| Creatinine clearance (ml/min) | 88.7 (68.2–113.5) | 89.4 (68.3–114.1) | 87.4 (68.1, 111.8) | 0.26 |
| Killip class 2–4 | 8.6% | 8.7% | 8.4% | 0.83 |
| Ventricular tachycardia/fibrillation | 2.2% | 2.2% | 2.4% | 0.67 |
| Cardiopulmonary resuscitation | 0.6% | 0.6% | 0.6% | 0.91 |
| Bradycardia (<40 beats/min) | 3.9% | 3.9% | 4.1% | 0.74 |
| Respiratory failure | 0.5% | 0.6% | 0.3% | 0.35 |
| Complete heart block | 1.6% | 1.3% | 2.7% | 0.007 |
| Atrial fibrillation or supraventricular tachycardia | 1.5% | 1.4% | 2.0% | 0.22 |
| Hypotension | 2.4% | 2.4% | 2.3% | 0.83 |
| Left ventricular ejection fraction (%) | 50 (42–59) | 50 (42–57) | 50 (45–61) | <0.0001 |
| Culprit artery | ||||
| Left anterior descending | 42.2% | 41.9% | 43.3% | 0.52 |
| Left circumflex | 16.4% | 17.4% | 12.6% | 0.003 |
| Left main coronary artery | 0.5% | 0.5% | 0.6% | 0.78 |
| Right | 45.5% | 44.8% | 48.0% | 0.13 |
| Number of stents implanted | 1 (1–2) | 1 (1–2) | 1 (1–2) | 0.03 |
Results
The baseline and procedural characteristics of all patients who underwent PCI for STEMI are listed in Table 1 . Presentation occurred during workday hours (Monday through Friday, 8 a.m. to 6 p.m. ) for 53% of patients, and 75% presented directly to a PCI-capable hospital. The baseline and procedural characteristics of the patients for whom the first door-to-balloon time was (n = 2,639) versus was not (n = 701) available were similar in most, but not all, respects ( Table 1 ).
The median first hospital door-to-balloon time was 99 minutes (interquartile range 73 to 135), and varied widely across the 123 PCI hospitals in the present study ( Figure 1 ). The median first hospital door-to-balloon time was longer for patients presenting first to a non-PCI hospital and then transferred to a PCI hospital (n = 781) than for to the patients presenting directly to a PCI hospital (n = 1,858; median 136 minutes, interquartile range 108 to 178, vs median 86 minutes, interquartile range 65 to 113, p <0.0001). Of the patients transferred from a non-PCI to a PCI hospital, the median first hospital door-to-balloon times in the United States (n = 150) were similar to those outside (n = 631) the United States (median 135.5 minutes, interquartile range 111 to 171, vs 136 minutes, interquartile range 108 to 179; p = 0.92). For patients presenting to a PCI hospital, the median first hospital door-to-balloon times in the United States (n = 520) were similar to those outside (n = 1,338) the United States (median 83 minutes, interquartile range 67 to 106, vs 89 minutes, interquartile range 65 to 115; p = 0.30).
The univariate and multivariate correlates of first hospital door-to-balloon time among patients with complete time data are listed in Tables 2 and 3 . The strongest correlate of delay in the door-to-balloon time was presentation to a non-PCI hospital requiring transfer, which, on linear regression analysis, was associated with a 54-minute incremental delay to reperfusion. Other important independent correlates of prolonged door-to-balloon times were presentation with respiratory failure (42-minute incremental delay), previous congestive heart failure (15-minute incremental delay), and hospital arrival during off-work hours (11-minute incremental delay).
| Variable | Estimate | Standard Error | p Value |
|---|---|---|---|
| Demographic | |||
| Age (years) | 0.26 | 0.09 | 0.005 |
| Women | 4.71 | 2.52 | 0.06 |
| Body mass index | 0.47 | 0.23 | 0.04 |
| Enrolled outside United States | 7.00 | 2.45 | 0.004 |
| Hypertension | 3.52 | 2.13 | 0.099 |
| Diabetes mellitus | 10.72 | 2.90 | <0.001 |
| Previous bypass surgery | 12.99 | 6.77 | 0.06 |
| Previous angina | 12.33 | 2.60 | <0.001 |
| Previous myocardial infarction | 7.98 | 3.49 | 0.02 |
| Previous heart failure | 23.01 | 6.62 | <0.001 |
| Peripheral vascular disease | 11.77 | 5.48 | 0.03 |
| Creatinine clearance <60 ml/min | 5.77 | 2.98 | 0.053 |
| Preprocedural status | |||
| Ventricular fibrillation | −15.01 | 7.33 | 0.04 |
| Respiratory failure | 38.48 | 14.17 | 0.007 |
| Presentation to non-PCI hospital | 52.81 | 2.10 | <0.001 |
| Presentation during off-hours | 11.89 | 2.13 | <0.001 |
| Procedural data | |||
| Infarct-related artery: circumflex | 7.36 | 2.15 | <0.001 |
| Left ventricular ejection fraction (%) | −0.21 | 0.10 | 0.03 |
| Variable | Estimate | Standard Error | p Value |
|---|---|---|---|
| Presentation to nonangioplasty hospital | 53.94 | 2.18 | <0.001 |
| Respiratory failure | 41.55 | 13.79 | 0.003 |
| Previous congestive heart failure | 15.17 | 6.57 | 0.02 |
| Presentation during daytime weekday hours | −10.52 | 1.99 | <0.001 |
| Diabetes mellitus | 9.30 | 2.75 | <0.001 |
| Previous angina | 8.02 | 2.44 | 0.001 |
| Men | −7.08 | 2.36 | 0.003 |
| Infarct-related artery: circumflex | 6.26 | 2.02 | 0.002 |
| Left ventricular ejection fraction | −0.25 | 0.09 | 0.003 |
Independent correlates of delayed symptom onset to the first hospital arrival time included the presence of diabetes (associated with a 28-minute incremental delay) and previous angina (23-minute incremental delay), enrollment outside the United States (22-minute incremental delay) and presentation during daytime weekday hours (13-minute incremental delay; Table 4 ). Many of the same factors were also independent correlates of delay from symptom onset to arrival at the PCI hospital ( Table 4 ). However, by far, the most important variable correlating with a delay to arrival at the PCI hospital was previous presentation at a non-PCI hospital (59-minute incremental delay).
| Variable | Estimate | Standard Error | p Value |
|---|---|---|---|
| Symptom onset to first hospital door (n = 2,522) | |||
| Diabetes mellitus | 28.27 | 7.46 | <0.001 |
| Previous angioplasty | −27.95 | 8.75 | 0.001 |
| Previous angina pectoris | 23.12 | 6.47 | <0.001 |
| Enrolled in United States | −21.94 | 6.27 | <0.001 |
| Hypertension | 13.50 | 5.53 | 0.02 |
| Presentation during daytime weekday hours | 13.31 | 5.29 | 0.01 |
| Age (years) | 1.80 | 0.30 | <0.001 |
| Weight (kg) | −0.77 | 0.22 | <0.001 |
| Creatinine clearance (ml/min) | 0.65 | 0.12 | <0.001 |
| Symptom onset to angioplasty hospital door (n = 2,169) | |||
| Presentation to nonangioplasty hospital | 58.79 | 7.55 | <0.001 |
| Diabetes mellitus | 32.68 | 8.06 | <0.001 |
| Previous coronary angioplasty | −29.87 | 9.11 | 0.001 |
| Previous angina pectoris | 26.65 | 7.14 | <0.001 |
| Enrolled in United States | −19.27 | 6.85 | 0.005 |
| Age (years) | 2.02 | 0.33 | <0.001 |
| Weight (kg) | −0.77 | 0.24 | 0.001 |
| Creatinine clearance (ml/min) | 0.67 | 0.13 | <0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
