Measurement of left ventricular end-diastolic pressure (LVEDP) is readily obtainable in patients with ST-segment elevation myocardial infarction (STEMI) undergoing primary percutaneous coronary intervention (PCI). However, the prognostic utility of LVEDP during primary PCI has never been studied. LVEDP was measured in 2,797 patients during primary PCI in the Harmonizing Outcomes with RevascularIZatiON and Stents in Acute Myocardial Infarction (HORIZONS-AMI) trial. Outcomes were assessed at 30 days and 2 years stratified by medians of LVEDP. Multivariable analysis was performed to determine whether LVEDP was an independent determinate of adverse outcomes. The median (interquartile range) for LVEDP was 18 mm Hg (12 to 24). For patients with LVEDP >18 mm Hg versus those with ≤18 mm Hg, hazard ratios (95% confidence intervals) for death and death or reinfarction at 30 days were 2.00 (1.20 to 3.33, p = 0.007) and 1.84 (1.24 to 2.73, p = 0.002), respectively, and at 2 years were 1.57 (1.12 to 2.21, p = 0.009) and 1.45 (1.14 to 1.85, p = 0.002), respectively. Patients in the highest quartile of LVEDP (≥24 mm Hg) were at the greatest risk of mortality. Only a weak correlation was present between LVEDP and left ventricular ejection fraction (LVEF; R 2 = 0.03, p <0.01). By multivariable analysis increased LVEDP was an independent predictor of death or reinfarction at 2 years (hazard ratio 1.20, 95% confidence interval 1.02 to 1.42, p = 0.03) even after adjustment for baseline LVEF. In conclusion, baseline increased LVEDP is an independent predictor of adverse outcomes in patients with STEMI undergoing primary PCI even after adjustment for baseline LVEF. Patients with LVEDP ≥24 mm Hg are at the greatest risk for early and late mortality.
Variables known to affect prognosis in ST-segment elevation myocardial infarction (STEMI) include age, associated co-morbidities, infarct size and location, left ventricular systolic function, ventricular arrhythmias, ischemic mitral regurgitation, and cardiogenic shock. The prognostic utility of left ventricular ejection fraction (LVEF) measured during or after STEMI is also well established. However, although diastolic dysfunction precedes the onset of systolic dysfunction in acute ischemia, the prognostic utility of diastolic indexes in patients with STEMI has rarely been examined. Moreover, although left ventricular end-diastolic filling pressure (LVEDP), which reflects global ventricular compliance, is routinely measured during left heart catheterization, no studies have evaluated the implications of LVEDP in patients with STEMI undergoing primary percutaneous coronary intervention (PCI). We therefore examined the prognostic value of LVEDP and the interrelation between LVEDP and LVEF as assessed during the primary PCI procedure in patients with STEMI enrolled in the Harmonizing Outcomes with RevascularIZatiON and Stents in Acute Myocardial Infarction (HORIZONS-AMI) trial.
Methods
The HORIZONS-AMI trial design has been previously described in detail ( http://www.ClinicalTrials.gov , identifier NCT00433966 ). In brief, HORIZONS-AMI was a prospective open-label 2-by-2 factorial randomized multicenter trial in which patients with STEMI who were undergoing primary PCI were randomized in a 1:1 ratio to anticoagulation with unfractionated heparin plus a glycoprotein IIb/IIIa inhibitor or bivalirudin monotherapy (with bail-out glycoprotein IIb/IIIa). After emergency coronary arteriography patients were triaged to PCI (performed in 92.7% of patients), coronary artery bypass grafting, or medical management. After patency was restored patients were randomly assigned again in a 3:1 ratio to TAXUS EXPRESS paclitaxel eluting stents or identical uncoated EXPRESS bare-metal stents (Boston Scientific, Natick, Massachusetts). Dual antiplatelet therapy was recommended for ≥1 year.
Contrast left ventriculography was performed during the index procedure; LVEDP was recorded just before contrast injection. LVEDP was measured at the Z-point, which is identified on the LV pressure trace as the point at which the slope of the ventricular pressure upstroke changes, which coincides with the R wave on the electrocardiographic tracing. An independent angiographic core laboratory (Cardiovascular Research Foundation, New York, New York) blinded to randomization assignment and clinical events assessed all angiograms for baseline and final lesion and flow characteristics and determined LVEF and severity of mitral regurgitation. Thrombosis In Myocardial Infarction (TIMI) flow was assessed as previously reported.
Analyses were performed for all patients who underwent randomization according to an intention to treat. Outcomes were stratified by median LVEDP. The primary end point in the present analysis was all-cause death and death or MI at 30 days and 2 years of follow-up. Continuous variables were expressed as median and interquartile range and were compared using Wilcoxon rank-sum test. Categorical variables were compared using chi-square test or Fisher’s exact test. Kaplan–Meier methods were used to estimate event rates at follow-up and to plot time-to-event curves; comparisons were made using log-rank test. Cox proportional hazards regression was performed to identify independent predictors of death or MI at 30 days and 2 years. The multivariable model was built by stepwise variable selection with entry and exit criteria set at the p <0.1 level. The following potential covariates were included in the model: age, gender, history of diabetes, smoking, Killip class, baseline TIMI flow, creatinine level, hemoglobin, randomization to bivalirudin versus unfractionated heparin plus glycoprotein IIb/IIIa inhibitors, symptom to first balloon time, left anterior descending coronary artery infarct, mitral regurgitation, clopidogrel loading dose, and LVEDP. Given the expected correlation between LVEF and LVEDP, multivariable models were created with and without LVEF to examine the incremental prognostic utility of LVEDP. Pearson correlation was used to evaluate the linear correlation between LVEDP and LVEF. All analyses were 2-sided and significance was established at the 0.05 level.
Results
Of 3,602 patients enrolled in HORIZONS-AMI, LVEDP was measured during the index procedure in 2,797 patients (77.6%), comprising the study cohort for the present analysis. Left ventriculography and LVEDP measurement were performed before PCI in 52% of cases, after PCI in 39.5%, and at an unknown time in 8.5%.
Patients with compared to those without measured LVEDP data were well matched in baseline characteristics, except those with LVEDP data were slightly younger (median age 60 vs 61 years, p = 0.03) and more frequently men (77% vs 74%, p = 0.04). Median (interquartile range) LVEDP and LVEF measured during the index procedure were 18 mm Hg (12 to 24) and 50% (43 to 60), respectively. Baseline clinical and angiographic characteristics of the study cohort stratified by median LVEDP are presented in Table 1 . Patients with LVEDP >18 mm Hg compared to those with ≤18 mm Hg had more diabetes, previous congestive heart failure, higher Killip class and lower LVEF, shorter door-to-balloon time, more culprit left anterior descending coronary artery disease, and were more likely to have baseline TIMI grade 0/1 flow on presentation. There were no significant differences in number of lesions and vessels treated, antithrombotic regimens, or stent types used in the catheter laboratory.
| Variable | LVEDP | p Value | |
|---|---|---|---|
| ≤18 mm Hg | >18 mm Hg | ||
| (n = 1,491) | (n = 1,306) | ||
| Age (years) | 59.7 (51.9–68.7) | 60.2 (53.0–70.2) | 0.06 |
| Women | 21.6% (322/1,491) | 23.7% (310/1,306) | 0.18 |
| Body mass index (kg/m 2 ) | 27.0 (24.4–30.0) | 27.1 (24.6–30.4) | 0.15 |
| Diabetes mellitus | 15.1% (225/1,491) | 17.8% (233/1,306) | 0.05 |
| Insulin treated | 4.2% (62/1,491) | 4.5% (59/1,306) | 0.64 |
| Hypertension | 54.1% (807/1,491) | 53.3% (696/1,306) | 0.66 |
| Hyperlipidemia | 41.9% (625/1,491) | 44.1% (576/1,306) | 0.24 |
| Smoking | 63.3% (942/1,488) | 63.1% (821/1,301) | 0.91 |
| Previous myocardial infarction | 10.2% (152/1,491) | 10.6% (138/1,306) | 0.75 |
| Previous coronary angioplasty | 9.8% (146/1,491) | 10.6% (138/1,305) | 0.49 |
| Previous coronary grafting | 2.7% (41/1,491) | 2.8% (37/1,306) | 0.89 |
| Congestive heart failure | 1.6% (24/1,491) | 4.0% (52/1,306) | 0.0001 |
| Peripheral arterial disease | 3.9% (58/1,491) | 4.6% (60/1,306) | 0.36 |
| Renal insufficiency | 2.3% (35/1,491) | 3.1% (41/1,306) | 0.20 |
| Door-to-balloon time (hours) | 1.65 (1.22–2.28) | 1.58 (1.18–2.07) | 0.007 |
| Symptom onset-to-balloon time (hours) | 3.70 (2.63–5.62) | 3.50 (2.63–5.32) | 0.08 |
| Killip classes II to IV | 5.8% (86/1,488) | 11.0% (144/1,305) | <0.0001 |
| Clopidogrel loading dose | |||
| 300 mg | 32.1% (479/1,491) | 32.4% (423/1,305) | 0.87 |
| 600 mg | 67.7% (1,010/1,491) | 67.6% (882/1,305) | 0.93 |
| Hemoglobin (g/dl) | 14.6 (13.7–15.5) | 14.6 (13.6–15.7) | 0.49 |
| Creatinine clearance (ml/min) | 91.1 (70.2–114.9) | 88.7 (66.0–113.1) | 0.08 |
| Randomized to bivalirudin | 50.7% (756/1,491) | 48.6% (635/1,306) | 0.27 |
| Randomized to unfractionated heparin + glycoprotein IIb/IIIa | 49.3% (735/1,491) | 51.4% (671/1,306) | 0.27 |
| Left ventricular ejection fraction (%) | 54 (45.60) | 46 (40–55) | <0.0001 |
| Primary management strategy | |||
| Primary percutaneous coronary intervention | 92.2% (1,375/1,491) | 94.6% (1,235/1,306) | 0.01 |
| Deferred percutaneous coronary intervention | 0.1% (1/1,491) | 0.0% (0/1,306) | 1.00 |
| Coronary bypass grafting without percutaneous coronary intervention | 1.3% (19/1,491) | 2.1% (28/1,306) | 0.07 |
| Patients undergoing percutaneous coronary intervention | |||
| Number of vessels treated | 1.0 ± 0.2 | 1.0 ± 0.2 | 0.57 |
| Number of lesions treated | 1.1 ± 0.4 | 1.1 ± 0.4 | 0.52 |
| Number of stents implanted | 1.5 ± 0.9 | 1.5 ± 0.8 | 0.19 |
| Left anterior descending coronary artery culprit lesion | 34.9% (512/1,465) | 46.8% (623/1,331) | <0.0001 |
| Thrombolysis In Myocardial Infarction flow before percutaneous coronary intervention | |||
| 0/1 | 59.8% (875/1,463) | 68.6% (909/1,325) | <0.0001 |
| 2/3 | 40.2% (588/1,463) | 31.4% (416/1,325) | <0.0001 |
| Randomized to TAXUS paclitaxel-eluting stent | 74.4% (936/1,258) | 76.1% (863/1,134) | 0.34 |
| Randomized to EXPRESS bare-metal stent | 25.6% (322/1,258) | 23.9% (271/1,134) | 0.34 |
| Any drug-eluting stent implanted | 72.5% (943/1,301) | 72.8% (863/1,185) | 0.85 |
| Bare-metal stent only implanted | 28.5% (371/1,301) | 28.7% (340/1,185) | 0.92 |
Of patients undergoing PCI, TIMI grade 3 flow was established slightly less frequently in patients with LVEDP >18 mm Hg compared to those with ≤18 mm Hg (91.2% vs 93.6%, p = 0.02). Patients with LVEDP >18 mm Hg were also more likely to have ≥25% decrease in hemoglobin level from baseline to nadir after PCI (10.5% vs 7.9%, p = 0.02) and were more likely to have ≥25% increase in serum creatinine (17.5% vs 13.0%, p = 0.002).
As presented in Table 2 , patients with LVEDP >18 mm Hg had significantly higher unadjusted rates of death, reinfarction, and major bleeding at 30 days and 2 years. Mortality was greatest in those with an LVEDP ≥24 mm Hg (p = 0.02; Figure 1 ) ; the incremental risk of death with increased LVEDP peaked within the first 6 weeks of the STEMI and was stable thereafter. There was no significant difference in survival among the 3 lower quartiles of LVEDP. Similarly, 2-year unadjusted rates of death or reinfarction were highest in patients with LVEDP ≥24 mm Hg ( Figure 1 ). Unadjusted rates of stroke, ischemic target vessel revascularization, and stent thrombosis were unrelated to LVEDP ( Table 2 ).
| Variable | LVEDP | HR (95% CI) | p Value | |
|---|---|---|---|---|
| ≤18 mm Hg | >18 mm Hg | |||
| (n = 1,491) | (n = 1,306) | |||
| 30-day events | ||||
| Net adverse clinical events (major adverse cardiovascular events or major bleeding) | 9.1% (135) | 12.9% (168) | 1.44 (1.15–1.81) | 0.001 |
| Major adverse cardiovascular events | 3.7% (55) | 6.3% (82) | 1.72 (1.22–2.42) | 0.002 |
| Death, all causes | 1.5% (23) | 3.1% (40) | 2.00 (1.20–3.33) | 0.007 |
| Cardiac | 1.4% (21) | 2.6% (34) | 1.86 (1.08–3.20) | 0.02 |
| Noncardiac | 0.1% (2) | 0.5% (6) | 3.45 (0.70–17.08) | 0.11 |
| Reinfarction | 1.3% (19) | 2.3% (29) | 1.76 (0.98–3.13) | 0.053 |
| Death or reinfarction | 2.7% (40) | 4.9% (64) | 1.84 (1.24–2.73) | 0.002 |
| Stroke | 0.5% (8) | 0.9% (12) | 1.72 (0.70–4.22) | 0.23 |
| Ischemic target vessel revascularization | 1.7% (25) | 2.3% (30) | 1.38 (0.81–2.34) | 0.23 |
| Major bleeding (non coronary-artery bypass surgery related) | 6.1% (91) | 8.6% (112) | 1.42 (1.08–1.87) | 0.01 |
| Major bleeding, all | 7.8% (116) | 10.9% (141) | 1.40 (1.10–1.80) | 0.006 |
| Stent thrombosis (definite or probable) | 1.9% (25) | 2.3% (28) | 1.23 (0.72–2.11) | 0.45 |
| 2-year events | ||||
| Net adverse clinical events | 22.0% (317) | 25.5% (327) | 1.21 (1.03–1.41) | 0.02 |
| Major adverse cardiovascular events | 17.3% (246) | 20.1% (255) | 1.20 (1.01–1.43) | 0.04 |
| Death, all causes | 4.0% (57) | 6.1% (78) | 1.57 (1.12–2.21) | 0.009 |
| Cardiac | 2.2% (32) | 3.9% (50) | 1.79 (1.15–2.79) | 0.009 |
| Noncardiac | 1.8% (25) | 2.3% (28) | 1.29 (0.75–2.20) | 0.36 |
| Reinfarction | 5.2% (72) | 7.0% (86) | 1.38 (1.01–1.89) | 0.04 |
| Death or reinfarction | 8.4% (121) | 11.9% (152) | 1.45 (1.14–1.85) | 0.002 |
| Stroke | 1.4% (20) | 2.0% (24) | 1.38 (0.76–2.50) | 0.28 |
| Ischemic target vessel revascularization | 11.6% (161) | 12.5% (153) | 1.10 (0.88–1.37) | 0.41 |
| Major bleeding (non coronary-artery bypass surgery related) | 7.1% (104) | 9.4% (121) | 1.34 (1.03–1.75) | 0.03 |
| Major bleeding, all | 8.7% (128) | 12.0% (155) | 1.40 (1.11–1.77) | 0.004 |
| Stent thrombosis (definite or probable) | 3.7% (47) | 4.7% (55) | 1.30 (0.88–1.92) | 0.19 |
Linear regression analysis demonstrated a very weak correlation between LVEDP and LVEF (R 2 = 0.03, p <0.01; Figure 2 ) . Stratified by medians, 2-year rates of death or reinfarction in patients with LVEF ≥50% and LVEDP ≤18 mm Hg (n = 971), LVEF ≥50% and LVEDP >18 mm Hg (n = 725), LVEF <50% and LVEDP ≤18 mm Hg (n = 208), and LVEF <50% and LVEDP >18 mm Hg (n = 307) were 7.4%, 9.1%, 12.3%, and 15.8%, respectively (p <0.01).
By multivariable analysis ( Table 3 ) LVEDP was an independent predictor of death or reinfarction at 30 days and 2 years when LVEF was not included in the model. When LVEF was added to the model, LVEDP was a borderline predictor of death or reinfarction at 30 days (p = 0.07) but remained a significant predictor of death or reinfarction at 2 years (p = 0.03).
