SAN FRANCISCO – Sending patients home the same day that they undergo percutaneous coronary intervention appears to be as safe as watching patients overnight in the hospital, based on a meta-analysis of 37 studies involving 12,803 patients.
Most patients underwent PCI for stable angina: 66% in the seven randomized, controlled trials and 98% in the 30 observational studies in the meta-analysis.
In a meta-analysis of results for 2,738 of the patients who were in seven randomized, controlled trials, rates for a composite of death, myocardial infarction (MI), and target lesion revascularization did not differ significantly between the same-day discharge group and the overnight observation group, Dr. Kimberly Brayton and her associates reported.
The two groups also did not differ significantly in the rate of another composite endpoint in the randomized trials: major bleeding or vascular complications, she said at the annual meeting of the American College of Cardiology.
Pooled results for 10,065 of the patients in 30 observational studies showed a 1% rate of the combined endpoint of death, MI, and target lesion revascularization and a 0.7% rate of major bleeding or vascular complications, said Dr. Brayton, of Stanford (Calif.) University.
Of the 15 deaths in the observational studies, 11 occurred more than 24 hours after PCI and "presumably were not modifiable by overnight observation," she added. Data on the remaining four deaths did not specify when they occurred and did not suggest the patients died between 6 and 24 hours following PCI.
The similarities of results between groups in the randomized controlled trials held steady in a secondary analysis that excluded three trials of patients with acute coronary syndrome, and in another secondary analysis that excluded studies conducted outside the United States.
Restricting the analysis to studies using only transfemoral access for PCI did not produce any significant differences between groups in major bleeding or vascular complications.
The findings support consideration of same-day discharge for carefully selected patients undergoing PCI, though larger studies are needed for more robust comparisons. "PCI is now much safer than it was in the past," she said.
Complications from PCI predominantly happen in the first 6 hours after the procedure, previous studies have shown. Outside the United States, patients commonly go home the day of PCI, but only approximately 1% of U.S. patients get same-day discharge after PCI, one study suggests (JAMA 2011;306:1461-1467).
Dr. Brayton and her associates developed a sample protocol for deciding which patients undergoing PCI to send home the same day and which to admit for observation, based on existing PCI guidelines and the protocols used in the studies in the meta-analyses.
Some preprocedure criteria would exclude patients from same-day discharge, including age older than 75 years, a glomerular filtration rate less than 60 mL/min per 1.73 m2, an ejection fraction less than 30%, the presence of acute coronary syndrome, allergy to contrast media, inadequate social support at home, or living more than 30-60 minutes away from the hospital.
If none of those factors are present, consider procedural factors that would exclude same-day discharge. These include significant left main disease, use of a glycoprotein IIb/IIIa inhibitor, performance of a complex PCI, or procedure-related complications.
Barring any of those, observe the patient for 6 hours and admit for overnight observation if there is any recurrent chest pain, hemodynamic instability, or bleeding or vascular complications. If not, the patient may go home the same day, but schedule a follow-up visit or phone call for 24 hours after PCI, Dr. Brayton suggested.
Patients had a mean age of 61 years in the randomized controlled trials and 62 years in the observational studies. Before PCI, 27% in the randomized studies and 29% in the observational studies had multivessel disease. As part of PCI, 39% in the randomized studies and 17% in the observational studies received glycoprotein inhibitors, and 29% in the observational studies received bivalirudin (with no data on bivalirudin for the randomized trials). A closure device was used in 4% of patients in the randomized trials and 54% in the observational studies.
The findings were limited by the heterogeneity of the studies in the meta-analysis, limited statistical power in the small numbers studied, the unavailability of patient-level data, and possible bias, especially in the observational studies, she said.
Dr. Brayton reported having no financial disclosures.
On Twitter @sherryboschert