Full Text

Nearly 250,000 patients have an ST-segment elevation myocardial infarction (STEMI) each year in the United States, and this condition results in an estimated 1 million hospital days and more than $6 billion in hospital-related costs (1,2). Reducing length of hospital stay (LOS) in this population, while maintaining quality and outcomes, would likely result in dramatic cost savings for the U.S. health care system. Success in the management of STEMI over the past several decades, in conjunction with early mobilization of patients, has led to a steady decrease in the LOS (3,4), which has paralleled a dramatic decline in overall mortality after STEMI (5,6). Opportunities to decrease LOS further have been suggested by geographic variation in LOS after STEMI (7,8). Although there is ever-growing financial pressure to shorten hospitalization, a lurking concern is that goals for hospital efficiency may outpace medical evidence and may place patients at risk for harm from hospitalization that is too short after STEMI. The evidence available to answer the question of the optimal LOS after primary percutaneous coronary intervention (PCI) for STEMI is incomplete. This is reflected in professional society recommendations that do not provide clear guidance on the topic (9,10).

Nonrandomized data suggest that an early discharge strategy in low-risk patients is safe (11,12), but randomized trials investigating the safety of early discharge are small and significantly underpowered (13–15). The largest of these trials had a statistical power of <20% to detect a difference in cardiovascular events by 6 months (14). PAMI-II (Primary Angioplasty in Myocardial Infarction II) randomized 471 low-risk patients to an early discharge or usual care (mean 4.7 days vs. 7.1 days, respectively) and found no difference in mortality or adverse clinical events (14). In a smaller trial, Prague-5 randomized 56 low-risk patients to either a next-day discharge strategy (n = 37, mean duration of hospitalization 29 h) or usual care (n = 19, mean duration of hospitalization 105 h) and found no serious complications by 30 days (15). Notably, more than 2,000 patients were screened for this study, but only 3.6% met the strict inclusion criteria.

In this issue of the Journal, Swaminathan et al. (16) provide supportive evidence that a short hospitalization (≤3 days) after primary PCI for STEMI is safe. Data from more than 33,000 patients from the National Cardiovascular Data Registry (NCDR) CathPCI Registry from 2004 to 2009 were linked to U.S. Center for Medicare & Medicaid Services (CMS) claims; therefore, only patients 65 years old or older were included. Patients’ characteristics and 30-day outcomes were stratified by LOS. Of particular importance, rather than using the more traditionally accepted “number of midnights” to define LOS, these investigators use the CMS definition. CMS calculates LOS as the discharge date minus the admission date plus 1; therefore, a patient admitted on a Monday and discharged on a Wednesday would be considered to have an LOS of 3 days as defined by the CMS, but the same patient would have an LOS of 2 days (2 midnights) using the more conventional definition.

Swaminathan et al. (16) observed that patients with a long LOS (>5 days) were older, had more comorbidities, and were more likely to have multivessel coronary disease or cardiogenic shock compared with patients with a medium (4 to 5 days) or short (≤3 days) LOS. Compared with a short LOS, patients with a long LOS had a higher 30-day mortality rate (unadjusted 0.9% vs. 3.5%, respectively). These findings are not unexpected. Differences in baseline characteristics or differences in the severity of and complications related to myocardial infarction are likely to have driven the major adverse cardiovascular event (MACE) and mortality rates, as well as the LOS. These investigators attempted to adjust for differences in severity of illness and comorbidities, but undoubtedly unmeasured confounders exist. The principal finding of this observational study, even with the limitations noted, was that there appeared to be no difference in 30-day mortality and MACE between patients with a short (≤3 days) LOS and patients with a medium (4 to 5 days) LOS.

An additional important finding is the significant geographic and hospital-related variation that seemed to affect LOS. Smaller hospitals and those in the West and Midwest regions of the United States were much more likely to have a short LOS. Specifically, 35% of patients in the West compared with only 16% of patients in the Northeast had a short LOS. This variation in practice may provide an opportunity for a substantial reduction in resource utilization at those hospitals and regions with a generally longer LOS after primary PCI for STEMI. Although clinical outcomes and quality must be carefully monitored with such change in practice, the adjusted analysis by Swaminathan et al. (16) indicates that such outcomes can be achieved with a shorter LOS.

In a secondary analysis, these investigators examine the mortality rates in a small subset of patients who were hospitalized for 1 to 2 days (representing patients with a same-day or next-day discharge). These 1,244 patients had a 30-day mortality rate twice that of patients discharged after 3 to 4 days of hospitalization. These results are intriguing, but they may simply highlight the limitations of the dataset analyzed. This group of patients represents only 3.7% of the entire study population. In addition, just as confounders such as comorbidities and severity of illness are likely to affect mortality in patients with a very long LOS, certain confounders may also play a role in patients with a very short LOS. Patients leaving against medical advice or patients transferred to another health care facility (potentially as a result of clinical instability) comprise potential factors that may increase the risk profile of the cohort with a very short LOS. These limitations significantly affect the ability to make any conclusions about this small subset of patients discharged after a very short LOS.

Unfortunately, the lack of a large randomized trial assessing the safety of an early discharge after primary PCI for STEMI leaves clinicians with incomplete evidence with which to make decisions. The paper by Swaminathan et al. (16) helps affirm that the current practice to discharge lower-risk patients early (≤3 days) is likely as safe as longer hospital stays. Given the observed geographic variation, there is probably an opportunity to apply this practice more broadly, thereby achieving significant health care cost savings while maintaining the quality of STEMI care.

Please click here to read the full text and references.

Copyright © 2015 American College of Cardiology Foundation. Published by Elsevier Inc. All rights reserved.