The implications of long-term acute care hospital transfer practices for measures of in-hospital mortality and length of stay.

RATIONALE: The National Quality Forum recently endorsed in-hospital mortality and intensive care unit length of stay (LOS) as quality indicators for patients in the intensive care unit. These measures may be affected by transferring patients to long-term acute care hospitals (LTACs). OBJECTIVES: To quantify the implications of LTAC transfer practices on variation in mortality index and LOS index for patients in academic medical centers. METHODS: We used a cross-sectional study design using data reported to the University HealthSystem Consortium from 2008-2009. Data were from patients who were mechanically ventilated for more than 96 hours. MEASUREMENTS AND MAIN RESULTS: Using linear regression, we measured the association between mortality index and LTAC transfer rate, with the hospital as the unit of analysis. Similar analyses were conducted for LOS index and cost index. A total of 137 hospitals were analyzed, averaging 534 transfers to LTAC per hospital during the study period. Mean±SD in-hospital mortality was 24±6.4%, and observed LOS was 30.4±8.2 days. The mean LTAC transfer rate was 15.7±13.7%. Linear regression demonstrated a significant correlation between transfer rate and mortality index (R2=0.14; P<0.0001) and LOS index (R2=0.43; P<0.0001). CONCLUSIONS: LTAC hospital transfer rate has a significant impact on reported mortality and LOS indices for patients requiring prolonged acute mechanical ventilation. This is an example of factors unrelated to quality of medical care or illness severity that must be considered when interpreting mortality and LOS as quality indicators.

Ionizing radiation exposure to patients admitted with acute myocardial infarction in the United States.

BACKGROUND: Invasive and noninvasive cardiovascular imaging is beneficial in the care of patients admitted with acute myocardial infarction. Little is known about patients' cumulative radiation exposure. METHODS AND RESULTS: All patients admitted with an acute myocardial infarction to any of 49 University HealthSystem Consortium member hospitals from 2006 to 2009 were reviewed for inpatient procedures involving ionizing radiation that included chest radiograph, computed tomogram scans, radionuclide imaging, diagnostic cardiac catheterization, and percutaneous coronary intervention. The average cumulative effective radiation dose per patient was estimated on the basis of published typical effective radiation doses for imaging procedures. Patients (n=64 071) admitted for acute myocardial infarction had a median age of 64.9 years. A total of 276 651 procedures involving ionizing radiation were performed during the study period, a median of 4.3 procedures per patient per admission. The majority of patients had invasive catheterization (77%), followed by computed tomogram scans (52%), mostly body examinations. The median cumulative effective radiation dose delivered was 15.02 mSv per patient per acute myocardial infarction admission. Postprocedural bleeding was a significant predictor of radiation exposure (odds ratio, 2.01; 95% confidence interval, 1.85 to 2.18), together with postprocedural mechanical complications resulting from device implantation (odds ratio, 2.86; 95% confidence interval, 2.61 to 3.13). Patients with higher underlying clinical complexity (defined by severity of illness scores) had higher radiation exposure and higher mortality (P<0.0001). There was also significant geographic variation in radiation exposure; patients in New England received the lowest cumulative exposure (odds ratio, 0.78; 95% confidence interval, 0.74 to 0.81). CONCLUSIONS: Acute myocardial infarction inpatients are exposed to an approximate median radiation dose of 15 mSv. This exposure is a result of multiple cardiovascular and noncardiovascular procedures. Efforts should be made to understand the risks and benefits of radiation exposure per episode of care for acute myocardial infarction.

Benchmarking risk-adjusted adult antibacterial drug use in 70 US academic medical center hospitals.

BACKGROUND: Antimicrobial stewardship programs are advised to measure and risk-adjust antimicrobial use to facilitate interhospital comparisons, a process called benchmarking. The purpose of this investigation was to evaluate a new benchmarking strategy for antibacterials. METHODS: Hospital-wide adult antibacterial drug use in 2009 was measured as days of therapy (DOT) and length of therapy (LOT) from billing records in 70 US academic medical centers (AMCs). Patients were assigned to 1 of 35 clinical service lines (CSL) based on their Medicare Severity Diagnosis Related Group. Expected (E) use was determined by indirect standardization and compared with observed (O) use. RESULTS: Of 1,791 ,180 discharged adults, 63.7% received antibacterial drugs; the range by CSL was 14.3% (psychiatry) to 99.7% (lung transplant). Mean ± SD hospital-wide use was 839 ± 106 DOTs (range, 594-1109) and 536 ± 53.0 LOT (range, 427-684) per 1000 patient-days. The ventilator support CSL had the most DOT per discharge, 39.4 ± 9.4 days; the LOT was 21.5 ± 4.5 days. The hospital-wide O/E ratio range was 0.7-1.45; in 5 AMCs the ratio exceeded the 90% confidence interval (CI) and was below the 90% CI in 6. Variability in use was explained by the proportion of treated patients within each CSL and mean LOT and DOT per discharge. CONCLUSIONS: Adult antibacterial drug use was benchmarked to expected use adjusted for patient mix, and outlier hospitals were identified. Differences between expected and observed use reflect usage patterns that were benchmarked and are targets for evaluation and intervention.

Diabetes mellitus does not adversely affect outcomes from a critical illness.

OBJECTIVE: Chronic diabetes mellitus (DM) is a known cause of multisystem injury. The effect of DM in acute critical illness may also be detrimental, but is not specifically known. We hypothesized that the preexisting diagnosis of DM is an independent risk factor for mortality in critically ill patients. DESIGN: Parallel retrospective and prospective cohort study. SETTING: Two large patient datasets were used: the retrospective University HealthSystem Consortium database (UHC) and the prospective Mayo Clinic Acute Physiology And Chronic Health Evaluation III critical care database (Mayo). PATIENTS: Inclusion criteria were admission to an intensive care unit and age > or =18 yrs. Patients with diabetic ketoacidosis or hyperosmolar nonketotic coma were excluded. A total of 1,509,890 patients (including 143,078 deaths) in the UHC cohort and 36,414 patients (including 3562 deaths) in the Mayo cohort were included in the study analysis. MEASUREMENTS AND MAIN RESULTS: The primary outcome was in-hospital mortality compared between patients with a history of DM and all other patients. Other outcomes included in-hospital mortality in prespecified subgroups. In the UHC dataset, patients with DM had a lower unadjusted odds ratio (0.90, 95% confidence interval 0.89-0.91, p < .001) and a lower adjusted effect on mortality (odds ratio 0.75, 0.74-0.76, p < .001) compared with that seen in patients without DM. In the Mayo dataset, patients with DM had a comparable unadjusted odds ratio (1.07, 0.97-1.17, p = NS) and a lower adjusted effect on mortality (odds ratio 0.88, 0.79-0.98, p = .022) compared with that seen in patients without DM. A lower mortality in diabetic patients held across multiple demographic subgroups, including patients who underwent coronary-artery bypass grafting (UHC data: unadjusted odds ratio 0.66, 0.62-0.71, p < .001). CONCLUSIONS: Critically ill adults with DM do not have an increased mortality compared with that seen in patients without DM, and may have a decreased mortality. Further investigation needs to be done to determine the mechanism for this effect.

Racial disparity in the relationship between hospital volume and mortality among patients undergoing coronary artery bypass grafting.

OBJECTIVE: To examine whether the volume-mortality relationship in coronary artery bypass grafting (CABG) differs by race and operative risk. SUMMARY BACKGROUND DATA: In-hospital mortality after CABG is inversely associated with hospital volume. Racial disparities exist in the outcomes of CABG, possibly due to blacks' high operative risk. METHODS: We analyzed 71,949 CABG procedures performed between 2002 and 2005 at 93 academic medical centers participating in the University HealthSystem Consortium. In-hospital mortality was examined across hospital volume categories (very low, <100/yr; low, 100-299/yr; medium, 300-499/yr; and high, > or =500/yr) via logistic regression. RESULTS: In-hospital mortality was 2.0% in whites and 2.8% in blacks. Controlling for patient risk, geographic region, and proportion of African American patients treated at the hospital, the benefit of higher volume was substantial for blacks but only modest for whites (race-by-volume interaction; P = 0.033). Odds ratios of mortality for increasing volume categories (compared with very low volume) were 0.46, 0.37, and 0.47 among blacks but only 0.85, 0.77, and 0.75 among whites. Racial disparities in mortality existed mostly in very low-volume hospitals. The differential volume effect across the 2 racial groups seemed to be primarily driven by regional patterns, as the volume effect was much more pronounced in the South and the Midwest (region by volume interaction; P = 0.033). CONCLUSIONS: Blacks have greater reduction in mortality than whites by undergoing CABG at higher-volume hospitals, regardless of operative risk. Because of limited generalizability, these findings should be confirmed using more representative database.

The status of adult inpatient care by family physicians at US academic medical centers and affiliated teaching hospitals 2003 to 2012: the impact of the hospitalist movement.

BACKGROUND AND OBJECTIVES: Over the past 10--15 years, the number of hospitalists has grown from a few hundred to over 30,000, and hospitalists have assumed a greater proportion of the care of hospitalized patients. No existing studies report on the impact this movement has had on the characteristics of the hospital practice of family physicians in teaching hospitals. To explore this impact we examined the volume and scope of discharges by family physicians at teaching hospitals in 2003 and 2012, the most recent decade of hospitalist growth. We also compared the characteristics of family physicians' hospital practices in 2012 with hospitalists in 2012 to assess for differences in hospital performance. METHODS: We used the University Health Consortium (UHC) Clinical Database to capture adult non-pregnancy-related inpatient discharges in US teaching hospitals by family physicians and hospitalists in 2003 and 2012. We calculated the proportion of inpatient discharges by specialty in all UCH hospitals; did a qualitative comparison of frequent discharge diagnoses; and analyzed length of stay, case mix index, 7- and 30-day readmission rates, and mortality by specialty using UHCs risk adjustment methodology. RESULTS: The proportion of all inpatient discharges by family physicians in UHC hospitals was stable between 2003 and 2012 (2.7% versus 2.6%) though the volume increased. Over the same time, the proportion of discharges attributable to hospitalists increased (0.0% to 13.9%) with a concomitant decrease in proportion of discharges from general medicine (18.4% to 13.9%) and all other specialties (78.9% to 73.2%). Fourteen of the top 20 discharge diagnoses by family physicians from UHC hospitals were the same between 2003 and 2012. Family physicians and hospitalists shared 17 of the top 20 discharge diagnoses in 2012. Length of hospital stay was stable for family medicine across time and lower than that of hospitalists (4.5 versus 5.5 days; P<.001). Seven- and 30-day readmission rates for any cause were lower for hospitalists but there was no difference in either rate when limited to readmission for the same DRG. Hospitalists cared for a somewhat more complex patient mix. CONCLUSIONS: The growth of hospitalists has had little to no impact on the proportion of inpatient discharges in teaching hospitals by family physicians. Quality of care as judged by length of stay, mortality, and readmission rates was comparable between family physicians and hospitals in 2012.

Incidence of hospital-acquired venous thromboembolic codes in medical patients hospitalized in academic medical centers.

BACKGROUND: Hospital-acquired venous thromboembolism (HA-VTE, VTE occurring during a hospitalization) codes in hospital billing data are often used as a surrogate for hospital-associated VTE events occurring during or up to 30 days after a hospitalization, which are more difficult to measure. OBJECTIVE: Establish the incidence and composition of HA-VTE/superficial venous thrombosis (SVT) coded in a large cohort of medical patients. DESIGN: Retrospective analysis of discharges. SETTING: Eighty-three academic medical centers in UHC (formerly University HealthSystem Consortium). PATIENTS: Patients with medical diagnoses hospitalized >2 days between October 1, 2009, and March 31, 2011. MEASUREMENTS: Incidence and anatomic location of HA-VTE codes, defined as International Classification of Diseases, Ninth Revision, Clinical Modification codes for VTE coupled to a present-on-admission indicator flag set to "No." RESULTS: Among 2,525,068 medical hospitalizations, 12,847 (0.51%) cases had ≥1 thrombotic code; 2449 (19.1%) with pulmonary embolism (PE), and 3848 (30%) with lower-extremity deep venous thrombosis (LE-DVT) without PE. Upper-extremity DVT (2893; 22.5%) and SVT (3248; 25.3%) comprised the bulk of remaining cases. Among cases with HA-PE/LE-DVT, 34.3% had cancer, 47.8% received care in an intensive care unit, 78% had severe or extreme severity of illness, and 16.5% died in the hospital. Overall, 54.9% of the patients who developed a HA-PE/LE-DVT had been started on VTE pharmacoprophylaxis on hospital day 1 or 2. CONCLUSION: At academic centers, HA-VTE/SVT is coded in 0.51% of medical inpatients, and HA-PE/LE-DVT is coded in half of those. Most patients with HA-PE/LE-DVT are severely ill and develop VTE despite receiving prophylaxis.

Outpatient thyroid surgery data from the University Health System (UHC) Consortium.

OBJECTIVE: Describe data from patients undergoing outpatient thyroid surgeries for benign and malignant disease at academic medical centers in the United States. STUDY DESIGN: Retrospective database search. SETTING: The University Health System Consortium (UHC), Oak Brook, Illinois, data compiled from discharge summaries. SUBJECTS AND METHODS: Discharge data were collected from the first quarter of 2005 through the fourth quarter of 2010. Searching strategy was based on diagnosis of thyroid disease and patients undergoing thyroid surgery across all UHC facilities. Demographic information was collected as well as charges. Complications were also evaluated in this analysis. RESULTS: During the study period, 38,362 outpatient thyroidectomies were performed from our sample, 32% for thyroid cancer. More total thyroidectomies (43%) and fewer hemithyroidectomies (36%) were being performed overall; 64.1% of patients stayed 23 hours. CONCLUSION: This is one of the largest series reporting outcomes for outpatient thyroid surgery. Since these surgeries appear to be shifting to an outpatient setting, this report reflects the experience with the majority of endocrine surgeries from the UHC database being performed presently. These results are derived from teaching hospitals and their affiliates and may not reflect the entirety of thyroid surgery in the United States.

High resource utilization does not affect mortality in acute respiratory failure patients managed with tracheostomy.

BACKGROUND: Tracheostomy practice in patients with acute respiratory failure (ARF) varies greatly among institutions. This variability has the potential to be reflected in the resources expended providing care. In various healthcare environments, increased resource expenditure has been associated with a favorable effect on outcome. OBJECTIVE: To examine the association between institutional resource expenditure and mortality in ARF patients managed with tracheostomy. METHODS: We developed analytic models employing the University Health Systems Consortium (Oakbrook, Illinois) database. Administrative coding data were used to identify patients with the principal diagnosis of ARF, procedures, complications, post-discharge destination, and survival. Mean resource intensity of participating academic medical centers was determined using risk-adjusted estimates of costs. Mortality risk was determined using a multivariable approach that incorporated patient-level demographic and clinical variables and institution-level resource intensity. RESULTS: We analyzed data from 44,124 ARF subjects, 4,776 (10.8%) of whom underwent tracheostomy. Compared to low-resource-intensity settings, treatment in high-resource-intensity academic medical centers was associated with increased risk of mortality (odds ratio 1.11, 95% CI 1.05-1.76), including those managed with tracheostomy (odds ratio high-resource-intensity academic medical center with tracheostomy 1.10, 95% CI 1.04-1.17). We examined the relationship between complication development and outcome. While neither the profile nor number of complications accumulated differed comparing treatment environments (P > .05 for both), mortality for tracheostomy patients experiencing complications was greater in high-resource-intensity (95/313, 30.3%) versus low-resource-intensity (552/2,587, 21.3%) academic medical centers (P < .001). CONCLUSIONS: We were unable to demonstrate a positive relationship between resource expenditure and outcome in ARF patients managed with tracheostomy.

Inappropriate prescribing of proton pump inhibitors in hospitalized patients.

BACKGROUND: Proton pump inhibitors have numerous important side effects, yet they are prescribed for outpatients who do not have recognized indications. Less is known with respect to prescribing for inpatients. OBJECTIVE: To determine the rate of inappropriate prescribing of protein pump inhibitors and to assess reasons why they are prescribed. DESIGN AND PARTICIPANTS: The study was a retrospective review of administrative data for adult hospital patients discharged from the Medicine service of Denver Health (DH) and from the University HealthSystem Consortium (UHC) between January 1, 2008 and December 31, 2009. MEASUREMENTS: Valid indications for proton pump inhibitors were sought from discharge diagnoses, prescription records, and, in a randomly selected group of patients from DH, from direct review of records. RESULTS: Inclusion criteria were met by 9875 DH patients and 6,592,100 UHC patients; of patients receiving a proton pump inhibitor, 61% and 73%, respectively, did not have a valid indication. Increased rates of Clostridium difficile infection were found in both groups of patients receiving proton pump inhibitors. Chart reviews found valid indications for proton pump inhibitors in 19% of patients who did not have a valid indication on the basis of the administrative data, and "prophylaxis" was the justification for inappropriate prescribing in 56%. CONCLUSION: Proton pump inhibitors are frequently inappropriately prescribed to Medicine inpatients who do not have a valid indication and this practice is associated with an increase in C. difficile infection. Interventions are needed to curtail this inappropriate prescribing practice.

Outpatient parathyroid surgery data from the University Health System Consortium.

OBJECTIVE: To determine demographics and cost for outpatients undergoing parathyroid surgery at hospitals belonging to the University Health System Consortium (UHC). STUDY DESIGN: UHC data were accessed in 2011 and reflected data collected from 2005 through 2010 (24 quarters). Searching strategy was based on diagnoses of parathyroid disease and patients undergoing parathyroidectomy across all UHC member facilities. Complications evaluated in this analysis included: hypocalcemia, hypoparathyroidism, aspiration pneumonia, hematoma, wound infection, stroke, myocardial infarction, deep venous thrombosis/pulmonary embolism (PE), and death. SETTING: The University Health System Consortium, Oak Brook, Illinois, was formed in 1984 and consists of 112 academic medical centers and 250 of their affiliated hospitals. This represents 90% of the nonprofit academic medical centers in the United States (www.uhc.edu). SUBJECTS AND METHODS: Patients enrolled in the UHC database were studied retrospectively. Data were compiled from discharge summaries into a secure, interactive, Web-based database. The outpatient data collection set has been a recent addition to the originally established UHC inpatient discharge database. RESULTS: There were 21,057 patients who had outpatient parathyroid surgery. The average age was 59.0 (0.8-96.2) yrs. Seventy-six percent of patients were female. Outpatient parathyroidectomy had lower charges than inpatient surgery ($12,738 and $14,657, respectively; P = 0.004, Wilcoxon signed-rank test). Complications were low but were likely underreported. CONCLUSION: Parathyroid surgery is increasingly being done in the outpatient setting in the United States. By virtue of omitting inpatient hospitalization, the outpatient approach becomes a more economical way to manage parathyroid disease. This is the largest known series reporting experience with outpatient parathyroid surgery.

Increasing burden of methicillin-resistant Staphylococcus aureus hospitalizations at US academic medical centers, 2003-2008.

OBJECTIVE: The incidence of invasive methicillin-resistant Staphylococcus aureus (MRSA) infections in the United States decreased during 2005-2008, but noninvasive community-associated MRSA (CA-MRSA) infections also frequently lead to hospitalization. We estimated the incidence of all MRSA infections among inpatients at US academic medical centers (AMCs) per 1,000 admissions during 2003-2008. DESIGN: Retrospective cohort study. SETTING AND PARTICIPANTS: Hospitalized patients at 90% of nonprofit US AMCs during 2003-2008. METHODS: Administrative data on MRSA infections from a hospital discharge database (University HealthSystem Consortium [UHC]) were adjusted for underreporting of the MRSA V09.0 International Classification of Diseases, Ninth Revision, Clinical Modification code and validated using chart reviews for patients with known MRSA infections in 2004-2005, 2006, and 2007. RESULTS: The mean sensitivity of administrative data for MRSA infections at the University of Chicago Medical Center in three 12-month periods during 2004-2007 was 59.1%. On the basis of estimates of billing data sensitivity from the literature and the University of Chicago Medical Center, the number of MRSA infections per 1,000 hospital discharges at US AMCs increased from 20.9 (range, 11.1-47.7) in 2003 to 41.7 (range, 21.9-94.0) in 2008. At the University of Chicago Medical Center, among infections cultured more than 3 days prior to hospital discharge, CA-MRSA infections were more likely to be captured in the UHC billing-derived data than were healthcare-associated MRSA infections. CONCLUSIONS: The number of hospital admissions for any MRSA infection per 1,000 hospital admissions overall increased during 2003-2008. Use of unadjusted administrative hospital discharge data or surveillance for invasive disease far underestimates the number of MRSA infections among hospitalized patients.

Characteristics of inpatient thyroid surgery at US academic and affiliated medical centers.

OBJECTIVE: Describe data from patients undergoing thyroid surgeries for benign and malignant disease at US academic medical centers. STUDY DESIGN: Retrospective, database search. SETTING: The University Health System (UHC) Consortium (Oak Brook, Illinois) data compiled from discharge summaries. SUBJECTS AND METHODS: Discharge data were collected from the first quarter of 2002 through the fourth quarter of 2009. Searching strategy was based on diagnosis of thyroid disease and patients undergoing thyroid surgery across all UHC facilities. Demographic information was collected as well as length of stay (LOS) and costs. Complications were evaluated in this analysis. RESULTS: During the study period, 68,014 thyroidectomies were performed, with 27,200 for thyroid cancer. During the same period 6365 neck dissections were performed, with 1539 as stand-alone procedures. Total thyroidectomy was the procedure of choice for malignant disease. More total thyroidectomies and fewer hemithyroidectomies were being performed for benign thyroid disease in the inpatient setting. Almost all postoperative complications were more frequent after surgery for cancer except myocardial infarction and aspiration pneumonia. On average, LOS was longer for benign disease, but costs were higher for malignant disease. CONCLUSION: This is the largest series reporting inpatient LOS and mortality for thyroid surgery. The limitation of this study is that it reports patients whose stays were more than 23 hours, leaving out a significant number of thyroid surgeries that are performed as outpatients. Although the results contribute greatly to characterizing inpatient surgery, the results may not reflect current US trends for thyroid surgery.

Safety of selective serotonin reuptake inhibitor in adults undergoing coronary artery bypass grafting.

Selective serotonin reuptake inhibitors (SSRIs) are commonly used in patients with coronary artery disease and depression, but they have been reported to increase the risk for bleeding. However, data on the short-term outcomes comparing SSRI and non-SSRI antidepressant use after coronary artery bypass grafting (CABG) are limited. A retrospective analysis was conducted of 1,380 adults who received any antidepressants before CABG from 2003 to 2006 at academic medical centers participating in the University HealthSystem Consortium. The primary end point was defined as a composite of in-hospital mortality or any bleeding events, including postprocedural hemorrhage or hematoma, gastrointestinal hemorrhage, and reopening of surgical site. A total of 1,076 adults (78%) received SSRIs. After controlling for propensity of receiving SSRIs compared with non-SSRIs, no significant differences were found in the primary end point (9.4% vs 8.2%, adjusted odds ratio [OR] 1.03, 95% confidence interval [CI] 0.60 to 1.78), any bleeding events (6.5% vs 7.2%, OR 0.93, 95% CI 0.50 to 1.76), or in-hospital mortality (3.1% vs 2.3%, OR 0.88, 95% CI 0.47 to 1.65). There was no increased risk associated with SSRI use when the analysis was restricted to patients who received antiplatelet and anticoagulant therapy for acute coronary syndromes (OR 1.03, 95% CI 0.40 to 2.61) and when examined by age, gender, nonsteroidal anti-inflammatory drug use, and type of CABG (on pump or off pump). In conclusion, compared with non-SSRIs, the preoperative use of SSRIs does not seem to increase the risk for bleeding or in-hospital mortality after CABG.

Aspirin and clopidogrel use in the early postoperative period following on-pump and off-pump coronary artery bypass grafting.

OBJECTIVE: Preoperative use of clopidogrel increases the risk of bleeding, but its postoperative use has not been studied. We studied early postoperative clopidogrel use in on-pump and off-pump coronary artery bypass grafting. METHODS: Data were obtained from the University HealthSystem Consortium database. We conducted a retrospective analysis of data of 15,067 adults who had coronary artery bypass grafting between 2003 and 2006 and received perioperative aspirin alone or in combination with clopidogrel, with the latter administered within 2 days after coronary artery bypass grafting. Logistic regression was used to analyze in-hospital mortality, 30-day readmission, ischemic or thrombotic events, and bleeding events, with propensity score adjustment for clopidogrel treatment. RESULTS: Combined aspirin and clopidogrel were used in 3268 patients (22%). Compared with aspirin alone, aspirin plus clopidogrel was associated with reductions of in-hospital mortality (0.95% vs 1.78%; adjusted odds ratio: 0.50; 95% confidence interval: 0.25, 0.99) and bleeding events (4.19% vs 5.17%; adjusted odds ratio: 0.70; 95% confidence interval: 0.51, 0.97). Ischemic or thrombotic events were not significantly different (1.29% vs 1.53%; adjusted odds ratio, 0.99; 95% confidence interval: 0.59, 1.64). The relative effect of combined treatment did not differ between on-pump and off-pump coronary artery bypass grafting. CONCLUSIONS: Early postoperative clopidogrel combined with aspirin may be safe and beneficial compared with perioperative aspirin treatment alone, in both on-pump and off-pump coronary artery bypass grafting. However, a possibility of selection bias calls for randomized controlled trials to confirm our findings.