Sage Journals: Discover world-class research

Abstract

Background:

Several clinical trials have documented clinical benefits associated with prophylactic corticosteroid administration at the time of coronary artery bypass graft (CABG) surgery, including a reduction in the risk of atrial fibrillation and hospital length of stay. Despite the published data, the extent to which providers have adopted the perioperative use of corticosteroids remains unknown.

Objectives:

To assess temporal trends, between-hospital variation, and determinants of perioperative intravenous corticosteroid use during CABG surgery.

Methods:

We identified all patients admitted for CABG surgery in the Premier Healthcare Database (2003-2014), a large US-based inpatient database. We determined the proportion of patients administered prophylactic corticosteroids on the day of CABG surgery. Linear time-series models were used to estimate the rate and trend of corticosteroid use over time. Separate multivariable generalized estimating equation models were used to quantify the variation in and determinants of perioperative corticosteroid use.

Results:

Of 401 788 eligible patients who underwent a CABG surgery between 2003 and 2014, 20% (n = 80 681) were administered intravenous prophylactic perioperative corticosteroids (methylprednisolone, dexamethasone, or hydrocortisone). Corticosteroid use increased from 17.5% in 2003 to 22.6% in 2014 (annual rate = 0.42%; P < .001). Individual hospitals accounted for >50% of variation in corticosteroid use. High between-hospital variation was also observed, and the probability of utilization was ≥32.4% in the upper versus ≤3.4% in the bottom quartiles of hospitals.

Conclusion:

Prophylactic corticosteroid administration during CABG has increased gradually since 2003. To further evaluate the risk–benefit trade-off associated with their use, we believe a large-scale outcomes study is warranted to assess this highly variable practice.

Keywords

coronary artery bypass graft (CABG)preoperative steroids use

Introduction

Nearly half a million coronary artery bypass graft (CABG) surgery procedures are performed annually in the United States.¹ Cardiopulmonary bypass and injury related to ischemia and reperfusion in cardiac surgery can induce a severe systemic inflammatory response.^2,3,4 This inflammatory process is associated with an extensive release of pro- and anti-inflammatory cytokines^5,6 and can lead to respiratory and multiple organ failure.^7,8

The prophylactic administration of corticosteroids has been studied and debated for more than 3 decades as a means to attenuate the acute phase response in cardiac surgery.^9,10 Corticosteroids have potent anti-inflammatory effects and have been shown to reduce the release of pro-inflammatory cytokines, inhibit macrophage production, and have favorable effects on C-reactive protein.^11,12 Existing evidence suggests that corticosteroids during cardiac surgery significantly reduce the risk of atrial fibrillation and shorten hospital length of stay, without increasing the risk of infection or other complications.¹³ Importantly, corticosteroids increase blood sugar levels through gluconeogenesis, alter fluid and electrolyte balance, and suppress the immune system.^14,15 Moreover, the impact of corticosteroids on perioperative mortality is unclear, and confirmation regarding the safety of their perioperative use in cardiac surgery is lacking.^10,16 In 2004, the American College of Cardiology/American Heart Association (ACC/AHA) guidelines for CABG supported the “liberal prophylactic use” of corticosteroids for the reduction of systemic inflammation among patients undergoing CABG, especially for those receiving on-pump CABG with cardiopulmonary bypass (ie, extracorporeal circulatory support).¹⁷ However, the most recent update, published in 2011, no longer recommended prophylactic corticosteroids.¹⁸

There are currently no national data available regarding the trends and patterns of utilization of perioperative prophylactic corticosteroids. Furthermore, it remains uncertain as to which factors influence the administration of corticosteroids during CABG surgery. Therefore, we sought to assess trends in the use of perioperative intravenous corticosteroids in CABG surgery in the United States and to identify patient- and hospital-level factors that determine their use. We aimed to explore between-hospital variation in the prophylactic perioperative administration of intravenous corticosteroids.

Methods

Data Source and Cohort Selection

We generated a cohort of patients who underwent CABG surgery (both on-pump and off-pump) using 2003 to 2014 data from the Premier Healthcare Database. This database captures data from over 700 geographically diverse hospitals and health systems in the United States, including community and teaching hospitals in both rural and urban settings. A large inpatient database, Premier Healthcare Database contains information on ∼1 in 6 hospitalizations in the United States, including data for ∼148 million patients.¹⁹ The database contains a record of daily charges for all medications, procedures, and diagnostic tests during each patient’s hospitalization, as well as patient demographic and hospital characteristics, discharge diagnoses, and discharge status.

Figure 1 is a schematic of how patients were selected for analyses. The analyses included all patients who underwent CABG surgery. To focus on patients who received corticosteroids prophylactically, rather than for the treatment of other preexisting conditions, we excluded those who received oral or inhaled corticosteroids on any hospital day prior to or on the day of CABG surgery and those who received intravenous corticosteroids on any hospital day prior to surgery.

Figure 1.

Sample selection of patients undergoing coronary artery bypass graft (CABG) surgery.

Outcome

The study outcome was intravenous corticosteroid use on the day of surgery, which was ascertained through hospital charge codes. For the time trend analysis, we calculated the proportion of CABG surgeries in which patients received intravenous corticosteroids on the day of surgery. We calculated yearly proportions both overall and separately for each intravenous corticosteroid. For the utilization prediction analysis, we modeled intravenous corticosteroid use as a binary outcome.

Study Variables

We ascertained both patient- and facility-level characteristics using a combination of International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) discharge diagnosis codes, procedure codes, and hospital charge codes. All charge codes used to define covariates were assessed in the period between initial hospitalization and the date of CABG surgery (inclusive). All patient characteristics, demographic information, preexisting medical conditions, medical procedures, and medication use were defined based on available preoperative data, except for valve repair/replacement and on-/off-pump procedures, which were assessed on the day of CABG. The hospital-level characteristics included number of hospital beds, regional location, teaching hospital status, and whether the hospital was located in a rural or urban area. A full list of study variables is provided in Table 1. All charge and ICD-9 codes are provided in Appendix A.

Table 1.

Distribution of Patient- and Hospital-Level Characteristics by Perioperative Intravenous Corticosteroid Administration Status.

Characteristics	Intravenous Corticosteroid Used						No Intravenous Corticosteroid Used (N = 321 107, 79.9%)
	Dexamethasone (N = 26 705, 6.6%)		Hydrocortisone (N = 14 685, 3.7%)		Methylprednisolone (N = 39 291, 9.8%)		No Intravenous Corticosteroid Used (N = 321 107, 79.9%)
	n	%	n	%	n	%	n	%
Demographics
Race, %
White	20 377	76.3	11 062	75.3	30 253	77.0	236 087	73.5
Black	1073	4.0	596	4.1	2873	7.3	21 584	6.7
Hispanic	623	2.3	151	1.0	857	2.2	7225	2.3
American Indian	58	0.2	91	0.6	61	0.2	878	0.3
Asian Pacific	151	0.6	48	0.3	216	0.5	1769	0.6
Other	4423	16.6	2737	18.6	5031	12.8	53 564	16.7
Age, %
<46	724	2.7	426	2.9	1164	3.0	9358	2.9
45-54	3361	12.6	1771	12.1	5117	13.0	41 136	12.8
55-64	7497	28.1	3910	26.6	10 880	27.7	88 002	27.4
65-75	8524	31.9	4978	33.9	13 083	33.3	103 978	32.4
>75	6599	24.7	3600	24.5	9047	23.0	78 633	24.5
Gender, %
Female	7148	26.8	4265	29.0	11 118	28.3	89 349	27.8
Clinical factors prior to surgery
Previous CABG	676	2.5	431	2.8	1253	2.7	2055	2.6
Percutaneous coronary intervention during the index admission	1106	4.1	573	3.9	1517	3.9	12 140	3.8
Echocardiogram	12 442	46.6	6752	46.0	18 948	48.2	152 157	47.4
Telemetry	6181	23.1	1745	11.9	8446	21.5	71 281	22.2
Chronic/end-stage renal failure	3195	12.0	1716	11.7	5106	13.0	38 874	12.1
ICU stay	24 768	92.7	13 883	94.5	36 116	91.9	300 000	93.4
Clinical factors on day of surgery
Off-pump surgery	296	1.1	235	1.6	494	1.3	3859	1.2
Valve repair/replacement	2224	8.3	893	6.1	2874	7.3	22 971	7.2
Perioperative corticosteroid use
ACE inhibitors	6894	25.8	3909	26.6	10 722	27.3	84 858	26.4
ARB	1802	6.7	907	6.2	2803	7.1	22 523	7.0
Aggrenox dipyridamole	173	0.6	109	0.7	656	1.7	3007	0.9
Aldosterone agonist	501	1.9	308	2.1	798	2.0	5617	1.7
Aspirin only	14 325	53.6	7799	53.1	22 537	57.4	181 895	56.6
β-blockers	19 325	72.4	11 225	76.4	27 783	70.7	218 634	68.1
Calcium channel blockers	9535	35.7	3165	21.6	13 819	35.2	99 035	30.8
Clopidogrel	3079	11.5	2474	16.8	5062	12.9	45 808	14.3
Digoxin	1773	6.6	700	4.8	2180	5.5	18 524	5.8
Fibrate	738	2.8	382	2.6	1080	2.7	8940	2.8
H2 blocker	18 977	71.1	10 324	70.3	28 079	71.5	178 613	55.6
Heparin (IV)	13 427	50.3	5739	39.1	18 958	48.3	119 692	37.3
Low-molecular-weight heparin	4524	16.9	2665	18.1	7385	18.8	58 217	18.1
Loop diuretic	10 702	40.1	4908	33.4	13 869	35.3	94 220	29.3
Nicotine replacement	683	2.6	349	2.4	1179	3.0	7271	2.3
Oxygen used	17 758	66.5	8290	56.5	23 925	60.9	178 113	55.5
Proton pump inhibitors	10 204	38.2	4795	32.7	14 336	36.5	131 136	40.8
Rhythm control drugs^a	7789	29.2	3262	22.2	10 361	26.4	81 678	25.4
Statin	13 447	50.4	6905	47.0	19 603	49.9	160 982	50.1
Thiazide diuretic	1718	6.4	917	6.2	2632	6.7	21 512	6.7
Varenicline	30	0.1	23	0.2	86	0.2	598	0.2
Vitamin K	627	2.3	249	1.7	1324	3.4	4606	1.4
Warfarin	259	1.0	143	1.0	347	0.9	3283	1.0
Hospital-level characteristics
Geographic region, %
West	8780	32.9	1466	10.0	3566	9.1	49 309	15.4
Midwest	6755	25.3	1848	12.6	9324	23.7	58 098	18.1
Northeast	3025	11.3	1087	7.4	4795	12.2	58 717	18.3
South	8145	30.5	10 284	70.0	21 606	55.0	154 983	48.3
Location of hospital, %
Rural	1769	6.6	3127	21.3	3087	7.9	33 795	10.5
Urban	24 936	93.4	11 558	78.7	36 204	92.1	287 312	89.5
Teaching hospital status, %
No	13 908	52.1	6804	46.3	20 611	52.5	147 270	45.9
Yes	12 797	47.9	7881	53.7	18 680	47.5	173 837	54.1
Number of beds, %
<51	7909	29.6	7605	51.8	15 354	39.1	148 731	46.3
50-99	5346	20.0	1611	11.0	7886	20.1	53 757	16.7
100-199	6837	25.6	1152	7.8	10 415	26.5	59 674	18.6
200-299	4423	16.6	3617	24.6	3692	9.4	37 693	11.7
300-399	969	3.6	610	4.2	1336	3.4	16 028	5.0
400-499	1148	4.3	88	0.6	527	1.3	4505	1.4
>500	73	0.3	2	0.0	81	0.2	719	0.2

Abbreviations: ACE, Angiotensin converting enzyme; ARB, Angiotensin II receptor blocker; CABG, coronary artery bypass graft; ICU, Intensive care unit; IV, intravenous.

^aRhythm control drugs: amiodarone, dronedarone, sotalol, procainamide, or propafenone.

Statistical Analysis

Statistical analyses were performed to describe the preoperative characteristics of patients who did and did not receive corticosteroids, to evaluate temporal trends in perioperative corticosteroid use, and to explore sources of variation in perioperative corticosteroid use. Analyses were performed with SAS v 9.4 (SAS Institute Inc, Cary, North Carolina) statistical software.

Descriptive and temporal trend analysis

We characterized those who did and did not receive corticosteroids based on baseline demographics, clinical conditions, preexisting chronic conditions, perioperative corticosteroid use, and hospital-level characteristics. Trends in annual utilization of corticosteroids were evaluated visually with time-series plots of the proportion of CABG patients who received these agents prior to surgery. Linear time-series analyses were used to model annual rates of perioperative intravenous corticosteroids utilization between 2003 and 2014. Specifically, an autoregressive model was used to correct for potential first-order autocorrelation among annual data points, since we modeled aggregated intravenous corticosteroid use per yearly intervals. The models were further stratified by geographical region to assess trends in the perioperative corticosteroid administration by region.

Between-hospital variation

We plotted the probability of intravenous corticosteroid administration (percentage of CABG patients administered corticosteroids) across hospitals. Multilevel hierarchical models were²⁰ used to quantify the proportion of variation in intravenous corticosteroid use explained by the clustering of patients within hospitals as well as to identify patient- and hospital-level factors associated with intravenous corticosteroid use. Specifically, these models were built with multivariable generalized linear mixed models (PROC GLIMMIX function in SAS v 9.4) with a binary indicator of intravenous corticosteroid use as the dependent variable and hospitals as the clustering variable. Models were built in the following sequential fashion to quantify the incremental explanatory effects of patient- and hospital-level factors: model 1: hospital (random effects intercept) plus calendar year of intravenous corticosteroid use (random effects); model 2: model 1 plus patient demographics (fixed effects); model 3: model 2 plus patient clinical characteristics (fixed effects); model 4: model 3 plus hospital-level factors (fixed effects). The intracluster correlation coefficient (ICC), a measure of how much of the total variation in the probability of intravenous corticosteroid use was accounted for by the hospitals, was estimated from each model as: (intercept variance estimate)/(intercept variance estimate + 3.29 [residual variance]).^21,22 Additionally, we created caterpillar plots of the variation in intravenous corticosteroid use to visualize the extent to which each group of prespecified factors explained the observed variation.

Determinants of perioperative corticosteroid use

We estimated multivariable-adjusted odds ratios (ORs) for the association between intravenous corticosteroid use and each patient- and hospital-level characteristic in model 4 and described above.

Results

We identified and included 401 788 eligible patients who underwent CABG between 2003 and 2014. Of these, 80 681 were administered intravenous methylprednisolone (n = 39 291), dexamethasone (n = 26 705), or hydrocortisone (n = 14 685). We restricted our analyses to these 3 corticosteroids because <1% of patients were administered other intravenous corticosteroids. The median ages of both who received or did not receive intravenous corticosteroid were similar at 66 years (interquartile range: 58-74 years). Table 1 shows patient- and hospital-level characteristics for those who did and did not receive intravenous corticosteroids.

Temporal Trends in Intravenous Corticosteroid Administration

The proportion of patients administered methylprednisolone, dexamethasone, or hydrocortisone intravenously on the day of CABG surgery increased from 17.5% in 2003 to 22.6% in 2014 at an annual rate of 4.2 additional exposed patients per every 1000 CABG surgeries each year (annual rate = 0.42%; 95% confidence interval [CI]: 0.29%-0.55%; P < .001; Figure 2). This increase over time was largely driven by an increase in methylprednisolone use (annual increase = 0.46%; 95% CI: 0.36%-0.50%; P < .001). There was also substantial regional variation in corticosteroid use, with the highest and lowest observed in Midwest (23.6%) and the Northeast regions (13.2%), P < .001, respectively (Supplemental Figure 1).

Figure 2.

Perioperative intravenous corticosteroid utilization over time by total and individual medication use.

Determinants of Intravenous Corticosteroid Administration

Figure 3 shows the multivariable-adjusted OR and 95% CIs for the association between perioperative intravenous corticosteroid administration and preoperative patient- and hospital-level characteristics. Hispanic and black patients, as well as patients of other ethnicities, were more likely to be administered intravenous corticosteroids, as compared to white patients. Males were more likely to be administered these medications than females. The use of several medications during the perioperative period was predictive of intravenous corticosteroid administration: warfarin, statins, aspirin, fibrates, angiotensin-converting enzyme inhibitors, and β-blockers were positively associated with intravenous corticosteroid administration, whereas clopidogrel, heparin, proton pump inhibitors, aspirin/extended-release dipyridamole, digoxin, loop diuretics, vitamin K, heart rhythm control drugs, nicotine replacement therapy, and H2 blockers were inversely associated with intravenous corticosteroid exposure. Prior medical procedures such as oxygen use, valve repair/replacement, intensive care unit (ICU) stay, and previous PCI were inversely associated with use of intravenous corticosteroid. Patients undergoing CABG surgery in hospitals located in the South and Midwest (vs West) were more likely to be administered intravenous corticosteroids, while patients admitted to hospitals in urban areas (vs rural) were less likely to be administered an intravenous corticosteroid.

Figure 3.

Determinants of perioperative intravenous corticosteroid use adjusted for patient clusters within hospitals.

Between-Hospital Variation in Intravenous Corticosteroid Use

Figure 4 shows that there was substantial variability in intravenous corticosteroid administration across the 503 hospitals where patients underwent CABG surgery. Among the upper quartile of hospitals with highest use of intravenous corticosteroid administration during CABG surgery, ≥32.4% of patients were administered an intravenous corticosteroid compared to no more than 3.4% of patients in the lower quartile. From the generalized linear mixed-effects models, hospitals alone, conditioned on calendar year, accounted for over half (ICC = 51%) of all observed variation in perioperative utilization of intravenous corticosteroids between 2003 and 2014. Plots of the variance in intravenous corticosteroid administration (Figure 5) show that the observed variability in intravenous corticosteroid administration attributable to hospitals did not change with the addition of patient demographics (ICC = 51%), clinical conditions (ICC = 52%), and hospital-level (ICC = 52%) characteristics. Hence, about half of the variation in intravenous corticosteroid administration between hospitals could not be explained by patient- and hospital-level characteristics.

Figure 4.

Interhospital variation in perioperative intravenous corticosteroid utilization.

Figure 5.

Panel plot of the proportion of variability in perioperative intravenous corticosteroid utilization explained by hospitals, patient demographic and clinical factors, and hospital characteristics.

Discussion

We observed a modest but steady rise in perioperative administration of intravenous corticosteroid among patients undergoing CABG surgery from 17.5% in 2003 to 22.6% in 2014. Our analyses found that prophylactic perioperative intravenous corticosteroid administration varied considerably across hospitals and by geographical location of these hospitals. Our study also provides insights into patient- and hospital-level characteristics that are significantly associated with the use of intravenous corticosteroids during CABG. However, patient characteristics appear to explain little variation in intravenous corticosteroid administration in CABG above and beyond variation across hospitals and geographic regions.

There may be several explanations for the increase in intravenous corticosteroid use over time. First, there was an increase between 2004 and 2005 that corresponded with the timing of the publication of the 2004 ACC/AHA guidelines for CABG, which recommended use of corticosteroids for the reduction of systemic inflammation. There was another increase between 2009 and 2010, which occurred following the publication of a meta-analysis of randomized trials by Whitlock et al, which suggested that corticosteroids reduced the incidence of atrial fibrillation and postoperative bleeding and reduced duration of ICU stay and length of total hospitalization.¹⁰

Second, we observed differences in trends in intravenous corticosteroid administration across geographic regions. Notably, intravenous corticosteroid administration increased dramatically in the South between 2003 and 2014. Many studies have previously documented wide regional variation in use of surgery and other medical practices, even within fairly small geographic units, such as states.¹⁸ Large differences in corticosteroid use across regions persisted even after accounting for hospital characteristics in multivariable regression models.

Third, about half of the proportion of variation in intravenous corticosteroid administration was accounted for by hospitals alone. Neither the addition of hospital-level nor patient-level characteristics explained much additional variation. A potential explanation for the observed variability in intravenous corticosteroid administration across hospitals may relate to institution-based differences in knowledge of, beliefs about, and preferences toward use of these medications in CABG. In the absence of clear guidelines about the role of intravenous corticosteroids in CABG, different hospitals may make different decisions on the basis of the same evidence in the literature. Providers in different hospitals likely also have different degrees of familiarity with guidelines and published literature. This may explain why the prevalence of intravenous corticosteroid administration was below 34.6% in three-quarters of the hospitals, even after the AHA/ACC endorsed the use of these medications in 2004. Differences in formularies could also explain across-hospital variation. Also, the individual providers can be an important source of variability when there is substantial clustering of patients within providers. For example, different providers could have different experiences with these drugs. However, we could not assess the role of individual surgeons, anesthesiologists, or care teams on corticosteroid use. Other unobserved hospital-level traits, such as institutional practice culture, could also have accounted for the residual unexplained variation in intravenous corticosteroid use. Future studies should explore these factors when assessing variability in perioperative intravenous corticosteroid use across hospitals.

Our findings are strengthened by the large sample of patients undergoing CABG surgery drawn from 503 hospitals across the United States during a 12-year period. This makes our findings generalizable to patients undergoing CABG surgery in the United States. We were able to show that hospitals are a major determinant for perioperative corticosteroid use. This information is important and needs to be factored into the design of future studies evaluating the effect of corticosteroid use on health outcomes. Using inpatient data captured during the perioperative period allowed us to accurately capture prophylactic use of methylprednisolone, dexamethasone, and hydrocortisone around the time of CABG surgery. We were also able to measure and examine a large number of potential predictors of corticosteroid use.

Our findings are limited by the unavailability of some important potential sources of variation in perioperative intravenous corticosteroid, such as formulary and prescriber preferences. The Premier Healthcare Database does not contain information on the reason for perioperative corticosteroid use of these corticosteroids. Theoretically, it is possible that some of the observed intravenous corticosteroid use may have been for underlying clinical issues. Also, given that we tested for statistical significance for the associations between several potential predictors and perioperative corticosteroid administration, it is possible that some of the observed statistically significant associations were due to chance. Finally, the Premier database contains only inpatient data. As such, we were not able to fully capture comorbid conditions that may have been present upon admission but were not recorded among the discharge diagnoses.

Conclusion

Perioperative corticosteroid use among patients undergoing CABG surgery in the United States increased steadily between 2003 and 2014. Utilization differed by geographic region. Individual hospital was identified as the single major source of the variation, accounting for more than half of the observed variation. Nearly a quarter of patients undergoing CABG in the United States now receive perioperative intravenous corticosteroids. Given the substantial variation in practice patterns, we are currently conducting a large-scale study to assess the associations between the perioperative prophylactic use of corticosteroids and multiple postoperative safety end points.

Supplemental Material

Supplemental Material, Supplemental_data - Trends and Patterns of Corticosteroid Use During Coronary Artery Bypass Grafting Surgery in the United States

Supplemental Material, Supplemental_data for Trends and Patterns of Corticosteroid Use During Coronary Artery Bypass Grafting Surgery in the United States by Macarius M. Donneyong, Alexander Kulik, and Joshua J. Gagne in Journal of Cardiovascular Pharmacology and Therapeutics

Footnotes

Appendix A

List of Covariates

Covariates	Codes (Charge/ICD-9 diagnosis)	Period Assessed
Diabetes drugs		CAP
COPD drugs		CAP
Antibiotics		CAP
Date of hospital admission
Age		Day of CABG
Sex		CAP
Race		CAP
Marital status		CAP
Region		CAP
Year		Day of CABG
Indicator for hospital day of CABG		Day of CABG
Hospital ID		Day of CABG
IV hydrocortisone		Day of CABG
IV methylprednisolone		Day of CABG
IV dexamethasone		Day of CABG
History of CABG	ICD-9 diagnosis = V4581	CAP
Valve repair or replacement on day of surgery		Day of CABG
On-pump surgery		Day of CABG
Off-pump surgery		Day of CABG
Chronic renal disease or end-stage renal disease (hemodialysis, peritoneal dialysis, erythropoietin, and darbepoetin)		CAP
PCI (PTCA or stent) on the index admission		CAP
Rhythm-control drugs (amiodarone, dronedarone, sotalol, procainamide, and propafenone)		CAP
Calcium channel blocker		CAP
β-blocker		CAP
Thiazide diuretic		CAP
Loop diuretic		CAP
ACE inhibitor		CAP
Angiotensin receptor blocker		CAP
Warfarin		CAP
IV or low-molecular-weight heparin (bemiparin, nadroparin, reviparin, enoxaparin, parnaparin, certoparin, dalteparin, and tinzaparin)		CAP
Proton pump inhibitors (charge codes available in spreadsheet)		CAP
H2 blockers (charge codes available in spreadsheet)		CAP
Vitamin K (charge codes available in spreadsheet)		CAP
Aldosterone antagonist		CAP
Digoxin		CAP
Aspirin		CAP
Clopidogrel		CAP
Dipyridamole		CAP
Statin		CAP
Fibrate		CAP
Corticosteroid		CAP
Telemetry		CAP
Echocardiogram		CAP
Oxygen use		CAP
Intensive nursing care		CAP
Intensive physician visits		CAP
Nicotine replacement therapy		CAP
Varenicline		CAP

Abbreviations: ACE, angiotensin-converting enzyme; CABG, coronary artery bypass graft; CAP, covariates assessment period; COPD, chronic obstructive pulmonary disease; ICD, International Classification of Diseases; IV, intravenous; PCI, percutaneous intervention; PTCA, percutaneous transluminal coronary angioplasty.

^aCovariates assessment period was defined as the period between hospital admission date to date of CABG (inclusive). Also assess and define covariates occurring only on day of CABG.

Authors’ Note

Dr Gagne is principal investigator of an investigator initiated, unrestricted research grant from Novartis Pharmaceuticals Corporation to the Brigham and Women’s Hospital for unrelated work.

Author Contributions

M. Donneyong contributed to conception and design; contributed to acquisition, analysis, and interpretation; drafted the manuscript; critically revised the manuscript; and gave final approval. A. Kulik contributed to design, contributed to interpretation, drafted the manuscript, critically revised the manuscript, and gave final approval. J. Gagne contributed to conception and design; contributed to acquisition, analysis, and interpretation; drafted the manuscript; critically revised the manuscript; and gave final approval. All authors agree to be accountable for all aspects of work ensuring integrity and accuracy.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr Gagne is a consultant to Aetion, Inc and to Optum, Inc.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Supplemental Material

Supplementary material for this article is available online.

References

Mozaffarian

Roger

; American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics—2014 update: a report from the American Heart Association. Circulation. 2014;129(3):e28–e292. doi:10.1161/01.cir.0000441139.02102.80.

Sawa

Shimazaki

Kadoba

. Attenuation of cardiopulmonary bypass-derived inflammatory reactions reduces myocardial reperfusion injury in cardiac operations. J Thorac Cardiovasc Surg. 1996;111(1):29–35.

Kharazmi

Andersen

Baek

. Endotoxemia and enhanced generation of oxygen radicals by neutrophils from patients undergoing cardiopulmonary bypass. J Thorac Cardiovasc Surg. 1989;98(3):381–385.

Kirklin

Westaby

Blackstone

. Complement and the damaging effects of cardiopulmonary bypass. J Thorac Cardiovasc Surg. 1983;86(3):845–857.

Butler

Parker

Pillai

Westaby

Shale

Rocker

. Effect of cardiopulmonary bypass on systemic release of neutrophil elastase and tumor necrosis factor. J Thorac Cardiovasc Surg. 1993;105(1):25–30.

Rinder

Bonan

Rinder

Mathew

Hines

Smith

. Cardiopulmonary bypass induces leukocyte-platelet adhesion. Blood. 1992;79(5):1201–1205.

Casey

. Role of cytokines in the pathogenesis of cardiopulmonary-induced multisystem organ failure. Ann Thorac Surg. 1993;56(5 suppl):S92–S96.

Windsor

Mullen

Fowler

Sugerman

. Role of the neutrophil in adult respiratory distress syndrome. Br J Surg. 1993;80(1):10–17.

Cappabianca

Rotunno

Schinosa

. Protective effects of steroids in cardiac surgery: a meta-analysis of randomized double-blinded trials. J Cardiothorac Vasc Anesth. 2011;25(1):156–165.

10.

Whitlock

Chan

Devereaux

. Clinical benefit of steroid use in patients undergoing cardiopulmonary bypass: a meta-analysis of randomized trials. Eur Heart J. 2008;29(21):2592–2600.

11.

Wan

LeClerc

Vincent

. Inflammatory response to cardiopulmonary bypass. Chest. 1997;112(3):676–692.

12.

Weis

Beiras-Fernandez

Schelling

. Stress doses of hydrocortisone in high-risk patients undergoing cardiac surgery: effects on interleukin-6 to interleukin 10 ratio and early outcome. Crit Care Med. 2009;37(5):1685–1690.

13.

Ali-Hassan-Sayegh

Mirhosseini

Haddad

. Protective effects of corticosteroids in coronary artery bypass graft surgery alone or combined with valvular surgery: an updated and comprehensive meta-analysis and systematic review. Interact Cardiovasc Thorac Surg. 2015;20(6):825–836.

14.

Liakopoulos

Schmitto

Kazmaier

. Cardiopulmonary and systemic effects of methylprednisolone in patients undergoing cardiac surgery. Ann Thorac Surg. 2007;84(1):110–118.

15.

Chaney

Durazo-Arvizu

Nikolov

Blakeman

Bakhos

. Methylprednisolone does not benefit patients undergoing coronary artery bypass grafting and early tracheal extubation. J Thorac Cardiovasc Surg. 2001;121(3):561–569.

16.

Tan

. Benefits and risks of corticosteroid prophylaxis in adult cardiac surgery: a dose-response meta-analysis. Circulation. 2009;119(14):1853–1866.

17.

Eagle

Guyton

Davidoff

. ACC/AHA 2004 guideline update for coronary artery bypass graft surgery: summary article: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Update the 1999 Guidelines for Coronary Artery Bypass Graft Surgery). Circulation. 2004;110(9):1168–1176.

18.

Fleisher

Fleischmann

Auerbach

. 2014 ACC/AHA guideline on perioperative cardiovascular evaluation and management of patients undergoing noncardiac surgery: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines. Circulation. 2014;130(24):e278–e333.

19.

Premier, Inc. Premier Healthcare Database: Data that Informs and Performs. Charlotte, NC: Premier, Inc; 2017.

20.

Gelman

. Data Analysis Using Regression and Multilevel/Hierarchical Models. New York, NY: Cambridge University Press; 2007.

21.

Snijders

Bosker

. Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling. Thousand Oaks, CA: SAGE; 1999.

22.

Long

Freese

. Regression Models for Categorical Dependent Variables Using Stata. 2nd ed. College Station, TX: Stata Press; 2006.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.12 MB