Patient Outcomes Associated with 2-Octyl Cyanoacrylate Topical Skin Adhesive in Coronary Artery Bypass Graft Surgery

Abstract

Background:

Surgical site infection (SSI) is an important cause of morbidity and occasionally death after coronary artery bypass graft (CABG) surgery. These infections also are associated with higher costs and poorer surgical outcomes. We used a retrospective observational database to examine and quantify the effects of the topical skin adhesive 2-octyl cyanoacrylate, used as the final layer of site closure in patients undergoing CABG surgery, on the incidence of post-operative SSI.

Methods:

All patients in the Premier Perspective™ Comparative Database of inpatient hospitalizations who underwent CABG surgery in 2005 and 2006 were identified. Qualifying patients were classified into four groups according to the method of surgical site closure on the basis of detailed hospital billing charges: Sutures only; sutures and 2-octyl cyanoacrylate; sutures and staples; and sutures, 2-octyl cyanoacrylate, and staples. Site infections were identified by a combination of International Classification of Diseases (ICD-9-CM) diagnosis codes, patterns of post-operative use of antibiotics, and All Patient Refined Diagnostic Related Group (APR-DRG) and Diagnostic Related Group (DRG) codes indicating post-operative infections at hospital re-admission.

Results:

A total of 59,006 patients qualified for the study: 38,799 who had sutures only; 10,262 having sutures and 2-octyl cyanoacrylate; 8,180 having sutures and staples; and 1,765 having sutures, 2-octyl cyanoacrylate, and staples. The lowest unadjusted rate of post-CABG SSI was found in the sutures and 2-octyl cyanoacrylate group (4.3%; 95% confidence interval [CI] 3.9–4.7), followed by sutures only (5.3%; 95% CI 5.1–5.5); sutures and staples (6.2%; 95% CI 5.7–6.8); and sutures, staples, and 2-octyl cyanoacrylate (7.1%; 95% CI 6.0%–8.4%). A logistic regression model that controlled for selected baseline patient, hospital, and surgical characteristics showed significantly lower rates of post-CABG SSI (odds ratio 0.76; 95% CI 0.68–0.85) in patients closed with sutures and 2-octyl cyanoacrylate relative to patients who had only sutures.

Conclusions:

The observed rates of post-CABG SSI were consistent with the rates observed in the literature. The SSI rate for patients who had sutures and 2-octyl cyanoacrylate used as the final layer of site closure was significantly lower than the rates for patients having other types of closure.

Surgical site infection (SSI) is an important cause of morbidity and occasionally death after coronary artery bypass graft (CABG) surgery [1,2]. Such infections also are associated with higher costs and poorer patient outcomes. It is important to quantify the risk of SSIs in patients undergoing CABG surgery and the clinical and economic impact of hospitalizations and re-hospitalizations [3].

DERMABOND^® Topical Skin Adhesive is a synthetic 2-octyl cyanoacrylate tissue adhesive approved by the U.S. Food and Drug Administration for skin closure. Several advantages of 2-octyl cyanoacrylate have been observed in clinical studies, including shorter operative time, formation of a barrier to microbial penetration as long as the adhesive film remains intact, simplified incision care, no need for suture removal, and better scar appearance. These benefits suggest that 2-octyl cyanoacrylate is an effective option for surgical site closure. This study used a retrospective observational database to examine and quantify the independent, real-world, clinical and economic effects of 2-octyl cyanoacrylate on the incidence of post-operative SSI when used as the final layer of site closure in patients undergoing CABG surgery. Observational studies such as those via automated claims databases provide an effective means for identifying rates and differences of clinical outcomes and adverse events across products, as well as evaluating variables associated with these events. In the area of SSI, such data present methodologic challenges that, when understood and addressed, can contribute to medical decision-making [4].

Materials and Methods

Data source

Data for this study were extracted from the Premier Perspective™ Comparative Database (PCD). The PCD is the largest U.S. hospital, clinical, and economic database developed for quality and utilization benchmarking. Annually, more than five million hospital discharges are processed and recorded in the database. For these discharges, the PCD contains more than five billion daily service records, and about 65 million service records are added each month. In addition to the data elements available in most of the standard hospital discharge files, the PCD contains a date-stamped log of all billed items, including procedures, medications, laboratory, and diagnostic and therapeutic services for individual patients. The database provides the principal discharge diagnosis for each hospital admission; secondary diagnosis codes also may be recorded, depending on the hospital. Hospitals submit data to the PCD quarterly or monthly. The data undergo quality checks, and cost information is reconciled with the hospitals' financial statements. The data are used by hospitals to benchmark their clinical and financial performance against that of their peers. The quality of the data is enhanced by the intense scrutiny they undergo daily by the hospital-based users.

Each hospitalization in the Premier database has a unique identification number that is encrypted to protect confidentiality when the files are released for research. Multiple hospitalizations for the same patient can be linked by another unique identification number when the hospitalizations were at the same facility. The data extracted from all hospitalizations were linked using these two patient identifiers.

Patient Identification

All patients in the Premier hospital database who underwent CABG surgery from January 1, 2005, through December 31, 2006, were identified by a procedure code indicating CABG (International Classification of Diseases, 9^th revision, Clinical Modification [ICD-9-CM] codes 36.10 to 36.2). The codes used for cohort selection are summarized in Table 1. Although CABG operations also can be identified in the Premier data by Current Procedural Terminology, 4^th edition (CPT4) codes and billing charges specific to CABG surgery, a detailed examination of the Premier data indicated that the ICD-9-CM procedure codes identified all eligible patients.

Table 1.

Measures Used for Identification of Surgical Site Infection after Coronary Artery Bypass Grafting in Premier Data

Indicator	Values	Initial hospitalization	Re-hospitalizations up to 2 months after surgery
Pharmacy billing charges for oral and parenteral antibiotics	Aminoglycosides (oral and parenteral) Anti-infectives systemic misc (oral and parenteral) Cephalosporins and related (oral) Macrolides and related (oral) Penicillins (oral and parenteral) Quinolones (oral and parenteral) Tetracyclines (oral and parenteral)	New antibiotic regimen from day 9 to day 60 after CABG surgery	Not used
ICD-9-CM Diagnosis Code	998.5x—Post-operative infection	Used	Used
DRG codes	415—Operating room procedure for infectious and parasitic diseases 418—Post-operative and post-traumatic infections	Not used	In-patient only
APR-DRG codes	711—Procedures for post-operative and post-traumatic infection 721—Post-operative and post-traumatic infection	Not used	In-patient only

APR-DRG=All Patient Refined – Diagnosis Related Groups; DRG=Diagnosis Related Groups; ICD-9-CM=The International Classification of Diseases, Ninth Revision, Clinical Modification.

Additional inclusion and exclusion criteria were used to define the study population. The inclusion criteria were as follows:

• Patient age ≥18 years;

• Inpatient hospitalization;

• CABG surgery between hospitalization days 1 and 14;

• Appropriate All Patient Refined Diagnostic Related Group (APR-DRG) codes (165, 166) and Diagnostic Related Group (DRG) codes (106–109, 547–550) for CABG.

Because post-operative prescription of antibiotics was used as an indicator of SSI, patients observed to have infections other than SSI that were treated with antibiotics were excluded from the analyses. Patients also were excluded if they had any indication of the following:

• Childbirth, obstetric, newborn, or pediatric care;

• Admission to the hospital from a trauma center or an unknown source;

• Additional non-cardiovascular operations that could provide sources of infection;

• Secondary non-SSI post-operative infections treated with antibiotics, such as pneumonia and urinary tract infections.

Patients also were excluded from the analysis for the following:

• Billing charges for Indermil^® tissue adhesive (Covidien, Norwalk, CT);

• Inconsistent or missing relevant data in the Premier database.

Baseline characteristics and co-morbidities

The study population was characterized by baseline demographics found in the Premier database, including age, sex, race, geographic region, and primary payer. Co-morbidity information is not available directly in the Premier database because physician office visits and outpatient prescriptions are not included: The database includes only diagnostic, procedural, and billing information recorded during the hospitalization. Chronic diseases and other co-morbid conditions at baseline can be identified only if the patient was treated for those conditions during the CABG hospitalization or during a prior hospital stay.

Many studies have identified several co-morbid conditions, especially diabetes mellitus and obesity, as significant risk factors for poor surgical site healing and post-CABG SSI [1,5], so identifying these risk factors was essential. To help identify patient risk factors, all inpatient and outpatient hospitalizations within three months prior to the CABG hospitalization were identified, and these data were analyzed for indications of relevant risk factors.

Given the nature of the Premier data, the absence of co-morbidity indicators during specific hospitalizations does not guarantee the absence of the co-morbidity in the patient. However, an assumption was made that this method would be sufficient to capture risk factors in this cohort of CABG patients because the serious nature of their surgery would warrant the inclusion of relevant co-morbidity data. An examination of the resulting prevalence of co-morbidities was conducted by a cardiac surgeon, who found that the results were consistent with the expected rates in the general CABG surgical population and relevant published literature.

Diabetic status was determined by looking at both patient ICD-9-CM diagnosis codes (250.x) and treatment with oral hypoglycemic agents, such as sulfonylureas, thiazolidinediones, biguanides, and meglitinides, during the CABG hospitalization or any prior hospitalizations. Insulin use during any hospitalization prior to the CABG surgery also was used to define diabetic status. However, insulin use only during the CABG hospitalization was not considered an indication of diabetes mellitus because of its use in glycemic control in both diabetic and non-diabetic patients undergoing CABG procedures.

Obesity also is a significant risk factor for post-CABG SSI, but Premier billing data do not contain the patient's weight or body mass index. The Premier data were examined to find methods of identifying obese patients, such as ICD-9-CM diagnosis codes for overweight and obesity (278.00 to 278.02) or billing for supplies and treatments specific to obese patients.

Hospital characteristics

The characteristics of the hospitals in which the patients were treated were determined. These characteristics were:

• Size (number of beds);

• Teaching or non-teaching;

• Location (urban/rural);

• Derived annual CABG volume (number of CABG procedures per year).

Patient hospital-admitting characteristics included admission type (emergency, urgent, elective) and transfer from another hospital.

Surgical risk factors

Characteristics of both the CABG surgery and the immediate post-operative care have been identified as significant risk factors for post-operative infections [1,5]. The surgical risk factors include the duration of surgery, the number of grafts, on-pump/off-pump surgery, and surgical re-exploration for bleeding. The Premier data were not able to differentiate between endoscopic and open vein harvesting, so the method of harvesting could not be included as a surgical risk factor for SSI.

Exposure definitions

Within the CABG cohort, patients were classified into exposure groups by the method of surgical site closure:

• Sutures only;

• Sutures and 2-octyl cyanoacrylate;

• Sutures and staples;

• Sutures, 2-octyl cyanoacrylate, and staples.

Sutures are the standard of care for site closure after CABG surgery, and the sutures and 2-octyl cyanoacrylate group was the reference used in analyses.

Patients were assigned to these exposure groups on the basis of itemized central supply billing charges for staples and 2-octyl cyanoacrylate on the day of the CABG operations. Sutures are a low-cost item, and approximately 25% of CABG patients did not have a specific itemized billing charge for sutures on the day of their surgery, although many of these patients did have itemized billing charges for staples or 2-octyl cyanoacrylate. An assumption was made that the suture billing charges may have been absorbed in a general supply charge for these patients, so CABG patients did not need to have an itemized supply charge for sutures. We conducted a sensitivity analysis, which included only patients who had an itemized billing charge for sutures on the day of CABG surgery, and the results (not presented) did not differ from these results, where the assumption was made that all CABG patients had sutures.

Patients undergoing CABG surgery usually have two or more incision sites: A sternal incision and one or more incisions in the leg or other site for vein harvesting. Different closure techniques can be used at the different sites, and the Premier data do not differentiate between methods of surgical closure at the various sites. Because we were unable to differentiate among closure techniques at the two incision sites, patients were classified into one closure group on the basis of the billing charges on the day of surgery.

Outcomes

The primary outcome measure was the occurrence of an SSI. The methodology described by Yokoe et al. was used to identify post-operative infections [5]. This study evaluated antibiotic use following three separate surgical procedures, including CABG, and concluded that post-operative use of antibiotics could provide a reliable way to identify SSIs.

Surgical site infections were identified by ICD-9-CM diagnosis codes that suggest such infections (998.5, 998.51, and 998.59) during the initial hospital admission for the CABG procedure or during any re-admission to the same hospital within two months. The DRG and APR-DRG codes indicating post-operative infections assigned to hospital re-admissions also were used as an indication of post-CABG SSI.

Patient exposure to antimicrobial drugs initiated 9–60 days post-operatively during the initial hospital admission also was used to identify patients with post-operative SSI. Use of antimicrobial drugs during hospital re-admissions was not used because we were unable to determine when and why antimicrobial use was initiated. The DRG and APR-DRG codes associated with hospital re-admissions were employed to identify post-CABG SSI for in-patient re-hospitalizations. The methods for identifying post-CABG SSI are summarized in Table 1.

Additional outcomes included hospitalization length of stay (LOS), intensive care unit (ICU) and routine care days, additional CABG surgery, hospital discharge disposition, hospital re-admission, and cost of CABG hospitalization. Mediastinitis (ICD-9-CM diagnosis code 519.2) was considered for analysis but was excluded because only six patients in the study cohort had this medical diagnosis assigned.

Analyses

To determine whether there were any differences in the types of patients who were closed with each of the techniques, we summarized each of the baseline, hospital, and surgery and outcomes separately for patients with the different types of site closures. The patients in the suture and 2-octyl cyanoacrylate group were compared with each of the other closure groups on all baseline characteristics. Continuous variables are presented as means with standard deviations and medians with ranges, and categorical variables are presented as counts and percentages. The closure group pairs were compared on all baseline characteristics, using t-tests for continuous variables and χ² tests for categorical variables. Significance was determined by a two-tailed p value (<0.05 for all tests).

Multivariate logistic regression analysis was used to determine the predictors of post-CABG SSI. The model included site closure techniques and confounders from among baseline patient, hospital, and surgical characteristics.

Results

A total of 59,006 patients qualified for the study: 10,262 closed with sutures and 2-octyl cyanoacrylate; 38,799 with sutures only; 8,180 with sutures and staples; and 1,765 with sutures, staples, and 2-octyl cyanoacrylate. Overall, there were few differences among the groups Table 2. There were no significant differences in age. The two closure groups that included staples had a higher percentage of female patients (30.1% for sutures and staples; 30.0% for sutures, staples, and 2-octyl cyanoacrylate) than the groups that did not include staples (26.1% for sutures and 2-octyl cyanoacrylate; 25.5% for sutures only). A higher percentage of the 2006 CABG surgeons used 2-octyl cyanoacrylate than those operating in 2005, and the highest percentages of procedures that used staples were found in the South. Patients in either of the two closure groups with staples (staples and sutures; staples, sutures, and 2-octyl cyanoacrylate) had overall higher rates of co-morbid conditions: Diabetes mellitus (sutures and staples), obesity (sutures), chronic obstructive pulmonary disease (sutures and staples), kidney disease (sutures and staples), and congestive heart failure (sutures and staples).

Table 2.

Baseline Characteristics of Exposure Groups

Patient characteristics	Sutures+2-octyl cyanoacrylate skin adhesive (n=10,262)	Sutures only (n=38,799)	Sutures+staples (n=8,180)	Sutures+staples+2-octyl cyanoacrylate skin adhesive (n=1,765)
Age
Mean (SD)	64.58 (10.72)	64.71 (10.73)	64.76 (10.86)	64.31 (10.87)
Median (range)	65.00 (25.00–89.00)	65.00 (18.00–89.00)	65.00 (23.00–89.00)	65.00 (24.00–89.00)
Age category, years (%)
<55	1,891 (18.4)	7,023 (18.1)	1,515 (18.5)	317 (18.0)
55–64	3,140 (30.6)	11,741 (30.3)	2,368 (28.9)	557 (31.6)
65–74	3,139 (30.6)	12,078 (31.1)	2,572 (31.4)	545 (30.9)
>74	2,092 (20.4)	7,957 (20.5)	1,725 (21.1)	346 (19.6)
Gender (%)
Male	7,582 (73.9)	28,898 (74.5)	5,720 (69.9)^a	1,235 (70.0)^a
Female	2,680 (26.1)	9,901 (25.5)	2,460 (30.1)	530 (30.0)
Race (%)
White	7,385 (72.0)	28,839 (74.3)^a	6,357 (77.7)^a	1,298 (73.5)^a
Black	683 (6.7)	2,056 (5.3)	703 (8.6)	221 (12.5)
Hispanic	202 (2.0)	1,426 (3.7)	170 (2.1)	23 (1.3)
Other	1,992 (19.4)	6,478 (16.7)	950 (11.6)	223 (12.6)
Year of CABG (%)
2005	4,833 (47.1)	18,906 (48.7)^a	4,372 (53.4)^a	768 (43.5)^a
2006	5,429 (52.9)	19,893 (51.3)	3,808 (46.6)	997 (56.5)
Geographic region (%)
Northeast	1,655 (16.1)	6,116 (15.8)^a	833 (10.2)^a	104 (5.9)^a
South	5,377 (52.4)	20,238 (52.2)	5,296 (64.7)	1,228 (69.6)
Midwest	1,264 (12.3)	6,694 (17.3)	1,575 (19.3)	281 (15.9)
West	1,966 (19.2)	5,751 (14.8)	476 (5.8)	152 (8.6)
Primary payer (%)
Medicare	5,126 (50.0)	19,772 (51.0)^a	4,262 (52.1)^a	854 (48.4)
Medicaid	407 (4.0)	1,644 (4.2)	322 (3.9)	61 (3.5)
Private	3,923 (38.2)	14,637 (37.7)	2,941 (36.0)	689 (39.0)
Other	806 (7.9)	2,746 (7.1)	655 (8.0)	161 (9.1)
Coronary disease (%)
MI all	2,870 (28.0)	10,454 (26.9)^a	2,325 (28.4)	498 (28.2)
MI admitting or primary diagnosis	2,526 (24.6)	9,163 (23.6)^a	2,055 (25.1)	421 (23.9)
MI secondary diagnosis	503 (4.9)	1,841 (4.7)	407 (5.0)	106 (6.0)
Other non-MI coronary disease	5,582 (54.4)	21,895 (56.4)^a	4,621 (56.5)^a	901 (51.0)^a
Co-morbidities (%)
Diabetes mellitus	4,135 (40.3)	15,597 (40.2)	3,337 (40.8)	765 (43.3)^a
Obesity	1,683 (16.4)	6,425 (16.6)	1,459 (17.8)^a	270 (15.3)
COPD	2,509 (24.4)	8,912 (23.0)^a	1,888 (23.1)^a	497 (28.2)^a
Hypertension	8,986 (87.6)	33,984 (87.6)	7,312 (89.4)^a	1,600 (90.7)^a
Kidney disease	1,141 (11.1)	4,220 (10.9)	904 (11.1)	233 (13.2)^a
Congestive heart failure	1,653 (16.1)	6,002 (15.5)	1,416 (17.3)^a	400 (22.7)^a
Peripheral vascular disease	1,446 (14.1)	5,640 (14.5)	1,225 (15.0)	274 (15.5)
Connective tissue disease	140 (1.4)	528 (1.4)	97 (1.2)	20 (1.1)
Cancer	399 (3.9)	1,485 (3.8)	346 (4.2)	74 (4.2)
Prior hospitalization (%)
No	7,317 (71.3)	28,547 (73.6)^a	5,800 (70.9)	1,272 (72.1)
Yes	2,945 (28.7)	10,252 (26.4)	2,380 (29.1)	493 (27.9)

Significantly different from suture and 2-octyl cyanoacrylate group.

CABG=coronary artery bypass graft; COPD=chronic obstructive pulmonary disease; MI=myocardial infarction; SD=standard deviation.

Additional analyses were conducted to understand better the characteristics of the relatively small number of patients who had 2-octyl cyanoacrylate and staples with sutures for site closure, and no significant predictors of closure type were identified (data not shown).

Table 3 presents the hospital, admitting, and surgical characteristics by site closure group. Although the hospital characteristics show statistically significant differences in the different closure groups, no obvious trend was observed across hospitals. The closure groups with the highest percentage of urban hospital locations included 2-octyl cyanoacrylate, and additional analysis of the data (not shown) indicated that 10% of all CABG operations in rural hospitals included 2-octyl cyanoacrylate compared with 21% of all procedures in urban hospitals.

Table 3.

Hospital, Admitting, and Surgical Characteristics

Hospital characteristics	Sutures+2-octyl cyanoacrylate skin adhesive (n=10,262)	Sutures only (n=38,799)	Sutures+staples (n=8,180)	Sutures+staples+2-octyl cyanoacrylate skin adhesive (n=1,765)
Beds (%)
50–249	1,284 (12.5)	3,458 (8.9)^a	908 (11.1)^a	273 (15.5)^a
250–-499	5,647 (55.0)	16,014 (41.3)	2,739 (33.5)	611 (34.6)
500–749	1,990 (19.4)	12,020 (31.0)	2,837 (34.7)	652 (36.9)
750–999	1,327 (12.9)	7,128 (18.4)	1,553 (19.0)	172 (9.7)
≥1,000	14 (0.1)	179 (0.5)	143 (1.7)	57 (3.2)
Teaching hospital (%)
No	5,145 (50.1)	15,780 (40.7)^a	4,274 (52.2)^a	1,040 (58.9)^a
Yes	5,117 (49.9)	23,019 (59.3)	3,906 (47.8)	725 (41.1)
Hospital location (%)
Urban	9,769 (95.2)	34,984 (90.2)^a	7,613 (93.1)^a	1,727 (97.8)^a
Rural	493 (4.8)	3,815 (9.8)	567 (6.9)	38 (2.2)
Annual CABG volume (%)
Low (<250/year)	6,494 (63.3)	20,224 (52.1)^a	4,686 (57.3)^a	1,082 (61.3)^a
Mid (250–499/year)	2,867 (27.9)	12,305 (31.7)	2,977 (36.4)	669 (37.9)
High (>500/year)	901 (8.8)	6,270 (16.2)	517 (6.3)	14 (0.8)
Admission type (%)
Emergency	2,809 (27.4)	9,950 (25.6)^a	2,206 (27.0)^a	542 (30.7)^a
Urgent	2,615 (25.5)	10,446 (26.9)	1,834 (22.4)	331 (18.8)
Elective	4,838 (47.1)	18,403 (47.4)	4,140 (50.6)	892 (50.5)
Transfer from another hospital (%)
No	8,609 (83.9)	32,213 (83.0)^a	6,661 (81.4)^a	1,536 (87.0)^a
Yes	1,636 (15.9)	6,354 (16.4)	1,316 (16.1)	228 (12.9)
Unknown	17 (0.2)	232 (0.6)	203 (2.5)	1 (0.1)
Characteristics of surgery (%)
On pump	7,804 (76.0)	31,423 (81.0)^a	7,127 (87.1)^a	1,511 (85.6)^a
With catheterization	6,083 (59.3)	22,370 (57.7)^a	4,890 (59.8)	1,087 (61.6)
Internal mammary artery graft: single	8,770 (85.5)	32,934 (84.9)	6,604 (80.7)^a	1,408 (79.8)
Internal mammary artery graft: double	422 (4.1)	1,177 (3.0)^a	334 (4.1)	54 (3.1)
Number of vessels (%)
One	1,362 (13.3)	5396 (13.9)^a	1,205 (14.7)^a	221 (12.5)^a
Two	3,523 (34.3)	13,240 (34.1)	2,851 (34.9)	570 (32.3)
Three	3,333 (32.5)	12,204 (31.5)	2,479 (30.3)	628 (35.6)
Four or more	1,708 (16.6)	6,163 (15.9)	1,328 (16.2)	303 (17.2)
Unknown	336 (3.3)	1,796 (4.6)	317 (3.9)	43 (2.4)
Duration of surgery, hours
Mean (SD)	4.88 (2.20)	4.81 (2.02)^a	4.91 (2.00)	4.77 (1.99)^a
Median (range)	4.83, 0.00–156.25	4.75, 0.00–110.00	4.75, 0.00–30.52	4.92, 0.00–17.50
<1 (%)	357 (3.5)	1,719 (4.4)^a	392 (4.8)^a	113 (6.4)^a
1–<4 (%)	1,464 (14.3)	8,088 (20.8)	1,553 (19.0)	383 (21.7)
4–<7 (%)	7,630 (74.4)	25,512 (65.8)	5,333 (65.2)	1,063 (60.2)
7–<10 (%)	739 (7.2)	3,144 (8.1)	740 (9.0)	180 (10.2)
10–<13 (%)	58 (0.6)	262 (0.7)	119 (1.5)	25 (1.4)
≥13 (%)	14 (0.1)	74 (0.2)	43 (0.5)	1 (0.1)
Hospitalization day of CABG surgery
Mean (SD)	2.93 (2.28)	2.88 (2.22)	3.07 (2.39)^a	3.09 (2.39)^a
Median (range)	2.00 (1.00–14.00)	2.00 (1.00–14.00)	2.00 (1.00–14.00)	2.00 (1.00–14.00)

Significantly different from sutures and 2-octyl cyanoacrylate group.

CABG=coronary artery bypass graft; SD=standard deviation.

The characteristics of the CABG operations by closure group are shown in Table 3. The differences among surgery characteristics were found more in a comparison of closure with and without staples than a comparison of groups with and without 2-octyl cyanoacrylate. A higher percentage of on-pump operations involved staples (87.1% for sutures and staples; 85.6% for sutures, staples, and 2-octyl cyanoacrylate) than did not (76.0% for sutures and 2-octyl cyanoacrylate; 81.0% for sutures only). A CABG with a single internal mammary artery graft was slightly less common in the groups closed with staples (80.7% and 79.8%) than in the groups without staples (85.5% and 84.9%).

The rates and corresponding 95% binomial confidence intervals of post-CABG SSIs by wound closure group are presented in Table 4. The lowest rate of post-CABG SSI was found in the suture and 2-octyl cyanoacrylate group (4.3%; 95% confidence interval [CI] 3.9–4.7), followed by sutures only (5.3%; 95% CI 5.1–5.5); sutures and staples (6.2%; 95% CI 5.7–6.8); and sutures, staples, and 2-octyl cyanoacrylate (7.1%; 95% CI 6.0–8.4). The rate of post-CABG SSI for the suture and 2-octyl cyanoacrylate group was significantly lower than that in the other three treatment groups. The difference in the infection rates for the sutures and staples group and the sutures, staples, and 2-octyl cyanoacrylate group was not statistically significant.

Table 4.

Surgical Site Infections and Other Outcomes of Coronary Artery Bypass Grafting

Outcome	Sutures+2-octyl cyanoacrylate skin adhesive (n=10,262)	Sutures only (n=38,799)	Sutures+staples (n=8,180)	Sutures+staples+2-octyl cyanoacrylate skin adhesive (n=1,765)
Surgical site infection (%)
Yes	439 (4.3)	2,056 (5.3)^a	508 (6.2) ^a	126 (7.1)^a
95% binomial confidence interval (%)	3.9–4.7	5.1–5.5	5.7–6.8	6.0–8.4
Total hospitalization days
Mean (SD)	8.00 (4.20)	8.19 (4.70)^a	8.63 (5.18)^a	8.98 (5.36)^a
95% confidence interval	7.92–8.08	8.15–8.24	8.52–8.74	8.73–9.23
Median (range)	7.0 (1.0–64.0)	7.0 (1.0–117.0)	7.0 (1.0–105.0)	8.0 (1.0–74.0)
Intensive care unit days
Mean (SD)	3.02 (2.87)	2.80 (3.10)^a	3.59 (3.66)^a	4.19 (4.24)^a
95% confidence interval	2.97–3.08	2.77–2.84	3.51–3.67	3.99–4.39
Median (range)	2.0 (0.0–35.0)	2.0 (0.0–87.)	2.00 (0.0–56.0)	3.00 (0.0–67.0)
Routine care days
Mean (SD)	4.98 (3.42)	5.39 (3.83)^a	5.04 (4.04)	4.79 (3.97)
95% confidence interval	4.91–5.04	5.35–5.43	4.95–5.13	4.61–4.98
Median (range)	4.0 (0.0–42.0)	5.0 (0.0–108.0)	4.00 (0.0–72.0)	4.0 (0.0–35.0)
Additional CABG surgery (%)	12 (0.1)	59 (0.2)	13 (0.2)	1 (0.1)
On ventilator post-CABG (%)	3,443 (33.6)	13,503 (34.8)	3,246 (39.7)	669 (37.9)^a
Hospitalization disposition (%)
Died	132 (1.3)	546 (1.4)^a	174 (2.1)^a	46 (2.6)^a
Home	8,856 (86.3)	33,832 (87.2)	7,073 (86.5)	1,482 (84.0)
Rehab/SNF	1,135 (11.1)	3,994 (10.3)	858 (10.5)	216 (12.2)
Transfer to in-patient facility	72 (0.7)	164 (0.4)	31 (0.4)	15 (0.8)
Other/unknown	67 (0.7)	263 (0.7)	44 (0.5)	6 (0.3)
Re-admission within 2 months (%)	3,927 (38.3)	13,844 (35.7)^a	2,898 (35.4)^a	700 (39.7)

Significantly different from sutures and 2-octyl cyanoacrylate group.

CABG=coronary artery bypass graft; SD=standard deviation; SNF =skilled nursing facility.

The mean hospital LOS increased slightly as the SSI rate increased, although the median LOS for all closure groups was seven days with the exception of the sutures, staples, and 2-octyl cyanoacrylate group, which had a median LOS of eight days. Rates of additional CABG operations were low in each of the closure groups, ranging from 0.1% to 0.2%. The in-hospital death rate increased among closure groups in the same pattern as increasing SSI rates, with the lowest rate being found in the sutures and 2-octyl cyanoacrylate group (1.3%); the rates of the other groups were: Sutures only (1.4%); sutures and staples (2.1%); and sutures, staples, and 2-octyl cyanoacrylate (2.6%).

Table 5 presents the mean and median costs associated with the CABG hospitalization for each closure group, as well as the differences for the sutures and 2-octyl cyanoacrylate group relative to the other groups. Median costs are used because the cost data distributions were highly skewed. The magnitude of the difference between the two groups was assessed via the Kruskal-Wallis statistic, a non-parametric test for equality of medians, as well as bootstrapped confidence intervals for the difference in means. An analysis of the mean hospitalization costs showed no significant difference between the suture and 2-octyl cyanoacrylate and the sutures groups, although all other comparisons of cost means and medians showed a significantly lower cost for the suture and 2-octyl cyanoacrylate group. The costs of CABG hospitalization increased correspondingly as the rates of SSI increased.

Table 5.

Total Hospitalization Cost (US$) for Coronary Artery Bypass Grafting According to Type of Closure

	Sutures+2-octyl cyanoacrylate skin adhesive (n=10,262)	Sutures only (n=38,799)	Sutures+staples (n=8,180)	Sutures+staples+2-octyl cyanoacrylate skin adhesive (n=1,765)
Mean (SD)	28,828.81 (12,503.84)	28,908.53 (14,836.14)	29,980.71 (15,145.79)	31,182.32 (15,379.17)
Median (range)	26,252.12 (3,258.71–197,352.95)	25,616.66 (1,416.64–696,797.97)	26,449.42 (7,465.26–217,526.03)	27,492.15 (8,387.74–162,032.67)
P value (t-test)		0.57	<0.0001	<0.0001
P value (Kruskal-Wallis)		<0.0001	0.027	<0.0001
Mean diff^a		−79.72	−1,151.90	−2,353.51
Bootstrap 90% confidence interval (mean diff)		−335, 167	−1,514, −802	−3,019, −1,696

CABG=coronary artery bypass graft; SD=standard deviation.

Mean diff=mean cost of sutures and 2-octyl cyanoacrylate skin adhesive vs. mean cost of comparator.

Table 6 shows the distribution of other unadjusted outcomes in non-SSI and SSI patients. Patients with SSI had significantly longer hospital stays (16.0 vs. 7.8 days), more days spent in the ICU (6.8 vs. 2.8 days), higher re-admission rates (64% vs. 35%), more use of a ventilator after surgery (59.5% vs. 34.0%), and significantly higher hospitalization costs ($47,874 vs. $28,061). Patients who had SSI also had a significantly higher hospitalization death rate than patients who did not develop SSI (4.1% vs. 1.4%).

Table 6.

Distribution of Outcomes in Patients with and without Post-Operative Surgical Site Infection

Risk factor	No SSI (n=55,877)	SSI (n=3,129)^a
Total hospitalization days
Mean (SD)	7.81 (3.73)	16.00 (10.36)
Median (range)	7.0 (1.0–74.0)	14.0 (3.0–117.0)
Intensive care unit days
Mean (SD)	2.78 (2.65)	6.77 (7.30)
Median (range)	2.0 (0.0–48.0)	4.0 (0.0–87.0)
Routine care days
Mean (SD)	5.03 (3.31)	9.23 (7.72)
Median (range)	4.0 (0.0–45.0)	7.00 (0.0–108.0)
On ventilator post-CABG (%)	18,998 (34.0)	1,863 (59.5)
Died during hospitalization (%)	770 (1.4)	128 (4.1)
Re-admission within 3 months (%)	19,378 (34.7)	1,991 (63.6)
Cost of CABG hospitalization (US%)
Mean (SD)	28,061 (12,236.59),	47,874 (30,574.35)
Median (range)	25,527 (1,417–317,079)	40,062 (4,924–696,798)

All measures are significantly different from the no SSI group.

CABG=coronary artery bypass graft; SD=standard deviation; SSI=surgical site infection.

The logistic regression results for development of SSI after CABG surgery are presented in Table 7. Method of closure variables and selected baseline patient characteristics, hospital characteristics, and surgical characteristics that showed a significant (p ≤0.05) effect using backward selection were included in the model.

Table 7.

Logistic Regression Models of Surgical Site Infection after Coronary Artery Bypass Grafting

Predictors	Odds ratio (95% confidence interval)
No 2-octyl cyanoacrylate skin adhesive or staples	Reference
Sutures+2-octyl cyanoacrylate skin adhesive	0.760 (0.682–0.846)
Sutures+staples	1.096 (0.988–1.215)
Sutures+staples+2-octyl cyanoacrylate skin adhesive	1.154 (0.951–1.401)
Age <55 years	References
55–64	0.944 (0.839–1.062)
65–74	1.076 (0.959–1.207)
>74	1.260 (1.116–1.423)
Male	References
Female	1.436 (1.328–1.554)
Region—West	References
Northeast	1.275 (1.104–1.473)
South	1.196 (1.056–1.350)
Midwest	1.144 (0.992–1.320)
Rural hospital	References
Urban hospital	1.194 (1.033–1.380)
CABG in 2005	References
CABG in 2006	0.914 (0.849–0.985)
CABG without catheterization	References
CABG with catheterization	1.135 (1.044–1.235)
CABG without internal mammary artery graft	References
CABG with internal mammary artery graft: Single	0.849 (0.767–0.939)
CABG with internal mammary artery graft: Double	0.985 (0.780–1.243)
Surgery <4 hours	References
Surgery 4 to >7 hours	1.252 (1.137–1.378)
Surgery ≥7 hours	1.716 (1.503–1.960)
Co-morbidities	Reference is absence of co-morbidity
Non-MI coronary disease	0.846 (0.785–0.911)
Diabetes mellitus	1.303 (1.207–1.407)
Obesity	1.279 (1.164–1.406)
COPD	1.569 (1.449–1.698)
Kidney disease	2.803 (2.565–3.059)
CHF	1.834 (1.685–1.996)
Connective tissue disease	1.397 (1.072–1.820)
Cancer	1.784 (1.540–2.066)

CABG=coronary artery bypass graft; CHF=congestive heart failure; COPD=chronic obstructive pulmonary disease; MI=myocardial infarction; SSI=surgical site infection.

Patients who had sutures and 2-octyl cyanoacrylate used for site closure showed significantly lower rates of post-CABG SSI (odds ratio [OR] 0.76; 95% CI 0.68, 0.85) relative to patients who had only sutures. Patients who had staples and sutures for closure did not have a higher risk of post-CABG SSI (OR 1.10; 95% CI 0.99–1.21). Patients who had both 2-octyl cyanoacrylate and staples with sutures had the highest risk of post-CABG SSI (OR 1.15; 95% CI 0.95–1.40), although this slightly elevated risk was not statistically significant, probably because of the relatively small number of patients in this group.

Additional modeling techniques such as stepwise selection and inclusion of all available covariates were conducted to ensure the validity of the original model. These models produced consistent results for the rates of post-CABG SSI relative to the different closure groups (data not shown).

The risk of post-CABG SSI increased with age, and women were more likely to develop such infections (OR 1.44; 95% CI 1.33–1.55). The SSI rates were slightly lower in 2006 (OR 0.91; 95% CI 0.85–0.99), and SSI rates were consistently higher in patients with all baseline co-morbidities, especially renal disease (OR 2.80; 95% CI 2.57–3.06), congestive heart failure (OR 1.83; 95% CI 1.69–2.00), and cancer (OR 1.78; 95% CI 1.54–2.07). The likelihood of post-CABG SSI increased as the duration of surgery increased and was lower in patients who had CABG with a single internal mammary graft and without catheterization.

Discussion

The observed rates of post-CABG infection were consistent with infection rates described in the literature [1,2,5 –7]. The unadjusted post-CABG rate of SSI for patients who had sutures and 2-octyl cyanoacrylate used as the final layer of site closure was lower than the rates in patients who received sutures and staples (4.3%; 95% CI 3.9–4.7 vs. 6.2%; 95% CI 5.7–6.8); sutures only (5.3%; 95% CI 5.1–5.5); or sutures, staples, and 2-octyl cyanoacrylate (7.1%; 95% CI 6.0–8.4). The primary comparison probably should be the difference between the sutures and 2-octyl cyanoacrylate group and the sutures and staples group. This analysis showed a significant (approximately one-third) reduction in the incidence of SSI in the sutures and 2-octyl cyanoacrylate group. There was no statistically significant difference in the SSI rate between the sutures and staples group and the sutures, staples, and 2-octyl cyanoacrylate group. These rates would be the same if the use of staples had a negative effect on the microbial barrier properties of 2-octyl cyanoacrylate if the two closure methods were used together. Alternatively, the similar infection rates could be explained by the fact that most patients undergoing CABG surgery have two or three incisions—a sternal incision and an incision or incisions at the site(s) of graft harvesting. Different closure techniques can be used at these sites, and we were unable to differentiate among methods of surgical closure at the various sites. If the patients whose incisions were closed with sutures, staples, and 2-octyl cyanoacrylate had staples and sutures only at either site, then their overall infection rate would resemble the higher infection rate of the sutures and staples group.

The overall mortality rate based on in-hospital coding was lower among patients who were closed with sutures and 2-octyl cyanoacrylate (1.29%) than for those patients who had sutures only (1.41%); sutures and staples (2.13%); or sutures, staples, and 2-octyl cyanoacrylate (2.61%). The overall hospitalization costs also were lower among patients who received sutures and 2-octyl cyanoacrylate (mean $28,829; median $26,252) than for those patients who had sutures only (mean $28,909; median $26,617); sutures and staples (mean $28,981; median $26,449), or sutures, staples, and 2-octyl cyanoacrylate (mean $31,182; median $27,492).

A logistic regression model for post-CABG SSI, controlling for confounding baseline variables, showed a significantly lower risk (OR 0.76; 95% CI 0.68–0.85) of SSI in patients treated with 2-octyl cyanoacrylate only. Patients treated with staples only (OR 1.10; 95% CI 0.99–1.22) and both 2-octyl cyanoacrylate and staples (OR 1.17; 95% CI 0.96–1.42) showed no elevated risk of SSI, in that these differences were not statistically significant. The results of the logistic model indicated other significant risk factors for post-CABG SSI, such as gender, obesity, diabetes mellitus, and other co-morbid conditions that have already been identified and discussed in the literature [1, 6, 7].

The Premier inpatient hospitalization data pose some limitations, but these data do provide a valuable tool for research. The data provide a large sample of patients and are well representative of national hospitals, patient characteristics, and risk factors. It also should be noted that there may be a variety of factors, including the clinical severity of the patient's condition and full range (and severity) of co-morbidities, associated with the risk of bleeding or subsequent infection that were not identifiable from this database. Hence, there may be variables associated with the clinical decision for specific treatment approaches or associated with the risk of outcomes of interest that could not be examined and controlled fully in this analysis. The database includes a nationally representative sample of hospitals that likely have a variety of treatment protocols for CABG and a large number of patients, and we did have the ability to control for many identifiable risk factors, including patient demographics and co-morbidities, hospital characteristics, and surgical characteristics, for which the prevalence in the study population and the association with the treatment and outcomes of interest were clinically sensible. The most important variable that could not be examined or controlled was the reason for the choice of closure technique, which may have provided some insight into the extent of potential patient channeling to or from specific treatments.

Hospitalization billing data did not permit differentiation between types of surgical closure at the sternal and graft surgical sites, nor did the data distinguish between sternal and graft harvest post-operative SSIs. We believe we captured overall site closure techniques based on hospital billing data, and any lack of co-morbidity information for these patients would, conservatively, result in underestimation of the risk.

Hospital billing data do not provide clinical verification of any post-operative infections using medical charts. Patients whose surgical sites become infected after discharge from the CABG hospitalization of interest can be identified only if they are re-admitted to the same hospital for treatment of the infection. Hospitalization billing data also did not permit classification of infection severity and were unable to differentiate between deep and superficial infections. In this study, post-operative infections were identified by a combination of antimicrobial exposure through hospital billing charges, diagnosis codes, and APR-DRG and DRG codes of inpatient re-hospitalizations. This methodology was based on the coding algorithm used for the determination of post-operative SSI after comparing hospital billing data with actual hospital records [4]; and the post-CABG SSI rates calculated in this study are consistent with those in the literature that are based on actual hospital surveillance. Finally, these results are specific to the adhesive studied, DERMABOND^® Topical Skin Adhesive, which is a 2-octyl cyanoacrylate formulation, and may not carry over to other products because of differences in chemical formulations and clinical properties.

Footnotes

Author Disclosure Statement

This study was funded by Ethicon, Inc.

KHF and MWR are employees of United BioSource Corporation, which received funding from Ethicon Inc. for this study. BBV and JCH are employees of Ethicon Inc.

References

Bhatia

, Pandey

, Rodrigues

et al. Postoperative wound infection in patients undergoing coronary artery bypass graft surgery: A prospective study with evaluation of risk factors. Indian J Med Microbiol, 2003; 21:246–251.

Swenne

, Lindholm

, Boroweic

, Carlsson

. Surgical-site infections within 60 days of coronary artery by-pass graft surgery. J Hosp Infect, 2002; 54:14–24.

Hollenbeak

, Murphy

, Koenig

et al. The clinical and economic impact of deep chest surgical site infections following coronary artery bypass graft surgery. Chest, 2000; 118:397–402.

Ligthelm

, Borzì

, Gumprecht

et al. Importance of observational studies in clinical practice. Clin Ther, 2007; 29,Spec No:1284–1292.

Yokoe

, Noskin

, Cunningham

et al. Enhanced identification of postoperative infections among inpatients. Emerg Infect Dis, 2004; 10:1924–1930.

Hassan

, Smith

, Engel

. Predictors and outcomes of sternal wound complications in patients after coronary artery bypass graft surgery. Am Surg, 2006; 72:515–520.

Harrington

, Russo

, Spelman

et al. Surgical-site infection rates and risk factor analysis in coronary artery bypass graft surgery. Infect Control Hosp Epidemiol, 2004; 25:472–476.