Predictors of Surgical Site Infections Among Patients With Diabetes Mellitus Post Coronary Artery Bypass Graft Surgery: A Quasi-experimental Study

Background. Surgical site infections (SSIs) are considered to be some of the most serious postoperative health concerns among patients with diabetes mellitus (DM) who undergo coronary artery bypass graft (CABG) surgery. Objective. The purpose of this study is to identify and explore the predictors of SSIs among patients with DM after CABG surgery. Methods. A quasi-experimental study was conducted using a convenience sample of 144 adult patients who were scheduled to undergo CABG at the main referral heart institute in Amman, Jordan from September 1, 2018, through November 30, 2018. Eligibility criteria stipulated participants should be adult (43–74 years) Jordanian patients with DM who underwent CABG surgery with or without cardiopulmonary bypass (CPB) and elective or urgent CABG surgery as well as being able to read and write Arabic. Key patient demographics, health history, baseline lab work, surgical characteristics, and clinical outcomes were collected from medical records. Data were collected to spreadsheets, anonymized, and entered into statistical software for bivariate and multivariate negative binomial regression analyses. Mean and standard deviation were used to describe continuous variables, and frequencies and percentages were used to describe categorically measured variables, along with chi-squared calculations. Results. Of the 144 participants, the majority of the patients (130; 90.3%) were male (mean age, 59.66 years [SD = 9.3]). Data revealed the most significant predictors for the development of SSIs post-CABG surgery included higher body mass index (bivariate mean 32 ± 4.6, P = .028; multivariate: [1-1.186] x 100 = 18.6% times more likely to experience SSI), lower preoperative serum cholesterol level (bivariate: P = .005; multivariate: [1-0.973] x 100 = 2.7% times less likely), and higher preoperative serum blood urea nitrogen level (bivariate: P = .011; multivariate: [1-1.191] x 100 = 19.1% times more likely). Conclusions. Three key factors were found to predict the occurrence of SSIs in patients with DM undergoing CABG. These findings underscore the necessity for health care providers to adhere to and employ meticulous infection control practices when managing at-risk CABG patients. 

Surgical site infection (SSI) is defined as an infection that arises after a surgical procedure. It may only affect the superficial layers of an incision or may extend into deeper tissues that were handled during the operation.¹ Surgical site infections are considered one of the most serious postoperative health concerns among patients undergoing cardiac surgery; these infections are accompanied by increased health care costs, rates of readmission, length of hospital stay (LOS), and considerable morbidity and mortality.^2-8 A 2011 study conducted in Jordan² showed SSI incidence was 16.8% postoperatively among patients undergoing coronary artery bypass grafting (CABG) surgery. A report issued in 2017 by the Jordanian Royal Medical Services⁹ (JRMS) indicated that the rate of SSIs was 6.8% postoperatively among patients undergoing CABG surgery. Increased health costs and high rates of morbidity and mortality are associated with sternal wound infections (SWIs). Most SWIs are classified as superficial. In SWIs, wound debridement and antibiotics are less effective. The postoperative incidence of superficial SWIs among patients undergoing cardiac surgery was approximately 3.3% to 4.65%^10,11 and the mortality rate was 4.6%.¹⁰

Hyperglycemia occurs when the blood glucose level is >140 mg/dL in patients with or without diabetes mellitus (DM)¹² and is a common occurrence in patients undergoing CABG. About 93% of patients with and 83% of patients without diabetes who underwent CABG had hyperglycemia.¹³ Several studies addressing the effects of hyperglycemia during the preoperative, intraoperative, or postoperative period indicated the incidence of higher postoperative infections, particularly SSIs, longer intensive care unit (ICU) and hospital stays, and higher morbidity and mortality after CABG surgery.^14-19

Diabetes mellitus is regarded as an independent risk factor of SSIs among patients undergoing cardiac surgery and can worsen the condition of these patients postoperatively.^10,11,20-24 The literature shows a significant correlation between uncontrolled glycemic level and adverse health outcomes.^25-27 Glycemic control for patients with DM undergoing cardiac surgery is the main focus for improving clinical outcomes, including infectious complications. Several studies have demonstrated positive outcomes with the use of insulin infusion as a standard of care to prevent potential complications of DM.^28-30 

Numerous studies have revealed that hyperglycemia during the perioperative period, including preoperative, intraoperative, and postoperative phases, is one of the major predictors of adverse health outcomes, such as SSIs, among patients^14-19; additional variables that have been shown to be predictors of SWIs include obesity, female sex, having bilateral internal mammary artery grafts, and the need for blood products, such as packed red blood cells and platelets.²¹ 

Infection assessment and prevention is a vital part of the role of health care providers, such as physicians and nurses who manage patients with DM following CABG surgery. The purpose of this study was to identify the predictors of SSIs among patients with DM after CABG surgery.

Design and sample
A quasi-experimental design was used in this study. The target population included adult Jordanian patients with DM after CABG surgery. The participants were recruited from the study population of interest and treated at a main facility located in the central part of Jordan. The participants were recruited from September 1, 2018, through November 30, 2018. This facility handles most cardiac conditions (~50%) in Jordan, including cardiac surgery. The sample size was calculated using G-Power. Given power = 0.80, level of significance α = 0.05, and a medium effect size = 0.3 determined that the sample size required was 128 patients. Ten percent of the total sample size was added due to the attrition rate. Eligibility criteria stipulated participants should be adult (43–74 years old) Jordanian patients with DM who underwent CABG surgery with or without cardiopulmonary bypass (CPB) and elective or urgent CABG surgery as well as being able to read and write Arabic. 

Data collection procedure
Four categories of data were collected and anonymized: (1) demographic data (age, sex, height, weight, and body mass index [BMI]) were collected at baseline and during the preoperative phase from medical records on spreadsheets; (2) historical medical variables (preoperative diabetic control [triglyceride, hemoglobin A1C, blood glucose, white blood cell, hemoglobin, hematocrit, platelets, creatinine, sodium, and potassium], smoking, comorbidities, and prior cardiac surgery) were collected during the preoperative phase; (3) surgical clinical data (first or subsequent surgery, number and type of harvest site grafts [anastomoses], CPB surgery, CPB time, cross-clamp time, surgery time, intra-aortic balloon pump (IABP) used, IABP time, and hospital and ICU length of stay) were collected on the day of the surgery; and (4) outcome data (absence or presence of superficial and/or deep SSIs were collected throughout the 30 days post-CABG surgery). The incidence rate of SSIs in both groups was determined at 4 timepoints: during ICU stay, at hospital discharge, 1-week post-hospital discharge, and at a 30-day follow-up after discharge. 

Data analysis
Statistical analysis was performed using SPSS version 22 (IBM Corp). Mean and standard deviation were used to describe continuous variables, and frequencies and percentages were used to describe categorically measured variables. An adjusted χ² test of association was used to assess the presence of cells with less than expected counts, violating the statistical assumptions of the χ² test, such as the Yates-corrected χ² test and the likelihood ratio (LR) χ² test used for 2 by 2 and larger contingency tables with fewer than expected counts. In addition, for predictors that were not normally distributed, nonparametric statistical procedures such as the Mann-Whitney U test were employed. Multivariate negative binomial regression analysis was used to assess the association between selected relevant variables and count variables, such as SSI rate. An α level of 0.05 was set as the level of significance for all statistical procedures executed in this study.

Ethical considerations
Ethical approval was obtained; approval from the Jordan University of Science and Technology Institutional Review Board (IRB) was obtained before the data collection process started. Another ethical approval was obtained from the IRB of the selected setting. Participation in the study was voluntary, and participant confidentiality was maintained. Study participants provided written informed consent; the study was explained to participants. 

Among the 144 participants, mean participant age was 59.66 ± 9.3 years; 46 (31.9%) of the patients were younger than 54 years, 59 (41%) were 55 to 64 years old, and 39 (27.1%) were 65 or more years old. The majority of the participants (130; 90.3%) were male. Only 10 participants, all of whom were male, developed an SSI. An LR χ² test of association indicated that a person’s age was not significantly associated with SSIs (P = .334). This also was confirmed with a non-parametric Mann-Whitney U test comparing the mean ranked ages of the patients with and without SSIs (P = .637). The Yates-corrected χ² test of association suggested there was no statistically significant association between patient sex and the likelihood of developing SSIs after CABG surgery (P = .282). However, results of a Mann-Whitney U non-parametric test suggested the patients who developed SSIs had a significantly higher BMI than patients who did not develop SSIs post-CABG surgery (mean BMI = 32 ± 4.6 kg/m² vs. 28 ± 4.2 kg/m²; P = .028). In addition, the LR-adjusted χ² test suggested class I obese (30–34.9 kg/m²) and class II obese (35–39.9 kg/m²) patients (6 total) were significantly more likely to incur SSIs than the 4 patients who were overweight (25–29.9 kg/m²) or who had a normal weight-to-height BMI (20–24.9 kg/m²) (0; P = .054). 

Moreover, patients with a health history of smoking (n = 6), myocardial infarction (MI; n = 2), and hypertension (HTN; n = 7) were not at a greater risk of having SSIs after CABG surgery. However, a Yates-corrected χ² test indicated that patients with a positive history of chronic obstructive pulmonary disease (COPD; n = 3) were slightly more inclined to have a higher incidence of SSIs than patients without COPD (n = 7; P = .083) (Table 1).

A non-parametric Mann-Whitney U test suggested that patients who developed SSIs post-CABG surgery had a significantly lower preoperative serum cholesterol than those who did not have postoperative SSIs (132.7 ± 23.2 vs. 166.1 ± 39.7; P = .005). The patients’ preoperative serum lab results did not differ significantly between those who did or did not develop postoperative SSIs (P > .05). However, a statistically significant difference in the mean blood urea nitrogen (BUN) was noted between patients who did and did not develop SSIs after CABG surgery (20.2 ± 4.6 vs. 16 ± 4.4; P = .011). According to a Mann-Whitney U non-parametric test, persons who developed SSIs after CABG surgery had a significantly higher preoperative BUN than those who did not develop SSIs. In addition, a Mann-Whitney U test suggested patients who developed SSIs also had significantly lower preoperative serum calcium than those who did not develop SSIs (8.7 ± 0.5 vs. 9 ± 0.6, respectively; P = .033) (Table 2).

In addition, the yield analysis results showed no statistically significant association between the variables surgical priority, harvested graft sites, underwent CPB, use of IABP, number of cardiac grafts used intraoperatively, CPB time, aortic cross-clamp time, postoperative complications, ICU LOS, and mortality and postoperative SSIs (P > .05). However, a Mann-Whitney U test showed patients who developed SSIs spent significantly more time on an IABP compared with those who had no SSIs (3.6 ± 7.60 vs. 0.7 ± 3.8, respectively; P = .032). Similarly, patients who developed SSIs had significantly more total operative hours than those who did not (5.3 ± 1.6 vs. 4.6 ± 0.9; P = .050), according to a Mann-Whitney U non-parametric test. Moreover, the patients who developed SSIs received significantly more mechanical ventilation hours than those who did not develop SSIs (10 ± 3.78, vs. 8 ± 3.2; P = .017).

Additional statistical analyses were performed to understand how the risk factors of SSIs were correlated with post-CABG SSI by examining the number of surgical sites measured for statistically significant associations with the patients’ sociodemographic, medical history, and laboratory work-up. Only 10 patients developed at least 1 SSI (6.9%); the remainder of the patients (134) had no SSIs. That is, the count of those infections was zero-inflated. As a result, a multivariate negative binomial regression analysis was used to assess the combined and individual association between selected relevant outcomes found in the bivariate analysis to determine their multivariate association with the cumulative probability density of the SSI rate (Table 3). The yielded iterative analysis showed patient age, sex, baseline laboratory hematological findings, hospital LOS, ICU LOS, IABP time and usage, and operative time were not significant factors (P > .05). Given the number of measured surgical infections when analyzed simultaneously with the other predictor variables such as sex, the analysis showed complete separation due to the absence of female patients with SSIs, and such predictors were dismissed from the analysis. The final model (Table 3) showed that patients’ preoperative serum cholesterol levels were significantly and negatively correlated with the rate of SSIs among patients who underwent CABG surgery (P = .03), accounting for the other predictors. This indicated that as the patients’ preoperative serum cholesterol tended to rise by 1 mmol/L, the rise in SSI rate tended to decline by a factor equal to 0.973 (ie, it declined by 1-0.973 * 100 = 2.7% times less) on average. 

However, the yielded analysis showed the patients’ preoperative serum BUN level (mg/dL) was significantly and positively correlated (P = .030) with the SSI rate, adjusting for other factors in the analysis. This indicated that a rise in the patients’ preoperative serum BUN level (mg/dL) by 1 mmol/L could be associated with a significant increase in the infection rate by a factor equal to (1-1.191) * 100 = 19.1% times higher on average. The patients’ hours of mechanical ventilation in the ICU as well as their comorbidities did not correlate significantly with their measured SSI (P > .05). However, the patients’ BMIs were a significant and positive factor with regard to SSI count (P = .019), accounting for the other predictors in the analysis. Therefore, a single additional unit increase in the patients’ BMIs correlates with a substantive rise in their risk of developing SSIs post CABG surgery, which was 18.6% times higher on average.

Bivariate analysis and multivariate negative binomial regression analysis were employed to determine the most significant predictors of postoperative SSIs using key patient demographics, health history, baseline lab work, surgical characteristics, and clinical outcomes. The results showed the most significant predictors for developing post-CABG surgery SSIs among adult Jordanian patients who had a CABG included a BMI ≥30 kg/m², higher preoperative BUN, and higher preoperative serum cholesterol. 

The current study indicated that patients with higher BMIs tended to develop SSIs more often than patients with lower BMIs on average. In addition, patients with class I and class II obesity were significantly more inclined to incur SSIs than those who were overweight or normal weight. The underlying mechanism of obesity remains unclear; obese patients may be exposed to increased tension across the sternum incision, leading to sternal instability, and this may predispose patients to the development of postoperative SSIs. Furthermore, decreased vascularity due to fatty tissue may lead to impaired wound nutrition and, therefore, poor healing, which increases the risk of postoperative SSIs. An alternative suggestion is that the use of prophylactic antibiotics for patients who are obese and undergoing cardiac surgery may be inadequate, thus increasing their susceptibility to developing postoperative SSIs. It has been suggested that factors such as technical difficulties and prolonged operative time may contribute to the development of postoperative SSIs.^{10,20-22,24,31}

The second predictor was higher preoperative BUN level (BUN >20). The current study revealed that patients who developed postoperative SSIs had a significantly higher preoperative BUN compared with those who did not have SSIs. The increased BUN levels reflect impaired kidney function as well as conditions that lead to decreased blood flow to the kidneys, such as congestive heart failure, recent heart attack, stress, shock, and severe burns.³² Patients with a higher preoperative BUN level, which could be a sign of impaired renal function, suggests they may be more susceptible to postoperative SSIs, which, in turn, may lead to weak immunity, an impaired wound healing process through hyperuremia, transfer of multiple blood products, and the presence of a long-term dialysis catheter.³³

The third predictor was preoperative cholesterol level. The current study found that patients who had developed SSIs following CABG surgery had significantly lower preoperative serum cholesterol than those who had no postoperative SSIs. To the authors’ knowledge, there is no association between patients’ preoperative low cholesterol and the development of postoperative SSIs in the literature. This result might be because patients with obesity had significantly higher preoperative triglyceride levels and higher preoperative creatinine levels, but slightly higher total cholesterol. Low-density lipoprotein (LDL) and high-density lipoprotein (HDL) were not identified in the current study, which, in turn, might not support the predictive level of total cholesterol. Total cholesterol might not be an accurate representation of the lipid profile of the participants before undergoing CABG. In this study, a higher BMI significantly predisposed patients with DM to developing SSIs after CABG surgery. 

In addition, preoperative treatment of hyperlipidemia (anti-hyperlipidemia drugs) was not assessed, nor was whether physicians preferred to reduce patient blood cholesterol before surgery. As a result, although it was not ascertained, data seemed to indicate that a lower total cholesterol level was associated with SSIs. The lack of data regarding LDL and HDL preoperatively might be a limitation in the current study and supports the need for future studies that elucidate the impact of anti-hyperlipidemia treatment on SSIs among patients with DM undergoing CABG surgery. However, the results of the current study showed that patients with a positive history of COPD and/or MI had significantly lower preoperative cholesterol levels (P = .023 and P = .050, respectively). This finding is consistent with the results of previous studies^{2,10,20,22,31} that indicated that patients with a positive history of COPD and/or MI were more likely to develop postoperative SSIs. As a result, identifying predictors among patients with DM assists health care providers, including physicians and nurses, in anticipating which patients undergoing CABG surgery are at high risk for postoperative SSIs.

The limitations of the study included the use of observational methods and use of a single institution database, which limited the generalizability of the study findings. Convenient sampling technique was used in the current study that causes limitation in the external validity and limits the generalizability of the results. Lastly, this study could be prone to the interobserver reliability issues that may occur as a result of personal opinion differences between the physicians during assessing the SSIs.

A quasi-experimental study conducted among 144 Jordanian patients undergoing CABG surgery found that higher BMIs, lower preoperative serum cholesterol levels, and higher preoperative serum BUN levels are predictors of SSIs among patients with DM post-CABG surgery. These issues must be considered by health care professionals, including physicians and nurses, to detect conditions that put patients at risk for SSIs following CABG surgery. Health care providers also should use these predictors to develop an SSI prevention protocol, increasing adherence to and employing meticulous infection control practices when working with persons undergoing CABG surgery, particularly patients with DM. Future studies that elucidate the impact of presurgical regulation of noted risk factors on SSIs among patients with DM undergoing CABG surgery are warranted. 

Authors: Audai A. Hayajneh, PhD, RN, CPT¹; Issa M. Hweidi, RN, MSN, DNSc¹; and Ala M. Zytoon, RN, MSN²

Affiliations: ¹Adult Health Nursing Department, Jordan University of Science and Technology, Irbid, Jordan; and ²Jordan Royal Medical Services, Queen Alia Heart Institute, Amman, Jordan

Correspondence: Audai A. Hayajneh, PhD, RN, CPT, Assistant Professor, Adult Health Nursing Department, Faculty of Nursing, Jordan University of Science and Technology, P.O. Box 3030, Irbid, Jordan 22110; aahayajneh@just.edu.jo

Disclosure: The authors would like to thank Jordan University of Science and Technology for funding this study (Research Grant No: 20180441). The authors disclose no financial or other conflicts of interest.