Infections Following Cardiac Surgery: An Opportunity for Clarity∗
There remains little doubt that morbidity from health care–associated infections (HAI) following cardiac surgery exacts substantial clinical and economic impact, justifying ongoing targeted strategies for quality improvement (1). However, despite the rising comorbidity profile of patients requiring cardiac surgery in the United States, there has been a substantive decline in the prevalence of major postoperative infections in recent years. In the current era of transparency of outcomes and public reporting, augmented data-driven awareness of the impact of preventive measures, such as appropriate-use criteria for antibiotics, perioperative blood sugar control, and blood conservation, has justly made infection deterrence a priority of all programs performing cardiac surgery (2–4). Although center-level variance exists, the current overall HAI rates following cardiac surgery are at an all-time low of <1% for septicemia and deep sternal wound infection (DSWI) and <5% for pneumonia (4–6). Despite these achievements, opportunity still exists for clarity between clinical data, claims data, and homogeneity when approaching the occasionally challenging diagnostic dilemma of infection in postoperative cardiac surgical patients.
In this issue of the Journal, Greco et al. (7) elegantly present an analysis of data from 4,320 adult patients undergoing a broad array of cardiac operations between February 2010 and October 2010, in order to evaluate the cost and prevalence of HAI for up to 65 days following surgery. In an attempt to address some of the important known limitations of a purely claims-based examination of infection using International Classification of Diseases-Ninth Edition codes, the authors from the Cardiothoracic Surgery Trials Network pooled clinical data from cardiac operations from nine U.S. academic institutions and linked independent, de-identified, patient-level financial charges. They applied infection definitions to the clinical data using criteria set forth by the Centers for Disease Control and Prevention (CDC) and the National Healthcare Safety Network. They estimated costs from charges by using institutional cost-to-charge ratios on the basis of annual hospital Medicare cost reports. Through information available to the authors, they applied a best-in-class, generalized linear model to statistically estimate the cost of defined infections. The result was that, despite the relatively low 2.8% overall incidence of HAI, the assailing cost of index admissions and readmissions highlight the importance of infection following cardiac surgery in value-based health care delivery.
Center-level quality-improvement protocols to mitigate infectious complications following cardiac surgery stem from national quality initiatives on the basis of detailed clinical analyses of homogeneous, nonemergent, primary risk-adjusted operations. In the current study, the authors included all forms of cardiac operations in the analysis. Elective, urgent, emergent coronary artery bypass grafting (CABG) operations; valve plus CABG; cardiac transplantation; and ventricular assist device (VAD) operations were all examined in those patients without acute pre-operative infections. Major well-known predictors of infections were identified. Patients with heart failure, lower ejection fraction, systemic corticosteroid use, repeat operations, urgent or emergent status, and/or transplantation or VAD ex- or implantation all had significantly increased infection rates. In fairness to the authors, their objective was to estimate the broad economic burden of HAI following cardiac surgery. This first step in affecting quality improvement appears to have been nicely realized. Although the authors attempt to statistically justify the validity of including transplant and VAD operations through cost estimations from generalized linear modeling, the clinical reality remains that the inclusion of these patients known to be at exceedingly high risk for infection and readmission was an apples-to-oranges, or at least an apples-to-pears, comparison. Before recommendations are made about how the development of infections following cardiac surgery might be addressed in health care policy and hospital-level reimbursement, further procedure-specific granularity is required.
We therefore applied the similar criteria outlined in current study by Greco et al. (7) to an examination of the most recent annual data from the Society of Thoracic Surgeons’ (STS) Adult Cardiac Surgery Database. After excluding cases of active endocarditis, we reviewed 217,829 cardiac operations performed between January 1, 2013, and December 31, 2013, inclusive. The overall prevalence of HAI across all operation types was 3.8%. However, expected significant differences in infections were observed when classified by operation category (Table 1). The overall prevalence of DSWI remained remarkably low, at <1%, across all operations (transplantation, 0.92%; VAD, 0.33%; isolated CABG, 0.21%; and isolated valve, 0.09%). There were significant differences in the prevalence of septicemia defined by positive blood cultures, the highest being for transplantation and VAD operation (transplantation, 5.49%; VAD, 4.28%; valve, 0.53%; and CABG, 0.43%; p < 0.0001). A similar important trend was observed for postoperative pneumonia (9.46%, VAD; 7.32%, transplant; CABG, 2.81%; and valve, 2.53%; p < 0.0001).
(n = 146,498)
(n = 43,565)
|Valve + CABG†|
(n = 25,226)
(n = 437)
|Ventricular Assist Device‡|
(n = 2,103)
|Deep sternal wound infection|
|Yes||309 (0.21)||40 (0.09)||57 (0.23)||4 (0.92)||7 (0.33)|
|No||145,878 (99.58)||43,421 (99.67)||25,091 (99.46)||394 (90.16)||2,020 (96.05)|
|Data unavailable||311 (0.21)||104 (0.24)||78 (0.31)||39 (8.92)||76 (3.61)|
|Yes||4,116 (2.81)||1,103 (2.53)||1,239 (4.91)||32 (7.32)||199 (9.46)|
|No||142,090 (96.99)||42,359 (97.23)||23,912 (94.79)||366 (83.75)||1,832 (87.11)|
|Data unavailable||292 (0.20)||103 (0.24)||75 (0.30)||39 (8.92)||72 (3.42)|
|Sepsis (positive blood cultures)|
|Yes||633 (0.43)||233 (0.53)||214 (0.85)||24 (5.49)||90 (4.28)|
|No||145,304 (99.18)||43,136 (99.02)||24,877 (98.62)||374 (85.58)||1,935 (92.01)|
|Data unavailable||561 (0.38)||196 (0.45)||135 (0.54)||39 (8.92)||78 (3.71)|
These new data from the STS Adult Cardiac Surgery Database and the observations of the current study together highlight another important future consideration for the interpretation and study of infections after cardiac operations. The diagnosis of infections defined as clinically and economically important to cardiac surgical programs and cardiothoracic surgeons, and those currently defined by the CDC/National Healthcare Safety Network for all forms of surgery as used in this study (8), have room for specialty-specific clarity.
To illustrate, let us review a common clinical scenario. Following thoracic incisions for cardiac surgery, patients commonly develop pulmonary atelectasis often mixed with basilar effusions. This fairly routine situation may often occur a few days following cardiac surgery, when a patient may still be on oxygen, with low-grade fever and leukocytosis, with a radiologic appearance of consolidation or atelectasis while expectorating sputum. To the experienced cardiothoracic surgeon, this scenario is completely consistent with routine postoperative atelectasis that will resolve with physical therapy of the chest and conservative management. However, the factors outlined in this everyday example also may meet the current CDC definition of clinically defined pneumonia (PNU1). Health care providers not experienced with patients following cardiac surgery, let alone immunosuppressed transplant recipients, may be quick to treat with antibiotics only for these to be discontinued rapidly once the atelectasis or effusion resolves. Similarly, hospital coders and allied health infection control providers may wish to code this as pneumonia, as it may been seen as administratively advantageous for morbidity indices.
This relatively common incongruence of clinical and administrative data objectives is one example of potential misinterpretation of the presence and relevance of infection following cardiac operations (9). To collaboratively address this unique situation, the STS and CDC recently entered into discussions with the goal of harmonizing definitions of infections specific to cardiac surgical patients, such as those of pneumonia and DSWI. These harmonized definitions will involve both the numerator (definition of cardiac surgical infections) and the denominator (definition of eligible cardiac surgical operations) necessary to calculate the rates of infection associated with adult and pediatric cardiac surgery.
Once this opportunity for clarity is realized, further risk-adjusted, operation-specific study of the impact of infection following cardiac surgery may be unambiguously translated to tangible hospital-level improvements in outcome, cost, and quality.
1. : "Management practices and major infections after cardiac surgery". J Am Coll Cardiol 2014; 64: 372.
2. : "Adherence to surgical care improvement project measures and the association with postoperative infections". JAMA 2010; 303: 2479.
3. : "Blood product conservation is associated with improved outcomes and reduced costs after cardiac surgery". J Thorac Cardiovasc Surg 2013; 145: 796.
4. : "Sustained reduction in methicillin-resistant Staphylococcus aureus wound infections after cardiothoracic surgery". Arch Intern Med 2010; 171: 68.
5. : "Trends in isolated coronary artery bypass grafting: an analysis of the Society of Thoracic Surgeons Adult Cardiac Surgery Database". J Thorac Cardiovasc Surg 2012; 143: 273.
6. : "Center-level variation in infection rates after coronary artery bypass grafting". Circ Cardiovasc Qual Outcomes 2014; 7: 567.
7. : "Costs associated with health care-associated infections in cardiac surgery". J Am Coll Cardiol 2015; 65: 15.
8. CDC/NHSN Surveillance Definitions for Specific Types of Infections. January 1, 2014. Centers for Disease Control and Prevention. Available at: http://www.cdc.gov/nhsn/pdfs/pscmanual/17pscnosinfdef_current.pdf. Accessed October 7, 2014.
9. : "Use of administrative data for surgical site infection surveillance after congenital cardiac surgery results in inaccurate reporting of surgical site infection rates". Ann Thorac Surg 2014; 97: 651.
Both authors have reported that they have no relationships relevant to the contents of this paper to disclose.