Causal determinants of postoperative length of stay in cardiac surgery using causal graphical learning.

OBJECTIVE: We aimed to learn the causal determinants of postoperative length of stay in cardiac surgery patients undergoing isolated coronary artery bypass grafting or aortic valve replacement surgery. METHODS: For patients undergoing isolated coronary artery bypass grafting or isolated aortic valve replacement surgeries between 2011 and 2016, we used causal graphical modeling on electronic health record data. The Fast Causal Inference (FCI) algorithm from the Tetrad software was used on data to estimate a Partial Ancestral Graph (PAG) depicting direct and indirect causes of postoperative length of stay, given background clinical knowledge. Then, we used the latent variable intervention-calculus when the directed acyclic graph is absent (LV-IDA) algorithm to estimate strengths of causal effects of interest. Finally, we ran a linear regression for postoperative length of stay to contrast statistical associations with what was learned by our causal analysis. RESULTS: In our cohort of 2610 patients, the mean postoperative length of stay was 219 hours compared with the Society of Thoracic Surgeons 2016 national mean postoperative length of stay of approximately 168 hours. Most variables that clinicians believe to be predictors of postoperative length of stay were found to be causes, but some were direct (eg, age, diabetes, hematocrit, total operating time, and postoperative complications), and others were indirect (including gender, race, and operating surgeon). The strongest average causal effects on postoperative length of stay were exhibited by preoperative dialysis (209 hours); neuro-, pulmonary-, and infection-related postoperative complications (315 hours, 89 hours, and 131 hours, respectively); reintubation (61 hours); extubation in operating room (-47 hours); and total operating room duration (48 hours). Linear regression coefficients diverged from causal effects in magnitude (eg, dialysis) and direction (eg, crossclamp time). CONCLUSIONS: By using retrospective electronic health record data and background clinical knowledge, causal graphical modeling retrieved direct and indirect causes of postoperative length of stay and their relative strengths. These insights will be useful in designing clinical protocols and targeting improvements in patient management.

Heparin-Induced Thrombocytopenia After Cardiac Surgery-A Statewide Review of Health Care Utilization.

BACKGROUND: Despite its severe consequences, clinical and economic impacts of heparin-induced thrombocytopenia (HIT) after cardiac operations have not been well characterized. This study assessed statewide incidence, outcomes, and resource consumption associated with HIT after cardiac operations. METHODS: This was a retrospective investigation of cardiac surgery patients using the Maryland Health Services Cost Review Commission's database from 2012 to 2020. Health care costs, utilization, and outcomes for those who experienced postoperative HIT were compared with propensity score-matched controls. RESULTS: Of 33 583 cardiac surgery patients, 184 (0.55%) were diagnosed with postoperative HIT. Compared with non-HIT patients, HIT patients were significantly more likely to be in the oldest age group (>80 years; P < .001) and to have greater severity of illness at admission (P < .001). HIT was associated with a longer hospitalization (21 vs 7 days; P < .001), greater mortality (13.6% vs 2.3%; P < .001), and greater hospital charges ($123 160 vs $45 303; P < .001), even after propensity score matching. Readmission rates were not significantly different, however, and readmission hospital charges did not significantly differ between HIT and non-HIT patients. CONCLUSIONS: In addition to worse outcomes, HIT was associated with higher costs during index admissions but not during readmissions of cardiac surgery patients. Strategies to minimize HIT could yield better outcomes and reduced costs, particularly relevant for health care systems seeking improved value-based care while reducing unnecessary hospitalizations.

Leveraging Machine Learning to Predict 30-Day Hospital Readmission After Cardiac Surgery.

BACKGROUND: Hospital readmission within 30 days of discharge is a well-studied outcome. Predicting readmission after cardiac surgery, however, is notoriously challenging; the best-performing models in the literature have areas under the curve around .65. A reliable predictive model would enable clinicians to identify patients at risk for readmission and to develop prevention strategies. METHODS: We analyzed The Society of Thoracic Surgeons (STS) Adult Cardiac Surgery Database at our institution, augmented with electronic medical record data. Predictors included demographics, preoperative comorbidities, proxies for intraoperative risk, indicators of postoperative complications, and time series-derived variables. We trained several machine learning models, evaluating each on a held-out test set. RESULTS: Our analysis cohort consisted of 4924 cases from 2011 to 2016. Of those, 723 (14.7%) were readmitted within 30 days of discharge. Our models included 141 STS-derived and 24 electronic medical records-derived variables. A random forest model performed best, with test area under the curve 0.76 (95% confidence interval, 0.73 to 0.79). Using exclusively preoperative variables, as in STS calculated risk scores, degraded the area under the curve, to 0.64 (95% confidence interval, 0.60 to 0.68). Key predictors included length of stay (12.5 times more important than the average variable) and whether the patient was discharged to a rehabilitation facility (11.2 times). CONCLUSIONS: Our approach, augmenting STS variables with electronic medical records data and using flexible machine learning modeling, yielded state-of-the-art performance for predicting 30-day readmission. Separately, the importance of variables not directly related to inpatient care, such as discharge location, amplifies questions about the efficacy of assessing care quality by readmissions.

Cerebral autoregulation in the operating room and intensive care unit after cardiac surgery.

BACKGROUND: Cerebral autoregulation monitoring is a proposed method to monitor perfusion during cardiac surgery. However, limited data exist from the ICU as prior studies have focused on intraoperative measurements. Our objective was to characterise cerebral autoregulation during surgery and early ICU care, and as a secondary analysis to explore associations with delirium. METHODS: In patients undergoing cardiac surgery (n=134), cerebral oximetry values and arterial BP were monitored and recorded until the morning after surgery. A moving Pearson's correlation coefficient between mean arterial proessure (MAP) and near-infrared spectroscopy signals generated the cerebral oximetry index (COx). Three metrics were derived: (1) globally impaired autoregulation, (2) MAP time and duration outside limits of autoregulation (MAP dose), and (3) average COx. Delirium was assessed using the 3-Minute Diagnostic Interview for CAM-defined Delirium (3D-CAM) and the Confusion Assessment Method for the ICU (CAM-ICU). Autoregulation metrics were compared using χ2 and rank-sum tests, and associations with delirium were estimated using regression models, adjusted for age, bypass time, and logEuroSCORE. RESULTS: The prevalence of globally impaired autoregulation was higher in the operating room vs ICU (40% vs 13%, P<0.001). The MAP dose outside limits of autoregulation was similar in the operating room and ICU (median 16.9 mm Hg×h; inter-quartile range [IQR] 10.1-38.8 vs 16.9 mm Hg×h; IQR 5.4-35.1, P=0.20). In exploratory adjusted analyses, globally impaired autoregulation in the ICU, but not the operating room, was associated with delirium. The MAP dose outside limits of autoregulation in the operating room and ICU was also associated with delirium. CONCLUSIONS: Metrics of cerebral autoregulation are altered in the ICU, and may be clinically relevant with respect to delirium. Further studies are needed to investigate these findings and determine possible benefits of autoregulation-based MAP targeting in the ICU.

Persistent Opioid Use After Open Aortic Surgery: Risk Factors, Costs, and Consequences.

BACKGROUND: The incidence and financial impact of persistent opioid use (POU) after open aortic surgery is undefined. METHODS: Insurance claim data from opioid-naïve patients who underwent aortic root replacement, ascending aortic replacement, or transverse arch replacement from 2011 to 2017 were evaluated. POU was defined as filling an opioid prescription in the perioperative period and between 90 and 180 days postoperatively. Postoperative opioid prescriptions, emergency department visits, readmissions, and health care costs were quantified. Multivariable logistic regression identified risk factors for POU, and quantile regression quantified the impact of POU on postoperative health care costs. RESULTS: Among 3240 opioid-naïve patients undergoing open aortic surgery, 169 patients (5.2%) had POU. In the univariate analysis, patients with POU were prescribed more perioperative opioids (375 vs 225 morphine milligram equivalents, P < .001), had more emergency department visits (45.6% vs 25.4%, P < .001), and had significantly higher health care payments in the 6 months postoperatively ($10,947 vs $7223, P < .001). Independent risk factors for POU in the multivariable logistic regression included preoperative nicotine use and more opioids in the first perioperative prescription (all P < .05). After risk adjustment, POU was associated with a $2439 increase in total health care costs in the 6 months postoperatively. CONCLUSIONS: POU is a challenge after open aortic operations and can have longer-term impacts on health care payments and emergency department visits in the 6 months after surgery. Strategies to reduce outpatient opioid use after aortic surgery should be encouraged when feasible.

Risk Factors of Ischemic and Hemorrhagic Strokes During Venovenous Extracorporeal Membrane Oxygenation: Analysis of Data From the Extracorporeal Life Support Organization Registry.

OBJECTIVES: Stroke is commonly reported in patients receiving venovenous extracorporeal membrane oxygenation, but risk factors are not well described. We sought to determine preextracorporeal membrane oxygenation and on-extracorporeal membrane oxygenation risk factors for both ischemic and hemorrhagic strokes in patients with venovenous extracorporeal membrane oxygenation support. DESIGN: Retrospective analysis. SETTING: Data reported to the Extracorporeal Life Support Organization by 366 extracorporeal membrane oxygenation centers from 2013 to 2019. PATIENTS: Patients older than 18 years supported with a single run of venovenous extracorporeal membrane oxygenation. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Of 15,872 venovenous extracorporeal membrane oxygenation patients, 812 (5.1%) had at least one type of acute brain injury, defined as ischemic stroke, hemorrhagic stroke, or brain death. Overall, 215 (1.4%) experienced ischemic stroke and 484 (3.1%) experienced hemorrhagic stroke. Overall inhospital mortality was 36%, but rates were higher in those with ischemic or hemorrhagic stroke (68% and 73%, respectively). In multivariable analysis, preextracorporeal membrane oxygenation pH (adjusted odds ratio = 0.10; 95% CI, 0.03-0.35; p < 0.001), hemolysis (adjusted odds ratio = 2.27; 95% CI, 1.22-4.24; p = 0.010), gastrointestinal hemorrhage (adjusted odds ratio = 2.01; 95% CI 1.12-3.59; p = 0.019), and disseminated intravascular coagulation (adjusted odds ratio = 3.61; 95% CI, 1.51-8.66; p = 0.004) were independently associated with ischemic stroke. Pre-extracorporeal membrane oxygenation pH (adjusted odds ratio = 0.28; 95% CI, 0.12-0.65; p = 0.003), preextracorporeal membrane oxygenation Po2 (adjusted odds ratio = 0.96; 95% CI, 0.93-0.99; p = 0.021), gastrointestinal hemorrhage (adjusted odds ratio = 1.70; 95% CI, 1.15-2.51; p = 0.008), and renal replacement therapy (adjusted odds ratio=1.57; 95% CI, 1.22-2.02; p < 0.001) were independently associated with hemorrhagic stroke. CONCLUSIONS: Among venovenous extracorporeal membrane oxygenation patients in the Extracorporeal Life Support Organization registry, approximately 5% had acute brain injury. Mortality rates increased two-fold when ischemic or hemorrhagic strokes occurred. Risk factors such as lower pH and hypoxemia during the pericannulation period and markers of coagulation disturbances were associated with acute brain injury. Further research on understanding preextracorporeal membrane oxygenation and on-extracorporeal membrane oxygenation risk factors and the timing of acute brain injury is necessary to develop appropriate prevention and management strategies.

Nutrition Support After Cardiac Surgery: Lessons Learned From a Prospective Study.

The importance of postoperative nutrition support (NS) has been poorly recognized in cardiac surgery. In this population, we aim to describe the delivery of NS, factors affecting calorie/protein delivery and NS-associated morbidity. From January 2015 to January 2017, we prospectively observed all cardiac surgery patients at a single institution who could not take nutrition orally, requiring postoperative NS, either enteral or parenteral, for the duration of NS up to 14 days. We compared outcomes to patients without NS and examined NS indications, factors affecting its delivery and its associated complications. Nine percent of patients (232/2603) required NS for a total of 1938 NS-days. The most common indication was mechanical ventilation. NS met 69% of daily caloric needs. On days when tube feeds (TFs) were held (mean of 13 hours), this decreased to 43%, compared to 96% when TFs were not held (P < 0.001). The most common reason for holding TFs was procedures. When TFs were supplemented with parenteral nutrition (TFs + PN), 86% of daily caloric needs were met. Even on days when TFs were held, this only dropped to 77% (TFs + PN), compared to 36% (TFs-only). By multivariable logistic regression, elemental and semielemental formulas, TF volume, and postpyloric feeds increased the risk of diarrhea, occurring in 28% of patients and 18% of TF-days. In cardiac surgery patients given postoperative NS, mortality and morbidity were an order of magnitude higher than those able to be fed orally. Enteral feeding delivered approximately two-thirds of needs, but PN supplementation dramatically improved this. Diarrhea was common, associated with the postpyloric route, increasing TF volume, and nonintact formula.

Assessment of Patient Ambulation Profiles to Predict Hospital Readmission, Discharge Location, and Length of Stay in a Cardiac Surgery Progressive Care Unit.

Importance: Promoting patient mobility during hospitalization is associated with improved outcomes and reduced risk of hospitalization-associated functional decline. Therefore, accurate measurement of mobility with high-information content data may be key to improved risk prediction models, identification of at-risk patients, and the development of interventions to improve outcomes. Remote monitoring enables measurement of multiple ambulation metrics incorporating both distance and speed. Objective: To evaluate novel ambulation metrics in predicting 30-day readmission rates, discharge location, and length of stay using a real-time location system to continuously monitor the voluntary ambulations of postoperative cardiac surgery patients. Design, Setting, and Participants: This prognostic cohort study of the mobility of 100 patients after cardiac surgery in a progressive care unit at Johns Hopkins Hospital was performed using a real-time location system. Enrollment occurred between August 29, 2016, and April 4, 2018. Data analysis was performed from June 2018 to December 2019. Main Outcomes and Measures: Outcome measures included 30-day readmission, discharge location, and length of stay. Digital records of all voluntary ambulations were created where each ambulation consisted of multiple segments defined by distance and speed. Ambulation profiles consisted of 19 parameters derived from the digital ambulation records. Results: A total of 100 patients (81 men [81%]; mean [SD] age, 63.1 [11.6] years) were evaluated. Distance and speed were recorded for more than 14 000 segments in 840 voluntary ambulations, corresponding to a total of 127.8 km (79.4 miles) using a real-time location system. Patient ambulation profiles were predictive of 30-day readmission (sensitivity, 86.7%; specificity, 88.2%; C statistic, 0.925 [95% CI, 0.836-1.000]), discharge to acute rehabilitation (sensitivity, 84.6%; specificity, 86.4%; C statistic, 0.930 [95% CI, 0.855-1.000]), and length of stay (correlation coefficient, 0.927). Conclusions and Relevance: Remote monitoring provides a high-information content description of mobility, incorporating elements of step count (ambulation distance and related parameters), gait speed (ambulation speed and related parameters), frequency of ambulation, and changes in parameters on successive ambulations. Ambulation profiles incorporating multiple aspects of mobility enables accurate prediction of clinically relevant outcomes.

Understanding Characteristics of Acute Brain Injury in Adult Extracorporeal Membrane Oxygenation: An Autopsy Study.

OBJECTIVES: Current studies lack information on characteristics of acute brain injury in patients with extracorporeal membrane oxygenation. We sought to characterize the types, timing, and risk factors of acute brain injury in extracorporeal membrane oxygenation. DESIGN: Retrospective analysis. SETTING: We reviewed the extracorporeal membrane oxygenation patients who had undergone brain autopsy with gross and microscopic examinations from January 2009 to December 2018 from a single tertiary center. PATIENTS: Twenty-five patients (median age 53 yr) had postmortem brain autopsy. INTERVENTIONS: Description and analysis of neuropathologic findings. MEASUREMENT AND MAIN RESULTS: Of 25, 22 had venoarterial extracorporeal membrane oxygenation (88%) (nine cardiac arrest; 13 cardiogenic shock) and three had venovenous extracorporeal membrane oxygenation cannulation (12%). The median extracorporeal membrane oxygenation support time was 96 hours (interquartile range, 26-181 hr). The most common acute brain injury was hypoxic-ischemic brain injury (44%), followed by intracranial hemorrhage (24%), and ischemic infarct (16%). Subarachnoid hemorrhage (20%) was the most common type of intracranial hemorrhage, followed by intracerebral hemorrhage (8%), and subdural hemorrhage (4%). Only eight patients (32%) were without acute brain injury after extracorporeal membrane oxygenation. The most common involved location for hypoxic-ischemic brain injury was cerebral cortices (82%) and cerebellum (55%). The pattern of ischemic infarct was territorial in cerebral cortices. The risk factors for acute brain injury included hypertension history (11 vs 1; p = 0.01), preextracorporeal membrane oxygenation antiplatelet use (7 vs 0; p = 0.03), and a higher day 1 lactate level (10.0 vs 5.1; p = 0.02). Patients with hypoxic-ischemic brain injury had more hypertension (8 vs 4; p = 0.047), a higher day 1 lactate level (12.6 vs 5.8; p = 0.02), and a lower pH level (7.09 vs 7.24; p = 0.027). Extracorporeal membrane oxygenation duration, cannulation methods, hemoglobin level, coma, renal impairment, and hepatic impairment were not associated with acute brain injury. CONCLUSIONS: In the population who underwent postmortem neuropathologic evaluation, 68% of extracorporeal membrane oxygenation nonsurvivors developed acute brain injury. Hypoxic-ischemic brain injury was the most common type of injury suggesting that patients sustained acute brain injury as a consequence of cardiogenic shock and cardiac arrest. Further research with a systematic neurologic monitoring is necessary to define the timing of acute brain injury in patients with extracorporeal membrane oxygenation.

Vasoactive and Inotropic Support, Tube Feeding, and Ischemic Gut Complications After Cardiac Surgery.

BACKGROUND: Vasoactive and inotropic support (VIS) may predispose cardiac surgery patients to ischemic gut complications (IGCx). The purpose of this study was to describe the effect of VIS on the manner in which we deliver tube feeds (TFs) and determine its relationship with IGCx in cardiac surgery patients. METHODS: We reviewed cardiac surgery patients at a single institution and examined the effect of VIS (none, low, medium, high) on TF administration and evaluated IGCx. RESULTS: Of 3088 cardiac surgery patients, 249 (8%) required TFs, comprising 2151 total TF-days. Increasing VIS was associated with decreased amounts of TF administered per day (P = .001) and an increase in time that TF was held per day (P < .001). High VIS was associated with less intact, more semi-elemental/elemental formula use (P < .001) and increased use of gastric route (P < .001). Of all cardiac surgery patients, 11 of 3125 suffered IGCx (0.4%), with a mortality of 73%. Of the 3 receiving TF, 2 IGCx were focal and consistent with acute embolus, whereas one was diffuse, on high VIS and an intra-aortic balloon pump. Of the 8 IGCx in the patients not receiving TF, 5 were focal, whereas 3 were diffuse and not embolic (P = .21). CONCLUSIONS: Despite 32% of TF-days on moderate to high VIS, non-embolic IGCx were not increased compared with patients not receiving TF. As delivered at this institution, TF in even those requiring moderate to high inotropic and pressor support were not associated with an increase in attributable IGCx.

New Persistent Opioid Use After Aortic and Mitral Valve Surgery in Commercially Insured Patients.

BACKGROUND: Deaths from prescription opioid overdose are dramatically increasing. This study evaluates the incidence, risk factors, and cost of new persistent opioid use after aortic valve replacement, mitral valve replacement, and mitral valve repair. METHODS: Insurance claims from commercially insured patients who underwent aortic valve replacement, mitral valve replacement, mitral valve repair, or aortic valve replacement and mitral valve replacement/repair from 2014 to 2016 were evaluated. New persistent opioid use was defined as opioid-naive patients who filled an opioid prescription in the perioperative period and filled opioid prescriptions between 90 and 180 days postoperatively. Multivariable logistic regression identified risk factors for new persistent opioid use. Quantile regression evaluated the impact of new persistent opioid use on total healthcare payments in the 6 months after discharge. RESULTS: Among 3404 opioid-naive patients undergoing aortic valve replacement, mitral valve replacement, or mitral valve repair, 188 (5.5%) had new persistent opioid use. Living in the southern United States (odds ratio, 1.89; 95% confidence interval, 1.35-2.63; P < .001) and increased opioids prescribed in the perioperative period (odds ratio, 1.009; 95% confidence interval, 1.006-1.012; P < .001) were independently associated with new persistent opioid use. After risk adjustment, new persistent opioid use was associated with a 2-fold higher number of emergency department visits (odds ratio, 2.21; 95% confidence interval, 1.61-3.03; P < .001) and a $5422 increase in healthcare payments in the 6 months after discharge. CONCLUSIONS: New persistent opioid use is a significant and costly complication after aortic and mitral valve surgery in privately insured patients. Variation in regional susceptibility and opioid prescribing suggests that standardization may help prevent this complication.

Predictors of new persistent opioid use after coronary artery bypass grafting.

OBJECTIVE: Deaths from prescription opioid overdose have quadrupled in the past 15 years, and no studies have evaluated appropriate opioid prescribing after cardiac surgery. The aim of this study is to quantify the amount of outpatient opioids prescribed to patients after coronary artery bypass grafting and determine the incidence and risk factors for new persistent opioid use after coronary artery bypass grafting. METHODS: Insurance claim data from privately insured opioid-naïve patients who underwent coronary artery bypass grafting from 2014 to 2016 were evaluated. New persistent opioid use was defined as patients who filled an opioid prescription in the perioperative period and filled opioid prescriptions between 90 and 180 days after surgery. Multivariable logistic regression was used to determine the preoperative and operative factors associated with new persistent opioid use. RESULTS: Among 7292 opioid-naïve patients undergoing coronary artery bypass grafting, 5628 (77.2%) filled opioid prescriptions in the perioperative period, and 590 (8.1%) had new persistent opioid use. Female gender (odds ratio [OR], 1.30; confidence interval [CI], 1.05-1.61; P = .018), anxiety (OR, 1.40; CI, 1.09-1.81; P = .009), tobacco use (OR, 1.34; CI, 1.08-1.65; P = .007), prior substance abuse (OR, 1.99; CI, 1.16-3.41; P = .013), chronic obstructive pulmonary disease (OR, 1.29; CI, 1.02-1.63; P = .037), living in the Southern United States (OR, 1.46; CI, 1.21-1.77; P < .001), and increased amount of opioids prescribed in the perioperative period (OR, 1.016; CI, 1.014-1.018; P < .001) were independently associated with new persistent opioid use. CONCLUSIONS: New persistent opioid use after coronary artery bypass grafting is surprisingly common. Prospective studies are needed to determine the opioid requirements of patients after coronary artery bypass grafting to prevent opioid dependence.

Prospective Image Quality Analysis and Control for Prior-Image-Based Reconstruction of Low-Dose CT.

PURPOSE: Prior-image-based reconstruction (PIBR) is a powerful tool for low-dose CT, however, the nonlinear behavior of such approaches are generally difficult to predict and control. Similarly, traditional image quality metrics do not capture potential biases exhibited in PIBR images. In this work, we identify a new bias metric and construct an analytical framework for prospectively predicting and controlling the relationship between prior image regularization strength and this bias in a reliable and quantitative fashion. METHODS: Bias associated with prior image regularization in PIBR can be described as the fraction of actual contrast change (between the prior image and current anatomy) that appears in the reconstruction. Using local approximation of the nonlinear PIBR objective, we develop an analytical relationship between local regularization, fractional contrast reconstructed, and true contrast change. This analytic tool allows prediction bias properties in a reconstructed PIBR image and includes the dependencies on the data acquisition, patient anatomy and change, and reconstruction parameters. Predictions are leveraged to provide reliable and repeatable image properties for varying data fidelity in simulation and physical cadaver experiments. RESULTS: The proposed analytical approach permits accurate prediction of reconstructed contrast relative to a gold standard based on exhaustive search based on numerous iterative reconstructions. The framework is used to control regularization parameters to enforce consistent change reconstructions over varying fluence levels and varying numbers of projection angles - enabling bias properties that are less location- and acquisition-dependent. CONCLUSIONS: While PIBR methods have demonstrated a substantial ability for dose reduction, image properties associated with those images have been difficult to express and quantify using traditional metrics. The novel framework presented in this work not only quantifies this bias in an intuitive fashion, but it gives a way to predict and control the bias. Reliable and predictable reconstruction methods are a requirement for clinical imaging systems and the proposed framework is an important step translating PIBR methods to clinical application.

A Novel Risk Score to Predict the Need for Nutrition Support After Cardiac Surgery.

BACKGROUND: In specific patients, early postoperative nutrition mitigates malnutrition-related morbidity and mortality. The goal of this study was to develop and validate a prediction score designed to stratify patients immediately after cardiac surgery according to risk for nutrition support (NS). METHODS: We identified adult cardiac surgery patients at our institution in 2012 requiring postoperative NS, enteral or parenteral. Using multivariable logistic regression modeling, we developed a Johns Hopkins Hospital Nutrition Support (JHH NS) score from relative odds ratios generated by variables that independently predicted the need for NS. The JHH NS score was then prospectively validated using all patients undergoing cardiac surgery in 2015. RESULTS: Among 1,056 patients in the derivation cohort, 87 (8%) required postoperative NS. Seven variables were identified on multivariable analysis as independent predictors of NS need and were used to create the JHH NS score. Scores ranged from 0 to 36. Each 1-point increase in the JHH NS score was associated with a 20% increase in the risk of requiring NS (odds ratio 1.20, p < 0.001). The c-statistic of the regression model for NS was 0.85. In all, 115 of 1,336 patients (8.6%) in the validation cohort required NS. Observed rates of NS in the validation group correlated positively with predicted rates (r = 0.89). CONCLUSIONS: The JHH NS score reliably stratified patients at risk for the need for postoperative NS. This easily calculable and highly predictive screening tool may expedite timing of initiation of NS in patients at high risk for not being able to physically take in adequate nutrition.

Preoperative Staphylococcus Aureus Screening and Targeted Decolonization in Cardiac Surgery.

BACKGROUND: We assessed the impact of preoperative Staphylococcus aureus screening and targeted decolonization on the incidence of postoperative methicillin-resistant S aureus (MRSA) colonization, intensive care unit MRSA transmission, and surgical site infections in cardiac surgery patients. METHODS: We reviewed medical records for all adult patients during two periods: preintervention (January 2007 to April 2010) and intervention (January 2011 to December 2014). In the intervention period, we performed nasal screening for methicillin-sensitive S aureus and MRSA using polymerase chain reaction within 30 days of the operation. Colonized patients received intranasal mupirocin twice daily and chlorhexidine baths daily for 5 days; patients colonized with MRSA also received prophylactic vancomycin plus cefazolin with contact isolation precautions. Nasal surveillance for MRSA was performed on intensive care unit admission and weekly thereafter. Multivariable logistic regression models were constructed to determine risk factors for postoperative MRSA colonization, and surgical site infections and the impact of our screening program was assessed in these models. Poisson regression was used to assess MRSA transmission. RESULTS: Comparing 2,826 preintervention and 4,038 intervention patients, cases differed in age, diabetes mellitus, preoperative infection, preoperative length of stay, and bypass time (all p ≤ 0.03). Intervention patients had risk-adjusted reductions in MRSA colonization (odds ratio 0.53, 95% confidence interval [CI]: 0.37 to 0.76, p < 0.001), transmission (incidence rate ratio 0.29, 95% CI: 0.13 to 0.65, p = 0.002), and surgical site infections (odds ratio 0.58, 95% CI: 0.40 to 0.86, p = 0.007). Increased duration of preoperative decolonization therapy was associated with decreased postoperative MRSA colonization (odds ratio 0.73, 95% CI: 0.53 to 1.00, p = 0.05). CONCLUSIONS: Preoperative S aureus screening with targeted decolonization was associated with reduced MRSA colonization, transmission, and surgical site infections. Duration of preoperative therapy correlated with decreased frequency of postoperative MRSA colonization.

A novel score to estimate the risk of pneumonia after cardiac surgery.

OBJECTIVE: The purpose of this study was to derive and validate a risk score for pneumonia (PNA) after cardiac surgery. METHODS: Adults undergoing cardiac surgery between 2005 and 2012 were identified in a single-institution database. The primary outcome was postoperative PNA. Patients were randomly assigned to training and validation sets in a 3:1 ratio. A multivariable model was constructed incorporating univariate pre- and intraoperative predictors of PNA in the training set. Points were assigned to significant risk factors in the multivariable model based on their associated regression coefficients. RESULTS: A total of 6222 patients were included. The overall rate of postoperative PNA was 4.5% (n = 282). A 33-point score incorporating 6 risk factors (age, chronic lung disease, peripheral vascular disease, cardiopulmonary bypass time, intraoperative red blood cell transfusion, and pre- or intraoperative intra-aortic balloon pump) was generated. The model used to generate the score in the training set was robust in predicting PNA (c = 0.72, P < .001). Predicted rates of PNA increased exponentially with increasing risk score, ranging from 1.2% (score = 0) to 59% (score = 33). There was significant correlation between predicted rates of PNA based on the training cohort and actual rates of pneumonia in the validation cohort in weighted regression analysis (r = 0.74, P < .001). The composite score outperformed the STS prolonged ventilation model in predicting PNA in the validation cohort (c-index 0.76 vs 0.71, respectively). CONCLUSIONS: This 33-point risk score is strongly predictive of postoperative PNA after cardiac surgery. The composite score has utility in tailoring perioperative management and in targeting diagnostic and preventative interventions.

Early Extubation: A Proposed New Metric.

Shorter intubation periods after cardiac surgery are associated with decreased morbidity and mortality. Although the Society of Thoracic Surgeons uses a 6-hour benchmark for early extubation, the time threshold above which complications increase is unknown. Using an institutional Society of Thoracic Surgeons database, we identified 3007 adult patients who underwent 1 of 7 index cardiac operations from 2010-2014. Patients were stratified by the duration of time to extubation after surgery-0-6, 6-9, 9-12, and 12-18 hours. Aggregate outcomes were compared among time-to-extubation cohorts. Primary outcomes included operative mortality and a composite of major postoperative complications; secondary outcomes included prolonged postoperative hospital length of stay (PLOS) (> 14 days) and reintubation. Multivariable logistic regression analysis was used to control for case mix. In results, extubation percentages in each time cohort were hours 0-6-36.4%, 6-9-25.6%, 9-12-12.5%, and 12-18-10.5%. Patients extubated in hours 12-18 vs < 12 experienced a significantly higher risk of operative mortality (odds ratio = 2.7, 95% CI: 1.0-7.5, P = 0.05) and the composite complication outcome (odds ratio = 3.6, 95% CI: 2.2-6.1, P < 0.01); however, insignificant differences were observed in those extubated in hours 6-9 vs 0-6 nor in hours 9-12 vs 0-9. An identical trend was observed for our secondary outcomes of PLOS and reintubation. In conclusion, our results indicate that the risks of operative mortality, major morbidity, and PLOS do not significantly increase until the time interval to extubation exceeds 12 hours. Cardiac surgery programs should be evaluated on their ability to extubate patients within this time interval.

A Predictive Model and Risk Score for Unplanned Cardiac Surgery Intensive Care Unit Readmissions.

BACKGROUND: Readmissions or "bounce back" to the intensive care unit (ICU) following cardiac surgery is associated with an increased risk of morbidity and mortality. We sought to identify clinical and system-based factors associated with ICU bounce backs in order to generate a Bounce Back After Transfer (BATS) prediction score. METHODS: We prospectively collected the clinical and financial records of all patients undergoing coronary artery bypass grafting (CABG) or surgical aortic valve replacement (AVR) between May 2013 and March 2014. Multivariable logistic regression was used to identify independent predictors of bounce backs to the ICU which served as the basis for our BATS score. RESULTS: Of the 532 patients that underwent CABG or AVR during the study period, 35 (6.6%) were readmitted to the ICU. After risk adjustment, female sex, NYHA class III/IV, urgent or emergent operative status, and postoperative renal failure were the predictors of ICU bounce backs utilized to create the BATS score. Patients in the low (<5), moderate (5-10), and high-risk (>10) score cohorts experienced bounce back rates of 3.0%, 10.4%, and 42%, respectively. After adjusting for preoperative patient risk, ICU bounce back resulted in an increase in $68,030 to a patient's total hospital charges. CONCLUSIONS: A predictive model (BATS) can determine the risk of a bounce back to the ICU after transfer to the floor. We speculate that determination of a patient's BATS upon ICU transfer would allow targeted floor care and decrease bounce back rates, along with postoperative morbidity, mortality, and cost of care.

A Novel Risk Score to Predict Dysphagia After Cardiac Surgery Procedures.

BACKGROUND: Although the exact cause of dysphagia after cardiac operations is unknown, timely diagnosis is critical to avoid a devastating aspiration event. Accordingly, we sought to generate the risk of dysphagia in cardiac surgery (RODICS) score to identify patients at risk for its development after heart surgery. METHODS: All adult heart surgery patients at our institution between January 2011 and March 2012 were analyzed. A videofluoroscopic swallow study stratified patients into two groups based on the presence or absence of dysphagia. Covariates (p < 0.20) were included in a multivariable model to determine the strongest independent predictors of postoperative dysphagia. Based on the relative odds ratios of significant variables, the RODICS score was generated. Risk cohorts were then created based on easily applicable, whole-integer score cutoffs. RESULTS: During the study period, 115 of 1,314 patients (8.8%) undergoing heart surgery were diagnosed with clinically significant dysphagia. The 38-point RODICS score comprises seven patient-specific characteristics and perioperative factors. The low risk (less than 4), intermediate risk (5 to 9), and high risk (more than 9) cohorts had postoperative dysphagia rates of 3.0%, 6.8%, and 21.6%, respectively (p < 0.001). The intermediate-risk cohort (odds ratio 2.3, 95% confidence interval: 1.33 to 4.27, p = 0.01) and high-risk cohort (odds ratio 8.9, 95% confidence interval: 5.22 to 15.32, p < 0.001) were at significantly higher risk of dysphagia developing. The RODICS score demonstrated excellent discriminatory ability (area under the curve 0.75). CONCLUSIONS: The incidence and impact of dysphagia after open cardiac operations is significant. This novel scoring system could lead to prompt identification of patients at high risk for postoperative dysphagia and potentially minimize the complications of aspiration.

dPIRPLE: a joint estimation framework for deformable registration and penalized-likelihood CT image reconstruction using prior images.

Sequential imaging studies are conducted in many clinical scenarios. Prior images from previous studies contain a great deal of patient-specific anatomical information and can be used in conjunction with subsequent imaging acquisitions to maintain image quality while enabling radiation dose reduction (e.g., through sparse angular sampling, reduction in fluence, etc). However, patient motion between images in such sequences results in misregistration between the prior image and current anatomy. Existing prior-image-based approaches often include only a simple rigid registration step that can be insufficient for capturing complex anatomical motion, introducing detrimental effects in subsequent image reconstruction. In this work, we propose a joint framework that estimates the 3D deformation between an unregistered prior image and the current anatomy (based on a subsequent data acquisition) and reconstructs the current anatomical image using a model-based reconstruction approach that includes regularization based on the deformed prior image. This framework is referred to as deformable prior image registration, penalized-likelihood estimation (dPIRPLE). Central to this framework is the inclusion of a 3D B-spline-based free-form-deformation model into the joint registration-reconstruction objective function. The proposed framework is solved using a maximization strategy whereby alternating updates to the registration parameters and image estimates are applied allowing for improvements in both the registration and reconstruction throughout the optimization process. Cadaver experiments were conducted on a cone-beam CT testbench emulating a lung nodule surveillance scenario. Superior reconstruction accuracy and image quality were demonstrated using the dPIRPLE algorithm as compared to more traditional reconstruction methods including filtered backprojection, penalized-likelihood estimation (PLE), prior image penalized-likelihood estimation (PIPLE) without registration, and prior image penalized-likelihood estimation with rigid registration of a prior image (PIRPLE) over a wide range of sampling sparsity and exposure levels.