Search

Machine learning (ML) algorithms may enhance outcomes prediction and help guide clinical decision making. This study aimed to develop and validate a ML model that predicts postoperative outcomes and costs after cardiac surgery.The Society of Thoracic Surgeons registry data from 4874 patients who underwent cardiac surgery (56% coronary artery bypass grafting, 42% valve surgery, 19% aortic surgery) at our institution were divided into training (80%) and testing (20%) datasets. The Extreme Gradient Boosting decision-tree ML algorithms were trained to predict three outcomes: operative mortality, major morbidity or mortality, and Medicare outlier high hospitalization cost. Algorithm performance was determined using accuracy, F1 score, and area under the precision-recall curve (AUC-PR). The ML algorithms were validated in index surgery cases with The Society of Thoracic Surgeons risk scores for mortality and major morbidities and with logistic regression and were then applied to nonindex cases.The ML algorithms with 25 input parameters predicted operative mortality (accuracy 95%; F1 0.31; AUC-PR 0.21), major morbidity or mortality (accuracy 71%, F1 0.47; AUC-PR 0.47), and high cost (accuracy 84%; F1 0.62; AUC-PR 0.65). Preoperative creatinine, complete blood count, patient height and weight, ventricular function, and liver dysfunction were important predictors for all outcomes. For patients undergoing nonindex cardiac operations, the ML model achieved an AUC-PR of 0.15 (95% CI, 0.05-0.32) for mortality and 0.59 (95% CI, 0.51-0.68) for major morbidity or mortality.The extreme gradient boosting ML algorithms can predict mortality, major morbidity, and high cost after cardiac surgery, including operations without established risk models. These ML algorithms may refine risk prediction after cardiac surgery for a wide range of procedures.

Patients with severe respiratory failure from COVID-19 refractory to conventional therapies may be treated with extracorporeal membrane oxygenation (ECMO). ECMO requirement is associated with high mortality and prolonged hospital course. ECMO is a high-resource intervention with significant burdens placed on caregivers and families with limited data on the integration of palliative care consultation (PCC).To explore the role of standard vs. automatic PCC in the management of COVID patients on ECMO.Retrospective chart review of all COVID patients on ECMO admitted from March 2020 to May 2021 at a large volume academic medical center with subsequent analysis.Forty-eight patients were included in the analysis. Twenty-six (54.2%) received PCC of which 42% of consults were automatically initiated. PCC at any point in admission was associated with longer duration on ECMO (24.5 vs. 37 days; Plt; 0.05). Automatic PCC resulted in more family meetings than standard PCC (0 vs. 3; Plt; 0.05) and appears to trend with reduced time on ECMO, shorter length of stay, and higher DNAR rates at death, though results were not significant. Decedents not receiving PCC had higher rates of no de-escalation of interventions at time of death (31% vs. 11%), indicating full intensive care measures continued through death.Among patients with COVID-19 receiving ECMO, PCC may be associated with a shift to DNAR status particularly with automatic PCC. There may be a further impact on length of stay, duration of time on ECMO and care plan at end of life.

Machine learning may enhance prediction of outcomes after coronary artery bypass grafting (CABG). We sought to develop and validate a dynamic machine learning model to predict CABG outcomes at clinically relevant pre- and postoperative time points.The Society of Thoracic Surgeons (STS) registry data elements from 2086 isolated CABG patients were divided into training and testing datasets and input into Extreme Gradient Boosting decision-tree machine learning algorithms. Two prediction models were developed based on data from preoperative (80 parameters) and postoperative (125 parameters) phases of care. Outcomes included operative mortality, major morbidity or mortality, high cost, and 30-day readmission. Machine learning and STS model performance were assessed using accuracy and the area under the precision-recall curve (AUC-PR).Preoperative machine learning models predicted mortality (accuracy, 98%; AUC-PR = 0.16; F1 = 0.24), major morbidity or mortality (accuracy, 75%; AUC-PR = 0.33; F1 = 0.42), high cost (accuracy, 83%; AUC-PR = 0.51; F1 = 0.52), and 30-day readmission (accuracy, 70%; AUC-PR = 0.47; F1 = 0.49) with high accuracy. Preoperative machine learning models performed similarly to the STS for prediction of mortality (STS AUC-PR = 0.11; P = .409) and outperformed STS for prediction of mortality or major morbidity (STS AUC-PR = 0.28; P.001). Addition of intraoperative parameters further improved machine learning model performance for major morbidity or mortality (AUC-PR = 0.39; P.01) and high cost (AUC-PR = 0.64; P.01), with cross-clamp and bypass times emerging as important additive predictive parameters.Machine learning can predict mortality, major morbidity, high cost, and readmission after isolated CABG. Prediction based on the phase of care allows for dynamic risk assessment through the hospital course, which may benefit quality assessment and clinical decision-making.

Safety-net hospitals provide essential services to vulnerable patients with complex medical and socioeconomic circumstances. We hypothesized that matched patients at safety-net hospitals and non-safety-net hospitals would have comparable outcomes, costs, and readmission rates after isolated surgical aortic valve replacement (AVR) or mitral valve replacement (MVR).The National Readmissions Database was queried to identify patients who underwent isolated AVR (n = 109 744) or MVR (n = 31 475) from 2016 to 2018. Safety-net burden was defined as the percentage of patients who were uninsured or insured with Medicaid, with hospitals in the top quartile designated as safety-net hospitals. After propensity score matching, outcomes for AVR and MVR at safety-net hospitals vs non-safety-net hospitals were compared.Overall, 17 925 AVRs (16%) and 5516 MVRs (18%) were performed at safety-net hospitals, and these patients had higher comorbidity rates, had lower socioeconomic status, and more frequently required urgent surgery. Observed inhospital mortality was similar between safety-net hospitals and non-safety-net hospitals (AVR 2.2% vs 2.1%, P = .4; MVR 4.8% vs 4.3%, P = .1). After matching, rates of inhospital mortality, major morbidity, and readmission were similar; however, safety-net hospitals had longer length of stay after AVR (7 vs 6 days, P = .001) and higher total cost after AVR ($49 015 vs $42 473, P.001) and MVR ($59 253 vs $52 392, P.001).Isolated surgical AVR and MVR are both performed at safety-net hospitals with outcomes comparable to those at non-safety-net hospitals, supporting efforts to expand access to these procedures for underserved populations. Investment in care coordination resources to reduce length of stay and curtail cost at safety-net hospitals is warranted.

The coronavirus disease 2019 (COVID-19) pandemic continues to disrupt the provision of cardiac procedural services due to overwhelming interval surges in COVID-19 cases and the associated crisis of cardiac intervention deferment. Despite the availability of widespread testing, highly efficacious vaccines, and intensive public health efforts, the pandemic is entering its third year, where new severe acute respiratory syndrome-coronavirus-2 variants have increased the likelihood that patients scheduled for a cardiac intervention will contract COVID-19 in the perioperative period. The Society of Thoracic Surgeons (STS) Workforce on Critical Care, the STS Workforce on Adult Cardiac and Vascular Surgery, and the Canadian Society of Cardiac Surgeons have developed this document, endorsed by the STS and affirmed by the Society of Cardiovascular Angiography and Interventions and the Canadian Association of Interventional Cardiology, to provide guidance for cardiac procedure deferment and intervention timing for preoperative patients diagnosed with COVID-19. This document is intended for the perioperative cardiac surgical team and outlines the present state of the pandemic, the impact of COVID-19 on intervention outcome, and offers a recommended algorithm for individualized cardiac procedure triage and timing.

Investigations into readmissions after surgical repair of acute Stanford type A aortic dissection (TAAD) remain scarce. We analyzed potential risk factors for readmission after TAAD.The 2013 to 2014 US Nationwide Readmissions Database was queried for TAAD index hospitalizations and 90-day readmissions indicated by diagnostic and procedural codes. Multivariable analysis was completed to identify risk factors and the most common reasons for readmission.We identified 6975 patients (65% men; mean age, 60.0 ± 0.4 years) who underwent surgical repair for TAAD. Overall 2062 patients (29.6%) were readmitted within 90 days: 634 (30.7%) during the first 30 days and 1428 (69.3%) during days 31 through 90. Readmitted patients had a higher prevalence of chronic kidney disease at index admission (18.0% vs 11.6%, P = .002), greater overall index length of stay (17.8 ± 0.6 vs 15. 5 ± 0.4 days; P = .0003), and greater index hospitalization cost ($90,637 ± $2691 vs $80,082 ± $2091; P = .0003). Mortality during readmission was 3.6% (n = 74). Indications for readmission were most commonly cardiac (26.2%), infectious (17.8%), and pulmonary (11.7%). Multivariate analysis identified 2 independent risk factors for readmission: acute kidney injury (odds ratio, 1.49; 95% confidence interval, 1.24-1.78; P.0001) and an Elixhauser comorbidity index4 (odds ratio, 1.26; 95% confidence interval, 1.06-1.49; P = .009).After surgical repair of TAAD, approximately 30% of patients were readmitted within 90 days, two-thirds of them during the 31- to 90-day period. Targeted improvements in perioperative care and postdischarge follow-up of patients with multiple comorbidities could mitigate readmission rates. Efforts to reduce readmissions should be continued throughout the 90-day period.

In the 2016 ISHLT listing criteria guidelines for heart transplantation, recipients were recommended to have a body mass index (BMI)35 kg/m². However, outcomes data for subgroups of transplant recipients with a BMI35 kg/m² are limited. We examined the outcomes of heart transplant recipients who had a BMI of 35 to 39.9 kg/m² or ≥40 kg/m² and compared their outcomes with recipients who had a BMI35 kg/mUsing data from the United Network for Organ Sharing database, we performed a retrospective cohort analysis of 23,009 adults who underwent cardiac transplantation between 2009 and 2018. Transplant recipients were stratified by BMI categories (35 kg/m², 35-39.9 kg/m², and ≥40 kg/m²). Patient survival was depicted by Kaplan-Meier curves. Cox proportional-hazards modeling was used to determine the prognostic factors associated with mortality within 90 days, 1 year, and 5 years after transplantation.Survival at 90 days, 1 year, and 5 years after transplantation was better in recipients who had a BMI35 kg/m² than in those who had a BMI of 35 to 39.9 kg/m² (P values ranged from 0.01 to0.001) or ≥40 kg/m² (P0.001). Additionally, survival at 90 days (P0.001) and 1 year (P = 0.002) was significantly better in recipients who had a BMI of 35 to 39.9 kg/m² than in those who had a BMI ≥40 kg/m². In multivariate analysis, a BMI of 35 to 39.9 was significantly associated with increased 90-day mortality (HR = 1.53; 95% CI 1.12, 2.08; P = 0.01) but not increased 1-year (HR = 1.28; 95% CI 0.99, 1.66; P = 0.06) or 5-year mortality (HR = 1.11; 95% CI 0.91, 1.36; P = 0.29).Although heart transplant recipients with class II obesity (BMI 35-39.9 kg/m²) may have suboptimal survival compared with those who have a BMI35 kg/m², these patients have better outcomes than do those with class III obesity (BMI ≥40 kg/m²). Thus, contrary to current guidelines, selected patients with class II obesity should be considered for transplantation.

Background We investigated the relationship of sex with clinical outcomes after proximal aortic (ascending and arch) operations, and whether sex-specific preoperative factors are associated with mortality. Methods Of 3745 patients who underwent elective, urgent, and emergency proximal aortic operations over a 20-year period, 1153 pairs of men and women were propensity-matched, and their early and long-term outcomes were compared. Kaplan-Meier survival analysis was used to estimate late survival. Results Women and men had similar operative mortality (9.1% vs 8.8%, P = .8), stroke (5.7% vs 5.6%, P = .9), and renal failure rates (7.0% vs 6.6%, P = .7). Thirty-day mortality was 7.5% vs 5.6% (P = .06), respectively. Results were less favorable for women than for men regarding respiratory failure (34.3% vs 29.2%, P=0.008) and intensive care unit length of stay (9.11 ± 11.9 vs 7.87 ± 12.48 days; P = .023). Long-term survival was not significantly different between women and men: 66.3% (95% confidence interval [CI] 62.8%-69.5%) vs 67.1% (95% CI 63.6%-70.4%) at 5 years, and 45.9% (95% CI 41.76%-50.0%) vs 46.2% (95% CI 41.7%-50.6%) at 10 years (P = .4). Preoperative factors including diabetes, prior stroke, prior renal insufficiency, and peripheral vascular disease were associated with operative mortality in men, whereas chronic obstructive pulmonary disease was the main risk factor in women. Conclusions No differences were seen between the sexes in life-changing adverse outcomes after ascending aortic and arch procedures, although specific preoperative variables were associated with specific adverse events. Recognizing differences in preoperative risk factors for mortality between the sexes may facilitate targeted preoperative assessment, preparation, and counseling.

OBJECTIVES: Safety-net hospitals (SNHs) provide essential services to predominantly underserved patients regardless of their ability to pay. We hypothesized that patients who underwent coronary artery bypass grafting (CABG) would have inferior observed outcomes at SNHs compared with non-SNHs but that matched cohorts would have comparable outcomes. METHODS: We queried the Nationwide Readmissions Database for patients who underwent isolated CABG from 2016 to 2018. We ranked hospitals by the percentage of all admissions in which the patient was uninsured or insured with Medicaid; hospitals in the top quartile were designated as SNHs. We used propensity-score matching to mitigate the effect of confounding factors and compare outcomes between SNHs and non-SNHs. RESULTS: A total of 525,179 patients underwent CABG, including 96,133 (18.3%) at SNHs, who had a greater burden of baseline comorbidities (median Elixhauser score 8 vs 7; P = .04) and more frequently required urgent surgery (57.1% vs 52.8%; P

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.

We evaluated the relationship among community socioeconomic factors (poverty, income, and education), gender, and outcomes in patients who underwent ascending aortic, root, and arch surgery.For 2634 consecutive patients, we associated patients' ZIP codes with community socioeconomic factors. The composite adverse outcome comprised death, persistent neurological injury, and renal failure necessitating dialysis at discharge. Multivariable analysis and Kaplan-Meier survival curves were used. Men and women from the full cohort and from the elective patients were propensity matched.Median follow-up was 3.6 years (interquartile range, 1.2-9.3). Men lived in areas characterized by less poverty (P = .03), higher household income (P = .01), and more education (P = .02) than women; likewise, in the elective cohort, all community socioeconomic factors favored men (P ≤ .009). Female gender predicted composite adverse outcome (P = .006). In the propensity-matched women and men (820 pairs), the composite adverse outcome rates were 14.2% and 11%, respectively (P = .06). In 583 propensity-matched pairs of elective patients, men had less composite adverse outcome (P = .02), operative mortality (P = .04), and renal (P = .02) and respiratory failure (P = .0006). The 5- and 10-year survivals for these men and women were 74.2% versus 71.4% and 50.2% versus 48.2%, respectively (P = .06). All community socioeconomic factors in both propensity-matched groups nonsignificantly favored men.This study is among the first to examine the association among community socioeconomic factors, gender, and outcomes in patients who undergo proximal aortic surgery. Female gender predicted a composite adverse outcome. In the elective patients, most adverse outcomes were significantly less in men. In the propensity-matched patients, all community socioeconomic factors favored men, although not significantly. Larger studies with patient-level socioeconomic information are needed.

Background: Open surgical repair of a failed valve-sparing aortic root replacement (VSARR) or stentless bioroot aortic root replacement (bio-ARR) entails significant operative risks. Whether valve-in-valve transcatheter aortic valve replacement (ViV-TAVR) is feasible in patients with a previous VSARR or stentless bio-ARR remains unclear, given lingering concerns about the ill-defined aortic annulus in these patients and the potential for coronary obstruction. We present our experience with patients who had a previous VSARR or stentless bio-ARR and underwent ViV-TAVR to repair a degenerated aortic valve with combined valvular disease, aortic insufficiency and aortic stenosis. Methods: In this retrospective data review, we identified and analyzed consecutive patients with a previous VSARR or stentless bio-ARR who underwent ViV-TAVR between December 1, 2014 and August 31, 2019. Results: ViV-TAVR was performed in twelve high-risk patients with previous VSARR or bio-ARR during the study period. Of these, seven received Medtronic Freestyle porcine stentless bioprosthetic aortic roots, three received homograft aortic roots, one underwent a Ross procedure and one underwent VSARR. ViV-TAVR restored satisfactory valve function in all patients, and technical success was 100%. No patient had more than mild regurgitation after implantation. No thirty-day mortality was seen. One patient had major bleeding after transapical access, one patient had a transient ischemic stroke, and one patient needed permanent pacemaker implantation. At a median last follow-up of 21.5 months (interquartile range, 9.0–69.0 months), all patients remained alive and had satisfactory valve function. Conclusions: In this study, ViV-TAVR was a clinically effective option for treating patients with a failed stentless bio-ARR or previous VSARR. Short-term and intermediate-term results after these procedures were favorable. These findings may have important implications for treating high-risk patients with structural aortic root deterioration and call for better transcatheter heart valves that are suitable for treating aortic insufficiency.

Background The current representation of female patients and racial and ethnic minorities in cardiovascular trials is unclear. We evaluated these groups’ inclusion in U.S. cardiovascular trials. Methods Using publicly available data from ClinicalTrials.gov , we evaluated cardiovascular trials pertaining to coronary artery bypass grafting (CABG), heart valve disease, aortic aneurysm, ventricular assist devices, and heart transplantation. This yielded 178 U.S. trials (159 completed, 19 active but not recruiting) started between September 1998 and May 2017, with 237,132 participants. To examine female patients’ and racial and ethnic minorities’ representation in these trials, we calculated participation-to-prevalence ratios (PPRs). Values of 0.8 to 1.2 reflect similar representation. Results All 178 trials reported sex distribution, whereas only 76 (42.7%) trials reported racial distribution and 52 (29.2%) trials reported ethnic (Hispanic vs non-Hispanic) distribution. Among all trials, participants were 28.3% female, 11.2% were Hispanic/Latino, 4.0% were African American, 10.4% were Asian, and 2.3% were other. The CABG PPR for female patients was 0.64, for Hispanic patients was 0.72, for African American patients was 0.28, and for Asian patients was 3.20. Between 2008-2012 and 2013-2017, the CABG PPR decreased for female patients (0.67→0.50) and African American patients (0.37→0.17) but increased for Hispanic patients (0.38→1.32) and Asian patients (3.51→4.57). Conclusions Participation in cardiovascular trials by female patients and minorities (except Asian patients) remains low. Given that inherent differences among the previously mentioned groups could affect outcomes, balance is clearly needed. The engagement of our surgical leadership, community, and industry to address these disparities is vitally important.

Patients with postcardiotomy cardiogenic shock refractory to conventional support can be successfully supported with extracorporeal membrane oxygenation. Management considerations are discussed to aid clinicians caring for these patients.

Patients with severe refractory hypoxemic respiratory failure may benefit from extracorporeal membrane oxygenation (ECMO) for salvage therapy. The Coronavirus disease 2019 (COVID-19) pandemic offered three high-volume independent ECMO programs at a large medical center the chance to collaborate to optimize ECMO care at the beginning of the pandemic in Spring 2020. Between March 15, 2020, and May 30, 2020, 3,615 inpatients with COVID-19 were treated at the Texas Medical Center. During this time, 35 COVID-19 patients were cannulated for ECMO, all but one in a veno-venous configuration. At hospital discharge, 23 (66%) of the 35 patients were alive. Twelve patients died of vasodilatory shock (n = 9), intracranial hemorrhage (n = 2), and cannulation-related bleeding and multiorgan dysfunction (n = 1). The average duration of ECMO was 13.6 days in survivors and 25.0 days in nonsurvivors ( p0.04). At 1 year follow-up, all 23 discharged patients were still alive, making the 1 year survival rate 66% (23/35). At 2 years follow-up, the overall rate of survival was 63% (22/35). Of those patients who survived 2 years, all were at home and alive and well at follow-up.

Choosing between a bioprosthetic and a mechanical mitral valve is an important decision for both patients and surgeons. We compared patient outcomes and readmission rates after bioprosthetic mitral valve replacement (Bio-MVR) vs mechanical mitral valve replacement (Mech-MVR).The Nationwide Readmissions Database was queried to identify 31 474 patients who underwent isolated MVR (22 998 Bio-MVR, 8476 Mech-MVR) between January 1, 2016, and December 31, 2018. Propensity score matching by age, sex, elective status, and comorbidities was used to compare outcomes between matched cohorts by prosthesis type. Freedom from readmission within the first calendar year was estimated by Kaplan-Meier analysis and compared between matched cohorts.Bio-MVR patients were older (median age, 69 vs 57 years; P.001) and had more comorbidities (median Elixhauser score, 14 vs 11; P.001) compared with Mech-MVR patients. After propensity score matching (n = 15 549), Bio-MVR patients had similar operative mortality (3.5% vs 3.4%; P = .97) and costs ($50 958 vs $49 782; P = .16) but shorter lengths of stay (8 vs 9 days; P.001) and fewer 30-day (16.0% vs 18.1%; P = .04) and 90-day (23.8% vs 26.8%; P = .01) readmissions compared with Mech-MVR patients. The difference in readmissions persisted at 1 year (P = .045). Readmission for bleeding or coagulopathy complications was less common with Bio-MVR (5.7% vs 10.1%; P .001).Readmission was more common after Mech-MVR than after Bio-MVR. Identifying and closely observing patients at high risk for bleeding complications may bridge the readmissions gap between Bio-MVR and Mech-MVR.

Objective Sarcopenia (core muscle loss) has been used as a surrogate marker of frailty. We investigated whether sarcopenia would adversely affect survival after thoracoabdominal aortic aneurysm repair. Methods We retrospectively reviewed prospectively collected data from patients aged 60 years or older who underwent thoracoabdominal aortic aneurysm repairs from 2006 to 2016. Imaging was reviewed by 2 radiologists blinded to clinical outcomes. The total psoas index was derived from total psoas muscle cross-sectional area (cm2) at the mid-L4 level, normalized for height (m2). Patients were divided by sex-specific total psoas index values into sarcopenia (lower third) and nonsarcopenia (upper two-thirds) groups. Multivariable modeling identified operative mortality and spinal cord injury predictors. Unadjusted and adjusted survival curves were analyzed. Results Of 392 patients identified, those with sarcopenia (n = 131) were older than nonsarcopenic patients (n = 261) (70.0 years vs 68.0 years; P = .02) and more frequently presented with aortic rupture or required urgent/emergency operations. Operative mortality was comparable (sarcopenia 13.7% vs nonsarcopenia 10.0%; P = .3); sarcopenia was not associated with operative mortality in the multivariable model (odds ratio, 1.40; 95% confidence interval, 0.73-2.77; P = .3). Sarcopenic patients experienced more frequent delayed (13.0% vs 4.6%; P = .005) and persistent (10.7% vs 3.4%; P = .008) paraplegia. Sarcopenia independently predicted delayed paraplegia (odds ratio, 3.17; 95% confidence interval, 1.42-7.08; P = .005) and persistent paraplegia (odds ratio, 3.29; 95% confidence interval, 1.33-8.13; P = .01) in the multivariable model. Adjusted for preoperative/operative covariates, midterm survival was similar for sarcopenic and nonsarcopenic patients (P = .3). Conclusions Sarcopenia did not influence early mortality or midterm survival after thoracoabdominal aortic aneurysm repair but was associated with greater risk for delayed and persistent paraplegia.

Acute kidney injury after cardiac surgery increases morbidity and mortality. Diagnosis relies on oliguria or increased serum creatinine, which develop 48 to 72 hours after injury. We hypothesized machine learning incorporating preoperative, operative, and intensive care unit data could dynamically predict acute kidney injury before conventional identification.Cardiac surgery patients at a tertiary hospital (2008-2019) were identified using electronic medical records in the Medical Information Mart for Intensive Care IV database. Preoperative and intraoperative parameters included demographics, Charlson Comorbidity subcategories, and operative details. Intensive care unit data included hemodynamics, medications, fluid intake/output, and laboratory results. Kidney Disease: Improving Global Outcomes creatinine criteria were used for acute kidney injury diagnosis. An ensemble machine learning model was trained for hourly predictions of future acute kidney injury within 48 hours. Performance was evaluated by area under the receiver operating characteristic curve and balanced accuracy.Within the cohort (n = 4267), there were approximately 7 million data points. Median baseline creatinine was 1.0 g/dL (interquartile range, 0.8-1.2), with 17% (735/4267) of patients having chronic kidney disease. Postoperative stage 1 acute kidney injury occurred in 50% (2129/4267), stage 2 occurred in 8% (324/4267), and stage 3 occurred in 4% (183/4267). For hourly prediction of any acute kidney injury over the next 48 hours, area under the receiver operating characteristic curve was 0.82, and balanced accuracy was 75%. For hourly prediction of stage 2 or greater acute kidney injury over the next 48 hours, area under the receiver operating characteristic curve was 0.95 and balanced accuracy was 86%. The model predicted acute kidney injury before clinical detection in 89% of cases.Ensemble machine learning models using electronic medical records data can dynamically predict acute kidney injury risk after cardiac surgery. Continuous postoperative risk assessment could facilitate interventions to limit or prevent renal injury.

There are limited data on the use of extracorporeal membrane oxygenation (ECMO) for pregnant and peripartum women with COVID-19 associated acute respiratory distress syndrome (ARDS). Pregnant women may exhibit more severe infections with COVID-19, requiring intensive care. We supported nine pregnant or peripartum women with COVID-19 ARDS with ECMO, all surviving and suffering no major complications from ECMO. Our case series demonstrates high-maternal survival rates with ECMO support in the management of COVID-19 associated severe ARDS, highlighting that these pregnant and postpartum patients should be supported with ECMO during this pandemic.

Visual Abstract

Hyponatremia is associated with adverse outcomes in heart failure and after cardiac surgery. We hypothesized that hyponatremia is associated with poorer short-term and longer term survival in patients after continuous-flow left ventricular assist device (CF-LVAD) placement.We reviewed a single-center database of patients who received a CF-LVAD during 2012-2017. Sodium (Na) values obtained within 14 days before CF-LVAD insertion were averaged; patients (n = 332) were divided into hyponatremia (mean Na135 mEq/L; n = 160; 48.2%) and normonatremia groups (mean Na 135-145 mEq/L; n = 172; 51.8%). Patients requiring preoperative dialysis or pump exchange were excluded. We compared outcomes between preoperative hyponatremia and normonatremia groups.The two groups' baseline characteristics were similar, although hyponatremia patients more often had preoperative mechanical circulatory support (44.4% vs. 31.4%, p = 0.002). Although hyponatremic and normonatremic patients did not differ in 30-day mortality (7.5% vs. 6.5%, p = 0.7), preoperative hyponatremia was associated with greater 5-year mortality (61% vs. 44%, p = 0.03). On binary logistic regression analysis, the strongest independent predictors of late mortality were hyponatremia (odds ratio [OR] 1.88, 95% CI [1.07-3.31], p = 0.02), older age (OR 1.03, 95% CI [1.01-1.05], p = 0.01), and elevated mean right atrial pressure/pulmonary capillary wedge pressure ratio (OR 4.69, 95% CI [1.76-12.47], p = 0.002).Hyponatremia was not associated with greater early mortality but was associated with poorer late survival. The optimal timing of LVAD implantation in relation to hyponatremia, and whether correcting hyponatremia perioperatively improves long-term survival, should be investigated.