Informatics infrastructure for syndrome surveillance, decision support, reporting, and modeling of critical illness

Epidemiology and outcomes of acute kidney injury in cardiac intensive care unit patients

PURPOSE

To describe the epidemiology and outcomes of acute kidney injury (AKI) among contemporary non-surgical cardiac intensive care unit (CICU) patients.

MATERIALS AND METHODS

We reviewed adult non-surgical CICU patients admitted from 2007 to 2015. The highest AKI stage during hospitalization was defined using modified Kidney Disease: Improving Global Outcomes (KDIGO) criteria, based on changes in serum creatinine. Hospital and 5-year mortality were examined using logistic regression and Cox proportional-hazards models, respectively.

RESULTS

We included 9311 patients with a mean age of 67.5 years, including 37% females. AKI was present in 51%: stage 1 AKI in 34%, stage 2 AKI in 9%, and stage 3 AKI in 8%. Hospital mortality was associated with AKI stage (adjusted OR for each AKI stage 1.17, 95% CI 1.04-1.31, p = 0.007). Five-year mortality was incrementally associated with AKI stage (adjusted HR per AKI stage 1.13, 95% CI 1.08-1.18; p < 0.001), particularly post-discharge mortality among hospital survivors (adjusted HR per AKI stage 1.20, 95% CI 1.15-1.25, p < 0.001). Patients with stage 3 AKI (especially requiring dialysis) had the highest adjusted hospital and five-year mortality.

CONCLUSION

AKI severity is incrementally associated with higher short-term and long-term mortality in CICU patients, especially severe AKI requiring dialysis.

Timeline of sepsis bundle component completion and its association with septic shock outcomes

PURPOSE

To assess the impact of the timeline of sepsis bundle completion with clinical outcomes in septic shock.

MATERIALS AND METHODS

We retrospectively studied adult (≥18 years) patients with septic shock from January 1, 2006, through May 31, 2018, who were admitted to the intensive care unit in Mayo Clinic, Rochester. We divided patients into three groups based on the SSC compliant 1) <1h, 2) 1.1 to 3 h, 3) >3 h after the time of septic shock diagnosis.

RESULTS

We enrolled 1052 septic shock patients, among 8% were in group 1, 26% in group 2, and the remaining in group 3. Those who completed all bundle components within 3 h had the lowest 28-day mortality (17.5% vs. 31.4%, p < .001) and higher survival at 90 days (HR = 0.67; 95% CI 0.55-0.80; p < .001). Sepsis bundle completion in <1 h had no significant advantage in 28-day mortality (21.5% vs.15.9%, p = .4) or 90-day survival compared with group 2 (HR = 1.08; 95% CI 0.77-1.53; p = .6).

CONCLUSIONS

We showed an association between the completion of SSC bundle components within three hours with lower mortality or earlier shock reversal. This relationship was not evident when compared to bundle completion in 1 h vs. within 3 h.

Shock Severity and Hospital Mortality In Out of Hospital Cardiac Arrest Patients Treated With Targeted Temperature Management

BACKGROUND

Shock in patients resuscitated after out of hospital cardiac arrest (OHCA) is associated with an increased risk of mortality. We sought to determine the associations between lactate level, mean arterial pressure (MAP), and vasopressor/inotrope doses with mortality.

METHODS

Retrospective cohort study of adult patients with OHCA of presumed cardiac etiology treated with targeted temperature management (TTM) between December 2005 and September 2016. Multivariable logistic regression was performed to determine predictors of hospital death.

RESULTS

Among 268 included patients, the median age was 64 (55, 71.8) years, including 27% females. OHCA was witnessed in 89%, OHCA rhythm was shockable in 87%, and bystander CPR was provided in 64%. Vasopressors were required during the first 24 h in 60%. Hospital mortality occurred in 104 (38.8%) patients. Higher initial lactate, peak Vasoactive-Inotropic Score (VIS), and lower mean 24-h MAP were associated with higher hospital mortality (all P < 0.001). After multivariable regression, both higher initial lactate (adjusted OR 1.15 per 1 mmol/L higher, 95% CI 1.00-1.31, P = 0.03) and higher peak VIS (adjusted OR 1.20 per 10 units higher, 95% CI 1.10-1.54, P = 0.003) were associated with higher hospital mortality, but mMAP was not (P = 0.92). However, patients with a mMAP < 70 mm Hg remained at higher risk of hospital mortality after multivariable adjustment (adjusted OR 9.30, 95% CI 1.39-62.02, P = 0.02).

CONCLUSIONS

In patients treated with TTM after OHCA, greater shock severity, as reflected by higher lactate levels, mMAP < 70 mmHg, and higher vasopressor requirements during the first 24 h was associated with an increased rate of hospital mortality.

Patterns of Cystatin C Uptake and Use Across and Within Hospitals

OBJECTIVE

To characterize the use of cystatin C (cysC) across and within hospitals.

PATIENTS AND METHODS

This 2-part study first evaluated access to cysC testing across 129 hospitals in the state of Minnesota, using a telephone-based survey. Second, granular data from a single center (Mayo Clinic) with on-site, rapid-turnaround testing (<1 day) and automated estimated glomerular filtration rate (eGFR) reporting was used to describe temporal patterns. The characteristics of hospitals that offered cysC testing and of patients who underwent rapid cysC testing at Mayo Clinic between January 1, 2011, and March 31, 2018, were described. Poisson regression analyzed temporal trends in cysC testing.

RESULTS

Of the 114 hospitals (88%) that responded to the statewide survey, cysC was available in 91 (80%), but only 3 of 91 (3%) reported a turnaround time of <1 day. At Mayo Clinic, cysC use increased from 0.74 tests per 1000 patient-days in 2011 to 14 tests per 1000 patient-days in 2018 (P=.004). Of the 3774 patients with cysC tests, the mean first available eGFR was 46 mL/min per 1.73 m2 using cysC and 59 mL/min per 1.73 m2 using serum creatinine (P<.001). CysC testing was used across all intensities of care and was ordered by a variety of specialties. Nephrology was consulted in only 42% of cases.

CONCLUSION

In the hospital, rapid-turnaround cysC testing is necessary for practical use but was not widely available in Minnesota. When available, a marked increase in cysC testing was observed over the study timeframe. Additional research is needed to determine optimal strategies for implementation of cysC within hospitals.

Hyperkalemia Is Associated With Increased Mortality Among Unselected Cardiac Intensive Care Unit Patients

Background

Hyperkalemia has been associated with increased mortality in patients with myocardial infarction, but few data exist regarding hyperkalemia in cardiac intensive care unit (CICU) patients. We hypothesize that hyperkalemia is associated with increased mortality in unselected CICU patients.

Methods and Results

We retrospectively reviewed a historical cohort of 9681 CICU patients admitted from January 2007 to December 2015. Hyperkalemia was defined as admission potassium ≥5.0 mEq/L and hypokalemia as admission potassium <3.5 mEq/L. Multivariate logistic regression was used to determine predictors of in‐hospital mortality. Postdischarge survival was assessed using Kaplan–Meier analysis and Cox proportional hazards models. The mean age of included patients was 67±15 years, with 36% females, and in‐hospital mortality was 9%. Hyperkalemia occurred in 1187 (12.3%) and hypokalemia occurred in 719 (7.4%) patients. Both patients with hyperkalemia (unadjusted odds ratio, 2.85; 95% CI, 2.40–3.39; P<0.001) and patients with hypokalemia (unadjusted odds ratio, 2.31; 95% CI, 1.85–2.88; P<0.001) were at increased risk of unadjusted in‐hospital mortality. After adjustment for illness severity and renal function, only patients with hyperkalemia demonstrated increased risk of in‐hospital death (adjusted odds ratio, 1.44; 95% CI, 1.11–1.87; P=0.006). Among hospital survivors, only patients with hyperkalemia had lower postdischarge survival by Kaplan–Meier analysis (P<0.001). After adjustment for illness severity and renal function, hospital survivors with admission hyperkalemia remained at increased risk for postdischarge mortality (adjusted hazard ratio, 1.20; 95% CI, 1.08–1.34; P<0.001).

Conclusions

Hyperkalemia on CICU admission is associated with higher in‐hospital and postdischarge mortality, independent of renal function and illness severity. These findings emphasize the importance of potassium abnormalities as a risk predictor in patients admitted to the CICU.

Cefepime is Associated with Acute Encephalopathy in Critically Ill Patients: A Retrospective Case-Control Study

BACKGROUND

Acute encephalopathy (AE) is a common complication of critical illness and is associated with increased short and long-term mortality. In this study, we evaluated the role of cefepime in causing AE.

METHODS

Retrospective case-control study involving consecutive patients enrolled in the intensive care units (ICUs) of Mayo Clinic Rochester, MN between July 1, 2004 and December 31, 2015. AE was defined by the presence of delirium or depressed level of consciousness in the absence of deep sedation. Controls were identified as patients not developing AE and were matched by propensity score for age, Charlson Comorbidity Index, 24-h Apache III score and invasive ventilation use.

RESULTS

The total number of eligible ICU admissions during our study period was 152,999. AE was present in 57,726 (37.7%) with a median AE duration of 17 (interquartile range [IQR] 4.0-51.8) hours. We matched 14,645 cases with AE with the same number of controls. Cefepime was used in 1241 (4.2%) patients and its use was associated with greater incidence of AE [713 (4.9%) vs 528 (3.6%), p < 0.001] and duration [unit estimate 0.73; (95% CI 0.542-0.918)]. On multivariate analysis, cefepime was associated with an increased likelihood of AE after controlling for shock, midazolam infusion and acute kidney injury [OR 1.24 (95% CI 1.10-1.27)]. These associations were also present after controlling for prior chronic kidney disease.

CONCLUSION

The use of cefepime is associated with increased likelihood and duration of AE. These associations are stronger among patients with impaired renal function, but can also occur in patients without renal impairment.

Inpatient Kidney Function Recovery among Septic Shock Patients Who Initiated Kidney Replacement Therapy in the Hospital

BACKGROUND

Sepsis and septic shock are life-threatening causes of acute kidney injury (AKI) frequently seen and managed in intensive care units (ICUs). Sepsis-associated AKI (SA-AKI) independently contributes to the mortality of sepsis. Understanding the potential factors involved in kidney function recovery may further aid in the prevention and management of SA-AKI. This study aimed to describe the clinical characteristics of septic shock patients who required kidney replacement therapy and factors associated with kidney function recovery.

METHODS

We conducted a retrospective cohort study of adult septic shock patients who received in-hospital kidney replacement therapy at medical intensive care unit (MICU) at the Mayo Clinic, Rochester, from January 1, 2006, to May 31, 2018. Kidney function recovery was defined as liberation from kidney replacement therapy before hospital discharge. Associations between clinical features and kidney recovery were analyzed using multivariable Fine and Gray regression accounting for death as a competing event.

RESULTS

Our retrospective cohort consisted of 229 patients with a median (interquartile range [IQR]) age of 64 (52-74) years: 55% were men, 89% were Caucasians, 39% had diabetes mellitus (DM), 16% had heart failure, APACHE (Acute Physiology and Chronic Health Evaluation) III score was 105 (84-123), and SOFA (Sequential [Sepsis-related] Organ Failure Assessment) score was 12 (9-14). The patients received 1,567 (524-4,108) mL of intravenous fluids in the first 3 h, 92% required vasopressor support, and 83% required mechanical ventilation. The median MICU and hospital stays were 7 (4-13) and 19 (10-31) days, respectively. Median (IQR) kidney replacement therapy duration was 7 (3.5-17.1) days. Among 158 ICU survivors, 73 (46%) patients were weaned from RRT in ICU and 85 (54%) were transitioned to intermittent RRT. A higher volume of fluid resuscitation in the first 3 h (hazard ratio [HR] = 1.07 per 1 L, CI: 1.01-1.14, p = 0.04) and a history of DM (HR = 1.70, CI: 1.14-2.54, p = 0.009) were associated with kidney function recovery.

CONCLUSION

Among septic shock patients who initiated kidney replacement therapy in the MICU, 41% recovered kidney function before discharge. A higher initial fluid resuscitation volume was associated with recovery, and interestingly, patients with DM had a higher chance of recovery.

Admission diagnosis and mortality risk prediction in a contemporary cardiac intensive care unit population

BACKGROUND

Critical care risk scores can stratify mortality risk among cardiac intensive care unit (CICU) patients, yet risk score performance across common CICU admission diagnoses remains uncertain.

METHODS

We evaluated performance of the Acute Physiology and Chronic Health Evaluation (APACHE)-III, APACHE-IV, Sequential Organ Failure Assessment (SOFA) and Oxford Acute Severity of Illness Score (OASIS) scores at the time of CICU admission in common CICU admission diagnoses. Using a database of 9,898 unique CICU patients admitted between 2007 and 2015, we compared the discrimination (c-statistic) and calibration (Hosmer-Lemeshow statistic) of each risk score in patients with selected admission diagnoses.

RESULTS

Overall hospital mortality was 9.2%. The 3182 (32%) patients with a critical care diagnosis such as cardiac arrest, shock, respiratory failure, or sepsis accounted for >85% of all hospital deaths. Mortality discrimination by each risk score was comparable in each admission diagnosis (c-statistic 95% CI values were generally overlapping for all scores), although calibration was variable and best with APACHE-III. The c-statistic values for each score were 0.85-0.86 among patients with acute coronary syndromes, and 0.76-0.79 among patients with heart failure. Discrimination for each risk score was lower in patients with critical care diagnoses (c-statistic range 0.68-0.78) compared to non-critical cardiac diagnoses (c-statistic range 0.76-0.86).

CONCLUSIONS

The tested risk scores demonstrated inconsistent performance for mortality risk stratification across admission diagnoses in this CICU population, emphasizing the need to develop improved tools for mortality risk prediction among critically-ill CICU patients.

Effect of initial infusion rates of fluid resuscitation on outcomes in patients with septic shock: a historical cohort study

Background

Fluid resuscitation has become the cornerstone of early septic shock management, but the optimal fluid rate is still not well studied. The goal of this investigation is to examine the relationship between fluid resuscitation rate and septic shock resolution.

Method

We retrospectively studied adult (≥ 18 years) patients with septic shock, defined based on sepsis III definition, from January 1, 2006, through May 31, 2018, in the medical intensive care unit (MICU) of Mayo Clinic Rochester. The fluid resuscitation time was defined as the time required to infuse the initial fluid bolus of 30 ml/kg, based on the recommendations of the 2016 surviving sepsis campaign. The cohort was divided into four groups based on the average fluid rate (group 1 ≥ 0.5, group 2 0.25–0.49, group 3 0.17–0.24, and group 4 < 0.17 ml/kg/min). The primary outcome was the time to shock reversal. Multivariable regression analyses were conducted to account for potential confounders.

Result

A total of 1052 patients met eligibility criteria and were included in the analysis. The time-to-shock reversal was significantly different among the groups (P < .001). Patients in group 1 who received fluid resuscitation at a faster rate had a shorter time to shock reversal (HR = 0.78; 95% CI 0.66–0.91; P = .01) when compared with group 4 with a median (IQR) time-to-shock reversal of 1.7 (1.5, 2.0) vs. 2.8 (2.6, 3.3) days, respectively. Using 0.25 ml/kg/min as cutoff, the higher fluid infusion rate was associated with a shorter time to shock reversal (HR = 1.22; 95% CI 1.06–1.41; P = .004) and with decreased odds of 28-day mortality (HR = 0.71; 95% CI 0.60–0.85; P < .001).

Conclusion

In septic shock patients, initial fluid resuscitation rate of 0.25–0.50 ml/kg/min (i.e., completion of the initial 30 ml/kg IV fluid resuscitation within the first 2 h), may be associated with early shock reversal and lower 28-day mortality compared with slower rates of infusion.

Influence of cardiac arrest and SCAI shock stage on cardiac intensive care unit mortality

BACKGROUND

Patients with concomitant cardiac arrest (CA) and shock are at increased risk of mortality, even when stratified according to shock severity. We sought to determine whether the presence of ventricular fibrillation (VF) modified the relationship between CA and mortality in cardiac intensive care unit (CICU) patients.

METHODS

We retrospectively analyzed unique Mayo Clinic CICU patients admitted between 2007 and 2015. Society for Cardiovascular Angiography and Intervention (SCAI) shock stages A through E were classified at admission. Hospital mortality in each SCAI shock stage was stratified by the presence of CA, VF CA, or non-VF CA.

RESULTS

We included 9,898 patients with a mean age of 68 years (38% females). CA was present in 12%, including 53% with VF CA and 47% with non-VF CA. Hospital mortality was higher in patients with CA compared to patients without CA (34% vs. 6%; adjusted odds ratio [OR] = 3.1, 95% CI [2.4, 4.0], p < .001), and patients with non-VF CA had higher hospital mortality than patients with VF CA (44% vs. 25%; adjusted OR = 2.1, 95% CI [1.4, 3.0], p < .001). After adjustment, patients with any CA or non-VF CA had higher hospital mortality at each SCAI stage, except stage E (all other p < .05), whereas patients with VF CA did not (all p > .1).

CONCLUSIONS

CA rhythm modifies the relationship between CA and mortality in CICU patients, when accounting for coma, shock, and organ failure. Outcome studies examining CA in patients with cardiogenic shock need to account for important differences such as CA rhythm.

Risk of respiratory failure among hospitalized patients with various admission serum potassium levels

BACKGROUND

The objective of this study was to assess the relationship between admission serum potassium and the risk of respiratory failure requiring mechanical ventilation in all hospitalized patients.

METHODS

All non-dialysis and non-mechanically ventilated patients who had serum potassium measurement at admission from 2011 to 2013 were studied. Serum potassium levels were stratified into five groups; ≤3.4, 3.5 to 3.9, 4.0 to 4.4, 4.5 to 4.9, 5.0 to 5.4, and ≥5.5 mEq/L. The outcome of interest was the respiratory failure requiring mechanical ventilation during hospitalization. Logistic regression analysis was performed to assess the independent risk of in-hospital respiratory failure requiring mechanical ventilation based on various admission serum potassium, using serum potassium of 4.0 to 4.4 mEq/L as the reference group.

RESULTS

Of 67,034 eligible patients, with the mean admission serum potassium of 4.2 ± 0.5 mEq/L, 2,886 (4.3%) patients developed respiratory failure requiring mechanical ventilation during hospitalization. As demonstrated by U-shaped association, increased risk of in-hospital respiratory failure was significantly associated with low admission serum potassium ≤ 3.4 mEq/L (odds ratio 1.36, p-value <0.001) and high admission serum potassium ≥5.5 mEq/L (odds ratio 1.37, p-value = 0.01).

CONCLUSION

Increased risk of in-hospital respiratory failure requiring mechanical ventilation was noted when serum potassium was below 3.5 mEq/L or above 5.4 mEq/L at the time of hospital admission. Patients with either hypokalemia or hyperkalemia are at risk of respiratory failure requiring mechanical ventilation.

Use of diuretics in shock: Temporal trends and clinical impacts in a propensity-matched cohort study

OBJECTIVE

Fluid overload is common among critically ill patients and is associated with worse outcomes. We aimed to assess the effect of diuretics on urine output, vasopressor dose, acute kidney injury (AKI) incidence, and need for renal replacement therapies (RRT) among patients who receive vasopressors.

PATIENTS AND METHODS

This is a single-center retrospective study of all adult patients admitted to the intensive care unit between January 2006 and December 2016 and received >6 hours of vasopressor therapy and at least one concomitant dose of diuretic. We excluded patients from cardiac care units. Hourly urine output and vasopressor dose for 6 hours before and after the first dose of diuretic therapy was compared. Rates of AKI development and RRT initiation were assessed with a propensity-matched cohort of patients who received vasopressors but did not receive diuretics.

RESULTS

There was an increasing trend of prescribing diuretics in patients receiving vasopressors over the course of the study. We included 939 patients with median (IQR) age of 68(57, 78) years old and 400 (43%) female. The average hourly urine output during the first six hours following time zero in comparison with average hourly urine output during the six hours prior to time zero was significantly higher in diuretic group in comparison with patients who did not receive diuretics [81 (95% CI 73-89) ml/h vs. 42 (95% CI 39-45) ml/h, respectively; p<0.001]. After propensity matching, the rate of AKI within 7 days of exposure and the need for RRT were similar between the study and matched control patients (66 (15.6%) vs. 83 (19.6%), p = 0.11, and 34 (8.0%) vs. 37 (8.7%), p = 0.69, respectively). Mortality, however, was higher in the group that received diuretics. Ninety-day mortality was 191 (45.2%) in the exposed group VS 156 (36.9%) p = .009.

CONCLUSIONS

While the use of diuretic therapy in critically ill patients receiving vasopressor infusions augmented urine output, it was not associated with higher vasopressor requirements, AKI incidence, and need for renal replacement therapy.

Early noncardiovascular organ failure and mortality in the cardiac intensive care unit

Background

Noncardiac organ failure has been associated with worse outcomes among a cardiac intensive care unit (CICU) population.

Hypothesis

We hypothesized that early organ failure based on the sequential organ failure assessment (SOFA) score would be associated with mortality in CICU patients.

Methods

Adult CICU patients from 2007 to 2015 were reviewed. Organ failure was defined as any SOFA organ subscore ≥3 on the first CICU day. Organ failure was evaluated as a predictor of hospital mortality and postdischarge survival after adjustment for illness severity and comorbidities.

Results

We included 10 004 patients with a mean age of 67 ± 15 years (37% female). Admission diagnoses included acute coronary syndrome in 43%, heart failure in 46%, cardiac arrest in 12%, and cardiogenic shock in 11%. Organ failure was present in 31%, including multiorgan failure in 12%. Hospital mortality was higher in patients with organ failure (22% vs 3%, adjusted OR 3.0, 95% CI 2.5‐3.7, P < .001). After adjustment, each failing organ system predicted twofold higher odds of hospital mortality (adjusted OR 1.9, 95% CI 1.1‐2.1, P < .001). Mortality risk was highest with cardiovascular, coagulation and liver failure. Among hospital survivors, organ failure was associated with higher adjusted postdischarge mortality risk (P < .001); multiorgan failure did not confer added long‐term mortality risk.

Conclusions

Early noncardiovascular organ failure, especially multiorgan failure, is associated with increased hospital mortality in CICU patients, and this risk continues after hospital discharge, emphasizing the need to promote early recognition of organ failure in CICU patients.

Predictive Value of the Get With The Guidelines Heart Failure Risk Score in Unselected Cardiac Intensive Care Unit Patients

Background

The cardiac intensive care unit (CICU) population is no longer composed of only patients with acute coronary syndromes, and includes those with acute heart failure and multiple comorbidities. We hypothesized that the GWTG‐HF (Get With The Guidelines–Heart Failure) risk score that predicts inpatient mortality in hospitalized patients with heart failure would predict mortality in CICU patients.

Methods and Results

We retrospectively analyzed CICU patients at a tertiary care hospital from 2007 to 2015. The GWTG‐HF risk score was calculated at CICU admission. As a secondary analysis, the EFFECT (Enhanced Feedback for Effective Cardiac Treatment), OPTIMIZE‐HF (Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients With Heart Failure), and ADHERE (Acute Decompensated Heart Failure National Registry) risk scores were calculated. Kaplan–Meier survival analysis and the area under the receiver operating characteristic curve value were determined for inpatient and 1‐year mortality. The GWTG‐HF risk score was calculated in 9532 (95%) patients, with a median value of 40 (interquartile range, 35–47). Inpatient mortality occurred in 824 (8.6%) patients, and 2075 (21.8%) patients died by 1 year. Patients who died in hospital had a significantly higher mean GWTG‐HF score (47.7 versus 40.2; P<0.001). Inpatient and 1‐year mortality increased in each GWTG‐HF risk score quartile (P<0.0001). Discrimination of the GWTG‐HF, EFFECT, OPTIMIZE‐HF, and ADHERE risk scores was assessed using area under the receiver operating characteristic curve values for hospital mortality, and were similar for all risk scores (0.72–0.74; P>0.05). The Hosmer–Lemeshow statistic suggested poor calibration for hospital mortality by the GWTG‐HF risk score (P<0.001).

Conclusions

The GWTG‐HF risk score and other heart failure prediction tools demonstrate good discrimination for inpatient and 1‐year mortality in a heterogeneous cohort of CICU patients. Our study emphasizes that prognostic variables overlap in cardiac patients, regardless of the admission diagnosis.

Abnormal Serum Sodium is Associated With Increased Mortality Among Unselected Cardiac Intensive Care Unit Patients

Background

Abnormal serum sodium levels have been associated with higher mortality among patients with acute coronary syndromes and heart failure. We sought to describe the association between sodium levels and mortality among unselected cardiac intensive care unit (CICU) patients.

Methods and Results

We retrospectively reviewed consecutive adult patients admitted to our cardiac intensive care unit from 2007 to 2015. Hyponatremia and hypernatremia were defined as admission serum sodium <135 and >145 mEq/L, respectively. In‐hospital mortality was assessed by multivariable regression, and postdischarge mortality was evaluated by Cox proportional‐hazards analysis. We included 9676 patients with a mean age of 68±15 years (37.5% females). Hyponatremia occurred in 1706 (17.6%) patients, and hypernatremia occurred in 322 (3.3%) patients; these groups had higher illness severity and a greater number of comorbidities. Risk of hospital mortality was higher with hyponatremia (15.5% versus 7.5%; unadjusted odds ratio, 2.41; 95% CI, 2.06–2.82; P<0.001) or hypernatremia (17.7% versus 8.6%; unadjusted odds ratio, 2.82; 95% CI, 2.09–3.80; P<0.001), with a J‐shaped relationship between admission sodium and mortality. After multivariate adjustment, only hyponatremia was significantly associated with in‐hospital mortality (adjusted odds ratio, 1.42; 95% CI, 1.14–1.76; P=0.002). Among hospital survivors, risk of postdischarge mortality was higher in patients with hyponatremia (adjusted hazard ratio, 1.28; 95% CI, 1.17–1.41; P<0.001) or hypernatremia (adjusted hazard ratio, 1.36; 95% CI, 1.12–1.64; P=0.002).

Conclusions

Hyponatremia and hypernatremia on admission to the cardiac intensive care unit are associated with increased unadjusted short‐ and long‐term mortality. Further studies are needed to determine whether correcting abnormal sodium levels can improve outcomes in cardiac intensive care unit patients.

Admission Society for Cardiovascular Angiography and Intervention shock stage stratifies post-discharge mortality risk in cardiac intensive care unit patients

BACKGROUND

The five-stage Society for Cardiovascular Angiography and Intervention (SCAI) cardiogenic shock classification scheme can stratify hospital mortality risk in patients admitted to the cardiac intensive care unit (CICU). We sought to evaluate the SCAI shock classification for prediction of post-discharge mortality in CICU survivors.

METHODS

We retrospectively analyzed hospital survivors admitted to a single CICU between 2007 and 2015. SCAI CS stages A through E were classified using CICU admission data using a previously published algorithm. All-cause post-discharge mortality was compared across SCAI stages using Kaplan-Meier analysis and Cox proportional hazards models.

RESULTS

Among 9096 unique hospital survivors, 43.2% had acute coronary syndrome (ACS), 44.6% had heart failure (HF), and 8.7% had cardiac arrest (CA) on admission. The proportion of patients in each SCAI shock stage was: A, 49.1%; B, 30.6%; C, 15.2; D/E 5.2%. Kaplan-Meier survival at 5 years in each SCAI shock stage was: A, 88.2%; B, 81.6%; C, 76.7%; D/E, 71.7% (P < .001 by log-rank). Each higher SCAI shock stage was associated with increased adjusted post-discharge mortality compared to SCAI shock stage A (all P < .001); results were consistent among patients with ACS or HF. Late hemodynamic deterioration after 24 hours, but not an admission diagnosis of CA, was associated with higher post-discharge mortality.

CONCLUSIONS

The SCAI shock classification assessed at the time of CICU admission was predictive of post-discharge mortality risk among hospital survivors, although an admission diagnosis of CA was not. The SCAI shock classification can be used for post-discharge mortality risk stratification.

Risk Factors for Persistent Cognitive Impairment After Critical Illness, Nested Case-Control Study

OBJECTIVES

Persistent cognitive impairment after critical illness is an important healthcare problem forecasted to worsen in the near future. However, the epidemiology is insufficiently explored. We aimed to determine potentially modifiable risk factors during ICU hospitalization that play a significant role in developing persistent cognitive impairment.

DESIGN

An observational case-control study.

SETTINGS

Mayo Clinic ICUs between July 1, 2004, and November 20, 2015.

PATIENTS

We conducted a study nested in a large cohort of 98,227 adult critically ill patients. Using previously validated computable phenotypes for dementia and cognitive impairment, we determined the onset of cognitive impairment relative to ICU hospitalization and associated risk factors. The primary endpoint of the study was new and persistent cognitive impairment documented between 3 and 24 months after ICU discharge.

INTERVENTIONS

Unadjusted and adjusted analyses were performed to identify potentially modifiable risk factors during ICU hospitalization.

MEASUREMENTS AND MAIN RESULTS

Among 21,923 unique patients identified as cognitively impaired (22% of the entire ICU cohort), 2,428 (2.5%) developed incident new and persistent cognitive dysfunction after the index ICU admission. Compared with age- and sex-matched ICU controls (2,401 pairs), cases had higher chronic illness burden (Charlson Comorbidity Index, 6.2 vs 5.1; p < 0.01), and were more likely to have multiple ICU stays (22% vs 14%; p < 0.01). After adjustment for baseline differences, new and persistent cognitive dysfunction was associated with higher frequency of acute brain failure in the ICU, a higher exposure to severe hypotension, hypoxemia, hyperthermia, fluctuations in serum glucose, and treatment with quinolones or vancomycin. Association with sepsis observed in univariate analysis did not persist after adjustment.

CONCLUSIONS

Cognitive dysfunction is highly prevalent in ICU patients. Incident new and persistent cognitive impairment is less common but important, potentially preventable problem after critical illness. Chronic comorbidities and number of ICU stays increase the risk of post-ICU cognitive dysfunction irrespective of age. Modifiable ICU exposures were identified as potential targets for future prevention trials.

Hemodynamic profiles following digoxin use in patients with sepsis in the ICU

PURPOSE

To explore the impact of digoxin on hemodynamic parameters in patients with sepsis and tachycardia admitted to the intensive care unit.

MATERIALS AND METHODS

Retrospective review of adult patients admitted to the medical and mixed ICU at Mayo Clinic Rochester, Minnesota from March 2008 to February 2018, initiated on digoxin within 24 h of ICU stay. Hemodynamic parameters were reviewed before digoxin administration and at 6, 12 and 24 h after. Adverse events including new onset conduction abnormalities or arrhythmias during the first 48 h after digoxin administration were reviewed by a critical care cardiologist.

RESULTS

Study included 180 patients. We observed significant decrease in heart rate from 124 (115-138) beats/min 1 h before digoxin to 101 (87-117) 6 h after digoxin and 94 (84-112) 12 h after (p < .01). Median systolic blood pressure increased from 100 (91-112) mm Hg 1 h before to 110 (100-122) (p < .01) and 111 (103-124) at 6 and 12 h respectively after digoxin.

CONCLUSIONS

Early digoxin administration in patients with sepsis and tachycardia is uncommon but associated with improvements of hemodynamic parameters. These preliminary results will help formulate future hypotheses for focused trials on utility, efficacy and safety of digoxin in sepsis.

Role of Loop Diuretic Challenge in Stage 3 Acute Kidney Injury

OBJECTIVE

To assess whether loop diuretic challenge predicts the need for dialysis among critically ill patients with acute kidney injury (AKI) stage 3.

PATIENTS AND METHODS

Adult patients admitted to intensive care units between January 1, 2004, and December 31, 2016, were screened. Acute kidney injury stage 3 was identified by an electronic surveillance tool, and patients who received loop diuretics in a dosage of at least 1mg/kg intravenous bolus furosemide equivalent were included. Urine output following loop diuretic challenge was modeled as a restricted cubic spline. We then compared the area under the receiver operating characteristic curve for urine outputs at 2 hours and 6 hours after loop diuretic challenge to predict the need for dialysis within the next 24 hours.

RESULTS

Of 687 patients included in the study, those who received dialysis were younger and had higher Sequential Organ Failure Assessment scores on the day of loop diuretic challenge. Urine outputs at 2 hours and 6 hours were lower in patients who needed dialysis, but urine output by 6 hours was better in predicting dialysis initiation within 24 hours (area under the curve, 0.71 vs 0.67; P=.02). The sensitivity and specificity of 6-hour urine output cutoff of 600 mL or less to predict dialysis was 80.9% and 50.5%, respectively, and that for 300 mL or less was 64.2% and 68.2%, respectively.

CONCLUSION

Among patients with stage 3 AKI, 6-hour urine output after the loop diuretic challenge had a modest discriminant capacity to identify dialysis initiation within the next 24 hours.

Improvements in red blood cell transfusion utilization following implementation of a single-unit default for electronic ordering

BACKGROUND

The prevention of excessive allogeneic red blood cell (RBC) transfusion is an important component of patient blood management initiatives. In this investigation, changes in transfusion behaviors following modification of computerized physician order entry (CPOE) procedures for RBC transfusions to a single-unit default quantity were assessed.

STUDY DESIGN AND METHODS

This is an observational cohort study of adults for whom nonemergency allogeneic RBC transfusions were ordered in the 2 years before and 2 years after the date of modification of the CPOE system to a single-unit default (June 18, 2015). Changes in the frequency of single- versus multiunit RBC transfusion orders and other transfusion metrics were compared between preintervention and postintervention cohorts.

RESULTS

A total of 52,773 unique transfusion orders for 61,989 RBC units were included, of which 60,045 (96.9%) were transfused. Single-unit orders increased annually, from 10,404 (74.1%) in the first year to 11,645 (88.6%) in the last year, while multiunit orders decreased by more than half (p < 0.0001). The number of RBC units transfused decreased by 13.9% from 32,528 in the preintervention cohort to 27,497 in the post intervention cohort (p < 0.0001) with an estimated reduction in transfusion-related expenditures of nearly $4 million. The percentage of transfusions associated with a posttransfusion hemoglobin of10 g/dL or greater decreased by 34.5% (p < 0.0001).

CONCLUSION

Modification of the CPOE process such that nonemergency RBC transfusion orders were defaulted to a single unit was associated with decreased rates of multiunit RBC transfusion orders, lower transfusion volumes, and substantial cost savings.

Predictive value of individual Sequential Organ Failure Assessment sub-scores for mortality in the cardiac intensive care unit

Purpose

To determine the impact of Sequential Organ Failure Assessment (SOFA) organ sub-scores for hospital mortality risk stratification in a contemporary cardiac intensive care unit (CICU) population.

Materials and methods

Adult CICU admissions between January 1, 2007 and December 31, 2015 were reviewed. The SOFA score and organ sub-scores were calculated on CICU day 1; patients with missing SOFA sub-score data were excluded. Discrimination for hospital mortality was assessed using area under the receiver-operator characteristic curve (AUROC) values, followed by multivariable logistic regression.

Results

We included 1214 patients with complete SOFA sub-score data. The mean age was 67 ± 16 years (38% female); all-cause hospital mortality was 26%. Day 1 SOFA score predicted hospital mortality with an AUROC of 0.72. Each SOFA organ sub-score predicted hospital mortality (all p <0.01), with AUROC values of 0.53 to 0.67. On multivariable analysis, only the cardiovascular, central nervous system, renal and respiratory SOFA sub-scores were associated with hospital mortality (all p <0.01). A simplified SOFA score containing the cardiovascular, central nervous system and renal sub-scores had an AUROC of 0.72.

Conclusions

In CICU patients with complete SOFA sub-score data, risk stratification for hospital mortality is determined primarily by the cardiovascular, central nervous system, renal and respiratory SOFA sub-scores.

Automated Continuous Acute Kidney Injury Prediction and Surveillance: A Random Forest Model

OBJECTIVE

To develop and validate a prediction model of acute kidney injury (AKI) of any severity that could be used for AKI surveillance and management to improve clinical outcomes.

PATIENTS AND METHODS

This retrospective cohort study was conducted in medical, surgical, and mixed intensive care units (ICUs) at Mayo Clinic in Rochester, Minnesota, including adult (≥18 years of age) ICU-unique patients admitted between October 1, 2004, and April 30, 2011. Our primary objective was prediction of AKI using extant clinical data following ICU admission. We used random forest classification to provide continuous AKI risk score.

RESULTS

We included 4572 and 1958 patients in the training and validation mutually exclusive cohorts, respectively. Acute kidney injury occurred in 1355 patients (30%) in the training cohort and 580 (30%) in the validation cohort. We incorporated known AKI risk factors and routinely measured vital characteristics and laboratory results. The model was run throughout ICU admission every 15 minutes and achieved an area under the receiver operating characteristic curve of 0.88 on validation. It was 92% sensitive and 68% specific and detected 30% of AKI cases at least 6 hours before the criterion standard time (AKI stages 1-3). For discrimination of AKI stages 2 to 3, the model had 91% sensitivity, 71% specificity, and 53% detection of AKI cases at least 6 hours before AKI onset.

CONCLUSION

We developed and validated an AKI prediction model using random forest for continuous monitoring of ICU patients. This model could be used to identify high-risk patients for preventive measures or identifying patients of prospective interventional trials.

Impact of time to antibiotic on hospital stay, intensive care unit admission, and mortality in febrile neutropenia

PURPOSE

To determine if time to antibiotics (TTA) improves outcomes of hospital length of stay, admission to the intensive care unit, and 30-day mortality in adult patients with febrile neutropenia.

METHODS

This retrospective cohort study evaluated the impact of time to antibiotic, in the treatment of febrile neutropenia, on hospital length of stay, admission to the intensive care unit, and 30-day mortality. Cases included were patients 18 years or older hospitalized with febrile neutropenia from August 1, 2006 to July 31, 2016. To adjust for other characteristics associated with hospital length of stay, admission to the intensive care unit, and 30-day mortality, a multivariate analysis was performed.

RESULTS

A total of 3219 cases of febrile neutropenia were included. The median hospital length of stay was 7.0 days (IQR 4.1-13.3), rate of intensive care unit admission was 13.6%, and 30-day mortality was 6.6%. Multivariate analysis demonstrated time to antibiotics was not associated with hospital length of stay but was associated with admission to the intensive care unit admission and 30-day mortality. Delays in time to antibiotic of up to 3 hours did not impact outcomes.

CONCLUSIONS

A shorter time to antibiotic is important in treatment of febrile neutropenia; however, moderate delays in antibiotic administration did not impact outcomes. Further investigation is needed in order to determine if other indicators of infection, in addition to fever, or other supportive management, in addition to antibiotics, are indicated in the early identification and management of infection in patients with neutropenia.

New-Onset Heart Failure and Mortality in Hospital Survivors of Sepsis-Related Left Ventricular Dysfunction

BACKGROUND

The association between new-onset left ventricular (LV) dysfunction during sepsis with long-term heart failure outcomes is lesser understood.

METHODS

Retrospective cohort study of all adult patients with severe sepsis and septic shock between 2007 and 2014 who underwent echocardiography within 72 h of admission to the intensive care unit. Patients with prior heart failure, LV dysfunction, and structural heart disease were excluded. LV systolic dysfunction was defined as LV ejection fraction <50% and LV diastolic dysfunction as ≥grade II. Primary composite outcome included new hospitalization for acute decompensated heart failure and all-cause mortality at 2-year follow-up. Secondary outcomes included persistent LV dysfunction, and hospital mortality and length of stay.

RESULTS

During this 8-year period, 434 patients with 206 (48%) patients having LV dysfunction were included. The two groups had similar baseline characteristics, but those with LV dysfunction had worse function as demonstrated by worse LV ejection fraction, cardiac index, and LV diastolic dysfunction. In the 331 hospital survivors, new-onset acute decompensated heart failure hospitalization did not differ between the two cohorts (15% vs. 11%). The primary composite outcome was comparable at 2-year follow-up between the groups with and without LV dysfunction (P = 0.24). Persistent LV dysfunction was noted in 28% hospital survivors on follow-up echocardiography. Other secondary outcomes were similar between the two groups.

CONCLUSIONS

In patients with severe sepsis and septic shock, the presence of new-onset LV dysfunction did not increase the risk of long-term adverse heart failure outcomes.

Echocardiographic parameters of patients in the intensive care unit undergoing continuous renal replacement therapy

MAIN OBJECTIVES

Echocardiographic parameters have been used to predict outcomes for specific intensive care unit (ICU) populations. We sought to define echocardiographic parameters for ICU patients receiving continuous renal replacement therapy (CRRT).

DESIGN, SETTING, PARTICIPANTS, AND MEASUREMENTS

This is a historical cohort study of consecutive ICU patients at Mayo Clinic (Rochester, Minnesota) who received CRRT from December 9, 2006, through November 13, 2015. Only patients with an echocardiographic examination within 7 days of CRRT initiation were considered.

RESULTS

The study included 1,276 patients. Decreased left ventricular ejection fraction (LVEF; ≤45%) was noted in 361/1,120 (32%) and increased right ventricular systolic pressure (RVSP; ≥40 mm Hg) was noted in 529/798 (66%). Right ventricular systolic dysfunction was observed in 320/820 (39%). The most common valvular abnormality was tricuspid regurgitation (244/1,276 [19%]). Stratification of these parameters by ICU type (medical, surgical, cardiothoracic, cardiac) showed that most echocardiographic abnormalities were significantly more prevalent among cardiac ICU patients: LVEF ≤45% (67/105 [64%]), RVSP ≥40 mm Hg (63/79 [80%]) and tricuspid regurgitation (50/130 [38%]). We compared patients with acute kidney injury (AKI) vs end-stage renal disease and showed that decreased LVEF (284/921 [31%] vs 78/201 [39%]), was significantly less prevalent among patients with AKI, but increased RVSP was more prevalent (445/651 [68%] vs 84/147 [57%]) with AKI.

CONCLUSIONS

ICU patients who required CRRT had increased prevalence of pulmonary hypertension and right and left ventricular systolic dysfunction. Prediction of adverse outcomes with echocardiographic parameters in this patient population can lead to identification of modifiable risk factors.

Pharmacologic Treatment of Intensive Care Unit Delirium and the Impact on Duration of Delirium, Length of Intensive Care Unit Stay, Length of Hospitalization, and 28-Day Mortality

OBJECTIVE

To determine whether treatment of delirium affects outcomes.

PATIENTS AND METHODS

A retrospective cohort study of patients admitted to the medical intensive care unit (ICU) from July 1, 2015, through June 30, 2016, was conducted. Patients with ICU delirium, defined by a positive Confusion Assessment Method for the ICU score, were included. Patients were stratified into 4 treatment groups based on exposure to melatonin and antipsychotic agents during ICU stay: no pharmacologic treatment, melatonin only, antipsychotics only, and both melatonin and antipsychotics. A time-dependent cause-specific hazards model with death as a competing risk was used to evaluate the effect of melatonin or antipsychotic drug use for delirium on duration of ICU delirium, length of ICU stay, and length of hospitalization. A logistic regression was used to evaluate 28-day mortality. Covariates significantly associated with exposure to melatonin and antipsychotics were included in the minimally adjusted model. Covariates significantly associated in the minimally adjusted model were included in a final adjusted model.

RESULTS

A total of 449 admissions to the medical ICU were included in the analysis. Exposure to melatonin or antipsychotic agents did not reduce the duration of ICU delirium, ICU/hospital length of stay, or 28-day mortality. However, antipsychotic use only was associated with longer hospitalization.

CONCLUSION

Antipsychotic drugs for the treatment ICU delirium may not provide the benefit documented in earlier literature. Further investigation on patient selection, type of antipsychotic, and dosing is needed.

Improving Quality of Life in Patients at Risk for Post-Intensive Care Syndrome

Objective

To improve quality of life (QOL) in patients at risk for post-intensive care syndrome (PICS).

Patients and Methods

We conducted a mixed-method, prospective, observational, pre-post interventional study in an adult medical and mixed medical/surgical/transplant intensive care unit (ICU) at a tertiary academic hospital. Preintervention included patients admitted from October 1 through October 31, 2016, and postintervention included patients admitted from January 15 through February 14, 2017. First, a multidisciplinary team of stakeholders identified barriers associated with decreased QOL in patients at risk for PICS. Next, interventions were designed and implemented. The effect of interventions was assessed using a mixed-method analysis. The qualitative analysis used a modified grounded theory approach. The quantitative analysis included assessment of preexisting symptoms and risk factors associated with PICS. The 36-Item Short-Form Health Status Survey (SF-36), which surveys physical and mental composite scores, was used to assess QOL.

Results

Barriers identified were lack of awareness and understanding of PICS. Interventions included educational videos, paper and online education and treatment materials, and online and in-person support groups for education and treatment. After interventions, the qualitative analysis found that patients who participated in the interventions after hospital discharge showed improved QOL, whereas education during hospitalization alone was not effective. The quantitative analysis did not find improvement in QOL, as defined by SF-36 physical or mental composite scores.

Conclusion

Interventions targeted to patients after hospitalization may offer subjective improvement in QOL for those at risk for PICS.

Development and performance of a novel vasopressor-driven mortality prediction model in septic shock

Background

Vasoactive medications are essential in septic shock, but are not fully incorporated into current mortality prediction risk scores. We sought to develop a novel mortality prediction model for septic shock incorporating quantitative vasoactive medication usage.

Methods

Quantitative vasopressor use was calculated in a cohort of 5352 septic shock patients and compared using norepinephrine equivalents (NEE), cumulative vasopressor index and the vasoactive inotrope score models. Having best discrimination prediction, log₁₀NEE was selected for further development of a novel prediction model for 28-day and 1-year mortality via backward stepwise logistic regression. This model termed ‘MAVIC’ (Mechanical ventilation, Acute Physiology And Chronic Health Evaluation-III, Vasopressors, Inotropes, Charlson comorbidity index) was then compared to Acute Physiology And Chronic Health Evaluation-III (APACHE-III) and Sequential Organ Failure Assessment (SOFA) scores in an independent validation cohort for its accuracy in predicting 28-day and 1-year mortality.

Measurements and main results

The MAVIC model was superior to the APACHE-III and SOFA scores in its ability to predict 28-day mortality (area under receiver operating characteristic curve [AUROC] 0.73 vs. 0.66 and 0.60) and 1-year mortality (AUROC 0.74 vs. 0.66 and 0.60), respectively.

Conclusions

The incorporation of quantitative vasopressor usage into a novel ‘MAVIC’ model results in superior 28-day and 1-year mortality risk prediction in a large cohort of patients with septic shock.

Vasopressor magnitude predicts poor outcome in adults with congenital heart disease after cardiac surgery

BACKGROUND

High levels of vasoactive inotrope support (VIS) after congenital heart surgery are predictive of morbidity in pediatric patients. We sought to discern if this relationship applies to adults with congenital heart disease (ACHD).

METHODS

We retrospectively studied adult patients (≥18 years old) admitted to the intensive care unit after cardiac surgery for congenital heart disease from 2002 to 2013 at Mayo Clinic. Vasoactive medication dose values within 96 hours of admission were examined to determine the relationship between VIS score and poor outcome of early mortality, early morbidity, or complication related morbidity.

RESULTS

Overall, 1040 ACHD patients had cardiac surgery during the study time frame; 243 (23.4%) met study inclusion criteria. Sixty-two patients (25%), experienced composite poor outcome [including eight deaths within 90 days of hospital discharge (3%)]. Thirty-eight patients (15%) endured complication related early morbidity. The maximum VIS (maxVIS) score area under the curve was 0.92 (95% CI: 0.86-0.98) for in-hospital mortality; and 0.82 (95% CI: 0.76-0.89) for combined poor clinical outcome. On univariate analysis, maxVIS score ≥3 was predictive of composite adverse outcome (OR: 14.2, 95% CI: 7.2-28.2; P < 0.001), prolonged ICU LOS ICU LOS (OR: 19.2; 95% CI: 8.7-42.1; P < 0.0001), prolonged mechanical ventilation (OR: 13.6; 95% CI: 4.4-41.8; P < 0.0001) and complication related morbidity (OR: 7.3; 95% CI: 3.4-15.5; P < 0.0001).

CONCLUSIONS

MaxVIS score strongly predicted adverse outcomes and can be used as a risk prediction tool to facilitate early intervention that may improve outcome and assist with clinical decision making for ACHD patients after cardiac surgery.

Comparison of Mortality Risk Prediction Among Patients ≥70 Versus <70 Years of Age in a Cardiac Intensive Care Unit

Older adults account for an increasing number of cardiac intensive care unit (CICU) admissions. This study sought to determine the predictive value of illness severity scores for mortality in CICU patients ≥70 years of age. Adult patients admitted to the CICU from 2007 to 2015 at one tertiary care hospital were reviewed. Severity of illness scores were calculated on the first CICU day. Area under the receiver-operator characteristic curve (AUROC) values were used to assess discrimination for hospital mortality in patients ≥70 versus <70 years of age. We included 10,004 patients with a mean age of 67.4 ± 15.2 years (37.4% female); 4,771 patients (47.7%) were ≥70 years of age. Patients ≥70 years of age had greater illness severity and more extensive co-morbidities compared with patients <70 years of age. Patients ≥70 years of age had higher hospital mortality (11.6% vs 6.8%, odds ratio 1.80, 95% confidence interval 1.57 to 2.07, p <0.001), with a progressive increase in mortality as a function of decade. Severity of illness scores had lower AUROC values for hospital mortality in patients ≥70 years of age compared with patients <70 years of age (all p <0.05 by DeLong test). The Braden skin score on CICU admission predicted hospital mortality with an AUROC value only slightly lower than these scores. Increasing age decade was associated with decreased postdischarge survival by Kaplan-Meier analysis (p <0.001 by log-rank). In conclusion, contemporary CICU patients ≥70 years of age have greater illness severity, more co-morbidities and higher mortality than patients <70 years of age, yet severity of illness scores are less accurate for predicting mortality in CICU patients ≥70 years of age, emphasizing the need for more effective risk-stratification methods in this population.

Big data and targeted machine learning in action to assist medical decision in the ICU

Historically, personalised medicine has been synonymous with pharmacogenomics and oncology. We argue for a new framework for personalised medicine analytics that capitalises on more detailed patient-level data and leverages recent advances in causal inference and machine learning tailored towards decision support applicable to critically ill patients. We discuss how advances in data technology and statistics are providing new opportunities for asking more targeted questions regarding patient treatment, and how this can be applied in the intensive care unit to better predict patient-centred outcomes, help in the discovery of new treatment regimens associated with improved outcomes, and ultimately how these rules can be learned in real-time for the patient.

Features of Adult Hyperammonemia Not Due to Liver Failure in the ICU

OBJECTIVES

To evaluate the epidemiology of hyperammonemia unrelated to liver failure in the critical care setting.

DESIGN

Retrospective case series.

SETTING

Critically ill patients admitted to ICUs at Mayo Clinic, Rochester, MN (medical ICU, two mixed medical-surgical ICUs, coronary care unit, or the cardiosurgical ICU) between July 1, 2004, and October 31, 2015.

PATIENTS

Adult critically ill patients with hyperammonemia not related to acute or chronic liver failure. We excluded patients with diagnosis of moderate or severe liver disease, hyperbilirubinemia, and patients who denied the use of their medical records.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Of 3,908 ICU patients with hyperammonemia, 167 (4.5%) had no evidence of acute or chronic liver failure. One-hundred one patients (60.5%) were male with median age of 65.7 years (interquartile range, 50-74.5 yr) and median serum ammonia level of 68 µg/dL (interquartile range, 58-87 µg/dL). Acute encephalopathy was present in 119 patients (71%). Predisposing conditions included malnutrition 27 (16%), gastric bypass six (3.6%), total parenteral nutrition four (2.4%); exposure to valproic acid 17 (10%); status epilepticus 11 (6.6%), high tumour burden 19 (11.3%), and renal failure 82 (49.1%). Urea cycle defects were diagnosed in seven patients (4.1%). Hospital mortality was high (30%), and median ammonia level was higher among the nonsurvivors (74 vs 67 µg/dL; p = 0.05). Deaths were more likely in hyperammonemic patients who were older (p = 0.016), had greater illness severity (higher Acute Physiology and Chronic Health Evaluation III score, p < 0.01), malignancy (p < 0.01), and solid organ transplantation (p = 0.04), whereas seizure disorder was more common in survivors (p = 0.02). After adjustment, serum ammonia level was not associated with increased mortality.

CONCLUSIONS

Hyperammonemia occurs in a substantial minority of critically ill patients without liver failure. These patients have a poor prognosis, although ammonia level per se is not independently associated with mortality. Serum ammonia should be measured when risk factors are present, such as nutritional deficiencies and protein refeeding, treatment with valproic acid, high tumour burden, and known or suspected urea cycle abnormalities.

Delirium in Postoperative Patients Admitted to the Intensive Care Unit

Postoperative delirium (POD) is common among surgical patients admitted to the intensive care unit (ICU) and is associated with increased resource utilization, morbidity, and death. Our primary aim was to compare rates of POD using administrative International Classification of Diseases, Ninth Revision, records and automated interrogation of electronic health records from Confusion Assessment Method for the ICU (CAM-ICU) screening. The secondary aim was to assess POD risk associated with patient and perioperative characteristics. Electronic health records of surgical patients admitted to the ICU during 2011 through 2014 were abstracted for POD assessment by CAM-ICU and by administrative codes, Charlson comorbidity index, surgical characteristics, and Acute Physiology, Age, Chronic Health Evaluation III scores. Of 6338 patients, CAM-ICU identified 606 (9.6%) and administrative records identified 55 (0.9%) POD cases, with agreement on 50 cases. In multivariable logistic regression based on POD identified with CAM-ICU, preexisting dementia had the strongest association with POD (odds ratio [95% confidence interval], 6.47 [3.68–11.37]; P < 0.001). Other associations found were older age, congestive heart failure, chronic pulmonary disease, increased surgical duration, emergency cases, blood transfusions, postoperative ventilation, and higher Acute Physiology, Age, Chronic Health Evaluation III scores (all P ≤ 0.01). POD cases had lengthier ICU and hospital stays and a higher mortality rate (all P < 0.001). CAM-ICU scores identified higher rates of POD than a search for POD based on administrative codes. Preoperative presence of dementia and major comorbidities were associated with POD. Delirium in surgical patients is associated with worse outcomes.

Seeking out SARI: an automated search of electronic health records

The definition of severe acute respiratory infection (SARI) - a respiratory illness with fever and cough, occurring within the past 10 days and requiring hospital admission - has not been evaluated for critically ill patients. Using integrated electronic health records data, we developed an automated search algorithm to identify SARI cases in a large cohort of critical care patients and evaluate patient outcomes. We conducted a retrospective cohort study of all admissions to a medical intensive care unit from August 2009 through March 2016. Subsets were randomly selected for deriving and validating a search algorithm, which was compared with temporal trends in laboratory-confirmed influenza to ensure that SARI was correlated with influenza. The algorithm was applied to the cohort to identify clinical differences for patients with and without SARI. For identifying SARI, the algorithm (sensitivity, 86.9%; specificity, 95.6%) outperformed billing-based searching (sensitivity, 73.8%; specificity, 78.8%). Automated searching correlated with peaks in laboratory-confirmed influenza. Adjusted for severity of illness, SARI was associated with more hospital, intensive care unit and ventilator days but not with death or dismissal to home. The search algorithm accurately identified SARI for epidemiologic study and surveillance.

Clinical Outcomes of Weight-Based Norepinephrine Dosing in Underweight and Morbidly Obese Patients: A Propensity-Matched Analysis

BACKGROUND

Weight-based dosing strategy for norepinephrine in septic shock patients with extremes of body mass index has been lesser studied.

METHODS

This historical study of adult septic shock patients was conducted from January 1, 2010, to December 31, 2015, at all intensive care units (ICUs) in Mayo Clinic, Rochester. Patients with documented body mass index were classified into underweight (body mass index <18.5 kg/m²), normal weight (18.5-24.9 kg/m²), and morbidly obese (≥40 kg/m²) patients. Patients with repeat ICU admissions, ICU stay <1 day, and body mass index 25 to 39.9 kg/m² were excluded. The primary outcome was in-hospital mortality, and secondary outcomes included cumulative norepinephrine exposure acute kidney injury, cardiac arrhythmias, and 1-year mortality. Two-tailed P < .05 was considered statistically significant.

RESULTS

From 2010 to 2015, 2016 patients met inclusion-145, 1406, and 466 patients, respectively, in underweight, normal weight, and morbidly obese cohorts. Underweight patients used the highest peak dose and absolute exposure was greatest for morbidly obese patients. In-hospital mortality decreased with increasing log₁₀ body mass index: 41.4% (underweight), 28.4% (normal weight), and 24.7% (morbidly obese), respectively (P < .001); however, this relationship was not noted at 1 year. Unadjusted log₁₀ norepinephrine cumulative exposure (mg) was associated with higher in-hospital mortality, acute kidney injury, cardiac arrhythmias, and 1-year mortality. After adjustment for demographics, body mass index, comorbidity, and illness severity, log₁₀ norepinephrine exposure was an independent predictor of in-hospital mortality (odds ratio 2.4 [95% confidence interval, 2.0-2.8]; P < .001) and 1-year mortality (odds ratio 1.7 [95% confidence interval, 1.5-2.0]; P < .001). In a propensity-matched analysis of 1140 patients, log₁₀ norepinephrine was an independent predictor of in-hospital mortality (odds ratio 2.2 [95% confidence interval, 1.8-2.6]; P < .001).

CONCLUSIONS

Morbidly obese patients had lower in-hospital mortality but had higher 1-year mortality compared to normal weight and underweight patients. Cumulative norepinephrine exposure was highest in morbidly obese patients. Total norepinephrine exposure was an independent mortality predictor in septic shock.

The Incremental Burden of Acute Respiratory Distress Syndrome: Long-term Follow-up of a Population-Based Nested Case-Control Study

OBJECTIVE

To evaluate the long-term survival of patients at similar risk for hospital-acquired acute respiratory distress syndrome (ARDS) who did and did not develop ARDS.

METHODS

We conducted long-term follow-up of a population-based nested case-control study in a consecutive cohort of adult Olmsted County, Minnesota, patients admitted from January 1, 2001, through December 31, 2010. Patients in whom ARDS developed during their hospital stay (cases) were matched to similar-risk patients without ARDS (controls) by 6 characteristics: age, sex, sepsis, high-risk surgery, ratio of oxygen saturation to fraction of inspired oxygen, and ARDS risk according to the Lung Injury Prediction Score. Hospital mortality, discharge disposition, and long-term survival were compared.

RESULTS

Patients who developed hospital-acquired ARDS (n=400) had higher hospital mortality than at-risk controls (n=400) (35% vs 5%; P<.001). Among hospital survivors (252 matched pairs), ARDS cases were more likely to be discharged to rehabilitation (13% vs 4%) and long-term care (30% vs 15%) facilities, whereas more controls were discharged home (71% vs 41%). After discharge, differences in survival persisted beyond 90 days (adjusted hazard ratio [HR], 1.76; 95% CI, 1.2-2.5; P=.002) and 6 months (adjusted HR, 1.73; 95% CI, 1.2-2.6; P<.001).

CONCLUSION

These results suggest that in a population-based matched case-control study of patients with similar characteristics at the time of hospital admission, those who developed hospital-acquired ARDS had worse long-term survival.

Predictive Value of the Sequential Organ Failure Assessment Score for Mortality in a Contemporary Cardiac Intensive Care Unit Population

BACKGROUND

Optimal methods of mortality risk stratification in patients in the cardiac intensive care unit (CICU) remain uncertain. We evaluated the ability of the Sequential Organ Failure Assessment (SOFA) score to predict mortality in a large cohort of unselected patients in the CICU.

METHODS AND RESULTS

Adult patients admitted to the CICU from January 1, 2007, to December 31, 2015, at a single tertiary care hospital were retrospectively reviewed. SOFA scores were calculated daily, and Acute Physiology and Chronic Health Evaluation (APACHE)-III and APACHE-IV scores were calculated on CICU day 1. Discrimination of hospital mortality was assessed using area under the receiver-operator characteristic curve values. We included 9961 patients, with a mean age of 67.5±15.2 years; all-cause hospital mortality was 9.0%. Day 1 SOFA score predicted hospital mortality, with an area under the receiver-operator characteristic curve value of 0.83; area under the receiver-operator characteristic curve values were similar for the APACHE-III score, and APACHE-IV predicted mortality (P>0.05). Mean and maximum SOFA scores over multiple CICU days had greater discrimination for hospital mortality (P<0.01). Patients with an increasing SOFA score from day 1 and day 2 had higher mortality. Patients with day 1 SOFA score <2 were at low risk of mortality. Increasing tertiles of day 1 SOFA score predicted higher long-term mortality (P<0.001 by log-rank test).

CONCLUSIONS

The day 1 SOFA score has good discrimination for short-term mortality in unselected patients in the CICU, which is comparable to APACHE-III and APACHE-IV. Advantages of the SOFA score over APACHE include simplicity, improved discrimination using serial scores, and prediction of long-term mortality.

Informatics for the Modern Intensive Care Unit

Advanced informatics systems can help improve health care delivery and the environment of care for critically ill patients. However, identifying, testing, and deploying advanced informatics systems can be quite challenging. These processes often require involvement from a collaborative group of health care professionals of varied disciplines with knowledge of the complexities related to designing the modern and "smart" intensive care unit (ICU). In this article, we explore the connectivity environment within the ICU, middleware technologies to address a host of patient care initiatives, and the core informatics concepts necessary for both the design and implementation of advanced informatics systems.

Platform for Efficient Switching between Multiple Devices in the Intensive Care Unit

Introduction: This article is part of the Focus Theme of Methods of Information in Medicine on “Managing Interoperability and Complexity in Health Systems”.

Objectives: Handheld computers, such as tablets and smartphones, are becoming more and more accessible in the clinical care setting and in Intensive Care Units (ICUs). By making the most useful and appropriate data available on multiple devices and facilitate the switching between those devices, staff members can efficiently integrate them in their workflow, allowing for faster and more accurate decisions. This paper addresses the design of a platform for the efficient switching between multiple devices in the ICU. The key functionalities of the platform are the integration of the platform into the workflow of the medical staff and providing tailored and dynamic information at the point of care.

Methods: The platform is designed based on a 3-tier architecture with a focus on extensibility, scalability and an optimal user experience. After identification to a device using Near Field Communication (NFC), the appropriate medical information will be shown on the selected device. The visualization of the data is adapted to the type of the device. A web-centric approach was used to enable extensibility and portability.

Results: A prototype of the platform was thoroughly evaluated. The scalability, performance and user experience were evaluated. Performance tests show that the response time of the system scales linearly with the amount of data. Measurements with up to 20 devices have shown no performance loss due to the concurrent use of multiple devices.

Conclusions: The platform provides a scalable and responsive solution to enable the efficient switching between multiple devices. Due to the web-centric approach new devices can easily be integrated. The performance and scalability of the platform have been evaluated and it was shown that the response time and scalability of the platform was within an acceptable range.

Clinical profile and outcomes of acute cardiorenal syndrome type-5 in sepsis: An eight-year cohort study

BACKGROUND

To evaluate the clinical features and outcomes of acute cardiorenal syndrome type-5 in patients with severe sepsis and septic shock.

METHODS

Historical cohort study of all adult patients with severe sepsis and septic shock admitted to the intensive care units (ICU) at Mayo Clinic Rochester from January 1, 2007 through December 31, 2014. Patients with prior renal or cardiac dysfunction were excluded. Patients were divided into groups with and without cardiorenal syndrome type-5. Acute Kidney Injury (AKI) was defined by both serum creatinine and urine output criteria of the AKI Network and the cardiac injury was determined by troponin-T levels. Outcomes included in-hospital mortality, ICU and hospital length of stay, and one-year survival.

RESULTS

Of 602 patients meeting the study inclusion criteria, 430 (71.4%) met criteria for acute cardiorenal syndrome type-5. Patients with cardiorenal syndrome type-5 had higher severity of illness, greater vasopressor and mechanical ventilation use. Cardiorenal syndrome type-5 was associated higher unadjusted in-hospital mortality, ICU and hospital lengths of stay, and lower one-year survival. When adjusted for age, gender, severity of illness and mechanical ventilation, cardiorenal syndrome type-5 was independently associated with 1.7-times greater odds of in-hospital mortality (p = .03), but did not predict one-year survival (p = .06) compared to patients without cardiorenal syndrome.

CONCLUSIONS

In sepsis, acute cardiorenal syndrome type-5 is associated with worse in-hospital mortality compared to patients without cardiorenal syndrome.

Spurious Hyperchloremia in the Presence of Elevated Plasma Salicylate: A Cohort Study

BACKGROUND

Acute metabolic acidosis is rarely associated with a reduced or negative anion gap (AG), but several case reports have described such an abnormality occurring in the setting of acute salicylate intoxication. The underlying cause of this phenomenon is unclear.

METHODS

In this retrospective cohort study, we reviewed our institutional database to identify all patients admitted for salicylate intoxication at Mayo Clinic (Rochester, MN, USA) from January 2010 through December 2012. Serum chloride was measured with the Cobas INTEGRA 400 plus electrode (expedited laboratory test) or Cobas 6000 (routine laboratory test). We compared blood chloride levels measured by the 2 devices in the presence of positive blood salicylate level.

RESULTS

Twelve adult patients with salicylate levels >20 mg/dL had markedly elevated chloride concentrations. The median (interquartile range) chloride level at admission was 120 (107-145) mmol/L on their initial laboratory studies, resulting in reduced or even negative AGs. None of the patients had bromide toxicity, nor did they have any other identifiable cause of hyperchloremia or decreased AG. Further testing of the same blood samples with an alternative measurement system (Roche Cobas 6000) yielded normal chloride values, indicating that falsely elevated chloride values with the initial testing led to the diminished or negative AG values.

CONCLUSION

Circulating levels of salicylate can interfere with chloride measured by using routine techniques, resulting in spurious hyperchloremia outcomes and erroneous AG values. In patients with acute metabolic acidosis and abnormally reduced or negative AG, salicylate interference with chloride measurement should be suspected.

Precision newborn screening for lysosomal disorders

PURPOSE

The implementation of newborn screening for lysosomal disorders has uncovered overall poor specificity, psychosocial harm experienced by caregivers, and costly follow-up testing of false-positive cases. We report an informatics solution proven to minimize these issues.

METHODS

The Kentucky Department for Public Health outsourced testing for mucopolysaccharidosis type I (MPS I) and Pompe disease, conditions recently added to the recommended uniform screening panel, plus Krabbe disease, which was added by legislative mandate. A total of 55,161 specimens were collected from infants born over 1 year starting from February 2016. Testing by tandem mass spectrometry was integrated with multivariate pattern recognition software (Collaborative Laboratory Integrated Reports), which is freely available to newborn screening programs for selection of cases for which a biochemical second-tier test is needed.

RESULTS

Of five presumptive positive cases, one was affected with infantile Krabbe disease, two with Pompe disease, and one with MPS I. The remaining case was a heterozygote for the latter condition. The false-positive rate was 0.0018% and the positive predictive value was 80%.

CONCLUSION

Postanalytical interpretive tools can drastically reduce false-positive outcomes, with preliminary evidence of no greater risk of false-negative events, still to be verified by long-term surveillance.

Determinants of Endotracheal Intubation in Critically Ill Patients Undergoing Gastrointestinal Endoscopy Under Conscious Sedation

OBJECTIVES

Our primary aim was to determine the factors leading to prophylactic endotracheal intubation in intensive care unit (ICU) patients undergoing gastrointestinal endoscopy. Secondary aims were to determine the rate of unplanned endotracheal intubations during endoscopy and to determine the rate of aspiration following endoscopy for patients admitted to the ICU.

METHODS

Critically ill adult (≥18 years) patients who underwent upper and lower endoscopic procedures from January 2012 to July 2016 in a medical/surgical ICU were included. Determinants of prophylactic endotracheal intubation prior to endoscopy as well as other postprocedure outcomes were electronically captured by a validated data mart system. Given our focus on aspiration in those who were not endotracheally intubated prior to endoscopy, we used a validated definition a priori.

RESULTS

A total of 320 patients were included in the final analysis: 76(24%) were intubated prior to endoscopy and 244 (76%) were not. The endotracheally intubated group had a significantly higher Acute Physiologic and Chronic Health Evaluation III (44.5 [16.2] vs 39.5 [15.5]; P = .02) and Sequential Organ Failure Assessment (6.9 [4.4] vs 3.8 [3]; P ≤ .01) scores, higher rate of hematemesis within 24 hours of endoscopy (28 [37%] vs 45 [18%]; P ≤ .01), and higher rate of upper endoscopy (72 [96%] vs 181 [74%]; P ≤ .01). We composed a composite outcome for multivariable analyses, which demonstrated the rate of any complication was significantly higher among those who were intubated prior to the procedure versus those who were not intubated previously (odds ratio: 2.80, 95% confidence interval (CI): 1.16-6.72, P = .02).

CONCLUSION

Endoscopy performed in the ICU without endotracheal intubation is safe. However, patient selection for prophylactic intubation prior to endoscopy is of critical importance as illustrated in this study with higher illness severity, planned upper endoscopy, and hematemesis 24 hours prior being key factors on deciding to perform endotracheal intubation. Prophylactic intubation for endoscopy and preexisting cardiac disease were associated with a higher rate of adverse outcomes.

Can a Novel ICU Data Display Positively Affect Patient Outcomes and Save Lives?

The aim of this study was to quantify the impact of ProCCESs AWARE, Ambient Clinical Analytics, Rochester, MN, a novel acute care electronic medical record interface, on a range of care process and patient health outcome metrics in intensive care units (ICUs). ProCCESs AWARE is a novel acute care EMR interface that contains built-in tools for error prevention, practice surveillance, decision support and reporting. We compared outcomes before and after AWARE implementation using a prospective cohort and a historical control. The study population included all critically ill adult patients (over 18 years old) admitted to four ICUs at Mayo Clinic, Rochester, MN, who stayed in hospital at least 24 h. The pre-AWARE cohort included 983 patients from 2010, and the post-AWARE cohort included 856 patients from 2014. We analyzed patient health outcomes, care process quality, and hospital charges. After adjusting for patient acuity and baseline demographics, overall in-hospital and ICU mortality odds ratios associated with AWARE intervention were 0.45 (95% confidence interval 0.30 to 0.70) and 0.38 (0.22, 0.66). ICU length of stay decreased by about 50%, hospital length of stay by 37%, and total charges for hospital stay by 30% in post AWARE cohort (by $43,745 after adjusting for patient acuity and demographics). Better organization of information in the ICU with systems like AWARE has the potential to improve important patient outcomes, such as mortality and length of stay, resulting in reductions in costs of care.

Role of Admission Troponin-T and Serial Troponin-T Testing in Predicting Outcomes in Severe Sepsis and Septic Shock

BACKGROUND

Troponin-T elevation is seen commonly in sepsis and septic shock patients admitted to the intensive care unit. We sought to evaluate the role of admission and serial troponin-T testing in the prognostication of these patients.

METHODS AND RESULTS

This was a retrospective cohort study from 2007 to 2014 on patients admitted to the intensive care units at the Mayo Clinic with severe sepsis and septic shock. Elevated admission troponin-T and significant delta troponin-T were defined as ≥0.01 ng/mL and ≥0.03 ng/mL in 3 hours, respectively. The primary outcome was in-hospital mortality. Secondary outcomes included 1-year mortality and lengths of stay. During this 8-year period, 944 patients met the inclusion criteria with 845 (90%) having an admission troponin-T ≥0.01 ng/mL. Serial troponin-T values were available in 732 (78%) patients. Elevated admission troponin-T was associated with older age, higher baseline comorbidity, and severity of illness, whereas significant delta troponin-T was associated with higher severity of illness. Admission log10 troponin-T was associated with unadjusted in-hospital (odds ratio 1.6; P=0.003) and 1-year mortality (odds ratio 1.3; P=0.04), but did not correlate with length of stay. Elevated delta troponin-T and log10 delta troponin-T were not significantly associated with any of the primary or secondary outcomes. Admission log10 troponin-T remained an independent predictor of in-hospital mortality (odds ratio 1.4; P=0.04) and 1-year survival (hazard ratio 1.3; P=0.008).

CONCLUSIONS

In patients with sepsis and septic shock, elevated admission troponin-T was associated with higher short- and long-term mortality. Routine serial troponin-T testing did not add incremental prognostic value in these patients.