Clinical examination, critical care ultrasonography and outcomes in the critically ill: cohort profile of the Simple Intensive Care Studies-I

Deep embedded clustering generalisability and adaptation for integrating mixed datatypes: two critical care cohorts

We validated a Deep Embedded Clustering (DEC) model and its adaptation for integrating mixed datatypes (in this study, numerical and categorical variables). Deep Embedded Clustering (DEC) is a promising technique capable of managing extensive sets of variables and non-linear relationships. Nevertheless, DEC cannot adequately handle mixed datatypes. Therefore, we adapted DEC by replacing the autoencoder with an X-shaped variational autoencoder (XVAE) and optimising hyperparameters for cluster stability. We call this model "X-DEC". We compared DEC and X-DEC by reproducing a previous study that used DEC to identify clusters in a population of intensive care patients. We assessed internal validity based on cluster stability on the development dataset. Since generalisability of clustering models has insufficiently been validated on external populations, we assessed external validity by investigating cluster generalisability onto an external validation dataset. We concluded that both DEC and X-DEC resulted in clinically recognisable and generalisable clusters, but X-DEC produced much more stable clusters.

External Validation of Mortality Prediction Models for Critical Illness Reveals Preserved Discrimination but Poor Calibration

OBJECTIVES

In a recent scoping review, we identified 43 mortality prediction models for critically ill patients. We aimed to assess the performances of these models through external validation.

DESIGN

Multicenter study.

SETTING

External validation of models was performed in the Simple Intensive Care Studies-I (SICS-I) and the Finnish Acute Kidney Injury (FINNAKI) study.

PATIENTS

The SICS-I study consisted of 1,075 patients, and the FINNAKI study consisted of 2,901 critically ill patients.

MEASUREMENTS AND MAIN RESULTS

For each model, we assessed: 1) the original publications for the data needed for model reconstruction, 2) availability of the variables, 3) model performance in two independent cohorts, and 4) the effects of recalibration on model performance. The models were recalibrated using data of the SICS-I and subsequently validated using data of the FINNAKI study. We evaluated overall model performance using various indexes, including the (scaled) Brier score, discrimination (area under the curve of the receiver operating characteristics), calibration (intercepts and slopes), and decision curves. Eleven models (26%) could be externally validated. The Acute Physiology And Chronic Health Evaluation (APACHE) II, APACHE IV, Simplified Acute Physiology Score (SAPS)-Reduced (SAPS-R)' and Simplified Mortality Score for the ICU models showed the best scaled Brier scores of 0.11' 0.10' 0.10' and 0.06' respectively. SAPS II, APACHE II, and APACHE IV discriminated best; overall discrimination of models ranged from area under the curve of the receiver operating characteristics of 0.63 (0.61-0.66) to 0.83 (0.81-0.85). We observed poor calibration in most models, which improved to at least moderate after recalibration of intercepts and slopes. The decision curve showed a positive net benefit in the 0-60% threshold probability range for APACHE IV and SAPS-R.

CONCLUSIONS

In only 11 out of 43 available mortality prediction models, the performance could be studied using two cohorts of critically ill patients. External validation showed that the discriminative ability of APACHE II, APACHE IV, and SAPS II was acceptable to excellent, whereas calibration was poor.

Right ventricular strain measurements in critically ill patients: an observational SICS sub-study

BACKGROUND

Right ventricular (RV) dysfunction is common in critically ill patients and is associated with poor outcomes. RV function is usually evaluated by Tricuspid Annular Plane Systolic Excursion (TAPSE) which can be obtained using critical care echocardiography (CCE). Myocardial deformation imaging, measuring strain, is suitable for advanced RV function assessment and has widely been studied in cardiology. However, it is relatively new for the Intensive Care Unit (ICU) and little is known about RV strain in critically ill patients. Therefore, the objectives of this study were to evaluate the feasibility of RV strain in critically ill patients using tissue-Doppler imaging (TDI) and explore the association between RV strain and conventional CCE measurements representing RV function.

METHODS

This is a single-center sub-study of two prospective observational cohorts (Simple Intensive Care Studies (SICS)-I and SICS-II). All acutely admitted adults with an expected ICU stay over 24 h were included. CCE was performed within 24 h of ICU admission. In patients in which CCE was performed, TAPSE, peak systolic velocity at the tricuspid annulus (RV s') and TDI images were obtained. RV free wall longitudinal strain (RVFWSL) and RV global four-chamber longitudinal strain (RV4CSL) were measured during offline analysis.

RESULTS

A total of 171 patients were included. Feasibility of RVFWSL and RV4CSL was, respectively, 62% and 56% in our population; however, when measurements were performed, intra- and inter-rater reliability based on the intraclass correlation coefficient were good to excellent. RV dysfunction based on TAPSE or RV s' was found in 56 patients (33%) and 24 patients (14%) had RV dysfunction based on RVFWSL or RV4CSL. In 14 patients (8%), RVFWSL, RV4CSL, or both were reduced, despite conventional RV function measurements being preserved. These patients had significantly higher severity of illness scores. Sensitivity analysis with fractional area change showed similar results.

CONCLUSIONS

TDI RV strain imaging in critically ill patients is challenging; however, good-to-excellent reproducibility was shown when measurements were adequately obtained. Future studies are needed to elucidate the diagnostic and prognostic value of RV strain in critically ill patients, especially to outweigh the difficulty and effort of imaging against the clinical value.

Identifying and characterizing high-risk clusters in a heterogeneous ICU population with deep embedded clustering

Critically ill patients constitute a highly heterogeneous population, with seemingly distinct patients having similar outcomes, and patients with the same admission diagnosis having opposite clinical trajectories. We aimed to develop a machine learning methodology that identifies and provides better characterization of patient clusters at high risk of mortality and kidney injury. We analysed prospectively collected data including co-morbidities, clinical examination, and laboratory parameters from a minimally-selected population of 743 patients admitted to the ICU of a Dutch hospital between 2015 and 2017. We compared four clustering methodologies and trained a classifier to predict and validate cluster membership. The contribution of different variables to the predicted cluster membership was assessed using SHapley Additive exPlanations values. We found that deep embedded clustering yielded better results compared to the traditional clustering algorithms. The best cluster configuration was achieved for 6 clusters. All clusters were clinically recognizable, and differed in in-ICU, 30-day, and 90-day mortality, as well as incidence of acute kidney injury. We identified two high mortality risk clusters with at least 60%, 40%, and 30% increased. ICU, 30-day and 90-day mortality, and a low risk cluster with 25–56% lower mortality risk. This machine learning methodology combining deep embedded clustering and variable importance analysis, which we made publicly available, is a possible solution to challenges previously encountered by clustering analyses in heterogeneous patient populations and may help improve the characterization of risk groups in critical care.

Incidence, prognostic factors, and outcomes of venous thromboembolism in critically ill patients: data from two prospective cohort studies

Background

The objective of this study was to describe the prevalence, incidence, prognostic factors, and outcomes of venous thromboembolism in critically ill patients receiving contemporary thrombosis prophylaxis.

Methods

We conducted a pooled analysis of two prospective cohort studies. The outcomes of interest were in-hospital pulmonary embolism or lower extremity deep vein thrombosis (PE-LDVT), in-hospital nonleg deep vein thrombosis (NLDVT), and 90-day mortality. Multivariable logistic regression analysis was used to evaluate the association between predefined baseline prognostic factors and PE-LDVT or NLDVT. Cox regression analysis was used to evaluate the association between PE-LDVT or NLDVT and 90-day mortality.

Results

A total of 2208 patients were included. The prevalence of any venous thromboembolism during 3 months before ICU admission was 3.6% (95% CI 2.8–4.4%). Out of 2166 patients, 47 (2.2%; 95% CI 1.6–2.9%) developed PE-LDVT and 38 patients (1.8%; 95% CI 1.2–2.4%) developed NLDVT. Renal replacement therapy (OR 3.5 95% CI 1.4–8.6), respiratory failure (OR 2.0; 95% CI 1.1–3.8), and previous VTE (OR 3.6; 95% CI 1.7–7.7) were associated with PE-LDVT. Central venous catheters (OR 5.4; 95% CI 1.7–17.8) and infection (OR 2.2; 95% CI 1.1–4.3) were associated with NLDVT. Occurrence of PE-LDVT but not NLDVT was associated with increased 90-day mortality (HR 2.7; 95% CI 1.6–4.6, respectively, 0.92; 95% CI 0.41–2.1).

Conclusion

Thrombotic events are common in critically ill patients, both before and after ICU admittance. Development of PE-LDVT but not NLDVT was associated with increased mortality. Prognostic factors for developing PE-LDVT or NLDVT despite prophylaxis can be identified at ICU admission and may be used to select patients at higher risk in future randomized clinical trials.

Trial registration

NCT03773939.

Temporal artery temperature measurements versus bladder temperature in critically ill patients, a prospective observational study

Purpose

Accurate measurement of body temperature is important for the timely detection of fever or hypothermia in critically ill patients. In this prospective study, we evaluated whether the agreement between temperature measurements obtained with TAT (test method) and bladder catheter-derived temperature measurements (BT; reference method) is sufficient for clinical practice in critically ill patients.

Methods

Patients acutely admitted to the Intensive Care Unit were included. After BT was recorded TAT measurements were performed by two independent researchers (TAT_1; TAT₂). The agreement between TAT and BT was assessed using Bland-Altman plots. Clinical acceptable limits of agreement (LOA) were defined a priori (<0.5°C). Subgroup analysis was performed in patients receiving norepinephrine.

Results

In total, 90 critically ill patients (64 males; mean age 62 years) were included. The observed mean difference (TAT-BT; ±SD, 95% LOA) between TAT and BT was 0.12°C (-1.08°C to +1.32°C) for TAT₁ and 0.14°C (-1.05°C to +1.33°C) for TAT₂. 36% (TAT₁) and 42% (TAT₂) of all paired measurements failed to meet the acceptable LOA of 0.5°C. Subgroup analysis showed that when patients were receiving intravenous norepinephrine, the measurements of the test method deviated more from the reference method (p = NS).

Conclusion

The TAT is not sufficient for clinical practice in critically ill adults.

Clinical Examination for the Prediction of Mortality in the Critically Ill: The Simple Intensive Care Studies-I

OBJECTIVES

Caregivers use clinical examination to timely recognize deterioration of a patient, yet data on the prognostic value of clinical examination are inconsistent. In the Simple Intensive Care Studies-I, we evaluated the association of clinical examination findings with 90-day mortality in critically ill patients.

DESIGN

Prospective single-center cohort study.

SETTING

ICU of a single tertiary care level hospital between March 27, 2015, and July 22, 2017.

PATIENTS

All consecutive adults acutely admitted to the ICU and expected to stay for at least 24 hours.

INTERVENTIONS

A protocolized clinical examination of 19 clinical signs conducted within 24 hours of admission.

MEASUREMENTS MAIN RESULTS

Independent predictors of 90-day mortality were identified using multivariable logistic regression analyses. Model performance was compared with established prognostic risk scores using area under the receiver operating characteristic curves. Robustness of our findings was tested by internal bootstrap validation and adjustment of the threshold for statistical significance. A total of 1,075 patients were included, of whom 298 patients (28%) had died at 90-day follow-up. Multivariable analyses adjusted for age and norepinephrine infusion rate demonstrated that the combination of higher respiratory rate, higher systolic blood pressure, lower central temperature, altered consciousness, and decreased urine output was independently associated with 90-day mortality (area under the receiver operating characteristic curves = 0.74; 95% CI, 0.71-0.78). Clinical examination had a similar discriminative value as compared with the Simplified Acute Physiology Score-II (area under the receiver operating characteristic curves = 0.76; 95% CI, 0.73-0.79; p = 0.29) and Acute Physiology and Chronic Health Evaluation-IV (using area under the receiver operating characteristic curves = 0.77; 95% CI, 0.74-0.80; p = 0.16) and was significantly better than the Sequential Organ Failure Assessment (using area under the receiver operating characteristic curves = 0.67; 95% CI, 0.64-0.71; p < 0.001).

CONCLUSIONS

Clinical examination has reasonable discriminative value for assessing 90-day mortality in acutely admitted ICU patients. In our study population, a single, protocolized clinical examination had similar prognostic abilities compared with the Simplified Acute Physiology Score-II and Acute Physiology and Chronic Health Evaluation-IV and outperformed the Sequential Organ Failure Assessment score.

Non-invasive oscillometric versus invasive arterial blood pressure measurements in critically ill patients: A post hoc analysis of a prospective observational study

PURPOSE

The aim was to compare non-invasive blood pressure measurements with invasive blood pressure measurements in critically ill patients.

METHODS

Non-invasive blood pressure was measured via automated brachial cuff oscillometry, and simultaneously the radial arterial catheter-derived measurement was recorded as part of a prospective observational study. Measurements of systolic arterial pressure (SAP), diastolic arterial pressure (DAP), and mean arterial pressure (MAP) were compared using Bland-Altman and error grid analyses.

RESULTS

Paired measurements of blood pressure were available for 736 patients. Observed mean difference (±SD, 95% limits of agreement) between oscillometrically and invasively measured blood pressure was 0.8 mmHg (±15.7 mmHg, -30.2 to 31.7 mmHg) for SAP, -2.9 mmHg (±11.0 mmHg, -24.5 to 18.6 mmHg) for DAP, and -1.0 mmHg (±10.2 mmHg, -21.0 to 18.9 mmHg) for MAP. Error grid analysis showed that the proportions of measurements in risk zones A to E were 78.3%, 20.7%, 1.0%, 0%, and 0.1% for MAP.

CONCLUSION

Non-invasive blood pressure measurements using brachial cuff oscillometry showed large limits of agreement compared to invasive measurements in critically ill patients. Error grid analysis showed that measurement differences between oscillometry and the arterial catheter would potentially have triggered at least low-risk treatment decisions in one in five patients.

Plasma neutrophil gelatinase-associated lipocalin at intensive care unit admission as a predictor of acute kidney injury progression

Background

Acute kidney injury (AKI) is a common complication in patients during intensive care unit (ICU) admission. AKI is defined as an increase in serum creatinine (SCr) and/or a reduction in urine output. SCr is a marker of renal function with several limitations, which led to the search for biomarkers for earlier AKI detection. Our aim was to study the predictive value of plasma neutrophil gelatinase-associated lipocalin (NGAL) at admission as a biomarker for AKI progression during the first 48 h of ICU admission in an unselected, heterogeneous ICU patient population.

Methods

We conducted a prospective observational study in an academic tertiary referral ICU population. We recorded AKI progression in all ICU patients during the first 48 h of ICU admission in a 6-week period. Plasma NGAL was measured at admission but levels were not reported to the attending clinicians. As possible predictors of AKI progression, pre-existing AKI risk factors were recorded. We examined the association of clinical parameters and plasma NGAL levels at ICU admission with the incidence and progression of AKI within the first 48 h of the ICU stay.

Results

A total of 361 patients were included. Patients without AKI progression during the first 48 h of ICU admission had median NGAL levels at admission of 115 ng/mL [interquartile range (IQR) 81–201]. Patients with AKI progression during the first 48 h of ICU admission had median NGAL levels at admission of 156 ng/mL (IQR 97–267). To predict AKI progression, a multivariant model with age, sex, diabetes mellitus, body mass index, admission type, Acute Physiology and Chronic Health Evaluation score and SCr at admission had an area under the receiver operating characteristics (ROC) curve of 0.765. Adding NGAL to this model showed a small increase in the area under the ROC curve to 0.783 (95% confidence interval 0.714–0.853).

Conclusions

NGAL levels at admission were higher in patients with progression of AKI during the first 48 h of ICU admission, but adding NGAL levels at admission to a model predicting this AKI progression showed no significant additive value.

Should the ultrasound probe replace your stethoscope? A SICS-I sub-study comparing lung ultrasound and pulmonary auscultation in the critically ill

Background

In critically ill patients, auscultation might be challenging as dorsal lung fields are difficult to reach in supine-positioned patients, and the environment is often noisy. In recent years, clinicians have started to consider lung ultrasound as a useful diagnostic tool for a variety of pulmonary pathologies, including pulmonary edema. The aim of this study was to compare lung ultrasound and pulmonary auscultation for detecting pulmonary edema in critically ill patients.

Methods

This study was a planned sub-study of the Simple Intensive Care Studies-I, a single-center, prospective observational study. All acutely admitted patients who were 18 years and older with an expected ICU stay of at least 24 h were eligible for inclusion. All patients underwent clinical examination combined with lung ultrasound, conducted by researchers not involved in patient care. Clinical examination included auscultation of the bilateral regions for crepitations and rhonchi. Lung ultrasound was conducted according to the Bedside Lung Ultrasound in Emergency protocol. Pulmonary edema was defined as three or more B lines in at least two (bilateral) scan sites. An agreement was described by using the Cohen κ coefficient, sensitivity, specificity, negative predictive value, positive predictive value, and overall accuracy. Subgroup analysis were performed in patients who were not mechanically ventilated.

Results

The Simple Intensive Care Studies-I cohort included 1075 patients, of whom 926 (86%) were eligible for inclusion in this analysis. Three hundred seven of the 926 patients (33%) fulfilled the criteria for pulmonary edema on lung ultrasound. In 156 (51%) of these patients, auscultation was normal. A total of 302 patients (32%) had audible crepitations or rhonchi upon auscultation. From 130 patients with crepitations, 86 patients (66%) had pulmonary edema on lung ultrasound, and from 209 patients with rhonchi, 96 patients (46%) had pulmonary edema on lung ultrasound. The agreement between auscultation findings and lung ultrasound diagnosis was poor (κ statistic 0.25). Subgroup analysis showed that the diagnostic accuracy of auscultation was better in non-ventilated than in ventilated patients.

Conclusion

The agreement between lung ultrasound and auscultation is poor.

Trial registration

NCT02912624. Registered on September 23, 2016.

Feasibility of cardiac output measurements in critically ill patients by medical students

BACKGROUND

Critical care ultrasonography (CCUS) is increasingly applied also in the intensive care unit (ICU) and performed by non-experts, including even medical students. There is limited data on the training efforts necessary for novices to attain images of sufficient quality. There is no data on medical students performing CCUS for the measurement of cardiac output (CO), a hemodynamic variable of importance for daily critical care.

OBJECTIVE

The aim of this study was to explore the agreement of cardiac output measurements as well as the quality of images obtained by medical students in critically ill patients compared to the measurements obtained by experts in these images.

METHODS

In a prospective observational cohort study, all acutely admitted adults with an expected ICU stay over 24 h were included. CCUS was performed by students within 24 h of admission. CCUS included the images required to measure the CO, i.e., the left ventricular outflow tract (LVOT) diameter and the velocity time integral (VTI) in the LVOT. Echocardiography experts were involved in the evaluation of the quality of images obtained and the quality of the CO measurements.

RESULTS

There was an opportunity for a CCUS attempt in 1155 of the 1212 eligible patients (95%) and in 1075 of the 1212 patients (89%) CCUS examination was performed by medical students. In 871 out of 1075 patients (81%) medical students measured CO. Experts measured CO in 783 patients (73%). In 760 patients (71%) CO was measured by both which allowed for comparison; bias of CO was 0.0 L min-1 with limits of agreement of - 2.6 L min-1 to 2.7 L min-1. The percentage error was 50%, reflecting poor agreement of the CO measurement by students compared with the experts CO measurement.

CONCLUSIONS

Medical students seem capable of obtaining sufficient quality CCUS images for CO measurement in the majority of critically ill patients. Measurements of CO by medical students, however, had poor agreement with expert measurements. Experts remain indispensable for reliable CO measurements. Trial registration Clinicaltrials.gov; http://www.clinicaltrials.gov; registration number NCT02912624.

Clinical examination findings as predictors of acute kidney injury in critically ill patients

BACKGROUND

Acute Kidney Injury (AKI) in critically ill patients is associated with a markedly increased morbidity and mortality. The aim of this study was to establish the predictive value of clinical examination for AKI in critically ill patients.

METHODS

This was a sub-study of the SICS-I, a prospective observational cohort study of critically ill patients acutely admitted to the Intensive Care Unit (ICU). Clinical examination was performed within 24 hours of ICU admission. The occurrence of AKI was determined at day two and three after admission according to the KDIGO definition including serum creatinine and urine output. Multivariable regression modeling was used to assess the value of clinical examination for predicting AKI, adjusted for age, comorbidities and the use of vasopressors.

RESULTS

A total of 1003 of 1075 SICS-I patients (93%) were included in this sub-study. 414 of 1003 patients (41%) fulfilled the criteria for AKI. Increased heart rate (OR 1.12 per 10 beats per minute increase, 98.5% CI 1.04-1.22), subjectively cold extremities (OR 1.52, 98.5% CI 1.07-2.16) and a prolonged capillary refill time on the sternum (OR 1.89, 98.5% CI 1.01-3.55) were associated with AKI. This multivariable analysis yielded an area under the receiver-operating curve (AUROC) of 0.70 (98.5% CI 0.66-0.74). The model performed better when lactate was included (AUROC of 0.72, 95%CI 0.69-0.75), P = .04.

CONCLUSION

Clinical examination findings were able to predict AKI with moderate accuracy in a large cohort of critically ill patients. Findings of clinical examination on ICU admission may trigger further efforts to help predict developing AKI.

Burden of acute kidney injury and 90-day mortality in critically ill patients

Background

Mortality rates associated with acute kidney injury (AKI) vary among critically ill patients. Outcomes are often not corrected for severity or duration of AKI. Our objective was to analyse whether a new variable, AKI burden, would outperform 1) presence of AKI, 2) highest AKI stage, or 3) AKI duration in predicting 90-day mortality.

Methods

Kidney Diseases: Improving Global Outcomes (KDIGO) criteria using creatinine, urine output and renal replacement therapy were used to diagnose AKI. AKI burden was defined as AKI stage multiplied with the number of days that each stage was present (maximum five), divided by the maximum possible score yielding a proportion. The AKI burden as a predictor of 90-day mortality was assessed in two independent cohorts (Finnish Acute Kidney Injury, FINNAKI and Simple Intensive Care Studies I, SICS-I) by comparing four multivariate logistic regression models that respectively incorporated either the presence of AKI, the highest AKI stage, the duration of AKI, or the AKI burden.

Results

In the FINNAKI cohort 1096 of 2809 patients (39%) had AKI and 90-day mortality of the cohort was 23%. Median AKI burden was 0.17 (IQR 0.07–0.50), 1.0 being the maximum. The model including AKI burden (area under the receiver operator curve (AUROC) 0.78, 0.76–0.80) outperformed the models using AKI presence (AUROC 0.77, 0.75–0.79, p = 0.026) or AKI severity (AUROC 0.77, 0.75–0.79, p = 0.012), but not AKI duration (AUROC 0.77, 0.75–0.79, p = 0.06). In the SICS-I, 603 of 1075 patients (56%) had AKI and 90-day mortality was 28%. Median AKI burden was 0.19 (IQR 0.08–0.46). The model using AKI burden performed better (AUROC 0.77, 0.74–0.80) than the models using AKI presence (AUROC 0.75, 0.71–0.78, p = 0.001), AKI severity (AUROC 0.76, 0.72–0.79. p = 0.008) or AKI duration (AUROC 0.76, 0.73–0.79, p = 0.009).

Conclusion

AKI burden, which appreciates both severity and duration of AKI, was superior to using only presence or the highest stage of AKI in predicting 90-day mortality. Using AKI burden or other more granular methods may be helpful in future epidemiological studies of AKI.

DEBATE-statistical analysis plans for observational studies

Background

All clinical research benefits from transparency and validity. Transparency and validity of studies may increase by prospective registration of protocols and by publication of statistical analysis plans (SAPs) before data have been accessed to discern data-driven analyses from pre-planned analyses.

Main message

Like clinical trials, recommendations for SAPs for observational studies increase the transparency and validity of findings. We appraised the applicability of recently developed guidelines for the content of SAPs for clinical trials to SAPs for observational studies. Of the 32 items recommended for a SAP for a clinical trial, 30 items (94%) were identically applicable to a SAP for our observational study. Power estimations and adjustments for multiplicity are equally important in observational studies and clinical trials as both types of studies usually address multiple hypotheses. Only two clinical trial items (6%) regarding issues of randomisation and definition of adherence to the intervention did not seem applicable to observational studies. We suggest to include one new item specifically applicable to observational studies to be addressed in a SAP, describing how adjustment for possible confounders will be handled in the analyses.

Conclusion

With only few amendments, the guidelines for SAP of a clinical trial can be applied to a SAP for an observational study. We suggest SAPs should be equally required for observational studies and clinical trials to increase their transparency and validity.

Disagreement in cardiac output measurements between fourth-generation FloTrac and critical care ultrasonography in patients with circulatory shock: a prospective observational study

Background

Cardiac output measurements may inform diagnosis and provide guidance of therapeutic interventions in patients with hemodynamic instability. The FloTrac™ algorithm uses uncalibrated arterial pressure waveform analysis to estimate cardiac output. Recently, a new version of the algorithm has been developed. The aim was to assess the agreement between FloTrac™ and routinely performed cardiac output measurements obtained by critical care ultrasonography in patients with circulatory shock.

Methods

A prospective observational study was performed in a tertiary hospital from June 2016 to January 2017. Adult critically ill patients with circulatory shock were eligible for inclusion. Cardiac output was measured simultaneously using FloTrac™ with a fourth-generation algorithm (CO_AP) and critical care ultrasonography (CO_CCUS). The strength of linear correlation of both methods was determined by the Pearson coefficient. Bland-Altman plot and four-quadrant plot were used to track agreement and trending ability.

Result

Eighty-nine paired cardiac output measurements were performed in 17 patients during their first 24 h of admittance. CO_AP and CO_CCUS had strong positive linear correlation (r² = 0.60, p < 0.001). Bias of CO_AP and CO_CCUS was 0.2 L min⁻¹ (95% CI − 0.2 to 0.6) with limits of agreement of − 3.6 L min⁻¹ (95% CI − 4.3 to − 2.9) to 4.0 L min⁻¹ (95% CI 3.3 to 4.7). The percentage error was 65.6% (95% CI 53.2 to 77.3). Concordance rate was 64.4%.

Conclusions

In critically ill patients with circulatory shock, there was disagreement and clinically unacceptable trending ability between values of cardiac output obtained by uncalibrated arterial pressure waveform analysis and critical care ultrasonography.

Trial registration

Clinicaltrials.gov, NCT02912624, registered on September 23, 2016

Electronic supplementary material

The online version of this article (10.1186/s40560-019-0373-5) contains supplementary material, which is available to authorized users.

Associations between tricuspid annular plane systolic excursion to reflect right ventricular function and acute kidney injury in critically ill patients: a SICS-I sub-study

BACKGROUND

Acute kidney injury (AKI) occurs in up to 50% of all critically ill patients and hemodynamic abnormalities are assumed to contribute, but their nature and share is still unclear. We explored the associations between hemodynamic variables, including cardiac index and right ventricular function, and the occurrence of AKI in critically ill patients.

METHODS

In this prospective cohort study, we included all patients acutely admitted to an intensive care unit (ICU). Within 24 h after ICU admission clinical and hemodynamic variables were registered including ultrasonographic measurements of cardiac index and right ventricular function, assessed using tricuspid annular plane systolic excursion (TAPSE) and right ventricular systolic excursion (RV S'). Maximum AKI stage was assessed according to the KDIGO criteria during the first 72 h after admission. Multivariable logistic regression modeling was used including both known predictors and univariable significant predictors of AKI. Secondary outcomes were days alive outside ICU and 90-day mortality.

RESULTS

A total of 622 patients were included, of which 338 patients (54%) had at least AKI stage 1 within 72 h after ICU admission. In the final multivariate model higher age (OR 1.01, 95% CI 1.00-1.03, for each year), higher weight (OR 1.03 CI 1.02-1.04, for each kg), higher APACHE IV score (OR 1.02, CI 1.01-1.03, per point), lower mean arterial pressure (OR 1.02, CI 1.01-1.03, for each mmHg decrease) and lower TAPSE (OR 1.05, CI 1.02-1.09 per millimeter decrease) were all independent predictors for AKI in the final multivariate logistic regression model. Sepsis, cardiac index, RV S' and use of vasopressors were not significantly associated with AKI in our data. AKI patients had fewer days alive outside of ICU, and their mortality rate was significantly higher than those without AKI.

CONCLUSIONS

In our cohort of acutely admitted ICU patients, the incidence of AKI was 54%. Hemodynamic variables were significantly different between patients with and without AKI. A worse right ventricle function was associated with AKI in the final model, whereas cardiac index was not.

The diagnostic accuracy of clinical examination for estimating cardiac index in critically ill patients: the Simple Intensive Care Studies-I

PURPOSE

Clinical examination is often the first step to diagnose shock and estimate cardiac index. In the Simple Intensive Care Studies-I, we assessed the association and diagnostic performance of clinical signs for estimation of cardiac index in critically ill patients.

METHODS

In this prospective, single-centre cohort study, we included all acutely ill patients admitted to the ICU and expected to stay > 24 h. We conducted a protocolised clinical examination of 19 clinical signs followed by critical care ultrasonography for cardiac index measurement. Clinical signs were associated with cardiac index and a low cardiac index (< 2.2 L min^-1 m²) in multivariable analyses. Diagnostic test accuracies were also assessed.

RESULTS

We included 1075 patients, of whom 783 (73%) had a validated cardiac index measurement. In multivariable regression, respiratory rate, heart rate and rhythm, systolic and diastolic blood pressure, central-to-peripheral temperature difference, and capillary refill time were statistically independently associated with cardiac index, with an overall R² of 0.30 (98.5% CI 0.25-0.35). A low cardiac index was observed in 280 (36%) patients. Sensitivities and positive and negative predictive values were below 90% for all signs. Specificities above 90% were observed only for 110/280 patients, who had atrial fibrillation, systolic blood pressures < 90 mmHg, altered consciousness, capillary refill times > 4.5 s, or skin mottling over the knee.

CONCLUSIONS

Seven out of 19 clinical examination findings were independently associated with cardiac index. For estimation of cardiac index, clinical examination was found to be insufficient in multivariable analyses and in diagnostic accuracy tests. Additional measurements such as critical care ultrasonography remain necessary.

Expert statement for the management of hypovolemia in sepsis

Hypovolemia is frequent in patients with sepsis and may contribute to worse outcome. The management of these patients is impeded by the low quality of the evidence for many of the specific components of the care. In this paper, we discuss recent advances and controversies in this field and give expert statements for the management of hypovolemia in patients with sepsis including triggers and targets for fluid therapy and volumes and types of fluid to be given. Finally, we point to unanswered questions and suggest a roadmap for future research.