Development and validation of the creatinine clearance predictor machine learning models in critically ill adults

BACKGROUND

In critically ill patients, measured creatinine clearance (CrCl) is the most reliable method to evaluate glomerular filtration rate in routine clinical practice and may vary subsequently on a day-to-day basis. We developed and externally validated models to predict CrCl one day ahead and compared them with a reference reflecting current clinical practice.

METHODS

A gradient boosting method (GBM) machine-learning algorithm was used to develop the models on data from 2825 patients from the EPaNIC multicenter randomized controlled trial database. We externally validated the models on 9576 patients from the University Hospitals Leuven, included in the M@tric database. Three models were developed: a "Core" model based on demographic, admission diagnosis, and daily laboratory results; a "Core + BGA" model adding blood gas analysis results; and a "Core + BGA + Monitoring" model also including high-resolution monitoring data. Model performance was evaluated against the actual CrCl by mean absolute error (MAE) and root-mean-square error (RMSE).

RESULTS

All three developed models showed smaller prediction errors than the reference. Assuming the same CrCl of the day of prediction showed 20.6 (95% CI 20.3-20.9) ml/min MAE and 40.1 (95% CI 37.9-42.3) ml/min RMSE in the external validation cohort, while the developed model having the smallest RMSE (the Core + BGA + Monitoring model) had 18.1 (95% CI 17.9-18.3) ml/min MAE and 28.9 (95% CI 28-29.7) ml/min RMSE.

CONCLUSIONS

Prediction models based on routinely collected clinical data in the ICU were able to accurately predict next-day CrCl. These models could be useful for hydrophilic drug dosage adjustment or stratification of patients at risk.

TRIAL REGISTRATION

Not applicable.

Development and validation of clinical prediction models for acute kidney injury recovery at hospital discharge in critically ill adults

PURPOSE

Acute kidney injury (AKI) recovery prediction remains challenging. The purpose of the present study is to develop and validate prediction models for AKI recovery at hospital discharge in critically ill patients with ICU-acquired AKI stage 3 (AKI-3).

METHODS

Models were developed and validated in a development cohort (n = 229) and a matched validation cohort (n = 244) from the multicenter EPaNIC database to create prediction models with the least absolute shrinkage and selection operator (Lasso) machine-learning algorithm. We evaluated the discrimination and calibration of the models and compared their performance with plasma neutrophil gelatinase-associated lipocalin (NGAL) measured on first AKI-3 day (NGAL_AKI3) and reference model that only based on age.

RESULTS

Complete recovery and complete or partial recovery occurred in 33.20% and 51.23% of the validation cohort patients respectively. The prediction model for complete recovery based on age, need for renal replacement therapy (RRT), diagnostic group (cardiac/surgical/trauma/others), and sepsis on admission had an area under the receiver operating characteristics curve (AUROC) of 0.53. The prediction model for complete or partial recovery based on age, need for RRT, platelet count, urea, and white blood cell count had an AUROC of 0.61. NGAL_AKI3 showed AUROCs of 0.55 and 0.53 respectively. In cardiac patients, the models had higher AUROCs of 0.60 and 0.71 than NGAL_AKI3's AUROCs of 0.52 and 0.54. The developed models demonstrated a better performance over the reference models (only based on age) for cardiac surgery patients, but not for patients with sepsis and for a general ICU population.

CONCLUSION

Models to predict AKI recovery upon hospital discharge in critically ill patients with AKI-3 showed poor performance in the general ICU population, similar to the biomarker NGAL. In cardiac surgery patients, discrimination was acceptable, and better than NGAL. These findings demonstrate the difficulty of predicting non-reversible AKI early.

Technical failure of the EZ‐blocker™ causing serious adverse events during one lung ventilation: a case series

We present a case series of intra-operative adverse events while using a specific type of bronchial blocker, designed to facilitate device positioning and minimise the risk of dislocation. The Rüsch^® EZ-blocker™ (Teleflex Life Sciences Ltd., Athlone, Ireland) is a Y-shaped catheter equipped with two separately inflatable cuffs at the tip - one for each bronchial lumen. In this report, we describe four cases where the use of the EZ-blocker was associated with the development of high airway pressures, hypoxaemia and expansion of the non-dependent lung. Bronchoscopic evaluation showed spontaneous inflation of the cuff within the dependent (i.e. ventilated) bronchus, causing bronchial obstruction, and volume loss of the cuff within the non-dependent (i.e. unventilated) bronchus, causing unintended expansion of the non-dependent lung. After removal of the bronchial blocker, the catheter showed no visible defect, but a bench test revealed a functional connection inside the catheter which allowed air to pass slowly from one bronchial cuff to the other. This technical defect relates to the unique design of the EZ-blocker as it is the only bronchial blocker equipped with two bronchial cuffs. Clinicians should be aware of this inherent risk since complications may develop insidiously and affect both lungs simultaneously. Early recognition and prompt intervention can prevent life-threatening intra-operative deterioration.

Postoperative Cerebral Oxygen Saturation in Children After Congenital Cardiac Surgery and Long-Term Total Intelligence Quotient: A Prospective Observational Study

OBJECTIVES

During the early postoperative period, children with congenital heart disease can suffer from inadequate cerebral perfusion, with possible long-term neurocognitive consequences. Cerebral tissue oxygen saturation can be monitored noninvasively with near-infrared spectroscopy. In this prospective study, we hypothesized that reduced cerebral tissue oxygen saturation and increased intensity and duration of desaturation (defined as cerebral tissue oxygen saturation < 65%) during the early postoperative period, independently increase the probability of reduced total intelligence quotient, 2 years after admission to a PICU.

DESIGN

Single-center, prospective study, performed between 2012 and 2015.

SETTING

The PICU of the University Hospitals Leuven, Belgium.

PATIENTS

The study included pediatric patients after surgery for congenital heart disease admitted to the PICU.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Postoperative cerebral perfusion was characterized with the mean cerebral tissue oxygen saturation and dose of desaturation of the first 12 and 24 hours of cerebral tissue oxygen saturation monitoring. The independent association of postoperative mean cerebral tissue oxygen saturation and dose of desaturation with total intelligence quotient at 2-year follow-up was evaluated with a Bayesian linear regression model adjusted for known confounders. According to a noninformative prior, reduced mean cerebral tissue oxygen saturation during the first 12 hours of monitoring results in a loss of intelligence quotient points at 2 years, with a 90% probability (posterior β estimates [80% credible interval], 0.23 [0.04-0.41]). Similarly, increased dose of cerebral tissue oxygen saturation desaturation would result in a loss of intelligence quotient points at 2 years with a 90% probability (posterior β estimates [80% credible interval], -0.009 [-0.016 to -0.001]).

CONCLUSIONS

Increased dose of cerebral tissue oxygen saturation desaturation and reduced mean cerebral tissue oxygen saturation during the early postoperative period independently increase the probability of having a lower total intelligence quotient, 2 years after PICU admission.

Change in QRS area by cardiac resynchronization therapy is associated with clinical outcomes and echocardiographic response

Background

Cardiac Resynchronization Therapy (CRT) is the cornerstone of treatment in patients with dyssynchronous heart failure. Recently, baseline QRS area proved to predict outcomes after CRT better than QRS duration and morphology.

Purpose

It was the aim of the study to investigate whether the change in QRS area (ΔQRS area) by CRT-pacing further improves the prediction of CRT outcomes.

Methods

We conducted a retrospective analysis on 1,299 patients, who were included in a CRT-registry from three Dutch University hospitals with both pre- (baseline) and post-implantation 12-lead ECGs. ΔQRS area and ΔQRS duration were defined as the decrease in their respective values after CRT. Optimal cut offs for ΔQRS area and ΔQRS duration by means of Youden indices were found at 62μVs and −11ms, respectively. Primary endpoint was a combination of all-cause mortality, heart transplantation, and left ventricular assist device implantation. Secondary endpoint was the relative reduction in left ventricular end-systolic volume (LVESV), and echocardiographic response being defined as ≥15% LVESV reduction.

Results

The primary endpoint occurred in 408 patients (31%). ΔQRS area was superior to ΔQRS duration for the primary and secondary endpoints. Primary endpoint analysis showed a lower risk in the ΔQRS area ≥62μVs than in the &lt;62μVs group (HR 0.43; 0.33–0.56, p&lt;0.001). In the multivariable analysis, both baseline QRS area and ΔQRS area remained significantly associated with both primary and secondary endpoints. Clinical outcome (left panel of figure) and echocardiographic response (right panel) were significantly worse in patients with baseline QRS area &lt;109μVs (group 3) than in those with QRS area ≥109μVs. Within the latter group, outcomes were significantly better in patients with ΔQRS area ≥62μVs (group 1) as compared to ΔQRS area &lt;62μVs (group 2) (figure). Baseline QRS duration and ΔQRS duration were not independently associated with both clinical outcome and echocardiographic response.

Conclusion

The combination of baseline QRS area and ΔQRS area has a stronger association with CRT response than baseline QRS area alone, and (Δ)QRS duration. These results suggest that especially in patients with a good electrical substrate (large baseline QRS area) it is worthwhile to adjust CRT to achieve the largest decrease in QRS area.

Funding Acknowledgement

Type of funding source: None

Health-related quality of life of children and their parents 2 years after critical illness: pre-planned follow-up of the PEPaNIC international, randomized, controlled trial

Background

Pediatric intensive care unit (PICU) survivors are at risk for prolonged morbidities interfering with daily life. The current study examined parent-reported health-related quality of life (HRQoL) in former critically ill children and parents themselves and aimed to determine whether withholding parenteral nutrition (PN) in the first week of critical illness affected children’s and parents’ HRQoL 2 years later.

Methods

Children who participated in the pediatric early versus late parenteral nutrition in critical illness (PEPaNIC) trial and who were testable 2 years later (n = 1158) were included. Their HRQoL outcomes were compared with 405 matched healthy controls. At PICU admission, children had been randomly assigned to early-PN or late-PN. In the early-PN group, PN was initiated within 24 h after PICU admission. In the late-PN group, PN was withheld for up to 1 week in the PICU. Parents completed the Infant Toddler Quality of Life Questionnaire (ITQOL; age 2–3 years) or the Child Health Questionnaire-Parent Form 50 (CHQ-PF50; age 4–18 years). Besides, they completed the Health Utility Index (HUI) and the Short Form Health Survey (SF-12) regarding their child’s and their own HRQoL, respectively.

Results

For the total age group of 786 post-PICU survivors, parents reported lower scores for almost all HRQoL scales compared to healthy children. Age-specifically, younger critically ill children (2.5 to 3 years old) scored worse for growth and development and older children (4–18 years old) scored worse for role functioning and mental health. Parents’ own mental and physical HRQoL was comparable to that of healthy control parents. No HRQoL differences were found between children in the late-PN and those in the early-PN group.

Conclusions

Parent-reported HRQoL of children 2 years after critical illness was impaired compared with healthy controls. In relation to their child’s HRQoL, parents reported impairments in emotions, personal time, and family activities; however, their own HRQoL was not impaired. Withholding PN in the first week during critical illness had no impact on longer-term HRQoL of the child.

Trial registration

Clinical trials, NCT01536275. Registered 22 February 2012

Predicting patient nurse-level intensity for a subsequent shift in the intensive care unit: A single-centre prospective observational study

BACKGROUND

A dynamic optimized nurse staffing model for the Intensive Care Unit (ICU), requires a tool for monitoring nurse-level intensity with validated cut-offs to identify patients requiring 1:1, 2:1 or 3:1 patient-to-nurse ratios.

OBJECTIVES

We aimed to determine the Nursing Activities Score (NAS) cut-off values which can best distinguish between high, average and lower nurse-level intensity as unanimously perceived by care providers, and to evaluate whether these NAS cut-offs allow to predict nurse-level intensity in the next shift or the same shift the next day.

DESIGN

A prospective observational study.

SETTING

9 ICUs in a Belgian tertiary care center.

PARTICIPANTS

All 3295 patients admitted between March 20, 2013, and September 12, 2013 were included. NAS was quantified at the end of each shift using automatically derived and manually entered care information. Additionally, 412 nurses, 24 nurse managers and 37 physicians rated perceived nurse-level intensity.

METHODS

We first assessed concordance between nurses', nurse managers' and physicians' perceptions of lower (3:1 patient-to-nurse ratio), average (2:1 patient-to-nurse ratio) and high (1:1 patient-to-nurse ratio) nurse-level intensity. Next, receiver operating characteristic (ROC) analysis was applied to determine the NAS cut-offs that best distinguish between different levels of perceived intensity for cases with concordant opinions. Last, logistic regression analysis was applied to estimate the ability of these NAS cut-offs to predict low and high perceived intensity during the next shift and during the same shift the next day.

RESULTS

Nurses', nurse managers' and physicians' perceptions were concordant in 57.1% (n = 4693) of cases, mostly concerning lower or average intensity. Optimal NAS cut-offs for lower and high intensity patients equaled 52.7% and 69.8%, respectively. The lower intensity NAS cut-off showed 74.0% accuracy to predict lower intensity in the next shift and 75.9% accuracy to predict lower intensity for the same shift the next day. The high intensity NAS cut-off showed 67.9% accuracy to predict high intensity in the next shift and 72.0% accuracy to predict high intensity for the same shift the next day.

CONCLUSIONS

NAS cut-offs could contribute considerably in predicting patient nurse-level intensity, and thus patient-to-nurse staffing ratios, in the next shift or day. Identification or prediction of high intensity, nevertheless, appears most complex and requires further study. Future studies need to account for the many confounding variables which complicate nurse staffing planning.

MON-LB012 Long-Term Developmental Impact of Withholding Parenteral Nutrition in Pediatric-ICU: A 4-Year Follow-Up of the Pepanic Randomized Controlled Trial

Aim: Between 2012-2015, the PEPaNIC randomized controlled trial, which included 1440 critically ill infants and children, showed that withholding parenteral nutrition during the first week in the pediatric intensive care unit (PICU) (late-PN), as compared with initiating supplemental PN early (early-PN), improved PICU outcomes (1) and improved neurocognitive development assessed 2 years later (2). The latter was explained by avoiding early-PN induced adversely altered DNA-methylation of 37 CpG sites (3). As a large number of patients were younger than 1 year of age at randomization and given that assessment of most neurocognitive domains is only possible from 4 years of age onwards, we performed a 4-year follow-up to determine the impact of late-PN versus early-PN on physical, neurocognitive, and emotional/behavioral development. This pre-planned, 4-year follow-up study of the 1440 PEPaNIC patients and of 369 matched healthy children was blinded for treatment allocation (ClinicalTrials.gov-NCT01536275). Methods: Studied clinical outcomes included anthropometrics, health status, parent/caregiver-reported executive functions, and emotional/behavioral problems, and clinical tests for intelligence, visual-motor integration, alertness, motor coordination and memory. Univariable and multivariable linear and logistic regression analyses adjusted for risk factors assessed the impact of late-PN versus early-PN on the outcomes and investigated a potential mediation role of the adversely altered DNA-methylation of 37 CpG sites previously shown to be evoked by late-PN as compared with early-PN (3). Results: Overall, at 4 years follow-up, patients (356 late-PN patients, 328 early-PN patients) could be tested neurocognitively. They revealed worse anthropometric, health status, neurocognitive and emotional/behavioral developmental outcomes than the healthy control children. Outcomes of late-PN patients were never worse than those of early-PN patients. In contrast, late-PN patients had fewer internalizing (P=0.042) and externalizing problems (P=0.046), and fewer total emotional/behavioral problems (P=0.007) than early-PN patients, which were normalized by late-PN. Avoiding the early-PN induced adversely altered DNA-methylation status of the 37 CpG sites statistically explained its impact on the behavioral outcomes. Conclusion: Four years after randomization to late-PN or early-PN in the PICU, late-PN did not show harm, and was found to protect against emotional/behavioral problems, with altered DNA-methylation as a potential biological mediator hereof. These data further support de-implementation of PN-use early during critical illness in infants and children. (1) Fivez et al. N Eng J Med 2016 (2) Verstraete et al. Lancet Respir Med 2019 (3) Guiza et al. Lancet Respir Med 2020 (in press)

SAT-724 Endocrine Disruption by Phthalate Exposure in the Pediatric Intensive Care Unit

Aim: Pediatric intensive care relies on plastic indwelling medical devices softened by phthalates. Phthalates leach into the circulation and concerns about toxicity were raised. Exceeding a certain threshold of di(2-ethylhexyl)phthalate (DEHP) exposure in the pediatric intensive care unit (PICU) has been associated with an attention deficit 4 years later (1). Moreover, DEHP and its metabolites have endocrine disrupting properties. Critically ill children reveal the non-thyroidal illness syndrome (2) and unexplained relatively low cortisol (3). Whether DEHP exposure in PICU has endocrine disruptive effects is unknown. We investigated whether DEHP exposure in the PICU, exceeding the previously identified “toxic” threshold for attention, is independently associated with thyroid- and HPA-axis alterations upon PICU discharge. Methods: In this preplanned secondary analysis of the PEPaNIC RCT (N=1440) (4), plasma DEHP metabolite concentrations (MEHP, 5OH-MEHP, 5cx-MEPP, 5oxo-MEHP) were quantified for all patients with a last PICU day sample (N=920). Minimal DEHP exposure was defined as the product of the total DEHP metabolite concentrations on the last PICU day and duration of PICU stay, with 0.551 µmol/L.days identified as “toxic” threshold (1). Serum TSH, total T4, total T3 and rT3 concentrations were quantified for patients with an available last day sample (N=913). For patients with a last day plasma sample and who did not receive corticosteroids (N=391), plasma ACTH, total cortisol, albumin and CBG concentrations were quantified and free cortisol calculated. Multivariable linear regression analyses, adjusted for baseline risk factors and for duration of PICU stay, assessed whether exceeding the previously determined threshold of toxic DEHP exposure was independently associated with the hormone levels on the last PICU day. Main results: Median total DEHP metabolite concentration was 0.101 (IQR 0.049 - 0.279) µmol/L on the last PICU day. Minimal DEHP exposure was 0.337 (IQR 0.161 - 0.880) µmol/l.days, and 328 patients (35.7%) exceeded the toxic threshold. Exceeding this threshold was independently associated with lower total T4 (P=0.002), total T3 (P=0.02) and total cortisol (P=0.001), and higher rT3 (P=0.01) concentrations on the last PICU day, but not with TSH, ACTH or free cortisol. Conclusion: Critically ill children had DHEP metabolites in plasma upon PICU discharge and more than a third were exposed to toxic levels. Toxic DEHP exposure was an independent contributor to the severity of the non-thyroidal illness phenotype and to lower cortisol upon PICU discharge. Future research should assess whether such endocrine-disruptive impact of DHEP exposure in the PICU plays a role in the long-term developmental legacy of critical illness in children. 1 Verstraete et al Intensive Care Med 2016 2 Jacobs et al Thyroid 2019 3 Jacobs et al Intensive Care Med 2019 4 Fivez et al N Engl J Med 2016

OR16-4 The Growth Hormone Axis in Relation to Muscle Weakness in the ICU: Effect of Early Macronutrient Deficit

Background: The activity of the growth hormone (GH) axis is altered by critical illness. In the acute phase, GH resistance, as reflected by increased GH and decreased insulin-like growth factor-I (IGF-I), mimics fasting-induced changes in health. Although early full feeding in ICU has long been assumed to prevent muscle wasting and weakness, the EPaNIC RCT observed fewer complications and faster recovery with accepting the macronutrient deficit in the first ICU week, as compared with early full feeding, including less acquired muscle weakness [1,2]. We previously observed that accepting the early macronutrient deficit attenuated rather than aggravated the rise in GH as compared with early full feeding [3]. We have now further characterized its impact on the GH axis, in relation to the risk of acquiring muscle weakness in ICU. Methods: This was a preplanned sub-analysis of the EPaNIC RCT. For 10 matched patients per group, serum GH was quantified every 10 min between 9 PM and 6 AM, followed by deconvolution analyses to estimate GH secretion. For 564 patients per group, matched for baseline characteristics, and for all patients investigated for muscle weakness (n=600), serum IGF-I, IGF binding protein 3 (IGFBP3) and IGFBP1 were measured upon ICU admission, at day 4 (if still in ICU), and on the last ICU day (LD). Matched healthy subjects (n=65) were included as controls. Groups were compared with Wilcoxon test or repeated-measures ANOVA. Associations between changes in concentrations from baseline to day 4 or LD for patients with shorter ICU stay (d4/LD) and risk of muscle weakness were assessed with nominal logistic regression analysis, adjusted for baseline risk factors, baseline hormone concentrations and randomization. Results: Upon ICU admission, patients revealed low IGF-I and IGFBP3 and high IGFBP1 as compared with controls (p<0.001). Tolerating an early macronutrient deficit in ICU decreased basal (non-pulsatile) GH secretion (p=0.005) without affecting pulsatile GH secretion. From admission to d4/LD IGF-I and IGFBP3 increased, whereas IGFBP1 decreased (all p<0.001) in the fully fed group. Compared to full feeding, tolerating the early macronutrient deficit prevented the rise in IGF-I (p<0.001), did not affect IGFBP3 and attenuated the decrease in IGFBP1 (p<0.001). A stronger rise in GH and IGF-I from admission to d4/LD was independently associated with a lower risk of acquiring muscle weakness (OR (95%CI) per ng/ml change 0.88 (0.81-0.96) for GH, p=0.001; 0.98 (0.97-0.99) for IGF-I, p=0.002). Conclusion: Tolerating the early macronutrient deficit suppresses basal but not pulsatile GH secretion and alters IGF-I bioavailability during critical illness. These effects may counteract the protection of the intervention against the development of muscle weakness. 1. Casaer et al. N Engl J Med 2011;365:506-17 2. Hermans et al. Lancet Respir Med 2013;1:621-9 3. Van Dyck et al. ENDO 2018;SUN601

OR17-1 Changes in the Hypothalamic-Pituitary-Adrenal Axis during Pediatric Critical Illness

Critical illness in adults is hallmarked by prolonged hypercortisolism and a rapid decrease in ACTH. The high cortisol is explained by reduced cortisol breakdown and a decrease in cortisol binding proteins, rather than by adrenal production. In critically ill children, the evolution of the changes in the HPA axis and underlying mechanisms remain to be studied. Also, the impact hereon of nutritional management is unknown. In the PEPaNIC RCT, accepting low macronutrient intake up to day 8 in the pediatric intensive care unit (PICU) by withholding early supplemental parenteral nutrition (PN) accelerated recovery as compared with initiating supplemental PN early. In this preplanned secondary analysis of the PEPaNIC RCT, we assessed the changes over time in the HPA axis and their prognostic value during critical illness in children and investigated the impact on these changes of withholding early PN. We quantified plasma ACTH, total cortisol, CBG and albumin and calculated free cortisol upon PICU admission, day 3 and last PICU day for 223 short-stay (<3 days) and 309 long-stay patients (≥3 days) who did not receive corticosteroids before sampling, in comparison with 64 matched healthy children. Upon admission, ACTH was elevated in short-stay patients (P=0.02) and comparable with healthy children in long-stay patients (P=0.75), whereas total and free cortisol were elevated in both groups (P<0.0001). In short-stay patients, ACTH became subnormal on the last day (P<0.0001). Also total and free cortisol decreased, but remained higher than in healthy children (P=0.003-0.005). In long-stay patients, ACTH decreased towards day 3 (P<0.0001) and remained low on the last day (P<0.0001), whereas total and free cortisol were normal on day 3 (P=0.37-0.62) and on the last day (P=0.30-0.74). CBG and albumin were low throughout PICU stay in both groups. The rapid decrease in cortisol over time in PICU likely excludes negative feedback by cortisol on pituitary ACTH secretion as cause of the decrease in ACTH. In multivariable analysis, the decrease in ACTH was also not associated with drugs commonly used in PICU, that have been suggested to affect the HPA axis in adults. High total and free cortisol upon admission were associated with 90-day mortality and prolonged length of stay in univariable analysis. This association disappeared when adjusting for baseline risk factors. Withholding early PN did not affect the changes in the HPA axis from admission to day 3 or to last day for short-stay patients. In conclusion, the time course of the changes in the HPA axis during critical illness in children differs from that in adults, with a rapid normalization of cortisol together with a decrease in ACTH over time. Further investigation of the mechanism underlying this ACTH decrease is warranted, since it might be iatrogenic or related to a possibly more immature central secretion during critical illness in children as compared with adults.

Non-Thyroidal Illness Syndrome in Critically Ill Children: Prognostic Value and Impact of Nutritional Management

INTRODUCTION

Non-thyroidal illness (NTI), which occurs with fasting and in response to illness, is characterized by thyroid hormone inactivation with low triiodothyronine (T3) and high reverse T3 (rT3), followed by suppressed thyrotropin (TSH). Withholding supplemental parenteral nutrition early in pediatric critical illness (late-PN), thus accepting low/no macronutrient intake up to day 8 in the pediatric intensive care unit (PICU), accelerated recovery compared to initiating supplemental parenteral nutrition early (early-PN). Whether NTI is harmful or beneficial in pediatric critical illness and how it is affected by a macronutrient deficit remains unclear. This study investigated the prognostic value of NTI, the impact of late-PN on NTI, and whether such impact explains or counteracts the outcome benefit of late-PN in critically ill children.

METHODS

This preplanned secondary analysis of the Early versus Late Parenteral Nutrition in the Pediatric Intensive Care Unit randomized controlled trial quantified serum TSH, total thyroxine (T4), T3, and rT3 concentrations in 982 patients upon PICU admission versus 64 matched healthy children and in 772 propensity score-matched early-PN and late-PN patients upon admission and at day 3 or last PICU day for shorter PICU stay. Associations between thyroid hormone concentrations upon admission and outcome, as well as impact of late-PN on NTI in relation with outcome, were assessed with univariable analyses and multivariable logistic regression, linear regression, or Cox proportional hazard analysis, adjusted for baseline risk factors.

RESULTS

Upon PICU admission, critically ill children revealed lower TSH, T4, T3, and T3/rT3 and higher rT3 than healthy children (p < 0.0001). A more pronounced NTI upon admission, with low T4, T3, and T3/rT3 and high rT3 was associated with higher mortality and morbidity. Late-PN further reduced T4, T3, and T3/rT3 and increased rT3 (p ≤ 0.001). Statistically, the further lowering of T4 by late-PN reduced the outcome benefit (p < 0.0001), whereas the further lowering of T3/rT3 explained part of the outcome benefit of late-PN (p ≤ 0.004). This effect was greater for infants than for older children.

CONCLUSION

In critically ill children, the peripheral inactivation of thyroid hormone, characterized by a decrease in T3/rT3, which is further accentuated by low/no macronutrient intake, appears beneficial. In contrast, the central component of NTI attributable to suppressed TSH, evidenced by the decrease in T4, seems to be a harmful response to critical illness. Whether treating the central component with TSH releasing hormone infusion in the PICU is beneficial requires further investigation.

Mechanical Unloading Properties of Axial Flow Pumps and their Effect on Myocardial Stunning

Postischemic myocardial dysfunction affects morbidity and mortality in patients with coronary artery disease. It is known that mechanical unloading of the left heart ventricle can positively influence postischemic myocardial dysfunction. In this respect we tested two miniaturised axial flow pumps, i.e. the 14-F and the 21-F Hemopump®. An experimental study was carried out on 30 open chest sheep where regional myocardial wall motion was followed using sonomicrometry in a preparation of transient coronary artery occlusion. Only the larger 21-F Hemopump® showed hemodynamically significant unloading of the left ventricle. Furthermore, as far as stunning is concerned, systolic wall thickening recovered better when this type of pump was used during reperfusion. Also postejection thickening, which is an indication of diastolic postischemic dysfunction, is reduced significantly in the postischemic area (ANOVA, p<0.05). Thus, the 21F Hemopump®, but not the 14F Hemopump®, provides adequate mechanical unloading in order to beneficially influence myocardial stunning.

Software-guided versus nurse-directed blood glucose control in critically ill patients: the LOGIC-2 multicenter randomized controlled clinical trial

Background

Blood glucose control in the intensive care unit (ICU) has the potential to save lives. However, maintaining blood glucose concentrations within a chosen target range is difficult in clinical practice and holds risk of potentially harmful hypoglycemia. Clinically validated computer algorithms to guide insulin dosing by nurses have been advocated for better and safer blood glucose control.

Methods

We conducted an international, multicenter, randomized controlled trial involving 1550 adult, medical and surgical critically ill patients, requiring blood glucose control. Patients were randomly assigned to algorithm-guided blood glucose control (LOGIC-C, n = 777) or blood glucose control by trained nurses (Nurse-C, n = 773) during ICU stay, according to the local target range (80–110 mg/dL or 90–145 mg/dL). The primary outcome measure was the quality of blood glucose control, assessed by the glycemic penalty index (GPI), a measure that penalizes hypoglycemic and hyperglycemic deviations from the chosen target range. Incidence of severe hypoglycemia (<40 mg/dL) was the main safety outcome measure. New infections in ICU, duration of hospital stay, landmark 90-day mortality and quality of life were clinical safety outcome measures.

Results

The median GPI was lower in the LOGIC-C (10.8 IQR 6.2–16.1) than in the Nurse-C group (17.1 IQR 10.6–26.2) (P < 0.001). Mean blood glucose was 111 mg/dL (SD 15) in LOCIC-C versus 119 mg/dL (SD 21) in Nurse-C, whereas the median time-in-target range was 67.0% (IQR 52.1–80.1) in LOGIC-C versus 47.1% (IQR 28.1–65.0) in the Nurse-C group (both P < 0.001). The fraction of patients with severe hypoglycemia did not differ between LOGIC-C (0.9%) and Nurse-C (1.2%) (P = 0.6). The clinical safety outcomes did not differ between groups. The sampling interval was 2.3 h (SD 0.5) in the LOGIC-C group versus 3.0 h (SD 0.8) in the Nurse-C group (P < 0.001).

Conclusions

In a randomized controlled trial of a mixed critically ill patient population, the use of the LOGIC-Insulin blood glucose control algorithm, compared with blood glucose control by expert nurses, improved the quality of blood glucose control without increasing hypoglycemia.

Trial registration

ClinicalTrials.gov, NCT02056353. Registered on 4 February 2014.

Electronic supplementary material

The online version of this article (doi:10.1186/s13054-017-1799-6) contains supplementary material, which is available to authorized users.

AKIpredictor, an online prognostic calculator for acute kidney injury in adult critically ill patients: development, validation and comparison to serum neutrophil gelatinase-associated lipocalin

PURPOSE

Early diagnosis of acute kidney injury (AKI) remains a major challenge. We developed and validated AKI prediction models in adult ICU patients and made these models available via an online prognostic calculator. We compared predictive performance against serum neutrophil gelatinase-associated lipocalin (NGAL) levels at ICU admission.

METHODS

Analysis of the large multicenter EPaNIC database. Model development (n = 2123) and validation (n = 2367) were based on clinical information available (1) before and (2) upon ICU admission, (3) after 1 day in ICU and (4) including additional monitoring data from the first 24 h. The primary outcome was a comparison of the predictive performance between models and NGAL for the development of any AKI (AKI-123) and AKI stages 2 or 3 (AKI-23) during the first week of ICU stay.

RESULTS

Validation cohort prevalence was 29% for AKI-123 and 15% for AKI-23. The AKI-123 model before ICU admission included age, baseline serum creatinine, diabetes and type of admission (medical/surgical, emergency/planned) and had an AUC of 0.75 (95% CI 0.75-0.75). The AKI-23 model additionally included height and weight (AUC 0.77 (95% CI 0.77-0.77)). Performance consistently improved with progressive data availability to AUCs of 0.82 (95% CI 0.82-0.82) for AKI-123 and 0.84 (95% CI 0.83-0.84) for AKI-23 after 24 h. NGAL was less discriminant with AUCs of 0.74 (95% CI 0.74-0.74) for AKI-123 and 0.79 (95% CI 0.79-0.79) for AKI-23.

CONCLUSIONS

AKI can be predicted early with models that only use routinely collected clinical information and outperform NGAL measured at ICU admission. The AKI-123 models are available at http://akipredictor.com/ . Trial registration Clinical Trials.gov NCT00512122.

Emergency Department Crowding in Relation to In-hospital Adverse Medical Events: A Large Prospective Observational Cohort Study

BACKGROUND

Emergency department (ED) crowding has been linked with adverse medical events. However, this association was inadequately controlled for potential confounding variables.

OBJECTIVES

To investigate whether ED crowding is independently associated with risk of in-hospital death and morbidity, and longer total hospital stay.

METHODS

Prospective observational cohort study of all patients (≥ 18 years) presenting to the ED of an academic teaching hospital in Leuven, Belgium from June 21, 2010 to July 20, 2012. Multivariate logistic regression and proportional hazard analysis was used to control for risk factors. ED occupancy was determined for 108,229 included patients and labeled "ED crowding" when occupancy was within the highest quartile of occupancy. Outcomes within 10 days of ED admission included in-hospital death, hospital-acquired morbidities, and total hospital stay.

RESULTS

During ED crowding, a median of 58 (interquartile range 55-63) patients were present for 40 licensed treatment bays. After controlling for all baseline risk factors and as compared with the lowest quartile of ED occupancy (30 [26-32] patients), ED crowding was not independently associated with mortality (odds ratio [OR] 0.94, 95% confidence interval [CI] 0.74-1.19; p = 0.6), but tended to be associated with higher incidence of hospital-acquired pneumonia (OR 1.24, 95% CI 0.96-1.62; p = 0.09).

CONCLUSIONS

Failing to control for baseline risk factors may have led to false-positive associations between ED crowding and mortality in previous studies. After controlling for risk factors, we showed that ED crowding was associated with longer hospital stays but not with increased mortality.

Interobserver agreement of Medical Research Council sum-score and handgrip strength in the intensive care unit

INTRODUCTION

Muscle weakness often complicates critical illness and is associated with devastating short- and long-term consequences. For interventional studies, reliable measurements of muscle force in the intensive care unit (ICU) are needed.

METHODS

To examine interobserver agreement, two observers independently measured Medical Research Council (MRC) sum-score (n = 75) and handgrip strength (n = 46) in a cross-sectional ICU sample.

RESULTS

The intraclass correlation coefficient (ICC) for MRC sum-score was 0.95 (0.92-0.97). The kappa coefficient for identifying "significant weakness" (MRC sum-score <48, MRC subtotal upper limbs <24) and "severe weakness" (MRC sum-score <36) was 0.68 ± 0.09, 0.88 ± 0.07, and 0.93 ± 0.07, respectively. The ICC for left and right handgrip strength was 0.97 (0.94-0.98) and 0.93 (0.86-0.97), respectively.

CONCLUSIONS

Interobserver agreement on MRC sum-score and handgrip strength in the ICU was very good. Agreement on "severe weakness" (MRC sum-score <36) was excellent and supports its use in interventional studies. Agreement on "significant weakness" (MRC sum-score <48) was good, but even better using the equivalent cut-off in the upper limbs. It remains to be determined whether this may serve as a substitute.

Endotracheal intubation using videolaryngoscopy causes less cardiovascular response compared to classic direct laryngoscopy, in cardiac patients according a standard hospital protocol

INTRODUCTION

Previous studies comparing Glidescope and classic direct laryngoscopy did not show an attenuation of CV responses to endotracheal intubation. In the present study, we hypothesize that indirect videolaryngoscopy can attenuate cardiovascular responses to endotracheal intubation.

METHODS

In a randomized cross-over study, eighty adults (ASA PS II-III) were included. Both direct and indirect videolaryngoscopies were used in a random order, in the same patient. Cardiovascular responses to intubation were recorded as a relative change in rate pressure product (RPP = systolic blood pressure times heart rate) from baseline values. A linear mixed model was used to study the association between the outcome variable RPP and the type of laryngoscope used.

RESULTS

The relative increase of the RPP at intubation was significantly smaller (i.e. 27%, P < 0.001) using videolaryngoscopy compared to the classic direct laryngoscopy. Cardiovascular responses were blunted by an additional 10.2% (P = 0.029), when the patient was on beta blockade.

CONCLUSIONS

Our study shows less hemodynamic responses during endotracheal intubation using indirect videolaryngoscopy compared to classic direct laryngoscopy.

Prospective evaluation of intraoperative hemodynamics in liver transplantation with whole, partial and DCD grafts

The interaction of systemic hemodynamics with hepatic flows at the time of liver transplantation (LT) has not been studied in a prospective uniform way for different types of grafts. We prospectively evaluated intraoperative hemodynamics of 103 whole and partial LT. Liver graft hemodynamics were measured using the ultrasound transit time method to obtain portal (PVF) and arterial (HAF) hepatic flow. Measurements were recorded on the native liver, the portocaval shunt, following reperfusion and after biliary anastomosis. After LT HAF and PVF do not immediately return to normal values. Increased PVF was observed after graft implantation. Living donor LT showed the highest compliance to portal hyperperfusion. The amount of liver perfusion seemed to be related to the quality of the graft. A positive correlation for HAF, PVF and total hepatic blood flow with cardiac output was found (p = 0.001). Portal hypertension, macrosteatosis >30%, warm ischemia time and cardiac output, independently influence the hepatic flows. These results highlight the role of systemic hemodynamic management in LT to optimize hepatic perfusion, particularly in LDLT and split LT, where the highest flows were registered.

Near-infrared spectroscopy (NIRS) monitoring in contemporary anesthesia and critical care

Near-infrared spectroscopy (NIRS) is a noninvasive technology that continuously monitors regional tissue oxygenation. Originally used for assessment of oxygen saturation of the brain, its use has now been expanded to evaluation of oxygenation of tissues other than the brain. There is also growing evidence for the larger applicability of NIRS as an estimate of systemic venous saturation in correspondence with the adequacy of the circulatory status. New and promising advances may further this technology to become part of our standard armamentarium, in order to optimize patient care in daily anesthesia practice. The present paper briefly reviews the basic principles of operation, the inherent limitations of the technology and the clinical data that have been acquired with NIRS monitoring in the broad field of acute clinical medicine.

Anesthesia and perioperative management for a patient with Ullrich syndrome undergoing surgery for scoliosis

Ullrich syndrome is a rare congenital hypotonic-sclerotic muscular disorder in which affected children develop a slowly progressive scoliosis and contractures and limpness of joints. The disease causes increasingly invalidating contractures and hardening of the muscles of the neck and trunk. While this neuromuscular type of scoliosis is progressive, patients rarely attain the point of surgery due to their compromised general medical condition. This may explain the current lack of outcome data and the paucity of information on perioperative management for patients with Ullrich syndrome undergoing major surgery. The purpose of this report was therefore to describe our first experience with the perioperative and anesthetic management of a 15-year-old boy presenting with Ullrich syndrome and a secondary invalidating scoliosis. The specific challenges of this condition characterized by severe restrictive lung disease and a challenging airway abnormality are discussed.

Na-P(i) cotransporter type I activity causes a transient intracellular alkalinization during ATP depletion in rabbit medullary thick ascending limb cells

The cellular pathophysiology of renal ischemia-reperfusion injury was investigated in primary cell cultures from rabbit medullary thick ascending limb (MTAL). Metabolic inhibition (MI) was achieved with cyanide and 2-deoxyglucose. Sixty minutes of MI caused a profound but reversible decrease in intracellular concentration of ATP ([ATP]i). Intracellular pH (pHi) first decreased after initiation of MI, followed by a transient alkalinization. When [ATP]i reached its lowest value (<1% of control), the cells slowly acidified to reach a stable pHi of 6.92 after 50 min of MI. In the presence of EIPA (10 micromol/L), the pattern of changes in pHi was unchanged and acidification was not increased, indicating that the Na+/H+ exchangers were inactive during ATP depletion. When inorganic phosphate (P(i)) or Na+ was omitted from the apical solutions during MI, the transient alkalinization was no longer observed and the cytosol slowly acidified. Experiments on Na+-dependent alkalinizations revealed the presence of a Na-P(i) cotransporter in the apical cell membrane. With indirect immunofluorescence, the Na-P(i) cotransporter expressed in these primary cell cultures could be identified as Na-P(i) type I. Although the exact physiological role of Na-P(i) type I still is unresolved, these experiments demonstrate that apical Na-P(i) type I activity is increased at the onset of ATP depletion in MTAL cells.

Thyroid hormone receptor isoform expression in livers of critically ill patients

OBJECTIVE

The THRA gene encodes two isoforms of the thyroid hormone receptor (TR), TRalpha1 and TRalpha2. The ratio of these splice variants could have a marked influence on T3-regulated gene expression, especially during illness.

DESIGN

We studied the expression of the isoforms TRbeta1, TRalpha1, and TRalpha2 and 5'-deiodinase in postmortem liver biopsies of 58 patients who were critically ill and died in the intensive care unit (ICU). All mRNA levels were determined using real-time PCR.

MAIN OUTCOME

All ratios of the biopsies were higher than those found in three normal liver biopsies due to an increased TRalpha1 level. The TRalpha1/TRalpha2 ratio increased with age and severity of illness following the equation: TRalpha1/TRalpha2 ratio = - 1.854 + (0.0323 x age) + (0.0431 x Therapeutic Intervention Scoring System score) indicating that 28% of the changed TRalpha1/TRalpha2 ratio can be predicted by these clinical variables. There was no effect of randomization to intensive insulin therapy or glucocorticoid or thyroid hormone treatment on the TRalpha1/TRalpha2 ratio or TRbeta1. Furthermore, no relation was seen between the expression levels of the 5'-deiodinase mRNA and TR isoforms or the triiodothyronine T3 levels.

CONCLUSION

It appears that in critically ill patients the ratio of TRalpha1/TRalpha2 expression increases with age and severity of illness, possibly indicating a mechanism to enhance sensitivity to T3 in the oldest and sickest patients.

An adaptive input-output modeling approach for predicting the glycemia of critically ill patients

In this paper we apply system identification techniques in order to build a model suitable for the prediction of glycemia levels of critically ill patients admitted to the intensive care unit. These patients typically show increased glycemia levels, and it has been shown that glycemia control by means of insulin therapy significantly reduces morbidity and mortality. Based on a real-life dataset from 15 critically ill patients, an initial input-output model is estimated which captures the insulin effect on glycemia under different settings. To incorporate patient-specific features, an adaptive modeling strategy is also proposed in which the model is re-estimated at each time step (i.e., every hour). Both one-hour-ahead predictions and four-hours-ahead simulations are executed. The optimized adaptive modeling technique outperforms the general initial model. To avoid data selection bias, 500 permutations, in which the patients are randomly selected, are considered. The results are satisfactory both in terms of forecasting ability and in the clinical interpretation of the estimated coefficients.