Severity-Adjusted ICU Mortality Only Tells Half the Truth—The Impact of Treatment Limitation in a Nationwide Database

Trends in hospital mortality of patients with status epilepticus in the ICU before and during the COVID-19 pandemic

The evidence on the temporal trends of short-term mortality among critically ill patients with status epilepticus (SE) is limited and relatively dated, with data lacking on the prognostic impact of the coronavirus disease 2019 (COVID-19) pandemic in these patients. We used statewide data to examine intensive care unit (ICU) admissions with SE aged ≥ 18 years in Texas during 2016 to 2022. Interrupted time series analyses using segmented hierarchical logistic models were fit to estimate trends in hospital mortality during the prepandemic and COVID-19 pandemic periods overall and within age, sex, and race and ethnicity strata, expressed as average marginal effects (AME). Separate hierarchical models were fit to forecast hospital mortality during the pandemic period had the pandemic not occurred (counterfactual). There were 27,885 ICU admissions with SE during the study period (33.1% aged ≥ 65 years, 58.8% racial and ethnic minority, 2.4% with COVID-19, 51.8% mechanically ventilated). Overall hospital mortality was 10.6%. On interrupted time series analyses, hospital mortality decreased during the prepandemic period overall (AME -0.31%/quarter [95% confidence interval -0.39 to -0.23]) and within all demographic strata, except those aged 18 to 44 years, whose hospital mortality was unchanged. During the pandemic period hospital mortality remained unchanged over time overall (AME -0.03%/quarter [95% confidence interval -0.18 to 0.11]) and for all demographic strata, and was consistently higher than counterfactual estimates, including following the exclusion of ICU admissions with COVID-19. Hospital mortality has decreased among critically ill patients with SE prior to the pandemic, except among younger adults. However, these outcome gains were stalled by the COVID-19 pandemic, with increased hospital deaths in this population. Determination of the specific contributions of the disruption in neurological and neuro-critical care support systems and of general health system factors to the adverse outcomes of critically ill patients with SE is paramount for future mitigation and recovery efforts. These data can inform effective implementation of care protocols in SE under population-wide health crises, as well as health policy and resource allocation. Finally, as the pandemic is receding, it is crucial to determine whether there is corresponding recovery of the prepandemic mortality trends in this population.

Advance directives in the intensive care unit: An eight-year vanguard cohort study

PURPOSE

To investigate the frequency, content, and clinical translation of advance directives in intensive care units (ICUs).

MATERIAL AND METHODS

Retrospective cohort study in a Swiss tertiary ICU, including patients with advance directives treated in ICUs ≥48 h. The primary endpoint was the violation of directives. Key secondary endpoints were the directives' prevalence and their translation into clinical practice.

RESULTS

Of 5'851 patients treated ≥48 h in ICUs, 2.7 % had documented directives. Despite 92 % using templates, subjective or contradictory wording was found in 19 % and 12 %. Nine percent of directives were violated. Patients with directive violations had worse in-hospital outcomes (p = 0.012). At admission, 64 % of patients experiencing violations could not communicate, and directives were missing/unrecognized in 30 %. Mostly, directives were not followed regarding life-prolonging measures (6 %), ICU admission (5 %), and mechanical ventilation (3 %). Kaplan Meier statistics revealed a lower survival rate with directives recognized at admission (p = 0.04) and when treatment was withheld (p < 0.001).

CONCLUSIONS

Advance directives are available in a minority of ICU patients and often contain subjective/contradictory wording. Physicians respected directives in 90 % of patients, with treatment adapted following their wishes. However, violation of directives may have serious consequences with unfavorable in-hospital outcomes and decreased long-term survival with treatment adaption following directives.

Associations between Late Lactate Clearance and Clinical Outcomes in Adults with Hyperlactataemia in the Setting of Diabetic Ketoacidosis

Objective: This study aimed to determine the associations between lactate clearance in hyperlactataemic patients with diabetic ketoacidosis (DKA) and intensive care unit (ICU), hospital length of stay (LOS), and case-fatality. Methods: A retrospective, multicentre, cohort study of adult patients admitted to ICU with hyperlactataemia and a primary diagnosis of DKA from twelve sites in Queensland, Australia was conducted utilising pre-existing datasets that were linked for research purposes. The patients were divided into early and late lactate clearance groups; the early lactate clearance group included patients whose lactate returned to <2.0 mmol/L within 12 h, and the remainder were classified as late lactate clearance group. Results: The final dataset included 511 patients, 427 in the early lactate clearance group and 84 in the late lactate clearance group. Late lactate clearance was associated with increasing ICU LOS (β = +15.82, 95% CI +0.05 to +31.59, p < 0.049), increasing hospital LOS (β = +7.24, 95% CI +0.11 to 14.37, p = 0.048) and increasing Acute Physiology and Chronic Health Evaluation(APACHE) III score (ICU LOS outcome variable β = +1.05, 95% CI +0.88 to +1.22, p < 0.001; hospital LOS outcome variable β = +3.40, 95% CI +2.22 to 4.57, p < 0.001). Hospital case-fatality was not significantly different (2.2% in the early clearance group vs. 1.7% in the late clearance group, p = 0.496). Conclusions: In hyperlactataemic patients with DKA, late lactate clearance was associated with a statistically significant increase in both ICU and hospital LOS, though the clinical significance in both is minor.

Sex differences in the SOFA score of ICU patients with sepsis or septic shock: a nationwide analysis

BACKGROUND

The Sequential Organ Failure Assessment (SOFA) score is an important tool in diagnosing sepsis and quantifying organ dysfunction. However, despite emerging evidence of differences in sepsis pathophysiology between women and men, sex is currently not being considered in the SOFA score. We aimed to investigate potential sex-specific differences in organ dysfunction, as measured by the SOFA score, in patients with sepsis or septic shock and explore outcome associations.

METHODS

Retrospective analysis of sex-specific differences in the SOFA score of prospectively enrolled ICU patients with sepsis or septic shock admitted to one of 85 certified Swiss ICUs between 01/2021 and 12/2022.

RESULTS

Of 125,782 patients, 5947 (5%) were admitted with a clinical diagnosis of sepsis (2244, 38%) or septic shock (3703, 62%). Of these, 5078 (37% women) were eligible for analysis. A statistically significant difference of the total SOFA score on admission was found between women (mean 7.5 ± SD 3.6 points) and men (7.8 ± 3.6 points, Wilcoxon rank-sum p < 0.001). This was driven by differences in the coagulation (p = 0.008), liver (p < 0.001) and renal (p < 0.001) SOFA components. Differences between sexes were more prominent in younger patients < 52 years of age (women 7.1 ± 4.0 points vs men 8.1 ± 4.2 points, p = 0.004). No sex-specific differences were found in ICU length of stay (women median 2.6 days (IQR 1.3-5.3) vs men 2.7 days (IQR 1.2-6.0), p = 0.13) and ICU mortality (women 14% vs men 15%, p = 0.17).

CONCLUSION

Sex-specific differences exist in the SOFA score of patients admitted to a Swiss ICU with sepsis or septic shock, particularly in laboratory-based components. Although the clinical meaningfulness of these differences is unclear, a reevaluation of sex-specific thresholds for SOFA score components is warranted in an attempt to make more accurate and individualised classifications.

Critically ill patients with high predicted mortality: Incidence and outcome

OBJECTIVE

As calculated by the severity scores, an unknown number of patients are admitted to the Intensive Care Unit (ICU) with a very high risk of death. Clinical studies have poorly addressed this population, and their prognosis is largely unknown.

DESIGN

Post hoc analysis of a multicenter, cohort, longitudinal, observational, retrospective study (CIMbA).

SETTING

Sixteen Portuguese multipurpose ICUs.

PATIENTS

Patients with a Simplified Acute Physiology Score II (SAPS II) predicted hospital mortality above 80% on admission to the ICU (high-risk group); A comparison with the remaining patients was obtained.

INTERVENTIONS

None.

MAIN VARIABLES OF INTEREST

Hospital, 30 days, 1 year mortality.

RESULTS

We identified 4546 patients (59.9% male), 12.2% of the whole population. Their SAPS II predicted hospital mortality was 89.0±5.8%, whilst the observed mortality was lower, 61.0%. This group had higher mortality, both during the first 30 days (aHR 3.52 [95% CI 3.34-3.71]) and from day 31 to day 365 after ICU admission (aHR 1.14 [95%CI 1.04-1.26]), respectively. However, their hospital standardized mortality ratio was similar to the other patients (0.69 vs. 0.69, P=.92). At one year of follow-up, 30% of patients in the high-risk group were alive.

CONCLUSIONS

Roughly 12% of patients admitted to the ICU for more than 24h had a SAPS II score predicted mortality above 80%. Their hospital standardized mortality was similar to the less severe population and 30% were alive after one year of follow-up.

Sex Differences in Vital Organ Support Provided to ICU Patients

OBJECTIVES

Critically ill women may receive less vital organ support than men but the mortality impact of this differential treatment remains unclear. We aimed to quantify sex differences in vital organ support provided to adult ICU patients and describe the relationship between sex, vital organ support, and mortality.

DESIGN

In this retrospective observational study, we examined the provision of invasive ventilation (primary outcome), noninvasive ventilation, vasoactive medication, renal replacement therapy, extracorporeal membrane oxygenation (ECMO), or any one of these five vital organ supports in women compared with men. We performed logistic regression investigating the association of sex with each vital organ support, adjusted for illness severity, diagnosis, preexisting treatment limitation, year, and hospital. We performed logistic regression for hospital mortality adjusted for the same variables, stratified by vital organ support (secondary outcome).

SETTING AND PATIENTS

ICU admissions in the Australia and New Zealand Intensive Care Society Adult Patient Database 2018-2021. This registry records admissions from 90% of ICUs in the two nations.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We examined 699,535 ICU admissions (43.7% women) to 199 ICUs. After adjustment, women were less likely than men to receive invasive ventilation (odds ratio [OR], 0.64; 99% CI, 0.63-0.65) and each other organ support except ECMO. Women had lower adjusted hospital mortality overall (OR, 0.94; 99% CI, 0.91-0.97). Among patients who did not receive any organ support, women had significantly lower adjusted hospital mortality (OR, 0.82; 99% CI, 0.76-0.88); among patients who received any organ support women and men were equally likely to die (OR, 1.01; 99% CI, 0.97-1.04).

CONCLUSIONS

Women received significantly less vital organ support than men in ICUs in Australia and New Zealand. However, our findings suggest that women may not be harmed by this conservative approach to treatment.

Sex-related differences in adult patients with status epilepticus: a seven-year two-center observation

BACKGROUND

Conflicting findings exist regarding the influence of sex on the development, treatment, course, and outcome of status epilepticus (SE). Our study aimed to investigate sex-related disparities in adult SE patients, focusing on treatment, disease course, and outcome at two Swiss academic medical centers.

METHODS

In this retrospective study, patients treated for SE at two Swiss academic care centers from Basel and Geneva from 2015 to 2021 were included. Primary outcomes were return to premorbid neurologic function, death during hospital stay and at 30 days. Secondary outcomes included characteristics of treatment and disease course. Associations with primary and secondary outcomes were assessed using multivariable logistic regression. Analysis using propensity score matching was performed to account for the imbalances regarding age between men and women.

RESULTS

Among 762 SE patients, 45.9% were women. No sex-related differences were found between men and women, except for older age and lower frequency of intracranial hemorrhages in women. Compared to men, women had a higher median age (70 vs. 66, p = 0.003), had focal nonconvulsive SE without coma more (34.9% vs. 25.5%; p = 0.005) and SE with motor symptoms less often (52.3% vs. 63.6%, p = 0.002). With longer SE duration (1 day vs. 0.5 days, p = 0.011) and a similar proportion of refractory SE compared to men (36.9% vs. 36.4%, p = 0.898), women were anesthetized and mechanically ventilated less often (30.6% vs. 42%, p = 0.001). Age was associated with all primary outcomes in the unmatched multivariable analyses, but not female sex. In contrast, propensity score-matched multivariable analyses revealed decreased odds for return to premorbid neurologic function for women independent of potential confounders. At hospital discharge, women were sent home less (29.7% vs. 43.7%, p < 0.001) and to nursing homes more often (17.1% vs. 10.0%, p = 0.004).

CONCLUSIONS

This study identified sex-related disparities in the clinical features, treatment modalities, and outcome of adult patients with SE with women being at a disadvantage, implying that sex-based factors must be considered when formulating strategies for managing SE and forecasting outcomes.

Treatment limitations and clinical outcomes in critically ill frail patients with and without COVID-19 pneumonitis

BACKGROUND

The presence of treatment limitations in patients with frailty at intensive care unit (ICU) admission is unknown. We aimed to evaluate the presence and predictors of treatment limitations in patients with and without COVID-19 pneumonitis in those admitted to Australian and New Zealand ICUs.

METHODS

This registry-based multicenter, retrospective cohort study included all frail adults (≥16 years) with documented clinical frailty scale (CFS) scores, admitted to ICUs with admission diagnostic codes for viral pneumonia or acute respiratory distress syndrome (ARDS) over 2 years between January 01, 2020 and December 31, 2021. Frail patients (CFS ≥5) coded as having viral pneumonitis or ARDS due to COVID-19 were compared to those with other causes of viral pneumonitis or ARDS for documented treatment limitations.

RESULTS

884 frail patients were included in the final analysis from 129 public and private ICUs. 369 patients (41.7%) had confirmed COVID-19. There were more male patients in COVID-19 (55.3% vs 47.0%; p = 0.015). There were no differences in age or APACHE-III scores between the two groups. Overall, 36.0% (318/884) had treatment limitations, but similar between the two groups (35.8% [132/369] vs 36.1% [186/515]; p = 0.92). After adjusting for confounders, increasing frailty (OR = 1.72; 95%-CI 1.39-2.14), age (OR = 1.05; 95%-CI 1.04-1.06), and presence of chronic respiratory condition (OR = 1.58; 95%-CI 1.10-2.27) increased the likelihood of instituting treatment limitations. However, the presence of COVID-19 by itself did not influence treatment limitations (odds ratio [OR] = 1.39; 95%-CI 0.98-1.96).

CONCLUSIONS

The proportion of treatment limitations was similar in patients with frailty with or without COVID-19 pneumonitis at ICU admission.

Impact of Antibiotic Time-Outs in Multidisciplinary ICU Rounds for Antimicrobial Stewardship Program on Patient Survival: A Controlled Before-and-After Study

The antimicrobial stewardship program (ASP) is an important quality improvement initiative that is recommended in the ICU. However, the shortage of infectious disease physicians in Japan has led to the need for simpler methods for implementing ASPs. We investigated whether antibiotic time-outs (ATOs) during multidisciplinary rounds as part of an ASP can improve patient survival and reduce the number of days of therapy (DOT) with antibiotics.

DESIGN

Single-center controlled before-and-after study.

SETTING

Medical/surgical ICU in a tertiary university medical center in Tokyo, Japan.

PATIENTS

All patients 16 years old or older admitted consecutively in the ICU between October 2016 and March 2020.

INTERVENTIONS

An intensivist-driven ICU multidisciplinary round was introduced in October 2016, and ATOs with ICU rounds were implemented in June 2018. ATOs were conducted 3, 7, and 14 days after initiation of antibiotics.

MEASUREMENTS AND MAIN RESULTS

The primary outcome was the subdistribution hazard ratio (SHR) of survival to hospital discharge compared between multidisciplinary rounds (phase 1) and ATO during multidisciplinary rounds (phase 2) using the multivariable Fine-Gray model. The secondary outcomes were the SHR of survival to ICU discharge and the trends in the DOT with IV antibiotics per 1,000 patient-days between October 2016 and March 2020 by using interrupted time-series analysis. The number of patients in phases 1 and 2 was 777 and 796, respectively. The group that underwent ATO during multidisciplinary rounds showed a significant increase in the survival to hospital discharge in comparison with the multidisciplinary round-only group (SHR, 1.13; 95% CI, 1.02-1.25); however, the SHR of survival to ICU discharge showed no significant intergroup difference. The DOT with total IV antibiotics decreased after ATO implementation (change in intercept, -178.26; 95% CI, -317.74 to -38.78; change in slope, -7.00; 95% CI, -15.77 to 1.78).

CONCLUSIONS

ATOs during multidisciplinary rounds are associated with improved patient survival and reduced DOT.

Development and validation of a deep learning model to predict the survival of patients in ICU

BACKGROUND

Patients in the intensive care unit (ICU) are often in critical condition and have a high mortality rate. Accurately predicting the survival probability of ICU patients is beneficial to timely care and prioritizing medical resources to improve the overall patient population survival. Models developed by deep learning (DL) algorithms show good performance on many models. However, few DL algorithms have been validated in the dimension of survival time or compared with traditional algorithms.

METHODS

Variables from the Early Warning Score, Sequential Organ Failure Assessment Score, Simplified Acute Physiology Score II, Acute Physiology and Chronic Health Evaluation (APACHE) II, and APACHE IV models were selected for model development. The Cox regression, random survival forest (RSF), and DL methods were used to develop prediction models for the survival probability of ICU patients. The prediction performance was independently evaluated in the MIMIC-III Clinical Database (MIMIC-III), the eICU Collaborative Research Database (eICU), and Shanghai Pulmonary Hospital Database (SPH).

RESULTS

Forty variables were collected in total for model development. 83 943 participants from 3 databases were included in the study. The New-DL model accurately stratified patients into different survival probability groups with a C-index of >0.7 in the MIMIC-III, eICU, and SPH, performing better than the other models. The calibration curves of the models at 3 and 10 days indicated that the prediction performance was good. A user-friendly interface was developed to enable the model's convenience.

CONCLUSIONS

Compared with traditional algorithms, DL algorithms are more accurate in predicting the survival probability during ICU hospitalization. This novel model can provide reliable, individualized survival probability prediction.

Critical illness in patients with metastatic cancer: a population-based cohort study of epidemiology and outcomes

The appropriateness of intensive care unit (ICU) admission of patients with metastatic cancer remains debated. We aimed to examine the short-term outcomes and their temporal pattern in critically ill patients with metastatic disease. We used state-wide data to identify hospitalizations aged ≥18 years with metastatic cancer admitted to ICU in Texas during 2010-2014. Multivariable logistic regression modeling was used to examine the factors associated with short-term mortality and its temporal trends among all ICU admissions and those undergoing mechanical ventilation. Among 136,644 ICU admissions with metastatic cancer, 50.8% were aged ≥65 years, with one or more organ failures present in 53.3% and mechanical ventilation was used in 11.1%. The crude short-term mortality among all ICU admissions and those mechanically ventilated was 28.1% and 62.0%, respectively. Discharge to home occurred in 57.1% of all ICU admissions. On adjusted analyses, short-term mortality increased with rising number of organ failures (adjusted OR (aOR) 1.399, 95% CI 1.374 to 1.425), while being lower with chemotherapy (aOR 0.467, 95% CI 0.432 to 0.506) and radiation therapy (aOR 0.832, 95% CI 0.749 to 0.924), and decreased over time (aOR 0.934 per year, 95% CI 0.924 to 0.945). Predictors of short-term mortality were largely similar among those undergoing mechanical ventilation. Most ICU admissions with metastatic cancer survived hospitalization, although short-term mortality was very high among those undergoing mechanical ventilation. Short-term mortality decreased over time and was lower among those receiving chemotherapy and radiation therapy. These findings support consideration of critical care in patients with metastatic cancer, but underscore the need to address patient-centered goals of care ahead of ICU admission.

Gender differences in the provision of intensive care: a Bayesian approach

PURPOSE

It is currently unclear whether management and outcomes of critically ill patients differ between men and women. We sought to assess the influence of age, sex and diagnoses on the probability of intensive care provision in critically ill cardio- and neurovascular patients in a large nationwide cohort in Switzerland.

METHODS

Retrospective analysis of 450,948 adult patients with neuro- and cardiovascular disease admitted to all hospitals in Switzerland between 01/2012 and 12/2016 using Bayesian modeling.

RESULTS

For all diagnoses and populations, median ages at admission were consistently higher for women than for men [75 (64;82) years in women vs. 68 (58;77) years in men, p < 0.001]. Overall, women had a lower likelihood to be admitted to an intensive care unit (ICU) than men, despite being more severely ill [odds ratio (OR) 0.78 (0.76-0.79)]. ICU admission probability was lowest in women aged > 65 years (OR women:men 0.94 (0.89-0.99), p < 0.001). Women < 45 years had a similar ICU admission probability as men in the same age category [OR women:men 1.03 (0.94-1.13)], in spite of more severe illness. The odds to die were significantly higher in women than in men per unit increase in Simplified Acute Physiology Score (SAPS) II (OR 1.008 [1.004-1.012]).

CONCLUSION

In the care of the critically ill, our study suggests that women are less likely to receive ICU treatment regardless of disease severity. Underuse of ICU care was most prominent in younger women < 45 years. Although our study has several limitations that are imposed by the limited data available from the registries, our findings suggest that current ICU triage algorithms could benefit from careful reassessment. Further, and ideally prospective, studies are needed to confirm our findings.

Phosphate abnormalities and outcomes among admissions to the intensive care unit: A retrospective multicentre cohort study

PURPOSE

We investigated the effect of serum phosphate abnormalities at intensive care unit (ICU) admission on risk of death and length of stay in critically ill patients.

MATERIALS AND METHODS

A retrospective cohort of patients admitted to three adult ICUs in Queensland, Australia from April 2014 to 2019 was studied. Hypophosphataemia, normophosphataemia and hyperphosphataemia were defined as serum phosphate level of <0.8, 0.8-1.5 and >1.5 mmol/L respectively. Univariable and logistic regression analyses were performed to investigate the association between the phosphate groups and the risk of death.

MEASUREMENTS AND MAIN RESULTS

We included 13,155 patients in the analysis, of which 1424 (10.8%) patients had hypophosphataemia and 2544 (19.3%) hyperphosphataemia. The mean admission phosphate level was 1.25 (SD, ±0.43) mmol/L. Both hypophosphatemia (OR 1.29; 95% CI, 1.02-1.64; p = 0.034) and hyperphosphataemia (OR 1.39; 95% CI, 1.15-1.68; p = 0.001) at admission were independently associated with increased risk of death after adjusting for covariables using logistic regression analysis.

CONCLUSION

Hypophosphatemia and hyperphosphatemia were both independently associated with an increased case fatality rate and ICU length of stay in a large multicentre ICU cohort.