Association between hematocrit and the 30-day mortality of patients with sepsis: A retrospective analysis based on the large-scale clinical database MIMIC-IV

Association between lactate to hematocrit ratio and 30-day all-cause mortality in patients with sepsis: a retrospective analysis of the Medical Information Mart for Intensive Care IV database

Background

The lactate to hematocrit ratio (LHR) has not been assessed for predicting all-cause death in sepsis patients. This study aims to evaluate the relationship between LHR and 30-day all-cause mortality in sepsis patients.

Methods

This retrospective study used the data from Medical information mart for intensive care IV (MIMIC-IV, version 2.0). Our study focused on adult sepsis patients who were initially hospitalized in the Intensive care unit (ICU). The prognostic significance of admission LHR for 30-day all-cause mortality was evaluated using a multivariate Cox regression model, ROC curve analysis, Kaplan-Meier curves, and subgroup analyses.

Results

A total of 3,829 sepsis patients participated in this study. Among the cohort, 8.5% of individuals died within of 30 days (p < 0.001). The area under the curve (AUC) for LHR was 74.50% (95% CI: 71.6-77.50%), higher than arterial blood lactate (AUC = 71.30%), hematocrit (AUC = 64.80%), and shows no significant disadvantage compared to qSOFA, SOFA, and SAPS II. We further evaluated combining LHR with qSOFA score to predict mortality in sepsis patients, which shows more clinical significance. ROC curve analysis showed that 6.538 was the optimal cutoff value for survival and non-survival groups. With LHR ≥6.538 vs. LHR <6.538 (p < 0.001). Subgroup analysis showed significant interactions between LHR, age, sex, and simultaneous acute respiratory failure (p = 0.001-0.005).

Conclusion

LHR is an independent predictor of all-cause mortality in sepsis patients after admission, with superior predictive ability compared to blood lactate or hematocrit alone.

Risk and mediation analyses of hemoglobin glycation index and survival prognosis in patients with sepsis

An increasing number of studies have reported the close relation of the hemoglobin glycation index (HGI) with metabolism, inflammation, and disease prognosis. However, the prognostic relationship between the HGI and patients with sepsis remains unclear. Thus, this study aimed to analyze the association between the HGI and all-cause mortality in patients with sepsis using data from the MIMIC-IV database. In this study, 2605 patients with sepsis were retrospectively analyzed. The linear regression equation was established by incorporating glycated hemoglobin (HbA1c) and fasting plasma glucose levels. Subsequently, the HGI was calculated based on the difference between the predicted and observed HbA1c levels. Furthermore, the HGI was divided into the following three groups using X-tile software: Q1 (HGI ≤  - 0.50%), Q2 (- 0.49% ≤ HGI ≤ 1.18%), and Q3 (HGI ≥ 1.19%). Kaplan-Meier survival curves were further plotted to analyze the differences in 28-day and 365-day mortality among patients with sepsis patients in these HGI groups. Multivariate corrected Cox proportional risk model and restricted cubic spline (RCS) were used. Lastly, mediation analysis was performed to assess the factors through which HGI affects sepsis prognosis. This study included 2605 patients with sepsis, and the 28-day and 365-day mortality rates were 19.7% and 38.9%, respectively. The Q3 group had the highest mortality risk at 28 days (HR = 2.55, 95% CI: 1.89-3.44, p < 0.001) and 365 days (HR = 1.59, 95% CI: 1.29-1.97, p < 0.001). In the fully adjusted multivariate Cox proportional hazards model, patients in the Q3 group still displayed the highest mortality rates at 28 days (HR = 2.02, 95% CI: 1.45-2.80, p < 0.001) and 365 days (HR = 1.28, 95% CI: 1.08-1.56, p < 0.001). The RCS analysis revealed that HGI was positively associated with adverse clinical outcomes. Finally, the mediation effect analysis demonstrated that the HGI might influence patient survival prognosis via multiple indicators related to the SOFA and SAPS II scores. There was a significant association between HGI and all-cause mortality in patients with sepsis, and patients with higher HGI values had a higher risk of death. Therefore, HGI can be used as a potential indicator to assess the prognostic risk of death in patients with sepsis.

Nonlinear relationship of red blood cell indices (MCH, MCHC, and MCV) with all-cause and cardiovascular mortality: A cohort study in U.S. adults

BACKGROUND

In recent years, increasing attention has been focused on the impact of red blood cell indices (RCIs) on disease prognosis. We aimed to investigate the association of mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), and mean corpuscular volume (MCV) with mortality.

METHODS

The study used cohort data from U.S. adults who participated in the 1999-2008 National Health and Nutrition Examination Survey. All-cause mortality was the primary outcome during follow-up, with secondary cardiovascular mortality outcomes. COX regression was applied to analyze the connection between RCIs and mortality. We adopted three models to minimize potential bias. Smooth-fit curves and threshold effect analyses were utilized to observe the dose-response relationship between RCIs and all-cause and cardiovascular mortality. In addition, we performed sensitivity analyses.

RESULTS

21,203 individuals were enrolled in our research. During an average 166.2 ± 54.4 months follow-up, 24.4% of the population died. Curve fitting indicated a U-shaped relationship between MCV and MCH with all-cause mortality, and the relationship of MCHC to all-cause mortality is L-shaped. We identified inflection points in the relationship between MCV, MCH, and MCHC and all-cause mortality as 88.56732 fl, 30.22054 pg, 34.34624 g/dl (MCV <88.56732 fl, adjusted HR 0.99, 95 CI% 0.97-1.00; MCV >88.56732 fl, adjusted HR 1.05, 95 CI% 1.04-1.06. MCH <30.22054 pg, adjusted HR 0.95, 95 CI% 0.92-0.98; MCH >30.22054 pg, adjusted HR 1.08, 95 CI% 1.04-1.12. MCHC <34.34624 g/dl, adjusted HR 0.88, 95 CI% 0.83-0.93). Besides, the MCV curve was U-shaped in cardiovascular mortality (MCV <88.56732 fl, adjusted HR 0.97, 95 CI% 0.94-1.00; MCV >88.56732 fl, adjusted HR 1.04, 95 CI% 1.01-1.06).

CONCLUSION

This cohort study demonstrated that RCIs (MCH, MCHC, and MCV) were correlated with mortality in the general population. Three RCIs were nonlinearly correlated with all-cause mortality. In addition, there were nonlinear relationships between MCH and MCV and cardiovascular mortality.

Acute blood loss anemia in hospitalized patients is associated with adverse outcomes: An analysis of the Nationwide Inpatient Sample

BACKGROUND

Acute blood loss anemia is the most common form of anemia and often results from traumatic injuries or gastrointestinal bleeding. There are limited studies analyzing outcomes associated with acute blood loss anemia in hospitalized patients.

METHODS

The Nationwide Inpatient Sample (NIS) was analyzed from 2010 to 2014 (n = 133,809). The impact of acute blood loss anemia on in-hospital mortality, length of stay (LOS), healthcare cost, and disposition was determined using regression modeling adjusted for age, gender, race, and comorbidities.

RESULTS

Hospitalized patients with acute blood loss anemia had significantly higher healthcare cost (adj OR 1.04; 95% CI: 1.04-1.05), greater lengths of stay (adj OR 1.18; 95% CI: 1.17-1.18), and were less likely to be discharged home compared to the general medical population (adj OR 0.27; 95% CI: 0.26-0.28). Acute blood loss anemia was associated with increased risk for mortality in unadjusted models (unadj 1.16; 95% CI: 1.12-1.20) but not in adjusted models (adj OR 0.91; 95% CI: 0.88-0.94). When analyzing comorbidities, a "muscle loss phenotype" had the strongest association with mortality in patients with acute blood loss anemia (adj OR 4.48; 95% CI: 4.35-4.61). The top five primary diagnostic codes associated with acute blood loss anemia were long bone fractures, GI bleeds, cardiac repair, sepsis, and OB/Gyn related causes. Sepsis had the highest association with mortality (18%, adj OR 2.59; 95% CI: 2.34-2.86) in those with acute blood loss anemia.

CONCLUSIONS

Acute blood loss anemia is associated with adverse outcomes in hospitalized patients.

Postoperative SIRS after thermal ablation of HCC: Risk factors and short-term prognosis

Background

We aimed to explore the potential risk factors and short-term prognosis for SIRS after thermal ablation of hepatocellular carcinoma (HCC).

Methods

Data from patients with HCC who underwent thermal ablation in the Third Affiliated Hospital of Sun Yat-sen University between January 2015 and August 2021 were retrieved from the perioperative database. Pre-, intra- and postoperative data between SIRS group and non-SIRS group were compared and multivariate logistic regression analysis was performed to identify the risk factors for SIRS after thermal ablation.

Results

A total of 1491 patients were enrolled and 234 (15.7 %) patients developed SIRS after thermal ablation. Compared with those without SIRS, patients with SIRS had a longer hospital stay, higher hospitalization costs and higher risk of more severe postoperative complications. In the multivariate logistic regression analysis, current smoking (OR 1.58, 95 %CI 1.09-2.29), decreased HCT (OR 1.51,95 %CI 1.11-2.04), NEUT < 1.5 × 109/L(OR 1.74, 95 %CI 1.14-2.65), NEUT% < 0.5 or > 0.7 (OR 1.36, 95 %CI 1.01-1.83) and PT > 16.3s (OR 2.42, 95 %CI 1.57-3.74) were significantly associated with postoperative SIRS.

Conclusions

Current smoking, decreased HCT, neutropenia, abnormal percentage of neutrophils and prolonged PT are the independent risk factors for SIRS after thermal ablation of HCC, which worsens outcomes of patients. This study can help identify high-risk population and guide appropriate care so as to reduce the incidence of postoperative SIRS.

Applying artificial neural network in predicting sepsis mortality in the emergency department based on clinical features and complete blood count parameters

A complete blood count (CBC) is routinely ordered for emergency department (ED) patients with infections. Certain parameters, such as the neutrophil-to-lymphocyte ratio (NLR), might have prognostic value. We aimed to evaluate the prognostic value of the presenting CBC parameters combined with clinical variables in predicting 30-day mortality in adult ED patients with infections using an artificial neural network (ANN). We conducted a retrospective study of ED patients with infections between 17 December 2021 and 16 February 2022. Clinical variables and CBC parameters were collected from patient records, with NLR, monocyte-to-lymphocyte ratio (MLR), and platelet-to-lymphocyte ratio (PLR) calculated. We determined the discriminatory performance using the area under the receiver operating characteristic curve (AUROC) and performed a 70/30 random data split and supervised ANN machine learning. We analyzed 558 patients, of whom 144 (25.8%) had sepsis and 60 (10.8%) died at 30 days. The AUROCs of NLR, MLR, PLR, and their sum were 0.644 (95% CI 0.573-0.716), 0.555 (95% CI 0.482-0.628), 0.606 (95% CI 0.529-0.682), and 0.610 (95% CI 0.534-0.686), respectively. The ANN model based on twelve variables including clinical variables, hemoglobin, red cell distribution width, NLR, and PLR achieved an AUROC of 0.811 in the testing dataset.

Construction and validation of a risk assessment model for acute kidney injury in patients with acute pancreatitis in the intensive care unit

BACKGROUND

To construct and validate a risk assessment model for acute kidney injury (AKI) in patients with acute pancreatitis (AP) in the intensive care unit (ICU).

METHODS

A total of 963 patients diagnosed with acute pancreatitis (AP) from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database was included. These patients were randomly divided into training set (N = 674) and validation set (N = 289) at a ratio of 7:3. Clinical characteristics were utilized to establish a nomogram for the prediction of AKI during ICU stay. These variables were selected by the least absolute shrinkage and selection operation (LASSO) regression and included in multivariate logistic regression analysis. Variables with P-values less than 0.05 were included in the final model. A nomogram was constructed based on the final model. The predicted accuracy of the nomogram was assessed by calculating the receiver operating characteristic curve (ROC) and the area under the curve (AUC). Moreover, calibration curves and Hosmer-Lemeshow goodness-of-fit test (HL test) were performed to evaluate model performance. Decision curve analysis (DCA) evaluated the clinical net benefit of the model.

RESULTS

A multivariable model that included 6 variables: weight, SOFA score, white blood cell count, albumin, chronic heart failure, and sepsis. The C-index of the nomogram was 0.82, and the area under the receiver operating characteristic curve (AUC) of the training set and validation set were 0.82 (95% confidence interval:0.79-0.86) and 0.76 (95% confidence interval: 0.70-0.82), respectively. Calibration plots showed good consistency between predicted and observed outcomes in both the training and validation sets. DCA confirmed the clinical value of the model and its good impact on actual decision-making.

CONCLUSION

We identified risk factors associated with the development of AKI in patients with AP. A risk prediction model for AKI in ICU patients with AP was constructed, and improving the treatment strategy of relevant factors in the model can reduce the risk of AKI in AP patients.

Association between ferritin to albumin ratio and 28-day mortality in patients with sepsis: a retrospective cohort study

OBJECTIVES

The ratio of ferritin to albumin (FAR) has been proposed as a novel prognostic indicator for COVID-19. However, the role of FAR in predicting the all-cause mortality rate in patients with sepsis has not been evaluated. Therefore, the aim of this study is to elucidate the correlation between FAR and the 28-day all-cause mortality rate in patients with sepsis.

METHODS

This study used data from the Medical Information Mart for Intensive Care IV database (v2.0) for a retrospective cohort analysis. The study focused on adult patients with sepsis who were admitted to the intensive care unit. The primary objective was to assess the predictive capability of FAR in determining the 28-day all-cause mortality rate among patients with sepsis.

RESULTS

The study involved 1553 sepsis patients in total. Based on the survival status of sepsis patients within 28 days, they were divided into two groups: a survival group consisting of 973 patients, and a death group consisting of 580 patients. The results revealed a 28-day mortality rate of 37.35% among sepsis patients. The multivariable Cox regression analysis revealed that FAR was an independent predictor of the 28-day all-cause mortality rate in patients with sepsis (hazard ratio [HR]: 1.17-1.19; 95% confidence interval 1.11-1.26; P < 0.001). The FAR demonstrated a higher area under the curve (AUC) of 61.01% (95% confidence interval 58.07-63.96%), compared to serum ferritin (60.48%), serum albumin (55.56%), and SOFA score (56.97%). Receiver operating characteristic curve (ROC) analysis determined the optimal cutoff value for FAR as 364.2215. Kaplan-Meier analysis revealed a significant difference in the 28-day all-cause mortality rate between patients with FAR ≥ 364.2215 and those with FAR < 364.2215 (P < 0.001). Furthermore, subgroup analysis showed no significant interaction between FAR and each subgroup.

CONCLUSIONS

This study revealed a significant correlation between FAR and the 28-day mortality rate in patients with sepsis. Higher FAR values were strongly associated with increased mortality rates within 28 days.

The predictive values of admission characteristics for 28-day all-cause mortality in septic patients with diabetes mellitus: a study from the MIMIC database

Background

Septic patients with diabetes mellitus (DM) are more venerable to subsequent complications and the resultant increase in associated mortality. Therefore, it is important to make tailored clinical decisions for this subpopulation at admission.

Method

Data from large-scale real-world databases named the Medical Information Mart for Intensive Care Database (MIMIC) were reviewed. The least absolute selection and shrinkage operator (LASSO) was performed with 10 times cross-validation methods to select the optimal prognostic factors. Multivariate COX regression analysis was conducted to identify the independent prognostic factors and nomogram construction. The nomogram was internally validated via the bootstrapping method and externally validated by the MIMIC III database with receiver operating characteristic (ROC), calibration curves, decision curve analysis (DCA), and Kaplan-Meier curves for robustness check.

Results

A total of 3,291 septic patients with DM were included in this study, 2,227 in the MIMIC IV database and 1,064 in the MIMIC III database, respectively. In the training cohort, the 28-day all-cause mortality rate is 23.9% septic patients with DM. The multivariate Cox regression analysis reveals age (hazard ratio (HR)=1.023, 95%CI: 1.016-1.031, p<0.001), respiratory failure (HR=1.872, 95%CI: 1.554-2.254, p<0.001), Sequential Organ Failure Assessment score (HR=1.056, 95%CI: 1.018-1.094, p=0.004); base excess (HR=0.980, 95%CI: 0.967-0.992, p=0.002), anion gap (HR=1.100, 95%CI: 1.080-1.120, p<0.001), albumin (HR=0.679, 95%CI: 0.574-0.802, p<0.001), international normalized ratio (HR=1.087, 95%CI: 1.027-1.150, p=0.004), red cell distribution width (HR=1.056, 95%CI: 1.021-1.092, p=0.001), temperature (HR=0.857, 95%CI: 0.789-0.932, p<0.001), and glycosylated hemoglobin (HR=1.358, 95%CI: 1.320-1.401, p<0.001) at admission are independent prognostic factors for 28-day all-cause mortality of septic patients with DM. The established nomogram shows satisfied accuracy and clinical utility with AUCs of 0.870 in the internal validation and 0.830 in the external validation cohort as well as 0.820 in the septic shock subpopulation, which is superior to the predictive value of the single SOFA score.

Conclusion

Our results suggest that admission characteristics show an optimal prediction value for short-term mortality in septic patients with DM. The established model can support intensive care unit physicians in making better initial clinical decisions for this subpopulation.

Blood microsampling technologies: Innovations and applications in 2022

With the development of highly sensitive bioanalytical techniques, the volume of samples necessary for accurate analysis has reduced. Microsampling, the process of obtaining small amounts of blood, has thus gained popularity as it offers minimal-invasiveness, reduced logistical costs and biohazard risks while simultaneously showing increased sample stability and a potential for the decentralization of the approach and at-home self-sampling. Although the benefits of microsampling have been recognised, its adoption in clinical practice has been slow. Several microsampling technologies and devices are currently available and employed in research studies for various biomedical applications. This review provides an overview of the state-of-the-art in microsampling technology with a focus on the latest developments and advancements in the field of microsampling. Research published in the year 2022, including studies (i) developing strategies for the quantitation of analytes in microsamples and (ii) bridging and comparing the interchangeability between matrices and choice of technology for a given application, is reviewed to assess the advantages, challenges and limitations of the current state of microsampling. Successful implementation of microsampling in routine clinical care requires continued efforts for standardization and harmonization. Microsampling has been shown to facilitate data-rich studies and a patient-centric approach to healthcare and is foreseen to play a central role in the future digital revolution of healthcare through continuous monitoring to improve the quality of life.

The correlation of hemoglobin and 28-day mortality in septic patients: secondary data mining using the MIMIC-IV database

BACKGROUND

Previous studies found minimal evidence and raised controversy about the link between hemoglobin and 28-day mortality in sepsis patients. As a result, the purpose of this study was to examine the association between hemoglobin and 28-day death in sepsis patients by analyzing the Medical Intensive Care IV (MIMIC-IV) database from 2008 to 2019 at an advanced medical center in Boston, Massachusetts.

METHODS

We extracted 34,916 sepsis patients from the MIMIC-IV retrospective cohort database, using hemoglobin as the exposure variable and 28-day death as the outcome variable, and after adjusting for confounders (demographic indicators, Charlson co-morbidity index, SOFA score, vital signs, medication use status (glucocorticoids, vasoactive drugs, antibiotics, and immunoglobulins, etc.)), we investigated the independent effects of hemoglobin and 28-day risk of death by binary logistic regression as well as two-piecewise linear model, respectively.

RESULTS

Hemoglobin levels and 28-day mortality were shown to be non-linearly related.The inflection points were 104 g/L and 128 g/L, respectively. When HGB levels were between 41 and 104 g/L, there was a 10% decrease in the risk of 28-day mortality (OR: 0.90; 95% CI: 0.87 to 0.94, p-value = 0.0001). However, in the range of 104-128 g/L, we did not observe a significant association between hemoglobin and 28-day mortality (OR: 1.17; 95% CI: 1.00 to 1.35, P value = 0.0586). When HGB was in the range of 128-207 g/L, there was a 7% increase in the risk of 28-day mortality for every 1 unit increase in HGB (OR: 1.07; 95% CI: 1.01 to 1.15, P value = 0.0424).

CONCLUSION

In patients with sepsis, baseline hemoglobin was related to a U-shaped risk of 28-day death. When HGB was in the range of 12.8-20.7 g/dL, there was a 7% increase in the risk of 28-day mortality for every 1 unit increase in HGB.

The PaO2/FiO2 is independently associated with 28-day mortality in patients with sepsis: a retrospective analysis from MIMIC-IV database

BACKGROUND

To clarify the relationship between the PaO₂/FiO₂ and 28-day mortality in patients with sepsis.

METHODS

This was a retrospective cohort study regarding MIMIC-IV database. Nineteen thousand two hundred thirty-three patients with sepsis were included in the final analysis. PaO₂/FiO₂ was exposure variable, 28-day mortality was outcome variable. PaO₂/FiO₂ was log-transformed as LnPaO₂/FiO₂. Binary logistic regression was used to explore the independent effects of LnPaO₂/FiO₂ on 28-day mortality using non-adjusted and multivariate-adjusted models. A generalized additive model (GAM) and smoothed curve fitting was used to investigate the non-linear relationship between LnPaO₂/FiO₂ and 28-day mortality. A two-piecewise linear model was used to calculate the OR and 95% CI on either side of the inflection point.

RESULTS

The relationship between LnPaO₂/FiO₂ and risk of 28-day death in sepsis patients was U-shape. The inflection point of LnPaO₂/FiO₂ was 5.30 (95%CI: 5.21-5.39), which indicated the inflection point of PaO₂/FiO₂ was 200.33 mmHg (95%CI: 183.09 mmHg-219.20 mmHg). On the left of inflection point, LnPaO₂/FiO₂ was negatively correlated with 28-day mortality (OR: 0.37, 95%CI: 0.32-0.43, p < 0.0001). On the right of inflection point, LnPaO₂/FiO₂ was positively correlated with 28-day mortality in patients with sepsis (OR: 1.53, 95%CI: 1.31-1.80, p < 0.0001).

CONCLUSIONS

In patients with sepsis, either a high or low PaO₂/FiO₂ was associated with an increased risk of 28-day mortality. In the range of 183.09 mmHg to 219.20 mmHg, PaO₂/FiO₂ was associated with a lower risk of 28-day death in patients with sepsis.

Identification and control for the effects of bioinformatic globin depletion on human RNA-seq differential expression analysis

When profiling blood samples by RNA-sequencing (RNA-seq), RNA from haemoglobin (Hgb) can account for up to 70% of the transcriptome. Due to considerations of sequencing depth and power to detect biological variation, Hgb RNA is typically depleted prior to sequencing by hybridisation-based methods; an alternative approach is to deplete reads arising from Hgb RNA bioinformatically. In the present study, we compared the impact of these two approaches on the outcome of differential gene expression analysis performed using RNA-seq data from 58 human tuberculosis (TB) patient or contact whole blood samples-29 globin kit-depleted and 29 matched non-depleted-a subset of which were taken at TB diagnosis and at six months post-TB treatment from the same patient. Bioinformatic depletion of Hgb genes from the non-depleted samples (bioinformatic-depleted) substantially reduced library sizes (median = 57.24%) and fewer long non-coding, micro, small nuclear and small nucleolar RNAs were captured in these libraries. Profiling published TB gene signatures across all samples revealed inferior correlation between kit-depleted and bioinformatic-depleted pairs when the proportion of reads mapping to Hgb genes was higher in the non-depleted sample, particularly at the TB diagnosis time point. A set of putative "globin-fingerprint" genes were identified by directly comparing kit-depleted and bioinformatic-depleted samples at each timepoint. Two TB treatment response signatures were also shown to have decreased differential performance when comparing samples at TB diagnosis to six months post-TB treatment when profiled on the bioinformatic-depleted samples compared with their kit-depleted counterparts. These results demonstrate that failure to deplete Hgb RNA prior to sequencing has a negative impact on the sensitivity to detect disease-relevant gene expression changes even when bioinformatic removal is performed.

A non-linear positive relationship between serum phosphate and clinical outcomes in sepsis

Objective

This study aimed to evaluate the possible relationship between serum phosphate and short-term outcomes in sepsis.

Methods

This was a retrospective study. Sepsis patients in MIMIC-IV database were included. Based on the quartiles of serum phosphate, all sepsis patients were divided into four groups. Univariable and multivariable regression analyses were constructed for discussing the relationship between different parameters and 30-day mortality in sepsis. A generalized additive model was performed for exploring the association of serum phosphate with 30-day mortality.

Results

6251 sepsis patients including 4368 survivors and 1883 non-survivors were included. A significant relationship between serum phosphate and 30-day mortality was found after adjusting for all potential confounders (OR = 1.19, 95%CI:1.13-1.26, P < 0.0001). The relationship was non-linear with an inflection point of 6.8 mg/dl. On the left side of the inflection point (≤6.8 mg/dl, n = 5911 (94.56%)), the OR was 1.24 (95%CI: 1.17-1.31, P < 0.0001). On the right side of the inflection point (>6.8 mg/dl, n = 340 (5.44%)), the OR was 0.94 (95%CI:0.78-1.13, P = 0.5038).

Conclusion

A non-linear positive relationship was found between serum phosphate and 30-day mortality in sepsis. Serum phosphate was associated with mortality in sepsis. Our results could be used for screening out those sepsis patients with higher risk of worse outcomes.

Predicting length of stay in ICU and mortality with temporal dilated separable convolution and context-aware feature fusion

In healthcare, Intensive Care Unit (ICU) bed management is a necessary task because of the limited budget and resources. Predicting the remaining Length of Stay (LoS) in ICU and mortality can assist clinicians in managing ICU beds efficiently. This study proposes a deep learning method based on several successive Temporal Dilated Separable Convolution with Context-Aware Feature Fusion (TDSC-CAFF) modules, and a multi-view and multi-scale feature fusion for predicting the remaining LoS and mortality risk for ICU patients. In each TDSC-CAFF module, temporal dilated separable convolution is used to encode each feature separately, and context-aware feature fusion is proposed to capture comprehensive and context-aware feature representations from the input time-series features, static demographics, and the output of the last TDSC-CAFF module. The CAFF outputs of each module are accumulated to achieve multi-scale representations with different receptive fields. The outputs of TDSC and CAFF are concatenated with skip connection from the output of the last module and the original time-series input. The concatenated features are processed by the proposed Point-Wise convolution-based Attention (PWAtt) that captures the inter-feature context to generate the final temporal features. Finally, the final temporal features, the accumulated multi-scale features, the encoded diagnosis, and static demographic features are fused and then processed by fully connected layers to obtain prediction results. We evaluate our proposed method on two publicly available datasets: eICU and MIMIC-IV v1.0 for LoS and mortality prediction tasks. Experimental results demonstrate that our proposed method achieves a mean squared log error of 0.07 and 0.08 for LoS prediction, and an Area Under the Receiver Operating Characteristic Curve of 0.909 and 0.926 for mortality prediction, on eICU and MIMIC-IV v1.0 datasets, respectively, which outperforms several state-of-the-art methods.

Factors associated with severe childhood community-acquired pneumonia: a retrospective study from two hospitals

Background

Community-acquired pneumonia (CAP) is the leading cause of death in children globally. Indonesia is ranked 1st in South East Asia with the highest burden of pneumonia. Identification of risk factors is necessary for early intervention and better management. This study intended to describe CAP’s clinical signs and laboratory findings and explore the risk factors of severe CAP among children in Indonesia.

Methods

This was a retrospective study of childhood hospitalizations in Siloam General Hospitals and Siloam Hospitals Lippo Village from December 2015 to December 2019. Demographic data, clinical signs, and laboratory findings were collected and processed using IBM SPSS 26.0.

Results

This study included 217 participants with 66 (30.4%) severe pneumonia cases. Multivariate analysis shows that fever that lasts more than 7 days (ORadj = 4.95; 95%CI 1.61–15.21, Padj = 0.005) and increase in respiratory rate (ORadj = 1.05, 95%CI 1.01–1.08, Padj = 0.009) are two predictors of severe pneumonia. Meanwhile, a normal hematocrit level (ORadj = 0.9; 95%CI 0.83–0.98, Padj = 0.011) and children with normal BMI (ORadj = 0.7; 95%CI 0.57–0.84, Padj < 0.001) are significant independent predictors of severe pneumonia. The Hosmer-Lemeshow test shows that this model is a good fit with a P-value of 0.281. The AUC for this model is 0.819 (95%CI = 0.746–0.891, P-value < 0.001) which shows that this model has good discrimination.

Conclusion

Pediatric CAP hospitalizations with fever lasting > 7 days and tachypnea were at higher risk for progressing to severe pneumonia. A normal hematocrit level and a normal BMI are protective factors for severe pneumonia.