Association between Body Mass Index and Short-Term Clinical Outcomes in Critically Ill Patients with Sepsis: A Real-World Study

Development and validation of an interpretable machine learning for mortality prediction in patients with sepsis

Introduction

Sepsis is a leading cause of death. However, there is a lack of useful model to predict outcome in sepsis. Herein, the aim of this study was to develop an explainable machine learning (ML) model for predicting 28-day mortality in patients with sepsis based on Sepsis 3.0 criteria.

Methods

We obtained the data from the Medical Information Mart for Intensive Care (MIMIC)-III database (version 1.4). The overall data was randomly assigned to the training and testing sets at a ratio of 3:1. Following the application of LASSO regression analysis to identify the modeling variables, we proceeded to develop models using Extreme Gradient Boost (XGBoost), Logistic Regression (LR), Support Vector Machine (SVM), and Random Forest (RF) techniques with 5-fold cross-validation. The optimal model was selected based on its area under the curve (AUC). Finally, the Shapley additive explanations (SHAP) method was used to interpret the optimal model.

Results

A total of 5,834 septic adults were enrolled, the median age was 66 years (IQR, 54-78 years) and 2,342 (40.1%) were women. After feature selection, 14 variables were included for developing model in the training set. The XGBoost model (AUC: 0.806) showed superior performance with AUC, compared with RF (AUC: 0.794), LR (AUC: 0.782) and SVM model (AUC: 0.687). SHAP summary analysis for XGBoost model showed that urine output on day 1, age, blood urea nitrogen and body mass index were the top four contributors. SHAP dependence analysis demonstrated insightful nonlinear interactive associations between factors and outcome. SHAP force analysis provided three samples for model prediction.

Conclusion

In conclusion, our study successfully demonstrated the efficacy of ML models in predicting 28-day mortality in sepsis patients, while highlighting the potential of the SHAP method to enhance model transparency and aid in clinical decision-making.

Role of obesity in lower mortality risk in sepsis: a meta-analysis of observational studies

This meta-analysis aims to explore the correlation between obesity and mortality in patients with sepsis. Data were gathered from various sources, including PubMed, the Cochrane Library, and Embase (no language restrictions). Clinical studies, both retrospective and prospective ones, were selected to analyze mortality due to sepsis in patients with or without obesity. The Newcastle-Ottawa Scale was used to assess the quality of the studies included. In data synthesis, odds ratio (OR) and 95% confidence interval (CI) were meta-analyzed using the DerSimonian-Laird random-effects model, followed by sensitivity and heterogeneity analyses. Two cohort studies were included to investigate survival in inpatients with obesity and sepsis, with pooled analysis indicating a lowered mortality rate (OR=0.88; 95% CI: 0.81-0.95; I2=0.00%; P=0.000). This meta-analysis lends support to the obesity paradox, suggesting a reduced mortality from sepsis in obese patients. However, further prospective trials and research on mechanisms are needed to test this hypothesis.

The association between body mass index and mortality in septic older adults

BACKGROUND

There is limited evidence on the association between body mass index (BMI) and outcomes in older adults with sepsis.

OBJECTIVES

The purpose of this study was to explore the relationship between BMI and the clinical outcome in the older adults with sepsis.

METHODS

All older adults (age 65 or older) with sepsis were analyzed retrospectively in the Intensive Care Medical Information Mart (MIMIC)- IV database. The primary outcome was 30-day mortality. The relationship between BMI and 30-day mortality was estimated by multivariate logistic regression model. The interaction and stratified analysis were performed by sex, race, renal disease, congestive heart failure (CHF), diabetes, and chronic pulmonary disease.

RESULTS

The total number of participants was 6604. After adjustment for potential covariates, there was a significant correlation between BMI and 30-day mortality. A 1 kg/m2 increase in BMI was associated with a 3 % decrease in 30-day mortality (adjusted HR = 0.97, 95 % CI: 0.96-0.98, P < 0.001). The correlation between BMI and 30-day mortality showed a statistically significant nonlinear association with an l-shaped curve (p = 0.001). Significant interactions were observed only for sex in the stratified analyses (P = 0.001).

CONCLUSIONS

In this study, it was observed that a higher BMI is linked with better survival rates among older adults suffering from sepsis, while being underweight raises the risk of mortality. Notably, male patients with a higher BMI had a lower mortality risk compared to female patients.

Association of obesity and mortality in sepsis patients: A meta-analysis from observational evidence

The project aims to investigate the correlation between obesity, overweight, or low body weight and the risk of mortality in sepsis patients. We performed a rigorous and thorough search of major electronic databases, including PubMed, Web of Science, EMBASE, and Cochrane Library, from the inception of these databases up to March 28, 2023. The data were analyzed with Stata software (version 16.0). Twelve studies incorporating 521,207 individuals were enrolled. The results demonstrated that obesity (OR = 0.82; 95% CI: 0.69-0.97; P < 0.001) or overweight (OR = 0.83; 95% CI: 0.73-0.94; P < 0.001) decreased the risk of mortality in sepsis patients. Instead, the reverse phenomena existed in patients with a low weight (OR = 1.43; 95%CI: 1.16-1.76; P = 0.038). There is an "obesity paradox" phenomenon in the mortality of obese and overweight patients with sepsis, but low body weight is an independent risk factor for the mortality of sepsis patients. This study demonstrated that the mortality in sepsis patients and obesity or overweight were negatively correlated, but displayed a significant positive relation to low weight.

Association of body mass index with mortality of sepsis or septic shock: an updated meta-analysis

BACKGROUND

The effects of body mass index (BMI) on mortality of sepsis remain unknown, since previous meta-analyses have reported conflicting results. Several observational studies published recently have provided new evidence. Thus, we performed this updated meta-analysis.

METHODS

PubMed, Embase, Web of Science, and Cochran Library were searched for articles published before February 10, 2023. Observational studies that assessed the association of BMIs with mortality of sepsis patients aged > 18 years were selected. We excluded studies of which data were unavailable for quantitative synthesis. Odds ratios (OR) with 95% confidence interval (CI) were the effect measure, which were combined using fixed-effect or random-effect models. The Newcastle-Ottawa Scale was applied for quality assessment. Subgroups analyses were conducted according to potential confounders.

RESULTS

Fifteen studies (105,159 patients) were included in the overall analysis, which indicated that overweight and obese BMIs were associated with lower mortality (OR: 0.79, 95% CI 0.70-0.88 and OR: 0.74, 95% CI 0.67-0.82, respectively). The association was not significant in patients aged ≤ 50 years (OR: 0.89, 95% CI 0.68-1.14 and OR: 0.77, 95% CI 0.50-1.18, respectively). In addition, the relationship between morbidly obesity and mortality was not significant (OR: 0.91, 95% CI 0.62-1.32).

CONCLUSIONS

Overweight and obese BMIs (25.0-39.9 kg/m²) are associated with reduced mortality of patients with sepsis or septic shock, although such survival advantage was not found in all crowds. Trial registration The protocol of this study was registered in PROSPERO (registration number CRD42023399559).

Body mass index and risk of clostridioides difficile infection: a systematic review and meta-analysis

OBJECTIVE

To comprehensively investigate the association between obesity/high body mass index (BMI) and risk of Clostridioides difficile infection (CDI) using systematic review and meta-analysis.

METHODS

Potentially eligible studies were identified from Medline and EMBASE databases from inception to February 2021 using search strategy consisting of terms for "Body Mass Index" and "Clostridioides Difficile". We only included studies that consist of a group of individuals with CDI and another group without CDI. Then, the studies must report their BMI or history of obesity. Odds ratio (OR) and 95% CIs of the association between BMI status and CDI were retrieved from each study and combined using the generic inverse variance method. Funnel plot was used to assess publication bias.

RESULTS

A total of 4609 articles were identified. After two rounds of systematic review, 17 studies met the eligibility criteria and were included into the meta-analysis. Pooled analysis showed that individuals with high BMI had a significantly decreased odds of CDI with the pooled OR of 0.88 (95% CI 0.80-0.97). This meta-analysis had high statistical heterogeneity with I² of 74%. Funnel plot was symmetric, which was not suggestive of presence of publication bias.

CONCLUSION

This meta-analysis revealed a significant negative association between BMI and CDI.

Early supplemental parenteral nutrition is associated with reduced mortality in critically ill surgical patients with high nutritional risk

BACKGROUND & AIMS

Adequate nutritional provision is important for critically ill patients to improve clinical outcomes. Starting enteral nutrition (EN) as early as possible is recommended and preferred to parenteral nutrition (PN). However, patients who undergo emergency abdominal operations may have alterations in their intra-abdominal environment and gastrointestinal motility leading to limitation in starting an enteral diet. Therefore, our study was designed to evaluate the benefit of early supplemental PN to achieve adequate calorie and protein supply in critically ill patients undergoing surgery who are not eligible for early EN.

METHODS

We reviewed the medical records of 317 patients who underwent emergency abdominal surgery for complicated intra-abdominal infection (cIAI) between January 2013 and December 2018. The nutritional data of the patients were collected for 7 days in maximum, starting on the day of intensive care unit (ICU) admission. The patients were divided by low or high malnutrition risk using the modified Nutrition Risk in Critically ill (mNUTRIC) score and body mass index. The low- and high-risk groups were subdivided into the following two categories: those who received PN within 48 h ("early") and those who did not ("usual"). Data regarding the baseline characteristics, initial severity of illness, morbidity, and mortality rates were also obtained. The average calorie and protein supply per day were calculated in these groups.

RESULTS

Patients in all groups showed no significant differences in baseline characteristics, initial status, and infectious complications. In terms of outcomes, patients with low malnutrition risk had no significant difference in mortality. However, among patients with high malnutrition risk, the "Early" group had lower rates of 30-day mortality (7.6% vs. 26.7%, p = 0.006) and in-hospital mortality (13.6% vs. 28.9%, p = 0.048) than those of the "Usual" group. Kaplan-Meier survival curves for 30-day mortality in these groups also showed a statistically significant difference (p = 0.001). The caloric adequacy of the "Early" group and the "Usual" group were 0.88 ± 0.34 and 0.6 ± 0.29, respectively. Amounts of protein received were 0.94 ± 0.39 g/kg in the "Early" group and 0.47 ± 0.34 g/kg in the "Usual" group, respectively. There was no significant difference in infectious complications between both groups.

CONCLUSIONS

Mortality in patients with high malnutrition risk who received early PN supply within 48 h after emergency surgery for cIAI was lower than those who did not receive PN earlier. PN may be necessary to fulfill the caloric and protein requirements for critically ill patients who cannot achieve their nutritional requirements to the fullest with EN alone.

Sex Differences in Short- and Long-Term Survival Among Critically Ill Patients with Sepsis

Background

Currently, there have been studies showing a correlation between sex differences and prognosis. Nevertheless, the conclusions of clinical studies on sex-based differences are controversial. We aimed to evaluate the effect of sex on the short- and long-term survival of critically ill patients with sepsis.

Methods

We use the critical care database of the healthcare information mart. Cox models were conducted to determine the relationship of 28-day and 1-year mortality with a different sex. Interaction and stratified analyses were conducted to test whether the effect of sex differed across age and sequential organ failure assessment (SOFA) score subgroups.

Results

A total of 12,321 patients were enrolled in this study. The Cox regression analysis showed that the 28-day and 1-year mortality rates of female patients were significantly lower than those of male patients by 10% and 8%, respectively (hazard ratio [HR]=0.90, 95% confidence interval [CI] 0.83–0.98, and HR=0.92, 95% CI 0.87–0.97, respectively). The effects of the association between sex and 28-day and 1-year mortality were broadly consistent for age and the SOFA subgroup variables. Only age was observed to have significant interactions in the 1-year mortality (P=0.0177). Compared with male patients, female patients aged <50 years had a long-term survival advantage (HR=0.77, 95% CI 0.62–0.95). In contrast, we did not find sex-based differences in the short- and long-term survival for patients aged ≥50 years.

Conclusion

In the current retrospective large database review, the 28-day and 1-year mortality were significantly lower in females than in male patients among critically ill patients with sepsis. Notably, there was an interaction between age and sex, and whether female-associated hormones or other contributing factors affect the clinical outcomes of patients with sepsis needs to be further researched.

Association of Diabetes and Admission Blood Glucose Levels with Short-Term Outcomes in Patients with Critical Illnesses

Background

Association of diabetes and admission glucose on the short-term prognosis in patients with critical illnesses are currently ambiguous. We aimed to determine whether diabetes and admission glucose affects short-term prognosis of critically ill patients.

Methods

We performed a retrospective analysis of data on 46,476 critically ill patients from the critical care database. Association of diabetes with 28-day mortality was assessed by inverse probability weighting based on the propensity score. Smoothing splines and threshold effect analysis were applied to explore the relationship between admission glucose and clinical outcomes.

Results

Of the 33,680 patients enrolled in the study, 8,701 (25.83%) had diabetes. In the main analysis, the 28-day mortality was reduced by 29% (hazard ratio (HR)=0.71, 95% confidence interval (CI) 0.67–0.76) in patients with diabetes compared to those without diabetes. The E-value of 2.17 indicated robustness to unmeasured confounders. Significant interactions were observed for glucose at ICU admission, admission type, and insulin use (Interaction P <0.05). A V-shaped relationship was observed between admission glucose and 28-day mortality in non-diabetic patients, with the lowest 28-day mortality corresponding to a glucose level of 101.75 mg/dl (95% CI 94.64–105.80 mg/dl), and admission hypoglycemia or hyperglycemia should be avoided, especially in patients admitted to the surgical intensive care unit (SICU), cardiac surgery recovery unit (CSRU), and coronary care unit (CCU); for diabetic patients, elevated admission glucose does not appear to be associated with a poor prognosis and perhaps may be beneficial except for CCU and CSRU.

Conclusion

The non-detrimental effect of diabetes on the short-term prognosis of critically ill patients was further confirmed, which would reduce 28-day mortality by approximately 29%. For non-diabetic patients, the admission glucose level corresponding to the lowest 28-day mortality was 101.75 mg/dl (95% CI 94.64–105.80 mg/dl); however, for diabetics, the appropriate admission glucose threshold remains unresolved.