Srivali, Narat

Background: Lactic acidosis is a heterogeneous condition with multiple underlying causes and associated outcomes. The use of multi-dimensional patient data to subtype lactic acidosis can personalize patient care. Machine learning consensus clustering may identify lactic acidosis subgroups with unique clinical profiles and outcomes. Methods: We used the Medical Information Mart for Intensive Care III database to abstract electronic medical record data from patients admitted to intensive care units (ICU) in a tertiary care hospital in the United States. We included patients who developed lactic acidosis (defined as serum lactate ≥ 4 mmol/L) within 48 h of ICU admission. We performed consensus clustering analysis based on patient characteristics, comorbidities, vital signs, organ supports, and laboratory data to identify clinically distinct lactic acidosis subgroups. We calculated standardized mean differences to show key subgroup features. We compared outcomes among subgroups. Results: We identified 1919 patients with lactic acidosis. The algorithm revealed three best unique lactic acidosis subgroups based on patient variables. Cluster 1 (n = 554) was characterized by old age, elective admission to cardiac surgery ICU, vasopressor use, mechanical ventilation use, and higher pH and serum bicarbonate. Cluster 2 (n = 815) was characterized by young age, admission to trauma/surgical ICU with higher blood pressure, lower comorbidity burden, lower severity index, and less vasopressor use. Cluster 3 (n = 550) was characterized by admission to medical ICU, history of liver disease and coagulopathy, acute kidney injury, lower blood pressure, higher comorbidity burden, higher severity index, higher serum lactate, and lower pH and serum bicarbonate. Cluster 3 had the worst outcomes, while cluster 1 had the most favorable outcomes in terms of persistent lactic acidosis and mortality. Conclusions: Consensus clustering analysis synthesized the pattern of clinical and laboratory data to reveal clinically distinct lactic acidosis subgroups with different outcomes.

Background: Lactic acidosis is the most common cause of anion gap metabolic acidosis in the intensive care unit (ICU), associated with poor outcomes including mortality. We sought to compare machine learning (ML) approaches versus logistic regression analysis for prediction of mortality in lactic acidosis patients admitted to the ICU. Methods: We used the Medical Information Mart for Intensive Care (MIMIC-III) database to identify ICU adult patients with lactic acidosis (serum lactate ≥4 mmol/L). The outcome of interest was hospital mortality. We developed prediction models using four ML approaches consisting of random forest (RF), decision tree (DT), extreme gradient boosting (XGBoost), artificial neural network (ANN), and statistical modeling with forward stepwise logistic regression using the testing dataset. We then assessed model performance using area under the receiver operating characteristic curve (AUROC), accuracy, precision, error rate, Matthews correlation coefficient (MCC), F1 score, and assessed model calibration using the Brier score, in the independent testing dataset. Results: Of 1919 lactic acidosis ICU patients, 1535 and 384 were included in the training and testing dataset, respectively. Hospital mortality was 30%. RF had the highest AUROC at 0.83, followed by logistic regression 0.81, XGBoost 0.81, ANN 0.79, and DT 0.71. In addition, RF also had the highest accuracy (0.79), MCC (0.45), F1 score (0.56), and lowest error rate (21.4%). The RF model was the most well-calibrated. The Brier score for RF, DT, XGBoost, ANN, and multivariable logistic regression was 0.15, 0.19, 0.18, 0.19, and 0.16, respectively. The RF model outperformed multivariable logistic regression model, SOFA score (AUROC 0.74), SAP II score (AUROC 0.77), and Charlson score (AUROC 0.69). Conclusion: The ML prediction model using RF algorithm provided the highest predictive performance for hospital mortality among ICU patient with lactic acidosis.

AbstractBackgroundWe aimed to report the incidence of hospital-acquired hypophosphataemia and hyperphosphataemia along with their associated in-hospital mortality.MethodsWe included 15 869 adult patients hospitalised at a tertiary medical referral centre from January 2009 to December 2013, who had normal serum phosphate levels at admission and at least two serum phosphate measurements during their hospitalisation. The normal range of serum phosphate was defined as 2.5–4.2 mg/dL. In-hospital serum phosphate levels were categorised based on the occurrence of hospital-acquired hypophosphataemia and hyperphosphataemia. We analysed the association of hospital-acquired hypophosphataemia and hyperphosphataemia with in-hospital mortality using multivariable logistic regression.ResultsFifty-three per cent (n=8464) of the patients developed new serum phosphate derangements during their hospitalisation. The incidence of hospital-acquired hypophosphataemia and hyperphosphataemia was 35% and 27%, respectively. Hospital-acquired hypophosphataemia and hyperphosphataemia were associated with odds ratio (OR) of 1.56 and 2.60 for in-hospital mortality, respectively (p value<0.001 for both). Compared with patients with persistently normal in-hospital phosphate levels, patients with hospital-acquired hypophosphataemia only (OR 1.64), hospital-acquired hyperphosphataemia only (OR 2.74) and both hospital-acquired hypophosphataemia and hyperphosphataemia (ie, phosphate fluctuations; OR 4.00) were significantly associated with increased in-hospital mortality (all p values <0.001).ConclusionHospital-acquired serum phosphate derangements affect approximately half of the hospitalised patients and are associated with increased in-hospital mortality rate.

Background and objectives: Goodpasture’s syndrome (GS) is a rare, life-threatening autoimmune disease. Although the coexistence of anti-neutrophil cytoplasmic antibody (ANCA) with Goodpasture’s syndrome has been recognized, the impacts of ANCA vasculitis on mortality and resource utilization among patients with GS are unclear. Materials and Methods: We used the National Inpatient Sample to identify hospitalized patients with a principal diagnosis of GS from 2003 to 2014 in the database. The predictor of interest was the presence of ANCA-associated vasculitis. We tested the differences concerning in-hospital treatment and outcomes between GS patients with and without ANCA-associated vasculitis using logistic regression analysis with adjustment for other clinical characteristics. Results: A total of 964 patients were primarily admitted to hospital for GS. Of these, 84 (8.7%) had a concurrent diagnosis of ANCA-associated vasculitis. Hemoptysis was more prevalent in GS patients with ANCA-associated vasculitis. During hospitalization, GS patients with ANCA-associated required non-significantly more mechanical ventilation and non-invasive ventilation support, but non-significantly less renal replacement therapy and plasmapheresis than those with GS alone. There was no significant difference in in-hospital outcomes, including organ failure and mortality, between GS patients with and without ANCA-associated vasculitis. Conclusions: Our study demonstrated no significant differences between resource utilization and in-hospital mortality among hospitalized patients with coexistence of ANCA vasculitis and GS, compared to those with GS alone.

Background: Goodpasture’s syndrome is a rare, life-threatening, small vessel vasculitis. Given its rarity, data on its inpatient burden and resource utilization are lacking. We conducted this study aiming to assess inpatient prevalence, mortality, and resource utilization of Goodpasture’s syndrome in the United States. Methods: The 2003–2014 National Inpatient Sample was used to identify patients with a principal diagnosis of Goodpasture’s syndrome. The inpatient prevalence, clinical characteristics, in-hospital treatment, end-organ failure, mortality, length of hospital stay, and hospitalization cost were studied. Multivariable logistic regression was performed to identify independent factors associated with in-hospital mortality. Results: A total of 964 patients were admitted in hospital with Goodpasture’s syndrome as the principal diagnosis, accounting for an overall inpatient prevalence of Goodpasture’s syndrome among hospitalized patients in the United States of 10.3 cases per 1,000,000 admissions. The mean age of patients was 54 ± 21 years, and 47% were female; 52% required renal replacement therapy, whereas 39% received plasmapheresis during hospitalization. Furthermore, 78% had end-organ failure, with renal failure and respiratory failure being the two most common end-organ failures. The in-hospital mortality rate was 7.7 per 100 admissions. The factors associated with increased in-hospital mortality were age older than 70 years, sepsis, the development of respiratory failure, circulatory failure, renal failure, and liver failure, whereas the factors associated with decreased in-hospital mortality were more recent year of hospitalization and the use of therapeutic plasmapheresis. The median length of hospital stay was 10 days. The median hospitalization cost was $75,831. Conclusion: The inpatient prevalence of Goodpasture’s syndrome in the United States is 10.3 cases per 1,000,000 admissions. Hospitalization of patients with Goodpasture’s syndrome was associated with high hospital inpatient utilization and costs.

Background: Lung transplantation has been increasingly performed worldwide and is considered an effective therapy for patients with various causes of end-stage lung diseases. We performed a systematic review to assess the incidence and impact of acute kidney injury (AKI) and severe AKI requiring renal replacement therapy (RRT) in patients after lung transplantation. Methods: A literature search was conducted utilizing Ovid MEDLINE, EMBASE, and Cochrane Database from inception through June 2019. We included studies that evaluated the incidence of AKI, severe AKI requiring RRT, and mortality risk of AKI among patients after lung transplantation. Pooled incidence and odds ratios (ORs) with 95% confidence interval (CI) were obtained using random-effects meta-analysis. The protocol for this meta-analysis is registered with PROSPERO (International Prospective Register of Systematic Reviews; no. CRD42019134095). Results: A total of 26 cohort studies with a total of 40,592 patients after lung transplantation were enrolled. Overall, the pooled estimated incidence rates of AKI (by standard AKI definitions) and severe AKI requiring RRT following lung transplantation were 52.5% (95% CI: 45.8–59.1%) and 9.3% (95% CI: 7.6–11.4%). Meta-regression analysis demonstrated that the year of study did not significantly affect the incidence of AKI (p = 0.22) and severe AKI requiring RRT (p = 0.68). The pooled ORs of in-hospital mortality in patients after lung transplantation with AKI and severe AKI requiring RRT were 2.75 (95% CI, 1.18–6.41) and 10.89 (95% CI, 5.03–23.58). At five years, the pooled ORs of mortality among patients after lung transplantation with AKI and severe AKI requiring RRT were 1.47 (95% CI, 1.11–1.94) and 4.79 (95% CI, 3.58–6.40), respectively. Conclusion: The overall estimated incidence rates of AKI and severe AKI requiring RRT in patients after lung transplantation are 52.5% and 9.3%, respectively. Despite advances in therapy, the incidence of AKI in patients after lung transplantation does not seem to have decreased. In addition, AKI after lung transplantation is significantly associated with reduced short-term and long-term survival.

Background: Although acute kidney injury (AKI) is a frequent complication in patients receiving extracorporeal membrane oxygenation (ECMO), the incidence and impact of AKI on mortality among patients on ECMO remain unclear. We conducted this systematic review to summarize the incidence and impact of AKI on mortality risk among adult patients on ECMO. Methods: A literature search was performed using EMBASE, Ovid MEDLINE, and Cochrane Databases from inception until March 2019 to identify studies assessing the incidence of AKI (using a standard AKI definition), severe AKI requiring renal replacement therapy (RRT), and the impact of AKI among adult patients on ECMO. Effect estimates from the individual studies were obtained and combined utilizing random-effects, generic inverse variance method of DerSimonian-Laird. The protocol for this systematic review is registered with PROSPERO (no. CRD42018103527). Results: 41 cohort studies with a total of 10,282 adult patients receiving ECMO were enrolled. Overall, the pooled estimated incidence of AKI and severe AKI requiring RRT were 62.8% (95%CI: 52.1%–72.4%) and 44.9% (95%CI: 40.8%–49.0%), respectively. Meta-regression showed that the year of study did not significantly affect the incidence of AKI (p = 0.67) or AKI requiring RRT (p = 0.83). The pooled odds ratio (OR) of hospital mortality among patients receiving ECMO with AKI on RRT was 3.73 (95% CI, 2.87–4.85). When the analysis was limited to studies with confounder-adjusted analysis, increased hospital mortality remained significant among patients receiving ECMO with AKI requiring RRT with pooled OR of 3.32 (95% CI, 2.21–4.99). There was no publication bias as evaluated by the funnel plot and Egger’s regression asymmetry test with p = 0.62 and p = 0.17 for the incidence of AKI and severe AKI requiring RRT, respectively. Conclusion: Among patients receiving ECMO, the incidence rates of AKI and severe AKI requiring RRT are high, which has not changed over time. Patients who develop AKI requiring RRT while on ECMO carry 3.7-fold higher hospital mortality.

Background: Several immune-mediated inflammatory disorders, such as rheumatoid arthritis, psoriatic arthritis, and systemic lupus erythematosus have been linked to an increased risk of venous thromboembolism (VTE). However, the data on ankylosing spondylitis (AS) are limited. Methods: We conducted a systematic review and meta-analysis of observational studies that reported odds ratio, relative risk, hazard ratio, or standardized incidence ratio comparing the risk of VTE and possible pulmonary embolism (PE) in patients with AS versus non-AS participants. Pooled risk ratio and 95% confidence intervals were calculated using a random-effect, generic inverse variance method of DerSimonian and Laird. Results: Of 423 potentially relevant articles, three studies met our inclusion criteria and thus, were included in the data analysis. The pooled risk ratio of VTE in patients with AS was 1.60 (95% confidence interval: 1.05–2.44). The statistical heterogeneity of this study was high with an I 2 of 93%. Conclusion: Our study demonstrated a statistically significant increased VTE risk among patients with AS.