Insulin resistance is a cardiovascular risk factor in hypertensive adults without type 2 diabetes mellitus.

BACKGROUND: Metabolic syndrome refers to the association among several cardiovascular risk factors: obesity, dyslipidemia, hyperglycemia, and hypertension. It is associated with increased cardiovascular risk and the development of type 2 diabetes mellitus. Insulin resistance is the underlying mechanism of metabolic syndrome, although its role in increased cardiovascular risk has not been directly identified. OBJECTIVE: We investigated the association between insulin resistance and increased cardiovascular risk in hypertensive adults without diabetes mellitus. DESIGN AND PARTICIPANTS: We enrolled participants without diabetes from an outpatient setting in a retrospective, longitudinal study. Several demographic, clinical, and laboratory parameters were recorded during the observation period. Plasma insulin and homeostatic model assessment for insulin resistance (HOMA-IR) were used to determine insulin resistance and four cardiovascular events (acute coronary disease, acute cerebrovascular disease, incident heart failure, and cardiovascular mortality) were combined into a single outcome. Logistic regression and Cox proportional hazards models were fitted to evaluate the association between covariates and outcomes. RESULTS: We included 1899 hypertensive adults without diabetes with an average age of 53 years (51.3% women, 23% had prediabetes, and 64.2% had metabolic syndrome). In a logistic regression analysis, male sex (odds ratio, OR = 1.66) having high levels of low-density lipoprotein (LDL, OR = 1.01), kidney function (OR = 0.97), and HOMA-IR (OR = 1.06) were associated with the incidence of cardiovascular events; however, in a survival multivariate analysis, only HOMA-IR (hazard ratio, HR 1.4, 95% confidence interval, CI: 1.05-1.87, p = 0.02) and body mass index (HR 1.05, 95% CI: 1.02-1.08, p = 0.002) were considered independent prognostic variables for the development of incident cardiovascular events. CONCLUSION: Insulin resistance and obesity are useful for assessing cardiovascular risk in hypertensive people without diabetes but with preserved kidney function. This work demonstrates the predictive value of the measurement of insulin, and therefore of insulin resistance, in an outpatient setting and attending to high-risk patients.

Insulin Resistance and Metabolic Syndrome as Risk Factors for Hospitalization in Patients with COVID-19: Pilot Study on the Use of Machine Learning.

Aim: Conditions linked to metabolic syndrome, such as obesity, hypertension, insulin resistance, and dyslipidemia, are common in patients with severe coronavirus disease 2019 (COVID-19). These conditions can act synergistically to contribute to negative outcomes. We describe and analyze the relationship between metabolic syndrome and COVID-19 severity in terms of risk of hospitalization. Methods: We designed a retrospective, cross-sectional study, including patients with confirmed COVID-19 diagnosis. Clinical and laboratory parameters regarding metabolic syndrome were collected. The Homeostatic Model Assessment of Insulin Resistance (HOMA-IR) was used to assess insulin resistance. The outcome was needed for hospitalization. Logistic regression was used to calculate odds ratios, and to determine the association between variables and risk of hospitalization. Advanced approaches using machine learning were also used to identify and interpret the effects of predictors on the proposed outcome. Results: We included 2716 COVID-19 patients with a mean age of 61.8 years. Of these, 48.9% were women, 28.9% had diabetes, and 50.6% were diagnosed with metabolic syndrome. Overall, 212 patients required hospitalization. Patients with metabolic syndrome had a 58% greater chance of hospitalization if they were men, 32% if they had metabolic syndrome, and 23% if they were obese. Machine learning methods identified body mass index, metabolic syndrome, systolic blood pressure, and HOMA-IR as the most relevant features for our predictive model. Conclusion: Metabolic syndrome and its related biomarkers increase the odds for a severe clinical course of COVID-19 and the need for hospitalization. Machine learning methods can aid understanding of the effects of single features when assessing risks for a given outcome.

Hospitalization burden and epidemiology of the COVID-19 pandemic in Spain (2020-2021).

BACKGROUND: Spain had some of Europe's highest incidence and mortality rates for coronavirus disease 2019 (COVID-19). Here we describe the epidemiology and trends in hospitalizations, the number of critical patients, and deaths in Spain in 2020 and 2021. METHODS: We performed a descriptive, retrospective, nationwide study using an administrative database, the Minimum Basic Data Set at Hospitalization, which includes 95-97% of discharge reports for patients hospitalized in Spain in 2020 and 2021. We analyzed the number of hospitalizations, admissions to intensive care units, and deaths and their geographic distribution across regions of Spain. RESULTS: As of December 31, 2021, a total of 498,789 patients (1.04% of the entire Spanish population) had needed hospitalization. At least six waves of illness were identified. Men were more prone to hospitalization than women. The median age was 66. A total of 54,340 patients (10.9% of all hospitalizations) had been admitted to the intensive care unit. We identified 71,437 deaths (mortality rate of 14.3% among hospitalized patients). We also observed important differences among regions, with Madrid being the epicenter of hospitalizations and mortality. CONCLUSIONS: We analyzed Spain's response to COVID-19 and describe here its experiences during the pandemic in terms of hospitalizations, critical illness, and deaths. This research highlights changes over several months and waves and the importance of factors such as vaccination, the predominant variant of the virus, and public health interventions in the rise and fall of the outbreaks.

Differences in Trends in Admissions and Outcomes among Patients from a Secondary Hospital in Madrid during the COVID-19 Pandemic: A Hospital-Based Epidemiological Analysis (2020-2022).

Spain had some of Europe's highest incidence and mortality rates for coronavirus disease 2019 (COVID-19). This study highlights the impact of the COVID-19 pandemic on daily health care in terms of incidence, critical patients, and mortality. We describe the characteristics and clinical outcomes of patients, comparing variables over the different waves. We performed a descriptive, retrospective study using the historical records of patients hospitalized with COVID-19. We describe demographic characteristics, admissions, and occupancy. Time series allowed us to visualize and analyze trends and patterns, and identify several waves during the 27-month period. A total of 3315 patients had been hospitalized with confirmed COVID-19. One-third of these patients were hospitalized during the first weeks of the pandemic. We observed that 4.6% of all hospitalizations had been admitted to the intensive care unit, and we identified a mortality rate of 9.4% among hospitalized patients. Arithmetic- and semi-logarithmic-scale charts showed how admissions and deaths rose sharply during the first weeks, increasing by 10 every few days. We described a single hospital's response and experiences during the pandemic. This research highlights certain demographic profiles in a population and emphasizes the importance of identifying waves when performing research on COVID-19. Our results can extend the analysis of the impact of COVID-19 and can be applied in other contexts, and can be considered when further analyzing the clinical, epidemiological, or demographic characteristics of populations with COVID-19. Our findings suggest that the pandemic should be analyzed not as a whole but rather in different waves.

Hospitalization Burden Related to Herpes Zoster Infection in Spain (2016-2019).

INTRODUCTION: Herpes zoster (HZ) and its complications still represent a significant burden for patients and health care systems. In Spain, vaccination is progressively being introduced and recommended for patients between 65 and 80 years old and patients > 18 years of age suffering from certain immunosuppression conditions. The aim of this study is to estimate the number of hospital admissions related to HZ from 2016 to 2019 in Spain. METHODS: Data were collected from the Minimum Basic DataSet (MBDS) and codified according to the Spanish version of the 10th International Classification of Disease (ICD-10-CM codes B02-B02.9). Among others, variables such as sex, age and presence of complications were included. RESULTS: A total of 27,642 hospitalizations were identified (90% in patients > 50 and 45.8% in patients > 80). Women represented 51.2% of the patients, and 59.9% of patients presented complications related to HZ. The hospitalization rate was 17.74, the mortality rate was 1.2, and the case fatality rate was 6.75%. All rates were significantly higher with age, among men and in complicated HZ. Immunosuppression status for which vaccination had been recommended represented 22.7% of the total cases, affecting mostly individuals > 65 and causing more deaths in those > 80 years. The estimated annual cost of hospitalization for herpes zoster was €35,738,285, and the mean cost per patient was €5172. CONCLUSION: The hospitalization burden for HZ is still important in Spain. Data on the current epidemiology are important to evaluate future vaccination strategies.

Delayed Immune Reconstitution Inflammatory Syndrome in an Immunosuppressed Patient With SARS-CoV-2.

Both immune reconstitution inflammatory syndrome (IRIS) and severe coronavirus disease 2019 (COVID-19) are marked by hyperinflammation as a consequence of dysfunction in myeloid cells and increased production of proinflammatory cytokines. Although these features are common to both diseases, their physiopathology remains unclear. Here we report the case of a 63-year-old woman admitted for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. In her clinical course, she developed acute respiratory distress syndrome, probably triggered by the use of granulocyte colony-stimulating factor (G-CSF). We hypothesize that G-CSF unmasked IRIS.

Delayed Diagnosis of an Atypical Pneumonia Resembling a Solitary Pulmonary Nodule.

Atypical pneumonia shows clinical features that are different from those of typical pneumonia, and it can mimic other entities. We report the case of a 42-year-old male with a solitary pulmonary nodule found in an X-ray for a preoperative evaluation. Our patient was asymptomatic, and a pulmonary neoplasm was the first diagnostic suspicion. The round-shaped nodule seen in the X-ray turned out to be a linear ground glass opacity in a thoracic CT scan. Viral pneumonia due to SARS-CoV-2 was diagnosed. We emphasize here the educational value of this case report. We do not report a new radiological finding because lung nodules resembling neoplasms have already been reported in the medical literature. However, some clinical features of COVID-19 are relatively new and can mimic other entities, and the results of some investigations and clinicians' interpretations of them can be misleading. Atypical radiological findings make it necessary to widen the spectrum of alternative diagnoses.

Assessment of Classification Models and Relevant Features on Nonalcoholic Steatohepatitis Using Random Forest.

Nonalcoholic fatty liver disease (NAFLD) is the hepatic manifestation of metabolic syndrome and is the most common cause of chronic liver disease in developed countries. Certain conditions, including mild inflammation biomarkers, dyslipidemia, and insulin resistance, can trigger a progression to nonalcoholic steatohepatitis (NASH), a condition characterized by inflammation and liver cell damage. We demonstrate the usefulness of machine learning with a case study to analyze the most important features in random forest (RF) models for predicting patients at risk of developing NASH. We collected data from patients who attended the Cardiovascular Risk Unit of Mostoles University Hospital (Madrid, Spain) from 2005 to 2021. We reviewed electronic health records to assess the presence of NASH, which was used as the outcome. We chose RF as the algorithm to develop six models using different pre-processing strategies. The performance metrics was evaluated to choose an optimized model. Finally, several interpretability techniques, such as feature importance, contribution of each feature to predictions, and partial dependence plots, were used to understand and explain the model to help obtain a better understanding of machine learning-based predictions. In total, 1525 patients met the inclusion criteria. The mean age was 57.3 years, and 507 patients had NASH (prevalence of 33.2%). Filter methods (the chi-square and Mann-Whitney-Wilcoxon tests) did not produce additional insight in terms of interactions, contributions, or relationships among variables and their outcomes. The random forest model correctly classified patients with NASH to an accuracy of 0.87 in the best model and to 0.79 in the worst one. Four features were the most relevant: insulin resistance, ferritin, serum levels of insulin, and triglycerides. The contribution of each feature was assessed via partial dependence plots. Random forest-based modeling demonstrated that machine learning can be used to improve interpretability, produce understanding of the modeled behavior, and demonstrate how far certain features can contribute to predictions.

Machine learning approaches to constructing predictive models of vitamin D deficiency in a hypertensive population: a comparative study.

Objective: Given the association between vitamin D deficiency and risk for cardiovascular disease, we used machine learning approaches to establish a model to predict the probability of deficiency. Determination of serum levels of 25-hydroxy vitamin D (25(OH)D) provided the best assessment of vitamin D status, but such tests are not always widely available or feasible. Thus, our study established predictive models with high sensitivity to identify patients either unlikely to have vitamin D deficiency or who should undergo 25(OH)D testing.Methods: We collected data from 1002 hypertensive patients from a Spanish university hospital. The elastic net regularization approach was applied to reduce the dimensionality of the dataset. The issue of determining vitamin D status was addressed as a classification problem; thus, the following classifiers were applied: logistic regression, support vector machine (SVM), random forest, naive Bayes, and Extreme Gradient Boost methods. Classification accuracy, sensitivity, specificity, and predictive values were computed to assess the performance of each method.Results: The SVM-based method with radial kernel performed better than the other algorithms in terms of sensitivity (98%), negative predictive value (71%), and classification accuracy (73%).Conclusion: The combination of a feature-selection method such as elastic net regularization and a classification approach produced well-fitted models. The SVM approach yielded better predictions than the other algorithms. This combination approach allowed us to develop a predictive model with high sensitivity but low specificity, to identify the population that could benefit from laboratory determination of serum levels of 25(OH)D.

Retrospective Cohort Noninferiority Analysis of Barrier Cream in Preventing Pressure Injuries.

OBJECTIVE: To assess the effectiveness of a dimethicone- and zinc-based barrier cream compared with hyperoxygenated fatty acids in preventing pressure injuries (PIs) in patients at high or very high risk. METHODS: Researchers conducted a retrospective noninferiority study in an inpatient acute care hospital in Spain that included hospitalized patients in nonsurgical departments with impaired mobility. RESULTS: The study authors reviewed 522 patients in a control group (hyperoxygenated fatty acids) and an experimental group (barrier cream) over a period of 7 days. The incidence of PI was 31% in the control group and 31.1% in the experimental group. The hazard ratio for developing PI was 0.84 (confidence interval, 0.61-1.17; P = .32) in the experimental group compared with the control group, meeting the criteria for noninferiority. The Kaplan-Meier estimator indicated no statistically significant difference between groups (log-rank = 0.654). CONCLUSIONS: Dimethicone- and zinc-based barrier cream was not inferior to hyperoxygenated fatty acids in preventing PIs in hospitalized patients at high or very high risk of developing them during their hospital stay.

The Use of Machine Learning Techniques to Determine the Predictive Value of Inflammatory Biomarkers in the Development of Type 2 Diabetes Mellitus.

Background: Certain inflammatory biomarkers, such as interleukin-6, interleukin-1, C-reactive protein (CRP), and fibrinogen, are prototypical acute-phase parameters that can also be predictors of cardiovascular disease. However, this inflammatory response can also be linked to the development of type 2 diabetes mellitus (T2DM). Methods: We performed a cross-sectional, retrospective study of hypertensive patients in an outpatient setting. Demographic, clinical, and laboratory parameters, such as the Homeostatic Model Assessment of Insulin Resistance (HOMA-IR), CRP, and fibrinogen, were recorded. The outcome was progression to overt T2DM over the 12-year observation period. Results: A total of 3,472 hypertensive patients were screened, but 1,576 individuals without T2DM were ultimately included in the analyses. Patients with elevated fibrinogen, CRP, and insulin resistance had a significantly greater incidence of progression to T2DM. During follow-up, 199 patients progressed to T2DM. Multivariate logistic regression analyses showed that body mass index [odds ratio (OR) 1.04, 95% confidence interval (CI): 1.01-1.07], HOMA-IR (OR 1.13, 95% CI: 1.08-1.16), age (OR 1.05, 95% CI: 1.03-1.07), log(CRP) (OR 1.37, 95% CI: 1.14-1.55), and fibrinogen (OR 1.44, 95% CI: 1.23-1.66) were the most important predictors of progression to T2DM. The area under the receiver operating characteristic curve (AUC) of this model was 0.76. Using machine learning methods, we built a model that included HOMA-IR, fibrinogen, and log(CRP) that was more accurate than the logistic regression model, with an AUC of 0.9. Conclusion: Our results suggest that inflammatory biomarkers and HOMA-IR have a strong prognostic value in predicting progression to T2DM. Machine learning methods can provide more accurate results to better understand the implications of these features in terms of progression to T2DM. A successful therapeutic approach based on these features can avoid progression to T2DM and thus improve long-term survival.

Listeria monocytogenes infections: Analysis of 41 patients / Infección por Listeria monocytogenes: análisis de 41 pacientes

OBJECTIVE: We describe and analyze Listeria-related demographics and clinical features to determine the predisposing conditions for severe infections. METHODS: We performed a retrospective study using positive isolation of Listeria monocytogenes from blood, cerebrospinal fluid, and other organic fluids. Electronic health records were used to determine the epidemiological and clinical features of infections caused by L. monocytogenes. Mortality and sepsis were considered dependent variables in the statistical analyses. RESULTS: We included 41 patients in an observation period of 15 years (2003-2018), with an annual incidence rate of 1.3 cases per 100,000 population. Three main population profiles were identified: newborns, pregnant women, and other adults (17.1%, 12.2%, and 82.9%, respectively). Neuroinvasive infection was present in 17 patients (41.5%). In both univariate and multivariate analyses, neurological infections, whether meningoencephalitis, rhombencephalitis, or brain abscesses, were the main risk factors for severe forms of Listeria-related infections (odds ratio 1.8, 95% CI 1.52-2.14, p = 0.01). Malignancies, whether solid tumors or hematological neoplasms, immunosuppression, and chronic diseases were not related to either mortality or severe clinical syndromes. CONCLUSION: Infections caused by L. monocytogenes were uncommon but could cause severe sepsis and mortality, especially in susceptible populations. Our study focused on neurological involvement and severe invasive forms of listeriosis. Neuroinvasive forms were the most important risk factors for severe illness but not for mortality

INTRODUCCIÓN: Describir y analizar las características demográficas y clínicas de las infecciones por Listeria para determinar los factores predisponentes para infecciones severas. MÉTODOS: Diseñamos un estudio retrospectivo utilizando los aislamientos positivos de Listeria monocytogenes en sangre, líquido cefalorraquídeo u otros fluidos orgánicos. Se obtuvieron los registros electrónicos para conseguir las características clínicas y epidemiológicas de las infecciones causadas por L. monocytogenes. Mortalidad y sepsis fueron las variables dependientes en los análisis estadísticos. RESULTADOS: Se incluyeron 41 pacientes en un período de 15 años (2003-2018), con una incidencia anual de 1,3 casos por cada 100.000 habitantes. Identificamos tres perfiles de población: neonatos, mujeres embarazadas y resto de adultos (el 17,1%, el 12,2% y el 82,9%, respectivamente). Las formas neuroinvasivas se identificaron en 17 pacientes (41,5%). Tanto en los análisis univariados como en los multivariados, las infecciones neurológicas, bien meningoencefalitis, rombencefalitis o abscesos cerebrales, fueron los principales factores de riesgo para considerar formas severas de infección por Listeria (odds ratio 1,8; IC 1,52-2,14, p = 0,01). Las neoplasias sólidas o hematológicas, la inmunosupresión o las enfermedades crónicas no estuvieron relacionadas ni con la mortalidad ni con la presencia de severidad en la infección. CONCLUSIÓN: Las infecciones causadas por L. monocytogenes son infrecuentes, pero son causa de sepsis severa y mortalidad en poblaciones susceptibles. Nuestro estudio estuvo dirigido a la infección neuroinvasiva y otras formas graves. La forma neuroinvasiva fue el factor de riesgo más importante asociado a la infección severa, pero no a la mortalidad

Use of a K-nearest neighbors model to predict the development of type 2 diabetes within 2 years in an obese, hypertensive population.

Prediabetes is a type of hyperglycemia in which patients have blood glucose levels above normal but below the threshold for type 2 diabetes mellitus (T2DM). Prediabetic patients are considered to be at high risk for developing T2DM, but not all will eventually do so. Because it is difficult to identify which patients have an increased risk of developing T2DM, we developed a model of several clinical and laboratory features to predict the development of T2DM within a 2-year period. We used a supervised machine learning algorithm to identify at-risk patients from among 1647 obese, hypertensive patients. The study period began in 2005 and ended in 2018. We constrained data up to 2 years before the development of T2DM. Then, using a time series analysis with the features of every patient, we calculated one linear regression line and one slope per feature. Features were then included in a K-nearest neighbors classification model. Feature importance was assessed using the random forest algorithm. The K-nearest neighbors model accurately classified patients in 96% of cases, with a sensitivity of 99%, specificity of 78%, positive predictive value of 96%, and negative predictive value of 94%. The random forest algorithm selected the homeostatic model assessment-estimated insulin resistance, insulin levels, and body mass index as the most important factors, which in combination with KNN had an accuracy of 99% with a sensitivity of 99% and specificity of 97%. We built a prognostic model that accurately identified obese, hypertensive patients at risk for developing T2DM within a 2-year period. Clinicians may use machine learning approaches to better assess risk for T2DM and better manage hypertensive patients. Machine learning algorithms may help health care providers make more informed decisions.

Logistic LASSO and Elastic Net to Characterize Vitamin D Deficiency in a Hypertensive Obese Population.

Aim: The primary objective of our research was to compare the performance of data analysis to predict vitamin D deficiency using three different regression approaches and to evaluate the usefulness of incorporating machine learning algorithms into the data analysis in a clinical setting. Methods: We included 221 patients from our hypertension unit, whose data were collected from electronic records dated between 2006 and 2017. We used classical stepwise logistic regression, and two machine learning methods [least absolute shrinkage and selection operator (LASSO) and elastic net]. We assessed the performance of these three algorithms in terms of sensitivity, specificity, misclassification error, and area under the curve (AUC). Results: LASSO and elastic net regression performed better than logistic regression in terms of AUC, which was significantly better in both penalized methods, with AUC = 0.76 and AUC = 0.74 for elastic net and LASSO, respectively, than in logistic regression, with AUC = 0.64. In terms of misclassification rate, elastic net (18%) outperformed LASSO (22%) and logistic regression (25%). Conclusion: Compared with a classical logistic regression approach, penalized methods were found to have better performance in predicting vitamin D deficiency. The use of machine learning algorithms such as LASSO and elastic net may significantly improve the prediction of vitamin D deficiency in a hypertensive obese population.

Listeria monocytogenes infections: Analysis of 41 patients.

OBJECTIVE: We describe and analyze Listeria-related demographics and clinical features to determine the predisposing conditions for severe infections. METHODS: We performed a retrospective study using positive isolation of Listeria monocytogenes from blood, cerebrospinal fluid, and other organic fluids. Electronic health records were used to determine the epidemiological and clinical features of infections caused by L. monocytogenes. Mortality and sepsis were considered dependent variables in the statistical analyses. RESULTS: We included 41 patients in an observation period of 15 years (2003-2018), with an annual incidence rate of 1.3 cases per 100,000 population. Three main population profiles were identified: newborns, pregnant women, and other adults (17.1%, 12.2%, and 82.9%, respectively). Neuroinvasive infection was present in 17 patients (41.5%). In both univariate and multivariate analyses, neurological infections, whether meningoencephalitis, rhombencephalitis, or brain abscesses, were the main risk factors for severe forms of Listeria-related infections (odds ratio 1.8, 95% CI 1.52-2.14, p=0.01). Malignancies, whether solid tumors or hematological neoplasms, immunosuppression, and chronic diseases were not related to either mortality or severe clinical syndromes. CONCLUSION: Infections caused by L. monocytogenes were uncommon but could cause severe sepsis and mortality, especially in susceptible populations. Our study focused on neurological involvement and severe invasive forms of listeriosis. Neuroinvasive forms were the most important risk factors for severe illness but not for mortality.

Pulse Wave Velocity and Machine Learning to Predict Cardiovascular Outcomes in Prediabetic and Diabetic Populations.

Few studies have addressed the predictive value of arterial stiffness determined by pulse wave velocity (PWV) in a high-risk population with no prevalent cardiovascular disease and with obesity, hypertension, hyperglycemia, and preserved kidney function. This longitudinal, retrospective study enrolled 88 high-risk patients and had a follow-up time of 12.4 years. We collected clinical and laboratory data, as well as information on arterial stiffness parameters using arterial tonometry and measurements from ambulatory blood pressure monitoring. We considered nonfatal, incident cardiovascular events as the primary outcome. Given the small size of our dataset, we used survival analysis (i.e., Cox proportional hazards model) combined with a machine learning-based algorithm/penalization method to evaluate the data. Our predictive model, calculated with Cox regression and least absolute shrinkage and selection operator (LASSO), included body mass index, diabetes mellitus, gender (male), and PWV. We recorded 16 nonfatal cardiovascular events (5 myocardial infarctions, 5 episodes of heart failure, and 6 strokes). The adjusted hazard ratio for PWV was 1.199 (95% confidence interval: 1.09-1.37, p < 0.001). Arterial stiffness was a predictor of cardiovascular disease development, as determined by PWV in a high-risk population. Thus, in obese, hypertensive, hyperglycemic patients with preserved kidney function, PWV can serve as a prognostic factor for major adverse cardiac events.

Relevant Features in Nonalcoholic Steatohepatitis Determined Using Machine Learning for Feature Selection.

Aim: We investigated the prevalence and the most relevant features of nonalcoholic steatohepatitis (NASH), a stage of nonalcoholic fatty liver disease, (NAFLD) in which the inflammation of hepatocytes can lead to increased cardiovascular risk, liver fibrosis, cirrhosis, and the need for liver transplant. Methods: We analyzed data from 2239 hypertensive patients using descriptive statistics and supervised machine learning algorithms, including the least absolute shrinkage and selection operator and random forest classifier, to select the most relevant features of NASH. Results: The prevalence of NASH among our hypertensive patients was 11.3%. In univariate analyses, it was associated with metabolic syndrome, type 2 diabetes, insulin resistance, and dyslipidemia. Ferritin and serum insulin were the most relevant features in the final model, with a sensitivity of 70%, specificity of 79%, and area under the curve of 0.79. Conclusion: Ferritin and insulin are significant predictors of NASH. Clinicians may use these to better assess cardiovascular risk and provide better management to hypertensive patients with NASH. Machine-learning algorithms may help health care providers make decisions.

Identification of clinically relevant features in hypertensive patients using penalized regression: a case study of cardiovascular events.

Appropriate management of hypertensive patients relies on the accurate identification of clinically relevant features. However, traditional statistical methods may ignore important information in datasets or overlook possible interactions among features. Machine learning may improve the prediction accuracy and interpretability of regression models by identifying the most relevant features in hypertensive patients. We sought the most relevant features for prediction of cardiovascular (CV) events in a hypertensive population. We used the penalized regression models least absolute shrinkage and selection operator (LASSO) and elastic net (EN) to obtain the most parsimonious and accurate models. The clinical parameters and laboratory biomarkers were collected from the clinical records of 1,471 patients receiving care at Mostoles University Hospital. The outcome was the development of major adverse CV events. Cox proportional hazards regression was performed alone and with penalized regression analyses (LASSO and EN), producing three models. The modeling was performed using 10-fold cross-validation to fit the penalized models. The three predictive models were compared and statistically analyzed to assess their classification accuracy, sensitivity, specificity, discriminative power, and calibration accuracy. The standard Cox model identified five relevant features, while LASSO and EN identified only three (age, LDL cholesterol, and kidney function). The accuracies of the models (prediction vs. observation) were 0.767 (Cox model), 0.754 (LASSO), and 0.764 (EN), and the areas under the curve were 0.694, 0.670, and 0.673, respectively. However, pairwise comparison of performance yielded no statistically significant differences. All three calibration curves showed close agreement between the predicted and observed probabilities of the development of a CV event. Although the performance was similar for all three models, both penalized regression analyses produced models with good fit and fewer features than the Cox regression predictive model but with the same accuracy. This case study of predictive models using penalized regression analyses shows that penalized regularization techniques can provide predictive models for CV risk assessment that are parsimonious, highly interpretable, and generalizable and that have good fit. For clinicians, a parsimonious model can be useful where available data are limited, as such a model can offer a simple but efficient way to model the impact of the different features on the prediction of CV events. Management of these features may lower the risk for a CV event. Graphical Abstract In a clinical setting, with numerous biological and laboratory features and incomplete datasets, traditional statistical methods may ignore important information and overlook possible interactions among features. Our aim was to identify the most relevant features to predict cardiovascular events in a hypertensive population, using three different regression approaches for feature selection, to improve the prediction accuracy and interpretability of regression models by identifying the relevant features in these patients.