Explore the vast range of titles available.

Polygenic Risk Scores (PRS) for AD offer unique possibilities for reliable identification of individuals at high and low risk of AD. However, there is little agreement in the field as to what approach should be used for genetic risk score calculations, how to model the effect of APOE , what the optimal p- value threshold (pT) for SNP selection is and how to compare scores between studies and methods. We show that the best prediction accuracy is achieved with a model with two predictors ( APOE and PRS excluding APOE region) with pT<0.1 for SNP selection. Prediction accuracy in a sample across different PRS approaches is similar, but individuals’ scores and their associated ranking differ. We show that standardising PRS against the population mean, as opposed to the sample mean, makes the individuals’ scores comparable between studies. Our work highlights the best strategies for polygenic profiling when assessing individuals for AD risk. While polygenic risk scores have been shown to be correlated with disease risk, there is little agreement on how the score should be calculated. Here the authors investigate risk scores for Alzheimer’s disease, finding that the most effective approach includes an APOE score and a polygenic score excluding APOE.

Thyroid cancer incidence increased with the greatest change in adults aged ≥65 years. To determine the relationship between area-level use of imaging and thyroid cancer incidence over time. Longitudinal imaging patterns in Medicare patients aged ≥65 years residing in Surveillance, Epidemiology, and End Results (SEER) regions were assessed in relationship to differentiated thyroid cancer diagnosis in patients aged ≥65 years included in SEER-Medicare. Linear mixed-effects modeling was used to determine factors associated with thyroid cancer incidence over time. Multivariable logistic regression was used to determine patient characteristics associated with receipt of thyroid ultrasound as initial imaging. Thyroid cancer incidence. Between 2002 and 2013, thyroid ultrasound use as initial imaging increased (P < 0.001). Controlling for time and demographics, use of thyroid ultrasound was associated with thyroid cancer incidence (P < 0.001). Findings persisted when cohort was restricted to papillary thyroid cancer (P < 0.001), localized papillary thyroid cancer (P = 0.004), and localized papillary thyroid cancer with tumor size ≤1 cm (P = 0.01). Based on our model, from 2003 to 2013, at least 6594 patients aged ≥65 years were diagnosed with thyroid cancer in the United States due to increased use of thyroid ultrasound. Thyroid ultrasound as initial imaging was associated with female sex and comorbidities. Greater thyroid ultrasound use led to increased diagnosis of low-risk thyroid cancer, emphasizing the need to reduce harms through reduction in inappropriate ultrasound use and adoption of nodule risk stratification tools.

At the beginning of June 2020, there were nearly 7 million reported cases of coronavirus disease 2019 (COVID-19) worldwide and over 400,000 deaths in people with COVID-19. The objective of this study was to determine associations between comorbidities listed in the Charlson comorbidity index and mortality among patients in the United States with COVID-19. A retrospective cohort study of adults with COVID-19 from 24 healthcare organizations in the US was conducted. The study included adults aged 18-90 years with COVID-19 coded in their electronic medical records between January 20, 2020, and May 26, 2020. Results were also stratified by age groups (<50 years, 50-69 years, or 70-90 years). A total of 31,461 patients were included. Median age was 50 years (interquartile range [IQR], 35-63) and 54.5% (n = 17,155) were female. The most common comorbidities listed in the Charlson comorbidity index were chronic pulmonary disease (17.5%, n = 5,513) and diabetes mellitus (15.0%, n = 4,710). Multivariate logistic regression analyses showed older age (odds ratio [OR] per year 1.06; 95% confidence interval [CI] 1.06-1.07; p < 0.001), male sex (OR 1.75; 95% CI 1.55-1.98; p < 0.001), being black or African American compared to white (OR 1.50; 95% CI 1.31-1.71; p < 0.001), myocardial infarction (OR 1.97; 95% CI 1.64-2.35; p < 0.001), congestive heart failure (OR 1.42; 95% CI 1.21-1.67; p < 0.001), dementia (OR 1.29; 95% CI 1.07-1.56; p = 0.008), chronic pulmonary disease (OR 1.24; 95% CI 1.08-1.43; p = 0.003), mild liver disease (OR 1.26; 95% CI 1.00-1.59; p = 0.046), moderate/severe liver disease (OR 2.62; 95% CI 1.53-4.47; p < 0.001), renal disease (OR 2.13; 95% CI 1.84-2.46; p < 0.001), and metastatic solid tumor (OR 1.70; 95% CI 1.19-2.43; p = 0.004) were associated with higher odds of mortality with COVID-19. Older age, male sex, and being black or African American (compared to being white) remained significantly associated with higher odds of death in age-stratified analyses. There were differences in which comorbidities were significantly associated with mortality between age groups. Limitations include that the data were collected from the healthcare organization electronic medical record databases and some comorbidities may be underreported and ethnicity was unknown for 24% of participants. Deaths during an inpatient or outpatient visit at the participating healthcare organizations were recorded; however, deaths occurring outside of the hospital setting are not well captured. Identifying patient characteristics and conditions associated with mortality with COVID-19 is important for hypothesis generating for clinical trials and to develop targeted intervention strategies.

To evaluate the performance of machine learning (ML) algorithms and to compare them with logistic regression for the prediction of risk of cardiovascular diseases (CVDs), chronic kidney disease (CKD), diabetes (DM), and hypertension (HTN) and in a prospective cohort study using simple clinical predictors. We conducted analyses in a population-based cohort study in Asian adults (n = 6,762). Five different ML models were considered—single-hidden-layer neural network, support vector machine, random forest, gradient boosting machine, and k-nearest neighbor—and were compared with standard logistic regression. The incidences at 6 years of CVD, CKD, DM, and HTN cases were 4.0%, 7.0%, 9.2%, and 34.6%, respectively. Logistic regression reached the highest area under the receiver operating characteristic curve for CKD (0.905 [0.88, 0.93]) and DM (0.768 [0.73, 0.81]) predictions. For CVD and HTN, the best models were neural network (0.753 [0.70, 0.81]) and support vector machine (0.780 [0.747, 0.812]), respectively. However, the differences with logistic regression were small (less than 1%) and nonsignificant. Logistic regression, gradient boosting machine, and neural network were systematically ranked among the best models. Logistic regression yields as good performance as ML models to predict the risk of major chronic diseases with low incidence and simple clinical predictors. •Low-dimensional settings include low number of events and predictors.•In such settings, logistic regression yields as good performance as ML models.•ML techniques may not be warranted in such cases.

Here we identify hundreds of new drug-disease associations for over 900 FDA-approved drugs by quantifying the network proximity of disease genes and drug targets in the human (protein–protein) interactome. We select four network-predicted associations to test their causal relationship using large healthcare databases with over 220 million patients and state-of-the-art pharmacoepidemiologic analyses. Using propensity score matching, two of four network-based predictions are validated in patient-level data: carbamazepine is associated with an increased risk of coronary artery disease (CAD) [hazard ratio (HR) 1.56, 95% confidence interval (CI) 1.12–2.18], and hydroxychloroquine is associated with a decreased risk of CAD (HR 0.76, 95% CI 0.59–0.97). In vitro experiments show that hydroxychloroquine attenuates pro-inflammatory cytokine-mediated activation in human aortic endothelial cells, supporting mechanistically its potential beneficial effect in CAD. In summary, we demonstrate that a unique integration of protein-protein interaction network proximity and large-scale patient-level longitudinal data complemented by mechanistic in vitro studies can facilitate drug repurposing. Repurposing approved drugs could accelerate treatment options for various diseases. Here, the authors use network proximity of disease gene products and drug targets in the human protein interactome to identify drug-disease associations for cardiovascular disease, and validate these using longitudinal healthcare data.

ObjectiveTo examine the relationship between Mediterranean diet and risk of later-onset Crohn’s disease (CD) or ulcerative colitis (UC).DesignWe conducted a prospective cohort study of 83 147 participants (age range: 45–79 years) enrolled in the Cohort of Swedish Men and Swedish Mammography Cohort. A validated food frequency questionnaire was used to calculate an adherence score to a modified Mediterranean diet (mMED) at baseline in 1997. Incident diagnoses of CD and UC were ascertained from the Swedish Patient Register. We used Cox proportional hazards modelling to calculate HRs and 95% CI.ResultsThrough December of 2017, we confirmed 164 incident cases of CD and 395 incident cases of UC with an average follow-up of 17 years. Higher mMED score was associated with a lower risk of CD (Ptrend=0.03) but not UC (Ptrend=0.61). Compared with participants in the lowest category of mMED score (0–2), there was a statistically significant lower risk of CD (HR=0.42, 95% CI 0.22 to 0.80) but not UC (HR=1.08, 95% CI 0.74 to 1.58). These associations were not modified by age, sex, education level, body mass index or smoking (all Pinteraction >0.30). The prevalence of poor adherence to a Mediterranean diet (mMED score=0–2) was 27% in our cohorts, conferring a population attributable risk of 12% for later-onset CD.ConclusionIn two prospective studies, greater adherence to a Mediterranean diet was associated with a significantly lower risk of later-onset CD.

Effects of treatment or other exposure on outcome events are commonly measured by ratios of risks, rates, or odds. Adjusted versions of these measures are usually estimated by maximum likelihood regression (eg, logistic, Poisson, or Cox modelling). But resulting estimates of effect measures can have serious bias when the data lack adequate case numbers for some combination of exposure and outcome levels. This bias can occur even in quite large datasets and is hence often termed sparse data bias. The bias can arise or be worsened by regression adjustment for potentially confounding variables; in the extreme, the resulting estimates could be impossibly huge or even infinite values that are meaningless artefacts of data sparsity. Such estimate inflation might be obvious in light of background information, but is rarely noted let alone accounted for in research reports. We outline simple methods for detecting and dealing with the problem focusing especially on penalised estimation, which can be easily performed with common software packages.

The more infectious SARS-CoV-2 lineage B.1.1.7 rapidly spread in Europe after December, 2020, and a concern that B.1.1.7 could cause more severe disease has been raised. Taking advantage of Denmark's high RT-PCR testing and whole genome sequencing capacities, we used national health register data to assess the risk of COVID-19 hospitalisation in individuals infected with B.1.1.7 compared with those with other SARS-CoV-2 lineages. We did an observational cohort study of all SARS-CoV-2-positive cases confirmed by RT-PCR in Denmark, sampled between Jan 1 and March 24, 2021, with 14 days of follow-up for COVID-19 hospitalisation. Cases were identified in the national COVID-19 surveillance system database, which includes data from the Danish Microbiology Database (RT-PCR test results), the Danish COVID-19 Genome Consortium, the National Patient Registry, the Civil Registration System, as well as other nationwide registers. Among all cases, COVID-19 hospitalisation was defined as first admission lasting longer than 12 h within 14 days of a sample with a positive RT-PCR result. The study population and main analysis were restricted to the proportion of cases with viral genome data. We calculated the risk ratio (RR) of admission according to infection with B.1.1.7 versus other co-existing lineages with a Poisson regression model with robust SEs, adjusted a priori for sex, age, calendar time, region, and comorbidities. The contribution of each covariate to confounding of the crude RR was evaluated afterwards by a stepwise forward inclusion. Between Jan 1 and March 24, 2021, 50 958 individuals with a positive SARS-CoV-2 test and at least 14 days of follow-up for hospitalisation were identified; 30 572 (60·0%) had genome data, of whom 10 544 (34·5%) were infected with B.1.1.7. 1944 (6·4%) individuals had a COVID-19 hospitalisation and of these, 571 (29·4%) had a B.1.1.7 infection and 1373 (70·6%) had an infection with other SARS-CoV-2 lineages. Although the overall number of hospitalisations decreased during the study period, the proportion of individuals infected with B.1.1.7 increased from 3·5% to 92·1% per week. B.1.1.7 was associated with a crude RR of hospital admission of 0·79 (95% CI 0·72–0·87; p<0·0001) and an adjusted RR of 1·42 (95% CI 1·25–1·60; p<0·0001). The adjusted RR was increased in all strata of age and calendar period—the two covariates with the largest contribution to confounding of the crude RR. Infection with SARS-CoV-2 lineage B.1.1.7 was associated with an increased risk of hospitalisation compared with that of other lineages in an analysis adjusted for covariates. The overall effect on hospitalisations in Denmark was lessened due to a strict lockdown, but our findings could support hospital preparedness and modelling of the projected impact of the epidemic in countries with uncontrolled spread of B.1.1.7. None.

Cancer patients have a higher risk of cardiovascular disease (CVD) mortality than the general population. Low dose computed tomography (LDCT) for lung cancer screening offers an opportunity for simultaneous CVD risk estimation in at-risk patients. Our deep learning CVD risk prediction model, trained with 30,286 LDCTs from the National Lung Cancer Screening Trial, achieves an area under the curve (AUC) of 0.871 on a separate test set of 2,085 subjects and identifies patients with high CVD mortality risks (AUC of 0.768). We validate our model against ECG-gated cardiac CT based markers, including coronary artery calcification (CAC) score, CAD-RADS score, and MESA 10-year risk score from an independent dataset of 335 subjects. Our work shows that, in high-risk patients, deep learning can convert LDCT for lung cancer screening into a dual-screening quantitative tool for CVD risk estimation. Low dose computed tomography (LDCT) for lung cancer screening offers an opportunity for simultaneous CVD risk estimation in at-risk patients. Here, the authors develop a deep learning model to perform this task, showing human-level performance.

Policies that stimulate active transportation (walking and bicycling) have been related to heath benefits. This study aims to assess the potential health risks and benefits of promoting active transportation for commuting populations (age groups 16-64) in six European cities. We conducted a health impact assessment using two scenarios: increased cycling and increased walking. The primary outcome measure was all-cause mortality related to changes in physical activity level, exposure to fine particulate matter air pollution with a diameter <2.5 μm, as well as traffic fatalities in the cities of Barcelona, Basel, Copenhagen, Paris, Prague, and Warsaw. All scenarios produced health benefits in the six cities. An increase in bicycle trips to 35% of all trips (as in Copenhagen) produced the highest benefits among the different scenarios analysed in Warsaw 113 (76-163) annual deaths avoided, Prague 61 (29-104), Barcelona 37 (24-56), Paris 37 (18-64) and Basel 5 (3-9). An increase in walking trips to 50% of all trips (as in Paris) resulted in 19 (3-42) deaths avoided annually in Warsaw, 11(3-21) in Prague, 6 (4-9) in Basel, 3 (2-6) in Copenhagen and 3 (2-4) in Barcelona. The scenarios would also reduce carbon dioxide emissions in the six cities by 1,139 to 26,423 (metric tonnes per year). Policies to promote active transportation may produce health benefits, but these depend of the existing characteristics of the cities. Increased collaboration between health practitioners, transport specialists and urban planners will help to introduce the health perspective in transport policies and promote active transportation.