Extreme Weights Research Articles

The impact of body mass on cardiovascular outcomes in patients with acute coronary syndrome

Abstract Background Patients with acute coronary syndrome (ACS) and extreme body weights exhibit varying risks of bleeding and thrombotic outcomes. Furthermore, extreme body weights can influence the pharmacodynamic response to potent P2Y12 inhibitors (ticagrelor/prasugrel) and clopidogrel. Purpose This study aims to assess the impact of extreme body weights on cardiovascular outcomes, as well as the comparative efficacy and safety of potent P2Y12 inhibitors versus clopidogrel. Methods Data was extracted from the FORCE-ACS registry, a prospective, multicentre, registry, enrolling patients with ACS. Patients were stratified (in kg/m2) into the categories: underweight (body mass index [BMI] ≤20.0), normal weight (20.1–24.9), overweight (25.0–29.9), class 1 (30.0–34.9) and class 2 obesity (≥35.0). The primary thrombotic endpoint was major adverse cardiac events (MACE), a composite of cardiovascular death, myocardial infarction (MI), or stroke. The primary bleeding endpoint was defined as Bleeding Academic Research Consortium (BARC) 2, 3 or 5 bleeding. Cox proportional hazard models and Kaplan-Meier survival curves were used to estimate the risk of clinical events across the BMI categories. Restricted cubic splines were made to assess the continuous association of BMI with the risk of the primary endpoints, stratified for P2Y12-inhibitor therapy (potent vs. clopidogrel) and gender. Results The total included population consisted of 5,864 patients with ACS (mean age 65.2 years, 72.4% men), and were stratified into underweight (N=112 [2%]), normal weight (N=1,627 [28%]), overweight (N=2,697 [46%]), class 1 obesity (N=1,081 [18%]) and class 2 obesity (N=347 [6%]). Compared with normal weight (10.8%), the rate of primary bleeding outcome was higher in underweight (15.2%) patients, but lower in the overweight (8.9%), class 1 obesity (7.9%) and class 2 obesity (6.3%) categories. MACE occurred in 9.8% for the underweight group, 7.9% for normal weight, 7.0% for overweight, 8.0% for class, and 6.1% for class 2 obesity. After adjusting for confounding factors such as age, gender, cardiovascular risk factors, access site, and medical therapy, each unit increase in BMI was associated with a decreased bleeding risk (hazard ratio [HR] 0.97, 95% CI 0.95-0.99, p = 0.02), with no significant association found for MACE. Similar findings are demonstrated in the Kaplan Meier analysis (Figure 1A-1B). Subgroup analysis regarding bleeding risk and BMI did not show a significant interaction for potent P2Y12-inhibitor therapy versus clopidogrel and gender (P-value for interaction: 0.04 and 0.32, respectively)(Figure 2A-2D). Conclusions Our findings indicate a stronger correlation between BMI and bleeding risk compared to thrombotic risk, showing an increased risk for bleeding in patients with a low BMI. These results underscore the importance of careful monitoring and tailored interventions to mitigate bleeding risk in underweight patients.

Impact of body weight and body mass index on clinical outcomes of edoxaban therapy in atrial fibrillation patients included in the ETNA-AF-Global registry

Abstract Background In atrial fibrillation (AF) patients (pts) receiving non-vitamin K antagonist oral anticoagulant therapy, extreme body weights, both obesity and low body weight, may impact clinical outcomes. Purpose To determine the impact of body weight and body mass index (BMI) on clinical outcomes in AF pts receiving edoxaban from the Global ETNA program (Europe [NCT02944019], Japan [UMIN000017011], Korea/Taiwan [NCT02951039], Hong Kong [NCT03247582] and Thailand [NCT03247569]). Methods ETNA-AF-Global, a multinational, prospective, observational study collected demographic and clinical outcomes data for AF pts receiving edoxaban. In this subanalysis, pts were categorised into the following body weight groups: ≤60kg, >60 to 80kg (reference body weight), >80 to ≤100kg and >100kg; and BMI groups: <18.5kg/m², ≥18.5 to <25.0kg/m² (reference BMI), ≥25.0 to <30.0kg/m², ≥30.0 to <40.0kg/m² and ≥40.0kg/m². Clinical outcomes at 2-year follow-up were compared across body weight and BMI groups. Results 26,805 pts were included in this analysis. Mean ± standard deviation body weight was 72.2 ± 18.1kg and mean BMI was 26.4 ± 5.0kg/m². 43.7% of pts were categorised into the >60 to 80kg reference body weight group. [other groups: ≤60kg, 28.1%; >80 to ≤100kg, 21.7%; >100kg, 6.5%] and 40.0% of pts into the ≥18.5 to <25.0kg/m² reference BMI group [other groups: <18.5kg/m², 3.0%; ≥25.0 to <30.0kg/m², 37.3%; ≥30.0 to <40.0kg/m², 18.0%; ≥40.0kg/m², 1.7%]. Clinical outcomes are shown in Fig. 1 as unadjusted hazard ratios vs. reference group according to weight and in Fig. 2 according to BMI. Among body weight groups, pts weighing ≤60kg had the highest annualised rates of any stroke or systemic embolic event (SEE; 1.24%), ischemic stroke (0.97%), all-cause death (3.96%), major bleeding events (1.49%), and major bleeding or clinical-relevant non-major (CRNM) bleeding events (3.45%). Pts in the reference body weight group had the lowest annualised rates of all-cause death (2.81%). Annualised rates of major bleeding, major bleeding or CRNM bleeding events and any stroke or SEE were lowest in pts weighing >100kg. Among BMI groups, annualised rates of all effectiveness and safety outcomes were highest in pts with a BMI <18.5kg/m². Pts in the ≥25.0 to <30.0kg/m² BMI group had the lowest annualised rates of all-cause death (2.85%) and pts in the ≥18.5 to <25.0kg/m² BMI group had the lowest annualised rates of CV death (0.73%). Conclusions In the Global ETNA-AF program, clinical outcomes varied between different body weight and BMI subgroups. These findings highlight the need to consider the clinical implications of low body weight and/or low BMI, as AF pts who belong to these categories may be at higher risk of adverse outcomes during long-term anticoagulant treatment.Figure 1.Figure 2.

Portfolio resampling in the Brazilian stock market

Markowitz optimization plays a crucial role in modern portfolio theory. However, it is well known that Markowitz optimization is highly affected by estimation errors in the mean vector and covariance matrix, resulting in extreme and/or unrealistic portfolio weights, lack of diversification, and poor out-of-sample performance. In response to these challenges, Michaud and Michaud (1998) proposed a heuristic portfolio resampling approach aimed to deliver more diversified and better out-of-sample portfolio performance. This approach has received both criticism and praise among academics and practitioners, with the main critique being the unclear economic benefits of applying this method to empirical data. To contribute to this ongoing debate, our study assess the performance of the Michaud and Michaud (1998) portfolio resampling approach using data from the Brazilian stock market. Additionally, we evaluate whether a resampling approach based on factor structure can yield superior out-of-sample performance. Our findings indicate no evidence of superiority over Markowitz optimization, reinforcing criticisms of the portfolio resampling approach.

Propensity scores used as overlap weights provide exact covariate balance.

Overlap weighting, using weights defined as the probability of receiving the opposite treatment, is a relatively new, alternative propensity score-based weighting technique used to adjust for confounding when estimating causal treatment effects. It has preferable properties compared to inverse probability of treatment weighting, such as exact covariate balance, safeguards against extreme weights and emphasis on medical equipoise, where treatment decisions are most uncertain. In this article, we introduce the overlap weighting methodology, compare it to inverse probability of treatment weighting and provide some strategies for assessing weighting impact, through an applied example of hospital mortality. When the propensity score distributions have large separation, inverse probability of treatment weighting has been shown to produce biased and less efficient estimates of the treatment effect, making overlap weighting a preferred method in such cases.

Assessment of centanafadine in adults with ADHD: a matching adjusted indirect comparison versus methylphenidate hydrochloride extended release (Concerta)

Objective To compare safety and efficacy of centanafadine versus methylphenidate hydrochloride extended release (ER; Concerta) in adults with ADHD. Methods Without head-to-head trials, anchored matching-adjusted indirect comparisons (MAIC) of adverse event rates reported across trials and mean change from baseline in Adult ADHD Investigator Symptom Rating Scale (AISRS) score between centanafadine and methylphenidate hydrochloride ER were conducted. Pooled patient-level data from two centanafadine trials (NCT03605680/NCT03605836) and aggregate data from one published methylphenidate hydrochloride ER trial (NCT00937040) were used. Characteristics of individual patients from the centanafadine trials were matched to aggregate baseline characteristics from the methylphenidate hydrochloride ER trial using propensity score weighting. A sensitivity analysis assessed the robustness of the results to the capping of extreme weights (i.e. >99th percentile). Results Compared with methylphenidate hydrochloride ER, centanafadine was associated with significantly lower risk of dry mouth (risk difference [RD] in percentage points: −11.95), initial insomnia (−11.10), decreased appetite (−8.05), anxiety (−5.39), palpitations (−5.25), and feeling jittery (−4.73) though a significantly smaller reduction in AISRS score (4.16-point). In the sensitivity analysis, the safety results were consistent with the primary analysis but there was no significant difference in efficacy between centanafadine and methylphenidate hydrochloride ER. Conclusion In this MAIC, centanafadine had better safety and possibly lower efficacy than methylphenidate hydrochloride ER. While safety results were robust across analyses, there was no efficacy difference between centanafadine and methylphenidate hydrochloride ER in the sensitivity analysis. Considering its favorable safety profile, centanafadine may be preferred among patients for whom treatment-related adverse events are a concern.

Optimal Weight-Based Dosing of Vancomycin to Achieve an Area Under the Curve of 400 to 600 Stratified by Body Mass Index

Background: In 2020, the American Society of Health System Pharmacist (ASHP) and Infectious Diseases Society of America (IDSA) published a consensus guideline for vancomycin management, recommending area under the curve (AUC) as the preferred monitoring strategy. These guidelines recommend doses of 15-20 mg/kg every 8 to 12 hours for most patients with normal renal function. However, in extreme body weights, standard dosing may deviate to provide a therapeutic AUC. The primary objective of this pharmacokinetic study is to evaluate the optimal vancomycin weight-based dosing strategy that achieves a therapeutic AUC of 400-600 stratified by body mass index (BMI). The secondary objective is to evaluate the incidence of acute kidney injury (AKI) based on BMI. Methods: Patients were identified from two sites within the Department of Veterans Affairs who received vancomycin for at least 48 hours and had at least one steady-state level from January 2015 through July 2022. Regimens with a frequency of ≤8 hours or patients with baseline creatine clearance of < 5 0 ml/min were excluded. Patients were categorized based on the Center for Disease Control BMI groups: healthy weight, overweight, or obese. The online vancomycin calculator, VancoPK©, was utilized to calculate AUC. Renal function at baseline and during vancomycin therapy was collected. Descriptive statistics were used for data analysis. Continuous outcomes were summarized using mean and standard deviation. The primary and secondary endpoints were analyzed using the analysis of variance and Fisher’s exact tests, respectively. Statistical significance was established at a p-value of < 0 .05. Results: A total of 347 unique vancomycin regimens were included: 120 in the healthy weight group, 101 in the overweight group, and 126 in the obese group. The average total daily doses that achieved a therapeutic AUC were 1971mg (15.6mg/kg/dose), 2298mg (15mg/kg/dose), and 2524mg (12.6mg/kg/dose) for the healthy weight, overweight, and obese groups, respectively. There was a statistically significant difference among these groups. AKI occurred in 10/254 (3.9%) unique patients: 2/89 (2.2%) in the healthy weight group, 3/71 (4.2%) in the overweight group, and 5/94 (5.3%) in the obese group. This did not reach statistical significance. Conclusions: Vancomycin dosing regimens largely followed guideline recommendations. However, the average vancomycin mg/kg/dose that achieved a therapeutic AUC decreased as BMI increased, which was a statistically significant trend. While further research is needed to draw clinically impactful conclusions, these findings suggest that a lower mg/kg vancomycin dose in obesity may be needed to achieve therapeutic targets.

Balancing versus modelling in weighted analysis of non-randomised studies with survival outcomes: A simulation study.

Weighting methods are widely used for causal effect estimation in non-randomised studies. In general, these methods use the propensity score (PS), the probability of receiving the treatment given the covariates, to arrive at the respective weights. All of these "modelling" methods actually optimize prediction of the respective outcome, which is, in the PS model, treatment assignment. However, this does not match with the actual aim of weighting, which is eliminating the association between covariates and treatment assignment. In the "balancing" approach, covariates are thus balanced directly by solving systems of numerical equations, explicitly without fitting a PS model. To compare modelling, balancing and hybrid approaches to weighting we performed a large simulation study for a binary treatment and a survival outcome. For maximal practical relevance all simulation parameters were selected after a systematic review of medical studies that used PS methods for analysis. We also introduce a new hybrid method that uses the idea of the covariate balancing propensity score and matching weights, thus avoiding extreme weights. In addition, we present a corrected robust variance estimator for some of the methods. Overall, our simulations results indicate that balancing approach methods work worse than expected. However, among the considered balancing methods, entropy balancing consistently outperforms the variance balancing approach. All methods estimating the average treatment effect in the overlap population perform well with very little bias and small standard errors even in settings with misspecified propensity score models. Finally, the coverage using the standard robust variance estimator was too high for all methods, with the proposed corrected robust variance estimator improving coverage in a variety of settings.

A matching-based machine learning approach to estimating optimal dynamic treatment regimes with time-to-event outcomes.

Observational data (e.g. electronic health records) has become increasingly important in evidence-based research on dynamic treatment regimes, which tailor treatments over time to patients based on their characteristics and evolving clinical history. It is of great interest for clinicians and statisticians to identify an optimal dynamic treatment regime that can produce the best expected clinical outcome for each individual and thus maximize the treatment benefit over the population. Observational data impose various challenges for using statistical tools to estimate optimal dynamic treatment regimes. Notably, the task becomes more sophisticated when the clinical outcome of primary interest is time-to-event. Here, we propose a matching-based machine learning method to identify the optimal dynamic treatment regime with time-to-event outcomes subject to right-censoring using electronic health record data. In contrast to the established inverse probability weighting-based dynamic treatment regime methods, our proposed approach provides better protection against model misspecification and extreme weights in the context of treatment sequences, effectively addressing a prevalent challenge in the longitudinal analysis of electronic health record data. In simulations, the proposed method demonstrates robust performance across a range of scenarios. In addition, we illustrate the method with an application to estimate optimal dynamic treatment regimes for patients with advanced non-small cell lung cancer using a real-world, nationwide electronic health record database from Flatiron Health.

Mapping hazardous locations on a road network due to extreme gross vehicle weights

This study investigates the application of hazard maps in identifying critical locations associated with extreme vehicle load events. Examining extreme vehicle loads can offer valuable insights into the reliability of road infrastructure. Weigh-In-Motion (WIM) systems are commonly employed for this purpose; however, their limited availability often necessitates the use of less-sophisticated traffic counters (LSTCs). Unfortunately, LSTCs are unable to measure vehicular axle loads, posing a significant drawback. To overcome this limitation, we propose a methodology that utilizes data from LSTCs to estimate axle loads and map extreme gross vehicle weights. To demonstrate the feasibility of this approach, we present a case study concentrating on the major highway corridors in Mexico. While WIM stations are absent, data from a network of 1,777 counting stations and origin–destination surveys are available. By quantifying Gaussian copula-based Bayesian networks using the existing information, we generate synthetic site-specific axle loads. Subsequently, we calculate gross vehicle weights for selected return periods. These study findings facilitate the identification of hazardous locations. Additionally, we provide an interactive web map and a graphical user interface to generate synthetic axle loads. These tools, along with the proposed methodology, can serve as the foundation for maintenance strategies for existing roads and bridges.

Safety and efficacy of oral anticoagulants in extreme weights

BackgroundThe 2021 International Society on Thrombosis and Haemostasis' (ISTH) recommends standard doses of apixaban and rivaroxaban regardless of high body mass index (BMI) and weight, but had not compare DOACs head-to-head in obesity or address underweight patients. MethodsOur aim is to evaluate the safety and efficacy of DOACs in underweight and obese patients compared to warfarin. The primary endpoints include incidence of thromboembolic and bleeding events. Descriptive statistics was used for continuous variables. The Kruskal-Wallis test was used to compare the four-groups for continuous measures and the chi-square test or Fisher's exact test was used to analyze categorical data. The chi-square test or Fisher's exact test, was used for categorical variables, and the Mann-Whitney test (the non-parametric counterpart to the two-sample t-test) for continuous data. ResultsOf 2940 patients receiving anticoagulation for venous thromboembolism (VTE) treatment or atrial fibrillation (AF), 492 met eligibility criteria. Within each group, 248 patients received warfarin, 101 received apixaban, 100 received rivaroxaban and 43 received dabigatran. Patients were characterized in 4 body mass index (BMI) categories, in which 80 were underweight and 412 were obese. ConclusionsWhen each DOAC was compared to warfarin in rates of VTE, apixaban showed statistically significant lower rate of VTE (p = 0.0149). However, no statistical significance was identified in the rate of VTE between DOACs combined vs. warfarin (p = 0.1529). When each DOAC was compared to warfarin, apixaban showed the lowest rate of overall bleeding (p = 0.0194). However, no statistical difference in the rate of bleeding was observed between DOACs combined vs. warfarin (p = 0.3284). Patients with extreme body weights requiring anticoagulation for VTE and AF may safety benefit from DOAC therapy. This evaluation showed apixaban with the lowest rate of VTE and bleeding compared to warfarin, rivaroxaban, and dabigatran. These results provide experience for the clinician to use DOACs, particularly apixaban, in underweight and obese populations.

Reference Values of Cardiopulmonary Exercise Test Parameters in the Contemporary Paediatric Population

BackgroundThe evaluation of health status by cardiopulmonary exercise test (CPET) has shown increasing interest in the paediatric population. Our group recently established reference Z-score values for paediatric cycle ergometer VO2max, applicable to normal and extreme weights, from a cohort of 1141 healthy children. There are currently no validated reference values for the other CPET parameters in the paediatric population. This study aimed to establish, from the same cohort, reference Z-score values for the main paediatric cycle ergometer CPET parameters, apart from VO2max.ResultsIn this cross-sectional study, 909 healthy children aged 5–18 years old underwent a CPET. Linear, quadratic, and polynomial mathematical regression equations were applied to identify the best CPET parameters Z-scores, according to anthropometric parameters (sex, age, height, weight, and BMI). This study provided Z-scores for maximal CPET parameters (heart rate, respiratory exchange ratio, workload, and oxygen pulse), submaximal CPET parameters (ventilatory anaerobic threshold, VE/VCO2 slope, and oxygen uptake efficiency slope), and maximum ventilatory CPET parameters (tidal volume, respiratory rate, breathing reserve, and ventilatory equivalent for CO2 and O2).ConclusionsThis study defined paediatric reference Z-score values for the main cycle ergometer CPET parameters, in addition to the existing reference values for VO2max, applicable to children of normal and extreme weights. Providing Z-scores for CPET parameters in the paediatric population should be useful in the follow-up of children with various chronic diseases. Thus, new paediatric research fields are opening up, such as prognostic studies and clinical trials using cardiopulmonary fitness outcomes.Trial registration NCT04876209—Registered 6 May 2021—Retrospectively registered, https://clinicaltrials.gov/ct2/show/NCT04876209.

Multi-View Adjacency-Constrained Hierarchical Clustering

This paper explores the problem of multi-view clustering, which aims to promote clustering performance with multi-view data. The majority of existing methods have problems with parameter adjustment and high computational complexity. Moreover, in the past, there have been few works based on hierarchical clustering to learn the granular information of multiple views. To overcome these limitations, we propose a simple but efficient framework: Multi-view adjacency-Constrained Hierarchical Clustering (MCHC). Specifically, MCHC mainly consists of three parts: including the Fusion Distance matrices with Extreme Weights (FDEW); adjacency-Constrained Nearest Neighbor Clustering (CNNC); and the internal evaluation Index based on Rawls' Max-Min criterion (MMI). FDEW aims to learn a fusion distance matrix set, which not only uses complementary information among multiple views, but exploits the information from each single view. CNNC is utilized to generate multiple partitions based on FDEW, and MMI is designed for choosing the best one from the multiple partitions. In addition, we propose a parameter-free version of MCHC (MCHC-PF). Without any parameter selection, MCHC-PF can give partitions at different granularity levels with a low time complexity. Comprehensive experiments tested on eight real-world datasets validate the superiority of the proposed methods compared with the 13 current state-of-the-art methods.

Using Balancing Weights to Target the Treatment Effect on the Treated when Overlap is Poor.

Inverse probability weights are commonly used in epidemiology to estimate causal effects in observational studies. Researchers often focus on either the average treatment effect or the average treatment effect on the treated with inverse probability weighting estimators. However, poor overlap in the baseline covariates between the treated and control groups can produce extreme weights that can result in biased treatment effect estimates. One alternative to inverse probability weights are overlap weights, which target the population with the most overlap on observed covariates. Although estimates based on overlap weights produce less bias in such contexts, the causal estimand can be difficult to interpret. An alternative to model-based inverse probability weights are balancing weights, which directly target imbalances during the estimation process, rather than model fit. Here, we explore whether balancing weights allow analysts to target the average treatment effect on the treated in cases where inverse probability weights lead to biased estimates due to poor overlap. We conduct three simulation studies and an empirical application. We find that balancing weights often allow the analyst to still target the average treatment effect on the treated even when overlap is poor. We show that although overlap weights remain a key tool, more familiar estimands can sometimes be targeted by using balancing weights.

DOACs use in extreme body-weighted patients: results from the prospective START-register.

Direct oral anticoagulants (DOACs) are widely used for the treatment of venous thromboembolism (VTE) and for stroke prevention in atrial fibrillation (AF). However, evidence in obese and underweight patients is limited. We assessed the safety and effectiveness of DOACs and vitamin K antagonists (VKAs) in patients ≥ 120kg or ≤ 50kg enrolled in an observational prospective cohort study, the START-Register. Adult patients started on anticoagulant therapy were followed up for a median of 1.5years (IQR 0.6-2.8). Primary efficacy outcome was the occurrence of VTE recurrence, stroke and systemic embolism. Primary safety outcome was major bleeding (MB). 10,080 AF and VTE patients were enrolled between March 2011 and June 2021, 295 patients weighted ≤ 50kg and 82 patients ≥ 120kg. Obese patients were significantly younger than underweight patients. Rates of thrombotic events were low and similar between DOACs and VKAs in underweight patients (1 event on DOACs therapy [0.9% 95% CI 0.11-5.39] and 2 on VKAs [1.1% 95% CI 0.01-47.68]) and in overweight patients (0 events on DOACs, 1 on VKAs [1.6%, 95% CI 0.11-5.79]. Two MB events occurred on DOACs (1.9%, 95% CI 0.38-6.00) and 3 on VKAs (1.6%, 95% CI 0.04-22.06) in the underweight group; 1MB on DOACs (5.3% 95% CI 0.33-16.68) and 2 on VKAs (3.3%, 95% CI 0.02-130.77) in the overweight group. DOACs seem to be effective and safe also for the treatment of patients with extreme body weights, both underweight and overweight. Further prospective studies are needed to support these findings.

Causal inference in survival analysis using longitudinal observational data: Sequential trials and marginal structural models.

Longitudinal observational data on patients can be used to investigate causal effects of time-varying treatments on time-to-event outcomes. Several methods have been developed for estimating such effects by controlling for the time-dependent confounding that typically occurs. The most commonly used is marginal structural models (MSM) estimated using inverse probability of treatment weights (IPTW) (MSM-IPTW). An alternative, the sequential trials approach, is increasingly popular, and involves creating a sequence of "trials" from new time origins and comparing treatment initiators and non-initiators. Individuals are censored when they deviate from their treatment assignment at the start of each "trial" (initiator or noninitiator), which is accounted for using inverse probability of censoring weights. The analysis uses data combined across trials. We show that the sequential trials approach can estimate the parameters of a particular MSM. The causal estimand that we focus on is the marginal risk difference between the sustained treatment strategies of "always treat" vs "never treat." We compare how the sequential trials approach and MSM-IPTW estimate this estimand, and discuss their assumptions and how data are used differently. The performance of the two approaches is compared in a simulation study. The sequential trials approach, which tends to involve less extreme weights than MSM-IPTW, results in greater efficiency for estimating the marginal risk difference at most follow-up times, but this can, in certain scenarios, be reversed at later time points and relies on modelling assumptions. We apply the methods to longitudinal observational data from the UK Cystic Fibrosis Registry to estimate the effect of dornase alfa on survival.

Perinatal and Maternal Outcomes According to the Accurate Term Antepartum Ultrasound Estimation of Extreme Fetal Weights.

(1) Background: The accuracy of ultrasound estimation of fetal weight (EFW) at term may be useful in addressing obstetric complications since birth weight (BW) is a parameter that represents an important prognostic factor for perinatal and maternal morbidity. (2) Methods: In a retrospective cohort study of 2156 women with a singleton pregnancy, it is verified whether or not perinatal and maternal morbidity differs between extreme BWs estimated at term by ultrasound within the seven days prior to birth with Accurate EFW (difference < 10% between EFW and BW) and those with Non-Accurate EFW (difference ≥ 10% between EFW and BW). (3) Results: Significantly worse perinatal outcomes (according to different variables such as higher rate of arterial pH at birth < 7.20, higher rate of 1-min Apgar < 7, higher rate of 5-min Apgar < 7, higher grade of neonatal resuscitation and need for admission to the neonatal care unit) were found for extreme BW estimated by antepartum ultrasounds with Non-Accurate EFW compared with those with Accurate EFW. This was the case when extreme BWs were compared according to percentile distribution by sex and gestational age following the national reference growth charts (small for gestational age and large for gestational age), and when they were compared according to weight range (low birth weight and high birth weight). (4) Conclusions: Clinicians should make a greater effort when performing EFW by ultrasound at term in cases of suspected extreme fetal weights, and need to take an increasingly prudent approach to its management.

Response to: Association of Intraoperative Red Blood Cell Transfusions With Venous Thromboembolism and Adverse Outcomes After Cardiac Surgery.

We are grateful for the interest and comments of Butt et al1 with regard to our recently published study.2 First, Butt et al1 noted that the overall exposure rate to intraoperative packed red blood cells (PRBCs) was 17.8% in our study compared to the reported rate of 12.8% to 60.3% in previous studies in similar patients. They then speculated that the relatively lower rate of intraoperative PRBC transfusion observed in our study was due to our decision to label missing information on intraoperative PRBC transfusion as “no transfusion.” In support of their argument, they cited a retrospective study of 8141 patients from 10 centers who had cardiac surgery between 2011 and 2014; these patients underwent either isolated coronary artery bypass grafting (CABG) surgery or aortic valve replacement, and Butt et al1 did not cite the risk-adjusted rate for intraoperative PRBC transfusion, which was found to be 4% to 59%.3 Thus, our reported rate of 17.8% is likely reflective of the more contemporary and average national rate of intraoperative PRBC transfusion for isolated CABG surgery. Furthermore, Butt et al1 highlighted that labeling missing data for intraoperative PRBC transfusions as “no transfusion” as justified in the methodology section of our manuscript, and omission of any postoperative transfusion before any hospital-acquired venous thromboembolic (HA-VTE) complications might have biased the association between intraoperative PRBC transfusion and HA-VTE complications. To address this, we performed a sensitivity analysis which indicated that after removing patients with missing data for intraoperative PRBC transfusion the association between intraoperative PRBC transfusion and HA-VTE remained (odds ratios [ORs] for 1, 2, 3, and ≥4 units of PRBC versus none: 1.16 [1.09 to 1.23], 1.19 [1.12 to 1.26], 1.59 [1.49 to 1.68], 1.49 [1.40 to 1.58]). Postoperative transfusion before any HA-VTE complications was not collected by the Society of Thoracic Surgeons Adult Cardiac Surgery Database and was not available for analysis. Second, Butt et al1 also indicated that a causality relationship between intraoperative PRBC transfusion and HA-VTE complications was not supported by the lower odds of HA-VTE complications in women. To address this comment, we reiterate the following: i) given the retrospective nature of our study it was not meant to definitively ascertain a causality, but rather reinforce a previously described association between intraoperative PRBC transfusion and HA-VTE complications in a large-scale national registry; ii) our propensity score-weighted, and covariate-adjusted analysis including adjusting for sex indicated that intraoperative PRBC transfusion was independently associated with HA-VTE complications after isolated CABG surgery; iii) only a small proportion of patients had missing data for height (0.03%) and weight (0.02%), and thus, the calculated body mass index (BMI) was observed in nearly all patients (missing, 0.03%). Therefore, our conditional imputation for 0.03% of the missing BMI is highly unlikely that it may have impacted the association of BMI and intraoperative PRBC transfusion in the development of the propensity score. We agree with Butt et al1 that it would have been potentially valuable to study the impact of geographic location on our selected patient- and procedure-related characteristics, but this information was not available through the participant user file research mechanism via which our study was approved. Third, we agree with the comment that delayed thromboprophylaxis during a prolonged hospital length of stay may also exacerbate HA-VTE complications independently or in combination with intraoperative and postoperative PRBC transfusions. As indicated by Butt and colleagues1 the Society of Thoracic Surgeons Adult Cardiac Surgery Database does not capture data on delayed thromboprophylaxis during a prolonged hospital length of stay, and thus we were not able to study its respective association with prolonged length of stay. Butt et al1 speculate that our approach of implementing a propensity score-weighted method using matching inverse probability of treatment weighting to account for potential confounding of the association between intraoperative PRBC transfusion and our outcomes may have been impacted by handling of extreme weights.4 To address the possibility of residual confounding, we considered the evidence for causality (E value),5 which is considered as an alternative approach to sensitivity analyses for potential unmeasured confounding, and were able to demonstrate that there was a low potential for unmeasured confounding significantly influencing the association between intraoperative PRBC transfusion and our primary outcome of HA-VTE. Finally, we agree with Butt et al1 that there is a need to better predict HA-VTE. We also strongly believe that there is a potentially unmet need for current and future “studies to explore interventions aimed at improving perioperative anemia and blood management while maintaining appropriate postoperative thromboprophylaxis, thereby significantly decreasing the need for PRBC transfusions in this setting and potentially decreasing the incidence of postoperative venous thromboembolism after cardiac surgery.”2

Comment on "Association of Intraoperative Red Blood Cell Transfusions With Venous Thromboembolism and Adverse Outcomes After Cardiac Surgery".

We read with great interest a recently published work by Raphael et al.1 Their primary hypothesis was to establish independent causality of intraoperative transfusion of red blood cells (RBCs) on venous thromboembolism (VTE) after coronary bypass grafting surgery (n = 751,893) using the database of the Society of Thoracic Surgeons. In their secondary analyses, they further explored the types of VTE, mortality, readmission rate, and length of hospital stay. Their analyses of these data were extensive; however, there are issues that warrant further discussion. First, the overall exposure to RBCs was reported as 17.8% in the current study, but it is important to note the authors handled missing data on intraoperative/postoperative blood products as “no transfusion.” Previous studies in similar patients have shown that intraoperative transfusion rates varied between 12.8% and 60.3% due to variations in institutional practice.2 Missing data labeled as “no transfusion” and omission of any postoperative transfusion before any VTE might have skewed the relationship between intraoperative RBCs and VTE. Second, 37% of women versus 12% of men received blood transfusion. The women were 3 times more likely to receive more than 3–4 units of transfusion than men; however, the adjusted odds ratio for VTE was 0.90 for women. Others also demonstrated that women were more likely to receive >4 U RBC transfusion.3 Women tend to have a lower body mass index (BMI) compared with men, and thus they are more prone to dilutional anemia, and subsequently RBC transfusion. A causality relationship between RBCs and VTE was not supported by the lower odds of VTE in women. As a related issue, we would like to discuss their use of conditional imputation for handling missing data. The extent or presumed pattern of missing data was not specified, but BMI was mentioned as an example of conditional imputation. BMI is known to have a wide variation, particularly by geographic location. Conditional imputation without consideration of regional clustering might have distorted the relationship of BMI (and other predictors) and transfusion in development of the propensity score.4,5 Third, the authors attempted to isolate the effect of intraoperative RBCs by addressing or adjusting all the other risk factors for VTE. From Table 3, the exposures to non-RBC components, recombinant activated factor VII, and factor VIII inhibitor bypassing agent were increased in those with increased RBC transfusion. Hemostatic components were treated as a confounder in the model, but there may be causal associations with both acute and subacute VTE and arterial thromboses.6,7 Patients with severe postoperative bleeding tend to have prolonged hospital stay. The authors showed that cohorts with ≥3 units of RBC had longer hospital stay (mean, 9.8–11.2 days) compared with nontransfused patients (mean, 6.6 days). Although not captured in the database, delayed thromboprophylaxis during a prolonged hospital stay may also exacerbate VTE risk. Lastly, the authors utilized inverse probability treatment weighting method using inverse probability of treatment weighting based on the matching as shown in their Table 2; however, the authors did not mention the handling of extreme weights, which may lead to undue influence of otherwise less common groups (ie, patients with 3 or 4+ units of RBCs) (Tables S1–S10, https://links.lww.com/SLA/E307). While inverse probability treatment weighting adequately accounted for measured confounders pertinent to transfusion (eg, cardiopulmonary bypass [CPB] time), it is speculated that there are residual confounders that impact VTE rates.8 For example, extensive tissue injury and CPB are well known to trigger activations of endothelial cells and leukocytes. Cross-talks between coagulation and immune systems are referred to as “thrombo-inflammation,” and it plays a pivotal role in the pathogenesis of VTE.9 Patient-level inflammation parameters are neither routinely measured nor captured in the database. The statistical adjustment for CPB time does not account for inter-individual variability in inflammation and VTE risk. Given the modest area under the curve statistics they reported for these models, it seems clear there remains an unmet need to improve precision of VTE prediction by including patient-level biomarkers with sufficient sensitivity and specificity. ACKNOWLEDGMENTS A.L.B. conceptualized the letter, and writing up of the first draft of the letter. E.C.W., K.A.T., and K.S. have helped write the letter, and approved the final article. All authors have written and reviewed the article, and approved the final article.

Robustifying Markowitz

Markowitz mean–variance portfolios with sample mean and covariance as input parameters feature numerous issues in practice. They perform poorly out of sample due to estimation error, they experience extreme weights together with high sensitivity to change in input parameters. The heavy-tail characteristics of financial time series are in fact the cause for these erratic fluctuations of weights that consequently create substantial transaction costs. In robustifying the weights we present a toolbox for stabilizing costs and weights for global minimum Markowitz portfolios. Utilizing a projected gradient descent (PGD) technique, we avoid the estimation and inversion of the covariance operator as a whole and concentrate on robust estimation of the gradient descent increment. Using modern tools of robust statistics we construct a computationally efficient estimator with almost Gaussian properties based on median-of-means uniformly over weights. This robustified Markowitz approach is confirmed by empirical studies on equity markets. We demonstrate that robustified portfolios reach the lowest turnover compared to shrinkage-based and constrained portfolios while preserving or slightly improving out-of-sample performance.

Addressing substantial covariate imbalance with propensity score stratification and balancing weights: connections and recommendations.

Propensity score (PS) weighting methods are commonly used to adjust for confounding in observational treatment comparisons. However, in the setting of substantial covariate imbalance, PS values may approach 0 and 1, yielding extreme weights and inflated variance of the estimated treatment effect. Adaptations of the standard inverse probability of treatment weights (IPTW) can reduce the influence of extremes, including trimming methods that exclude people with PS values near 0 or 1. Alternatively, overlap weighting (OW) optimizes criteria related to bias and variance, and performs well compared to other PS weighting and matching methods. However, it has not been compared to propensity score stratification (PSS). PSS has some of the same potential advantages; being insensitive extreme values. We sought to compare these methods in the setting of substantial covariate imbalance to generate practical recommendations. Analytical derivations were used to establish connections between methods, and simulation studies were conducted to assess bias and variance of alternative methods. We find that OW is generally superior, particularly as covariate imbalance increases. In addition, a common method for implementing PSS based on Mantel-Haenszel weights (PSS-MH) is equivalent to a coarsened version of OW and can perform nearly as well. Finally, trimming methods increase bias across methods (IPTW, PSS and PSS-MH) unless the PS model is re-fit to the trimmed sample and weights or strata are re-derived. After trimming with re-fitting, all methods perform similarly to OW. These results may guide the selection, implementation and reporting of PS methods for observational studies with substantial covariate imbalance.