Verification of Pharmacogenetics-Based Warfarin Dosing Algorithms in Han-Chinese Patients Undertaking Mechanic Heart Valve Replacement

Objective

To study the performance of pharmacogenetics-based warfarin dosing algorithms in the initial and the stable warfarin treatment phases in a cohort of Han-Chinese patients undertaking mechanic heart valve replacement.

Methods

We searched PubMed, Chinese National Knowledge Infrastructure and Wanfang databases for selecting pharmacogenetics-based warfarin dosing models. Patients with mechanic heart valve replacement were consecutively recruited between March 2012 and July 2012. The predicted warfarin dose of each patient was calculated and compared with the observed initial and stable warfarin doses. The percentage of patients whose predicted dose fell within 20% of their actual therapeutic dose (percentage within 20%), and the mean absolute error (MAE) were utilized to evaluate the predictive accuracy of all the selected algorithms.

Results

A total of 8 algorithms including Du, Huang, Miao, Wei, Zhang, Lou, Gage, and International Warfarin Pharmacogenetics Consortium (IWPC) model, were tested in 181 patients. The MAE of the Gage, IWPC and 6 Han-Chinese pharmacogenetics-based warfarin dosing algorithms was less than 0.6 mg/day in accuracy and the percentage within 20% exceeded 45% in all of the selected models in both the initial and the stable treatment stages. When patients were stratified according to the warfarin dose range, all of the equations demonstrated better performance in the ideal-dose range (1.88–4.38 mg/day) than the low-dose range (<1.88 mg/day). Among the 8 algorithms compared, the algorithms of Wei, Huang, and Miao showed a lower MAE and higher percentage within 20% in both the initial and the stable warfarin dose prediction and in the low-dose and the ideal-dose ranges.

Conclusions

All of the selected pharmacogenetics-based warfarin dosing regimens performed similarly in our cohort. However, the algorithms of Wei, Huang, and Miao showed a better potential for warfarin prediction in the initial and the stable treatment phases in Han-Chinese patients undertaking mechanic heart valve replacement.     

A Novel Admixture-Based Pharmacogenetic Approach to Refine Warfarin Dosing in Caribbean Hispanics

Aim

This study is aimed at developing a novel admixture-adjusted pharmacogenomic approach to individually refine warfarin dosing in Caribbean Hispanic patients.

Patients & Methods

A multiple linear regression analysis of effective warfarin doses versus relevant genotypes, admixture, clinical and demographic factors was performed in 255 patients and further validated externally in another cohort of 55 individuals.

Results

The admixture-adjusted, genotype-guided warfarin dosing refinement algorithm developed in Caribbean Hispanics showed better predictability (R² = 0.70, MAE = 0.72mg/day) than a clinical algorithm that excluded genotypes and admixture (R² = 0.60, MAE = 0.99mg/day), and outperformed two prior pharmacogenetic algorithms in predicting effective dose in this population. For patients at the highest risk of adverse events, 45.5% of the dose predictions using the developed pharmacogenetic model resulted in ideal dose as compared with only 29% when using the clinical non-genetic algorithm (p<0.001). The admixture-driven pharmacogenetic algorithm predicted 58% of warfarin dose variance when externally validated in 55 individuals from an independent validation cohort (MAE = 0.89 mg/day, 24% mean bias).

Conclusions

Results supported our rationale to incorporate individual’s genotypes and unique admixture metrics into pharmacogenetic refinement models in order to increase predictability when expanding them to admixed populations like Caribbean Hispanics.

Trial Registration

ClinicalTrials.gov {"type":"clinical-trial","attrs":{"text":"NCT01318057","term_id":"NCT01318057"}}NCT01318057     

Comparison of Nine Statistical Model Based Warfarin Pharmacogenetic Dosing Algorithms Using the Racially Diverse International Warfarin Pharmacogenetic Consortium Cohort Database

Objective

Multiple linear regression (MLR) and machine learning techniques in pharmacogenetic algorithm-based warfarin dosing have been reported. However, performances of these algorithms in racially diverse group have never been objectively evaluated and compared. In this literature-based study, we compared the performances of eight machine learning techniques with those of MLR in a large, racially-diverse cohort.

Methods

MLR, artificial neural network (ANN), regression tree (RT), multivariate adaptive regression splines (MARS), boosted regression tree (BRT), support vector regression (SVR), random forest regression (RFR), lasso regression (LAR) and Bayesian additive regression trees (BART) were applied in warfarin dose algorithms in a cohort from the International Warfarin Pharmacogenetics Consortium database. Covariates obtained by stepwise regression from 80% of randomly selected patients were used to develop algorithms. To compare the performances of these algorithms, the mean percentage of patients whose predicted dose fell within 20% of the actual dose (mean percentage within 20%) and the mean absolute error (MAE) were calculated in the remaining 20% of patients. The performances of these techniques in different races, as well as the dose ranges of therapeutic warfarin were compared. Robust results were obtained after 100 rounds of resampling.

Results

BART, MARS and SVR were statistically indistinguishable and significantly out performed all the other approaches in the whole cohort (MAE: 8.84–8.96 mg/week, mean percentage within 20%: 45.88%–46.35%). In the White population, MARS and BART showed higher mean percentage within 20% and lower mean MAE than those of MLR (all p values < 0.05). In the Asian population, SVR, BART, MARS and LAR performed the same as MLR. MLR and LAR optimally performed among the Black population. When patients were grouped in terms of warfarin dose range, all machine learning techniques except ANN and LAR showed significantly higher mean percentage within 20%, and lower MAE (all p values < 0.05) than MLR in the low- and high- dose ranges.

Conclusion

Overall, machine learning-based techniques, BART, MARS and SVR performed superior than MLR in warfarin pharmacogenetic dosing. Differences of algorithms’ performances exist among the races. Moreover, machine learning-based algorithms tended to perform better in the low- and high- dose ranges than MLR.     

A Pharmacogenetics-Based Warfarin Maintenance Dosing Algorithm from Northern Chinese Patients

Inconsistent associations with warfarin dose were observed in genetic variants except VKORC1 haplotype and CYP2C9*3 in Chinese people, and few studies on warfarin dose algorithm was performed in a large Chinese Han population lived in Northern China. Of 787 consenting patients with heart-valve replacements who were receiving long-term warfarin maintenance therapy, 20 related Single nucleotide polymorphisms were genotyped. Only VKORC1 and CYP2C9 SNPs were observed to be significantly associated with warfarin dose. In the derivation cohort (n = 551), warfarin dose variability was influenced, in decreasing order, by VKORC1 rs7294 (27.3%), CYP2C9*3(7.0%), body surface area(4.2%), age(2.7%), target INR(1.4%), CYP4F2 rs2108622 (0.7%), amiodarone use(0.6%), diabetes mellitus(0.6%), and digoxin use(0.5%), which account for 45.1% of the warfarin dose variability. In the validation cohort (n = 236), the actual maintenance dose was significantly correlated with predicted dose (r = 0.609, P<0.001). Our algorithm could improve the personalized management of warfarin use in Northern Chinese patients.     

An acenocoumarol dosing algorithm using clinical and pharmacogenetic data in spanish patients with thromboembolic disease

Appropriate dosing of coumarins is difficult to establish, due to significant inter-individual variability in the dose required to obtain stable anticoagulation. Several genetic and other clinical factors have been associated with the coumarins dose, and some pharmacogenetic-guided dosing algorithms for warfarin and acenocoumarol have been developed for mixed populations. We recruited 147 patients with thromboembolic disease who were on stable doses and with an international normalized ratio (INR) between 2 and 3. We ascertained the influence of clinical and genetic variables on the stable acenocoumarol dose by multiple linear regression analysis in a derivation cohort (DC; n = 117) and developed an algorithm for dosing that included clinical factors (age, body mass index and concomitant drugs) and genetic variations of VKORC1, CYP2C9, CYP4F2 and APOE. For purposes of comparison, a model including only clinical data was created. The clinical factors explained 22% of the dose variability, which increased to 60.6% when pharmacogenetic information was included (p<0.001); CYP4F2 and APOE variants explained 4.9% of this variability. The mean absolute error of the predicted acenocoumarol dose (mg/week) obtained with the pharmacogenetic algorithm was 3.63 vs. 5.08 mg/week with the clinical algorithm (95% CI: 0.88 to 2.04). In the testing cohort (n = 30), clinical factors explained a mere 7% of the dose variability, compared to 39% explained by the pharmacogenetic algorithm. Considering a more clinically relevant parameter, the pharmacogenetic algorithm correctly predicted the real stable dose in 59.8% of the cases (DC) vs. only 37.6% predicted by the clinical algorithm (95% CI: 10 to 35). Therefore the number of patients needed to genotype to avoid one over- or under-dosing was estimated to be 5.     

Development and Validation of a Clinical Scoring System for Predicting Risk of HCC in Asymptomatic Individuals Seropositive for Anti-HCV Antibodies

Background

The development of a risk assessment tool for long-term hepatocellular carcinoma risk would be helpful in identifying high-risk patients and providing information of clinical consultation.

Methods

The model derivation and validation cohorts consisted of 975 and 572 anti-HCV seropositives, respectively. The model included age, alanine aminotransferase (ALT), the ratio of aspirate aminotransferase to ALT, serum HCV RNA levels and cirrhosis status and HCV genotype. Two risk prediction models were developed: one was for all-anti-HCV seropositives, and the other was for anti-HCV seropositives with detectable HCV RNA. The Cox''s proportional hazards models were utilized to estimate regression coefficients of HCC risk predictors to derive risk scores. The cumulative HCC risks in the validation cohort were estimated by Kaplan-Meier methods. The area under receiver operating curve (AUROC) was used to evaluate the performance of the risk models.

Results

All predictors were significantly associated with HCC. The summary risk scores of two models derived from the derivation cohort had predictability of HCC risk in the validation cohort. The summary risk score of the two risk prediction models clearly divided the validation cohort into three groups (p<0.001). The AUROC for predicting 5-year HCC risk in the validation cohort was satisfactory for the two models, with 0.73 and 0.70, respectively.

Conclusion

Scoring systems for predicting HCC risk of HCV-infected patients had good validity and discrimination capability, which may triage patients for alternative management strategies.     

Warfarin Anticoagulant Therapy: A Southern Italy Pharmacogenetics-Based Dosing Model

Background and Aim

Warfarin is the most frequently prescribed anticoagulant worldwide. However, warfarin therapy is associated with a high risk of bleeding and thromboembolic events because of a large interindividual dose-response variability. We investigated the effect of genetic and non genetic factors on warfarin dosage in a South Italian population in the attempt to setup an algorithm easily applicable in the clinical practice.

Materials and Methods

A total of 266 patients from Southern Italy affected by cardiovascular diseases were enrolled and their clinical and anamnestic data recorded. All patients were genotyped for CYP2C9*2,*3, CYP4F2*3, VKORC1 -1639 G>A by the TaqMan assay and for variants VKORC1 1173 C>T and VKORC1 3730 G>A by denaturing high performance liquid chromatography and direct sequencing. The effect of genetic and not genetic factors on warfarin dose variability was tested by multiple linear regression analysis, and an algorithm based on our data was established and then validated by the Jackknife procedure.

Results

Warfarin dose variability was influenced, in decreasing order, by VKORC1-1639 G>A (29.7%), CYP2C9*3 (11.8%), age (8.5%), CYP2C9*2 (3.5%), gender (2.0%) and lastly CYP4F2*3 (1.7%); VKORC1 1173 C>T and VKORC1 3730 G>A exerted a slight effect (<1% each). Taken together, these factors accounted for 58.4% of the warfarin dose variability in our population. Data obtained with our algorithm significantly correlated with those predicted by the two online algorithms: Warfarin dosing and Pharmgkb (p<0.001; R² = 0.805 and p<0.001; R² = 0.773, respectively).

Conclusions

Our algorithm, which is based on six polymorphisms, age and gender, is user-friendly and its application in clinical practice could improve the personalized management of patients undergoing warfarin therapy.     

Readmissions and Death after ICU Discharge: Development and Validation of Two Predictive Models

Introduction

Early discharge from the ICU is desirable because it shortens time in the ICU and reduces care costs, but can also increase the likelihood of ICU readmission and post-discharge unanticipated death if patients are discharged before they are stable. We postulated that, using eICU® Research Institute (eRI) data from >400 ICUs, we could develop robust models predictive of post-discharge death and readmission that may be incorporated into future clinical information systems (CIS) to assist ICU discharge planning.

Methods

Retrospective, multi-center, exploratory cohort study of ICU survivors within the eRI database between 1/1/2007 and 3/31/2011. Exclusion criteria: DNR or care limitations at ICU discharge and discharge to location external to hospital. Patients were randomized (2∶1) to development and validation cohorts. Multivariable logistic regression was performed on a broad range of variables including: patient demographics, ICU admission diagnosis, admission severity of illness, laboratory values and physiologic variables present during the last 24 hours of the ICU stay. Multiple imputation was used to address missing data. The primary outcomes were the area under the receiver operator characteristic curves (auROC) in the validation cohorts for the models predicting readmission and death within 48 hours of ICU discharge.

Results

469,976 and 234,987 patients representing 219 hospitals were in the development and validation cohorts. Early ICU readmission and death was experienced by 2.54% and 0.92% of all patients, respectively. The relationship between predictors and outcomes (death vs readmission) differed, justifying the need for separate models. The models for early readmission and death produced auROCs of 0.71 and 0.92, respectively. Both models calibrated well across risk groups.

Conclusions

Our models for death and readmission after ICU discharge showed good to excellent discrimination and good calibration. Although prospective validation is warranted, we speculate that these models may have value in assisting clinicians with ICU discharge planning.     

Nomogram to Assess the Risk of Central Cervical Lymph Node Metastasis in Patients With Clinical N0 Papillary Thyroid Carcinoma

ObjectiveTo develop and validate an individualized risk prediction model for the need for central cervical lymph node dissection in patients with clinical N0 papillary thyroid carcinoma (PTC) diagnosed using ultrasound.MethodsUpon retrospective review, derivation and internal validation cohorts comprised 1585 consecutive patients with PTC treated from January 2017 to December 2019 at hospital A. The external validation cohort consisted of 406 consecutive patients treated at hospital B from January 2016 to June 2020. Independent risk factors for central cervical lymph node metastasis (CLNM) were determined through univariable and multivariable logistic regression analysis. An individualized risk prediction model was constructed and illustrated as a nomogram, which was internally and externally validated.ResultsThe following risk factors of CLNM were established: a solitary primary thyroid nodule’s diameter, shape, calcification, and capsular abutment-to-lesion perimeter ratio. The areas under the receiver operating characteristic curves of the risk prediction model for the internal and external validation cohorts were 0.921 and 0.923, respectively. The calibration curve showed good agreement between the nomogram-estimated probability of CLNM and the actual CLNM rates in the 3 cohorts. The decision curve analysis confirmed the clinical usefulness of the nomogram.ConclusionThis study developed and validated a model for predicting the risk of CLNM in individual patients with clinical N0 PTC, which should be an efficient tool for guiding clinical treatment.     

The pharmocogenomics of warfarin: closing in on personalized medicine

Warfarin, a coumarin anticoagulant, is used worldwide for the treatment and prevention of thromboembolic disease. Warfarin therapy, however, can be difficult to manage because of the drug's narrow therapeutic index and the wide interindividual variability in patient response. It is now clear that genetic polymorphisms in genes influencing metabolism (CYP2C9) and pharmacodynamic response (VKORC1) are strongly associated with warfarin responsiveness. Optimal warfarin dosing in turn drives other positive anticoagulation-related outcomes. Therefore, a strong basic science argument is emerging for prospective genotyping of warfarin patients. Effective clinical translation would establish warfarin pharmacogenomics as a heuristic model for personalized medicine.     

Dextran 70 in Prophylaxis of Postoperative Venous Thrombosis. A Controlled Trial

A controlled prospective trial was carried out in a group of 80 women undergoing gynaecological surgery and thought to be at risk of developing postoperative venous thrombosis. The patients, who had been randomly allocated to prophylaxis with either dextran 70 or warfarin, were well matched in age, weight and other predisposing factors.In the warfarin group, 12 out of 40 patients developed deep vein thrombosis, six of these episodes being classified as major and six as minor. In the dextran 70 group, 4 out of 40 patients developed deep vein thrombosis, all of them minor. The protective effect of dextran 70 is significantly better than that of warfarin (P<0·01) as used in the present study.     

Establishment and Validation of GV-SAPS II Scoring System for Non-Diabetic Critically Ill Patients

MethodsTraining and validation cohorts were exacted from the Multiparameter Intelligent Monitoring in Intensive Care database III version 1.3 (MIMIC-III v1.3). The GV-SAPS II score was constructed by Cox proportional hazard regression analysis and compared with the original SAPS II, Sepsis-related Organ Failure Assessment Score (SOFA) and Elixhauser scoring systems using area under the curve of the receiver operator characteristic (auROC) curve.Results4,895 and 5,048 eligible individuals were included in the training and validation cohorts, respectively. The GV-SAPS II score was established with four independent risk factors, including hyperglycemia, hypoglycemia, standard deviation of blood glucose levels (Glu_SD), and SAPS II score. In the validation cohort, the auROC values of the new scoring system were 0.824 (95% CI: 0.813–0.834, P< 0.001) and 0.738 (95% CI: 0.725–0.750, P< 0.001), respectively for 30 days and 9 months, which were significantly higher than other models used in our study (all P < 0.001). Moreover, Kaplan-Meier plots demonstrated significantly worse outcomes in higher GV-SAPS II score groups both for 30-day and 9-month mortality endpoints (all P< 0.001).ConclusionsWe established and validated a modified prognostic scoring system that integrated glucose variability for non-diabetic critically ill patients, named GV-SAPS II. It demonstrated a superior prognostic capability and may be an optimal scoring system for prognostic evaluation in this patient group.     

Clinical and genetic determinants of warfarin pharmacokinetics and pharmacodynamics during treatment initiation

Variable warfarin response during treatment initiation poses a significant challenge to providing optimal anticoagulation therapy. We investigated the determinants of initial warfarin response in a cohort of 167 patients. During the first nine days of treatment with pharmacogenetics-guided dosing, S-warfarin plasma levels and international normalized ratio were obtained to serve as inputs to a pharmacokinetic-pharmacodynamic (PK-PD) model. Individual PK (S-warfarin clearance) and PD (I(max)) parameter values were estimated. Regression analysis demonstrated that CYP2C9 genotype, kidney function, and gender were independent determinants of S-warfarin clearance. The values for I(max) were dependent on VKORC1 and CYP4F2 genotypes, vitamin K status (as measured by plasma concentrations of proteins induced by vitamin K absence, PIVKA-II) and weight. Importantly, indication for warfarin was a major independent determinant of I(max) during initiation, where PD sensitivity was greater in atrial fibrillation than venous thromboembolism. To demonstrate the utility of the global PK-PD model, we compared the predicted initial anticoagulation responses with previously established warfarin dosing algorithms. These insights and modeling approaches have application to personalized warfarin therapy.     

Systematic analysis of survival-associated alternative splicing signatures in clear cell renal cell carcinoma

Alternative splicing (AS) constitutes a major reason for messenger RNA (mRNA) and protein diversity. Increasing studies have shown a link to splicing dysfunction associated with malignant neoplasia. Systematic analysis of AS events in kidney cancer remains poorly reported. Therefore, we generated AS profiles in 533 kidney renal clear cell carcinoma (KIRC) patients in The Cancer Genome Atlas (TCGA) database using RNA-seq data. Then, prognostic models were developed in a primary cohort (N = 351) and validated in a validation cohort (N = 182). In addition, splicing networks were built by integrating bioinformatics analyses. A total of 11 268 and 8083 AS variants were significantly associated with patient overall survival time in the primary and validation KIRC cohorts, respectively, including STAT1, DAZAP1, IDS, NUDT7, and KLHDC4. The AS events in the primary KIRC cohorts served as candidate AS events to screen the independent risk factors associated with survival in the primary cohort and to develop prognostic models. The area under the curve of the receiver-operator characteristic curve for prognostic prediction in the primary and validation KIRC cohorts was 0.84 and 0.82 at 2500 days of overall survival, respectively. In addition, splicing correlation networks revealed key splicing factors (SFs) in KIRC, such as HNRNPH1, HNRNPU, KHDBS1, KHDBS3, SRSF9, RBMX, SFQ, SRP54, HNRNPA0, and SRSF6. In this study, we analyzed the AS landscape in the TCGA KIRC cohort and detected predictors (prognostic) based on AS variants with high performance for risk stratification of the KIRC cohort and revealed key SFs in splicing networks, which could act as underlying mechanisms.     

Warfarin Dosage Response Related Pharmacogenetics in Chinese Population

Objectives

As the most frequently prescribed anticoagulant, warfarin has large inter-individual variability in dosage. Genetic polymorphisms could largely explain the differences in dosage requirement. rs9923231 (VKORC1), rs7294 (VKORC1), rs1057910 (CYP2C9), rs2108622 (CYP4F2), and rs699664 (GGCX) involved in the warfarin action mechanism and the circulatory vitamin K were selected to investigate their polymorphism characteristics and their effects on the pharmacodynamics and pharmacokinetics of warfarin in Chinese population.

Methods

220 patients with cardiac valve replacement were recruited. International normalized ratio and plasma warfarin concentrations were determined. The five genetic polymorphisms were genotyping by pyro-sequencing. The relationships of maintenance dose, plasma warfarin concentration and INR were assessed among groups categorized by genotypes.

Results

rs9923231 and rs7294 in VKORC1 had the analogous genotype frequencies (D’: 0.969). 158 of 220 recruited individuals had the target INR (1.5–2.5). Patients with AA of rs9923231 and CC of rs7294 required a significantly lower maintenance dose and plasma concentration than those with AG and TC, respectively. The mean weekly maintenance dose was also significantly lower in CYP2C9 rs1057910 mutated heterozygote than in patients with the wild homozygote. Eliminating the influence from environment factors (age, body weight and gender), rs9923231 and rs1057910 could explain about 32.0% of the variability in warfarin maintenance dose; rs7294 could explain 26.7% of the variability in plasma concentration. For patients with allele G of rs9923231 and allele T of rs7294, higher plasma concentration was needed to achieve the similar goal INR.

Conclusions

A better understanding of the genetic variants in individuals can be the foundation of warfarin dosing algorithm and facilitate the reasonable and effective use of warfarin in Chinese.     

Testing the generalizability of cfDNA fragmentomic features across different studies for cancer early detection

cfDNA fragmentomic features have been used in cancer detection models; however, the generalizability of the models needs to be tested. We proposed a type of cfDNA fragmentomic feature named chromosomal arm-level fragment size distribution (ARM-FSD), evaluated and compared its performance and generalizability for lung cancer and pan-cancer detection with existing cfDNA fragmentomic features (as reference) by using cohorts from different institutions. The ARM-FSD lung cancer model outperformed the reference model by ∼10% when being tested by two external cohorts (AUC: 0.97 vs. 0.86; 0.87 vs. 0.76). For pan-cancer detection, the performance of the ARM-FSD based model is consistently higher than the reference (AUC: 0.88 vs. 0.75, 0.98 vs. 0.63) in a pan-cancer and a lung cancer external validation cohort, indicating that ARM-FSD model produces stable performance across multiple cohorts. Our study reveals ARM-FSD based models have a higher generalizability, and highlights the necessity of cross-study validation for predictive model development.     

Validation of a proteomic biomarker panel to diagnose minor-stroke and transient ischaemic attack: phase 2 of SpecTRA,a large scale translational study

Abstract

OBJECTIVE: To validate our previously developed 16 plasma-protein biomarker panel to differentiate between transient ischaemic attack (TIA) and non-cerebrovascular emergency department (ED) patients.

METHOD: Two consecutive cohorts of ED patients prospectively enrolled at two urban medical centers into the second phase of SpecTRA study (training, cohort 2A, n?=?575; test, cohort 2B, n?=?528). Plasma samples were analyzed using liquid chromatography/multiple reaction monitoring-mass spectrometry. Logistic regression models which fit cohort 2A were validated on cohort 2B.

RESULTS: Three of the panel proteins failed quality control and were removed from the panel. During validation, panel models did not outperform a simple motor/speech (M/S) deficit variable. Post-hoc analyses suggested the measured behaviour of L-selectin and coagulation factor V contributed to poor model performance. Removal of these proteins increased the external performance of a model containing the panel and the M/S variable.

CONCLUSIONS: Univariate analyses suggest insulin-like growth factor-binding protein 3 and serum paraoxonase/lactonase 3 are reliable and reproducible biomarkers for TIA status. Logistic regression models indicated L-selectin, apolipoprotein B-100, coagulation factor IX, and thrombospondin-1 to be significant multivariate predictors of TIA. We discuss multivariate feature subset analyses as an exploratory technique to better understand a panel’s full predictive potential.     

A Multi-Center Prospective Derivation and Validation of a Clinical Prediction Tool for Severe Clostridium difficile Infection

Background and Aims

Prediction of severe clinical outcomes in Clostridium difficile infection (CDI) is important to inform management decisions for optimum patient care. Currently, treatment recommendations for CDI vary based on disease severity but validated methods to predict severe disease are lacking. The aim of the study was to derive and validate a clinical prediction tool for severe outcomes in CDI.

Methods

A cohort totaling 638 patients with CDI was prospectively studied at three tertiary care clinical sites (Boston, Dublin and Houston). The clinical prediction rule (CPR) was developed by multivariate logistic regression analysis using the Boston cohort and the performance of this model was then evaluated in the combined Houston and Dublin cohorts.

Results

The CPR included the following three binary variables: age ≥ 65 years, peak serum creatinine ≥2 mg/dL and peak peripheral blood leukocyte count of ≥20,000 cells/μL. The Clostridium difficile severity score (CDSS) correctly classified 76.5% (95% CI: 70.87-81.31) and 72.5% (95% CI: 67.52-76.91) of patients in the derivation and validation cohorts, respectively. In the validation cohort, CDSS scores of 0, 1, 2 or 3 were associated with severe clinical outcomes of CDI in 4.7%, 13.8%, 33.3% and 40.0% of cases respectively.

Conclusions

We prospectively derived and validated a clinical prediction rule for severe CDI that is simple, reliable and accurate and can be used to identify high-risk patients most likely to benefit from measures to prevent complications of CDI.     

Multiwave validation sampling for error-prone electronic health records

Electronic health record (EHR) data are increasingly used for biomedical research, but these data have recognized data quality challenges. Data validation is necessary to use EHR data with confidence, but limited resources typically make complete data validation impossible. Using EHR data, we illustrate prospective, multiwave, two-phase validation sampling to estimate the association between maternal weight gain during pregnancy and the risks of her child developing obesity or asthma. The optimal validation sampling design depends on the unknown efficient influence functions of regression coefficients of interest. In the first wave of our multiwave validation design, we estimate the influence function using the unvalidated (phase 1) data to determine our validation sample; then in subsequent waves, we re-estimate the influence function using validated (phase 2) data and update our sampling. For efficiency, estimation combines obesity and asthma sampling frames while calibrating sampling weights using generalized raking. We validated 996 of 10,335 mother-child EHR dyads in six sampling waves. Estimated associations between childhood obesity/asthma and maternal weight gain, as well as other covariates, are compared to naïve estimates that only use unvalidated data. In some cases, estimates markedly differ, underscoring the importance of efficient validation sampling to obtain accurate estimates incorporating validated data.     

External validation and sub-cohort analysis of stochastic forecasting models in NICU cohorts

Hyperglycaemia is a prevalent complication in the neonatal intensive care unit (NICU) and is associated with worsened outcomes. It occurs as a result of prematurity, under-developed endogenous glucose regulatory systems, and clinical stress. The stochastic targeting (STAR) framework provides patient-specific, model-based glycaemic control with a clinically proven level of confidence on the outcome of treatment interventions, thus directly managing the risk of hypo- and hyper-glycaemia. However, stochastic models that are over conservative can limit control performance. Retrospective clinical data from 61 episodes (25 retrospective to STAR, and 36 from a prospective-STAR blood glucose control study) of insulin therapy in very-low birth weight (VLBW) and extremely-low birth weight (ELBW) neonates are used to create a new stochastic model of model-based insulin sensitivity (S_I [L/mU/min]). Sub-cohort models based on gestational age (GA) and birth weight (BW) are also created. Performance is assessed by the percentage of patients who have 90% of actual intra-patient variability in S_I captured by the 90% confidence bands of the cohort based (inter-patient) stochastic variability model created. This assessment measures per-patient accuracy for any given cohort model.Per-patient coverage trends were very similar between prospective and retrospective cohorts, providing a measure of external validation of cohort similarity. Per-patient coverage was improved through the use of BW and GA dependent stochastic models, which ensures that the stochastic models more accurately capture both inter- and intra-patient variability. Stochastic models based on insulin sensitivities during insulin treatment periods are tighter, and give better and safer glycaemic control. Overall it seems that inter-patient variation is more significant than intra-patient variation as a limiting factor in this stochastic forecasting model, and a small number of patients are essentially different in behaviour. More patient specific methods, particularly in the modelling of endogenous insulin and glucose production, will be required to further improve forecasting and glycaemic control.