Publications by authors named "Anne-Sophie Jannot"

70 Publications

One-Anastomosis Gastric Bypass Revision for Gastroesophageal Reflux Disease: Long Versus Short Biliopancreatic Limb Roux-en-Y Gastric Bypass.

Obes Surg 2022 Jan 17. Epub 2022 Jan 17.

Université de Paris, 75015, Paris, France.

Purpose: One-anastomosis gastric bypass (OAGB) may be associated with refractory gastroesophageal reflux disease (GERD). The nature of this GERD remains unclear. This complication can be treated either with an additional enteroenterostomy without shortening of gastric pouch (long biliopancreatic limb Roux-en-Y gastric bypass (L-BPL-RYGB)) or revision to conventional short biliopancreatic limb Roux-en-Y gastric bypass (S-BPL-RYGB). The objective of this study is to compare the aforementioned procedures in terms of efficacy on GERD symptoms.

Materials And Methods: Retrospective analysis between October 2012 and June 2020.

Results: Fifty-two patients underwent OAGB revision to S-BPL-RYGB (n = 21) or L-BPL-RYGB (n = 31) secondary to GERD. Investigation with pH impedance prior to revision was performed in 15 patients showing biliary reflux (BR) in 7 (46.6%), acid reflux (AR) in 6 (40%), and no confirmation in 2. Patients with AR had a revision to S-BPL-RYGB, whereas patients with BR underwent L-BPL-RYGB. Among the patients without pH metry results (n = 37), S-BPL-RYGB was performed for associated disabling digestive disorders or nutritional deficiencies. GERD was treated in 68% of patients with L-BPL-RYGB versus 95% of patients after S-BPL-RYGB. Patients, whose decision for revisional procedure was based on the results of pH impedance testing, did not reveal refractory GERD.

Conclusion: L-BPL-RYGB seems appropriate in patients with BR, whereas conversion to S-BPL-RYGB should be preferred if AR is present.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1007/s11695-022-05892-6DOI Listing
January 2022

Clinical diagnosis, outcomes and treatment of thiamine deficiency in a tertiary hospital.

Clin Nutr 2022 Jan 11;41(1):33-39. Epub 2021 Nov 11.

Hôpital Européen Georges Pompidou, Service de Nutrition, Centre Spécialisé Obésité (CSO) Ile-de-France-Sud, APHP-centre, Université de Paris, France; Université de Paris, F-75015, Paris, France. Electronic address:

Background: Acute thiamine deficiency can occur in patients with or without history of alcohol abuse and can lead to life-threatening complications. Clinical diagnosis is challenging, often resulting in delayed recognition and treatment. Patients may present with heterogenous symptoms, more diverse than the historical neurological description. Cerebral MRI can contribute to the diagnosis in patients with neurological signs but it is not always feasible in emergency settings. Prompt parenteral supplementation is required to obtain the improvement of symptoms and avoid chronic complications.

Aims: To describe the clinical presentation of reported cases of thiamine deficiency, assess prescription and results of cerebral imaging, review treatments that had been prescribed in accordance or not with available guidelines, and study the short-term outcome of these patients.

Methods: This is a monocentric retrospective analysis of all reported cases of thiamine deficiency in a French tertiary hospital between January 1st 2008 and December 31st 2018.

Results: Fifty-six cases were identified during the study period. Forty-five (80%) patients had a history of alcohol abuse. Most patients were diagnosed based on neurological symptoms but non-specific and digestive symptoms were frequent. Thirty-four percent of patients fulfilled clinical criteria for malnutrition. A brain MRI was performed in 54% of patients and was abnormal in 63% of these cases. Eighty-five percent of patients were treated by parenteral thiamine administration and the supplementation was continued orally in 55% of them. The majority of patients initially received 1000 mg daily of IV thiamine but the dose and duration of thiamine supplementation were variable. At the time of discharge, partial or complete improvement of symptoms was noted in 59% of patients.

Conclusion: This study highlights the clinical and radiological heterogeneity of thiamine deficiency. These observations should encourage starting thiamine supplementation early in patients with risk factors or suggestive symptoms even in non-alcoholic patients, and underline the importance of early nutritional support.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.clnu.2021.10.021DOI Listing
January 2022

Aetiological classification and prognosis in patients with heart failure with preserved ejection fraction.

ESC Heart Fail 2021 Nov 29. Epub 2021 Nov 29.

Université de Paris, INSERM, PARCC, Paris, F-75006, France.

Aims: Heart failure with preserved ejection fraction (HFpEF) is a heterogeneous syndrome with various causes that may influence prognosis.

Methods And Results: We extracted the electronic medical records for 2180 consecutive patients hospitalized between 2016 and 2019 for decompensated heart failure. Using a text mining algorithm looking for a left ventricular ejection fraction ≥50% and plasma brain natriuretic peptide level >100 pg/mL, we identified 928 HFpEF patients. We screened for a prevailing cause of HFpEF according to European guidelines and found that 418 (45.0%) patients had secondary HFpEF due to either myocardial (n = 125, 13.5%) or loading condition abnormalities (n = 293, 31.5%), while the remaining 510 (55.0%) patients had idiopathic HFpEF. We assessed the association between the causes of HFpEF and survival collected up to 31 December 2020 using Cox proportional hazards analysis. Even though patients with idiopathic HFpEF were older, frequently female, and had frequent co-morbidities and a higher crude mortality rate compared with secondary HFpEF patients, their prognosis was similar after adjustment for age and sex. Unsupervised clustering analysis revealed three main phenogroups with different distribution of idiopathic vs. secondary HFpEF. The phenogroup with the highest proportion of idiopathic HFpEF (69%) had (i) an excess rate of non-cardiac co-morbidities including chronic obstructive pulmonary disease (31%) or obesity (41%) and (ii) a better prognosis compared with the two other phenogroups enriched with secondary HFpEF.

Conclusions: Aetiological classification provides clinical and prognostic information and may be useful to better decipher the clinical heterogeneity of HFpEF.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/ehf2.13717DOI Listing
November 2021

The ongoing French BaMaRa-BNDMR cohort: implementation and deployment of a nationwide information system on rare disease.

J Am Med Inform Assoc 2021 Nov 6. Epub 2021 Nov 6.

Banque Nationale de Données Maladies Rares, DSI-I&D, APHP, Paris, France.

Background: BaMaRa allows the secure collection and deidentified centralization of medical data from all patients followed-up in a rare disease expert network in France, based on a minimum data set (SDM-MR). The present article describes BaMaRa information system implementation and development across the whole national territory as well as data access requests through BNDMR, the data warehouse which centralizes all BaMaRa data, during the 2015-2020 period.

Materials And Methods: SDM-MR is made up of 60 interoperable items and is routinely collected through BaMaRa in rare disease centers as part of care and discharged into BNDMR after deidentification and data reconciliation. Data access is regulated by a scientific committee.

Results: In total, 668 002 affected patients had an SDM-MR recorded in BNDMR by the end of 2020 with a mean value of 3.4 activities per patients. Data access was provided for 66 projects.

Conclusion: The BaMaRa-BNDMR infrastructure provides an administrative and epidemiological resources for rare diseases in France.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1093/jamia/ocab237DOI Listing
November 2021

Cardiometabolic Disorders and the Risk of Critical COVID-19 as Compared to Influenza Pneumonia.

J Clin Med 2021 Oct 8;10(19). Epub 2021 Oct 8.

Université de Paris, F-75006 Paris, France.

We aimed to compare the influence of cardiometabolic disorders on the incidence of severe COVID-19 vs. non-COVID pneumonia. We included all consecutive patients admitted with SARS-CoV-2-positive pneumonia between 12 March 2020 and 1 April 2020 and compared them to patients with influenza pneumonia hospitalized between December 2017 and December 2019 at the same tertiary hospital in Paris. Patients with COVID-19 were significantly younger and more frequently male. In the analysis adjusted for age and sex, patients with COVID-19 were more likely to be obese (adjOR: 2.25; 95% CI 1.24-4.09; = 0.0076) and receive diuretics (adjOR: 2.13; 95% CI 1.12-4.03; = 0.021) but were less likely to be smokers (adjOR: 0.40; 95% CI 0.24-0.64; = 0.0002), have COPD (adjOR: 0.25; 95% CI 0.11-0.56; = 0.0008), or have a previous or active cancer diagnosis (adjOR: 0.54, 95% CI 0.32-0.91; = 0.020). The rate of ICU admission was significantly higher in patients with COVID-19 (32.4% vs. 5.2% < 0.0001). Obesity was significantly associated with the risk of direct ICU admission in patients with COVID-19 but not in patients with influenza pneumonia. Likewise, pre-existing hypertension was significantly associated with mortality in patients with COVID-19 but not in patients with influenza pneumonia. Cardiometabolic disorders differentially influenced the risk of presenting with severe COVID-19 or influenza pneumonia.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3390/jcm10194618DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8509116PMC
October 2021

Binacox: automatic cut-point detection in high-dimensional Cox model with applications in genetics.

Biometrics 2021 Aug 18. Epub 2021 Aug 18.

LaMME, UEVE and UMR 8071, Paris Saclay University, Evry, France.

We introduce binacox, a prognostic method to deal with the problem of detecting multiple cut-points per feature in a multivariate setting where a large number of continuous features are available. The method is based on the Cox model and combines one-hot encoding with the binarsity penalty, which uses total-variation regularization together with an extra linear constraint, and enables feature selection. Original nonasymptotic oracle inequalities for prediction (in terms of Kullback-Leibler divergence) and estimation with a fast rate of convergence are established. The statistical performance of the method is examined in an extensive Monte Carlo simulation study, and then illustrated on three publicly available genetic cancer data sets. On these high-dimensional data sets, our proposed method outperforms state-of-the-art survival models regarding risk prediction in terms of the C-index, with a computing time orders of magnitude faster. In addition, it provides powerful interpretability from a clinical perspective by automatically pinpointing significant cut-points in relevant variables.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1111/biom.13547DOI Listing
August 2021

10 years of CEMARA database in the AnDDI-Rares network: a unique resource facilitating research and epidemiology in developmental disorders in France.

Orphanet J Rare Dis 2021 08 4;16(1):345. Epub 2021 Aug 4.

Centre de Référence Anomalies du Développement et Syndromes Malformatifs, CHU de Dijon, Dijon, France.

Background: In France, the Ministry of Health has implemented a comprehensive program for rare diseases (RD) that includes an epidemiological program as well as the establishment of expert centers for the clinical care of patients with RD. Since 2007, most of these centers have entered the data for patients with developmental disorders into the CEMARA population-based registry, a national online data repository for all rare diseases. Through the CEMARA web portal, descriptive demographic data, clinical data, and the chronology of medical follow-up can be obtained for each center. We address the interest and ongoing challenges of this national data collection system 10 years after its implementation.

Methods: Since 2007, clinicians and researchers have reported the "minimum dataset (MDS)" for each patient presenting to their expert center. We retrospectively analyzed administrative data, demographic data, care organization and diagnoses.

Results: Over 10 years, 228,243 RD patients (including healthy carriers and family members for whom experts denied any suspicion of RD) have visited an expert center. Among them, 167,361 were patients affected by a RD (median age 11 years, 54% children, 46% adults, with a balanced sex ratio), and 60,882 were unaffected relatives (median age 37 years). The majority of patients (87%) were seen no more than once a year, and 52% of visits were for a diagnostic procedure. Among the 2,869 recorded rare disorders, 1,907 (66.5%) were recorded in less than 10 patients, 802 (28%) in 10 to 100 patients, 149 (5.2%) in 100 to 1,000 patients, and 11 (0.4%) in > 1,000 patients. Overall, 45.6% of individuals had no diagnosis and 6.7% had an uncertain diagnosis. Children were mainly referred by their pediatrician (46%; n = 55,755 among the 121,136 total children referrals) and adults by a medical specialist (34%; n = 14,053 among the 41,564 total adult referrals). Given the geographical coverage of the centers, the median distance from the patient's home was 25.1 km (IQR = 6.3 km-64.2 km).

Conclusions: CEMARA provides unprecedented support for epidemiological, clinical and therapeutic studies in the field of RD. Researchers can benefit from the national scope of CEMARA data, but also focus on specific diseases or patient subgroups. While this endeavor has been a major collective effort among French RD experts to gather large-scale data into a single database, it provides tremendous potential to improve patient care.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s13023-021-01957-4DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8335940PMC
August 2021

Change in Birth Rate Before and After Bariatric Surgery in France.

Obes Surg 2021 10 12;31(10):4657-4661. Epub 2021 Jul 12.

Assistance Publique-Hôpitaux de Paris (AP-HP), Service de Nutrition, Centre Spécialisé Obésité, Hôpital Européen Georges Pompidou, Paris, France.

Importance: Bariatric surgery worldwide has grown significantly over the past years and is performed, in a vast majority, in women of childbearing age. The impact of these procedures on birth rates remains largely unknown.

Objective: The main objective was to study the evolution of the birth rates in these women, before and after bariatric surgery and in comparison with women from the general population. The secondary objectives were to compare the birth rates before and after surgery according to the Body Mass Index (BMI) (BMI 40-50 versus >50 kg/m).

Method: Our analyses are based on a national medico-administrative database. All women of childbearing age and who had bariatric surgery between 2012 and 2016 were included, and we included all deliveries between 2012 and 2018 in this population. We compared the birth rates before and after bariatric surgery. National statistics (INSEE, Institut National de la Statistique et des Etudes Economiques) were used for comparison with women from the general population.

Findings: A total of 69,932 women were included between 2012 and 2016, with a median age at surgery of 33 years [27; 39]. Among them, 9391 (13%) had a BMI over 50 kg/m before surgery, 46,818 (67%) benefited from a sleeve gastrectomy (SG), and 23,376 (33%) from a gastric bypass (GBP). Birth rates tend to be smaller after surgery compared to before surgery for women under 27 years (OR: 0.92, CI95% [0.88; 0.96]), while being not different after 27 in both groups (OR: 1.00, CI95% [0.97; 1.03]). Birth rates of obese women who have benefited or will benefit from the bariatric surgery were higher than the general population, with a peak at a younger age, and then decline earlier than the general population. Trends were different according to BMI class. Birth rates were higher for women with BMI between 40 and 50 kg/m compared to women with BMI greater than 50 kg/m (OR=1.28, CI95% [1.21; 1.36]) before surgery while slightly lower after surgery (OR=0.95, CI95% [0.91; 0.99]).

Conclusions And Relevance: We showed that birth rates after bariatric surgery were not different before and after bariatric surgery except for women under 27 years old who had a smaller birth rate. Interestingly, we observed an improvement of birth rates after surgery for women with a BMI of 50 kg/m or more.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1007/s11695-021-05561-0DOI Listing
October 2021

Low-income neighbourhood was a key determinant of severe COVID-19 incidence during the first wave of the epidemic in Paris.

J Epidemiol Community Health 2021 12 30;75(12):1143-1146. Epub 2021 Jun 30.

Hôpital Européen Georges Pompidou, Service d'informatique médicale, biostatistiques et santé publique, AP-HP, Paris, France.

Background: Previous studies have demonstrated that socioeconomic factors are associated with COVID-19 incidence. In this study, we analysed a broad range of socioeconomic indicators in relation to hospitalised cases in the Paris area.

Methods: We extracted 303 socioeconomic indicators from French census data for 855 residential units in Paris and assessed their association with COVID-19 hospitalisation risk.

Findings: The indicators most associated with hospitalisation risk were the third decile of population income (OR=9.10, 95% CI 4.98 to 18.39), followed by the primary residence rate (OR=5.87, 95% CI 3.46 to 10.61), rate of active workers in unskilled occupations (OR=5.04, 95% CI 3.03 to 8.85) and rate of women over 15 years old with no diploma (OR=5.04, 95% CI 3.03 to 8.85). Of note, population demographics were considerably less associated with hospitalisation risk. Among these indicators, the rate of women aged between 45 and 59 years (OR=2.17, 95% CI 1.40 to 3.44) exhibited the greatest level of association, whereas population density was not associated. Overall, 86% of COVID-19 hospitalised cases occurred within the 45% most deprived areas.

Interpretation: Studying a broad range of socioeconomic indicators using census data and hospitalisation data as a readily available and large resource can provide real-time indirect information on populations with a high incidence of COVID-19.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1136/jech-2020-216068DOI Listing
December 2021

Differential association between inflammatory cytokines and multiorgan dysfunction in COVID-19 patients with obesity.

PLoS One 2021 26;16(5):e0252026. Epub 2021 May 26.

INSERM UMRS 1138, Cordeliers Research Center, Team Inflammation, Complement, and Cancer, Paris, France.

To investigate the mechanisms underlying the SARS-CoV-2 infection severity observed in patients with obesity, we performed a prospective study of 51 patients evaluating the impact of multiple immune parameters during 2 weeks after admission, on vital organs' functions according to body mass index (BMI) categories. High-dimensional flow cytometric characterization of immune cell subsets was performed at admission, 30 systemic cytokines/chemokines levels were sequentially measured, thirteen endothelial markers were determined at admission and at the zenith of the cytokines. Computed tomography scans on admission were quantified for lung damage and hepatic steatosis (n = 23). Abnormal BMI (> 25) observed in 72.6% of patients, was associated with a higher rate of intensive care unit hospitalization (p = 0.044). SARS-CoV-2 RNAaemia, peripheral immune cell subsets and cytokines/chemokines were similar among BMI groups. A significant association between inflammatory cytokines and liver, renal, and endothelial dysfunctions was observed only in patients with obesity (BMI > 30). In contrast, early signs of lung damage (ground-glass opacity) correlated with Th1/M1/inflammatory cytokines only in normal weight patients. Later lesions of pulmonary consolidation correlated with BMI but were independent of cytokine levels. Our study reveals distinct physiopathological mechanisms associated with SARS-CoV-2 infection in patients with obesity that may have important clinical implications.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0252026PLOS
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8153504PMC
June 2021

Correction to: Assessing rare diseases prevalence using literature quantification.

Orphanet J Rare Dis 2021 May 10;16(1):213. Epub 2021 May 10.

Department of Medical Informatics, Hôpital Européen Georges Pompidou, AP-HP, 20 Rue Leblanc, 75015, Paris, France.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s13023-021-01854-wDOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8111929PMC
May 2021

Influence of the clindamycin administration route on the magnitude of clindamycin-rifampicin interaction: a prospective pharmacokinetic study.

Clin Microbiol Infect 2021 Dec 26;27(12):1857.e1-1857.e7. Epub 2021 Apr 26.

Unité Fonctionnelle de Pharmacologie, Groupe Hospitalier Paris-Seine-Saint-Denis, Bondy, France.

Objectives: An important clindamycin-rifampicin pharmacokinetic (PK) interaction has been reported, but the potential influence of the clindamycin administration route on that interaction is unknown. This prospective, observational, comparative PK study was undertaken to characterize and analyse the impact of the route, comparing the rifampicin enzyme-inductor effects on clindamycin clearance (CLclin) for oral versus intravenous (IV) administration.

Methods: Patients with bone-and-joint infections (BJIs) were treated with clindamycin monotherapy (n = 20) or clindamycin-rifampicin combination therapy (n = 19). Patients received continuous IV clindamycin infusion for 2-6 weeks, followed by an oral regimen. Liquid chromatography-mass spectrometry was used to measure plasma clindamycin concentrations at the end of IV and after 2 weeks of oral treatment. The ratios of the mean CLclin for the combination and monotherapy groups were calculated for IV (Riv) and oral (Rpo) routes, with the final ratio, Rf = Rpo/Riv, representing the fold change of the rifampicin-inducing effect from the IV to the oral route.

Results: Comparing monotherapy with combination-therapy groups, the former's median steady-state concentration was two-fold higher after IV administration (8.49 versus 3.82 mg/L, p < 0.001) and its median AUC was 12 times higher after oral intake (37.7 versus 3.1 mg.h/L, p < 0.001). Riv, Rpo and Rf were 2.68, 18.8 and 7.0 respectively.

Conclusion: The magnitude of this interaction was markedly increased by oral intake, questioning the use of oral treatment for difficult-to-treat infections like BJIs. Nevertheless, the clindamycin-rifampicin combination seems possible provided that clindamycin is administered by continuous IV infusion.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.cmi.2021.04.017DOI Listing
December 2021

Assessing rare diseases prevalence using literature quantification.

Orphanet J Rare Dis 2021 03 20;16(1):139. Epub 2021 Mar 20.

Department of Medical Informatics, Hôpital Européen Georges Pompidou, AP-HP, 20 Rue Leblanc, 75015, Paris, France.

Introduction: Estimating the prevalence of diseases is crucial for the organization of healthcare. The amount of literature on a rare pathology could help differentiate between rare and very rare diseases. The objective of this work was to evaluate to what extent the number of publications can be used to predict the prevalence of a given pathology.

Methods: We queried Orphanet for the global prevalence class for all conditions for which it was available. For these pathologies, we cross-referenced the Orphanet, MeSH, and OMIM vocabularies to assess the number of publication available on Pubmed using three different query strategies (one proposed in the literature, and two built specifically for this study). We first studied the association of the number of publications obtained by each of these query strategies with the prevalence class, then their predictive ability.

Results: Class prevalence was available for 3128 conditions, 2970 had a prevalence class < 1/1,000,000, 41 of 1-9/1,000,000, 84 of 1-9/100,000, and 33 of 1-9/10,000. We show a significant association and excellent predictive performance of the number of publication, with an AUC over 94% for the best query strategy.

Conclusion: Our study highlights the link and the excellent predictive performance of the number of publications on the prevalence of rare diseases provided by Orphanet.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s13023-020-01639-7DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7980535PMC
March 2021

Proteinuria and Clinical Outcomes in Hospitalized COVID-19 Patients: A Retrospective Single-Center Study.

Clin J Am Soc Nephrol 2021 Feb 23. Epub 2021 Feb 23.

Université de Paris, Paris, France.

Background And Objectives: Kidney involvement is frequent among patients with coronavirus disease 2019 (COVID-19), and occurrence of AKI is associated with higher mortality in this population. The objective of this study was to describe occurrence and significance of proteinuria in this setting. DESIGN, SETTING, PARTICIPANTS & MEASUREMENTS : We conducted a single-center retrospective study to describe the characteristic features of proteinuria measured within 48 hours following admission among patients with COVID-19 admitted in a tertiary care hospital in France, and to evaluate its association with initiation of dialysis, intensive care unit admission, and death.

Results: Among 200 patients with available data, urine protein-creatinine ratio at admission was ≥1 g/g for 84 (42%), although kidney function was normal in most patients, with a median serum creatinine of 0.94 mg/dl (interquartile range, 0.75-1.21). Median urine albumin-creatinine ratio was 110 mg/g (interquartile range, 50-410), with a urine albumin-protein ratio <50% in 92% of patients. Urine retinol binding protein concentrations, available for 85 patients, were ≥0.03 mg/mmol in 62% of patients. Urine protein-creatinine ratio ≥1 g/g was associated with initiation of dialysis (odds ratio, 4.87; 95% confidence interval, 2.03 to 13.0; <0.001), admission to the intensive care unit (odds ratio, 3.55; 95% confidence interval, 1.93 to 6.71; <0.001), and death (odds ratio, 3.56; 95% confidence interval, 1.90 to 6.54; <0.001).

Conclusions: Proteinuria is very frequent among patients admitted for COVID-19 and may precede AKI. Low levels of albuminuria suggest a predominant tubular origin, confirmed by the elevated levels of urine retinol binding protein. Urine protein-creatinine ratio ≥1 g/g at admission is strongly associated with poor kidney and patient outcome.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.2215/CJN.09130620DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8092053PMC
February 2021

Observational study of haloperidol in hospitalized patients with COVID-19.

PLoS One 2021 19;16(2):e0247122. Epub 2021 Feb 19.

Département de Psychiatrie, Hôpital Corentin-Celton, Centre Université de Paris, AP-HP, Issy-les-Moulineaux, France.

Background: Haloperidol, a widely used antipsychotic, has been suggested as potentially useful for patients with COVID-19 on the grounds of its in-vitro antiviral effects against SARS-CoV-2, possibly through sigma-1 receptor antagonist effect.

Methods: We examined the associations of haloperidol use with intubation or death and time to discharge home among adult patients hospitalized for COVID-19 at Assistance Publique-Hôpitaux de Paris (AP-HP) Greater Paris University hospitals. Study baseline was defined as the date of hospital admission. The primary endpoint was a composite of intubation or death and the secondary endpoint was discharge home among survivors in time-to-event analyses. In the primary analyses, we compared these two outcomes between patients receiving and not receiving haloperidol using univariate Cox regression models in matched analytic samples based on patient characteristics and other psychotropic medications. Sensitivity analyses included propensity score analyses with inverse probability weighting and multivariable Cox regression models.

Results: Of 15,121 adult inpatients with a positive COVID-19 PT-PCR test, 39 patients (0.03%) received haloperidol within the first 48 hours of admission. Over a mean follow-up of 13.8 days (SD = 17.9), 2,024 patients (13.4%) had a primary end-point event and 10,179 patients (77.6%) were discharged home at the time of study end on May 1st. The primary endpoint occurred in 9 patients (23.1%) who received haloperidol and 2,015 patients (13.4%) who did not. The secondary endpoint of discharge home occurred in 16 patients (61.5%) who received haloperidol and 9,907 patients (85.8%) who did not. There were no significant associations between haloperidol use and the primary (HR, 0.80; 95% CI, 0.39 to 1.62, p = 0.531) and secondary (HR, 1.30; 95% CI, 0.74 to 2.28, p = 0.355) endpoints. Results were similar in multiple sensitivity analyses.

Conclusion: Findings from this multicenter observational study suggest that haloperidol use prescribed at a mean dose of 4.5 mg per day (SD = 5.2) for a mean duration of 8.4 days (SD = 7.2) may not be associated with risk of intubation or death, or with time to discharge home, among adult patients hospitalized for COVID-19.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0247122PLOS
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7895415PMC
March 2021

Observational Study of Chlorpromazine in Hospitalized Patients with COVID-19.

Clin Drug Investig 2021 Mar 9;41(3):221-233. Epub 2021 Feb 9.

DMU Psychiatrie et Addictologie, AP-HP. Centre-Université de Paris, Hôpital Corentin-Celton, 4 parvis Corentin Celton, 92130, Issy-les-Moulineaux, France.

Introduction: Chlorpromazine has been suggested as being potentially useful in patients with coronavirus disease 2019 (COVID-19) on the grounds of its potential antiviral and anti-inflammatory effects.

Objective: The aim of this study was to examine the association between chlorpromazine use and mortality among adult patients hospitalized for COVID-19.

Methods: We conducted an observational, multicenter, retrospective study at Assistance Publique-Hôpitaux de Paris (AP-HP) Greater Paris University hospitals. Study baseline was defined as the date of first prescription of chlorpromazine during hospitalization for COVID-19. The primary endpoint was death. Among patients who had not been hospitalized in intensive care units (ICUs), we compared this endpoint between those who received chlorpromazine and those who did not, in time-to-event analyses adjusted for patient characteristics, clinical markers of disease severity, and other psychotropic medications. The primary analysis used a Cox regression model with inverse probability weighting. Multiple sensitivity analyses were performed.

Results: Of the 14,340 adult inpatients hospitalized outside ICUs for COVID-19, 55 patients (0.4%) received chlorpromazine. Over a mean follow-up of 14.3 days (standard deviation [SD] 18.2), death occurred in 13 patients (23.6%) who received chlorpromazine and 1289 patients (9.0%) who did not. In the primary analysis, there was no significant association between chlorpromazine use and mortality (hazard ratio [HR] 2.01, 95% confidence interval [CI] 0.75-5.40; p = 0.163). Sensitivity analyses included a Cox regression in a 1:5 ratio matched analytic sample that showed a similar result (HR 1.67, 95% CI 0.91-3.06; p = 0.100) and a multivariable Cox regression that indicated a significant positive association (HR 3.10, 95% CI 1.31-7.34; p = 0.010).

Conclusion: Our results suggest that chlorpromazine prescribed at a mean daily dose of 70.8 mg (SD 65.3) was not associated with reduced mortality.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1007/s40261-021-01001-0DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7871023PMC
March 2021

Association between antidepressant use and reduced risk of intubation or death in hospitalized patients with COVID-19: results from an observational study.

Mol Psychiatry 2021 09 4;26(9):5199-5212. Epub 2021 Feb 4.

AP-HP.Centre-Université de Paris, Hôpital Corentin-Celton, Département Médico-Universitaire de Psychiatrie et Addictologie, 92130, Issy-les-Moulineaux, France.

A prior meta-analysis showed that antidepressant use in major depressive disorder was associated with reduced plasma levels of several pro-inflammatory mediators, which have been associated with severe COVID-19. Recent studies also suggest that several antidepressants may inhibit acid sphingomyelinase activity, which may prevent the infection of epithelial cells with SARS-CoV-2, and that the SSRI fluoxetine may exert in-vitro antiviral effects on SARS-CoV-2. We examined the potential usefulness of antidepressant use in patients hospitalized for COVID-19 in an observational multicenter retrospective cohort study conducted at AP-HP Greater Paris University hospitals. Of 7230 adults hospitalized for COVID-19, 345 patients (4.8%) received an antidepressant within 48 h of hospital admission. The primary endpoint was a composite of intubation or death. We compared this endpoint between patients who received antidepressants and those who did not in time-to-event analyses adjusted for patient characteristics, clinical and biological markers of disease severity, and other psychotropic medications. The primary analysis was a multivariable Cox model with inverse probability weighting. This analysis showed a significant association between antidepressant use and reduced risk of intubation or death (HR, 0.56; 95% CI, 0.43-0.73, p < 0.001). This association remained significant in multiple sensitivity analyses. Exploratory analyses suggest that this association was also significant for SSRI and non-SSRI antidepressants, and for fluoxetine, paroxetine, escitalopram, venlafaxine, and mirtazapine (all p < 0.05). These results suggest that antidepressant use could be associated with lower risk of death or intubation in patients hospitalized for COVID-19. Double-blind controlled randomized clinical trials of antidepressant medications for COVID-19 are needed.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41380-021-01021-4DOI Listing
September 2021

Measured glomerular filtration rate (GFR) significantly and rapidly decreases after radical cystectomy for bladder cancer.

Sci Rep 2020 09 30;10(1):16145. Epub 2020 Sep 30.

Department of Physiology, Functional Explorations Unit, Hôpital Européen Georges Pompidou, Paris, France.

Precise determination of glomerular filtration rate (GFR) is essential for the management of patients with muscle-invasive bladder cancer (MIBC). We aim to describe the early evolution of measured GFR (mGFR) after radical cystectomy and urinary diversion (RCUD) and to identify risk factors for GFR decline. GFR measurement using Cr-EDTA continuous infusion, estimated GFR (eGFR) from five published equations and renal scintigraphy with split renal function determination were performed before and 6 months after RCUD. Chronic Kidney Disease (mGFR < 60 mL/min/1.73 m) and GFR stages were defined according to the KDIGO guidelines using mGFR. Twenty-seven patients (men 85%, median age 65, IQR 59; 68 years) were included. A total of 20 (74%) patients experienced significant mGFR decline at 6 months postoperatively. Median mGFR decreased from 84.1 pre-operatively (IQR 65.3; 97.2) to 69.9 mL/min/1.73 m (IQR 55.0; 77.9) 6 months after surgery (p < 0.001). Thirteen (48%) patients had a progression to a worse GFR stage. Of the 22 patients without pre-operative CKD, 5 (23%) developed post-operative CKD. Diabetes mellitus was more frequent in patients in the highest tertile of relative mGFR decline (44% vs. 11%, p = 0.02) and platinum-based adjuvant chemotherapy tended to be more frequently used in these patients (44% vs. 17%, p = 0.06). Importantly, pre-operative weight was independently and negatively associated with post-operative mGFR and with mGFR slope in multivariable analyses. In this prospective series, we demonstrated that early and significant mGFR decline occurred after RCUD and perioperative platinum-based chemotherapy, especially in patients with diabetes mellitus and overweight.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41598-020-73191-0DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7528003PMC
September 2020

Diagnostic Algorithm to Differentiate Benign Atypical Leiomyomas from Malignant Uterine Sarcomas with Diffusion-weighted MRI.

Radiology 2020 11 15;297(2):361-371. Epub 2020 Sep 15.

From the Departments of Radiology (C.A.W., C.B., A.B., L.S.F.), Medical Informatics and Public Health (A.S.J.), Gynecologic and Breast Oncologic Surgery (C.C., A.S.B.), and Pathology (M.A.L.B.), AP-HP, Hôpital Européen Georges Pompidou, 20 Rue Leblanc, Université de Paris, F-75015 Paris, France; Department of Radiology McGill University Health Centre, Montreal, Canada (P.A.B., C.R.); Department of Radiology, AP-HP, Hôpital Tenon, Sorbonne Université, Paris, France (I.T.N.); and Université de Paris, PARCC, INSERM, France (A.B., L.F.).

Background Improving the differentiation of uterine sarcomas from atypical leiomyomas remains a clinical challenge and is needed to avoid inappropriate surgery. Purpose To develop a diagnostic algorithm including diffusion-weighted MRI criteria to differentiate malignant uterine sarcomas from benign atypical leiomyomas. Materials and Methods This case-control retrospective study identified women with an atypical uterine mass at MRI between January 2000 and April 2017, with surgery or MRI follow-up after 1 year or longer. A diagnostic algorithm including T2-weighted MRI and diffusion-weighted imaging (DWI) signal and apparent diffusion coefficient (ADC) values was developed to predict for sarcoma. The training set consisted of 51 sarcomas and 105 leiomyomas. Two external validation sets were used to evaluate interreader reproducibility (16 sarcomas; 26 leiomyomas) and impact of reader experience (29 sarcomas; 30 leiomyomas). Wilson confidence intervals (CIs) were calculated for sensitivity and specificity. Results Evaluated were 156 women (median age, 50 years; interquartile range, 44-63 years). Predictive MRI criteria for malignancy were enlarged lymph nodes or peritoneal implants, high DWI signal greater than that in endometrium, and ADC less than or equal to 0.905 × 10 mm/sec. Conversely, a global or focal area of low T2 signal intensity and a low or an intermediate DWI signal less than that in endometrium or lymph nodes allowed readers to confidently diagnose as benign a uterine mass demonstrating one or more of these signs ( < .001) in 100% cases in all three data sets. The sensitivities and specificities of the algorithm for diagnosis of malignancy were 98% (50 of 51 masses; 95% CI: 90%, 100%) and 94% (99 of 105 masses; 95% CI: 88%, 98%) in the training set; 88% (14 of 16 masses; 95% CI: 64%, 97%) and 100% (26 of 26 masses; 95% CI: 87%, 100%) in the validation set; and 83% (24 of 29 masses; 95% CI: 65%, 92%) and 97% (29 of 30 masses; 95% CI: 83%, 99%) for the less experienced reader, respectively. Conclusion A diagnostic algorithm with predictive features including lymphadenopathy, high diffusion-weighted imaging signal with reference to endometrium, and low apparent diffusion coefficient enabled differentiation of malignant sarcomas from atypical leiomyomas, and it may assist inexperienced readers. © RSNA, 2020 See also the editorial by Méndez in this issue.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1148/radiol.2020191658DOI Listing
November 2020

International electronic health record-derived COVID-19 clinical course profiles: the 4CE consortium.

NPJ Digit Med 2020 19;3:109. Epub 2020 Aug 19.

Biomedical Informatics Center, Medical University of South Carolina, Charleston, SC USA.

We leveraged the largely untapped resource of electronic health record data to address critical clinical and epidemiological questions about Coronavirus Disease 2019 (COVID-19). To do this, we formed an international consortium (4CE) of 96 hospitals across five countries (www.covidclinical.net). Contributors utilized the Informatics for Integrating Biology and the Bedside (i2b2) or Observational Medical Outcomes Partnership (OMOP) platforms to map to a common data model. The group focused on temporal changes in key laboratory test values. Harmonized data were analyzed locally and converted to a shared aggregate form for rapid analysis and visualization of regional differences and global commonalities. Data covered 27,584 COVID-19 cases with 187,802 laboratory tests. Case counts and laboratory trajectories were concordant with existing literature. Laboratory tests at the time of diagnosis showed hospital-level differences equivalent to country-level variation across the consortium partners. Despite the limitations of decentralized data generation, we established a framework to capture the trajectory of COVID-19 disease in patients and their response to interventions.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41746-020-00308-0DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7438496PMC
August 2020

Obesity Doubles Mortality in Patients Hospitalized for Severe Acute Respiratory Syndrome Coronavirus 2 in Paris Hospitals, France: A Cohort Study on 5,795 Patients.

Obesity (Silver Spring) 2020 12 6;28(12):2282-2289. Epub 2020 Nov 6.

Université de Paris, Paris, France.

Objective: Preliminary data from different cohorts of small sample size or with short follow-up indicate poorer prognosis in people with obesity compared with other patients. This study aims to precisely describe the strength of association between obesity in patients hospitalized with coronavirus disease 2019 (COVID-19) and mortality and to clarify the risk according to usual cardiometabolic risk factors in a large cohort.

Methods: This is a prospective cohort study including 5,795 patients aged 18 to 79 years hospitalized from February 1 to April 30, 2020, in the Paris area, with confirmed infection by severe acute respiratory syndrome coronavirus 2. Adjusted regression models were used to estimate the odds ratios (ORs) and 95% CIs for the mortality rate at 30 days across BMI classes, without and with imputation for missing BMI values.

Results: Eight hundred ninety-one deaths had occurred at 30 days. Mortality was significantly raised in people with obesity, with the following ORs for BMI of 30 to 35 kg/m , 35 to 40 kg/m , and >40 kg/m : 1.89 (95% CI: 1.45-2.47), 2.79 (95% CI: 1.95-3.97), and 2.55 (95% CI: 1.62-3.95), respectively (18.5-25 kg/m was used as the reference class). This increase holds for all age classes.

Conclusions: Obesity doubles mortality in patients hospitalized with COVID-19.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/oby.23014DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7461006PMC
December 2020

Clinical and Pharmacological Aspects of Hospital-Acquired Acute Kidney Injuries Outside the Intensive Care Unit: A Phenome-Wide Association Study.

Kidney Dis (Basel) 2019 Oct 6;5(4):272-280. Epub 2019 Aug 6.

Assistance Publique Hôpitaux de Paris, Paris, France.

Introduction: Acute kidney injury (AKI) occurring in the hospital in noncritically ill patients involves a broad spectrum of clinical conditions and medical scenarios that are better appreciated by systematic association studies.

Methods: We extracted all diagnoses and drug prescriptions from an i2b2 clinical data warehouse for patients who stayed in an academic hospital between 2013 and 2017, and had at least two plasma creatinine measurements performed during the first week of their stay, and analyzed the association between AKI occurring outside the intensive care unit (ICU), as identified using the AKIN classification criteria, and International Classification of Diseases (ICD)-10 diagnosis codes and drug categories.

Results: 16,662 hospital stays for unique individuals were extracted. The prevalence of AKI outside the ICU was 8%, with a distribution of frequencies that greatly varied according to the departments. 4% of patients with AKI died during their hospital stay (OR 6.17, 95% CI [2.59-17.9]). ICD-10 diagnosis codes were related to infections, kidney cancer, heart failure, respiratory failure, and chronic kidney disease. Drugs targeting the renin angiotensin system and loop diuretics had the larger size effect on AKI. The ICD-10 code N17/"Acute kidney failure" was recorded in average in only 16% of the cases with AKI, and its frequency ranged from 0 to 80%, according to the hospital department; the lack of encoding did not impact mortality.

Conclusion: A systematic search for the associations of AKI with prescribed drugs and medical diagnosis using a phenome-wide approach allows to describe in depth the epidemiology of AKI outside the ICU.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1159/000501432DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC6872991PMC
October 2019

Conditional prescriptions of oral antihypertensive drugs for the management of hypertension urgencies in the inpatient setting: An observational study.

J Clin Pharm Ther 2020 Apr 28;45(2):282-289. Epub 2019 Sep 28.

Pharmacy Department, Hôpital Européen Georges Pompidou, Assistance Publique-Hôpitaux de Paris, Paris, France.

What Is Known And Objectives: The management of hypertension urgencies during hospitalization may generally not necessitate urgent care. However, physicians frequently prescribe 'as needed' antihypertensive drugs for which administration is triggered by blood pressure thresholds. The lack of rationale for this hospital practice led us to study oral conditional antihypertensive (OCA) prescriptions. We aimed to estimate the prevalence of OCA prescriptions and to establish their characteristics.

Methods: In our institution, prescriptions are computerized. The study was retrospectively performed using a hospital clinical data warehouse over a 5-year period.

Results And Discussion: The prevalence of OCA prescriptions was 6.9% among subjects treated with an antihypertensive drug. The median duration of these prescriptions was 4 days, until the day of the patient discharge in 78.8% stays. The calcium channel inhibitors were the main (79.9%) pharmacological class prescribed, with mostly prescriptions of nicardipine. OCA prescriptions were associated with another antihypertensive medication in 58.8% of the prescriptions; for 19.3%, it was a medication belonging to the same pharmacological class than the OCA drug prescribed. Regarding the computerized drafting, 39.6% of the conditional prescriptions were considered uninterpretable. At least one administration by nurses concerned 65.1% of the OCA prescriptions. The mean SBP and DBP before the initiation of an OCA drug was 142.9 ± 28.2 and 75.8 ± 24.5 mm Hg, respectively, relative to 143.0 ± 24.9 and 77.6 ± 19.9 mm Hg after the initiation (P = .8 for SBP and P = .06 for DBP).

What Is New And Conclusion: The originality of this study lies in the use of a clinical data warehouse to evaluate OCA prescriptions in hospital. These prescriptions are current, often uninterpretable and mostly ordered until patient discharge. Such drug orders could be associated with an increased risk of iatrogenic events and/or administration errors. This underlies the need for developing decision support tools and computerized protocols to manage hypertension urgencies.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1111/jcpt.13059DOI Listing
April 2020

Trajectories of Biological Values and Vital Parameters: An Observational Cohort Study of Adult Patients with Sickle Cell Disease Hospitalized for a Non-Complicated Vaso-Occlusive Crisis.

J Clin Med 2019 Sep 19;8(9). Epub 2019 Sep 19.

Internal Medicine Department, Sickle Cell Referral Center, Georges Pompidou European Hospital, AP-HP, 75015 Paris, France.

Hospital admission of patients with sickle-cell disease (SCD) presenting with a vaso-occlusive crisis (VOC) can be justified by pain refractory to usual outpatient care and/or the occurrence of a complication. Yet, the trajectories of vital parameters and standard biomarkers throughout a non-complicated VOC has not been established. In this observational cohort study, we describe the course of routine parameters throughout 329 hospital stays for non-complicated VOC. We used a new spline-based approach to study and visualize non-specific time-dependent variables extracted from the hospital clinical data warehouse. We identified distinct trends during the VOC for hemoglobin level, leukocytes count, C-Reactive Protein (CRP) level and temperature. Hemoglobin decreased after admission and rarely returned to steady state levels before discharge. White blood cell counts were elevated at admission before immediately decreasing, whereas eosinophils increased slowly throughout the first five days of the stay. In over 95% of non-complicated VOC-related stays, the CRP value was below 100 mg/L within the first day following admission and above normal after 48 hours, and the temperature was below 38 °C throughout the entire stay. Knowing the typical trajectories of these routine parameters during non-complicated VOC may urge the clinicians to be more vigilant in case of deviation from these patterns.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3390/jcm8091502DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC6780589PMC
September 2019

Integration of elicited expert information via a power prior in Bayesian variable selection: Application to colon cancer data.

Stat Methods Med Res 2020 02 9;29(2):541-567. Epub 2019 Apr 9.

INSERM U1138, University Paris Descartes, Sorbonne University, Paris, France.

Background: Building tools to support personalized medicine needs to model medical decision-making. For this purpose, both expert and real world data provide a rich source of information. Currently, machine learning techniques are developing to select relevant variables for decision-making. Rather than using data-driven analysis alone, eliciting prior information from physicians related to their medical decision-making processes can be useful in variable selection. Our framework is electronic health records data on repeated dose adjustment of Irinotecan for the treatment of metastatic colorectal cancer. We propose a method that incorporates elicited expert weights associated with variables involved in dose reduction decisions into the Stochastic Search Variable Selection (SSVS), a Bayesian variable selection method, by using a power prior.

Methods: Clinician experts were first asked to provide numerical clinical relevance weights to express their beliefs about the importance of each variable in their medical decision making. Then, we modeled the link between repeated dose reduction, patient characteristics, and toxicities by assuming a logistic mixed-effects model. Simulated data were generated based on the elicited weights and combined with the observed dose reduction data via a power prior. We compared the Bayesian power prior-based SSVS performance to the usual SSVS in our case study, including a sensitivity analysis using the power prior parameter.

Results: The selected variables differ when using only expert knowledge, only the usual SSVS, or combining both. Our method enables one to select rare variables that may be missed using only the observed data and to discard variables that appear to be relevant based on the data but not relevant from the expert perspective.

Conclusion: We introduce an innovative Bayesian variable selection method that adaptively combines elicited expert information and real world data. The method selects a set of variables relevant to model medical decision process.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1177/0962280219841082DOI Listing
February 2020

Comparison of methods for early-readmission prediction in a high-dimensional heterogeneous covariates and time-to-event outcome framework.

BMC Med Res Methodol 2019 03 6;19(1):50. Epub 2019 Mar 6.

Assistance Publique-Hôpitaux de Paris, Biomedical Informatics and Public Health Department, European Georges Pompidou Hospital, 20 Rue Leblanc, Paris, 75015, France.

Background: Choosing the most performing method in terms of outcome prediction or variables selection is a recurring problem in prognosis studies, leading to many publications on methods comparison. But some aspects have received little attention. First, most comparison studies treat prediction performance and variable selection aspects separately. Second, methods are either compared within a binary outcome setting (where we want to predict whether the readmission will occur within an arbitrarily chosen delay or not) or within a survival analysis setting (where the outcomes are directly the censored times), but not both. In this paper, we propose a comparison methodology to weight up those different settings both in terms of prediction and variables selection, while incorporating advanced machine learning strategies.

Methods: Using a high-dimensional case study on a sickle-cell disease (SCD) cohort, we compare 8 statistical methods. In the binary outcome setting, we consider logistic regression (LR), support vector machine (SVM), random forest (RF), gradient boosting (GB) and neural network (NN); while on the survival analysis setting, we consider the Cox Proportional Hazards (PH), the CURE and the C-mix models. We also propose a method using Gaussian Processes to extract meaningfull structured covariates from longitudinal data.

Results: Among all assessed statistical methods, the survival analysis ones obtain the best results. In particular the C-mix model yields the better performances in both the two considered settings (AUC =0.94 in the binary outcome setting), as well as interesting interpretation aspects. There is some consistency in selected covariates across methods within a setting, but not much across the two settings.

Conclusions: It appears that learning withing the survival analysis setting first (so using all the temporal information), and then going back to a binary prediction using the survival estimates gives significantly better prediction performances than the ones obtained by models trained "directly" within the binary outcome setting.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s12874-019-0673-4DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC6404305PMC
March 2019
-->