The future of digital health with federated learning

Data-driven machine learning (ML) has emerged as a promising approach for building accurate and robust statistical models from medical data, which is collected in huge volumes by modern healthcare systems. Existing medical data is not fully exploited by ML primarily because it sits in data silos and privacy concerns restrict access to this data. However, without access to sufficient data, ML will be prevented from reaching its full potential and, ultimately, from making the transition from research to clinical practice. This paper considers key factors contributing to this issue, explores how federated learning (FL) may provide a solution for the future of digital health and highlights the challenges and considerations that need to be addressed.

A Deep Learning Architecture for Temporal Sleep Stage Classification Using Multivariate and Multimodal Time Series

Sleep stage classification constitutes an important preliminary exam in the diagnosis of sleep disorders. It is traditionally performed by a sleep expert who assigns to each 30 s of the signal of a sleep stage, based on the visual inspection of signals such as electroencephalograms (EEGs), electrooculograms (EOGs), electrocardiograms, and electromyograms (EMGs). We introduce here the first deep learning approach for sleep stage classification that learns end-to-end without computing spectrograms or extracting handcrafted features, that exploits all multivariate and multimodal polysomnography (PSG) signals (EEG, EMG, and EOG), and that can exploit the temporal context of each 30-s window of data. For each modality, the first layer learns linear spatial filters that exploit the array of sensors to increase the signal-to-noise ratio, and the last layer feeds the learnt representation to a softmax classifier. Our model is compared to alternative automatic approaches based on convolutional networks or decisions trees. Results obtained on 61 publicly available PSG records with up to 20 EEG channels demonstrate that our network architecture yields the state-of-the-art performance. Our study reveals a number of insights on the spatiotemporal distribution of the signal of interest: a good tradeoff for optimal classification performance measured with balanced accuracy is to use 6 EEG with 2 EOG (left and right) and 3 EMG chin channels. Also exploiting 1 min of data before and after each data segment offers the strongest improvement when a limited number of channels are available. As sleep experts, our system exploits the multivariate and multimodal nature of PSG signals in order to deliver the state-of-the-art classification performance with a small computational cost.

Performance of an Ambulatory Dry-EEG Device for Auditory Closed-Loop Stimulation of Sleep Slow Oscillations in the Home Environment

Recent research has shown that auditory closed-loop stimulation can enhance sleep slow oscillations (SO) to improve N3 sleep quality and cognition. Previous studies have been conducted in lab environments. The present study aimed to validate and assess the performance of a novel ambulatory wireless dry-EEG device (WDD), for auditory closed-loop stimulation of SO during N3 sleep at home. The performance of the WDD to detect N3 sleep automatically and to send auditory closed-loop stimulation on SO were tested on 20 young healthy subjects who slept with both the WDD and a miniaturized polysomnography (part 1) in both stimulated and sham nights within a double blind, randomized and crossover design. The effects of auditory closed-loop stimulation on delta power increase were assessed after one and 10 nights of stimulation on an observational pilot study in the home environment including 90 middle-aged subjects (part 2).The first part, aimed at assessing the quality of the WDD as compared to a polysomnograph, showed that the sensitivity and specificity to automatically detect N3 sleep in real-time were 0.70 and 0.90, respectively. The stimulation accuracy of the SO ascending-phase targeting was 45 ± 52°. The second part of the study, conducted in the home environment, showed that the stimulation protocol induced an increase of 43.9% of delta power in the 4 s window following the first stimulation (including evoked potentials and SO entrainment effect). The increase of SO response to auditory stimulation remained at the same level after 10 consecutive nights. The WDD shows good performances to automatically detect in real-time N3 sleep and to send auditory closed-loop stimulation on SO accurately. These stimulation increased the SO amplitude during N3 sleep without any adaptation effect after 10 consecutive nights. This tool provides new perspectives to figure out novel sleep EEG biomarkers in longitudinal studies and can be interesting to conduct broad studies on the effects of auditory stimulation during sleep.

Clinical pharmacokinetics during continuous haemofiltration

Continuous haemofiltration is an extracorporeal technique that is increasingly used to remove fluid, electrolytes, and other waste products from the blood supply of critically ill patients with acute renal failure. Continuous arteriovenous haemofiltration (CAVH), where the blood exits the body from an artery and re-enters through a vein, is widely used. Continuous venovenous haemofiltration (CVVH), where blood both exits and enters through a vein by way of a mechanical pump, avoids problems that result from the variable ultrafiltration rate found during CAVH. Continuous arteriovenous or venovenous haemodiafiltration (CAVHD or CVVHD) combine continuous haemofiltration and haemodialysis. All methods involve ultrafiltration of the patient's blood through a filter that is highly permeable to water and small molecules. Drug elimination by haemofiltration depends mainly on the rate of ultrafiltration, the drug protein binding and the sieving coefficient of the membrane. Because patients undergoing continuous haemofiltration have impaired renal function, dosage reduction is often recommended so that adverse drug reactions are avoided. In contrast, if drug removal by haemofiltration is significant, dosage supplementation may be required to ensure therapeutic efficacy of the drug. Therefore, knowledge of the impact of continuous haemofiltration on drug elimination and the pharmacokinetic profile of drugs is essential to good clinical management. The currently available information on the clinical pharmacokinetic aspects of drug therapy during continuous haemofiltration are summarised. Drugs commonly associated with haemofiltration therapy are tabulated with updated pharmacokinetics and drug-monitoring information.

Placental leptin receptor isoforms in normal and pathological pregnancies

Alternate mRNA splicing of human leptin receptor generates four membrane isoforms with different C-terminal sequences. They differ by the length of their intracellular domain which include specific motifs crucial for the specificity of leptin signalling. As a step towards functional studies, we have characterized leptin receptors in human placenta from normal pregnancies and pregnancies associated with diabetes and pre-eclampsia. Leptin and leptin receptors were visualized by immunohistochemistry of placentas obtained from first and third trimester pregnancies. Antibodies against N and C-terminal epitopes showed signals in the apical membrane of the syncytiotrophoblast in early and term placental villi as well as in JAr and BeWo derived trophoblast cells. In addition, a distinct isoform recognized by its extracellular juxtamembrane epitope was exclusively localized in cytotrophoblast cells and likely stains the soluble receptor. At contrast with the transmembrane receptors, the expression of this isoform is increased in placentas of pre-eclamptic and diabetic women which synthesize more leptin than placenta from uncomplicated pregnancy. These data demonstrate that short and long transmembrane leptin receptors are expressed in the trophoblast and indicate that leptin synthetized within the placenta can act locally through both receptor isoforms. Being also accessible to leptin from maternal origin, these transmembrane receptors may signal differently in pregnancy with normal and increased leptin production. The co-localization of leptin and the soluble receptor isoform suggests that this isoform serves for modulating maternal free leptin levels through modification of leptin binding capacities.

Federated learning for predicting histological response to neoadjuvant chemotherapy in triple-negative breast cancer

Triple-negative breast cancer (TNBC) is a rare cancer, characterized by high metastatic potential and poor prognosis, and has limited treatment options. The current standard of care in nonmetastatic settings is neoadjuvant chemotherapy (NACT), but treatment efficacy varies substantially across patients. This heterogeneity is still poorly understood, partly due to the paucity of curated TNBC data. Here we investigate the use of machine learning (ML) leveraging whole-slide images and clinical information to predict, at diagnosis, the histological response to NACT for early TNBC women patients. To overcome the biases of small-scale studies while respecting data privacy, we conducted a multicentric TNBC study using federated learning, in which patient data remain secured behind hospitals' firewalls. We show that local ML models relying on whole-slide images can predict response to NACT but that collaborative training of ML models further improves performance, on par with the best current approaches in which ML models are trained using time-consuming expert annotations. Our ML model is interpretable and is sensitive to specific histological patterns. This proof of concept study, in which federated learning is applied to real-world datasets, paves the way for future biomarker discovery using unprecedentedly large datasets.

Endotracheal and aerosol administrations of ceftazidime in patients with nosocomial pneumonia: pharmacokinetics and absolute bioavailability

Pharmacokinetic studies on ceftazidime, an aminothiazole cephalosporin with a wide spectrum of antibacterial activity, including activity against Pseudomonas aeruginosa, were performed in patients with nosocomial pneumonia. The concentration-time profiles of ceftazidime in plasma, urine, and bronchial secretions of 12 patients were investigated after intravenous (i.v.) (n = 12), endotracheal (n = 10), and aerosol (n = 5) administrations. In all cases a 1-g dose was administered. Concentrations of drug in all samples were assayed by high-performance liquid chromatography with UV detection. The elimination of the drug from the blood followed a biexponential (i.v. administration) or a monoexponential (endotracheal and aerosol administrations) decay, with an elimination half-life of 6 h and a total body clearance of 4.2 liters/h. The apparent volume of distribution was 0.36 liter/kg of body weight. Renal clearance of the drug accounted for 58% of the total clearance; 66% +/- 17.7%, 33.5% +/- 17.3%, and 6.59% +/- 3.45% of the administered dose were eliminated in urine as parent drug after i.v., endotracheal, and aerosol administrations, respectively. The absolute bioavailabilities were 0.47 and 0.08 for endotracheal and aerosol administrations, respectively. Very high concentrations were found in bronchial secretions after local administration. The MICs for 90% of the most important pathogens responsible for nosocomial infections were exceeded by concentrations in bronchial secretion for up to 12 h after i.v. infusion and for up to 24 h after endotracheal and aerosol administrations.

Doxorubicin and doxorubicinol: intra- and inter-individual variations of pharmacokinetic parameters

Doxorubicin was given by short i.v. infusion (dose range 25-72 mg/m2) to 18 patients who underwent three to seven successive courses of chemotherapy (total, 57 courses). Plasma levels of doxorubicin and its major metabolite doxorubicinol were determined by high-performance liquid chromatography over a 48-h period after the infusion. Pharmacokinetic parameters for the parent drug and its metabolite were calculated for each course of treatment. The results show considerable inter- and intraindividual variations for most parameters. The coefficients of variation (CV) ranged from 37% to 93% (inter-individual) and from 6% to 59% (intra-individual). Nevertheless, we observed a good stability over successive courses for terminal half-life in six patients (CV, 6%-25%) and for clearance and AUC in four subjects (CV, 10%-22%). The ratio of the AUCs for doxorubicinol: doxorubicin averaged 0.514. The pharmacokinetic pattern of doxorubicinol was biphasic in plasma of the majority of patients. We propose a model for curve-fitting of these metabolite plasma concentrations that is based on two successive releases of the compound in the plasma compartment, separated by a lag time.

Steady-state pharmacokinetics of ciprofloxacin in plasma from patients with nosocomial pneumonia: penetration of the bronchial mucosa

A 30-min intravenous (i.v.) infusion of 200 mg of ciprofloxacin was administered twice daily to 12 patients with nosocomial pneumonia scheduled to undergo diagnostic fiberoptic bronchoscopy. The pharmacokinetics of ciprofloxacin were examined at the presumed steady state after 5 days of treatment. Eleven successive plasma samples were collected in the interval from 0 to 12 h after administration, and bronchial mucosa samples were taken 2 h after administration. Concentrations of drug in all samples were assayed by high-performance liquid chromatography with fluorimetric detection. The results showed that the kinetics in plasma did not differ from those determined previously in healthy volunteers. The mean concentrations in plasma peaked at 4.94 +/- 2.90 mg/liter at the end of infusion. The terminal half-life was 4.95 +/- 2.81 h, and the mean residence time 6.13 +/- 3.17 h. A large volume of distribution was calculated: 2.59 +/- 1.43 liters/kg. Mean total body clearance was 23.3 +/- 10.1 liters/h. The concentrations in bronchial mucosa reached 21.6 +/- 5.63 micrograms/g 2 h after drug intake. The tissue-versus-plasma concentration ratios ranged from 10.1 to 26.3 (mean value, 16.9 +/- 5.43). After 6 to 12 days of i.v. treatment, four patients were switched to oral ciprofloxacin. We propose a model for the simultaneous fit of the concentration-time curves obtained after i.v. infusion and oral dosing. The concentrations in tissue observed in this study were in excess of the MICs for bacteria considered to be susceptible to ciprofloxacin.

Immunocytological evidence for hematopoiesis in the early human placenta

Hematopoiesis has previously been observed in the human yolk sac, in placental villi and in the embryonic aorta. Here, our immunocytological study at 24 and 35 days showed packed erythroblasts in the placental vessels, mitotic figures and anti-Ki-67 reactions within these cells. Morphologically, the erythroblasts and vessels were similar to those found in the yolk sac during primitive hematopoiesis. In addition, numerous extravascular erythroblasts were found in the villous core. Positive reactions were obtained in erythroblasts using antibodies against glycophorin-A, GATA-2 and C-kit that characterize the hematopoietic cells. However, erythroblasts did not react with anti-CD34 and anti-CD45. In this respect, they differ from the hematopoietic cell clusters observed in the aorta of the human embryo. The staining for glycophorin-A was maintained in erythroblasts at 6-7 weeks and 12-14 weeks. Anti-GATA-2 reaction was decreased in erythroblasts and appeared in the perivillous cytotrophoblast. Anti-C-kit signal was detected in endothelial cells at 6-7 weeks and switched to stromal and perivascular cells at 12-14 weeks. By term, anti-GATA-2 staining was still present in the trophoblast and appeared in vessels while anti-C-kit was negative. For the leukocytes marker CD15, a staining was found in the endothelium at 35 days, 6-7 and 12-14 weeks and in leukocytes at term. CD45 antibody decorated the leukocytes at 12-14 weeks and at term. Erythroblasts undergo a primitive hematopoiesis in the early placental vessels that may be of value for the embryo in a period of low oxygen environment.

MELLODDY: Cross-pharma Federated Learning at Unprecedented Scale Unlocks Benefits in QSAR without Compromising Proprietary Information

Federated multipartner machine learning has been touted as an appealing and efficient method to increase the effective training data volume and thereby the predictivity of models, particularly when the generation of training data is resource-intensive. In the landmark MELLODDY project, indeed, each of ten pharmaceutical companies realized aggregated improvements on its own classification or regression models through federated learning. To this end, they leveraged a novel implementation extending multitask learning across partners, on a platform audited for privacy and security. The experiments involved an unprecedented cross-pharma data set of 2.6+ billion confidential experimental activity data points, documenting 21+ million physical small molecules and 40+ thousand assays in on-target and secondary pharmacodynamics and pharmacokinetics. Appropriate complementary metrics were developed to evaluate the predictive performance in the federated setting. In addition to predictive performance increases in labeled space, the results point toward an extended applicability domain in federated learning. Increases in collective training data volume, including by means of auxiliary data resulting from single concentration high-throughput and imaging assays, continued to boost predictive performance, albeit with a saturating return. Markedly higher improvements were observed for the pharmacokinetics and safety panel assay-based task subsets.

Methotrexate concentrations in synovial membrane and trabecular and cortical bone in rheumatoid arthritis patients

OBJECTIVE

To determine methotrexate (MTX) concentrations in the synovial membrane (SM) and cortical and trabecular bone of rheumatoid arthritis (RA) patients.

METHODS

Ten RA patients (9 women, 1 man; mean +/- SD age 49.2 +/- 10.6, mean disease duration 13.2 +/- 9.9 years) undergoing surgical procedures for rheumatoid articular lesions participated in this study. Mean +/- SD MTX treatment duration was 26.4 +/- 21.3 months. The day preceding surgery, 10 mg of MTX was administered intramuscularly. During surgery, a mean +/- SD of 19.7 +/- 2.6 hours after MTX administration, SM, bone fragments, and blood were collected simultaneously. MTX was assayed by fluorescence polarization immunoassay in plasma and tissues.

RESULTS

The mean +/- SD plasma concentration was 0.0252 +/- 0.01 nmoles/ml at the time of tissue sampling. The mean MTX concentration in SM was 0.285 +/- 0.159 nmoles/gm. The mean MTX concentrations in trabecular and cortical bone were 0.292 +/- 0.164 and 0.286 +/- 0.126 nmoles/gm, respectively.

CONCLUSION

After intramuscular administration, high MTX concentrations are found in SM and cortical and trabecular bone of RA patients.

Multiple-dose pharmacokinetics of amikacin and ceftazidime in critically ill patients with septic multiple-organ failure during intermittent hemofiltration

The pharmacokinetic parameters of amikacin and ceftazidime were assessed in four patients undergoing hemofiltration for septic shock. The parameters were assessed during hemofiltration and in the interim period. The concentration-time profiles of these two drugs in plasma, urine, and ultrafiltrate were investigated after intravenous perfusion (30 min). In all cases a 1-g dose of ceftazidime was administered; for amikacin, the dosage regimen was adjusted according to the patient's amikacin levels (250 to 750 mg). Concentrations of drug in all samples were assayed by high-performance liquid chromatography with UV detection for ceftazidime and by enzyme multiplied immunoassay for amikacin. The elimination half-life (t1/2) and the total clearance of amikacin ranged from 31.1 to 138.2 h and from 5.4 to 8.9 ml/min, respectively, during the interhemofiltration period in anuric patients. Hemofiltration substantially decreased the t1/2 (3.5 +/- 0.49 h) and increased the total clearance (89.5 +/- 11.8 ml/min). The hemofiltration clearance of amikacin represented 71% of the total clearance, and the hemofiltration process removed, on average, 60% of the dose. During hemofiltration, the elimination t1/2 of ceftazidime (2.8 +/- 0.69 h) was greatly reduced and the total clearance increased (74.2 +/- 11.2 ml/min) compared with those in the interhemofiltration period (9 to 43.7 h and 7.4 to 16.8 ml/min, respectively). About 55% of the administered dose was recovered in the filtrate, and the hemofiltration clearance of ceftazidime was 46 +/- 14.3 ml/min. A redistribution phenomenon (rebound) in the amikacin and ceftazidime concentrations in plasma (35 and 28%, respectively) was reported after hemofiltration in two patients. The MICs for 90% of the most important pathogens were exceeded by the concentrations of the two drugs in plasma during the whole treatment of these patients.

A reproducible, simple and sensitive HPLC assay for determination of ofloxacin in plasma and lung tissue. Application in pharmacokinetic studies

A high-performance liquid chromatographic method with fluorometric detection was developed for the analysis of ofloxacin in plasma and lung tissue. The detection was performed at 280 nm for excitation and 500 nm for emission. The procedure involves the addition of an internal standard followed by treatment of the samples with acetonitrile and dichloromethane. The proposed technique is reproducible, selective, reliable and sensitive. Linear detector response was observed for the calibration curve standards in the range of 0.1-5 micrograms ml-1 for plasma and 0.025-2.5 micrograms g-1 for lung tissue. The limit of quantitation is 5 ng ml-1 or 5 ng g-1. The accuracy of the method is good; that is, the relative error is < 10%. This method was applied to the pharmacokinetic study of ofloxacin in 24 chronic obstructive pulmonary disease patients.

High-performance liquid chromatographic determination of cis-dichlorodiammineplatinum(II) in plasma ultrafiltrate

A reproducible, simple and sensitive high-performance liquid chromatographic method was described for the quantitative analysis of cis-diamminedichloroplatinum(II) (CDDP) in ultrafiltrate plasma in the presence of nickel chloride as internal standard. CDDP and the internal standard were chelated by exchange with diethyldithiocarbamate. After derivatization, the mixture was directly injected into the column. Chromatography was performed on an Ultrasphere column and the eluent measured spectrophotometrically at 260 nm for CDDP and at 250 nm for the internal standard. The peak area ratio of CDDP to the internal standard varied linearly with concentration over the range 0.05-10 micrograms ml-1. Precision and reproducibility were both excellent and the limit of quantification was 0.03 micrograms ml-1 using only 0.5 ml of ultrafiltrate. The present method, without extraction, should be entirely automated. This assay may be suitable for therapeutic drug monitoring in patients receiving CDDP.

Identification of patients with impaired hepatic drug metabolism using a limited sampling procedure for estimation of phenazone (antipyrine) pharmacokinetic parameters

Phenazone (antipyrine) 1g was given by short intravenous infusion to 62 study participants (10 healthy drug-free volunteers and 52 patients with chronic liver disease). A Bayesian approach was developed to determine the individual pharmacokinetic parameters of phenazone. Statistical characteristics of the population pharmacokinetic parameters were first evaluated for 30 patients. When combined with 1 plasma drug concentration from members of the second group, these led to a Bayesian estimation of individual pharmacokinetic parameters for the remaining 32 individuals. Total clearance computed by Bayesian estimation was compared with maximal likelihood estimation of this parameter, the classical procedure. No statistically significant differences were found. Performance of the developed methodology was evaluated by computing bias and precision. The mean error was 0.0477 L/h. The precision of the prediction of this parameter (0.155 L/h) remained lower than the interindividual standard deviation (0.765 L/h). This procedure enables the estimation of individual pharmacokinetic parameters for phenazone. In this study, numerous laboratory tests were performed. A highly significant correlation (p < 0.001) was found between phenazone clearance and the prothrombin time, albumin, gamma-globulin, factor V, antithrombin III, fibrinogen and total bilirubin. Discriminant analysis determined that protein, alkaline phosphatase, creatininaemia and gamma-globulin had more significant discriminating power and gave better prognostic results than those seen with the Child-Pugh test.

Epidural analgesia with a bupivacaine-fentanyl mixture in obstetrics: comparison of repeated injections and continuous infusion

We compared the efficacy and side-effects of continuous infusion versus repeated injections of epidural bupivacaine-fentanyl during labour. Forty-four parturients were randomly distributed into two groups balanced for population size, morphology and parity. Analgesia was begun at the same stage of labour with a mixture of 20 ml 0.25 per cent plain bupivacaine and 2 ml (100 micrograms) fentanyl. In Group I the initial dose ranged from 8-12 ml as a function of height; an injection of the same dose was repeated immediately upon recurrence of pain. In Group II, after an initial dose of 5-7 ml, a continuous infusion of 3 ml.h-1 was begun, and continued until full dilatation. Analgesia was rated using a pain scale; effects on maternal blood pressure, respiratory rate and neonatal status were noted. Bupivacaine and fentanyl assays were carried out on maternal venous blood in 30 parturients during the course of labour. There was a longer latency to onset of analgesia in Group II (approximately five minutes), followed by a more constant degree of analgesia. This better analgesia cannot be accounted for by a difference in dosage; doses were significantly lower in Group II, despite the fact that labour was of the same duration. The course of labour, and maternal and neonatal status were comparable in the two groups. Assays showed no difference in bupivacaine blood concentrations between the two groups nor signs of drug accumulation. The constant infusion technique is advantageous since it provides a more regular degree of analgesia with lower doses than those required for patients having repeated injections.

A sensitive and reproducible HPLC assay for doxorubicin and pirarubicin

A high-performance liquid chromatographic method with spectrofluorometric detection has been developed for the analysis of doxorubicin (DOX), pirarubicin (PIRA) and their metabolite, doxorubicinol, in plasma. The detection was performed at 480 nm for excitation, and 590 nm for emission. The proposed technique is selective, reliable, and sensitive. The limit of quantification was 2 ng ml-1 for DOX and 5 ng ml-1 for PIRA. The reproducibility of the analytical method through statistical coefficients is approximately 5%. The accuracy of the method is good; the relative error is less than 5%.

Ultrastructural and immunocytochemical localization of immunoglobulin synthesis in tumor cells in Hodgkin's disease

An immunocytochemical and ultrastructural study using the Fab fragment of an anti-human Ig antibody labelled with peroxidase was carried out on affected lymph nodes from five Hodgkin's disease patients. The tumor cells (Reed-Sternberg cells and Hodgkin cells) showed an exclusively hyaloplasmic granular staining. By comparing these grains with ribisome staining. By comparing these grains with ribosome staining of the endoplasmic reticulum of plasma cells it could be suggested that they are free risobomes. This ribosomal Ig synthesis is a major argument for the B lymphocyte nature of Reed-Sternberg and Hodgkin cells. The total absence of vacuole staining allows us to conclude that these cells are not histiocytic or macrophage derivatives.

Culture of endothelial cells from human placental microvessels

Endothelial cells are known to participate in angiogenesis, adaptation of vascular tonus and maintenance of blood fluidity in the microcirculation. To investigate these functions in the placenta, we devised a method of isolation and culture of endothelial cells from villous microvessels. In primary culture, these intraplacental endothelial cells exhibited many features observed in microvascular endothelium from other organs: spindle-shape, rosette associations, circular arrangements and confluence. In contrast to the confluent endothelial cells derived from the umbilical vein, cells from microvessels did not form a cobblestone network. After trypsin digestion of microvessels, magnetic microbeads coated with S-Endol immunoglobulin, antithrombomodulin and Ulex europaeus-I lectins were tested for sorting endothelial cells. Only the microbeads coated with antithrombomodulin allowed a suitable magnetic cell separation after trypsinization. By contrast, the microbeads coated with each of these antibodies or with lectins attached to confluent cells from the second passage. The microbeads detached from the cells at different rates. Their examination by scanning microscopy indicates that a portion of these microbeads was phagocytosed. Microvascular endothelial cells from the second passage were intensively stained by the anti-von Willebrand reaction and only weakly by the anti-smooth muscle alpha-actin reaction. They incorporated acetylated-low density lipoproteins coupled to a fluorescent probe. The positive reactions against the anti-von Willebrand factor and the uptake of the fluorescent acetylated-low density lipoproteins were modified after eight passages.

Pefloxacin clinical pharmacokinetics

Pefloxacin has a broad spectrum of activity against a great number of Gram-negative and Gram-positive bacteria. It is also capable of penetration into cells, yielding high tissue:serum ratios, with implications for the treatment of infections caused by intracellular pathogens. Pefloxacin is well absorbed from the gastrointestinal tract. Its elimination half-life ranges from 6.2 to 12.4 hours. After repeated administration, a major change in pharmacokinetic parameters is observed. Pharmacokinetic parameters are minimally altered or not altered in patients with impaired renal function. Altered plasma pharmacokinetics in patients with liver insufficiency and in elderly patients are observed, so dosage adjustments are necessary. In addition, pefloxacin interacts with a number of other compounds at hepatic (e.g. theophylline and cimetidine) and gastrointestinal (e.g. antacids) sites. With the exception of saliva, cerebrospinal fluid, aqueous humor, vitreous fluid and amniotic fluid, body fluid concentrations reach plasma concentrations. Studies on tissue penetration show that concentrations exceeding plasma concentrations are obtained in most tissues. The highest tissue:plasma concentration ratios are achieved in lung and kidney, whereas concentrations in fat are considerably lower than those in plasma.

Bayesian estimation of doxorubicin pharmacokinetic parameters

Doxorubicin was given by brief i.v. infusion (doses ranging from 25 to 72 mg/m2) to 28 patients for 2-7 successive courses of chemotherapy (68 courses studied in all). A Bayesian approach was developed to determine the individual pharmacokinetic parameters of doxorubicin. Statistical characteristics of the population pharmacokinetic parameters were first evaluated for 19 patients and a total of 30 courses, which, when combined with 4 individual plasma concentrations of drug, led to a Bayesian estimation of individual pharmacokinetic parameters for the remaining 38 courses. The estimated parameters for the elimination phase (A3/V1 and t1/2 elimination) and the residual plasma level at 48 h as computed by Bayesian estimation on this reduced sub-optimal sampling protocol were compared with a maximal likelihood estimation of these parameters. No statistically significant differences were found. Performance of the developed methodology was evaluated by computing bias and precision. The mean errors were -0.0315 x 10(-4) l-1 for A3/V1, 0.0839 h for t1/2 elimination, and -0.22 ng/ml for c(48 h). The precision of the prediction of these three parameters (0.304 x 10(-5) l-1, 3.34 h, and 0.659 ng/ml, respectively) remained lower than the interindividual standard deviation (1.42 x 10(-4) l-1, 14.9 h, and 4.54 ng/ml, respectively). This procedure enables the estimation of individual pharmacokinetic parameters for doxorubicin at minimal cost and minimal disturbance of the patient.