Preeclampsia has two phenotypes which require different treatment strategies.

The opinion on the mechanisms underlying the pathogenesis of preeclampsia still divides scientists and clinicians. This common complication of pregnancy has long been viewed as a disorder linked primarily to placental dysfunction, which is caused by abnormal trophoblast invasion, however, evidence from the previous two decades has triggered and supported a major shift in viewing preeclampsia as a condition that is caused by inherent maternal cardiovascular dysfunction, perhaps entirely independent of the placenta. In fact, abnormalities in the arterial and cardiac functions are evident from the early subclinical stages of preeclampsia and even before conception. Moving away from simply observing the peripheral blood pressure changes, studies on the central hemodynamics reveal two different mechanisms of cardiovascular dysfunction thought to be reflective of the early-onset and late-onset phenotypes of preeclampsia. More recent evidence identified that the underlying cardiovascular dysfunction in these phenotypes can be categorized according to the presence of coexisting fetal growth restriction instead of according to the gestational period at onset, the former being far more common at early gestational ages. The purpose of this review is to summarize the hemodynamic research observations for the two phenotypes of preeclampsia. We delineate the physiological hemodynamic changes that occur in normal pregnancy and those that are observed with the pathologic processes associated with preeclampsia. From this, we propose how the two phenotypes of preeclampsia could be managed to mitigate or redress the hemodynamic dysfunction, and we consider the implications for future research based on the current evidence. Maternal hemodynamic modifications throughout pregnancy can be recorded with simple-to-use, noninvasive devices in obstetrical settings, which require only basic training. This review includes a brief overview of the methodologies and techniques used to study hemodynamics and arterial function, specifically the noninvasive techniques that have been utilized in preeclampsia research.

Peri-implantation urinary hormone monitoring distinguishes between types of first-trimester spontaneous pregnancy loss.

BACKGROUND: Lutenising hormone (LH) and human chorionic gonadotropin (hCG) hormone are useful biochemical markers to indicate ovulation and embryonic implantation, respectively. We explored "point-of-care" LH and hCG testing using a digital home-testing device in a cohort trying to conceive. OBJECTIVE: To determine conception and spontaneous pregnancy loss rates, and to assess whether trends in LH-hCG interval which are known to be associated with pregnancy viability could be identified with point-of-care testing. METHODS: We recruited healthy women aged 18-44 planning a pregnancy. Participants used a home monitor to track LH and hCG levels for 12 menstrual cycles or until pregnancy was conceived. Pregnancy outcomes (viable, clinical miscarriage, or biochemical pregnancy loss) were recorded. Monitor data were analysed by a statistician blinded to pregnancy outcome. RESULTS: From 387 recruits, there were 290 pregnancies with known outcomes within study timeline. Adequate monitor data for analysis were available for 150 conceptive cycles. Overall spontaneous first-trimester pregnancy loss rate was 30% with clinically recognised miscarriage rate of 17%. The difference to LH-hCG interval median had wider spread for biochemical losses (0.5-8.5 days) compared with clinical miscarriage (0-5 days) and viable pregnancies (0-6 days). Fixed effect hCG profile change distinguished between pregnancy outcomes from as early as day-2 post-hCG rise from baseline. CONCLUSIONS: The risk of first-trimester spontaneous pregnancy loss in our prospective cohort is comparable to studies utilising daily urinary hCG collection and laboratory assays. A wider LH-hCG interval range is associated with biochemical pregnancy loss and may relate to late or early implantation. Although early hCG changes discriminate between pregnancies that will miscarry from viable pregnancies, this point-of-care testing model is not sufficiently developed to be predictive.

Early and late preeclampsia are characterized by high cardiac output, but in the presence of fetal growth restriction, cardiac output is low: insights from a prospective study.

BACKGROUND: Preeclampsia and fetal growth restriction are considered to be placentally mediated disorders. The clinical manifestations are widely held to relate to gestation age at onset with early- and late-onset preeclampsia considered to be phenotypically distinct. Recent studies have reported conflicting findings in relation to cardiovascular function, and in particular cardiac output, in preeclampsia and fetal growth restriction. OBJECTIVE: We conducted this study to examine the possible relation between cardiac output and peripheral vascular resistance in preeclampsia and fetal growth restriction. STUDY DESIGN: We investigated maternal cardiovascular function in relation to clinical subtype in 45 pathological pregnancies (14 preeclampsia only, 16 fetal growth restriction only, 15 preeclampsia and fetal growth restriction) and compared these with 107 healthy person observations. Cardiac output was the primary outcome measure and was assessed using an inert gas-rebreathing method (Innocor), from which peripheral vascular resistance was derived; arterial function was assessed by Vicorder, a cuff-based oscillometric device. Cardiovascular parameters were normalized for gestational age in relation to healthy pregnancies using Z scores, thus allowing for comparison across the gestational range of 24-40 weeks. RESULTS: Compared with healthy control pregnancies, women with preeclampsia had higher cardiac output Z scores (1.87 ± 1.35; P = .0001) and lower peripheral vascular resistance Z scores (-0.76 ± 0.89; P = .025); those with fetal growth restriction had higher peripheral vascular resistance Z scores (0.57 ± 1.18; P = .04) and those with both preeclampsia and fetal growth restriction had lower cardiac output Z scores (-0.80 ± 1.3 P = .007) and higher peripheral vascular resistance Z scores (2.16 ± 1.96; P = .0001). These changes were not related to gestational age of onset. All those affected by preeclampsia and/or fetal growth restriction had abnormally raised augmentation index and pulse wave velocity. Furthermore, in preeclampsia, low cardiac output was associated with low birthweight and high cardiac output with high birthweight (r = 0.42, P = .03). CONCLUSION: Preeclampsia is associated with high cardiac output, but if preeclampsia presents with fetal growth restriction, the opposite is true; both conditions are nevertheless defined by hypertension. Fetal growth restriction without preeclampsia is associated with high peripheral vascular resistance. Although early and late gestation preeclampsias are considered to be different diseases, we show that the hemodynamic characteristics of preeclampsia were unrelated to gestational age at onset but were strongly associated with the presence or absence of fetal growth restriction. Fetal growth restriction more commonly coexists with preeclampsia at early gestation, thus explaining the conflicting results of previous studies. Furthermore, antihypertensive agents act by reducing cardiac output or peripheral vascular resistance and are administered without reference to cardiovascular function in preeclampsia. The underlying pathology (preeclampsia, fetal growth restriction, preeclampsia and fetal growth restriction) defines cardiovascular phenotype, providing a rational basis for choice of therapy in which high or low cardiac output or peripheral vascular resistance is the predominant feature.

Maternal gestational diabetes is associated with genome-wide DNA methylation variation in placenta and cord blood of exposed offspring.

Exposure of a developing foetus to maternal gestational diabetes (GDM) has been shown to programme future risk of diabetes and obesity. Epigenetic variation in foetal tissue may have a mechanistic role in metabolic disease programming through interaction of the pregnancy environment with gene function. We aimed to identify genome-wide DNA methylation variation in cord blood and placenta from offspring born to mothers with and without GDM. Pregnant women of South Asian origin were studied and foetal tissues sampled at term delivery. The Illumina HumanMethylation450 BeadChip was used to assay genome-wide DNA methylation in placenta and cord blood from 27 GDM exposed and 21 unexposed offspring. We identified 1485 cord blood and 1708 placenta methylation variable positions (MVPs) achieving genome-wide significance (adjusted P-value <0.05) with methylation differences of >5%. MVPs were disproportionately located within first exons. A bioinformatic co-methylation algorithm was used to detect consistent directionality of methylation in 1000 bp window around each MVP was observed at 74% of placenta and 59% of cord blood MVPs. KEGG pathway analysis showed enrichment of pathways involved in endocytosis, MAPK signalling and extracellular triggers to intracellular metabolic processes. Replication studies should integrate genomics and transcriptomics with longitudinal sampling to elucidate stability, determine causality for translation into biomarker and prevention studies.

Towards More Reliable Confidence Estimation.

As a task that aims to assess the trustworthiness of the model's prediction output during deployment, confidence estimation has received much research attention recently, due to its importance for the safe deployment of deep models. Previous works have outlined two important characteristics that a reliable confidence estimation model should possess, i.e., the ability to perform well under label imbalance and the ability to handle various out-of-distribution data inputs. In this work, we propose a meta-learning framework that can simultaneously improve upon both characteristics in a confidence estimation model. Specifically, we first construct virtual training and testing sets with some intentionally designed distribution differences between them. Our framework then uses the constructed sets to train the confidence estimation model through a virtual training and testing scheme leading it to learn knowledge that generalizes to diverse distributions. Besides, we also incorporate our framework with a modified meta optimization rule, which converges the confidence estimator to flat meta minima. We show the effectiveness of our framework through extensive experiments on various tasks including monocular depth estimation, image classification, and semantic segmentation.

GradMDM: Adversarial Attack on Dynamic Networks.

Dynamic neural networks can greatly reduce computation redundancy without compromising accuracy by adapting their structures based on the input. In this paper, we explore the robustness of dynamic neural networks against energy-oriented attacks targeted at reducing their efficiency. Specifically, we attack dynamic models with our novel algorithm GradMDM. GradMDM is a technique that adjusts the direction and the magnitude of the gradients to effectively find a small perturbation for each input, that will activate more computational units of dynamic models during inference. We evaluate GradMDM on multiple datasets and dynamic models, where it outperforms previous energy-oriented attack techniques, significantly increasing computation complexity while reducing the perceptibility of the perturbations https://github.com/lingengfoo/GradMDM.

Cardiac output changes from prior to pregnancy to post partum using two non-invasive techniques.

OBJECTIVES: We aimed to describe cardiac output (CO) trend from prepregnancy to post partum using an inert gas rebreathing (IGR) device and compare these measurements with those obtained by a pulse waveform analysis (PWA) technique, both cross-sectionally and longitudinally. METHODS: Non-smoking healthy women, aged 18-44 years, with body mass index <35 were included in this prospective observational study. CO measurements were collected at different time points (prepregnancy, at four different gestational epochs and post partum) using IGR and PWA. A linear mixed model analysis tested whether the longitudinal change in CO differed between the techniques. Bland-Altman analysis and intraclass correlation coefficient (ICC) were used for cross-sectional and a four-quadrant plot for longitudinal comparisons. RESULTS: Of the 413 participants, 69 had a complete longitudinal assessment throughout pregnancy. In this latter cohort, the maximum CO rise was seen at 15.2 weeks with IGR (+17.5% from prepregnancy) and at 10.4 weeks with PWA (+7.7% from prepregnancy). Trends differed significantly (p=0.0093). Cross-sectional analysis was performed in the whole population of 413 women: the mean CO was 6.14 L/min and 6.38 L/min for PWA and IGR, respectively, the percentage of error was 46% and the ICC was 0.348, with similar results at all separate time points. Longitudinal concordance was 64%. CONCLUSIONS: Despite differences between devices, the maximum CO rise in healthy pregnancies is more modest and earlier than previously reported. The two methods of CO measurement do not agree closely and cannot be used interchangeably. Technique-specific reference ranges are needed before they can be applied in research and clinical settings.

An early report: a modified porphyrin-linked metronidazole targeting intracellular Porphyromonas gingivalis in cultured oral epithelial cells.

Porphyromonas gingivalis (P. gingivalis) has a strong association with the pathogenesis of periodontal disease. Recurrence of periodontal disease following therapy is attributed to numerous factors, and of growing interest is the potential problem of intracellular bacteria that are able to persist and multiply within the host cell, thereby facilitating relapse of infection. The effect of antibiotic therapy in controlling P. gingivalis is questionable. Accordingly, while metronidazole is very effective against anaerobic extracellular P. gingivalis by disrupting the DNA of anaerobic microbial cells, this antibiotic does not effectively penetrate into mammalian cells to inhibit intracellular bacteria. Therefore in the present study, a modified porphyrin-linked metronidazole adducts, developed in our laboratory, was used to kill intracellular P. gingivalis. A series of experiments were performed, including cytotoxicity assays and cellular uptake of adducts by flow cytometry coupled with live cell imaging analysis, P. gingivalis invasion and elimination assays, and the analysis of colocalization of P. gingivalis and porphyrin-linked metronidazole by confocal laser scanning microscopy. Findings indicated that P. gingivalis and porphyrin-linked metronidazole were colocalized in the cytoplasm, and this compound was able to kill P. gingivalis intracellular with a sufficient culture time. This is a novel antimicrobial approach in the elimination of P. gingivalis from the oral cavity.

Maternal death from stroke: a thirty year national retrospective review.

OBJECTIVE: In the United Kingdom (UK), the maternal mortality rate from stroke is reported at 0.3/100,000 deliveries, but only antenatal data have previously been reviewed. We hypothesise that the true rate is much higher due to a propensity for stroke occurring in the post-partum period, and that the rate will rise in parallel with trends of increasing maternal age and medical co-morbidities. Our objectives are to investigate the UK stroke mortality rate in pregnancy and the puerperium, and to examine temporal changes in fatal maternal strokes over a 30 year period. STUDY DESIGN: Retrospective review of stroke-related maternal deaths reported to the UK confidential enquiries into maternal death between 1979 and 2008, encompassing 21,514,457 maternities. In accordance with the ICD.10 classification, cases were divided into direct or indirect deaths. Late and coincidental deaths were not included in analyses. Lessons from sub-standard care associated with maternal death from stroke were collated. RESULTS: In 1979-2008 there were 347 maternal deaths from stroke: 139 cases were direct deaths, i.e. the fatal stroke was a direct result of pregnancy. The incidence of fatal stroke is relatively constant at 1.61/100,000 maternities, with a 13.9% (95% CI 12.6-15.3) proportional mortality rate. Intracranial haemorrhage was the single greatest cause of maternal death from stroke. CONCLUSION: This is the largest UK study examining the incidence of fatal maternal stroke in pregnancy and the puerperium. Our results highlight the high proportion of women who die from stroke in the puerperium. Sub-standard care featured especially in regard to management of dangerously high systolic blood pressure levels. These deaths highlight the importance of education in managing rapid-onset hypertension and superimposed coagulopathies.