Linear Effects Models Research Articles

Comparing the Responses of Osseous Versus Soft-Tissue Metastases of Renal Cell Carcinoma to Receptor Tyrosine Kinase Inhibitors and Immunotherapy

BACKGROUND: Checkpoint inhibitors and receptor tyrosine kinase inhibitors (RTKIs) have changed the standard of care for metastatic renal cell carcinoma (mRCC). Anecdotal evidence suggests these therapies may be less effective for treating bone than soft-tissue metastases. PURPOSE: We performed a retrospective review evaluating the relative clinical responses in soft-tissue and bone metastases in patients undergoing therapy using RTKIs and anti-programmed death-1 (PD-1) agents for mRCC. METHODS: Of the 2,212 patients in our institutional cancer registry with renal cell carcinoma (1997–2017), 68 (82 disease courses) were identified with measurable bone and soft-tissue metastases treated with RTKIs and/or PD-1s. Extent of metastasis was quantified at the time of therapy initiation (baseline) and at 3 months, 6 months, and 1 year. Changes in disease status were categorized as complete response, partial response, stable, mixed, or progression of disease according to RECIST v1.1 and MD Anderson criteria. These categories were further organized into “response to treatment” or “evidence of progression” to generate a generalized linear effects model with soft-tissue response as the independent variable and bone response as the dependent variable. Alpha = 0.05. RESULTS: Soft-tissue response correlated with bone response at 3 months (76 disease courses, p = 0.005) and 6 months (48 disease courses, p = 0.017). Of the patients with controlled soft-tissue disease, only 14 (19%) and 15 (32%) had progression in bone at 3 and 6 months, respectively. CONCLUSION: Contrary to anecdotal reports, osseous metastases do not appear to respond worse than soft-tissue metastases to treatment with these agents.

Factors affecting depth perception and comparison of depth perception measured by the three-rods test in monocular and binocular vision

PurposeThe purpose of this study was to explore the effects of factors affecting depth perception of moving objects using a modified three-rods test, which can be used at longer distances than the conventional one, and to compare differences in the results between binocular and monocular vision. MethodsThis study included 24 volunteers (10 women, 14 men; mean age, 35.2 years; standard deviation, 6.8 years; range, 22–56 years). We measured depth perception using a modified three-rods test under eight different conditions and investigated the factors affecting depth perception using a linear-effect model. ResultsThe results identified test distance, binocularity, masking, and direction of movement as significant factors affecting depth perception of a moving object. ConclusionsThe current study successfully determined factors affecting depth perception using the three-rods test with a moving object and the results should contribute to further clinical and social applications of the three-rods test.

Workshop 8: Beyond the farebox: Sustainable funding of public transport by better understanding service values

This Workshop was built around papers that aimed to identify ways of bridging the funding gap in public transport provision using methods other than the farebox. A number of traditional approaches such as land value capture and road pricing to provide additional sources of finance were discussed as were more general approaches such as the development of “Mobility as a Service”. The main thrust of the discussion was, however, the need to take an even broader approach recognising behavioural and governance factors that shape individual and social norms and trust in institutions. The creation of shared social values was seen as an essential precursor to successful public-private cooperation. The enhancement of efficiency in public transport provision to help reduce deficits rather than reductions in service would contribute to this. Above all the focus should be on a sustainable transport system embracing all modes to meet the needs of the population rather than a focus on sustainable financing of one element. Such a system depends on a complex set of interacting factors and not on a simple linear cause and effect model.

Determinants of change in CD4 count and relationship with survival among children with HIV in Ethiopia

ABSTRACT We identify the determinants of the change in CD4 cell count and survival time of children living with HIV taking antiretroviral therapy. In a retrospective study, a cohort of 201 children living with HIV/AIDS aged less than 15 years were followed from October 2013 to March 2017. Separate linear effect models were done for the longitudinal outcome of CD4 cell count. The Cox PH model was conducted for the time-to-event outcome. Joint modeling was performed both for the longitudinal and survival outcomes, results were compared with the separate analysis. However, with the specific interest in identifying the determinants and characterizing the relationship between longitudinal CD4 cell count and time-to-event outcomes, the study focused on joint modeling. The finding from the joint model indicated that the estimated association parameter was −0.10, this shows lower values of CD4 cells associated with higher death. Seemingly; observation time, age, WHO clinical stages, history of Tuberculosis (assuming that Tuberculosis is active), and functional status were determinants for the mean change in the square root of CD4 cell count. Furthermore, WHO clinical stages, functional status, history of TB, and Body Mass Index have a significant negative impact on the survival probabilities of children living with HIV.

The effect of referral to an open-group behavioural weight-management programme on the relative risk of normoglycaemia, non-diabetic hyperglycaemia and type 2 diabetes: Secondary analysis of the WRAP trial.

To examine the impact of open-group behavioural weight-management programmes on the risk of diabetes among those with a body mass index (BMI) of ≥28 kg/m2 and those with non-diabetic hyperglycaemia (NDH). This was a secondary analysis of data from the WRAP trial, in which participants (N = 1267; aged ≥18 years, BMI ≥ 28 kg/m2 ) were randomized to brief intervention (BI; self-help booklet), a weight-management programme (WW; formerly Weight Watchers) for 12 weeks, or WW for 52 weeks. We used multinomial logistic regression to examine the effect of intervention group on the risk of hyperglycaemia and diabetes at 12 months in all participants with glycaemic status at both time points (N = 480; 38%) and those with NDH at baseline (N = 387; 31%). We used mixed effects models and linear fixed effects models to examine the effect of intervention group on body weight and HbA1c at 12 months in people with NDH. There was a 61% relative reduction in the risk of NDH at the 12-month follow-up (12 weeks vs. BI: relative risk ratio [RRR] = 0.39 [95% CI 0.18, 0.87], P = .021; 52 weeks vs. BI: RRR = 0.38 [95% CI 0.17, 0.86], P = .020). For intervention effects on the risk of diabetes, confidence intervals were wide and overlapped 1 [12 weeks vs. BI: RRR = 0.49 [95% CI 0.12, 1.96], P = .312; 52 weeks vs. BI: RRR = 0.40 [95% CI 0.10, 1.63], P = .199). Participants with hyperglycaemia at baseline in the weight-management programme were more probable to have normoglycaemia at the 12-month follow-up [12-week programme vs. BI: RRR = 3.57 [95% CI 1.24, 10.29], P = .019; 52-week programme vs. BI: RRR = 4.14 [95% CI 1.42, 12.12], P = .009). Open-group behavioural weight-management programmes can help to prevent the development of NDH in people with overweight and obesity and to normalize glycaemia in people with NDH.

Changes in D-dimer after initiation of antiretroviral therapy in adults living with HIV in Kenya

BackgroundIncreased coagulation biomarkers are associated with poor outcomes among people living with HIV (PLHIV). There are few data available from African cohorts demonstrating the effect of antiretroviral therapy (ART) on coagulation biomarkers.MethodsFrom March 2014 to October 2014, ART-naïve PLHIV initiating non-nucleoside reverse transcriptase inhibitor-based ART were recruited from seven clinics in western Kenya and followed for up to 12 months. Demographics, clinical history and blood specimens were collected. Logistic regression models adjusted for intrasite clustering examined associations between HIV viral load and D-Dimer at baseline. Mixed linear effects models were used to estimate mean change from baseline to 6 months overall, and by baseline viral load, sex and TB status at enrollment. Mean change in D-dimer at 6 months is reported on the log10 scale and as percentage change from baseline.ResultsAmong 611 PLHIV enrolled, 66% were female, median age was 34 years (interquartile range (IQR) 29–43 years), 31 (5%) participants had tuberculosis and median viral load was 113,500 copies/mL (IQR: 23,600-399,000). At baseline, 311 (50.9%) PLHIV had elevated D-dimer (> 500 ng/mL) and median D-dimer was 516.4 ng/mL (IQR: 302.7–926.6) (log baseline D-dimer: 2.7, IQR: 2.5–3.0). Higher baseline D-dimer was significantly associated with higher viral load (p < 0.0001), female sex (p = 0.02) and tuberculosis (p = 0.02). After 6 months on ART, 518 (84.8%) PLHIV had achieved viral load < 1000 copies/mL and median D-dimer was 390.0 (IQR: 236.6–656.9) (log D-dimer: 2.6, IQR: 2.4–2.8). Mean change in log D-dimer from baseline to 6 months was − 0.12 (95%CI −0.15, − 0.09) (p < 0.0001) indicating at 31.3% decline (95%CI −40.0, − 23.0) in D-dimer levels over the first 6 months on ART. D-dimer decline after ART initiation was significantly greater among PLHIV with tuberculosis at treatment initiation (− 172.1, 95%CI −259.0, − 106.3; p < 0.0001) and those with log viral load > 6.0 copies/mL (− 91.1, 95%CI −136.7, − 54.2; p < 0.01).ConclusionsIn this large Kenyan cohort of PLHIV, women, those with tuberculosis and higher viral load had elevated baseline D-dimer. ART initiation and viral load suppression among ART-naïve PLHIV in Kenya were associated with significant decrease in D-dimer at 6 months in this large African cohort.

Calf Muscle Oxygenation Limits Maximal Respiratory Performance During Incremental Walking Test In Obese Children

PURPOSE: To study the interrelationship between respiratory and muscle haemodynamic responses during an incremental exercise in overweight children and confirm the limitations clinically observed that lead to the premature termination of an aerobic fitness test. METHODS: Twenty-four overweight children (12 girls and 12 boys) performed a progressive maximal treadmill test during which the slope increases every minute. Changes in muscle oxygen (O2) supply were continuously assessed using near-infrared spectroscopy. Pulmonary gas exchanges and heart rate (HR) were measured. For each participant, changes in deoxyhemoglobin level (Δ[HHb]) in the medial head of the gastrocnemius muscle were expressed as a function of time, then as a percentage of peak of oxygen uptake (VO2 peak). The influence of the sex, BMI and maturation on breakdown point in muscle desoxyhemoglobin (BPHHb) was assessed with linear model effects. RESULTS: Girls reach lower VO2 peak than boys (p = 0.07). A BPHHb occured during the test in 9 girls and 6 boys and was strongly correlated with VO2 peak (r = 0.80, p < 0.001). Expressed as a percentage of maximal theorical value, HR peak tended to be higher in non BPHHB (94.7 ± 5) than in BPHHB (89.9 ± 8) participants (p = 0.1). Current data suggests that a lower maturation and lower BMI are the two main factors that delay the onset of BPHHb. CONCLUSIONS: Oxygenation of the calf muscle could limit maximal respiratory performance during an incremental walking test with a slope, especially in prepubere obese children who have lower BMI. Muscle limitations should be considered when choosing a protocol to assess VO2 peak in this population.

On the Margins: Considering the Relationship between Informal Work and Reoffending

Research suggests that job stability and job quality are vital in enhancing the crime suppression effects of employment. Unfortunately, with the erosion of the manufacturing economy and the increase in the service-dominated economy, offenders who are typically on the margins of society, are pushed towards the informal economy now more than before. The current study attends to the relationship between informal work and crime by analyzing data from the Pathways to Desistance study. Results from fixed effects linear probability models show that informal work is associated with a higher probability of engaging in expressive crimes, but not instrumental crimes. Neither informal nor formal work arrangements seem to work as crime suppressants, but informal arrangements appear more criminogenic.

THU0100 FATIGUE: A PREVALENT AND PERSISTENT SYMPTOM IN EARLY RHEUMATOID ATHRITIS

Background: Fatigue is associated with poor quality of life in people with Rheumatoid Arthritis (RA). The exact burden of fatigue in RA is uncertain. Evidence shows that fatigue may persist, even in people with well-controlled inflammatory disease. However, this is largely based on cross sectional data, and data from people with long standing or refractory disease. This study aims to examine the nature of fatigue in early RA. Objectives: Describe the prevalence of fatigue and longitudinal course of fatigue Investigate heterogeneity in the course of fatigue and identify risk factors associated with fatigue heterogeneity Methods: Data were from the early Rheumatoid Arthritis Network (ERAN), an inception cohort of people with a disease duration of Fatigue was measured using the Vitality subscale of the Short Form Health Survey questionnaire (SF36VT). Fatigue and severe fatigue were classified as SF36VT values lower than 1 and 2 standard deviations (SD) below the UK healthy population average respectively. Baseline prevalence rates standardized to Eurostat 2013 by age and sex were calculated. The course of fatigue was examined using linear mixed effect models. Group Based Trajectory Modelling (GBTM) was used to examine heterogeneity in the course of fatigue. Baseline characteristics were then used to identify predictors of group membership using univariate and multiple regression analysis. Results: Baseline characteristics include female sex (67%), mean age = 57(SD±14) yrs. Mean SF36VT score = 41(SD±11), median disease duration was 11 mths (IQR:7 – 18). The age and sex standardized prevalence rates of fatigue and severe fatigue were 44%(CI:38-50) and 18% (CI:15 – 22) respectively. 729 (59%) participants were included in the longitudinal analysis. Over the course 4 years follow up, and after accounting for the effect of age, sex, patient’s global assessment of disease activity, BMI, pain and mental health, there was a reduction in the population vitality levels from baseline,(β: -0.14CI: -0.26 to -0.02, p ≤ 0.001). Inflammation measured by erythrocyte sedimentation rate (ESR) was not significantly associated with the course of fatigue. GBTM analysis identified 2 sub-groups. These groups were named ‘Fatigue’ and ‘No-fatigue’ groups and comprised about 52% and 47% of the population respectively (Fig 1). Females, participants with, at baseline, higher BMI, higher disability score (HAQ), disease activity score (DAS28), worse pain, mental health scores and higher comorbidity score (RDCI) were more likely to belong to the Fatigue group (each p ≤ 0.05) in univariate analysis. However, higher BMI (OR 1.05 (CI: 1.0 – 1.1), HAQ (OR 2: CI: 1.5 - 2.7), RDCI(OR 1.3, CI: 1.2 – 1.5), worse SF36 pain and mental health scores (OR 0.93 CI: 0.92 – 0.95) and (OR 0.97, CI: 0.95 – 0.99) were collectively associated with fatigue group membership (AUROC=0.81). Conclusion: Fatigue is a prevalent symptom in RA, even in early disease. Embedded within the RA population are distinct sub-populations, with or without fatigue. Those with fatigue at baseline were likely to continue to report fatigue over 4 years of follow up. Unlike our previous data on pain trajectories within this cohort, a ‘resolving fatigue’ was not found. Diverse baseline characteristics, including pain, were associated with persistent fatigue. Management of fatigue might require strategies additional to disease modification, and people who require such interventions might be identified at presentation with early RA. Disclosure of Interests: Onosi Ifesemen: None declared, Daniel McWilliams Grant/research support from: Grant support from Pfizer Ltd, Adam Young: None declared, Patrick Kiely: None declared, David Walsh Grant/research support from: Grant support from Pfizer Ltd and Eli Lilly, Consultant of: Consultancy to Eli Lilly, Pfizer, Abbvie and GSK (all payments made to University of Nottingham). Consultancy to Love Productions(all payments made to the University of Nottingham).

Morphological changes in the subthalamic nucleus of people with mild-to-moderate Parkinson\u2019s disease: a 7T MRI study

This project investigated whether structural changes are present in the subthalamic nucleus (STN) of people with mild-to-moderate severity of Parkinson’s disease (PD). Within-subject measures of STN volume and fractional anisotropy (FA) were derived from high-resolution 7Tesla magnetic resonance imaging (MRI) for 29 subjects with mild-to-moderate PD (median disease duration = 2.3±1.9 years) and 18 healthy matched controls. Manual segmentation of the STN was performed on 0.4 mm in-plane resolution images. FA maps were generated and FA values were averaged over the left and right STN separately for each subject. Motor sign severity was assessed using the Movement Disorders Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS). Linear effects models showed that STN volume was significantly smaller in the PD subjects compared to controls (p = 0.01). Further, after controlling for differences in STN volumes within or between groups, the PD group had lower FA values in the STN compared to controls (corrected p ≤ 0.008). These findings demonstrate that morphological changes occur in the STN, which likely impact the function of the hyperdirect and indirect pathways of the basal ganglia and movement control.

0539 Are Short Sleepers Uncoupled Sleepers?

Abstract Introduction While short sleep durations (&amp;lt;7h/night) are associated with increased diabetes risk, there is limited evidence that increasing the habitual sleep duration of short sleepers is either feasible, or will reliably improve metabolic health outcomes. Furthermore, in the absence of insomnia disorder, it remains unclear whether habitual short sleep mainly reflects a genetic predisposition or a lifestyle choice. In a randomized controlled study we delivered a sleep extension protocol based on CBTi principles to overweight ‘short sleepers’ at increased risk of Type II diabetes. Methods 18 male short sleepers (Mage=41.4; MBMI=29.57; baseline mean TST=5.8 h/ night) with no complaints of insomnia were randomized to the sleep extension intervention or control condition (printed sleep hygiene advice). The 6-week intervention commenced with personalized sleep re-scheduling negotiated in a 60-minute 1-to-1 session, and supported by elements of sleep hygiene, stimulus control, relaxation and cognitive strategies. Outcomes included sleep duration (actigraphy), fasting insulin, Mean Amplitude of Glycemic Excursions (MAGE) from continuous glucose monitoring, and blood pressure. Data were analyzed in linear fixed effects models including time, group and baselines values. Results Adherence to the 6-week protocol was high. Relative to controls (n=8), intervention participants (n = 10) showed a significant increase in TST (95%CI 46.91min, 101.64min, p&amp;lt;0.001; MDiff = 79.4min, p&amp;lt;0.001) and significant reductions in fasting insulin (95%CI -32.08 pmol/L, -.97.0 pmol/L; p=0.04; MDiff = -10.2 pmol/L, p=0.06); MAGE (95%CI -0.77, -0.08, p=0.02; MDiff -0.35, p=0.05) and diastolic (95%CI -22, -5, p=0.004; MDiff=-12, p=0.004); and systolic blood pressure (95%CI -20, -2, p=0.03; MDiff=-10, p=0.006). Conclusion CBTi-based sleep extension protocols offer feasible and effective lifestyle interventions in the management of metabolic health in overweight short sleepers who fit published categorization of non-complaining poor sleepers with an undeveloped insomnia identity whose subjective sleep experience and objective sleep characteristics are ‘uncoupled’. Support School of Sport, Exercise and Health Sciences, Loughborough University

The impact of therapeutic mediastinal radiation on stent healing in cancer patients: An optical coherence tomography study.

e19096 Background: Cancer is a prothrombotic and proinflammatory state. Understanding stent healing in cancer patients could modify the approach to antiplatelet regimen and facilitate cancer treatment. Mediastinal radiation therapy brings an additional layer of complexity to an already challenging clinical scenario in cancer patients undergoing percutaneous coronary intervention (PCI). We sought to compare stent healing patterns with and without a history of mediastinal radiation therapy using optical coherence tomography (OCT) data. Methods: All cancer patients that underwent PCI and follow-up OCT less than 12 months post-PCI between 2009-2018 were retrospectively identified from our institutional cardiac catheterization laboratory registry. Stent healing, which was defined by a combination of stent and strut coverage, apposition, expansion, in-stent restenosis, and neo-intimal hyperplasia, was compared among 2 groups of oncological PCI patients: those with and without a history of mediastinal radiation therapy. Differences among the two categories were assessed using linear-effect models. Results: There were 86 cancer patients who received PCI with follow-up OCT within 12 months. Of these, 22% received mediastinal radiation. There was no statistical difference in baseline characteristics (hypertension, dyslipidemia, diabetes, etc.) between the two studied groups. In the non-radiation group, 59% were late stage or advanced cancer patients, while 63% of patients in the radiation group were advanced cancer patients. Stent healing was similar among those who had received radiation therapy and those who had not (Table). Conclusions: A history of mediastinal radiation does not appear to affect early stent healing within 1 year of PCI. Given stent healing appears to be similar with or without radiation, dual anti-platelet therapy regimens may not need to change based on radiation exposure alone. Future studies should address the duration of dual antiplatelet regimens and long-term outcomes (i.e. in-stent restenosis) based on stent healing properties. [Table: see text]

Multisite randomized trial of integrated palliative and oncology care for patients with acute myeloid leukemia (AML).

12000 Background: Patients with AML receiving intensive chemotherapy experience substantial decline in their quality of life (QOL) and mood during their hospitalization for induction chemotherapy and often receive aggressive care at the end of life (EOL). We sought to examine the effect of integrated palliative and oncology care on QOL, mood, post-traumatic stress (PTSD) symptoms, and EOL outcomes in patients with AML. Methods: We conducted a multi-site randomized trial of integrated palliative and oncology care (n = 86) versus usual oncology care (n = 74) for patients with AML undergoing intensive chemotherapy. Patients assigned to the intervention were seen by palliative care clinicians at least twice per week during their hospitalization for induction chemotherapy and all subsequent hospitalizations. Patients completed the Functional Assessment of Cancer Therapy-Leukemia, the Hospital Anxiety and Depression Scale, and the PTSD Checklist to assess their QOL, mood, and PTSD symptoms at baseline, weeks 2, 4, 12, and 24. The primary endpoint was QOL at week-2. We used analysis of covariance and mixed linear effect models, controlling for baseline scores, to assess the effect of the intervention on patient-reported outcomes. Results: Between 1/2017 and 7/2019, we enrolled 160/235 (68.1%) of eligible patients. Compared to those receiving usual care, intervention patients reported better QOL (107.59 vs. 116.45, P = 0.039) and lower depression (7.20 vs. 5.68, P = 0.021), anxiety (5.94 vs. 4.53, P = 0.018), and PTSD symptoms (31.69 vs. 27.79, P = 0.009) at week 2. Intervention effects were sustained up to week 24 for QOL (B = 2.35, P = 0.048), depression (B = -0.42, P = 0.039), anxiety (B = -0.38, P = 0.042), and PTSD symptoms (B = -1.43, P = 0.002). Among deceased participants, those receiving the intervention were more likely to report discussing their EOL care preferences with their clinicians (75.0% vs. 40.0%, P = 0.009) and less likely to receive chemotherapy in the last 30 days of life (34.9% vs. 65.9%, P = 0.008). There was no difference in hospice utilization or hospitalization at the EOL. Conclusions: The integrated palliative and oncology care model for patients with AML receiving intensive chemotherapy led to substantial improvements in patients’ QOL, psychological distress, and EOL care. Thus, palliative care should be considered a new standard of care for patients with AML. Clinical trial information: NCT02975869 .

HOW MANY DOSEMETERS ARE NEEDED FOR CORRECT MEAN ORGAN DOSE ASSESSMENT WHEN PERFOMING PHANTOM DOSIMETRY? A PHANTOM STUDY EVALUATING LIVER ORGAN DOSE AND INVESTIGATING TLD NUMBERS AND WAYS OF DOSEMETER PLACEMENT.

This study evaluated repeated mean organ dose measurements of the liver by phantom dosimetry and statistical modelling in order to find a way to reduce the number of dosemeters needed for precise organ dose measurements. Thermoluminescent dosemeters were used in an adult female phantom exposed to a biplanar x-ray source at three different axial phantom rotations. Generalised mixed linear effect modelling was used for statistical analysis. A subgroup of five to six organ-specific locations out of 28 yielded mean liver organ doses within 95% confidence intervals of measurements based on all 28 liver-specific dosemeter locations. No statistical difference of mean liver dose was observed with rotation of the phantom either 10° clockwise or counter-clockwise as opposed to the coronal plane. Phantom dosimetry handling time during organ dose measurements can be markedly reduced, in this case the liver, by 79% (22/28), while still providing precise mean organ dose measurements.

Venous blood flow alterations in glaucoma patients.

To evaluate the correlation between retinal venous blood flow parameters and glaucomatous visual field damage in a retrospective analysis. Fifty-five (24 male, 31 female) glaucoma patients, under intraocular pressure (IOP) reducing treatment, aged (mean ± SD) 69 ± 10years, 29 with primary open-angle and 26 with normal-tension glaucoma, were evaluated with regard to the correlation between IOP, color Doppler retinal venous blood flow velocity and glaucomatous damage. Univariate regression analysis disclosed statistically significant correlations of the visual field index MD with age, IOP and venous blood flow (p ≤ 0.026 for each parameter). A mixed linear effects model disclosed a significant correlation of MD with age, IOP and venous blood flow (p ≤ 0.002 for each parameter), but not with sex, side (right eye versus left eye) and diagnosis (primary open-angle glaucoma versus normal-tension glaucoma). Finally, interocular difference (right eye vs. left eye of the same patient) in MD correlated with interocular differences in venous blood flow velocities (p < 0.001), but not with interocular differences in IOP. Glaucomatous damage correlated negatively with retinal venous blood flow velocity, be it between subjects or between eyes within individuals. This study is limited by its cross-sectional design, and it is not possible to draw any conclusion with regard to the origin of the correlations.

Detection of latent heteroscedasticity and group-based regression effects in linear models via Bayesian model selection

–Standard linear modeling approaches make potentially simplistic assumptions regarding the structure of categorical effects that may obfuscate more complex relationships governing data. For example, recent work focused on the two-way unreplicated layout has shown that hidden groupings among the levels of one categorical predictor frequently interact with the ungrouped factor. We extend the notion of a “latent grouping factor” to linear models in general. The proposed work allows researchers to determine whether an apparent grouping of the levels of a categorical predictor reveals a plausible hidden structure given the observed data. Specifically, we offer a Bayesian model selection-based approach to reveal latent group-based heteroscedasticity, regression effects, and/or interactions. Failure to account for such structures can produce misleading conclusions. Since the presence of latent group structures is frequently unknown a priori to the researcher, we use fractional Bayes factor methods and mixture g-priors to overcome lack of prior information.

Effects of Lisinopril on Arterial Stiffness, Cerebral Blood Flow and Cortical Thickness in Hypertensive Dahl‐S Rats

Study objectivesCentral arterial stiffness (CAS), associated with hypertension, is likely underlies the stiffening of cerebral arterial wall resulting in cerebral hypoperfusion, neuronal density loss and cognitive decline. We previously found associations between pulse wave velocity (PWV), a marker of CAS, and hippocampal cerebral blood flow (CBF) and neuronal density in 6‐mo old hypertensive Dahl salt‐sensitive (Dahl‐S) rats. This study examined the effect of ACE inhibitor, Lisinopril, on cardiovascular and brain parameters in hypertensive Dahl‐S rats.MethodsMale Dahl‐S rats (n=32) were fed with a normal salt diet (0.5% NaCl) for a duration of study. Systolic blood pressure (SBP), PWV (by echocardiography); hippocampal N‐acetyl aspartate (NAA) concentration, hippocampal CBF were measured at 6‐mo of age (baseline) and after 3‐mo and 6‐mo of Lisinopril (15mg/kg body weight/day in drinking water; n=14) or control (n=18) treatment. NAA was measured using proton spectroscopy in a 7T Bruker Biospec MRI scanner; CBF was measured using continuous arterial spin labeling (CASL) in the coronal slice with the largest hippocampal area; cortical thickness (CT) analysis was done using a T2 weighted RARE 3D image (n=12). Advanced Normalization Tools (ANTs) were used for the CT estimation. T‐test and 2‐way ANOVA linear mix effect model analyses were performed in R, EXCEL and MATLAB. Data presented as mean±SE.ResultsAt 6‐mo of age, the Dahl‐S rats developed advanced hypertension vs. 3‐mo (SBP: 164±4 vs. 141±3 mmHg; P&lt;0.01, t‐test). SBP of non‐treated (NT) animals was steadily high at all time points while treated animals exhibited a decrease in SBP at 9‐mo vs. 6‐mo of age. SBP remained low from 9‐mo to 12‐mo of age (Fig. 1a). In treated animals, PWV decreased at all timepoints after 6‐mo of age. For NT animals, PWV remained steady from 6‐mo to 9‐mo but increases from 9‐mo to12‐mo of age (Fig. 1b). Hippocampal CBF of NT animals decreased in a downwards trend but remain stabilized in the treated group across timepoints (Fig. 1c). Likewise, the hippocampal NAA, among treated animals remained stabilized, while in NT animals this parameter fluctuated with an eventual downwards trend (Fig. 1d). The mean difference maps of the CT for the NT and treated rats are similar except for the regions denoted by the red arrows when comparing treated to NT (Fig. 2a–b). The occipital region showed cortical thickening (blue) and parietal region showed cortical thinning (yellow). The P value maps show that those regions of significant thickening and thinning are greater in the treated vs. NT rats (Fig. 2c–d).ConclusionThe posterior cerebral artery is a dominant source of perfusion to the hippocampus and the occipital lobe; thus, our findings suggest a coherent pattern through which CAS can affect the cerebral arteries in hypertensive Dahl‐S rats. Lisinopril treatment decrease CAS and improve regional CBF and neuronal density. Therefore, CAS may be a therapeutic target for vascular dementia.Support or Funding InformationThe work was supported by the Intramural Research Program of the National Institute on Aging of the National Institutes of Health.Longitudinal changes in cardiovascular parameters in Lisinopril treated Dahl‐S rats. Systolic blood pressure SBP (a), pulse wave velocity PWV (b), hippocampal CBF (c), and NAA (d) in Lisinopril treated (T) and nontreated (NT) rats. By 2‐way ANOVA Linear Mixed Effects: *p &lt; 0.001 vs. age‐matched treated rats; #p &lt; 0.03 vs. 6‐mo, ## p &lt;0.001 vs. 6‐mo; by paired T‐test (c, d): * p = 0.07 vs. 6‐mo; # p = 0.06 vs. 9‐mo NT.Figure 1Mean Differences in Cortical Thickness and their p value maps (6‐mo subtracted by 12‐mo). Mean difference between age 6‐mo and 12‐mo of (a) nontreated animals and (b) lisinopril treated animals. Blue indicates regions of thickening. Yellow indicated regions of thinning. Occipital and parietal regions indicated by the red arrows. P value map of nontreated animals (c) and lisinopril treated animals (d). All insignificant p values &gt; 0.05 were revalued as 0.05 for visualization purposes.Figure 2

Comparison of HIV prevalence among antenatal clinic attendees estimated from routine testing and unlinked anonymous testing.

In 2015, WHO and UNAIDS released new guidance recommending that countries transition from conducting antenatal clinic (ANC) unlinked anonymous testing (ANC-UAT) for tracking HIV prevalence trends among pregnant women to using ANC routine testing (ANC-RT) data, which are more consistent and economic to collect. This transition could pose challenges for distinguishing whether changes in observed prevalence are due to a change in underlying population prevalence or due to a change in the testing approach. We compared the HIV prevalence measured from ANC-UAT and ANCRT in 15 countries that had both data sources in overlapping years. We used linear mixed-e effects model (LMM) to estimate the RT-to-UAT calibration parameter as well as other unobserved quantities. We summarized the results at different levels of aggregation (e.g., country, urban, rural, and province). Based on our analysis, the HIV prevalence measured by ANC-UAT and ANC-RT data are consistent in most countries. Therefore, if large discrepancy is observed between ANC-UAT and ANC-RT at the same location, we recommend that people should be cautious and investigate the reason. For countries that lack information to estimate the calibration parameter, we propose an informative prior distribution of mean 0 and standard deviation 0.2 for the RT-to-UAT calibration parameter.

The Inconsistent Nature of Heart Rate Variability During Sleep in Normal Children and Adolescents.

Introduction: Cardiac function is modulated by multiple factors including exogenous (circadian rhythm) and endogenous (ultradian 90–110 min sleep cycle) factors. By evaluating heart rate variability (HRV) during sleep, we will better understand their influence on cardiac activity. The aim of this study was to evaluate HRV in the dark phase of the circadian rhythm during sleep in healthy children and adolescents.Methods: One 3 min segment of pre-sleep electrocardiography (EEG) and 3, 6 min segments of electrocardiography recorded during polysomnography from 75 healthy children and adolescents were sampled during progressive cycles of slow wave sleep (SWS1, SWS2, SWS3). Three, 3 min segments of rapid eye movement sleep (REM) were also assessed, with REM1 marked at the last REM period before awakening. Studies that recorded REM3 prior to SWS3 were used for assessment. HRV variables include the following time domain values: mean NN (average RR intervals over given time), SDNN (Standard Deviation of RR intervals), and RMSSD (root Mean Square of beat-to-beat Differences). Frequency domain values include: low frequency (LF), high frequency (HF), and LF:HF.Results: Mixed linear effects model analysis revealed a significant difference in time and frequency domain values between sleep cycles and stages. Mean NN was lowest (highest heart rate) during pre—sleep then significantly increased across SWS1-3. Mean NN in SWS1 was similar to all REM periods which was significantly lower than both SWS2 and SWS3. SDNN remained at pre-sleep levels until SWS3, and then significantly increased in REM1&2. There was a large drop in LF from pre-sleep to SWS1. As cycles progressed through the night, LF remains lower than awake but increases to awake like levels by REM2. RMSSD and HF were lowest in pre-sleep and increased significantly by SWS1 and remain high and stable across stages and cycles except during the REM3 period where RMSSD decreased.Conclusion: Our results demonstrate that there are considerable changes in the spectral analysis of cardiac function occurring during different sleep stages and between sleep cycles across the night. Hence, time of night and sleep stage need to be considered when reporting any HRV differences.

Changes in bone mineral density over 10 years in patients with early rheumatoid arthritis

ObjectivesTo investigate changes in bone mineral density (BMD) in patients with early rheumatoid arthritis (RA) over a 10-year period.MethodsConsecutive patients with early RA (symptom duration <12 months) were followed according...