Supplementation of vitamin E as an addition to a commercial renal diet does not prolong survival of cats with chronic kidney disease.

BACKGROUND: The aim of this double-blind, placebo-controlled study was to investigate the effect of vitamin E supplementation as an addition to a commercial renal diet on survival time of cats with different stages of chronic kidney disease (CKD). In addition, we were interested whether vitamin E supplementation affects selected oxidative stress and clinical parameters. Thirty-four cats with CKD and 38 healthy cats were included in the study. Cats with CKD were classified according to the IRIS Guidelines; seven in IRIS stage 1, 15 in IRIS stage 2, five in IRIS stage 3 and seven in IRIS stage 4. Cats with CKD were treated according to IRIS Guidelines. Cats with CKD were randomly assigned to receive vitamin E (100 IU/cat/day) or placebo (mineral oil) for 24 weeks in addition to standard therapy. Plasma malondialdehyde (MDA) and protein carbonyl (PC) concentrations, DNA damage of peripheral lymphocytes and plasma vitamin E concentrations were measured at baseline and four, eight, 16 and 24 weeks thereafter. Routine laboratory analyses and assessment of clinical signs were performed at each visit. RESULTS: Vitamin E supplementation had no effect on the survival time and did not reduce the severity of clinical signs. Before vitamin E supplementation, no significant differences in vitamin E, MDA and PC concentrations were found between healthy and CKD cats. However, plasma MDA concentration was statistically significantly higher (p = 0.043) in cats with early CKD (IRIS stages 1 and 2) than in cats with advanced CKD (IRIS stages 3 and 4). Additionally, DNA damage was statistically significantly higher in healthy cats (p ≤ 0.001) than in CKD cats. Plasma vitamin E concentrations increased statistically significantly in the vitamin E group compared to the placebo group four (p = 0.013) and eight (p = 0.017) weeks after the start of vitamin E supplementation. During the study and after 24 weeks of vitamin E supplementation, plasma MDA and PC concentrations and DNA damage remained similar to pre-supplementation levels in both the placebo and vitamin E groups. CONCLUSIONS: Vitamin E supplementation as an addition to standard therapy does not prolong survival in feline CKD.

Statistical analysis of high-dimensional biomedical data: a gentle introduction to analytical goals, common approaches and challenges.

BACKGROUND: In high-dimensional data (HDD) settings, the number of variables associated with each observation is very large. Prominent examples of HDD in biomedical research include omics data with a large number of variables such as many measurements across the genome, proteome, or metabolome, as well as electronic health records data that have large numbers of variables recorded for each patient. The statistical analysis of such data requires knowledge and experience, sometimes of complex methods adapted to the respective research questions. METHODS: Advances in statistical methodology and machine learning methods offer new opportunities for innovative analyses of HDD, but at the same time require a deeper understanding of some fundamental statistical concepts. Topic group TG9 "High-dimensional data" of the STRATOS (STRengthening Analytical Thinking for Observational Studies) initiative provides guidance for the analysis of observational studies, addressing particular statistical challenges and opportunities for the analysis of studies involving HDD. In this overview, we discuss key aspects of HDD analysis to provide a gentle introduction for non-statisticians and for classically trained statisticians with little experience specific to HDD. RESULTS: The paper is organized with respect to subtopics that are most relevant for the analysis of HDD, in particular initial data analysis, exploratory data analysis, multiple testing, and prediction. For each subtopic, main analytical goals in HDD settings are outlined. For each of these goals, basic explanations for some commonly used analysis methods are provided. Situations are identified where traditional statistical methods cannot, or should not, be used in the HDD setting, or where adequate analytic tools are still lacking. Many key references are provided. CONCLUSIONS: This review aims to provide a solid statistical foundation for researchers, including statisticians and non-statisticians, who are new to research with HDD or simply want to better evaluate and understand the results of HDD analyses.

Unique Clinical, Immune, and Genetic Signature in Patients with Borrelial Meningoradiculoneuritis1.

Lyme neuroborreliosis (LNB) in Europe may manifest with painful meningoradiculoneuritis (also known as Bannwarth syndrome) or lymphocytic meningitis with or without cranial neuritis (peripheral facial palsy). We assessed host immune responses and the prevalence of TLR1 (toll-like receptor 1)-1805GG polymorphism to gain insights into the pathophysiology of these conditions. Regardless of LNB manifestation, most mediators associated with innate and adaptive immune responses were concentrated in cerebrospinal fluid; serum levels were unremarkable. When stratified by specific clinical manifestation, patients with meningoradiculoneuritis had higher levels of B-cell chemoattractants CXC motif chemokine ligand (CXCL) 12 and CXCL13 and T-cell-associated mediators CXCL9, CXCL10, and interleukin 17, compared with those without radicular pain. Moreover, these patients had a higher frequency of TLR1-1805GG polymorphism and more constitutional symptoms. These findings demonstrate that meningoradiculoneuritis is a distinct clinical entity with unique immune and genetic pathophysiology, providing new considerations for the study of LNB and borrelial meningoradiculitis.

Hidden analyses: a review of reporting practice and recommendations for more transparent reporting of initial data analyses.

BACKGROUND: In the data pipeline from the data collection process to the planned statistical analyses, initial data analysis (IDA) typically takes place between the end of the data collection and do not touch the research questions. A systematic process for IDA and clear reporting of the findings would help to understand the potential shortcomings of a dataset, such as missing values, or subgroups with small sample sizes, or shortcomings in the collection process, and to evaluate the impact of these shortcomings on the research results. A clear reporting of findings is also relevant when making datasets available to other researchers. Initial data analyses can provide valuable insights into the suitability of a data set for a future research study. Our aim was to describe the practice of reporting of initial data analyses in observational studies in five highly ranked medical journals with focus on data cleaning, screening, and reporting of findings which led to a potential change in the analysis plan. METHODS: This review was carried out using systematic search strategies with eligibility criteria for articles to be reviewed. A total of 25 papers about observational studies were selected from five medical journals published in 2018. Each paper was reviewed by two reviewers and IDA statements were further discussed by all authors. The consensus was reported. RESULTS: IDA statements were reported in the methods, results, discussion, and supplement of papers. Ten out of 25 papers (40%) included a statement about data cleaning. Data screening statements were included in all articles, and 18 (72%) indicated the methods used to describe them. Item missingness was reported in 11 papers (44%), unit missingness in 15 papers (60%). Eleven papers (44%) mentioned some changes in the analysis plan. Reported changes referred to missing data treatment, unexpected values, population heterogeneity and aspects related to variable distributions or data properties. CONCLUSION: Reporting of initial data analyses were sparse, and statements on IDA were located throughout the research articles. There is a lack of systematic reporting of IDA. We conclude the article with recommendations on how to overcome shortcomings in the practice of IDA reporting in observational studies.

Identifying the Italian provinces with increased mortality during COVID-19 epidemics using the data made available by the Italian National Institute of Statistics. A methodological challenge.

OBJECTIVES: to identify the Italian provinces with excess mortality during the COVID-19 epidemics using the mortality data provided in April 2020 by the Italian National Institute of Statistics (Istat) that, by design, included only the municipalities with at least 20% mortality increase compared to the same period in 2015-19. Inference with the aim to identify increased mortality at provincial level was a very important task when the Istat data were released in April, but the naïve aggregation of the selected municipalities was not sensible to due to the selection criteria of the municipalities used by Istat. DESIGN: use of a permutation-based approach to identify the Italian provinces with excess mortality during the first month of the COVID-19 epidemics using the data made available from Istat and taking into account the biased inclusion criteria. SETTING AND PARTICIPANTS: the number of deaths from any cause from 1 January was available for each year of the 2015-2020 period. Data were stratified by municipality, sex and 21 age categories. The third data release (R3) included 1,686 of the 7,904 Italian municipalities with increased mortality in 2020, covering about 40% of the Italian population. Results were compared with those obtainable with the fifth data release (R5), made available in June, when the selection of the municipalities was no longer based on increased mortality and which included more than 90% of the Italian population. R5 was considered the gold standard. MAIN OUTCOME MEASURES: excess of deaths from any cause in the Italian provinces between 1 March and 4 April; relative risk (RR); permutation p-values; permutation-based adjusted relative risk; population coverage. RESULTS: the results of this study, which are based on two different test statistics, identify 17 and 33 provinces (out of 103) with increased overall mortality, respectively, controlling the family-wise error rate at 0.05 level. Most of the identified provinces are neighbouring provinces in the northern regions of Lombardy, Emilia-Romagna, Piedmont, Liguria, Marche and Tuscany, where most of the COVID-19 cases and deaths were identified. The comparison with data from R5 shows that all the identified provinces had an increase in overall mortality, mostly (31/34) above 25%. On average, the adjusted RR slightly underestimates the RR from R5, underestimating the large RR and overestimating the small RR. CONCLUSIONS: this was, to the best of the authors' knowledge, the first attempt to aggregate the Istat data at province level and obtain a reliable and generalizable statistical inference. This permutation-based approach provides a feasible approach to take into account the selection bias that was present in the data and could be used for analysing other types of data that present some type of selection bias.

Classification of weekly provincial overall age- and gender-specific mortality patterns during the COVID-19 epidemics in Italy.

OBJECTIVES: to provide a time-varying classification of the Italian provinces based on the weekly age- and gender-specific relative risks (RR) for overall mortality, obtained comparing the number of deaths from 13 weeks from the beginning of the COVID-19 epidemics, with the average number of deaths from the same period in 2015-19. DESIGN: population overall mortality data provided by the Italian National Statistical Office (Istat). SETTING AND PARTICIPANTS: Italian residents 60 years or older from 7,357/7,904 Italian municipalities. For the included municipalities, the number of deaths from any cause from 1 January to 30 May 2020 was available for each day of the 2015-2020 period. Data were stratified by gender, 4 age categories (60-69, 70-79, 80-89, 90+), week, and province. MAIN OUTCOME MEASURES: province- and gender-specific weekly RR curves (age category vs RR), obtained for 13 weeks between 26 February and 26 May; excess mortality; time-varying/weekly classification of provinces. RESULTS: these results provide a weekly classification of the Italian provinces based on their RR curves in 5 groups, 2 of which had high and very high excess mortality during the epidemics. Most of the provinces that appeared at least once in the highest-risk group are neighbouring provinces in the Northern Regions of Lombardy, Emilia-Romagna, Piedmont, and Marche (in central Italy), where most of the COVID-19 cases and deaths were identified. Temporally, most of these provinces remained in the highest-risk group for 4 or 5 weeks; those that entered the group later, improved faster. The overall RR curves for groups differed in magnitude, but also in the shape, which varied markedly also between men and women and, most importantly, in the highest-risk group. CONCLUSIONS: this study gives timely re-analysis of the Istat data at weekly level and provides a classification of the geographical and temporal characteristics of the excess mortality in the Italian provinces during the COVID-19 epidemics. As expected, the used clustering method groups the provinces that have similar RR values in the two gender-specific curves. The results facilitate the presentation of the spatio-temporal mortality patterns of the epidemics and provide evidence of high heterogeneity in the group of provinces that were defined as high-risk groups by others, based on their geographical position or on the time of the observed spread of the virus.

Virus RNA Load in Patients with Tick-Borne Encephalitis, Slovenia.

We determined levels of tick-borne encephalitis (TBE) virus (TBEV) RNA in serum samples obtained from 80 patients during the initial phase of TBE in Slovenia. For most samples, levels were within the range of 3-6 log10 copies RNA/mL. Levels were higher in female patients than in male patients, but we found no association between virus load and several laboratory and clinical parameters, including severity of TBE. However, a weak humoral immune response was associated with a more severe disease course, suggesting that inefficient clearance of virus results in a more serious illness. To determine whether a certain genetic lineage of TBEV had a higher virulence potential, we obtained 56 partial envelope protein gene sequences by directly sequencing reverse transcription PCR products from clinical samples of patients. This method provided a large set of patient-derived TBEV sequences. We observed no association between phylogenetic clades and virus load or disease severity.

Firth's logistic regression with rare events: accurate effect estimates and predictions?

Firth's logistic regression has become a standard approach for the analysis of binary outcomes with small samples. Whereas it reduces the bias in maximum likelihood estimates of coefficients, bias towards one-half is introduced in the predicted probabilities. The stronger the imbalance of the outcome, the more severe is the bias in the predicted probabilities. We propose two simple modifications of Firth's logistic regression resulting in unbiased predicted probabilities. The first corrects the predicted probabilities by a post hoc adjustment of the intercept. The other is based on an alternative formulation of Firth's penalization as an iterative data augmentation procedure. Our suggested modification consists in introducing an indicator variable that distinguishes between original and pseudo-observations in the augmented data. In a comprehensive simulation study, these approaches are compared with other attempts to improve predictions based on Firth's penalization and to other published penalization strategies intended for routine use. For instance, we consider a recently suggested compromise between maximum likelihood and Firth's logistic regression. Simulation results are scrutinized with regard to prediction and effect estimation. We find that both our suggested methods do not only give unbiased predicted probabilities but also improve the accuracy conditional on explanatory variables compared with Firth's penalization. While one method results in effect estimates identical to those of Firth's penalization, the other introduces some bias, but this is compensated by a decrease in the mean squared error. Finally, all methods considered are illustrated and compared for a study on arterial closure devices in minimally invasive cardiac surgery. Copyright © 2017 John Wiley & Sons, Ltd.

Borrelial Lymphocytoma in Adult Patients.

BACKGROUND: Information on the course and outcome of borrelial lymphocytoma (BL) is limited. METHODS: The study comprised 144 adult patients (75 female, 69 male; median age, 49 years) who had BL diagnosed at a single center between 1986 and 2014 and were followed up for 1 year. RESULTS: BL was located on the breast in 106 patients (73.6%), on the ear lobe in 27 (18.8%), and elsewhere in 11 (7.6%). The median duration of BL before diagnosis was 27 days (interquartile range [IQR], 9-68 days). Concomitant erythema migrans was registered in 104 of 144 patients (72.2%); other objective manifestations of Lyme borreliosis (LB) were present in 11 (7.6%). Immunoglobulin M and/or G borrelial serum antibodies were present in 72 patients (50%). Borreliae were isolated from BL lesions in 14 of 42 patients (33.3%) who had not received antibiotics before skin biopsy. Of 13 typed Borrelia strains, 11 were B. afzelii, 1 was B. garinii, and 1 was B. bissettii The median duration of BL after starting antibiotic treatment was 21 days ([IQR], 10-30 days); the average duration was longer in patients who were older, had longer BL duration before treatment, or had signs of disseminated LB. Treatment failure occurred in 14 of 144 patients (9.7%). Patients with signs or symptoms of disseminated LB before treatment had nearly 4 times higher odds of treatment failure (95% confidence interval, 1.22-13.07) than those without such symptoms. All patients with treatment failure had uneventful outcome after retreatment. CONCLUSIONS: BL is a rare manifestation of early localized LB. Fourteen-day antibiotic treatment, as used for erythema migrans, is effective.

Course and Outcome of Early European Lyme Neuroborreliosis (Bannwarth Syndrome): Clinical and Laboratory Findings.

BACKGROUND: Information on the course and outcome of early European Lyme neuroborreliosis is limited. METHODS: The study comprised 77 patients (38 males, 39 females; median age, 58 years) diagnosed with painful meningoradiculitis (Bannwarth syndrome) who were followed up for 1 year at a single center. RESULTS: Duration of neurological symptoms before diagnosis was 30 (interquartile range, 14-50) days. The most frequent symptoms/signs were radicular pain (100%), sleep disturbances (75.3%), erythema migrans (59.7%), headache (46.8%), fatigue (44.2%), malaise (39%), paresthesias (32.5%), peripheral facial palsy (PFP) (36.4%), meningeal signs (19.5%), and pareses (7.8%). Cerebrospinal fluid (CSF) analysis revealed lymphocytic/monocytic pleocytosis, elevated protein concentration, and intrathecal synthesis of borrelial immunoglobulin M and immunoglobulin G antibody in 100%, 81.1%, 63%, and 88.7% of patients, respectively. Borreliae (predominantly Borrelia garinii) were isolated from CSF, skin, and blood in 15.6%, 40.6%, and 2.7% of patients, respectively. The outcome after 14-day treatment with ceftriaxone was favorable in 87.8% of patients. Control CSF examination at 3 months showed decreased leukocyte counts in all patients; however, 23.3% still had pleocytosis (>10 × 10(6) cells/L). A model based on pretreatment data and the findings at the end of 14-day antibiotic treatment accurately predicted which patients would have an unfavorable outcome 6 or 12 months after treatment. CONCLUSIONS: Our patients had fewer pretreatment neurological complications (PFP, pareses) than reported for Bannwarth syndrome decades ago, probably as the result of earlier recognition and prompt antibiotic treatment. Unfavorable outcome was rare and was predicted by the continued presence of symptoms 14 days after commencement of treatment.

Viral respiratory infections in a nursing home: a six-month prospective study.

BACKGROUND: The knowledge on viral respiratory infections in nursing home (NH) residents and their caregivers is limited. The purpose of the present study was to assess and compare the incidence of acute respiratory infections (ARI) in nursing home (NH) residents and staff, to identify viruses involved in ARI and to correlate viral etiology with clinical manifestations of ARI. METHODS: The prospective surveillance study was accomplished in a medium-sized NH in Slovenia (central Europe). Ninety NH residents and 42 NH staff were included. Nasopharyngeal swabs were collected from all participants at enrollment (December 5th, 2011) and at the end of the study (May 31st, 2012), and from each participant that developed ARI within this timeframe. Molecular detection of 15 respiratory viruses in nasopharyngeal swab samples was performed. RESULTS: The weekly incidence rate of ARI in NH residents and NH staff correlated; however, it was higher in staff members than in residents (5.9 versus 3.8/1,000 person-days, P = 0.03), and was 2.5 (95 % CI: 1.36-4.72) times greater in residents without dementia than in residents with dementia. Staff members typically presented with upper respiratory tract involvement, whereas in residents lower respiratory tract infections predominated. Respiratory viruses were detected in 55/100 ARI episodes. In residents, influenza A virus, respiratory syncytial virus, and human metapneumovirus were detected most commonly, whereas in NH staff rhinovirus and influenza A virus prevailed. 38/100 ARI episodes (30/56 in residents, 8/44 in staff) belonged to one of three outbreaks (caused by human metapneumovirus, influenza A virus and respiratory syncytial virus, respectively). NH residents had higher chances for virus positivity within outbreak than HN staff (OR = 7.4, 95 % CI: 1.73-31.48, P < 0.01). CONCLUSIONS: ARI are common among NH residents and staff, and viruses were detected in a majority of the episodes of ARI. Many ARI episodes among NH residents were outbreak cases and could be considered preventable. TRIAL REGISTRATION: The study was registered on the 1th of December 2011 at ClinicalTrials ( NCT01486160 ).

Boosting for high-dimensional two-class prediction.

BACKGROUND: In clinical research prediction models are used to accurately predict the outcome of the patients based on some of their characteristics. For high-dimensional prediction models (the number of variables greatly exceeds the number of samples) the choice of an appropriate classifier is crucial as it was observed that no single classification algorithm performs optimally for all types of data. Boosting was proposed as a method that combines the classification results obtained using base classifiers, where the sample weights are sequentially adjusted based on the performance in previous iterations. Generally boosting outperforms any individual classifier, but studies with high-dimensional data showed that the most standard boosting algorithm, AdaBoost.M1, cannot significantly improve the performance of its base classier. Recently other boosting algorithms were proposed (Gradient boosting, Stochastic Gradient boosting, LogitBoost); they were shown to perform better than AdaBoost.M1 but their performance was not evaluated for high-dimensional data. RESULTS: In this paper we use simulation studies and real gene-expression data sets to evaluate the performance of boosting algorithms when data are high-dimensional. Our results confirm that AdaBoost.M1 can perform poorly in this setting, often failing to improve the performance of its base classifier. We provide the explanation for this and propose a modification, AdaBoost.M1.ICV, which uses cross-validated estimates of the prediction errors and outperforms the original algorithm when data are high-dimensional. The use of AdaBoost.M1.ICV is advisable when the base classifier overfits the training data: the number of variables is large, the number of samples is small, and/or the difference between the classes is large. To a lesser extent also Gradient boosting suffers from similar problems. Contrary to the findings for the low-dimensional data, shrinkage does not improve the performance of Gradient boosting when data are high-dimensional, however it is beneficial for Stochastic Gradient boosting, which outperformed the other boosting algorithms in our analyses. LogitBoost suffers from overfitting and generally performs poorly. CONCLUSIONS: The results show that boosting can substantially improve the performance of its base classifier also when data are high-dimensional. However, not all boosting algorithms perform equally well. LogitBoost, AdaBoost.M1 and Gradient boosting seem less useful for this type of data. Overall, Stochastic Gradient boosting with shrinkage and AdaBoost.M1.ICV seem to be the preferable choices for high-dimensional class-prediction.

Joint use of over- and under-sampling techniques and cross-validation for the development and assessment of prediction models.

BACKGROUND: Prediction models are used in clinical research to develop rules that can be used to accurately predict the outcome of the patients based on some of their characteristics. They represent a valuable tool in the decision making process of clinicians and health policy makers, as they enable them to estimate the probability that patients have or will develop a disease, will respond to a treatment, or that their disease will recur. The interest devoted to prediction models in the biomedical community has been growing in the last few years. Often the data used to develop the prediction models are class-imbalanced as only few patients experience the event (and therefore belong to minority class). RESULTS: Prediction models developed using class-imbalanced data tend to achieve sub-optimal predictive accuracy in the minority class. This problem can be diminished by using sampling techniques aimed at balancing the class distribution. These techniques include under- and oversampling, where a fraction of the majority class samples are retained in the analysis or new samples from the minority class are generated. The correct assessment of how the prediction model is likely to perform on independent data is of crucial importance; in the absence of an independent data set, cross-validation is normally used. While the importance of correct cross-validation is well documented in the biomedical literature, the challenges posed by the joint use of sampling techniques and cross-validation have not been addressed. CONCLUSIONS: We show that care must be taken to ensure that cross-validation is performed correctly on sampled data, and that the risk of overestimating the predictive accuracy is greater when oversampling techniques are used. Examples based on the re-analysis of real datasets and simulation studies are provided. We identify some results from the biomedical literature where the incorrect cross-validation was performed, where we expect that the performance of oversampling techniques was heavily overestimated.

Course and outcome of Early Lyme borreliosis in patients with hematological malignancies.

Patients with erythema migrans and underlying hematological malignancy more often had signs of disseminated Lyme borreliosis and more frequently needed antibiotic retreatment than sex-, age-, and antibiotic treatment-matched immunocompetent persons with erythema migrans. However, the outcome was excellent in both groups.

Intrathecally produced IgG and IgM antibodies to recombinant VlsE, VlsE peptide, recombinant OspC and whole cell extracts in the diagnosis of Lyme neuroborreliosis.

Detection of intrathecally produced specific antibodies (AI) is essential in the diagnosis of Lyme neuroborreliosis (LNB); however, the performance of various newer AI detection methods has not been systematically assessed. Here we assessed and compared advanced test systems for detecting borrelia IgG-AI and IgM-AI. Serum and cerebrospinal fluid (CSF) samples from well-defined LNB and tick-borne encephalitis (TBE) patients, 25 each, were tested with three antibody detection systems, one based on chemiluminescence (CLA) and two based on enzyme-linked immunosorbent assays (ELISA), employing different antigens for detection of IgG and IgM antibodies. In samples from patients with LNB, IgG-AI was detected in 20 samples by CLA, 19 by ELISA1, and 22 by ELISA2, and IgM-AI was detected in 16 samples by CLA, six by ELISA1, and 11 by ELISA2. In samples from TBE patients, IgG-AI was positive in one case by CLA and ELISA2, and in 7 cases by ELISA1, whereas IgM-AI was positive in one case by CLA and in none by ELISA. IgG-AI and IgM-AI were not detected within the first week of disease. Duration of disease correlated with IgG-AI while IgM-AI results were heterogeneous for each test assay. Moreover, the levels of IgG-AI, but not IgM-AI, correlated with protein concentration in CSF. IgG is the relevant immunoglobulin isotype for detecting intrathecal synthesis of borrelia antibodies. The highest sensitivity and specificity were achieved by the antibody detection assay using VlsE IR6 peptide. Detection of IgM-AI yielded heterogenous results and did not support the laboratory diagnosis of LNB.

Detection of post-exercise stunning by early gated SPECT myocardial perfusion imaging: results from the IAEA multi-center study.

BACKGROUND: Transient post-ischemic LV dysfunction due to myocardial stunning in patients with coronary artery disease can be missed by conventional gated SPECT (GSPECT) acquisitions. The aim of this IAEA-sponsored multi-center study was to determine whether early post-exercise imaging is more likely to detect stunning than conventional without adversely affecting image quality or perfusion information. METHODS AND RESULTS: Patients undergoing exercise/rest GSPECT were enrolled in this international multicenter study. Post-exercise studies were acquired at 15 ± 5 minutes after radiotracer injection (Stress-1) and repeated at 60 ± 15 minutes (Stress-2). Rest studies (R) were acquired at 60 minutes post injection. A core laboratory quantitatively assessed perfusion pattern and LV blinded to the acquisition time. Ischemia was defined as summed stress score (SDS) ≥4, and stunning was defined as the difference between rest and post-stress LVEF (Δ-LVEF). In the 229 patients enrolled into the study, both image quality and perfusion information were similar between Stress-1 and Stress-2. Post-stress LVEF was associated with both ischemia and time of acquisition, with a significant correlation between SDS and Δ-LVEF, which was stronger at Stress-1 than Stress-2 in the ischemic compared to the non-ischemic population (r = 0.23 vs 0.08, P = 0.10). CONCLUSIONS: Early post-exercise imaging is feasible, and can potentially improve the detection of post-ischemic stunning without compromising image quality and perfusion data.

Resting heart rate variability and heart rate recovery after submaximal exercise.

PURPOSE: Aerobic training accelerates Heart Rate Recovery after exercise in healthy subjects and in patients with coronary disease. As shown by pharmacological autonomic blockade, HRR early after exercise is dependent primarily on parasympathetic reactivation. Thus, accelerated HRR early after exercise in endurance-trained athletes may be attributed to augmented parasympathetic reactivation. In the present study, we tested the hypothesis that the HRR early after submaximal exercise is related to the pre-exercise parasympathetic modulation. METHODS: Thirty endurance-trained athletes (20 males, 50 ± 7 years) and thirty control subjects (20 males, 52 ± 6 years) performed a submaximal exercise on a cyclo-ergometer. Pre-exercise resting short-term heart rate variability (HRV) parameters in time and frequency-domains were correlated with HRR during the first 30 s, 1 and 2 min after cessation of exercise. RESULTS: We found that HRR was statistically significantly faster in athletes than in controls at all examination time points (p < 0.05). HF, SDNN and RMSSD were statistically significantly higher in athletes than in controls (p < 0.05), but other resting HRV parameters were not statistically different between groups. After 30 s, 1 and 2 min of recovery, HRR correlation with total power, HF, HFnu and RMSSD was positive, while the correlation with LF/HF was negative for small and positive for larger values. The opposite was true for SDNN. CONCLUSIONS: These findings support the hypothesis that HRR early after submaximal exercise is related to resting parasympathetic modulation in the middle-aged subjects. In addition, they suggested an optimal range of HRV for maximal HRR after exercise.

Initial data analysis for longitudinal studies to build a solid foundation for reproducible analysis.

Initial data analysis (IDA) is the part of the data pipeline that takes place between the end of data retrieval and the beginning of data analysis that addresses the research question. Systematic IDA and clear reporting of the IDA findings is an important step towards reproducible research. A general framework of IDA for observational studies includes data cleaning, data screening, and possible updates of pre-planned statistical analyses. Longitudinal studies, where participants are observed repeatedly over time, pose additional challenges, as they have special features that should be taken into account in the IDA steps before addressing the research question. We propose a systematic approach in longitudinal studies to examine data properties prior to conducting planned statistical analyses. In this paper we focus on the data screening element of IDA, assuming that the research aims are accompanied by an analysis plan, meta-data are well documented, and data cleaning has already been performed. IDA data screening comprises five types of explorations, covering the analysis of participation profiles over time, evaluation of missing data, presentation of univariate and multivariate descriptions, and the depiction of longitudinal aspects. Executing the IDA plan will result in an IDA report to inform data analysts about data properties and possible implications for the analysis plan-another element of the IDA framework. Our framework is illustrated focusing on hand grip strength outcome data from a data collection across several waves in a complex survey. We provide reproducible R code on a public repository, presenting a detailed data screening plan for the investigation of the average rate of age-associated decline of grip strength. With our checklist and reproducible R code we provide data analysts a framework to work with longitudinal data in an informed way, enhancing the reproducibility and validity of their work.

Improved shrunken centroid classifiers for high-dimensional class-imbalanced data.

BACKGROUND: PAM, a nearest shrunken centroid method (NSC), is a popular classification method for high-dimensional data. ALP and AHP are NSC algorithms that were proposed to improve upon PAM. The NSC methods base their classification rules on shrunken centroids; in practice the amount of shrinkage is estimated minimizing the overall cross-validated (CV) error rate. RESULTS: We show that when data are class-imbalanced the three NSC classifiers are biased towards the majority class. The bias is larger when the number of variables or class-imbalance is larger and/or the differences between classes are smaller. To diminish the class-imbalance problem of the NSC classifiers we propose to estimate the amount of shrinkage by maximizing the CV geometric mean of the class-specific predictive accuracies (g-means). CONCLUSIONS: The results obtained on simulated and real high-dimensional class-imbalanced data show that our approach outperforms the currently used strategy based on the minimization of the overall error rate when NSC classifiers are biased towards the majority class. The number of variables included in the NSC classifiers when using our approach is much smaller than with the original approach. This result is supported by experiments on simulated and real high-dimensional class-imbalanced data.