A double-Cox model for non-proportional hazards survival analysis with frailty.

The Cox regression, a semi-parametric method of survival analysis, is extremely popular in biomedical applications. The proportional hazards assumption is a key requirement in the Cox model. To accommodate non-proportional hazards, we propose to parameterize the shape parameter of the baseline hazard function using the additional, separate Cox-regression term which depends on the vector of the covariates. This parametrization retains the general form of the hazard function over the strata and is similar to one in Devarajan and Ebrahimi (Comput Stat Data Anal. 2011;55:667-676) in the case of the Weibull distribution, but differs for other hazard functions. We call this model the double-Cox model. We formally introduce the double-Cox model with shared frailty and investigate, by simulation, the estimation bias and the coverage of the proposed point and interval estimation methods for the Gompertz and the Weibull baseline hazards. For real-life applications with low frailty variance and a large number of clusters, the marginal likelihood estimation is almost unbiased and the profile likelihood-based confidence intervals provide good coverage for all model parameters. We also compare the results from the over-parametrized double-Cox model to those from the standard Cox model with frailty in the case of the scale-only proportional hazards. The model is illustrated on an example of the survival after a diagnosis of type 2 diabetes mellitus. The R programs for fitting the double-Cox model are available on Github.

Risk-adjusted cUSUM control charts for shared frailty survival models with application to hip replacement outcomes: a study using the NJR dataset.

BACKGROUND: Continuous monitoring of surgical outcomes after joint replacement is needed to detect which brands' components have a higher than expected failure rate and are therefore no longer recommended to be used in surgical practice. We developed a monitoring method based on cumulative sum (CUSUM) chart specifically for this application. METHODS: Our method entails the use of the competing risks model with the Weibull and the Gompertz hazard functions adjusted for observed covariates to approximate the baseline time-to-revision and time-to-death distributions, respectively. The correlated shared frailty terms for competing risks, corresponding to the operating unit, are also included in the model. A bootstrap-based boundary adjustment is then required for risk-adjusted CUSUM charts to guarantee a given probability of the false alarm rates. We propose a method to evaluate the CUSUM scores and the adjusted boundary for a survival model with the shared frailty terms. We also introduce a unit performance quality score based on the posterior frailty distribution. This method is illustrated using the 2003-2012 hip replacement data from the UK National Joint Registry (NJR). RESULTS: We found that the best model included the shared frailty for revision but not for death. This means that the competing risks of revision and death are independent in NJR data. Our method was superior to the standard NJR methodology. For one of the two monitored components, it produced alarms four years before the increased failure rate came to the attention of the UK regulatory authorities. The hazard ratios of revision across the units varied from 0.38 to 2.28. CONCLUSIONS: An earlier detection of failure signal by our method in comparison to the standard method used by the NJR may be explained by proper risk-adjustment and the ability to accommodate time-dependent hazards. The continuous monitoring of hip replacement outcomes should include risk adjustment at both the individual and unit level.

Associations between inflammation-related biomarkers and depressive symptoms in individuals with recently diagnosed type 1 and type 2 diabetes.

Depressive disorders represent a frequent comorbidity of both type 1 (T1D) and type 2 diabetes (T2D). Inflammation-related processes have been implicated in the development of both diabetes and depression. This study aimed to investigate whether biomarkers of subclinical inflammation were associated with depressive symptoms in individuals with recently diagnosed diabetes and if such associations differed by diabetes type. This cross-sectional study was based on 295 individuals with T2D (67% men, mean age 53years) and 139 individuals with T1D (60% men, mean age 36years) of the German Diabetes Study. The main inclusion criterion was a known disease duration of <1year. Depressive symptoms were assessed with the Allgemeine Depressionsskala, Langversion (ADS-L) questionnaire, the German version of the Center for Epidemiological Studies Depression scale (CES-D) questionnaire. Associations between biomarkers of subclinical inflammation and the ADS-L as continuous score were assessed using multiple linear regression models adjusting for age, sex, body mass index, HbA1c, lipids, hypertension, medication and comorbidities. Serum high-sensitivity C-reactive protein (hsCRP) and the ratio of high-molecular-weight (HMW)/total adiponectin were positively associated with ADS-L in T2D (both P<0.01), but not in T1D. In contrast, serum levels of soluble intercellular adhesion molecule (sICAM)-1 were positively associated with ADS-L only in T1D (P=0.035). The latter association was significantly different between both diabetes types (Pinteraction=0.036). No associations were observed for interleukin (IL)-6, IL-18 and soluble E-selectin. Only the association between HMW/total adiponectin and ADS-L in T2D remained significant after correction for multiple testing. In conclusion, our study shows that the ratio HMW/total adiponectin is associated with depressive symptoms in individuals with recently diagnosed T2D. It also provides suggestive evidence that further biomarkers of subclinical inflammation and endothelial activation may be associated with depressive symptoms in individuals with recently diagnosed T1D and T2D.

Joint analysis of bivariate competing risks survival times and genetic markers data.

Bivariate survival models with discretely distributed frailty based on the major gene concept and applied to the data on related individuals such as twins and sibs can be used to estimate the underlying hazard, the relative risk and the frequency of the longevity allele. To determine the position of the longevity gene, additional genetic markers data are needed. If the action of the longevity allele does not depend on its position in the genome, these two problems can be solved separately using a two-step procedure. We proposed an extension of this method allowing us to search the position of two longevity genes at a chromosome using the bivariate survival data with correlated competing risks combined with genetic markers data. We have studied the properties of the model with two longevity genes located on the same and on different chromosomes using simulated data sets.

Twin study indicates loss of interaction between microbiota and mucosa of patients with ulcerative colitis.

BACKGROUND & AIMS: Interactions between genetic and environmental factors are believed to be involved in onset and initiation of inflammatory bowel disease. We analyzed the interaction between gastrointestinal mucosal microbiota and host genes in twin pairs discordant for ulcerative colitis (UC) to study the functional interaction between microbiota and mucosal epithelium. METHODS: Biopsy were collected from sigmoid colon of UC patients and their healthy twins (discordant twin pairs) and from twins without UC. Microbiota profiles were determined from analysis of 16S ribosomal DNA libraries; messenger RNA profiles were determined by microarray analysis. RESULTS: Patients with UC had dysbiotic microbiota, characterized by less bacterial diversity and more Actinobacteria and Proteobacteria than that of their healthy siblings; healthy siblings from discordant twins had more bacteria from the Lachnospiraceae and Ruminococcaceae families than twins who were both healthy. In twins who were both healthy, 34 mucosal transcripts correlated with bacterial genera, whereas only 25 and 11 correlated with bacteria genera in healthy individuals and their twins with UC, respectively. Transcripts related to oxidative and immune responses were differentially expressed between patients with UC and their healthy twins. CONCLUSIONS: The transcriptional profile of the mucosa appears to interact with the colonic microbiota; this interaction appears to be lost in colon of patients with UC. Bacterial functions, such as butyrate production, might affect mucosal gene expression. Patients with UC had different gene expression profiles and lower levels of biodiversity than their healthy twins, as well as unusual aerobic bacteria. Patients with UC had lower percentages of potentially protective bacterial species than their healthy twins.

Exploring Structural Flexibility and Stability of α-Synuclein by the Landau-Ginzburg-Wilson Approach.

α-Synuclein (αS) is the principal protein component of the Lewy body and Lewy neurite deposits that are found in the brains of the victims of one of the most prevalent neurodegenerative disorders, Parkinson's disease. αS can be qualified as a chameleon protein because of the large number of different conformations that it is able to adopt: it is disordered under physiological conditions in solution, in equilibrium with a minor α-helical tetrameric form in the cytoplasm, and is α-helical when bound to a cell membrane. Also, in vitro, αS forms polymorphic amyloid fibrils with unique arrangements of cross-ß-sheet motifs. Therefore, it is of interest to elucidate the origins of the structural flexibility of αS and what makes αS stable in different conformations. We address these questions here by analyzing the experimental structures of the micelle-bound, tetrameric, and fibrillar αS in terms of a kink (heteroclinic standing wave solution) of a generalized discrete nonlinear Schrödinger equation. It is illustrated that without molecular dynamics simulations the kinks are capable of identifying the key residues causing structural flexibility of αS. Also, the stability of the experimental structures of αS is investigated by simulating heating/cooling trajectories using the Glauber algorithm. The findings are consistent with experiments.

On topology and knotty entanglement in protein folding.

We investigate aspects of topology in protein folding. For this we numerically simulate the temperature driven folding and unfolding of the slipknotted archaeal virus protein AFV3-109. Due to knottiness the (un)folding is a topological process, it engages the entire backbone in a collective fashion. Accordingly we introduce a topological approach to model the process. Our simulations reveal that the (un)folding of AFV3-109 slipknot proceeds through a folding intermediate that has the topology of a trefoil knot. We observe that the final slipknot causes a slight swelling of the folded AFV3-109 structure. We disclose the relative stability of the strands and helices during both the folding and unfolding processes. We confirm results from previous studies that pointed out that it can be very demanding to simulate the formation of knotty self-entanglement, and we explain how the problems are circumvented: The slipknotted AFV3-109 protein is a very slow folder with a topologically demanding pathway, which needs to be properly accounted for in a simulation description. When we either increase the relative stiffness of bending, or when we decrease the speed of ambient cooling, the rate of slipknot formation rapidly increases.

Investigation of Phosphorylation-Induced Folding of an Intrinsically Disordered Protein by Coarse-Grained Molecular Dynamics.

Apart from being the most common mechanism of regulating protein function and transmitting signals throughout the cell, phosphorylation has an ability to induce disorder-to-order transition in an intrinsically disordered protein. In particular, it was shown that folding of the intrinsically disordered protein, eIF4E-binding protein isoform 2 (4E-BP2), can be induced by multisite phosphorylation. Here, the principles that govern the folding of phosphorylated 4E-BP2 (pT37pT46 4E-BP218-62) are investigated by analyzing canonical and replica exchange molecular dynamics trajectories, generated with the coarse-grained united-residue force field, in terms of local and global motions and the time dependence of formation of contacts between Cαs of selected pairs of residues. The key residues involved in the folding of the pT37pT46 4E-BP218-62 are elucidated by this analysis. The correlations between local and global motions are identified. Moreover, for a better understanding of the physics of the formation of the folded state, the experimental structure of the pT37pT46 4E-BP218-62 is analyzed in terms of a kink (heteroclinic standing wave solution) of a generalized discrete nonlinear Schrödinger equation. It is shown that without molecular dynamics simulations the kinks are able to identify not only the phosphorylated sites of protein, the key players in folding, but also the reasons for the weak stability of the pT37pT46 4E-BP218-62.

Detecting genes contributing to longevity using twin data.

Searching for genes contributing to longevity is a typical task in association analysis. A number of methods can be used for finding this association - from the simplest method based on the technique of contingency tables to more complex algorithms involving demographic data, which allow us to estimate the genotype-specific hazard functions. The independence of individuals is the common assumption in all these methods. At the same time, data on related individuals such as twins are often used in genetic studies. This paper proposes an extension of the relative risk model to encompass twin data. We estimate the power and also discuss what happens if we treat the twin data using the univariate model.

Dynamic early identification of hip replacement implants with high revision rates. Study based on the NJR data from UK during 2004-2012.

BACKGROUND: Hip replacement and hip resurfacing are common surgical procedures with an estimated risk of revision of 4% over 10 year period. Approximately 58% of hip replacements will last 25 years. Some implants have higher revision rates and early identification of poorly performing hip replacement implant brands and cup/head brand combinations is vital. AIMS: Development of a dynamic monitoring method for the revision rates of hip implants. METHODS: Data on the outcomes following the hip replacement surgery between 2004 and 2012 was obtained from the National Joint Register (NJR) in the UK. A novel dynamic algorithm based on the CUmulative SUM (CUSUM) methodology with adjustment for casemix and random frailty for an operating unit was developed and implemented to monitor the revision rates over time. The Benjamini-Hochberg FDR method was used to adjust for multiple testing of numerous hip replacement implant brands and cup/ head combinations at each time point. RESULTS: Three poorly performing cup brands and two cup/ head brand combinations have been detected. Wright Medical UK Ltd Conserve Plus Resurfacing Cup (cup o), DePuy ASR Resurfacing Cup (cup e), and Endo Plus (UK) Limited EP-Fit Plus Polyethylene cup (cup g) showed stable multiple alarms over the period of a year or longer. An addition of a random frailty term did not change the list of underperforming components. The model with added random effect was more conservative, showing less and more delayed alarms. CONCLUSIONS: Our new algorithm is an efficient method for early detection of poorly performing components in hip replacement surgery. It can also be used for similar tasks of dynamic quality monitoring in healthcare.

Genetic control of global gene expression levels in the intestinal mucosa: a human twin study.

Phenotypic variation between individuals, such as different mRNA expression levels, is influenced by genetic and nongenetic factors. Although several studies have addressed the interplay between genotypes and expression profiles in various model organisms in the recent years, the detailed and relative contributions of genetic and nongenetic factors in regulating plasticity of gene expression in barrier organs (e.g., skin, gut), which are exposed to continuous environmental challenge, are still poorly understood. Here we systematically monitored the level of genetic control over genomewide mRNA expression profiles in the healthy intestinal mucosa of 10 monozygotic and 10 dizygotic human twin pairs with microarray analyses. Our results, which are supported by real-time PCR and analysis of molecular phylogenetic conservation, indicate that genes associated with energy metabolism and cell and tissue regeneration pathways are under strong genetic control. Conversely, genes associated with immune response seem to be mainly controlled by exogenous factors. Further insights into the relative extent of genetic and nongenetic determinants of transcriptomal profiles and their influence on physiological and pathophysiological mechanisms are crucial to understanding the key role played by gene-environment interactions in health and disease.

Protein tertiary structure and the myoglobin phase diagram.

We develop an effective theory approach to investigate the phase properties of globular proteins. Instead of interactions between individual atoms or localized interaction centers, the approach builds directly on the tertiary structure of a protein. As an example we construct the phase diagram of (apo)myoglobin with temperature (T) and acidity (pH) as the thermodynamical variables. We describe how myoglobin unfolds from the native folded state to a random coil when temperature and acidity increase. We confirm the presence of two molten globule folding intermediates, and we predict an abrupt transition between the two when acidity changes. When temperature further increases we find that the abrupt transition line between the two molten globule states terminates at a tricritical point, where the helical structures fade away. Our results also suggest that the ligand entry and exit is driven by large scale collective motions that destabilize the myoglobin F-helix.

Dynamics of the mucosa-associated flora in ulcerative colitis patients during remission and clinical relapse.

The colonic mucosa-associated flora (MAF) in patients with active ulcerative colitis (UC) (n = 13) was investigated by examining 16S rRNA gene signatures during remission and relapse against levels for controls (n = 5). Baseline reduction, temporal instability, and decrease of bacterial richness toward relapse were observed for UC patients, whereas the MAF for controls was stable over time.

How do patients with diabetes report their comorbidities? Comparison with administrative data.

AIMS: Patients with diabetes are probably often unaware of their comorbidities. We estimated agreement between self-reported comorbidities and administrative data. METHODS: In a random sample of 464 diabetes patients, data from a questionnaire asking about the presence of 14 comorbidities closely related to diabetes were individually linked with statutory health insurance data. RESULTS: Specificities were >97%, except cardiac insufficiency (94.5%), eye diseases (93.8%), peripheral arterial disease (92.6%), hypertension (90.9%), and peripheral neuropathy (85.8%). Sensitivities were <60%, except amputation (100%), hypertension (83.1%), and myocardial infarction (67.2%). A few positive predictive values were >90% (hypertension, myocardial infarction, and eye disease), and six were below 70%. Six negative predictive values were >90%, and two <70% (hypertension and eye disease). Total agreement was between 42.7% (eye disease) and 100% (dialysis and amputation). Overall, substantial agreement was observed for three morbidities (kappa 0.61-0.80: hypertension, myocardial infarction, and amputation). Moderate agreement (kappa 0.41-0.60) was estimated for angina pectoris, heart failure, stroke, peripheral neuropathy, and kidney disease. Factors associated with agreement were the number of comorbidities, diabetes duration, age, sex, and education. CONCLUSIONS: Myocardial infarction and amputation were well reported by patients as comorbidities; eye diseases and foot ulceration rather poorly, particularly in older, male, or less educated patients. Patient information needs improving.

What information needs do people with recently diagnosed diabetes mellitus have and what are the associated factors? A cross-sectional study in Germany.

OBJECTIVES: This study aimed to identify: (1) information needs of people with recently diagnosed type 1 or type 2 diabetes mellitus (DM); (2) information needs within different subgroups; and (3) factors associated with information needs concerning DM such as current level of information, health-related quality of life or participation preferences. DESIGN: A mixed-method approach combining quantitative and qualitative methods was used. Information needs for different topics and estimated associated factors were described using logistic regression models. Additionally, a qualitative content analysis was performed. SETTING: Monocentre study. PARTICIPANTS: Information needs were assessed and analysed in 138 consecutive participants with DM who took part in the German Diabetes Study (54% type 2 diabetes, 64% male, mean age 46.3±12.3 years, known diabetes duration <1 year). RESULTS: Most participants displayed a need for information in all topics provided, especially in diabetes research (86%) and treatment/therapy (80%). Regarding those topics, participants wished for information regarding new treatments that simplify their everyday life. In general, participants preferred topics that focus on the management or handling of DM over topics related to clinical factors of DM, such as causes and complications. A low current level of information and treatment with antihyperglycaemic medication were significantly associated with higher information needs, and diabetes-related comorbidity and higher mental component summary score in the 36-Item Short-Form Health Survey (SF-36) with lower information needs. CONCLUSION: People with recently diagnosed DM display high information needs, which differ according to the current level of information, mode of diabetes treatment, diabetes-related comorbidity and mental component summary score in the SF-36. There appears to be a preference for information, which can help to simplify life with diabetes and for information that corresponds to their level of knowledge. This should be considered in patient information activities. TRIAL REGISTRATION NUMBER: NCT01055093.

Robust method for detecting differential gene expression in twin studies.

MOTIVATION: A steadily increasing number of experiments with microarrays stimulate the further development of the statistical methods of the analysis of gene expression data. One of the central problems in this area is detecting differential gene expression under two or more conditions. Unfortunately, up to now it has not been studied how the correlations between related individuals, such as twins influence the estimates of differential gene expression. RESULTS: In this paper, we discuss this problem and propose a new method that is robust with respect to correlations of gene expression data for twins.

Time cost of diabetes: Development of a questionnaire to assess time spent on diabetes self-care.

BACKGROUND: Methods to measure patient time spent on health-related activities are currently not well elaborated or standardized. AIM: The purpose of this study was to develop a recall questionnaire measuring patient time devoted to diabetes self-care and to examine its feasibility and validity under field conditions. METHODS: The initial questionnaire was developed on the basis of instruments frequently used to assess self-care behavior in patients with diabetes, evaluated in two focus groups with patients with type 2 diabetes (N=15) and tested in a random sample of patients with type 2 diabetes (N=178). To assess the validity of the questionnaire, four hypotheses about expected differences in self-care time across various patient sub-groups were tested. RESULTS: The final questionnaire includes thirteen items estimating time spent on regular diabetes-related activities undertaken in the previous seven days. 78% of respondents completed the questionnaire without item non-response or other evident problems. As hypothesized, respondents receiving insulin treatment, those with poor self-rated health and those with diabetes-related emotional distress (PAID-5 score ≥8) reported spending more time on diabetes self-care than the rest of the sample. Contrary to our assumption, no differences in time spent on diabetes self-care between employed and retired individuals were detected by the questionnaire. CONCLUSION: The recall questionnaire measuring patient time devoted to a broad range of regular diabetes self-care activities was developed and its feasibility was proved under field conditions. Ideally, the questionnaire should be further validated within a variety of populations. Exploration of the convergent validity between the recall method and prospective diary may be also useful.

Agreement found between self-reported and health insurance data on physician visits comparing different recall lengths.

OBJECTIVE: To analyze the impact of different recall lengths on agreement between self-reported physician visits and those documented in health insurance data applying an experimental design. STUDY DESIGN AND SETTING: We randomly assigned 432 patients with diabetes to one of two versions of a written survey, each asking about the number of physician visits over a 3- or 6-month recall period. Health insurance data were linked individually. RESULTS: In both groups, the mean number of self-reported physician visits per month was lower than in the insurance data, with a larger difference in the 6-month group (-0.9; 95% CI -1.0, -0.7) than in the 3-month group (-0.5; -0.7; -0.2), difference between the two groups: 0.4 (0.1-0.7; P = 0.009). The percentage of participants with correct reporting was small and did not differ largely between the two groups (6.5% and 9.3%). However, there was more overreporting in the 3-month group (25.6% vs. 11.1%). CONCLUSIONS: Shorter recall periods may produce more accurate results when estimating the mean number of physician visits. However, this may be driven not by a more accurate reporting, but by a higher proportion of respondents that overreported and a lower proportion of respondents that underreported, when compared to the longer reporting period.

Study of Disease Progression and Relevant Risk Factors in Diabetic Foot Patients Using a Multistate Continuous-Time Markov Chain Model.

The diabetic foot is a lifelong disease. The longer patients with diabetes and foot ulcers are observed, the higher the likelihood that they will develop comorbidities that adversely influence ulcer recurrence, amputation and survival (for example peripheral arterial disease, renal failure or ischaemic heart disease). The purpose of our study was to quantify person and limb-related disease progression and the time-dependent influence of any associated factors (present at baseline or appearing during observation) based on which effective prevention and/or treatment strategies could be developed. Using a nine-state continuous-time Markov chain model with time-dependent risk factors, all living patients were divided into eight groups based on their ulceration (previous or current) and previous amputation (none, minor or major) status. State nine is an absorbing state (death). If all transitions are fully observable, this model can be decomposed into eight submodels, which can be analyzed using the methods of survival analysis for competing risks. The dependencies of the risk factors (covariates) were included in the submodels using Cox-like regression. The transition intensities and relative risks for covariates were calculated from long-term data of patients with diabetic foot ulcers collected in a single specialized center in North-Rhine Westphalia (Germany). The detected estimates were in line with previously published, but scarce, data. Together with the interesting new results obtained, this indicates that the proposed model may be useful for studying disease progression in larger samples of patients with diabetic foot ulcers.

Genetic markers data in survival studies of twins: the results of a simulation study.

Previous longevity studies of related individuals such as twins or siblings based on the major gene model have shown that the frequency and the relative risk of mortality of a beneficial allele in the population could be estimated. If, in addition to survival data for related individuals, the genetic markers data are available, one could try to locate the longevity allele in the genome. In the case where the phenotypic trait is life span or age at onset of disease, a two-step procedure can be used. First, the parameters of bivariate survival functions must be estimated from bivariate survival data for twins without markers. The second step is focused on determining the position of longevity genes between respective markers. To calculate the joint distribution of inheritance vector and genetic markers, the hidden Markov chain technique is used. This approach is illustrated with a simulation example for one longevity gene.