Multistate Model Research Articles

Longitudinal patterns of use of stimulants and opioids in the AIDS linked to the IntraVenous experience cohort, 2005–2019

BackgroundOverdoses involving opioids and stimulants are on the rise, yet few studies have examined longitudinal trends in use of both substances. We sought to describe use and co-use of opioids and stimulants, 2005–2019, in the AIDS Linked to the Intravenous Experience (ALIVE) cohort – a community-based cohort of people with a history of injection drug use living in or near Baltimore, MD. MethodsWe included 2083 ALIVE participants, who had at least two visits during the study period. Our outcome was based on self-reported use of opioids and stimulants in the prior 6 months. We estimated prevalence of 4 categories of use (neither stimulants nor opioids, only stimulants, only opioids, stimulants and opioids), using a non-parametric multi-state model, accounting for the competing event of death and weighting for informative loss to follow-up. All analyses were stratified by enrollment cohort, with the main analysis including participants who enrolled prior to 2015 and a sub-analysis including participants who enrolled 2015–2018. ResultsIn the main analysis, prevalence of using stimulants and opioids decreased from 38 % in 2005 to 12 % 2013 but stabilized from 2014 onwards (13–19 %). The prevalence of using only stimulants (7–11 %) and only opioids (5–10 %) was stable across time. Participants who reported using both were more likely to report homelessness, depression, and other substance use (e.g., marijuana and heavy alcohol use) than participants in the other use categories. On average, 65 % of visits with use of both were followed by a subsequent visit with use of both; of participants transitioning out of using both, 13% transitioned to using neither. ConclusionsWhile use of stimulants and opioids declined in the cohort through 2013, a meaningful proportion of participants persistently used both. More research is needed to understand and develop strategies to mitigate harms associated with persistent use of both stimulants and opioids.

Association between diet quality and ovarian cancer risk and survival.

Research on diet quality and ovarian cancer is limited. We examined the association between diet quality and ovarian cancer risk and survival in a large prospective cohort. We used data from women in the prospective National Institutes of Health-AARP Diet and Health Study enrolled from 1995 to 1996 who were aged 50-71 years at baseline with follow-up through December 31, 2017. Participants completed a 124-item food frequency questionnaire at baseline, and diet quality was assessed via the Healthy Eating Index-2015, the alternate Mediterranean diet score, and the Dietary Approaches to Stop Hypertension score. Primary outcomes were first primary epithelial ovarian cancer diagnosis from cancer registry data and among those diagnosed with ovarian cancer all-cause mortality. We used a semi-Markov multistate model with Cox proportional hazards regression to account for semicompeting events. Among 150 643 participants with a median follow-up time of 20.5 years, 1107 individuals were diagnosed with a first primary epithelial ovarian cancer. There was no evidence of an association between diet quality and ovarian cancer risk. Among those diagnosed with epithelial ovarian cancer, 893 deaths occurred with a median survival of 2.5 years. Better prediagnosis diet quality, according to the Healthy Eating Index-2015 (quintile 5 vs quintile 1: hazard ratio [HR] = 0.75, 95% confidence interval [CI] = 0.60 to 0.93) and alternate Mediterranean diet score (quintile 5 vs quintile 1: HR = 0.68, 95% CI = 0.53 to 0.87), was associated with lower all-cause mortality. There was no evidence of an association between Dietary Approaches to Stop Hypertension score and all-cause mortality. Better prediagnosis diet quality was associated with lower all-cause mortality after ovarian cancer diagnosis but was not associated with ovarian cancer risk.

An Italian Real-World Study Highlights the Importance of Some Clinicopathological Characteristics Useful in Identifying Metastatic Breast Cancer Patients Resistant to CDK4/6 Inhibitors and Hormone Therapy.

Background: Cyclin-dependent kinase 4 and 6 (CDK4 and CDK6) inhibitors have changed the therapeutic management of hormone receptor-positive (HR+) metastatic breast cancer (mBC) by targeting the cell cycle machinery and overcoming endocrine resistance. However, a large number of patients present disease progression due to cancer cells resisting CDK4/6 inhibitors. Our research considers which clinicopathological characteristics could be useful in identifying patients who might respond to CDK4/6 inhibitors by analyzing a retrospective case series of patients with HR+ mBC who were treated with hormone therapy plus CDK4/6 inhibitors. Methods: Approximately 177 mBC patients were enrolled, of whom 66 were treated with CD4/6 inhibitors plus letrozole and 111 were treated with CDK4/6 inhibitors and fulvestrant. A multistate model was used. Results: A low body surface area and older age were associated with an increased risk of developing neutropenia. A high Ki67 index, the presence of visceral metastases, and not having previously undergone adjuvant chemotherapy were prognostic factors of disease progression/death. As expected, some of the neutropenic patients who had previously undergone multiple lines of treatment were at a higher risk of disease progression/death. Furthermore, neutropenia status was associated with a more than doubled risk of progression/death compared to patients without neutropenia (HR = 2.311; p = 0.025). Conclusions: Having identified certain factors that could be associated with the development of neutropenia and considering that neutropenia itself is associated with an increased risk of progression, we believe that the baseline characteristics should be taken into account to reduce cases of neutropenia and disease progression.

Determinants of COVID-19 Infection Among Employees of an Italian Financial Institution.

Understanding the trend of the severe acute respiratory syndrome Coronavirus 2 (SARS-CoV-2) is becoming crucial. Previous studies focused on predicting COVID-19 trends, but few papers have considered models for disease estimation and progression based on large real-world data. We used de-identified data from 60,938 employees of a major financial institution in Italy with daily COVID-19 status information between 31 March 2020 and 31 August 2021. We consider six statuses: (i) concluded case, (ii) confirmed case, (iii) close contact, (iv) possible-probable contact, (v) possible contact, and (vi) no-COVID-19 or infection. We conducted a logistic regression to assess the odds ratio (OR) of transition to confirmed COVID-19 case at each time point. We also fitted a general model for disease progression via the multi-state transition probability model at each time point, with lags of 7 and 15 days. Employment in a branch versus in a central office was the strongest predictor of case or contact status, while no association was detected with gender or age. The geographic prevalence of possible-probable contacts and close contacts was predictive of the subsequent risk of confirmed cases. The status with the highest probability of becoming a confirmed case was concluded case (12%) in April 2020, possible-probable contact (16%) in November 2020, and close contact (4%) in August 2021. The model based on transition probabilities predicted well the rate of confirmed cases observed 7 or 15 days later. Data from industry-based surveillance systems may effectively predict the risk of subsequent infection.

Competing risk multistate censored data modeling by propensity score matching method

The potential contribution of the paper is the use of the propensity score matching method for updating censored observations within the context of multi-state model featuring two competing risks.The competing risks are modelled using cause-specific Cox proportional hazard model.The simulation findings demonstrate that updating censored observations tends to lead to reduced bias and mean squared error for all estimated parameters in the risk of cause-specific Cox model.The results for a chemoradiotherapy real dataset are consistent with the simulation results.

Discord in the voter model for complex networks.

Online social networks have become primary means of communication. As they often exhibit undesirable effects such as hostility, polarization, or echo chambers, it is crucial to develop analytical tools that help us better understand them. In this paper we are interested in the evolution of discord in social networks. Formally, we introduce a method to calculate the probability of discord between any two agents in the multistate voter model with and without zealots. Our work applies to any directed, weighted graph with any finite number of possible opinions, allows for various update rates across agents, and does not imply any approximation. Under certain topological conditions, the opinions are independent and the joint distribution can be decoupled. Otherwise, the evolution of discord probabilities is described by a linear system of ordinary differential equations. We prove the existence of a unique equilibrium solution, which can be computed via an iterative algorithm. The classical definition of active links density is generalized to take into account long-range, weighted interactions. We illustrate our findings on real-life and synthetic networks. In particular, we investigate the impact of clustering on discord and uncover a rich landscape of varied behaviors in polarized networks. This sheds lights on the evolution of discord between, and within, antagonistic communities.

Weight Loss Treatment and Longitudinal Weight Change Among Primary Care Patients With Obesity

Among individuals with obesity, 5% or greater weight loss can improve health. Weight management treatments (WMT) include nutrition counseling, very low-calorie meal replacement (MR), antiobesity medications (AOM), and bariatric surgery; however, little is known about how these WMT are associated with weight change among individual patients and populations. To characterize weight status and WMT use among primary care patients and assess associations between WMT and weight trajectories. Retrospective, population-based cohort study of primary care patients from 1 academic health system in Michigan between October 2015 and March 2020 using cross-sectional analysis to compare obesity prevalence and WMT utilization. For patients with obesity and WMT exposure or matched controls, a multistate Markov model assessing associations between WMT and longitudinal weight status trajectories was used. Data were analyzed from October 2021 to October 2023. Cross-sectional exposure was year: 2017 or 2019. Trajectory analysis exposures were WMT: nutrition counseling, MR, AOM, and bariatric surgery. Cross-sectional analysis compared mean body mass index (BMI), obesity prevalence, and, among patients with obesity, prospective WMT use. The trajectory analysis examined longitudinal weight status using thresholds of ±5% and 10% of baseline weight with primary outcomes being the 1-year probabilities of 5% or greater weight loss for each WMT. Adult patients (146 959 participants) consisted of 83 636 female participants (56.9%); 8940 (6.1%) were Asian, 14 560 (9.9%) were Black, and 116 664 (79.4%) were White. Patients had a mean (SD) age of 49.6 (17.7) years and mean (SD) BMI of 29.2 (7.2). Among 138 682 patients, prevalence of obesity increased from 39.2% in 2017 to 40.7% in 2019; WMT use among patients with obesity increased from 5.3% to 7.1% (difference: 1.7%; 95% CI, 1.3%-2.2%). In a multistate model (10 180 patients; 33 549 patient-years), the 1-year probability of 5% or greater weight loss without WMT exposure was 15.6% (95% CI, 14.3%-16.5%) at reference covariates. In contrast, the probability of 5% or greater weight loss was more likely with year-long exposures to any WMT (nutrition counseling: 23.1%; 95% CI, 21.3%-25.1%; MR: 54.6%; 95% CI, 46.5%-61.2%; AOM: 27.8%; 95% CI, 25.0%-30.5%; bariatric surgery: 93.0%; 95% CI, 89.7%-95.0%). In this cohort study of primary-care patients with obesity, all WMT increased the patient-level probability of achieving 5% or greater weight loss, but current rates of utilization are low and insufficient to reduce weight at the population level.

Impact of early cyclosporine A levels on acute graft-versus-host disease in allogeneic hematopoietic stem cell transplantation using in vivo T-cell depletion

Background aimsCyclosporin A (CsA) remains a major component of immunosuppressive regimens applied in allogeneic hematopoietic stem cell transplantation (HSCT). The impact of CsA trough levels during the first weeks after HSCT has not yet been investigated specifically in anti-T-lymphocyte globulin (ATLG)-based HSCT from matched related and unrelated donors. MethodsTo address this issue, we have retrospectively examined 307 consecutive matched related (n = 145) and unrelated (n = 162) HSCTs, using peripheral blood stem cells or bone marrow. HSCTs for active, uncontrolled malignancies were excluded. The initial three weeks’ average mean CsA trough levels were analyzed in landmark and multi-state models, using a cut-off of 200 ng/mL. ResultsCsA levels >200 ng/mL were associated with a reduced risk of acute graft-versus-host disease (GVHD) grade 3–4 at the first-week landmark (subdistribution hazard ratio [SHR] 0.59, P = 0.03) and the second-week landmark (SHR 0.48, P = 0.004), whereas there was no impact at the third-week landmark (HR 0.87, P = 0.69). This was supported by a multi-state model, in which week 1 (hazard ratio [HR] 0.53, P = 0.006) and week 2 (HR 0.48, P = 0.003), but not week 3 (HR 0.80, P = 0.44) CsA levels >200 ng/mL were associated with a reduced acute GVHD 3–4 risk. Relapse incidence was not significantly affected by week 1 through 3 CsA levels. Despite ATLG's inherent GVHD-preventive properties, week 1 CsA trough levels >200 ng/mL following ATLG-based HSCT (n = 220) were associated with a significantly reduced risk of non-relapse mortality (SHR 0.52, P = 0.02) and improved overall survival (HR 0.61, P = 0.02). ConclusionsOur findings emphasize the continuing importance of ensuring CsA levels ≥200 ng/mL immediately post-transplant in the setting of ATLG-based HSCT.

Factors Associated With Treatment Pathways in Early Axial Spondyloarthritis: A Multistate Analysis of the 10-Year Follow-Up of the DESIR Cohort.

Current recommendations for the management of patients with axial spondyloarthritis (axSpA) emphasize the need of an individualized strategy in therapeutic decision-making. The study objectives were to describe therapeutic strategies observed in axSpA, and to assess the factors associated with treatment intensification over time. We included patients with axSpA from the French prospective cohort DESIR (Devenir des Spondylarthropathies Indifférenciées Récentes), with a scheduled 10-year follow-up. A multistate model with 4 ordered treatment states (no treatment, nonsteroidal antiinflammatory drugs [NSAIDs], conventional synthetic disease-modifying antirheumatic drugs [csDMARDs], and tumor necrosis factor inhibitors [TNFi]) was defined, with 6 possible transitions. Restricted mean sojourn times in each state were estimated. Then, predictors of those transitions were assessed by multivariable Cox models. A total of 686/708 (96.9%) patients who had > 1 visit were analyzed. At cohort entry, 199 (29%) were untreated, 427 (62.2%) were receiving NSAIDs, 60 (8.7%) csDMARDs, and none were receiving TNFi. Over the follow-up period, patients mostly (46.4% of the time) received NSAIDs, followed by TNFi (24.4% of the time). The presence of sacroiliitis on radiographs, inflammatory bowel disease, and articular index were jointly associated with the transition to NSAIDs. Longer duration in the previous state often decreased the hazard of the transition to csDMARDs or TNFi. Worse disease activity outcomes increased the hazard of most transitions. To our knowledge, this was the first study using a multistate model to easily represent different treatment states, detailing the transitions across them and their associated factors. Different time profiles for the management of patients with axSpA were identified, including in those abstaining from treatment up to a significant proportion of patients treated with csDMARDs.

Multistate models with nested frailty for lifetime analysis: Application to bone marrow transplantation recovery patients

Multistate models allow individuals to move through a series of states over time, making it possible to estimate the transition probabilities and intensities between states and the effect of covariates associated with each transition. Multistate models are useful for studying the evolution of individuals over time, particularly in clustered data, where individuals are grouped together and may share common risk factors. Frailties can be used in multistate models to capture the heterogeneity between groups at risk for different types of transitions, as well as the dependency structure between transitions of individuals in the same group. Frailties are unobserved random variables that are assumed to be independent between different types of transitions but shared by groups of individuals. In this paper, we propose a multistate model with frailties to capture the heterogeneity between groups at risk for different types of transitions, as well as the dependency structure between transitions of individuals in the same group. We assume that the frailties are independent between different types of transitions but shared by groups of individuals. We discuss both parametric and semiparametric estimation of the model parameters, and we evaluate the performance of the model in simulation studies. Finally, we apply the model to a dataset of bone marrow transplantation recovery patients.

Role of hospitalisation for substance misuse in marital status transitions: a 47-year follow-up of a Swedish birth cohort

BackgroundSubstantial research suggests that the risk of substance misuse is higher among individuals who remain unmarried or experience marital dissolution, whereas marriage tends to be protective. However, few studies have...

A multistate transition model of changes in loneliness and carer stress among community-dwelling older adults in Aotearoa New Zealand.

To identify changes in loneliness and carer stress between two time points for older people of different ethnicities who had repeated interRAI home care assessments. Participants consisted of community-dwelling older adults across New Zealand who received two interRAI-HC assessments between 5 July 2012 and 31 December 2019. Two multistate models were developed: the first model was not lonely versus lonely, and the second model was no carer stress versus carer stress. The one-year transition probabilities were calculated. Mean sojourn times were calculated for each state except death. Paired t-tests assessed the differences in transition probabilities between the different ethnic groups. The mean age of the cohort was 82.5 years (SD 7.7 years). At first assessment, 14,646 (21%) older people stated they were lonely and 26,789 carers (38%) experienced stress. The most common first transition type was not lonely to not lonely: Māori 42%, Pacific 54%, Asian, 48% and Other 40%. The highest one-year transition probability in the loneliness model was living in aged residential care to death (0.79). The most common first transition type for the carer stress was no carer stress to no carer stress: Māori 35%, Pacific, 46%, Asian, 43% and Other 33%. The highest one-year transition probability in the carer stress model was living in aged residential care to death (0.80). The statuses not lonely and no carer stress had a mean sojourn time of approximately one year, and eight months to one year, respectively. Loneliness can change over time due to circumstances and an individual's perception of loneliness at the time of assessment. Carer stress is enduring and has a low probability of improvement.

Comorbidity risk and distribution characteristics of chronic diseases in the elderly population in China

BackgroundThe risk of comorbid chronic diseases in elderly people is an important problem affecting their health and quality of life. We analyzed the incidence of chronic diseases for combinations of chronic diseases analyzed.MethodsWe used the original data to construct hypothetical cohorts of elderly individuals that evolved with age. The complex network was used to reduce the dimensionality of disease. The multistate transition model is used to calculate the incidence of each chronic disease, exploring comorbidity characteristics and rules.Results(1) By using complex network, seven chronic diseases were screened out in men, including hypertension, diabetes, heart disease, stroke, chronic lung disease, arthritis and dyslipidemia; six chronic diseases in women showed significant comorbidity except chronic lung disease. (2) Incidence show differences in age and sex; incidence of chronic diseases generally increased with age. (3) The marginal risk increases with the number of basic chronic diseases associated with comorbidities. (4) When hypertension is present as a basic disease, its impact on the risk of other chronic diseases is much less than that of other chronic diseases. (5) When diseases occur as basic chronic diseases, hypertension–heart disease and diabetes–dyslipidemia are combinations that have the greatest impact on each other in men; hypertension–heart disease in women.ConclusionsThe incidence of chronic diseases in patients who have chronic diseases and will form comorbidities differs from that in healthy states, and the related effects of different chronic diseases also differ. Among these conditions, hypertension is caused by a special mechanism.

Applying a multistate survival model to explore the role of fine particles in promoting frailty in the Medicare cohort.

Fine particle pollution is a well-established risk to human health. Observational epidemiology generally treats events as though they are independent of one another and so do not examine the role air pollution may play in promoting the progression of disease. Multistate survival models account for the complex pathway of disease to death. We employ a multistate survival model to characterize the role of chronic exposure to PM2.5 in affecting the rate at which Medicare beneficiaries transition to first hospitalization for cardiovascular disease and then subsequently death. We use an open cohort of Medicare beneficiaries and PM2.5 concentrations estimated with photochemical model predictions, satellite-based observations, land-use data, and meteorological variables. The multistate model included three transitions: (1) entry to cardiovascular hospital admission; (2) entry to death; and (3) cardiovascular hospital admission to death. The transition intensity was modeled using a Cox proportional hazards model. For a 1 µg/m3 increase in annual mean PM2.5, we estimate a nationally pooled hazard ratio of 1.022 (95% confidence interval [CI] = 1.018, 1.025) for the transition from entry to first cardiovascular hospital admission; 1.054 (95% CI = 1.039, 1.068) for the transition from entry to death; 1.036 (95% CI = 1.027, 1.044) for the transition from first cardiovascular hospital admission to death. The hazard ratios exhibited some heterogeneity within each of nine climatological regions and for each of the three transitions. We find evidence for the role of PM in both promoting chronic illness and increasing the subsequent risk of death.

Abstract IA016: Bayesian phylodynamics to quantify cancer growth across space and time

Abstract Phylogenetics has traditionally inferred shared genetic histories of species over macroevolutionary time. With the expansion of cancer multi-region and single-cell sequencing, phylogenetic approaches have found new uses mapping the evolutionary histories of tumors as well. Recently developed “phylodynamic” methods that integrate phylogenetic and population dynamic modeling offer further promise - uncovering cancer growth dynamics from the topologies and branch lengths of tumor trees. Here, I discuss two recent efforts in this area: 1) Heterogeneous growth rates across a tumor become encoded in the branching patterns of phylogenetic trees built from multi-region sequencing. We introduce a cancer-specific extension to a class of multi-state birth-death models to infer these growth rate differences across space. We apply our model (SDevo) to a sample of hepatocellular carcinomas and infer that these tumor peripheries grow 2-6 times faster than their interiors. 2) Tumors experience differential growth rates over time as they acquire new driver mutations and invade their surrounding tissues, and these growth dynamics similarly leave characteristic signatures in their phylogenies. We employ Bayesian skyline models to dissect tumor growth dynamics from a dataset of diverse cancers and reveal variation in cancer growth rates over time. As cancer sequencing continues to proliferate, phylodynamic approaches are poised to reveal new quantitative insights into cancer growth dynamics across time and space. Citation Format: Maya Lewinsohn, Trevor Bedford, Nicola F. Müller, Alison Feder. Bayesian phylodynamics to quantify cancer growth across space and time [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Translating Cancer Evolution and Data Science: The Next Frontier; 2023 Dec 3-6; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(3 Suppl_2):Abstract nr IA016.

Computing a Bayesian multi-state model of a protein structure from an X-ray diffraction pattern

Computing a Bayesian multi-state model of a protein structure from an X-ray diffraction pattern

Premorbid personality traits as predictors for incident predementia syndromes: a multistate model approach.

Associations have been found between five-factor model (FFM) personality traits and risk of developing specific predementia syndromes such as subjective cognitive decline (SCD) and mild cognitive impairment (MCI). The aims of this study were to: 1) Compare baseline FFM traits between participants who transitioned from healthy cognition or SCD to amnestic MCI (aMCI) versus non-amnestic MCI (naMCI); and 2) Determine the relationship between FFM traits and risk of transition between predementia cognitive states. Participants were 562 older adults from the Einstein Aging Study, 378 of which had at least one follow-up assessment. Baseline data collected included levels of FFM personality traits, anxiety and depressive symptoms, medical history, performance on a cognitive battery, and demographics. Follow-up cognitive diagnoses were also recorded. Mann-Whitney U tests revealed no differences in baseline levels of FFM personality traits between participants who developed aMCI compared to those who developed naMCI. A four-state multistate Markov model revealed that higher levels of conscientiousness were protective against developing SCD while higher levels of neuroticism resulted in an increased risk of developing SCD. Further, higher levels of extraversion were protective against developing naMCI. FFM personality traits may be useful in improving predictions of who is at greatest risk for developing specific predementia syndromes. Information on these personality traits could enrich clinical trials by permitting trials to target individuals who are at greatest risk for developing specific forms of cognitive impairment. These results should be replicated in future studies with larger sample sizes and younger participants.

Roles of general and central adiposity in cardiometabolic multimorbidity: revisiting the obesity paradox using a multistate model.

The objective of this study was to evaluate the associations of general and central obesity with risk of first cardiometabolic disease (FCMD), cardiometabolic multimorbidity (CMM), and death. A total of 86,169 participants who were CMD-free were included from the Kailuan cohort and categorized into four groups by quartiles of BMI, waist to hip ratio (WHR), weight-adjusted waist index, and waist to height ratio. We defined FCMD as the first onset of diabetes, stroke, or myocardial infarction and CMM as co-occurrence of at least two CMDs. Multistate models were used to estimate hazard ratios and 95% CI. A total of 18,461 participants developed FCMD, of whom 1476 progressed to CMM, and 10,009 died during follow-ups. Both general and central adiposity indices increased the risk of transition from baseline to FCMD and from FCMD to CMM. However, compared with the first quartile, the hazard ratio (95% CI) of the fourth quartile of BMI was 0.86 (95% CI: 0.80-0.91) for transition from health to death and 0.66 (95% CI: 0.59-0.74) from FCMD to death, whereas the corresponding estimates of WHR were 1.22 (95% CI: 1.14-1.31) and 1.16 (95% CI: 1.02-1.32), respectively. Central adiposity indices such as WHR were associated with an increased risk of CMD and mortality, showing no evidence for the obesity paradox and thereby supporting a shift of public focus from BMI only to both general obesity and adiposity distribution.

Polyphenism predicts actuarial senescence and lifespan in tiger salamanders.

Actuarial senescence (called 'senescence' hereafter) often shows broad variation at the intraspecific level. Phenotypic plasticity likely plays a central role in among-individual heterogeneity in senescence rate (i.e. the rate of increase in mortality with age), although our knowledge on this subject is still very fragmentary. Polyphenism-the unique sub-type of phenotypic plasticity where several discrete phenotypes are produced by the same genotype-may provide excellent study systems to investigate if and how plasticity affects the rate of senescence in nature. In this study, we investigated whether facultative paedomorphosis influences the rate of senescence in a salamander, Ambystoma mavortium nebulosum. Facultative paedomorphosis, a unique form of polyphenism found in dozens of urodele species worldwide, leads to the production of two discrete, environmentally induced phenotypes: metamorphic and paedomorphic individuals. We leveraged an extensive set of capture-recapture data (8948 individuals, 24 years of monitoring) that were analysed using multistate capture-recapture models and Bayesian age-dependent survival models. Multistate models revealed that paedomorphosis was the most common developmental pathway used by salamanders in our study system. Bayesian age-dependent survival models then showed that paedomorphs have accelerated senescence in both sexes and shorter adult lifespan (in females only) compared to metamorphs. In paedomorphs, senescence rate and adult lifespan also varied among ponds and individuals. Females with good body condition and high lifetime reproductive success had slower senescence and longer lifespan. Late-breeding females also lived longer but showed a senescence rate similar to that of early-breeding females. Moreover, males with good condition had longer lifespan than males with poor body condition, although they had similar senescence rates. In addition, late-breeding males lived longer but, unexpectedly, had higher senescence than early-breeding males. Overall, our work provides one of the few empirical cases suggesting that environmentally cued polyphenism could affect the senescence of a vertebrate in nature, thus providing insights on the ecological and evolutionary consequences of developmental plasticity on ageing.

Sex and size shape the ontogeny of partial migration.

The arrival-time hypothesis of partial seasonal migration proposes that over-winter residence is driven by reproductive benefits of early presence on the breeding grounds. Thus, it predicts increased occurrence of residence at reproductive age. In contrast, the body size hypothesis proposes age-independent benefits of residence for large individuals, who should exhibit greater winter tolerance. Despite different expectations in age patterns for the two hypotheses in long-lived partially migrant species, there is little empirical work investigating the ontogeny of migratory phenotypic expression, that is the expression of residence or migration. We investigated the influence of age, sex and body size on migratory phenotype throughout ontogeny (from first year to early adulthood) in a long-lived partially migrant species, the red kite Milvus milvus. We GPS-tracked 311 individuals tagged as juveniles and 70 individuals tagged as adults over multiple years, yielding 881 observed annual cycles. From this data, we estimated age-dependent probabilities of the transition to residence and of survival in migrants and residents using a Bayesian multistate capture-recapture model, as well as the probability of resuming migration once resident. We then calculated the resulting proportion of residents per age class. In both sexes, almost all juveniles migrated in their first winter, after which the probability of becoming resident gradually increased with age class to approximately 0.3 in adults (>3 calendar years). A size effect in third calendar year females suggests that large females adopt residence earlier in life than small females. The transition from residence back to migration only occurred with a probability of 0.15 across all resident individuals. In addition, survival was notably reduced in adult male migrants compared to adult male residents. These results are largely consistent with the arrival-time and body size hypotheses, simultaneously. Our results reveal a plastic, yet primarily directional within-individual change in migratory phenotype towards more residence with increasing age, varying between sexes and between individuals of different size. This study highlights that different individual characteristics can jointly shape the ontogeny of migratory behaviour and result in complex within-population patterns and persistence of migratory phenotypes.