Postmenopausal Shrinkage of Uterine Myomas: A Retrospective Study of 97 Cases Monitored Annually for 10 Years.

Introduction and aim Both patients and gynecologists are concerned about how much and how quickly myomas shrink after menopause. This study aimed to elucidate clinical findings that may be associated with postmenopausal shrinkage of uterine myomas. Materials and methods This study included 97 patients who underwent menopause by August 2012, had myoma nodules with the longest diameter between 50 mm and 160 mm, and visited our specialized myoma clinic annually for at least 10 years after menopause. They underwent transabdominal ultrasonography at least once per year. An experienced gynecologist measured the longest diameter of myoma nodules with a maximum diameter between 50 mm and 160 mm. The shrinkage rate of myoma diameters after menopause compared to premenopausal diameters was calculated each year for 10 years. The shrinkage rate of the longest diameter of the largest nodule 10 years after menopause (10-year shrinkage rate) and its relationship with clinical findings (the age at menopause, parity, body mass index {BMI}, number of nodules, MRI findings on T2-weighted image, location of the nodule, and longest diameter of the largest nodule before menopause) were analyzed. Additionally, we examined annual changes in shrinkage rate of myomas over a 10-year period after menopause (annual trend), and the relationship between annual trends and factors such as BMI and the number of nodules. Results In this examination of 10-year shrinkage rate, the group with a BMI of less than 25 showed a significantly greater shrinkage rate compared to the group with a BMI of 25 or more (25.0% vs 15.7%, p=0.023). Additionally, the group with a single nodule showed a significantly greater 10-year shrinkage rate compared to the group with four or more nodules (26.3% vs 15.2%, p=0.036). For annual trends, the rate of change in the first two years after menopause was significantly faster compared to the trend from the third to the 10th year (difference in slope: 3.888 points per year, p<0.001). When divided into two groups based on the number of nodules (one or two nodules group and three or more nodules group), the group with one or two nodules showed a significant difference in the shrinkage rate between up to two years after menopause and from the period from the third to the 10th year (difference in slope: 4.590 points per year, p<0.001). However, for the group with three or more nodules, there was no significant difference in the annual trend between the first two years after menopause and the rate from the third to the 10th year (difference in slope: 1.626 points per year, p=0.107). Conclusion BMI and the number of myoma nodules were significantly related to the 10-year shrinkage rate. Although myomas shrank significantly faster within the first two years after menopause compared to the later period, the early annual trend did not differ significantly from the trend in the later period when there were multiple nodules with a maximum diameter of 50 mm or more.

A Hybrid Model Simulating Multi-Stage Continuous Fermentation of Saccharomyces cerevisiae.

The performance of industrial strains has gradually improved with the rapid development of synthetic biotechnology. The production efficiency of traditional batch and fed-batch culture is limited and product quality varies since both are dynamic processes, whereas multi-stage continuous culture can maximise the production efficiency of specific fermentation processes and achieve consistent product quality. However, each single-stage fermentation under multi-stage continuous fermentation requires accurate steady-state control, and a model with adequate accuracy is required for designing and controlling a multi-stage continuous fermentation process. At present, there are few reports on kinetic models for the control of multi-stage continuous fermentation. In this work, we constructed a hybrid model for Saccharomyces cerevisiae multi-stage continuous culture, taking both oxygen limitation and Crabtree effect. The accuracy of the model was ∼80%, the advantages and limitations of the model are discussed and a potential improvement strategy is proposed.

Plasma Proteomic Biomarkers in Alzheimer's Disease and Cardiovascular Disease: A Longitudinal Study.

The co-occurrence of Alzheimer's disease (AD) and cardiovascular diseases (CVDs) in older adults highlights the necessity for the exploration of potential shared risk factors. A total of 566 adults were selected from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, including 111 individuals with AD, 383 with mild cognitive impairment (MCI), and 410 with CVD. The multivariable linear mixed model (LMM) was used to investigate the associations of AD and CVD with longitudinal changes in 146 plasma proteomic biomarkers (measured at baseline and the 12-month follow-up). The LMM showed that 48 biomarkers were linked to AD and 46 to CVD (p < 0.05). Both AD and CVD were associated with longitudinal changes in 14 biomarkers (α1Micro, ApoH, ß2M, BNP, complement C3, cystatin C, KIM1, NGAL, PPP, TIM1, THP, TFF3, TM, and VEGF), and both MCI and CVD were associated with 12 biomarkers (ApoD, AXL, BNP, Calcitonin, CD40, C-peptide, pM, PPP, THP, TNFR2, TTR, and VEGF), suggesting intricate connections between cognitive decline and cardiovascular health. Among these, the Tamm Horsfall Protein (THP) was associated with AD, MCI, CVD, and APOE-ε4. This study provides valuable insights into shared and distinct biological markers and mechanisms underlying AD and CVD.

Trajectory Modeling and Response Prediction in Transcranial Magnetic Stimulation for Depression.

Background: Repetitive transcranial magnetic stimulation (rTMS) therapy could be improved by more accurate and earlier prediction of response. Latent class mixture (LCMM) and non-linear mixed effects (NLME) modeling have been applied to model the trajectories of antidepressant response (or non-response) to TMS, but it is not known whether such models are useful in predicting clinically meaningful change in symptom severity, i.e. categorical (non)response as opposed to continuous scores. Methods: We compared LCMM and NLME approaches to model the antidepressant response to TMS in a naturalistic sample of 238 patients receiving rTMS for treatment resistant depression, across multiple coils and protocols. We then compared the predictive power of those models. Results: LCMM trajectories were influenced largely by baseline symptom severity, but baseline symptoms provided little predictive power for later antidepressant response. Rather, the optimal LCMM model was a nonlinear two-class model that accounted for baseline symptoms. This model accurately predicted patient response at 4 weeks of treatment (AUC = 0.70, 95% CI = [0.52 - 0.87]), but not before. NLME offered slightly improved predictive performance at 4 weeks of treatment (AUC = 0.76, 95% CI = [0.58 - 0.94], but likewise, not before. Conclusions: In showing the predictive validity of these approaches to model response trajectories to rTMS, we provided preliminary evidence that trajectory modeling could be used to guide future treatment decisions.

Sensitivity analysis for publication bias in meta-analysis of sparse data based on exact likelihood.

Meta-analysis is a powerful tool to synthesize findings from multiple studies. The normal-normal random-effects model is widely used to account for between-study heterogeneity. However, meta-analyses of sparse data, which may arise when the event rate is low for binary or count outcomes, pose a challenge to the normal-normal random-effects model in the accuracy and stability in inference since the normal approximation in the within-study model may not be good. To reduce bias arising from data sparsity, the generalized linear mixed model can be used by replacing the approximate normal within-study model with an exact model. Publication bias is one of the most serious threats in meta-analysis. Several quantitative sensitivity analysis methods for evaluating the potential impacts of selective publication are available for the normal-normal random-effects model. We propose a sensitivity analysis method by extending the likelihood-based sensitivity analysis with the $t$-statistic selection function of Copas to several generalized linear mixed-effects models. Through applications of our proposed method to several real-world meta-analyses and simulation studies, the proposed method was proven to outperform the likelihood-based sensitivity analysis based on the normal-normal model. The proposed method would give useful guidance to address publication bias in the meta-analysis of sparse data.

Development of mixed linear models to analyze and describe the impact of malathion on the larval growth of Megaselia scalaris (Diptera: Phoridae) under various feeding media and environmental conditions.

Forensic entomology plays a crucial role in estimating the minimum postmortem interval through the study of insect larvae found at crime scenes. The precision of this estimation relies on various biotic and abiotic elements that simultaneously influence insect growth and development, encompassing factors such as temperature, humidity, photoperiod, diet, and the existence of xenobiotics in decomposing tissues. Despite numerous studies on the influence of these factors, including the impact of xenobiotics, there are currently no robust tools available for making corrections to this estimation considering concurrently all variables. In an attempt to propose an exploratory and descriptive statistical model to analyze the simultaneous effect and interaction of different variables on larval growth, this study aimed to compare the effect of malathion on the growth of Megaselia scalaris (Loew, 1866) (Diptera: Phoridae) raised in malathion-spiked porcine muscle, under controlled and uncontrolled temperature and humidity conditions (environmental conditions). Larvae were also reared using various growth media. A split-plot design that combined crossed and nested factors was employed; 2 linear mixed models were developed to assess the relationships between the variables. The model provides valuable insights into the complex interactions among xenobiotics, growth media, and environmental conditions in the size and development of M. scalaris.

Assessment of household preferences for net textile type (polyester versus polyethylene) for decision-making of the National Malaria Control Programme in Burkina Faso: methods for a quasi-experimental study.

BACKGROUND: A quasi-experimental comparative trial will be designed in Burkina Faso. The study will compare the use and preferences for two groups types of insecticide-treated nets textile: polyester-based and polyethylene-based, according to their use and preferences in selected health districts. These health districts will be selected in three eco-climate zones (Sahelian, dry savannah and wet savannah) in the country. These findings will inform decisions on future net procurements for national malaria control programme in 2025. METHODS: Quantitative surveys and qualitative data collection will be carried out to gather information on the type of net textile most commonly used and preferred by the community. They will be performed between the end of the dry season and the early rainy season. The quantitative surveys involved household interviews with households and individuals' questionnaires, while the qualitative data collection involved in-depth individual interviews and focus group discussions to explore and clarify some key evaluation criteria. A total of 9450 insecticide-treated nets were surveyed for quantitative survey purposes. For the qualitative study, 48 in-depth individual interviews and 12 focus group discussions were carried out. A mixed model approach combining the results from quantitative surveys and qualitative studies will be used for decision-making on the type of insecticide-treated net preference. CONCLUSION: This methodological approach will be used by the National Malaria Control Programme to conduct this study on determinants of net use in Burkina Faso in order to provide robust evidence across diverse settings. This mixed-methods approach for data collection and analysis could be used in other countries to provide evidence that would help to increase the uptake of insecticide-treated nets, the main vector control tool in Africa.

Joint meta-analysis of two diagnostic tests accounting for within and between studies dependence.

There is an extensive literature on methods for meta-analysis of diagnostic test accuracy, but it mainly focuses on a single test. A multinomial generalised linear mixed model was recently proposed for the joint meta-analysis of studies comparing two tests on the same participants in a paired tests design with a gold standard. In this setting, we propose a novel model for joint meta-analysis of studies comparing two diagnostic tests which assumes independent multinomial distributions for the counts of each combination of test results in diseased and non-diseased patients, conditional on the latent vector of probabilities of each combination of test results in diseased and non-diseased patients. For the random effects distribution of the latent proportions, we employ a one-truncated D-vine copula that can provide tail dependence or asymmetry. The proposed model includes the multinomial generalised linear mixed model as a special case, accounts for the within-study dependence induced because the tests are applied to the same participants, allows for between-studies dependence, and can also operate on the original scale of the latent proportions. The latter enables the derivation of summary receiver operating characteristic curves. Our methodology is demonstrated with simulation studies and a meta-analysis of screening for Down's syndrome with two tests: shortened humerus and shortened femur.

Genetic regulation of B cell receptor signaling pathway: Insights from expression quantitative trait locus analysis using a mixed model.

The B cell receptor (BCR) signaling pathway regulates non-immune cellular response through various pathways like MAPK, NF-kB, and PI3K-Akt. This study aimed to identify expression quantitative trait loci (eQTL) and their regulatory functions on BCR signaling pathway genes. A mixed model was employed to analyze eQTL using RNA expression levels in lymphoblastoid from 376 Europeans in the GEUVADIS dataset. In total, 266 SNPs, including 115 cis-acting SNPs, were found for association with transcription of 13 genes (P < 5 × 10-8), revealing 19 independent signals for five genes through linkage disequilibrium analysis. Functional analysis, aligning them with DNase sensitive sites, transcription factor binding sites, histone modification, promoters/enhancers, CpG islands, and ChIA-PET, identified regulatory variants targeting SYK, VAV2, and PLCG2. Notably, rs2562397 was validated as a SYK promoter variant, and rs694505, rs636667, and rs4889409 were confirmed as enhancer variants for VAV2 and PLCG2. Their allelic differences in gene expression were also confirmed using ENCODE ChIP-seq and Sei neural network prediction. Persistent differential expression of these genes by alleles might impact the adaptive immune system, vascular development, and/or relevant diseases that have been previously associated with other variants of the genes. Comprehensive genetic architecture studies of the BCR signaling pathway, along with experiments demonstrating related mechanisms, will greatly contribute to understanding the underlying mechanisms of relevant disease development and implementing precision medicine.

Asymptotically exact fit for linear mixed model in genetic association studies.

The linear mixed model (LMM) has become a standard in genetic association studies to account for population stratification and relatedness in the samples to reduce false positives. Much recent progresses in LMM focused on approximate computations. Exact methods remained computationally demanding and without theoretical assurance. The computation is particularly challenging for multiomics studies where tens of thousands of phenotypes are tested for association with millions of genetic markers. We present IDUL and IDUL† that use iterative dispersion updates to fit LMMs, where IDUL† is a modified version of IDUL that guarantees likelihood increase between updates. Practically, IDUL and IDUL† produced identical results, both are markedly more efficient than the state-of-the-art Newton-Raphson method, and in particular, both are highly efficient for additional phenotypes, making them ideal to study genetic determinants of multiomics phenotypes. Theoretically, the LMM likelihood is asymptotically unimodal, and therefore the gradient ascent algorithm IDUL† is asymptotically exact. A software package implementing IDUL and IDUL† for genetic association studies is freely available at https://github.com/haplotype/IDUL.

Gait Speed Reserve in the general population-based 'Good Aging in Skåne' cohort study-distribution and associated factors.

Gait Speed Reserve (GSR) expresses a difference between fast and comfortable gait speed and may have an impact on everyday functioning. It was also hypothesized as a useful proxy measure of physiological reserve. However, height-normalizing values of GSR and its associated factors have not been evaluated in a general population of older adults. Therefore, we aimed to investigate the distribution of height-normalized GSR (HN-GSR) in an elderly population-based cohort from urban and rural areas (n = 4342) aged 60-93 years and evaluate associated physiological and lifestyle factors. Using linear mixed models, we identified gender and nine modifiable factors as significantly associated with HN-GSR across four age groups. Better handgrip strength, cognition and standing balance, higher physical activity level, larger calf circumference, and less smoking had positive associations with HN-GSR, while female gender, more leg pain, higher weight and, alcohol consumption had opposite effects. The Marginal R2 imply that this model explained 26% of the variance in HN-GSR. Physical activity and handgrip strength varied across age groups in impact on HN-GSR. The differences were however comparatively minor. In this large cohort study of older adults, we proposed for the first time that factors associated with HN-GSR represented multi-domain features that are in line with previous findings reported for GSR. Measuring HN-GSR/GSR may help clinicians identify early physiological impairments or unhealthy lifestyle habits, especially among older women, and may also have safety implications in daily life. Further work is needed to find out if measuring HN-GSR/GSR may be useful in identifying adverse health outcomes and overall physiological reserve.

Edaphic influences on soil organic carbon in the forest systems of southern Western Ghats, India.

Spatial distribution and edaphic influences on soil organic carbon (SOC) are key determinants of carbon sequestration potential, and analysis of aggregate-protected SOC gives an in-depth understanding of the stability of carbon stored in soils. The present study evaluated the edaphic effects on the SOC in four different forest types-tropical evergreen forest, tropical moist deciduous forest, tropical dry deciduous forest and shola forest-in the southern high hills agro-ecological zone of Western Ghats, India. SOC stocks at depths of up to 1 m varied significantly across forest types, with the highest levels observed in the shola forest type (441.08 Mg C/ha) and the lowest in the dry deciduous forest (138.17 Mg C/ha). Around 70% of SOC was found in upper layers (0-30 cm) in all the studied forest types. Evaluation by a fixed-effect model showed that forest type, soil depth and aggregate size significantly affected SOC storage in these systems. An assessment of the relative importance and effect of 14 edaphic factors on SOC content in different forest types using the random forest model showed that the algorithm could explain 93.68%, 41.72%, 45.53% and 75.2% variability of SOC concentration across shola, dry deciduous, moist deciduous and evergreen systems, respectively. Across all forest types, except for dry deciduous forests, soil texture was found to be the primary factor influencing SOC, surpassing all other edaphic parameters. Ionic interactions by way of metal oxides like Ca2+, Al3+, Fe3+, Mg and H+ influenced the SOC in tropical forest systems. The insights into SOC dynamics and the edaphic factors regulating them offer valuable guidance for forest management in tropical regions, particularly regarding climate change mitigation.

Inflammatory Trajectory and Anti-Inflammatory Pharmacotherapy in Frozen Elephant Trunk-Treated Acute Type I Aortic Dissection.

Background: Acute DeBakey type I aortic dissection is associated with high morbidity and mortality. Little is known regarding the role of leukocyte trajectory in prognosis. Methods: We included adult acute DeBakey type I aortic dissection patients with emergency frozen elephant trunk and total arch replacement in 2 cardiovascular centers (2020-2022). We used latent class mixed model to group patients according to their leukocyte patterns from hospital admission to the first 5 days after surgery. We investigated the association of leukocyte trajectory and 30-day and latest follow-up mortality (October 31, 2023), exploratorily analyzing the effects of ulinastatin treatment on outcome. Results: Of 255 patients included, 3 distinct leukocyte trajectories were identified: 196 in group I (decreasing trajectory), 34 in group II (stable trajectory), and 25 in group III (rising trajectory). Overall, 30-day mortality was 25 (9.8%), ranging from 8.2% (16/196) in group I, 8.8% (3/34) in group II, to 24.0% (6/25) in group III (P for trend = .036). Group III was associated with higher mortality both at 30 days (adjusted hazard ratio, 3.260; 95% CI, 1.071-9.919; P = .037) and at the last follow-up (adjusted hazard ratio, 2.840; 95% CI, 1.098-7.345; P = .031) compared with group I. Conclusions: Distinct and clinically relevant groups can be identified by analyzing leukocyte trajectories, and a rising trajectory was associated with higher short-term and midterm mortality.

Profiling people with Parkinson's disease at risk of cognitive decline: Insights from PPMI and ICICLE-PD data.

Introduction: A subset of people with Parkinson's disease (PD) develop dementia faster than others. We aimed to profile PD cognitive subtypes at risk of dementia based on their rate of cognitive decline. Method: Latent class mixed models stratified subtypes in Parkinson's Progression Markers Initiative (PPMI) (N = 770) and ICICLE-PD (N = 212) datasets based on their decline in the Montreal Cognitive Assessment over at least 4 years. Baseline demographic and cognitive data at diagnosis were compared between subtypes to determine their clinical profile. Results: Four subtypes were identified: two with stable cognition, one with steady decline, and one with rapid decline. Performance on Judgement of Line Orientation, but not category fluency, was associated with a steady decline in the PPMI dataset, and deficits in category fluency, but not visuospatial function, were associated with a steady decline in the ICICLE-PD dataset. Discussion: People with PD susceptible to cognitive decline demonstrate unique clinical profiles at diagnosis, although this differed between cohorts. Highlights: Four cognitive subtypes were revealed in two Parkinson's disease samples.Unique profiles of cognitive impairment were related to cognitive decline.Judgement of Line Orientation/category fluency predictive of steady decline.Global deficits related to rapid cognitive decline and increased dementia risk.

Who can Benefit From Volunteer Participation? Analysis of Wellbeing Heterogeneity in Volunteer Participation of Older People.

The basic premise of this study is that the traditional method to treating all older people as coming from the same distribution misspecifies the true model and ignores potentially important information in wellbeing outcomes of social participation. Using data from the China Longitudinal Aging Social Survey (CLASS), this paper proposes a finite mixture model (FMM) to identify the heterogeneous relationship between volunteer participation and older people's subjective well-being (SWB) and then explore the determinants of wellbeing heterogeneity in volunteer participation. The results reveal that older people can be classified into two latent subgroups, that is the volunteering beneficiary group (accounting for about 42%) and the volunteering non-beneficiary group (accounting for about 58%). The FMM is therefore more appropriate in estimating the complex impact of volunteering. Rural older people with poorer health, weaker social networks, better economic status, and better community environments are more likely to benefit from volunteer participation. Our findings have suggested some practical implications to increase the probability of benefit from volunteer participation.

Variability in rat weight gain during development.

The rat is one of the most employed animal models in biomedicine. Traditionally, weight gain has been utilized to gauge development and compare across species. Numerous studies have conducted longitudinal analyses of rat development, with emphasis on weight gain analysis. Given the high variability in these patterns, experimental data from a single laboratory may not be reliable for generalized estimation. This study aimed to analyze the effect of different factors on the pattern of weight gain during rat development. A literature survey was conducted to compile a database comprising nearly 300 data points of age and weight from 15 longitudinal studies. The database comprised both pre- and postnatal data. Utilizing the Gompertz equation, the data was analyzed to formulate a comprehensive model describing rat development. Differences in growth patterns became increasingly evident at later developmental stages, when significant differences in the maximum asymptote between sexes and strains were reached.

Effects of high-intensity interval training on sleep disturbances associated with posttraumatic stress disorder.

Sleep disturbances are common in individuals with posttraumatic stress disorder. Exercise interventions are a promising approach in the treatment of sleep disorders, but little is known about the efficacy of exercise interventions for sleep disturbances associated with posttraumatic stress disorder. A total of 40 individuals with posttraumatic stress disorder were randomized to six sessions of either high-intensity interval training or low-to-moderate-intensity training, administered within 12 days. Sleep quality was assessed over 24 days from baseline to post with the Pittsburgh Sleep Quality Index, a sleep log, and a waist-worn actigraphy. Analyses revealed that, regardless of group allocation, Pittsburgh Sleep Quality Index score improved significantly by 2.28 points for high-intensity interval training and 1.70 points for low-to-moderate-intensity training (d = 0.56 for high-intensity interval training; 0.49 for low-to-moderate-intensity training) over time, while there were no significant changes in any sleep log or actigraphy measure. Analysis of a subsample of those affected by clinically significant sleep disturbances (n = 24) revealed a significant time effect with no difference between exercise interventions: Pittsburgh Sleep Quality Index improved significantly by 2.65 points for high-intensity interval training and 2.89 points for low-to-moderate-intensity training (d = 0.53 for high-intensity interval training; 0.88 for low-to-moderate-intensity training), and actigraphy measure of wake after sleep onset was reduced significantly by 14.39 minutes for high-intensity interval training and 6.96 minutes for low-to-moderate-intensity training (d = 0.47 for high-intensity interval training; 0.11 for low-to-moderate-intensity training) from baseline to post. In our pilot study, we found an improvement in sleep quality from pre- to post-assessment. There were no significant differences between exercise groups. Further studies are needed to investigate whether the found time effects reflect the exercise intervention or unrelated factors.

A longitudinal study of the influence of air pollutants on children: a robust multivariate approach.

This paper aims to evaluate the statistical association between exposure to air pollution and forced expiratory volume in the first second (FEV1) in both asthmatic and non-asthmatic children and teenagers, in which the response variable FEV1 was repeatedly measured on a monthly basis, characterizing a longitudinal experiment. Due to the nature of the data, an robust linear mixed model (RLMM), combined with a robust principal component analysis (RPCA), is proposed to handle the multicollinearity among the covariates and the impact of extreme observations (high levels of air contaminants) on the estimates. The Huber and Tukey loss functions are considered to obtain robust estimators of the parameters in the linear mixed model (LMM). A finite sample size investigation is conducted under the scenario where the covariates follow linear time series models with and without additive outliers (AO). The impact of the time-correlation and the outliers on the estimates of the fixed effect parameters in the LMM is investigated. In the real data analysis, the robust model strategy evidenced that RPCA exhibits three principal component (PC), mainly related to relative humidity (Hmd), particulate matter with a diameter smaller than 10 µm (PM10) and particulate matter with a diameter smaller than 2.5 µm (PM2.5).

Processing causal structure sentences in Mandarin Chinese: an eye movements study.

It remains uncertain whether causal structure prediction can improve comprehension in Chinese sentences and whether the position of the headword mediates the prediction effect. We conducted an experiment to explore the effect of causal prediction and headword position in Chinese sentence reading. Participants were asked to read sentences containing causal connectives with their eye movements recorded. In the experiment, we manipulated the causal structure of the sentence and the position of the headword. We found a promoting effect of causal structure on first-pass reading time and a hindering impact on total reading time. However, the effect was not mediated by the headword position. The results show that causal syntactic prediction facilitated early-stage processing and increased the integration cost in the late stage of Chinese sentence processing. These findings also support the constraint-based approach, which suggests an isolation between semantic and syntactic processing.

Trends in mercury, lead and cadmium concentrations in 27 European streams and rivers: 2000-2020.

Temporal trends for concentrations of mercury (Hg), lead (Pb) and cadmium (Cd) were evaluated from year 2000-2020 in 20 (Hg), 23 (Pb) and 11 (Cd) watercourses in remote forest catchments in Europe. Decreasing trends were observed in 15% (Hg), 39% (Pb) and 45% (Cd) of the watercourses during the period of evaluation. Decreasing trends were mainly observed between 2000 and 2005 for Hg and between 2000 and 2015 for Pb and Cd. For the last five years of the studied time period (2015-2020), more watercourses showed significant increasing, rather than decreasing Hg, Pb and Cd trends. This was interpreted as a legacy effect of metals still retained in catchment soils. The overall negative trends during the earlier part of the study period were likely driven by declining deposition of metals over Europe, especially for Pb and Cd. Other changes related to metal transport and chemistry may have contributed to the observed trends as well, including recovery from acidification and the ongoing browning of surface waters at northern latitudes. Here we found that organic carbon could explain the seasonal variation in Hg and Pb, but was not related the interannual trends. This study highlights the need for long-term monitoring and robust statistical methods that can detect multidirectional, long-term change in water chemistry.