Subtle magnesium liberation of self-fabricated functional filler actuates highly efficient phosphorus removal from source-separated urine by SBBR.

Domestic wastewater source-separated treatment has attracted wide attention due to the efficiency improvement of sewage treatment systems, energy saving, resource reuse, and the construction and operation cost saving of pipeline networks. Nonetheless, the excess source-separated urine still demands further harmless treatment. Sequencing batch biofilm reactor (SBBR), a new type of composite biofilm reactor developed by filling different fillers into the sequential batch reactor (SBR) reactor, has higher pollutant removal performance and simpler operation and maintenance. However, the phosphorus removal ability of the SBBR filling with conventional fillers is still limited and needs further improvement. In this study, we developed two new fillers, the self-fabricated filler A and B (SFA/SFB), and compared their source-separated urine treatment performance. Long-term treatment experimental results demonstrated that the SBBR systems with different fillers had good removal performance on the COD and TN in the influent, and the removal rate increased with the increasing HRT. However, only the SBBR system with the SFA showed excellent PO43--P and TP removal performance, with the removal rates being 83.7 ± 11.9% and 77.3 ± 13.7% when the HRT was 1 d. Microbial community analysis results indicated that no special bacteria with strong phosphorus removal ability were present on the surface of the SFA. Adsorption experimental results suggested that the SFA had better adsorption performance for phosphorus than the SFB, but it could not always have stronger phosphorus adsorption and removal performance during long-term operation due to the adsorption saturation. Through a series of characterizations such as SEM, XRD, and BET, it was found that the SFA had a looser structure due to the use of different binder and production processes, and the magnesium in the SFA gradually released and reacted with PO43- and NH4+ in the source-separated urine to form dittmarite and struvite, thus achieving efficient phosphorus removal. This study provides a feasible manner for the efficient treatment of source-separated urine using the SBBR system with self-fabricated fillers.

Printed smart devices for anti-counterfeiting allowing precise identification with household equipment.

Counterfeiting has become a serious global problem, causing worldwide losses and disrupting the normal order of society. Physical unclonable functions are promising hardware-based cryptographic primitives, especially those generated by chemical processes showing a massive challenge-response pair space. However, current chemical-based physical unclonable function devices typically require complex fabrication processes or sophisticated characterization methods with only binary (bit) keys, limiting their practical applications and security properties. Here, we report a flexible laser printing method to synthesize unclonable electronics with high randomness, uniqueness, and repeatability. Hexadecimal resistive keys and binary optical keys can be obtained by the challenge with an ohmmeter and an optical microscope. These readout methods not only make the identification process available to general end users without professional expertise, but also guarantee device complexity and data capacity. An adopted open-source deep learning model guarantees precise identification with high reliability. The electrodes and connection wires are directly printed during laser writing, which allows electronics with different structures to be realized through free design. Meanwhile, the electronics exhibit excellent mechanical and thermal stability. The high physical unclonable function performance and the widely accessible readout methods, together with the flexibility and stability, make this synthesis strategy extremely attractive for practical applications.

Engineering Hyaluronic Acid Microneedles Loaded with Mn2+ and Temozolomide for Topical Precision Therapy of Melanoma.

Topical therapy has received worldwide attention for in situ tumors owing to its higher efficacy of drug delivery. Herein, this work reports a dissolvable multifunctional hyaluronic acid microneedles (HMNs) patch coloaded with temozolomide (TMZ) and MnCl2 (TMZ/MnCl2@HMN) for chemoimmunotherapy of melanoma. HMNs can ensure the stability of TMZ over time, and exhibit fewer side effects with a localized release way. In particular, TMZ not only promotes dendritic cell maturation by triggering immunogenic cell death in tumor cells, but also induces DNA damage that can further enhance the Mn2+-activated cGAS-STING (stimulator of interferon genes pathway). As a result, the TMZ/MnCl2@HMN multifunctional platform significantly inhibits lung metastases for melanoma, providing a practical strategy for precision therapy of melanoma.

5-aminolevulinic acid photodynamic therapy using 560-1200 nm followed by 420-1200 nm broadband light in the treatment of moderate-to-severe acne.

BACKGROUND: Moderate-to-severe acne vulgaris, which is a chronic inflammatory skin disease, seriously impacts millions of people. However, traditional therapies may cause severe adverse reactions that are unacceptable to many patients, thus limiting the further application of these therapies. Novel therapeutic approaches to effectively treat moderate-to-severe acne vulgaris with minimal adverse reactions are urgently needed. In this retrospective study, we investigated the efficacy and adverse reactions of photodynamic therapy (PDT) using 560-1200 nm followed by 420-1200 nm broadband light (BBL). METHODS: Twenty-four patients with moderate-to-severe acne vulgaris were included in the study and all patients expressed a strong desire for beauty. After aminolevulinic acid (ALA) gel applied, the entire face was sequentially irradiated by using BBL with a 560 nm cut-off filter (560-1200 nm), followed by BBL with a 420 nm cut-off filter (420-1200 nm). The clinical efficacy was evaluated by the proportion of patients achieving cured response and excellent response (effective rate), based on the percentage of lesions reduction (treatment rate). The fluorescent images and photographs of acne vulgaris were recorded. Pain and other common local adverse reactions during the treatment were also recorded and evaluated. RESULTS: In patients with moderate acne, the mean treatment rates were 57.74 ± 16.40 (%) and 87.40 ± 8.521 (%) at the 6th week and 12th week of treatment, respectively. In patients with severe acne, the mean treatment rates were 60.95 ± 12.06 (%) and 85.04 ± 9.115 (%) at the 6th week and 12th week of treatment, respectively. At the 6th and 12th weeks of treatment, the effective rates of patients were 20.00 % and 93.33 % in patients with moderate acne, and 0.000 % and 88.89 % in patients with severe acne, respectively. Pain scores were significantly higher in patients with severe acne compared to patients with moderate acne when receiving 560-1200 nm BBL-PDT. Additionally, patients when receiving 420-1200 nm BBL-PDT exhibited significantly higher pain scores than those when receiving 560-1200 nm BBL-PDT. The degree of erythema was more severe in patients with severe acne than in those with moderate acne. The pigmentation was observed in one patient with moderate acne and one patient with severe acne. CONCLUSION: The 560-1200 nm and 420-1200 nm BBL-PDT therapy can effectively treat moderate-to-severe acne vulgaris with tolerable adverse reactions, providing a new option for patients with higher esthetic requirements.

The role of CD4+ T cells in tumor and chronic viral immune responses.

Immunotherapies are mainly aimed to promote a CD8+ T cell response rather than a CD4+ T cell response as cytotoxic T lymphocytes (CTLs) can directly kill target cells. Recently, CD4+ T cells have received more attention due to their diverse roles in tumors and chronic viral infections. In antitumor and antichronic viral responses, CD4+ T cells relay help signals through dendritic cells to indirectly regulate CD8+ T cell response, interact with B cells or macrophages to indirectly modulate humoral immunity or macrophage polarization, and inhibit tumor blood vessel formation. Additionally, CD4+ T cells can also exhibit direct cytotoxicity toward target cells. However, regulatory T cells exhibit immunosuppression and CD4+ T cells become exhausted, which promote tumor progression and chronic viral persistence. Finally, we also outline immunotherapies based on CD4+ T cells, including adoptive cell transfer, vaccines, and immune checkpoint blockade. Overall, this review summarizes diverse roles of CD4+ T cells in the antitumor or protumor and chronic viral responses, and also highlights the immunotherapies based on CD4+ T cells, giving a better understanding of their roles in tumors and chronic viral infections.

Albendazole induces an anti-tumor effect and potentiates PD-L1 blockade immunotherapy.

BACKGROUND: Previously, albendazole (ABZ) has been reported as an anti-parasitic drug rather than anti-tumor drug. Our study aim to investigate whether ABZ also has a potential anti-tumor effect by shaping the tumor immune microenvironment and interrogate whether ABZ could synergize with the PD-L1 blockade. METHODS: C57BL/6 mice (C57) were intravenously injected with B16F10-luciferase (B16-luc) cells to establish a lung metastatic melanoma model and subcutaneously inoculated with B16-luc cells to establish a subcutaneous tumor model. The tumor volume and tumor metastasis loci of the mice were measured by a vernier caliper and in vivo imaging. RNA sequencing was performed to analyze the different genes and pathways of immune cells in the tumors. Flow cytometry and immunofluorescence were used to analyze the different subsets of tumor-infiltrating immune cells. RESULTS: The results suggested that ABZ significantly inhibited lung melanoma metastasis with decreased fluorescence intensity and nodule score and mediated the regression of subcutaneous melanoma in mice with decreased tumor volume. Moreover, RNA sequencing results showed that ABZ regulated the gene expression levels and pathways of immune cells in the tumor microenvironment (TME). Meanwhile, flow cytometry and immunofluorescence showed that the number and percentage of CD8+ T cells, CD4+ T cells, and TH1 cells were enhanced in tumors after ABZ treatment. Furthermore, the combination of ABZ and anti-PD-L1 treatment significantly potentiated anti-tumor efficacy in both lung metastasis and subcutaneous melanoma models and mediated an increase in the percentage of CD8+ T cells, CD4+ T cells, and TH1 cells as compared to the control group. CONCLUSION: ABZ inhibits melanoma growth and metastasis. Moreover, ABZ synergized with PD-L1 blockade mediates tumor regression.

Improving Robustness of Intent Detection Under Adversarial Attacks: A Geometric Constraint Perspective.

Deep neural networks (DNNs)-based natural language processing (NLP) systems are vulnerable to being fooled by adversarial examples presented in recent studies. Intent detection tasks in dialog systems are no exception, however, relatively few works have been attempted on the defense side. The combination of linear classifier and softmax is widely used in most defense methods for other NLP tasks. Unfortunately, it does not encourage the model to learn well-separated feature representations. Thus, it is easy to induce adversarial examples. In this article, we propose a simple, yet efficient defense method from the geometric constraint perspective. Specifically, we first propose an M-similarity metric to shrink variances of intraclass features. Intuitively, better geometric conditions of feature space can bring lower misclassification probability (MP). Therefore, we derive the optimal geometric constraints of anchors within each category from the overall MP (OMP) with theoretical guarantees. Due to the nonconvex characteristic of the optimal geometric condition, it is hard to satisfy the traditional optimization process. To this end, we regard such geometric constraints as manifold optimization processes in the Stiefel manifold, thus naturally avoiding the above challenges. Experimental results demonstrate that our method can significantly improve robustness compared with baselines, while retaining the excellent performance on normal examples.

High-normal serum uric acid predicts macrovascular events in patients with type 2 diabetes mellitus without hyperuricemia based on a 10-year cohort.

BACKGROUND AND AIMS: The upper limits of normal serum uric acid (SUA) or the lower limits of hyperuricemia were frequently set at 420 or 360 µmol/L (7.0 or 6.0 mg/dL). We aimed to explore the association between high-normal SUA (360 ≤ SUA≤420 µmol/L) and incidence of macrovascular and renal events based on a 10-year cohort with type 2 diabetes mellitus (T2DM) to explore which cut-off was more appropriate. METHODS AND RESULTS: A total of 2988 patients with T2DM without hyperuricemia (SUA≤420 µmol/L) were included and followed up. Cox proportional hazards models and restricted cubic spline regression were used to evaluate the relationship between baseline SUA (as continuous and categorical variable) and macrovascular and renal events. Patients were grouped as low-normal (SUA<360 µmol/L) and high-normal groups based on baseline SUA, and the latter group had higher incidence of macrovascular events. Multivariate Cox regression analysis indicated that baseline levels of SUA were significantly associated with cardiovascular (HR = 1.385, 95%CI:1.190-1.613, P < 0.001) and peripheral vascular events (HR = 1.266, 95%CI:1.018-1.574, P = 0.034), and the linear association existed. Moreover, fully adjusted multivariable Cox analyses indicated high-normal SUA increased the risks of cardiovascular (HR = 1.835, 95%CI:1.319-2.554, P < 0.001) and peripheral vascular events (HR = 1.661, 95%CI:1.000-2.760, P = 0.050) compared to low-normal SUA. CONCLUSIONS: Baseline SUA levels were positively associated with cardiovascular and peripheral vascular events, and high-normal SUA increased the risks of these events in patients with T2DM even without hyperuricemia. A threshold value for SUA of 360 µmol/L should be more appropriate in terms of predicting macrovascular events risks compared to the value of 420 µmol/L.

MULocDeep web service for protein localization prediction and visualization at subcellular and suborganellar levels.

Predicting protein localization and understanding its mechanisms are critical in biology and pathology. In this context, we propose a new web application of MULocDeep with improved performance, result interpretation, and visualization. By transferring the original model into species-specific models, MULocDeep achieved competitive prediction performance at the subcellular level against other state-of-the-art methods. It uniquely provides a comprehensive localization prediction at the suborganellar level. Besides prediction, our web service quantifies the contribution of single amino acids to localization for individual proteins; for a group of proteins, common motifs or potential targeting-related regions can be derived. Furthermore, the visualizations of targeting mechanism analyses can be downloaded for publication-ready figures. The MULocDeep web service is available at https://www.mu-loc.org/.

Time-Series Representation Learning in Topology Prediction for Passive Optical Network of Telecom Operators.

The passive optical network (PON) is widely used in optical fiber communication thanks to its low cost and low resource consumption. However, the passiveness brings about a critical problem that it requires manual work to identify the topology structure, which is costly and prone to bringing noise to the topology logs. In this paper, we provide a base solution firstly introducing neural networks for such problems, and based on that solution we propose a complete methodology (PT-Predictor) for predicting PON topology through representation learning on its optical power data. Specifically, we design useful model ensembles (GCE-Scorer) to extract the features of optical power with noise-tolerant training techniques integrated. We further implement a data-based aggregation algorithm (MaxMeanVoter) and a novel Transformer-based voter (TransVoter) to predict the topology. Compared with previous model-free methods, PT-Predictor is able to improve prediction accuracy by 23.1% in scenarios where data provided by telecom operators is sufficient, and by 14.8% in scenarios where data is temporarily insufficient. Besides, we identify a class of scenarios where PON topology does not follow a strict tree structure, and thus topology prediction cannot be effectively performed by relying on optical power data alone, which will be studied in our future work.

Exercise Changes Gut Microbiota: A New Idea to Explain that Exercise Improves Autism.

The effect of exercise interventions on autism spectrum disorder (ASD) has been demonstrated in many studies, and the discovery of a bidirectional relationship between the gut microbiome (GM) and the central nervous system (CNS) has led to the concept of the microbial gut-brain axis (MGBA) and has linked the abnormal GM to a variety of neuropsychiatric disorders, autism being one of them. Research on improving the GM through exercise is also starting to come into focus. However, there are currently few studies on exercise intervention in the GM of autism. The purpose of this review was to find evidence to explore the possible potential effects of exercise to improve the behavior of individuals with autism in the MGBA in this treatment, as well as the potential of GM as an exercise treatment for autism. We will explore (1) changes in GM components of ASD and their relationship to the pathophysiology of ASD; (2) the relationship between exercise and changes in GM components, and (3) the effect of exercise on GM in CNS disorders. Ultimately, we concluded that Streptococcus, Bifidobacterium, Clostridium, Bacteroides, and Blautia may be potential effectors through the MGBA network during exercise to ameliorate ASD targeting microbiotas. They deserve high attention in the follow-up studies.

Swimming exercise is a promising early intervention for autism-like behavior in Shank3 deletion rats.

INTRODUCTION: SHANK3 is an important excitatory postsynaptic scaffold protein, and its mutations lead to genetic cause of neurodevelopmental diseases including autism spectrum disorders (ASD), Philan McDermid syndrome (PMS), and intellectual disability (ID). Early prevention and treatment are important for Shank3 gene mutation disease. Swimming has been proven to have a positive effect on neurodegenerative diseases. METHODS: Shank3 gene exon 11-21 knockout rats were intervened by a 40 min/day, 5 day/week for 8-week protocol. After the intervention, the rats were tested to behavioral measures such as learning and memory, and the volume and H-spectrum of the brain were measured using MRI; hippocampal dendritic spines were measured using Golgi staining and laser confocal. RESULTS: The results showed that Shank3-deficient rats had significant deficits in social memory, object recognition, and water maze learning decreased hippocampal volume and number of neurons, and lower levels of related scaffold proteins and receptor proteins were found in Shank3-deficient rats. CONCLUSION: It is suggested that early swimming exercise has a positive effect on Shank3 gene-deficient rats, which provides a new therapeutic strategy for the prevention and recovery of neurodevelopmental disorders.

Association between physical health and physical activity behaviors for children aged 3-6 years in kindergarten: A cross-sectional study from China.

PURPOSE: To explore the association between the physical health (PH) and physical activity behavior (PAB) of children aged 3 to 6 years, and to provide a basis for the scientific planning of children's physical activity behavior in kindergarten. METHODS: A total of 18041 children in China aged 3 to 6 years were selected as the research subject. The PH monitoring indicators were measured according to "The National Physical Fitness Measurement Standards Manual" (Preschool Children Version) (NPFMSM), and PAB was reported by teachers. The chi-square test was used to test for differences in PH and PAB across groups, the percentile method was used to rank PAB, correlation analysis was used to analyze the association between components of PH and PAB, and multiple linear regression was used to determine the independent association between PAB and PH. RESULTS: Participants in this study exhibited poor PH status, and the detection rates of failure were 16.4% for boys and 16.8% for girls, and showed a decreasing trend with increasing age. The grade difference and sex difference in PH components were statistically significant (P<0.01). PAB changes to static activities with increasing age, mainly to static indoor physical activities (SIPAs). The regression effects of the total duration of physical activity (TDPA) and dynamic physical activity (DPA) on PH score and physical fitness (PF) score were significant (P < 0.01) but not statistically significant with static physical activity (SPA) (P>0.05). Dynamic outdoor physical activity (DOPA) is the core factor affecting children's PH, and is significantly positively correlated with the components of PH. CONCLUSION: PAB in kindergarten can improve children's PH, and reasonable planning of PAB in kindergarten has a targeted effect on PH promotion.

Irisin as a predictor of bone metabolism in Han Chinese Young Men with pre-diabetic individuals.

BACKGROUND: Irisin is a novel myokine both in mice and humans, and it can also be secreted by adipose tissue and the liver in a small amounts. There are few studies on irisin and bone metabolism. The aim of this study was to assess the relationship between serum irisin levels and bone metabolism and analyze its related factors in Han young male with pre-diabetic individuals. METHODS: This cross-sectional study included 41 pre-diabetes and 45 normal glucose tolerance (NGT). Anthropometric measurements, including height, weight, waist circumference (WC), and bone mineral content (BMC), were performed. All patients underwent an oral glucose tolerance test (OGTT) after 8 h of fasting, and the levels of glucose, insulin, lipids, serum irisin and bone turnover markers were measured. RESULTS: The levels of serum irisin (4.4 ± 1.4 vs. 6.3 ± 1.5 µg/mL), P1NP and OC were significantly lower and CTX was significantly higher in the pre-diabetes group (P < 0.05). BMC did not differ in the two groups (P > 0.05). Serum irisin levels negatively correlated with BMI (r =-0.325), FPG (r =-0.329), TG (r =-0.339) (P < 0.05) in NGT individuals. Serum irisin levels positively correlated with P1NP (r = 0.398), OC (r = 0.351), HDL-C (r = 0.432) and negatively correlated with FPG (r = -0.725), 2 h-PG (r = -0.360) (P < 0.05) in pre-diabetic individuals. Multiple regression analysis revealed that Serum irisin (ß = 9.768, P = 0.025) and WC (ß = -2.355, P = 0.002) were significant independent predictors for P1NP. CONCLUSION: Bone turnover markers were changed rather than bone mineral content in young men with pre-diabetes. In pre-diabetes individuals, serum irisin levels were reduced and close relationship with P1NP. Falling irisin levels may be a predictor of decreased bone formation in Han young men with pre-diabetes individuals.

A gradient boosting tree model for multi-department venous thromboembolism risk assessment with imbalanced data.

Venous thromboembolism (VTE) is the world's third most common cause of vascular mortality and a serious complication from multiple departments. Risk assessment of VTE guides clinical intervention in time and is of great importance to in-hospital patients. Traditional VTE risk assessment methods based on scaling tools, which always require rules carefully designed by human experts, are difficult to apply to large-population scenarios since the manually designed rules are not guaranteed to be accurate to all populations. In contrast, with the development of the electronic health record (EHR) datasets, data-driven machine-learning-based risk assessment methods have proven superior predictability in many studies in recent years. This paper uses the gradient boosting tree model to study the VTE risk assessment problem with multi-department data. There exist two distinct characteristics of VTE data collected at the level of the entire hospital: its wide distribution and heterogeneity across multiple departments. To this end, we consider the prediction task over multiple departments as a multi-task learning process, and introduce the algorithm of a task-aware tree-based method TSGB to tackle the multi-task prediction problem. Although the introduction of multi-task learning improves overall across-department performance, we reveal the problem of task-wise performance decline while dealing with imbalanced VTE data volume. According to the analysis, we finally propose two variants of TSGB to alleviate the problems and further boost the prediction performance. Compared with state-of-the-art rule-based and multi-task tree-based methods, the experimental results show the proposed methods not only improve the overall across-department AUC performance effectively, but also ensure the improvement of performance over every single department prediction.

Association between Sleep Duration and Physical Fitness in Children Aged 3-6 Years: A Cross-Sectional Study from China.

Aim: To explore associations between sleep duration and physical fitness (PF) of children aged 3−6 years. Methods: This study investigated the sleep duration and PF data of children aged 3−6 years by stratified random sampling. The restricted cubic spline model and binary logistic regression analysis were mainly used for the empirical analysis of the correlation effect between sleep duration and PF. The final data had a total of 21,857 children, of which 11,245 (51.45%) were boys and 10,612 (48.55%) were girls. Results: The PF level of the children in this study showed a relatively positive level (pass rate = 93.6%), and 19.7% of them had abnormal sleep duration; the results of the restricted cubic spline showed an inverted U-shaped association between the level of PF and the risk of abnormal sleep duration (X2 = 28.13, p < 0.0001). The results of logistic regression analysis showed that children with abnormal sleep duration were more likely to have a low PF, body morphology and motor ability levels at an OR (95% CI) of 1.077 (1.023−1.133), 1.077 (1.016−1.142) and 1.035 (1.08−1.062), respectively. The results of the bias correlation analysis showed varying degrees of correlation between sleep duration and various components of children's PF. Conclusion: Insufficient or excessive amounts of sleep were significantly associated with PF in children, with abnormal sleep duration leading to reduced levels of PF and its components.

Visualizing attention zones in machine reading comprehension models.

The attention mechanism plays an important role in the machine reading comprehension (MRC) model. Here, we describe a pipeline for building an MRC model with a pretrained language model and visualizing the effect of each attention zone in different layers, which can indicate the explainability of the model. With the presented protocol and accompanying code, researchers can easily visualize the relevance of each attention zone in the MRC model. This approach can be generalized to other pretrained language models. For complete details on the use and execution of this protocol, please refer to Cui et al. (2022).

Multilingual multi-aspect explainability analyses on machine reading comprehension models.

Achieving human-level performance on some of the machine reading comprehension (MRC) datasets is no longer challenging with the help of powerful pre-trained language models (PLMs). However, the internal mechanism of these artifacts remains unclear, placing an obstacle to further understand these models. This paper focuses on conducting a series of analytical experiments to examine the relations between the multi-head self-attention and the final MRC system performance, revealing the potential explainability in PLM-based MRC models. To ensure the robustness of the analyses, we perform our experiments in a multilingual way on top of various PLMs. We discover that passage-to-question and passage understanding attentions are the most important ones in the question answering process, showing strong correlations to the final performance than other parts. Through comprehensive visualizations and case studies, we also observe several general findings on the attention maps, which can be helpful to understand how these models solve the questions.

Effect of Hemoglobin A1c Trajectories on Future Outcomes in a 10-Year Cohort With Type 2 Diabetes Mellitus.

Background: Hemoglobin A1c (HbA1c) variability may be a predictor of diabetic complications, but the predictive values of HbA1c trajectories remain unclear. We aimed to classify long-term HbA1c trajectories and to explore their effects on future clinical outcomes in a 10-year cohort with type 2 diabetes mellitus (T2DM). Methods: A total of 2,161 participants with T2DM from the Beijing Community Diabetes Study were included. The 10-year follow-up was divided into two stages for the present data analysis. Stage I (from 2008 to 2014) was used to identify the HbA1c trajectories and to calculate the adjusted SD of HbA1c (HbA1c-adjSD), or the coefficient of variation of HbA1c (HbA1c-CV). Stage II (from 2014 to 2018) was used to collect the records of the new occurrence of diabetes-related clinical outcomes. Latent growth mixture models were used to identify HbA1c trajectories. Cox proportional hazards models were used to explore the relationship between HbA1c trajectories, HbA1c-adjSD, or HbA1c-CV and the future outcomes. Results: Three HbA1c trajectories were identified, including low stable (88.34%), gradual decreasing (5.83%), and pre-stable and post-increase (5.83%). Either the risk of death or the chronic complications were significantly higher in the latter two groups compared to the low stable group after adjustment for average HbA1c and other traditional risk factors, the adjusted hazard ratios (HRs) for renal events, composite endpoint, and all-cause death for the pre-stable and post-increase group were 2.83 [95%CI: 1.25-6.41, p = 0.013], 1.85 (95%CI: 1.10-3.10, p = 0.020), and 3.01 (95%CI: 1.13-8.07, p = 0.028), respectively, and the adjusted HR for renal events for the gradual decreasing group was 2.37 (95%CI: 1.08-5.21, p = 0.032). In addition, both univariate and multivariate Cox HR models indicated that participants in the fourth and third quartiles of HbA1c-CV or HbA1c-adjSD were at higher risk of renal events compared to participants in the first quartile. Conclusions: HbA1c trajectories, HbA1c-CV, and HbA1c-adjSD could all predict diabetes-related clinical outcomes. HbA1c trajectories could reflect long-term blood glucose fluctuation more intuitively, and non-stable HbA1c trajectories may predict increased risk of renal events, all-cause death, and composite endpoint events, independent of average HbA1c.