High-dimensional generalized median adaptive lasso with application to omics data.

Recently, there has been a growing interest in variable selection for causal inference within the context of high-dimensional data. However, when the outcome exhibits a skewed distribution, ensuring the accuracy of variable selection and causal effect estimation might be challenging. Here, we introduce the generalized median adaptive lasso (GMAL) for covariate selection to achieve an accurate estimation of causal effect even when the outcome follows skewed distributions. A distinctive feature of our proposed method is that we utilize a linear median regression model for constructing penalty weights, thereby maintaining the accuracy of variable selection and causal effect estimation even when the outcome presents extremely skewed distributions. Simulation results showed that our proposed method performs comparably to existing methods in variable selection when the outcome follows a symmetric distribution. Besides, the proposed method exhibited obvious superiority over the existing methods when the outcome follows a skewed distribution. Meanwhile, our proposed method consistently outperformed the existing methods in causal estimation, as indicated by smaller root-mean-square error. We also utilized the GMAL method on a deoxyribonucleic acid methylation dataset from the Alzheimer's disease (AD) neuroimaging initiative database to investigate the association between cerebrospinal fluid tau protein levels and the severity of AD.

Greenness and its composition and configuration in association with allergic rhinitis in preschool children.

BACKGROUND: Few studies focus on the associations of green space composition and configuration with children's allergic rhinitis (AR). METHODS: A multi-center population-based cross-sectional study was performed in 7 cities in mainland of China between 2019 and 2020, recruiting 36,867 preschool children. Information on the current AR symptoms and demographics were collected by questionnaire. Exposure to residential greenness was estimated by Normalized Difference Vegetation Index (NDVI, 1000 m buffer) around the residences. Greenness composition was estimated in 3 main categories: forest, grassland, shrubland. Configuration of each category and total greenness (a spatial resolution of 10 m × 10 m) was estimated by 6 landscape pattern metrics to quantify their area, shape complexity, aggregation, connectivity, and patch density. Exposure to daily ambient particulate matter (PM1, PM2.5 and PM10, a spatial resolution of 1 km × 1 km) was estimated. Multilevel logistic regression models were applied to analyze the associations of greenness and its composition and configuration with AR, and mediation effects by PMs were examined by mediation analysis models. RESULTS: The prevalence of self-reported current AR in preschool children was 33.1%. Two indicators of forest, Aggregation Index of forest patches (AIforest) (odds ratio (OR):0.92, 95% Confidential Interval (CI): 0.88-0.97), and Patch Cohesion of forest (COHESIONforest) (OR: 0.93, 95% CI:0.89-0.98) showed significantly negative associations with AR symptoms. Mediation analyses found the associations were partially mediated by PMs. Age, exclusive breastfeed duration and season were the potential effect modifiers. The associations varied across seven cities. CONCLUSION: Our findings suggest the inverse associations of the aggregation and connectivity of forest patches surrounding residence addresses with AR symptoms. Since the cross-sectional study only provides associations rather than causation, further studies are needed to confirm our results as well as the underlying mechanisms.

ASSOCIATION OF TESSELLATION DENSITY WITH PROGRESSION OF AXIAL LENGTH AND REFRACTION IN CHILDREN: An Artificial Intelligence-Assisted 4-Year Study.

PURPOSE: To investigate fundus tessellation density (TD) and its association with axial length (AL) elongation and spherical equivalent (SE) progression in children. METHODS: The school-based prospective cohort study enrolled 1,997 individuals aged 7 to 9 years in 11 elementary schools in Mojiang, China. Cycloplegic refraction and biometry were performed at baseline and 4-year visits. The baseline fundus photographs were taken, and TD, defined as the percentage of exposed choroidal vessel area in the photographs, was quantified using an artificial intelligence-assisted semiautomatic labeling approach. After the exclusion of 330 ineligible participants because of loss to follow-up or ineligible fundus photographs, logistic models were used to assess the association of TD with rapid AL elongation (>0.36 mm/year) and SE progression (>1.00 D/year). RESULTS: The prevalence of tessellation was 477 of 1,667 (28.6%) and mean TD was 0.008 ± 0.019. The mean AL elongation and SE progression in 4 years were 0.90 ± 0.58 mm and -1.09 ± 1.25 D. Higher TD was associated with longer baseline AL (ß, 0.030; 95% confidence interval: 0.015-0.046; P < 0.001) and more myopic baseline SE (ß, -0.017; 95% confidence interval: -0.032 to -0.002; P = 0.029). Higher TD was associated with rapid AL elongation (odds ratio, 1.128; 95% confidence interval: 1.055-1.207; P < 0.001) and SE progression (odds ratio, 1.123; 95% confidence interval: 1.020-1.237; P = 0.018). CONCLUSION: Tessellation density is a potential indicator of rapid AL elongation and refractive progression in children. TD measurement could be a routine to monitor AL elongation.

Birth Weight, Gestational Age, and Risk of Cardiovascular Disease in Early Adulthood: Influence of Familial Factors.

The association between intrauterine growth restriction and cardiovascular disease (CVD) later in life might be confounded by familial factors. We conducted a binational register-based cohort study to assess associations of birth weight for gestational age (GA), a proxy for intrauterine growth restriction, and GA with CVD risk in early adulthood, before and after addressing familial factors via sibling comparison. We included 3,410,334 live nonmalformed singleton births from Sweden (1973-1996) and Denmark (1978-1998). During a median follow-up period of 10 years from age 18 years onwards, 29,742 individuals developed incident CVD (hypertension, ischemic heart disease, or cerebrovascular disease). Compared with individuals born with appropriate birth weight for GA (AGA; 10th-90th percentiles) or full term (39-40 gestational weeks), individuals born severely small for GA (SGA; ≤3rd percentile) or preterm (22-36 weeks) were at increased risk of CVD (hazard ratio (HR) = 1.38 (95% confidence interval (CI): 1.32, 1.45) and HR = 1.31 (95% CI: 1.25, 1.38), respectively). The association was attenuated when comparing individuals born SGA with their AGA siblings (HR = 1.11, 95% CI: 0.99, 1.25) but remained robust when comparing individuals born preterm with their term siblings (HR = 1.21, 95% CI: 1.07, 1.37). Our findings suggest that both SGA and preterm birth are associated with CVD risk in early adulthood, with greater familial confounding noted for SGA birth.

Maternal Inflammatory Bowel Disease During Pregnancy and Infectious Disease in Offspring Younger Than 5 Years: A Population-Based Cohort Study.

INTRODUCTION: Maternal inflammatory bowel disease (IBD) during pregnancy may be associated with increased susceptibility to infection in offspring. We aimed to assess this association, taking into consideration the mediating role of anti-tumor necrosis factor α (anti-TNFα) agents and adverse birth outcomes. METHODS: This population-based cohort study included all live-born singletons born in Denmark during 1995-2016 (n = 1,343,960). The exposure was maternal IBD. Main outcome of interest was offspring infection younger than 5 years, defined by either infection-related hospitalization or systemic antibiotic prescription, whose corresponding risk estimates were hazard ratios (HRs) and incidence rate ratios (IRRs), respectively. We applied an inverse probability-weighted marginal structural model for mediation analysis. RESULTS: Offspring born to mothers with Crohn's disease (CD) had an 18% increased risk of infection-related hospitalization (HR 1.18, 95% confidence interval 1.10-1.26) and a 16% increased frequency of prescribed antibiotics (IRR 1.16, 95% confidence interval 1.11-1.21). Anti-TNFα agents could explain 10% and 3% of the 2 estimated total associations, respectively, while a composite of preterm birth, low birth weight, and small for gestational age could explain 4% and 0%, respectively. The association between prenatal anti-TNFα and frequency of antibiotics attenuated after additional adjustment for maternal CD (IRR from 1.23 [0.98-1.55] to 1.10 [0.87-1.40]). Maternal ulcerative colitis, however, was not associated with offspring infection. DISCUSSION: Maternal CD, but not ulcerative colitis, was associated with an increased risk of infection in offspring younger than 5 years, of which adverse birth outcomes and anti-TNFα had a minor role. The association between anti-TNFα agents and pediatric infection could be partially explained by maternal CD.

Maternal hypertensive disorders during pregnancy and the risk of offspring diabetes mellitus in childhood, adolescence, and early adulthood: a nationwide population-based cohort study.

BACKGROUND: Maternal hypertensive disorders during pregnancy (HDP) have been suggested to contribute to the development of offspring cardiovascular disease later in life, but empirical evidence remains inconsistent. This study was aimed to assess the association of maternal overall and type-specific HDPs with diabetes in offspring from childhood to early adulthood. METHODS: Using Danish national health registers, a total of 2,448,753 individuals born in Denmark from 1978 to 2018 were included in this study. Maternal HDP included chronic hypertension, gestational hypertension, and preeclampsia. The outcome of interest was diabetes in offspring (including type 1, type 2, and gestational diabetes). The follow-up of offspring started at birth and ended at the first diagnosis of diabetes, emigration from Denmark, death, or time end on 31 December 2018, whichever came first. Cox proportional hazards regression was used to evaluate the hazard ratios (HRs) with 95% confidence intervals (CIs) of the association between maternal HDP and diabetes (including type 1, type 2, and gestational diabetes) in offspring from birth to young adulthood (up to 41 years), with the offspring's age as the time scale. RESULTS: During a follow-up of up to 41 (median: 19.3) years, 1247 offspring born to mothers with HDP and 23,645 offspring born to mothers without HDP were diagnosed with diabetes. Compared with offspring born to mothers without HDP, those born to mothers with HDP had an increased risk for overall diabetes (HR=1.27, 95% CI=1.20-1.34), as well as for type 2 diabetes (HR=1.57, 95% CI=1.38-1.78) and gestational diabetes (HR=1.37, 95% CI=1.25-1.49). We did not observe obvious increased risk for type 1 diabetes (HR=1.08, 95% CI=0.98-1.18). Offspring of mothers with gestational hypertension (HR=1.37, 95% CI=1.00-1.88) or preeclampsia (HR=1.62, 95% CI=1.41-1.87) had higher risks of type 2 diabetes. The strongest association was observed for severe preeclampsia, with a 2-fold risk of type 2 diabetes (HR=2.00, 95% CI=1.42-2.82). The association between maternal HDP and type 1 diabetes did not reach statistical significance, except for maternal gestational hypertension (HR=1.41, 95%CI=1.17-1.71). In addition, we found that offspring born to mothers with any subtypes of maternal HDP had higher risk of gestational diabetes, and the corresponding HRs (95%CIs) for chronic hypertension, gestational hypertension, and preeclampsia were 1.60 (1.06-2.41), 1.29 (1.04-1.59), and 1.38 (1.24-1.53), respectively. We also observed stronger associations among offspring of mothers with HDP and comorbid diabetes (HR=4.64, 95%CI=3.85-5.60) than offspring of mothers with HDP or diabetes alone. CONCLUSIONS: Offspring of mothers with HDP, especially mothers with comorbid diabetes, had an increased risk of diabetes later in their life. Our findings suggest that timely and effective prevention of HDP in women of childbearing age should be taken into consideration as diabetes prevention and control strategies for their generations.

Improved ambient air quality is associated with decreased prevalence of childhood asthma and infancy shortly after weaning is a sensitive exposure window.

BACKGROUND: The urban ambient air quality has been largely improved in the past decade. It is unknown whether childhood asthma prevalence is still increasing in ever top-ranking city of Shanghai, whether the improved air quality is beneficial for children's asthma and what time window of exposure plays critical roles. METHODS: Using a repeat cross-sectional design, we analyzed the association between early life exposure to particles and wheezing/asthma in each individual and combined surveys in 2011 and 2019, respectively, in 11,825 preschool children in Shanghai. RESULTS: A significantly lower prevalence of doctor-diagnosed asthma (DDA) (6.6% vs. 10.5%, p < 0.001) and wheezing (10.5% vs. 23.2%, p < 0.001) was observed in 2019 compared to 2011. Exposure to fine particulate matter (PM2.5 ), coarse particles (PM2.5-10 ) and inhalable particles (PM10 ) was decreased in 2019 by 6.3%, 35.4%, and 44.7% in uterus and 24.3%, 20.2%, and 31.8% in infancy, respectively. Multilevel log-binomial regression analysis showed exposure in infancy had independent association with wheezing/DDA adjusting for exposure in uterus. For each interquartile range (IQR) increase of infancy PM2.5 , PM2.5-10 and PM10 exposure, the odds ratios were 1.39 (95% confidence interval (CI): 1.24-1.56), 1.51 (95% CI:1.15-1.98) and 1.53 (95% CI:1.27-1.85) for DDA, respectively. The distributed lag non-linear model showed the sensitive exposure window (SEW) was 5.5-11 months after birth. Stratified analysis showed the SEWs were at or shortly after weaning, but only in those with <6 months of exclusive breastfeeding. CONCLUSIONS: Improved ambient PM benefits in decreasing childhood asthma prevalence. We firstly reported the finding of SEW to PM at or closely after weaning on childhood asthma.

Machine learning to predict occult metastatic lymph nodes along the recurrent laryngeal nerves in thoracic esophageal squamous cell carcinoma.

PURPOSE: Esophageal squamous cell carcinoma (ESCC) metastasizes in an unpredictable fashion to adjacent lymph nodes, including those along the recurrent laryngeal nerves (RLNs). This study is to apply machine learning (ML) for prediction of RLN node metastasis in ESCC. METHODS: The dataset contained 3352 surgically treated ESCC patients whose RLN lymph nodes were removed and pathologically evaluated. Using their baseline and pathological features, ML models were established to predict RLN node metastasis on each side with or without the node status of the contralateral side. Models were trained to achieve at least 90% negative predictive value (NPV) in fivefold cross-validation. The importance of each feature was measured by the permutation score. RESULTS: Tumor metastases were found in 17.0% RLN lymph nodes on the right and 10.8% on the left. In both tasks, the performance of each model was comparable, with a mean area under the curve ranging from 0.731 to 0.739 (without contralateral RLN node status) and from 0.744 to 0.748 (with contralateral status). All models showed approximately 90% NPV scores, suggesting proper generalizability. The pathology status of chest paraesophgeal nodes and tumor depth had the highest impacts on the risk of RLN node metastasis in both models. CONCLUSION: This study demonstrated the feasibility of ML in predicting RLN node metastasis in ESCC. These models may potentially be used intraoperatively to spare RLN node dissection in low-risk patients, thereby minimizing adverse events associated with RLN injuries.

Association of Self-Reported Nighttime Sleep Duration with Chronic Kidney Disease: China Health and Retirement Longitudinal Study.

INTRODUCTION: The cohort study aimed to assess the association of nighttime sleep duration and the change in nighttime sleep duration with chronic kidney disease (CKD) and whether the association between nighttime sleep duration and CKD differed by daytime napping. METHODS: This study included 11,677 individuals from the China Health and Retirement Longitudinal Study (CHARLS) and used data from the 2011 baseline survey and four follow-up waves. Nighttime sleep duration was divided into three groups: short (<7 h per night), optimal (7-9 h), and long nighttime sleep duration (>9 h). Daytime napping was divided into two groups: no nap and with a nap. We used Cox proportional hazards model to examine the effect of nighttime sleep duration at baseline and change in nighttime sleep duration on incident CKD and a joint effect of nighttime sleep duration and nap time on onset CKD. RESULTS: With a follow-up of 7 years, the incidence of CKD among those with short, optimal, and long nighttime sleep duration was 9.89, 6.75, and 9.05 per 1,000 person-years, respectively. Compared to individuals with optimal nighttime sleep duration, short nighttime sleepers had a 44% higher risk of onset CKD (hazard ratio [HR]: 1.44, 95% confidence interval [CI]: 1.21-1.72). Compared to participants with persistent optimal nighttime sleep duration, those with persistent short or long nighttime sleep duration had an increased risk of incident CKD (HR: 1.44, 95% CI: 1.15-1.80). We found a lower incidence of CKD in participants with short nighttime sleep duration and a nap (HR: 0.74, 95% CI: 0.60-0.93), compared to those with short nighttime sleep duration and no nap. CONCLUSION: Short nighttime sleep duration and persistent long or short nighttime sleep duration were associated with a higher risk of onset CKD. Keeping persistent optimal nighttime sleep duration may help reduce CKD risk later in life. Daytime napping may be protective against CKD incidence.

Multiply robust estimator for the difference in survival functions using pseudo-observations.

BACKGROUND: When estimating the causal effect on survival outcomes in observational studies, it is necessary to adjust confounding factors due to unbalanced covariates between treatment and control groups. There is no study on multiple robust method for estimating the difference in survival functions. In this study, we propose a multiply robust (MR) estimator, allowing multiple propensity score models and outcome regression models, to provide multiple protection. METHOD: Based on the previous MR estimator (Han 2014) and pseudo-observation approach, we proposed a new MR estimator for estimating the difference in survival functions. The proposed MR estimator based on the pseudo-observation approach has several advantages. First, the proposed estimator has a small bias when any PS and OR models were correctly specified. Second, the proposed estimator considers the advantage pf the pseudo-observation approach, which avoids proportional hazards assumption. A Monte Carlo simulation study was performed to evaluate the performance of the proposed estimator. And the proposed estimator was used to estimate the effect of chemotherapy on triple-negative breast cancer (TNBC) in real data. RESULTS: The simulation studies showed that the bias of the proposed estimator was small, and the coverage rate was close to 95% when any model for propensity score or outcome regression is correctly specified regardless of whether the proportional hazard assumption holds, finite sample size and censoring rate. And the simulation results also showed that even though the propensity score models are misspecified, the bias of the proposed estimator was still small when there is a correct model in candidate outcome regression models. And we applied the proposed estimator in real data, finding that chemotherapy could improve the prognosis of TNBC. CONCLUSIONS: The proposed estimator, allowing multiple propensity score and outcome regression models, provides multiple protection for estimating the difference in survival functions. The proposed estimator provided a new choice when researchers have a "difficult time" choosing only one model for their studies.

An improved multiply robust estimator for the average treatment effect.

BACKGROUND: In observational studies, double robust or multiply robust (MR) approaches provide more protection from model misspecification than the inverse probability weighting and g-computation for estimating the average treatment effect (ATE). However, the approaches are based on parametric models, leading to biased estimates when all models are incorrectly specified. Nonparametric methods, such as machine learning or nonparametric double robust approaches, are robust to model misspecification, but the efficiency of nonparametric methods is low. METHOD: In the study, we proposed an improved MR method combining parametric and nonparametric models based on the previous MR method (Han, JASA 109(507):1159-73, 2014) to improve the robustness to model misspecification and the efficiency. We performed comprehensive simulations to evaluate the performance of the proposed method. RESULTS: Our simulation study showed that the MR estimators with only outcome regression (OR) models, where one of the models was a nonparametric model, were the most recommended because of the robustness to model misspecification and the lowest root mean square error (RMSE) when including a correct parametric OR model. And the performance of the recommended estimators was comparative, even if all parametric models were misspecified. As an application, the proposed method was used to estimate the effect of social activity on depression levels in the China Health and Retirement Longitudinal Study dataset. CONCLUSIONS: The proposed estimator with nonparametric and parametric models is more robust to model misspecification.

Covariate balance-related propensity score weighting in estimating overall hazard ratio with distributed survival data.

BACKGROUND: When data is distributed across multiple sites, sharing information at the individual level among sites may be difficult. In these multi-site studies, propensity score model can be fitted with data within each site or data from all sites when using inverse probability-weighted Cox regression to estimate overall hazard ratio. However, when there is unknown heterogeneity of covariates in different sites, either approach may lead to potential bias or reduced efficiency. In this study, we proposed a method to estimate propensity score based on covariate balance-related criterion and estimate the overall hazard ratio while overcoming data sharing constraints across sites. METHODS: The proposed propensity score was generated by choosing between global and local propensity score based on covariate balance-related criterion, combining the global propensity score fitted in the entire population and the local propensity score fitted within each site. We used this proposed propensity score to estimate overall hazard ratio of distributed survival data with multiple sites, while requiring only the summary-level information across sites. We conducted simulation studies to evaluate the performance of the proposed method. Besides, we applied the proposed method to real-world data to examine the effect of radiation therapy on time to death among breast cancer patients. RESULTS: The simulation studies showed that the proposed method improved the performance in estimating overall hazard ratio comparing with global and local propensity score method, regardless of the number of sites and sample size in each site. Similar results were observed under both homogeneous and heterogeneous settings. Besides, the proposed method yielded identical results to the pooled individual-level data analysis. The real-world data analysis indicated that the proposed method was more likely to find a significant effect of radiation therapy on mortality compared to the global propensity score method and local propensity score method. CONCLUSIONS: The proposed covariate balance-related propensity score in multi-site distributed survival data outperformed the global propensity score estimated using data from the entire population or the local propensity score estimated within each site in estimating the overall hazard ratio. The proposed approach can be performed without individual-level data transfer between sites and would yield the same results as the corresponding pooled individual-level data analysis.

Age-specific association of stage of hypertension at diagnosis with cardiovascular and all-cause mortality among elderly patients with hypertension: a cohort study.

BACKGROUND: Hypertension affects 31.1% of adults worldwide, with higher prevalence of great than 60% in elderly. Advanced hypertension stage was associated with the higher risk of mortality. However, little is known about the age-specific association of stage of hypertension at diagnosis on cardiovascular mortality or all-cause mortality. Therefore, we aim to explore this age-specific association among the hypertensive elderly through stratified and interaction analyses. METHODS: This cohort study included 125,978 elderly hypertensive patients aged 60+ years from Shanghai of China. Cox regression was used to estimate the independent and joint effect of hypertension stage and age at diagnosis on cardiovascular and all-cause mortality. Interactions were evaluated both additively and multiplicatively. Multiplicative interaction was examined by the Wald test of the interaction term. Additive interaction was assessed by relative excess risk due to interaction (RERI). All analyses were performed stratified by sex. RESULTS: 28,250 patients died during the follow-up up to 8.85 years, and 13,164 died of cardiovascular events. Older age and advanced hypertension stage were risk factors of cardiovascular mortality and all-cause mortality. Besides, smoking, rarely exercise, BMI < 18.5 and diabetes were also the risk factors. When we compared stage 3 hypertension with stage 1 hypertension, hazard ratios (95% confidence interval) of cardiovascular mortality and all-cause mortality were 1.56(1.41-1.72) and 1.29(1.21-1.37) for males aged 60-69 years, 1.25(1.14-1.36) and 1.13(1.06-1.20) for males aged 70-85 years, 1.48(1.32-1.67) and 1.29(1.19-1.40) for females aged 60-69 years, and 1.19(1.10-1.29) and 1.08(1.01-1.15) for females aged 70-85 years, respectively. Negative multiplicative interaction and positive additive interaction between age at diagnosis and stage of hypertension at diagnosis on cardiovascular mortality were observed in males (HR: 0.81, 95% CI: 0.71-0.93 RERI: 0.59, 95% CI: 0.09-1.07) and females (HR: 0.81, 95% CI: 0.70-0.93 RERI: 0.66, 95% CI: 0.10-1.23). CONCLUSIONS: Diagnosed with stage 3 hypertension was associated with higher risks of cardiovascular mortality and all-cause mortality, which were stronger among patients with age at diagnosis of 60-69 years compared with those with age at diagnosis of 70-85 years. Therefore, for the younger part of the elderly, the Department of Health should pay more attention to treating patients with stage 3 hypertension.

Childhood diabetes mellitus and early-onset kidney diseases later in life: a nationwide population-based matched cohort study.

BACKGROUND: The empirical evidence remains inconclusive for an association between diabetes mellitus (DM) in children and early-onset kidney disease later in life, and little is known about the effects of DM types (i.e., type 1 diabetes [T1DM] and type 2 diabetes [T2DM]) in childhood on type-specific kidney diseases. We aimed to evaluate the association of childhood DM with overall and type-specific early-onset kidney diseases later in life. METHODS: The population-based matched cohort study included 9356 individuals with DM (T1DM: 8470, T2DM: 886) diagnosed in childhood (< 18 years) who were born between 1977 and 2016, and 93,560 individuals without DM matched on sex and year of birth in Denmark. The main outcomes were overall and type-specific early-onset kidney diseases. The follow-up period of all included participants was from the date of DM diagnosis in the exposure group until the first diagnosis of kidney disease, emigration, or 31 December 2018, whichever came first. RESULTS: During a median follow-up of 13 years, children with DM had a 154% increased risk of early-onset kidney diseases than children without DM (adjusted hazard ratios 2.54, 95% confidence intervals 2.38-2.72), and T1DM (2.48, 2.31-2.67) and T2DM (2.75, 2.28-3.31) showed similar results. Children with DM also had a higher risk of multiple specific kidney diseases including glomerular diseases, renal tubulo-interstitial diseases, renal failure, and urolithiasis. The risks of type-specific kidney diseases including glomerular diseases and renal failure tended to be higher for children with T2DM (glomerular diseases: 5.84, 3.69-9.24; renal failure: 14.77, 8.53-25.59) than those with T1DM (glomerular diseases: 3.14, 2.57-3.83; renal failure: 8.24, 6.66-10.20). CONCLUSIONS: Children with DM had a higher increased risk of early-onset overall and specific kidney diseases later in life. Early prevention and treatment of both T1DM and T2DM in childhood may significantly reduce the risk of kidney diseases later in life.

Time-varying association between body mass index and all-cause mortality in patients with hypertension.

BACKGROUND: Relationship between BMI and all-cause mortality in patients with hypertension remains controversial. This study aimed to evaluate the time-varying association between BMI in patients with hypertension and all-cause mortality. METHODS: This population-based cohort study included 212,394 Chinese adults with hypertension from 2007 to 2015 and was followed up until death, loss-to-follow-up, or December 31, 2018. According to the World Health Organization criteria for Asians, BMI was categorized into five groups: underweight (BMI < 18.5 kg/m2), normal weight (18.5-22.9 kg/m2), overweight (23-24.9 kg/m2), class I obesity (25-29.9 kg/m2) and class II obesity (BMI ≥ 30 kg/m2). Cox model was used to estimate the time-varying association of BMI on the risk of mortality by including the interaction term between BMI and time using restricted cubic spline. RESULTS: Compared with normal weight, underweight and class II obesity were associated with higher mortality (Hazard ratio [HRs] at 1 and 10 years of follow-up: 1.51 [95% CI: 1.39-1.65], and 1.27 (1.15-1.41) for underweight, respectively; 1.08 (0.96-1.21), and 1.16 (1.03-1.30) for class II obesity, respectively). However, overweight and class I obesity were associated with lower mortality, although the protective effects gradually attenuated over time (HRs at 1 and 10 years of follow-up: 0.85 (0.81-0.90), and 0.96 (0.91-1.02) for overweight, respectively; 0.80 (0.76-0.84), and 1.04 (0.99-1.10) for class I obesity, respectively). CONCLUSIONS: We found increased mortality among hypertensive patients with underweight and class II obesity while decreased mortality with overweight and class I obesity was observed during the first 5 years of follow-up. Management efforts for hypertension may target controlling body weight in a reasonable range for patients, and probably more attention should be given to underweight patients.

Separate and combined effect of visit-to-visit glycaemic variability and mean fasting blood glucose level on all-cause mortality in patients with type 2 diabetes: A population-based cohort study.

AIMS: To assess the independent and combined impacts of visit-to-visit fasting blood glucose variability (VVV-FBG) and mean fasting blood glucose level (M-FBG) on all-cause mortality. MATERIALS AND METHODS: This prospective cohort study included 48 843 Chinese patients with type 2 diabetes. Cox proportional hazards regression models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) to evaluate the association of VVV-FBG and M-FBG with all-cause mortality. The potential nonlinear associations were examined using restricted cubic splines, and additive interaction was evaluated using relative excess risk due to interaction (RERI). Cox generalized additive models (CGAMs) and bivariate response surface models were further used to assess the combined effects of VVV-FBG and M-FBG. RESULTS: A total of 4087 deaths were observed during a median follow-up of 6.99 years. Compared with patients with values at the 5th percentile of average real variability (ARV) and M-FBG, we observed a 23% and 38% increased risk of premature deaths among those with values at the 95th percentile of ARV (HR 1.23, 95% CI 1.10, 1.37) and M-FBG (HR 1.38, 95% CI 1.26, 1.51), respectively. The interaction between glycaemic variability (ARV) and M-FBG was significant on both the additive scale (RERI 0.80 [0.29, 1.32]) and the multiplicative scale (HR 1.90 [1.10, 3.28]). High VVV-FBG and high M-FBG conferred the highest risk of all-cause mortality (HR 1.89, 95% CI 1.64, 2.17), compared to low VVV-FBG and low M-FBG. The CGAMs showed significant synergistic effects between glycaemic variability and M-FBG (P < 0.05). Moreover, a bivariate surface plot showed that risk of death increased more rapidly in type 2 diabetes patients with lower M-FBG combined with lower VVV-FBG. CONCLUSIONS: The coexistence of high glycaemic variability and high glucose level might exacerbate the independent risk of premature mortality in type 2 diabetes patients, highlighting the importance of achieving normal and stable glucose levels simultaneously in the management of glucose.

Estimation of average treatment effect based on a multi-index propensity score.

BACKGROUND: Estimating the average effect of a treatment, exposure, or intervention on health outcomes is a primary aim of many medical studies. However, unbalanced covariates between groups can lead to confounding bias when using observational data to estimate the average treatment effect (ATE). In this study, we proposed an estimator to correct confounding bias and provide multiple protection for estimation consistency. METHODS: With reference to the kernel function-based double-index propensity score (Ker.DiPS) estimator, we proposed the artificial neural network-based multi-index propensity score (ANN.MiPS) estimator. The ANN.MiPS estimator employed the artificial neural network to estimate the MiPS that combines the information from multiple candidate models for propensity score and outcome regression. A Monte Carlo simulation study was designed to evaluate the performance of the proposed ANN.MiPS estimator. Furthermore, we applied our estimator to real data to discuss its practicability. RESULTS: The simulation study showed the bias of the ANN.MiPS estimators is very small and the standard error is similar if any one of the candidate models is correctly specified under all evaluated sample sizes, treatment rates, and covariate types. Compared to the kernel function-based estimator, the ANN.MiPS estimator usually yields smaller standard error when the correct model is incorporated in the estimator. The empirical study indicated the point estimation for ATE and its bootstrap standard error of the ANN.MiPS estimator is stable under different model specifications. CONCLUSIONS: The proposed estimator extended the combination of information from two models to multiple models and achieved multiply robust estimation for ATE. Extra efficiency was gained by our estimator compared to the kernel-based estimator. The proposed estimator provided a novel approach for estimating the causal effects in observational studies.

Developing an immune signature for triple-negative breast cancer to predict prognosis and immune checkpoint inhibitor response.

Aim: We aimed to develop a new signature based on immune-related genes to predict prognosis and response to immune checkpoint inhibitors in patients with triple-negative breast cancer (TNBC). Materials & methods: Single-sample gene set enrichment was used to develop an immune-based prognostic signature (IPRS) for TNBC patients. We conducted multivariate Cox analysis to evaluate the prognosis value of the IPRS. Result: An IPRS based on 66 prognostic genes was developed. Multivariate Cox analysis indicated that the IPRS was an independent factor for prognosis. PD-1, PD-L1, PD-L2 and CTLA4 gene expression was higher in the low-risk group, suggesting IPRS could predict the response to immune checkpoint inhibitors. Conclusion: The IPRS might be a reliable signature to predict TNBC patients' prognosis and response to immune checkpoint inhibitors, but needs prospective validation.

Association of labour epidural analgesia with neurodevelopmental disorders in offspring: a Danish population-based cohort study.

BACKGROUND: Whether labour epidural analgesia impacts risk of neurodevelopmental disorders in offspring is unsettled, raising public and scientific concerns. We explored the association between maternal labour epidural analgesia and autism spectrum disorder, and specific developmental disorder, attention-deficit hyperactivity disorder, intellectual disability, and epilepsy in offspring. METHODS: This nationwide population-based cohort study included 624 952 live-born singletons delivered by women who intended to deliver vaginally (i.e. vaginal and intrapartum Caesarean deliveries) in Denmark from 2005 to 2016. A total of 80 862 siblings discordant for exposure to labour epidural analgesia were analysed in a sibling-matched analysis. Both full-cohort and sibling-matched analyses were performed to estimate hazard ratios (HRs) of offspring risk of autism spectrum disorder, specific developmental disorder, attention-deficit hyperactivity disorder, intellectual disability, and epilepsy, according to exposure to labour epidural analgesia, adjusted for maternal socio-economic, pregnancy, and perinatal covariates. RESULTS: In the full cohort, maternal labour epidural analgesia was associated with autism spectrum disorder in offspring (HR 1.11; 95% confidence interval [CI]: 1.04-1.18); however, in the sibling-matched analysis, no association with autism spectrum disorder was found (HR 1.03; 95% CI: 0.84-1.27). The association between labour epidural analgesia and specific developmental disorder (HR 1.12; 95% CI: 1.03-1.22) in the full cohort also disappeared in the sibling-matched analysis (HR 1.01; 95% CI: 0.78-1.31). No association between maternal labour epidural analgesia and the remaining neurodevelopmental disorders was found overall (attention-deficit hyperactivity disorder, HR 0.98; 95% CI: 0.92-1.03; intellectual disability, HR 0.98; 95% CI: 0.85-1.14; epilepsy, HR 0.89; 95% CI: 0.79-1.00) or in the sibling-matched analyses. CONCLUSIONS: Our findings did not support an association between maternal attention-deficit hyperactivity disorder and autism spectrum disorder, specific developmental disorder, attention-deficit hyperactivity disorder, intellectual disability, or epilepsy.

Death of a child and the risk of atrial fibrillation: a nationwide cohort study in Sweden.

AIMS: The role of psychological stress in the aetiology of atrial fibrillation (AF) is unclear. The death of a child is one of the most severe sources of stress. We aimed to investigate whether the death of a child is associated with an increased risk of AF. METHODS AND RESULTS: We studied parents with children born during 1973-2014 included the Swedish Medical Birth Register (n = 3 924 237). Information on death of a child, AF and socioeconomic, lifestyle and health-related covariates was obtained through linkage to nationwide population and health registers. We examined the link between death of a child and AF risk using Poisson regression. Parents who lost a child had a 15% higher risk of AF than unexposed parents [incidence rate ratio (IRR) and 95% confidence intervals (CI): 1.15 (1.10-1.20)]. An increased risk of AF was observed not only if the child died due to cardiovascular causes [IRR (95% CI): 1.35 (1.17-1.56)], but also in case of deaths due to other natural [IRR (95% CI): 1.15 (1.09-1.21)] or unnatural [IRR (95% CI): 1.10 (1.02-1.19)] causes. The risk of AF was highest in the 1st week after the loss [IRR (95% CI): 2.87 (1.44-5.75)] and remained 10-40% elevated on the long term. CONCLUSIONS: Death of a child was associated with a modestly increased risk of AF. Our finding that an increased risk was observed also after loss of a child due to unnatural deaths suggests that stress-related mechanisms may also be implicated in the development of AF.