BMC Medicine最新文献

Background: Metabolic dysfunction-associated steatohepatitis (MASH) is a major precursor of hepatocellular carcinoma (HCC), yet the molecular mechanisms linking steatohepatitis to malignancy remain poorly defined. The STE20-type kinase MST3 associates with hepatocellular lipid droplets and regulates metabolic homeostasis and stress responses in the liver. Here, we investigated whether pharmacologic inhibition of MST3 could attenuate the initiation and progression of MASH-associated HCC in vivo.

Methods: The therapeutic potential of MST3 inhibition was evaluated in a mouse model in which MASH-HCC was induced by a single diethylnitrosamine injection followed by 30 weeks of Western-style diet feeding. Liver tumor burden was assessed after 12, 21, or 30 weeks of treatment with Mst3-targeting antisense oligonucleotide (ASO) or a non-targeting control ASO, and histological, biochemical, and mechanistic analyses were performed in the 30-week cohort. In parallel, proteomic profiling of CRISPR/Cas9-generated MST3 knockout and wild-type Huh7 cells was conducted to gain molecular insight into MST3-regulated pathways.

Results: Mst3 ASO therapy had no impact on the onset or aggravation of experimentally induced MASH-associated HCC in mice, despite markedly improving the whole-body metabolic profile and suppressing all key features of MASH. Proteomic profiling of MST3-deficient hepatocytes revealed coordinated activation of mitochondrial and lysosomal pathways, consistent with enhanced fatty acid degradation and catabolic clearance.

Conclusions: Our findings in the selected mouse model of MASH-HCC suggest that MST3 is dispensable for hepatocarcinogenesis, yet its antagonism dampens diet-induced metabolic dysfunction and effectively attenuates MASH severity, challenging the prevailing assumption that targeting MASH driver genes alone is sufficient to prevent HCC development in the context of obesity.

Background: A fundamental public health goal is that all individuals have the opportunity to reach old age with adequate care and support. Japan is the global leader in longevity, and understanding whether this advantage exists primarily in healthy older adults or those relying on long-term care (LTC) can reveal if it stems from a healthier population or more extensive, and potentially higher-quality, healthcare provision. This study examined Japan's mortality advantage by comparing life expectancy and death rates in Japan and Sweden across different levels of LTC.

Methods: We included the entire population aged 75 + in Sweden (n = 858,595) and nine Japanese municipalities (n = 334,873), categorizing individuals into three groups: no care, home care, and care home residence. We compared age-specific death rates, remaining life expectancy, and expected time spent in each LTC state. Finally, we quantified how much of the overall mortality differences could be explained by LTC state-specific mortality difference.

Results: Japanese older adults had lower death rates and longer life expectancy than Swedish counterparts, with more pronounced differences among individuals utilizing LTC. At age 75, total life expectancy was 12.0 vs. 11.7 years for men and 15.5 vs. 13.7 years for women in Japan and Sweden, respectively. Expected time without LTC was 9.8 vs. 9.6 years for men and 10.4 vs. 9.9 years for women. The difference (95% CI) in total life expectancy [men, 0.3 (0.2, 0.4); women, 1.8 (1.7, 1.9)] exceeded the difference in time without LTC [men, 0.2 (0.2, 0.3); women, 0.5 (0.4, 0.5)], particularly for women. Higher mortality in home care and care home populations in Sweden substantially increased Japan's advantage.

Conclusions: Our findings show that Japan's longevity advantage in old age is primarily driven by lower mortality in the segment of the population utilizing LTC. This indicates that the overall advantage in life expectancy may not stem solely from a healthier population, but rather from more extensive, or possibly higher-quality, care, including life-sustaining treatments. However, since we were unable to control for differences in health status in the two populations, future studies should explore if the threshold for entering LTC is different in Sweden and Japan.

Background: Aberrant functional connectivity (FC) between the left dorsolateral prefrontal cortex (DLPFC) and subgenual anterior cingulate cortex (sgACC) is a replicated neural correlate of major depressive disorder (MDD). Emerging evidence suggests that individualized DLPFC-sgACC peak connectivity profiles may optimize transcranial magnetic stimulation (TMS) targeting and therapeutic outcomes. We hypothesized that the heterogeneity of DLPFC-sgACC peak FC locations could serve as a neurological basis for classifying distinct MDD subgroups.

Methods: We recruited 120 patients with MDD and used resting-state functional magnetic resonance imaging (MRI) to identify the peak of DLPFC-sgACC FC. Using the personalized peaks' spatial distribution, we clustered patients with MDD into subgroups and compared between-subgroup depressive and anxiety profiles. The classification performance of different clinical profiles between the subgroups was evaluated. The TMS therapeutic outcomes were retrospectively compared between these two subgroups in a small subsample of 37 patients who completed TMS treatment.

Results: The personalized DLPFC-sgACC peaks of patients with MDD were spread widely within the left DLPFC, clustering into the anterior (73.3%) and posterior (26.7%) subgroups. Peaks in the anterior subgroup were closer to the group-averaged peak, whereas peaks in the posterior subgroup were localized in the posterior DLPFC. The two subgroups showed comparable depressive symptom severity; however, the posterior subgroup showed more severe anxiety symptoms. The clustering of MDD subgroups demonstrated robust consistency across two datasets and different radii parameters. The logistic regression classifier based on the clinical profiles with between-subgroup differences showed good discrimination (AUC = 0.83). The anterior subgroup exhibited greater improvements in the anxiety and depressive symptoms than the posterior subgroup after receiving anterior DLPFC-targeted TMS treatment in the subsample.

Conclusions: Our findings suggest two distinct MDD subgroups characterized by differential spatial distribution of personalized DLPFC-sgACC peak FC. These subgroups demonstrate clinical differences in both anxiety symptom severity and TMS therapeutic outcomes, warranting validation through large-sample prospective studies in the future.

Background: The impacts of social isolation on diverse health conditions and how it contributes to health risks remain unclear. We aimed to investigate the associations of social isolation with 59 health conditions among older adults.

Methods: Participants from the UK Biobank without baseline diagnosis of the included diseases were selected. Social isolation was assessed with three questions. The 59 health conditions included all-cause mortality, 5 cause-specific mortalities, and 53 diseases. We used an instrumental variable from multivariable common factor GWAS in Mendelian randomization (MR) to explore causal links of social isolation with diseases. Omics analyses were conducted to assess the roles of Olink plasma proteins and metabolomics, and PERM was calculated to evaluate the influence of other factors.

Results: A total of 489,741 individuals [266,706 (54.5%) women; mean age 56.5 years (SD 8.1)] were included. During a median follow-up of 12.5 years, social isolation was uncorrelated with the majority of 59 health conditions. Significantly, it was associated with increased risks of all-cause [adjusted HR (aHR) 1.28, 95% CI 1.25-1.32], 5 cause-specific mortalities (aHR range, 1.18-1.38), and 11 specific diseases (aHR range, 1.08-1.17). Living alone was the strongest item of isolation in predicting mortality (aHR range, 1.18-1.45) and selected diseases. MR analyses offered little evidence to support a causal link between social isolation and these diseases. The proteins involved in these associations are predominantly related to "response to stimulus". Proteomic signatures (PERM, 36%-49%), health behaviours (32%-59%), and socioeconomic factors (22%-42%) were the main explanatory factors linking social isolation to 8 health outcomes.

Conclusions: Social isolation is associated with elevated risks of 17 out of the 59 examined adverse health outcomes, predominantly mortality-related conditions; however, MR analyses indicate an absence of evidence supporting causality for these associations.

Background: Alcohol use disorder (AUD) represents a tremendous societal burden, yet few efficacious therapies are available and widely used. Pre-clinical and human observational data support fibroblast growth factor 21 (FGF21) as a promising therapeutic target for the treatment of AUD. The objective of this study is to identify a robust genetic instrument for FGF21 agonism and leverage it to explore the effects of FGF21 agonism on AUD and related traits, as well as metabolic outcomes more widely.

Methods: We first compared associations with the positive control outcomes of liver fat and liver cirrhosis risk for the FGF21 cis-protein quantitative trait locus (cis-pQTL) (rs838131) to those for the common allele FGF21 L174P missense variant (rs739320). Having identified the L174P missense variant as a plausible genetic instrument, we subsequently performed association analyses investigating effects on AUD, related traits, and metabolic outcomes more widely. Finally, we performed colocalisation analyses to test whether observed association results reflect a causal mechanism that overlaps with the clinical effects of FGF21 on liver fat and liver cirrhosis.

Results: Consistent association and colocalisation evidence support a protective association between genetically predicted FGF21 agonism and alcohol consumption (association p = 1 × 10^-18, colocalisation posterior probability = 0.90), problematic alcohol use (association p = 0.02, posterior probability = 0.64), and AUD (association p = 9 × 10^-8, posterior probability = 0.97). Similar evidence was also observed for favourable effects of FGF21 on improving kidney function, lowering triglyceride levels, lowering proportional energy intake from carbohydrates, increasing proportional energy intake from protein and fat, increasing body weight and lowering waist-to-hip ratio.

Conclusions: This study identifies a genetic instrument for FGF21 effects to provide causal human evidence supporting favourable effects of FGF21 analogues for the treatment of AUD and related traits, as well as on metabolic outcomes more broadly. Further clinical study is duly warranted.

Background: Fetal hypoxia is a leading cause of neonatal morbidity and mortality. Cardiotocography (CTG) is widely used to predict fetal hypoxia during labor, but its interpretation remains suboptimal. Artificial intelligence (AI) models have been developed for CTG interpretation, but their clinical utility is limited by two major challenges: demonstrating superiority over human experts and ensuring explainability in real-world settings.

Methods: A large dataset containing CTG traces from three tertiary hospitals between January 2014 and May 2022 was built for model development. Deep learning architectures, named Cardiotocography Artificial-intelligence Predictors (CAPs), were trained to predict fetal hypoxia from CTG traces based on CNN (CAP-C), Transformer (CAP-T), LSTM (CAP-L), and CfC (CAP-CfC) algorithms. The outcome was fetal hypoxia, determined by either low Apgar score (≤ 7 at 1 or 5 min) or umbilical artery acidemia (grade 1: pH of umbilical artery (pHa) < 7.20; grade 2: pHa < 7.15; grade 3: pHa < 7.10). Model performance was determined by area under the receiver operating characteristic curve (AUROC), evaluated through nationwide AI-human comparison and validated on the CTU-UHB dataset. Gradient-weighted class activation mapping (Grad-CAM) was applied to highlight the CTG regions that contributed most to the model's predictions.

Results: A total of 20,780 CTG traces were obtained for model development, and 467 cases were held out for the nationwide AI-human comparison. Among all models, CAP-L achieved highest AUROC in predicting fetal hypoxia (grade 1: 0.758, 95% CI: 0.754-0.761; grade 2: 0.770, 95% CI: 0.764-0.776; grade 3: 0.716, 95% CI: 0.700-0.732). In comparison with 10,571 expert responses, all CAP models achieved higher AUROC (0.757-0.789 vs. 0.715, P values in Delong test < 0.05). On the public CTU-UHB dataset, CAP-L achieved AUROC of 0.709, 0.727, and 0.730 in predicting fetal hypoxia with grade 1, 2, and 3 acidemia. Grad-CAM analysis showed that the CAP models leveraged variable and prolonged decelerations to predict fetal hypoxia, verified by perturbation-based faithfulness test.

Conclusions: The CAP algorithms developed in this study showed superior performance in detecting fetal hypoxia from CTG traces compared to human experts, and demonstrated promising explainability, supporting clinical CTG interpretation.

Trial registration: Clinical trial registration number: ChiCTR2100045316, ChiCTR2100052695, ChiCTR2400085338.

Background: Immigrant women can experience high rates of domestic violence and abuse (DVA) and migration trauma. Family or general practitioners (GPs) have limited DVA training or support to manage culturally competent DVA practice and associated trauma. The HARMONY study aimed to increase culturally competent DVA identification and referral among all, but especially migrant/refugee women from South-Asia, attending Australian GP clinics.

Methods: Twenty-four GP clinics were recruited among two South-Asian communities in Northwest and Southeast Melbourne for a pragmatic cluster randomised controlled trial. Eligible clinics (i) employed ≥ 1 South-Asian GPs, (ii) used 1 of 2 electronic software programs, and (iii) agreed to anonymised, aggregated data extraction from computerised records. The intervention comprised (a) GP DVA educator and bilingual South-Asian DVA advocate co-delivering 4 h of online accredited culturally competent DVA training, and (b) 12 months follow-up support by the DVA advocate to intervention clinics. Comparison clinics offered routine care and were offered DVA training following the intervention's completion. Investigators and statistician were blinded to allocation, but clinics and frontline staff were not. Aggregated, anonymised routine data were extracted for primary outcomes of DVA identification and referral at 12 and 15 months. Per-protocol adjusted, intention-to-treat analysis using Poisson regression.

Results: Five of 24 GP recruited clinics withdrew before the trial began due to COVID-19. At baseline, GPs recorded DVA in 0.6% of 45,438 women, (but 0.4% among South-Asian women) and none recorded DVA referrals. Identification trended up in both arms, but we found no evidence of difference in DVA identification at 12 months in Intervention (0.98%, 252/25816) vs Comparison (0.88%, 199/22546), IRR 1.17 (95% CI 0.60 to 2.28) or at 15 months Intervention 1.01% (287/26218) vs Comparison 0.96% (217/22643), IRR 1.17 (95% CI 0.60-2.67). Referrals were rare (Int 14/252 vs Comp 6/199). 6/14 Intervention referrals were South-Asian women. No adverse events were reported.

Conclusions: While we found no evidence of HARMONY effectiveness, COVID-19 may have undermined its implementation. The model is promising for future research and refinement for clinics motivated to improve DVA management with diaspora minority ethnic communities.

Trial registration: ACTRN12618001845224p registered:13/11/2018.

Background: Digital health interventions offer a promising opportunity to innovate traditional cessation services. Multi-component digital approaches show potential for greater effectiveness than traditional methods, though optimal integration strategies and engagement dynamics are still unclear. This study evaluated 'Way to Quit' (WQ)-a theory-guided, integrated modality based on WeChat (a popular social media platform)-compared with a quitline, with particular attention to long-term engagement and its impact on abstinence.

Methods: A two-arm, open-label, randomised controlled trial was conducted in Beijing, China, from December 1, 2022 to June 13, 2024. We recruited adult smokers intending to quit within one month via online advertisements. Participants were randomly assigned (1:1; using various block sizes) to either a multicomponent WeChat-based intervention (WQ)-integrating a mini-programme (application-like service), a chat group (peer support) and an official account (information channel)-or telephone counselling (Quitline), both lasting four weeks (one pre-quit week and three post-quit weeks). All participants received a four-week supply of nicotine gum by mail for on-demand use. Follow-up assessments were completed at one, three, six, and twelve months after the initiation of the intervention via telephone or online. The primary outcome was biochemically verified 7-day point prevalence abstinence (PPA) at one month. Analyses followed a modified intention-to-treat (mITT) approach.

Results: A total of 460 adult smokers were enrolled and randomised (WQ: n = 230; Quitline: n = 230), with 414 included in the mITT analysis (WQ: 195; Quitline: 219). At one month, biochemically verified 7-day PPA was 41.5% in the WQ group and 27.9% in the Quitline group (RR, 1.49; 95% CI, 1.14-1.95; P = 0.004). Mini-programme engagement peaked initially but declined after three months, from 99.0% in one month to 6.7% in eight months. Greater usage during the first three months was dose-dependently associated with higher long-term abstinence rates (3-, 6-, and 12-month 7-day PPA), with all P_trend < 0.001.

Conclusions: The WQ modality showed superior short-term efficacy compared with the quitline, supporting its potential as an alternative or adjunct to traditional cessation services. Although long-term engagement remained challenging, early engagement intensity within the first three months strongly predicted twelve-month abstinence.

Trial registration: The trial was preregistered at Chinese Clinical Trial Registry Identifier (https://www.chictr.org.cn/bin/project/edit?pid=179013, ID: ChiCTR2200066427). Registration date: December 5, 2022.

Background: Most climate and health studies are based on a single extreme weather event and do not consider multiple physical processes across spatial and temporal scales. Our aim was to explore the current landscape of epidemiologic evidence on health effects linked to compound climate extreme events, compared with individual climate extreme events. We placed particular attention on exposure characterization methods and identifying opportunities for future research.

Methods: We searched the Web of Science, Scopus, and Medline/OVID databases for original research articles, using key concepts related to compound climate events, extreme events, and health risks between January 1950 and November 2025. Three reviewers conducted title/abstract screening and double-reviewed the remaining articles. We also developed a conceptual framework to outline the pathways between compound climate extremes and health outcomes.

Results: Our review identified 37 studies exploring the health impacts of compound climate events. The health outcomes assessed were mortality, mental health, cardiovascular, respiratory, adverse pregnancy, and others. Most study designs were time-series or case-crossover, and several studies identified vulnerable subpopulations. Most quantitative studies reveal that compound climate extreme events exhibit an elevated risk of health impacts (n = 34), especially compared to non-compound events (n = 31). Among the 27 studies that compared health risks of compound extremes with their individual component events, 20 found an elevated risk of the compound event. Only seven studies assessed synergistic effects formally, with four indicating a greater joint risk of compound climate events compared to what would be expected from the combined effects of those events in isolation.

Conclusions: There is an emerging body of literature characterizing the substantial health impacts of compound climate extremes. Research has increasingly explored spatial heterogeneity and complex exposure-response associations emerging from compound climate exposures; however, definitions used to characterize compound climate extremes vary widely. Future research should further examine the synergistic mechanisms of compound climate extreme events and determinants of associated health outcomes to enhance public health and disaster preparedness, using a flexible exposure definition framework.

Background: Autoimmune conditions are common in women of reproductive age. They are associated with an increased risk of adverse pregnancy outcomes; whether this is causal is unclear. Our aim was to explore the causal effects of genetic liability to autoimmune conditions on pregnancy outcomes.

Methods: We conducted two-sample Mendelian randomization (MR) to estimate effects of genetic liability to ten autoimmune conditions on nine primary and seven secondary pregnancy outcomes. We used data from the MR-PREG collaboration including up to 714,889 pregnancies. Main analyses used the inverse variance weighted method to pool effects across genetic variants. Sensitivity analyses explored bias due to pleiotropic variants and fetal genetics.

Results: We found evidence for 7 effects of autoimmune condition genetic liability on primary pregnancy outcomes that were robust across all sensitivity analyses. Higher genetic liability to Hashimoto's thyroiditis was protective against large-for-gestational-age and increased the risk of hypertensive disorders of pregnancy (HDP) and preterm birth. For example, risk of preterm birth increased by 6% (OR = 1.06 (95%CI: 1.02, 1.11)) per doubling in log odds of Hashimoto's thyroiditis. Genetic liability to type 1 diabetes and systemic lupus erythematosus each increased the risk of preterm birth only. Higher genetic liability to rheumatoid arthritis increased the risk of HDP, while higher ankylosing spondylitis genetic liability reduced the risk of HDP. For multiple sclerosis, systemic sclerosis, coeliac disease, inflammatory bowel disease, and psoriasis, we did not detect any robust effects of increased genetic liability.

Conclusions: We observed higher genetic liability to Hashimoto's thyroiditis, type 1 diabetes, and rheumatic conditions causes increased risk of adverse pregnancy outcomes. Further research is warranted to explore the immune mechanisms underlying these relationships, and identify targets for prevention.