Remotely monitoring patients after they have been released from the hospital for serious infections does not reduce their risk of readmission, particularly for older adults, a randomized trial suggests.
The findings, released simultaneously in JAMA Network Open and at the Critical Care Reviews annual meeting in Belfast, emphasize the importance of studying potentially beneficial care models in real-world scenarios.
Remote monitoring involves the use of resources such as smartphone-based questionnaires, internet-connected scales, or blood pressure monitoring to allow patients to leave hospital.
These enable clinicians to monitor health from afar, while acting on alerts to ensure timely care.
While patients might prefer this type of care, the trial suggested that remote monitoring should be re-evaluated and individually tailored after acute care for sepsis and lower respiratory tract infections.
It did not increase the days patients spent at home and did not support U.S. federal agency efforts to reduce hospital readmissions by reimbursing remote monitoring.
“Health systems, insurers and policymakers all want to reduce hospital readmissions, and most patients prefer to recover safely at home,” explained lead researcher Sachin Yende, PhD, from the University of Pittsburgh.
“Remote monitoring has been held up as a solution, is reimbursed by the Centers of Medicare & Medicaid and its use has grown. But, aside from a few conditions, there’s a dearth of high-quality data to show it reduces readmissions.”
In an attempt to source this information, Yende and team conducted a complex trial in 1286 adults, with a median age of 63 years, who were recovering at home after hospitalization for sepsis or lower respiratory tract infections at 19 institutions. Of these patients, 32.6% received intensive care.
Participants were randomly assigned to one of four remote monitoring strategies or usual care.
Usual care generally involved a phone call from a nurse after discharge followed by continuing management from a primary care physician. The remote monitoring interventions consisted of short or long health questionnaire with standard or enhanced clinical response teams.
The standard response involved nurses who responded to alerts highlighted in the questionnaires together with coordinated care involving the patient’s primary and specialty clinicians.
Enhanced care included certified registered nurse practitioners with palliative care expertise who could coordinate some of the care response and also access social workers.
The primary endpoint was days at home at 90 days after discharge.
The trial showed that remote monitoring programs with high- or low-intensity questionnaires combined with a nurse response team or a nurse practitioner-led response team did not increase the time spent at home.
At least one readmission occurred among 37.8% of patients in the control group, receiving usual care.
This compared with 39.7% in patients receiving the shorter questionnaire and standard response team, 44.2% for those receiving the long questionnaire with standard clinical response, 37.3% receiving the shorter questionnaire but enhanced response team, and 36.3% for those receiving the long questionnaire with enhanced clinical response.
These differences did not reach statistical significance.
However, among patients aged 65 years and older, the remote monitoring programs did reduced time spent at home and increased readmission rates compared with usual care.
In this older age group, both the standard and enhanced response groups had fewer home days compared with usual care, with corresponding cumulative odds ratios of 0.56 and 0.67.
“These findings support reevaluating and tailoring remote monitoring in post–acute care following sepsis and lower respiratory tract infection to support further alignment with patients’ needs and desire for personalized monitoring,” the researchers concluded.
Scientists at the University of California San Diego have reported the results of a genome-wide association study in rats that identified key biological drivers of cocaine addiction. Using a genetically diverse group of nearly 900 rats to map genetic markers associated with compulsive drug use, the researchers uncovered a potential new therapeutic target that resides in the liver rather than in the brain.
Current research in this field often focuses on the brain, but the UC San Diego team’s findings suggest that how the body metabolizes cocaine may be just as critical in determining whether somebody develops an addiction.
“Finding a liver-based enzyme that shapes cocaine-taking behavior was a real ‘aha’ moment for us,” said Olivier George, PhD, a professor of psychiatry at UC San Diego School of Medicine. The George lab led the addiction behavioral studies that provided the foundation for the research. “It reminds us that addiction isn’t only in the brain. It’s a complex puzzle involving how the entire body processes the drug.”
Cocaine use disorder (CUD) has a strong genetic component, the authors noted. “Twin studies estimate the heritability of cocaine dependence to be as high as 70%, a finding supported by recent comprehensive reviews,” they wrote. GWAS have also uncovered a significant heritable component, the team continued, with single nucleotide polymorphism (SN)-based heritability estimated at 27-30%. However, scientists have struggled to pinpoint the specific genes that make certain individuals more vulnerable to addiction.
“The paucity of significant and replicated associations for CUD limits our understanding of this disorder, hampering our ability to identify novel pharmacological targets,” the investigators added. Co-corresponding author Abraham A. Palmer, PhD, professor of psychiatry at UC San Diego School of Medicine, who led the project’s intensive genetic modeling and analysis, further commented, “Identifying those genes in an important goal, because drugs could then be developed to target those genes, shifting genetically susceptible people to become more like genetically resistant people.”
To investigate further, the team carried out a GWAS in nearly 900 outbred Heterogeneous Stock (HS) rats—a model system capable of mimicking the vast genetic diversity found in human populations. By using HS rats the team was able to capture the critical differences between individuals who are genetically susceptible to addiction and those who are naturally more resistant. “Prior work has established the phenotypic diversity of HS rats across a broad range of addiction-relevant behaviors, including cocaine self-administration,” the researchers commented.
“The extended access model allowed us to characterize escalating intake, increased motivation to take the drug, and compulsive-like behavior despite negative consequences.” In addition to the GWAS results the researchers carried out a range of secondary analysis strategies to uncover what they describe as novel genetic drivers of cocaine self-administration behaviors.
Analyzing millions of genetic markers in each animal, the team discovered six major genetic regions linked to addiction-like behaviors, such as the escalation of drug intake and the time elapsed between doses. The researchers identified in the rats a specific group of carboxylesterase genes that are orthologous to the human CES1 gene, which are responsible for creating the enzyme that metabolizes cocaine. The study found that variations in these genes are closely linked to how frequently and compulsively rats self-administered the drug.
The findings also replicated a known genetic link found in humans (Trak2), providing a vital translational bridge between animal research and human medicine. This replication strengthens the argument that the biological pathways identified in the lab could eventually lead to real-world therapies. “Genes associated with CUD in humans remain limited, however our GWAS identified one gene (Trak2) that has also been identified by human GWAS of CUD, and the novel identification of Ces1 offers a fresh avenue for future studies,” they stated.
The collective findings suggest that by targeting the enzymes that metabolize cocaine with medicines, scientists might be able to alter how the drug affects the body, potentially reducing its addictive impact. In their paper they concluded “Our results replicate previous loci associated with CUD in humans and provide several novel biological insights including the potential of pharmacological strategies targeting carboxylesterases.”
Palmer said, “This work showcases the power of long-term, team-science collaboration that pairs experts in rodent behavior with quantitative geneticists. A decade of coordinated effort across multiple cohorts and federal partners made possible a discovery that no single lab could achieve alone.”
First author Montana Kay Lara, PhD, a postdoctoral researcher at UC San Diego School of Medicine, who helped bridge the gap between the study’s behavioral and genetic components, said, “Seeing the Ces1 signal validate a hypothesis that has been circulating for decades is incredibly exciting. It gives us a concrete target to test whether changing how cocaine is metabolized can blunt the drive toward compulsive use.”
The research team is now moving into the next phase of the project, which involves investigating exactly how these genetic mutations change function of the enzyme. They also hope to use the study’s extensive Preclinical Addiction Biobanks—collections of blood, urine, brain and other tissue samples—to identify biological markers that could one day help predict an individual’s risk of developing a substance use disorder.
The researchers hope that by leveraging this resource, they and other scientists working in this space will be able to translate genetic discoveries into diagnostic tools and new treatments that can help stabilize individuals struggling with addiction.
Background: Depression is a prevalent mental disorder, and it remains one of the leading causes of disability in Canada and globally. Mobile app–based physical activity interventions may offer an effective and accessible treatment option for individuals with depression who cannot or prefer not to access supervised exercise programs. Objective: This study aims to investigate the feasibility, acceptability, and proof of concept of a 9-week, theory-guided, app-based physical activity promotion intervention (MoodMover) developed for people with depression. Methods: We conducted a single-arm, pre-post study from November 2024 to May 2025, following the phase IIa: Proof-of-concept and phase IIb: Pilot and Preliminary Testing of the Obesity-Related Behavioural Intervention Trials model. Physically inactive adults who either self-reported a diagnosis of major depressive disorder or reported at least mild depressive symptoms, operationalized as a minimum score of 5 on the Patient Health Questionnaire, 9-Item, were recruited. The intervention spanned 9 weeks, with the first week serving as a run-in period and including a 15-minute orientation session on the first day. Participants were instructed to use the MoodMover program, delivered via the Pathverse app. Feasibility was assessed based on 4 primary criteria: recruitment, adherence, usability, and retention. Proof-of-concept was evaluated by assessing changes in physical activity behavior and depressive symptoms over the intervention period. Results: From November 2024 to March 2025, 32 of the 51 adults who met eligibility criteria consented to participate in this study, resulting in a recruitment rate of 63%. Twenty-eight participants completed baseline assessments, with a mean age of 39.8 (SD 13.4) years. A total of 21 participants attended the orientation session and received the intervention. Retention, adherence, and usability rates were 57% (16/28), 67% (14/21), and 50% (8/16), respectively. Regression analyses found that age consistently associated with app engagement, usability, and satisfaction. Two-tailed paired tests indicated significant pre-post changes in self-reported moderate to vigorous physical activity and depressive symptoms across the 75%, 85%, and 95% CIs. Among participants with clinically elevated depressive symptoms at baseline (Patient Health Questionnaire, 9-Item ≥10), 75% (9/12) achieved a clinically meaningful reduction in symptom severity. Conclusions: Our findings suggest that MoodMover holds potential for promoting physical activity behavior among individuals with depression and supporting depression management at scale. However, the feasibility of the tested version remains suboptimal. Necessary modifications (eg, improvements to enhance the accuracy of step tracking) should be implemented and reevaluated before progressing to a more rigorous efficacy trial. Trial Registration: ClinicalTrials.gov NCT06573125; https://clinicaltrials.gov/study/NCT06573125
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/3b1cfec4cf77a239ea121adef6075f91" />
Background: People with intellectual disability experience rates of mental illness up to 3 times higher than the general population, yet face significant barriers to care, including limited clinician expertise, diagnostic overshadowing, and exclusion from key mental health services. Electronic mental health interventions have demonstrated effectiveness in the general population and may address these barriers for people with intellectual disability by providing accessible, tailored treatment. Objective: This study examined the ability of Healthy Mind, a self-guided, web-based transdiagnostic mental health program designed specifically for people with borderline to mild intellectual disability, to reduce symptoms of anxiety and depression and improve functioning in people with intellectual disability. Methods: In a single-arm uncontrolled design, Australian residents aged ≥16 years with a diagnosis of borderline or mild intellectual disability, and mild-to-moderate symptoms of depression or anxiety on the Anxiety, Depression, and Mood Scale (ADAMS) were recruited online and offered access to Healthy Mind. Assistance from a nominated supporter was optional. Primary outcomes were symptoms of depression and anxiety (ADAMS). Secondary outcomes were psychological distress (Kessler Psychological Distress Scale [K10]) and functional impairment (World Health Organization Disability Assessment Schedule 2.0 [WHO-DAS]). Outcomes were assessed at baseline, 8 weeks, and 3 months. In total, 80 participants (mean age 27.8, SD 7 y; n=37, 46% female; n=39, 49% identifying as Aboriginal and/or Torres Strait Islander) enrolled; 61% (n=49) nominated a supporter. Data were analyzed using multilevel models with random intercepts for participants. Results: No significant changes were found in ADAMS depression or anxiety scores from baseline to postintervention or follow-up. Similarly, no significant effects were found for K10 or WHO-DAS scores, except for an improvement in K10 scores between baseline and 3 months when controlling for cognitive functioning (n=15). Having a supporter was associated with lower baseline distress but did not moderate treatment effects. Engagement with Healthy Mind was low; 42.5% (n=34) did not access the program, and among those who did, the median number of completed modules was 7 (IQR 3-11). Greater module completion was associated with slightly higher WHO-DAS scores post intervention. Conclusions: This trial did not demonstrate significant improvements in mental health or functioning associated with Healthy Mind, likely due to low engagement, reduced statistical power, and the absence of a control group. Nonetheless, the study demonstrates the feasibility of recruiting and retaining people with intellectual disability in fully online trials and highlights the urgent need for strategies to improve engagement, including gamification, personalized content, and integrated social features. Electronic mental health remains a promising avenue for addressing the substantial mental health service gap for people with intellectual disability. Trial Registration: ANZCTR ACTRN12620000113954; https://tinyurl.com/2ecdmde8 International Registered Report Identifier (IRRID): RR2-10.3390/ijerph18052473
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/7e2e8f60171580f8872a038a0a1d42e0" />
Background: The use of large language models (LLMs)–powered chatbots has reshaped how people seek information and advice, including for emotional and mental health support. While LLMs can offer scalable support, their ability to safely detect and respond to acute mental health crises—including suicidal ideation, self-harm, and violent thoughts—remains poorly understood. Progress is hampered by the absence of unified mental health crisis taxonomies, annotated benchmarks, and empirical evaluations grounded in clinical best practices. Objective: We addressed these gaps by introducing (1) a unified taxonomy of 6 clinically informed mental health crisis categories; (2) an evaluation dataset of over 2000 user inputs drawn from 12 publicly available conversational mental health datasets, classified into crisis categories; and (3) an expert-designed protocol for assessing response appropriateness. We also used LLMs to automatically identify crisis-indicative inputs and conducted an auditing study of 5 LLMs to evaluate the safety and appropriateness of their responses. Methods: We developed a taxonomy of mental health crisis categories informed by clinical experts and established literature. From over 239,000 mental health–related user inputs collected from 12 Hugging Face datasets, we curated 2252 examples (206 for validation, 2046 for testing) covering all taxonomy categories. We evaluated 3 LLMs on their ability to classify inputs into crisis categories, selecting the model with the strongest agreement with human annotators as the judge to label the test set. We then audited 5 LLMs on their ability to generate safe and appropriate responses to the 2046 test examples. Response quality was measured using a clinically informed 5-point Likert scale (1=harmful and 5=fully appropriate), relying on an LLM-as-a-judge validated against human expert feedback. Results: Several LLMs exhibited high consistency and generally reliable behavior when responding to explicit crisis disclosures, but significant risks remain. A nonnegligible proportion of responses was rated as inappropriate or harmful, particularly in the self-harm and suicidal ideation categories. Substantial performance differences were observed across models: gpt-5-nano and deepseek-v3.2-exp achieved very low harmful response rates, whereas gpt-4o-mini, Llama-4-Scout-17B-16E-Instruct, and grok-4-fast-non-reasoning generated markedly higher rates of unsafe outputs. All models exhibited systemic weaknesses, including poor handling of indirect or ambiguous risk signals, reliance on formulaic responses, and frequent misalignment with user context. Conclusions: These findings underscore the urgent need for enhanced safeguards, improved crisis detection, and context-aware interventions in LLM deployments and highlight the central role of alignment and safety engineering—beyond model scale or openness—in determining crisis response reliability. Our taxonomy, dataset, and evaluation framework lay the groundwork for ongoing research in artificial intelligence–driven mental health support, helping to minimize harm and protect vulnerable users.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/6d38d493c2ebfaa303202d21ad5c3a7d" />
Although gene therapy has shown promise for the treatment of Duchenne muscular dystrophy (DMD), the limitations of viral vectors have proven challenging to clinical advancement. Now, a new treatment platform delivered skeletal-muscle-targeted full-length DMD mRNA systemically in a murine model of DMD, successfully restoring the production of dystrophin, and dramatically improve muscle strength, endurance, and function in vivo.
The approach uses allogenically engineered targeting extracellular vesicles (DMD t-EVs)— which offer distinct benefits over current viral-based gene therapies, including reduced side effects and the ability to transfer the entire DMD gene. The researchers engineered the EVs with special tags that directly target skeletal muscles after being injected into the bloodstream. The work also demonstrated the safety and biocompatibility of DMD t-EVs in non-human primates, supporting their translational potential.
“Our new platform overcomes the limitations of current viral-based gene therapies, allowing for the delivery of full-length mRNA, restoring wild-type translation of dystrophin and significantly improving muscle function,” said Betty Kim, MD, PhD, in the department of neurosurgery at UT MD Anderson. “We are highly encouraged by these results, which provide a blueprint for mRNA-loaded EVs as a next-generation therapeutic strategy.”
The study, published today in Nature Biomedical Engineering, is entitled, “Skeletal-muscle-targeted non-viral delivery of full-length DMD mRNA for Duchenne muscular dystrophy.”
DMD is a severe genetic disorder caused by mutations in the DMD gene that prevent dystrophin production, which helps stabilize and protect muscle cells during contractions in healthy individuals. Without dystrophin, the muscles become easily damaged, leading to eventual inflammation and cell death. DMD primarily affects males, with symptoms such as delayed walking and waddling usually appearing in early childhood. As the disease progresses, it leads to loss of walking ability, scoliosis, heart problems and eventual respiratory failure.
Because DMD is the longest known gene in the human genome, current viral-based gene therapies are unable to carry the full length. These limitations result in the loss of the gene’s full function and prevent challenges like dose-limiting toxicities, immune reactions, and other adverse reactions including death.
These side effects have resulted in the removal of at least one Food and Drug Administration-approved gene therapy from the market and are why researchers have been trying to develop alternative ways of safely delivering the full-length DMD gene.
In this study, the researchers loaded the full-length DMD mRNA into EVs that were engineered to specifically target and bind to skeletal muscles. Injection of these mRNA-loaded EVs led to an increase in dystrophin protein expression as well as improved muscle strength and function in preclinical models, with no serious side effects.
Importantly, the treatment stayed on target inside of skeletal muscles and did not trigger any immune responses or toxicities commonly seen with viral-based treatments, even after repeated dosage.
Future studies are needed to determine the full safety of EV-mediated mRNA platforms for clinical trials, including whether they can be delivered to cardiac muscles, as heart conditions are commonly seen in advanced disease. However, based on these results, the authors point out this could be a promising method beyond treating Duchenne muscular dystrophy, also potentially serving as a broader “protein restoration” or cellular reprogramming platform.
“Given that we are now able to replace very large proteins, this platform- and disease-agnostic approach could potentially open doors far beyond rare genetic disorders and traditional gene therapy applications,” Kim said. “It’s possible this could ultimately enable restoration of proteins lost not only through inherited diseases but also from acquired or degenerative processes, including cancer, autoimmune disorders, neurodegeneration, fibrosis and other chronic diseases.”
Background: Primary health care (PHC) delivery in Pakistan is constrained by persistent workforce shortages, which are further exacerbated by the attrition of trained female physicians following marriage or childbirth. Telehealth platforms, such as Sehat Kahani, have emerged as one response to this gap, enabling female physicians to provide remote primary care from home. Within this model, a digital decision support system (DDSS) was recently piloted for selected febrile illnesses to strengthen clinical decision-making. However, evidence on how such systems are perceived by female PHC providers in low-income and middle-income country settings remains limited. In particular, there is limited understanding of how perceived usefulness, ease of use, and perceived impact on quality of care shape the early adoption of DDSS within tele-PHC workflows. Objective: This pilot study aimed to explore the acceptability, perceived ease of use, and perceived quality of care associated with a DDSS among female PHC providers delivering teleconsultations through a large-scale telehealth platform in Sindh, Pakistan. Methods: An exploratory pilot study using a sequential explanatory mixed methods design was conducted. Quantitative data were collected through an online survey of female health care providers (N=30) across 5 telehealth clinics using the DDSS. The survey assessed experiences related to DDSS utilization, usability, technical facilitation, satisfaction, and perceived diagnostic and treatment accuracy. This was followed by 3 focus group discussions to further examine facilitators and barriers to DDSS use. Survey data were analyzed descriptively, and qualitative data were analyzed thematically. Qualitative findings were used to explain and contextualize quantitative patterns. Results: Survey findings indicated frequent integration of DDSS into routine teleconsultation practice, with 43.3% (n=13) of providers using the system multiple times per day. DDSS was primarily used alongside clinical judgment (n=16, 53.3%) rather than as a standalone decision-making tool. Half (n=15, 50%) of participants reported confidence in the accuracy of DDSS-supported recommendations, while 46.7% (n=14) reported occasional reassessment of system outputs. Usability perceptions were generally positive, with 50% (n=15) reporting moderate satisfaction and 46.7% (n=14) finding the system easy to navigate. Qualitative findings contextualized these patterns, highlighting that DDSS enhanced decision-making confidence, supported care in unfamiliar clinical domains, and promoted standardized practice, while also revealing concerns about algorithm completeness and workflow burden. Conclusions: This pilot study suggests that DDSS embedded within telehealth platforms is acceptable to female PHC providers and can support clinical decision-making in resource-constrained PHC settings. By providing early implementation evidence from a real-world tele-PHC setting in a low- or middle-income country, this study contributes context-specific insights to inform iterative refinement and responsible scale-up of DDSS-enabled care models.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/20d5a081a7d5180003e727932eb12446" />
Antimicrobial resistance (AMR) is a significant global health threat, with its impact felt across all regions of the world.1 According to the World Health Organization (WHO), AMR is responsible for an estimated 700,000 deaths annually worldwide, and this figure is projected to rise to 10 million deaths per year by 2050 if current trends persist.
Notably, the number of deaths attributable to AMR in many countries surpasses those caused by diabetes, kidney diseases, digestive disorders, and other non-communicable diseases. AMR has profound implications for clinical practice, affecting the management of infections across various healthcare settings. Here, we will discuss recent advances using novel biomaterials and phage for combating AMR.
Wounds, in particular, are susceptible to colonization by AMR bacteria, complicating wound healing and increasing the risk of serious complications such as sepsis and amputation, which in turn exacerbates the chronic wound burden. Chronic wounds impact the healthcare system because of their increasing prevalence and cost. The rapid growth of AMR further limits the effectiveness of standard antibiotic therapies, necessitating the use of more potent and costly antimicrobial agents, which may have adverse effects and contribute to further resistance development.
The most common bacteria isolated from chronic wounds include species of Staphylococcus (47–55%), primarily S. aureus and S. epidermidis, P. aeruginosa (25–33.6%), Acinetobacter spp., Enterococcus faecalis, and Enterobacteriaceae such as Escherichia coli, Klebsiella pneumoniae, and Enterobacter spp.2 Many of these bacteria have developed persistent AMR, such as methicillin-resistant S. aureus (MRSA). They are highly resistant to commonly used antibiotics and, therefore, limit treatment options for wound infection. To combat the growing threat of infected wounds with AMR bacteria, Han and colleagues devised a creative approach to directly degrade proteins responsible for bacterial growth.3
UDP-N-acetylmuramoyl-L-alanine-D-glutamate ligase (MurD) is a prime target for combating antibiotic resistance in bacteria as it catalyzes the synthesis of peptidoglycan, the predominant structural component in bacterial cell walls. Han et al. developed a bacterial nanoinducer (bacNID) designed to specifically degrade MurD, effectively inhibiting the growth of both Gram-positive and Gram-negative bacteria.
Two critical, interconnected challenges
“Our paper addresses two critical, interconnected challenges in global public health and antibacterial therapy: 1) The crisis of antibiotic resistance. Bacteria rapidly evolve resistance to conventional antibiotics through mechanisms such as membrane permeability changes, target mutations, enzymatic inactivation, and efflux pumps; and 2) The failure of the traditional drug development model. The pharmaceutical industry faces a >95% failure rate in developing new antibiotics. Even when new drugs are found, bacteria often develop resistance quickly, and many candidates suffer from poor pharmacokinetics, systemic toxicity, and an inability to selectively target bacteria over healthy host cells,” says Guangjun Nie, PhD, senior author of the paper and professor at the National Center for Nanoscience and Technology, Beijing, China.
Nie adds that by moving away from the “one-target-one-drug” inhibition model, their study solves the problem of how to kill bacteria without giving them a chance to evolve resistance. “It achieves this by hijacking the bacteria’s own protein degradation machinery to destroy essential proteins such as MurD that are necessary for cell wall synthesis.”
The team first conjugated MurD-targeting peptides (pMurD) on gold nanoparticles, alongside the addition of a rapidly degradable SsrA peptide tag. The role of the SsrA tag is to bind MurD, thereby “tricking” the bacterial ClpXP protease into degrading MurD. Gold nanoparticles function as a peptide delivery vehicle that is taken up directly by the bacteria to circumvent the potential membrane permeability barrier. In this way, bacNID can destroy MurD, which is needed to synthesize the cell wall, leading to bacterial death.
The rapid growth of AMR limits the effectiveness of standard antibiotic therapies, necessitating the use of more potent and costly antimicrobial agents, which may have adverse effects and contribute to further resistance development. [10174593_258/Getty Images]
The team showed that bacNID was able to specifically inhibit model Gram-positive and Gram-negative bacteria with a dose-dependent degradation profile while exhibiting low cytotoxicity towards nontargeted mammalian cells. BacNID also specifically targeted MurD while sparing other Mur ligases. This approach can also be used with other nanoparticle vehicles, such as platinum, making it a versatile method for universal inhibition of diverse AMR bacteria.
To improve mechanistic understanding, the team utilized a variety of techniques and discovered that bacNID-treated bacteria suffered from cell wall damage, leading to leakage of a significant amount of DNA and ATP. Interestingly, compared with conventional antibiotics, treatment with bacNID did not lead to the formation of resistance after sustained treatment.
Using an in vivo infected skin wound model, the authors showed that bacNID treatment not only reduced infection burden but also promoted better wound healing outcomes, including greater skin cell proliferation, neo-angiogenesis, and lower inflammation. To expand the applicability of their method, bacNID was also tested in a S. aureus-infected nonhealing keratitis model and S. typhimurium-induced colitis model, showing great efficacy in both diseases.
“While our current study focuses on MurD, a major future step is to apply the bacNID platform to degrade other essential bacterial proteins. Readers can expect the team to develop bacNIDs against different targets in various pathogenic bacteria (e.g., targeting virulence factors or other metabolic enzymes). Future iterations of bacNIDs may incorporate stimuli-responsive nanotechnology (e.g., pH or enzyme-sensitive linkages) to ensure that the degradation-inducing activity is activated only within the specific microenvironment of the infection site, further minimizing off-target effects,” says Nie.
“We will also conduct more in-depth mechanistic studies to definitively elucidate why targeted protein degradation fails to induce the antibiotic resistance observed with conventional therapies.”
Guangjun Nie, PhD, professor at the National Center for Nanoscience and Technology, Beijing, China, with his team of scientists in the lab. [Guangjun Nie]
Trick bacteria with bacteria
Owing to an aging population, there is a rise in the use of implants, but these implants are prone to the formation of bacterial biofilm. The extracellular polymeric material in biofilm is known to reduce antibiotic penetration while creating an immunosuppressive environment, leading to impaired antimicrobial responses. In particular, orthopedic implants provide a conducive habitat for hematogenous bacteria for growth and formation of biofilm. Yang and colleagues hypothesized that bacteria that cause implant infection can be repackaged as drug carriers to penetrate biofilm for intra-film drug delivery.4
“Genetically modified bacteria have emerged as a promising delivery platform for diverse biomedical applications, ranging from cancer immunotherapy to infectious disease treatment. However, the clinical translation of current live bacterial biotherapeutics remains hindered by two major bottlenecks: unresolved in vivo safety concerns and the requirement for sophisticated species-specific genetic engineering. By exploiting the inherent life cycle of biofilms, our chemically primed bacterial triggers enable localized drug release deep inside biofilm structures, achieving effective biofilm eradication across genetically distinct bacterial and fungal infection models,” says Wei Tao, PhD, senior author and professor at Harvard Medical School.
The team first prepared bacteria by subjecting them to calcium chloride to increase membrane porosity and enhance their ability to uptake exogenous drugs like antibiotics. Ultraviolet radiation was then used to deactivate the bacterial membrane repair mechanism, creating irreversible membrane pores. They found that the modified bacteria, i.e., tricker, was able to migrate and thrive in biofilm and eventually, release exogenous drugs that are otherwise, challenging for delivery.
As a biofilm matures, surrounding bacteria are known to be attracted to and integrated into it. The team first labeled their tricker bacteria with a fluorescent dye and found that the bacteria were integrated throughout the biofilm with 80% coverage. However, a caveat is that the integration is most effective if the tricker and biofilm bacterial species are the same. The team discovered that while modified S. aureus can penetrate the core of S. aureus biofilm in 60 minutes, modified E. coli barely penetrates S. aureus biofilm. Likewise, modified E. coli can penetrate the core of E. coli biofilm, while modified S.aureus can penetrate E. coli biofilm with a much lower efficiency. Once in the biofilm, the chemically modified and inactivated bacteria were found to lyse, especially at hypoxic and acidic conditions.
Besides preventing antibiotic penetration, biofilm can also release bacterial-derived materials that suppress the immune system, particularly macrophages. For instance, it is well-characterized that S. aureus biofilms can bias macrophages towards an anti-inflammatory M2 phenotype, characterized by impaired antimicrobial peptide production, elevated arginase-1 (Arg-1), and attenuated inducible nitric oxide synthase (iNOS) expression. Interestingly, Yang and colleagues found that tricker bacteria were able to modify the metabolic states of the biofilm, resulting in enhanced production of I-arginine via iNOS to generate nitric oxide to improve bacterial clearance capacity.
Using an in vivo model of subcutaneous implant infection, the team found that there was an observable increase in mature dendritic cell and M1-like macrophage activation in the lymph nodes. The amount of memory B cells and antibodies with antimicrobial immune memory functions was also increased. After primary bacterial inoculation and intervention, the team reintroduced MRSA and found that 86% of treated mice rejected MRSA while all mice in the control group succumbed to the infection. This finding suggested that treatment with tricker bacteria was able to evoke innate and adaptive immune system endogenously for better control of AMR, with potential for bacterial-specific systemic memory to prevent relapse.
Finally, the strategy was tested in a murine bone infection model. By tracking cytokine levels and tissue histology, the team showed that their strategy was biologically safe. An MRSA rechallenge to the contralateral knee also led to a significant drop in biofilm burden in treated mice, providing convincing evidence of immune memory.
“To advance its clinical translation, the antibacterial efficacy of this approach will be further validated in large animal models, including rabbits, pigs, and dogs. This strategy exhibits enormous potential for future clinical translation of personalized antibacterial therapeutics, which enables highly efficient and precise treatment by profiling patient-derived pathogens and designing tailored “tricker” bacteria. Moreover, the current approach is adaptable to polymicrobial infections. Future work will also explore the feasibility of combining modified bacteria with other antibacterial agents or functional materials to optimize therapeutic performance,” adds Tao.
Wei Tao, PhD, professor at the Harvard Medical School (far right, first row) and his research team. [Wei Tao]
Using phage cocktail in clinical trials
Biofilm-related vascular graft infections (VGIs) are a major therapeutic challenge attributing to persistent, antibiotic-resistant bacteria residing in retained grafts. Graft explant is not always possible due to patient factors and surgical technical challenges. To effectively preserve the graft, treatment of VGI is typically a prolonged course of parenteral antibiotics followed by long-term suppressive antimicrobial therapy. Yet, graft survival rate is low, and recurrent infection is common.
Phages are viruses that specifically infect bacterial cells and can cause bacterial lysis. They have been shown to be active against both biofilm-forming bacteria and can even enhance antibiotic activity by eliciting phage-antibiotic synergies to combat AMR. Chung and colleagues made use of a phage cocktail to treat a 36-year-old female patient with refractory P. aeruginosa mediastinitis and vascular graft infection.5
“Our paper addresses key translational barriers to effective treatment of VGI caused by multidrug-resistant, biofilm-forming pathogens. Firstly, antibiotic failure in biofilm-associated infections as VGI pathogens embedded within biofilms exhibit marked tolerance to antibiotics, leading to persistent infection and relapse despite prolonged therapy. Secondly, escalating AMR as resistant subpopulations emerge under antibiotic pressure, further limiting treatment options in already complex infections. Thirdly, the lack of timely, personalized therapy as conventional phage therapy workflows are slow, making timely intervention difficult in acute or deteriorating cases.
Next, unpredictable phage–antibiotic interactions, such as phage-antibiotic synergy, are not reliably identified or optimised in routine clinical workflows. Finally, fragmented clinical-laboratory integration, as there is limited integration between real-time microbiology, pharmacology, and clinical decision-making to enable adaptive therapy,” says Andrea Kwa, PhD, senior author and associate professor at the SingHealth-Duke-National University of Singapore Medical School.
The team set up a multidimensional evaluation workflow to identify the most suitable therapeutic phages from the Singapore Phage Repository. Screening began with phage susceptibility testing of the four P. aeruginosa clinical isolates using spot and plaque assays, before other assays to identify the most potent cocktail. As phages are highly immunogenic when administered intravenously, the team also performed systemic inflammation monitoring and found that the patient tolerated the phages well.
The team found that their phage cocktail was able to restore antibiotic susceptibility by altering the efflux capacity of the bacteria. This positively impacted the antibiotic options for the patient. For instance, fluoroquinolone susceptibility was restored, resuscitating its use as an oral suppressive antibiotic for the long-term management of VGI.
Kwa adds that building on this proof-of-concept, her team’s next phase focuses on scaling, standardization, and integration of timely bespoke phage–antibiotic therapy into routine clinical practice. “Our key future directions include scaling up of rapid-response phage platforms, such as expansion of phage libraries/repositories with well-characterized, clinically ready phages with faster turnaround for matching and deployment, overcoming current procurement delays. We will also develop standardized precision workflows for phage susceptibility testing and phage–antibiotic synergy testing. Our team will also enhance regulatory and translational readiness of our technology for GMP-compatible production pipelines to enable scalable clinical deployment.”
AMR is a serious healthcare issue affecting the world. With the rising use of antibiotics in farms and clinical settings, this problem needs to be taken seriously. Unfortunately, the development of antibiotics is slow and mostly unsuccessful, thus requiring a new approach for society to effectively treat AMR. Biomaterials offer a new avenue to deliver tricker bacteria into biofilm to improve intra-film drug delivery and to activate the suppressed immune system, while also inhibiting intra-bacterial growth mechanisms. Phage is also becoming a popular option, especially for personalized medicine, and this therapy may see even greater efficacy when combined with biomaterials such as hydrogel to improve its delivery and reduce systemic immunogenicity.
References
Bertagnolio S, Dobreva Z, Centner CM, et al. WHO global research priorities for antimicrobial resistance in human health. Lancet Microbe. Elsevier Ltd. 2024;5(11). doi:10.1016/S2666-5247(24)00134-4
Uberoi A, McCready-Vangi A, Grice EA. The wound microbiota: microbial mechanisms of impaired wound healing and infection. Nat Rev Microbiol. Nature Research. 2024;22(8):507-521. doi:10.1038/s41579-024-01035-z
Han L, Huang W, Pan X, et al. Utilizing nanoinducers for precision degradation of bacterial protein to mitigate antibiotic resistance. Nature Communications . 2025;16(1). doi:10.1038/s41467-025-66221-w
Yang C, Saiding Q, Chen W, et al. Chemically modified and inactivated bacteria enable intra-biofilm drug delivery and long-term immunity against implant infections. Nat Biomed Eng. Published online January 16, 2026. doi:10.1038/s41551-025-01600-8
Chung SJ, Liu Y, Thong S, et al. Timely bespoke phage-antibiotic combination to treat refractory Pseudomonas aeruginosa mediastinitis and vascular graft infection. Nat Commun. Published online January 9, 2026. doi:10.1038/s41467-025-68136-y
For Anna Jeter, the story behind AOA Dx’s promising ovarian cancer diagnostic begins with a misconception. Despite serving more than half the population, women’s health continues to be viewed by many investors as a niche category—an assumption Jeter believes has contributed to decades of underinvestment in some of medicine’s most pressing unmet needs.
“Whenever you speak to investors about women’s health, there’s this perception that it’s somehow a smaller market,” Jeter, co-founder and chief regulatory officer of AOA Dx, told Inside Precision Medicine. “Women represent 51% of the population, but historically women’s health has been treated as a specialty category rather than a major healthcare market.”
That perception has shaped innovation patterns across healthcare, particularly in diagnostics. Despite advances in genomics, liquid biopsy, and AI in oncology, ovarian cancer detection has stagnated. The biomarker Cancer Antigen 125 (CA-125), introduced in 1987, remains the primary blood-based tool used to evaluate women presenting with symptoms suggestive of ovarian malignancy.
“There have been very few advancements in ovarian cancer,” Jeter said. “The standard of care today is still largely based on technology that became available almost 40 years ago.”
Early detection could have massive, life-changing consequences, as ovarian cancer is often highly treatable when diagnosed early. More than 90% of women survive when the disease is identified at stage 1, yet nearly 80% of cases are diagnosed only after the cancer has progressed to advanced stages.
After more than 15 years of developing diagnostics companies centered around women’s health, Jeter, Oriana Papin-Zoghbi, and Alex Fischer decided to take action and co-founded AOA Dx. Through an early partnership with McGill University, the team began investigating whether emerging advances in multi-omics could uncover biological signals that existing ovarian cancer diagnostics were missing.
Anna Jeter, co-founder and chief regulatory officer of AOA Dx
AOA Dx recently announced that its investigational blood test, AKRIVIS GD, achieved 92% sensitivity for Stage I and II ovarian cancer detection in symptomatic women—nearly double the sensitivity historically associated with CA-125 alone.
Yet Jeter argues the most important aspect of the findings is not the performance metric itself, but what enabled it: the successful translation of lipid biomarkers into a clinical oncology assay. “We are now looking at lipids as very specific biomarkers, which have made a major impact on our ability to detect ovarian cancer significantly earlier,” she said. “And we believe this approach can be translated into other cancers in the future.”
The announcement, which was recently made at the 2026 American Society of Clinical Oncology (ASCO) annual meeting, reflects a growing shift within cancer diagnostics away from single-analyte approaches and toward integrated biological modeling. Over the past decade, liquid biopsy innovation has largely centered on circulating tumor DNA (ctDNA), driven by advances in next-generation sequencing and the promise of minimally invasive genomic profiling. However, ctDNA performance has varied substantially across tumor types and disease stages, particularly in malignancies characterized by low tumor shedding during early progression.
Jeter contends that ovarian cancer exemplifies these limitations. “Certain cancers shed ctDNA very effectively into circulation, and in those cases the technology performs very well,” Jeter said. “But in other cancers, either the shedding is insufficient for the sensitivity required at early stages, or the specificity profile becomes challenging. That has led to a broader realization across the field that a single biological layer cannot fully capture disease progression.”
Increasingly, oncology researchers are converging around multi-omic frameworks that integrate orthogonal biological signals, including proteomic, metabolic, genomic, and transcriptomic features. Within that context, lipid metabolism has emerged as an area of renewed interest.
Lipid signaling, long associated primarily with cardiovascular disease and metabolic disorders, now plays a deeply intertwined role in oncogenesis. Alterations in lipid metabolism contribute to membrane remodeling, inflammatory signaling, tumor proliferation, immune modulation, and metastatic adaptation. Tumors frequently undergo profound metabolic reprogramming early in disease development, creating detectable shifts in circulating lipid and metabolite profiles before conventional biomarkers become clinically informative.
“What we have learned is that lipid metabolism appears to be highly active in the earliest stages of disease,” Jeter said. “That biology has been described in the literature for some time. The challenge has been translating those discoveries into a clinical-grade diagnostic assay.”
That translation problem is nontrivial. Lipids present substantial analytical complexity due to structural diversity, isomeric overlap, and variability in sample handling and quantitation. Discovery-level lipidomics frequently produces promising signals that prove difficult to standardize under regulated clinical laboratory conditions. AOA Dx’s central claim is that it has successfully navigated that transition.
The company’s platform integrates targeted lipidomic and metabolomic profiling using high-resolution mass spectrometry with protein biomarker analysis via immunoassays. Machine-learning models are then applied to derive composite signatures associated with ovarian malignancy.
New clinical biomarker class
According to the company, the assay was developed through biomarker discovery efforts involving more than 2,200 patient samples across multiple cohorts and demographic populations.
Importantly, the ASCO dataset focused specifically on the intended-use population: symptomatic women presenting in routine clinical settings with nonspecific but concerning symptoms such as abdominal pain, bloating, urinary changes, or altered bowel habits.
That population reflects a longstanding clinical challenge in gynecologic oncology. Although the majority of ovarian cancer patients report symptoms beginning relatively early in disease progression, those symptoms are often diffuse and overlap extensively with benign gastrointestinal or genitourinary conditions. As a result, diagnostic delays remain common, with patients frequently cycling through multiple evaluations before referral to gynecologic oncology.
“About 85–95% of women begin experiencing symptoms as early as stage 1,” Jeter noted. “But the symptoms are nonspecific enough that providers often do not initially suspect ovarian cancer.”
As a result, nearly 80% of ovarian cancers are still diagnosed at advanced stages, despite the dramatic survival differential associated with early intervention. When detected at stage 1 or 2, five-year survival rates exceed 90%. Once metastatic dissemination occurs, however, outcomes deteriorate sharply.
The company’s results suggest that multi-omic integration may provide clinically meaningful improvements over historical CA-125 performance, particularly in early-stage disease where existing diagnostic sensitivity has remained inadequate. According to AOA Dx, AKRIVIS GD achieved nearly double the early-stage sensitivity historically associated with CA-125 alone.
The broader implication is not simply improved biomarker performance but a reframing of how cancer detection may evolve over the next decade. Rather than relying on single molecular features, next-generation diagnostics are increasingly attempting to model disease as a systems-level biological phenomenon.
“There’s a growing recognition that different biological pathways become informative at different stages of disease progression,” Jeter explained. “Some protein markers rise later. Some genomic signals are difficult to detect early. By integrating lipidomic, metabolomic, and proteomic information simultaneously, we’re able to observe the disease from multiple biological angles.”
The company describes this infrastructure as a broader discovery engine known as GlycoLocate, which is designed to interrogate thousands of lipid and metabolite signatures across multiple disease states. Ovarian cancer represents the company’s first targeted clinical application, but executives indicate the platform is already being expanded into additional oncology indications, particularly within women’s health.
The women’s health investment gap
AOA Dx’s emergence also reflects a broader maturation within women’s health biotechnology, a sector that historically struggled to attract sustained institutional investment despite its substantial market opportunity.
Jeter has spoken openly about what she views as a longstanding disconnect between investor perception and commercial reality in women’s health innovation. Earlier this year, she presented an investment analysis at the 2026 JP Morgan Healthcare Conference examining historical returns within the category.
“Women’s health has traditionally been viewed as a niche market, despite women representing more than half the population,” she said. “Part of the issue is that many successful companies serving women have not historically been classified as women’s health investments, so the return profile of the sector has been systematically underestimated.”
That dynamic may now be shifting as women’s health increasingly intersects with other high-growth sectors, including artificial intelligence, precision medicine, molecular diagnostics, and computational biology. AOA Dx itself sits squarely at that convergence point, combining AI-driven biomarker modeling with high-complexity laboratory diagnostics and oncology applications.
The company is now advancing AKRIVIS GD toward validation studies and an early-access launch program. Equally important will be demonstrating clinical utility and securing reimbursement pathways, both of which remain major barriers for emerging diagnostics platforms. “It’s one thing to develop a test,” Jeter said. “It’s another thing entirely to ensure that patients can actually access it through reimbursement and coverage.”
To support commercialization efforts, AOA Dx has expanded its leadership team with executives experienced in diagnostic platform deployment and LC-MS assay translation, including chief product officer, Chris Roberts, and senior director of biomarker and analytical development, Cory Bystrom, PhD.
Whether lipidomics ultimately becomes a foundational pillar of oncology diagnostics remains uncertain. However, the field appears to be entering a transitional period in which metabolic biology is moving from peripheral exploratory science toward regulated clinical application.
For AOA Dx, the ambition extends beyond ovarian cancer itself. The company sees lipid-based multi-omics as an entirely new diagnostic layer capable of complementing—and in some settings potentially surpassing—the limitations of genomics-only approaches. “What I often say,” Jeter reflected, “is that the situation feels very similar to where next-generation sequencing was 10 or 15 years ago.”
If the company’s early data continues to hold in larger validation cohorts, AOA Dx may not simply be introducing another ovarian cancer assay. It may help to define the next biological frontier of liquid biopsy diagnostics.
![Guangjun Nie, PhD, professor at the National Center for Nanoscience and Technology, Beijing, China, with his team of scientists in the lab. [Guangjun Nie]](https://www.genengnews.com/wp-content/uploads/2026/06/Guangjun-2-300x201.jpg)
![Wei Tao, PhD, professor at the Harvard Medical School (far right, first row) and his research team. [Wei Tao]](https://www.genengnews.com/wp-content/uploads/2026/06/Wei-Tao-300x219.jpg)
