Artificial intelligence (AI) shows promise in identifying psychopathology through language, but replicability in AI models remains challenging. We develop an AI-based language assessment of posttraumatic-stress-disorder (PTSD) severity and introduce the sequential evaluation with model preregistration to rigorously evaluate its validity and replicability. This design includes two phases: development with preregistration and evaluation. Data included development (N = 1,437) and prospective (N = 346) samples, in which participants described their lives during automated interviews. In the prospective sample, preregistered models correlated with PTSD CheckList scores (r = .38, p < .001) and converged with PTSD diagnosis (area under the curve [AUC] = .76; outperforming demographics and trauma exposures: AUC = .61, p < .01). We found that for each standard-deviation increase, mental-health-care expenditure rose by $696.50 (p < .001). Our preregistered PTSD model assessments are replicable in prospectively collected clinical data and showed external validity against expense criteria. With further development, such models can be used to screen for PTSD or monitor treatment response, especially in telehealth or automated interviews, in which deployment can be seamless.

Language-based assessments of health have demonstrated validity in terms of convergence with self-report questionnaire scores, but recent work has shown they lack an important psychometric property: discriminant validity, the ability of the measure to distinguish the target construct from a related one. For example, a language-based general mental health assessment may accurately assess the degree of mental health, but it may also produce scores that correlate with general physical health well beyond theoretical expectations. Here, we propose and evaluate two methods for altering the standard loss function to directly penalize off-target correlations while retaining convergent validity. Evaluated across two clinical language datasets spanning physical and mental health, the augmentations substantially reduced off-target correlations with minimal loss in convergent validity (on-target correlations). Limiting loss in convergent validity to 0.005 Pearson correlation points, the loss function using Squared Cosine Similarity Discrimination significantly improved discriminant validity of language-based mental and physical health assessments (r = 0.454 → r = 0.322; p &lt; 0.05); and fundamental psychopathology dimensions (r = 0.442 → 0.430; p &lt; 0.05). These findings demonstrate that enforcing discriminant validity as a training objective is effective, moving language-based assessments closer to the specificity required for differential clinical use.

Background. Mental and physical health are correlated, yet are described differently in natural language. Standard symptom checklists may limit individuals from fully describing how they experience health problems. This study examines the distinction between how individuals linguistically describe mental and physical health. Method. We used the Sequential Evaluation with Model Pre-registration (SEMP) design to analyze language from automated, open-ended clinical interviews with 9/11 first responders. We developed language-based models assessing Mental and Physical Component Summary (M/PCS) scores from automated interview transcripts (N=1290). We then evaluated these models on held-out prospective data (N = 310) against M/PCS, medical record diagnoses, and healthcare expenditures. Results. Language-based assessment correlated significantly with M/PCS scores (r = .36 and r = .37, respectively, all p&lt;.001), exceeding SEMP pre-registered thresholds (r&gt;.315, and r&gt;.348, respectively). Analyses revealed both shared and distinct linguistic markers for mental and physical health, with language reflecting emotional, somatic, and functional themes. The language-based assessments demonstrated external validity through significant convergence with clinical diagnoses (AUCs=.74 and .65 for mental and physical health, respectively). Furthermore, language-based assessments exhibited incremental validity; when added to models containing traditional self-reports, they explained significant unique variance in both mental (ΔR²=.063, p&lt;.001) and physical (ΔR²=.032, p=.010) healthcare costs. Conclusions. Natural language from automated interviews captures general health through emotional, somatic, and functional themes. Validated against clinical diagnoses and healthcare costs, these models demonstrate that open-ended language provides interpretable insights beyond traditional rating scales, potentially augmenting clinical evaluations and understanding.

Abstract Introduction World Trade Center (WTC) responders suffer from high rates of chronic diseases such cancer, cardiovascular disease, respiratory disease, and mental health conditions. Lifestyle medicine approaches — which focus on promoting healthy behaviors such as regular exercise, Mediterranean diet, and adequate sleep — can prevent and reverse WTC-associated diseases. Lifestyle medicine interventions are effective but are constrained by high attrition rates (~50%). Behavioral predictors of attrition from these interventions are not well-understood. To address this gap, we examined: 1) Fitbit estimates of sleep and physical activity, 2) self-reported personality and mental health, and 3) large language model (LLM) derived estimates of sleep and physical activity as predictors of attrition from a lifestyle medicine intervention. Methods Participants were 323 WTC first responders enrolled in the WTC Health Program in Long Island, NY between 2023-2025. Participants completed a 6-month lifestyle medicine intervention, which involved 4 visits with a registered dietician, wearing a Fitbit Inspire 3 device for one week at baseline, as well as completing self-report surveys to assess personality and mental health. LLM derived variables were computed using ChatGPT5 estimates of sleep and physical activity from intake visit audio transcripts between participants and dieticians. Results Together, predictors explained 9% of the variance in attrition (AUC = 67%). Higher Fitbit sleep efficiency at baseline predicted lower odds of attrition from the lifestyle medicine intervention (OR = 0.58, p = 0.004). Higher self-reported extraversion predicted higher odds of attrition (OR = 1.54, p = 0.005). Other Fitbit, self-report, and LLM-derived variables did not predict attrition. Conclusion Results suggest fragmented sleep may be a strong risk factor for study dropout. Targeting sleep via cognitive behavioral therapy for insomnia may improve study engagement and adherence to lifestyle medicine practices. Support (if any) CDC/NIOSH R21OH012614 (PI: Kotov)

World Trade Center (WTC) responders endured exposures to neurotoxic dust particulate matter. This neuroimaging study examined the presence of amyloidosis in Alzheimer’s disease (AD) regions of interest (ROIs) and associations with exposure duration. Simultaneous positron-emission tomography with [ 18 F]-florbetaben and magnetic resonance neuroimaging was acquired on 34 middle aged WTC responders. Centiloid scale used mean standardized uptake value ratio (SUVR) with between subjects comparisons of cognitively unimpaired (n=17) and early-onset dementia (n=17) responders. Pathway analyses investigated potential neuropathological cascades. Only one subject was Centiloid positive. However, considerable SUVRs in non-AD ROIs associated with changes in cortical mean diffusivity, cortical thickness, cognition, and exposure duration, mediated by the olfactory bulb. Fibrillar amyloidosis may have originated in, and spread from the olfactory cortex because of prolonged exposures to WTC neurotoxic dust. Our study also identifies a non-AD neuropathological topology of amyloidosis, which associated with lower measures of brain health.

STUDY OBJECTIVES:Insufficient sleep costs the US economy over $411 billion per year. However, most studies investigating the economic costs of sleep rely on one-time measures of sleep, which may be prone to recall bias and cannot capture variability in sleep. To address these gaps, we examined how sleep metrics captured from daily sleep diaries predicted medical expenditures. METHODS:Participants were 391 World Trade Center (WTC) responders enrolled in the WTC Health Program (mean age = 54.97 years, 89% men). At baseline, participants completed 14 days of self-reported sleep and stress measures. Mean sleep, variability in sleep, and a novel measure of sleep reactivity (i.e. how much people's sleep changes in response to daily stress) were used to predict the subsequent year's medical expenditures, covarying for age, race/ethnicity, sex, medical diagnoses, and body mass index. RESULTS:Mean sleep efficiency did not predict mental healthcare utilization. However, greater sleep efficiency reactivity to stress (b = $191.75, p = .027), sleep duration reactivity to stress (b = $206.33, p = .040), variability in sleep efficiency (b = $339.33, p = .002), variability in sleep duration (b = $260.87, p = .004), and quadratic mean sleep duration (b = $182.37, p = .001) all predicted greater mental healthcare expenditures. Together, these sleep variables explained 12% of the unique variance in mental healthcare expenditures. No sleep variables were significantly associated with physical healthcare expenditures. CONCLUSIONS:People with more irregular sleep, more sleep reactivity, and either short or long sleep engage in more mental healthcare utilization. It may be important to address these individuals' sleep problems to improve mental health and reduce healthcare costs.

Clonal hematopoiesis of indeterminate potential (CHIP) represents the presence of clonal somatic mutations in blood cells in otherwise healthy individuals. While CHIP is known to increase risk for hematologic malignancies and cardiovascular disease, its association with airborne carcinogens remains largely unknown. We investigated CHIP mutations in 9/11 World Trade Center (WTC) disaster responders (n=350), who experienced substantial exposure to a complex mix of airborne carcinogens. Ultra-deep whole-exome sequencing at 250X was performed on banked blood samples. We characterized CHIP mutations and their associations with clinical factors (age, ancestry, gender, body mass index, cardiovascular disease, stroke), laboratory parameters (peripheral blood counts), mental and cognitive assessments, exposure data, and HLA zygosity. Statistical methods included Fisher's exact test, Wilcoxon rank sum test, and multivariate logistic regression. Findings were compared to unexposed controls (n=293) analyzed using identical methods. Among WTC participants, CHIP prevalence was 34.2%, with myeloid-CHIP (M-CHIP) at 16.2% and lymphoid-CHIP (L-CHIP) at 21.4%. M-CHIP prevalence correlated positively with age (p=0.02), smoking history (p=0.01), and lower platelet counts (p=0.03). The most frequent M-CHIP mutations were in DNMT3A, TET2, PPM1D, while L-CHIP mutations were in EEF1A1, DDX11 and KMT2D. Notably, DDX11 mutations were associated with lower Montreal Cognitive Assessment scores (p=6.57e-03). Comparison with unexposed controls demonstrated higher CHIP prevalence in WTC responders, particularly in those 55 or younger. Study highlights the potential utility of deep sequencing for CHIP detection with clinical, laboratory and exposure data to develop personalized risk-adapted screening programs for cancer and other CHIP-related conditions in individuals exposed to airborne carcinogens.

Background After surviving Coronavirus Disease 2019 (COVID-19), some people develop symptoms known as post- acute sequelae of COVID-19 (PASC). PASC is an emerging phenomenon yet to be fully understood, and identifying risk factors has been challenging. This study investigated the association between the number of COVID-19 episodes and the incidence of PASC among essential workers. Methods We analyzed data from 2511 essential workers, mainly first responders, with confirmed polymerase chain reaction, antibody, or antigen-positive test results for SARS-CoV-2 infection from March 2020 to February 2024. Data were collected through in-person questionnaires and surveys sent via text and email, internal medical records, follow-up calls, and external medical records. Participants who reported continuation or the development of new symptoms three months after the initial SARS-CoV-2 infection, with symptoms lasting for at least two months, were categorized as having PASC, while those without any COVID-19 or whose symptoms resolved were classified as non-PASC. PASC was common in this cohort so we used a Poisson regression model to compute multivariable-adjusted Relative Risk (RR) for the association between risk of PASC and SARS-CoV-2 re-infection, severity, and vaccination status at first infection. Findings A total of 475 (prevalence = 18.9%, [95% confidence interval] = [17.4-20.5]) PASC patients were identified. The mean (standard deviation (SD)) age of participants who experienced PASC (54.8 (7.2) years) was similar to those who did not (54.2 (7.4) years). There were 403 (16.1% [14.6-17.5]) participants who experienced multiple instances of COVID-19. After adjusting for relevant demographic, lifestyle, and clinical variables, we found a significant association between the risk of experiencing PASC and multiple SARS-COV-2 infections (RR = 1.41 [1.14-1.74]), severe COVID-19 (RR = 3.17 [2.41-4.16]), and being unvaccinated at first infection (RR = 3.29 [2.46-4.41]). Interpretation Although the pathogenetic mechanism for PASC remains unclear, identifying risk factors such as lack of vaccination or re-infection can assist in better understanding and managing the condition. Copyright (c) 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Posttraumatic stress disorder (PTSD) is persistent over time, thus identifying risk factors for chronic PTSD is crucial for clinical research. Trauma exposure severity and polygenic liability are two established predictors of PTSD onset and severity, but their contributions to the long-term course of PTSD remain largely unknown. In this prospective longitudinal cohort study, we tested whether severity of trauma exposure and polygenic risk for symptoms of PTSD independently predict long-term trajectories of PTSD symptoms. Data included 49,402 observations, spanning July 2002 to December 2022, from n = 5687 World Trade Center responders who had predominately European ancestry (baseline mean age = 37.74, SD = 8.19, range = 16–75; 92.89

BACKGROUND:People participating in the rescue, recovery, and clean-up effort after the September 11, 2001 attack on the World Trade Center (WTC) were exposed to a complex mix of noxious substances and subsequently experienced elevated gastroesophageal reflux disease (GERD) incidence, the second-most-common WTC-related condition. METHODS:Longitudinal WTC Health Program data, collected between July 2002 and December 2022, were used to describe the sample characteristics, diagnostic procedures, and treatment of consenting cohort members with self-reported GERD who reported incident GERD for a year or longer (n = 19,067). Cross-tabulations and binomial logistic regression, adjusted for confounders including comorbidities, assessed the associations with intermittent and resolved, compared with unresolved, GERD. RESULTS:12.6% of the study cohort reported intermittent GERD; 5.5% reported GERD resolution. Analyses indicated that most GERD resolution was reported by people of color and those with body mass index <25, and by cohort members who had longer postdiagnosis follow-up and implemented dietary modifications together with proton pump inhibitors or Program-approved antacids. GERD-certified members who underwent endoscopy, used medications without dietary modifications, or used bed head-elevation, and those with Barrett's disease (5.8%) or esophageal cancer (0.1%) may have had more severe GERD and reported little resolution. CONCLUSIONS:The use of GERD services was consistent with clinical guidelines. Members' implementing dietary modifications in conjunction with proton pump inhibitors or Program-approved antacids reported more resolution and may have had less severe GERD. Earlier diagnosis and intervention might increase earlier therapeutic resolution.

OBJECTIVE:In analyses without adjustment for World Trade Center (WTC) noise exposure, people with WTC and neighborhood exposures from the September 11, 2001, attacks have experienced slightly elevated risks of hearing loss. We investigated incident hearing loss in the WTC General Responder Cohort by their levels of WTC exposure, their WTC noise exposure and previous occupation. METHODS:Adjusted multivariable log binomial regression models assessed persistent (≥10 months) hearing loss associated with WTC exposures using 22 years of the monitoring visit data ( n = 45,537). RESULTS:Compared with the lowest exposure level without WTC noise exposure, WTC noise exposure increased hearing loss risk (adjusted relative risk range: 1.19 [95% confidence interval 1.08, 1.30] to 1.58 [1.43, 1.76]). CONCLUSIONS:The results clarify the importance of WTC noise when evaluating the associations of WTC exposures on hearing loss.

Objective: Posttraumatic stress disorder (PTSD) is defined by the assumption that qualifying traumatic events lead to a syndrome distinct from other internalizing disorders, while stressful life events play a prominent role in etiologic theories of major depressive disorder (MDD). We examined whether the environmental etiology of PTSD and MDD are distinct by evaluating the relative contributions of traumatic and stressful life events to both conditions. Harmful alcohol use and physical limitations served as noninternalizing comparators expected to show weaker associations with environmental factors. Methods: Longitudinal cohort study of World Trade Center disaster responders who completed annual assessments of mental health and physical functioning from July 1, 2002, to December 31, 2020. Psychiatric diagnoses were ascertained in clinical interviews. Multivariate regression and multilevel modeling quantified the percentage of variance in psychopathology and physical limitations attributable to trauma versus life stress. Results: 11,153 responders (mean age on September 11, 2001: 37.5 years; 91% male) completed 61,244 visits. The combined environmental effect of 9/11-trauma and life stress on PTSD and MDD was nearly identical (14.3% and 14.8% of between person variability), but much weaker for alcohol use and physical limitations (0.8% and 9.1%). Life stress explained the most variance in all diagnoses and symptoms across longitudinal and cross-sectional analytic strategies. Conclusions: In the longest study to date coexamining the environmental etiology of PTSD and MDD, trauma and life stress contributed to both conditions. Considering a spectrum of exposures from stressful life events to trauma and integrating knowledge across internalizing conditions may advance understanding and treatment of stress related psychopathology.

Facial expressions are an essential component of emotions that may reveal mechanisms maintaining posttraumatic stress disorder (PTSD). However, most research on emotions in PTSD has relied on self-reports, which only capture subjective affect. The few studies on outward emotion expressions have been hampered by methodological limitations, including low ecological validity and failure to capture the dynamic nature of emotions and symptoms. Our study addresses these limitations with an approach that has not been applied to psychopathology: person-specific models of day-to-day facial emotion expression and PTSD symptom dynamics. We studied a sample of World Trade Center responders (N=112) with elevated PTSD pathology who recorded a daily video diary and self-reported symptoms for 90 days (8,953 videos altogether). Facial expressions were detected from video recordings with a facial emotion recognition model. In data-driven, idiographic network models, most participants (80%) had at least one, reliable expression-symptom link. Six expression-symptom dynamics were significant for >10% of the sample. Each of these dynamics had statistically meaningful heterogeneity, with some people’s symptoms related to over-expressivity and others to under-expressivity. Our results provide the foundation for a more complete understanding of emotions in PTSD that not only includes subjective feelings but also outward emotion expressions.

Post-traumatic stress disorder (PTSD) is a debilitating condition with serious implications for mental and physical health. While previous studies have documented PTSD prevalence and short-term trajectories, detailed examination of long-term symptom patterns remains limited, especially in large cohorts. This study investigates 20-year PTSD symptom trajectories in World Trade Center (WTC) responders, constituting one of the largest and longest studies of PTSD to date. We aimed to answer six questions: (1) How stable are symptoms? (2) What is the average trajectory? (3) What are atypical trajectories? (4) How quickly do symptoms improve and worsen? (5) Does symptom course differ across demographics or (6) predict functional impairments or mental health care utilization? Data include 81,298 observations from 12,822 responders. Symptoms were stable in the short term but changed significantly over two decades, peaking over a decade after exposure and declining modestly thereafter. Demographic differences in average trajectories were small. Median time before symptoms improved was 8 to 10 years for PTSD cases (median = 8.88, 95% CI = 8.01, 9.79). Most experienced improvement after a decade, but approximately 10% reported elevated symptoms two decades after trauma. Changes in symptoms predicted higher functional impairments and mental health care utilization. Our findings highlight the enduring impact of PTSD among WTC responders, with substantial variability in individual trajectories. Despite overall modest declines, a subset remained highly symptomatic, underscoring the need for continued treatment. These results emphasize the importance of long-term monitoring and highlight the need for tailored treatment strategies for trauma-exposed populations.

Background: Posttraumatic stress disorder (PTSD) is often chronic and impairing. Mechanisms that maintain symptoms remain poorly understood due to heterogenous presentation. We parsed this heterogeneity by examining how individual differences in stress-symptom dynamics relate to the long-term maintenance of PTSD.Methods: We studied 7,308 trauma-exposed World Trade Center responders who self-reported PTSD symptoms and stressful life events at annual monitoring visits for up to 20 years (average=8.8 visits; [range=4–16]). We used multilevel structural equation models to separate the stable and time-varying components of symptoms and stressors. At the within-person level, we modeled stress reactivity by cross-lagged associations between stress and future symptoms, stress generation by cross-lagged associations between symptoms and future stress, and autoregressive effects represented symptom persistence and stress persistence. Clinical utility of the stress-symptom dynamics was evaluated by associations with PTSD chronicity and mental health care use.Results: Stress reactivity, stress generation, and symptom persistence were significant on average (bs=.03–.16). There were significant individual differences in the strength of each dynamic (interquartile ranges=.06–.12). Correlations among within-person processes showed some dynamics are intertwined (e.g., more reactive people also generate stress in a vicious cycle) and others represent distinct phenotypes (e.g., people are reactive or have persistent symptoms). Initial trauma severity amplified some dynamics. People in the top deciles of most dynamics had clinically significant symptom levels across the monitoring period and their health care cost 6-17x more per year than people at median levels. Conclusions: Individual differences in stress-symptom dynamics contribute to the chronicity and clinical burden of PTSD.

The test-negative design (TND) has been widely used to assess postmarketing COVID-19 vaccine effectiveness but requires further evaluation for this application. To determine whether the TND reliably evaluates vaccine effectiveness against symptomatic COVID-19 using placebo-controlled vaccine efficacy randomized clinical trials (RCTs). This secondary cross-protocol analysis constructed TND study datasets from study sites in 16 countries across 5 continents using the blinded phase cohorts of 5 harmonized phase 3 COVID-19 Prevention Network RCTs: COVE (Coronavirus Vaccine Efficacy and Safety), AZD1222, ENSEMBLE, PREVENT-19 (Prefusion Protein Subunit Vaccine Efficacy Novavax Trial COVID-19), and VAT00008. Participants included adults who received the intended number of doses, experienced COVID-19–like symptoms, and obtained SARS-CoV-2 testing. Start dates ranged from July 27, 2020, to October 19, 2021; data cutoff dates ranged from March 26, 2021, to March 15, 2022. Statistical analysis was performed from May 11, 2023, to February 25, 2025. Participants received vaccines consisting of messenger RNA-1273 (COVE; 2 doses 28 days apart), ChAdOx1 nCoV-19 (AZD1222; 2 doses 28 days apart), Ad26.COV2.S (ENSEMBLE; 1 dose), NVX-CoV2373 (PREVENT-19; 2 doses 21 days apart), CoV2 preS dTM-AS03 (VAT00008; D614) (2 doses 21 days apart), or CoV2 preS dTM-AS03 (D614 plus B.1.351) (VAT00008; 2 doses 21 days apart) or placebo. Main outcomes were symptomatic COVID-19 according to each trial’s primary efficacy definition and the Centers for Disease Control and Prevention definition. Vaccine effectiveness was estimated using targeted maximum likelihood estimation under a semiparametric logistic regression model and ordinary logistic regression. Noncase exchangeability, a core TND assumption for unbiased estimation, was also assessed by estimating vaccine efficacy against non–COVID-19 illness. Among the 12 157 participants included in the analysis, mean (SD) age was 45 (15) years, 6414 were female (53%), 5858 were vaccinated (48%), 2835 experienced primary COVID-19 (23%), and 2992 experienced Centers for Disease Control and Prevention–defined COVID-19 (25%). TND vaccine effectiveness estimates were concordant with RCT vaccine efficacy estimates (concordance correlation coefficient, 0.86 [95% CI, 0.58-0.96] for both outcomes). The semiparametric method had 48% smaller variance estimates than ordinary logistic regression. Noncase exchangeability was generally supported with a median vaccine efficacy against non–COVID-19 illness of 7.7% (IQR, 2.7%-16.8%) across trial cohorts and most 95% CIs including 0. In this cross-protocol analysis, the TND provided reliable inferences on COVID-19 vaccine effectiveness in health care–seeking populations for multiple vaccines and symptom definitions when confounding and selection bias were absent. A machine-learning approach for robust confounding control in postmarketing TND studies was also introduced.

Importance:Responders involved in rescue and recovery operations after the collapse of the World Trade Center (WTC) on September 11, 2001, were exposed to airborne carcinogens. Objectives:To examine the incidence of lung cancer after the WTC attacks and to compare the incidence of lung cancer among responders with varying degrees of exposure severity. Design, Setting, and Participants:In this prospective cohort study, data were collected between July 1, 2012, and December 31, 2023, from individuals who were enrolled in a medical monitoring program available to WTC responders residing on Long Island, New York. This study was restricted to people who survived and were followed up for incident lung cancer after a 10-year latency period. Exposures:Types and durations of exposures were based on responses to a detailed questionnaire about on-site work conditions, which included information about the type and duration of work, smells, and sights while working; exposure to dust; and the use of protective equipment. World Trade Center exposure characteristics and overall severity were measured as mild, moderate, and severe exposure using a validated approach. Main Outcomes and Measures:The incidence of lung cancer was the primary outcome. Diagnosis of lung cancer was ascertained following a standardized approach by trained clinicians, and diagnoses were verified by clinicians at the Centers for Disease Control and Prevention. Cox proportional hazards regression was used to estimate multivariable-adjusted hazard ratios. Result:Among 12 334 eligible responders (mean [SD] age at study inclusion, 49.3 [10.2] years; 11 213 men [90.9%]), 118 incident lung cancers were identified between July 1, 2012, and December 31, 2023 (incidence rate, 8.7/10 000 person-years [95% CI, 7.3-10.5 person-years]). When compared with mild exposures, the incidence of lung cancer was higher among moderately (adjusted hazard ratio [AHR], 1.86 [95% CI, 1.19-2.91]; P = .007) and severely (AHR, 2.90 [95% CI, 1.69-4.99]; P < .001) exposed groups. Specific WTC exposures, including smelling fumes (AHR, 1.05 [95% CI, 1.01-1.09]; P = .007) or sewage (AHR, 1.03 [95% CI, 1.01-1.05]; P = .004), were also associated with higher incidence of lung cancer after adjusting for demographics and measures of tobacco use. Conclusions and Relevance:In this cohort study of WTC responders, the incidence of lung cancer was higher among those with greater exposure severity. Future studies may investigate specific WTC exposures and histologic changes and clarify the role of WTC exposure for prognosis.

Objective This study explores the relationship between World Trade Center (WTC) response activities (WRAs) and cognitive impairment (CI) and uses a moderated-mediation model to examine the role of wearing a surgical/nuisance dust mask. Methods This study includes 3285 WTC responders. Responders were placed into eight WRA groups based on self-report structured responses and free-text descriptions of activities at the WTC. The presence/absence of surgical/nuisance dust mask usage was self-reported. The outcome was CI as determined using a Montreal Cognitive Assessment score < 23. Robust Poisson regression was used to examine the main effect, and counterfactual moderated-mediation analysis was used to determine the role of mask usage. Results The risk of CI was higher across most WRAs when compared to supervision. Mask usage was reported by 63 % of responders and varied across WRAs and was associated with a reduced risk of CI (adjusted risk ratio [aRR]=0.77, p = 0.008) after controlling for WRAs. Moderation effects indicated that responders are more likely to wear masks when encountering more dangerous exposures, even within the same WRA group. Responders in the WRA-enclosed group had a lower risk of CI through a moderated intermediary effect of mask usage (aRR=0.92, p = 0.05). Conclusion Surgical/nuisance dust mask usage provided mild protection against air pollution exposures during WTC response activities when compared to not wearing a mask. Results suggest that response workers at disaster sites might benefit from wearing surgical/nuisance dust masks when respirators are unavailable even when the air seems safe.

Importance:The test-negative design (TND) has been widely used to assess postmarketing COVID-19 vaccine effectiveness but requires further evaluation for this application. Objective:To determine whether the TND reliably evaluates vaccine effectiveness against symptomatic COVID-19 using placebo-controlled vaccine efficacy randomized clinical trials (RCTs). Design, Setting, and Participants:This secondary cross-protocol analysis constructed TND study datasets from study sites in 16 countries across 5 continents using the blinded phase cohorts of 5 harmonized phase 3 COVID-19 Prevention Network RCTs: COVE (Coronavirus Vaccine Efficacy and Safety), AZD1222, ENSEMBLE, PREVENT-19 (Prefusion Protein Subunit Vaccine Efficacy Novavax Trial COVID-19), and VAT00008. Participants included adults who received the intended number of doses, experienced COVID-19-like symptoms, and obtained SARS-CoV-2 testing. Start dates ranged from July 27, 2020, to October 19, 2021; data cutoff dates ranged from March 26, 2021, to March 15, 2022. Statistical analysis was performed from May 11, 2023, to February 25, 2025. Interventions:Participants received vaccines consisting of messenger RNA-1273 (COVE; 2 doses 28 days apart), ChAdOx1 nCoV-19 (AZD1222; 2 doses 28 days apart), Ad26.COV2.S (ENSEMBLE; 1 dose), NVX-CoV2373 (PREVENT-19; 2 doses 21 days apart), CoV2 preS dTM-AS03 (VAT00008; D614) (2 doses 21 days apart), or CoV2 preS dTM-AS03 (D614 plus B.1.351) (VAT00008; 2 doses 21 days apart) or placebo. Main Outcomes and Measures:Main outcomes were symptomatic COVID-19 according to each trial's primary efficacy definition and the Centers for Disease Control and Prevention definition. Vaccine effectiveness was estimated using targeted maximum likelihood estimation under a semiparametric logistic regression model and ordinary logistic regression. Noncase exchangeability, a core TND assumption for unbiased estimation, was also assessed by estimating vaccine efficacy against non-COVID-19 illness. Results:Among the 12 157 participants included in the analysis, mean (SD) age was 45 (15) years, 6414 were female (53%), 5858 were vaccinated (48%), 2835 experienced primary COVID-19 (23%), and 2992 experienced Centers for Disease Control and Prevention-defined COVID-19 (25%). TND vaccine effectiveness estimates were concordant with RCT vaccine efficacy estimates (concordance correlation coefficient, 0.86 [95% CI, 0.58-0.96] for both outcomes). The semiparametric method had 48% smaller variance estimates than ordinary logistic regression. Noncase exchangeability was generally supported with a median vaccine efficacy against non-COVID-19 illness of 7.7% (IQR, 2.7%-16.8%) across trial cohorts and most 95% CIs including 0. Conclusions and Relevance:In this cross-protocol analysis, the TND provided reliable inferences on COVID-19 vaccine effectiveness in health care-seeking populations for multiple vaccines and symptom definitions when confounding and selection bias were absent. A machine-learning approach for robust confounding control in postmarketing TND studies was also introduced.

Background: Emerging evidence suggests that better cognition is associated with a lower risk of chronic kidney disease (CKD). However, whether early-onset cognitive impairment (CI) at baseline is linked to rapid estimated glomerular filtration rate (eGFR) decline or incident CKD remains unclear. Methods: We conducted a prospective cohort study of 5,761 World Trade Center (WTC) responders (mean age: 53.8 ± 7.9 years) without CKD at baseline, followed for a mean of 4.2 ± 1.9 years. CI was defined as a Montreal Cognitive Assessment (MoCA) score ≤23, with a subgroup analysis for baseline dementia (MoCA ≤18). Primary outcomes included annual eGFR change and rapid eGFR decline (< -5 mL/min/1.73 m 2 per year). The secondary outcome was incident CKD (eGFR <60 mL/min/1.73 m 2 or diagnosis code). Multivariable Cox proportional hazards models and linear regressions were used for binary and continuous outcomes, respectively. Sensitivity analyses included looking at the effect of baseline mild cognitive impairment (MCI) (MoCA score 19-23), propensity matching for demographics, baseline age <60 years, removal of baseline post-traumatic stress disorder (PTSD)/ depression or baseline head trauma/stroke/cardiovascular disease and after exclusion of those who died during follow-up. Results: At baseline, 1,446 (25%) individuals had CI, while 89 (2%) had dementia. The mean baseline eGFR was 91.1 mL/min/1.73 m 2 , with an overall decline of –1.2 mL/min/1.73 m 2 per year. Rapid eGFR decline occurred in 550 (10%) individuals. After adjusting for age, sex, race/ethnicity, comorbidities, WTC exposure, screened PTSD, and baseline eGFR, CI and dementia were significantly associated with rapid eGFR decline (adjusted hazard ratio [aHR]: 1.63 and 2.42, respectively; both p < 0.001) and faster annual eGFR decline. Findings were consistent across all sensitivity analyses. Additionally, 248 (4%) individuals developed incident CKD. Both baseline CI (aHR: 1.72, p < 0.001) and dementia (aHR: 2.77, p = 0.010) were significantly associated with incident CKD. Conclusions: Among middle-aged individuals without CKD, early-onset cognitive impairment was independently associated with rapid eGFR decline and incident CKD. These findings warrant validation in other cohorts.