Subdistribution Hazard Model Research Articles

Joint modeling of longitudinal and competing-risk data using cumulative incidence functions for the failure submodels accounting for potential failure cause misclassification through double sampling.

Most of the literature on joint modeling of longitudinal and competing-risk data is based on cause-specific hazards, although modeling of the cumulative incidence function (CIF) is an easier and more direct approach to evaluate the prognosis of an event. We propose a flexible class of shared parameter models to jointly model a normally distributed marker over time and multiple causes of failure using CIFs for the survival submodels, with CIFs depending on the "true" marker value over time (i.e., removing the measurement error). The generalized odds rate transformation is applied, thus a proportional subdistribution hazards model is a special case. The requirement that the all-cause CIF should be bounded by 1 is formally considered. The proposed models are extended to account for potential failure cause misclassification, where the true failure causes are available in a small random sample of individuals. We also provide a multistate representation of the whole population by defining mutually exclusive states based on the marker values and the competing risks. Based solely on the assumed joint model, we derive fully Bayesian posterior samples for state occupation and transition probabilities. The proposed approach is evaluated in a simulation study and, as an illustration, it is fitted to real data from people with HIV.

Post-radiation primary hypothyroidism in patients with head and neck cancer: External validation of thyroid gland dose–volume constraints with long-term endocrine outcomes

BackgroundPost-radiation primary hypothyroidism is a common late complication in head and neck cancer (HNC) survivors. No radiation dose-volume constraint of the thyroid gland has been externally validated for predicting long-term thyroid function outcomes. Materials and MethodsThis external validation study evaluated the diagnostic properties of 22 radiation dose–volume constraints of the thyroid gland proposed in the literature. Radiation dosimetric data from 488 HNC patients who underwent neck irradiation from January 2013 to December 2015 at two tertiary oncology centers were reviewed. The diagnostic metrics of candidate constraints were computed by inverse probability of censoring weighting and compared using time-dependent receiver operating characteristic (ROC) curves with death designated as a competing event. Multivariable regression analyses were performed using the Fine–Gray sub-distribution hazard model. ResultsOver a median follow-up period of 6.8 years, 205 (42.0 %) patients developed post-radiation primary hypothyroidism. The thyroid volume spared from 60 Gy (VS60) had the largest area under ROC curve of 0.698 at 5 years after radiotherapy. Of all evaluated constraints, VS60 at a cutoff value of 10 cc had the highest F-score of 0.53. The 5-year hypothyroidism risks of patients with thyroid VS60 ≥ 10 cc and < 10 cc were 14.7 % and 38.2 %, respectively (p < 0.001). The adjusted sub-hazard ratio for post-radiation primary hypothyroidism for VS60 < 10 cc was 1.87 (95 % confidence interval, 1.22–2.87; p < 0.001). ConclusionThyroid VS60 is the best radiation dose–volume parameter to predict the long-term risk of primary hypothyroidism in patients with HNC who underwent neck irradiation. VS60 ≥ 10 cc is a robust constraint that limits the 5-year primary hypothyroidism risk to less than 15 % and should be routinely employed during radiotherapy optimization.

Diabetic Patients With Rosacea Increase the Risks of Diabetic Macular Edema, Dry Eye Disease, Glaucoma, and Cataract.

Inflammation plays a role in diabetic eye diseases, but the association between rosacea and eye diseases in patients with diabetes remains unknown. This retrospective cohort study used claims data from the National Health Insurance Research Database in Taiwan to investigate the association between rosacea and eye diseases in patients with diabetes. Taiwanese patients diagnosed as having diabetes mellitus between January 1, 1997, and December 31, 2013, and using any hypoglycemic agents were included and divided into rosacea and nonrosacea groups. After applying 1:20 sex and age matching and exclusion criteria, 1:4 propensity score matching (PSM) was conducted to balance the covariate distribution between the groups. The risk of time-to-event outcome between rosacea and nonrosacea groups in the PSM cohort was compared using the Fine and Gray subdistribution hazard model. A total of 4096 patients with rosacea and 16,384 patients without rosacea were included in the analysis. During a mean follow-up period of 5 years, diabetic patients with rosacea had significantly higher risks of diabetic macular edema [subdistribution hazard ratio (SHR): 1.31, 95% confidence interval (CI): 1.05-1.63], glaucoma with medical treatment (SHR: 1.11, 95% CI: 1.01-1.21), dry eye disease (SHR: 1.55, 95% CI: 1.38-1.75), and cataract surgery (SHR: 1.13, 95% CI: 1.02-1.25) compared with diabetic patients without rosacea. A cumulative incidence analysis performed up to 14 years after the index date revealed that the risks of developing ocular diseases consistently increased over time. No significant differences in diabetic retinopathy, age-related macular degeneration, retinal vascular occlusion, ischemic optic neuropathy, optic neuritis, uveitis, or retinal detachment were identified according to rosacea diagnosis. However, we observed significant associations between rosacea and psoriasis, irritable bowel syndrome, anxiety, and major depressive disorder among patients with diabetes. Rosacea is associated with diabetic macular edema, glaucoma, dry eye disease, and cataract development in diabetic patients, as well as increased risks of psoriasis, irritable bowel syndrome, anxiety, and depression in diabetic patients.

Tumor Specific Growth Rate as a Predictor of Outcomes in Oligo-Progressive Disease Treated with Stereotactic Body Radiotherapy

<h3>Purpose/Objective(s)</h3> Histologic classification alone does not fully represent the distribution of tumor behaviors due to within histology heterogeneity. The aim of this study is to assess the variation and the impact of pre- and post- tumor specific growth rate (SGR) on outcomes of metastatic patients with oligo-progressive (OP) disease treated with SBRT. <h3>Materials/Methods</h3> Patients with known metastatic disease and ≤3 radiologically OP were enrolled in a prospective phase II study. For the purpose of this analysis, treated metastases were retrospectively contoured on pre-SBRT diagnostic images (GTV1 - within 3-months), at time of SBRT (GTV2) and post-SBRT (GTV3 -within 3-months). SGR was calculated using the time interval (t=days) and volume changes (GTVx/GTVy) for each metastasis (<i>ln</i>(GTV2/GTV1)/t). Pre-SBRT growth (SGR1) was defined as <i>GTV2 – GTV1</i>, and post-SBRT (SGR2) as <i>GTV3 – GTV2</i>. High SGR was defined as greater than the median SGR value for each population. Progression was defined as local failure (LF) or distant progression. The impact of SGR1 and SGR2 (as continuous variables) on LF and progression was evaluated using a subdistribution hazards model to account for the competing risk of death. Cox Proportional Hazard was used to assess the association of SGR1 and SGR2 with overall survival (OS). For patients with multiple lesions, the highest SGR was used for progression and OS analyses. <h3>Results</h3> Thirty-five patients with 55 metastases from breast (29%), gastro-intestinal (GI) (40%) and genito-urinary (GU) (31%) cancers were analyzed. Median follow-up was 11.74 (interquartile range (IQR): 8.05-15.65 months). The median volume of GTV1, GTV2 and GTV3 was 3.91 (IQR 0.94 - 6.71)cc, 7.68 (IQR 2.04-15.43)cc and 2.72 (IQR 0.8-8.08)cc respectively. Median SGR1 and SGR2 was 0.007 (IQR 0.004 - 0.013) and -0.009 (IQR -0.01 to -0.002), with median SGR1/SGR2 per histology 0.005 (IQR 0.003 - 0.007)/-0.010 (IQR -0.019 to -0.006) breast, 0.011 (IQR 0.007 - 0.014)/-0.010 (IQR-0.013 to -0.002) GI and 0.006 (IQR 0.001 - 0.015)/-0.009 (IQR-0.012 to -0.001) GU. There was no statistically significant difference seen between histologies and SGR1 (p 0.15) and SGR2 (p 0.98). 48% of cases had a high SGR1 and 50% a high SGR2. At 12 months the LF and distant progression was11% and 76.8% respectively, while the probability of OS was 59%. There was a trend significant association between SGR1 and the probability of progression (p 0.055) while no significant association was found with the LF (p 0.28) or OS (p=0.35). Patients with low vs. high SGR1 had a higher probability of being alive but the difference did not reach statistical significance [SGR1 71% vs. 47% (p 0.35)]. SGR2 did not have a significant impact LF (p=0.59) or OS (p 0.39). <h3>Conclusion</h3> Lower SGR measurements seems to have better response after SBRT and may be used in the future for patient selection. These findings require validation in a larger cohort.

The Toxicity of Low-Dose Whole Abdomen Radiotherapy in Advanced Gynecologic Malignancies Receiving Stereotactic Body Radiotherapy and Immunotherapy

<h3>Purpose/Objective(s)</h3> The objective of this retrospective study was to evaluate the safety of low-dose whole abdomen radiotherapy (LD-WART) in patients with gynecologic malignancies receiving stereotactic body radiotherapy (SBRT) and immunotherapy (IO). We hypothesized that LD-WART volume would not be associated with increased adverse events (AEs). <h3>Materials/Methods</h3> Study design was a pooled secondary analysis of two prospective phase I trials, in which patients with advanced solid malignancies were treated with multi-site SBRT (≤4 sites) with IO. The study cohort (N = 22) was limited to patients with a gynecologic primary (ovarian or uterine) that received SBRT to ≥1 abdomino-pelvic site. IO was administered either concurrently (N = 8) or sequentially (i.e., within 7 days of SBRT, N = 14). IO regimens included pembrolizumab (N = 14), nivolumab plus urelumab (N = 6), and nivolumab plus cabiralizumab (N = 2). Whole abdominal cavities (with inclusion of the pelvis) were contoured for all patients. The kidneys and a 1cm isocentric contraction of the liver were excluded. Based on recent literature supporting immune reprogramming of the tumor microenvironment with doses of 0.5 to 2 Gy, the V0.5Gy (%), V1Gy (%), and V2Gy (%) for the whole abdomen were recorded. Time to CTCAE v4.0 grade 2-4 AEs and overall survival (OS) were defined as the time from start of SBRT to either toxicity or death, respectively. Toxicity was analyzed using the cumulative incidence function with death as a competing risk. Correlation between dose metrics and toxicity were analyzed using the Fine-Gray subdistribution hazard model, including the primary site and number of irradiated abdomino-pelvic sites as covariates. <h3>Results</h3> 22 patients met criteria for inclusion (73% ovarian, 27% endometrial). Median OS was 13.3 months. The median whole abdomen V0.5Gy, V1Gy, and V2Gy was 42% (range, 12% - 97.5%), 32% (range, 8% - 86%), and 26% (range, 6% - 77%), respectively. At 12 months (with 12 of 22 patients evaluable), the cumulative incidence of grade 3+ AEs was 9%, 9%, and 0% when whole abdomen V0.5Gy, V1Gy, and V2Gy exceeded the cohort median, in contrast to 18%, 18%, and 27%, respectively, when less than the cohort median. As a continuous variable in the multivariable Fine-Gray model, grade 3+ AEs were not increased with increasing whole abdomen V0.5Gy (subdistribution hazard ratio [SHR] 0.96, <i>P</i> = 0.26), V1Gy (SHR 0.95, <i>P</i> = 0.11), and V2Gy (SHR 0.95, <i>P</i> = 0.14). Similarly, the grade 2+ AEs were not increased with increasing whole abdomen V0.5Gy (SHR 0.98, <i>P</i> = 0.12), V1Gy (SHR 0.97, <i>P</i> = 0.08), and V2Gy (SHR 0.98, <i>P</i> = 0.17). There were no grade 2+ hepatic or renal AEs. <h3>Conclusion</h3> In this pooled analysis of two phase I trials, increasing LD-WART volume receiving 0.5-2 Gy was not associated with increased toxicity in the setting of SBRT and immunotherapy. These data support the potential safety of a future pilot study being explored by our institution for advanced gynecologic malignancies combining low-dose whole abdominal irradiation, SBRT to gross tumor sites, and IO.

Long-Term Thyroid Function Outcome after Head and Neck Radiotherapy: External Validation and Comparisons of Thyroid Gland Dose-Volume Constraints

<h3>Purpose/Objective(s)</h3> Primary hypothyroidism is a common late complication in patients who underwent head and neck radiotherapy, occurring at incidences as high as 20-50% among long-term survivors. Various radiation dose-volume constraints for the thyroid gland have been proposed in retrospective series, but none has been externally validated for predicting long-term thyroid function outcome. <h3>Materials/Methods</h3> We undertook a systematic review of the literature and identified 22 distinct radiation dose-volume constraints of the thyroid gland for head and neck radiotherapy planning. We reviewed the thyroid gland dosimetry of 488 consecutive head and neck cancer patients who had normal baseline thyroid function and underwent definitive radiotherapy between January 2013 and December 2015 in two tertiary oncology centers. Thyroid function was monitored at least yearly after radiotherapy. Primary hypothyroidism was defined as serum TSH levels above the upper limit of normal. Time to hypothyroidism was estimated using cumulative incidence function with death designated as a competing event. Univariable and multivariable analyses were conducted using Gray's test the Fine-Gray sub-distribution hazard model. Diagnostic performances of the 22 constraints were compared using time-dependent receiver operating characteristic curves. <h3>Results</h3> With a median follow-up of 6.8 years, 205 (42.0%) patients developed post-radiation primary hypothyroidism. The 5-year primary hypothyroidism rate was 30.5%. Among all candidate dose-volume parameters (Dmin, Dmax, Dmean, V25, V30, V30, V35, V40, V45, V50, V30-60, VS45 and VS60), thyroid volume spared from 60Gy (VS60) had the largest areas under curve of 0.715, 0.694 and 0.698, at 1 year, 3 years and 5 years after radiotherapy, respectively. The sensitivity, specificity, positive predictive value and negative predictive value for VS60 ≥10cc on 5-year hypothyroidism risk were 84.4%, 40.3%, 38.5% and 85.4%, respectively. Among all evaluated constraints, VS60 ≥10cc had the highest F-score of 0.59, suggesting a superior overall diagnostic accuracy. The 5-year hypothyroidism risk for patients with thyroid VS60 ≥10cc and VS60 <10cc were 14.7% and 38.2%, respectively (p <0.001). In multivariate regression, young age, VS60 <10cc and small thyroid volume and were independent predictors for post-radiation primary hypothyroidism. The adjusted sub-hazard ratio for post-radiation primary hypothyroidism for VS60 <10cc was 2.00 (95% C.I., 1.33 – 3.07; p <0.001). <h3>Conclusion</h3> Our external validation study identified thyroid VS60 as the best radiation dose-volume parameter to predict long-term thyroid gland function in patients who underwent head and neck radiotherapy. VS60 >10cc is a potentially useful constraint that limits 5-year primary hypothyroidism risk to below 15%. As far as it is feasible during radiotherapy optimization, higher VS60 targets should be considered to further reduce the risk of this late complication among head and neck cancer patients.

Covert Brain Infarction as a Risk Factor for Stroke Recurrence in Patients With Atrial Fibrillation.

We aimed to evaluate covert brain infarction (CBI), frequently encountered during the diagnostic work-up of acute ischemic stroke, as a risk factor for stroke recurrence in patients with atrial fibrillation (AF). For this prospective cohort study, from patients with acute ischemic stroke hospitalized at 14 centers between 2017 and 2019, we enrolled AF patients without history of stroke or transient ischemic attack and divided them into the CBI (+) and CBI (-) groups. The 2 groups were compared regarding the 1-year cumulative incidence of recurrent ischemic stroke and all-cause mortality using the Fine and Gray subdistribution hazard model with nonstroke death as a competing risk and the Cox frailty model, respectively. Each CBI lesion was also categorized into either embolic-appearing (EA) or non-EA pattern CBI. Adjusted hazard ratios and 95% CIs of any CBI, EA pattern CBI only, non-EA pattern CBI only, and both CBIs were estimated. Among 1383 first-ever stroke patients with AF, 578 patients (41.8%) had CBI. Of these 578 with CBI, EA pattern CBI only, non-EA pattern CBI only, and both CBIs were 61.8% (n=357), 21.8% (n=126), and 16.4% (n=95), respectively. The estimated 1-year cumulative incidence of recurrent ischemic stroke was 5.2% and 1.9% in the CBI (+) and CBI (-) groups, respectively (P=0.001 by Gray test). CBI increased the risk of recurrent ischemic stroke (adjusted hazard ratio [95% CI], 2.91 [1.44-5.88]) but did not the risk of all-cause mortality (1.32 [0.97-1.80]). The EA pattern CBI only and both CBIs elevated the risk of recurrent ischemic stroke (2.76 [1.32-5.77] and 5.39 [2.25-12.91], respectively), while the non-EA pattern only did not (1.44 [0.40-5.16]). Our study suggests that AF patients with CBI might have increased risk of recurrent stroke. CBI could be considered when estimating the stroke risk in patients with AF.

Immune thrombocytopenia and risk of stroke: Evidence from a nationwide population-based cohort study.

Research investigating differences in the overall stroke risk between individuals with and without immune thrombocytopenia (ITP) is lacking. This real-world study used the National Health Insurance Research Database (NHIRD). Risk of stroke was compared between 13,085 individuals with ITP enrolled between 1 January 2000 and 31 December 2015 and a control cohort of 52,340 individuals without ITP (1:4 ratio propensity score-matched by age, sex, index year, relevant comorbidities, and medications). Sub-distribution hazards models were used to estimate adjusted sub-distribution hazard ratio (SHR) and 95% confidence intervals (CIs), with the non-ITP group as the control group. Of the 65,425 participants, 13,085 had ITP, 63.3% were women, and the mean age was 52.59 years. The risk of both ischemic and hemorrhagic stroke was 1.14 times (adjusted SHR 1.14, 95% CI, 1.07-1.22) and 1.93 times (adjusted SHR 1.93, 95% CI, 1.70-2.20) higher in the ITP group than in controls. Patients with ITP in the 20- to 29-year subgroup had a higher risk of new-onset stroke (adjusted SHR, 4.06 (95% CI, 2.72-6.07), p value for interaction <0.01) than those aged 20-29 years without ITP. Individuals with severe ITP with splenectomy had a 1.79 times higher overall stroke risk than those without. ITP is associated with increased risk of both ischemic and hemorrhagic stroke.

Thrombosis-Related Loss of Arterial Lines in the First Wave of COVID-19 and Non-COVID-19 Intensive Care Unit Patients.

Patients with coronavirus disease 2019 (COVID-19) can present with severe respiratory distress requiring intensive care unit (ICU)-level care. Such care often requires placement of an arterial line for monitoring of pulmonary disease progression, hemodynamics, and laboratory tests. During the first wave of the COVID-19 pandemic in March 2020, experienced physicians anecdotally reported multiple attempts, decreased insertion durations, and greater need for replacement of arterial lines in patients with COVID-19 due to persistent thrombosis. Because invasive procedures in patients with COVID-19 may increase the risk for caregiver infection, better defining difficulties in maintaining arterial lines in COVID-19 patients is important. We sought to explore the association between COVID-19 infection and arterial line thrombosis in critically ill patients. In this primary exploratory analysis, a multivariable Fine-Gray subdistribution hazard model was used to retrospectively estimate the association between critically ill COVID-19 (versus sepsis/acute respiratory distress syndrome [ARDS]) patients and the risk of arterial line removal for thrombosis (with arterial line removal for any other reason treated as a competing risk). As a sensitivity analysis, we compared the number of arterial line clots per 1000 arterial line days between critically ill COVID-19 and sepsis/ARDS patients using multivariable negative binomial regression. We retrospectively identified 119 patients and 200 arterial line insertions in patients with COVID-19 and 54 patients and 68 arterial line insertions with non-COVID ARDS. Using a Fine-Gray subdistribution hazard model, we found the adjusted subdistribution hazard ratio (95% confidence interval [CI]) for arterial line clot to be 2.18 (1.06-4.46) for arterial lines placed in COVID-19 patients versus non-COVID-19 sepsis/ARDS patients ( P = .034). Patients with COVID-19 had 36.3 arterial line clots per 1000 arterial line days compared to 19.1 arterial line clots per 1000 arterial line days in patients without COVID-19 (adjusted incidence rate ratio [IRR] [95% CI], 1.78 [0.94-3.39]; P = .078). Our study suggests that arterial line complications due to thrombosis are more likely in COVID-19 patients and supports the need for further research on the association between COVID-19 and arterial line dysfunction requiring replacement.

Joint modeling under competing risks: Application to survival prediction in patients admitted in Intensive Care Unit for sepsis with daily Sequential Organ Failure Assessment score assessments.

Joint models of longitudinal process and time-to-event data have recently gained attention, notably to provide individualized dynamic predictions. In the presence of competing risks, models published mostly involve cause-specific hazard functions jointly estimated with a linear or generalized linear model. Here we propose to extend the modeling to full parametric joint estimation of a nonlinear mixed-effects model and a subdistribution hazard model. We apply this approach on 6046 patients admitted in intensive care unit (ICU) for sepsis with daily Sequential Organ Failure Assessment (SOFA) score measurements. The joint model is built on a randomly selected training set of two thirds of patients and links the current predicted SOFA measurement to the instantaneous risks of ICU death and discharge from ICU, both adjusted on the patient age. Stochastic Approximation Expectation Maximization algorithm in Monolix is used for estimation. SOFA evolution is significantly associated with both risks: 0.37, 95% confidence interval (CI)=[0.35, 0.39] for the risk of death and -0.38, 95% CI=[-0.39, -0.36] for the risk of discharge. A simulation study, inspired from the real data, shows the good estimation properties of the parameters. We assess on the validation set the added value of modeling the longitudinal SOFA follow-up for the prediction of death compared with a model that includes only SOFA at baseline. Time-dependent receiver operating characteristic area under the curve and Brier scores show that when enough longitudinal individual information is available, joint modeling provides better predictions. The methodology can easily be applied to other clinical applications because of the general form of the model.

REAL-WORLD OUTCOMES &amp; HEALTH CARE COSTS OF REGION-WIDE COMMUNITY-BASED CARDIAC REHABILITATION

REAL-WORLD OUTCOMES &amp; HEALTH CARE COSTS OF REGION-WIDE COMMUNITY-BASED CARDIAC REHABILITATION

Home, but Not Homebound: A Prospective Analysis of Persons Living With Dementia.

Homebound persons living with dementia may have increased difficulty accessing needed care in the community. This study identifies factors associated with becoming homebound among a national sample of Medicare beneficiaries with newly identified dementia. Prospective cohort analysis. We used the National Health and Aging Trends Study (NHATS) 2011-2018 to identify community-dwelling older adults at the time of a new dementia diagnosis (n= 939). Dementia status was determined based on cognitive testing and self and proxy reporting. We compared characteristics of homebound (ie, those who never or rarely left home) and non-homebound participants at the time of dementia identification. Among non-homebound participants, we used a Fine-Gray subdistribution hazard model to identify factors associated with becoming homebound over follow-up (median follow-up 4years), accounting for competing risks of death and moving to a nursing home. 20% of individuals with newly identified dementia were homebound and this group was more functionally impaired, medically complex, and socioeconomically disadvantaged as compared to the non-homebound. Over time, depression [subhazard ratio (SHR) 2.19, 95% CI 1.36, 3.54], living in an assisted living facility (SHR 2.60, 95% CI 1.35, 4.97), and Hispanic ethnicity (SHR 1.91, 95% CI 1.05, 3.47) were associated with becoming homebound. Most adults are not homebound at the time of dementia diagnosis. Identifying and addressing modifiable factors like depression may slow progression to homebound status and enable persons living with dementia to access needed care in the community. In order to accommodate diverse individual and family preferences for long-term care, robust systems of home-based clinical and long-term care are necessary for those who do become homebound.

Derivation and External Validation of a Clinical Model to Predict Heart Failure Onset in Patients With Incident Diabetes.

Heart failure (HF) often develops in patients with diabetes and is recognized for its role in increased cardiovascular morbidity and mortality in this population. Most existing models predict risk in patients with prevalent rather than incident diabetes and fail to account for sex differences in HF risk factors. We derived sex-specific models in Ontario, Canada to predict HF at diabetes onset and externally validated these models in the U.K. Retrospective cohort study using international population-based data. Our derivation cohort comprised all Ontario residents aged ≥18 years who were diagnosed with diabetes between 2009 and 2018. Our validation cohort comprised U.K. patients aged ≥35 years who were diagnosed with diabetes between 2007 and 2017. Primary outcome was incident HF. Sex-stratified multivariable Fine and Gray subdistribution hazard models were constructed, with death as a competing event. A total of 348,027 Ontarians (45% women) and 54,483 U.K. residents (45% women) were included. At 1, 5, and 9 years, respectively, in the external validation cohort, the C-statistics were 0.81 (95% CI 0.79-0.84), 0.79 (0.77-0.80), and 0.78 (0.76-0.79) for the female-specific model; and 0.78 (0.75-0.80), 0.77 (0.76-0.79), and 0.77 (0.75-0.79) for the male-specific model. The models were well-calibrated. Age, rurality, hypertension duration, hemoglobin, HbA1c, and cardiovascular diseases were common predictors in both sexes. Additionally, mood disorder and alcoholism (heavy drinker) were female-specific predictors, while income and liver disease were male-specific predictors. Our findings highlight the importance of developing sex-specific models and represent an important step toward personalized lifestyle and pharmacologic prevention of future HF development.

Racial Disparities in Rates of Revision and use of Modern Features in Total Knee Arthroplasty, a National Registry Study

BackgroundThe purpose of our study was to investigate the association of race and ethnicity with rates of modern implant use and postoperative outcomes in total knee arthroplasty (TKA) using the American Academy of Orthopaedic Surgeons American Joint Replacement Registry. MethodsAdult TKAs from 2012 to 2020 were queried from the American Joint Replacement Registry. A total of 1,121,457 patients were available for analysis for surgical features and 1,068,210 patients for analysis of outcomes. Mixed-effects multivariable logistic regression models were used to examine the association of race with each individual surgical feature (unicompartmental knee arthroplasty (UKA) and robotic-assisted TKA (RA-TKA)) and 30- and 90-day readmission. A proportional subdistribution hazard model was used to model the risk of revision TKA. ResultsOn multivariate analyses, compared to White patients, Black (odds ratio (OR): 0.52 P < .0001), Hispanic (OR 0.75 P < .001), and Native American (OR: 0.69 P = .0011) patients had lower rates of UKA, while only Black patients had lower rates of RA-TKA (OR = 0.76 P < .001). White (hazard ratio (HR) = 0.8, P < .001), Asian (HR = 0.51, P < .001), and Hispanic-White (HR = 0.73, P = .001) patients had a lower risk of revision TKA than Black patients. Asian patients had a lower revision risk than White (HR = 0.64, P < .001) and Hispanic-White (HR = 0.69, P = .011) patients. No significant differences existed between groups for 30- or 90-day readmissions. ConclusionBlack, Hispanic, and Native American patients had lower rates of UKA compared to White patients, while Black patients had lower rates of RA-TKA compared to White, Asian, and Hispanic patients. Black patients also had higher rates of revision TKA than other races.

Marginal semiparametric transformation models for clustered multivariate competing risks data.

Multivariate survival models are often used in studying multiple outcomes for right-censored data. However, the outcomes of interest often have competing risks, where standard multivariate survival models may lead to invalid inferences. For example, patients who had stem cell transplantation may experience multiple types of infections after transplant while reconstituting their immune system, where death without experiencing infections is a competing risk for infections. Such competing risks data often suffer from cluster effects due to a matched pair design or correlation within study centers. The cumulative incidence function (CIF) is widely used to summarize competing risks outcomes. Thus, it is often of interest to study direct covariate effects on the CIF. Most literature on clustered competing risks data analyses is limited to the univariate proportional subdistribution hazards model with inverse probability censoring weighting which requires correctly specifying the censoring distribution. We propose a marginal semiparametric transformation model for multivariate competing risks outcomes. The proposed model does not require modeling the censoring distribution, accommodates nonproportional subdistribution hazards structure, and provides a platform for joint inference of all causes and outcomes.

Mortality after the 9/11 terrorist attacks among world trade center health registry enrollees with cancer.

While several studies have reported the association between 9/11 exposure and cancer risk, cancer survival has not been well studied in the World Trade Center (WTC) exposed population. We examined associations of 9/11-related exposures with mortality in WTC Health Registry enrollees diagnosed with cancer before and after 9/11/2001. This is a longitudinal cohort study of 5061 enrollees with a first-ever primary invasive cancer diagnosis between 1995 and 2015 and followed through 2016. Based on the timing of first cancer diagnosis, pre-9/11 (n=634) and post-9/11 (n=4427) cancer groups were examined separately. 9/11-related exposures included witnessing traumatic events, injury on 9/11, and 9/11-related post-traumatic stress disorder (PTSD). Associations of exposures with all-cause mortality were examined using Cox proportional hazards regression. In the post-9/11 group, cancer-specific mortality was evaluated by enrollee group (WTC rescue/recovery workers vs. non-workers) using Fine and Gray's proportional sub-distribution hazard models, adjusting for baseline covariates, tumor characteristics, and treatment. In the pre-9/11 group, 9/11-related exposures were not associated with all-cause mortality. In the post-9/11 group, increased risk of all-cause mortality was associated with PTSD (adjusted HR=1.35; 95% CI=1.11-1.65), but not with injury or witnessing traumatic events. Cancer-specific mortality was not statistically significantly associated with 9/11-related exposures. In rescue/recovery workers, increased non-cancer mortality risk was associated with PTSD (aHR=2.13, 95% CI=1.13-4.00) and witnessing ≥3 traumatic events (aHR=2.00, 95% CI=1.13-3.55). We did not observe associations between 9/11-related exposures and cancer-specific mortality. Similar to findings in the non-cancer WTC exposed population, PTSD was associated with increased risk of all-cause mortality in cancer patients.

Clinical and programmatic outcomes of HIV-exposed infants enrolled in care at geographically diverse clinics, 1997-2021: A cohort study.

BackgroundAlthough 1·3 million women with HIV give birth annually, care and outcomes for HIV-exposed infants remain incompletely understood. We analyzed programmatic and health indicators in a large, multidecade global dataset of linked mother–infant records from clinics and programs associated with the International epidemiology Databases to Evaluate AIDS (IeDEA) consortium.Methods and findingsHIV-exposed infants were eligible for this retrospective cohort analysis if enrolled at <18 months at 198 clinics in 10 countries across 5 IeDEA regions: East Africa (EA), Central Africa (CA), West Africa (WA), Southern Africa (SA), and the Caribbean, Central, and South America network (CCASAnet). We estimated cumulative incidences of DNA PCR testing, loss to follow-up (LTFU), HIV diagnosis, and death through 24 months of age using proportional subdistribution hazard models accounting for competing risks. Competing risks were transfer, care withdrawal, and confirmation of negative HIV status, along with LTFU and death, when not the outcome of interest. In CA and EA, we quantified associations between maternal/infant characteristics and each outcome. A total of 82,067 infants (47,300 EA, 10,699 CA, 6,503 WA, 15,770 SA, 1,795 CCASAnet) born from 1997 to 2021 were included. Maternal antiretroviral therapy (ART) use during pregnancy ranged from 65·6% (CCASAnet) to 89·5% (EA), with improvements in all regions over time. Twenty-four-month cumulative incidences varied widely across regions, ranging from 12·3% (95% confidence limit [CL], 11·2%,13·5%) in WA to 94·8% (95% CL, 94·6%,95·1%) in EA for DNA PCR testing; 56·2% (95% CL, 55·2%,57·1%) in EA to 98·5% (95% CL, 98·3%,98·7%) in WA for LTFU; 1·9% (95% CL, 1·6%,2·3%) in WA to 10·3% (95% CL, 9·7%,10·9%) in EA for HIV diagnosis; and 0·5% (95% CL, 0·2%,1·0%) in CCASAnet to 4·7% (95% CL, 4·4%,5·0%) in EA for death. Although infant retention did not improve, HIV diagnosis and death decreased over time, and in EA, the cumulative incidence of HIV diagnosis decreased substantially, declining to 2·9% (95% CL, 1·5%,5·4%) in 2020. Maternal ART was associated with decreased infant mortality (subdistribution hazard ratio [sdHR], 0·65; 95% CL, 0·47,0·91 in EA, and sdHR, 0·51; 95% CL, 0·36,0·74 in CA) and HIV diagnosis (sdHR, 0·40; 95% CL, 0·31,0·50 in EA, and sdHR, 0·41; 95% CL, 0·31,0·54 in CA). Study limitations include potential misclassification of outcomes in real-world service delivery data and possible nonrepresentativeness of IeDEA sites and the population of HIV-exposed infants they serve.ConclusionsWhile there was marked regional and temporal heterogeneity in clinical and programmatic outcomes, infant LTFU was high across all regions and time periods. Further efforts are needed to keep HIV-exposed infants in care to receive essential services to reduce HIV infection and mortality.

Association between cause of kidney failure and fracture incidence in a national US dialysis population cohort study.

Whether fracture rates, overall and by fracture site, vary by cause of kidney failure in patients receiving dialysis is unknown. Using the US Renal Data System, we compared fracture rates across seven causes of kidney failure in patients who started dialysis between 1997 and 2014. We computed unadjusted and multivariable adjusted proportional sub-distribution hazard models, with fracture events (overall, and by site) as the outcome and immunoglobulin A nephropathy as the reference group. Kidney transplantation and death were competing events. Among 491 496 individuals, with a median follow-up of 2.0 (25%, 75% range 0.9-3.9) years, 62 954 (12.8%) experienced at least one fracture. Patients with diabetic nephropathy, vasculitis or autosomal polycystic kidney disease (ADPKD) had the highest (50, 46 and 40 per 1000 person-years, respectively), and patient with lupus nephritis had the lowest (20 per 1000 person-years) fracture rates. After multivariable adjustment, diabetic nephropathy [hazard ratio (HR) 1.43, 95% confidence interval 1.33-1.53], ADPKD (HR 1.37, 1.26-1.48), vasculitis (HR 1.22, 1.09-1.34), membranous nephropathy (HR 1.16, 1.02-1.30) and focal segmental glomerulosclerosis (FSGS) (HR 1.13, 1.02-1.24) were associated with a significantly higher, and lupus nephritis with a significantly lower (HR 0.85, 0.71-0.98) fracture hazard. The hazards for upper extremity and lower leg fractures were significantly higher in diabetic nephropathy, ADPKD, FSGS and membranous nephropathy, while the hazard for vertebral fracture was significantly higher in vasculitis. Our findings were limited by the lack of data on medication use and whether fractures were traumatic or non-traumatic, among other factors. Fracture risk, overall and by fracture site, varies by cause of end-stage kidney disease. Future work to determine underlying pathogenic mechanisms contributing to differential risks might inform more tailored treatment strategies. Our study was limited by lack of data regarding numerous potential confounders or mediators including medications and measures or bone biomarkers.

Neighborhood disadvantage and lung cancer risk in a national cohort of never smoking Black women

BackgroundCompared to women of other races who have never smoked, Black women have a higher risk of lung cancer. Whether neighborhood disadvantage, which Black women experience at higher rates than other women, is linked to never-smoking lung cancer risk remains unclear. This study investigates the association of neighborhood disadvantage and lung cancer risk in Black never-smoking women. Methods and materialsThis research utilized data from the Black Women’s Health Study, a prospective cohort of 59,000 Black women recruited from across the US in 1995 and followed by biennial questionnaires. Associations of lung cancer incidence with neighborhood-level factors (including two composite variables derived from Census Bureau data: neighborhood socioeconomic status and neighborhood concentrated disadvantage), secondhand smoke exposure, and PM2.5 were estimated using Fine-Gray subdistribution hazard models. ResultsAmong 37,650 never-smokers, 77 were diagnosed with lung cancer during follow-up from 1995 to 2018. The adjusted subdistribution hazard ratio (sHR) of lung cancer incidence with ten unit increase in neighborhood concentrated disadvantage index was 1.30 (95 % CI: 1.04, 1.63, p = 0.023). Exposure to secondhand smoke at work was associated with increased risk (sHR = 1.93, 95 % CI: 1.21, 3.10, p = 0.006), but exposure to secondhand smoke at home and PM2.5 was not. ConclusionWorse neighborhood concentrated disadvantage was associated with increased lung cancer risk in Black women who never smoked. These findings suggest that non-tobacco-related factors in disadvantaged neighborhoods may be linked to lung cancer risk in Black women and that these factors must be understood and targeted to achieve health equity.

Development and validation of a nomogram to predict cancer-specific survival in elderly patients with papillary thyroid carcinoma: a population-based study

ObjectiveThyroid carcinoma (TC) is the most common endocrine tumor in the human body. Papillary thyroid carcinoma (PTC) accounts for more than 80% of thyroid cancers. Accurate prediction of elderly PTC can help reduce the mortality of patients. We aimed to construct a nomogram predicting cancer-specific survival (CSS) in elderly patients with PTC.MethodsPatient information was downloaded from the Surveillance, Epidemiology, and End Results (SEER) program. Univariate and multivariate Cox regression models were used to screen the independent risk factors for patients with PTC. The nomogram of elderly patients with PTC was constructed based on the multivariate Cox regression model. We used the concordance index (C-index), the area under the receiver operating characteristic curve (AUC) and the calibration curve to test the accuracy and discrimination of the prediction model. Decision curve analysis (DCA) was used to test the clinical value of the model.ResultsA total of 14,138 elderly patients with PTC were included in this study. Patients from 2004 to 2015 were randomly divided into a training set (N = 7379) and a validation set (N = 3141), and data from 2016 to 2018 were divided into an external validation set (N = 3618). Proportional sub-distribution hazard model showed that age, sex, tumor size, histological grade, TNM stage, surgery and chemotherapy were independent risk factors for prognosis. In the training set, validation set and external validation set, the C-index was 0.87(95%CI: 0.852–0.888), 0.891(95%CI: 0.866–0.916) and 0.931(95%CI:0.894–0.968), respectively, indicating that the nomogram had good discrimination. Calibration curves and AUC suggest that the prediction model has good discrimination and accuracy.ConclusionsWe constructed a new nomogram to predict CSS in elderly patients with PTC. Internal cross-validation and external validation indicate that the model has good discrimination and accuracy. The predictive model can help doctors and patients make clinical decisions.