Time-dependent Risk Factors Research Articles

Risk factors and evolution of weight loss in Parkinson’s disease: a 9-year population-based study

Introduction:Weight loss is considered a common complication of Parkinson`s disease (PD), but there are few prospective longitudinal studies on weight loss in patients followed from time of PD diagnosis. We sought to determine the frequency, evolution and risk factors of weight loss in a representative incident PD cohort. Methods: In this prospective population-based observational study, we followed 180 newly-diagnosed, initially drug-naïve PD patients and 161 controls with repetitive weight examinations over 9 years. We used Cox regression models with adjustment for potential confounders to identify independent risk factors of clinically significant (>10%) weight loss. Results: Mean % weight change during follow-up was -3.9 (±11.2) in patients and -1.4 (±8.1) in controls (p=0.016). Clinically significant weight loss was observed in 26.7% of patients and 10.6% of controls (RR 2.53; 95% CI 1.52-4.21; p<0.001). Age was the only independent baseline risk factor for weight loss (HR 1.06 per year; 95% CI 1.03-1.10; p<0.001). Additional time-dependent risk factors were presence of olfactory impairment (HR 2.42; 95% CI 1.14-5.15; p=0.021), presence of dyskinesias (HR 3.14; 95% CI 1.58-6.23; p=0.001), and cognitive impairment (HR per MMSE unit 0.90; 95% CI 0.82-0.99; p=0.036). Dopamine agonist use reduced the risk of weight loss during follow-up (HR 0.44; 95% CI 0.24-0.82; p=0.007) Conclusion: The risk of weight loss is more than doubled in the general PD population and associated with both disease-related features and drug-related complications. This suggests a multifactorial nature of weight loss in PD, which is important to consider in research and clinical practice.

Quantitative risk assessment of coal mine gas explosion based on a Bayesian network and computational fluid dynamics

A gas explosion is one of the most serious coal mine accidents. The complexity and dynamics of gas explosion accidents, resulting from both multiple risk factors and time-dependent physical parameters, necessitates the use of risk assessment techniques to analyze the accident. A quantitative risk assessment method is proposed for coal mine gas explosions based on a Bayesian Network (BN) and Computational Fluid Dynamics (CFD). The results show that, in addition to the problems of “low vigilance” and “substandard safety education”, the “irrational design of mine gas extraction” has become another major risk factor at the new stage of development of mine excavation technology. Three typical gas explosion areas are selected to analyze the possible consequences of an accident. Based on the probabilistic risk assessment model, time-dependent physical parameters such as accident conditions, location and time are analyzed. The probability of occurrence of time-dependent risk factors is quantified to calculate the risk values of the consequences of explosion accidents corresponding to different times. The proposed method helps to improve the calculation accuracy of updating the risk probability value over time and provides theoretical support for the risk analysis of coal mine gas explosion accidents.

Incidence of Malignancies and the Association with Biological Disease-Modifying Antirheumatic Drugs in Japanese Patients with Rheumatoid Arthritis: A Time-Dependent Analysis from the IORRA Patient Registry.

Patients with rheumatoid arthritis (RA) may have an increased malignancy risk versus the general population, potentially elevated by biological disease-modifying antirheumatic drug (bDMARD) use. Using patient registry data, we determined malignancy risk, stratified by bDMARD use, among Japanese patients with RA versus the Japanese general population and investigated whether bDMARD use is a time-dependent risk factor for the development of malignancy. Patients aged ≥ 18years with ≥ 2 data entries of RA in the IORRA (Institute of Rheumatology, Rheumatoid Arthritis) patient registry, enrolled from January 2013-December 2018, were identified ('All RA' cohort). Patients were stratified into bDMARD (≥ 1bDMARD received) or non-bDMARD (no history of bDMARDs) sub-cohorts. Malignancy incidence rates and standardized incidence ratios (SIRs) with 95% confidence intervals (CIs) versus the Japanese general population were calculated. Risk of RA medication use was analyzed using a time-dependent Cox proportional hazards model, after adjusting for covariates. A total of 8020 patients were identified for the All RA cohort; 2187 and 5833 for the bDMARD and non-bDMARD sub-cohorts, respectively. For all three cohorts, incidence of overall malignancies was similar versus the Japanese general population. Incidence of specific malignancies was also similar, but incidence of lymphoma was higher for all three cohorts (SIRs[95% CIs] 3.72 [2.71-4.93], 5.97 [3.34-9.59], and 2.79 [1.82-4.02], respectively). In the bDMARD sub-cohort, no increase in SIRs was observed for other site-specific malignancies. Inthe All RA cohort, use of methotrexate, tacrolimus, glucocorticoids, non-steroidal anti-inflammatory drugs, and bDMARDs were not associated with the risk of overall malignancy; the hazard ratio (95% CI) was 1.36(0.96-1.93) for bDMARD use. Increased disease activity was a time-dependent risk factor of overall malignancy with a hazard ratio (95% CI) of 1.35 (1.15-1.59). The use of bDMARDs was not a time-dependent risk factor for malignancy.

Predictive Partly Conditional Models for Longitudinal Ordinal Outcomes with Application to Alzheimer’s Disease Progression

AbstractAssessing time-dependent risk factors in relation to the risk of disease progression is challenging, yet important, especially for chronic diseases with slow progression. In this paper, the partly conditional model is extended for characterizing disease progression at time t with longitudinal ordinal outcomes in the presence of time-dependent covariates at time s$$(s &lt; t)$$ ( s &lt; t ) and time-dependent effects. Advantages of the method include direct modeling of disease progression, use of longitudinal risk factors, and flexibility in target period of progression $$u = t - s$$ u = t - s . A generalized estimating equation approach is adopted for parameter estimation, and a new regularity condition requiring a fixed prediction time window $$u_0$$ u 0 is established to consistently estimate the covariance of the estimated parameters. Extensive simulation studies are conducted to assess the properties of the proposed model alongside with an explicit demonstration of the implementation using existing statistical software. The proposed method is applied to a longitudinal Alzheimer’s disease dataset from the National Alzheimer’s Coordinating Center to assess effect of time-dependent cognitive complaint on Alzheimer’s Disease progression.

Perinatal risk factors associated with acute kidney injury severity and duration among infants born extremely preterm.

We evaluated time-varying perinatal risk factors associated with early (≤7 post-natal days) and late (>7 post-natal days) severe acute kidney injury (AKI) occurrence and duration. A secondary analysis of Preterm Erythropoietin Neuroprotection Trial data. We defined severe AKI (stage 2 or 3) per neonatal modified Kidney Disease: Improving Global Outcomes criteria. Adjusted Cox proportional hazards models were conducted with exposures occurring at least 72 h before severe AKI. Adjusted negative binomial regression models were completed to evaluate risk factors for severe AKI duration. Of 923 participants, 2% had early severe AKI. In the adjusted model, gestational diabetes (adjusted HR (aHR) 5.4, 95% CI 1.1-25.8), non-steroidal anti-inflammatory drugs (NSAIDs) (aHR 3.2, 95% CI 1.0-9.8), and vancomycin (aHR 13.9, 95% CI 2.3-45.1) were associated with early severe AKI. Late severe AKI occurred in 22% of participants. Early severe AKI (aHR 2.5, 95% CI 1.1-5.4), sepsis (aHR 2.5, 95% CI 1.4-4.4), vasopressors (aHR 2.9, 95% CI 1.8-4.6), and diuretics (aHR 2.6, 95% CI 1.9-3.6) were associated with late severe AKI. Participants who had necrotizing enterocolitis or received NSAIDs had longer severe AKI duration. We identified major risk factors for severe AKI that can be the focus of future research. Time-dependent risk factors for severe acute kidney injury (AKI) and its duration are not well defined among infants born <28 weeks' gestation. Over 1 in 5 infants born <28 weeks' gestation experienced severe AKI, and this study identified several major time-dependent perinatal risk factors occurring within 72 h prior to severe AKI. This study can support efforts to develop risk stratification and clinical decision support to help mitigate modifiable risk factors to reduce severe AKI occurrence and duration.

A dynamic time-to-event model for prediction of acute graft-versus-host disease in patients after allogeneic hematopoietic stem cell transplantation.

Acute graft-versus-host disease (aGvHD) is a major cause of death for patients following allogeneic hematopoietic stem cell transplantation (HSCT). Effective management of moderate to severe aGvHD remains challenging despite recent advances in HSCT, emphasizing the importance of prophylaxis and risk factor identification. In this study, we analyzed data from 1479 adults who underwent HSCT between 2005 and 2017 to investigate the effects of aGvHD prophylaxis and time-dependent risk factors on the development of grades II-IV aGvHD within 100 days post-HSCT. Using a dynamic longitudinal time-to-event model, we observed a non-monotonic baseline hazard overtime with a low hazard during the first few days and a maximum hazard at day 17, described by Bateman function with a mean transit time of approximately 11 days. Multivariable analysis revealed significant time-dependent effects of white blood cell counts and cyclosporine A exposure as well as static effects of female donors for male recipients, patients with matched related donors, conditioning regimen consisting of fludarabine plus total body irradiation, and patient age in recipients of grafts from related donors on the risk to develop grades II-IV aGvHD. Additionally, we found that higher cumulative hazard on day 7 after allo-HSCT are associated with an increased incidence of grades II-IV aGvHD within 100 days indicating that an individual assessment of the cumulative hazard on day 7 could potentially serve as valuable predictor for later grades II-IV aGvHD development. Using the final model, stochastic simulations were performed to explore covariate effects on the cumulative incidence over time and to estimate risk ratios. Overall, the presented model showed good descriptive and predictive performance and provides valuable insights into the interplay of multiple static and time-dependent risk factors for the prediction of aGvHD.

Time-Dependent Factors Influencing Cardiocerebral Vascular Events in Chronic Hemodialysis Patients: Insights from a Longitudinal Study.

BACKGROUND Cardiocerebral vascular events (CVCs) are significant complications in patients undergoing hemodialysis (HD). Given the increased morbidity and mortality associated with CVCs in this population, understanding the factors influencing CVC occurrence over time is crucial. This study aimed to investigate these time-dependent factors in HD patients. MATERIAL AND METHODS A total of 228 HD patients from 2 dialysis centers, with at least 3 months of treatment between 2017 and 2021, were included. Annual clinical data were collected, and patients were monitored until CVC development. Kaplan-Meier analysis and a time-dependent Cox regression model were used for data analysis. RESULTS The mean age of 228 patients was 55.0±15.0 years, and 64.76% were male. For 5 years of monitoring, the mean follow-up interval was 3.1±1.0 years for patients to develop CVCs. The 1-year, 3-year, and 5-year CVC-free rates were 97.47%, 81.31%, and 70.71%, respectively. Time-dependent Cox regression revealed that C-reactive protein was an independent time-dependent risk factor in HD patients and blood flow rate was an independent time-dependent protective factor. The male subgroup and non-diabetic subgroup had these same results. The following were was the independent time-dependent risk factors: white blood cell count for the female subgroup; blood flow rate for the non-elderly subgroup; and C-reactive protein for the diabetic subgroup. None were risk factors for the elderly subgroup. CONCLUSIONS It took an average of 3.1±1.0 years for patients with HD to develop CVCs. C-reactive protein and blood flow rate emerged as key time-dependent influencing factors for CVCs in this population.

Risk as a pattern over time: Delineation of time-dependent risk factors in biological, psychological, and social variables in cancer patients.

Survival in cancer patients is associated with a multitude of biological, social, and psychological factors. Although it is well established that all these factors add to overall mortality, it is not well understood how the predictive power of these parameters changes in a comprehensive model and over time. Patients who attended the authors' outpatient clinic were invited to participate. The authors followed 5180 mixed cancer patients (51.1% female; mean age, 59.1years [SD=13.8]) for up to 16years and analyzed biological (age, sex, cancer site, anemia), psychological (anxiety, depression), and social variables (marital status, education, employment status) potentially predicting overall survival in a Cox proportional hazards model. The median survival time for the entire sample was 4.3years (95% confidence interval, 4.0-4.7). The overall survival probabilities for 1 and 10 years were 76.8% and 38.0%, respectively. Following an empirical approach, the authors split the time interval into five periods: acute, subacute, short-term, medium-term, and long-term. A complex pattern of variables predicted overall survival differently in the five periods. Biological parameters were important throughout most of the time, social parameters were either time-independent predictors or tended to be more important in the longer term. Of the psychological parameters, only depression was a significant predictor and lost its predictive power in the long-term. The findings of this study allow the development of comprehensive patient-specific models of risk and resilience factors addressing biopsychosocial needs of cancer patients, paving the way for a personalized treatment plan that goes beyond biomedical cancer care.

Time dependent risk factors for epileptic seizures in glioblastoma patients: A retrospective analysis of 520 cases.

Epilepsy is a common comorbidity of glioblastoma. Seizures may occur in various phases of the disease. We aimed to assess potential risk factors for seizures in accordance with the point of time at which they occurred. We retrospectively analysed medical files of adult patients with de novo glioblastoma treated at our institution between 01/2006 and 01/2020. We categorized seizures as preoperative seizures (POS), early postoperative seizures (EPS; before initiation of radio(chemo)therapy, RCT), seizures during radiotherapy (SDR; during or < 30 days after RCT) and post-therapeutic seizures (PTS; ≥ 30 days after completion of RCT). We addressed associations between patients' characteristics and their seizures. In the final cohort (N = 520), 292 patients experienced seizures. POS, EPS, SDR and/or PTS occurred in 29.6 % (154/520), 6.0 % (31/520), 13.8 % (70/509) and 36.1 % (152/421) of patients, respectively. POS occurred more frequently in patients with higher Karnofsky performance scores (odds ratio (OR) 3.27, p = 0.001) and tumour location in the temporal lobe (OR 1.51, P = 0.034). None of the parameters we analysed was related to the occurrence of EPS. SDR were independently associated with tumour location (parietal lobe, OR 1.86, P = 0.027) and POS, but not EPS, and were independent of RCT. PTS were independently associated with tumour progression (OR 2.32, P < 0.001) and with occurrence of SDR (OR 3.36, P < 0.001), and negatively correlated with temporal lobe location (OR 0.58, p < 0.014). In patients with tumours exclusively located in the temporal lobe, complete tumour resection was associated with a decreased risk of postoperative seizures. Seizures in glioblastoma patients have various, time-dependent risk factors. Temporal lobe localisation was a risk factor for preoperative seizures; surgery may have had a protective effect in these patients. RCT did not have dose-dependent pro- or anticonvulsive effects. PTS were associated with tumour progression.

Conditional Cancer-Specific Survival for Inflammatory Breast Cancer: Analysis of SEER, 2010 to 2016

BackgroundConditional survival takes into account the time that has elapsed since diagnosis and may have additional informative value. Compared with the static traditional survival evaluation method, conditional survival predictions can be adapted to incorporate the dynamic changes during the disease and provide a more suitable way of identifying time-evolved prognoses. MethodsOf 3333 patients diagnosed with inflammatory breast cancer between 2010 and 2016 were extracted from the Surveillance, Epidemiology, and End Results database. The trend of the hazard rate over time was represented by the kernel density smoothing curve. The traditional cancer-specific survival (CSS) rate was estimated by the Kaplan-Meier method. Conditional CSS assessment was defined as the probability that a patient will survive y years given the x years who already survived after diagnosis, and the formula is as follows: CS(y)=CSS(x + y)/CSS(x). 3-year cancer-specific survival (CSS3) and 3-year conditional cancer-specific survival (CS3) were estimated. The Fine-Gray proportional subdistribution hazard model was constructed to screen for time-dependent risk factors associated with cancer-specific death. Subsequently, a nomogram was applied to predict a 5-year survival rate based on the number of years already survived. ResultsOf 3333 patients, the cancer-specific survival (CSS) rate decreased from 57% in the 4th year to 49% in the 6th year, while the comparable 3-year CS (CS3) rate improved from 65% in the first year to 76% in the third year. Overall, the CS3 rate was superior to actuarial cancer-specific survival, which was also found in subgroup analysis, especially in patients with high-risk characteristics. The Fine-Gray's model indicated that remote organ metastasis (M stage), lymph node metastasis (N stage), and surgery all significantly impacted the prognosis for cancer-specific survival. The Fine-Gray's model-based nomogram was constructed to predict 5-year cancer-specific survival immediately after diagnosis and given survival for 1, 2, 3, and 4 years after diagnosis. ConclusionHigh-risk patients had a significantly improved cancer-specific survival prognosis after surviving for 1 or more years after diagnosis with inflammatory breast cancer. The probability of reaching 5-year cancer-specific survival following diagnosis improves with each additional year survived. More effective follow-up is required for patients diagnosed at an advanced N stage, remote organ metastasis, or not received surgery. Additionally, a nomogram and web-based calculator may be helpful for patients with inflammatory breast cancer during follow-up counseling (https://ibccondsurv.shinyapps.io/dynnomapp/).

Ambient PM2.5, ozone and mortality in Chinese older adults: A nationwide cohort analysis (2005–2018)

BackgroundCohort evidence linking long-term survival with exposure to multiple air pollutants (e.g., fine particulate matter [PM2.5] and ozone) was extensively sparse in low- and middle-income countries, especially among older adults. This study aimed to investigate potential associations of long-term exposures to PM2.5 and ozone with all-cause mortality in Chinese older adults. MethodsA dynamic nationwide prospective cohort comprising 20,352 adults aged ≥65 years were enrolled from the Chinese Longitudinal Healthy Longevity Study and followed up through 2005–2018. Participants’ annual exposures to warm-season ozone and year-round PM2.5 were assigned using satellite-derived spatiotemporal estimates. A directed acyclic graph (DAG) was developed to identify confounding variables. Associations of annual mean exposures to PM2.5 and ozone with mortality were evaluated using single- and two-pollutant Cox proportional hazards models, adjusting for time-dependent individual risk factors and ambient temperature. ResultsDuring 100 thousand person-years of follow-up (median: 3.6 years), a total of 14,313 death events occurred. The participants were averagely aged 87.1 years at baseline and exposed to a wide range of annual average concentrations of warm-season maximum 8-hour ozone (mean, 54.4 ppb; range, 23.3–81.6 ppb) and year-round PM2.5 (mean, 65.5 μg/m3; range, 10.1–162.9 μg/m3). Approximately linear concentration-response relationship was identified for ozone, whereas significant increases in PM2.5-associated mortality risks were observed only when concentrations were above 60 μg/m3. Rises of 10 ppb in ozone and 10 µg/m3 in PM2.5 above 60 µg/m3 were associated with increases in all-cause mortality of 13.2% (95% confidence interval [CI]: 10.2–16.2%) and 6.2% (95% CI: 4.6–7.7%) in DAG-based single-pollutant model, and of 9.7% (95% CI: 6.6–13.0%) and 5.3% (95% CI: 3.7–6.9%) in DAG-based two-pollutant model, respectively. We detected significant effect modification by temperature in associations of mortality with ozone (P <0.001 for interaction), suggesting greater ozone-related risks among participants in warmer locations. ConclusionsThis study provided longitudinal evidence that long-term exposure to ambient PM2.5 and ozone significantly and independently contributed to elevated risks of all-cause mortality among older adults in China.

The time-varying association between previous antibiotic use and antibiotic resistance

ObjectivesThe objective of the study was to estimate how the time elapsed from previous antibiotic use is associated with antibiotic resistance. MethodsData comprised electronic medical records of all patients in an Israeli hospital who had a positive bacterial culture from 2016 to 2019. These included susceptibility testing results and clinical and demographic data. Mixed-effects time-varying logistic models were fitted to estimate the association between the time elapsed since the last use of aminoglycosides and gentamicin resistance (n = 13 095), cephalosporins and ceftazidime resistance (n = 13 051), and fluoroquinolones and ciprofloxacin resistance (n = 15 364) while adjusting for multiple covariates. ResultsFor all examined antibiotics, previous antibiotic use had a statistically significant association with resistance (p < 0.001). These associations exhibited a clear decreasing pattern over time, which we present as a flexible function of time. Nonetheless, previous antibiotic use remained a significant risk factor for resistance for at least 180 days when the adjusted ORs were 1.94 (95% CI, 1.40–2.69), 1.33 (95% CI, 1.10–1.61), and 2.25 (95% CI, 1.49–3.41) for gentamicin, ceftazidime, and ciprofloxacin, respectively. DiscussionThe association between prior antibiotic use and resistance decreases over time. Commonly used cut-offs for prior antibiotic use can either misclassify patients still at higher risk when too recent or provide a diluted estimate of the effects of antibiotic use on future resistance when too distant. Hence, prior antibiotic use should be considered a time-dependent risk factor for resistance in both epidemiological research and clinical practice.

Determinants of Early and Late In-Hospital Mortality After Acute Myocardial Infarction: A Subanalysis of the OBTAIN Registry

BackgroundPredictors of in-hospital mortality after myocardial infarction (MI) have been reported dichotomously: survival vs death. Predictors of time from admission to death have not been reported. MethodsA total of 7335 patients were enrolled in a prospective multicentre registry of acute MI. In-hospital mortality was classified by time from admission as acute (≤ 2 days), subacute (3 to 7 days), late (8 to 14 days), and very late (≥ 15 days) to identify factors associated with time to death in patients who died before discharge. Patient and MI characteristics, in-hospital interventions, and electrocardiographic findings were screened for differences in time to in-hospital death. ResultsIn-hospital death affected 351 patients (4.8%). Mean age was 72.0 ± 12.4 years, and 40.5% were female patients. Median survival was 5 days (interquartile range: 2-12), and 41% of in-hospital deaths occurred after 1 week. Cardiac biomarkers and ejection fraction were not related to time to in-hospital death. Previous MI, systolic blood pressure, pharmacologic therapy, and interventional treatments were different among the 4 groups. The factors associated with late in-hospital death were coronary artery bypass graft surgery (CABG), new-onset atrial fibrillation or flutter, heart failure or pulmonary edema, bleeding, and lung disease. Acute and subacute in-hospital death was associated with ST-elevation MI, lower systolic blood pressure, and cardiac arrest on admission. CABG was performed in 12% of post-MI patients who died in hospital. ConclusionsClinical risk factors for in-hospital mortality evolve over time immediately after acute MI. Understanding the time-dependent risk factors may allow for the development of new approaches to curtail the “later” in-hospital mortality.

Cholesterol and Hypertension Treatment Improve Coronary Risk Prediction but Not Time-Dependent Covariates or Competing Risks.

Background and AimsCardiovascular (CV) risk functions are the recommended tool to identify high-risk individuals. However, their discrimination ability is not optimal. While the effect of biomarkers in CV risk prediction has been extensively studied, there are no data on CV risk functions including time-dependent covariates together with other variables. Our aim was to examine the effect of including time-dependent covariates, competing risks, and treatments in coronary risk prediction.MethodsParticipants from the REGICOR population cohorts (North-Eastern Spain) aged 35–74 years without previous history of cardiovascular disease were included (n = 8470). Coronary and stroke events and mortality due to other CV causes or to cancer were recorded during follow-up (median = 12.6 years). A multi-state Markov model was constructed to include competing risks and time-dependent classical risk factors and treatments (2 measurements). This model was compared to Cox models with basal measurement of classical risk factors, treatments, or competing risks. Models were cross-validated and compared for discrimination (area under ROC curve), calibration (Hosmer–Lemeshow test), and reclassification (categorical net reclassification index).ResultsCancer mortality was the highest cumulative-incidence event. Adding cholesterol and hypertension treatment to classical risk factors improved discrimination of coronary events by 2% and reclassification by 7–9%. The inclusion of competing risks and/or 2 measurements of risk factors provided similar coronary event prediction, compared to a single measurement of risk factors.ConclusionCoronary risk prediction improves when cholesterol and hypertension treatment are included in risk functions. Coronary risk prediction does not improve with 2 measurements of covariates or inclusion of competing risks.

A Prospective Longitudinal Study of Early Childhood Caries Onset in Initially Caries-Free Children.

Early childhood caries (ECC) is a complex oral disease that is prevalent in US children. The purpose of this 2-y prospective cohort study was to examine baseline and time-dependent risk factors for ECC onset in initially caries-free preschool children. A cohort of 189 initially caries-free children aged 1 to 3 y was recruited. At each 6-mo study visit, children were examined using the ICDAS index; salivary samples were collected to assess mutans streptococci (MS), lactobacilli, Candida species, salivary cortisol (prior and after a stressor), and salivary IgA. Diet and oral health behavior were assessed from parent report. Child and family stress exposure was assessed from measures of psychological symptoms, stressful life event exposure, family organization and violence exposure, and social support. Sociodemographic factors were also considered. A Kaplan-Meier estimator of survival function of time to ECC and a Cox proportional hazards model were used to identify predictors of ECC onset. Onset of ECC was associated with high salivary MS levels at baseline (log-rank test, P < 0.0001). Cox proportional hazards regression showed that the risk of dental caries significantly increased with salivary MS in log scale over the 6-mo period (hazard ratio, 1.08; P = 0.01). Other risk factors in the model did not reach statistical significance. Our results provide prospective evidence that an increase in salivary MS predicts ECC onset in young, initially caries-free children, confirming that a high salivary MS count likely plays a causal role in ECC onset, independent of covariates. These results suggest that we must focus on reducing salivary MS counts in young children and preventing or delaying MS colonization in infants and young children determined to be at risk for ECC.

Cholesterol and hypertension treatment but not time-dependent covariates or competing risks improve coronary risk prediction: the REGICOR study

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Carlos III Health Institute Agency for Management of University and Research Grants Background Cardiovascular (CV) risk functions are the recommended tool to identify high-risk individuals. However, discrimination of CV risk functions is not optimal. While the effect of biomarkers in CV risk prediction has been extensively studied, there is no data of CV risk functions including time-dependent covariates, competing risks and treatments. Aim To examine the effect of including time-dependent covariates, competing risks and treatments in CV risk prediction. Methods Participants from the REGICOR population cohorts (North-Eastern Spain) aged 35-74 years without previous history of cardiovascular disease were included (n=8,470). Coronary and stroke events, and mortality due to other CV causes or to cancer were recorded during the follow-up (median=12.6 years). A multi-state Markov model was constructed to include competing risks and time-dependent classical risk factors and treatments (2 measurements). This model was compared to Cox models with the basal measurement of classical risk factors, treatments or competing risks. Models were cross-validated and compared by their discrimination (area under the ROC curve), calibration (Hosmer-Lemeshow test) and reclassification (categorical net reclassification index). Results Cancer mortality was the event with the highest cumulative incidence. In coronary event prediction, cholesterol and hypertension treatment addition to classical risk factors, improved significantly discrimination by 2% and reclassification by 7-9%. In stroke event prediction, inclusion of time-dependent covariates decreased significantly discrimination by 3-5%. Conclusion Coronary risk prediction improves when cholesterol and hypertension treatment are included in risk functions. Coronary/stroke prediction does not improve with 2 measurements of covariates or with competing risks.

Time-dependent incidence rates and risk factors for transferring to hemodialysis in patients on peritoneal dialysis under the Thai PD-First Policy.

Haemodialysis (HD) transfer (HDT) is the major challenge of peritoneal dialysis (PD). This study aimed to analyse the time-dependent incidence rates and risk factors for permanent HDT in patients under Thailand's PD First policy. The records of 20,545 patients from January 2008 to June 2018 were studied. The time on therapy (TOT) was divided into 0-3, 3-12, 12-24, 24-36, 36-48 and more than 48 months. The time-dependent incidence rates and causes of PD dropout were investigated. The risk factors for HDT were analysed by multivariable Poisson regression model and presented as incidence rate ratios (IRRs) and 95% confidence intervals (CIs). The main cause of PD dropout was death (45.7%) with 17.4% of the patients transferred to HD. The median (25th to 75th interquartile range) dialysis vintage was 1.4 (0.5-2.7) years. The incidence rates of HDT increased with TOT. Patients with universal coverage were transferred to HD less frequently than those with other health schemes. Patients who were illiterate or only had primary school education had a higher risk of being transferred to HD after 48 months of TOT (IRR 1.41 (95% CI 1.07-1.89)). Peritonitis within the first year of PD was the risk for HDT during 13-48 months of PD. The reasons for HDT changed with TOT. Mechanical complications followed by peritonitis were the main causes of HDT during the first 3 months, and after that peritonitis was the main reason. The incidence of HDT increased with TOT. The risks for HDT changed over time on PD.

Estimation of fracture risk by the FRAX tool in patients with systemic lupus erythematosus: a 10-year longitudinal validation study.

Background:The fracture risk assessment tool has been widely used to stratify the 10-year fracture risk to guide therapy. Using the actual fracture data of a 10-year longitudinal cohort of older patients with systemic lupus erythematosus, we reported an underestimation of the tool in predicting major symptomatic osteoporotic fractures. Treatment of osteoporosis in systemic lupus erythematosus should not be based on fracture risk estimation alone. Relevant time-dependent risk factors should be taken into account for an individualized decision.Objective:To compare the observed fracture incidence in a 10-year longitudinal cohort of patients with systemic lupus erythematosus (SLE) with the fracture risk prediction from the fracture risk assessment (FRAX) tool.Methods:Adult patients (⩾40 years) with SLE who had a first DEXA scan performed in 2005–2009 were studied. The 10-year rates of major osteoporotic and hip fractures were estimated by FRAX using clinical data at DEXA with adjustment for prednisolone dosage. The actual incidence of clinical fractures at 10 years was compared with the estimated rates. Factors associated with new fractures were studied by logistic regression.Results:A total of 229 SLE patients were studied (age: 50.2 ± 6.6 years, 93% women). Glucocorticoid was used in 148 (65%) patients at baseline (mean dose: 7.3 ± 6.9 mg/day; 34% ⩾ 7.5 mg/day). Osteoporosis (bone mineral density T score ⩽ –2.5) at the hip, femoral neck, or spine was present in 61 (27%) patients. The estimated 10-year risk of major osteoporotic and hip fractures by FRAX was 3.4 ± 4.5% and 0.95 ± 2.3%, respectively. After 10 years, three patients developed hip fracture, 6 patients had limb fractures and 20 patients had symptomatic vertebral fractures (major osteoporotic fracture 12.7%, hip fracture 1.3%). The actual major osteoporotic fracture rate was significantly higher than the FRAX estimation (12.7% vs 3.4%; p < 0.001). Logistic regression revealed that osteoporosis (odds ratio (OR): 4.07 [1.51–10.9]), previous fragility fracture (OR: 3.18 [1.02–9.90]), and a parental history of fracture (OR: 4.44 [1.16–17.0]) were independently associated with new clinical fractures at 10 years.Conclusion:The FRAX tool underestimates the major clinical fracture risk at 10 years in patients with SLE.

Dynamic Risk Prediction for Hospital-Acquired Pressure Injury in Adult Critical Care Patients

To develop and validate a dynamic risk prediction model to estimate the risk of developing a hospital-acquired pressure injury among adult ICU patients. ICU admission data were split into training and validation sets. With death as a competing event, both static and dynamic Fine-Gray models were developed to predict hospital-acquired pressure injury development less than 24, 72, and 168 hours postadmission. Model performance was evaluated using Wolbers' concordance index, Brier score, net reclassification improvement, and integrated discrimination improvement. We performed a retrospective cohort study of ICU patients in a tertiary care hospital located in San Francisco, CA, from November 2013 to August 2017. Data were extracted from electronic medical records of 18,019 ICU patients (age ≥ 18 yr; 21,220 encounters). Record of hospital-acquired pressure injury data was captured in our institution's incident reporting system. The information is periodically reviewed by our wound care team. Presence of hospital-acquired pressure injury during an encounter and hospital-acquired pressure injury diagnosis date were provided. The dynamic model predicting hospital-acquired pressure injury more than 24 hours postadmission, including predictors age, body mass index, lactate serum, Braden scale score, and use of vasopressor and antifungal medications, had adequate discrimination ability within 6 days from time of prediction (c = 0.73). All dynamic models produced more accurate risk estimates than static models within 26 days postadmission. There were no significant differences in Brier scores between dynamic and static models. A dynamic risk prediction model predicting hospital-acquired pressure injury development less than 24 hours postadmission in ICU patients for up to 7 days postadmission was developed and validated using a large dataset of clinical variables readily available in the electronic medical record.

Impact of ABO Blood Group Type on Risk of Venous Thromboembolism in Patients with Cancer

Introduction: Venous thromboembolism (VTE) is common in patients with cancer. Non-O blood type is associated with higher levels of factor FVIII activity and von Willebrand factor compared to blood type O, and has been identified as a risk factor for VTE in the general population. However, the impact of ABO blood type on risk of cancer-associated VTE has not been clarified.Methods: To determine the influence of non-O blood type on risk of cancer-associated VTE, we utilized the dataset of the Vienna Cancer and Thrombosis Study (CATS), which is a single center, prospective observational cohort study including patients with newly diagnosed or recurrent cancer. Patients were followed for objectively diagnosed, independently adjudicated VTE for a maximum of 2 years. VTE was quantified in competing risk analysis, accounting for all-cause mortality as competing outcome event. A proportional sub-hazard regression model according to Fine & Gray was used for between-group comparisons. Based on the violation of the proportional sub-hazard assumption, we explored potential time-dependent effects of non-O blood type on VTE risk in a restricted cubic spline analysis, modeling differences in risk estimates over follow-up time. Further, time-restricted subdistribution hazard ratios (SHR) were obtained specifically for the <3 months and ≥3 months follow-up intervals. In a subgroup analysis, differences in VTE risk according to ABO-blood type were analyzed for patients with very high thrombotic risk tumor types (pancreatic, gastric, glioblastoma), compared to the remainder of patients.Results: In total, 1,708 patients were included in our analysis (46% female, median age: 61 years [interquartile range, IQR: 52-68]). The most common tumor types were lung (19%), breast (16%), and brain (14%) cancer, with 32% of solid tumor patients having metastatic disease at study inclusion. Over a median follow-up of 24 months (IQR: 10-24), 151 patients were diagnosed with VTE (cumulative 2-year incidence: 9.2%, 95% confidence interval [CI]: 7.9-10.7) and 649 patients died (2-year mortality: 38%). Overall, blood type O was present in 38% of patients, A in 40%, B in 15%, and AB in 7%. The cumulative incidence of VTE at 3-, 6-, 12-, and 24-months for patients with blood type O was 3.8% (95% CI: 2.5-5.5), 5.7% (95% CI: 4.1-7.7), 7.0% (95% CI: 5.1-9.1), and 7.6% (95% CI: 5.7-9.9), compared to 3.4% (95% CI: 2.4-4.7), 6.5% (95% CI: 5.1-8.1), 8.4% (95% CI:6.8-10.2), and 10.2% (95% CI: 8.4-12.2) in patients with non-O blood type (Gray´s test: p=0.103, Figure 1). Upon visual inspection of cumulative incidence functions, a violation of the proportional sub-hazard assumption was suspected. In restricted cubic spline analysis, estimating hazard ratio (HR) for VTE of patients with non-O compared to O blood type, a time-varying effect of non-O blood type towards an increased VTE risk was observed (Figure 2). Based on that, time-restricted competing risk regression models were performed. During the first 3 months of follow-up, no differences in VTE risk were found (SHR for non-O vs. O blood type: 1.00, 95% CI: 0.60-1.67, p=0.992). Beyond the first 3-month follow-up, patients with non-O blood type had an increased VTE risk compared to patients with blood type O (SHR 1.79, 95%CI: 1.12-2.85, p=0.015). In a subgroup analysis, no association with VTE risk was found in patients with very high thrombotic risk tumor types (SHR 0.94, 95% CI: 0.55-1.61, p=0.824). In contrast, in patients with low/intermediate risk cancer, non-O blood type was associated with increased risk of VTE (SHR 1.73, 95% CI: 1.09-2.73, p=0.019).Conclusion: Non-O blood type was identified as a time dependent risk factor for cancer-associated VTE. In the first 3 months after study inclusion, characterized as the highest VTE risk period in our cancer cohort, no differences in VTE risk between blood types were found. Afterwards, beyond the first 3 months of follow-up, an increased VTE-risk in non-O blood types was observed, comparably in magnitude to the risk difference in the general non-cancer population. Further, an association of non-O blood type with VTE risk in patients with low/intermediate thrombotic risk cancers was observed, whereas no effect was present in those with very high-risk tumors. These findings indicate non-O blood type as a putative risk factor for VTE in patients with cancer in comparably low thrombotic risk scenarios. [Display omitted] DisclosuresPabinger: Pfizer: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Novartis: Consultancy, Honoraria; Alexion: Consultancy, Honoraria; Daiichi Sanchyo: Consultancy, Honoraria; Takeda: Consultancy, Honoraria; Bayer: Consultancy, Honoraria; NovoNordisk: Consultancy, Research Funding; CSL Behring: Consultancy, Honoraria, Research Funding.