Novel Risk Modeling Approach of Atrial Fibrillation With Restricted Mean Survival Times: Application in the Framingham Heart Study Community-Based Cohort

Bidirectional Association Between Atrial Fibrillation and Myocardial Infarction, and Relation to Mortality in the Framingham Heart Study

BACKGROUND

Individuals with both atrial fibrillation (AF) and myocardial infarction (MI) have higher mortality compared with individuals with only 1 condition. Whether mortality differs according to the temporal order of AF and MI is unclear.

METHODS AND RESULTS

We included participants from the FHS (Framingham Heart Study) from 1960 and onwards. We assessed the hazard ratio (HR) of new-onset AF and MI, and mortality according to MI and AF status (prevalent and interim) using multivariable-adjusted Cox proportional hazards models. Interim diseases were modeled as time-varying variables. For the analysis of new-onset AF, 10 923 participants (55% women; mean±SD age, 54±8 years) were included. For new-onset MI, 10 804 participants (55% women; mean±SD age, 54±8 years) were included. Compared with no MI, the hazard of new-onset AF was higher in participants with prevalent (HR, 1.60 [95% CI, 1.32-1.94]) and interim MI (HR, 3.96 [95% CI, 3.18-4.91]). Both ST-segment-elevation MI and non-ST-segment-elevation MI were associated with new-onset AF. Interim AF, not prevalent AF, was associated with higher hazard rate of new-onset MI (HR, 2.21 [95% CI, 1.67-2.92]). Interim AF was associated with both ST-segment-elevation MI and non-ST-segment-elevation MI. Mortality was significantly greater among participants with AF and MI compared with participants with 1 of the 2, regardless of temporal order.

CONCLUSIONS

We report a bidirectional association between AF and MI, which was observed for both non-ST-segment-elevation MI and ST-segment-elevation MI. Participants with both AF and MI had considerably higher mortality compared with participants with only 1 of the 2 conditions, regardless of order.

A Dynamic Prediction Model Supporting Individual Life Expectancy Prediction Based on Longitudinal Time-Dependent Covariates

In the field of clinical chronic diseases, common prediction results (such as survival rate) and effect size hazard ratio (HR) are relative indicators, resulting in more abstract information. However, clinicians and patients are more interested in simple and intuitive concepts of (survival) time, such as how long a patient may live or how much longer a patient in a treatment group will live. In addition, due to the long follow-up time, resulting in generation of longitudinal time-dependent covariate information, patients are interested in how long they will survive at each follow-up visit. In this study, based on a time scale indicator-restricted mean survival time (RMST)-we proposed a dynamic RMST prediction model by considering longitudinal time-dependent covariates and utilizing joint model techniques. The model can describe the change trajectory of longitudinal time-dependent covariates and predict the average survival times of patients at different time points (such as follow-up visits). Simulation studies through Monte Carlo cross-validation showed that the dynamic RMST prediction model was superior to the static RMST model. In addition, the dynamic RMST prediction model was applied to a primary biliary cirrhosis (PBC) population to dynamically predict the average survival times of the patients, and the average C-index of the internal validation of the model reached 0.81, which was better than that of the static RMST regression. Therefore, the proposed dynamic RMST prediction model has better performance in prediction and can provide a scientific basis for clinicians and patients to make clinical decisions.

Restricted mean survival time regression model with time-dependent covariates

In clinical or epidemiological follow‐up studies, methods based on time scale indicators such as the restricted mean survival time (RMST) have been developed to some extent. Compared with traditional hazard rate indicator system methods, the RMST is easier to interpret and does not require the proportional hazard assumption. To date, regression models based on the RMST are indirect or direct models of the RMST and baseline covariates. However, time‐dependent covariates are becoming increasingly common in follow‐up studies. Based on the inverse probability of censoring weighting (IPCW) method, we developed a regression model of the RMST and time‐dependent covariates. Through Monte Carlo simulation, we verified the estimation performance of the regression parameters of the proposed model. Compared with the time‐dependent Cox model and the fixed (baseline) covariate RMST model, the time‐dependent RMST model has a better prediction ability. Finally, an example of heart transplantation was used to verify the above conclusions.

Life-Years Lost After Newly Diagnosed Atrial Fibrillation in Patients with Heart Failure

Objective

Prior work estimated excess death rates associated with atrial fibrillation (AF) in heart failure (HF) with hazard ratios (HR). The aim was to estimate the life-years lost after newly diagnosed AF in HF patients.

Methods

Among patients diagnosed with HF in 2008–2018 in the nationwide Danish Heart Failure Registry, we compared patients with incident AF to referents matched on age, sex, and time since HF. We estimated the marginal hazard ratio (HR) for death and marginal difference in restricted mean survival times (RMST) between AF cases and referents at 10 years after AF diagnosis. We adjusted for sex, age at AF diagnosis, clinical and lifestyle risk factors, and medications.

Results

Among 4463 AF cases and 17,792 referents (mean age 73.7 years, 29% women), the HR was 1.41 (95% CI 1.38; 1.44) but there was evidence of non-proportional hazards. The difference in RMST was −18.2 months (95% CI −16.8; −19.6) at 10 years after AF diagnosis. There were differences in life-years lost between patients diagnosed with AF >1 year and ≤1 year after HF (−25.7 months, 95% CI −23.7; −27.7 vs −10.4 months, 95% CI −8.2; −12.5, p < 0.001), women and men (−20.3 months, 95% CI −17.7; −21.9 vs −17.2 months, 95% CI −15.5; −19.0, p = 0.05), patients with low, medium, and high CHA₂DS₂-VASc (10.3 months, 95% CI −4.6; −16.1 vs −18.5 months, 95% CI −16.7; −20.4 vs 22.1, 95% CI −18.8; −22.3, p = 0.002).

Conclusion

HF patients with incident AF lost on average 1.5 life-years over 10 years after AF. Life-years lost were larger among patients diagnosed with AF >1 year after HF, women, and patients with higher CHA₂DS₂-VASc.

Neck Circumference and Risk of Incident Atrial Fibrillation in the Framingham Heart Study

Background

Increased neck circumference, a proxy for upper‐body subcutaneous fat, is associated with cardiovascular risk and metabolic risk factors, accounting for body mass index (BMI) and waist circumference. The association between neck circumference and incident atrial fibrillation (AF) is unclear. The aim of current study was to evaluate the association between neck circumference and incident AF.

Methods and Results

We selected participants from the Framingham Heart Study aged ≥55 years without diagnosed AF and with available neck circumference, BMI, and waist circumference measurements. We defined high neck circumference as ≥14 inches in women and ≥17 inches in men on the basis of the Contal and O’Quigley changepoint method. We used Fine‐Gray models to estimate subdistribution hazards ratios (sHRs) for the association between neck circumference and incident AF accounting for the competing risk of death. We adjusted models for clinical risk factors. We then additionally adjusted separately for BMI, waist circumference, and height/weight. The study sample included 4093 participants (mean age 64±7 years, 55% female). During 11.2±5.7 mean years of follow‐up, incident AF occurred in 571 participants. High neck circumference was associated with incident AF (sHR for high versus low: 1.58; 95% CI, 1.32–1.90, P<0.0001). The association remained significant after adjustment for BMI (sHR, 1.51; 95% CI, 1.21–1.89; P=0.0003), waist circumference (sHR, 1.47; 95% CI, 1.18–1.83; P<0.0001), and height/weight (sHR, 1.37; 95% CI, 1.09–1.72; P=0.007).

Conclusions

High neck circumference was associated with incident AF adjusting for traditional adiposity measures such as BMI and waist circumference.

Implementation of an Alternative Method for Assessing Competing Risks: Restricted Mean Time Lost

In clinical and epidemiologic studies, hazard ratios are often applied to compare treatment effects between 2 groups for survival data. For competing-risks data, the corresponding quantities of interest are cause-specific hazard ratios and subdistribution hazard ratios. However, they both have some limitations related to model assumptions and clinical interpretation. Therefore, we recommend restricted mean time lost (RMTL) as an alternative measure that is easy to interpret in a competing-risks framework. Based on the difference in RMTL (RMTLd), we propose a new estimator, hypothetical test, and sample-size formula. Simulation results show that estimation of the RMTLd is accurate and that the RMTLd test has robust statistical performance (both type I error and statistical power). The results of 3 example analyses also verify the performance of the RMTLd test. From the perspectives of clinical interpretation, application conditions, and statistical performance, we recommend that the RMTLd be reported along with the hazard ratio in analyses of competing-risks data and that the RMTLd even be regarded as the primary outcome when the proportional hazards assumption fails.

Dynamic prediction and analysis based on restricted mean survival time in survival analysis with nonproportional hazards

BACKGROUND AND OBJECTIVE

In the process of clinical diagnosis and treatment, the restricted mean survival time (RMST), which reflects the life expectancy of patients up to a specified time, can be used as an appropriate outcome measure. However, the RMST only calculates the mean survival time of patients within a period of time after the start of follow-up and may not accurately portray the change in a patient's life expectancy over time.

METHODS

The life expectancy can be adjusted for the time the patient has already survived and defined as the conditional restricted mean survival time (cRMST). A dynamic RMST model based on the cRMST can be established by incorporating time-dependent covariates and covariates with time-varying effects. We analyzed data from a study of primary biliary cirrhosis (PBC) to illustrate the use of the dynamic RMST model, and a simulation study was designed to test the advantages of the proposed approach. The predictive performance was evaluated using the C-index and the prediction error.

RESULTS

Considering both the example results and the simulation results, the proposed dynamic RMST model, which can explore the dynamic effects of prognostic factors on survival time, has better predictive performance than the RMST model. Three PBC patient examples were used to illustrate how the predicted cRMST changed at different prediction times during follow-up.

CONCLUSIONS

The use of the dynamic RMST model based on the cRMST allows for the optimization of evidence-based decision-making by updating personalized dynamic life expectancy for patients.

The restricted mean survival time as a replacement for the hazard ratio and the number needed to treat in long-term studies

Aims

We applied the restricted mean survival time (RMST) to analyse the survival data reported in the PARADIGM‐HT trial in which sacubitril + valsartan was studied in comparison with enalapril in patients with heart failure. The estimates of this parameter were compared with the published values of hazard ratio (HR).

Methods

Two endpoints were evaluated: a composite of death or hospitalization and cardiovascular death. Our analyses were performed by considering the original follow‐up of 41.4 months and on the basis of a lifetime perspective. All statistical calculations were carried out using specific packages developed under the R‐platform.

Results

According to our RMST analysis, the results for the composite endpoint in the comparison of sacubitril + valsartan vs. enalapril showed an improvement from 32.9 to 34.2 months (gain of 1.25 months). This result is based on a time horizon of 41.4 months. The results for the cardiovascular mortality endpoint showed a RMST of 37.2 months for sacubitril + valsartan vs. 36.2 for enalapril (gain of 0.96 months). In the two lifetime analyses, the improvements were much more relevant and yielded a gain of 25.8 months for the composite endpoint and 27.6 months for survival free from cardiovascular death.

Conclusions

Using the data of the PARADIGM‐HT trial, our analysis confirmed that the RMST has documented advantages over the HR, particularly when the clinical study is characterized by a long follow‐up. The number needed to treat (NNT) has a more specific methodological role and cannot be replaced by the RMST.

Estimation and modeling of the restricted mean time lost in the presence of competing risks

Survival data with competing or semi-competing risks are common in observational studies. As an alternative to cause-specific and subdistribution hazard ratios, the between-group difference in cause-specific restricted mean times lost (RMTL) gives the mean difference in life expectancy lost to a specific cause of death or in disease-free time lost, in the case of a nonfatal outcome, over a prespecified period. To adjust for covariates, we introduce an inverse probability weighted estimator and its variance for the marginal difference in RMTL. We also introduce an inverse probability of censoring weighted regression model for the RMTL. In simulation studies, we examined the finite sample performance of the proposed methods under proportional and nonproportional subdistribution hazards scenarios. We illustrated both methods with competing risks data from the Framingham Heart Study. We estimated sex differences in atrial fibrillation (AF)-free times lost over 40 years. We also estimated sex differences in mean lifetime lost to cardiovascular disease (CVD) and non-CVD death over 10 years among individuals with AF.

Trends in excess mortality associated with atrial fibrillation over 45 years (Framingham Heart Study): community based cohort study

OBJECTIVE

To assess temporal trends in the association between newly diagnosed atrial fibrillation and death.

DESIGN

Community based cohort study.

SETTING

Framingham Heart Study cohort, in 1972-85, 1986-2000, and 2001-15 (periods 1-3, respectively), in Framingham, MA, USA.

PARTICIPANTS

Participants with no atrial fibrillation, aged 45-95 in each time period, and identified with newly diagnosed atrial fibrillation (or atrial flutter) during each time period.

MAIN OUTCOME MEASURES

The main outcome was all cause mortality. Hazard ratios for the association between time varying atrial fibrillation and all cause mortality were calculated with adjustment for time varying confounding factors. The difference in restricted mean survival times, adjusted for confounders, between participants with atrial fibrillation and matched referents at 10 years after a diagnosis of atrial fibrillation was estimated. Meta-regression was used to test for linear trends in hazard ratios and restricted mean survival times over the different time periods.

RESULTS

5671 participants were selected in time period 1, 6177 in period 2, and 6174 in period 3. Adjusted hazard ratios for all cause mortality between participants with and without atrial fibrillation were 1.9 (95% confidence interval 1.7 to 2.2) in time period 1, 1.4 (1.3 to 1.6) in period 2, and 1.7 (1.5 to 2.0) in period 3 (P_trend=0.70). Ten years after diagnosis of atrial fibrillation, the adjusted difference in restricted mean survival times between participants with atrial fibrillation and matched referents decreased by 31%, from -2.9 years (95% confidence interval -3.2 to -2.5) in period 1, to -2.1 years (-2.4 to -1.8) in period 2, to -2.0 years (-2.3 to -1.7) in period 3 (P_trend=0.03).

CONCLUSIONS

No evidence of a temporal trend in hazard ratios for the association between atrial fibrillation and all cause mortality was found. The mean number of life years lost to atrial fibrillation at 10 years had improved significantly, but a two year gap compared with individuals without atrial fibrillation still remained.

Incidence and Risk Assessment for Atrial Fibrillation at 5 Years: Hypertensive Diabetic Retrospective Cohort

(1) Background: The link between diabetes and hypertension is mutual and reciprocal, increasing the risks for the development of atrial fibrillation (AF). The main objective was to develop a prediction model for AF in a population with both diabetes and hypertension at five years of follow-up. (2) Methods: A multicenter and community-based cohort study was undertaken of 8237 hypertensive diabetic patients without AF between 1 January 2103 and 31 December 2017. Multivariate Cox proportional-hazards regression models were used to identify predictors AF and to stratify risk scores by quartiles. (3) Results: AF incidence was 10.5/1000 people/years (95% confidence interval (CI) 9.5-11.5), higher in men. The independent prognostic factors identified: age (hazard ratio (HR) 1.07 95% CI 1.05-1.09, p < 0.001), weight (HR 1.03 95% CI 1.02-1.04, p < 0.001), CHA₂DS₂VASc score (HR 1.57 95% CI 1.16-2.13, p = 0.003) and female gender (HR 0.55 95% CI 0.37-0.82, p = 0.004). Q4 (highest-risk group for AF) had the highest AF incidence, stroke and mortality, and the smallest number needed to screen to detect one case of AF. (4) Conclusions: Risk-based screening for AF should be used in high cardiovascular risk patients as the hypertensive diabetics, for treatment of modifiable cardiovascular risk, and monitoring AF detection.