Associations Of Frailty With Cardiometabolic Disease And Dose-dependent Effects Of Frailty-stratified Physical Activity

BACKGROUND: With the population aging,cardiometabolic disease(CMD,in this study including coronary heart disease,stroke,and diabetes)has become the leading cause of death among middle-aged and older people,and ≥2CMDs often coexist in an individual,leading to an emerging multimorbidity pattern,cardiometabolic multimorbidity(CMM),which contributes to a severe disease burden.Etiologically,aging was traditionally considered as a nonmodifiable risk factor for CMD,and was included in risk prediction systems in the form of chronological age.However,chronological age could not reflect aging-related decline in physiological function directly.Frailty is considered as a biological age marker reflecting aging-related decline in physiological reserve and function.There has been few systematic study on the role of frailty in CMD progression trajectory,its interaction with CMDs on the risk of mortality is unclear,and physical activity-related dose-effect relationship with mortality and CMD risk among people with different frailty status remains controversial.OBJECTIVE: To assess the transition-specific associations of frailty with CMD progression trajectory,explore the interactive effects between frailty and CMD status on mortality risk,and further describe the frailty-stratified doseeffect relationship of device-measured physical activity with mortality and CMD incidence.METHODS: Based on UK Biobank,we assessed participants’ frailty status using two tools,frailty phenotype and frailty index(FI).Based on 5 criteria(unintentional weight loss,exhaustion,physical activity,walking speed,and grip strength),frailty phenotype classified participants as robust(0 criteria),prefrail(1-2 criteria),or frail(≥3 criteria)according to the fulfilled number of criteria.FI was calculated using 49 self-reported items covering health,diseases,disabilities,and mental well-being traits,and defined frailty status as robust(<0.12),mildly frail(0.12-0.24),or moderate/severe frail(>0.24).Then we conducted the following analyses:(1)Based on 418 290 participants with complete information in frailty and covariates and without CMDs at baseline,we constructed multi-state models consisting of 4 health states(no CMD,single CMD,CMM,and death)and 5transitions((1)No CMD to single CMD;(2)No CMD to death;(3)Single CMD to CMM;(4)Single CMD to death;(5)CMM to death),and analyzed the transition-specific role of baseline frailty along the progression trajectory of CMD.Then,based on 17 106 participants with information in frailty and covariates for both baseline and repeated assessment visits and without CMD at repeated visit,we defined the changes in frailty status as stable,improved,or worsen according to fulfilled criteria of frailty phenotype or FI at two visits,and estimated the associations of changes in frailty status with the risk of CMD,CMM,and mortality in Cox proportional hazard models.(2)We included 467 406 participants with complete information in baseline frailty and covariates,and defined CMD status at baseline as no CMD,single CMD or CMM.With all-cause and cardiac mortality as outcomes,we examined the multiplicative and additive interactions between frailty and CMD status in Cox proportional hazard models,and further assessed the joint association of coexisting frailty and CMD with outcomes.(3)We included 84 355 participants who had complete information in accelerometer-measured physical activity,frailty and covariates and were free of CMDs when participating the accelerometer measurements.According to accelerometer-measured data,we calculated the duration of light intensity physical activity(LPA),moderate/vigorous intensity physical activity(MVPA)and levels of total physical activity(TPA)per week,and then described the doseeffect relationship of device-measured physical activity with mortality and CMD risk among people with different frailty status using restricted cubic spline(RCS)curves based on Cox proportional hazard model.RESULTS:(1)Compared with people with robust phenotype,frail patients had 89%,156% higher risk experiencing transitions from no CMD progression to single disease or death,50% and 65% higher risk from single CMD to CMM or death,and 53% higher risk of death with CMM.Consistent results were observed for FI.When considering specific diseases,frailty was most strongly associated with coronary heart disease and diabetes-related progression transitions.Compared with those with stable frailty status,participants experiencing improved status assessed by frailty phenotype had a 22% and 28% lower risk of single CMD and mortality,while those with a worsen frailty status had 54% and 82% increased risk of developing single CMD and CMM,and 68% increased risk of all-cause death.Similar associations were also observed for changes in frailty defined by FI with the risk of single CMD and death.(2)There was an additive interaction between frailty phenotype and CMD status on the risk of death.The combination of frailty phenotype and CMM had a 3.91 times higher risk of all-cause death,23% of which was attributable to the additive interaction,and a 7.33 times higher risk of cardiac death,36% of which was attributable to the additive interaction.A significant multiplicative interaction was observed between FI-derived frailty status and CMD,rather than additive interaction.Compared with those free of frailty and CMD at baseline,patients with moderate/severe frailty by FI and CMM had a 3.63 and 5.91 times higher risk of all-cause and cardiac death.(3)Overall,LPA,MVPA,and TPA-related mortality risk tended to decrease initially and then flat or slightly increased.For risk of CMD incidence,frailtystratified dose-effect curves varied according to intensity of physical activity.In robust or prefrail population assessed by frailty phenotype,the risk of all-cause mortality decreased until LPA and MVPA reached approximately 2 000 min/week and 500 min/week,while that of frail people did not decrease significantly after LPA ≥1 500 min/week and MVPA ≥300 min/week.Priorly,per 60-minute/week increase in LPA or MVPA among frail people was associated with approximately11% and 47% lower risk of death,respectively.Similar results were observed in analysis based on FI assessment.CONCLUSIONS:(1)Frailty was an independent risk factor for all transitions of CMD progression trajectory,and long-term improvement(worsening)of frailty was associated with lower(higher)risk of CMD progression and mortality.(2)There was an interaction between frailty and CMD status on the risk of death.Frailty combined with CMD was significantly associated with an increased risk of death,with a dose-dependent effect according to the degree of frailty and number of CMDs.(3)Increasing physical activity level was independently associated with a significantly lower risk of death,and the highest levels of physical activity that could benefit frail patients were lower than those for people who were less frail.Significant clinical benefit in mortality and CMD incidence could be achieved by increasing physical activity levels,especially MVPA levels,in frail population.