The Long Life Family Study cohort had a wider range of ages than any of the other cohorts (Table ). The mean age in CHS was similar to LLFS, though limited to age 65 and over at the younger end of the distribution. The Framingham Heart Study original cohort was older than both LLFS and CHS, while the FHS offspring cohort was more similar in age to LLFS. The NECS had the oldest individuals, with a mean age of 102.8 years, ranging from 90 to 119 years, similar to the LLFS probands generation. The Long Life Family Study cohort was more highly educated compared to CHS, NECS and FHS and this was more apparent in the LLFS offspring generation. History of smoking was similar among LLFS, CHS and FHS, with about 40-50% of participants with no history of smoking. In the NECS, 80% of participants reported that they never smoked.
Socio-demographic, health and functional characteristics of LLFS and comparison cohorts.
Table presents the odds ratios for disease prevalence in the LLFS proband generation cohort compared to similarly aged subsets of the other comparison cohorts. The odds ratio for heart disease, stroke and hypertension was similar to or lower in the comparison cohorts compared to LLFS probands except for the FHS original cohort which had had a significantly higher risk of hypertension compared to the LLFS cohort. In the case of diabetes, FHS offspring cohort had higher prevalence compared to LLFS probands; there was no difference in diabetes between CHS or NECS participants. The prevalence of cancer was significantly lower for FHS offspring cohort compared to the LLFS cohort, with no differences found between the other cohorts and LLFS. Self-reported history of lung disease was more than three times more prevalent in the CHS cohort, compared to the LLFS, p<0.0001. The risk of having peripheral artery disease (defined as an ankle-arm index <0.9) was three times more likely in the CHS compared with the LLFS participants. Further, the risk of gait speeds of <1.0 m/s were nearly four times more likely for the CHS participants as compared to LLFS, p<0.0001.
LLFS Probands: Odds ratios for disease and disability prevalence in comparison cohorts relative to LLFS; adjusted for age, sex, race, education and smoking status.
Table presents the odds ratios for disease prevalence in the LLFS offspring generation cohort and in the similarly aged subsets of CHS, FHS original, FHS offspring, and LLFS spousal controls. Heart disease prevalence among CHS participants was more than two times higher compared to the LLFS offspring. There were no differences in self-reported history of stroke among the CHS, FHS original, FHS offspring and the LLFS cohort. However, the odds of stroke were more than two-fold higher for LLFS spouse controls compared to LLFS offspring. CHS and FHS original cohorts had significantly lower risk of hypertension compared to the LLFS offspring. For diabetes, CHS and FHS offspring cohorts had higher prevalence compared to LLFS offspring. The prevalence of cancer was significantly lower by 20% for the FHS offspring cohort compared to the LLFS cohort, with no differences found between the other cohorts and LLFS offspring. The CHS cohort had a nearly 2.5-fold greater prevalence of self-reported history of lung disease compared to the LLFS. Peripheral artery disease risk was three times more likely in the CHS compared with the LLFS offspring cohort. In addition, the risk of gait speeds of ≤1.0 m/s were nearly five times more likely for the CHS participants as compared to LLFS offspring, p<0.0001.
LLFS Offspring: Odds ratios for disease and disability prevalence in comparison cohorts relative to LLFS, adjusted for age, sex, race, education and smoking status.
Adjusted mean differences in blood chemistries and physical characteristics were compared for LLFS, CHS, and FHS cohorts (Table and ). LLFS proband generation participants had higher BMI compared to all other cohorts, but this was only statistically significant compared to the CHS cohort. For the LLFS offspring generation, BMI was higher compared to both the CHS and FHS original cohorts. Mean grip strength was higher for LLFS probands compared to FHS original cohort however, it was lower compared to CHS. There were no statistically significant differences in grip strength between the LLFS offspring cohort and any of the comparison groups. Long Life Family Study participants in both the proband and offspring generation had significantly faster walk speeds compared to CHS. Similarly, digit symbol substitution test scores were significantly better for both LLFS cohorts compared to CHS. Additionally, there was a slight difference between the LLFS proband generation and FHS offspring cohort for the MMSE.
LLFS Probands: Measurement means in LLFS and comparison cohorts r adjusted for age, sex, race, education and smoking status.
LLFS Offspring: Measurement means in LLFS and comparison cohorts adjusted for age, sex, race, education and smoking status.
The Long Life Family Study probands and offspring had significantly higher DBP and lower pulse pressures compared to CHS (Tables and ). Pulse pressure was also significantly lower for LLFS proband generation compared with both FHS cohorts. Also, ankle-arm index was significantly better for both LLFS cohorts compared to those in the CHS and the FHS offspring cohort.
All lipid values were significantly better for LLFS proband and offspring generation compared to CHS, including lower total cholesterol, lower LDL-C and lower triglycerides. Triglycerides were lower for Long Life Family Study probands and offspring compared to both FHS cohorts. Both LLFS cohorts consistently exhibited higher levels of HDL-C, by about 5 mg/dL (Tables and ) compared to CHS. Glucose levels were significantly lower by about 13 mg/dL for LLFS probands and 28 mg/dL for LLFS offspring compared to CHS. Kidney function was better for LLFS proband and offspring generations when compared to CHS, FHS original and FHS offspring as exhibited by lower creatinine levels.
For each measure, interactions with age were tested to determine if the differences between LLFS and the other cohorts might be greater at either extreme of age. There were no significant interactions however, age-specific analyses suggested several interesting patterns of age-related change. For pulse pressure, the differences between LLFS and other cohorts were most apparent after age 65. (Figure ) Gait speed was higher in LLFS than the other cohorts (Figure ). Digit symbol substitution test showed that LLFS participants had higher scores at every age, but with a similar age-specific decline (Figure ).
Pulse Pressure by age, adjusted for sex and anti-hypertensive medication use.
Gait Speed (m/s) by age, adjusted for sex.
Digit Symbol Substitution Test (DSST) score by age.