Dynamic Determinants of Longevity and Exceptional Health

Table 4

Effect of “dynamic” risk factors calculated from individual trajectories of physiological indices with nonmonotonic patterns on risk of onset of “unhealthy life” in the Framingham Heart Study (original cohort) estimated by the Cox proportional hazards model.

Physiological Index	Risk Factor (RF)	Mean RF (St. Dev.)	iCox Regression Model
			Parameter (S.E.)	Hazard Ratio (95% C.I.)
BMI; (𝑁=1380,; 𝑁𝑒=782,; 𝑁𝑐=598)	Age Max	63.199 (8.535)	0.012 (0.006)	1.189 (0.997, 1.417)
	Max Index	27.383 (4.016)	−0.011 (0.024)	0.947 (0.757, 1.186)
	Intercept2L	25.696 (3.875)	0.037 (0.025)	1.184 (0.954, 1.469)
	Left Slope	0.194 (0.350)	0.340^∗ (0.155)	1.073 (1.007, 1.143)
	Right Slope	−0.235 (0.801)	−0.041 (0.044)	0.991 (0.972, 1.011)
	Variability2L	0.703 (0.359)	−0.007 (0.121)	0.997 (0.910, 1.092)
	Sex		0.497^† (0.102)	1.644 (1.346, 2.009)
DBP; (𝑁=2139,; 𝑁𝑒=1512,; 𝑁𝑐=627)	Age Max	55.325 (7.112)	−0.006 (0.004)	0.926 (0.827, 1.037)
	Max Index	85.806 (9.966)	0.017^† (0.004)	1.239 (1.129, 1.360)
	Intercept2L	79.309 (12.066)	0.001 (0.003)	1.013 (0.940, 1.093)
	Left Slope	0.912 (1.808)	0.009 (0.017)	1.008 (0.978, 1.040)
	Right Slope	−0.980 (2.390)	−0.014 (0.009)	0.986 (0.968, 1.004)
	Variability2L	3.850 (1.311)	0.039 (0.020)	1.064 (0.999, 1.133)
	Sex		0.488^† (0.053)	1.629 (1.470, 1.806)
HC; (𝑁=1254,; 𝑁𝑒=705,; 𝑁𝑐=549)	Age Max	66.094 (7.125)	−0.011 (0.006)	0.863 (0.732, 1.017)
	Max Index	46.088 (3.141)	0.028 (0.018)	1.129 (0.967, 1.319)
	Intercept2L	43.417 (3.810)	−0.002 (0.016)	0.990 (0.868, 1.130)
	Left Slope	0.372 (0.713)	−0.041 (0.078)	0.985 (0.933, 1.041)
	Right Slope	−0.775 (2.966)	0.0002 (0.013)	1.000 (0.985, 1.016)
	Variability2L	1.442 (0.534)	0.033 (0.074)	1.021 (0.930, 1.122)
	Sex		0.363^† (0.091)	1.438 (1.203, 1.718)
PR; (𝑁=1401,; 𝑁𝑒=1013,; 𝑁𝑐=388)	Age Max	47.167 (7.589)	0.004 (0.005)	1.057 (0.921, 1.213)
	Max Index	80.449 (10.535)	0.009^∗ (0.004)	1.129 (1.011, 1.262)
	Intercept2L	71.162 (13.289)	−0.001 (0.004)	0.987 (0.893, 1.091)
	Left Slope	1.443 (3.034)	0.007 (0.013)	1.010 (0.973, 1.049)
	Right Slope	−0.850 (1.650)	0.013 (0.023)	1.014 (0.968, 1.062)
	Variability2L	4.954 (1.936)	−0.002 (0.018)	0.994 (0.913, 1.082)
	Sex		0.569^† (0.070)	1.766 (1.539, 2.025)
SCH; (𝑁=1361,; 𝑁𝑒=913,; 𝑁𝑐=448)	Age Max	56.821 (8.200)	0.005 (0.006)	1.063 (0.908, 1.245)
	Max Index	261.843 (42.286)	0.002 (0.001)	1.120 (0.978, 1.282)
	Intercept2L	225.919 (62.186)	−0.001 (0.001)	0.943 (0.854, 1.040)
	Left Slope	5.147 (8.013)	0.002 (0.007)	1.012 (0.936, 1.095)
	Right Slope	−4.329 (9.443)	−0.004 (0.004)	0.984 (0.952, 1.017)
	Variability2L	13.452 (6.246)	0.010 (0.006)	1.070 (0.986, 1.161)
	Sex		0.595^† (0.097)	1.813 (1.499, 2.192)

Notes. ^∗0.01≤𝑃<.05, ^#.001≤𝑃<.01, ^x.0001≤𝑃<.001, ^†𝑃<.0001, for other estimates: 𝑃≥.05; Sex: 1—male, 0—female; the other Risk Factors are continuous and calculated as described in Section 2; 𝑁 denotes the total number of individuals; 𝑁_𝑒 is the total number of events (onset of “unhealthy life”); 𝑁_𝑐 is the total number of censored individuals; Hazard Ratios for continuous risk factors are for an increase from the first quartile to the third quartile of respective empirical distributions.