Table 2.

Cox Models Measuring Relationship Between Pregnancy History by Months and Parity and Alzheimer’s Disease Risk.^a

Model Number	Parameter, Natural Log Transformed	Coef	Exp(coef)	Se(coef)	P Value	95% CI
1	Cumulative months pregnant	−1.901	0.1495	0.821	.021b	0.030-0.747
2	Cumulative months pregnant, adjusted for parity	−2.049	0.1289	0.958	.032b	0.020-0.843
3	Parity	0.666	1.947	0.526	.205 (NS)	0.695-5.458
4	Parity, adjusted for cumulative months pregnant	1.693	5.436	1.067	.113 (NS)	0.671-44.019

Abbreviations: CI, confidence interval; NS, not significant.

^a Cox model analysis of the dependence of AD risk on cumulative months pregnant and parity defined as number of full-term pregnancies. All models are adjusted for age at first birth, reproductive span, and history of breastfeeding, marriages, and occupation. The table reports the partial likelihood point estimate for the effect of the parameter, the corresponding exponentiated value, the standard error, the P value for the relative sharp null hypothesis, and the 95% confidence interval for exp(coef). The partial likelihood ratio test P value for the null hypothesis of no effect for model 1 was .005, model 2 was .008, model 3 was .038, and model 4 was .039. The score log-rank test P value for model 1 was .004, model 2 was .007, model 3 was .055, and model 4 was .047. Models were fitted on the basis of 95 sample individuals (10 observations deleted due to ineligibility and 28 observations deleted due to lack of data), for a total of 39 observed failure events.

^b P < .05.