John H. Page

In the presence of time-varying confounders affected by prior treatment, standard statistical methods for failure time analysis may be biased. Methods that correctly adjust for this type of covariate include the parametric g-formula, inverse probability weighted estimation of marginal structural Cox proportional hazards models, and g-estimation of structural nested accelerated failure time models. In this article, we propose a novel method to estimate the causal effect of a time-dependent treatment on failure in the presence of informative right-censoring and time-dependent confounders that may be affected by past treatment: g-estimation of structural nested cumulative failure time models (SNCFTMs). An SNCFTM considers the conditional effect of a final treatment at time m on the outcome at each later time k by modeling the ratio of two counterfactual cumulative risks at time k under treatment regimes that differ only at time m. Inverse probability weights are used to adjust for informative censoring. We also present a procedure that, under certain “no-interaction” conditions, uses the g-estimates of the model parameters to calculate unconditional cumulative risks under nondynamic (static) treatment regimes. The procedure is illustrated with an example using data from a longitudinal cohort study, in which the “treatments” are healthy behaviors and the outcome is coronary heart disease.

Background— We prospectively evaluated the relationships between fasting plasma levels of vitamin B ₆ , as pyridoxal phosphate, and subsequent myocardial infarction risk in women. Methods and Results— Among 32 826 women who provided blood samples between 1989 and 1990 (27% of the original 1976 cohort), 239 were diagnosed with incident myocardial infarction (fatal and nonfatal) after blood collection but before July 1998. Of these women, 144 had provided a sample after fasting >10 hours. Cases were matched 1:2 by age, cigarette smoking status, and month of and fasting status at the time of blood collection with controls from the same cohort. Conditional logistic regression was used to adjust for potential confounders, including traditional coronary risk factors, anthropometric factors, dietary intake, and selected biomarkers. Median age at blood collection was 63 years. Plasma levels of pyridoxal phosphate were inversely associated with risk of myocardial infarction; the multivariable-adjusted rate ratio for the highest compared with the lowest quartiles (>70 versus P for trend=0.05). The association varied by age: among women who were Conclusions— Fasting plasma concentration of pyridoxal phosphate was inversely associated with myocardial infarction risk, an effect that was in part independent of dietary B ₆ intake. In addition to dietary vitamin B ₆ intake, there are other determinants of plasma vitamin B ₆ status, and these factors warrant further research.

Background: It is not yet established what specific measures of obesity might be most strongly associated with risk of coronary artery disease. We compared the waist-height ratio to waist-hip ratio, waist circumference, and body mass index as predictors of subsequent coronary heart disease (CHD) in a group of predominantly postmenopausal women. Methods: The data come from the prospective Nurses’ Health Study cohort. We included 45,563 women in 1986 who were aged 40–65 years and were free of heart disease, stroke, and cancer. Waist circumference, hip circumference, height, weight, age, and other covariates were collected by questionnaire. Our primary end point was incident coronary heart disease reported up to June 2002. Areas under the receiver operating characteristic curves (AUCs) were estimated nonparametrically for each of the anthropometric measures. We estimated differences between the AUCs for weight-height ratio and the other measures, with corresponding 95% confidence intervals. We used Cox proportional hazard models to estimate the relationships with risk of CHD. Results: Waist-height ratio, waist-hip ratio, and waist circumference were similar in predicting subsequent risk of CHD. All 3 waist-derived measures were superior to body-mass index (BMI) in predicting CHD. The unadjusted AUCs were 0.62 (95% confidence interval = 0.60–0.64) for waist-height ratio, 0.63 (0.61–0.65) for waist-hip ratio, 0.62 (0.60–0.64) for waist-circumference, and 0.57 (0.55–0.59) for BMI. Conclusion: Waist-height ratio is comparable with waist circumference and waist-hip ratio for prediction of coronary heart disease incidence among middle-aged and older women, but superior to BMI.

Background: The aim of this study was to prospectively evaluate relationships between plasma concentrations of insulinlike growth factor 1 (IGF1) and insulinlike growth factor binding protein 3 (IGFBP3) and subsequent myocardial infarction (MI) in women. Methods: We used case-control sampling to select study participants from women who had already been selected for inclusion in the prospective Nurses’ Health Study cohort. Blood samples were collected from 32 826 women in 1989–1990. During the follow-up period from sample collection through June 1998, MI (fatal and nonfatal) was diagnosed in 245 women. Cases were matched to controls 1:2 by age, cigarette-smoking status, and month and fasting status at the time of blood collection. Conditional logistic regression was used to adjust for potential confounders (menopausal status, parental history of MI, postmenopausal hormone use, diabetes mellitus, hypertension, hypercholesterolemia, aspirin use, alcohol use, body mass index, and physical activity). Results: Multivariable adjusted analyses did not reveal a statistically significant linear relationship between IGF1 or IGFBP3 concentrations or their molar ratio and risk of MI. Women in the highest IGF1 quartile had a multivariable-adjusted rate ratio of 1.46 (95% CI 0.79, 2.72; P for trend = 0.46) for MI, compared with those in the lowest. The corresponding rate ratios (95% CI) for IGFBP3 and the IGF1:IGFBP3 mol/L ratio were 1.24 (0.71, 2.17) and 1.29 (0.70, 2.37), respectively. Conclusions: We did not observe a monotonic relationship between IGF1 or IGFBP3 and MI among predominantly postmenopausal women. Future studies are warranted to evaluate these relationships in other demographic groups including younger women.