University of Minnesota

U.S. Railroad Study

Study Category: The Cohort Studies (1947-1972)
Year Begun: 1957
Location: United States of America
Principal Investigator: Taylor, Henry Longstreet
External Resource: View URL


The study was designed to study prospectively the relationship of occupational physical activity, other risk characteristics, and cardiovascular disease, hypothesizing a graded relation of activity to risk from active to sedentary jobs. It was assumed that answers relied on observational studies, that a large cohort with follow-up was needed to reduce bias attendant on a “natural” concentration of cardiac susceptibles in the more sedentary rail jobs, and finally, that an idealized experiment using exercise was infeasible.


The cohort was enrolled between 1957 and 1960 from rail companies in the midwest and northwest of the United States, based on job profiles from active maintenance-of-way workers, to switchmen, to clerks, to executives. In a mobile laboratory visiting the worksites, 2,571 men aged 40-59 were examined, a participation rate of 75 per cent, and measures made of blood pressure, blood cholesterol, anthropometrics, respiratory function, a treadmill stress test, and questionnaires on smoking and diet patterns. Subjects were reexamined after 5 years and followed for 40 years through the Railroad Retirement Board and the National Death Index, with no losses to follow-up. Final analyses were run on 2,376 CVD-free men using the Cox proportional hazards model, based on all events in 40 years and on partitioned 5 year blocks and censored cases.


Age, systolic blood pressure, serum cholesterol, and cigarette smoking have a constant, regular, monotonic relationship to coronary heart disease (CHD) deaths during the first 30 years of follow-up, but smoking and cholesterol tend to a lesser association afterward. The overall relation is attenuated by regression-dilution effects; use of anti-hypertensive medications probably influenced pressure associations over time, and smoking and cholesterol levels decreased with age.


Single measurements taken early in life hold predictive power for many years. This is separate from the issue of single measures taken at advanced ages, examined elsewhere. It is speculated that these characteristics measured in “middle life” represent cumulative experience and “damage” done which persist in effects and predictive value, despite later risk factor variation. (HB)


Menotti, A., Kromhout, D., Blackburn, H., Jacobs, D. and Lanti, M. 2004. ‘Early and late coronary deaths in the U.S. Railroad Study predicted by major coronary risk factors’. European Journal of Cardiovascular Prevention and Rehabilitation, vol. 11, pp. 382-388.

Taylor, H.L., Blackburn, H., Keys, A., Parlin, R.W., Vasquez, C. and Puchner, T. 1970. ‘Five-year follow-up of employees of selected U.S. railroad companies’. Supplement I to Circulation, vols. 41-42, pp. I-20-39.