Statistical criteria for using short-term measurements as an index of 24-hour mean arterial pressure in unanesthetized unrestrained dogs.

ABSTRACT

This study assessed the statistical validity of short time-interval measurements as estimators of true 24 hour mean arterial pressure in unanesthetized, unrestrained dogs. 24 hour intra-arterial pressure recordings were obtained using a stable FM telemetry system. The 24 hour pressure measurements approximated a normal distribution whose variance was inversely related to the selected averaging interval. Given the variance of a normal distribution one can calculate the 95% confidence interval for any single random measurement. Conversely the number of random samples necessary to be within a prescribed confidence interval can be determined. In this study, the 95% confidence interval for a single, random 30 minute arterial pressure average was calculated to be 11.2 mmHg. Only 4.8 +/- 1.4% of 480 individual 30 minute arterial pressure measurements fell beyond this confidence interval. These outlying values were distributed throughout the 24 hour period. The data suggest that randomly chosen short time-interval measurements may be a valid index of true 24 hour mean pressure if the average variance of a population is known and confidence intervals are defined.