Interpersonal and temporal variability of urinary cotinine in elderly subjects.

Although a single measurement of urinary cotinine is often used for biological monitoring, the validity of this measurement for estimating long-term exposure has not been well evaluated. The purpose of this study was to determine the efficacy of using a single measurement per person to estimate long-term smoking exposure in an elderly population. Analysis was conducted using a cohort of elderly subjects for whom multiple urinary cotinine measurements had been collected over eight weeks. The mixed-effects model of urinary cotinine indicated that interpersonal variability was greater than temporal variability. The efficacy of using single measurements to track the mean long-term exposure of a population is supported by the relatively consistent population averages. The classification of high and low exposure groups using a single measurement or using eight measurements produced similar group distributions. When using cut-off levels of 20 and 100 μg cotinine/g creatinine, 9.3% and 5.6% of individual exposures were misclassified, respectively, when using a single measurement. Urinary cotinine can be a useful biomarker for characterizing exposure to tobacco smoke. Although the incidence of misclassification decreased with an increase in the number of measurements, a single measurement of the urinary cotinine level may accurately represent long-term exposure to tobacco smoke.