Most analyses of urine ion excretion and pH are based on 24-hour urine collections. However, fasting morning urine, collected in the morning after an initial void and a short fast, is easier to collect and may be sensitive to dietary intake. The objective of this study was to determine whether the within-subject dietary acid load is stable by testing the levels of urine pH and ion excretion (ie, calcium, chloride, magnesium, phosphate, potassium, sodium, sulfate) in fasting morning urine specimens collected at a 5-year interval. Stable variables provide the best utility as potential risk factors. The subjects were 200 randomly selected adults (mean age = 61.5 +/- 11.1 years) from the 420 subjects who donated baseline and 5-year urine samples in the Canadian Multicentre Osteoporosis Study. The samples were collected in the morning after an initial void and a wait of 2 hours, while subjects maintained a fast from the evening before. The intraclass correlation coefficient was calculated to characterize the level of agreement between the baseline and 5-year urine samples as an indirect measure of diet stability. The stability of the within-subject urine measures over 5 years ranged from fair to moderate, and none were ranked as substantially stable. This fair-to-moderate stability of fasting morning urine measures of the diet acid load indicates a limited ability of a single sample of fasting morning urine to estimate subjects' actual long-term urine composition.