RESULTS
Table 1 shows the type 3 tests of fixed effects. The type 3 tests of hypotheses show differences (P < .0001) in the means of NERL adjusted plasma tCO2 values across time. Differences were also seen (P = .0016) in the means of NERL adjusted plasma tCO2 values ac- cording to furosemide dose. No significant differences
were found (P > .05) in the means of NERL adjusted plasma tCO2 values between weeks. However, a signifi- cant difference was found (P < .0001) according to treatment ϫ hour, suggesting that the change over time was influenced by the dose of furosemide given.
Table 2 shows the estimated means of NERL ad- justed values for plasma tCO2 by hour for 0, 250, and 500 mg furosemide doses, the pooled standard error (SE) and lower (mean – 1.96 ϫ SE) and upper (mean + 1.96 ϫ SE) confidence limits for the plasma tCO2 val- ues.The elevation of plasma tCO2 after the administra- tion of either 250 or 500 mg furosemide is apparent. A reduction in plasma tCO2 appears at all dosage levels, including the control, by hours 5 or 6 after treatment, which may suggest some cyclic variation in plasma tCO2 throughout the day in addition to the expected attenu- ation of the furosemide effect over several hours.
Table 3 shows the least squares means estimates for the NERL adjusted values for plasma tCO2 including all sampling times. Table 4 shows the comparisons of least squares means differences between each dose of furosemide using the Tukey adjustment for multiple comparisons. The adjusted P-values for differences of means show significant differences between 0-mg and 250-mg doses of furosemide (P = .022) and between 0- mg and 500-mg doses of furosemide (P = .0014); how- ever, no significant difference was seen between 250-mg and 500-mg doses of furosemide (P = .4308