ABSTRACT
The pharmacokinetic behavior of amikacin and predictive performance of Sawchuk-Zaske dosing method, have been prospectively evaluated in 31 [16 male, 15 female] critically ill septic patients of mean [ +/- SD] age of 58 +/- 23 years, mean ideal body weight of 59.6 +/- 6.4 kg, mean creatinine clearance of 52 +/- 21.5 ml/min, mean serum albumin of 3.1 +/- 0.5 mg/dl and median APACHE [acute physiology and chronic health evaluation] II score of 26 [with a range of 18 to 33]. In this cross-sectional study, critically ill patients who met the Bone criteria for spesis but had stable creatinine clearance [serum creatinine change <0.5 mg/dl of the baseline] received the ordered dose of amikacin in one hour infusions. Blood sample were collected 30 minute after the third dose, half an hour before the fourth dose which was 1.5 times of the predicted half life of amikacin after the third dose. Cirrhotic patients and patients with renal failure requiring any mode of dialysis were excluded. Vital signs were recorded at each time of blood sampling; serum Mg+, serum albumin and APACHE score were recorded at the time of the first blood sampling. Mean [ +/- SD] of the pharmacokinetic key parameters of amikacin in this population was as follow: Vd=0.390.045 l/kg; Ke=0.141 +/- 0.057 /h; half-life=5.7 +/- 2.06 h; Clearance=54.2 +/- 25.2 ml/min. There was a good correlation between Vd and serum albumin and also APACHE score II [r2=0.83, P=0.033;r2=0.82, P<0.001 respectively]. Mean measured peak and trough amikacin concentrations were 20.9 and 3.2 micro g/ml respectively which were significanthy different [P<0.05 paired t test] from levels, predicted by Sawchuk-Zaske method [33.5 and 4.6 micro g/ml respectively]. Ke, t0.5 and cledrane did not show any statistically significant changes [P>0.05 repeated measure test] amongst three times of blood sampling, but Vd was significanly different [P<0.05]. The overall predictive performance of Sawchuk-Zaske method was poor; in spite of good correlation between predicted and measured parameters when using pooled data