Gentamicin dosage schedules.

Sir, The paper by Yoshioka et al. (Archives, 1978, 53, 334), gives a guide to gentamicin dosage for an age group missed by the popular adult nomograms. Such a guide is necessary when choosing initial dosage, but we believe that subsequent maintenance therapy should be monitored by determining serum gentamicin levels in all patients rather than just in 'difficult' cases as suggested by the authors. The disadvantage of determining the half-life of gentamicin from creatinine clearance is that the latter should be corrected to the patient's lean body mass (Hull and Sarubbi, 1976). The lean body mass of an acutely-ill oedematous child with chronic renal disease is difficult to estimate. Gentamicin clearance is a sensitive indicator of renal function, and is logically best predicted by measuring serum levels of the drug itself. Sawchuck and Zaske (1976) described a method of calculating the half-life and distribution volume of gentamicin. We have adapted this method for a programmable calculator instead of the computer. Our experience suggests that the one-compartment model used in this method is perhaps too simple: the half-life derived from serum levels of gentamicin in the first 4 hours after a dose overestimates the subsequent rate of clearance and slightly underestimates the interval between doses. A better model could be made by fitting a 2or 3-term exponential function to later serum levels (Kahlmeter et al., 1978) but the one-compartment model does provide useful insight into the effects of varying dosage regimens. We question whether a dosage of 1 mg/kg is enough to treat serious infections, especially by IM injection. The peak serum levels achieved in the 3 patients treated by the schedule were all below 5 Cug/ml. The longer the dose takes to enter the circulation, the lower the interstitial fluid levels will be (Kozak et al., 1977). There appears to be no alternative to repeated measurement of serum gentamicin levels, and our efforts should be directed towards improving the accuracy and availability of our assays. Micromethods using small quantities of blood would be particularly valuable for very young children for whom venepuncture is often difficult or impossible. There is one factor in our favour: the interval between doses for many patients with renal failure is longer than our slowest bio-assay. R. T. MAYON-WHITE and ELIZABETH M. PERKS Regional Public Health Laboratory and Department of Bacteriology, Radcliffe Infirmary, Oxford