1 The hypothesis that a logarithmic correlation exists between the plasma theophylline concentrations 6 h after a test dose C(6) and the maintenance dose calculated to achieve a desired drug concentration during chronic oral administration (DM,CALC) was tested. A nomogram based on this relationship was evaluated as a means of predicting the optimum dose of theophylline in 14 patients with obstructive airways disease (DM,PMRD).2 Each patient was given 5 mg/kg theophylline intravenously (i.v.) and plasma theophylline concentrations were measured for 12 h thereafter including one exactly 6 h after commencing the i.v. infusion C(6). Pharmacokinetic parameters derived from the plasma concentration-time curve were used to establish DM,CALC for a concentration of 10 ,ug/ml. DM,PPED was obtained from the nomogram using both the optimum and the actual values for C(6). Subsequently oral sustained-release theophylline was administered and the dose adjusted to establish a trough concentration of approximately 10 ,ug/ml. This dose was then corrected arithmetically, assuming a linear relationship between dose and plasma level, to represent that required to achieve a plasma concentration of exactly 10 ,ug/ml (DM,AC-)- 3 The correlation between C(6) and log DM,CALC was confirmed (r = 0.97 P < 0.001), validating the hypothesis. DM,ACT was found to correlate significantly with DM,PRED (r = 0.90, P < 0.01) substantiating the value of the nomogram. In nine of the 14 patients, DM,ACr corresponded satisfactorily to DM,PRED. In the remaining five, for whom DM,ACT lay outside the 95% confidence limits for the predicted dose, DM,PRED in general underestimated DMACT, an advantage in a drug with a low therapeutic index. The predictive error for DM,PRED was lower than that for DM,CALC and the bias using either method was not significant. 4 The results suggest that a single plasma theophylline assay following a test dose, and the nomogram may be useful in simplifying optimal theophylline administration.