|dc.description.abstract||Theophylline is a dimethylated xanthine similar in structure to caffeine which is commonly found in tea, coffee and cola beverages (Hendeles and Weinberger, 1983; Rall, 1985). Clinically, its most important pharmacological action is the ability to relax bronchial smooth muscle throughout the bronchial tree (Persson, 1986). This effect has found extensive use in the treatment of asthma with the drug being recommended as the first line agent for chronic asthma (la/rate et ai, 1986). The observation that both beneficial effects as well as toxicity correlate with serum concentrations and that the drug displays a narrow therapeutic window (Finn et al, 1981; Hendeles and Matthay, 1986) has resulted in the recommendation that theophylline dosing be guided by serum concentration measurements (Hendeles and Weinberger, 1980; Whiting et al, 1984; Fitzpatrick and Moss-Barclay, 1985; Barlow et. al, 1988). However, this recommendation appears to have been largely ignored locally. In 1986, one of the first local Therapeutic Drug Monitoring Clinics for theophylline was established at R K Khan Provincial Hospital in Chatsworth, Durban. Preliminary results from this clinic confirmed the widespread use of standard theophylline dosing regimens and revealed that 68% (n = 44) of patients given these regimens had serum theophylline concentrations below the generally accepted therapeutic range (Pillai and Miller, 1988).
Previous studies have assessed the influence of Therapeutic Drug Monitoring programmes in terms of the attainment of 'therapeutic' serum concentrations (Whiting et aI, 1984; Fitzpatrick and Moss-Barclay, 1985). This approach has been criticised and it has been recommended that clinical assessment should be the criterion. The purpose of this study was to investigate the influence of serum concentration monitoring on theophylline utilisation at the R K Khan Hospital in terms of clinical control of asthma symptoms. A secondary purpose of this study was to determine population pharmacokinetic parameters in Indian patients. In order to interpret the serum concentrations and make recommendations on dosage design for individual patients, the Bayesian technique of drug dose optimisation is used (Sheiner et aI, 1972). This technique has been shown to be accurate, precise and easy to use (Sheiner and Beal, 1982; Hurley and McNeil, 1988) particularly with currently available computer software. It has been emphasised, however, that for satisfactory performance of this technique, good initial estimates of the population parameter distributions are important (Whiting et al, 1986). Since this information is not available for the Indian population this study was undertaken. A knowledge of population pharmacokinetics can help one to choose initial dosage, to modify dosage appropriately in response to observed drug levels, to make rational decisions regarding drug regulatory requirements and toinvestigate and elucidate certain research questions in pharmacokinetics (Sheiner, 1984). The NONMEM approach (Sheiner et aI, 1972; 1977), currently the mostsatisfactory method of population pharmacokinetic data analysis is utilised in this study.||en