Kingsley Coulthard (1), Ask Aabenhus (1), Kjersti Moen (1), Peter Baghurst (2), Craig Hirth (2), John Turnidge (3), James Martin (4). September, 2001. Distribution Pharmacokinetics of Once-daily Tobramycin in Children with Cystic Fibrosis [online]. Pharmacy Department (1), Public Health Research Unit (2), Department of Microbiology and Infectious Diseases (3), Department of Pulmonary Medicine (4), Adelaide Women's and Children's Hospital, South Australia, Australia. Available from http://www.cysticfibrosismedicine.com
The advantages, both theoretical and practical, associated with once-daily administration of aminoglycosides, are well known. The pharmacodynamic benefits may be even more relevant in the treatment of acute pulmonary exacerbations of cystic fibrosis (CF) given the characteristics of tobramycin against Pseudomonas aeruginosa and the airway environment associated with CF.
Given the relatively narrow therapeutic index and the need for adequate "peak" plasma concentrations, therapeutic drug monitoring (TDM) of aminoglycosides is recommended. The introduction of once-daily dosing has complicated the use of the "peak-trough" TDM traditionally associated with multiple-dosing per day of these drugs. The higher total body clearances in children, and even more so in patients with CF, mean that pre-dose concentrations will be well below the lower limit of sensitivity of routine laboratory assays for aminoglycosides.
An alternative to this traditional form of TDM is the area under the plasma drug concentration-time (AUC) method which is based on the premise that AUC reflects exposure of the organism to drug over time, and thus efficacy. A number of software programs are available to allow calculation of AUC. ALADDIN, developed by Prof. John Turnidge and Ms. Jan Bell, requires 2 samples to determine the AUC for that patient. Calculations of a number of pharmacokinetic parameters are performed on the assumption that the aminoglycoside follows mono-exponential pharmacokinetics.
The actual timing of the two samples needed for ALADDIN has a number of implications. Given the pharmacokinetic model used, the first post-dose sample should be taken after distribution of the drug is complete - traditionally thought to be over within one hour of the end of administration. The second sample should be taken at a time when it can be confidently assumed that the resulting plasma concentration will be above the lower limit of sensitivity of the assay. Practical factors such as convenient nursing and laboratory technician times also need to be considered. The Adelaide Women's and Children's Hospital (WCH) had adopted sampling guidelines of 1 hour and 6 hours post-dose.
A publication in 1997 suggested that the 1-hour sampling point was inappropriate since the higher, single doses used in once-daily therapy required longer to distribute. In 11 healthy adult volunteers, a 2mg/kg dose of gentamicin given over 1 hour reportedly had a mean distribution half-life of 21.8 minutes: 4-5 times this time is required for distribution to be complete (1). When these same patients received 7mg/kg, the average distribution half-life was 41.6 mins. Further support for the concept of a prolonged distribution came from Barclay et al when evaluating data from CF patients (2). If these conclusions were correct, then a 1 hour sampling point would be still within the distribution phase and thus programs such as ALADDIN would tend to overestimate AUC and underestimate the recommended dose.
The study of Bates et al in 1997 contradicted the above view. In a study of 18 CF patients given between 7-15mg/kg tobramycin as a single dose over 20 minutes, the derived pharmacokinetic data was compared to that from historical multiple daily dosing (3). This group concluded that distribution was complete by one hour.
Patients with CF admitted to the WCH for an acute pulmonary exacerbation are routinely treated with tobramycin given as a once-daily dose. Plasma concentrations obtained a 1 and 6 hours after the end of a 15 minute infusion are entered into ALADDIN to give a dose equating to an AUC of approx 100mg.hr/L. The majority of CF patients have implantable venous access devices such as a Port-A-Cath or a PICC line - given that the tobramycin is given via these and may contaminate samples for plasma level monitoring, TDM is performed on finger prick samples at 1 and 6 hours. The following study, with WCH Ethics Committee approval, was performed in 21 CF patients aged 6-16 years.
An additional (3rd) finger prick was taken on day 1 and day 8 of IV therapy - sampling was either at 1,3 and 6 hours post-dose or 2,4 and 6 hours post-dose. Each patient received both sampling regimens with that on day 8 being the alternative to that received on day 1. The times of administration and sampling were recorded independently of the nursing/laboratory staff by the investigators (AA,KM). The various 2-point sampling point concentrations (1&3, 1&6, 3&6, 2&4, 2&6, 4&6 hours) were entered into ALADDIN and the following pharmacokinetic data derived - elimination half-life, clearance, AUC, volume of distribution, plasma concentration extrapolated to time zero. Levene's test was used to compare the equality of variance between the 2 days: no statistically significant differences were found. The non-parametric Mann-Whitney U test and Wilcoxon's rank sum test were employed to compare differences in sample days for each pharmacokinetic parameter and sampling point combination.
No significant differences were found in the distribution of data. The differences between the means in sampling point combinations were compared by the Kruskal Wallis test. Significance was detected for the pharmacokinetic parameters: volume of distribution, elimination half-life, AUC and theoretical concentration at time zero. The day effect, the different patients and the use of different sampling point combinations were also analysed with a split-plot (box-plot) analysis of variance.
Elimination half-life
The split-plot analysis showed that there was no statistically significantly difference in the elimination half-lives between the two days. There were significant differences between patients and also between the different sampling combinations. This difference was statistically significant for the 1-3 h sampling combination.
Volume of distribution
The split-plot analysis showed that there was no statistically significant difference in the volume of distribution obtained on the two days. There were significant differences between patients and also between the different sampling combinations. There was a statistically significant difference between the 1-6 h combination and all other combinations except the 1-3 h combination. When the 1 h sample is used, the median volume of distribution is 0.08 L/kg lower than with the other sampling combinations.
Total body clearance
The split-plot analysis showed that there was no statistically significant difference in the total body clearance (adjusted for weight) obtained on the two different days. There were significant differences between patients and also between the sampling combinations. The clearance was significantly different when the 1-6 h sampling combination was compared to the other sampling combinations. These results demonstrate that the 1-6 h sampling combination gives a lower clearance than other sampling combinations.
Area under the concentration-time curve
The split-plot analysis of AUC adjusted for differences in dose between the two days for all patients showed that the adjusted AUCs were all significantly different from the 1-6 h sampling combination. The 1-6 h sampling combination was 13-18 mg.hr/L higher than sampling combinations that did not include a 1 h level.
Theoretical concentration at time zero
The box plot of the theoretical concentrations at time zero from the different sampling combinations showed that the 1-6 h sampling combination was significantly different from all other sampling combinations. Sampling combinations where both the levels were taken later than 1 h gave consistent median theoretical concentrations at time zero. With the 1-6 h sampling combination, the theoretical concentration at time zero is 6-9 mg/L higher than with the sampling combinations not including the 1 h sample.
Administration - sampling times
Accurate recording of drug administration and sampling times are critical components of TDM. The anecdotal view that recorded times are often "suspicious" was confirmed as per the following table based on investigator observation for the current study.
A syringe pump or similar device such as a Springfusor is routinely used for the 15 minute administration and, on the average, had an acceptable performance except in 3 of the 42 administrations where this was greater than 20 minutes.
Of greater concern are the large discrepancies between stated times of sample collection and that observed by the investigators. The 1-hour sample was collected anywhere from 25-68 minutes and the 2-hour sample anywhere between 49-142 minutes. It should be noted that, if it is assumed that distribution is complete at 2 hours, then a ± 5 minute discrepancy in this time results in a ± 3% variation in calculated AUC using ALADDIN. (The real (observed) times were used in the analyses for this study.)
Conclusion
In conclusion, we believe that our data confirms the earlier work of Demczar (1) and Barclay (2) that the higher doses of aminoglycoside given once-daily exhibit a prolonged distribution phase. Furthermore, since our study, Tsang et al have presented their work at the 2000 North American Cystic Fibrosis Conference in which they concluded that sampling should be delayed for 3 hours after a one hour infusion of once daily tobramycin in adults with cystic fibrosis (4). If post-distribution sampling is required for the particular TDM method used, then sampling should not occur within 2 hours of completion of administration of the dose. Although the clinical significance of an underestimated dose of approx 15% is likely to be minimal, there may be situations (eg research) when this variance may be relevant. Of more concern was the observation of significant variation between actual times relating to administration and/or sampling - these may significantly contribute to the results of TDM and lead to inappropriate dosing.
References:
1. Demczar DJ, Nafziger AN, Bertin JS. 1997. J Antimicrob Chemother 41: 1115-1119
2. Barclay ML, Robertshawe BJ, Pattemore P, Charles B. 1998: Australasian Society of Ethical and Experimental Pharmacologists Conference (Abstract)
3. Bates RD, Nahata MC, Jones JW, McCoy K et al. 1997. Chest 112: 1208-13 4. Tsang L, Aminimanizani A, Beringer PM, Jelliffe R. 2000. Pediatr Pulmonol Supl 20: 284
Copyright © cysticfibrosismedicine