Therapeutic drug monitoring

Editor, – The stated therapeutic range for lithium of 0.6-1.2 in the article 'Therapeutic drug monitoring' (Aust Prescr 1997;20:9-11) may give a wrong message to non-psychiatrists who are managing patients on lithium.

Although patients with an acute manic episode may be given doses of lithium carbonate which produce levels of up to 1.2 mmol/L, the recommended serum levels in the maintenance phase have, for many years, been between 0.5and 0.8.1,2

I have had occasion in the past to inform clinical pathology laboratories of their out-of-date recommended range, which on one occasion was 0.8-1.5 mmol/L, the higher figure liable to result in lithium toxicity in some patients.

I treat over 120 patients with lithium, and in only one current patient do I need to keep the serum level above 0.8 mmol/L.

The issue is not an academic one, for adverse effects, and perhaps nephrotoxicity, are dose-related. Because the vast majority of your readers are not psychiatrists, they may be encouraged by your article to maintain patients on higher lithium doses than necessary, and at some cost.

David Grounds
Richmond, Vic.

Editor, – This article contains some figures which are not congruent with modern literature and practice in the area of antibiotic levels and therapeutic monitoring.

In Table 2, the following antibiotics and their therapeutic ranges are listed:

Gentamicin, tobramycin, netilmicin trough <2 * ; peak >5
Amikacin trough <5 2; peak >15
Vancomycin trough <10; peak 20-40
* for 8-hourly dosing

The administration of 8-hourly aminoglycosides has been recognised as being associated with nephrotoxicity, particularly in patients with impairment of glomerular filtration.

High peak levels of aminoglycosides have been shown to beassociated with a better therapeutic outcome and the recognition of the post-antibiotic effect associated with aminoglycoside levels below the MIC of the target antibiotic have led to a shift away from 3 times daily dosing with these drugs to once daily dosing of up to 7 mg/kg of the drug. This allows a high post-dose peak and time for excretion of the antibiotic to minimise nephrotoxicity.

The 9th edition of the 'Antibiotic Guidelines' recommends the aminoglycoside levels quoted for non-respiratory infections only with a recommended peak level for respiratory infections of >8 mg/L.

The post-dose amikacin peak level quoted by Professor Birkettof >15 mg/L is not congruent with the 9th edition of the 'Antibiotic Guidelines' which recommends >25 mg/L.

Amikacin is usually administered twice daily and not 3 times daily, so the information in the superscript in the Table above is also erroneous for this antibiotic.

Vancomycin is a bacterio static drug and, in contrast to the aminoglycosides, a trough level of between 5 and 10 or 15 mg/L must be maintained for efficacy. Allowing the vancomycin level to fall lower than 5 mg/L will ensure lack of efficacy of the treatment regimen.

I am surprised that in this important area of drug monitoring, there was not a more careful evaluation of the data quoted and a more thoughtful analysis of the need for monitoring and the appropriate levels to achieve a cost efficient and therapeutically effective outcome.

J.L. Faoagali
Director of Microbiology
Royal Brisbane Hospital
Herston, Qld

Editor, – I read with interest the informative article by Professor Birkett. He lists and gives a therapeutic range for a number of drugs the use of which 'without monitoring would be difficult and often dangerous'.

For many patients, vancomycin can now be removed from this list. Recent reviews have described the lack of toxicity of current formulations of the drug, and the lack of correlation between serum levels and either toxicity or clinical efficacy.3

Professor Birkett gives a therapeutic range for aminoglycosides, such as gentamicin, based on 8-hourly dosing. Such therapeutic ranges are meaningless for once daily dosing. A number of monitoring methods for once daily dosing have been developed in Australasia, including at least two computer-based methods('Aladdin' and 'Dosecalc') and one nomogram.4 These are generally based on performance of one or two sets of levels after the dose has been given, rather than the traditional 'peak and trough' collections. While theoretically attractive, these methods have yet to be compared with each other to determine which approach is most useful clinically and most appealing financially.

Finally, Professor Birkett omits tacrolimus (FK-506) from his list of essential drugs to monitor. Tacrolimus is an immunosuppressive agent, rapidly replacing cyclosporin as the key immunosuppressant for transplant recipients in Australia. There is a clear correlation between elevated tacrolimus levels and nephrotoxicity and neurotoxicity.5 While the correlation between low tacrolimus concentrations and acute rejection is less clear-cut, it has been demonstrated in at least one study.6

David L. Paterson
Infectious Diseases Physician
Department of Veterans' Affairs Medical Centre
Pittsburgh, Pennsylvania

Professor D. Birkett, the author of the article, comments: These letters reveal the complexities and sometimes lack of good data in the use of therapeutic drug concentration ranges. For example, Dr Paterson states that therapeutic drug monitoring for vancomycin is not useful, whereas Dr Faoagali considers that a 'trough level between 5 and 10 or 15 mg/L must be maintained for efficacy'. With the move to once daily dosing for aminoglycosides, the best method of monitoring to avoid toxicity is not clear, and the current uncertainties are well discussed in the 9th edition of the 'Antibiotic Guidelines'. As noted by Dr Paterson, a number of methods for monitoring of once daily dosing have been proposed on pharmacokinetic grounds, but the clinical relevance of these has yet to be established. It does seem clear that a trough concentration of 2 mg/mL for patients with normal renal function is too high, but the situation is less clear for patients with reduced renal function where the dose size and peak concentrations are reduced. Also, once daily dosing is not suitable for all situations as noted in the 'Antibiotic Guidelines'.

Dr Grounds has taken me to task over the therapeutic range for lithium. In preparing the article, I found as many therapeutic ranges for lithium as there are laboratories and reference sources. There are several general points that are relevant. The first is that reference ranges are often established by small studies with small numbers of patients and often many years ago. The extrapolation to the wider patient population needs some caution. The second is that drug concentrations need to be interpreted in the clinical context of individual patients - there is substantial interindividual variability in the relationship between drug concentration and therapeutic response or adverse effects. Finally, as noted in the article, the concentration-response relationship is not all or nothing, but is a continuum (see 'Pharmacodynamics- the concentration-effect relationship' Aust Prescr 1995;18:102-4). The therapeutic range represents an attempt to choose a drug concentration range which optimises the chance in a patient population of having a therapeutic effect with an acceptable incidence of adverse effects.

I thank your correspondents for their remarks which draw attention to the need to consider therapeutic concentration ranges as a guide to be interpreted in the context of the individual patient.


  1. Johnson G. The role of lithium in the affective disorders. Aust NZ J Psychiatry 1996;30:715-9.
  2. Jefferson JW. Lithium encyclopedia for clinical practice. 2nd ed. Washington, DC: American Psychiatric Press, 1987:600.
  3. Cantu TG, Yamanaka-Yuen NA, Lietman PS. Serum vancomycin concentrations: re-appraisal of their clinical value. Clin Infect Dis 1994;18:533-43
  4. Victorian Drug Usage Advisory Committee. Antibiotic Guidelines. 9th ed. Melbourne: Victorian Medical Postgraduate Foundation,1996.
  5. Backman L, Nicar M, Levy M, Distant D, Eisenstein C, Renard T, et al. FK506 trough levels in whole blood and plasma in liver transplant recipients. Correlation with clinical events and side effects. Transplantation 1994;57:519-25.
  6. Jain AB, Todo S, Fung JJ, Venkataramanan R, Day R, Bryant J, et al. Correlation of rejection episodes with FK506 dosage,FK506 level, and steroids following primary orthotopic liver transplant. Transplant Proc 1991;23:3023-5.