- Email: [email protected]
Monitoring of serum aminoglycoside levels with once-daily dosing
Pathology (1998) 30, pp. 289-294
MONITORING OF SERUM AMINOGL YCOSIDE LEVELS WITH ONCE-DAILY DOSING DAVID L. PATERSON*t, JENNIFER M.B. ROBSON*, MARILYN M. WAGENERt AND MICHAEL PETERS*
Sullivan Nicolaides Pathology, Taringa, Queensland, Australia*, and Infectious Disease Section, VA Medical Center, Pittsburgh, Pennsylvania, USA t
Summary There is considerable confusion as to how to monitor serum aminoglycoside levels when using once-daily dosing. At least five methods are in use in Australia. We prospectively assessed 100 consecutive once-daily courses of gentamicin or tobramycin, during which 120 pre-dose and 213 sets of immediate post-dose and six hour post-dose levels were taken. By using the six hour post-dose level we were able to compare dosage recommendations made using methods known as ALADDIN, DOSECALC and the Australian Antibiotic Guidelines nomogram (AAGN). There were statistically significant differences in the doses recommended by each method. When comparing each of the three methods, at least 25% of dosage recommendations differed by more than 80 mg per dose. Although we have not been able to determine the clinical significance of these differences, we are concerned that methods used in dosage adjustment of aminoglycosides differ so widely in their recommendations. Presumably the ALADDIN method, which utilises two post-dose levels to determine an area under aminoglycoside concentrationtime curve, gives more accurate pharmacokinetic information than methods which rely on a single level. Comparative cost-effectiveness studies of different methods, although in practice difficult to perform, should be undertaken to resolve the optimal management of patients receiving aminoglycosides once-daily. Key words: Aminoglycosides, monitoring, pharmacokinetics, gentamicin, tobramycin. Abbreviations: AAGN, Australian Antibiotic Guidelines nomogram; ALADDIN, aminoglycoside levels and daily dose indicator; AVC, aminoglycoside concentration-time curve.
Accepted 5 February J998
INTRODUCTION Once-daily aminoglycoside dosing is as effective as, and probably results in a lower risk of nephrotoxicity than, multiple-daily dosing. 1- 3 The theoretical reasons why oncedaily aminoglycoside dosing is more useful include the findings that it takes advantage of the prolonged postantibiotic effect seen with high peak aminoglycoside concentrations and that it utilises the concentration dependent killing of this class of antibiotics. In addition, the reduction in nursing time required in its preparation and administra-
tion, and the reduced variability in the timing of doses, make once-daily dosing the method of choice for prescription of aminoglycosides. The applicability of traditional peak and trough aminoglycoside concentrations to once-daily dosing has been questioned by several authors.4-6 The following reasons have been offered as to why monitoring trough concentrations is inappropriate with once-daily dosing: (i) if the initial trough level is taken 24 hours after the initial dose (ie; just prior to the second dose), no alteration in dose will be made until the third dose (ie; 48 hours after the first dose); (ii) there has been no correlation made between trough levels and toxicity or effectiveness with once-daily dosing; (iii) patients with normal aminoglycoside pharmacokinetics would not be expected to have a trough concentration exceeding 0.05 mg/l at 24 hours, yet the lower limit of sensitivity of assays such as Abbot TDx is 0.27 mg/l, thus fivefold accumulation could be occurring before this would become detectable. 4 Similarly the usefulness of peak concentrations with once-daily dosing is doubtfu1. 6 Conventional target peak concentrations (eg; 6-10 mg/I) do not apply to once-daily dosing,4 and there have been few relationships made between efficacy or toxicity and peak concentrations with once-daily dosing. An alternative approach to using peak and trough levels has been to utilise the area under the 24 hour serum aminoglycoside concentration-time curve (AUC).4 This can be done by taking two levels after the end of an aminoglycoside infusion and using these in calculating the AUC. A dose recommendation can be made by comparing this measured AUC with a target AUC that would be expected in a patient with mean population values of volume of distribution and half-life of elimination. While the calculation can be done manually, it has been packaged into a software program developed in Australia by John Turnidge and Jan Bell and known as ALADDIN (Aminoglycoside Levels and Daily Dose Indicator).7 The ALADDIN method also aims for high peak levels to reduce the risk of selection for resistance. 7 A second Australian software, developed at Royal Perth Hospital and known as DOSECALC, utilises a single serum aminoglycoside level, six to 14 hours after the first dose to calculate the AUC. Instead of measuring an early post-dose level, a presumed concentration is derived based on the dose and known parameters such as calculated creatinine clearance. Finally, a simple graphical method for patients with normal renal function has been developed. 4 This method is based on a constructed serum aminoglycoside concentration curve calculated on the maximum AUC which
11031-3025/98/030289-06 © J998 Royal College of Pathologists of Australasia
290
PATERSON et al.
would have been permitted using a "conventional" aminoglycoside regimen that resulted in a peak of 10 mg/l and a trough of 2 mg/l. A single serum aminoglycoside concentration is determined between six and 14 hours post-dose and plotted onto a graph containing the curve below which drug accumulation would be unlikely. This method appears in a modified form in the of Australian Antibiotic Guidelines (1996/1997). 8 The aim of this study was to evaluate prospectively the ALADDIN software program in the management of a group of patients who were receiving once-daily aminoglycoside therapy and who were at high risk of nephrotoxicity. At the same time, since there are other methods of monitoring aminoglycoside use, we aimed to compare whether the doses predicted by the different methods were similar. In particular, we aimed to determine whether methods using just one blood level (which would be less expensive for laboratory services and more convenient for patients and health care professionals) predicted a dose comparable to the ALADDIN method which requires two levels to be drawn. Thus we compared the doses predicted by ALADDIN with those predicted by DOSECALC and the graphical method published in Antibiotic Guidelines. 8
MATERIALS AND METHODS Residents, registrars and visiting medical officers at Greenslopes Private Hospital, Brisbane, were given education (grand rounds, several written letters) regarding the benefits of once-daily aminoglycoside dosing in August and September, 1996. They were encouraged to use the following initial dose of either gentamicin or tobramycin from 25 September, 1996: Age 10--29 yrs 30--60 yrs Over 60yrs
Dose 6mg/kg 5 mg/kg 4mg/kg
The recommendations did not apply to patients in the Intensive Care Unit and Renal Unit, and patients in these units were not evaluated in this study. Nursing staff were asked to give the aminoglycoside dose over 30 min. Blood collectors were advised to perform 2 levels after the first dose of aminoglycoside. The first level was to be taken immediately the infusion ended. The timing of this first level was regarded as critical. A second level was to be performed about 6 h later. The timing of this second level was not regarded as critical, although the precise time of collection was to be recorded. These two levels were analysed by TDx (Abbott Laboratories, Illinois). The blood sample obtained for the first level was utilised for a serum creatinine level at the start of treatment ("baseline creatinine"). Demographic details (age, sex, height, weight) plus the serum creatinine and 2 serum aminoglyeoside levels were entered into the ALADDIN software program. The dosage and dosing interval recommended by this program were communicated to the clinician by way of the laboratory report containing the aminoglycoside levels. This information was available well in advance of the second dose. A recommendation was made to collect blood for serum creatinine and the 2 serum aminoglycoside levels after the third, fifth and seventh doses. In addition, in some patients a trough sample was collected immediately prior to these doses. The result of the trough level was added to the laboratory report but was not utilised by the authors in the recommendation of dose or dosing interval. Clinicians were however free to use this information in adjusting dose. The compliance of clinicians with the dosage and dosing interval on the laboratory report was recorded by way of a review of medication sheets for each patient. We advised use of an alternative antibiotic after 7 d of aminoglycoside use. If the aminoglycoside was continued we recommended monitoring of serum creatinine and the 2 serum aminoglycoside levels after every second dose. The serum aminoglycoside level obtained approximately 6 h after the
Pathology (1998), 30, August dose was utilised in dose recommendations by the DOSECALC program and the Australian Antibiotic Guidelines nomogram (AAGN). This information was not utilised in the recommended dosage and dosage interval communicated to the clinician by way of the laboratory report. Its only use was in a subsequent analysis comparing doses recommended by the 3 systems described in this report. Patient demographics and laboratory data were entered into Excel (Microsoft). The percentage of patients developing nephrotoxicity was recorded using a standard definition for drug-induced nephrotoxicity (defined as an increase in serum creatinine over baseline of greater than 0.04 mmolJl).2,9 Vestibular toxicity was defined as vertigo and gait instability with no other identifiable cause found. to Ototoxicity was defined as a 15 dB change in hearing at any frequency, as assessed by audiometry.2 Audiometry was only performed at the discretion of the patients's clinician. Efficacy of the aminoglycoside course was defined as a successful resolution of the infection without need to substitute another antibiotic. Dose recommendations made with ALADDIN, DOSECALC and the AAGN were compared by using Wilcoxon's signed-rank test for paired data. The effect of compliance with dosage recommendations on incidence of nephrotoxicity was compared by using the chi-squared test. Student's I-test was used to determine correlation between peak levels and nephrotoxicity.
RESULTS The ALADDIN program One hundred consecutive courses of aminoglycoside, where doses were calculated using ALADDIN, were evaluated from 25 September to 30 November, 1996. Ninetyfour patients were included in this evaluation-six patients had two courses of aminoglycoside in the study period. Tobramycin was used in 18 of the courses and gentamicin in 82. The mean course duration was 7.5 days (range one to 20 days). The study population was predominantly male (79% male, 21 % female) and elderly (mean age 74.9 years, range 33-91). Indications for aminoglycosides included infective exacerbations of chronic obstructive airways disease or bronchiectasis (41 %), pneumonia (17%), urinary tract infection (10%), peritonitis/intra-abdominal abscess (6%) and others (26%). Thirteen percent of the courses of aminoglycoside were accompanied by nephrotoxicity and four percent by otovestibular toxicity. The proportion of patients who developed oto-vestibular toxicity may be an underestimate; patients were assessed for vestibular toxicity as clinically appropriate and only two patients underwent audiometry (one of whom was found to have a significant reduction in hearing). The study population was at high risk for aminoglyco side toxicity. Sixty-two percent were on other nephrotoxic drugs during the aminoglycoside course (30% on two or more nephrotoxic drugs). These included vancomycin 15%, non-steroidal anti-inflammatory drugs 21 %, diuretics 32%, angiotensin-converting enzyme inhibitors 21 %, allopurinol 12% and amphotericin 2%. Eighteen percent received intravenous contrast radiological agents during the aminoglycoside course. Fourteen percent had significant hypotension (systolic blood· pressure less than 80 mm Hg) during the aminoglycoside course. The study population had generally suboptimal baseline renal function; 36% of patients had a baseline serum creatinine of more than 0.12 mmol/l. The mean baseline creatinine clearance of the study patients was 49 ml/min. The 13 patients who developed nephrotoxicity were older than those who did not (mean age 80.7 versus 75.9
AMINOGLYCOSIDE MONITORING
years, p 0.02), received a longer course of aminoglycoside (mean duration of aminoglycoside course 11.2 versus 7.0 days, p = 0.01) and were more likely to have received allopurinol concomitantly with the aminoglyco side (31% versus 9%; p 0.048). Patients who developed nephrotoxicity were also more likely to have elevated baseline serum creatinine measurements (0.15 versus 0.12 mmol/l; p = 0.09) and have developed significant hypotension (systolic blood pressure less than 80 mm Hg) during the aminogJycoside course (31 % versus 12%; p = 0.08), but these associations were not statistically significant. There was no statistically significant association between nephrotoxicity and the following variables: sex, type of aminoglycoside and concomitant therapy with vancomycin, non-steroidal anti-inflammatory agents, diuretics, angiotensin converting enzyme inhibitors or intravenous contrast agents (p > 0.20 for all). Two hundred and thirteen sets of levels were performed in the study period. Compliance with the recommendations made on the basis of the "ALADDIN" program was 87.5%. There was a statistically significant association between noncompliance with the dose recommendations made using these levels and nephrotoxicity. In seven of the 13 aminoglycoside courses which resulted in nephrotoxicity there was at least one instance of non-compliance with the dosage recommendations by the medical team caring for the patient, as compared to 15 of the 87 aminoglycoside courses which did not result in nephrotoxicity (p == 0.005). Thus those whose medical attendants were non-compliant at least once with the recommendations were at 4.1 times greater risk of toxicity (95% confidence intervals: 1.511.1). This risk of nephrotoxicity rose to 5.9 times greater (95% Cl: 2.4-14.4) if there were two or more episodes of non-compliance. In each of the seven courses in nephrotoxic patients in which non-compliance with the dosage recommendations occurred, a recommendation to lengthen the dosing interval was made but was ignored. In contrast, in 75% of cases of non-compliance with dosing recommendations in patients who did not subsequently develop nephrotoxicity, clinicians prescribed an aminoglycoside dose lower than that recommended by the laboratory. Clinical cure was documented in 85% of the aminoglycoside courses. In 6% of cases the aminoglycoside was ceased because of nephrotoxicity, before clinical cure was observed. In 7%, the aminoglycoside was changed because of failure to respond-in three of these seven cases there was no microbiological evidence found of an infection against which aminoglycosides were microbiologically active. Of the four cases of antibiotic failure, two had infective exacerbations of chronic obstructive airways disease, one had pneumonia and one had cellulitis. Two patients (2%) died during their aminoglycoside courseone had sepsis with methicillin resistant Staphylococcus aureus and the other had autopsy proven pulmonary thromboembolism.
291
1 Distribution of "peak" aminoglycoside serum concentrations in the study population (213 "peak" levels in 100 antibiotic courses)
TABLE
Peak serum concentration (mg/l) Less than 8 8-11.9 12-15.9 16-19.9 Greater than 20
Frequency (%) 5 37 73 60 38
(2.3%) (17.4%) (34.3%) (28.2%) (17.8%)
8 mg/I-these occurred in five separate patients, all of whom were under-dosed on a mg/kg basis (eg; a 102 kg, 33 year old man was given a dose of 1.6 mg/kg). When only the "peak" level taken after the first dose of aminoglycoside was considered, the mean first peak level was 17.2 mg/l in those who subsequently developed nephrotoxicity and 15.8 mg/l in those who did not (p 0.31). All 13 patients who developed nephrotoxicity had their first peak greater than 12 mg/l, as compared to 70/87 (80%) who did not (p = 0.12). The DOSECALC program The second post-dose level was used to predict a dose using the DOSECALC program. This dose was compared to the dose predicted by the ALADDIN program using both pre- and post-dose levels. The comparison between the doses predicted by the two programs is given in Fig. 1. The mean dose predicted by DOSECALC was' 207 mg, as compared to 210 mg predicted by ALADDIN. Sixty of the 213 (28.2%) levels of the predicted doses were identical when using the two programs. However 55 of the 213 (25.8%) levels of the predicted doses differed by 80 mg or more. Using Wilcoxon's signed-rank test for paired data, there was a statistically significant difference between the doses recommended by each method (Z= 3.18852, 0.001).
p
When only the 13 patients who developed nephrotoxicity were considered, the differences between the methods became even more pronounced. Sixty percent (21/35) of the dosage recommendations differed by 80 mg or greater. In 71 % (25/35) of dosage recommendations in patients 30 (f)
25
'0
~
20
~
15
0..
10
'5
1£
'0
c:
e (!)
5
(!)
Peak aminoglycoside levels The serum aminoglycoside level taken between the end of the infusion and 30 minutes post-infusion, equates to a "peak" level as measured in the traditional "peak and trough" style monitoring of aminoglycosides. Two hundred and thirteen such levels were taken. Their distribution is given in Table 1. Only 5/213 (2.3%) levels were less than
0...
Doses (in mg) predicted by DOSECALC less than those of ALADDIN
Doses (in mg) predicted by DOSECALC greater than those of ALADDIN
Fig. I Comparison of the doses (in mg) recommended by ALADDIN and those recommended by DOSECALC.
292
Pathology (1998), 30, August
PATERSON et al.
40
TABLE 2 Methods available for monitoring once-daily aminoglycosides
(/)
3l
.g
Traditional trough and pe,\\ measurements Trough measurement only ALADDIN program? DOSECALC program Australian Antibiotics Guidelines Nomogram 8 Nicolau Nom0l3.raml6 Prins schedule
30
"'0 Q)
t5
'g
0.
c Q)
~
20 10
Q)
a..
Doses (in mg) predicted by AAG less than that of ALADDIN
Doses (in mg) predicted by AAG greater than those of ALADDIN
Fig. 2 Comparison of the doses (in mg) recommended by ALADDlN and those by the AAGN.
who became nephrotoxic, DOSECALC predicted a lower dose than ALADDIN.
The Australian Antibiotic Guidelines Nomogram The second post-dose level was used to predict a dose using the AAGN. This predicted dose was compar:ed to the dose predicted by the ALADDIN program usmg both pre-and post-dose levels in 185 cases (Fig. 2). The mean dose predicted by the AAGN was 249 mg ~s compared to 210 mg predicted by ALADDIN. Forty-D1~e o~ the 185 (27%) cases of the predicted doses were Identlcal when using the two programs. However 29.2% (54/185) of the doses predicted by AAGN differed by 80 mg or more fr?m those predicted by ALADDIN using the same data. Usmg WiJcoxon's signed-rank test for paired data, there was a statistically significant difference between the doses recommended by each method (Z = 8.77422, p = 0.0001). Even when the use of the AAGN was restricted to patients whose creatinine clearance was greater than 60 ml/min, there was still a statistically significant difference between the doses recommended by each method (p < 0.01). When only patients who developed nephrotoxicity were considered, 28.6% (10/35) of the dosages recommended by ALADDIN and AAGN differed by 80 mg or more. In only two cases (5.7% of the 35 doses in patients who became nephrotoxic) was the dose recommended by AAGN lower than that recommended by ALADDIN. In patients who became nephrotoxic, the mean dose recommended by ALADDIN was 168 mg, as compared to 194 mg recommended by AAGN. Finally, when doses predicted by DOSECALC an~ the AAGN were compared, 20/185 (10.8 %) of the pred.lcted doses were identical and a statistically significant dlfference between the doses recommended by each method was observed (Z = 7.29393, p = 0.0001). DISCUSSION There are no clear recommendations available regarding how to monitor aminoglycosides when using once-daily dosing. An ideal method would be o~e which. requires the fewest blood collections (avoiding mconveDlence to the
patient and minimising costs to the blood c?l!e~tion servi:e and laboratory) and which allows the chmcIan to aVOId toxicity and provide a cure for the infection being treated. Numerous methods are in use in Australia at the present time (Table 2). The only way to determine which meth?d is closest to the ideal would be to perform a controlled trIal randomly allocating monitoring of aminoglycosides to each different method and determining which method could deliver the most cost-effective service (minimising toxicity and cost, and maximising clinical cure). Such a study would require a sample size unobtainable for most la?oratories and would be logistically very difficult to coordmate. In the absence of being able to perform such a trial, the aim of this study was to determine whether each of the three methods in common use (ALADDIN, DOSECALC and AAGN) gave a similar dose recommendation. If methods utilising one blood collection gave the same predicted dose as those utilising two blood collections, it would appear that methods needing a single blood collection would be more cost-effective and should be recommended for use with once-daily dosing. We found that the three different monitoring methods were not equivalent. Both methods using a single blood collection (DOSECALC and AAGN) gave a significantly different dose recommendation from that recommended by utilising two levels and the ALADDIN program. W,e recognise that a statistically significant difference IS not necessarily clinically significant. However we found that for at least 25% of the levels, a dose was recommended by either DOSECALC or AAGN which was at least 80 mg different from that recommended by the ALADDIN program. We have no way of determini~g whether .s~ch a difference in dose would increase likelIhood of tOXICIty or decrease chance of clinical cure, however we feel that it is likely to be clinically important. We would expect that the use of two levels compared to the use of one would give more accurate pharmacokinetic information and would therefore be advantageous in the care of the patient. We found that when using the ALADDIN program to direct once-daily aminoglycoside dosing, 13% of patients developed nephrotoxicity and 4% developed vesti?u.locochlear toxicity. The percentage of patients recelvmg once-daily dosing who develop nephrotoxicity has ranged in previous studies from 0 to 26.7%, I1 with the two ~~st recent studies of once-daily dosing showing nephrotOXICity in 14-24% of patients. 12.l 3 Both of these recent studies dealt with a vounger popUlation with substantially better baseline renal function than our own. 12•13 Both of these demographic details are demonstrated Ii~k factors f~r nephrotoxicity.14 The patients in these St~dl~S wer~ momtored utilisinG traditional methods. It IS ImpOSSible to compare pop~lations in different studies, but we believe that the 13% incidence of nephrotOXicity in our popu-
AMINOGLYCOSIDE MONITORING
lation, which had numerous risk factors for renal dysfunction, is at least no worse than that observed with traditional monitoring methods. FUlthermore, we observed that noncompliance with the recommendations of ALADDIN was significantly associated with development of nephrotoxicity. Likelihood of clinical cure was high. We feel therefore that this program has demonstrated usefulness in the management of patients on aminoglycosides. Despite the pleasing results of this program in the monitoring of aminoglycosides, we found some drawbacks with the ALADDIN system. The need for two blood collections and in particular, a second blood collection six hours after the aminoglycoside infusion, requires close coordination between nursing staff and blood collectors. A dose administered after 6 p.m. necessitates waking the patient for the second level six hours later. Although not absolutely required in the prediction of a dose, the desirability of requirement for information on height, weight and serum creatinine may be rarely provided by busy clinicians. We avoided this problem by asking the blood collectors to provide this information and by performing a creatinine assay on the serum sample submitted for aminoglycoside testing. As our laboratory report included a dosage recommendation, each report was reviewed by a medical microbiologist before being sent to the clinician. While this provided no difficulty during working hours, the issue of medical coverage on weekends and afterhours needs to be addressed when using the ALADDIN system. We found that liaison with laboratory scientists by way of telephone and modem after hours was generally sufficient. While use of a monitoring system based on a 24 hour serum aminoglycoside concentration-time curve is theoretically attractive, the use of trough and sometimes peak measurements appears to be widespread, with once-daily dosing in many institutions in Australia. Recent reviews have suggested that monitoring of trough concentrations has never been demonstrated to improve the outcome over empiric dosing regimens with dosage adjusted if a change in renal function occurs.5 ,14,15 A number of authors have suggested restricting monitoring of serum concentrations to only those patients with a creatinine clearance of below 50 ml/minute and then only performing trough levels. 12,16,17 Cost savings from this strategy of more than $100,000 per year were expected in one 850 bed hospital. 17 We found no correlation between peak serum aminoglycoside concentration after a dose and development of nephrotoxicity. This is in contrast to a recent study which showed that those who developed nephrotoxicity had a significantly higher initial serum peak concentration than those who did not develop this complication. 13 These authors suggested that an initial peak concentration of greater than 12 mg/l was predictive of nephrotoxicity. While we found that all patients who developed nephrotoxicity had a first peak of greater than 12 mg/l, 80% of those who did not develop nephrotoxicity also had levels greater than 12 mg/l and the association between nephrotoxicity and a first peak of greater than 12 mg/l did not reach statistical significance. Once-daily dosing using the dosage recommendations in this study, gave peak levels greater than 8 mg/l in more than 97% of cases. This is reassuring as at least one study has found a correlation between successful treatment of serious Gram-negative infections and peak levels greater than 8 mg/l.IS If a laboratory does not use a monitoring system based on AUC calculations (such as ALADDIN), it
293
may be prudent to check a peak concentration after the first dose is given, particularly if underdosing is frequent in their institution. An appropriate target range could be between 8 and 20 mg/l. We consider a target peak of 8-12 mg/l as too restrictive, particularly considering a range of less than 12 mg/l is based on only a single small study.13 There has never been a correlation between a peak level other than the first peak level and toxicity, and hence there is no indication to continue performing any more than one peak level. In conclusion, there are currently numerous methods available for the monitoring of aminoglycosides prescribed once per day. While monitoring of peaks and troughs has continued in many laboratories, there are theoretical reasons why this is not as appropriate as using AUC methods. If peaks and troughs are measured, the reference ranges should be different from those advocated previously when multiple-daily doses were given. We have found that the three methods currently available which use AUC differ significantly in their dosage recommendations. The methods are clearly not interchangable. We are unable to conclude whether one method is superior over another, but we have found a reasonably low rate of nephrotoxicity in high-risk patients monitored using the ALADDIN method. Furthermore, there was a close correlation between noncompliance with the doses recommended by this method and development of nephrotoxicity. Large randomised controlled trials would be necessary to resolve the question as to whether one method is more cost-effective and clinically superior to another, or if indeed levels need to be performed at all in most patients. ACKNOWLEDGMENTS The authors gratefully acknowledge the assistance of the medical and nursing staff at Greenslopes Private Hospital (Qld) and of the venepuncture and laboratory staff at Sullivan Nicolaides Pathology (Qld). Alasdair Miller is thanked for providing a free copy of DOSECALC software and its subsequent revision. Address for correspondence: D.L.P., Infectious Disease Section, VA Medical Center, University Drive C, Pittsburgh, PA 15240, USA.
References 1. Barza M, Ioannidis JPA, Cappelleri JC, Lau 1. Single or multiple daily doses of aminoglycosides: a meta-analysis. Br Med J 1996; 312: 338-45. 2. Hatala R, Dinh T, Cook DJ. Once-daily aminoglycoside dosing in immunocompetent adults: a meta-analysis. Ann Intern Med 1996; 124: 717-25. 3. Munckhof WJ, Grayson ML, Turnidge JD. A meta-analysis of studies on the safety and efficacy of aminoglycosides given either once daily or as divided doses. J Antimicrob Chemother 1996; 37: 645-63. 4. Begg EJ, Barclay ML, Duffull SB. A suggested approach to oncedaily aminoglycoside dosing. Br J CUn Pharmac 1995; 39: 605-9. 5. Cronberg S. Simplified monitoring of aminoglycosides. J Antimicrob Chemother 1994; 34: 819-27. 6. Konrad F, Wagner R, Neumeister B, Rommel H, Georgieff M. ~tudies on drug monitoring in three and once-daily treatment with aminoglycosides. Intensive Care Med 1993; 19: 215-20. 7. Turnidge J, Bell J, Mascaro L. The peak-AUC method for aminoglycoside dosage adjustment. Program and abstracts of the 36th Interscience Conference on Antimicrobial Agents and Chemotherapy, New Orleans, September 1996. 8. Victorian Drug Usage Advisory Committee. Antibiotic guidelines. 9th edn. North Melbourne: Victorian Medical Postgraduate Foundation Inc., 1996. 9. Smith CR, Lipsky 11, Laskin OL, Hellmann DB, Mellits ED, Longstreth J, et al. Double-blind comparison of the nephrotoxicity and auditory toxicity of gentamicin and tobramycin. N Engl J Med 1980; 302: 11 06-9.
294
PATERSON et al.
10. Longridge NS. Approach to the patient with dizziness and vertigo. In: Kelley WN, editor. Textbook of Internal Medicine. 3rd edn. Philadelphia: Lippincott-Raven, 1997. 11. Blaser J, Konig C. Once-daily dosing of aminoglycosides. Eur J Clin Microbial Infect Vis 1995; 14: 1029-38. 12. Prins JM, Weverling GJ, De Blok K, Van Ketel RJ, Speelman P. Validation and nephrotoxicity of a simplified once-daily aminoglycoside dosing schedule and guidelines for monitoring therapy. Antimicrob Agents Chemother 1996; 40: 2494-9. 13. Koo J, Tight R, Rajkumar V, Hawa Z. Comparison of once-daily versus pharmacokinetic dosing of aminoglycosides in elderly patients. Am J Med 1996; 101: 177-83. 14. McCormack JP, Jewesson PJ. A critical reevaluation of the "therapeutic range" of aminoglycosides. Clin Infect Dis 1992; 14: 320-9.
Pathology (1998), 30, August 15. Brown G, McCormack JP. Studies on drug monitoring in thrice and once daily treatment with aminoglycosides. Intensive Care Med 1993; 20: 243-4. 16. Nicolau DP, Freeman CD, Belliveau PP, Nightingale CH, Ross JW, Quintiliani R. Experience with a once-daily aminoglycoside program administered to 2184 adult patients. Amimia'ob Agents Chemother 1995; 39: 650-5. 17. Nicolau DP, WU AHB, Finocchiaro S, Udeh E, Chow MSS, Quintiliani R, Nightingale CH. Once-daily aminoglycoside dosing: impact on requests and costs for therapeutic drug monitoring. Ther Drug Monitor 1996; 18: 263-6. 18. Moore RD, Smith CR, Lietman PS. Association of aminoglycoside plasma levels with therapeutic outcome in Gram-negative pneumonia. Am J Med 1984; 77: 657-62.
MONITORING OF SERUM AMINOGL YCOSIDE LEVELS WITH ONCE-DAILY DOSING DAVID L. PATERSON*t, JENNIFER M.B. ROBSON*, MARILYN M. WAGENERt AND MICHAEL PETERS*
Sullivan Nicolaides Pathology, Taringa, Queensland, Australia*, and Infectious Disease Section, VA Medical Center, Pittsburgh, Pennsylvania, USA t
Summary There is considerable confusion as to how to monitor serum aminoglycoside levels when using once-daily dosing. At least five methods are in use in Australia. We prospectively assessed 100 consecutive once-daily courses of gentamicin or tobramycin, during which 120 pre-dose and 213 sets of immediate post-dose and six hour post-dose levels were taken. By using the six hour post-dose level we were able to compare dosage recommendations made using methods known as ALADDIN, DOSECALC and the Australian Antibiotic Guidelines nomogram (AAGN). There were statistically significant differences in the doses recommended by each method. When comparing each of the three methods, at least 25% of dosage recommendations differed by more than 80 mg per dose. Although we have not been able to determine the clinical significance of these differences, we are concerned that methods used in dosage adjustment of aminoglycosides differ so widely in their recommendations. Presumably the ALADDIN method, which utilises two post-dose levels to determine an area under aminoglycoside concentrationtime curve, gives more accurate pharmacokinetic information than methods which rely on a single level. Comparative cost-effectiveness studies of different methods, although in practice difficult to perform, should be undertaken to resolve the optimal management of patients receiving aminoglycosides once-daily. Key words: Aminoglycosides, monitoring, pharmacokinetics, gentamicin, tobramycin. Abbreviations: AAGN, Australian Antibiotic Guidelines nomogram; ALADDIN, aminoglycoside levels and daily dose indicator; AVC, aminoglycoside concentration-time curve.
Accepted 5 February J998
INTRODUCTION Once-daily aminoglycoside dosing is as effective as, and probably results in a lower risk of nephrotoxicity than, multiple-daily dosing. 1- 3 The theoretical reasons why oncedaily aminoglycoside dosing is more useful include the findings that it takes advantage of the prolonged postantibiotic effect seen with high peak aminoglycoside concentrations and that it utilises the concentration dependent killing of this class of antibiotics. In addition, the reduction in nursing time required in its preparation and administra-
tion, and the reduced variability in the timing of doses, make once-daily dosing the method of choice for prescription of aminoglycosides. The applicability of traditional peak and trough aminoglycoside concentrations to once-daily dosing has been questioned by several authors.4-6 The following reasons have been offered as to why monitoring trough concentrations is inappropriate with once-daily dosing: (i) if the initial trough level is taken 24 hours after the initial dose (ie; just prior to the second dose), no alteration in dose will be made until the third dose (ie; 48 hours after the first dose); (ii) there has been no correlation made between trough levels and toxicity or effectiveness with once-daily dosing; (iii) patients with normal aminoglycoside pharmacokinetics would not be expected to have a trough concentration exceeding 0.05 mg/l at 24 hours, yet the lower limit of sensitivity of assays such as Abbot TDx is 0.27 mg/l, thus fivefold accumulation could be occurring before this would become detectable. 4 Similarly the usefulness of peak concentrations with once-daily dosing is doubtfu1. 6 Conventional target peak concentrations (eg; 6-10 mg/I) do not apply to once-daily dosing,4 and there have been few relationships made between efficacy or toxicity and peak concentrations with once-daily dosing. An alternative approach to using peak and trough levels has been to utilise the area under the 24 hour serum aminoglycoside concentration-time curve (AUC).4 This can be done by taking two levels after the end of an aminoglycoside infusion and using these in calculating the AUC. A dose recommendation can be made by comparing this measured AUC with a target AUC that would be expected in a patient with mean population values of volume of distribution and half-life of elimination. While the calculation can be done manually, it has been packaged into a software program developed in Australia by John Turnidge and Jan Bell and known as ALADDIN (Aminoglycoside Levels and Daily Dose Indicator).7 The ALADDIN method also aims for high peak levels to reduce the risk of selection for resistance. 7 A second Australian software, developed at Royal Perth Hospital and known as DOSECALC, utilises a single serum aminoglycoside level, six to 14 hours after the first dose to calculate the AUC. Instead of measuring an early post-dose level, a presumed concentration is derived based on the dose and known parameters such as calculated creatinine clearance. Finally, a simple graphical method for patients with normal renal function has been developed. 4 This method is based on a constructed serum aminoglycoside concentration curve calculated on the maximum AUC which
11031-3025/98/030289-06 © J998 Royal College of Pathologists of Australasia
290
PATERSON et al.
would have been permitted using a "conventional" aminoglycoside regimen that resulted in a peak of 10 mg/l and a trough of 2 mg/l. A single serum aminoglycoside concentration is determined between six and 14 hours post-dose and plotted onto a graph containing the curve below which drug accumulation would be unlikely. This method appears in a modified form in the of Australian Antibiotic Guidelines (1996/1997). 8 The aim of this study was to evaluate prospectively the ALADDIN software program in the management of a group of patients who were receiving once-daily aminoglycoside therapy and who were at high risk of nephrotoxicity. At the same time, since there are other methods of monitoring aminoglycoside use, we aimed to compare whether the doses predicted by the different methods were similar. In particular, we aimed to determine whether methods using just one blood level (which would be less expensive for laboratory services and more convenient for patients and health care professionals) predicted a dose comparable to the ALADDIN method which requires two levels to be drawn. Thus we compared the doses predicted by ALADDIN with those predicted by DOSECALC and the graphical method published in Antibiotic Guidelines. 8
MATERIALS AND METHODS Residents, registrars and visiting medical officers at Greenslopes Private Hospital, Brisbane, were given education (grand rounds, several written letters) regarding the benefits of once-daily aminoglycoside dosing in August and September, 1996. They were encouraged to use the following initial dose of either gentamicin or tobramycin from 25 September, 1996: Age 10--29 yrs 30--60 yrs Over 60yrs
Dose 6mg/kg 5 mg/kg 4mg/kg
The recommendations did not apply to patients in the Intensive Care Unit and Renal Unit, and patients in these units were not evaluated in this study. Nursing staff were asked to give the aminoglycoside dose over 30 min. Blood collectors were advised to perform 2 levels after the first dose of aminoglycoside. The first level was to be taken immediately the infusion ended. The timing of this first level was regarded as critical. A second level was to be performed about 6 h later. The timing of this second level was not regarded as critical, although the precise time of collection was to be recorded. These two levels were analysed by TDx (Abbott Laboratories, Illinois). The blood sample obtained for the first level was utilised for a serum creatinine level at the start of treatment ("baseline creatinine"). Demographic details (age, sex, height, weight) plus the serum creatinine and 2 serum aminoglyeoside levels were entered into the ALADDIN software program. The dosage and dosing interval recommended by this program were communicated to the clinician by way of the laboratory report containing the aminoglycoside levels. This information was available well in advance of the second dose. A recommendation was made to collect blood for serum creatinine and the 2 serum aminoglycoside levels after the third, fifth and seventh doses. In addition, in some patients a trough sample was collected immediately prior to these doses. The result of the trough level was added to the laboratory report but was not utilised by the authors in the recommendation of dose or dosing interval. Clinicians were however free to use this information in adjusting dose. The compliance of clinicians with the dosage and dosing interval on the laboratory report was recorded by way of a review of medication sheets for each patient. We advised use of an alternative antibiotic after 7 d of aminoglycoside use. If the aminoglycoside was continued we recommended monitoring of serum creatinine and the 2 serum aminoglycoside levels after every second dose. The serum aminoglycoside level obtained approximately 6 h after the
Pathology (1998), 30, August dose was utilised in dose recommendations by the DOSECALC program and the Australian Antibiotic Guidelines nomogram (AAGN). This information was not utilised in the recommended dosage and dosage interval communicated to the clinician by way of the laboratory report. Its only use was in a subsequent analysis comparing doses recommended by the 3 systems described in this report. Patient demographics and laboratory data were entered into Excel (Microsoft). The percentage of patients developing nephrotoxicity was recorded using a standard definition for drug-induced nephrotoxicity (defined as an increase in serum creatinine over baseline of greater than 0.04 mmolJl).2,9 Vestibular toxicity was defined as vertigo and gait instability with no other identifiable cause found. to Ototoxicity was defined as a 15 dB change in hearing at any frequency, as assessed by audiometry.2 Audiometry was only performed at the discretion of the patients's clinician. Efficacy of the aminoglycoside course was defined as a successful resolution of the infection without need to substitute another antibiotic. Dose recommendations made with ALADDIN, DOSECALC and the AAGN were compared by using Wilcoxon's signed-rank test for paired data. The effect of compliance with dosage recommendations on incidence of nephrotoxicity was compared by using the chi-squared test. Student's I-test was used to determine correlation between peak levels and nephrotoxicity.
RESULTS The ALADDIN program One hundred consecutive courses of aminoglycoside, where doses were calculated using ALADDIN, were evaluated from 25 September to 30 November, 1996. Ninetyfour patients were included in this evaluation-six patients had two courses of aminoglycoside in the study period. Tobramycin was used in 18 of the courses and gentamicin in 82. The mean course duration was 7.5 days (range one to 20 days). The study population was predominantly male (79% male, 21 % female) and elderly (mean age 74.9 years, range 33-91). Indications for aminoglycosides included infective exacerbations of chronic obstructive airways disease or bronchiectasis (41 %), pneumonia (17%), urinary tract infection (10%), peritonitis/intra-abdominal abscess (6%) and others (26%). Thirteen percent of the courses of aminoglycoside were accompanied by nephrotoxicity and four percent by otovestibular toxicity. The proportion of patients who developed oto-vestibular toxicity may be an underestimate; patients were assessed for vestibular toxicity as clinically appropriate and only two patients underwent audiometry (one of whom was found to have a significant reduction in hearing). The study population was at high risk for aminoglyco side toxicity. Sixty-two percent were on other nephrotoxic drugs during the aminoglycoside course (30% on two or more nephrotoxic drugs). These included vancomycin 15%, non-steroidal anti-inflammatory drugs 21 %, diuretics 32%, angiotensin-converting enzyme inhibitors 21 %, allopurinol 12% and amphotericin 2%. Eighteen percent received intravenous contrast radiological agents during the aminoglycoside course. Fourteen percent had significant hypotension (systolic blood· pressure less than 80 mm Hg) during the aminoglycoside course. The study population had generally suboptimal baseline renal function; 36% of patients had a baseline serum creatinine of more than 0.12 mmol/l. The mean baseline creatinine clearance of the study patients was 49 ml/min. The 13 patients who developed nephrotoxicity were older than those who did not (mean age 80.7 versus 75.9
AMINOGLYCOSIDE MONITORING
years, p 0.02), received a longer course of aminoglycoside (mean duration of aminoglycoside course 11.2 versus 7.0 days, p = 0.01) and were more likely to have received allopurinol concomitantly with the aminoglyco side (31% versus 9%; p 0.048). Patients who developed nephrotoxicity were also more likely to have elevated baseline serum creatinine measurements (0.15 versus 0.12 mmol/l; p = 0.09) and have developed significant hypotension (systolic blood pressure less than 80 mm Hg) during the aminogJycoside course (31 % versus 12%; p = 0.08), but these associations were not statistically significant. There was no statistically significant association between nephrotoxicity and the following variables: sex, type of aminoglycoside and concomitant therapy with vancomycin, non-steroidal anti-inflammatory agents, diuretics, angiotensin converting enzyme inhibitors or intravenous contrast agents (p > 0.20 for all). Two hundred and thirteen sets of levels were performed in the study period. Compliance with the recommendations made on the basis of the "ALADDIN" program was 87.5%. There was a statistically significant association between noncompliance with the dose recommendations made using these levels and nephrotoxicity. In seven of the 13 aminoglycoside courses which resulted in nephrotoxicity there was at least one instance of non-compliance with the dosage recommendations by the medical team caring for the patient, as compared to 15 of the 87 aminoglycoside courses which did not result in nephrotoxicity (p == 0.005). Thus those whose medical attendants were non-compliant at least once with the recommendations were at 4.1 times greater risk of toxicity (95% confidence intervals: 1.511.1). This risk of nephrotoxicity rose to 5.9 times greater (95% Cl: 2.4-14.4) if there were two or more episodes of non-compliance. In each of the seven courses in nephrotoxic patients in which non-compliance with the dosage recommendations occurred, a recommendation to lengthen the dosing interval was made but was ignored. In contrast, in 75% of cases of non-compliance with dosing recommendations in patients who did not subsequently develop nephrotoxicity, clinicians prescribed an aminoglycoside dose lower than that recommended by the laboratory. Clinical cure was documented in 85% of the aminoglycoside courses. In 6% of cases the aminoglycoside was ceased because of nephrotoxicity, before clinical cure was observed. In 7%, the aminoglycoside was changed because of failure to respond-in three of these seven cases there was no microbiological evidence found of an infection against which aminoglycosides were microbiologically active. Of the four cases of antibiotic failure, two had infective exacerbations of chronic obstructive airways disease, one had pneumonia and one had cellulitis. Two patients (2%) died during their aminoglycoside courseone had sepsis with methicillin resistant Staphylococcus aureus and the other had autopsy proven pulmonary thromboembolism.
291
1 Distribution of "peak" aminoglycoside serum concentrations in the study population (213 "peak" levels in 100 antibiotic courses)
TABLE
Peak serum concentration (mg/l) Less than 8 8-11.9 12-15.9 16-19.9 Greater than 20
Frequency (%) 5 37 73 60 38
(2.3%) (17.4%) (34.3%) (28.2%) (17.8%)
8 mg/I-these occurred in five separate patients, all of whom were under-dosed on a mg/kg basis (eg; a 102 kg, 33 year old man was given a dose of 1.6 mg/kg). When only the "peak" level taken after the first dose of aminoglycoside was considered, the mean first peak level was 17.2 mg/l in those who subsequently developed nephrotoxicity and 15.8 mg/l in those who did not (p 0.31). All 13 patients who developed nephrotoxicity had their first peak greater than 12 mg/l, as compared to 70/87 (80%) who did not (p = 0.12). The DOSECALC program The second post-dose level was used to predict a dose using the DOSECALC program. This dose was compared to the dose predicted by the ALADDIN program using both pre- and post-dose levels. The comparison between the doses predicted by the two programs is given in Fig. 1. The mean dose predicted by DOSECALC was' 207 mg, as compared to 210 mg predicted by ALADDIN. Sixty of the 213 (28.2%) levels of the predicted doses were identical when using the two programs. However 55 of the 213 (25.8%) levels of the predicted doses differed by 80 mg or more. Using Wilcoxon's signed-rank test for paired data, there was a statistically significant difference between the doses recommended by each method (Z= 3.18852, 0.001).
p
When only the 13 patients who developed nephrotoxicity were considered, the differences between the methods became even more pronounced. Sixty percent (21/35) of the dosage recommendations differed by 80 mg or greater. In 71 % (25/35) of dosage recommendations in patients 30 (f)
25
'0
~
20
~
15
0..
10
'5
1£
'0
c:
e (!)
5
(!)
Peak aminoglycoside levels The serum aminoglycoside level taken between the end of the infusion and 30 minutes post-infusion, equates to a "peak" level as measured in the traditional "peak and trough" style monitoring of aminoglycosides. Two hundred and thirteen such levels were taken. Their distribution is given in Table 1. Only 5/213 (2.3%) levels were less than
0...
Doses (in mg) predicted by DOSECALC less than those of ALADDIN
Doses (in mg) predicted by DOSECALC greater than those of ALADDIN
Fig. I Comparison of the doses (in mg) recommended by ALADDIN and those recommended by DOSECALC.
292
Pathology (1998), 30, August
PATERSON et al.
40
TABLE 2 Methods available for monitoring once-daily aminoglycosides
(/)
3l
.g
Traditional trough and pe,\\ measurements Trough measurement only ALADDIN program? DOSECALC program Australian Antibiotics Guidelines Nomogram 8 Nicolau Nom0l3.raml6 Prins schedule
30
"'0 Q)
t5
'g
0.
c Q)
~
20 10
Q)
a..
Doses (in mg) predicted by AAG less than that of ALADDIN
Doses (in mg) predicted by AAG greater than those of ALADDIN
Fig. 2 Comparison of the doses (in mg) recommended by ALADDlN and those by the AAGN.
who became nephrotoxic, DOSECALC predicted a lower dose than ALADDIN.
The Australian Antibiotic Guidelines Nomogram The second post-dose level was used to predict a dose using the AAGN. This predicted dose was compar:ed to the dose predicted by the ALADDIN program usmg both pre-and post-dose levels in 185 cases (Fig. 2). The mean dose predicted by the AAGN was 249 mg ~s compared to 210 mg predicted by ALADDIN. Forty-D1~e o~ the 185 (27%) cases of the predicted doses were Identlcal when using the two programs. However 29.2% (54/185) of the doses predicted by AAGN differed by 80 mg or more fr?m those predicted by ALADDIN using the same data. Usmg WiJcoxon's signed-rank test for paired data, there was a statistically significant difference between the doses recommended by each method (Z = 8.77422, p = 0.0001). Even when the use of the AAGN was restricted to patients whose creatinine clearance was greater than 60 ml/min, there was still a statistically significant difference between the doses recommended by each method (p < 0.01). When only patients who developed nephrotoxicity were considered, 28.6% (10/35) of the dosages recommended by ALADDIN and AAGN differed by 80 mg or more. In only two cases (5.7% of the 35 doses in patients who became nephrotoxic) was the dose recommended by AAGN lower than that recommended by ALADDIN. In patients who became nephrotoxic, the mean dose recommended by ALADDIN was 168 mg, as compared to 194 mg recommended by AAGN. Finally, when doses predicted by DOSECALC an~ the AAGN were compared, 20/185 (10.8 %) of the pred.lcted doses were identical and a statistically significant dlfference between the doses recommended by each method was observed (Z = 7.29393, p = 0.0001). DISCUSSION There are no clear recommendations available regarding how to monitor aminoglycosides when using once-daily dosing. An ideal method would be o~e which. requires the fewest blood collections (avoiding mconveDlence to the
patient and minimising costs to the blood c?l!e~tion servi:e and laboratory) and which allows the chmcIan to aVOId toxicity and provide a cure for the infection being treated. Numerous methods are in use in Australia at the present time (Table 2). The only way to determine which meth?d is closest to the ideal would be to perform a controlled trIal randomly allocating monitoring of aminoglycosides to each different method and determining which method could deliver the most cost-effective service (minimising toxicity and cost, and maximising clinical cure). Such a study would require a sample size unobtainable for most la?oratories and would be logistically very difficult to coordmate. In the absence of being able to perform such a trial, the aim of this study was to determine whether each of the three methods in common use (ALADDIN, DOSECALC and AAGN) gave a similar dose recommendation. If methods utilising one blood collection gave the same predicted dose as those utilising two blood collections, it would appear that methods needing a single blood collection would be more cost-effective and should be recommended for use with once-daily dosing. We found that the three different monitoring methods were not equivalent. Both methods using a single blood collection (DOSECALC and AAGN) gave a significantly different dose recommendation from that recommended by utilising two levels and the ALADDIN program. W,e recognise that a statistically significant difference IS not necessarily clinically significant. However we found that for at least 25% of the levels, a dose was recommended by either DOSECALC or AAGN which was at least 80 mg different from that recommended by the ALADDIN program. We have no way of determini~g whether .s~ch a difference in dose would increase likelIhood of tOXICIty or decrease chance of clinical cure, however we feel that it is likely to be clinically important. We would expect that the use of two levels compared to the use of one would give more accurate pharmacokinetic information and would therefore be advantageous in the care of the patient. We found that when using the ALADDIN program to direct once-daily aminoglycoside dosing, 13% of patients developed nephrotoxicity and 4% developed vesti?u.locochlear toxicity. The percentage of patients recelvmg once-daily dosing who develop nephrotoxicity has ranged in previous studies from 0 to 26.7%, I1 with the two ~~st recent studies of once-daily dosing showing nephrotOXICity in 14-24% of patients. 12.l 3 Both of these recent studies dealt with a vounger popUlation with substantially better baseline renal function than our own. 12•13 Both of these demographic details are demonstrated Ii~k factors f~r nephrotoxicity.14 The patients in these St~dl~S wer~ momtored utilisinG traditional methods. It IS ImpOSSible to compare pop~lations in different studies, but we believe that the 13% incidence of nephrotOXicity in our popu-
AMINOGLYCOSIDE MONITORING
lation, which had numerous risk factors for renal dysfunction, is at least no worse than that observed with traditional monitoring methods. FUlthermore, we observed that noncompliance with the recommendations of ALADDIN was significantly associated with development of nephrotoxicity. Likelihood of clinical cure was high. We feel therefore that this program has demonstrated usefulness in the management of patients on aminoglycosides. Despite the pleasing results of this program in the monitoring of aminoglycosides, we found some drawbacks with the ALADDIN system. The need for two blood collections and in particular, a second blood collection six hours after the aminoglycoside infusion, requires close coordination between nursing staff and blood collectors. A dose administered after 6 p.m. necessitates waking the patient for the second level six hours later. Although not absolutely required in the prediction of a dose, the desirability of requirement for information on height, weight and serum creatinine may be rarely provided by busy clinicians. We avoided this problem by asking the blood collectors to provide this information and by performing a creatinine assay on the serum sample submitted for aminoglycoside testing. As our laboratory report included a dosage recommendation, each report was reviewed by a medical microbiologist before being sent to the clinician. While this provided no difficulty during working hours, the issue of medical coverage on weekends and afterhours needs to be addressed when using the ALADDIN system. We found that liaison with laboratory scientists by way of telephone and modem after hours was generally sufficient. While use of a monitoring system based on a 24 hour serum aminoglycoside concentration-time curve is theoretically attractive, the use of trough and sometimes peak measurements appears to be widespread, with once-daily dosing in many institutions in Australia. Recent reviews have suggested that monitoring of trough concentrations has never been demonstrated to improve the outcome over empiric dosing regimens with dosage adjusted if a change in renal function occurs.5 ,14,15 A number of authors have suggested restricting monitoring of serum concentrations to only those patients with a creatinine clearance of below 50 ml/minute and then only performing trough levels. 12,16,17 Cost savings from this strategy of more than $100,000 per year were expected in one 850 bed hospital. 17 We found no correlation between peak serum aminoglycoside concentration after a dose and development of nephrotoxicity. This is in contrast to a recent study which showed that those who developed nephrotoxicity had a significantly higher initial serum peak concentration than those who did not develop this complication. 13 These authors suggested that an initial peak concentration of greater than 12 mg/l was predictive of nephrotoxicity. While we found that all patients who developed nephrotoxicity had a first peak of greater than 12 mg/l, 80% of those who did not develop nephrotoxicity also had levels greater than 12 mg/l and the association between nephrotoxicity and a first peak of greater than 12 mg/l did not reach statistical significance. Once-daily dosing using the dosage recommendations in this study, gave peak levels greater than 8 mg/l in more than 97% of cases. This is reassuring as at least one study has found a correlation between successful treatment of serious Gram-negative infections and peak levels greater than 8 mg/l.IS If a laboratory does not use a monitoring system based on AUC calculations (such as ALADDIN), it
293
may be prudent to check a peak concentration after the first dose is given, particularly if underdosing is frequent in their institution. An appropriate target range could be between 8 and 20 mg/l. We consider a target peak of 8-12 mg/l as too restrictive, particularly considering a range of less than 12 mg/l is based on only a single small study.13 There has never been a correlation between a peak level other than the first peak level and toxicity, and hence there is no indication to continue performing any more than one peak level. In conclusion, there are currently numerous methods available for the monitoring of aminoglycosides prescribed once per day. While monitoring of peaks and troughs has continued in many laboratories, there are theoretical reasons why this is not as appropriate as using AUC methods. If peaks and troughs are measured, the reference ranges should be different from those advocated previously when multiple-daily doses were given. We have found that the three methods currently available which use AUC differ significantly in their dosage recommendations. The methods are clearly not interchangable. We are unable to conclude whether one method is superior over another, but we have found a reasonably low rate of nephrotoxicity in high-risk patients monitored using the ALADDIN method. Furthermore, there was a close correlation between noncompliance with the doses recommended by this method and development of nephrotoxicity. Large randomised controlled trials would be necessary to resolve the question as to whether one method is more cost-effective and clinically superior to another, or if indeed levels need to be performed at all in most patients. ACKNOWLEDGMENTS The authors gratefully acknowledge the assistance of the medical and nursing staff at Greenslopes Private Hospital (Qld) and of the venepuncture and laboratory staff at Sullivan Nicolaides Pathology (Qld). Alasdair Miller is thanked for providing a free copy of DOSECALC software and its subsequent revision. Address for correspondence: D.L.P., Infectious Disease Section, VA Medical Center, University Drive C, Pittsburgh, PA 15240, USA.
References 1. Barza M, Ioannidis JPA, Cappelleri JC, Lau 1. Single or multiple daily doses of aminoglycosides: a meta-analysis. Br Med J 1996; 312: 338-45. 2. Hatala R, Dinh T, Cook DJ. Once-daily aminoglycoside dosing in immunocompetent adults: a meta-analysis. Ann Intern Med 1996; 124: 717-25. 3. Munckhof WJ, Grayson ML, Turnidge JD. A meta-analysis of studies on the safety and efficacy of aminoglycosides given either once daily or as divided doses. J Antimicrob Chemother 1996; 37: 645-63. 4. Begg EJ, Barclay ML, Duffull SB. A suggested approach to oncedaily aminoglycoside dosing. Br J CUn Pharmac 1995; 39: 605-9. 5. Cronberg S. Simplified monitoring of aminoglycosides. J Antimicrob Chemother 1994; 34: 819-27. 6. Konrad F, Wagner R, Neumeister B, Rommel H, Georgieff M. ~tudies on drug monitoring in three and once-daily treatment with aminoglycosides. Intensive Care Med 1993; 19: 215-20. 7. Turnidge J, Bell J, Mascaro L. The peak-AUC method for aminoglycoside dosage adjustment. Program and abstracts of the 36th Interscience Conference on Antimicrobial Agents and Chemotherapy, New Orleans, September 1996. 8. Victorian Drug Usage Advisory Committee. Antibiotic guidelines. 9th edn. North Melbourne: Victorian Medical Postgraduate Foundation Inc., 1996. 9. Smith CR, Lipsky 11, Laskin OL, Hellmann DB, Mellits ED, Longstreth J, et al. Double-blind comparison of the nephrotoxicity and auditory toxicity of gentamicin and tobramycin. N Engl J Med 1980; 302: 11 06-9.
294
PATERSON et al.
10. Longridge NS. Approach to the patient with dizziness and vertigo. In: Kelley WN, editor. Textbook of Internal Medicine. 3rd edn. Philadelphia: Lippincott-Raven, 1997. 11. Blaser J, Konig C. Once-daily dosing of aminoglycosides. Eur J Clin Microbial Infect Vis 1995; 14: 1029-38. 12. Prins JM, Weverling GJ, De Blok K, Van Ketel RJ, Speelman P. Validation and nephrotoxicity of a simplified once-daily aminoglycoside dosing schedule and guidelines for monitoring therapy. Antimicrob Agents Chemother 1996; 40: 2494-9. 13. Koo J, Tight R, Rajkumar V, Hawa Z. Comparison of once-daily versus pharmacokinetic dosing of aminoglycosides in elderly patients. Am J Med 1996; 101: 177-83. 14. McCormack JP, Jewesson PJ. A critical reevaluation of the "therapeutic range" of aminoglycosides. Clin Infect Dis 1992; 14: 320-9.
Pathology (1998), 30, August 15. Brown G, McCormack JP. Studies on drug monitoring in thrice and once daily treatment with aminoglycosides. Intensive Care Med 1993; 20: 243-4. 16. Nicolau DP, Freeman CD, Belliveau PP, Nightingale CH, Ross JW, Quintiliani R. Experience with a once-daily aminoglycoside program administered to 2184 adult patients. Amimia'ob Agents Chemother 1995; 39: 650-5. 17. Nicolau DP, WU AHB, Finocchiaro S, Udeh E, Chow MSS, Quintiliani R, Nightingale CH. Once-daily aminoglycoside dosing: impact on requests and costs for therapeutic drug monitoring. Ther Drug Monitor 1996; 18: 263-6. 18. Moore RD, Smith CR, Lietman PS. Association of aminoglycoside plasma levels with therapeutic outcome in Gram-negative pneumonia. Am J Med 1984; 77: 657-62.