Pharmacokinetics of remifentanil and its major metabolite, remifentanil acid, in ICU patients with renal impairment{dagger}

M. Pitsiu*,1, A. Wilmer2, A. Bodenham3, D. Breen4, V. Bach5, J. Bonde6, P. Kessler7, S. Albrecht8, G. Fisher9 and A. Kirkham9

1 Medeval Ltd, Skelton House, Manchester Science Park, Lloyd Street North, Manchester M15 6SH, UK. 2 MICU, UZ Gasthuisberg, Leuven, Belgium. 3 ICU, Leeds General Infirmary, Leeds, UK. 4 ICU, Royal Hallamshire Hospital, Sheffield, UK. 5 ICU, Hilleroed Syngehus, Hilleroed, Denmark. 6 ICU, Amtssygehuset i Herlev, Herlev, Denmark. 7 ICU, J-W Goethe Universitat Zentrum der Anaesthesiologie und Wiederbelebung, Frankfurt, Germany. 8 ICU, Universitat Erlangen-Nurnberg.Klinik fur Anaesthesiologie, Erlangen, Germany. 9 GlaxoSmithKline R&D, Greenford, UK

*Corresponding author. E-mail: m.pitsiu@medeval.com
{dagger}Declaration of interest. This study (USA30212) was supported by a research grant from GlaxoSmithKline, UK.

Accepted for publication: November 12, 2003


    Abstract
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Background. The pharmacokinetics of remifentanil, an opioid analgesic metabolized by non-specific esterases, and its principal metabolite, remifentanil acid (RA), which is excreted via the kidneys, were assessed as part of an open-label safety study in intensive care unit (ICU) patients with varying degrees of renal impairment.

Methods. Forty adult ICU patients with normal/mildly impaired renal function (creatinine clearance [CLcr] 62.9 (SD) 14.5 ml min–1; n=10) or moderate/severe renal impairment (CLcr 14.7 (15.7) ml min–1; n=30) were included. Remifentanil was infused for up to 72 h, at a starting rate of 6–9 µg kg–1 h–1 titrated to achieve a target sedation level, with additional propofol (0.5 mg kg–1 h–1) if required. Intensive arterial sampling was performed for up to 72 h after infusion. Pharmacokinetic parameters obtained by simultaneous modelling of remifentanil and RA data were statistically compared between the two groups.

Results. Remifentanil pharmacokinetics were not significantly affected by renal status. RA clearance in the moderate/severe group was reduced to about 25% that of the normal/mild group (41 (29) vs 176 (49) ml kg–1 h–1, P<0.0001). Metabolic ratio, a predictor of the ratio of RA to remifentanil concentrations at steady state, was approximately eight-fold higher in the moderate/severe group relative to the normal/mild group (116 (110) vs 15 (4), P<0.0001). Maximum RA levels approached 700 ng ml–1 in the moderate/severe group.

Conclusions. Although RA accumulates in patients with moderate/severe renal impairment, pharmacokinetic modelling predicts that RA concentrations during a 9 µg kg–1 h–1 remifentanil infusion for up to 15 days would not exceed those reported in the present study, for which no associated prolongation of µ-opioid effects was observed.

Br J Anaesth 2004; 92: 493–503

Keywords: analgesics opioid, remifentanil; intensive care, renal; pharmacology, drug metabolism; pharmacology, pharmacokinetics


    Introduction
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Critically ill patients in the intensive care unit (ICU) often experience pain, anxiety, agitation and confusion, while being exposed to numerous potentially noxious stimuli attributable to diagnostic, therapeutic and nursing interventions. Most critically ill patients require a combination of analgesia (provided by an opioid) and sedation (provided by benzodiazepines or other hypnotics) during at least part of their stay in the ICU. Opioids such as fentanyl, sufentanil and morphine are traditionally used for the provision of analgesia. Elimination of opioids may be prolonged in critically ill patients, due to organ-dependent elimination and, in some cases, formation of active metabolites (e.g. morphine-6-glucuronide1 2). There is therefore a potential for accumulation of drug, metabolite or both and hence unpredictable or delayed recovery from analgesia and sedation, particularly during weaning from mechanical ventilation.

Remifentanil (remifentanil hydrochloride) is an opioid with a rapid onset of action of about 1 min which quickly achieves steady state. It has a context-sensitive half-life of 2–3 min, which is independent of the duration of infusion.3 Remifentanil is not a substrate for plasma cholinesterase and therefore its metabolism is not subject to genetic variance. Unlike existing opioids, remifentanil exhibits a predictable, rapid metabolism by non-specific esterases in the blood and tissues, principally to a carboxylic acid derivative, remifentanil acid (RA).4 This organ-independent elimination of remifentanil makes it a useful agent in the ICU setting, where patients commonly have some degree of organ dysfunction.5 These features of remifentanil combine to make it easy to titrate to the desired analgesic effect and allow it to be administered for long periods and at higher doses than are normally used with traditional opioids, without the risk of significant accumulation. A number of studies have looked at the potential role and use of remifentanil in the critically ill.613

Whilst remifentanil elimination is essentially independent of renal and liver function, RA is eliminated via the kidneys, and its elimination is prolonged in patients with severe renal impairment (predicted creatinine clearance (CLcr) <10 ml min–1).14 RA therefore accumulates in these patients. Since RA has never been administered to man, the concentration at which µ-agonist effects are likely to be seen has not been ascertained. A remifentanil infusion of 0.2 µg kg–1 min–1 results in respiratory depression and, according to simulations by Minto and colleagues,15 that equates to a remifentanil blood concentration of approximately 5 ng ml–1 in volunteers. RA has been demonstrated to be much less potent (1/4600) a µ-opioid agonist than the parent molecule in dogs,16 and assuming the same potency ratio applies to man, it is therefore thought not to result in any clinically relevant effects at concentrations below 900 ng ml–1. A clinical trial was designed primarily to determine any potential safety issues that might arise as a result of the accumulation of RA in renally impaired ICU patients given a continuous remifentanil infusion for up to 72 h. This length of infusion was chosen as a cautious approach, which would on average allow steady-state concentrations of RA to be approached and also complied with the 72 h restriction in the propofol data sheet in some countries. In addition to the pharmacodynamic effect, adverse events and haemodynamic profiles, the pharmacokinetic profiles of remifentanil and RA were assessed. The specific objectives of the analysis were to characterize the pharmacokinetics of remifentanil and RA using both model-independent and compartmental modelling methods, and to correlate the pharmacokinetic parameters obtained to the degree of renal impairment.


    Methods
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Study design
This was an open-label, non-comparator study assessing the safety and pharmacokinetic profiles of remifentanil and RA after administration of remifentanil by continuous i.v. infusion for up to 72 h in ICU patients with varying degrees of renal impairment. The study was conducted in accordance with good clinical practice and with the guidelines set out in the Declaration of Helsinki. Written informed consent/assent was obtained from all patients or their representatives. Following local and national ethics committee approvals, 40 patients were recruited altogether, from three centres in the UK, two in Denmark, two in Germany and one in Belgium.

Male or female (of non-child bearing potential or using contraception) post-surgical and medical patients were eligible for entry into the study if they were aged 18 yr or more, weighed 120 kg or less, and if they were expected to require mechanical ventilation for a further 24–72 h. Patients should have had a sedation–agitation scale (SAS) score in the range 2–4 at admission to the ICU (score of 2: patient is very sedated, can be roused by physical stimuli but does not communicate or follow commands, may move spontaneously; score of 3: patient is sedated, difficult to rouse, awakens to verbal stimuli or gentle shaking but drifts off again, will follow simple commands; score of 4: patient is calm and cooperative, easily rousable and follows commands).17

During an initial screening period, the patient’s renal function was measured by estimating CLcr from plasma and urinary creatinine concentrations and urine volumes over a minimum period of 4 h with extrapolation to 24 h.18 Patients with an estimated CLcr of 50 ml min–1 or higher were classified as having normal renal function or mild renal impairment; patients with values below 50 ml min–1 were classified as having moderate/severe renal impairment.

Remifentanil (lyophilized powder in sterile vials each containing 5 mg of the compound, provided by GlaxoSmithKline UK, reconstituted and diluted to 50 ml using standard diluents) was administered as a continuous i.v. infusion for a maximum of 72 h. The remifentanil infusion started at a dose of 6 or 9 µg kg–1 h–1, and was then titrated according to a dosing algorithm.19 Propofol (starting dose 0.5 mg kg–1 h–1) was administered as an additional infusion if required when the remifentanil infusion rate requirements reached 12 µg kg–1 h–1. During the treatment period, patients were continuously assessed and the remifentanil (and propofol if required) dose regimen was adjusted in order to maintain an SAS score of 2–4 (as considered clinically appropriate) with no or only mild pain. In addition, scheduled down-titrations of the remifentanil regimen by 25% of the initial rate at 10 min intervals were performed at 8, 24 and 48 h. The infusion rate was reduced until an offset of remifentanil’s pharmacodynamic effects was observed. If propofol was being infused at the time of a scheduled down-titration, it was maintained at a constant infusion rate. Study treatment was permanently discontinued after a maximum of 72 h, by reducing the remifentanil infusion rate by 25% at 10 min intervals. Any propofol which was infused at this time was maintained at a constant rate until 30 min after the remifentanil had been stopped and then decreased in decrements. After the remifentanil was finally discontinued, the patient was switched to standard therapy of the investigator’s choice. Although the intention was for patients to complete the 72 h of remifentanil infusion, there was no minimum infusion period for entry into the trial and all patients who received remifentanil and had measurable remifentanil concentrations were included.

Sample acquisition, handling and processing
Arterial blood samples (5 ml) were collected from all patients into tubes containing citric acid and were frozen for subsequent assay at the following time points: before the start of the remifentanil infusion and at 15 and 30 min, and 1, 2 and 4 h after starting the infusion; at 08:00 h and 20:00 h each day (unless these times were within 4 h of a scheduled down-titration); and at the 8, 24 and 48 h scheduled down-titrations (immediately before the first down-titration step and before the infusion rate was turned up again). If the remifentanil infusion was temporarily discontinued during a scheduled down-titration, a blood sample was obtained immediately before discontinuation and at 5, 10 and 30 min or just before the infusion was restarted, whichever occurred first. At final treatment discontinuation (72 h maximum), blood samples were obtained immediately before the first down-titration, before the infusion was decreased by 25% of the initial rate at 10 min intervals (three samples) and immediately before the infusion was finally discontinued. During the post-treatment period, blood samples were obtained at approximately 5, 10 and 30 min, and at 1, 2, 4, 8, 24, 40, 56 and 72 h after final discontinuation of remifentanil infusion.

On the occasions that renal replacement therapy (RRT) was performed, at the time of starting dialysis and after the system had been stabilized, two blood samples were obtained simultaneously from the line entering the dialysis machine and from the line exiting the dialysis machine.

Assay method
The concentrations of remifentanil and RA in whole blood were determined using validated assay procedures. The method, a specific liquid chromatography tandem mass spectrometry method, was a modification of a previously published method,20 which involved solid-phase extraction with methanol instead of dichloromethane and enables simultaneous quantification of free remifentanil and RA. The lower limit of quantification (LLQ) for both analytes was 0.1 ng ml–1. Intra-assay precision values at LLQ were less than 20% and accuracy values were within 100 (20)%, with reference to the nominal value. Overall intra- and inter-assay precision values were less than 15% and accuracy values within 100 (15)% of the nominal value.

Pharmacokinetic analysis
The following parameters were determined for remifentanil and RA using standard model-independent methods in WinNonlinTM professional version 3.121 maximum plasma concentration (Cmax), the first time to reach Cmax (tmax), and area under the plasma concentration–time curve (AUC) to the last sample time (AUClast). The metabolic ratio was calculated as the ratio of AUClast of RA to that of remifentanil.22 The extraction of remifentanil and RA during RRT was determined as the ratio (%) of the concentration in the samples taken from the lines entering and exiting the dialysis equipment.

Simultaneous pharmacokinetic modelling of the individual remifentanil (parent) and RA data was performed using NONMEM.23 Visual inspection of the data indicated that remifentanil concentrations during the post-infusion period were adequately characterized by a monoexponential decline. However, RA concentrations during the post-infusion period were characterized by either a mono- or a biexponential pattern. Consequently, simultaneous compartmental analysis of parent drug and metabolite involved fitting either a two- or a three-compartment model to the blood concentration–time data, with one compartment associated with the parent drug and one or two compartments with the metabolite, with a variable zero-order input rate to the parent compartment. The pharmacokinetic model is depicted in Figure 1.



View larger version (5K):
[in this window]
[in a new window]
 
Fig 1 Graphical representation of the combined pharmacokinetic model for remifentanil and its metabolite (remifentanil acid; RA). Compartment 1 is associated with remifentanil and compartments 2 and 3 with RA. k0 is the remifentanil zero-order infusion rate, k12 and k20 the first-order elimination rate constants of remifentanil and RA, respectively, and k23 and k32 the intercompartmental rate constants for RA between compartments 2 and 3.

 
The full structural pharmacokinetic model used in NONMEM is given below in differential equation form:

dR1/dt=k0/V–(R1k12)

dM2/dt=(R1k12)–(M2k23)+(M3k32)–(M2k20)

dM3/dt=(M2k23)–(M3k32)

where R1 is the concentration of remifentanil in compartment 1; M2 and M3 are the concentrations of RA in compartments 2 and 3, respectively; k0 is the remifentanil infusion rate, V is the volume of distribution of remifentanil (compartment 1), and k12 is the total elimination rate constant of remifentanil, assuming that all the parent drug is converted into the metabolite.

To facilitate the fitting process, the model was re-parameterized in terms of remifentanil clearance (CL) and volume of distribution (V), and RA clearance (CLm), central compartment volume (Vm1), intercompartmental clearance (Qm) and peripheral compartment volume (Vm2), which were estimated as primary parameters. From these the compartmental rate constants were estimated as follows:

k12=CL/V

k20=CLm/Vm1

k23=Qm/Vm1

k32=Qm/Vm2

The elimination half-life of remifentanil (t1/2)was estimated as t1/2= ln(2)/k12. The half-lives associated with the alpha ({alpha}) and beta (ß) of the exponential form of the model for RA[M(t)=Ae{alpha}t+Beßt] were derived from the appropriate rate constants.24 The phase that on average contributed most to the total AUC was termed the elimination phase and the corresponding half-life as the RA elimination half-life.

Since this analysis involved individual data, the variance model in NONMEM represented intra-individual variability ({epsilon}), with different {epsilon} values estimated for remifentanil ({epsilon}R) and RA ({epsilon}M), both assumed to arise from a log-normal distribution.

The reduced structural pharmacokinetic model (one compartment for remifentanil and one compartment for RA, i.e. excluding compartment 3 in Figure 1) and variants of the variance model were also tested. The same parameterization in terms of CL and V for remifentanil and CLm and Vm1 for RA was applied and the compartmental rate constants were estimated as follows:

k12=CL/V

k20=CLm/Vm1

For this model, the elimination half-life of RA was estimated as t1/2,m,k20 = ln(2)/k20.

In both models, it was assumed that all remifentanil was converted to RA. Since 95–98% of a remifentanil dose is recovered in urine as RA,14 this assumption was not expected to have led to any significant bias in the parameter estimates.

Model discrimination was based on the NONMEM objective function, taking into account the estimation error and the degree of correlation between the parameters. Residual plots were inspected for goodness of fit and lack of bias. The likelihood ratio was used to assess whether the difference in the NONMEM objective function between the model with a biexponential decline for the RA data (more complex) and that with a monoexponential decline (base model) indicated an improved fit of the model to the data, as follows. A change in the NONMEM objective function of more than 3.84 (based on the likelihood ratio, which is approximately {chi}2 distributed for one degree of freedom) when compared with the base model was considered significant (P<0.05). A superior model was also expected to reduce residual error terms.

Finally, any association between covariates likely to affect the pharmacokinetic parameters of remifentanil and RA (e.g. CLcr, body weight, age) was explored graphically.

Statistical analysis
Descriptive statistics were used to summarize the pharmacokinetic parameters of remifentanil and RA in the two groups and for the following three subgroups in the moderate/severe group: patients who had no RRT (subgroup A); patients having intermittent RRT (i.e. RRT starting or ending within the period of the pharmacokinetic sampling) (subgroup B); patients who had continuous RRT for the duration of pharmacokinetic sampling (subgroup C). As patients in this study received individualized remifentanil regimens for different lengths of time, only the ranges of maximum concentrations and AUC values are reported.

Appropriate statistical methods (ANOVA by group) were employed to test for differences in the pharmacokinetic parameters of remifentanil and RA between the renal function groups. P<0.05 was considered statistically significant. To elucidate any differences, an exploratory ANOVA and contrast analyses between the normal/mild group and the three moderate/severe subgroups were also performed. All statistical analyses were performed in SAS® System version 8.


    Results
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Study design
Forty patients were recruited into the study, 10 in the normal/mild group and 30 in the moderate/severe group; all were evaluable for the pharmacokinetic analyses. In a previous pilot study,26 12 patients with normal renal function or mild renal impairment were included, and their pharmacokinetic profiles were characterized. There was therefore thought to be no need to include more mild/normal patients in the present study. The ratio 3:1 still permits statistical comparisons. The baseline physical and clinical characteristics of these patients are summarized in Table 1. The mean age and weight values were similar between the two groups, although the proportion of male patients in the normal/mild group was higher than in the moderate/severe renal impairment group.


View this table:
[in this window]
[in a new window]
 
Table 1 Physical characteristics of patients. Renal function was assigned by the clinician, based on the patient’s estimated creatinine clearance (CLcr). Normal/mildly impaired renal function corresponded to an estimated CLcr of 50 ml min–1 or higher; moderate/severe renal impairment corresponded to an estimated CLcr below 50 ml min–1. SAPS II, simplified acute physiology score II.25 Ranges are given in parentheses
 
There were differences in CLcr between the two groups, as expected from the study stratification, although these were not tested for statistical significance. Mean CLcr at screening was 62.9 (SD) 14.5 ml min–1 in the normal/mild group and 14.7 (15.7) ml min–1 in the moderate/severe group. Only two patients in the normal/mild group had a CLcr value greater than 80 ml min–1, indicating normal renal function.

More patients with moderate/severe renal impairment had higher simplified acute physiology score II values25 (associated with higher hospital mortality rate) than those in the normal/mild group.

Of the 30 patients in the moderate/severe group, five patients underwent continuous RRT, nine underwent intermittent RRT, and 16 patients required no RRT. Four different methods of RRT were used: haemodialysis (n=3), continuous venovenous haemodialysis (n=2), continuous venovenous haemofiltration (n=4) and continuous venovenous haemodiafiltration (n=5). There was no evidence for remifentanil being extracted during RRT; RA appeared to be extracted by approximately 60% in the three cases of haemodialysis. None of the patients in the normal/mild group needed RRT.

The duration of the remifentanil infusion ranged from 45.4 to 72.8 h in the normal/mild group and from 4.8 to 72.5 h in the moderate/severe group; on average it was shorter for the moderate/severe group (mean 43.1 h) than for the normal/mild group (mean 70.4 h). The weighted mean remifentanil infusion rate requirement was slightly lower in the moderate/severe group (10.1 µg kg–1 h–1) than in the normal/mild group (13.1 µg kg–1 h–1). Requirement for supplementary propofol infusions was also lower in the moderate/severe group (53%) than in the normal/mild group (70%), with mean propofol infusion rates in the two groups being 0.7 and 1.3 mg kg–1 h–1, respectively. These differences in remifentanil and propofol requirements between the moderate/severe and the normal/mild group were not statistically significant, and within each group, remifentanil and propofol requirements did not change with time.

Pharmacokinetics
Remifentanil concentrations in both renal function groups rose immediately after starting the i.v. infusion and declined rapidly during the down-titration phase and after discontinuation of the remifentanil infusion. For more than half of the patients, there were fewer than three quantifiable concentrations in the post-infusion samples. Maximum levels of remifentanil were observed at various times during the infusion period, reflecting fluctuations in the remifentanil dosing rate (Fig. 2).



View larger version (23K):
[in this window]
[in a new window]
 
Fig 2 Semilogarithmic plots of individual remifentanil concentrations in the two groups: (A) normal/mild; (B) moderate/severe renal impairment.

 
RA concentrations showed a gradual increase after starting the remifentanil infusion. Maximum levels of RA in the blood were generally observed just before the final remifentanil infusion down-titration, indicating a slow accumulation towards steady state. RA profiles were not very sensitive to changes in the remifentanil dosing rate. After discontinuation of the remifentanil infusion, when comparing the profiles in patients for whom a similar number of post-infusion samples were available, RA levels appear to decline in a biexponential fashion for patients in the normal/mild group, and in a monoexponential fashion for patients in the moderate/severe group (Fig. 3), with RA concentrations in the blood samples from patients in the moderate/severe group generally much higher than those from the patients in the normal/mild group. The two profiles that show no decline of remifentanil or RA concentrations in Figures 2 and 3 correspond to subjects who continued to receive remifentanil, at the investigator’s decision, beyond the 72 h study regimen. This was taken into account in the pharmacokinetic analysis.



View larger version (24K):
[in this window]
[in a new window]
 
Fig 3 Semilogarithmic plots of individual remifentanil acid concentrations in the two groups: (A) normal/mild; (B) moderate/severe renal impairment.

 
A three-compartment model (parent drug one compartment; metabolite two compartments) was found to be statistically superior in providing the best fit for the combined parent and metabolite data for patients in the normal/mild group, while a two-compartment model (parent drug one compartment; metabolite one compartment) was found to be adequate for the majority of the patients in the moderate/severe group. Observed and predicted RA semilogarithmic concentration–time plots from representative patients demonstrating two- and one-compartment behaviour are shown in Figures 4 and 5.



View larger version (18K):
[in this window]
[in a new window]
 
Fig 4 Two examples of patients’ remifentanil and remifentanil acid (RA) concentration–time profiles demonstrating biexponential decline (normal/mild renal impairment group).

 


View larger version (18K):
[in this window]
[in a new window]
 
Fig 5 Two examples of patients’ remifentanil and remifentanil acid (RA) concentration–time profiles demonstrating monoexponential decline (moderate/severe renal impairment group).

 
The principal pharmacokinetic parameters for remifentanil and RA are summarized in Tables 2 and 3. Remifentanil CL and V appeared to be higher in the moderate/severe group overall compared with the normal/mild group, but the interindividual variability in the parameters was high. The mean t1/2 of remifentanil was longer in the moderate/severe group than in the normal/mild group.


View this table:
[in this window]
[in a new window]
 
Table 2 Model-independent pharmacokinetic parameters for remifentanil and remifentanil acid following remifentanil infusion in intensive care unit patients with various degrees of renal impairment. Subgroups A, B and C were undergoing no, intermittent or continuous renal replacement therapy, respectively. Data are mean (SD) or range. Pharmacokinetic terms are explained in the text
 

View this table:
[in this window]
[in a new window]
 
Table 3 Compartmental pharmacokinetic parameters for remifentanil and remifentanil acid following remifentanil infusion in intensive care unit patients with various degrees of renal impairment. Data are mean (SD). na, not applicable; {dagger}for the three-compartment model only; {ddagger}t1/2,m,{alpha} for normal/mild (best model three-compartment), t1/2,m,k20 for moderate/severe (best model two-compartment); §excluding patients 16142 (parameters not estimable) and 16141, 16160, 16166 (best model three-compartment). Pharmacokinetic terms are explained in the text
 
Mean metabolic ratio was about 8-fold higher in the patients in the moderate/severe group compared with the normal/mild group (Table 2). CLm in the moderate/severe group was approximately 25% of that in the normal/mild group (Table 3), and was linearly related to CLcr (Fig. 6, r=0.845, P=0.0962 and P<0.0001 for the intercept and slope, respectively). Two half-lives were estimable in the normal/mild group, the terminal-phase one being similar to the single half-life estimated for the moderate/severe group.



View larger version (16K):
[in this window]
[in a new window]
 
Fig 6 Scatter plot of remifentanil acid clearance (CLm) vs creatinine clearance (CLcr), including regression line.

 
Statistical analysis of remifentanil and RA pharmacokinetic parameters indicated that the differences in metabolic ratio, CLm and RA half-life (t1/2m,k20 vs t1/2m,{alpha}) between the moderate/severe and normal/mild groups were statistically significant, all at P<0.0001 (Table 4).


View this table:
[in this window]
[in a new window]
 
Table 4 Statistical comparisons between the normal/mild and moderate/severe groups. Pharmacokinetic terms are explained in the text. *Statistically significant
 

    Discussion
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
The wide variation in remifentanil and RA Cmax and AUClast values primarily reflected differences in remifentanil infusion rate and duration, with confounded interpatient variability in pharmacokinetic parameters.

The metabolic ratio is indicative of the ratio of the RA to remifentanil concentrations at steady state.22 The mean metabolic ratio for the normal/mild renal function group in this study (15.1) is consistent with the estimate obtained in a previous study in patients with mild renal impairment (17.0).26 The metabolic ratio increased to 116 in patients with moderate/severe renal impairment and the difference would have been even larger if the complete AUC of the metabolite was estimable. This increase is consistent with previous findings in patients with severe renal failure.14 The differences observed in the metabolic ratio between the moderate/severe subgroups (mean metabolic ratio 70, 132 and 192 for subgroups A, B and C, respectively) inversely correlated with the differences in Clcr in these subgroups (mean CLcr 24.0, 5.3 and 1.8 ml min–1 respectively).

One-, two- and three-compartment models have previously been used to describe the pharmacokinetics of remifentanil in humans,4 14 15 while the pharmacokinetics of RA have been previously described by a one-compartment model, reporting a half-life of 1.5 h in healthy patients.14 The present data only supported a one-compartment model for remifentanil. The concentrations of RA in the normal/mild group declined in a biexponential manner, suggesting a two-compartment model, with a much longer terminal half-life, as was also the case in a recent study of similar design using the same assay for analyte determination.26 The ability to observe a slow terminal phase, previously undetected,14 in patients with normal renal function was probably due to the high concentrations of RA, the increased assay sensitivity and the extended sampling period employed in that previous study26 and the present one. This terminal phase, apparently unaffected by renal function, accounted for about 25% of the total AUC and probably represents a slowly equilibrating distribution compartment, which becomes rate limiting at later times. A one-compartment model adequately described RA concentrations in patients in the moderate/severe group. Either a combined three-compartment model (normal/mild group) or a two-compartment model (moderate/severe group) was therefore used for simultaneous modelling of parent and metabolite data.

Mean remifentanil CL was 59.0 (SD 52.9) ml min–1 kg–1 in the moderate/severe group and 44.3 (14.4) ml min–1 kg–1 in the normal/mild group; the difference was not statistically significant. V and t1/2 were estimated to be 1.76 (2.34) litre kg–1 and 20.5 (17.8) min in the moderate/severe group and 0.737 (0.645) litre kg–1 and 11.4 (7.24) min in normal/mild group, respectively. Large variability was observed in both these variables and any apparent differences in the mean estimates were not statistically significant (P=0.0735 and P=0.0824, respectively), and are not expected to have a clinically significant effect on recovery. These observations are also consistent with a previous study in volunteers with renal failure where no significant differences in the pharmacokinetics of remifentanil were observed compared with healthy volunteers,14 although the dosage and duration of remifentanil treatment in that study were much lower. Exploratory contrast analysis between subgroups (results available on request) indicated that the trend in remifentanil volume mainly reflected extremely large differences in the volume estimates of two patients in the moderate/severe group undergoing continuous RRT (subgroup C) compared with all other subgroups. Such differences in volume of distribution might be due to an effect of any concomitant fluid replacement therapy these patients might have undergone and in addition, continuous RRT is known to cause large fluid shifts and hence transient hypervolaemia in such patients. The overall trend in the half-life of remifentanil was again found to mainly reflect differences between subgroup C and the other subgroups, and is attributable to differences in volume of distribution. In a recent study in 13 end-stage renal failure patients undergoing RRT,27 it was reported that remifentanil clearance was significantly reduced and terminal half-life increased relative to matched control patients with normal renal function. As a result, remifentanil blood concentrations were significantly higher in the renal failure group. However, the changes were clinically modest and it was suggested that they may be explained by a reduced volume of distribution in the period following haemodialysis. In the present study, although a (not significant) prolongation of remifentanil half-life was observed, volume of distribution was, if anything, increased in patients undergoing RRT and no consistent effect on remifentanil clearance, or remifentanil concentrations was observed.

Mean CLm was significantly decreased in the moderate/severe group compared with the normal/mild group: 41.4 (28.9) vs 176 (49.3) ml h–1 kg–1. The differences between the normal/mild group and all three subgroups of the moderate/severe group were statistically significant and it appears that CLm is linearly related to CLcr (Fig. 6). The mean elimination half-life of RA, as estimated by t1/2,m,{alpha} for the three-compartment combined model in the normal/mild group, was 2.48 (1.03) h, which is similar to that previously reported for patients with mild renal impairment (2 h).26 It was also significantly shorter than the single RA half-life estimated by the two-compartment combined model for the moderate/severe group, 18.5 (13.6) h. In contrast, the half-life of RA associated with the terminal phase in the normal/mild group (t1/2,m,ß, three-compartment model) was 16.6 h, and this was not significantly different from the single half-life estimated for the moderate/severe group. Due to the different compartmental models used, a comparison of these half-lives is not strictly statistically permissible, but the results of the statistical analyses can be used to elucidate underlying mechanisms. In renal impairment, it is the elimination of RA that is affected, as demonstrated by the decrease in CLm in the present study. The shift in the metabolite kinetics from two compartments to one is consistent with a prolongation of RA elimination half-life (which can be assumed to be given by t1/2,m,{alpha} in the normal/mild group) with increasing severity of renal impairment. In the study in patients with mild renal impairment,26 a 2-fold increase in t1/2,m,{alpha} has been observed relative to normal patients, while the terminal half-life t1/2,m,ß remained unaltered. It can be expected that, if t1/2,m,{alpha} is further prolonged in moderate/severe renal impairment, the two phases may no longer be distinguishable, and that would be consistent with the previously reported 20-fold increase of RA half-life in severe renal impairment.14 It can then be safely assumed that the single half-life estimated in the moderate/severe group in the present study reflects this prolongation in RA elimination half-life, as a result of the increased severity of renal impairment. Although this situation called for mixed-effects modelling in order to statistically evaluate covariate effects such as that of renal function and patient characteristics on RA pharmacokinetics, all attempts to estimate population pharmacokinetic parameters and their variability in NONMEM failed to converge.

In exploratory plots of pharmacokinetic parameters vs the covariates age and weight, no obvious trends were observed. It was not possible to illustrate the previously found effect of age15 on remifentanil pharmacokinetics since patients were unevenly distributed (three aged 31–40 yr, 37 aged 54–81 yr). In addition, the small number of female patients included in the study and their uneven distribution (9/40, one in the normal/mild group and eight in the moderate/severe group) did not permit any gender comparisons.

Pharmacodynamic results of the same clinical trial, presented by Breen and colleagues,28 showed no evidence of any clinically relevant prolongation of µ-opioid effects, as assessed by the time to the offset of the pharmacodynamic effects. The main finding relevant to the discussion of RA pharmacokinetics was that the difference in the mean time to offset between the normal/mild and moderate/severe groups after 72 h of remifentanil administration was only 17 min, even when RA concentrations reach up to 500 times those of remifentanil in individual patients with a significant degree of renal impairment.

In order to assist future study designs in ICU patients who might require remifentanil administration for longer than 72 h, remifentanil and metabolite simulations were performed using mean parameter estimates obtained here, with CLm expressed as a function of CLcr, using the equation given in Figure 6. Simulations for average patients with various degrees of renal impairment receiving a 9 µg kg–1 h–1 continuous remifentanil infusion for 15 days are presented in Figure 7A.



View larger version (24K):
[in this window]
[in a new window]
 
Fig 7 Simulated remifentanil acid concentrations for an average individual (weight 75 kg) with (A) various degrees of renal impairment ( CLcr 0, 5, 20, 50 and 80 ml min–1) and (B) for average individual with 0 CLcr together with the patients with the highest observed remifentanil acid Cmax (Patient A, CLcr=5 ml min–1) and the highest metabolic ratio (Patient B, CLcr=0 ml min–1).

 
Had we obtained a population pharmacostatistical model, it would have been possible to report confidence intervals on the RA concentrations predicted by the model. Instead, together with the average estimate for CLcr=0 ml min–1, we present in Figure 7B predictions based on two extreme individuals, one with the highest observed RA Cmax (Patient A, CLcr=5 ml min–1) and one with the highest metabolic ratio (Patient B, CLcr=0 ml min–1). These simulations predict that in the worst-case scenario, RA concentrations would reach steady state by 144 h (6 days) and would not be expected to exceed 600 ng ml–1. Such concentrations of RA were observed in the present study, when higher infusion rates were used, and according to the results cited above from Breen and colleagues,28 were not associated with any prolongation of µ-opioid effects, as assessed by the time to the offset of the pharmacodynamic effects.

In conclusion, the pharmacokinetics of remifentanil were not significantly altered in ICU patients with moderate/severe renal impairment compared with those with normal/mild impaired renal function, even after continuous i.v. administration for up to 3 days. CLm decreased in a linear fashion with decreasing CLcr, and in patients with moderate/severe renal impairment was reduced to 25% that of the normal/mild group. The metabolic ratio increased by 8-fold in the moderate/severe group relative to the normal/mild group, predicting average RA concentrations at steady state more than 100-fold those of remifentanil in patients with moderate/severe renal impairment. Simulations using the pharmacokinetic parameters obtained from the study predict that in patients with renal failure, RA concentrations would reach steady state after 6 days of continuous infusion and, for a remifentanil regimen of 9 µg kg–1 h–1, these concentrations would not exceed those reported in the present study, for which no associated prolongation of µ-opioid effects was observed.


    Acknowledgements
 
The authors would like to acknowledge the contribution of the following to the conduct of the study: Dr Leon Aarons, School of Pharmacy, Manchester University, UK, for his expert advice on population pharmacokinetic modelling; Dr M. Cross and Mr S. Elliot, Leeds General Infirmary, Leeds, UK; Ms S. Smith, Royal Hallamshire Hospital, Sheffield, UK; Dr L. Nielsen and Dr O. Christensen, Hilleroed Syngehus, Hilleroed, Denmark; Dr T. Faber, Amtssygehuset i Herlev, Herlev, Denmark; Dr D. Meininger and J.-W. Goethe, Universitat Zentrum der Anaesthesiologie und Wiederbelebung, Frankfurt, Germany; Dr M. Marsch, Dr M. Kirmse and Dr S. Goddon, Universitat Erlangen-Nurnberg.Klinik fur Anaesthesiologie, Erlangen, Germany, for their assistance with conducting the study.


    References
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
1 Portenoy KR, Thaler H, Inturrisi CE, Friedlander-Klar H, Foley K. The metabolite morphine-6-gluguronide contributes to the analgesia produced by morphine infusion in patients with pain and normal renal function. Clin Pharmacol Ther 1992; 51: 422–31[ISI][Medline]

2 Shimomura K, Kamata O, Ueki S, et al. Analgesic effect of morphine glucuronides. Tohoku J Exp Med 1971; 105: 45–52[ISI][Medline]

3 Kapila A, Glass PS, Jacobs JR, et al. Measured context-sensitive half-times of remifentanil and alfentanil. Anesthesiology 1995; 83: 968–75[CrossRef][ISI][Medline]

4 Westmoreland CL, Hoke JF, Sebel PS, Hug CC Jr, Muir KT. Pharmacokinetics of remifentanil (GI87084B) and its major metabolite (GR90291) in patients undergoing elective surgery. Anesthesiology 1993; 79: 893–903[ISI][Medline]

5 Dershwitz M, Hoke F, Rosow C, et al. Pharmacokinetics and pharmacodynamics of remifentanil in volunteer patients with severe liver disease. Anesthesiology 1996; 84: 812–20[CrossRef][ISI][Medline]

6 Park GR, Evans TN. Remifentanil in the critically ill – what will its place be? Br J Intensive Care 1996; 79: 893–903

7 Evans TN, Park GR. Remifentanil in the critically ill. Anaesthesia 1997; 52: 800–1[ISI][Medline]

8 Main A. Remifentanil as an analgesic in the critically ill. Anaesthesia 1998; 53: 823–4

9 Wilhelm W, Dorscheid E, Schlaich N, Niederprum P, Deller D. The use of remifentanil in critically ill patients. Clinical findings and early experience. Anaesthesist 1999; 48: 625–9[CrossRef][ISI][Medline]

10 Tipps LB, Coplin WM, Murry KR, Rhoney DH. Safety and feasibility of continuous infusion of remifentanil in the neurosurgical intensive care unit. Neurosurgery 2000; 46: 596–601[ISI][Medline]

11 López A, Muellejans B, Cross MH, Bonome C, Morrison L, Kirkham A. The safety and efficacy of remifentanil for the provision of optimal sedation in ICU patients (Abstract A407). Intensive Care Med 2001; 27(Suppl 2): S239

12 Kessler P, Chinachoti T, Van Deer Berg P, Stanley A, Kirkham A. Remifentanil vs. morphine for the provision of optimal sedation in ICU patients (Abstract A406). Intensive Care Med 2001; 27 (Suppl 2): S239

13 Karabinis A, Hantson P, Speelberg B, et al. A remifentanil-based technique for analgesia and sedation in ICU patients with neurotrauma: preliminary data (Abstract A549). Intensive Care Med 2001; 27 (Suppl 2): S275

14 Hoke JF, Shlugman D, Dershwitz M, et al. Pharmacokinetics and pharmacodynamics of remifentanil in persons with renal failure compared with healthy volunteers. Anesthesiology 1997; 87: 533–41[CrossRef][ISI][Medline]

15 Minto CF, Schnider TW, Egan TD, et al. Influence of age and gender on the pharmacokinetics and pharmacodynamics of remifentanil. Anesthesiology 1997; 86: 10–23[ISI][Medline]

16 Hoke JF, Cunningham F, James MK, Muir KT, Hoffman WE. Comparative pharmacokinetics and pharmacodynamics of remifentanil, its principle metabolite (GR90291) and alfentanil in dogs. J Pharmacol Exp Ther 1997; 281: 226–32[Abstract/Free Full Text]

17 Riker RR, Picard JT, Fraser GL. Prospective evaluation of the Sedation-Agitation Scale for adult critically ill patients. Crit Care Med 1999; 27: 1325–9[ISI][Medline]

18 Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976; 16: 31–41[ISI][Medline]

19 Kirkham A, Fisher G, Kessler P. A dosing algorithm for the use of remifentanil in providing optimal sedation and analgesia in ICU patients (Abstract A405). Intensive Care Med 2001; 27(Suppl 2): S238

20 Bender J, van den Elshout J, Selinger K, Broeders G, Dankers J, van der Heiden C. Determination of remifentanil in human heparinised whole blood by tandem mass spectrometry with short-column separation. J Pharm Biomed Anal 1999; 21: 559–67[CrossRef][ISI][Medline]

21 WinNonlinTM Professional Edition Version 3.1, Pharsight Corporation 800 West El Camino Real, Suite 200, Mountain View, CA 94040

22 Gibaldi M, Perrier D. Pharmacokinetics, 2nd edn. New York: Marcel Dekker, 1982; 345

23 NONMEM version V, NONMEM Project Group, University of California at San Francisco, CA 94143 USA

24 Gabrielsson J, Weiner D. Pharmacokinetic and Pharmacodynamic Data Analysis: Concepts and Applications, 2nd edn. Stockholm: Swedish Pharmaceutical Press, 1997; 88

25 LeGall J-R, Lemeshow S, Saulnier F. A new simplified acute physiology score (SAPS II) based on a European/North American multicenter study. JAMA 1993; 270: 2957–63[Abstract]

26 Binning S, Morrison L, Bodenham A, Snowden C, Pitsiu M, Kirkham A. The pharmacokinetics of remifentanil and RAin patients with mild renal impairment (Abstract A256). Intensive Care Med 2000; 26 (Suppl 3): S280

27 Dahaba AA, Oettl K, von Klobucar F, Reibnegger G, List WF. End stage renal failure reduces central clearance and prolongs the elimination half life of remifentanil. Can J Anaesth 2002; 49: 369–74[Abstract/Free Full Text]

28 Breen D, Wilmer A, Bodenham A, et al. The offset of pharmacodynamic effects and safety of remifentanil in ICU patients with various degrees of renal impairment. Crit Care 2004; 8: R21–30[CrossRef][ISI][Medline]