Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Morphine is metabolized extensively in humans to M3G and M6G (Yeh et al., 1977), both of which are excreted predominantly into urine (Osborne et al., 1990). Given the substantial evidence for the opioid agonist activity of M6G in animals and humans, it is now recognized that its formation in humans has important pharmacological and toxicological implications (Kroemer and Klotz, 1992; Milne et al., 1996; Mulder, 1992). Furthermore, from studies in animals there are reports suggestive of a functional antagonism by M3G of the antinociceptive effects of morphine and M6G and of its ability to stimulate ventilation and invoke hyperesthesia and hyperactive motor behavior. These effects of M3G may result from its interaction with receptors other than the known opioid receptors (Milne et al., 1996, and references therein).

Regnard and Twycross (1984) observed a requirement for lower doses of morphine in patients with renal failure, along with an increased incidence of side effects, when usual therapeutic doses were administered. Two studies using chromatographic methods specific for morphine (Säwe and Odar-Cederlöf, 1987; Woolner et al., 1986) reported that the total body clearance of morphine from patients in renal failure was not significantly different than those with normal renal function, but a more recent study (Osborne et al., 1993) suggested that the clearance of morphine may be significantly reduced by renal failure. Although the importance of the kidney for the elimination of morphine in humans remains unknown, we reported previously that the kidney contributes considerably to the total clearance of morphine in sheep with normal kidneys (Milne et al., 1993, 1995; Sloan et al., 1991). The kidney of humans (Osborne et al., 1990) and sheep (Milne et al., 1993, 1995) was the major organ for the elimination of M3G and M6G; in humans, it has been proposed that the prolonged opioid effects after the administration of morphine to patients with renal failure was due to accumulation of M6G (Bodd et al., 1990; Hasselström et al., 1989; Osborne et al., 1986; Shelly et al., 1986).

Our previous studies (Milne et al., 1993, 1995; Sloan et al., 1991) involving the administration of morphine and M3G to sheep with normal kidneys demonstrated that the liver and kidneys are responsible for the majority of the clearance of morphine, the kidney both metabolizes and excretes morphine, M3G is the predominant metabolite and there is net extraction of M3G by the kidney and net production by the gut. It was concluded that the kidney is the major site for the elimination of M3G and that morphine undergoes enterohepatic cycling, presumably involving biliary excretion of M3G followed by hydrolysis in the gut and reabsorption of morphine (Milne et al., 1993, 1995).

The aims of this study were to examine the disposition of morphine, M3G and M6G in sheep with renal failure and during infusion with morphine and make comparisons with data obtained previously from sheep with normal kidneys (Milne et al., 1995).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Approval for the study was obtained from the Animal Ethics Review Committee of the Flinders Medical Centre.

Preparation of the sheep. Sheep were obtained from Mortlock Farm (University of Adelaide, Australia). Their preparation and the methods used for the measurement of regional blood-flows were as previously described (Sloan et al., 1991), with modifications. During surgery for the placement of catheters, a left lumbar incision was made, and a 4F polyethylene catheter was inserted into the left renal artery, against the flow of blood, and fastened to a plastic plate that was tied to adjacent muscle. Measurement of renal blood flow by the Fick method was made with the rate of infusion of p-aminohippuric acid being one-half that administered to sheep with normal kidneys (Sloan et al., 1991). Sheep were given a lethal dose of pentobarbital sodium if they became ill or showed signs of distress.

Induction of renal failure. The renal artery to and renal vein from the right kidney were ligated, and the kidney was removed. Between 4 and 10 days later, induction of renal failure was commenced by the administration of sterile suspensions of polystyrene microspheres (25-30-µm diameter NEM 003 microspheres, DuPont-New England Nuclear, Boston, MA) in normal saline (120,000 microspheres/ml, 1-5 ml) and/or glass microspheres (75-150 µm diameter, acid-washed type 1-W, Sigma Chemical Co., St. Louis, MO) in 5% dextrose (100,000 microspheres/ml, 0.5-2.5 ml) into the left renal artery. The development of renal failure was monitored by daily measurement of the concentrations of creatinine in plasma (Jaffé rate method, Autoanalyzer, Department of Clinical Biochemistry, Flinders Medical Centre, Adelaide).

Administration of morphine. Seven merino ewes with a mean ± S.D. weight of 52 ± 4 kg and ages ranging from 1.5 to 2 yr were used. After surgical preparation and the induction of renal failure, two-stage infusions of morphine (as morphine sulfate B.P., David Bull Laboratories, Mulgrave, Australia) in normal saline were administered into a right atrial catheter. Initially, the rate was 80 mg/hr for 15 min, followed by a rate maintained at 10 mg/hr (morphine sulfate) for an additional 5.75 hr. Morphine sulfate B.P. (as the pentahydrate), 10 mg, is equivalent to 7.52 mg of morphine. The infusion flowed at ~0.5 ml/min. Four of the sheep (sheep 1, 2, 4 and 5) included in the present study had been administered morphine on a previous occasion when their kidneys were normal, and the results were included in an earlier report (Milne et al., 1995).

Collection of samples. Blood samples (1 ml; at 15-min intervals from 5 to 6 hr after beginning the infusion) were collected simultaneously from catheters previously inserted into the aorta, pulmonary artery, hepatic vein, hepatic portal vein, renal vein and posterior vena cava. Additional blood samples were collected from the aorta at 0.25, 0.5, 1, 2, 3 and 4 hr (sheep 1) or 5, 10, 15, 20, 25 and 30 min and 1, 2, 3 and 4 hr (sheep 2-7) and for all sheep at 24, 48, 72, 96, 120 and 144 hr. Blood samples were centrifuged within 30 min, and the plasma was transferred to 1.5-ml polypropylene centrifuge tubes for storage at 20°C.

Binding of morphine, M3G and M6G in plasma. Binding was determined by ultrafiltration as described previously (Milne et al., 1995) but without additional M6G.

Determination of C_b/C. The values for C_b/C were determined as previously described (Milne et al., 1993) in drug-free blood collected before commencing the infusions of morphine sulfate. Duplicate samples of blood were spiked with morphine (70.2 and 257 nM), M3G (1120 nM) and M6G (52.4 and 198 nM) and equilibrated at 37°C for 30 min. The nominal concentration in blood divided by the concentration measured in plasma by HPLC was taken as C_b/C and used to convert pharmacokinetic parameters determined with respect to the concentrations in plasma to equivalent values calculated with respect to concentrations in blood.

Determination of morphine, M3G and M6G. Concentrations of morphine, M3G and M6G in plasma, ultrafiltrate from plasma and urine were determined by paired-ion HPLC (Milne et al., 1991). Between-day accuracy and reproducibility for morphine in plasma at 13.3, 133 and 1060 nM were 103 ± 14.5%, 104 ± 5.4% and 96.2 ± 8.2%, respectively; for M3G at 108, 867 and 8670 nM, values were 102 ± 15.1%, 102 ± 6.7% and 95.7 ± 8.2%, respectively; and for M6G at 20.4, 102 and 204 nM, values were 97.6 ± 16.5%, 94.7 ± 9.5% and 92.3 ± 8.7%, respectively. The limits of quantification for morphine, M3G and M6G were 13.3, 108 and 20.4 nM, respectively.

Determination of creatinine in plasma and urine. Concentrations of creatinine in plasma collected from the aorta at 5 and 6 hr and in urine collected during this interval were determined by HPLC (Milne et al., 1995).

Data analysis. The AUC was calculated from 0 to 5 hr [AUC(0-5)] and 5 to 6 hr [AUC(5-6)] by the trapezoid rule. The CL_b of morphine was calculated from the quotient of the amount of morphine base infused during the 5- to 6-hr period and the product of the respective C_b/C and pulmonary arterial AUC(5-6). The terminal rate constants for M3G and M6G were determined by log-linear regression on time (Multi-Fit, Day Computing, Cambridge, UK) of their concentrations in plasma measured from 48 hr and beyond when the infusion of morphine was started and were used to calculate the corresponding half-lives. Regional net extraction ratios for morphine, M3G and M6G were calculated as the difference between the respective AUC(5-6) for plasma entering and leaving a given region divided by that entering. It was assumed again that 80% of the blood flowing to the liver came via the hepatic portal vein, with the remainder flowing via the hepatic artery (Milne et al., 1993; Sloan et al., 1991). Regional net clearances for morphine, M3G and M6G were estimated as the products of the respective extraction ratios and blood-flows. As an indicator of drug eliminated into bile and involved in an enterohepatic cycle, the fraction of summed morphine, M3G and M6G in blood entering the liver and not reappearing in the hepatic vein as either morphine, M3G or M6G (fractional retention) was calculated (Milne et al., 1993).

Statistical analysis. Comparisons of means between two or more groups were performed, as appropriate, by paired or unpaired Student's t test or by analysis of variance (StatView, Brainpower, Calabasas, CA). Values of P < .05 were considered significant. Standard methods were used to determine the power of a test (Winer, 1971). Variation about the mean is given as S.D. The data obtained from this study were compared with those reported previously from five sheep with normal kidneys that had received morphine (Milne et al., 1995), supplemented with data from an additional sheep that had received morphine. Associations or relationships between parameters were examined by Spearman rank-order correlation or linear regression, respectively (StatView).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Before the induction of renal failure, the concentrations of creatinine in plasma ranged from 0.07 to 0.10 mM. Sheep 1 was given morphine when the concentration of creatinine reached 0.16 mM; the other sheep had concentrations ranging from 0.40 to 0.54 mM on the day of morphine infusion. Figure 1 shows the concentrations of creatinine in sheep 5 during the development of renal failure. Complete collections of plasma and urine were obtained from sheep 1, 2 and 5. Plasma and urine were collected up to 48 hr from sheep 3 and for 72 hr from sheep 6 and 7. The collection of urine between 5 and 6 hr from sheep 4 was incomplete, and blood could not be taken from the catheter placed into the hepatic portal vein of sheep 6. 

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Concentrations of creatinine in the plasma of sheep 5 during the development of renal failure. a, Day on which the infusion of morphine was administered. Numbers above and below points on the line indicate the volume (ml) of suspension containing glass (100,000 microspheres/ml) or polystyrene (120,000 microspheres/ml), respectively, administered into the left renal artery on those days.

Concentration in plasma-time profiles. Figure 2 depicts concentrations of morphine, M3G and M6G in plasma-time profiles from sheep 5 during the 5- to 6-hr of infusion with morphine; these were typical of other sheep. The variability in the pulmonary arterial concentrations of morphine, M3G and M6G measured every 15 min between 5 and 6 hr, expressed as coefficients of variation, was comparable to the respective analytical reproducibility for each compound, and there were no obvious trends in the concentrations over the time interval. However, unlike morphine, the concentrations of M3G and M6G in arterial plasma clearly showed a gradual increase from 0 to 6 hr (fig. 3a). The half-lives of M3G and M6G, calculated from 48 hr after starting the infusion of morphine, ranged from 26 to 113 hr (n = 4) and 43 to 95 hr (n = 2), respectively (fig. 3b).

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Concentrations of morphine, M3G and M6G in the plasma of sheep 5 during the fifth to sixth hour of infusion with morphine sulfate at 10 mg/hr. Concentrations are in samples collected from the abdominal aorta (), pulmonary artery (), hepatic vein (), hepatic portal vein (), renal vein () and posterior vena cava ().

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Concentrations of morphine, M3G and M6G in the arterial plasma of sheep 5 between (a) 0 and 6 hr and (b) 0 and 144 hr from the start of a 6-hr infusion with morphine sulfate at 10 mg/hr. Concentrations are for morphine (), M3G () and M6G ().

Regional extraction. The mean regional net extraction ratios for morphine and M3G during administration of morphine are shown in figure 4. There was significant extraction of morphine by the liver (0.72 ± 0.05, P < .001, 5 df) and kidney (0.42 ± 0.15, P < .001, 6 df). There was significant net extraction of M3G by the kidney (0.044 ± 0.025, P = .004, 6 df), but there was no significant difference from zero in the net extraction of M3G by the gut (0.017 ± 0.062, P = .52, 5 df) or liver (0.027 ± 0.030, P = .054, 6 df). For M6G, there was significant net extraction by the kidney only (0.051 ± 0.043, P = .020, 6 df, data not shown).

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Mean ± S.D. regional net extraction ratios of morphine and M3G during infusion of sheep in renal failure with morphine sulfate. *, Significantly different from zero, P < .050.

C_b/C and binding in plasma. The overall mean ± S.D. (n = 6) values of C_b/C for morphine, M3G and M6G were 1.32 ± 0.17, 0.81 ± 0.16 and 0.74 ± 0.12, respectively. In the absence of individual values for sheep 2, the overall mean values were used. The mean hematocrit was 0.34 ± 0.07.

Total and regional clearances for morphine. Table 2 presents the CL_b of morphine and clearances by the liver and kidney and gives unpaired comparisons with corresponding values in sheep with normal kidneys. There was no difference in the mean CL_b of morphine between the two groups, but there was a significant difference in regional clearance between the normal and failed kidney. For the group with renal failure, the mean of the summed clearances for morphine by the liver and kidney was 1.8 ± 0.5 liters/min and was not significantly different (P = .29, 5 df) from the CL_b.

Renal clearance and urinary recovery. The mean renal clearance of creatinine was 0.011 ± 0.007 (range 0.0045-0.024) liter/min. Comparing the renal excretory clearances over the 0- to 5-hr and 5- to 6-hr intervals, there were no significant differences for morphine (P = .10, 5 df), M3G (P = .84, 5 df) and M6G (P = .13, 4 df). Based on data collected during the 5- to 6-hr period, the mean renal excretory clearances of morphine, M3G and M6G from plasma during infusion with morphine were 0.011 ± 0.007, 0.013 ± 0.009 and 0.014 ± 0.012 liter/min, respectively. There was a significant relationship between the renal excretory clearances of unbound morphine and creatinine (P = .039, 5 df), and the significance was increased (P < .001, 11 df) when data from sheep with normal kidneys were also included (fig. 5). Figure 5 also shows the corresponding significant relationships for unbound M3G (P < .001) and M6G (P = .001) during infusion with morphine, when the data from both groups of sheep were combined. Only the slopes of the relationships for M3G and M6G were significantly greater than unity.

View larger version (11K): [in this window] [in a new window]   Fig. 5.   Relationships between the renal excretory clearance of creatinine and the renal excretory clearances of unbound morphine (y = 0.0012 + 1.4x, P < .001, r = .82, 11 df), M3G (y = 0.0050 + 2.4x, P < .001, r = .96, 11 df) and M6G (y = 0.0078 + 2.0x, P = .001, r = .89, 9 df) during infusion with morphine of sheep with normal kidneys () and those in renal failure (). The 95% confidence intervals of the slopes for morphine, M3G and M6G were 0.70 to 2.0, 1.9 to 2.8 and 1.1 to 2.8, respectively.

View larger version (36K): [in this window] [in a new window]   Fig. 6.   Mean ± S.D. percentage urinary recovery from sheep with renal failure of the dose of morphine as morphine, M3G, and M6G.

Mass balance across the liver. The mean fraction of the sum of morphine, M3G and M6G entering the liver via the hepatic artery and portal vein that did not reappear in the hepatic vein (fractional retention) was 0.082 ± 0.024 and was significantly different from zero (P < .001, 6 df). However, the value was not significantly different (P = .070, 10 df) from the value of 0.12 ± 0.04 determined in sheep with normal renal function.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

After surgical removal of the right kidney, renal failure was successfully induced in all seven sheep by the introduction of a mixture of polystyrene and/or glass microspheres into the left renal artery. The degree of renal dysfunction was monitored by daily measurement of the concentrations of creatinine in plasma. Renal failure was induced more rapidly by the glass microspheres, and the concentrations of creatinine were maintained by repeated administration of both types of microspheres. Histopathological evaluation of the changes induced by the microspheres was not possible because glass present within the tissue precluded preparation of fine sections for histological examination. However, it is likely that occlusion of the smaller branches of the renal artery was followed by infarction and necrosis. We compared our method with that of others for inducing renal failure in sheep; English et al. (1977) used heminephrectomy and 50% infarction of the remaining kidney, but adaptive increases in renal function by the remaining renal tissue ensured only a 64% reduction in GFR and a 79% increase in the concentration of creatinine in plasma. Eschbach et al. (1980) induced renal failure by ligation of branches of the renal artery to one kidney, producing ~90% infarction, followed 2 weeks later by contralateral nephrectomy. The concentrations of creatinine were raised to values comparable to those produced by the administration of the polystyrene and glass microspheres, and GFR was reduced to <15 ml/min. The method described in the present study for the induction of renal failure required less surgery, but it appeared that Eschbach et al. (1980) were able to maintain survival for a longer period, albeit with the aid of hemodialysis for some sheep.

During infusion of sheep for 6 hr, essentially constant concentrations of morphine were achieved in plasma between 5 and 6 hr, similar to our previous observations when sheep with normal kidneys were infused with the same dose regimen (Milne et al., 1993, 1995). In contrast, the concentrations of M3G and M6G after 6 hr of infusion were still increasing (fig. 3); compared with sheep with normal kidneys, they were markedly higher and maintained for a longer period after the infusions had ceased. Given that the kidney was the major route for the elimination of M3G and M6G in sheep with normal kidneys (Milne et al., 1993, 1995), their clearance would be reduced during renal failure and, assuming an unchanged volume of distribution, their half-lives of elimination would increase. Previous investigators (Osborne et al., 1993; Säwe and Odar-Cederlöf, 1987) have reported similar findings for M3G and M6G when morphine was administered to humans with renal failure. Interestingly, the mean ratio of the AUC(5-6) for M3G to that for M6G during renal failure was not significantly different from the corresponding ratio in sheep with normal kidneys. Given that the clearances of both compounds were reduced to a similar degree by renal failure (table 3), their formation and subsequent appearance in plasma were either unaffected or affected equally by renal failure.

It was fortunate that four sheep were common to the experiments conducted when their kidneys were functioning normally (Milne et al., 1995) and when they were in renal failure. Consistent decreases in the CL_b of morphine were observed in each of these sheep, of a magnitude comparable to the reduction in its clearance by the kidney. A paired comparison, which is not possible in studies with humans, revealed a significant decrease in CL_b. When all sheep were used in an unpaired comparison (table 2), as was performed with the studies in humans (Osborne et al., 1993; Säwe and Odar-Cederlöf, 1987; Woolner et al., 1986), the induction of renal failure did not alter the CL_b of morphine significantly, despite our direct evidence from sheep for a sizable clearance by the kidney (Milne et al., 1995). The power of the unpaired comparison in sheep to detect a difference of 25% (the mean value observed in the paired comparison) was weak.

Osborne et al. (1993) compared the total body clearance of morphine from healthy humans with the values determined in three groups of patients with failed kidneys who had all undergone surgery: patients receiving dialysis, a nondialyzed group and another group who received transplanted kidneys. Although the mean total body clearance of morphine in the healthy subjects was greater than that in patients receiving dialysis, it was not different from that determined in the patients with transplanted kidneys. The findings may have been influenced by the anesthetic regimen used during surgery but, nevertheless, those given the transplant had a significantly greater mean clearance than the group receiving dialysis but not the nondialyzed group. Thus, although the study suggests a role for the kidney in the metabolic clearance of morphine in humans, as found in the sheep (Milne et al., 1995), a greater number of patients may have provided an unequivocal conclusion. Not unexpectedly, the interindividual variability in clearance in sheep (Milne et al., 1995; Sloan et al., 1991) or in humans (Osborne et al., 1990) weakened the power of the unpaired comparison performed in the present study and in those using humans (Osborne et al., 1993; Säwe and Odar-Cederlöf, 1987; Woolner et al., 1986).

Clearance of morphine by the kidney in sheep was significantly reduced by renal failure (table 2) and, using the renal clearance of creatinine as a marker of renal function, there clearly was an association between the reduction in the clearance of morphine by the kidney and the degree of renal failure. Reduced clearance of morphine was probably caused by a lesser blood flow and/or intrinsic clearance via excretion and/or metabolism to M3G (and M6G). Decreases in the extraction ratio for morphine due to renal failure were almost identical with the decreases in the extraction for p-aminohippuric acid. Both compounds are highly extracted by the renal tubular cells in normal sheep (Milne et al., 1995), but their extraction may have been reduced in renal failure by damage to, or reduced functional perfusion of, tubular cells and/or by competition from accumulated endogenous compounds of small molecular weight (McNay et al., 1976). Clearance of morphine by the kidney during renal failure was significantly greater than its renal excretory clearance, indicating that the kidney still excreted and metabolized morphine, as found in sheep with normal kidneys (Milne et al., 1995).

Significant relationships were observed between the renal clearance of creatinine and the renal clearances of unbound morphine, M3G and M6G (fig. 5). Significant corresponding relationships were also found previously in patients with diverse renal function when infused at a constant rate with morphine while in intensive care (Milne et al., 1992). The slope of the relationship for morphine determined in the patients suggested net tubular secretion of morphine by the kidney. Assuming a similar renal handling of creatinine in humans and sheep (Faix et al., 1988; Shemesh et al., 1985) and that clearance of creatinine is a reasonable estimate of GFR in sheep (Milne et al., 1995), the data from sheep are in contrast to those from humans: neither net secretion nor reabsorption of morphine is apparent. Conclusions regarding mechanisms for the renal excretion of M3G (and M6G) in sheep during infusion with morphine are complicated by the metabolic conversion of morphine to M3G (and possibly M6G) in renal tubular cells of sheep with normal kidneys (Milne et al., 1995) and, as observed in the present study, during renal failure. No definitive conclusions in respect of net secretion or reabsorption of M3G by tubular cells could be drawn from a previous study in which M3G was infused into sheep with normal kidneys (Milne et al., 1995).

From the comparable clearances of morphine by the liver in sheep with normal kidneys and sheep in renal failure, it can be concluded that renal failure had no detectable effect on the capacity of the liver to eliminate morphine. Routine monitoring of hepatic function suggested no overt hepatic failure in the presence of renal failure (data not shown). Studies examining the effect of cirrhosis on the total clearance of morphine in humans (Hasselström et al., 1990; Patwardhan et al., 1981) demonstrated that a reduction was detectable only when there was considerable damage to hepatic tissue.

The net retention of summed morphine, M3G and M6G by the liver, as found previously in sheep with normal kidneys (Milne et al., 1993, 1995), adds support for the existence of an enterohepatic cycle. An inability to detect a net appearance of M3G in the portal blood of sheep during renal failure, in contrast to the experiments in sheep with normal kidneys, was probably due to the assay being unable to discriminate M3G appearing in the portal vein from the relatively large arterial concentrations of M3G. Compared with sheep with normal kidneys, it is possible that a greater fraction of the dose of morphine was subject to cycling in sheep with renal failure because the kidney contributed a lesser fraction to the CL_b (see table 2). In sheep with normal kidneys, it was proposed that the majority of the morphine taken up by the kidney was metabolized to M3G rather than excreted unchanged and that the majority of the so-formed M3G was excreted directly into urine (Milne et al., 1993, 1995). Therefore, in sheep with renal failure, it is likely that a greater fraction of the overall M3G formed by the body was formed in the liver, excreted into bile and thus involved in an enterohepatic cycle. These events would have exaggerated any differences in CL_b between sheep in renal failure and those with normal kidneys.

The mean recovery of the dose of morphine as summed morphine, M3G and M6G in the urine of sheep with renal failure (fig. 6) was well below that from sheep with normal kidneys. Recovery up to 48 hr was only 39% compared with mean values of 85% (Milne et al., 1993) and 79% (Milne et al., 1995) over the same period in sheep with normal kidneys. In both latter studies, most of the dose was recovered as M3G. It was proposed that a large proportion of the M3G was formed in the liver and subsequently excreted by the kidney; a lesser but significant amount was formed in the kidney and excreted directly into urine (Milne et al., 1995). Therefore, with formation and excretion by the kidney diminished by renal failure, alternative routes of elimination for M3G would have become more important. Of the dose appearing in urine, most was present as M3G, suggesting extensive metabolism of morphine. However, in consequence of a greater direct biliary excretion of M3G from plasma and an increased fraction of the dose of morphine being metabolized by the liver, it is probable that fecal elimination of M3G has become more important. In support of this hypothesis, it was found previously with renally ligated rats that 50% of an intravenous dose of radiolabeled M3G was excreted into bile as unchanged M3G (Roerig et al., 1974) compared with only 20% recovered from bile duct-ligated rats with normally functioning kidneys (Ouellet and Pollack, 1995); in our laboratory, 45% of an intravenous dose of M3G administered to one sheep in renal failure was recovered in urine after 144 hr, with none measurable in the last 24-hr interval,² compared with a recovery of 84% of the dose of M3G from sheep with normal kidneys (Milne et al., 1995).

In summary, this study has found that compared with sheep with normal kidneys, the induction of renal failure markedly reduced the effectiveness of the kidney as an organ for the elimination of morphine. Furthermore, although remaining a major organ for the elimination of M3G and M6G, renal failure greatly reduced the clearance of these two compounds by the kidney and caused considerable retention of M3G and M6G in plasma. While a paired comparison demonstrated a significant decrease in the CL_b of morphine during renal failure, an unpaired comparison failed to demonstrate a decrease. The high concentrations and prolonged elimination from plasma of M3G derived from morphine, mainly in the liver, were due to a combination of its involvement in an enterohepatic cycle, as in sheep with normal kidneys, and, more importantly, a considerable reduction in its renal clearance compared with the normal group. Continued recirculation through the liver of the higher concentrations of M3G found in plasma during renal failure ensured its involvement in an enterohepatic cycle and the probability of a greater fraction being eliminated in feces.