Recovery of Hepatic Drug Extraction after Hypothermic Preservation 

The experimental protocol was approved by the institutional Committee on Animal Research. Male Sprague-Dawley rats (230–275 g, Benton & Kingman, Fremont, CA) had access to rat chow and water ad libitum. Isolated perfused rat livers were prepared using a standard method. [4,5] Ether was used for anesthesia because it is rapidly eliminated by the perfusion system lung, thus reducing the chance for interaction with drug-metabolizing pathways. Liver and portal structures were mobilized, and the common bile duct was cannulated. The hepatic artery was ligated and the liver perfused via a 16-G cannula in the portal vein, with oxygenated Krebs-Henseleit solution maintained at 37 degrees Celsius. The inferior vena cava was ligated and the confluence of hepatic veins cannulated via the right atrium. After isolation, livers were transferred to a temperature-controlled chamber and perfused for 30 min with a recirculating perfluorocarbon oxygen carrier (FC-43 emulsion, Alpha Therapeutic, Los Angeles). [5,6] Perfusate was equilibrated with oxygen and carbon dioxide using an artificial lung consisting of thin-walled silastic tubing. [7] Flows of oxygen and carbon dioxide to the lung were adjusted to maintain the inflow perfusate pH, measured via an indwelling probe, at 7.40 plus/minus 0.05. Liver viability was assessed during this period by visual inspection, bile production, oxygen consumption, and perfusion pressure. Livers in the preservation group (n = 5) were perfused for 2 min with iced lactated Ringer's solution containing 0.1 mg/ml procaine and 10 U/ml heparin, followed by perfusion for 3 min with iced Belzer UW cold storage solution (Viaspan, Du Pont, Wilmington, DE) containing 0.04 U/ml insulin and 0.02 mg/ml dexamethasone. Livers were immersed in Belzer solution, and the perfusion dish was sealed with plastic wrap. After storage at 4 degrees Celsius for 24 h, livers were drained via gravity, returned to the perfusion chamber, and flushed for 2 min via the portal vein with iced 6% hetastarch solution (Hespan, DuPont). After this sequence of hypothermic preservation and in control livers (n = 5) immediately after the 30-min perfusion, livers were perfused single-pass at a constant flow rate with 12–17% washed human erythrocytes, 1% albumin, and 30 mM sodium taurocholate in Krebs-Henselheit solution. Target drug concentrations in the perfusate were: fentanyl 100 ng/ml, morphine 600 ng/ml, and vecuronium 1 micro gram/ml. In control livers, the drugs were added to the erthyrocyte-containing perfusate immediately before surgery; in preserved livers, drugs were added to this perfusate before the hetastarch flush. Portal vein inflow pressure and hepatic temperature were continuously monitored during perfusion. Inflow perfusate samples were collected at 4, 16, 32, and 48 min after initiation of the erythrocyte-containing perfusate (reperfusion). Effluent perfusate samples were collected at 1, 2, 4, 8, 16, 32, 48, and 64 min. Bile was collected continuously over 16-min intervals in preweighed tubes. At 64 min, perfusion was stopped and the liver was trimmed of nonhepatic tissue, weighed, and homogenized to determine the tissue concentration of vecuronium and its metabolite, 3-desacetylvecuronium. After inflow and outflow perfusate samples were centrifuged to remove erythrocytes, drug concentrations in the supernatant were determined. Samples of perfusate, bile, and homogenized liver for analysis of vecuronium were acidified to prevent spontaneous degradation of vecuronium. Fentanyl was measured by radioimmunoassay (Research Diagnostics, Flanders, NJ), sensitive to 0.1 ng/ml with a coefficient of variation of < 10% at that concentration. Morphine also was measured by radioimmunoassay (Diagnostic Products, Los Angeles, CA), sensitive to 0.8 ng/ml with a coefficient of variation < 10% at that concentration. Vecuronium and 3-desacetylvecuronium were measured by capillary gas chromatography, sensitive to 10 ng/ml with a coefficient of variation of 11% at that concentration. [8] Hepatic oxygen consumption (V with dot O²) was calculated from the oxygen content of inflow and outflow perfusate samples measured at 4, 16, 32, and 48 min using a Clark electrode (Radiometer, Copenhagen, Denmark). Blood gases were determined at the same times (Stat Profile 5 Analyzer, Nova Biomedical, Waltham, MA). Perfusate concentrations of Sodium sup +, Potassium sup +, Calcium sup ++, and glucose were determined (Stat Profile 5 Analyzer) to confirm physiologic concentrations.

Because these initial experiments demonstrated net extraction (rather than production) of 3-desacetylvecuronium in control and preserved livers, we studied an additional nine livers (five control, four preserved) with identical technique except that the perfusate contained only 1 micro gram/ml 3-desacetylvecuronium rather than morphine, fentanyl, and vecuronium.

For each drug, hepatic drug metabolic function was estimated by determining the extraction ratio for each liver, defined as:Equation 1.

For fentanyl or morphine (whose inflow concentrations remained constant over time), inflow concentration for each liver was determined as the mean of measured inflow values. For vecuronium (whose inflow concentration decreased over time, see Results), measured inflow concentrations were fitted to a single exponential decay (r = 0.85–0.99) to estimate the inflow concentration at each time. Extraction ratios were determined at 1, 2, 4, 8, 16, 32, 48, and 64 min using the measured outflow concentrations and the mean or fitted inflow concentrations. Mass balance for vecuronium plus 3-desacetylvecuronium was calculated based on molar quantities of these drugs in perfusate, effluent, bile, and liver tissue (collected at completion of each experiment).

Values for preserved and control livers were compared using the Mann-Whitney U-test. Changes over time within each group were assessed by repeated measures analysis of variance and Student-Newman-Keuls test or with Student's t test for unpaired or paired data. Data are reported as mean plus/minus SD. P < 0.05 was considered statistically significant.

During the initial 30-min perfusion with perfluoro-carbon emulsion, hepatic temperature increased from 28 degrees Celsius to 37 degrees Celsius, whereas portal vein pressure remained constant (Table 1). There was no difference between groups in either of these values or in the rate of bile production during this period. During the preservation sequence (preserved livers only), hepatic temperature decreased to 4 degrees Celsius and hepatic inflow pressure increased.

During perfusion with the erythrocyte-containing solutions, inflow pH, oxygen content, and flow rate were similar in control and preserved livers. Portal venous pressure during this phase was similar to that recorded during the baseline perfusion. Four minutes after the beginning of erythrocyte perfusion, liver temperature and V with dot O sub 2 were significantly less in preserved livers (Table 1). By 16 min and for the remainder of the perfusion, V with dot O²and hepatic temperature were similar in both groups. During the first 16 min, preserved livers produced less bile than control livers, but bile production for the subsequent intervals was similar, and there was no difference in total bile production. Hepatic weight at the end of the experiment was similar in control and preserved livers.

Fentanyl inflow concentration (target = 100 ng/ml) remained constant at 92.8 plus/minus 34.2 ng/ml. For both groups, fentanyl extraction was greater during the first 4 min than during the 16–64 min interval (Figure 1and Table 2). After the initial warming phase, mean fentanyl extraction in preserved livers was similar to that in control livers, except at 48 min when it exceeded that in control livers.

Morphine inflow concentrations (target = 600 ng/ml) were constant at 575 plus/minus 204 ng/ml. For control livers, morphine extraction was greater in the initial 4 min than the later interval (Figure 2and Table 2). For preserved livers, morphine extraction did not change over time. Morphine extraction in preserved livers was similar to that in control livers, except at 4 min when it was less than that in control livers.

Inflow concentrations of vecuronium (target = 1 micro gram/ml) decreased during the perfusion (Figure 3). This was due to spontaneous conversion of vecuronium to 3-desacetylvecuronium, as evidenced by the presence of 3-desacetylvecuronium in the inflow perfusate in all studies. Because the interval between the addition of vecuronium to the perfusate and its administration to the isolated liver was greater in control livers, the inflow concentration of 3-desacetylvecuronium was greater than that for preserved livers.* In control livers, vecuronium extraction was greater in the initial 4 min than during the later interval (Figure 4and Table 2). In contrast, preserved livers had a lower extraction at 4 min than at subsequent times. Extraction was lower for preserved livers than for control livers during the first 4 min, but by 8 min there were no further differences. For both control and preserved livers, outflow concentrations of 3-desacetylvecuronium were lower than inflow concentrations (data not shown). During the 16–64-min interval, extraction of 3-desacetylvecuronium was similar in control livers (0.23 plus/minus 0.23) and preserved livers (0.21 plus/minus 0.20). Biliary excretion and tissue deposition of vecuronium plus 3-desacetylvecuronium were similar in preserved and control livers (Table 3). Nearly all vecuronium plus 3-desacetylvecuronium entering the liver could be accounted for in the effluent, bile, and liver tissue. The fraction of the perfused vecuronium metabolized by the liver to 3-desacetylvecuronium was similar in the two groups.

In separate experiments with 3-desacetylvecuronium (target = 1 micro gram/ml), inflow concentration was stable at 1.2 plus/minus 0.3 micro gram/ml. In preserved livers, hepatic extraction was 0.37 plus/minus 0.19 during the first 4 min of perfusion after hypothermic preservation, then increased to 0.47 plus/minus 0.15 during the final 48 min. Extraction of 3-desacetylvecuronium was greater in the control group for the first 4 min (Figure 5). However, by 8 min, there were no further differences.

One liver in the preservation group given morphine, fentanyl, and vecuronium (included in all data analyses) functioned poorly: bile production was minimal (one-fourth of the other preserved livers), oxygen consumption was low (one-half of the other livers, increasing only slightly with rewarming), but portal venous inflow pressure was normal. Consistent with these markers of poor overall function, hepatic drug extraction of fentanyl and morphine decreased over time, and the liver eventually failed to extract 3-desacetylvecuronium. In addition, only 1% of the injected vecuronium dose was excreted in bile, and only 5% remained unchanged in the liver.

Our results suggest that the standard method of preparing livers for transplantation—hepatectomy, hypothermic preservation, and reperfusion—did not influence hepatic extraction of fentanyl or morphine. Although hepatic extraction of vecuronium and 3-desacetylvecuronium were impaired during early reperfusion, preservation did not affect hepatic extraction after rewarming or overall biliary excretion and tissue concentration of vecuronium or its metabolite.

Data on the metabolic capacity of the newly transplanted human liver are limited. Bodenham et al. [1] gave morphine during the anhepatic phase and noted that concentrations of morphine metabolites were negligible until the liver was reperfused. Although that study demonstrated that the newly transplanted liver conjugates, it lacked a control group, thus providing qualitative, rather that quantitative, evidence of hepatic function. Shelly et al. [2] reported that, shortly after liver transplantation, morphine's plasma clearance exceeded published values for normal subjects. Although this suggests that liver function was not compromised, lack of a control group again limits application. A recent study compared muscle relaxant dose requirements during the dissection and neohepatic phases of liver transplantation. [9] However, that study compared native, diseased livers to transplanted livers, and therefore cannot differentiate effects of hepatic disease from those of preservation. Thus, for two important classes of intravenous drugs, opioids and muscle relaxants, effects of hepatic preservation are poorly defined.

Many alterations in the newly transplanted liver—including hypothermia-induced cell swelling, intracellular acidosis, expansion of the interstitial space during reperfusion, free-radical induced cellular injury, and depletion of high-energy phosphates—might impair hepatic metabolism. [10,11] To further define these documented and theoretical insults to hepatic homeostasis, we determined the effect of preservation on hepatic extraction of drugs commonly used during transplantation, selecting drugs with different metabolic pathways: fentanyl is oxidized by mixed function oxidases (cytochrome P450), morphine is glucuronidated, and vecuronium is both metabolized (via deacetylation) and excreted intact in bile. These drugs also differ in their hepatic extraction: fentanyl is highly extracted, [12] whereas both morphine [13] and vecuronium [14] have intermediate extraction values. We did not study extraction of a drug with a low extraction ratio that may have been more sensitive to changes in hepatic function.

By measuring hepatic extraction repeatedly after reperfusion, we assessed transient changes in hepatic function. In the first minutes of reperfusion, extraction of all drugs (except vecuronium in preserved livers) exceeded subsequent values, presumably reflecting rapid hepatic uptake while intracellular drug concentrations are small. The subsequent plateau in extraction presumably reflects increasing efflux of drug from the liver into the perfusate as intracellular drug concentrations increase. These data suggest that metabolism and/or biliary excretion determine hepatic extraction during the late reperfusion phase but are less important during early reperfusion. In addition, because effluent concentrations plateaued in all studies, hepatic uptake presumably was complete during the experiment, and our final values for extraction, when multiplied times perfusate flow, estimate drug clearance.

Contrary to our expectation of diminished hepatic function after preservation, hypothermic preservation did not affect extraction of fentanyl and morphine. Preservation for 24 h maximized our likelihood of observing impaired function: Preservation of this duration results in viable grafts in rats but would be expected to reveal more hepatic dysfunction than with shorter periods. We also used drug concentrations exceeding those attained clinically to maximize the likelihood of detecting decreased hepatic function. Although extraction of vecuronium and 3-desacetylvecuronium was transiently affected by preservation, this impairment was temporally related to low hepatic temperature. Thus, we cannot distinguish the effect of hypothermia from that of preservation-induced changes (e.g., alterations in pH, depletion of adenosine triphosphate, or membrane changes) during this early period. Despite early differences in hepatic extraction of vecuronium, the overall effect of preservation was minimal: biliary excretion and final tissue content were similar in control and preserved livers. These findings suggest that the liver's ability to extract vecuronium and 3-desacetylvecuronium is minimally impaired by hypothermic preservation.

We expected that hepatic metabolism of vecuronium to 3-desacetylvecuronium would produce greater outflow than inflow concentrations of 3-desacetylvecuronium. In contrast, initial experiments revealed net extraction of 3-desacetylvecuronium. This unexpected finding can be reconciled by considering that extraction of 3-desacetylvecuronium by our isolated perfused system was 50% when perfused alone but only 20% when perfused with vecuronium. Assuming that elimination pathways of the two drugs do not compete, this suggests that hepatic metabolism of vecuronium creates 3-desacetylvecuronium, which then is released into the effluent, thereby reducing the apparent extraction of 3-desacetylvecuronium.

Our model permitted limited assessment of hepatic functions other than drug extraction. Similarity of bile flow and biliary excretion of vecuronium in preserved and control livers suggests that biliary excretory pathways were intact; similarity of oxygen consumption suggests that global function was not compromised. We did not assess hepatocellular damage (e.g., by measuring alanine amino transferase in effluent) or synthetic function of hepatocytes or sinusoidal endothelial cells (e.g., by measuring albumin synthesis) because these tests are insensitive during single-pass perfusion. Thus, preserved livers may be functionally impaired yet metabolize drugs well. Other in vitro model systems have been used to study the effects of preservation, including isolated hepatocytes and liver tissue slices [15]; however, these studies have not examined quantitative differences in hepatic drug metabolism between normal and preserved livers.

Although our results suggest that hepatic drug metabolism is protected during hypothermic preservation, our in vitro model differs from liver transplantation. First, leukocytes, thought to mediate hepatic injury during reperfusion, are absent [16]; their presence in vivo may increase reperfusion injury and thereby impair hepatic drug extraction. Second, we perfused at a constant rate and only through the portal vein. During transplantation, many factors influence liver blood flow, [17,18] including the sequence of portal vein and hepatic artery anastomosis. Marked decreases in liver blood flow would compromise hepatic delivery of drugs, thereby decreasing clearance of highly extracted (perfusion-limited) drugs such as fentanyl and possibly decreasing clearance of drugs with lower extraction ratios. Third, our study examines only hepatic drug extraction but does not address extrahepatic drug distribution. If the neohepatic period were associated with other physiologic changes that alter drug distribution, this would influence drug concentrations in vivo. Fourth, animal models do not necessarily duplicate human responses. Although metabolic pathways for these drugs were qualitatively and quantitatively similar in rats and humans, sensitivity to preservation injury may differ between species. Finally, although hepatic function cannot be maintained indefinitely in our model, isolated perfused livers function adequately for 3–4 h at 37 degrees Celsius. [19] The similarity of our extraction ratios to those obtained in vivo supports our model as being valid to examine hepatic drug elimination. [12,13].

Some authors suggest that the newly transplanted liver metabolizes drugs poorly and advise extreme caution when administering intravenous anesthetic drugs during this period. [20] In contrast, our results suggest that drug metabolism recovers rapidly after hepatic preservation and reperfusion. However, our results must be interpreted with caution. Clinical experience indicates that some hepatic allografts never regain adequate function after preservation. One of our preserved livers may have demonstrated this “primary nonfunction”: decreased bile production, decreased oxygen consumption, and a progressive decrease in hepatic extraction of all three drugs. Recently, production of monoethylglycinexylide, a metabolite of lidocaine, was proposed as a marker of function of donor livers before harvesting. [21] In addition, reduced production of monoethylglycinexylide may indicate that patients with hepatic dysfunction require urgent transplantation. [22] Whether decreased metabolism of other drugs, such as those in the current study, can identify early hepatic dysfunction remains to be determined.

In summary, our animal model suggests that hepatic hypothermic preservation for transplantation affects hepatic extraction of morphine, fentanyl, and vecuronium minimally. Transient impairment of vecuronium extraction may represent hypothermia or preservation-dependent reduction in hepatic uptake but resolves rapidly. Our results suggest that, in the neohepatic period, hepatic metabolism of intravenous anesthetics is minimally affected in livers that function well. Future work may indicate whether compromised drug metabolism is a sensitive and/or specific marker of dysfunction in the transplanted liver.

The authors thank Linda Kendrick, for assistance with the perfusion preparation, and Larry Gruenke, Ph.D., for analytic work.

*This time-dependent conversion of vecuronium to 3-desacetylvecuronium has been observed in blood maintained at physiologic pH and temperature (M. Sharma: Personal observation, 1993.). In addition, because the perfusate for control livers remained at 37 degrees Celsius and pH 7.4 longer than the perfusate for preserved livers (see Methods), lower concentrations of vecuronium and higher concentrations of 3-desacetylvecuronium should have been anticipated.