INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

PLASMA FREE FATTY ACID (FFA) and glycerol concentrations are elevated in patients with end-stage liver disease. In general, plasma glycerol concentration has been interpreted as an indicator of accelerated lipolysis because glycerol liberated by lipolysis is not reincorporated into triglyceride and glycerol clearance is not rate limiting for its plasma concentration in healthy or injured humans (2). The observation that plasma glycerol concentration is not dictated by its clearance in healthy or injured humans has been extrapolated to patients with liver disease, in whom it has been proposed that increased plasma FFA and glycerol concentrations are caused by an accelerated rate of lipolysis (9, 21). This proposal is based on the assumption that glycerol clearance, which occurs mainly in the liver, is not impaired in patients with end-stage liver disease. Indirect evidence from two nonisotopic studies, however, demonstrates that the underlying assumption that glycerol clearance is not impaired by liver disease is untrue (10, 15). The only direct evidence that lipolysis may be accelerated in early cirrhosis, despite hyperinsulinemia, comes from a single study in which glycerol flux, measured by isotope dilution, was used to index lipolytic rate (11). This observation is consistent with the finding in hyperinsulinemic type 2 diabetes without liver disease (17) but differs from the findings in obese individuals in whom the ability of insulin to inhibit lipolysis is preserved (1, 22) and suggests that a compromised antilipolytic effect of insulin may be responsible for an increased rate of lipolysis in cirrhotic subjects. Whether lipolysis is accelerated in end-stage cirrhosis, to what extent this can be predicted by the circulating glycerol concentration, and the role of impaired insulin action as a mediator of the changes in lipid metabolism in end-stage cirrhosis remain uncertain.

After orthotopic liver transplantation (OLT), patients frequently develop a hypertriglyceridemia that persists for years (5, 8). The mechanism producing the hypertriglyceridemia remains unknown. Moreover, the extent to which the abnormalities of lipid metabolism associated with end-stage cirrhosis are reversed after OLT is untested. This study was designed to directly compare the flux and clearance of palmitate and glycerol, their plasma concentrations, and palmitate oxidation rates among patients with end-stage cirrhosis, patients 1 wk after OLT, and healthy volunteers.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Subjects. This study was approved by the Institutional Review Boards at Oregon Health Sciences University (OHSU) and the Portland, OR Veterans Affairs Medical Center. Twenty patients with stable end-stage cirrhosis were enrolled after written informed consent. Of these, 10 patients were studied in the OHSU Clinical Research Center while awaiting OLT. Inclusion criteria for cirrhotic patients were end-stage liver disease; hemodynamic stability; no history of insulin-dependent diabetes mellitus; no recent history of cigarette smoking; medications that did not include -adrenergic agonists, -adrenergic antagonists, or steroids; and adequate renal function, as indexed by plasma creatinine concentration <1.4 mg/dl. Severity of liver disease was assessed by the Pugh-Childs scoring system (20). The remaining 10 patients were studied on day 7 after OLT. The seventh postoperative day was chosen because it is a time of stability when many patients are ready for hospital discharge, the prednisone steroid taper is almost complete, and chronic oral immunosuppression therapy has optimized. An inclusion criterion for postoperative subjects, in addition to those for preoperative subjects, was good liver graft function as indicated by circulating albumin, prealbumin, and bilirubin concentrations and coagulation status in the absence of blood product transfusion. Ten healthy volunteers without clinical or laboratory evidence of liver disease were enrolled as controls. Control subjects were studied in the OHSU Clinical Research Center. Dietary intake on the day before the study was measured for cirrhotic and control subjects but not postoperative OLT subjects. All three subject groups were awake and alert and postabsorptive for at least 8 h, during which time they received no intravenous glucose or other sources of calories. Screening venous blood samples were drawn before the start of the stable isotope infusion protocol.

Infusion protocol. The protocol began at 0700, when one intravenous catheter was placed in an upper extremity for stable isotope tracer infusion and a second intravenous catheter was placed in the dorsum of the contralateral hand for sampling of "arterialized" blood. The sampling hand was warmed with a heating pad. An indwelling central venous catheter was used for isotope infusion in postoperative patients.

1

1

1

1

1

1

Analytical procedures. Plasma palmitate isotope ratio was determined by negative chemical ionization gas chromatography-mass spectroscopy (NCI-GC/MS) on a Hewlett-Packard 5989B GC/MS system (Hewlett Packard, Fullerton, CA) as described by Hachey et al. (6). The pentafluorobenzyl derivative was prepared and analyzed by selective ion monitoring at mass-to-charge ratios (m/z) 255 and 256. Plasma glycerol isotope ratio was also measured by NCI-GC/MS, on the pentafluorobenzyl derivative, with selective monitoring of ions at m/z 680-685 as described by Gilker et al. (4).

Calculations. Palmitate flux (R_a palmitate) was calculated by the equation
where F is the infusion rate of tracer palmitate (µmol · kg¹ · min¹), IE_infusate is the isotopic enrichment of the infused tracer palmitate (mole % excess), and IE_plateau is the isotopic enrichment of plasma palmitate (mole % excess) at isotopic steady state. A similar calculation was used for glycerol flux (R_a glycerol). Clearance of palmitate or glycerol (ml · kg¹ · h¹) was calculated as substrate flux divided by its plasma concentration.

1

1

Statistical analysis. Data are expressed as means ± SE. One-way ANOVA with Tukey's post hoc test was used to compare plasma substrate and hormone concentrations, R_a values for palmitate, FFA, and glycerol, palmitate and FFA oxidation rates, and clearance of palmitate, FFA, and glycerol among groups. Statistical tests were performed using a specialized software program (Crunch 4, Crunch Software, Oakland, CA). Differences were considered statistically significant at P < 0.05.

1

1

1

1

1

1

1

1

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Baseline characteristics. The control group, eight men and two women, were 28 ± 2 (SE) years old, weighed 70.3 ± 4.5 kg, and had a body mass index (BMI) of 23.4 ± 0.9 kg/m². Table 1 shows demographic characteristics of the two patient groups. The most common etiology for cirrhosis in preoperative patients was concomitant ethanol abuse and hepatitis C viral infection. A broad spectrum of initial presenting liver disease was represented in the postoperative OLT group. Compared with controls, the two patient groups were older (P < 0.001) but had similar body weight and BMI. Cirrhotic patients did not differ from postoperative patients in any demographic characteristic, including the severity of liver disease immediately before OLT. Mean time since last ethanol intake for the cirrhotic patients with a history of ethanol abuse was 41 mo, with median 42 mo and range 7-60 mo. For postoperative OLT subjects with a history of ethanol abuse, the mean time since last intake was 30 mo with median 24 mo and range 13-62 mo. Dietary intake on the day before the isotope study for the cirrhotic subjects was diminished compared with controls, in terms of carbohydrate (2.30 ± 0.30 vs. 3.09 ± 0.28 g/kg), protein (0.61 ± 0.08 vs. 0.95 ± 0.08 g/kg), fat (0.37 ± 0.04 vs. 0.66 ± 0.06 g/kg), and total calories (14.9 ± 1.6 vs. 21.8 ± 1.8 kcal/kg) (all P < 0.05).

Effect of cirrhosis and liver transplantation on bicarbonate kinetics and gas exchange. The primed-constant infusion of NaH¹³CO₃ led to a steady-state enrichment of expired CO₂ in all groups. Recovery of NaH¹³CO₃ as exhaled breath ¹³CO₂ was not statistically different among controls (0.772 ± 0.031%), cirrhotic patients (0.862 ± 0.036%), and OLT patients (0.793 ± 0.051%). Total CO₂ production was 119 ± 6, 114 ± 3, and 125 ± 4 µmol · kg¹ · min¹ for controls, cirrhotic patients and OLT patients, respectively (not different). Respiratory quotient in controls (0.83 ± 0.02), cirrhotic patients (0.80 ± 0.02), and OLT patients (0.84 ± 0.02) also did not differ among groups.

Hormonal profile in cirrhosis and after OLT. Both cirrhotic patients and liver transplant patients exhibited a marked hyperinsulinemia, with serum insulin concentration two- and threefold higher, respectively, than in controls (Table 3). Hyperinsulinemia was paralleled by an elevated circulating C-peptide concentration. Concomitantly, serum glucagon concentration was higher than control value by more than threefold in cirrhotic patients and by almost fivefold in OLT patients. OLT patients showed higher serum concentrations of insulin, C-peptide, and glucagon compared with preoperative cirrhotic patients. Thus the hormonal profile did not correct toward corresponding control values after OLT. The isotope infusion protocol did not change the hormonal profile of any subject group.

Effect of cirrhosis on lipid flux and oxidation in vivo. Cirrhotic subjects had increased plasma concentrations of palmitate (by 52%), FFA (by 76%), and glycerol (by 33%) compared with corresponding control values (Table 4). Plasma palmitate and glycerol isotopic enrichment reached a steady state during the isotope infusion protocol in both subject groups. R_a palmitate in cirrhotic patients was 71% higher and total FFA flux was almost twofold higher, compared with their respective control values. FFA clearance in cirrhotic patients was unchanged compared with that of controls.

Effect of liver transplantation on lipid flux and oxidation in vivo. Seven days after OLT, patients had plasma concentrations of palmitate, total FFA, and glycerol that did not differ from those of controls (Table 4). The isotopic infusion protocol led to steady-state plasma isotopic enrichment of palmitate and glycerol in all OLT patients, as it did in controls and preoperative cirrhotic patients. Palmitate and total FFA fluxes in OLT subjects did not differ from those in controls, although there was a trend toward faster flux of both substrates. Clearance of fatty acids by OLT patients was 79% higher than in controls. Glycerol flux and clearance by OLT patients were ~40% slower than those in healthy controls (P < 0.01). The R_a palmitate-to-R_a glycerol ratio in OLT subjects was twofold higher, and the R_a FFA-to-R_a glycerol ratio almost threefold higher, than in controls. This flux relationship was not reflected in the ratio of plasma substrate concentrations. Neither palmitate nor total FFA oxidation by OLT subjects differed from that of healthy controls, with either the NaHCO₃ or the acetate correction factor. Similarly, the fraction of palmitate flux that was oxidized to CO₂, which indicates the relationship between fatty acid production and oxidation, did not differ between groups.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Our results showed that, although both FFA (palmitate) and glycerol concentrations are elevated in patients with end-stage cirrhosis, the kinetics underlying these alterations differ. Whereas the flux of FFA is faster, the flux of glycerol is slower than that of controls. Glycerol flux is a more accurate index of triglyceride breakdown rate (lipolysis) than palmitate flux because glycerol, unlike palmitate, cannot be recycled into triglyceride within the adipocyte. The slower glycerol flux indicates that the rate of lipolysis is slower in patients with end-stage cirrhosis than in healthy volunteers; hence their hyperglycerolemia compared with controls results from decreased glycerol clearance rather than accelerated lipolysis. On the other hand, the faster FFA flux in cirrhotic patients, despite a slower rate of lipolysis, coupled with unimpaired FFA oxidation indicates that there is a defect in the nonoxidative disposal of FFA. Because nonoxidative disposal of FFA is reincorporation into triglyceride (reesterification), we propose that the cirrhosis-induced defect in lipid metabolism is not accelerated lipolysis but probably an impairment of FFA reesterification into triglyceride within the adipocyte. One week after OLT, normalization of plasma palmitate and glycerol concentrations masks a continued inhibition of lipolysis, as indicated by the lower glycerol flux compared with controls, and an apparent impairment of FFA reesterification into triglyceride.

We measured palmitate flux and oxidation as indices of FFA metabolism because palmitate kinetics reflect those of total FFA in healthy volunteers (7) and the higher plasma palmitate concentration in cirrhotic subjects parallels that of total FFA (21). Our findings of elevated plasma concentrations of palmitate and glycerol in cirrhotic subjects agree with previous observations (9, 11, 21). Similarly, higher plasma total FFA concentration in cirrhotic subjects has also been reported previously (18, 21, 23). The somewhat less marked increase in FFA and glycerol plasma concentrations of our cirrhotic patients may reflect the exclusion of cigarette smokersa large percentage of patients with end-stage cirrhosisand patients receiving -agonist therapy, because nicotine and -adrenergic stimulation directly stimulate lipolysis (25).

Our flux values for lipid kinetics in healthy volunteers are in general agreement with previous reports for R_a palmitate (27), R_a FFA (11), and R_a glycerol (1, 11, 17). With regard to the kinetics underlying the elevated plasma palmitate and FFA concentrations of cirrhotics, the increased R_a palmitate in the present study is consistent with the findings of two previous isotope dilution studies (11, 23). Romijn et al. (23) administered a constant infusion of [¹⁴C]palmitate, and Kaye et al. (11) used a [¹³C]palmitate infusion protocol similar to ours. Our finding that glycerol flux is slower in cirrhotic patients, however, contrasts with the faster glycerol flux reported by Kaye et al. (11), despite similar stable isotope tracer protocols. It is possible that the qualitative difference in glycerol fluxes reflects more severe liver disease in our patients. All 10 of our patients had end-stage disease, whereas Kaye et al. (11) studied five Pugh class A and three Pugh class B patients. Alternatively, it is possible that either nicotine from cigarette smoking or -adrenergic therapy in the cirrhotic patients stimulated lipolysis independently of liver disease in the study by Kaye et al. (11, 25). It is noteworthy that our patients and those of Kaye et al. exhibited a comparable hyperinsulinemia relative to healthy controls. Our observation in cirrhotic patients that plasma glycerol concentration is higher than that in controls despite slower glycerol flux indicates that plasma glycerol concentration is increased because of decreased clearance rather than increased production. The concomitantly higher plasma FFA concentration, FFA flux, and FFA oxidation despite slower glycerol flux indicate that there is a reduced rate of FFA reesterification within the adipocytes of cirrhotic patients.

One week after OLT, fatty acid flux and oxidation are intermediate between that of patients with end-stage cirrhosis and healthy controls. Lipolytic rate, as indexed by glycerol flux, is concomitantly slower than that of healthy volunteers. Together, these findings indicate that the rate of FFA reesterification within the adipocyte remains decreased, resulting in a faster FFA flux. Because the absolute rate of FFA oxidation is not increased to the same extent as FFA flux in the OLT patients, more FFA will be available to the liver to be reincorporated into triglyceride and released into the circulation in the form of very low-density lipoprotein-triglyceride. This may explain the hypertriglyceridemia reported in OLT patients by Jindal et al. (8) and Gisbert et al. (5).

The underlying mechanism that impairs FFA reesterification within the adipocytes of cirrhotic and OLT patients is unknown. It is possible that the supply of glycerol phosphate is limited in adipocytes of cirrhotic subjects, either by decreased availability of dihydroxyacetone secondary to impaired glucose uptake (11, 19) or consequent to a possible defect in glycerol-3-phosphate dehydrogenase activity. Alternatively, either acyl-CoA synthetase or any of the acyltransferase reactions, which catalyze fusion of acyl-CoA with glycerol-3-phosphate, may be impaired. We were surprised by the failure of OLT to reverse the effects of liver disease on lipid metabolism by the seventh postoperative day. Given the present results, it would be of interest to study patients further removed from their transplant surgery, perhaps at 6 or 12 mo postoperatively. It is possible that the immunosuppressive drugs used in OLT patientscyclosporine, azathioprine, and prednisonemay contribute to the observed derangement in lipid metabolism. Glucocorticoid treatment increases plasma FFA concentration and lipid oxidation (26). Our OLT patients were receiving prednisone 1.5 mg · kg¹ · day¹ at the time of study, as part of a taper protocol that gradually decreased to 5-10 mg/day. The actions of cyclosporine and azathioprine on lipid metabolism is less well understood. After our study was completed, the introduction of tacrolimus for immunosuppression has increased the incidence of post-OLT hypertriglyceridemia relative to that observed with cyclosporine (8). As a practical matter, it is difficult to evaluate the effect of a transplanted liver without the potentially confounding effects of immunosuppressive therapy.

The influence of prior ethanol and other dietary intake on lipid metabolism deserves mention. It had been many months since the subjects in either liver disease group had consumed ethanol, and we feel that the effects of acute ethanol ingestion were not a factor in the present study. What remained in our patients in whom ethanol had been an etiologic factor in developing end-stage liver disease was the liver disease itself rather than an acute nutritional alteration. The previous dietary intake of our preoperative cirrhotic subjects was less than that of controls. Unfortunately, we were unable to document the dietary intake of post-OLT subjects, but we surmise that it too may have been decreased relative to that of controls. It is important to emphasize that all subjects were studied in the postabsorptive state. Nevertheless, it could be argued that the decreased dietary intake of our cirrhotic and post-OLT subjects would mimic a protracted fast that extended the effects of the overnight fasting. However, altered diet would not explain the decreased lipolysis and impaired reesterification observed in our cirrhotic and post-OLT subjects because protracted starvation stimulates rather than inhibits lipolysis, with parallel increases in glycerol and palmitate fluxes (29).

The role of hyperinsulinemia as a mediator of the changes in lipid metabolism observed in cirrhotic and OLT patients is subject to speculation. On the basis of their findings that accelerated lipolysis in cirrhotic patients coincided with hyperinsulinemia, Kaye et al. (11) concluded that cirrhotic patients must be resistant to the antilipolytic action of insulin, but they did not rule out other potential stimulators of lipolysis such as cigarette smoking or -adrenergic therapy. Insulin inhibits glycerol flux in healthy humans in a dose-dependent manner, such that a modest increase in serum insulin concentration from 9.8 to 14-20 µU/ml suppresses R_a glycerol by 50% (16). Furthermore, the maximal restraining effect of insulin on lipolysis occurs well within physiological hyperinsulinemia (16). Hyperinsulinemia to the level exhibited by our cirrhotic and OLT patients could theoretically retard lipolysis in these patients relative to the rate in healthy controls. The ability of insulin to inhibit glycerol flux is preserved in obesity, when corrected for fat mass, despite "insulin resistance" of glucose metabolism (1, 22). In contrast, lipolysis is accelerated and peripheral tissue FFA uptake and oxidation are diminished in type 2 diabetic patients without liver disease, despite a well-documented hyperinsulinemia (12, 17). Both plasma concentrations and fluxes of FFA and glycerol decrease in cirrhotic patients during insulin infusion, indicating that at least some of the ability of insulin to inhibit lipolysis is retained (11). More extensive evaluation of the role of insulin as a modulator of lipolysis and FFA oxidation remains to be performed in patients with cirrhosis and after OLT.

Finally, it is clear that plasma glycerol concentration cannot be utilized as an index of lipolytic rate in the setting of liver disease. Glycerol utilization has been considered to occur predominantly in the liver (14). Although glycerol clearance is not the limiting factor in the relationship between its flux and plasma concentration in healthy or injured humans (2), the marked impairment of glycerol clearance secondary to cirrhosis causes an increase in its plasma concentration despite a decreased flux. The extent to which glycerol clearance must be impaired before plasma glycerol concentration no longer reflects glycerol flux remains to be determined. Johnston et al. (10) found a 35% decrease in glycerol clearance in 14 cirrhotic patients, most of whom had less marked liver disease than those in the present study. They further reported an inverse correlation between plasma glycerol concentration and glycerol clearance rate in their cirrhotic subjects. Nosadini et al. (15) reported decreased glycerol clearance in cirrhotic patients when hepatocellular perfusion was compromised either by marked fibrosis or portocaval shunting. It should be pointed out that all 10 of our cirrhotic patients had evidence of portal hypertension and that 2 of them had an indwelling portosystemic shunt at the time of study. The notion that 70-90% of glycerol clearance occurs across the liver has recently been challenged by data from Brunengraber and his associates (13), which indicates that the liver may directly account for only ~30% of whole body glycerol disposal. Further investigation is required to determine the mechanism for, including whether and to what extent extrahepatic tissue contributes to, the decreased whole body glycerol clearance in cirrhosis and after OLT.